Enterococcus Species, Streptococcus gallolyticus Group, and Leuconostoc Species

Revised January 7, 2021

Historical Background

The first time that the term entérocoque was used appears to have been in an article in the French literature in 1899. The article was referring to a diplococcus found in the gastrointestinal (GI) tract that had the potential to become pathogenic for humans. The first clinical and pathologic description of an enterococcal infection was published the same year (1899) and concerned a patient admitted to “Dr. Osler's Service” with a clinical picture of endocarditis who succumbed to the infection. The authors isolated gram-positive cocci in “pairs and short chains” from the patient's blood and several organs (postmortem). The virulence properties of the organism were confirmed after inoculating it into several animal models and reproducing the pathologic findings observed in the patient. This organism was initially designated Micrococcus zymogenes because of its fermentative properties. In 1906 Andrews and Horder described in detail a study of streptococci pathogenic for humans and used the name Streptococcus faecalis for the first time to denote the most common species of streptococci present in the intestine of humans and other vertebrates. They referred to previous environmental experiments performed by themselves and Houston in London, indicating that the most common microorganisms collected from London's air were “intestinal streptococci,” which probably originated from horse dung, “which forms so large of a part of the organic contamination of London's air.”

In December 1937 Sherman, addressing the Society of American Bacteriologists, indicated that “the enterococcus” was a nonspecific term used for streptococci isolated from the gut, which was “a screen behind which the investigator could hide his ignorance of the organisms with which he worked.” Sherman proposed to group the enterococci as S. faecalis, S. faecalis var. hemolyticus, S. faecalis var. zymogenes, S. faecalis var. liquefaciens, and S. durans, all of which displayed common phenotypic characteristics that included growth in the presence of 6.5% sodium chloride, pH 9.6, and high temperature. Studies in the 1940s and 1950s showed that an organism initially identified in 1919 as S. faecium had distinct characteristics that differentiated it from S. faecalis, and in 1970 a formal proposal that the enterococcal streptococci be considered a new genus based on their distinct phenotypic characteristics was put forward. However, it was not until 1984 that Enterococcus was widely confirmed as a separate genus from Streptococcus after nucleic acid hybridization experiments were performed, and genetic tools were subsequently applied to differentiate the varied species of enterococci. More recently, using next-generation sequencing technology and molecular clocks to analyze the genomes of an array of representatives of the genus, the emergence of the ancestors of the modern enterococci was estimated to coincide with the advent of terrestrial land animals some 400 million years ago, speaking to the long and successful history of this durable commensal and opportunistic pathogen.

Microbiology and Taxonomy

Organisms belonging to the genus Enterococcus are gram-positive facultatively anaerobic bacteria that usually appear oval in shape and can be seen as single cells, pairs, short chains, or very long chains ( Fig. 200.1 ). They are capable of growing in media containing 6.5% sodium chloride and at temperatures between 10°C and 45°C, and they are able to hydrolyze esculin in the presence of 40% bile salts and produce a leucine aminopeptidase and pyrrolidonyl arylamidase (PYR) (except for E. cecorum, E. columbae, E. pallens, and E. saccharolyticus ). Enterococci are usually α-hemolytic or γ-hemolytic on trypticase soy, 5% sheep blood agar ( Table 200.1 ), whereas some are β-hemolytic (due to the acquisition of a hemolysis/cytolysin gene) on horse, rabbit, or human blood. Most enterococci react with Lancefield group D antisera and some react with group Q antisera, and some of them are motile (e.g., E. casseliflavus and E. gallinarum ). Table 200.2 lists some enterococci spp. isolated from human infections; the vast majority of clinical infections are produced by two species ( E. faecalis and E. faecium ), and clinical laboratories usually do not identify enterococci to the species level. However, in certain clinical scenarios or in epidemiologic studies, it may be important to differentiate between these two species because they appear to differ in their virulence and antibiotic resistance profiles (see later).

