Enterococcus faecalis Mutations Affecting Virulence in the Caenorhabditis elegans Model Host

Abstract

Enterococcus faecalis transposon insertion mutants were screened for attenuated killing of the nematode model host Caenorhabditis elegans. The genes disrupted in the attenuated mutants encode a variety of factors including transcriptional regulators, transporters, and damage control and repair systems. Five of nine mutants tested were attenuated in a mouse peritonitis model.

Enterococcus faecalis has emerged over the past few decades as one of the leading causes of hospital-acquired infection, causing diseases such as endocarditis, urinary tract infections, and bloodstream infections (14). E. faecalis's natural ruggedness, which causes intrinsic resistance to many antibiotics, and its versatility in swapping genetic information to gain additional resistance, have played large roles in its advance (12). Several E. faecalis virulence factors have been identified by their distinct biochemical properties, by their antigenic phenotypes, or by homology searches using known virulence factors from other bacteria (8). Surprisingly, given the importance of this pathogen, no in vitro or in vivo screens have been performed to isolate new virulence determinants in E. faecalis in an unbiased manner. We previously demonstrated that Caenorhabditis elegans could be used as a model host to identify potential mammalian virulence determinants (6). Additionally, we built an ordered library of transposon insertion mutants of E. faecalis strain OG1RF with approximately 25% of the nonessential genes disrupted (7). In this work, we identified 23 insertion mutants in the ordered library with attenuated killing of C. elegans. Five of nine mutants tested were also less virulent in a mouse peritonitis model.

To identify strains of E. faecalis mutants from our ordered library of 540 mutants that were deficient in C. elegans killing, the following strategy was used. Plates containing a bacterial lawn of each transposon mutant were generated by using growth conditions previously found optimal for killing of C. elegans by E. faecalis (6). Approximately 30 to 40 worms, strain N2 (the wild type), were placed on each mutant, and survival over time was assayed for 7 days at 24-h intervals. With GraphPad Prism 3.0 or STATA 6.0, survival was plotted by the Kaplan-Meier method and differences between the mutant and parent strain were compared by using the log-rank test. Seventy-two mutants with a difference resulting in a P value of 0.1 or less were tested in a second assay with a larger population of C. elegans (n = 60 to 80). P values of less than 0.05 were considered statistically significant. Mutants that caused significant attenuation by this criterion in both experiments are listed in Table 1, and an example of a typical killing assay is presented in Fig. 1A. Because mutants with a growth defect could cause a reduction in killing in a nonspecific manner, log-phase growth in liquid brain heart infusion (BHI) medium of all of the mutants was compared to that of the parent strain. All mutants displayed growth similar to that of the wild type, except for mutant 4H12 (Fig. 1B and Table 1). We also assayed growth in 50% serum from human volunteers as described previously (5), with and without heat inactivation. Growth in serum may more closely parallel conditions found in the mammalian host environment, and the comparison with heat-inactivated serum addresses whether or not there is increased sensitivity to complement. None of the mutants displayed a growth defect under either condition, including 4H12 (data not shown). Perhaps the deficiency of 4H12 in BHI is specific to conditions that result in a short doubling time.

TABLE 1.

Attenuated E. faecalis mutants identified by screening with C. elegans

Type of mutation and mutant	V583 ORFa	Mouse model P valueb	Gene name	Gene product, function
Transcriptional regulator
4G3	EF_0382	NT	None	Helix-turn-helix, Fis type
1C4	EF_1302	0.13	lysR	Helix-turn-helix, lysR type
1D1	EF_1569c	0.14	psr	CpsA/LytR/Psr family of transcriptional regulators
5C11	EF_1604	0.19	scrR-1	Negative regulator of sucrose operon
4D4	EF_1821	0.001e	fsrB	Peptide processor of Fsr two-component system
Transporter
2A1	EF_0243	NT	brnQ	Branched-chain amino acid transporter
4C8	EF_1513	0.31	None	Pheromone binding protein
4G6	EF_2598	NT	None	PEP-PTSf, beta-glucoside-specific IIABC component
Damage control and repair
6A3	EF_1545	0.005	recQ-1	DNA helicase, SOS response
3G8	EF_1598	0.038	phrB	Deoxyribodipyrimidine photolyase, DNA repair
4G7	EF_2591	NT	None	glxI, glyoxalase, protective against electrophiles
Catabolism
4G8	EF_0737	NT	None	Putative amidase
6C9	EF_1603	0.01e	scrB-1	Sucrose-6-phosphate hydrolase
6E9	EF_1623c	NT	pduJ	Carboxysome protein, metabolism of ethanolamine
Anabolism
1C9	EF_1576	NT	thyA	Thymidylate synthase, deoxyribonucleotide biosynthesis
4H12	EF_2200d	NT	map	Methionyl aminopeptidase, protein maturation
Other
4G10	EF_0376	NT	None	Hypothetical protein
6C10	EF_0861c	NT	None	GNAT family acetyltransferase
6B3	EF_1542	NT	None	Conserved hypothetical membrane protein
4F8	EF_1590c	0.0003	paiA	SSAT acetyltransferase, negative regulator of sporulation
6D5	EF_1792	NT	None	Conserved hypothetical
4G9	EF_2675	NT	None	Putative competence protein, CoiA-like family
6D6	EF_2957	NT	None	Hexapeptide repeat acetyltransferase

^aORF, open reading frame.

