INTRODUCTION
In the current study, we investigated the epidemiology of third-generation cephalosporin- and carbapenem-resistant ST410
(The data were partly presented at DANMAP 2014, DANMAP 2015, and DANMAP 2016.)
RESULTS
Genotypic characterization.
Genotypic characterization was applied to the 127
Of the 127 ST410 genomes, 114 genomes contained at least one of 14 different ESBL genes, plasmid-mediated AmpC (pAmpC) genes, or carbapenemase genes or a truncated pAmpC gene, whereas no gene was detected in 13 genomes (Table 1). The majority of the genomes carried blaCTX-M-15 (66%), followed by blaOXA-181 (53%), blaCMY-2 (48%), blaNDM-5 (13%), and blaCMY-42 (10%) (Table 1).
TABLE 1
Combination of extended-spectrum β-lactamase, pAmpC, and carbapenemase enzymes in the 127
| Phenotype | Gene | No. of genomes | % of genomes |
|---|---|---|---|
| None | No gene detected | 13 | 10 |
| ESBL | blaCTX-M-15 | 21 | 17 |
| blaCTX-M-1 | 2 | 2 | |
| blaSHV-12 | 1 | <1 | |
| blaCTX-M-1/blaCTX-M-15 | 1 | <1 | |
| pAmpC | blaCMY-42 | 7 | 6 |
| blaCMY-2 | 1 | <1 | |
| Carbapenemase | blaOXA-181 | 1 | <1 |
| ESBL/pAmpC | blaCTX-M-15/blaCMY-2 | 1 | <1 |
| blaCTX-M-15/blaCMY-42/blaDHA-1 | 1 | <1 | |
| ESBL/carbapenemase | blaCTX-M-15/blaOXA-181 | 3 | 2 |
| blaCTX-M-15/blaKPC-2 | 2 | 2 | |
| blaCTX-M-15/blaNDM-1 | 1 | <1 | |
| blaCTX-M-15/blaKPC-3 | 2 | 2 | |
| blaCTX-M-15/blaOXA-48 | 1 | <1 | |
| blaTEM-30/blaKPC-3 | 2 | 2 | |
| pAmpC/carbapenemase | blaCMY-2/blaOXA-181 | 11 | 9 |
| blaCMY-2/blaNDM-5 | 1 | <1 | |
| blaCMY-42/blaOXA-181 | 1 | <1 | |
| blaDHA-1/blaOXA-48 | 1 | <1 | |
| blaCMY-2/blaNDM-5/blaOXA-181 | 2 | 2 | |
| ESBL/pAmpC/carbapenemase | blaCTX-M-15/blaCMY-2/blaOXA-181 | 29 | 23 |
| blaCTX-M-15/blaCMY-42/blaOXA-181 | 3 | 2 | |
| blaCTX-M-15/blaCMY-2-truncated/blaOXA-181 | 2 | 2 | |
| blaCTX-M-15/blaCMY-42/blaOXA-48 | 1 | <1 | |
| blaCTX-M-15/blaCMY-2/blaNDM-5/blaOXA-181 | 9 | 7 | |
| blaCTX-M-15/blaCMY-2/blaNDM-4/blaOXA-181 | 3 | 2 | |
| blaCTX-M-15/blaSHV-12/blaCMY-2/blaNDM-5 | 1 | <1 | |
| blaCTX-M-15/blaSHV-12/blaCMY-2/blaNDM-5/blaOXA-181 | 3 | 2 |
From the analysis of the quinolone resistance-determining regions (QRDRs) in the 127 ST410 genomes, 115 of the genomes had S83L and D87N amino acid substitutions in gyrA, S80I substitution in parC, and S458A substitution in parE. None of the 127 genomes had amino acid changes in the QRDR of gyrB (Table S1).
Two genomes from the United States (SAMN06645838 and SAMN02442857) and one genome from Denmark (SAMN01885884) belonged to fimH53 (H53). The remaining 124 genomes belonged to fimH24 (H24) (Table S1).
Of the 127 ST410 genomes, 66 (52%) had an IncX3 replicon and 125 (98%) had an IncF replicon, with the majority (58/125, 46%) belonging to pMLST F1:A1:B49, while 24 (19%) belonged to F36:A4:B1, seven (6%) belonged to F31:A4:B1, six (5%) belonged to F48:A1:B49, and three (2%) belonged to F2:A4:B1. For the remaining 27 genomes, 23 different pMLSTs were identified (Table S1).