Conventional methods to identify enterococci at the species level include manual biochemical differentiation based on several tests (e.g., acid formation and hydrolysis of arginine); nonetheless, because of the laboriousness of this approach, laboratories usually rely on automated methods or rapid biochemical methods such as the analytical profile index system, which appear to be accurate for E. faecalis, but not for other enterococcal species. Because E. faecium is capable of fermenting arabinose, some selective arabinose-containing agars have been used to differentiate E. faecium from other clinically relevant enterococcal species. In addition, several molecular techniques have been developed for species differentiation, although they are not used routinely by clinical laboratories; among the most popular, amplification of ddl genes, amplification of or a probe for the ace gene, and sequencing of the 16S ribosomal RNA (rRNA) gene appear to reliably differentiate the relevant enterococcal species. Automated nucleic acid hybridization–based rapid diagnostic technologies and matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry are now being used more commonly in the clinical microbiology laboratory; these techniques use species-specific sequence variation or protein fragment profiles to quickly identify E. faecalis and E. faecium . Furthermore, resistance to ampicillin usually indicates E. faecium, whereas nonsusceptibility to quinupristin-dalfopristin (Q-D) is usually seen in E. faecalis.

Update: Enterococcal Colonization

Colonization, Virulence, and Genomics

Enterococci have well-adapted mechanisms to survive in the GI tract of humans. The human colonic microbiota comprises approximately 10 commensal bacteria per gram of contents, encompassing more than 100 culturable bacterial species and many more nonculturable species, with a predominance of anaerobes. Enterococci are clearly a minor population in relation to the anaerobic commensals in a normal host; they also appear to have a symbiotic relationship with the immune system and the other bacteria. However, one of the main effects that antibiotics have in the human gut is to alter the dynamics of colonization in favor of enterococci, which are naturally tolerant to a number of antimicrobial compounds (see “ Therapy and Antimicrobial Resistance ” later). Antibiotics that are excreted in the bile or have substantial antianaerobic activity without inhibiting enterococci (e.g., certain cephalosporins) have been shown to dramatically increase colonization of the GI tract by enterococci (e.g., vancomycin-resistant enterococci [VRE]). Moreover, administration of broad-spectrum antibiotics favors VRE colonization of the murine GI tract in part by downregulating the intestinal expression of the antimicrobial peptide RegIIIγ (a bactericidal lectin produced by intestinal epithelial and Paneth cells), which has activity against gram-positive intestinal organisms. This effect appears to be due to the fact that broad-spectrum antibiotics suppress gram-negative organisms (including anaerobes) in the gut that are responsible for the activation of the signals necessary for the production of the RegIIIγ peptide through the lipopolysaccharide present in their outer membranes. Moreover, the dominance of enterococci (particularly VRE) in the GI tract of hospitalized neutropenic patients after the administration of antibiotics has been shown to be a predictor of subsequent bloodstream infection in these patients. Intestinal colonization of VRE also seems to be influenced by the presence of specific bacterial species of the gut anaerobic microbiota. Microbiota containing Barnesiella species conferred resistance to the domination by VRE in mice ; the presence of this species also correlated with decreased GI VRE colonization of patients undergoing allogeneic hematopoietic stem cell transplantation and receiving antibiotics. A consortium of bacteria isolated from ampicillin-treated mice resistant to VRE colonization highlighted the complexity of unraveling the contributions of individual species within the microbiome. Blautia producta, a commensal anaerobe, directly inhibited VRE growth in the mouse GI tract after oral challenge; however, this inhibitory action depended on the presence of Clostridium bolteae . Furthermore, the β-lactamase–producing anaerobes Bacteroides sartorii and Parabacteroides distasonis were also required, presumably to clear inhibitory concentrations of ampicillin from the local environment and allow B. producta to proliferate. Greater understanding of the interplay between the organisms in the microbiome and the host may allow for the therapeutic use of beneficial microbial cocktails to prevent VRE colonization in high-risk patients. Another factor that may also play a role in colonization and overgrowth of enterococci in the gut is increased stomach pH, usually secondary to the administration of proton pump inhibitors, a strategy commonly used in critically ill patients to reduce the incidence of aspiration pneumonitis.

