Development and Characterization of High-Throughput Caenorhabditis elegans - Enterococcus faecium Infection Model

Abstract

The genus Enterococcus includes two Gram-positive pathogens of particular clinical relevance: E. faecalis and E. faecium. Infections with each of these pathogens are becoming more frequent, particularly in the case of hospital-acquired infections. Like most other bacterial species of clinical importance, antimicrobial resistance (and, specifically, multi-drug resistance) is an increasing threat, with both species considered to be of particular importance by the World Health Organization and the US Centers for Disease Control. The threat of antimicrobial resistance is exacerbated by the staggering difference in the speeds of development for the discovery and development of the antimicrobials versus resistance mechanisms. In the search for alternative strategies, modulation of host-pathogen interactions in general, and virulence inhibition in particular, have drawn substantial attention. Unfortunately, these approaches require a fairly comprehensive understanding of virulence determinants. This requirement is complicated by the fact that enterococcal infection models generally require vertebrates, making them slow, expensive, and ethically problematic, particularly when considering the thousands of animals that would be needed for the early stages of experimentation. To address this problem, we developed the first high-throughput C. elegans-E. faecium infection model involving host death. Importantly, this model recapitulates many key aspects of murine peritonitis models, including utilizing similar virulence determinants. Additionally, host death is independent of peroxide production, unlike other E. faecium-C. elegans virulence models, which allows the assessment of other virulence factors. Using this system, we analyzed a panel of lab strains with deletions of targeted virulence factors. Although removal of certain virulence factors (e.g., Δfms15) was sufficient to affect virulence, multiple deletions were generally required to affect pathogenesis, suggesting that host-pathogen interactions are multifactorial. These data were corroborated by genomic analysis of selected isolates with high and low levels of virulence. We anticipate that this platform will be useful for identifying new treatments for E. faecium infection.

Keywords: C. elegans; E. faecium; high-throughput screen; host-pathogen interaction; virulence.

Figure 1

E. faecium E007 is pathogenic to C. elegans in the Enterococcus liquid-killing assay. (A) High-throughput, liquid-based C. elegans – E. faecium killing assay pipeline (right, compared to the conventional agar-based assay (left)). (B, C) Survival curves of C. elegans that were exposed to E. faecalis OG1RF, E. faecium E007, or non-pathogenic E. coli OP50 either in the (B) agar-based assay or (C) under liquid conditions. Three biological replicates were performed and analyzed; representative replicates are shown for (B, C).

Figure 2

E. faecium virulence in Enterococcus liquid killing assay does not depend upon bacterial growth. (A) Correlation plot of a subset of ~30 E. faecium strains’ growth (OD 600, x-axis) against their virulence (based on C. elegans’ survival, y-axis). (B) Correlation plot of ~120 of E. faecium strains’ growth (OD 600, x-axis) against their virulence (based on C. elegans’ survival, y-axis). Three biological replicates with ~80 worms/replicate were analyzed. The correlation plot was generated with R package ggplot2 (version 3.3.3).

Figure 3

PMK-1 did not confer resistance towards E. faecium in Enterococcus liquid-killing assay. (A) Fluorescence images and (B) quantification of C. elegans survival after exposure to E. faecium E007 in liquid. Worms were reared on E. coli expressing empty vector (EV) as control or mdt-15(RNAi) or pmk-1(RNAi). (C) Quantification of C. elegans survival at 120 h exposure to E. faecium E007 in agar. Three biological replicates with ~400 worms/replicate for liquid-based assay or ~180 worms/replicate for agar-based assay were analyzed. Error bars represent SEM. p values were determined from Student’s t-test. NS not significant, *p < 0.05, ***p < 0.001.

Figure 4

Pathogenicity distribution of E. faecium strains belonging to Clade A1, A2, or B C. elegans were exposed to E. faecium strains belonging to (A) Clade A1, (B) Clade A2, or (C) Clade B in the liquid-killing assay. At least three biological replicates with ~80 worms/replicate were analyzed. Error bars represent SEM. p values were determined from Student’s t-test against the median isolates (indicated with short black lines in the graphs). Arrows indicate all isolates that are more (red) or less (blue) virulent than the median isolates (p < 0.05).

Figure 5

Pathogenicity of E. faecium from distinct clades does not differ significantly. (A, B) Pathogenesis of (A) preliminary (~30) or (B) all available (~120) E. faecium strains belonging to Clade A1, A2, or B In (A, B), C. elegans survival (y-axis) were plotted against E. faecium OD 600 (x-axis). Clustering was performed with ‘kmeans’ function (R v4.0.3). All panels show pooled results from three biological replicates. p-values were determined from Student’s t-test and are indicated on graphs.

Figure 6

Highly virulent E. faecium strains showed higher colonization, compared to low virulence strains, in the Enterococcus liquid-killing assay. Six E. faecium strains (three with the highest virulence in the liquid-based assay and three with the lowest virulence) were exposed to C. elegans in liquid-based assay. Panel (A) shows strain pathogenesis (C. elegans survival), while panel (B) shows pooled average of CFU from C. elegans gut after colonization. CFU counts for individual strains are in Figure S3 . (C) The killing ability of bacteria-free filtrates (from isolates in panel A) was examined after 100 h incubation with C. elegans. Three biological replicates with ~60 worms/replicate for CFU assay or ~400 worms/replicate for pathogenesis assay were analyzed. Error bars represent SEM. p-values were determined from Student’s t-test between pooled low- and high virulence groups. NS not significant, ***p < 0.001.

Figure 7

Genes encoding surface proteins and pili are important for pathogenesis in the C. elegans – E. faecium liquid-killing assay. Pathogenicity of several strains harboring mutations in (A) fms10 (scm), acm, fms18, or fms15, or (B) fms, scm, wxl, or ccpA were assayed in the C. elegans – E. faecium liquid-based pathosystem. Three biological replicates with ~400 worms/replicate were analyzed. Error bars represent SEM. p values were determined from Student’s t-test. NS not significant, *p < 0.05, ** p < 0.01, ***p < 0.001.

Figure 8

Genomic analysis showed a relationship between high- and low virulence strains. (A) A maximum-likelihood phylogenetic tree of eight strains belonging to the highest virulence group (yellow), six strains belonging to the lowest virulence group (purple), and a reference strain (TX0016, black). The phylogenetic tree was created by performing roary (Galaxy version 3.13.0), followed by RAxML (Galaxy version 1.0.0), and visualized by using treeio (v1.14.3) and ggtree (v2.4.1) packages (RStudio, R version 4.0.3). (B) Kinetic growth measurement or optical densities (OD 600) of TX2046 and TX2051 in BHI media. (C) Kinetic growth measurement or optical densities (OD 600) of three high- (TX2416, TX2051, TX2029) and three low-virulence isolates (TX2022, S447, TX2025), and reference strain E007 in BHI media. For (B, C), three biological replicates were performed and analyzed; a representative replicate is shown.

Figure 9

The C. elegans – E. faecium pathogenesis model recapitulates E. faecium virulence. (A, B) Survival curves of (A) C. elegans or (B) mice that were exposed to two E. faecium strains TX0016 and TX1330. All panels show results from pooled biological replicates (each dot in A represents a group of 20 worms). p-values were determined from Student’s t-test and are indicated on graphs.