Carbohydrates

Each of the three routes of intermediary carbohydrate metabolism—the Embden-Meyerhof-Parnas (glycolysis), Entner-Doudoroff, and pentose phosphate (phosphogluconate) pathways—are present in enterococci, or at least in E. faecalis, where appropriate studies have been performed (Sokatch & Gunsalus, 1657). The Embden-Meyerhof-Parnas and Entner-Doudoroff pathways are similar in that hexoses are phosphorylated and subsequently cleaved by aldolases to form the triose phosphate intermediate, glyceraldehyde-3-phosphate, for subsequent metabolism by the glycolytic pathway. These pathways provide the cell ATP via substrate phosphorylation. In contrast, the multifunctional pentose phosphate pathway not only ferments hexoses, pentoses, and other sugar acids for energy (Goddard & Sokatch, 1964), but also generates NADPH for biosynthetic reactions and channels pentoses into nucleotide biosynthesis.

The branch point from the Embden-Meyerhof-Parnas pathway to the pentose phosphate pathway occurs when glucose-6-phosphate is oxidized to 6-phosphogluconate which, in turn, is decarboxylated and oxidized to D-ribulose-5-phosphate. Few enzymes in these pathways have been characterized for enterococci. One exception is the E. faecalis 6-phosphogluconate dehydrogenase (E.C. 1.1.1.44–EF3142 and EF1049), which comprises two distinct enzymes that use NADP or NAD as cofactors (Brown & Wittenberger, 1972). The NADP-linked dehydrogenase is inhibited by fructose-1,6-bisphosphate, but not by ATP, while the NAD-linked enzyme is inhibited by ATP, but not by fructose-1,6-bisphosphate. Although these enzymes are undoubtedly important to carbon flow through the pentose phosphate and the Embden-Meyerhof-Parnas pathways (Wittenberger, Palumbo, Bridges, & Brown, 1971), little work in this area has been performed using enterococci.

In comparison to the relative lack of investigation into the central carbon metabolism for enterococci, considerable effort has focused on the E. faecalis phosphoenolpyruvate phosphotransferase system (PTS). This system senses sugars outside the cell and couples their uptake with phosphorylation (Postma, Lengeler, & Jacobson, 1993). The sugars most commonly metabolized by enterococci are substrates for PTS. Novel PTS components have been observed on mobile elements in E. faecalis, which suggests their importance in virulence (Paulsen, et al., 2003). Most low-GC content Gram-positive bacteria that synthesize ATP by substrate-level phosphorylation under anaerobic conditions also express PTS. This ubiquitous system allows organisms to couple carbohydrate transport to phosphorylation using a mechanism that is more efficient for monosaccharides than non-PTS systems, which can ultimately expend more than one ATP per imported sugar versus the consumption of a single phosphoenolpyruvate. For enterococci and other bacteria, PTS helps regulate glycerol metabolism (via HPr[his15-P]) and is involved in inducer expulsion, inducer exclusion, and catabolite repression.

The initial reaction in PTS-mediated sugar translocation is phosphorylation of a small, soluble cytoplasmic protein by enzyme I (EI, E.C. 2.7.3.9–EF0710) (Figure 1). For E. faecalis, EI is a constitutive 140 kDa homodimer phosphorylated at a specific histidyl residue by the energy-rich glycolytic intermediate phosphoenolpyruvate (Alpert, Frank, Stüber, Deutscher, & Hengstenberg, 1985). Phosphorylated EI (EI-P) transfers its phosphate to a constitutive 9.6 kDa histidine-containing phosphocarrier protein (HPr–EF0709). The three-dimensional structure of the E. faecalis HPr is similar to that of other microorganisms, with the phosphate acceptor located on a histidine residue at position 15 (HPr[his15-P]) (Maurer, Döker , Görler, Hengstenberg, & Kalbitzer, 2001). Unlike subsequent reactions that use the sugar-specific enzymes II (EIIs), EI and HPr are general PTS proteins. EIIs can consist of up to four polypeptides or domains (A, B, and C, or A, B, C, and D) with EIIC and EIID being integral membrane proteins (Deutscher, Francke, & Postma, 2006). EII enzymes catalyze the last step in sugar transport by acting as a phosphorelay from HPr to the sugar, which adds a phosphoryl group at the 6 carbon during hexose sugar transport. The only EII enzymes that have thus far been characterized for enterococci include mannitol, maltose, and gluconate (Fischer, von Strandmann, & Hengstenberg, 1991; Le Breton, Pichereau, Sauvageot, Auffray, & Rincé , 2005; Brockmeier, et al., 2009).

Figure 1

Phosphoenolpyruvate phosphotransferase and catabolite repression. Phosphorylation of E. faecalis phosphocarrier protein (HPr) by enzyme I (EI) at histidine residue 15 forms HPr(his15-P) and couples sugar uptake to sugar phosphorylation through carbohydrate-specific (more...)

The site of HPr phosphorylation is an important regulatory mechanism for sugar metabolism in E. faecalis. EI-P exclusively phosphorylates HPr at his15 (HPr[his15-P]) to initiate a cycle of events that leads to sugar uptake. This phosphorylation occurs in both Gram-negative and Gram-positive bacteria. In Gram-positive bacteria, especially those with low G+C content, HPr can also be phosphorylated at a seryl residue at position 46 (HPr[ser46-P]) (Deutscher, Francke, & Postma, 2006). This phosphorylation is reversible and both reactions are catalyzed by the bifunctional enzyme HPr Kinase/Phosphatase (hprK, E.C. 2.7.1., and E.C. 3.1.3.–EF1749). Phosphorylation only occurs when ATP levels are elevated, as might happen during active sugar metabolism (Kravanja, et al., 1999). Phosphorylation at this site greatly attenuates EI-P phosphorylation at his15, and as a consequence inhibits PTS-mediated sugar uptake. Conversely, during periods of ATP limitation, a second enzymatic site on HprK hydrolyzes the serine phosphate to free HPr for phosphorylation at his15 by EI-P. This uniquely dual function of HprK under ATP control presumably coordinates the metabolic demand by adjusting the ratios of active (HPr[his15-P]) to inactive (HPr[ser46-P]) HPr.

Bacteria, including enterococci, repress and/or inhibit alternate carbon source metabolism during growth on rapidly fermentable carbon sources like glucose. This phenomenon is called catabolite repression (CR) (Fig. 1). Serine-phosphorylated HPr is a key negative regulator for low-GC content Gram-positive bacteria. Other components of CR include a trans-acting factor called the catabolite control protein A (CcpA–EF1741) and cis-acting nucleotide sequences that are termed catabolite responsive elements (cres). CcpA is a DNA-binding protein that was first identified in Bacillus subtilis and regulates expression via cre sequences either within or near promoters of target genes. Both gene activation and repression have been described as a consequence of CcpA binding. The cofactor required for CcpA binding to cre is HPr[ser46-P], but not non-phosphorylated or histidine-phosphorylated forms of HPr. A 36 kDa CcpA homologue for E. faecalis was recently identified (Leboeuf, Auffray, & Hartke, 2000). The ccpA gene product restored glucose repression for cre-responsive genes in B. subtilis, which demonstrated CR function. 2-D protein electrophoresis assays determined that E. faecalis CcpA regulated 22 individual gene products, including glycerol dehydrogenase and dihydroxyacetone kinase. Additionally, an E. faecalis ccpA knockout mutant de-regulated catabolite repression in the presence of glucose, which allowed this strain to simultaneously co-metabolize citrate and glucose (Rea & Cogan, 2003). Another study demonstrated the presence of 63 putative cre-sites in the promoters of genes involved with several metabolic pathways, including those for citrate, sucrose, lactose, galactose, serine, and arginine (Opsata, Nes, & Holo, 2010). Another form of catabolite-dependent regulation is inducer expulsion. This process involves de-phosphorylating and exporting other sugars during growth on glucose (Ye, Minarcik, & Saier, Jr., 1996). Inducer expulsion requires a small membrane-associated sugar-phosphate phosphatase (Pase II) to dephosphorylate cytosolic sugars prior to export, and is positively regulated by HPr[ser46-P].

Glycerol metabolism

Glycerol metabolism is of importance as a pathway for the synthesis of lipids and (lipo)teichoic acids in many Gram-positive bacteria, including E. faecalis (Coyette & Hancock, 2002). Glycerol can also be a carbon/energy source for several pathogenic bacteria. For example, Listeria monocytogenes use glycerol-catabolizing enzymes for intracellular growth (Joseph, et al., 2006). Furthermore, studies on Mycoplasma sp., bacteria that are adapted to life within eukaryotic hosts through reductive evolution as evident by a minimalistic genome, still rely on a handful of carbon sources, including glycerol (Halbedel, Hames, & Stülke, 2004).

