Enterococcus Infection Biology: Lessons from Invertebrate Host Models

Abstract

The enterococci are commensals of the gastrointestinal tract of many metazoans, from insects to humans. While they normally do not cause disease in the intestine, they can become pathogenic when they infect sites outside of the gut. Recently, the enterococci have become important nosocomial pathogens, with the majority of human enterococcal infections caused by two species, Enterococcus faecalis and Enterococcus faecium. Studies using invertebrate infection models have revealed insights into the biology of enterococcal infections, as well as general principles underlying host innate immune defense. This review highlights recent findings on Enterococcus infection biology from two invertebrate infection models, the greater wax moth Galleria mellonella and the free-living bacteriovorous nematode Caenorhabditis elegans.

Introduction

The Enterococci are a widely dispersed group of bacteria and more than 40 species have been described (Murray et al., 2013). Two enterococcal species, E. faecium and E. faecalis, are commonly found among the commensal microflora in the human gut and typically do not cause disease. However, they can be important agents of human disease, especially when they infect extra-intestinal sites. The first disease associated with Enterococcus was a case of infective endocarditis in 1899 (MacCallum, 1899). Enterococci are currently the third most common nosocomial pathogen (12% of all hospital infections) causing urinary tract and wound infections, infective endocarditis, endophthalmitis, and peritonitis, which are often complicated by antibiotic drug resistance (Fischetti and American Society for Microbiology., 2006). The majority of human enterococcal infections are caused by E. faecalis (80–90%), with E. faecium comprising most of the remainder (10–15%) (Jett et al., 1994; Johnson, 1994). It is important to understand the biology of enterococcal infections to combat the rising rates of nosocomial infections. Within the past 25 years, many important insights have been gained into enterococcal infection biology from the viewpoint of host-pathogen interactions. In this review, we will focus on the insights gleaned from invertebrate infection models that have revealed how enterococcal species have become such a successful opportunistic, nosocomial pathogen, as well as the mechanisms by which the host innate immune system defends itself against enterococcal infection.

Rationale for studying host-pathogen interactions in invertebrate models

Understanding the relationship between enterococcal virulence mechanisms and the host response has helped to form an integrated view of enterococcal pathogenesis. Given that multiple interconnected factors with complex inter-relationships between both the host and the microbe are at play (Table 1), the use of a genetically tractable host and pathogen resulting in a dual genetic model allows one to deconstruct host-pathogen interactions in greater detail than is possible in vertebrate models.

Table 1.

Facets of host immunity and pathogen virulence

Host	Pathogen
Innate immunity ○ O Sensors ■ Pattern recognition ■ Altered host/damage recognition ○ O Effectors ■ Antimicrobial peptides ■ Lysozyme ○ O Phagocytosis ○ O Sequestration of pathogen	Quorum sensing Toxic mediators ○ O Matrix-degrading enzymes ○ O Cytolysin ○ O Reactive oxygen species Surface adhesion molecules Biofilm Persistence in environment and host
Adaptive immunity ○ O Humoral immunity (antibody-mediated) ○ O Cellular immunity

Invertebrate infection models have revealed important insights into host-pathogen interactions, including guiding principles underlying both the infectious process and the host immune response that are relevant to mammalian infections (Casadevall, 2005; Lemaitre et al., 2012; Vogel, 2012). The ability to use simple invertebrate hosts to study innate immunity and host-pathogen interactions is enabled by the evolutionary conservation of both ancient innate host defenses and bacterial virulence mechanisms. In general, the shared use of virulence factors by a pathogen infecting different hosts suggests that the mechanisms of pathogenesis are often host-independent (Waterfield et al., 2004; Irazoqui et al., 2010). Similarly, if a particular pathogen activates similar defense responses in both invertebrate and vertebrate hosts, it suggests that the underlying mechanisms of innate immunity against the pathogen may be conserved across metazoans. Thus, while the specific immune responses activated in a model host are likely to be most representative of phylogenetically related species, the categories of responses and the signaling pathways involved may be more highly conserved and may reflect common features of metazoan infection biology (Irazoqui et al., 2010).

Two invertebrate organisms have been used most successfully as model hosts for the study of enterococcal pathogenesis and will be discussed at length in this review: the greater wax moth Galleria mellonella and the free-living bacteriovorous nematode Caenorhabditis elegans (Figure 1). Compared to C. elegans, G. mellonella has a relatively complex innate immune system that may correlate better with observations in the mammalian host. On the other hand, C. elegans is genetically tractable with many genetic and genomic tools are available, as well as an extensive worldwide community of scientists focused on all aspects of C. elegans biology. This facilitates the discovery of novel features of the host-pathogen interaction. The advantages and disadvantages of these two infection models will be discussed at length in this review. While a few preliminary studies of enterococcal infection have also been carried out using Drosophila melanogaster as a model host (Cox and Gilmore, 2007; Teixeira et al., 2013), there has been very limited follow-up on these experiments.

