Comprehensive Genome Analysis of Carbapenemase-Producing Enterobacter spp.: New Insights into Phylogeny, Population Structure, and Resistance Mechanisms

Abstract

Knowledge regarding the genomic structure of Enterobacter spp., the second most prevalent carbapenemase-producing Enterobacteriaceae, remains limited. Here we sequenced 97 clinical Enterobacter species isolates that were both carbapenem susceptible and resistant from various geographic regions to decipher the molecular origins of carbapenem resistance and to understand the changing phylogeny of these emerging and drug-resistant pathogens. Of the carbapenem-resistant isolates, 30 possessed bla_KPC-2, 40 had bla_KPC-3, 2 had bla_KPC-4, and 2 had bla_NDM-1 Twenty-three isolates were carbapenem susceptible. Six genomes were sequenced to completion, and their sizes ranged from 4.6 to 5.1 Mbp. Phylogenomic analysis placed 96 of these genomes, 351 additional Enterobacter genomes downloaded from NCBI GenBank, and six newly sequenced type strains into 19 phylogenomic groups-18 groups (A to R) in the Enterobacter cloacae complex and Enterobacter aerogenes Diverse mechanisms underlying the molecular evolutionary trajectory of these drug-resistant Enterobacter spp. were revealed, including the acquisition of an antibiotic resistance plasmid, followed by clonal spread, horizontal transfer of bla_KPC-harboring plasmids between different phylogenomic groups, and repeated transposition of the bla_KPC gene among different plasmid backbones. Group A, which comprises multilocus sequence type 171 (ST171), was the most commonly identified (23% of isolates). Genomic analysis showed that ST171 isolates evolved from a common ancestor and formed two different major clusters; each acquiring unique bla_KPC-harboring plasmids, followed by clonal expansion. The data presented here represent the first comprehensive study of phylogenomic interrogation and the relationship between antibiotic resistance and plasmid discrimination among carbapenem-resistant Enterobacter spp., demonstrating the genetic diversity and complexity of the molecular mechanisms driving antibiotic resistance in this genus.

Importance: Enterobacter spp., especially carbapenemase-producing Enterobacter spp., have emerged as a clinically significant cause of nosocomial infections. However, only limited information is available on the distribution of carbapenem resistance across this genus. Augmenting this problem is an erroneous identification of Enterobacter strains because of ambiguous typing methods and imprecise taxonomy. In this study, we used a whole-genome-based comparative phylogenetic approach to (i) revisit and redefine the genus Enterobacter and (ii) unravel the emergence and evolution of the Klebsiella pneumoniae carbapenemase-harboring Enterobacter spp. Using genomic analysis of 447 sequenced strains, we developed an improved understanding of the species designations within this complex genus and identified the diverse mechanisms driving the molecular evolution of carbapenem resistance. The findings in this study provide a solid genomic framework that will serve as an important resource in the future development of molecular diagnostics and in supporting drug discovery programs.

FIG 1

Enterobacter sp. genomes. Five completely sequenced Enterobacter genomes were compared with the E. hormaechei ST171 strain 34978 genome (pink inner circle). Genes that are present in 34978 but not in the other genomes are shown as blank spaces in the rings representing the five genomes. The GC content and GC skew of strain 34978 are plotted, and prophage locations are indicated by orange rectangles.

FIG 2

Phylogenetic SNP tree of E. cloacae complex genomes. A whole-genome core SNP tree was constructed for 379 E. cloacae complex genomes with kSNP (30) and RAxML (31) (see Materials and Methods). Groups identified by ANI (A to R) are noted as colored nodes, as well as the innermost circle (see group key). Hosts, continents of origin, and KPC type metadata are also plotted in concentric circles as noted. Host colors: human, dark gray; nonhuman, blue. Continent colors: North America, purple; South America, orange; Australia, green; Europe, blue; Asia, red. KPC type colors: 3, red; 2, blue; 4, green.