Considerable research has been performed in the investigation of pathogenic determinants that increase the ability of enterococci to cause disease by enhancing their virulence, survival, or colonizing capacity in human hosts. The cytolysin hemolysin is a bacterial toxin, often encoded by pheromone-responsive plasmids, that is capable of lysing eukaryotic (and prokaryotic) cells and has been shown to contribute to E. faecalis virulence. Gelatinase and serine protease are bacterial enzymes that contribute to virulence in E. faecalis by several potential mechanisms that include, among others, (1) facilitation of microbial invasion by altering immunoglobulins or complement molecules; (2) processing of virulence factors to regulate autolysis and release of high-molecular-weight extracellular DNA, a critical component for the development of E. faecalis biofilms ; and (3) degradation of host connective tissues, exposing ligands for bacterial attachment and possibly providing nutrients for the cell. The expression of gelatinase and serine protease genes is regulated by fsr, a two-component quorum-sensing global regulatory system that is similar to the agr system of Staphylococcus aureus. Gls24 of E. faecalis and Gls20/Gls33 of E. faecium are thought to be general stress proteins that have been shown to be important in virulence in both mouse peritonitis and rat endocarditis models. Although their function has not been fully elucidated, they are associated with resistance of enterococci to bile salts, and their homologues in S. aureus have been linked to upregulation of VraSR and PrsA, important members of the cell wall stress response. Using comparative genome analysis, a locus encoding a putative phosphotransferase system was shown to increase the ability of E. faecium to colonize the murine intestinal tract during antibiotic treatment, and this system has been implicated in both biofilm formation and pathogenesis in a rat model of endocarditis.

Cell surface components are important factors in bacterial virulence because they are usually the first molecules to interact with the host tissue or immune system or both. Aggregation substance of E. faecalis is encoded by pheromone-responsive plasmids and assists in a particular type of conjugative plasmid transfer as well as contributing to virulence (i.e., endocarditis). Aggregation substance proteins increase the adherence and internalization of enterococci into several eukaryotic cells (phagocytes, renal cells, intestinal cells, and epithelial cells) and enhance adherence of producing bacteria to serum and extracellular matrix proteins such as fibrin, fibronectin, thrombospondin, vitronectin, and collagen type I. The E. faecalis surface protein (Esp) is another property found in some strains (and its homologue in E. faecium ); it is a protein that appears to function as an adhesin involved in the formation of biofilms in a glucose-dependent manner. Ace (adhesin to collagen of E. faecalis ) belongs to the family designated as MSCRAMMs (microbial surface components recognizing adhesive matrix molecules) and is considered part of the E. faecalis core genome. The Ace protein binds to collagen via what has been called the collagen hug model, in which the protein embraces the collagen molecule after initial docking. A similar protein, Acm, identified in E. faecium, is important in the pathogenesis of endocarditis and implicated in the emergence of E. faecium as an important nosocomial pathogen.

Another surface protein found to be important in enterococcal pathogenesis is ElrA (enterococcal leucine-rich repeat-containing protein). This polypeptide is a member of the WxL family of surface proteins, and deletion of the encoding gene attenuated E. faecalis virulence in a mouse peritonitis model. The characterization of pili on the surface of gram-positive bacteria has been a major step in the understanding of bacterial virulence . In E. faecalis, the presence of pili (Ebp) was demonstrated (see Fig. 200.1 ), and the characterization of ebp genes encoding the pilus subunits led to establishing that these structures play a major role in biofilm formation and fibrinogen adhesion and are important in the pathogenesis of experimental endocarditis and urinary tract infections (UTIs). Genes encoding homologous pilus subunits have also been identified in E. faecium, indicating that the pili are ubiquitous structures of enterococci.

