Genomic and Functional Analysis of Emerging Virulent and Multidrug-Resistant Escherichia coli Lineage Sequence Type 648
ABSTRACT
The pathogenic extended-spectrum-beta-lactamase (ESBL)-producing Escherichia coli lineage ST648 is increasingly reported from multiple origins. Our study of a large and global ST648 collection from various hosts (87 whole-genome sequences) combining core and accessory genomics with functional analyses and in vivo experiments suggests that ST648 is a nascent and generalist lineage, lacking clear phylogeographic and host association signals. By including large numbers of ST131 (n = 107) and ST10 (n = 96) strains for comparative genomics and phenotypic analysis, we demonstrate that the combination of multidrug resistance and high-level virulence are the hallmarks of ST648, similar to international high-risk clonal lineage ST131. Specifically, our in silico, in vitro, and in vivo results demonstrate that ST648 is well equipped with biofilm-associated features, while ST131 shows sophisticated signatures indicative of adaption to urinary tract infection, potentially conveying individual ecological niche adaptation. In addition, we used a recently developed NFDS (negative frequency-dependent selection) population model suggesting that ST648 will increase significantly in frequency as a cause of bacteremia within the next few years. Also, ESBL plasmids impacting biofilm formation aided in shaping and maintaining ST648 strains to successfully emerge worldwide across different ecologies. Our study contributes to understanding what factors drive the evolution and spread of emerging international high-risk clonal lineages.
INTRODUCTION
The pandemic extraintestinal pathogenic Escherichia coli (ExPEC) ST131 clonal lineage and in particular its clade C, or H30R, has been the focus of numerous seminal studies (1). The continuous acquisition of antimicrobial resistance (AMR) and virulence factors, which for example convey the ability to colonize the host and compete against other pathogens, have enabled its spread (2, 3). While research has predominantly focused on ST131, fewer studies are available for emerging pandemic ST648, a growing resistance-associated lineage observed from multiple origins (4–6). AMR identified in ST648 includes resistance to fluoroquinolones, third-generation cephalosporins, and other antibiotics (7). Mobile genetic elements have contributed to its recent spread, which is probably due not only to the carriage of AMR genes but also to other fitness factors: extended-spectrum beta-lactamase (ESBL)-carrying plasmids have been recently associated with an enhanced virulence potential in E. coli ST131 and ST648 strains (8, 9). We acquired a large ST648 strain collection and a set of ST131 (phylogroup B2, mostly pathogenic) and ST10 (phylogroup A, mostly commensal) strains for comparative purposes to help elucidate the success of ST648 (phylogroup F). ST10 is an ancestral, ubiquitously occurring lineage that comprises both commensal and pathogenic strains. By combining genomics with a functional approach, we gained valuable insights into these dominating E. coli lineages and met our main goal to assess and characterize similarities and differences in antimicrobial resistance, virulence, and in vivo colonization. Our results elucidate the features that contribute to the emergence of successful clonal lineages across multiple hosts and settings.
RESULTS AND DISCUSSION
A total of 87 E. coli ST648 genomes were analyzed (Fig. 1). The pangenome matrix revealed 12,276 coding sequences, with 3,596 genes shared by all 87 genomes that represent the maximum common genome (MCG). Mapping all sequencing reads against the completely closed chromosomal genome of IMT16316, which was also constructed in this study, resulted in an alignment of 3,883 Mbp. This alignment contained 3,834 informative single-nucleotide polymorphism (SNP) sites, which were used to estimate a maximum-likelihood phylogeny (Fig. 1A). The resulting tree revealed a ST648 lineage with low genetic diversity, divided into only four clades. Clades 1 and 2 differed by approximately 200 SNPs, clades 1 and 4 differed by approximately by 550 SNPs, and clade 3 was the most distant (an ∼600-SNP difference compared to clades 1 and 2). These results were confirmed by a Bayesian clustering using a Bayesian analysis of the population structure (BAPS) (Fig. 1) and are in contrast to lineage ST131, for which studies have identified approximately 7,000 to 9,000 SNPs between the three main clades (A, B, and C, with the latter comprising subclades C1 and C2, corresponding to H30R and H30Rx) (1, 10). It should be noted that a selection bias in the strain set may have influenced our results; however, we estimate the odds to be rather low.
FIG 1
Notably, ST648 clade 2 strains almost ubiquitously carried an identical ESBL plasmid, as discussed below. Strains from humans and animals with different backgrounds were distributed throughout the phylogeny (Fig. 1A), suggesting ST648’s interspecies movement and zoonotic potential while lacking clear genomic signatures that would indicate ecological adaptation to individual host species and infection sites. The geographical distributions were also similar, where strains from different continents occurred throughout the phylogeny (Fig. 1A). These results suggest that ST648 is a host generalist rather than a specialist, with the capability of frequent cross-species transmission, thriving in different clinical and nonclinical contexts, similar to ST131 (8). By including strains from multiple geographic regions and wildlife (Fig. 1A), unique insights into this newly emerging clonal lineage were discovered beyond the usually narrower human, clinically centered focus.
We analyzed the accessory gene profiles of the 87 ST648 genomes based on a total size of a variable pangenome of 8,680 genes, which provides information about past evolutionary events in this population, in addition to the simpler focus on plasmid-related gene contents (Fig. 1A). The accessory genome did not cluster with any assigned metadata groups (i.e., country, host, and origin) and only to a limited extent with the core genome-based clades, where clade 2 showed a slightly different accessory gene profile (Fig. 1A). These results are in contrast to those observed with ST131, where McNally et al. reported several subtypes within ST131 clade C based on highly congruent accessory gene profiles (8).