Figure 1.

Key features of two invertebrate models

Invertebrate model systems of enterococcal infection

Galleria mellonella as a model host

Galleria mellonella (hereafter referred to as Galleria) larvae have been used to identify and characterize bacterial virulence factors that also cause disease in mammals, including humans, in part because of a good correlation between the role of bacterial virulence factors in the Galleria infection model and mice (Jander et al., 2000). Studies using Galleria have identified several novel enterococcal virulence factors. While Galleria lacks a fully sequenced genome and genetic tractability, the Galleria infection model has been successfully used to biochemically identify enterococcal virulence factors in the insect hemolymph, as well as for studying novel protective responses in the host. The Galleria infection model of Enterococcus infection is also biologically relevant. The predominant organism colonizing the intestines of wild-caught Galleria larvae is the enterococcal species E. hirae, which is thought to have the potential to escape from the intestinal lumen to the body cavity during the larval to pupal transition, thereby inducing the host defenses of Galleria (Bucher and Williams, 1967; Dunn et al., 1994)

There are three main ways in which Galleria combat bacterial infections. First, circulating phagocytic hemocytes patrol the hemolymph, primed to recognize, engulf and sequester invading microbes. Second, upon detection of the pathogen within the inner cavity of the insect, proteolytic cascades can be quickly triggered, activating the melanization response (the synthesis and deposition of melanin to sequester pathogens at a wound site), followed by hemolymph coagulation, and opsonization (Tang, 2009). The melanization response of insects is analogous to the formation of an abscess in mammals following an infection, in which neutrophils accumulate within a tissue to prevent the spreading of the pathogen. Third, Galleria is able to induce a menagerie of antimicrobial immune effectors such as lysozymes, as well as antimicrobial peptides, such as cecropins and moricins, which can be rapidly synthesized by the fat body (Goldsmith and Marec, 2010). The insect larval midgut is the functional equivalent of the mammalian intestine, as it is not only a major part of the digestive tract, but also regulates metabolism and the immune response (Apidianakis and Rahme, 2011).

An added benefit of using Galleria for pathogenesis studies is that infections can be carried out at 37°C or higher, as Galleria tolerates relatively high temperatures, unlike both D. melanogaster and C. elegans (maximum 25°C) (Glavis-Bloom et al., 2012). This allows Enterococcus to grow optimally, and at human body temperature, for experimental studies. The larger size of the Galleria larva, compared to other invertebrate models, also allows it to be infected with larger, more controlled doses of the pathogen. In contrast, infection of C. elegans is carried out by allowing the animals to feed ad libitum on the pathogen, which does not control for dosage.

C. elegans as a model host

C. elegans is an excellent, although relatively new, model for studying pathogenesis and immunity. In the wild, C. elegans inhabits decaying vegetation and fruit and thus encounters a diverse microcommunity, rich in microbes (Felix and Braendle, 2010). As their environment naturally contains noxious bacteria, C. elegans must recognize the presence of the pathogenic bacteria in their food and environment in order to mount effective immune and detoxification responses.

The simplicity of the small roundworm C. elegans has allowed it to become an attractive system for the study of innate immunity. The adult C. elegans worm has 20 non-renewing intestinal epithelial cells (IECs) that share many common morphological features with mammalian IECs, including a “brush border” of microvilli, anchored into a cellular structure called a terminal web (Irazoqui et al., 2010). As the worm has no motile immune cells, it is dependent on these IECs to mediate intestinal immunity. By replacing non-pathogenic E. coli, the normal laboratory food source for C. elegans, with a pathogenic microbe, one can activate the worm innate immune system. Its genetic tractability easily allows for thorough analysis of genetic pathways, especially using whole-genome RNAi screens (Kamath et al., 2001; Kamath and Ahringer, 2003).

The small size of C. elegans and its ability to survive in liquid culture makes it amenable to cost-effective and rapid scaling-up for biochemical analysis (Walhout, 2006), as well as for high-throughput screening of chemical libraries in whole-animals – a platform that would be impractical to implement in the case of Galleria larva (or even Drosophila), let alone vertebrates, due to their size (Moy et al., 2006; Moy et al., 2009). Additionally, testing the effect of small volumes of rare natural compounds on infection biology is only currently possible using the worm infection model.

Mechanisms of enterococcal virulence

There are a number of factors that contribute to the success of E. faecalis in establishing nosocomial infections (Table 2). In addition to its ability to readily acquire antibiotic resistance and persist in harsh environments, Enterococci also have a substantial toolkit of virulence factors (Kayaoglu and Orstavik, 2004; Fisher and Phillips, 2009; Sava et al., 2010; Arias and Murray, 2012). In the sections below, we discuss the virulence factors and virulence-related processes that Enterococcus uses to establish an infection and the role that these factors play in pathogenesis in Galleria and C. elegans (Figure 2).

