Plasmid genomic epidemiology of carbapenem-hydrolysing class D β-lactamase (CDHL)-producing Enterobacterales in Canada, 2010−2021

Abstract

Carbapenems are last-resort antibiotics for treatment of infections caused by multidrug-resistant Enterobacterales, but carbapenem resistance is a rising global threat due to the acquisition of carbapenemase genes. Oxacillinase-48 (bla_OXA-48)-type carbapenemases are increasing in abundance in Canada and elsewhere; these genes are frequently found on mobile genetic elements and are associated with specific transposons. This means that alongside clonal dissemination, bla_OXA-48-type genes can spread through plasmid-mediated horizontal gene transfer. We applied whole genome sequencing to characterize 249 bla_OXA-48-type-producing Enterobacterales isolates collected by the Canadian Nosocomial Infection Surveillance Program from 2010 to 2021. Using a combination of short- and long-read sequencing, we obtained 70 complete and circular bla_OXA-48-type-encoding plasmids. Using MOB-suite, four major plasmids clustered were identified, and we further estimated a plasmid cluster for 91.9 % (147/160) of incomplete bla_OXA-48-type-encoding contigs. We identified different patterns of carbapenemase mobilization across Canada, including horizontal transmission of bla_OXA-181/IncX3 plasmids (75/249, 30.1 %) and bla_OXA-48/IncL/M plasmids (47/249, 18.9 %), and both horizontal transmission and clonal transmission of bla_OXA-232 for Klebsiella pneumoniae ST231 on ColE2-type/ColKP3 plasmids (25/249, 10.0 %). Our findings highlight the diversity of OXA-48-type plasmids and indicate that multiple plasmid clusters and clonal transmission have contributed to bla_OXA-48-type spread and persistence in Canada.

Data Summary

Raw sequencing reads were deposited to the NCBI SRA archive under BioProject PRJNA855907 (https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA855907); see Table S1, available in the online version of this article for a list of accessions. Complete bla_OXA-48-type-encoding plasmid sequences were deposited to NCBI GenBank under the accessions listed in Table S2.

Impact Statement

Resistance to last-resort carbapenem antibiotics is a global public health threat. The dissemination of carbapenemase resistance genes is significantly influenced by plasmids, which are mobile genetic elements that can transfer between unrelated species and strains. Consequently, understanding the features and distribution of carbapenemase-encoding plasmids is crucial for pathogen surveillance and mitigation of resistance. In this work, we used long-read and short-read sequencing to characterize genomic epidemiology of OXA-48-type carbapenemase-encoding plasmids across more than a decade of Enterobacterales surveillance data in Canada. Examining the genetic and genomic context of OXA-48-type carbapenemases reveals trends in the plasmid reservoir in Canada and enhances future international pathogen surveillance.

Introduction

Carbapenems are one of the last resort antibiotics for treating serious infections caused by multidrug-resistant Gram-negative bacteria. Carbapenem-resistant pathogens have emerged following the clinical use of carbapenems, and they pose a significant threat to human health [1]. Carbapenem-resistant Enterobacterales have been reported worldwide largely as a consequence of acquisition of carbapenemase genes [2,3].

Of the different classes of carbapenemases, the oxacillinase-48 (OXA-48)-type carbapenemases are one of the most commonly identified in many countries, including Canada [3,7]. OXA-48-type carbapenemases encompass several different bla_OXA-48-type variants, including bla_OXA-48, bla_OXA-181, bla_OXA-232, and bla_OXA-204. These carbapenemases are clinically significant as they have activity against penicillins, narrow-spectrum cephalosporins, carbapenems, and provide resistance to many β-lactam inhibitors [8,9]. The first description of OXA-48-type producers in Canada was in 2011 wherein four patients harboured distinct species and sequence types with either bla_OXA-48 or bla_OXA-181 [10]. In recent years, the proportion of OXA-48-type-producing organisms in Canada has increased from 12.4 % (n=20) of carbapenem-resistant organisms isolated in 2016 to 21.4 % in 2020 (n=45) [7].

Clonal transmission of high-risk organisms carrying bla_OXA-48-type genes on persistent plasmids has played an important role in the dissemination of OXA-48-type carbapenemases. High-risk Klebsiella pneumoniae ST147, ST307, ST15, ST14 and E. coli ST410, ST38 clonal lineages have been linked to clonal OXA-48-type outbreaks [8,11, 12]. For example, Esherichia coli ST38 encoding a chromosomal bla_OXA-48 carbapenemase was found in 25 different hospitals sites around the United Kingdom [12]. The dominant sequence types and species carrying OXA-48-type carbapenemases in Canada between 2011 and 2014 were ST38 and ST410 in Escherichia coli and ST14 in the K. pneumoniae species complex [13]. OXA-48-type variants have also been described in Citrobacter, Enterobacter, Serratia, and Providencia species [5,8].

The global ascendency of OXA-48-type carbapenemases can be largely explained by horizontal plasmid spread [8]. As with many types of antimicrobial resistance genes, bla_OXA-48-type genes are frequently found on mobile genetic elements that can transfer between strains, species, and genera [4,14]. In Canada and elsewhere, bla_OXA-48 is associated with IncL/M plasmids, bla_OXA-181 is associated with IncX3 plasmids, and bla_OXA-232 is associated with ColKP3/ColE2-type plasmids [8,13, 15,17]. Certain plasmid types have also been associated with specific species and sequence types [8]; for example, bla_OXA-232 on ColKP3 plasmids has been associated with K. pneumoniae ST231 in India [18]. Each of the common OXA-48-type-encoding plasmids are additionally associated with different types of transposons which can further contribute to their mobilization [8,14]. For example, Tn1999 is associated with bla_OXA-48 on IncL/M plasmids and several structural variants have been described [8,19]. Similarly, Tn2013 is associated with both bla_OXA-181 on IncX3 plasmids and bla_OXA-232 on ColKP3/ColE2-type plasmids, although the transposon structure is truncated on IncX3 plasmids [8,20].