By estimating the substitution rates and creating a time-scaled tree with estimated dates of the internal nodes, we were able to date important events within the microevolution of this lineage (Fig. 1B). The first split of the lineage dates back to 2001, and the more distantly related clades 3 and 4 separated from the branch-containing clades 1 and 2 in 2004. The most recent event in forming the ST648 population structure was the separation between clusters 1 and 2 in 2006. Along with the observation that the accessory genome did not have sufficient time to diversify along the shallow core lineages, these results suggest a young, nascent ST648 lineage. For ST131, Stoesser et al. estimated the time to the most recent common ancestor to be significantly longer, about 130 years ago, when clade A diverged from clades B and C, followed by a split of clades B and C 25 years ago (1).
To investigate whether ST648 and ST131 strains were similarly virulent while being more virulent than ancestral ST10, we performed a genomic analysis of all ST648, ST131, and ST10 genomes for virulence-associated genes (VAGs) typically associated with pathogenic E. coli. Colonization, iron uptake, and biofilm formation are key enabling factors for the clinical success of ST131 (11). Although biofilm formation contributes to bacterial pathogenesis through features such as immune evasion and adhesion, bacterial iron acquisition is especially important during infection as host-associated proteins sequester and thus deplete metals at the infection site (12). Based on a manually curated database with 75 VAGs highly relevant in pathogenic E. coli (13–15), the results revealed the highest overall mean for ST131 (36.2) per strain, closely followed by ST648 (31.1, P = 0.681). In contrast, ST10 had significantly fewer VAGs (19.6, P = 0.004) (see Fig. S1A in the supplemental material).
While biofilm/adherence- and iron uptake-associated genes were differently distributed, we observed similarities in factors related to invasion, toxins, and capsule-associated virulence. ST131 strains carried 21.7 biofilm/adherence-associated genes on average, and ST648 strains harbored an equal number (mean, 20.4; P = 0.662), while the number was again significantly lower in ST10 strains (mean, 13.0; P = 0.046). Similar ordering was observed in iron uptake-related genes (ST131 [mean, 8.6], ST648 [mean, 5.6; P = 0.159], and ST10 [mean, 2.2; P = 0.007]), which aligns well with earlier findings showing that ST131 is an important cause of urinary tract infections (UTIs) and follows in the context that strains belonging to phylogroup B2 often carry virulence factors, particularly those associated with extraintestinal infection, in contrast to phylogroup A. Iron acquisition is a particularly crucial step for urine pathogen survival in this largely metal-depleted ecological niche (16).
To further elucidate the importance of observed differences, we investigated whether ST648, ST131, and ST10 carried clusters of particular biofilm/adherence- and iron uptake-related gene sets indicative of host niche adaptation (Fig. S2). This was true for adherence-associated genes (the pap operon, afaA, iha, nfaE, and focX) in ST131. We found two large clusters (nodes 4 and 5) pointing toward their sophisticated adaptation to ecological niches where these VAGs are necessary, such as colonizing the urinary tract (Fig. S2B) (17, 18). Actually, most node 4 and 5 ST131 strains were isolated from UTIs (87.5%). ST648 patterns were not as clearly associated with UTI-relevant adhesion factors (i.e., the lack of pap genes in nodes 1 and 2), providing no evidence for UTI-related niche adaption in ST648 strains. This observation coincides with the finding of Vangchhia et al. that phylogroup F strains seemingly tend to lack virulence factors typically associated with extraintestinal infections (19).
In contrast to the separation of adhesion-related gene clusters according to their ST (Fig. S2B), we observed mixed clusters for iron uptake-related genes with the same set of genes present in strains from different STs (Fig. S2A). The largest iron uptake cluster (node 1) consisted of both ST131 and ST648 strains (Fig. S2A, node 1: chuA, fyuA, irp2, iucD, iutA, sitA, sitB, sitC, and sitD) but no ST10 strains. Iron uptake is also important during infection outside the urinary tract. ChuA, for instance, enables bacteria to use hemoglobin as an iron source. The presence of a set of virulence genes such as chuA in both ST131 and ST648 underlines their virulence potential in different clinical contexts. Overall, the VAG analysis of ST648, ST131, and ST10 provides evidence that the differences in the overall numbers of VAGs are caused by certain clusters of adhesion and iron uptake-related genes indicative of pathogen niche adaption, especially in ST131 strains.
To functionally examine our genomic results on a phenotypic level, we performed siderophore secretion/iron uptake experiments for 60 randomly selected ST648, ST131, and ST10 strains. ST131 was superior (85%) to ST648 (25%) and ST10 (20%) in iron acquisition, which partly accords with the above-mentioned in silico results, while following in the context of its history as a major cause of UTIs (Fig. S1). However, since a basic set of iron uptake genes was shared by ST648 and ST131 (Fig. S2A, node 1: chuA, fyuA, irp2, iucD, and sitA, sitB, sitC, and sitD), the superiority of ST131 strains in iron uptake is a noteworthy finding.