Among enterococci, E. faecalis appears to have the most diverse glycerol metabolism. Members of this species can ferment glycerol under aerobic as well as anaerobic conditions (Bizzini A. , et al., 2009; Gunsalus & Sherman, 1943). Genome sequences for E. faecalis reveal two pathways for glycerol catabolism (Fig. 2). One begins with the ATP-dependent phosphorylation of glycerol by glycerol kinase (GlpK / EF1929) to yield glycerol-3-phosphate (glycerol-3-P). In Gram-positive bacteria, such as Bacillus subtilis, E. faecalis, and E. casseliflavus, GlpK activity is inhibited by fructose-1,6-bisphosphate and activated by phosphorylation through the general PTS (phosphoenolpyruvate (PEP):carbohydrate phosphotransferase system) regulator (HPr[his15-P]), which leads to 10- to 15-fold increased activity (Deutscher, Francke, & Postma, 2006). Detailed structural studies with the GlpK of E. casseliflavus in the presence or absence of glycerol have been performed (Yeh, et al., 2004). The glycerol-3-P formed is subsequently oxidized to dihydroxyacetone phosphate (DHAP) by glycerol-3-P oxidase (GlpO–EF1928). This enzyme uses molecular oxygen as an electron sink and leads to the formation of H₂O₂. Exogenous catalase increases the aerobic growth of E. faecalis on glycerol, which indicates that this oxidant can accumulate to growth-inhibiting concentrations under these conditions (Bizzini A. , et al., 2009). Inactivation of the E. faecalis npr gene (EF1211) that encodes NADH peroxidase or inactivation of the ahpCF operon (EF2739 and EF2738) that encodes alkyl hydroperoxide reductase led to reduced growth on glycerol under aerobic conditions. These enzymes catalyse the NADH-dependent dismutation of H₂O₂ to oxygen and water (La Carbona, et al., 2007). Genes that encode GlpK and GlpO are contained in an operon structure that also encodes the glpF gene (EF1927) for glycerol transport. This integral membrane protein belongs to the aquaporin family of molecular water channels and allows glycerol to enter the cell through facilitated diffusion (13). The PrfA-like transcriptional regulator Ers (EF0074) has been indirectly implicated in regulating the glpK-operon (Riboulet-Bisson, Hartke, Auffray, & Giard, 2009). The glpK-operon appears to be conserved among enterococci, with identical elements present in genome sequences for other enterococcal species.

Figure 2.

Glycerol metabolism. Glycerol uptake in E. faecalis occurs through an energy-independent diffusion facilitator (GlpF). Enzymes for glycerol dissimilation include: ATP-dependent glycerol kinase (GlpK); H₂O₂-producing L-α-glycerophosphate oxidase (more...)

The second pathway for glycerol catabolism in E. faecalis involves the oxidation of glycerol to dihydroxyacetone by a soluble NAD⁺-dependent glycerol dehydrogenase (GldA1/ EF1358). Dihydroxyacetone is then phosphorylated to DHAP by dihydroxyacetone kinase (DhaK/ EF1360). Anaerobic growth of E. faecalis on glycerol depends on fumarate as an electron acceptor to allow for the reoxidization of NADH. In this reaction, fumarate is reduced to succinate by a membrane-associated fumarate reductase (30). Genes for these activities are encoded in an operon structure (EF1358 to EF1361) that includes GldA1, a small hypothetical protein, and two subunits for DhaK (DhaK and DhaL). Interestingly, E. faecium harbours a similar operon that lacks gldA, and this likely explains why the species is unable to grow anaerobically on glycerol. The putative sequence of a small hypothetical polypeptide on the dhaK operon has significant homology to the DhaM protein of Lactococcus lactis (Zurbriggen, et al., 2008). L. Lactis DhaM contains an EIIA domain for the mannose family of PTS and is also conserved in the DhaM-like E. faecalis protein. Recently it has been demonstrated that the E. faecalis DHA kinase is phosphorylated in a PEP-dependent phosphotransfer reaction, which includes the DhaM enzyme (Sauvageot, et al., 2012) that is similar to L. lactis (Fig. 2).

Mutants involving both pathways have been analysed for growth under both aerobic and anaerobic conditions (Bizzini A. , et al., 2009). Comparisons using several E. faecalis isolates demonstrate an impressive diversity of growth behaviours on glycerol under these conditions, which is not due to differing gene content, but rather a modulation of gene expression. Some strains preferentially use the glpK pathway, others the dhaK pathway; or glycerol is catabolised simultaneously under aerobic conditions. These findings clearly demonstrate that E. faecalis displays heterogeneous modes of aerobic glycerol catabolism (Bizzini A. , et al., 2009). However, under anaerobic conditions, glycerol is exclusively metabolized via the DhaK pathway.

Two other genes with annotated roles in glycerol metabolism are found in genome sequences for E. faecalis and include a second putative glycerol dehydrogenase (EF0895) and a putative glycerol-3-phosphate dehydrogenase (EF1747). However, these activities do not appear to be involved in glycerol catabolism (Bizzini A. , et al., 2009). Finally, a second putative glycerol facilitator (EF1828) is found in few E. faecalis strains besides the V583 strain, but its significance to glycerol transport remains to be determined.

Citrate

Enterococcal citrate metabolism is commercially valuable and presumed to contribute to “flavor compound” production in numerous cheeses. Citrate can be fermented as the sole source of carbon and energy for E. faecalis and E. durans. Enterococcal growth on citrate requires transportation across the cell membrane, followed by cleavage into acetyl-coA and oxaloacetate by citrate lyase. Oxaloacetate is then decarboxylated to pyruvate. End-products produced by E. faecalis grown on citrate as a sole carbon source include acetate, formate, and, to a lesser degree, lactate, ethanol, and acetoin (Sarantinopoulos, Kalantzopoulos, & Tsakalidou, 2001; Campbell & Gunsalus, 1944). Other minor products include α-acetolactate, diacetyl, and 2,3-butanediol. These metabolites, in addition to the previously listed compounds, contribute to flavor production in commercial fermentations.

The regulation of citrate metabolism has been characterized in E. faecalis (Blancato, Repizo, Suárez, & Magni, 2008). Citrate is presumably transported by the CitH transporter (EF3327) and leads to the increased affinity of a CitO protein for non-coding regulatory regions upstream of the citHO-oadHDB-citCDEFX-oadA-citMG operons. These genes comprise the citrate regulon and have significant sequence homology to genes for citrate fermentation in other lactic acid bacteria. Together these genes encode products necessary for citrate transport (CitH–EF3327), citrate-dependent regulation (CitO), citrate lyase (CitD-F–EF3318 to EF1322), oxaloacetate decarboxylase (CitM–EF3317), and the oxaloacetate decarboxylase complex (OadABD H). Citrate is metabolized in the presence of lactose, although catabolism is inhibited by PTS via catabolite repression, and additional unidentified regulatory elements are involved in this process (Suárez, Blancato, Poncet, Deutscher, & Magni, 2011).

Mucin degradation

Enterococci likely derive carbon and energy by fermenting non-absorbed sugars in the gastrointestinal tract. Other sugars, however, are also available in this environment as heavily glycosylated mucins produced by specialized epithelial cells. Mucins are large, cysteine-rich proteins that have been highly decorated with carbohydrates (70-80% by weight). These complex molecules form the primary component of the adherent mucous layer that protects the intestinal epithelium from luminal contents and acts as a mechanical lubricant. The ability to degrade mucin and liberate their attached sugars would be advantageous to members of the intestinal microbiota. However, relatively few enteric bacteria can degrade mucins. This is due, in part, to poly-O-acetylation and O-sulfation of sialic acids on outer non-reducing ends of mucin chains. These modifications protect mucin from most forms of bacterial degradation.

Only a few enteric bacteria can completely hydrolyze mucins and use them as a source of carbon and energy, including Escherichia coli, Bifidobacterium spp., Bacteroides spp., and Ruminococcus spp. These bacteria overcome acetylated and sulfated sialic acid residues by expressing cell-bound and extracellular sialate O-acetylesterase and arylesterase, in addition to sialidase, β-N-acetylglucosaminidase, and β-N-acetylgalactosaminidase. This allows E. coli to rapidly grow in cecal mucus, as compared to luminal contents (Poulsen, Licht, Rang, Krogfelt, & Molin, 1995). Evidence for enterococcal growth on mucin as a sole carbon source is lacking, although the sulfatase components of mucin degradation are observed in E. faecalis (Bøhle, et al., 2001). The growth of vancomycin-resistant enterococci in mucin is enhanced by the generation of monosaccharides from mucin through the mucin-degrading activities of either Ruminococcus torques or a human fecal consortia (Pultz, Hoskins, & Donskey, 2006). Apart from these studies, little work is available on mechanisms by which enterococci might degrade and grow in mucin.

Plant Carbohydrates and Errata

The diverse habitat range of the Enterococcus genus suggests that these bacteria are exposed to vast array of carbon and energy sources, both as planktonic bacteria and in gastrointestinal environments. While this information is incomplete, enterococci appear able to degrade numerous biological polymers and modified carbohydrates and use them as carbon and energy sources. Many plant carbohydrate polymers are resistant to decomposition by the human digestive system and can therefore provide a source of nutrients to the gastrointestinal microbiota. Of these polymers, cellulose is the most abundant and is composed of repeating β(1-4) glucose linkages. E. saccharolyticus, E. faecalis and E. gallinarum can grow in a pure or mixed consortium on cellulose as a carbon source (Adav, Lee, Wang, & Ren, 2009; Chassard, Delmas, Robert, & Bernalier-Donadille, 2010; Wang, Gao, Ren, Xu, & Liu, 2009). At least four putative β-glucosidases necessary for cellulose decomposition are encoded on the E. faecalis V583 genome (EF1020, EF1238, EF1243, and EF1606). Many more are likely present in other strains and species within the genus. Hemicellulose-containing compounds such as lignin are second in abundance to cellulose, but its degradation and utilization by enterococci has yet to be investigated.

Raffinose is a trisaccharide found in many plant tissues that cannot be digested by human enzymatic activity. Recently, raffinose-utilization genes were identified on a megaplasmid in E. faecium (Zhang, Vrijenhoek, Bonten, Willems, & van Schaik, 2011). This activity may provide a means for establishing a niche in the complex gastrointestinal milieu. E. faecium also possesses enzymes for the initial steps in fructosamine catabolism, which potentially allows it to grow on this plant-derived carbon source (Wiame, Lamosa, Santos, & Van Schaftingen, 2005). Pectin is another abundant plant-derived polysaccharide that is only partially digested by most mammals, at best. Pectin is composed of several sugars that most commonly include D-galacturonate in α(1–4) linkages. Human enterococcal isolates can utilize pectin in vitro, and dietary pectin from heavy fruit consumption has been linked to increased intestinal enterococci (Shinohara, Ohashi, Kawasumi, Terada, & Fujisawa, 2010).