Here, we applied whole genome sequencing to characterize the molecular epidemiology of OXA-48-type carbapenemase-producing isolates collected by the Canadian Nosocomial Infection Surveillance Program from 2010 to 2021. Using combined short- and long-read sequencing of selected representatives to generate complete bla_OXA-48-type-encoding plasmids, we investigated the diversity of carbapenemase-encoding plasmids among these isolates across Canada and compared them to the global context of bla_OXA-48-type genes.

Methods

Surveillance period and PCR confirmation of bla_OXA-48-typecarbapenemase gene

The Canadian Nosocomial Infection Surveillance Program (CNISP) is a sentinel surveillance system which collects epidemiological and linked microbiology data from 90 Canadian acute-care hospitals across ten provinces and one territory. Enterobacterales organisms isolated from patients between 2010 and 2021 were eligible for inclusion by minimum inhibitory concentration (MIC) above clinical breakpoints [21] or if they tested positive using molecular (PCR) or phenotypic testing (mCIM, CARBA-NP) [22]. Serial isolates from the same patient were included if the organism or carbapenemase gene differed, and there were no restrictions on collection from particular body sites. Further information about CNISP can be found online (https://health-infobase.canada.ca/cnisp/). Eligible isolates were submitted to the National Microbiology Laboratory (NML; Winnipeg, Canada) by Canadian hospitals and provincial public health laboratories for bla_OXA-48-type carbapenemase gene confirmation which was conducted by multiplex PCR as previously described [13]. A total of 249 isolates encoding bla_OXA-48-type carbapenemases were collected from 2010 to 2021 from 33 hospital sites (Table S1), with one hospital submitting 19 % (47/249) of all isolates. Where applicable, the Central region refers to the provinces of Ontario and Québec, and the West region refers to the provinces of British Columbia, Alberta, Saskatchewan, and Manitoba. Only a few cases were detected in the East region (provinces of Nova Scotia, New Brunswick, Prince Edward Island, and Newfoundland and Labrador) so these were grouped in Central region for epidemiological investigation.

Species complex definitions

Organism genus was confirmed using the RefSeq Masher Matches tool [23]. We used the following definitions for species complexes: the K. pneumoniae species complex includes K. pneumoniae, K. quasipneumoniae, and K. variicola [24]; the Enterobacter cloacae complex includes E. cloacae, E. hormachei, E. asburiae, E. kobei, and E. ludwigii [25]; and the Citrobacter freundii complex includes C. freundii, C. portucalensis, C. werkmanii, and C. youngae [26]. To assign species, Kleborate v2.2.0 [24] was used for species identification of Klebsiella with default parameters. Genomic clades and clusters in the Enterobacter spp. were defined by pairwise average nucleotide identity-based distance matrix using FastANI v1.3 [27,28] and the clade was assigned when the mean average nucleotide identity value was >95 %.

Whole genome sequencing and assembly

All 249 isolates were sequenced with Illumina MiSeq platforms and 88 of these were additionally sequenced using Oxford Nanopore Technologies (ONT). Isolates for ONT long-read sequencing represented about 33 % (88/249) of all OXA-48-type cases and isolates were selected to maintain representative proportions of each province in Canada. Genomic DNA was extracted using Epicentre MasterPure Complete kits (Mandel Scientific, Guelph, ON, Canada). The same DNA extract was used for both short-read Illumina and long-read ONT sequencing where possible. Short-read libraries were created with TruSeq Nano DNA HT sample preparation kits (Illumina, San Diego, CA, USA). Paired-end, 301 bp indexed reads were generated on an Illumina MiSeqTM platform (Illumina). Long-read sequences were generated using the Rapid Barcoding Kit (SQK-RBK004) or the Rapid Barcoding Kit 96 (SQK-RBK110.96) on R9.4.1 flow cells with the MinION Mk1B (ONT, Oxford, Oxfordshire, UK). Read data was basecalled and demultiplexed with Guppy v6.3.7 using the Super High Accuracy model (ONT). Average Illumina depth of coverage was 105X and average ONT depth of coverage was 72X.

Bioinformatic analyses

The assembly workflow was managed using Snakemake [29]. ONT reads were trimmed with Porechop v0.2.3_seqan2.1.1 [30] and filtered for Q-score>10 and length>1000 bases with Filtlong v0.2.1 [31]. Illumina reads had adaptors trimmed and were filtered for an average Q-score>30 with trim-galore v0.6.7 [32]. FastQC v0.11.9 [33] and Nanoplot v1.28.2 [34] were used to assess quality control metrics for Illumina and ONT reads respectively. Isolates with Illumina-only data were assembled with Unicycler v0.5.0 using default settings [35]. Those with Illumina and ONT data were assembled with Flye v2.9.2 [36], Raven v1.8.1 [37], and Miniasm v0.3_r179 [38], with the consensus assembly generated by Trycycler v0.5.0 [39]. A subset of these failed to generate a consensus with Trycycler (low ONT coverage, incomplete chromosomes) so were assembled using hybrid Unicycler v0.5.0 with default settings. Assemblies were polished with short reads using Polypolish v0.5.0 [40] and POLCA from MaSuRCA v4.0.9 [41]. A total of 70 OXA-48-type plasmids were circularized from these assemblies (Table S2).

StarAMR v0.9.1 [42] was used to detect antimicrobial resistance genes using the ResFinder database v2022-05-24 [43] and sequence type using the MLST database v2.23.0 [44,45]. Panaroo v1.3.2 [46] was used to estimate the pangenome. Plasmid taxonomic unit (PTU) designations were obtained from COPLA [47]. The MOB-typer tool from MOB-suite v3.1.4 [48,49] was used to identify plasmid replicons and mobility class using the default databases. The dnadiff tool which is part of MUMmer v3.23 [50] was used to align plasmid sequences to reference transposons (Table 1) with default settings and ISFinder database version 25 July 2023 [51] was used to delineate boundaries of insertion elements. SNVPhyl v1.2.3 [52] was used to investigate single nucleotide variants within primary plasmid clusters using the following parameters: min_coverage=10, min_mean_mapping=30, SNV_abundance_ratio=0.75, density threshold cutoff=2, size of search window=20. PHASTER (prophage/virus database version: 22 December 2020) was used to check for known prophages [53].