As biofilm formation was also suggested by the in silico results to be of particular interest, we performed phenotypic macrocolony experiments for all available ST648, ST131, and ST10 strains to gain a general overview of curli fibers and cellulose production, which are major biofilm components (20) (Fig. 1A and Fig. 2). Curli fibers are considered virulence factors, since they contribute to the adherence of E. coli to epithelial cells (21). Bacteria mostly produce cellulose for protective reasons, and cellulose has been described in a survival context on abiotic surfaces (22), demonstrating their potential to persist in both host and environment when expressing the relevant genes.
FIG 2
We observed significant differences in biofilm formation between the different STs (Fig. 2B). The majority of ST648 (76%) produced curli fibers or cellulose, in contrast to ST10 (39%, P < 0.001) and, interestingly, also ST131 (66%, P = 0.027) (Fig. 2B). As for both curli and cellulose production, ST648 again outcompeted (48%) ST131 (19.6% [P < 0.001]) and ST10 (18.75% [P < 0.001]). Although lacking typical UTI-associated adherence genes such as the pap gene group, ST648 frequently produced major biofilm components. Furthermore, we observed a negative correlation between the ability to produce these components and the bacteria’s motile behavior, which confirms previous observations (9). Most nonmotile strains showed curli and/or cellulose production, whereas the lack of such production was correlated with motility (Fig. 1A). Particularly striking was clade 2, which most pronouncedly consisted of strains producing curli and/or cellulose while demonstrating reduced motility (Fig. 1A).
Based on a comparative genomics analysis using a PacBio-sequenced reference (IMT16316), most of these strains carried an identical IncF-type ESBL plasmid, except for IMT24495 and IMT33149 (Fig. 2A). These two strains did not produce biofilm-related factors. Since we have previously demonstrated that ESBL plasmids seemingly influenced biofilm formation in ST648 and ST131 E. coli (8), we constructed ESBL plasmid-cured variants (PCVs), complemented with the above-mentioned clade 2 plasmid (pEcIMT16316 and IncF[F1:A1:B49]), and repeated the biofilm and motility experiments. With this complemented ESBL plasmid, both strains produced curli fibers and cellulose, while losing their motile behavior. Other closely related ESBL plasmids of the same IncF subtype carried by strains outside cluster 2 showed similar phenotypes, supporting our previous hypothesis that ESBL plasmids confer advantageous traits in certain E. coli strains. Although the exact contribution of VAGs to colonization and fitness remains to be investigated in-depth for most E. coli lineages, they have been a part of enabling the success of ST131, and it appears plausible that the same applies to ST648.
Several publications have reported the frequent occurrence of ESBL-producing ST131 and ST648 strains in wild-bird populations and poultry, possibly pointing toward an increased colonization ability of ESBL E. coli in the avian gut (23–25). To test whether ST131 and ST648 would differ in this respect in vivo, we performed colonization experiments for ST648 and ST131 in chicken intestines. Four randomly selected ST648 strains of clade 2 (IMT16316, IMT17887, IMT21183, and IMT23463) were compared to four ST131 strains (IMT17433, IMT19205, IMT19224, and IMT27685). Thirty days after oral bacterial inoculation (104 CFU per animal), the cecum and jejunum contents of eight animals per strain were characterized and quantified. We did not detect significant differences (cecum, P = 0.429; jejunum, P = 0.219) in the overall numbers of ST648 compared to ST131 strains, which ranged from 106 to 108 CFU/g (cecum) and 102 to 105 CFU/g (jejunum) (Fig. 3). This underlines the similarly high potentials of both lineages to colonize birds in vivo.
FIG 3
Following colonization, host invasion presents the next crucial step for successful pathogenesis. It has been recently suggested that the abundance of certain E. coli lineages in invasive infections such as bacteremia is driven by negative frequency-dependent selection (NFDS). So far, this has resulted in a relatively steady population structure of different phylogenetic lineages only disrupted by the emergence of ST131, as demonstrated in a large-scale bacteremia study performed in the United Kingdom. To also predict the invasive success of ST648, we applied the NFDS model to the above-mentioned UK BSAC bacteremia sample set (26). Our results show that ST648 will expectedly increase to a stable frequency, with a maximum near 1.5% among all populations (Fig. S3A). Comparative simulations with a neutral model do not predict the reproductive success of ST648 (Fig. S3B), while the alternative NFDS model, which lacks migration, results in most simulations losing the low-frequency strain. Notably, however, if ST648 survives long enough, it will reach the same prevalence as in other NFDS simulations (Fig. S3C). This indicates that the predicted future prevalence is maintained by the NFDS model, in contrast to having arisen as an artifact of migration. ST648 is predicted to increase 10-fold in frequency as a cause of bacteremia over the next few years but to remain at much lower levels than ST131.
Since the combination of virulence with multidrug resistance (MDR) has been considered crucial to the global emergence of ST131 in clinical settings (2, 27, 28), we sought finally to investigate antimicrobial resistance properties of our strains. A total of 94% of ST648 strains were MDR, followed by ST131 (72%) and ST10 (50%) resulting from an in silico resistance screening (see Table S2 in the supplemental material). MDR was significantly increased in the ST648 group compared to ST10 (P < 0.001) and also ST131 (P < 0.001). Genes imparting resistance to non-beta-lactams included aminoglycosides [aac(3)-II a/d, aadA5, aph(3′)-Ia, aac(6′)Ib-cr, and strA/B], sulfonamides (sul1, sul2), trimethoprim (dfrA14, dfrA17), tetracycline (tetA and tetB), and fenicoles (catB3). The ESBL-type blaCTX-M-15 was the dominant gene in ST648 and ST131 strains. The frequent presence of MDR in ST648 strains makes them likely clinically competitive compared to ST131.