Maltose is an α(1–4) linked glucose disaccharide produced by amylase degradation of starches and can be readily utilized by E. faecalis as a growth substrate. Recent activity has demonstrated that this disaccharide is transported by a PTS system that uses the maltose permease MalT and generates maltose-6-P similar to that of Bacillus subtilis (Le Breton, Pichereau, Sauvageot, Auffray, & Rincé , 2005). However, E. faecalis does not hydrolyse maltose-6-P into glucose-1-P and glucose-6-P like B. subtilis. Instead, E. faecalis dephosphorylates maltose-6-P using a novel enzyme MapP (Mokhtari, et al., 2013). Maltose is then cleaved by MalP into glucose and glucose-1-P, which can enter glycolysis after conversion to glucose-6-P by phosphoglucomutase.

Trehalose is an α(1–1) linked glucose disaccharide that is found primarily in insects and plant materials. Trehalose metabolism resembles maltose uptake and utilization, and involves a trehalose PTS system that converts this disaccharide into trehalose-6-P during transport. Trehalose-6-P is split by a trehalose-6-phosphate phosphorylase (TPP) enzyme into glucose-6-P and β-glucose-1P. β-glucose-1P is isomerized by β-phosphoglucomutase into glucose-6-P, and both sugars then enter glycolysis (Andersson & Rådström, 2002). Enzymatic activity for this pathway in enterococci has been verified, and putative gene sequences that encode these enzymes are found within the E. faecalis genome.

In addition to trehalose, enterococci are exposed to the β(1-4) linked N-acetylglucosamine (GlcNac) polymer chitin during growth on insect-associated material, as well as fungi. A recent study showed that E. faecalis can grow on insoluble chitin as a sole source of carbon and energy (Vaaje-Kolstad, et al., 2012). Chitin degradation requires several enzymatic steps with individual GlcNac monomers imported through PTS prior to entry into glycolytic pathways. E. faecalis secretes two enzymes encoded by a putative operon that functions as a chitinase and chitin-binding protein with chitinase activity (EF0361–EF0362). These enzymes degrade chitin into the disaccharide GlcNac₂,which is then degraded into GlcNac monomers by a secreted chitobiase (E.C. 3.2.1.52 - EF0114). GlcNac is then transported by an undetermined PTS system prior to deacetylation by either NagA-1 (EF1317) or NagA-2 (E.C. 3.5.1.25, EF3044) deamination into fructose-6-P (NagB, E.C. 3.5.99.6, EF0466), as well as entry into glycolytic pathways.

Ethanolamine catabolism is specific to E. faecalis and is obtained from cell membrane lipids through phosphatidylethanolamine degradation. Further degradation of lipid-derived fatty acids has not been previously demonstrated, and a survey of enterococcal genomic databases does not reveal the presence of any homologous gene sequences that encode for known β-oxidation enzymes. Ethanolamine catabolism is extensively covered in Pathogenesis and models of enterococcal infection.

Pyruvate synthesis

For E. faecalis, at least four metabolic pathways lead to the formation of pyruvate. Sugar fermentation produces pyruvate through dephosphorylation of phosphoenolpyruvate by pyruvate kinase (pyk, EC 2.7.1.40–EF1046). This reaction forms ATP and is the final step in the glycolysis. Malic enzyme (E.C. 1.1.1.39) constitutes another pathway leading to pyruvate. This enzyme catalyzes the NAD⁺-dependent oxidative decarboxylation of malate and, unlike other bacteria, does not participate in biosynthetic reactions that utilize pyruvate as a carbon source. Instead, malate is only catabolized for energy by E. faecalis. Not surprisingly, E. faecalis does not express a malolactic enzyme often found in lactic acid bacteria that directly converts malate to lactate. Malic enzyme has been identified in E. faecalis (MaeE–EF3316) and is essential for malate catabolism, as well as pH homeostasis under acidic conditions (Espariz, et al., 2011). Malate metabolism is regulated by HPr(Ser-P)-dependent catabolite repression in the presence of sugars and glycolytic intermediates. Furthermore, this gene is induced by the activation of the maeR gene product encoded on the maeKR operon (EF1209-EF1210) (Mortera, et al., 2012). Growth on malate is promoted by oxygen or fumarate as terminal electron acceptors, as well as lipoic acid.

Oxaloacetate decarboxylase (oad, E.C. 4.1.1.3) is a sodium transport decarboxylase and a member of a family that includes methylmalonyl-coenzyme A (CoA), glutaconyl-CoA, and malonate decarboxylases (Bott, 1997). These multicomponent biotin enzymes couple exergonic decarboxylation with sodium export to produce a transmembrane sodium gradient (ΔμNa⁺). This potentially assists other sodium transport systems (see Ion Transport section). At least one oxaloacetate decarboxylase has been identified in E. faecalis (CitM–EF3317); however, it remains to be determined whether this is the sole enzyme responsible for oxaloacetate catabolism (Espariz, et al., 2011).

Pyruvate dehydrogenase complex

Dehydrogenation of pyruvate by E. faecalis can occur through one of three reactions, each of which produces acetyl-CoA. Acetyl-CoA then proceeds by a two-step reduction to generate ATP, or directly enters into fatty acid biosynthesis. The pyruvate dehydrogenase complex (pdh) is one dehydrogenation pathway for this methyl-α-keto acid. Pdh is a member of a family of multienzymes that catalyze oxidative decarboxylation of α-keto acids. The E. faecalis Pdh is an ordered collection of multiple copies of three enzymes, pyruvate dehydrogenase (E1, E.C. 1.2.4.1), dihydrolipoyl acyltransferase (E2, E.C. 1.8.1.4), and dihydrolipoamide dehydrogenase (E3, E.C. 1.8.1.3) (Perham, 1991). Substrate specificity for Pdh resides in E1 and E2 subunits, with the E3 subunit serving to reoxidize enzyme-bound dihydrolipoamide. The E. faecalis Pdh dihydrolipoyl acyltransferase (aceF–EF1355, E2 subunit) is a hollow dodecahedral 60-meric enzyme with icosahedral symmetry by x-ray crystallography (Izard, et al., 1999). The E2 subunit contains two lipoyl domains, a peripheral subunit-binding domain, and an acetyltransferase that is linked by a polypeptide enriched for alanine and proline (Allen & Perham, 1991). The lipoyl group is a swinging arm that conveys substrates between three successive active sites. The E. faecalis E2 subunit is similar to human E2 (two lipoyl domains), but not E2s for yeast (one lipoyl domain) or other bacteria (one or three lipoyl domains). In this respect, the E. faecalis Pdh differs from all other bacterial pyruvate dehydrogenases.

The E3 subunit is a homodimer that contains noncovalently bound FAD and a redox-active disulfide bond. The E1/E2/E3 stoichiometry for E. faecalis Pdh is 1.5:1:1 repeated 60 times to give a complex with an estimated mass of 14 million Da (Snoep, et al., 1992). A unique feature of the E. faecalis Pdh complex is continued activity, despite strong reducing conditions. NADH/NAD⁺ ratios that completely inhibit E. coli Pdh have no effect on E. faecalis Pdh. In this regard, the E. faecalis E3 subunit seems far less susceptible to over-reduction than other E3 subunits and this, coupled with full expression of E. faecalis Pdh under anaerobic conditions, may explain the resistance of Pdh to anaerobiosis (Snoep, et al., 1992). This unusual characteristic likely provides E. faecalis with considerable metabolic flexibility during aerobic-anaerobic transitions.

Pyruvate formate-lyase

Dehydrogenation of pyruvate by E. faecalis and E. faecium can occur under anaerobic conditions by pyruvate formate-lyase (or formate acetyltransferase; pflB, E.C. 2.3.1.54–EF1613). This enzyme converts pyruvate and CoA to formate and acetyl-CoA through the unusual mechanism of a protein radical intermediate. Pfl has been partially purified for E. faecalis and is similar to E. coli Pfl. Under anaerobic conditions, E. coli is activated through the introduction of a stable free radical onto glycine at position 734 (Sawers & Watson, 1998; Wagner, Frey, Neugebauer, Schäfer, & Knappe, 1992). Activated Pfl is extraordinarily sensitive to oxygen, with exposure causing peptide-bond cleavage near the C terminus of the subunit that contains the glycyl radical. This leads to irreversible inactivation of the enzyme. Pfl is also acid sensitive, with growth of E. faecalis on pyruvate possible only at pH 7.5 or higher, when Pdh has been rendered non-functional (Snoep, van Bommel, Lubbers, Teixeira de Mattos, & Neijssel, 1993). Activated Pfl is protected from oxygen by a specific activating enzyme (pflA, E.C. 1.97.1.4–EF1612).

Pyruvate flavodoxin/ferredoxin oxidoreductase

A third, distinct enterococcal pyruvate dehydrogenation pathway has been reported, but only for E. faecalis (Yamazaki, Watanabe, Nishimura, & Kamihara, 1976). This activity is detectable under reduced conditions, converts pyruvate to acetyl-CoA, is unlike Pfl, and is insensitive to catabolite repression. The enzyme most likely responsible for this activity is pyruvate flavodoxin/ferredoxin oxidoreductase (E.C. 1.2.7.1–EF2559). A gene that codes for this enzyme is present in the E. faecalis and E. faecium genome databases and is nearly identical to por for Entamoeba spp., Giardia spp., and many other Gram-negative bacteria (Samuelson, 1999). No reports, however, have characterized this gene or its product in enterococci. Much like Pdh, it is likely that this enzyme catalyzes the oxidative decarboxylation of pyruvate using thiamine pyrophosphate, followed by an acetyl transfer to form acetyl-CoA. In this fashion, it differs from Pdh in the use of ferredoxin or flavodoxin as an electron acceptor with reducing equivalents transferred to NAD⁺.