Table 1. Features of the complete OXA-48-type Canadian plasmids and their transposons.

Plasmid cluster	blaOXA-48-type gene (n/N)*	Tn type	Tn subtype (n/N)*	Plasmid and Tn reference
AA836 (IncX3/rep_cluster_1195)	181 (19/20)	Tn2013	Tn2013 (19/20)	KP400525.1
	1181 (1/20)	Tn2013	Tn2013 (1/20)	KP400525.1
AB871 (IncL/M)	48 (14/14)	Tn1999	invTn1999.1 (1/14)	AY236073.2
			Tn1999.2 (2/14)	JN714122.1
			invTn1999.2 (10/14)	JN714122.1
			Other (1/14)
AB484 (rep_cluster_1195)	181 (2/17)	Tn2013	truncTn2013 (1/17)	JX423831.1
			ISKpn26 insertion (1/17)
	232 (15/17)	Tn2013	truncTn2013 (14/17)	JX423831.1
			ISKpn26 insertion (1/17)
AC843 (IncC/IncF)	204 (4/4)	Tn2016	Tn2016 intact ISEcp1 (4/4)	CP047276.1
Chromosome	48 (9/27)	Tn6237	Tn6237 (2/9)	KT444705.1
			truncTn6237 (6/9)
			Other (1/9)
	204 (4/27)	prophage	C. freundii prophage	MG430338.1
	244 (12/27)	Tn51098	Tn51098 (2/12)	KR364794.1
			truncTn51098 (10/12)
	181 (2/27)	Tn2013	truncTn2013 (2/2)	JX423831.1

1*n/N indicates number of complete and circular Canadian plasmids with this feature (n) out of the total number of Canadian plasmids in the respective cluster (N).

Plots were created using R v4.3.0 [54] and the following packages: tidyverse packages [55], patchwork v1.1.2 [56], and ggpubr v0.6.0 [57].

Plasmid clustering and containment analysis

MOB-suite primary cluster designations are a useful way to broadly cluster plasmids for epidemiologically studies, and so plasmids assigned to different primary MOB-clusters are sufficiently unrelated to not be considered as part of an epidemiologically relevant transmission event [48,49]. However, plasmids that share the same primary cluster designation can be examined in more detail through higher resolution subtyping such as secondary cluster designations. If two plasmids are assigned to the same secondary cluster, they have near duplicate sequences and are sufficiently related to be strong candidates for outbreak investigations [48,49]. In addition to secondary cluster designation, epidemiological data is required to best assess direct plasmid transmission.

For plasmid clustering analysis, the PLSDB v2021_06_23v2 [58] database was downloaded and clustered alongside the 70 circular OXA-48-type plasmids in this study using MOB-cluster from the MOB-suite v3.1.4 package [48,49]. The primary and secondary cluster numbers generated here are unique to this manuscript and differ from those used in the default MOB-suite database. For plasmid containment analysis, all 575 circular plasmids completed in this study (including the 70 OXA-48-type plasmids) were clustered using MOB-cluster to create a custom Canadian plasmid database. All 249 isolates were screened for plasmids with MOB-recon using this custom database and the output was filtered to focus on the reconstructed plasmids containing bla_OXA-48-type genes.

Results

Characteristics of Canadian OXA-48-type carbapenemase-producing isolates

A total of 249 nosocomial OXA-48-type-producing isolates were submitted by 33 hospital sites across Canada (21 in the Central region, 12 in the West region) from 2010 to 2021 (see Methods for more detail) (Table S1). We performed short-read sequencing on all isolates and long-read sequencing on a selection of 88 isolates. The OXA-48-type-producing isolates belong to seven genera and twelve species, with the most common genera being the E. coli (128/249, 51.4 %), K. pneumoniae species complex (83/249, 33.3 %), and C. freundii complex (23/249, 9.2 %) (Fig. 1). The most common sequence types within species were E. coli ST38 (38/128, 30.0 %), E. coli ST410 (23/128, 18.0 %), C. freundii ST22 (18/23, 78.2 %), K. pneumoniae ST231 (11/83, 13.2 %), and K. pneumoniae ST147 (9/83, 10.8 %).

Fig. 1. Genera (inner ring), species (inner-middle ring), bla_OXA-48-type gene (outer-middle ring), and multi-locus sequence type (MLST; outer ring) of bla_OXA-48-type-encoding isolates in this study (249 total isolates). MLST profiles found in two or fewer isolates were grouped in to ‘other’. Not all labels are displayed. ‘Citrob.’ is abbreviation for Citrobacter.

Using the StarAMR tool for the detection of resistance genes in the whole genome sequencing data, we observed 92.0 % (229/249) of isolates harboured additional β-lactamase genes alongside the bla_OXA-48-type genes (Fig. 2). Of the 229 isolates harbouring additional β-lactamases, bla_TEM-1B (116/229, 50.7 %), bla_CTX-M-15 (112/229, 48.9 %), bla_OXA-1 (86/229, 37.6 %), bla_SHV-100 (25/229, 10.9 %), and bla_CMY-4 (21/229, 9.2 %) were the most common types. Sulfonamide, macrolide, quinolone, and aminoglycoside resistance genes were commonly observed among multiple genera; the most common genes included sul1 (137/249, 55.0 %), mph(A) (135/249, 54.2 %), qacE (135/249, 54.2 %), qnrS1 (108/249, 43.4 %), aph(6)-Id (103/249, 41.4 %), aph(3’)-Ib (102/249, 41.0 %), sul2 (101/249, 40.6 %), and aac(6’)-Ib-cr (99/249, 39.8 %). Klebsiella spp. and E. coli were equally likely to encode aminoglycoside or additional β-lactamase genes (P>0.05, 84 % vs 78% and 95 % vs 94 %). Fosfomycin resistance genes were more likely to be found in Klebsiella spp. than E. coli (P<0.001, 63 % vs 2 %), whereas tetracycline resistance genes were significantly more likely to be found in E. coli than Klebsiella spp. (P<0.001, 58 % vs 34 %).