ESBL-producing E. coli are leading MDR pathogens, headed by a few internationally relevant, high-risk clonal lineages with ST131 as the most prominent. ST648 presents an emerging lineage increasingly reported from multiple origins with the greatest potential to follow ST131’s success. We applied comparative genomics and functional experiments to a large collection of ESBL-producing E. coli strains to unravel ST648’s population structure and to elucidate the mechanisms that contributed to its emergence across different ecologies. Our work highlights its nascent, generalist character, while combining MDR and virulence on a level similar to ST131. Both have seemingly developed individually sophisticated features possibly conveying advantages in different ecologies: ST648 strains significantly more often produced biofilm-related curli fibers and cellulose and had similar numbers of biofilm/adherence virulence-associated genes, while ST131 strains efficiently acquired iron. The supposed “superiority” of ST131 in some virulence characteristics (e.g., the propensity to cause UTIs) might be outbalanced by ST648’s biofilm-forming properties and increased antimicrobial resistance.
Nevertheless, results stemming from in silico and in vitro experiments alone cannot serve to differentiate and fully explain the success of distinct E. coli lineages. The ability of ST131 and ST648 to effectively colonize in vivo, which has been previously reported for ST131 (3) and was confirmed in the present study, is another major contributor to virulence. In addition, ESBL plasmids conferring beneficial traits for colonization and environmental survival in different ecologies likely aided in shaping and maintaining ST648 strains to being similarly successful. However, in addition to possessing an advantageous geno- and phenotypic makeup, emerging novel lineages have to be common enough to be able to spread in clinical and extraclinical environments. Our ST648 NFDS bacteremia modeling clearly underlines the fact that ST648 will increase in frequency. Although the prediction points toward lower levels than ST131, ST648 is well on its way to becoming another internationally circulating, high-risk clonal lineage, worsening infection treatment possibilities not only due to antimicrobial resistance but also due to other features.
These findings provide strong incentive for further in-depth studies of ST648’s properties and careful surveillance of its emergence across different ecologies.
MATERIALS AND METHODS
Origin of strains.
We used a total of 290 E. coli strains of ST648 (n = 87), ST131 (n = 107), and ST10 (n = 96), which were isolated between 2006 and 2014 (Table S1). The strains originated from a variety of different hosts (humans [n = 95], companion animals [n = 113], livestock [n = 56], and wildlife, mainly wild birds [n = 26]). They were either isolated from healthy (mainly feces or food samples [n = 108]) or diseased hosts (mostly sampled from urinary tract, wound, and respiratory tract infections [n = 182]). We obtained the strains from all over the world from 11 countries or regions (Europe [n = 222], Canada [n = 27], Australia [n = 8], New Zealand [n = 7], the United States [n = 9], South Africa [n = 5], Brazil [n = 5], Tunisia [n = 3], China [n = 1], Mongolia [n = 2], and India [n = 1]). All strains were initially selected based on multilocus sequence typing (MLST) as ST648, ST131, or ST10. ESBL production was confirmed according to a standard CLSI method (29).
Whole-genome data.
DNA was extracted at the Institute of Microbiology and Epizootics at the Freie Universität Berlin and sequenced at the Wellcome Trust Sanger Institute on an Illumina HiSeq2000, with 96 samples per flow cell using 100-bp paired-end reads and a minimum of 90× coverage. Sequencing data were submitted to the ENA or NCBI SRA (genome accession numbers are provided in Table S1). We assembled sequence data de novo using Velvet (30), iCORN (31), and the Velvet Columbus module, following automated sequence quality control by the Sanger systems. All draft genomes were annotated using Prokka (32). In addition, we sequenced strain IMT16316 on an RS sequencer with SMRT technology PacBio RSII (Pacific Biosciences, Menlo Park, CA) at GATC Biotech (Constance, Germany), using standard protocols according to the manufacturer’s instructions resulting in 106,768 reads with a mean length of 12,147 bp. De novo assembly was performed with the PacBio De Novo Assembly Pipeline, which is integrated in the CLC Genomics Workbench (https://www.qiagenbioinformatics.com). The assembly resulted in a fully closed IMT16316 chromosome and plasmid (pEcIMT16316), which were used as references. Both sequences were submitted to GenBank (accession numbers CP023815 for IMT16316 chromosome and CP023816 for the pEcIMT16316 plasmid). Sequence data that support the findings of this study have been deposited in ENA or NCBI SRA (genome accession numbers are provided in Table S1).
Whole-genome alignments.
Raw reads were first trimmed using Trimmomatic (v0.36; default parameters) (33). Trimmed reads were then aligned to the reference sequence (IMT16316 and pIMT16316) using BWASW (0.7.15-r1140; default parameters). Bam files were converted using samtools mpileup (34). Afterward, variants were called using VarScan (v2.3.8) (35) with the following parameters: min-coverage, 10; min-reads, 8; min-avg-qual, 20; min-var-freq, 0.8; min_freq-for-hom, 0.8; P = 0.01; and strand-filter disabled. To build the consensus sequences, insertions were excluded, and calls of reference bases were only considered if supported by at least 80% of the reads. Based on SNP density plots over the length of the alignment, regions with a significantly increased SNP density compared to the rest of the genome were filtered.