Alcohol dehydrogenase

Fermentation of mannitol or glucose by E. faecalis under nutrient-limited conditions produces not just formate and lactate, but also ethanol (Sarantinopoulos, Kalantzopoulos, & Tsakalidou, 2001; Snoep, Joost, de Mattos, & Neijssel, 1991). Ethanol-forming enzymes have not been identified for enterococci. One likely candidate, however, is the multifunctional alcohol dehydrogenase (adhE, E.C. 1.1.1.1 and 1.2.1.10–EF0900). A gene with similarity to adhE for other bacteria can be found in the E. faecalis and E. faecium genomic databases (Arnau, Jørgensen, Madsen, Vrang, & Israelsen, 1998). AdhE catalyzes the conversion of acetyl-CoA to ethanol through acetaldehyde dehydrogenase and alcohol dehydrogenase. Reducing equivalents are provided by NADH. This enzyme, which is expressed under anaerobic conditions in E. coli and L. lactis, is a gene fusion product, with the N terminus homologous to the family of aldehyde:NAD⁺ oxidoreductases and C-terminus homologous to iron-dependent alcohol:NAD⁺ oxidoreductases.

Acetate metabolism

Acetyl-CoA formed by Pdh, Pfl, and the putative Por can generate one additional molecule of ATP through reactions catalyzed by phosphotransacetylase (eutD, E.C. 2.3.1.8–EF0949) and acetate kinase (ackA, E.C. 2.7.2.1–EF1983). The first step in this energy-yielding process involves the replacement of acetyl-CoA with phosphate to produce acetylphosphate and CoA. AckA catalyzes the second step to generate ATP from acetylphosphate. This pathway appears to be reversible in E. faecalis when cytosolic lactate dehydrogenases (ldh, E.C. 1.1.1.27) are inactivated (Jönsson, Saleihan, Nes, & Holo, 2009; Rana, Gente, Rincé, Auffray, & Laplace, 2012).

Branched chain α-keto acid catabolism

An energy-yielding metabolic pathway has been described for E. faecalis that generates ATP from non-pyruvate α-keto acids. This pathway is encoded by the bkd operon and decarboxylates derivatives of branched-chain amino acids to produce corresponding acyl-CoAs. Compounds used in this pathway include α-ketoisovalerate (valine), α-ketoisocaproate (leucine), and α-keto-β-methylvalerate (isoleucine) (Rüdiger, Langenbeck, & Goedde, 1972; Ward, Ross, van der Weijden, Snoep, & Claiborne, 1999). Keto acids, like pyruvate, are energy-rich molecules that produce ATP through sequential reactions. The initial step involves an α-keto acid dehydrogenase complex that is encoded by the bkd-DABC operon. This large multienzyme complex is similar to Pdh and requires NAD⁺, CoASH, and lipoic acid, much like Pdh requires these same cofactors. bkdA (EF1660), bdkB (EF1659), bkdC (EF1658) and bkdD (EF1661) encode proteins homologous to the E1α, E1β, E2, and E3 subunits of the branched-chain α-keto acid dehydrogenase complex from B. subtilis, and function to decarboxylate α-keto acids with the generation of corresponding acyl-CoAs. Amongst enterococci, the bkd operon appears to be unique to E. faecalis (Palmer, et al., 2012).

After the formation of branched-chain acyl-CoAs, reactions catalyzed by ptb and buk gene products form ATP (Ward, Ross, van der Weijden, Snoep, & Claiborne, 1999). These enzymes use acyl-CoAs to produce isovalerate, isobutyrate, and methylbutyrate as end-products. Substrate specificity of the ptb gene product is broad, and includes C₂ to C₈ straight chain acyl-CoAs, along with branched-chain acyl-CoAs. The buk gene product is not well characterized, but shows kinase activity coupled to ATP formation using acetyl-phosphate and butyrylphosphate as substrates. Growth of E. faecalis in pyruvate-limited culture with α-ketoisovalerate, α-ketoisocaproate, or α-keto-β-methylvalerate results in a significantly greater biomass, and corresponding branched-chain carboxylic acids are recovered in the growth medium. This pathway results in 0.5 mole of ATP per mole of catabolized α-keto acid (Ward, et al., 2000). Although one gene similar to the family of branched-chain amino acid aminotransferases is present in the E. faecalis genome, when branched-chain amino acids were tested for growth enhancement, none was observed (Ward, Ross, van der Weijden, Snoep, & Claiborne, 1999).

Acetoin metabolism

Acetoin (3-hydroxy 2-butanone) is a neutral compound that allows bacteria to degrade sugars without substantial acidification of the growth medium (Sarantinopoulos, Kalantzopoulos, & Tsakalidou, 2001; Dolin & Gunsalus, 1951). Production is measured by the Voges-Proskauer reaction and is positive for all enterococci (Fertally & Facklam, 1987). There is no evidence for a conversion of acetoin to diacetyl or 2,3-butanediol by enterococci. Acetoin is formed from pyruvate by α-acetolactate synthase (alsS, E.C. 4.1.3.18–EF1213) and α-acetolactate decarboxylase (alsD, E.C. 4.1.1.5–EF1214), with both present in the E. faecium and E. faecalis genomes. α-acetolactate synthase is a thiamine pyrophosphate-containing enzyme that forms a pyruvate-thiamine adduct. Pyruvate is initially decarboxylated prior to attack on the keto group by a second pyruvate molecule to form α-acetolactate. This enzyme has a low affinity for pyruvate and is only active when intracellular pyruvate concentrations are high. Acetoin is then formed from α-acetolactate by α-acetolactate decarboxylase. Regulatory mechanisms for these genes have been extensively studied in many bacteria, including B. subtilis (Cruz Ramos, et al., 2000), but have only recently been studied in E. faecalis (Mehmeti, et al., 2011).

Lactate Production

Lactate is the major end product of enterococcal fermentation during growth on excess glucose under reducing conditions. It is generated by the reduction of pyruvate to regenerate NAD⁺ for ongoing glycolysis. E. faecalis possesses two cytosolic NAD⁺-dependent L-(+)-lactate dehydrogenases (ldh, E.C. 1.1.1.27), which are deemed ldh-1 and ldh-2 (EF0255 and EF0641). These isoenzymes are activated by the glycolytic intermediate fructose-1,6-bisphosphate (Ldh-2 > Ldh-1) and are further regulated by the combined effects of intracellular pH and phosphate (Feldman-Salit, et al., 2013). This intermediate regulation is a specific need for these enzymes that otherwise show little or no activity without it as an activator. Ldh-1 is more active and is responsible for the bulk of lactate production during growth on hexoses. Recent evidence has suggested that Ldh-1 is extensively regulated and post-transcriptionally dependent upon culture growth rate, possibly to allow E. faecalis to rapidly adapt to nutrient variations in its environment (Mehmeti, et al., 2012). The deletion of Ldh-1 or both Ldh-1 and 2 causes a significant increase in mixed acid fermentation end-products (Mehmeti, et al., 2011). Ldh-2 contributes to lactate formation, but is insufficient for normal growth on glucose by itself (Jönsson, Saleihan, Nes, & Holo, 2009). Interestingly, although growth under laboratory conditions is comparable, ldh1 and ldh1/ldh2 double mutants demonstrate decreased fitness in stressing environments and are attenuated in virulence (Rana, et al., 2013).

Demethylmenaquinone

Quinones are membrane-embedded electron carriers essential to all respiratory processes. In bacteria, quinones are derivatives of either ubiquinone or menaquinone. E. faecalis synthesizes a modified menaquinone that lacks a 2-methyl group. This derivative is termed demethylmenaquinone (Baum & Dolin, 1965). A comprehensive survey of quinones in enterococci identified demethylmenaquinone in E. faecalis, E. casseliflavus, and E. gallinarum, but not E. faecium or E. durans (Collins & Jones, 1979). No enterococcal strain has been shown to express ubiquinone or menaquinone. The midpoint potential for demethylmenaquinone (ΔE_m,7 = +36 mV) is halfway between ubiquinone (ΔE_m,7 = +113 mV) and menaquinone (ΔE_m,7 = -74 mV) (Gennis & Stewart, 1996). This characteristic confers respiratory flexibility and allows redox reactions to use fumarate or oxygen as electron acceptors under anaerobic and microaerophilic conditions.

Cytochrome bd

Enterococci, like streptococci, lack the ability to synthesize heme because porphyrin precursors cannot be produced, due to a missing tricarboxylic acid cycle. As a result, E. faecalis does not ordinarily express cytochromes, but instead relies on fermentation for growth. Among streptococci, enterococci, and lactococci, only E. faecalis and L. lactis are known to express cytochromes. This occurs only in the presence of heme, where aerobic growth leads to oxidative phosphorylation and enhanced production of ATP (Ritchey & Seely, Jr. , 1976; Winstedt, Frankenberg, Hederstedt, & von Wachenfeldt, 2000).