Table 2.

Enterococcal virulence factors

Secreted factors	Description
GelE	Gelatinase; matrix metalloproteinase of broad proteolytic activity
SprE	Serine protease; homology to S. aureus
Cytolysin	Member of lantibiotic class of bacteriocins; able to lyse prokaryotic and eukaryotic cells
Surface adhesion molecules
Esp	Enterococcal surface protein; adhesion to epithelium, biofilm formation and conjugation
Ace	Adhesin to collagen
Fss1-3	Adhesin to fibrinogen
Ebp	Adhesin to fibrinogen
Acm	Adhesin to collagen
Aggregation substance	Group of proteins encoded on pheromone-response, conjugative plasmids; role in transfer of antibiotic resistance genes and extracellular matrix proteins
Other toxic mediators
Extracellular ROS	Superoxide and hydrogen peroxide; can induce oxidative stress in nearby host cells

Figure 2. Mechanisms of enterococcal virulence.

A. Enterococci produce a number of enzymes (e.g., hyaluronidase, gelatinase, and serine proteases) that degrade the host extracellular matrix, which allow the bacteria to invade host tissues.

B. E. faecalis expresses several adhesins that allow it to bind the extracellular matrix and cell surface of the host. Aggregation substance (AS) facilitates the adherence of E. faecalis to extracellular matrix proteins. Ace and Fss are bacterial adhesins that bind collagen and fibrin, respectively.

C. Cross-sections of C. elegans animals showing the intestinal lumen. After infection with E. faecalis, the C. elegans gut becomes distended with live enterococcal cells, forming a persistent infection in the worm (left). Most enterococcal cells are intact, with some even undergoing fission. In contrast, nearly all E. coli are degraded by the C. elegans pharyngeal grinder and intestine (right). Microvilli (mv) and intestinal epithelial cell (IEC) are labeled.

D. Enterococcal gelatinase GelE aids in evading the immune system of both G. mellonella and humans. GelE degrades cecropin, an insect antimicrobial peptide that is induced early in infection. Additionally, GelE is able to hydrolyze human C3a and degrade C3b, two complement proteins generated by C3 activation. This inhibits opsonization, as well as the formation of the membrane attack complex, two important roles of the complement cascade.

Quorum sensing

Quorum sensing is the means by which bacteria are able to sense a species- or population-specific cell density via the production and secretion of small signal molecules. Quorum sensing can ultimately activate signaling transduction and thus modulate physiological behavior within the bacteria (Podbielski and Kreikemeyer, 2004). In Enterococci, quorum-sensing regulation is driven by the Fsr regulon, which is homologous to the Agr system of staphylococci, perhaps the most well-described cyclic peptide-mediated quorum-sensing system among Gram-positive bacteria. The Agr quorum sensing system (Lyon and Novick, 2004; Kong et al., 2006; Novick and Geisinger, 2008; Gospodarek et al., 2009) and, to some extent, the Fsr system (Carniol and Gilmore, 2004; Cook and Federle, 2013), have been discussed in several reviews. Briefly, the fsr quorum sensing system is comprised of the four genes fsrA (a response regulator), fsrB (a processing enzyme), fsrC (a sensor histidine kinase), and fsrD (a pro-peptide of the autoinducer) (Nakayama et al., 2001). Following its production, the pro-peptide FsrD is processed and cyclized by FsrB to yield gelatinase biosynthesis activating pheromone (GBAP), an 11-amino acid cyclic peptide containing a lactone linkage (Nakayama et al., 2006). GBAP then interacts with sensor kinase FsrC, which leads to the phosphorylation of regulator FsrA, thereby inducing the transcription of not only the fsrBC operon, but also two nearby genes encoding secreted products, the serine protease SprE and the zinc-metalloprotease gelatinase GelE. The Fsr two-component regulatory system is a major regulator of virulence in E. faecalis, and disruption of the fsr operon impairs the transcription of the two downstream proteases (SprE and GelE), as well as the development of biofilms (Hancock and Perego, 2004; Pillai et al., 2004). E. faecalis fsrB mutants are attenuated in virulence in both the C. elegans and Galleria infection models, as well as in mammalian infection models (mouse peritonitis and rabbit endophthalmitis) (Singh et al., 2005; Gaspar et al., 2009).