Population structure.
The resulting alignment was used to infer a maximum likelihood phylogeny using RAxML version 8.1.14 (36) with a general-time-reversible model and gamma correction for “among site rate variation” and as input for BAPS (37) to identify phylogenetic clusters. Hierarchical clustering was performed for three levels using five replicate runs of the inference algorithm with the upper bound for the number of clusters ranging between 5 and 10. All runs converged to the same estimate of the population structure. Trees and corresponding metadata information were visualized using Phandango (38).
Estimation of substitution rates and dates.
Based on the ST648 phylogenetic tree and sample date information, we calculated the substitution rates, resulting in a time-corrected and rooted phylogenetic tree with dated inner nodes using least-squares dating (39). Temporal constraints were included, and variances were used to reduce the weights of estimated errors in the input branch lengths.
Accessory genome.
The Prokka-annotated coding sequences of each genome were used as input for the Roary pipeline (rapid large-scale prokaryote pangenome analysis) (40). Clustering into orthologues was performed for genes sharing at least 95% nucleotide identity. Accessory genome profiles were visualized with Phandango in combination with the corresponding core genome phylogeny and strain metadata (38).
Virulence, antimicrobial resistance, and plasmid genes.
We compiled a set of most important VAGs (n = 75) according to current ExPEC literature (13, 14) and excluding allelic genes from the same operon (with the exception of pap variants [17]). This collection included genes related to adherence/biofilm formation (42), iron uptake (10), toxicity (13), invasion (5), and capsule protection (5). Wholegenome sequences were then submitted to ResFinder 2.1 (41) and PlasmidFinder 1.3 (42) of the Center for Genomic Epidemiology (http://www.genomicepidemiology.org). Resistances to antimicrobial classes beyond beta-lactams were analyzed by grouping resistance genes according to their expected phenotype using ResFinder 2.1. To consider a strain resistant, a minimum of one resistance gene per antimicrobial class had to be present. For plasmid comparison, raw reads of clade 2 plasmid sequences were mapped against the complete reference sequence of pECIMT16316 (GenBank accession number CP023816). The resulting consensus sequences were individually compared for similarity by using NCBI BLAST+ v2.2.29 (43), and the combined results were visualized using BRIG v0.95. (44)
Clustering of adherence and iron acquisition genotypes.
Based on the binary matrix containing the information about the existence or absence of a particular gene in the genome, a minimum spanning tree network was calculated and visualized using the Bionumerics software package v7.5 (Applied Maths, Belgium).
Macrocolony formation.
We tested 79 ST648 strains (8 were excluded due to absence of actual strains), 107 (all) ST131 strains, and 96 (all) ST10 strains on macrocolony plates. Portions (3 μl) of overnight cultures grown at 37°C were dropped onto span agar plates (H. Carroux, Germany) with sodium chloride (5%) and a 0.5% Congo red–0.25% Coomassie brilliant blue solution. Plates were incubated for 5 days at 28°C (45) and screened for curli fibers and cellulose production. Experiments were repeated in three biological replicates. Strains AAEC189 (46), IMT26949, and W3110 (47) were included as references. Statistics were performed using chi-square tests (IBM SPSS Statistics for Windows). A purple macrocolony color indicates the production of curli fibers; elevated, structured surfaces indicate the production of cellulose. Plain white colonies indicate the absence of both.
Motility.
The same set of strains in triplicates was tested on motility plates. Overnight cultures grown at 37°C were set to an optical density at 600 nm of 1. A 1-ml portion was centrifuged, and the pellet washed with 1× phosphate-buffered saline (PBS). Five microliters were dropped on Luria-Bertani plates with 0.3% agar. Plates were incubated at 37°C. Strain MG1655 (47) was used as a positive control. After 48 h, the colony diameters were measured. Purple macrocolonies indicate the production of curli fibers; elevated, structured surfaces indicate the production of cellulose. Plain white colonies indicate the absence of both. Statistics were performed using chi-square tests (IBM SPSS Statistics for Windows).
ESBL plasmid curing and transformation.
ESBL plasmids of IMT24494 and IMT33149 were cured by using a heat technique (8). After the verification of the ESBL plasmid-free status via plasmid profile analysis and phenotypic resistance screening, the BAPS cluster 2 ESBL plasmid was transformed by electroporation.
Iron acquisition.
Sixty randomly selected ST648, ST131, and ST10 strains were screened for their ability to secrete siderophores using a method described by Schwyn and Neilands (48). These strains were plated on agar plates containing chrome azurol S-iron(lII)-hexadecyltrimethylammonium bromide. Iron uptake was determined visually by color shift from blue to orange. One ST73 strain (IMT18341) and one K-12 strain (IMT 20859), respectively, were included as references. This experiment was performed in three biological replicates.
In vivo chicken colonization experiments.