Cytochrome bd is the key respiratory enzyme for E. faecalis. This cytochrome is a widely distributed terminal quinol oxidase (Borisov, Gennis, Hemp, & Verkhovsky, 2011) that contains two subunits (CydA and CydB–EF2061 and EF2060) with three distinct cytochromes (b₅₅₈, b₅₉₅, and d). The low-spin b₅₅₈ heme in CydA is the site of quinol oxidation. Substrate protons that arise from quinol oxidation are released outside the cell and help provide a proton-motive force by transmembrane charge separation. The high-spin b₅₉₅ heme is located near heme d in CydB and forms a bimetallic center for the four-electron reduction of O₂ to H₂O. cydC and cydD (EF2059 and EF2058) are necessary for cytochrome bd expression. These genes code for proteins with similarity to ATP-binding cassette transporters, and are implicated in heme transport and/or assembly. The loss of cytochrome bd in E. coli leads to a pleiotropic phenotype that is characterized by stationary-phase arrest, as well as hypersensitivity to elevated temperatures and high concentrations of H₂O₂ and zinc (Goldman, Gabbert, & Kranz, 1996). Similar effects have not been seen with E. faecalis and seem unlikely, since this organism tolerates diverse and severe stresses (see below) despite in vitro growth without heme.

Fumarate Reductase

Nonoxidative respiration for prokaryotic organisms substitutes a variety of electron acceptors—nitrate, nitrite, sulphite, iron (III), CO₂, and fumarate—for O₂. The only system so far identified for E. faecalis, and a few strains of E. faecium, is fumarate reductase (Frd) (Deibel, 1964; Huycke M. M., et al., 2001; Aue & Diebel, 1967). This membrane-associated enzyme catalyzes the reduction of fumarate, a four-carbon dicarboxylic acid, to succinate. Although the term respiration is used loosely to describe Frd activity, this enzyme cannot catalyze a net transfer of protons across the cell membrane (Hederstedt, 1999). Although Frd can form fumarate from succinate in vitro, this reverse oxidation reaction has not been observed in vivo. Frd is composed of a large flavin adenine dinucleotide (FAD)-containing domain on the cytosolic surface of the cell membrane (subunit A; EF2556), three iron-sulfur clusters (subunit B), and small hydrophobic polypeptide anchors (subunits C and C). Frd appears to be constitutively expressed.

F0F1-ATP Synthase

A primary ion pump for mitochondria and many bacteria is the proton-translocating F0F1-ATP synthase, and F0F1-ATP synthases from diverse biological sources are nearly identical. These complexes are one of the three major classes of ion-motive ATP synthases that have the distinctive characteristic of coupling ATP synthesis to an electrochemical gradient of protons across the cell or mitochrondrial membrane. This system is reversible, and the enzyme can act to either synthesize ATP, or to use ATP to extrude protons and maintain a gradient for transport of other substrates or control cytoplasmic pH. Enterococci can use F0F1-ATP synthase for both purposes.

The first description of a membrane-associated ATP synthase was the F0F1-ATP synthase isolated from E. hirae (formerly Streptococcus faecalis) (Abrams, McNamara, & Bing, 1960). This large enzyme has an F1 moiety with five subunits (α, β, γ, δ, and ε) and a membrane-spanning F0 component that consists of three subunits (a, b, and c) (Shibata, Ehara, Tomura, Igarashi, & Kobayashi, 1992). The structure of the F0F1-ATP synthase for E. hirae is similar to other F0F1-ATP synthases, although the enterococcal enzyme is highly resistant to azide. This enzyme is also encoded by E. faecalis (AtpA) and its contribution to ATP synthesis through oxidative phosphorylation is certain, based on several studies (Ritchey & Seeley Jun, 1974; Pritchard & Wimpenny, 1978). E. hirae, in contrast, does not express a cytochrome and therefore cannot couple the enzymatic activity of the F0F1-ATP synthase to a proton motive force. Instead, for E. hirae and non-respiring enterococci, the F0F1-ATP synthase assists in cytosolic alkalization in acidic environments, like the intestine.

L-Lactate Oxidation

Lactate is not just a fermentation end-product, but yet another energy source for the aerobic growth of E. faecalis (Pritchard & Wimpenny, 1978; Clarke & Knowles, 1980; London, 1968). This unusual feat of metabolism is a phenotypic characteristic of E. faecalis and is distinct from the activity of cytosolic lactate dehydrogenases. Current evidence suggests that a membrane-associated L-lactate:quinone oxidoreductase, analogous to the E. coli D-lactate dehydrogenase, is responsible for lactate oxidation. Evidence for this metabolic activity includes: (i) heme-grown E. faecalis oxidizes lactate 10-20 times faster than bacteria grown without heme (Pritchard & Wimpenny, 1978); (ii) uncoupling ionophores repress lactate oxidation, which links this to the proton motive force (Pritchard & Wimpenny, 1978); and (iii) brisk production of O₂^– occurs when isolated E. faecalis membranes are exposed to L-lactate (Huycke M. M., et al., 2001). Genes that encode enzymes for lactate oxidation have not been identified in enterococci.

Sodium

Salt and alkali tolerance are enterococcal phenotypes due, in part, to several independent mechanisms of sodium transport. All cells must expel excess sodium to maintain cytosolic concentrations in a homeostatic range. E. hirae expresses both a Na⁺/H⁺ antiporter that uses a proton motive force for activity (102) and a vacuolar-type ATPase that uses ATP hydrolysis to pump out sodium (103). A proton gradient used by the Na⁺/H⁺ antiporter can be generated by the F₀F₁-ATPase in cytochrome-free E. hirae or from E. faecalis grown without hematin and that expresses a non-functional apo-cytochrome bd. Under alkaline conditions when the proton motive force is dissipated, the Na⁺/H⁺ antiporter is inhibited and the V₀V₁-ATPase functions instead to extrude sodium and maintain homeostasis (103, 104).

Several genes that code for putative enterococcal Na⁺/H⁺ antiporter have been identified in genome databases (nhaC-1–EF0402; nhaC-2–EF0636; and EF1574), and one has been cloned from E. hirae (105). This gene, termed napA, encodes an extremely hydrophobic protein with 11 or 12 predicted transmembrane helices that shows little in common to other bacterial Na⁺/H⁺ antiporters, but is instead strikingly similar to the glutathione-regulated K⁺ efflux system (KefC) of E. coli (106). E. hirae mutants that only express NapA, and not the vacuolar ATPase (see below), grow in 0.5 M sodium, but only at a pH less than 9.5. Sodium sensitivity at high pH values reflects an inadequate proton-motive force for proton-driven sodium extrusion. NapA expression in an E. coli strain with inactivated Na⁺/H⁺ antiporters restores growth under high sodium concentrations. Growth above pH values of 9.5, however, requires the ATP-driven V₀V₁-ATPase sodium extrusion system, because NapA is nonfunctional when the proton motive force has been dissipated (107).

Vacuolar-ATPases are members of a class of widely distributed proton pumps found in acidic vacuoles of fungi and plants, endosomes of animal cells, and selected bacteria. Both F₀F₁- and V₀V₁-ATPases are multisubunit enzymes that consist of a hydrophilic catalytic portion (F₁ and V₁) and a membrane-associated proteolipid segment that contains the proton or sodium channel (F₀ and V₀, respectively). Significant similarity for major subunits of these enzymes suggests that they have a close evolutionary relationship. The E. hirae V₀V₁-ATPase is an exceptional member of the vacuolar class of enzymes because it translocates sodium, rather than protons, out of the cell. The “Na⁺”-translocating activity is encoded by the ntpFIKECGABD operon and consists of nine Ntp proteins (Murata, Kawano, Igarashi, Yamato, & Kakinuma, 2001). The E. hirae ntp operon is transcribed as a single mRNA and is induced by high intracellular sodium concentrations that occur with excessive extracellular sodium, by high pH when the Na⁺/H⁺ antiporter is nonfunctional, and when the proton motive force has been dissipated by ionophores or a loss of the F₀F₁-ATPase synthase (Murata, Kawano, Igarashi, Yamato, & Kakinuma, 2001). V₀V₁-ATPase activity is maximal at pH 8.5 to 9.0 and undetectable at a pH of 6.0, which fits well with its importance in sodium homeostasis under alkaline conditions. These observations highlight the importance of this ATPase to sodium homeostasis under alkaline conditions. Finally, the mechanism for sodium (or proton) extrusion by V₀V₁-ATPases involves coupling ATP hydrolysis to cation extrusion through rotational catalysis (for a review, see (Murata, Yamato, & Kakinuma, 2005)).

Potassium

Potassium is the major intracellular cation. For E. hirae cytosolic concentrations range from 0.4 to 0.6 M. Potassium is essential for cellular metabolism—it neutralizes intracellular anions, activates diverse enzymes, and regulates cytosolic pH. Maintenance of high intracellular concentrations, especially when potassium is in limited supply, requires active uptake. E. hirae expresses at least three potassium transporters to perform this task: KtrI, KtrII, and a low-affinity transporter (Kawano, Igarashi, & Kakinuma, 1999; Kawano M. , Abuki, Igarashi, & Kakinuma, 2001). In addition, there is the Kep system for potassium extrusion (Kakinuma & Igarashi, 1988).

The primary potassium uptake system for E. hirae is KtrI (in E. faecalis, kdpABC–EF0567, EF0568, and EF0569). This activity is likely due to symport of K⁺ and H⁺, although this has yet to be directly demonstrated (Bakker & Harold, 1980). The pH values optimal for KtrI is 6 to 7, with an apparent K_m of 0.2 mM for K⁺, which allows KtrI to establish a gradient from inside to outside of 10⁵ (Bakker & Harold, 1980). KtrI is regulated by an ATP-dependent modification, requires a membrane potential, and is active under neutral or acidic, but not alkaline conditions. These features are strikingly similar to the Trk K⁺ potassium transport system of E. coli, which also has a low affinity, high-capacity mechanism for K⁺ accumulation, coupled to ATP hydrolysis and a membrane potential (Epstein, 2003). KtrI activity appears to be constitutive, although purification, cloning, and analysis of regulatory mechanisms are not reported.