Epidemiology of
The epidemiology of the 49 ST410 isolates from the Danish surveillance program was investigated by single nucleotide polymorphism (SNP) analysis and visualized with the metadata year of isolation, third-generation cephalosporin resistance (3GC-R) genes, and IncX3 and IncF replicons in a phylogenetic tree (Fig. 1). For construction of the phylogenetic tree, 721 SNPs were identified between the 49 genomes, based on 86% of the reference chromosome after removal of recombinant regions.
FIG 1
Rooted maximum likelihood phylogeny constructed with Gubbins based on SNP alignment of 49 Danish
The SNP analysis was used to identify possible clones (PC) within the ST410 national Danish strain collection. A pairwise comparison of the SNP differences across the isolates (Fig. S1A) showed a group of isolates with no more than 40 SNPs to the nearest neighbor (0 to 41 SNPs in pairwise comparison), two isolates with 80 SNPs between the genomes, and the remainder with more than 150 SNPs to the nearest neighbor. Adhering to an SNP distance of ≤40 SNPs to the nearest neighboring isolate for describing possible clones (PC40) (Table 2) resulted in four PC40s (PC40-1, -2, -3, and -4). Additionally, possible outbreaks (PO; clusters containing at least three individual isolates) were identified by adhering to an SNP distance of ≤10 SNPs to the nearest neighboring isolate (PO10). Based on this cutoff value, five possible outbreaks were identified (PO10-1, -2, -3, -4, and -5) (Fig. 1).
TABLE 2
Possible clones and outbreaks identified in the Danish ST410
| PC40 | No. of | Region(s) | No. of | PO10 | No. of | Region(s) | No. of | Yr(s) (no. of | Resistance enzyme(s) | Replicon(s) |
|---|---|---|---|---|---|---|---|---|---|---|
| PC40-1 | 3 | Ce (3) | 0–4 | PO10-1 | 3 | Ce (3) | 0–4 | 2015 (1), | CTX-M-15 | IncF |
| PC40-2 | 6 | Ce (6) | 0–2 | PO10-2 | 6 | Ce (6) | 0–2 | 2015 (5), | CMY-2 (5) | IncF |
| PC40-3 | 28 | Ca (13), | 0–41 | PO10-3 | 11 | Ca (1), | 0–12 | 2015 (1), | CMY-2, OXA-181 (10), | IncF, IncX3 |
| PO10-4 | 3 | So (1), | 3–6 | 2016 (3) | CMY-2, OXA-181, | IncF, IncX3 | ||||
| PO10-5 | 3 | Ca (1), | 2–3 | 2015 (1), | CMY-2, OXA-181, | IncF, IncX3 | ||||
| PC40-4 | 4 | Ca (4) | 15–38 |
a
Abbreviations: PC40, possible clone with ≤40 SNPs to the nearest neighboring isolate; PO10, possible outbreak with ≤10 SNPs to the nearest neighboring isolate; Ce, central region of Denmark; Ca, capital region of Denmark; Ze, Zealand region; So, southern region of Denmark.
PC40-1 encompassed three genomes with 0 to 4 SNP differences, thus also indicating a possible outbreak (PO10-1). The three isolates were from 2015 (n = 1) and 2016 (n = 2); all three genomes contained the blaCTX-M-15 gene and an IncF replicon with pMLST F36:A4:B1. Travel information for the three patients was not available, but the patients had all been hospitalized concurrently in the same hospital unit, R-2 (Fig. S3), in the central region of Denmark (map in Fig. S2).
The second possible clone, PC40-2, included six genomes with 0 to 2 SNP differences, also indicating a possible outbreak (PO10-2). The six isolates were from 2015 (n = 5) and 2016 (n = 1) and originated from five patients hospitalized in the central region of Denmark. An IncF replicon with pMLST F48:A1:B49 and blaCMY-42 were present in 5/6 genomes. Travel information was unavailable for all five patients associated with these six isolates. However, patients P1, P3, P4, and P5 were hospitalized concurrently at the same hospital unit (R-8) in the beginning of June, and patients P2 and P4 also simultaneously at R-8 in mid-June (Fig. S4).