The study plan describing animal trials (G 0235/14) was approved by the State Office for Health and Social Affairs, Berlin, Germany, in accordance with the Act on Experimental Animals. Eighteen-day-old embryonated specific-pathogen-free eggs (Valo Biomedia, Osterholz-Scharmbeck, Germany) were hatched at the experimental facility. After hatching, the chicks were tested for the absence of ESBL-producing Enterobacteriaceae. Cloacal swabs were taken and streaked onto chromogenic agar plates (Brilliance UTI agar; Oxoid, Ltd., Wesel, Germany) containing 2 mg/liter cefotaxime (Sigma-Aldrich, Germany). Animal trials were performed in controlled pens inside the experimental facilities, and each group of eight chicks was housed in a separate pen. On day 1 after hatching, each check of the nine groups of eight chicks was orally inoculated into the crop with 0.5 ml of PBS containing 108 CFU/ml of the different ESBL-producing inoculum strains. At 30 days after challenge, the animals were euthanized, and necropsy was performed to aseptically remove cecal and jejunal contents. To quantify the number of CFU of ESBL-producing E. coli in the organ contents, a 10-fold dilution series (10−1 to 10−5) was prepared in PBS solution. Then, 10 μl of each dilution was inoculated on chromogenic agar plates with cefotaxime (2 mg/liter), followed by incubation overnight at 37°C. By using a semiquantitative method and based on the highest dilution at which ESBL E. coli growth was observed, the number of CFU/g of feces was calculated.
NFDS model-based prediction of ST648 future trajectory in bacteremia.
To predict the frequency of ST648 in bacteremia, we used the recently developed multilocus negative frequency-dependent selection (NFDS) model (26) for the UK BSAC longitudinal survey data on bacteremia isolates of E. coli (3). It was used to predict the ST648 frequency for a decade forward from the end of the longitudinal survey in 2012 by forward-simulating 100 independent realizations of the frequency trajectories of the clones present in the data. For comparison, predictive simulations were also performed using a neutral model and a model lacking migration.
ACKNOWLEDGMENTS
This study was supported by grants from the German Research Foundation (DFG) to S.G. and C.E. entitled (Functional Analysis of Nonresistance Genes of Extended-Spectrum Beta-Lactamases-Associated Sequence Types of Escherichia coli [grants GU 1283/3-1 and EW116/2-1]). K.S. was supported through a Feodor Lynen Research Fellowship awarded by the Alexander von Humboldt Foundation, Germany. Whole-genome sequencing was partly financed by the Wellcome Trust Sanger Institute (Cambridge, United Kingdom; study 2433ILB). The foundations funding this work had no role in study design, data collection and interpretation, or the decision to submit for publication. In addition, this work was partially supported by project CEITEC 2020 (LQ1601) from the Ministry of Education, Youth, and Sports of the Czech Republic to I.L.
K.S., T.S., L.H.W., J.C., and S.G. designed and drafted the manuscript. T.S., K.S., S.G., A.M., and C.E. performed the experiments. D.J.T., A.M., J.P., G.P., J.B., M.D., I.L., S.F., N.A., M.G., C.T., and D.P. helped in analysis and discussion of the results. C.E., A.M., D.J.T., L.H.W., J.C., and N.A. helped in proofreading and editing of the manuscript. All authors read and approved the final manuscript.
Supplemental Material
REFERENCES
1.
Stoesser N, Sheppard AE, Pankhurst L, De Maio N, Moore CE, Sebra R, Turner P, Anson LW, Kasarskis A, Batty EM, Kos V, Wilson DJ, Phetsouvanh R, Wyllie D, Sokurenko E, Manges AR, Johnson TJ, Price LB, Peto TE, Johnson JR, Didelot X, Walker AS, Crook DW, Modernizing Medical Microbiology Informatics Group. 2016. Evolutionary history of the global emergence of the Escherichia coli epidemic clone ST131. mBio 7:e02162.
2.
Ben Zakour NL, Alsheikh-Hussain AS, Ashcroft MM, Khanh Nhu NT, Roberts LW, Stanton-Cook M, Schembri MA, Beatson SA. 2016. Sequential acquisition of virulence and fluoroquinolone resistance has shaped the evolution of Escherichia coli ST131. mBio 7:e00347-16.
3.
Kallonen T, Brodrick HJ, Harris SR, Corander J, Brown NM, Martin V, Peacock SJ, Parkhill J. 2017. Systematic longitudinal survey of invasive Escherichia coli in England demonstrates a stable population structure only transiently disturbed by the emergence of ST131. Genome Res.
4.
Ewers C, Bethe A, Stamm I, Grobbel M, Kopp PA, Guerra B, Stubbe M, Doi Y, Zong Z, Kola A, Schaufler K, Semmler T, Fruth A, Wieler LH, Guenther S. 2014. CTX-M-15-D-ST648 Escherichia coli from companion animals and horses: another pandemic clone combining multiresistance and extraintestinal virulence? J Antimicrob Chemother 69:1224–1230.
5.
Peirano G, van der Bij AK, Gregson DB, Pitout JD. 2012. Molecular epidemiology over an 11-year period (2000 to 2010) of extended-spectrum beta-lactamase-producing Escherichia coli causing bacteremia in a centralized Canadian region. J Clin Microbiol 50:294–299.
6.
Zong Z, Yu R. 2010. Escherichia coli carrying the blaCTX-M-15 gene of ST648. J Med Microbiol 59:1536–1537.
7.
Johnson JR, Johnston BD, Gordon DM. 2017. Rapid and specific detection of the Escherichia coli sequence type 648 complex within phylogroup F. J Clin Microbiol 55:1116–1121.
8.