A second potassium uptake system for E. hirae is KtrII. Unlike KtrI, this transporter does not require a membrane potential or ATP. The K_m for K⁺ is 0.5 mM and its optimal pH value is near 9. KtrII is not constitutively expressed, but is instead induced by high intracellular concentrations of sodium. This uptake system operates under conditions that otherwise render KtrI nonfunctional, such as a high pH. KtrII requires an integral membrane protein, called NtpJ. This protein exhibits strong similarity to KtrB. Interestingly, ntpJ is the final open reading frame in the V₀V₁-ATPase ntpFIKECGABDJ operon, but is not a functional subunit of this sodium transporter (Murata, Takase, Yamato, Igarashi, & Kakinuma, 1996). Studies with E. hirae strains that are defective in NtpJ suggest that this protein mediates not just K⁺, but also Na⁺, uptake. Indeed, KtrII may be a major reentry pathway for Na⁺ under alkaline conditions (Kawano M. , Abuki, Igarashi, & Kakinuma, 2000). A third low affinity/high rate K⁺ transport system has been reported that functions at pH values greater than 10, although corresponding genes remain to be identified (Kawano M. , Abuki, Igarashi, & Kakinuma, 2001).

Kep is a potassium expulsion system that exports K⁺ against a concentration gradient in exchange for H⁺. The Kep K⁺/H⁺ antiporter is constitutively expressed, but only functions at alkaline pH values. K⁺ extrusion stops at pH values less than 8. Mutants defective in antiport are unable to grow at a pH value greater than 8.5 (Kakinuma & Igarashi, 1999). Thus, in an alkaline environment, this system regulates not only cytosolic K⁺, but also pH. K⁺/H⁺ antiport activity requires ATP, but attempts to detect differences in K⁺-activated ATPase in membrane vesicles of wild-type and mutant strains of E. hirae have been unsuccessful. Isolation of the Kep antiporter and cloning its gene has yet to be reported.

Copper

Copper is an essential cofactor in a large number of redox-active respiratory, metabolic, and stress enzymes, because of redox activity between Cu⁺ and Cu²⁺ oxidation states. This reactivity can be toxic, especially for enterococci that generate O₂^–, as well as derivative-reactive oxygen species, such as H₂O₂ and hydroxyl radicals. Copper, like iron, readily promotes free radical reactions by Haber-Weiss and Fenton chemistry. To minimize damaging oxidation from copper (or iron), cytosolic concentrations must be tightly regulated. This is achieved through proteins called copper chaperones and copper-ATPases that regulate uptake and export. Copper chaperones bind free copper to protect against redox toxicity and deliver this metal to target proteins. A system of copper homeostasis was first identified and characterized for E. hirae by Solioz and colleagues and is now recognized as being widely distributed among prokaryotic and eukaryotic organisms (Solioz & Stoyanov, 2003). E. hirae is a model organism for copper metabolism and has been the subject of intense study (see (Magnani & Solioz, 2005) and (Solioz & Stoyanov, 2003) for reviews). Human copper ATPases are remarkably similar to the E. hirae transport enzyme. Defects in these genes have been implicated in Menkes’ and Wilson’s diseases of copper metabolism.

The cop operon consists of four genes, copYZAB (EF0297 to EF0299 and EF0875; Figure 6). Both copA and copB genes encode transporters that are CPx-type ATPases (E.C. 3.6.1.) that pump heavy metals and belong to a larger class of P-type ATPase ion pumps (Lutsenko & Kaplan, 1995). The “P-type” refers to a phosphorylated intermediate that forms during the catalytic cycle and distinguishes these enzymes from V- and F₀F₁-type ATPases. CopA catalyzes the uptake of Cu⁺ under copper-limiting conditions. Since this ion is largely insoluble at a neutral pH, an extracellular copper reductase appears to generate Cu⁺ from Cu²⁺. CopZ presumably accepts Cu⁺ from CopA, although interactions between CopZ and CopA have not been directly demonstrated, nor have mechanisms been elucidated for Cu⁺ transfer between chaperone and ATPases or target proteins. CopB is another P-type ATPase that functions to export Cu⁺ or Ag⁺ from cytosol to the cell exterior.

Figure 6.

Copper metabolism. The E. hirae CopZ copper chaperone is central to copper homeostasis. Cu²⁺ is extracellularly reduced to Cu⁺ by a putative Cu²⁺-reductase prior to import via a P-type ATPase, termed CopA. CopZ transports bound Cu⁺ to target proteins (more...)

CopY is a copper-responsive repressor that binds an upstream promoter and regulates the expression of all other cop genes. CopZ is a 69-amino-acid hydrophilic protein that belongs to a family of copper chaperones that includes MerP, ATX1 from yeast, and HAH1 for humans (Rosenzweig & O'Halloran, 2000). Excess or limited copper induces cop operon expression (Solioz & Stoyanov, 2003). When cytosolic copper rises, two Cu⁺-CopZ molecules deliver copper to CopY, a copper-responsive repressor that controls the expression of cop genes. CopY is a homodimer that, when coordinated with Zn²⁺, binds to two distinct 28-bp sequences near the translational start of the cop operon. CopZ transfers Cu⁺ ions to CopY, which then releases Zn²⁺ and causes CopY to dissociate from the promoter. This results in operon induction (Solioz & Stoyanov, 2003). In contrast, to control cop expression during copper excess, copper-limited conditions do not cause the release of CopY from the promoter, and operon induction must occur through a second, as yet unknown mechanism. The unusual coregulation of copA and copB is believed to be necessary to protect the cell from copper toxicity by ensuring that an export pump is available to expel excess copper when the import pump is induced. In E. hirae, copper-dependent proteolytic degradation of CopZ is an additional novel mechanism for copper homeostasis (Magnani & Solioz, 2005).

Iron

Iron is an essential nutrient for most microorganisms, including enterococci (Lisiecki & Mikucki, 2006). Despite the exceedingly low concentration of free iron at physiological pH (10^-18 M) (Andrews, Robinson, & Rodríguez-Quiñones, 2003), acquisition is possible for enterococci, because these microorganisms synthesize and secrete iron-binding compounds called siderophores. Siderophores are a varied group of low-molecular-weight chelators that specifically bind Fe⁺³, and are made by cells in response to iron deprivation. Membrane-associated siderophore receptors complete a high affinity system for iron acquisition. Most enterococci tolerate iron deprivation through the secretion of linear trihydroxamate or citrate hydroxamate siderophores (Efthymiou, Saadi, Young, & Helfand, 1987; Lisiecki, Wysocki, & Mikucki, 2000; Maskell, 1980). Although the structural diversity of hydroxamate siderophores is enormous, all use ornithine as a common precursor. In addition to siderophore secretion, several strains of E. faecalis and E. faecium assimilate iron by transporting 2-oxo acids when complexed to Fe⁺³ (e.g., pyruvic acid or 2-oxo-3-methylvaleric acid) (Lisiecki & Mikucki, 2006).

Beyond limited work on siderophore secretion by enterococci, there has been no other report on iron acquisition by these microorganisms. The E. faecalis V583 genome database includes several iron uptake mechanisms that are homologous to other bacteria (Paulsen, et al., 2003). At least three putative operons are present. One encodes Fe-chelator ABC transporters (EC 3.6.3.34), another is homologous to feoA (EF0475) and feoB (EF0476) that encode for ferrous iron uptake, and a third composed of fur (ferric uptake regulator)-like sequences. The ABC transporters feuA (EF0188) and fatB (EF3082), as well as the ferrous transporter feoB, all appear to be involved in iron acquisition during growth in blood (Vebø, Snipen, Nes, & Brede, 2009). While experimental evidence for the role of these genes in iron metabolism remains to be determined, E. faecalis appears to encode the necessary machinery for both ferric and ferrous iron transport.

Manganese

Manganese is a trace metal with variable oxidation states that serves as a cofactor for numerous enzymes and regulators in metabolic pathways, signal transduction, and response to oxidative stress (Jakubovics & Jenkinson, 2001). In stark contrast to iron and copper, this ion does not catalyze Fenton chemistry to generate damaging hydroxyl radicals. This attribute has important implications for the resistance of enterococci to desiccation and ionizing radiation. The accumulation of high concentrations of intracellular manganese in E. faecium results in a high manganese-to-iron ratio that renders the species nearly as resistant to γ-irradiation as the prototypic radiation-resistant Deinococcus radiodurans (Daly M. J., et al., 2004). This degree of radiation tolerance is 10-fold greater than for most other bacteria and involves intracellular manganese protecting DNA repair proteins from oxidative damage caused by ionizing radiation (Daly M. J., et al., 2007). This mechanism helps explain how non-spore-forming enterococci (and Deinococcus spp.) are also able to survive extreme oxidative stresses associated with desiccation (Krisko & Radman, 2013).

Manganese can also modulate the virulence of enterococci. The E. faecalis efaCBA operon encodes a putative ATP-binding cassette transporter that is regulated by manganese through EfaR, a Mn⁺⁺-responsive transcriptional regulator (Low, Jakubovics, Flatman, Jenkinson, & Smith, 2003). Genome-wide analysis of E. faecalis V583 that was grown in high concentrations of manganese shows the induction of many genes, including numerous transporters (Abrantes, Lopes, & Kok, 2011). Finally, EfaA is a lipoprotein component of this transporter, and was highly expressed in patients with E. faecalis endocarditis. A knockout of this gene was associated with delayed mortality in mice (Singh, Coque, Weinstock, & Murray, 1998). Additional investigations of manganese physiology are needed to help better define its role in enterococcal stress responses and virulence.