The third possible clone (PC40-3) comprised 28 isolates, collected from patients hospitalized in the capital region of Denmark (n = 13), the Zealand region (n = 10), the southern region of Denmark (n = 1), and the central region of Denmark (n = 4). The genomes carried blaCMY-2 (28/28) and blaOXA-181 (27/28); IncF replicons with pMLST F1:A1:B49 (25/28), F1:A6:B49 (2/28), or F2:A1:B49 (1/28); and an IncX3 replicon (28/28). PC40-3 spanned between 0 and 41 SNPs and featured three possible outbreaks, PO10-3, PO10-4, and PO10-5, described in an epidemiological context below.
PO10-3 comprised 11 genomes with 0 to 12 SNP differences, collected from nine patients hospitalized during 2015 (n = 1), 2016 (n = 5), and 2017 (n = 5) in the capital region of Denmark (n = 1) and the Zealand region (n = 10). For two patients, two isolates were included: patient P3 with 11 SNP differences between the isolates collected 13 months apart and patient P6 with three SNP differences between the isolates collected 5 months apart. The genomes in PO10-3 carried blaNDM-5 (11/11) and blaCTX-M-15 (8/11), and all genomes harbored IncF (pMLST F1:A1:B49) and IncX3 replicons. A history of travel to Egypt was registered for P1 from 2015, whereas travel history was unavailable for the remaining 10 patients. Patients P2 to P6 and P8 could be linked through overlapping stays within the same hospital units, whereas patient P7 had been hospitalized at the same hospitals units as P2 to P6 and P8 but 2 or more months apart from the other patients (Fig. S5). No link could be established between patient P1 and the remaining patients. For patient P9, no obvious link could be found to the remaining patients, but patient P9 was admitted to hospital H shortly after patient P3, though at different units.
PO10-4 consists of three isolates with 3 to 6 SNP differences, isolated in 2016 (n = 3) in the southern region of Denmark (n = 1) and the capital region of Denmark (n = 2). The genomes carried blaCTX-M-15 (2/3), and all harbored IncF (pMLST F1:A1:B49) and IncX replicons. The three patients did not travel prior to hospitalization. Patients P1 and P3 had been hospitalized 4 months apart in the same hospital unit, but no link could be found to patient P2 (Fig. S6).
PO10-5 consisted of three isolates with 2 to 3 SNP differences; they were isolated in 2015 (n = 1) and 2016 (n = 2), in the capital region of Denmark (n = 1) and the central region of Denmark (n = 2). The genomes contained blaCTX-M-15 (3/3), and all harbored IncF (pMLST F1:A1:B49) and IncX3 replicons. A history of travel to Lebanon was registered for P1 from 2015, while for the remaining two patients, one had no history of recent travel and for the other information on travel history was unavailable. Patients P1 and P3 were admitted at different hospitals, whereas patient P2 was not hospitalized within the investigated time period (Fig. S7).
The fourth possible clone (PC40-4) included four isolates with 15 to 38 SNP differences. As the pairwise comparison showed >10 SNPs for all combinations, no regional outbreaks were inferred for this clone. All four isolates carried an IncF replicon with various pMLST profiles (F36:A4:B1, 2/4; F31:A4:B1, 1/4; or F2:A6:B33, 1/4), and two of the isolates carried an IncX3 replicon. The isolates were from 2015 (n = 2), 2016 (n = 1), and 2017 (n = 1) from patients hospitalized in the capital region of Denmark (n = 4). Two of the patients had a history of travel to Pakistan, while travel history was unavailable for the remaining two patients.
The 127 ST410 international isolates spanned more than 4 decades (1975 to 2017) and were from 14 different countries: Denmark (n = 51); United Kingdom (n = 32); United States (n = 20); Germany (n = 10); Canada (n = 4); Brazil (n = 2); and Ireland, Japan, Nepal, Norway, Saudi Arabia, Singapore, Sweden, and Tanzania (one isolate each).