Does the host interfere with quorum sensing, and if so, how does the host counteract its effects? Because quorum sensing is a feature of live and active bacteria, It has been proposed that quorum sensing may be targeted by the host immune system, which may try to recognize and inhibit the quorum sensing molecules (Vance et al., 2009). In one study, a specific acyl homoserine lactone produced by P. aeruginosa (3OC12-HSL), but not its chemically-related counterparts, was found to induce the chemotaxis of human polymorphonuclear neutrophils (PMNs) via a host receptor on the PMNs that directly binds 3OC12-HSL (Zimmermann et al., 2006). It is possible that the immune system has been selected to recognize these quorum-sensing auto-inducers as they are released from bacteria during the initial phases of biofilm formation at a time in the infection process when it may still be early enough for innate immune cells to stave off the infection. It has also been reported that C. elegans recognizes P. aeruginosa quorum sensing molecules, which results in aversive feeding behavior (Beale et al., 2006). There may be a similar, corresponding interaction between enterococcal quorum sensing molecules and the host immune system. While there are no reports that this has been observed in Galleria or C. elegans, this interaction may be worthwhile to explore in the future.

Toxic Mediators

Matrix-degrading enzymes

Proteases secreted by bacteria that degrade host tissues serve to provide peptide nutrients for the bacteria. In the case of bacterial pathogens, the secretion of proteases is often concomitant with pathogenesis. Bacterial proteases damage the host through multiple mechanisms: direct destruction of both cells and the extracellular matrix in the host tissue, indirect damage by proteolytic activation of bacterial matrix metalloproteinases or host proteolytic cascades (Sorsa et al., 1992), the processing of pathogenic virulence factors, and inhibition of host immunity (Lantz, 1997).

Gelatinase (GelE) is a secreted zinc metalloproteinase that allows Enterococcus to degrade and invade host tissues by degrading the collagenous matrix and that can interfere with host inflammatory processes by hydrolyzing a broad range of substrates including fibrinogen, fibrin, endothelin-1, bradykinin, human cathelicidin (LL-37), and complement components C3 and C3a (Makinen et al., 1989; Makinen and Makinen, 1994; Schmidtchen et al., 2002; Waters et al., 2003; Park et al., 2007). GelE is also critical for biofilm development in E. faecalis, as it activates a peptidoglycan-degrading enzyme called autolysin, which is responsible for the release of extracellular DNA and the formation of biofilm (Thomas et al., 2009). The formation of biofilm allows Enterococcus to adhere to and survive on urethral catheters, urethral stents, and cardiac valves, and enhances its survival in an antibiotic-rich environment (Reid et al., 1992; Donlan and Costerton, 2002; Sabbuba et al., 2002; Tunney and Gorman, 2002; Carniol and Gilmore, 2004). Injection of purified GelE was found to be toxic to Galleria in a dose-dependent manner. E. faecalis gelE mutants are also attenuated in virulence in Galleria. Although the insect does not contain any of the canonical targets of GelE such as collagen, GelE acts on other targets in Galleria. For example, GelE was shown to directly hydrolyze a cecropin-like antimicrobial peptide in Galleria, analogous to its ability to degrade immune mediators in the human serum such as the complement component C3a. As C3a and C3a-derived peptides have broad antibacterial properties, the degradation of C3a by GelE may serve as an efficient way to cripple immune responses downstream of C3a. GelE is thus capable of subverting and inhibiting the immune system in both an insect and a mammalian infection model (Park et al., 2007). In contrast, although injection of purified enterococcal extracellular serine protease SprE does not cause any obvious damage to Galleria, SprE was shown to be necessary for the full virulence of E. faecalis in killing in C. elegans, as a sprE mutant was attenuated in virulence in the C. elegans infection model (Sifri et al., 2002).

Like gelatinase, hyaluronidase is also a degradative enzyme; it hydrolyzes hyaluronic acid, contributing to host tissue damage and invasiveness of the infection. It is thought that disruption of host tissue may also allow other secreted bacterial factors to permeate the tissue, magnifying the damage (Takao et al., 1997). The breakdown of host tissue products by hyaluronidase may also allow Enterococcus to access and metabolize the nutrients of its host (Toto et al., 1968; Hynes and Walton, 2000).

Cytolysin

The enterococcal cytolysin, a secreted two-peptide lytic toxin, is related to the extensive group of lanthionine-containing bacteriocins produced by Gram-positive bacteria and is capable of lysing both prokaryotic and eukaryotic cells in response to quorum sensing signals (Coburn and Gilmore, 2003). The molecular properties of enterococcal cytolysin and its regulation have been well studied (Coburn and Gilmore, 2003; Cox et al., 2005; Van Tyne et al., 2013). One retrospective study found that of nearly 200 clinical E. faecalis isolates, 45% expressed cytolysin, which is encoded on a pheromone-responsive plasmid on a pathogenicity island (Huycke et al., 1991). The presence of cytolysin in a human infection has been reported to render the infection five times more acutely lethal (Huycke et al., 1991). In the wild, Enterococci have been isolated from Galleria larvae, and some of these isolates have been observed to exhibit lytic activity (Jarosz, 1975; Jarosz, 1979). In the C. elegans infection model (Garsin et al., 2001), as well as in a Drosophila oral ingestion model (Cox and Gilmore, 2007), noncytolytic E. faecalis strains were attenuated in virulence, although they still remained lethal to C. elegans.