FIG 3

Phylogenetic universal marker tree of representative genomes of members of the family Enterobacteriaceae. A maximum-likelihood tree was constructed from a concatenated alignment of 26 conserved universal marker alleles from 248 genomes belonging to 48 different genera. If a genus was represented by one to three genomes, the genus was given a letter designation (A to Q; see key, top left). If a genus was represented by four or more genomes and all members clustered together, the genus was labeled directly on the tree; however, if they did not all cluster together, unique noncircular shapes were used to identify them. Each genus is assigned a unique color (see key, top right).

FIG 4

Analysis of the E. hormaechei pangenome. The distribution of protein cluster sizes (the number of genomes sharing each ortholog) generated from the comparison of 219 E. hormaechei genomes with PanOCT (42) indicates the numbers of singleton and core genes (A). The pangenome size (B) and the number of novel genes discovered with the addition of each new genome (C) were estimated by using a pangenome model as described previously (44). Purple circles are the median of each distribution (gray circles). Power law (red lines) and exponential (blue lines) regressions were plotted to determine α (open/closed status) and tg(θ), the average extrapolated number of strain-specific/novel genes, respectively (45).

FIG 5

Historical metabolic differences defining E. hormaechei subspecies explained by fGI content. Variability within the flexible genomic region between core pangenome gene clusters 3936 and 3949 can contain either the gat operon to metabolize galactitol (defining group E) or the aga operon to metabolize N-acetylgalactosamine (defining groups A to D) (A). The presence of an operon between core gene clusters 420 and 404 homologous to the Klebsiella rbt/dal operon (75), for d-arabinitol and ribitiol catabolism, distinguishes group B isolates from A, C, D, and E isolates (B). Circles and arrows represent PanOCT gene clusters and protein coding regions, respectively. The size of a circle is scaled to the gene length of its centroid. The locus identifiers of core genes from applicable PacBio genomes in this study are noted in gray under their respective core clusters. Colored connecting lines represent the levels of conservation between those cluster centroids and genes whose protein sequences match at a BLASTP identity of ≥35% (see key, upper right). Blue circles and arrows denote flanking core clusters and genes, respectively.

FIG 6

Representation of major KPC-harboring plasmids among 97 Enterobacter isolates. (Left) Core SNP phylogenetic tree generated by RAxML. Core SNPs were identified by kSNP v 3.0 (see Materials and Methods). (Middle) The metadata, including isolation location, ST, KPC variants, Tn4401 isoforms, and predicated bla_KPC-harboring plasmids. (Right) Plasmid composition is illustrated by showing the BLASTn matches to each Enterobacter genome across all of the genes on the three reference plasmids, pBK30683, pKPC_UVA01, and pKpQIL. The blue bar denotes a minimal 95% nucleotide sequence identity to the plasmid genes. Abbreviations: uva, pKPC_UVA01-like plasmid; FIA, pBK30683 or pBK30661-like plasmid; Bk., pBK28610-like plasmid; qil, pKpQIL-like plasmid; ukn, bla_KPC-harboring contigs could not be assigned to a known or novel plasmid group; n, pKp048-like non-Tn4401 mobile element (NTM_KPC). Blue bars denote a ≥95% nucleotide sequence match to the plasmid genes.

FIG 7

Representation of plasmid distribution among ST171 Enterobacter isolates. (Left) Core SNP phylogenetic tree generated by RAxML (see Materials and Methods). (Middle) The metadata, including isolation location, isolation year, Tn4401 isoforms, and predicated bla_KPC-harboring plasmids. (Right) Plasmid composition is illustrated by showing the BLASTn matches to each Enterobacter genome across all of the genes on the reference plasmids. Six plasmids from completely sequenced E. xiangfangensis ST171 strain 34978, along with IncX3 plasmid pIncX-SHV (JN247852) and IncX4 plasmid pMNCRE44_4 (CP010880), were used as references. Blue bars denote a ≥95% nucleotide sequence match to the plasmid genes.