To investigate ST410 in a global context and its temporal dissemination and evolution, a Bayesian coalescent method was applied for phylogenetic reconstruction of the ST410 lineage based on the genomic data and sampling time. By using a general time-reversible (GTR) model and a strict molecular clock with a Bayesian Skyline population model, the BEAST analysis estimated the age of the ST410 lineage to be approximately 214 years (1803; 95% highest posterior density [HPD], 1774 to 1832) (Fig. 2). The phylogenetic reconstruction revealed two distinct clonal lineages of ST410: lineage A with fimH53 (A/H53) and lineage B with fimH24 (B/H24) (Fig. 3). The B/H24 lineage was further divided into three sublineages: B2/H24R with the introduction of fluoroquinolone resistance by mutations in gyrA and parC, B3/H24Rx with the introduction of blaCTX-M-15, and B4/H24RxC with the introduction of blaOXA-181.
FIG 2
Introduction of OXA-181 and NDM-5 into
FIG 3
Time-scaled phylogeny of
PC40-1, PC40-2, and PC40-4 were all part of the B3/H24Rx complex (Fig. 3). The Danish ST410 isolates in PC40-1 and PC40-2 from the national analysis did not cluster together with any non-Danish isolates (Fig. 3). The four Danish isolates from PC40-4 clustered together with one isolate from the United Kingdom, one isolate from Canada, and two isolates from the United States. For PC40-3, an additional 35 isolates clustered together with the 28 Danish isolates, resulting in a distinct clade with 63 isolates (B4/H24RxC in Fig. 3).
PO10-4 and PO10-5 clustered in separate clades without any additional isolates, whereas PO10-3 clustered closely with four isolates from the United Kingdom (7 to 21 SNPs to the nearest neighbor). These 63 isolates were collected between 2013 and 2017, in Denmark (n = 29), the United Kingdom (n = 29), Canada (n = 3), Norway (n = 1), and Sweden (n = 1). Additionally, 47 of the isolates in B4/H24RxC carried blaOXA-181, two isolates carried blaNDM-5, and 14 isolates carried both carbapenemases.
Introduction of blaOXA-181 and blaNDM-5 into ST410.
The timing of introductions of blaOXA-181 and blaNDM-5 into B4/H24RxC was predicted using BEAST (Fig. 3). The introduction of blaOXA-181 was estimated to have occurred approximately 14 years ago (2003; 95% HPD, 2000 to 2005), with the subsequent introduction of blaNDM-5 3.5 years ago (2014; 95% HPD, 2013 to 2014).
To investigate if blaOXA-181 was carried on an IncX3 plasmid, the genome scaffolds of the 63 ST410 genomes in B4/H24RxC with blaOXA-181 and/or IncX3 and an additional six genomes with blaOXA-181 and/or IncX3 were compared with the complete plasmid sequence of pAMA1167-OXA-181 from
To investigate if blaNDM-5 was introduced via an IncF plasmid with pMLST F1:A1:B49, the 58 ST410 draft genomes with blaNDM-5 and/or IncF with pMLST F1:A1:B49 were compared with the complete plasmid sequence of pAMA1167-NDM-5 (11) (Fig. S8B). All 58 isolates carried genomic information, which covered major parts of the pAMA1167-NDM-5 plasmid backbone, regardless of their blaNDM-5 gene status. However, among the 16 blaNDM-5-positive genomes, the IncF backbone revealed even higher similarity. The 11 Danish ST410 isolates from PO10-3, and four United Kingdom isolates clustering near the PO10-3 group, all carried similar IncF backbones (15 innermost rings in Fig. S8B [see also Fig. S9]). However, some of the plasmids lacked various resistance genes, including one United Kingdom isolate (ring 12 from the middle, H15074073205) missing the blaNDM-5 gene. Finally, three of the United Kingdom isolates carried blaNDM-4 (from the middle; ring 22, H14148045301; 24, H14142058201; and 25, H14140075501) on a plasmid very similar to the blaNDM-5 plasmid found in the PO10-3 isolates.
DISCUSSION
Recent studies indicate that
The possible clone PC40-3, including PO10-3, -4, and -5, consisted of 28 isolates with 0 to 41 SNP differences, and most of the isolates carried blaCTX-M-15, blaCMY-2, and blaOXA-181. Regarding PO10-4 and PO10-5, no epidemiological link could be made between the patients within the possible outbreaks. In PO10-4, the patients did not travel prior to hospitalization, and in PO10-5, one patient had recent travel activity to Lebanon, one patient did not report travel, and travel history was unavailable for the last patient. Thus, it is unclear if multiple introductions occurred, or if unknown carriers or contaminated environmental sources could link the spread between patients from these two PO10s.
