Comparative Genome Analysis of Enterococcus cecorum Reveals Intercontinental Spread of a Lineage of Clinical Poultry Isolates

Abstract

Enterococcus cecorum is an emerging pathogen responsible for osteomyelitis, spondylitis, and femoral head necrosis causing animal suffering and mortality and requiring antimicrobial use in poultry. Paradoxically, E. cecorum is a common inhabitant of the intestinal microbiota of adult chickens. Despite evidence suggesting the existence of clones with pathogenic potential, the genetic and phenotypic relatedness of disease-associated isolates remains little investigated. Here, we sequenced and analyzed the genomes and characterized the phenotypes of more than 100 isolates, the majority of which were collected over the last 10 years from 16 French broiler farms. Comparative genomics, genome-wide association studies, and the measured susceptibility to serum, biofilm-forming capacity, and adhesion to chicken type II collagen were used to identify features associated with clinical isolates. We found that none of the tested phenotypes could discriminate the origin of the isolates or the phylogenetic group. Instead, we found that most clinical isolates are grouped phylogenetically, and our analyses selected six genes that discriminate 94% of isolates associated with disease from those that are not. Analysis of the resistome and the mobilome revealed that multidrug-resistant clones of E. cecorum cluster into a few clades and that integrative conjugative elements and genomic islands are the main carriers of antimicrobial resistance. This comprehensive genomic analysis shows that disease-associated clones of E. cecorum belong mainly to one phylogenetic clade. IMPORTANCE Enterococcus cecorum is an important pathogen of poultry worldwide. It causes a number of locomotor disorders and septicemia, particularly in fast-growing broilers. Animal suffering, antimicrobial use, and associated economic losses require a better understanding of disease-associated E. cecorum isolates. To address this need, we performed whole-genome sequencing and analysis of a large collection of isolates responsible for outbreaks in France. By providing the first data set on the genetic diversity and resistome of E. cecorum strains circulating in France, we pinpoint an epidemic lineage that is probably also circulating elsewhere that should be targeted preferentially by preventive strategies in order to reduce the burden of E. cecorum-related diseases.

Keywords: Enterococcus cecorum; antimicrobial resistance; avian pathogenesis; comparative genomics; poultry; veterinary pathogens.

FIG 1

Pangenome analysis of 148 E. cecorum genomes. (A) Accumulation curves of gene clusters of the pan- and core genomes. The pangenome size, in black, corresponds to the total number of gene clusters against the number of genomes included. The core-genome size, in red, corresponds to the number of gene clusters in common against the number of genomes included; numbers were averaged on 1,000 randomized orders for genome addition. (B) Gene frequency spectrum. Only one representative per gene cluster is considered.

FIG 2

Phylogenetic tree and clinical status of 148 E. cecorum isolates. A neighbor-joining (BioNJ) tree was built on the pairwise distances between genomes. The innermost circle shows clades A to E with subclades E1 to E13 (colored strips). The first external circle shows the clinical status of the isolates. The second external circle shows the geographic origin. The third external circle shows the year of isolation. The name underlined in red corresponds to the reference genome of NCTC12421. Isolates CIRMBP-1307, CIRMBP-1308, CIRMBP-1309, CIRMBP-1310, CIRMBP-1311, CIRMBP-1312, CIRMBP-1313, CIRMBP-1314, CIRMBP-1315, CIRMBP-1316, and CIRMBP-1317 correspond to purified clones of strains CE1, CE2, CE3, SA1, SA2, SA3, CB-32, BB-66, CL-1, G-29, and D-104, respectively (see Table S1 in the supplemental material). NA, not applicable.

FIG 3

Distribution of antimicrobial resistance genes in sequenced genomes. The clade, geographic origin (blue squares, France; orange circles, Germany; red stars, United States; red triangles, Poland; yellow squares, Belgium), isolation year, and clinical origin of the sample (CH, clinical human; NCP, nonclinical poultry; CP, clinical poultry) are specified. The following antibiotic (Ab) families are represented by alternating gray blocks, from left to right: tetracycline, MLS (macrolide, lincosamide, and streptogramin), aminoglycoside, glycopeptide, and bacitracin. Black strips represent the presence and white ones represent the absence of ARGs. Potential multiresistant isolates (>2 ARGs for different families) are highlighted in gray.

FIG 4

Mobile genetic elements in complete E. cecorum genomes. From the inner to outer circles are CIRMBP-1212, CIRMBP-1283, CIRMBP-1292, CIRMBP-1246, CIRMBP-1302, NCTC12421, CIRMBP-1287, CIRMBP-1320, CIRMBP-1274, CIRMBP-1281, CIRMBP-1261, and CIRMBP-1228. The dotted lines on the genome of CIRMBP-1320 correspond to predicted junctions. Each colored rectangle indicates the integration of an element according to the color key. Hatching indicates a complex genomic island. The dnaA, rpoE, and rpoD genes and the epa and cps loci are indicated.

FIG 5

Distribution of phenotypic expression for biofilm robustness, adhesion to type II collagen, and growth in chicken serum for 118 E. cecorum isolates. Values for biofilm and adhesion to collagen represent estimated marginal mean biovolumes. Values for growth in serum represent estimated marginal means of serum growth indices calculated at 6 h (see the supplemental material). Each bar corresponds to an isolate. Strains with similar phenotypic expression were grouped into clusters on a gray color scale, separated by white lines.

FIG 6

Comparison of the virulence of selected E. cecorum isolates in a chicken embryo model of infection. Shown is a Kaplan-Meier survival plot of chicken embryos (n = 15) infected with 13 different isolates of E. cecorum. The strain name, inoculum size, and phylogenetic group are indicated. The log rank (Mantel-Cox) test indicated significant differences between the positive-control strain CIRMBP-1311 (SA2) and strains CIRMBP-1309 (CE3) (P < 0.0001), CIRMBP-1212 (P = 0.0007), CIRMBP-1281 (P = 0.0364), CIRMBP-1283 (P = 0.0035), CIRMBP-1287 (P = 0.0339), CIRMBP-1294 (P < 0.0001), and CIRMBP-1320 (P < 0.0001). The virulence of strains CIRMBP-1292, CIRMBP-1304, CIRMBP-1228, CIRMBP-1274, and CIRMBP-1302 was not significantly different from that of CIRMBP-1311. Data from one representative experiment of two are shown.