Lactate

Lactate is the predominant fermentative end-product and is one of the most abundant ions secreted for enterococci under anaerobic conditions with excess glucose. Lactate, with a pK_aʹ of 3.8, is anionic at all metabolic pH values and thus cannot freely pass the cell membrane. As a result, lactate efflux is necessary and occurs in symport with protons by a carrier-mediated process that translocates the protonated species, lactic acid (Harold & Levin, 1974). Recent evidence using ¹H-NMR techniques that distinguish between intracellular and extracellular lactate compartments shows the concentration of free cytosolic lactate in E. faecalis is in exact balance with the proton electrochemical-potential gradient over a wide pH range (Hockings & Rogers, 1997). These data identify a pool of tightly bound intracellular lactate at high external pH values and dispel any notion that end-product efflux is an energy-yielding process.

Purine Biosynthesis

Despite a lack of the biochemical characterization of purine biosynthesis in enterococci, some information on this pathway is available in the E. faecalis genome database (Paulsen, et al., 2003). E. faecalis purine synthesis is similar in reaction schema to that of other prokaryotes. The first step in purine biosynthesis is catalyzed by amidophosphoribosyltransferase (PurF, E.C. 2.4.2.14–EF1781) that forms 5-phosphoribosylamine from 5-phospho-α-D-ribose-1-diphosphate and L-glutamine. The E. faecalis PurF amino acid sequence is 62% identical to PurF for B. subtilis (Makaroff, Zalkin, Switzer, & Vollmer, 1983). After the formation of 5-phosphoribosylamine, biosynthesis continues through ten additional reactions, until the central intermediate inosine monophosphate (IMP) is formed. Purine biosynthesis branches from IMP into pathways that produce guanosine and adenosine monophosphate (GMP and AMP). The genes for purine biosynthesis resemble the pur operon in B. subtilis, with many encoded in a single putative transcriptional unit (EF1787 to EF1777).

Pyrimidine Biosynthesis

The pyrimidine biosynthetic pathway is similar in all bacteria and begins with glutamine, bicarbonate, and ATP. A single branch point occurs at the initial reaction where carbamoyl phosphate, an intermediate in arginine biosynthesis and catabolism, is formed by carbamoyl phosphate synthetase (CarAB or aspartate transcarbamoylase–EF1716). The E. faecalis carbamoyl phosphate synthetase is unusual when compared to the enzymes for other bacteria, because ATP is not activating and pyrimidines are not inhibitory. Instead, this enzyme has an allosteric activator site that is sensitive to many anions, including substrates and products of the reaction (Chang & Jones, 1974).

Genes for E. faecalis pyrimidine biosynthesis were first characterized during a study of enterococcal virulence factors (Li, Weinstock, & Murray, 1995). Unlike E. coli, where pyrimidine biosynthesis genes are spread throughout the chromosome, E. faecalis pyr genes are clustered (EF1715-EF1721) with coordinate regulation that is similar to the B. subtilis pyr operon (Turnbough, Jr. & Switzer, 2008). The 5ʹ end of the pyr operon encodes a regulatory protein, termed PyrR, which attenuates pyr transcription by binding its own mRNA transcript (Ghim, et al., 1999). In comparison, there are three PyrR binding sites in the B. subtilis pyr operon. Exogenous uracil represses the pyr operon, while UMP stimulates PyrR binding to a conserved anti-antiterminator sequence in the 5ʹ leader mRNA.

The oxidation of dihydroorotate to orotate is an intermediate reaction in the biosynthesis of UMP from aspartate and carbamoyl phosphate (Turnbough, Jr. & Switzer, 2008). Dihydroorotate dehydrogenase catalyses the formation of this aromatic intermediate and is the only redox reaction in the pathway. Two families of dihydroorotate dehydrogenase enzymes are known. One is membrane-associated, uses ubiquinone as an electron acceptor, and is found in Gram-negative bacteria and eukaryotic cells. The other family consists of two enzymes (types A and B) described for Gram-positive bacteria, including E. faecalis. The E. faecalis type A dihydroorotate dehydrogenase is a homodimer encoded by pyrD-1 (EF0285), and uses fumarate, oxygen, and quinones as electron acceptors (Marcinkeviciene, et al., 2000). In contrast, the E. faecalis type B dihydroorotate dehydrogenase (pyrD-2–EF1714) is a heterotetramer, with single FMN, FAD, and Fe-S redox sites. Since these enzymes are essential for bacterial growth, effective inhibitors are potential candidates for drug development.

The stringent response

The stringent response is a global cellular signalling mechanism in which bacteria adapt to adverse environmental conditions, such as nutrient starvation. This response is mediated by an accumulation of alarmones. These small molecules are synthesized by the phosphorylation of GDP and GTP, using ATP as the pyrophosphate donor, and include guanosine-5ʹ-diphosphate-3ʹ-diphosphate (ppGpp) and guanosine-5ʹ-triphosphate-3ʹ-diphosphate (pppGpp). In E. coli, RelA and SpoT are two proteins with alarmone synthase activity (Dalebroux, Svensson, Gaynor, & Swanson, 2010). The SpoT protein additionally functions as a (p)ppGpp hydrolase. E. faecalis harbors a homologue to the bifunctional SpoT enzyme, termed RelA, and another small monofunctional RelA-like synthase fragment, termed relQ (EF1974 and EF2671, respectively). In two E. faecalis strains, OG1RF (Abranches, et al., 2009) and V583 (Yan, et al., 2009), RelA is responsible for (p)ppGpp accumulation. In contrast, RelQ appears responsible for maintaining basal levels of alarmone in OG1RF, but not V583, and may be important for the timely activation of the stringent response (Gaca, Abranches, Kajfasz, & Lemos, 2012).

E. faecalis accumulates (p)ppGpp during amino acid starvation, heat, and alkaline shock, but not following exposure to acid pH (Abranches, et al., 2009), high osmolarity, or hydrogen peroxide (Yan, et al., 2009). A strong stringent response is observed in cells treated with vancomycin, a glycopeptide antibiotic that blocks cell-wall synthesis, but not with ampicillin, which also inhibits cell-wall synthesis (Abranches, et al., 2009). relA mutants of E. faecalis strains OG1RF and V583 show adverse growth effects under high osmolarity. The relA mutant for OG1RF, but not V583, is also sensitive to growth inhibition at a pH value of 5.0 and with 2 mM H₂O₂. In a relAQ double mutant that lacks (p)ppGpp, wild-type resistance to high osmolarity and low pH is restored for strain OG1RF and the double knockout is more tolerant to H₂O₂ stress. These OG1RF rel-mutants show less biovolume and decreased long-term survival in biofilms, in comparison to wild-type strains (Chávez de Paz, Lemos, Wickström , & Sedgley, 2012). In contrast to this enhanced sensitivity to environmental stress, OG1RF relA mutants grew faster with subinhibitory concentrations of vancomycin and showed a higher tolerance to this glycopeptide in time-course experiments compared to a wild-type strain. On the other hand, the relAQ double mutant and, to a lesser extent the relQ mutant, were more sensitive in these experiments. Of note, the relA mutant of strain V583 was slightly more resistant to lethal concentrations of ethanol, acid, and bile salts in comparison to the parent and relA complemented strains. The virulence of the relAQ double mutant (but not single mutants) was significantly attenuated in Caenorhabiditis elegans (Abranches, et al., 2009) and Galleria mellonella infection models (Yan, et al., 2009; Gaca, Abranches, Kajfasz, & Lemos, 2012), and survived less well in murine macrophages (Gaca, Abranches, Kajfasz, & Lemos, 2012).

These results indicate that E. faecalis (p)ppGpp alarmones are key for adaption to diverse stresses, resistance and tolerance to glycopeptides, and virulence. Like other bacteria, the stringent response triggers a complex re-programming of gene expression in E. faecalis, such as repression of genes associated with cell growth and the replication and activation of genes involved in amino acid biosynthesis, nutrient transport, and stress survival (Gaca, Abranches, Kajfasz, & Lemos, 2012). However, the physiological behavior of relA, relQ, and relAQ mutants is complex and without an obvious correlation between the RelA-dependent accumulation of (p)ppGpp and stress resistance or antibiotic tolerance. Rather, it seems the inability of the relA mutant to hydrolyse (p)ppGpp (as synthesized by RelQ), which potentially increases basal levels of (p)ppGpp, is responsible for the sensitization of E. faecalis to sublethal stresses and might increase resistance or tolerance to antibiotics (Abranches, et al., 2009). Therefore, to better understand the relationship between the stringent responses, stress responses, and antibiotic tolerances, study of the additional mutants that are only affected in one of the two RelA activities might be helpful. Recently, a relA mutant with a large C-terminal deletion named relAsp was characterized. A mutant expressing this truncated protein accumulated (p)ppGpp during amino acid starvation. In addition, this mutant was resistant to environmental stresses and was more virulent in the G. mellonella model of enterococcal infection (Yan, et al., 2009).

Viable but nonculturable state

Among enterococci, the extremes of temperature, starvation, osmotic concentration, and solar radiation can induce a metabolic state that is termed viable but nonculturable (VBNC). Bacteria with this phenotype do not grow on laboratory media, but are otherwise alive and capable of resuscitation. This metabolic state was initially reported for E. coli and Vibrio cholera, but has since been described for many eubacteria, including E. faecalis, E. faecium, and E. hirae (Oliver, 2010). The VBNC phenotype is a primary survival strategy for bacteria in natural environments. Cells in the VBNC state have lowered metabolism, maintain cell membranes, contain high levels of intracellular ATP, and continue gene expression with the persistence of mRNAs for as long as 50 days. Metabolism during VBNC results in distinctive protein profiles for E. faecalis. E. faecalis, E. faecium, and E. hirae enter the VBNC state after approximately two to six weeks of exposure to an inducing stress, show extensive cross-linking of cell wall peptidoglycan, and increase the expression of autolysins (Pfeffer, Strating, Weadge, & Clarke, 2006). This stress response and key survival strategy results in a general increase in antibiotic resistance, although antibiotics also appear capable of inhibiting resuscitation when cells reenter favorable environments (Lleò, Benedetti, Tafi, Signoretto, & Canepari, 2007). The VBNC phenotype likely renders routine culture techniques inadequate for accurately monitoring enterococci in environmental samples (Signoretto & Canepari, 2008).The potential role of the VBNC state in nosocomial transmission of enterococci, however, has yet to be investigated. Finally, although the expression of stress regulators, DNA polymerases, and virulence factors have been reported during the VBNC state for Gram-negative pathogens, including E. coli, Helicobacter pylori, Vibrio vulnificus, and V. cholerae O1 (Oliver, 2010), the metabolic and genetic basis for this phenotype in enterococci remains unknown.