Fig. 2. Proportion of isolates encoding antimicrobial resistance genes identified by StarAMR then categorized by drug class presented across the top three genera. Values represent the proportion of isolates encoding genes belonging to a certain antimicrobial class. ‘N =’ indicates the number of isolates in that genus. ‘Other’ genus includes Enterobacter spp., Providencia spp., Raoultella spp., and Shigella spp.

A total of 259 bla_OXA-48-type genes were detected among the 249 isolates, indicating an occurrence of 3.6 % of isolates habouring two copies of bla_OXA-48-type genes. The bla_OXA-181 variant was the most common (101/259, 39.0 %), followed by bla_OXA-48 (92/259, 35.5 %), bla_OXA-204 (22/259, 8.5 %), bla_OXA-232 (20/259, 7.7 %), bla_OXA-244 (19/259, 7.3 %), bla_OXA-484 (5/259, 1.9 %), and bla_OXA-1181 (1/259, 0.4 %).

Certain bla_OXA-48-type genes were associated with certain sequence types (ST) (Fig. 1). bla_OXA-181 was found in diverse E. coli and K. pneumoniae species complex ST backgrounds, however we noted that bla_OXA-181 was associated with E. coli ST410 (17/92, 18.5 %). In contrast, bla_OXA-232 was found mainly in Klebsiella species with a large proportion from ST231 (11/19, 57.9 %). bla_OXA-204 was found exclusively in C. freundii ST22 all isolated from a single hospital site. We also noted that E. coli ST38 comprised 47.4 % (9/19) and 31.5 % (29/92) of reported bla_OXA-244 and bla_OXA-48 genes, respectively.

Situating the complete and circular bla_OXA-48-type Canadian plasmids within a global plasmid dataset

Of 249 sequenced isolates we obtained only two closed circular plasmids from Illumina-only assemblies. Using a hybrid assembly approach on a subset of 88 isolates, we obtained an additional 68 OXA-48-type complete circular plasmids and identified bla_OXA-48-type genes on 27 complete chromosomes. Complete OXA-48-type plasmids (n=70) ranged from 6.1 kb to 364.4 kb and were distributed among the top species as described above (Fig. 3 and Table S2). Common plasmid incompatibility groups included rep_cluster_1195 (also known as ColKP3 in PlasmidFinder [59] or ColE2-type [8]) (42/70, 60.0 %), IncX3 (21/70, 30.0 %), IncL/M (15/70, 21.4 %), IncF-type (16/70, 22.9 %), with half of the plasmids containing two or more replicons (35/70, 50.0 %). Plasmids were predicted to be conjugative (48/70, 68.6 %), mobilizable (20/70, 28.5 %), or non-mobilizable (2/70, 2.9 %) based on the presence of mating pair formation (MPF) proteins, relaxases, and self-encoded oriT sequences detected by MOB-suite [48].

Fig. 3. Characteristics of 70 complete OXA-48-type plasmids sequenced in this study. Groups on the x-axis correspond to primary cluster IDs generated by de novo clustering of the global dataset (PLSDB and Canadian OXA-48-type plasmids). Tn type indicates transposon type. other* indicates Enterobacter, Providencia, Raoultella, and Shigella genera; other** indicates rep_cluster_1418, IncHI1B, and ColKP3 replicons; unknown indicates an uncharacterized or truncated transposon.

To investigate where these plasmids fit within the global context, we clustered our 70 complete plasmids alongside 34 513 plasmids present in PLSDB [58] using the MOB-cluster tool from MOB-suite [48,49], referred to hereafter as the global dataset. Our OXA-48-type plasmids grouped into a subset of 14 primary clusters, all of which contained representative plasmids from the global dataset. Primary clusters are defined as having a pairwise Mash distance of <0.06, and can contain multiple secondary clusters with a pairwise Mash distance of <0.025.

We observed that four primary clusters represented 78.6 % (55/70) of Canadian plasmids and tightly clustered (Mash distance <0.025) with other known OXA-48-type plasmids isolated from multiple genera in the global dataset (Fig. 4 and Table 2). Within the global dataset, the presence of bla_OXA-48-type alleles varied between the top primary clusters. Of the top primary clusters, bla_OXA-48-type genes were found on 96.9 % (primary cluster AB871), 95.9 % (primary cluster AB484) and 22.3 % (primary cluster AA836) of plasmids, respectively. In three of four top primary clusters, plasmids carried only a few antimicrobial resistance genes (averages of 1.1, 1.5, and 1.7 resistance genes per plasmid); primary cluster AC843 was an exception with an average of 12.4 resistance genes per plasmid, although most of these plasmids (21/23, 91.3 %) were harboured by Citrobacter species indicating this may be a clonal trend. The percentage of core/soft core genes (defined as genes found in >95  % of plasmids) varied from 8 % (AA836) to 13 % (AB871). Primary cluster AA836 had the broadest host range with plasmids being found in twelve genera, whereas primary cluster AC843 plasmids were only found in three genera to date.

Fig. 4. Pangenome size and bla_OXA-48-type prevalence among top four MOB-suite primary clusters identified among the global plasmid dataset (PLSDB and Canadian OXA-48-type plasmids). Pangenome size was calculated for (a) the global plasmid dataset (Canadian and PLSDB) and (b) Canadian plasmids only. Gene categories represent genes found in 99–100 % of plasmids (core), 95–99 % of plasmids (soft core), 15–95 % of plasmids (shell) and 0–15 % of plasmids (cloud), and N indicates the total number of genes identified per cluster. (c) Plasmid length in base pairs and prevalence of bla_OXA-48-type genes among plasmids in each primary cluster in the global dataset. N indicates the total number of plasmids per cluster.

Table 2. Summary features of top four primary plasmid clusters containing Canadian OXA-48-type plasmids and PLSDB plasmids from the global dataset.