Further analysis revealed another possible outbreak (PO10-3) involving 11 isolates from nine patients (P1 to P9), where one patient (P6) contributed with two samples isolated 5 months apart and one patient (P3) contributed with two samples 13 months apart. Isolates from PO10-3 had an additional carbapenemase gene, blaNDM-5. An epidemiological link was found among patients P2 to P8. The patient P9 could possibly be linked to this outbreak via hospital H. We were unable to link patient P1 to the other patients. However, it is not possible to exclude an unknown carrier of the clone, nor an unregistered travel link for patients P2 to P9, whereas patient P1, with the isolate from 2015 (AMA1167), had traveled to Egypt prior to isolation of AMA1167 (10). Also in 2015, an
From the global analysis, the close genetic relationship between four ST410 isolates isolated in the United Kingdom and the genomes from Danish ST410 isolates from PO10-3 could indicate a similar introduction to the United Kingdom; however, two of the patients from the United Kingdom reported no travel prior to hospitalization, and for the remaining two, travel information was unavailable. Similar genetic profiles have previously been reported for NDM-producing
The coalescence analysis (Fig. 2 and 3) on the global collection of 127 ST410 isolates indicated a most recent common ancestor approximately 214 years ago (the early 1800s), with confidence intervals stretching 30 years backward and forward. In comparison, the time to the most recent common ancestor for the ST131 clade was estimated as the late 1800s (22). Therefore, compared with ST131, ST410 is predicted to be a slightly older lineage of
Several studies have investigated the population structure of ST131 with WGS data and found a sublineage, C/H30, defined by the presence of the fimH30 allele. Within the C/H30 lineage, the C1/H30R clade, possessing fluoroquinolone resistance conferred by mutations in the chromosomal genes gyrA and parC, and the C2/H30Rx clade, which in addition is associated with carriage of blaCTX-M-15, emerged around 1987 (3, 4, 22, 23). We identified a similar population structure of ST410, with two sublineages based on the fimH alleles: A/H53 (3/127 genomes) and B/H24 (124/127 genomes). In the B/H24 sublineage, the introduction of the fluoroquinolone resistance by mutations in gyrA and parC (B2/H24R) and the introduction of blaCTX-M-15 (B3/H24Rx) were also estimated to have occurred around 1987 concurrently with the widespread clinical introduction of extended-spectrum cephalosporins and fluoroquinolones (Fig. 3). Additionally, the introduction of blaCTX-M-15 seems to be related to the acquisition and persistence of an IncFII plasmid harboring FIA/FIB replicons, which was also observed in ST131 (22). In ST410, the introduction of blaOXA-181 on an IncX3 plasmid was predicted to have occurred around 2003 (Fig. 3). Furthermore, introduction of blaNDM‑5, likely on an IncF replicon, around 3.5 years ago, resulted in a highly resistant clone causing an outbreak (PO10-3) in Denmark.
The majority of the 127 international ST410 isolates in our study were collected from humans (119/127); however, eight isolates were from nonhuman sources. The genomes obtained from isolates from swine (n = 2), poultry (n = 2), and air from poultry houses (n = 2) were all located in the nonresistant clades A/H53 and B1/H24 (Fig. 3), whereas the genomes obtained from sewage and from a dog were both located in the B3/H24Rx cluster. The latter cases likely represent the shared environment with humans. In our study, isolates found in the carbapenemase clade B4/H24RxC were all of human origin. However, carbapenemase genes are often located on mobile genetic elements (13, 14), together with other resistance genes encoding resistance to antibiotic classes commonly used in veterinary medicine (e.g., tetracycline, sulfonamides, or phenicols). Thus, once introduced in animals, carbapenemase genes may be coselected by the use of other classes of antibiotics, which could increase the prevalence of CPE in animals.