Surface adhesion proteins

Enterococcal surface protein (Esp) is a surface adhesin that controls adherence to host tissues, some of which have been implicated in endocarditis. Following host heart tissue damage, Enterococcus can use its surface adhesins to adhere to the exposed extracellular matrix of injured tissue (Heikens et al., 2011). The esp gene encodes a large LPxTG-anchored surface protein, which contributes to biofilm formation on abiotic surfaces. One study identified a putative surface antigen structurally related to Esp in vancomycin-resistant E. faecalis strain V583 called EF3314. Mutations in EF3314 resulted in attenuated virulence in the C. elegans infection model, suggesting that this surface antigen may have targets in both mammals and nematodes (Creti et al., 2009).

Another potential E. faecalis adhesion is the protein Ace, a collagen- and laminin-binding microbial surface component recognizing adhesive matrix molecules (MSCRAMM). Ace-specific antibodies have been found in the sera of patients with enterococcal infections, and in particular, those patients with E. faecalis endocarditis, implying that the environment is an efficient trigger of Ace expression and that the expression of Ace is quite common (Nallapareddy et al., 2000). Several in vivo environmental triggers, including high temperatures and the presence of collagen or bile salts, can trigger transcription of ace. However, the first indication that Ace has a role in addition to its binding the extracellular matrix, came from Galleria infection studies, where ace mutants were found to be attenuated in virulence (Lebreton et al., 2009). As collagen and laminin are both absent in insects, it is possible that Ace has other targets in Galleria, and perhaps in mammals as well.

Enterococcal aggregation substance (AS) is a pheromone-responsive, surface-bound, plasmid-encoded protein that promotes clumping of bacterial cells to facilitate plasmid exchange, as well as adherence to intestinal epithelial cells by specifically binding extracellular matrix proteins. Additionally, AS allows E. faecalis to directly bind human neutrophils through complement receptor type 3 (CR3) binding (Vanek et al., 1999). Despite the close proximity of the bacteria to the neutrophils, AS-expressing E. faecalis strains are resistant to neutrophil killing (Rakita et al., 1999). Indeed, the oxidative burst (superoxide production) from the neutrophil may be responsible for causing collateral damage to neighboring tissues, without seriously damaging the adherent E. faecalis cells. Interestingly, E. faecalis AS is not a virulence factor in C. elegans killing (Garsin et al., 2001), suggesting that the host target of AS may be absent.

Novel concepts in host-pathogen interactions revealed by invertebrate studies

Stress and its relation to virulence

Bacterial virulence is closely related to the ability of pathogens to adapt to different stresses. As bacteria traverse between the outside environment and the host during the course of the free-living and infection stages of their life cycles, they encounter myriad stressors, such as changing pH, temperature, and osmolarity, as well as the presence of reactive oxygen species. To establish a successful infection, a pathogen must be able to rapidly adapt to a changing environment, which is accompanied by many changes in gene expression levels of regulators and effectors of the stress response. These stress response genes are often required for bacterial virulence and have been explored in the Galleria model. Mutants corresponding to several stress-response genes were found to be attenuated in virulence, including the clpB gene, which encodes a Clp ATPase that is critical for promoting protein folding, as well as the assembly and degradation of proteins (de Oliveira et al., 2011). Mutations in a second stress response gene that encodes methionine sulfoxide reductase A also results in attenuated Galleria killing, suggesting the importance of antioxidant repair enzymes in virulence (Zhao et al., 2010). Surprisingly, however, a deletion mutation in the E. faecalis slyA gene, which encodes a member of the MarR/SlyA family of bacterial transcription factors that is predicted to be a regulator of stress responses, was shown to enhance virulence in Galleria, an unexpected result for a transcriptional activator (Michaux et al., 2011). The enhanced virulence of the ΔslyA mutant in Galleria is consistent with increased persistence in macrophage assays and in vivo mouse infections. Given these data, it is possible that a small subset of genes that are overexpressed in the ΔslyA mutant may be responsible for the enhanced virulence phenotype.