McNally A, Oren Y, Kelly D, Pascoe B, Dunn S, Sreecharan T, Vehkala M, Valimaki N, Prentice MB, Ashour A, Avram O, Pupko T, Dobrindt U, Literak I, Guenther S, Schaufler K, Wieler LH, Zhiyong Z, Sheppard SK, McInerney JO, Corander J. 2016. Combined analysis of variation in core, accessory and regulatory genome regions provides a super-resolution view into the evolution of bacterial populations. PLoS Genet 12:e1006280.
9.
Schaufler K, Semmler T, Pickard DJ, de Toro M, de la Cruz F, Wieler LH, Ewers C, Guenther S. 2016. Carriage of extended-spectrum beta-lactamase-plasmids does not reduce fitness but enhances virulence in some strains of pandemic Escherichia coli lineages. Front Microbiol 7:336.
10.
Petty NK, Ben Zakour NL, Stanton-Cook M, Skippington E, Totsika M, Forde BM, Phan MD, Gomes Moriel D, Peters KM, Davies M, Rogers BA, Dougan G, Rodriguez-Bano J, Pascual A, Pitout JD, Upton M, Paterson DL, Walsh TR, Schembri MA, Beatson SA. 2014. Global dissemination of a multidrug-resistant Escherichia coli clone. Proc Natl Acad Sci U S A 111:5694–5699.
11.
Pitout JD, DeVinney R. 2017. Escherichia coli ST131: a multidrug-resistant clone primed for global domination. F1000Res 6:F1000.
12.
Hood MI, Skaar EP. 2012. Nutritional immunity: transition metals at the pathogen-host interface. Nat Rev Microbiol 10:525–537.
13.
Ewers C, Antao EM, Diehl I, Philipp HC, Wieler LH. 2009. Intestine and environment of the chicken as reservoirs for extraintestinal pathogenic Escherichia coli strains with zoonotic potential. Appl Environ Microbiol 75:184–192.
14.
Johnson JR, Russo TA. 2002. Extraintestinal pathogenic Escherichia coli: “the other bad E. coli.” J Lab Clin Med 139:155–162.
15.
Kohler CD, Dobrindt U. 2011. What defines extraintestinal pathogenic Escherichia coli? Int J Med Microbiol 301:642–647.
16.
Peirano G, Richardson D, Nigrin J, McGeer A, Loo V, Toye B, Alfa M, Pienaar C, Kibsey P, Pitout JD. 2010. High prevalence of ST131 isolates producing CTX-M-15 and CTX-M-14 among extended-spectrum-beta-lactamase-producing Escherichia coli isolates from Canada. Antimicrob Agents Chemother 54:1327–1330.
17.
Wright KJ, Seed PC, Hultgren SJ. 2007. Development of intracellular bacterial communities of uropathogenic Escherichia coli depends on type 1 pili. Cell Microbiol 9:2230–2241.
18.
Hagan EC, Lloyd AL, Rasko DA, Faerber GJ, Mobley HLT. 2010. Escherichia coli global gene expression in urine from women with urinary tract infection. PLoS Pathog 6:e1001187.
19.
Vangchhia B, Abraham S, Bell JM, Collignon P, Gibson JS, Ingram PR, Johnson JR, Kennedy K, Trott DJ, Turnidge JD, Gordon DM. 2016. Phylogenetic diversity, antimicrobial susceptibility, and virulence characteristics of phylogroup F Escherichia coli in Australia. Microbiology 162:1904–1912.
20.
Serra DO, Mika F, Richter AM, Hengge R. 2016. The green tea polyphenol EGCG inhibits E. coli biofilm formation by impairing amyloid curli fibre assembly and downregulating the biofilm regulator CsgD via the σE-dependent sRNA RybB. Mol Microbiol 101:136–151.
21.
Gophna U, Barlev M, Seijffers R, Oelschlager TA, Hacker J, Ron EZ. 2001. Curli fibers mediate internalization of Escherichia coli by eukaryotic cells. Infect Immun 69:2659–2665.
22.
Solano C, Garcia B, Valle J, Berasain C, Ghigo JM, Gamazo C, Lasa I. 2002. Genetic analysis of Salmonella enteritidis biofilm formation: critical role of cellulose. Mol Microbiol 43:793–808.
23.
Guenther S, Aschenbrenner K, Stamm I, Bethe A, Semmler T, Stubbe A, Stubbe M, Batsajkhan N, Glupczynski Y, Wieler LH, Ewers C. 2012. Comparable high rates of extended-spectrum-beta-lactamase-producing Escherichia coli in birds of prey from Germany and Mongolia. PLoS One 7:e53039.
24.
Hasan B, Olsen B, Alam A, Akter L, Melhus A. 2015. Dissemination of multidrug-resistant ESBL-producing Escherichia coli O25b-ST131 clone and role of house crow (Corvus splendens) foraging on hospital waste in Bangladesh. Clin Microbiol Infect.
25.
Mora A, Herrera A, Mamani R, López C, Alonso MP, Blanco JE, Blanco M, Dahbi G, García-Garrote F, Pita JM, Coira A, Bernárdez MI, Blanco J. 2010. Recent emergence of clonal group O25b:K1:H4-B2-ST131 ibeA strains among Escherichia coli poultry isolates, including CTX-M-9-producing strains, and comparison with clinical human isolates. Appl Environ Microbiol 76:6991–6997.
26.