Primary cluster ID	Num. Canadian plasmids	Median size (kb)	Predicted mobility*,†	Replicon type*,†	Relaxase type*,†	MPF type*,†	Number of genera	Core genes (%)§	blaOXA-48-type allele*	blaOXA-48-type freq	Mean number of ARGs per plasmid‡	Plasmid Taxonomic Unit (PTU)*,¶
AA836	20	46.3	conjugative	IncX3/rep_cluster_1195	MOBP	MPF_T	12	0.08	bla OXA-181	0.22	1.5	PTU-X3
AB871	14	63.6	conjugative	IncL/M	MOBP	MPF_I	6	0.13	bla OXA-48	0.97	1.7	PTU-L/M
AB484	17	6.1	mobilizable	rep_cluster_1195	MOBP	None	4	0.12	bla OXA-232	0.96	1.1	PTU-E37
AC843	4	206.6	conjugative	IncC,IncFIB,IncFIC,IncFII	MOBF	MPF_F	3	0.12	bla OXA-204	0.43	12.4	PTU-C

1v*Values indicate the most common genotype in the cluster and may not apply to all plasmids in the cluster.

2 †vValues obtained from MOB-suite. Mobility is assigned based on presence of relaxase (mobilizable) and/or MPF proteins (conjugative) or absence of both (non-mobilizable).

3‡ARGs=antimicrobial resistance genes, including bla_OXA-48-type genes.

4§Core genes represents the number genes present in >95 % of plasmids in the cluster, divided by the total number of non-redundant genes in the cluster.

5¶PTU values obtained from COPLA.

All top primary plasmid clusters except for AB484 contained plasmids encoding one or multiple genes involved in stability/transfer/defence to support their persistence in the host cell. Primary cluster AB871 plasmids encoded two plasmid partition stability genes (parM and an unnamed plasmid family stability protein), as well as the pemKI toxin-antitoxin system. All plasmids in this primary cluster also encode the relB antitoxin. Most plasmids encoded ssb, the single-stranded binding protein (175/195, 89.7 %), and most also encoded an antirestriction gene (174/195, 89.2 %). All but one plasmid in primary cluster AC843 (22/23, 95.7 %) encoded two plasmid partition stability proteins (both labelled parB), plasmid SOS inhibition genes psiA and psiB, an antirestriction gene, a single-stranded binding protein ssb, and the EcoRII restriction modification system (including both the restriction enzyme and methyltransferase). Primary cluster AA836 encoded a parA partition gene in 99.3 % (n=400/403) of plasmids. No other stability/transfer/defence genes were detected in >10 % of plasmids in the top primary clusters.

The remaining nine primary clusters contained two or fewer Canadian plasmids per cluster, which suggests that they are not dominant in the Canadian population of OXA-48-type plasmids. These typically had IncF-type replicons that often appeared to be hybrids with IncH, IncL/M, IncX3, rep_cluster_1195, or other replicons, and the majority (10/15, 66.7 %) were predicted to be conjugative due to the presence of relaxases and mating pair formation proteins. Given the high proportion of conserved genes among Canadian plasmids within each top primary cluster (Fig. 4), we focused on the features of the Canadian plasmids separately from plasmids in global dataset in more detail in the next section.

Canadian plasmids in primary cluster AA836: bla_OXA-181 on IncX3/rep_cluster_1195 replicons

The Canadian plasmids in primary cluster AA836 encoded two replicons (IncX3 and rep_cluster_1195) and were classified into two secondary clusters: AM772 (19/20) and AM770 (1/20). Plasmids in secondary cluster AM772 were 51.5 kb in length and were isolated from K. pneumoniae species complex, K. oxytoca, E. coli, Citrobacter werkmannii, and Raoultella ornithinolytica between 2015 and 2020 from sixteen hospital sites in four provinces. These plasmids were >99.99 % identical and had 0–4 SNVs relative to the first reported case of bla_OXA-181 on an IncX3/rep_cluster_1195 plasmid from China (pOXA181_EC14828, KP400525.1) [20]. One of these plasmids contained a SNV in bla_OXA-181 which corresponded to bla_OXA-1181. In addition to bla_OXA-181 in Tn2013, the only other antimicrobial resistance gene found on these plasmids was qnrS1. The plasmid assigned to secondary cluster AM770 is a hybrid plasmid which contains the 51.5 kb IncX3/rep_cluster_1195 replicon along with an additional IncFIB replicon and MOBC relaxase for a total size of 91.2 kb. This plasmid encodes two aminoglycoside resistance genes along with bla_OXA-181 and qnrS1. The prevalence of secondary cluster AM772 plasmids among multiple genera and provinces along with the presence of conjugation genes (MOBP relaxase and MPF_T mating pair formation protein) suggest that horizontal transmission has contributed to these plasmids’ persistence in Canada. bla_OXA-181 is known to be associated with the truncated Tn2013 transposon in IncX3/rep_cluster_1195 plasmids alongside qnrS and several IS elements: IS3000, ISKpn19, IS26, and IS2-type [8,20]. All twenty Canadian plasmids in primary cluster AA836 have this same Tn2013 structure [20] (Table 1).

We examined the location of the SNVs in the secondary cluster AM772 plasmids in our dataset relative to the reference plasmid KP400525.1 [20] and did not observe any notable trends. The SNVs (11 in total) are located in several transposases (ISKox3, IS3000, and ISKpn19), bla_OXA-181 (corresponding to bla_OXA-1181), two ATPases, a DUF4158 domain-containing protein, and several intergenic regions. There were two groups within the AM772 plasmids which had zero SNV differences between them; the first contained eight plasmids which were identical to the reference KP400525.1. This group contained plasmids from five different provinces, and were isolated between 2018 and 2021. The second group contained three plasmids isolated from only one province from various species between 2018 and 2020. The remaining plasmids have 1–3 SNVs relative to the others, with no other plasmids containing the same set of SNVs.