^bMouse peritonitis model (6, 22). P values are from representative experiments. NT, not tested.

^cRegulated by Fsr system.

^dGrowth defect.

^ePreviously found to be attenuated in C. elegans and mouse peritonitis models (6, 20).

^fPEP-PTS, phosphoenolpyruvate-dependent phosphotransferase system.

FIG. 1.

Example killing assay and growth curve of E. faecalis mutants with attenuated C. elegans killing. (A) Survival of C. elegans on OG1RF (squares) compared to mutants with disruptions in 4D4 (triangles) (P < 0.0001), 4G6 (diamonds) (P < 0.0001), and 4H12 (circles) (P < 0.0194). (B) Growth curves of E. faecalis mutants listed in panel A grown in BHI medium. OD 600, optical density at 600 nm.

The mutants found in the screen were classified by the probable function of the protein encoded by the disrupted gene according to the annotation provided by the V583 genome sequence (19). We subjected the sequences to BLAST analysis and researched the literature for additional information on possible functions. Several putative transcriptional regulators were identified. EF_1302, for example, resembles a lysR helix-turn-helix transcriptional regulator involved in virulence and stress response in other bacteria such as Pseudomonas aeruginosa (21). EF_1569 encodes Psr, which belongs to a family of negative transcriptional regulators that control cell surface properties such as cell wall and exopolysaccharide composition and synthesis (15). Such a mutant could be affected in adhesion or biofilm formation, properties that affect the infectivity of E. faecalis. We also identified EF_1604, which encodes the transcriptional repressor ScrR. In previous work (6) (as well as this study), we found that a mutation in scrB, a gene likely regulated by ScrR and encoding sucrose-6-hydrolase, caused attenuation in C. elegans killing. Sucrose utilization plays an important role in biofilm formation and pathogenicity in S. mutans during caries formation (16) and endocarditis (17), as it is a substrate for the synthesis of the extracellular polymers glucan and fructan (9). We also identified FsrB in this screen; the Fsr two-component regulatory system is a major regulator of virulence in E. faecalis, and mutation of FsrB attenuates virulence in a variety of infection models, including C. elegans (6, 11, 18, 20).

Transporters were found, including a pheromone binding protein possibly involved in quorum sensing and a phosphoenolpyruvate-dependent phosphotransferase system beta-glucoside-specific IIABC component. Genes involved in beta-glucoside metabolism were found to be upregulated in Streptococcus gordonii on infected heart valves and contributed to biofilm formation (13).

Some of the mutants with attenuated C. elegans killing have insertions in genes encoding enzymes that could be involved in damage control and repair (Table 1). For example, recQ (EF_1545) and phrB (EF_1598) homologs encode DNA repair enzymes. Loss of such enzymes in other pathogens such as Salmonella results in attenuated killing of mice (2) and sensitivity to the oxidative burst in macrophages (3). glxI (EF_2591) has been found to be upregulated in macrophage-engulfed Salmonella and may have roles in dealing with oxidative stress (4).

EF_1623 is one of a cluster of genes that are orthologs of the eut-pdu operons in Salmonella. The E. faecalis operon appears to be involved in the use of ethanolamine, a readily available lipid component, as a carbon and nitrogen source. This gene cluster was previously identified as being strongly regulated by the Fsr system (1), and some of the components have been found to be upregulated in Salmonella engulfed by macrophages (10).

To determine if any of the newly identified mutants caused loss of infectivity in a mammalian model, seven were tested in a mouse peritonitis model. Mice were inoculated intraperitoneally with E. faecalis as previously described (6, 22). The significance of differences in survival time compared to that of the wild type was evaluated in the same manner as described for C. elegans. The experiment was repeated twice and carried out under approved protocols. Three mutants, 6A3, 3G8, and 4F8, were attenuated (P < 0.05) (Fig. 2 and Table 1). Additionally, two of the mutants identified, 6C9 and 4D4, had insertions in genes characterized as causing attenuation in the mouse model in previous work (6, 20). It is conceivable that use of a different animal model, such as a model of endocarditis, would identify a different subset of attenuated mutants.

FIG. 2.

Survival of mice infected intraperitoneally with E. faecalis mutants. Shown are representative experiments with the mutants characterized as attenuated in this model but not previously published (6A3, 4F8, and 3G8). (A) Survival of mice infected with 7.35 × 10⁸ CFU of OG1RF, 7.37 × 10⁸ CFU of 6A3, and 7.41 × 10⁸ CFU of 4F8. (B) Survival of mice infected with 5.55 × 10⁸ CFU of OG1RF and 5.65 × 10⁸ CFU of 3G8.

In summary, we identified 23 insertion mutants with attenuated C. elegans killing, 2 of which were previously known to affect pathogenesis in the worm and the mouse (6, 20). Several orthologs of the genes disrupted by the transposon insertions are known to affect mammalian pathogenesis in other bacterial species, and more than half of those tested were attenuated in a mouse peritonitis model. In conclusion, C. elegans is an efficient way to identify potential virulence determinants and screening a complete ordered library would likely uncover additional factors.