CPE is a worrisome risk in public health due to the multidrug resistance and the potential for dissemination to multiple sources, including transmission from patient to patient. The mobile colistin resistance gene mcr-1 has been reported in ESBL-producing
ST410
The possible outbreaks presented here and previous observations of interspecies transmission of ST410 (5) indicate a potential for effective transmission between patients. Additionally, among the Danish ST410 isolates, 20/49 (41%) were from bloodstream infections, which may qualify for severe infections, and show a potential for enhanced pathogenicity. ST410 belongs to phylogroup A, generally linked to commensal colonization (26), whereas ST131
These characteristics are properties of an international high-risk clone (1) and underline that ST410 qualifies as a lineage with new international multidrug-resistant high-risk clones, which should be monitored closely in the future.
MATERIALS AND METHODS
Bacterial isolates.
Initially, the strain collection at Statens Serum Institut (SSI) containing all ESBL/pAmpC-producing
Additionally, EnteroBase (https://enterobase.warwick.ac.uk/) was searched for
Whole-genome sequencing.
For isolates that were part of DANMAP surveillance, genomic DNA was extracted (DNeasy Blood and Tissue kit; Qiagen, Copenhagen, Denmark), with subsequent library construction (Nextera kit; Illumina, Little Chesterford, United Kingdom) and finally WGS (MiSeq or NextSeq; Illumina) according to the manufacturer’s instructions to obtain paired-end reads of 2 by 250 or 2 by 150 bp in length.
The raw reads of all 127 isolates were assembled into draft genomes using the SPAdes assembler v. 3.10.1 (28).
Genotypic characterization.
Draft genomes for the 127 ST410 isolates were submitted to the Center for Genomic Epidemiology (CGE) server through the batch uploader developed for the Bacterial Analysis Platform (BAP) (https://cge.cbs.dtu.dk/services/cge/) (29) and analyzed using the CGE tools described below (analysis performed between 14 August 2017 and 24 August 2017).
(i) Resistance genes and mutations in the QRDRs.
Resistance genes were identified with ResFinder 2.1 (30) (included in the CGE BAP), using a threshold of 100% identity (ID) for identifying genes encoding β-lactamases and carbapenemases and 98.00% ID for all other genes encoding transferable antimicrobial resistance. For phenotypically resistant isolates without β-lactamase- and carbapenemase-encoding genes or isolates with less than 100% identity, KmerResistance 1.0 (https://cge.cbs.dtu.dk/services/KmerFinder/) (31) was used to confirm the presence/absence of the gene(s).
To identify mutations in the QRDRs, a nucleotide BLAST database of the 127 ST410 genomes was set up using BLAST 2.6.0+ (32). Gene sequences of gyrA, gyrB, parC, and parE were extracted from all genomes by performing a BLASTN query of a representative nucleotide sequence for each of the four genes against this database. The nucleotide sequences were translated into corresponding amino acid sequences based on standard genetic code, using the reading frame beginning from the start codon. Amino acid changes were identified in the QRDRs (GyrA, amino acid position 67 to 106 [33]; GyrB, amino acid position 331 to 480 [34]; ParC, amino acid position 38 to 169 [35]; and ParE, amino acid position 365 to 525 [36]) for each gene by comparison with the wild-type gene variants of the quinolone- and fluoroquinolone-susceptible
(ii) Typing of isolates.
Multilocus sequence type (MLST) was extracted with MLST 1.6 (37), as part of the CGE BAP, to verify that the genomes belonged to ST410 in the Achtman scheme (MLST 1) (38).
fimH subtypes were identified with FimTyper 1.0 (https://cge.cbs.dtu.dk/services/FimTyper/) (39).
(iii) Plasmid analysis.
IncF and IncX3 plasmid replicons were identified using PlasmidFinder 1.4 (40) (included in CGE BAP) and further subtyped by plasmid MLST (pMLST) with the pMLST 1.4 tool (40).
The genome of the Danish ST410 isolate AMA1167 has previously been completely sequenced (GenBank accession no. GCA_002803905.1) (10), and the complete sequences of the OXA-181 and NDM-5 plasmids were used for comparative plasmid analysis. BLAST analysis of draft genome against pAMA1167-OXA-181 (GenBank accession no. CP024806.1) and pAMA1167-NDM-5 (GenBank accession no. CP024805.1) was performed using the GView server (https://server.gview.ca/).
Genomic epidemiology. (i) Single nucleotide polymorphism analysis.