McNally A, Kallonen T, Connor C, Abudahab K, Aanensen DM, Horner C, Peacock SJ, Parkhill J, Croucher NJ, Corander J. 2019. Diversification of colonization factors in a multidrug-resistant Escherichia coli lineage evolving under negative frequency-dependent selection. mBio 10:e00644-19.
27.
Pitout JDD. 2012. Extraintestinal pathogenic Escherichia coli: a combination of virulence with antibiotic resistance. Front Microbiol 3:9.
28.
Calhau V, Ribeiro G, Mendonca N, Da Silva GJ. 2013. Prevalent combination of virulence and plasmidic-encoded resistance in ST131 Escherichia coli strains. Virulence 4:726–729.
29.
Clinical and Laboratory Standards Institute. 2008. Performance standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals; approved standard, 3rd ed. Clinical and Laboratory Standards Institute, Wayne, PA.
30.
Zerbino DR, Birney E. 2008. Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome Res 18:821–829.
31.
Otto TD, Sanders M, Berriman M, Newbold C. 2010. Iterative correction of reference nucleotides (iCORN) using second generation sequencing technology. Bioinformatics 26:1704–1707.
32.
Seemann T. 2014. Prokka: rapid prokaryotic genome annotation. Bioinformatics 30:2068–2069.
33.
Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30:2114–2120.
34.
Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R. 2009. The sequence alignment/map format and SAMtools. Bioinformatics 25:2078–2079.
35.
Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, Miller CA, Mardis ER, Ding L, Wilson RK. 2012. VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing. Genome Res 22:568–576.
36.
Stamatakis A. 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30:1312–1313.
37.
Corander J, Tang J. 2007. Bayesian analysis of population structure based on linked molecular information. Math Biosci 205:19–31.
38.
Hadfield J, Croucher NJ, Goater RJ, Abudahab K, Aanensen DM, Harris SR. 2018. Phandango: an interactive viewer for bacterial population genomics. Bioinformatics 34:292–293.
39.
To TH, Jung M, Lycett S, Gascuel O. 2016. Fast dating using least-squares criteria and algorithms. Syst Biol 65:82–97.
40.
Page AJ, Cummins CA, Hunt M, Wong VK, Reuter S, Holden MT, Fookes M, Falush D, Keane JA, Parkhill J. 2015. Roary: rapid large-scale prokaryote pan genome analysis. Bioinformatics 31:3691–3693.
41.
Zankari E, Hasman H, Cosentino S, Vestergaard M, Rasmussen S, Lund O, Aarestrup FM, Larsen MV. 2012. Identification of acquired antimicrobial resistance genes. J Antimicrob Chemother 67:2640–2644.
42.
Joensen KG, Scheutz F, Lund O, Hasman H, Kaas RS, Nielsen EM, Aarestrup FM. 2014. Real-time whole-genome sequencing for routine typing, surveillance, and outbreak detection of verotoxigenic Escherichia coli. J Clin Microbiol 52:1501–1510.
43.
Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL. 2009. BLAST plus: architecture and applications. BMC Bioinformatics 10:421.
44.
Alikhan NF, Petty NK, Ben Zakour NL, Beatson SA. 2011. BLAST Ring Image Generator (BRIG): simple prokaryote genome comparisons. BMC Genomics 12:402.
45.
Romling U. 2005. Characterization of the rdar morphotype, a multicellular behavior in Enterobacteriaceae. Cell Mol Life Sci 62:1234–1246.
46.
Blomfield IC, McClain MS, Princ JA, Calie PJ, Eisenstein BI. 1991. Type 1 fimbriation and fimE mutants of Escherichia coli K-12. J Bacteriol 173:5298–5307.
47.
Hayashi K, Morooka N, Yamamoto Y, Fujita K, Isono K, Choi S, Ohtsubo E, Baba T, Wanner BL, Mori H, Horiuchi T. 2006. Highly accurate genome sequences of Escherichia coli K-12 strains MG1655 and W3110. Mol Systems Biol 2:2006.2007.
48.
Schwyn B, Neilands JB. 1987. Universal chemical-assay for the detection and determination of siderophores. Anal Biochem 160:47–56.
49.
Cheng L, Connor TR, Siren J, Aanensen DM, Corander J. 2013. Hierarchical and spatially explicit clustering of DNA sequences with BAPS software. Mol Biol Evol 30:1224–1228.
Information & Contributors
Information
Published In
Antimicrobial Agents and Chemotherapy
Volume 63 • Number 6 • June 2019
eLocator: 10.1128/aac.00243-19
PubMed: 30885899
Copyright
© 2019 American Society for Microbiology. All Rights Reserved.
History
Received: 1 February 2019
Returned for modification: 18 February 2019
Accepted: 11 March 2019
Published online: 24 May 2019
Keywords
Contributors
Metrics & Citations
Metrics
Note:
- For recently published articles, the TOTAL download count will appear as zero until a new month starts.
- There is a 3- to 4-day delay in article usage, so article usage will not appear immediately after publication.
- Citation counts come from the Crossref Cited by service.
Citations
Citation
2019.
Genomic and Functional Analysis of Emerging Virulent and Multidrug-Resistant Escherichia coli Lineage Sequence Type 648. Antimicrob Agents Chemother
63:10.1128/aac.00243-19.
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. For an editable text file, please select Medlars format which will download as a .txt file. Simply select your manager software from the list below and click Download.