Polysaccharides on bacterial surfaces may be important pathogenic determinants and may affect leukocyte-mediated killing of bacteria. Certain E. faecium strains are resistant to polymorphonuclear cell killing, a characteristic that might be due to a carbohydrate-containing surface moiety. In addition, antibody to a capsular polysaccharide component purified from an E. faecalis strain enhanced phagocytosis and killing of some strains of both E. faecalis and E. faecium; the capsular material may have vaccine potential because a reduction in bacterial numbers recovered from different organs of immunized mice was obtained compared with nonimmunized control subjects. Using rabbit antisera for the typing of E. faecalis, four capsular serotypes (1, 2, 4, and 7) were found to be present in most clinical isolates, and at least two types of gene clusters for the production of polysaccharide have been characterized (designated epa and cps loci ). Passive immunization with antibodies against E. faecalis lipoteichoic acid promoted the clearance of E. faecalis bacteremia in mice. The antibodies bound lipoteichoic acid from other gram-positive organisms and opsonized Staphylocccus epidermidis, S. aureus, and group B streptococci, raising the possibility of using these antibodies for possible vaccine development.

In E. faecium, acquisition of a very large plasmid (approximately 250 kb) by a commensal strain increased the virulence of the organism in a mouse peritonitis model. These virulence plasmids have been highly associated with clinical strains versus commensal isolates. More recently, genes associated with the regulation of oxidative stress have been identified as important in the virulence of E. faecium . One of these regulators is AsrR, which uses cysteine oxidation to sense hydrogen peroxide mediating the activation of many genes potentially involved in pathogenesis (e.g., acm ), antibiotic and antimicrobial peptide resistance, oxidative stress, and adaptive responses. Among the genes regulated by AsrR was pbp5 , encoding a low-affinity penicillin-binding protein 5 (PBP5); deletion of asrR markedly decreased the bactericidal activity of ampicillin and vancomycin and increased the ability of the mutant to form biofilm and persistence in Galleria mellonella and mouse systemic infection models.

Sequencing of enterococcal genomes has been another useful tool to facilitate the understanding of the complicated pathway by which enterococci evolved from commensals to pathogens. Sequencing of the genome of the first vancomycin-resistant E. faecalis strain isolated in the United States (designated V583) indicated that more than one-quarter of its genome is mobile DNA. A pathogenicity island, which is a large genetic element carrying a set of putative virulence-associated genes, a transposon carrying the vanB gene cluster, three plasmids with antibiotic-resistance determinants, and insertion sequences were the most prominent, potentially mobile elements of V583. The pathogenicity island encodes factors that may enable enterococci to gain an advantage in the gut, such as the cytolysin that has antibacterial properties, several surface adhesins, several carbohydrate utilization pathways, and enzymes that may permit colonization of certain areas of the intestine. It has been postulated that this pathogenicity island was acquired by an ancestral E. faecalis clonal strain that evolved to acquire antibiotic resistance determinants, thus becoming equipped to cause problematic infections in humans. Importantly, V583 and other multidrug-resistant strains lack an active CRISPR (clustered regularly interspaced short palindromic repeats)/Cas system, which acts as an immune system protecting the bacterial cell from incoming foreign DNA (i.e., phages). This deficiency has been associated with an increased frequency of mobile DNA elements (including plasmids and transposons), which can carry resistance determinants and may explain their nosocomial predominance. Similarly, sequencing of the entire genome of another E. faecalis strain ( E. faecalis OG1RF) revealed considerable variation in gene content in this species. As opposed to V583, E. faecalis OG1RF lacks plasmids and the pathogenicity island but is still able to cause infection in animal models. This strain has an intact CRISPR/Cas system and lacks the mobile elements typical of V583 but harbors genes predicted to encode proteins involved in adherence, defense against bacteriophages, metabolism of myo -inositol, and novel surface proteins. Recent genome comparisons indicate that there are two ancestrally distinct clades of E. faecium (animal and human commensal) that diverged from each other at least thousands of years ago; a third subclade responsible for most human, health care–associated E. faecium infections is estimated to have evolved from the animal clade approximately 75 years ago, coinciding with the time that antibiotics were introduced in clinical medicine to become successful in the nosocomial environment. This hospital-associated E. faecium clade, which mostly contains clinical and outbreak-associated strains, characteristically contains pbp5-R, the allele encoding the ampicillin-resistant version of PBP5, and insertion sequences, such as IS 16 .