The “stringent response” is a bacterial response to nutritional stress (e.g. amino acid starvation) and is an important component of enterococcal virulence. During the bacterial stringent response, two modified guanine nucleotides, guanosine tetraphosphate (ppGpp) and guanosine pentaphosphate (pppGpp), together known as (p)ppGpp, accumulate in the bacterial cell and act as “alarmones” or chemical messengers indicating stress, shutting down global transcription and selectively activating the transcription of genes regulating amino acid biosynthesis and stress-related survival. In Enterococcus, (p)ppGpp synthesis is catalyzed by two enzymes, RelA (the bifunctional synthetase/hydrolase of the alarmone) and RelQ (the monofunctional synthetase responsible for basal levels of alarmone during homeostasis). The alarmone (p)ppGpp was found to accumulate during vancomycin treatment and amino acid starvation. While the E. faecalis relA and relQ single mutants do not show attenuated virulence in the C. elegans infection model, a relAQ double mutant, in which (p)ppGpp production is completed abrogated, was highly attenuated in its virulence (Abranches et al., 2009). A similar result was obtained in Galleria (Gaca et al., 2012), and relAQ deficient strains also exhibited lower survival rates in macrophages in vitro than either the relA or relQ singly-deficient strains. These data support the model that basal levels of (p)ppGpp are required for full E. faecalis virulence in both C. elegans and G. mellonella as well as in mammalian cells. It is possible that the host innate immune system may recognize (p)ppGpp: the immune system is able to recognize the bacterial secondary messenger cyclic di-GMP as a MAMP (Karaolis et al., 2007; McWhirter et al., 2009; Yin et al., 2012), which, like (p)ppGpp, is also derived from the modification of a nucleotide.

Environmental persistence and biofilm production

The ability of Enterococcus to survive in a multitude of environments, from the human gut to abiotic surfaces in hospitals, reflects its hardiness as a pathogen and capacity for environmental persistence. In addition to being thermo-tolerant (growing at 10–45°C), pH-tolerant (pH 4.8–9.6) and salt-resistant (up to 28% NaCl), Enterococcus is also able to survive extended desiccation (Gilmore, 2002), which is unusual for a bacterium that is incapable of sporulation. This may explain how Enterococcus is transferred from person to person in the hospital. Enterococcus may be able to enter a viable but non-culturable state when faced with nutrient-poor environments and may exist in this form on inanimate objects, such as bedrails and medical station keyboards (Lleo et al., 1998; Giard et al., 2001; Heim et al., 2002). These intrinsic properties allow Enterococcus to flourish in a variety of environments, and may have been evolutionarily selected for by its need to survive in harsh environments, such as the intestinal tracts of animals.

The ability of Enterococcus to form a biofilm also allows Enterococcus to persist on abiotic surfaces as well as within the host. Biofilm protects pathogenic bacteria from the host immune response, as well as antibiotics and antimicrobials, which leads to chronic enterococal infections that are difficult to treat (Lewis, 2001). Two potential E. faecalis regulators of biofilms may also control virulence in the C. elegans infection model. Whole-organism screening using the C. elegans infection model has helped identify novel regulators of enterococcal biofilm. In screens for E. faecalis transposon mutants that were attenuated in killing C. elegans, two related, potential virulence factors were identified: the transcriptional repressor ScrR, and a gene that is highly homologous to sucrose-6-phosphate hydrolases (ScrB) of other bacterial species and likely regulated by ScrR. This was intriguing, as sucrose utilization has been demonstrated to have a role in the formation of biofilms in dental caries and endocarditis by Streptococcus mutans, in which the sucrose catabolites are used to synthesize insoluble glucans to promote adherence to the tooth surface and heart valves (Garsin et al., 2001; Maadani et al., 2007).

The oral E. faecalis isolate, OG1RF, is devoid of many genes related to pathogenicity and is comparatively avirulent in humans. V583 is a virulent strain and was also the first sequenced vancomycin-resistant E. faecalis strain. The genome of V583 contains 25% mobile elements, including a pathogenicity-associated island (Shankar et al., 2002; Paulsen et al., 2003; Shankar et al., 2006). However, both OG1RF and V583 have comparable virulence in the C. elegans infection model, even though OG1RF is believed to lack many genes thought to be required for virulence in humans (Garsin et al., 2001; Bourgogne et al., 2008). One possibility is that the genes required for virulence in humans are not necessary to render Enterococcus pathogenic in C. elegans, whereas E. faecalis genes that promote commensal enterococcal colonization of the human gut in healthy individuals (seen in both OG1RF and V583) are sufficient to cause virulence in C. elegans infection. This latter conclusion is supported by studies of enterococcal metabolism. Enterococcus has an impressive flexibility to catabolize a smorgasbord of carbon sources (Deibel, 1964a; Farrow et al., 1983; Devriese et al., 1993). Enterococcus is also able to use certain amino acids as energy and carbon sources (Deibel, 1964b; Roon and Barker, 1972). Concomitantly, and almost paradoxically, Enterococcus also has a number of auxotrophies, requiring a number of amino acids (e.g., valine, leucine, histidine, and tryptophan), as well as vitamins (e.g., biotin, riboflavin and nicotinic acid) (Murray et al., 1993; Gilmore, 2002). These features of enterococcal metabolism are conserved across both virulent and avirulent E. faecalis strains and may have been selected for by competition as a commensal organism in the mammalian GI tract (Vebo et al., 2010).