Using MOB-recon to search against the clusters generated from our complete and circular OXA-48-type plasmids, we examined the Illumina-only data for remaining n=160 (of total 249) isolates. MOB-recon reconstructs plasmids from incomplete Illumina data and we were able to predict plasmid clusters for most isolates (147/160, 91.9 %) when using our complete and circular OXA-48-type plasmids as references. We found 34.4 % (55/160) of isolates with Illumina-only data were predicted to be part of IncX3/rep_cluster_1195 (primary cluster AA836 and secondary cluster AM772) plasmids, providing further support for the prevalence of this plasmid type in Canada. Indeed, between 2019 and 2022, 40 % (55/136) of all OXA-48-type plasmids were from cluster AA836. This type of plasmid was commonly observed in E. coli ST410 (18/75, 24.0 %) in our dataset, of which 14 were observed since 2019, indicating this clonal lineage has contributed to this plasmid’s persistence specifically since 2019.

Canadian plasmids in primary cluster AB871: bla_OXA-48 on IncL/M replicons

The fourteen Canadian plasmids in primary cluster AB871 all grouped in the same secondary cluster AO777. These plasmids varied from 62.8 kb to 69.0 kb in size and were isolated from E. coli, C. freundii, and K. pneumoniae species complex between 2015 and 2020 from eleven hospital sites in four provinces. No other antimicrobial resistance genes were found on these plasmids aside from bla_OXA-48 within Tn1999. These plasmids are >97 % identical to the reference IncL plasmid pOXA-48a (accession JN626286.1) [14,59, 60] with variable insertions or inversions (Fig. S1). These insertion/deletion patterns appear to be random and do not match epidemiological information (year, hospital site, province). The prevalence of secondary cluster AO777 plasmids among multiple genera and provinces along with the presence of conjugation genes (MOBP relaxase and MPF_I mating pair formation protein) suggest that horizontal transmission has contributed to these plasmids’ persistence in Canada.

The Tn1999 composite transposons are known to encode bla_OXA-48 in IncL/M plasmids [8,19]. We found 71 % (n=10/14) of IncL/M plasmids encode the inverted Tn1999.2, which is a Tn1999 variant with IS1R inserted into IS1999 upstream of bla_OXA-48 that has inverted between the tir gene insertion site (Table 1) [61,63]. In addition, two plasmids encoded the normal (uninverted) Tn1999.2, one encoded an inverted Tn1999, and one encoded a modified Tn1999.2 with a 13 kb insertion.

MOB-recon predicted 20.6 % (n=33/160) of isolates generated from Illumina-only data that grouped in this same IncL/M primary (AB871) and secondary cluster (AO777). No plasmid distribution trends across Canada related to hospital site, organism, or type were observed among isolates harbouring these contigs.

Canadian plasmids in primary cluster AB484: bla_OXA-232 and bla_OXA-181 on rep_cluster_1195 replicons

The seventeen Canadian plasmids in primary cluster in AB484 all grouped in the same secondary cluster AO082. These plasmids were 6.1 kb with two exceptions at 7.3 kb. All had a replicon identified as rep_cluster_1195 by MOB-suite (also named as ColKP3 by PlasmidFinder [8,59]), and were classified as mobilizable by MOB-suite due to the presence of a MOBP relaxase but did not contain MPF genes. The isolates containing these plasmids contained an average of six plasmids each, which suggests that the 6.1 kb plasmids could perhaps be mobilized by conjugation machinery in co-resident plasmids in those isolates. These plasmids mainly carried bla_OXA-232 (15/17, 88.2 %) with two exceptions carrying bla_OXA-181, and no other resistance genes were found. These plasmids are >99.99 % identical to reference plasmid pOXA-232 first isolated in 2012 (accession JX423831.1) [64]. These plasmids were isolated from E. coli, R. ornithinolytica, and K. pneumoniae species complex between 2015 and 2020 from thirteen hospital sites in four provinces.

The small 6.1 kb rep_cluster_1195/ColKP3/ColE2 plasmids are associated with Tn2013 which contains a large deletion on the 5′ end of ISEcp1 [8,64]. We found that this structure was preserved in all of our primary cluster AB484 plasmids (Table 1). Many (10/17, 58.8 %) have a single SNV relative to the reference sequence at positions 4282 (1/10), 4283 (1/10), or 4286 (8/10), which corresponds to an intergenic region between ΔereA and repA. Additional data on plasmid SNVs did not indicate patterns in space or time.

MOB-recon predicted an additional 5.0 % (n=8/160) of isolates generated from Illumina-only data that grouped in this same primary (AB484) and secondary cluster (AO777). This secondary cluster was observed for all K. pneumoniae ST231 isolates in our dataset (11/11, 100 %). Indeed, of all bla_OXA-232 observed across Canada, 52.4 % (11/21) were associated with ST231, with no overall epidemiological links.

Canadian plasmids in primary cluster AC843: bla_OXA-204 on IncC/IncF plasmids

We obtained four complete plasmids which were all classified into primary cluster AC843 and secondary cluster AQ458; all plasmids encoded bla_OXA-204, had IncC/IncFIB/IncFIC/IncFII replicons, were classified as conjugative by MOB-suite, and were found in C. freundii ST22 isolated from a single hospital site. We previously characterized an outbreak of C. freundii ST22 encoding bla_OXA-204 in Canada from 2016 to 2018 that was attributed to chromosomal genes and three distinct plasmid types [65]. These isolates were not included in our dataset; however, plasmids from this study are part of the global dataset, six of which grouped into the same primary and secondary cluster as the four new bla_OXA-204 plasmids sequenced here (AC843/AQ458). Given the primary cluster assignment is identical between the old and new plasmids, this strain/plasmid combination appears to still be circulating in the same region. On these plasmids, the bla_OXA-204 gene was found within the Tn6750, which is a variation of the Tn2016 transposon with an intact ISEcp1 [8,65]. In addition to encoding bla_OXA-204 on plasmids, these four isolates had a second copy of bla_OXA-204 encoded within a prophage on their chromosomes, which was also observed in the 2016 to 2018 outbreak isolates [65].

A total of 18 isolates in our dataset encoded bla_OXA-204 on complete or incomplete contigs, all of which were C. freundii ST22 isolated from the same hospital site in 2016 (n=2), 2019 (n=7), 2020 (n=6) and 2021 (n=3). Fourteen of these had Illumina-only data and all but one were predicted to encode bla_OXA-204 on a contig assigned to this same primary cluster (AC843; 17/18, 94.4 %), with the one exception remaining was unclassified. Further investigation is required to confirm these plasmid structures.