Epidemiology of Enterococcal Infections

The two most common species responsible for the vast majority of enterococcal infections are E. faecalis and E. faecium. The only other species of enterococci known to be responsible for outbreaks and nosocomial spread, albeit rare, is E. gallinarum. Taken together, enterococci were the third leading cause of nosocomial infections in the United States from 2011 to 2014, with VRE now accounting for approximately 30% of enterococcal infections with most VRE isolates being E. faecium (>90%). The first step in the infectious process appears to involve colonization of the GI tract by hospital-associated strains, which may persist for months or years, although direct inoculation onto intravenous or urinary catheters, onto intravenous stop-cock sets, or via thermometers has been reported. The hospital environment can be heavily colonized with VRE, including bed rails, linen, doorknobs, bedpans, urinals, blood pressure cuffs, stethoscopes, and monitoring equipment, among others. Risk factors associated with increased VRE colonization include the presence of immunosuppression or serious comorbid conditions (e.g., diabetes, renal failure, high Acute Physiology and Chronic Health Evaluation [APACHE] score); increased length of hospital stay; residence in a long-term care facility; proximity to another colonized or infected patient, including sharing a room or hospitalization in a room previously occupied by a patient colonized with VRE; and invasive procedures and administration of broad-spectrum antibiotics (e.g., cephalosporins), antianaerobic drugs (e.g., metronidazole), or vancomycin. The hands of health care workers seem to be the most common source of transmission of VRE, and the Society for Health Epidemiology of America has published specific guidelines to curtail this transmission. The organisms are capable of surviving on the hands, gloves, and gowns of health care workers for prolonged periods, and independent risk factors for glove and gown contamination include contact with a colonized patient's catheter or drain, trunk, or lower extremity.

Once a patient becomes colonized with VRE, the risk of developing a subsequent bloodstream infection with the same VRE-colonizing strain appears to increase, although some studies have not found an association. Rates of bloodstream infections in patients colonized with VRE range from 0% to 34% and seem to be higher among patients with cancer and solid-organ and bone marrow transplant recipients. Risk factors associated with developing a VRE bloodstream infection in a patient already colonized with VRE include cancer or diabetes (relative risk [RR], 3.91), GI procedures (RR, 4.56), acute renal failure (RR, 3.1), exposure to vancomycin (RR, 1.95), infection of a site other than blood (odds ratio, 3.9), and dominance of VRE in the GI tract following use of broad-spectrum antibiotics. Among patients with leukemia, concurrent Clostridioides difficile (formerly Clostridium difficile ) infection was associated with increased risk of developing a VRE bloodstream infection. In addition, two meta-analyses evaluated the mortality of patients with VRE bacteremia compared with patients with vancomycin-susceptible enterococci (VSE). Both studies concluded that patients with bacteremia due to VRE were approximately 2.5 times more likely to die than patients with VSE bacteremia, indicating that the development of vancomycin resistance is a poor prognostic sign in critically ill patients. Despite the introduction of antibiotics with in vitro bactericidal activity against VRE, such as daptomycin, this effect persists, with data showing VRE bacteremia associated with an increase in mortality (odds ratio, 1.8) and longer hospital stays (mean 5.01 days) compared with VSE bacteremia.

Clinical Manifestations of Enterococcal Disease