Why are E. faecalis infections more common than E. faecium infections? One reason may be that E. faecalis is simply more abundant in the gastrointestinal tract than E. faecium; indeed, one study found that E. faecalis was on average 100 times more prevalent than E. faecium in the gut (Noble, 1978). An interesting observation from enterococcal infection of C. elegans may hint at a reason for dominance of E. faecalis. Infection with E. faecalis or E. faecium rapidly distends the C. elegans intestine, which fills up with live enterococcal cells. However, only E. faecalis results in a persistent and lethal infection in the worm, even with a small inoculum of E. faecalis bacterial cells. This may suggest that E. faecalis has an inherent ability to establish a persistent infection, perhaps contributing to its virulence, but the unique characteristics of E. faecalis that allow it to form a persistent infection remain unknown. There does not appear to be a single virulence factor that contributes to the bulk of E. faecalis pathogenicity, as E. faecalis transposon mutant library screens have not picked up any single mutant that renders E. faecalis completely attenuated. Although the mechanism for E. faecalis persistence in C. elegans is unknown, we speculate that E. faecalis may be better at competing for and establishing a protected niche for itself in the C. elegans intestine.

Innate immune perception of enterococcal infection

The mammalian innate immune response involves the perception of a bacterial infection by pattern recognition receptors (PRRs), which recognize conserved microbial structures (e.g., flagellin, lipopolysaccharide, unmethylated CpG motifs) as pathogen- or microbe-associated molecular patterns (M/PAMPs). The detection of bacterial M/PAMPs by host PRRs activates downstream signaling pathways that culminate in the production of immune effectors that inhibit the infection (Medzhitov, 2001). However, the localized context of infection is likely to be just as important for innate immune recognition of a pathogen, since M/PAMPs are present on non-pathogenic bacteria as well as pathogenic bacteria. Additionally, M/PAMPs are often thought to be invariant bacterial features, but pathogens are often able to vary their M/PAMPs, such as in the case of the many variants of LPS among different bacteria. It has been proposed that the immune system may respond to M/PAMPs in the context of other signals, derived from either the host or the pathogen, that are only generated during a pathogenic infection, called patterns of pathogenesis (Vance et al., 2009).

These principles also apply to the innate immune system of invertebrates. In the case of C. elegans, several immune signaling pathways have been identified, although no specific M/PAMPs that activate C. elegans immunity have been discovered. Chemical genetics may be one approach to identifying novel and potentially conserved immune signaling pathways in C. elegans, which may ultimately lead to the identification of the most upstream innate immune receptors in C. elegans, and potentially, the identification of specific bacterial M/PAMPs recognized by such receptors. One chemical genetics study employed a liquid-based, high-throughput screen of small molecules and extracts to identify compounds that promoted the survival of E. faecalis-infected worms (Moy et al., 2006; Moy et al., 2009). One compound selectively acted upon the C. elegans host via the activation of the well-studied and conserved p38 MAP kinase, PMK-1 (Pukkila-Worley et al., 2012). However, the precise biological target of this molecule is unknown. It is possible that this activating compound is modified in the host to generate a molecule resembling a pattern of pathogenesis or may activate signals that are also transduced by receptors that recognize patterns of pathogenesis. In addition to identifying novel immune activators, this chemical genetics approach can also be used to identify antimicrobials (which inhibit the growth of or kill bacteria) and inhibitors of bacterial virulence (which may not have an effect on bacterial survival, but may modulate bacterial physiologic behavior).

The DAF-2/DAF-16 (homologous to IGF-1 receptor and FOXO, respectively) insulin signaling pathway is also involved in the C. elegans defense against Enterococcus. When the DAF-2 receptor is prevented from inhibiting DAF-16, genes involved in regulating lifespan, dauer formation, and pathogen and stress resistance are activated, leading to enhanced resistance to E. faecalis and other pathogens (Garsin et al., 2003; Murphy et al., 2003). A number of DAF-16-regulated genes are necessary for the neutralization and detoxification of ROS, such as superoxide dismutases (which reduce superoxide to hydrogen peroxide) and catalases (which convert hydrogen peroxide to water and oxygen) (Chavez et al., 2007). This suggested that countering ROS is crucial component of an effective immune response; however, it was unknown whether the ROS was primarily generated by E. faecalis, or the host in response to pathogenic infection. It was later shown that C. elegans host intestinal cells, following infection by E. faecalis, generate extracellular ROS via the NADPH oxidase BLI-3, the C. elegans homolog of mammalian dual oxidase Duox1, which contributed to pathogen resistance (Chavez et al., 2007). The production of ROS, in tandem with the infection, activates a p38 MAPK PMK-1 cascade to induce the Nrf-family transcription factor SKN-1, which in turn upregulates stress and antioxidant responses during infection to protect against the collateral damage inflicted by endogenous ROS production (Hoeven et al., 2011; Papp et al., 2012). This response may serve as an ancient form of immunity, much like the oxidative burst observed in macrophages and neutrophils. It remains unknown, however, by what mechanism the dual oxidase is activated to eventually activate the immune response.