Interestingly, no other isolate in our dataset had or was predicted to have a plasmid in this same AQ458 secondary cluster, indicating that this plasmid type has not been observed outside of this region.

Other bla_OXA-48-type genes on chromosomes in Canadian isolates

We observed four bla_OXA-48-type variants on the chromosome in 27 isolates: bla_OXA-48 (n=9), bla_OXA-244 (n=12), bla_OXA-204 (n=4), and bla_OXA-181 (n=2). Unlike the other bla_OXA-48-type variants, bla_OXA-244 was found exclusively on chromosomes and exclusively in E. coli in our dataset. Interestingly, E. coli ST38 isolates commonly harboured either bla_OXA-244 (n=6/12) or bla_OXA-48 (n=7/9) on the chromosome.

Chromosomal bla_OXA-48 was associated with Tn6237, which is a 21.9 kb IS1R-based composite transposon that contains an inverted Tn1999.2 and a plasmid-derived fragment of pOXA-48a. Tn6132 is 99.9 % identical to Tn51098, which is also a 21.9 kb IS1R-based composite transposon but contains bla_OXA-244 [8]. We found Tn6237 was present in some form in most isolates that had a chromosomal bla_OXA-48 (n=8/9) (Table 1); only two contained the full 21.9 kb length, whereas others (n=6/9) had a truncated form that was between 10 kb and 20 kb in length. One isolate contained a 3 kb fragment of lysR and bla_OXA-48 that was not associated with any other Tn6237 genes. There was not a conserved genomic locus among isolates where all Tn6237 genes were inserted, and none of the isolates here had the Tn6237 insertion in the same region as reference KT444705.1.

Chromosomal bla_OXA-244 genes in varying lengths of Tn51098 were located at the same genomic locus in all isolates (Fig. S2), all of which were E. coli (n=12) and represented seven different STs. This genomic locus is the same as reference KR364794.1 [66]. Several isolates had the 21.9 kb full length Tn51098 match (2/12) (corresponding to genetic environment A in [67]), but the majority had lysR interrupted by ISR1 (7/12) (corresponding to genetic environment G in [67]), resulting in a 3.3 kb truncated fragment. Others had a 15 kb truncated fragment of Tn51098 (1/12, corresponding to genetic environment D in [67]), a 7 kb truncated fragment of Tn51098 (1/12), or a duplication of the bla_OXA-244/ISR1-truncated lysR immediately downstream of the 3.3 kb fragment (1/12).

There were two instances of chromosomal bla_OXA-181 which were found in Tn1999.3 structures; one was inserted into a fimbrial adhesion operon and the other was integrated at an rRNA/tRNA site.

Some of the incomplete bla_OXA-48-type-encoding contigs were not predicted to be part of any primary plasmid cluster or chromosome (13/160, 8.1 %), some of which (5/13, 38.5 %) were less than 2.3 kb which is close to the length of IS10A which makes classification challenging. Further long-read sequencing would be required to verify if the genomic locus of bla_OXA-48-type genes in these isolates. This method is also unable to predict chromosomal insertions of bla_OXA-48-type genes, so contigs encoding genes that were primarily predicted to be chromosomal (i.e. bla_OXA-244) but were assigned plasmid clusters should be interpreted with caution.

Epidemiology of bla_OXA-48-type plasmids and contigs across Canada

We examined the temporal and geographic patterns of all complete OXA-48-type plasmids and included the predicted OXA-48-type plasmid clusters based off Illumina-only data in Canada from 2010 to 2021 (Fig. 5). Isolates from the central region dominated our dataset (Central: 201/249, 80.7 %; West: 48/249, 19.3 %). Occurrences of bla_OXA-48-type genes in Canada were low until 2016. Primary cluster AA836 (IncX3/rep_cluster_1195) was the most common plasmid type detected in both West and Central Canada since 2018. Primary cluster AB871 (IncL/M) was one of the first primary clusters that became dominant in Canada and has been isolated sporadically in the Central region since then. Primary cluster AB484 (rep_cluster_1195) plasmids were observed at low frequencies since 2013. Primary cluster AC843 (IncC/FIB/FIC/FII) were detected in 2016 and from 2019 onwards, indicating that this plasmid continues to circulate at this hospital site.

Fig. 5. Epidemiology of Canadian OXA-48-type plasmid primary clusters from 2010 to 2021. Date of positive culture was grouped into year-month bins. This data includes both complete circular plasmids and incomplete bla_OXA-48-type-encoding contigs. The Central region represents the provinces of Ontario and Quebec, and the West region represents the provinces of British Columbia, Alberta, and Manitoba. ‘Other’ indicates plasmids which did not group in the top four primary clusters. N indicates total number of isolates in each cluster. Chromosomal bla_OXA-48-type genes were excluded.

Discussion

We examined the prevalence and distribution of bla_OXA-48-type-producing Enterobacterales and their plasmids in Canada from 2010 to 2021. The majority of OXA-48-type carbapenemase occurrences in Canada during this time period were predominantly from E. coli and K. pneumoniae species, which agrees with other countries [15,68,70]. This contrasts with bla_KPC, which was found in more diverse genera in Canada during the same time period [71], and with bla_NDM, which also was found in diverse genera around the world [72].

We searched for patterns among plasmid clusters that would allow us to predict dissemination of OXA-48-type plasmids in Canada. We examined SNV prevalence in two primary plasmid clusters whose plasmids were near identical (AA836, AB484). For IncX3/rep_cluster_1195 plasmids carrying bla_OXA-181 (AA836), we found remarkably high conservation among plasmid sequences wherein many were 100 % identical to the reference sequence, agreeing with previous studies [8,15, 69, 73, 74]. Although we found 1–3 SNVs in some plasmid sequences, these were scattered across the genome and did not appear to be linked to any epidemiological information which limits the usefulness of creating a Tn2013 transposon typing scheme, as done for bla_KPC genes with Tn4401 [71,75]. Similarly for the 6.1 kb rep_cluster_1195 plasmids (AB484) plasmids, we observed 0–2 SNVs between plasmids isolated from organisms that have no epidemiological links. Overall, our dataset shows plasmid SNV data is not enough alone to track plasmid spread, and that epidemiological data is critical for tracing transmission of these plasmid clusters due to high levels of sequence conservation.