The Northern Arizona SNP Pipeline (NASP) v1.0.0 was used to estimate the molecular epidemiology by identifying SNPs. Briefly, duplicate regions of the reference chromosome of ST410 isolate YD786 (GenBank accession no. NZ_CP013112.1) were identified by aligning the reference against itself with NUCmer v3.1 and masked from downstream analysis. Raw reads of the analyzed genomes were aligned against the FASTA-formatted reference using the Burrows-Wheeler Aligner (BWA), and SNPs were identified using GATK UnifiedGenotyper (41).
Gubbins v2.3.1 was used to remove recombination signals in the SNP output (42). The SNP phylogenies were annotated with relevant metadata using iTOL (http://itol.embl.de) (43).
(ii) Identification of possible clones, outbreaks, and patient links.
Possible clones (PC) in the Danish collection were predicted based on the distribution of the pairwise SNP differences between the isolates. Possible clonal outbreaks were identified for clusters containing at least three individual isolates, by adhering to an SNP distance of ≤10 SNPs to the nearest neighboring isolate (PO10), as previously described (9).
Patient links in the different possible outbreaks (PO10s) were investigated 6 months prior to the first identified isolate in the respective PO10 and 6 months after the last identified isolate. Hospital admissions (hospital, unit, and time of admission) were used to construct the timeframe and possible routes of transmission for the PO10s.
(iii) Coalescence-based analyses.
BEAST analysis was conducted to reconstruct the global phylogeny of ST410 based on the SNP data generated from the 127 raw sequence data sets, for estimation of the time points wherein blaOXA-181 and blaNDM-5 were acquired by the ST410 lineage. For defining the two groups for blaOXA-181 and blaNDM-5, a maximum likelihood phylogeny was estimated with PhyML 3.0 by using the Bayesian information criterion with BIONJ as initial tree, nearest neighbor interchange (NNI) as type of tree improvement, and a bootstrap value of 100 (44). Based on the phylogeny, the two groups were defined.
The timing of the acquisitions of blaOXA-181 and blaNDM-5 was inferred using coalescence-based analyses in BEAST 1.8.4 at CIPRES Science Gateway (https://www.phylo.org) using duplicate runs. Analyses were performed using different substitution models (general time-reversible [GTR] and Hasegawa-Kishino-Yano [HKY]), strict and relaxed molecular clock, and different demographic models including Bayesian Skyline, constant population, and exponentially growing population tree priors. Chains were sampled every 40,000th step during the 400 million Markov chain Monte Carlo (MCMC) steps. The first 10% of each chain was discarded as burn-in. The best model was selected based on Bayes factor (BF) comparison (45). Data were analyzed with Tracer v1.5 and TreeAnnotator v1.8.4, and the trees were depicted by using iTOL (http://itol.embl.de/) (43).
b Department of Clinical Microbiology, Hvidovre University Hospital, Hvidovre, Denmark
c Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark
d Department of Clinical Microbiology, Aarhus University Hospital, Aarhus, Denmark
e Department of Clinical Microbiology, Slagelse Hospital, Slagelse, Denmark
f Department of Clinical Microbiology, Herlev and Gentofte Hospital, Herlev, Denmark
g Department of Clinical Microbiology, Odense University Hospital, Odense, Denmark
h Department of Clinical Microbiology, Rigshospitalet, Copenhagen, Denmark
i Department of Clinical Microbiology, Aalborg University Hospital, Aalborg, Denmark
j Antimicrobial Resistance and Healthcare Associated Infections (AMRHAI) Reference Unit, National Infection Service, Public Health England, London, United Kingdom
k Institute of Medical Microbiology, Justus Liebig University Giessen and German Centre for Infection Research (DZIF), Partner Site Giessen-Marburg-Langen, Giessen, Germany
l Norwegian National Advisory Unit on Detection of Antimicrobial Resistance, Department of Microbiology and Infection Control, University Hospital of North Norway, Tromsø, Norway
m Department of Pharmacy, Faculty of Health Sciences, UiT-The Arctic University of Norway, Tromsø, Norway
n Public Health Agency of Sweden, Stockholm, Sweden
o Department of Infectious Disease Epidemiology and Prevention, Statens Serum Institut, Copenhagen, Denmark
JMI Laboratories
Copyright © 2018 Roer et al. This work is published under https://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.