Comprehensive Functional Analysis of the Enterococcus faecalis Core Genome Using an Ordered, Sequence-Defined Collection of Insertional Mutations in Strain OG1RF

Abstract

Enterococcus faecalis is a common commensal bacterium in animal gastrointestinal (GI) tracts and a leading cause of opportunistic infections of humans in the modern health care setting. E. faecalis OG1RF is a plasmid-free strain that contains few mobile elements yet retains the robust survival characteristics, intrinsic antibiotic resistance, and virulence traits characteristic of most E. faecalis genotypes. To facilitate interrogation of the core enterococcal genetic determinants for competitive fitness in the GI tract, biofilm formation, intrinsic antimicrobial resistance, and survival in the environment, we generated an arrayed, sequence-defined set of chromosomal transposon insertions in OG1RF. We used an orthogonal pooling strategy in conjunction with Illumina sequencing to identify a set of mutants with unique, single Himar-based transposon insertions. The mutants contained insertions in 1,926 of 2,651 (72.6%) annotated open reading frames and in the majority of hypothetical protein-encoding genes and intergenic regions greater than 100 bp in length, which could encode small RNAs. As proof of principle of the usefulness of this arrayed transposon library, we created a minimal input pool containing 6,829 mutants chosen for maximal genomic coverage and used an approach that we term SMarT (sequence-defined mariner technology) transposon sequencing (TnSeq) to identify numerous genetic determinants of bile resistance in E. faecalis OG1RF. These included several genes previously associated with bile acid resistance as well as new loci. Our arrayed library allows functional screening of a large percentage of the genome with a relatively small number of mutants, reducing potential effects of bottlenecking, and enables immediate recovery of mutants following competitions. IMPORTANCE The robust ability of Enterococcus faecalis to survive outside the host and to spread via oral-fecal transmission and its high degree of intrinsic and acquired antimicrobial resistance all complicate the treatment of hospital-acquired enterococcal infections. The conserved E. faecalis core genome serves as an important genetic scaffold for evolution of this bacterium in the modern health care setting and also provides interesting vaccine and drug targets. We used an innovative pooling/sequencing strategy to map a large collection of arrayed transposon insertions in E. faecalis OG1RF and generated an arrayed library of defined mutants covering approximately 70% of the OG1RF genome. Then, we performed high-throughput transposon sequencing experiments using this library to determine core genomic determinants of bile resistance in OG1RF. This collection is a valuable resource for comprehensive, functional enterococcal genomics using both traditional and high-throughput approaches and enables immediate recovery of mutants of interest.

Keywords: TnSeq; functional genomics; gut fitness; opportunistic pathogen.

FIG 1

The Straight Three mapping approach to large-scale Tn site mapping in arrayed libraries. The Straight Three algorithm relies on 3D (row × column × plate) orthogonal pooling to map Tn insertion sites. Straight Three mapping is illustrated using the mutant in plate 1, well A1. (A) Pools of mutants were made from columns (yellow shading), rows (blue shading), and full plates. (B and C) Unambiguous assignment of a Tn insertion location to a specific well in a given plate occurs when that position is identified once in a column pool, once in a row pool, and once in a single-plate pool (see, for example, plate 1, well A1). Mutants occupying analogous wells in different plates (for example, plate 2, well A1, and plate 3, well A1) are sorted based on their presence in the plate pools. For examples of results obtained when multiple clones containing Tn insertions at the same site are present in a single plate or when multiple Tn insertions are present in the same well, see Fig. S1.

FIG 2

Library preparation and mapping of Tn insertion sites. (A) Next-generation sequencing libraries prepared from OG1RF mariner Tn mutant DNA potentially contain three groups of genomic DNA fragments: 1, genomic DNA without Tn sequence (gray); 2, Tn-only sequence (yellow); and 3, genomic DNA-Tn junctions (gray and yellow). Only genomic DNA (gDNA)-Tn junctions (≈0.01% of library fragments) will be effectively enriched via TnSeq. (B) TnSeq amplification using primers that recognize Illumina library adapters (blue) and the 3′ ends of mariner elements selectively amplifies regions that contain genomic DNA-Tn junctions. Sequences on either side of the Tn insertion are identified due to bidirectional amplification of the mariner element. (C) Mapping of reads obtained from Illumina sequencing to the OG1RF genome identifies the precise TA site of a Tn insertion.

FIG 3

Distribution of mapped Tn insertions in the E. faecalis OG1RF genome. (A) Linear distribution of all mapped Tn mutants (black dots) ordered by position in the chromosome. (B) The distance between Tn mutants (red dots) was plotted relative to chromosomal position. The interinsertional spacing between transposon mutants is relatively constant across the OG1RF genome (316 ± 580 bp). The black line represents the running average of Tn spacing (with a 50-bp sliding window) for any given chromosomal position.

FIG 4

Distribution of Tn insertions across a subset of genes. Tn insertions (dots) mapped to the genomic region from OG1RF_10198 (aldA, aldehyde dehydrogenase) through OG1RF_10217 (phosphoglycerate mutase). The y axis indicates the chromosome position. Each rectangle spanning the x axis represents a different genomic feature (characterized protein-encoding genes [orange], hypothetical genes [green], intergenic regions [white], rRNA [red], and tRNA [teal]). In loci containing insertions, Tn elements are evenly distributed (distribution throughout the length of each feature is represented by the position on the y axis). For simplicity, the intergenic identifiers are not shown.

FIG 5

SMarT TnSeq analysis of OG1RF growth in cholic acid. (A) Growth of the SMarT TnSeq library. Samples were taken from untreated cultures (solid lines) or cultures treated with 0.15% cholic acid (dashed lines). Cells were harvested at early and late time points (t₁ and t₂, respectively). Sampling points are shown with arrows. Black and gray lines indicate separate biological replicates. (B) The total number of TA sites significantly underrepresented in each comparison (P < 0.05, log₂ fold change < 0). The cholic acid-specific pairwise comparisons are shown as different colors. (C) TA sites shared between comparisons. Cholic acid-specific comparisons are colored as in panel B. Comparisons between untreated cells and the input sample were included to evaluate the number of mutants with general growth defects.

FIG 6

Cholic acid sensitivity of Tn mutants identified by SMarT TnSeq. All strains were grown in MM9-YEG–0.15% cholic acid (or MM9-YEG medium alone for untreated samples). (A) Strains with dnaK mutations (transposon insertion, dnaK-Tn; markerless deletion, ΔdnaK) do not grow in the presence of cholic acid (open purple and red circles). Expression of wild-type dnaK from a plasmid (dnaK-Tn pDnaK, open green circles; ΔdnaK pDnaK, open teal circles) significantly rescues growth relative to the vector control strains. (B) Mutations that disrupt the translation of CcfA (Tn insertion, ccfA-Tn; stop codon, ccfA2) significantly reduce survival in cholic acid. Altering the cCF10 pheromone sequence to an inactive variant via an alanine substitution (ccfA-LAT) does not alter survival in cholic acid compared to OG1RF (compare open purple and black circles). In the untreated condition, none of the mutant strains had significant differences in growth relative to OG1RF (filled circles). (C) Strains with Tn insertions in hypothetical proteins (10022-Tn and 10040-Tn) have significantly reduced survival in the presence of cholic acid (open red and purple circles). These growth defects are complemented with a plasmid-borne copy of the genes (open green and teal circles, respectively). Values plotted at each time point are the averages from three biological replicates (each with two technical replicates). Error bars are standard errors of the means. Significant differences between OD₆₀₀ values at the end of the experiment (t = 10 h) were determined using a two-tailed Student t test (P < 0.05).