How is Enterococcus infection perceived by the C. elegans host immune system? Is it through recognition of conserved structural motifs specific to pathogens, or by the recognition of perturbed cellular processes in the host? There is increasing evidence for the latter (Dunbar et al., 2012; McEwan et al., 2012; Melo and Ruvkun, 2012). Host innate immune defense appears to be closely related to surveillance pathways that monitor core cellular activities (e.g., organelle function, protein translation, etc.), which may also allow the host to perceive the presence of pathogens that produce virulence factors that subvert fundamental host functions.

E. faecium and its opportunistic traits

In comparison with other enterococcal species, strains of E. durans and E. faecium are either avirulent or weakly virulent in G. mellonella compared to E. faecalis. This intriguing interspecies difference in pathogenesis has also been observed in the C. elegans infection model, and may imply the presence of E. faecalis-specific virulence factors.

It is not always the case, however, that E. faecium is avirulent. In studies with C. elegans, E. faecium that is grown anaerobically and then exposed to aerobic conditions, produces hydrogen peroxide, which is able to kill C. elegans rapidly without a few hours. The explanation behind this phenomenon is related to E. faecium being a facultative anaerobe relying on glycolysis and fermentation to generate energy. In E. faecium, lactic acid fermentation is favored under anaerobic conditions. However, under aerobic conditions, mixed acid fermentation takes place, where NADH oxidases use molecular oxygen as an electron acceptor to regenerate NAD+ from NADH. In turn, in these aerobic conditions, molecular oxygen is reduced to hydrogen peroxide. The NADH peroxidase, which scavenges the hydrogen peroxide, is only expressed under aerobic conditions. Thus, the immediate exposure of anaerobically grown E. faecium to aerobic conditions allows for the production and accumulation of hydrogen peroxide, with levels of NADH peroxidase that are too low to scavenge efficiently (Moy et al., 2004). It was also found that hydrogen peroxide production increased as aerobically-grown E. faecium entered the stationary phase. Taken together, the production of hydrogen peroxide may be one mechanism by which E. faecium is able to impede or kill its fellow competing bacteria, as well as damage host tissue or immune cells via oxidative stress.

Oxidative stress may be a key signal used by E. faecium to promote opportunistic traits. The E. faecium AsrR protein was identified as an oxidative stress sensor that uses cysteine oxidation to sense hydrogen peroxide. AsrR and its putative homologs are present in E. faecium, E. gallinarum and E. casseliflavus, but absent in E. faecalis. AsrR was found to regulate 181 genes related to many features of pathogenesis aside from oxidative stress, including resistance to antibiotics and antimicrobial peptides (Lebreton et al., 2012). Two important adhesion-encoding genes, acm and ecbA, were also found to be upregulated, and contributed to the increased persistence and colonization observed in Galleria infection, which was supported further by the observed increased adhesion to human intestinal cells and biofilm formation in two different in vitro assays.

Conclusions

There remain several unanswered questions in Enterococcus infection biology, and many of these questions are especially tractable using invertebrate model systems, which often allow for the genetic dissection on both the host and pathogen fronts. The use of invertebrate infection models will prove especially useful in recapitulating facets of the Enterococcus infection that involve the interaction between Enterococcus and host tissues, rather than the interaction between Enterococcus and individual cells (e.g., macrophages), which can be recapitulated in cell culture. By virtue of their small size, the C. elegans allows high throughput genetic and chemical screens in whole-animal infections for novel virulence factors as well as anti-infective compounds. In addition, C. elegans is especially genetically tractable and thus can be used for forward- and reverse-genetic screens to identify host factors required for defense against Enterococcus. However, genetic tools are rapidly improving, and it will not be long before Galleria can also be manipulated genetically, perhaps through the use of the CRISPR-Cas system (Ran et al., 2013a; Ran et al., 2013b) and other genome editing tools. The use of bacterial genetics tools such as Tn-seq (van Opijnen and Camilli, 2010), which utilizes next-generation sequencing for identifying transposon insertion mutants of interest, will also help elucidate the unique features of Enterococcus virulence such as its ability to induce a persistent infection.