The IncX3/rep_cluster_1195 (AA836) was the most common plasmid cluster found in Canada. Accordingly, bla_OXA-181 and the Tn2013 transposon were also the most common in our dataset. This is in contrast to multiple European countries, wherein bla_OXA-48 is the most common bla_OXA-48-type variant observed paired with IncL/M plasmids [15,68,70, 76].

The IncL/M (AB871) were the second most common primary plasmid cluster in Canada when including plasmids predicted in Illumina-only data (Table 2). David et al. [69] observed 82 % of K. pneumoniae species complex isolates were predicted to carry IncL/M plasmids in the EUSCAPE study. This incidence is much higher than what we predicted here, wherein only 32.5 % (27/83) of K. pneumoniae species complex isolates are predicted to carry an IncL/M pOXA-48a-type plasmid. David et al. [69] also showed these plasmids were highly related, with most (76 %) within two SNVs of each other. We observed small-scale recombination and inversions in our plasmids (Fig. S1), which may be explained by larger time gap (2015 to 2020) in our dataset compared to the 2 years (2013–2014) in the David et al. publication [69].

There were some clear associations between certain plasmid and bla_OXA-48-type genes with certain species/sequence types (Fig. 1), where these clones are not restricted to a geographic region/hospital site and have been transmitted across Canada. For example, all K. pneumoniae ST231 isolates in this study, which were not linked epidemiologically, encoded bla_OXA-232 on the 6.1 kb rep_cluster_1195 (AB484) plasmids, which has been observed in other countries [18,64, 77]. Similarly, many isolates (18/23, 78.2 %) in the clonal lineage E. coli ST410 encoded bla_OXA-181 on IncX3/rep_cluster_1195 plasmids. This multi-drug resistant and diverse lineage has been circulating worldwide with this plasmid for since 2003 [78]. Finally, E.coli ST38 was noted in occurrences of bla_OXA-244 (47.4 %) which were all chromosomally integrated and in occurrences of bla_OXA-48 (31.5 %), both on plasmids and chromosomally integrated. E.coli ST38 is well documented as a high risk clone and has contributed to the dissemination of bla_OXA-48-type genes, often by chromosomal integration [79].

Chromosomal integration was observed in 10.8 % of isolates and may be linked to persistence of OXA-48-type carbapenemases in Canada. The persistence of bla_OXA-244 due to chromosomal integration has been suggested by Emeraud et al. [67], whom observed a tendency for more recent isolates to encode a truncated form of Tn51098, whereas older strains tended to encode a full length transposon. This was also what we observed at the same genomic locus. This truncation suggests that transposition may no longer be possible as bla_OXA-244 becomes a permanent part of the genome. In contrast, the bla_OXA-48 chromosomal insertion sites were not conserved, nor was the length of insertion, which suggests that multiple transposition and recombination events have occurred [15,63].

The ongoing occurrences of bla_OXA-204 in C. freundii ST22 at a single hospital site in Canada has not been observed elsewhere in the world [65]. bla_OXA-204 was first detected in C. freundii isolated from wastewater samples in Tunisia between 2013 and 2015 [80]. C. freundii ST22 isolates encoding bla_OXA-204 were observed in France between 2019 and 2020 but no analysis of the genomic location of the bla_OXA-204 genes was completed [81]. While these isolates comprise only a small proportion of the total bla_OXA-48-type-encoding isolates in this dataset, the longevity of this endemic lineage merits further investigation to determine why this strain/plasmid combination has remained localized and has not spread to other regions in Canada or elsewhere.

Multiple OXA-48-type primary clusters (AA836, AB871, AB484) that are dominant in Canada and spread by horizontal transmission notably lacked other antimicrobial resistance genes. Only bla_OXA-48 was found on IncL/M (AB871) and rep_cluster_1195 (AB484) plasmids, whereas only bla_OXA-181 and qnrS1 were found on IncX3/rep_cluster_1195 (AA836) plasmids. These findings agree with Hendrickx et al. [15], who also observed a lack of antimicrobial resistance genes on bla_OXA-48-type plasmids in the Netherlands, and with a recent analysis of IncL/M plasmids in NCBI RefSeq which found 82.7 % (162/196) of IncL/M plasmids encoded bla_OXA-48 as the only resistance gene [82]. The lack of antimicrobial resistance genes on OXA-48-type plasmids contrasts with other types of carbapenemase plasmids that frequently have multiple other types of antimicrobial resistance genes, such as bla_KPC [71] and bla_NDM genes [72,83].

Overall, our results suggest horizontal transmission (AB871), clonal transmission (AC843), and a combination of both horizontal and clonal transmission (AB484 and AA836) have contributed to OXA-48-type carbapenemase persistence in Canada. Specifically, the emergence of AA836 over the last 4 years indicates a significant driver of OXA-48-type dissemination in Canada. Given that primary cluster AA836 is mobilizable, associated with a wide variety of species, and also associated with the high-risk E.coli ST410 lineage, multiple factors are all likely contributing to its persistence. All dominant primary plasmid clusters have been observed elsewhere in the world, with the exception of bla_OXA-204 on IncC/IncF plasmids (AC843) which is unique to Canada. The use of SNV data (AA836, AB484) or trends in plasmid homology (AB871) were not successful in distinguishing top Canadian plasmids by location or time. We suggest maintaining high-quality reference plasmid databases for characterizing OXA-48-type plasmids in new nosocomial isolates. Tracking dissemination of specific bla_OXA-48-type genes requires plasmid clustering alongside molecular analysis of high-risk lineages in combination with epidemiological data when investigating plasmids with such high sequence homology.

Abbreviations

ARG

antimicrobial resistance gene

MPF

mating pair formation

ONT

Oxford Nanopore Technologies

OXA-48

oxacillinase-48

PTU

Plasmid Taxonomic Unit