Extracellular superoxide

The production of extracellular O₂^– by “group D Streptococcus” was initially described by Falcioni and colleagues in 1981 (164). These observations were extended to several other species of enterococci and lactococci (Winters, Schlinke, Joyce, Glore, & Huycke, 1998; Huycke, Joyce, & Gilmore, 1995), as well as Gram-negative bacteria, to a more limited extent (Korshunov & Imlay, 2006). O₂^– results from the univalent reduction of O₂. As an anion, this radical is impermeant to passive diffusion and remains extracellular when generated on the outer side of a cell membrane. Although O₂^– is relatively non-reactive, it can readily dismute to form more powerful oxidants, such as H₂O₂ and a hydroxyl radical. Disruption of the pathway for demethylmenaquinone synthesis blocks extracellular O₂^– production by E. faecalis and suggests membrane-associated quinones as the source of these radicals (Huycke M. M., et al., 2001). The most probable mechanism involves univalent oxidation of reduced quinone, or quinol-to-labile semiquinone radicals that spontaneously react with O₂ to form O₂^– (Figure 7).

Figure 7.

Extraceulluar superoxide production. Model of extracellular O₂^- production by E. faecalis demethylmenaquinone. Cytosolic reducing equivalents transfer to demethylmenaquinone (left) through oxidoreductases that form demethylmenaquinol (right). Normally, (more...)

For E. faecalis, functional cytochrome bd suppresses the production of extracellular O₂^– (Huycke M. M., et al., 2001). Thus, when these bacteria are grown with heme, the sole apo-cytochrome expressed by this species becomes functional and radical production is abolished. Similarly, inactivation of the cytochrome bd by gene knockout or withholding heme is permissive to extracellular O₂^–. Presumably, actively cycling cytochrome bd efficiently reoxidizes demethylmenaquinol and minimizes semiquinone reactivity with O₂. Frd is another E. faecalis quinol oxidase that also attenuates extracellular O₂^– production. This effect, however, only occurs in the presence of fumarate, the substrate for this membrane-bound enzyme (Huycke M. M., et al., 2001). Electron-spin resonance studies of intestinal contents and the detection of hydroylated isomers of tyrosine in rodents colonized with E. faecalis indicate that conditions exist in the mammalian intestinal tract that allow for O₂^– formation (Huycke & Moore, 2002; Moore, Kotake, & Huycke, 2004). These observations show that commensals can be potential sources for oxidant stress on the intestinal epithelium. Indeed, mice colonized with E. faecalis develop colonic epithelial cell DNA damage (Huycke, Abrams, & Moore, 2002; Wang & Huycke, 2007). Oxidative stress from E. faecalis also causes chromosomal instability in mammalian cells (Wang & Huycke, 2007; Wang, et al., 2008). Finally, interleukin-10 knockout mice colonized with these bacteria develop inflammation and colorectal cancer (Balish & Warner, 2002; Kim, et al., 2005). The formation of tumors in these mice depends on extracellular O₂^– production (Wang, et al., 2012). These findings have profound implications for the role of intestinal commensals in general, and enterococci in particular, in inflammatory bowel disease and colorectal cancer (Sinicrope, 2007).

Manganese superoxide dismutase

The univalent reduction of O₂ to produce O₂^– is a common minor byproduct of numerous intracellular sources that include autoxidizable small molecules, oxidoreductive enzymes, and subcellular organelles, like mitochrondria (Fridovich, 1999). If not properly scavenged, O₂^– can damage thiols, tetrahydropterins, ascorbate, poly-unsaturated fatty acids, [4Fe-4S] clusters in dehydratases, and DNA. O₂^– can also form other damaging oxygen and nitrogen compounds, like H₂O₂, hydroxyl radical, and peroxynitrite. Superoxide dismutase (SOD) and peroxidases are the primary defenses against the cascade of oxidative injuries initiated by O₂^–.

Virtually all facultative and aerobic organisms express one or more types of SOD. SOD rapidly converts O₂^– to H₂O₂ and molecular oxygen. Depending on the cofactor, these metalloenzymes are classified as iron, manganese, or copper-zinc SODs. Iron and manganese SODs share a common ancestor, while copper-zinc enzymes appear to have evolved independently (Zelko, Mariani, & Folz, 2002) and, among bacteria, are most often found in the periplasm of Gram-negative organisms. E. faecalis contains a single manganese SOD (SodA, EF0463) that is induced by O₂ (Britton, Malinowski, & Fridovich, 1978). No evidence exists for iron or copper-zinc forms. Mutants in sodA are more sensitive to oxidative stress and show reduced resistance to killing by macrophages (Verneuil, et al., 2006). E. faecalis is highly tolerant to many bactericidal drugs. Recently, SodA was functionally shown to eliminate the bactericidal activity of penicillin and vancomycin (Bizzini A. , Zhao, Auffray, & Hartke, 2009). This clinically important finding is presented in further detail at the end of this chapter (see the section on the relation between oxidative stress and antibiotic tolerance).

NADH oxidase

E. faecalis expresses an unusual FAD-dependent cytosolic enzyme that catalyzes the direct reduction of oxygen to H₂O. NADH oxidase (Nox, EC 1.6.99.3–EF1586) completes this reaction through a four-electron reduction of O₂, without the release of any O₂^- or H₂O₂ (Schmidt, Stöcklein , Danzer, Kirch, & Limbach, 1986). This reaction requires NADH as the preferred electron acceptor. The main physiological role for Nox is to regenerate NAD⁺ for glycolysis. This allows the synthesis of additional ATP by converting pyruvate to acetate instead of lactate. Since O₂ is consumed by this reaction, Nox can also act as an antioxidant. A Nox homologue is present in E. faecium, but appears to be absent in E. casseliflavus and E. gallinarum. The E. faecalis Nox is a homodimer containing one FAD and, like the NADH peroxidase (Npr, see below), a nonflavin cysteine-sulfenic acid (Cys-SOH) redox center (Mallett & Claiborne, 1998). The nox gene for E. faecalis is 44% identical to npr (see below) and, as with NADH peroxidase, conserves the Cys-SOH at position 42, which suggests similarity in catalysis. Coordinate regulation of NADH oxidase has not been described, although this enzyme is induced by oxygen and is partially repressed by exogenous heme (Pugh & Knowles, 1982). The potential role for Nox in enterococcal pathogenesis remains to be investigated.

Glutathione

Few Gram-positive bacteria synthesize glutathione (γ-GluCysGly, GSH), a thiol tripeptide that protects against oxidative stress and serves as an essential cofactor for many metabolic reactions (Masip, Veeravalli, & Georgiou, 2006). Instead, most bacteria simply import GSH. E. faecalis and E. faecium, however, can not only import GSH, but can also synthesize it de novo (Gopal, et al., 2005; Newton, et al., 1996). GSH is usually made by two enzymes in separate steps: γ-glutamylcysteine formation from glutamate and cysteine, and glutathione formation by the subsequent addition of glycine. In enterococci, however, these reactions appear to occur through an unusual bifunctional enzyme that exhibits both γ-glutamylcysteine synthetase and glutathione synthetase activity (Gopal, et al., 2005; Janowiak, Hayward, Peterson, Volkman, & Griffith, 2006). An enzyme with similar activity has been isolated from S. agalactiae, Listeria monocytogenes, and Pasteurella multocida (Gopal, et al., 2005). A homologous gene termed gshAB (or gshF–EF3089) is annotated in genome databases for E. faecalis and E. faecium. However, the enterococcal gene product remains to be characterized.

Many enzymes utilize GSH as a cofactor, including glutathione peroxidase, glutathione S-transferases, and glutathione reductase. E. faecalis expresses glutathione reductase (Gor, EF3270), an FAD-containing enzyme, and a poorly characterized glutathione peroxidase that scavenges H₂O₂ (Patel, Marcinkeviciene, & Blanchard, 1998). A survey of available genomic sequences indicates that genes that encode proteins with amino acid identities of 34% (E. casseliflavus, E. gallinarum) to 47% (E. faecium) to EF1211 are present in other enterococcal species. In contrast, Gor is conserved among enterococci and catalytically reduces GSSG to GSH. The reaction equilibrium for glutathione reductase is such that cytosolic ratios of GSH to GSSG remain high, which provides an anaerobic-like cytosolic environment. As such, glutathione reductase serves a primary role in the defense against oxidants. Glutathione reductase is induced by O₂, although aerobiosis does not significantly change the concentration of intracellular GSH, which suggests that GSH synthesis is not coordinately regulated.

Catalase

Catalase is a cytosolic hemoprotein that catalyzes the dismutation of H₂O₂ to molecular oxygen and water. Catalase (KatA, EF1597) activity is detected in E. faecalis during aerobic growth in the presence of heme (Whittenbury, 1964; Pugh & Knowles, 1983), but a homologous gene is absent in E. faecium, E. casseliflavus and E. gallinarum. Because E. faecalis cannot synthesize heme, and because the incorporation of heme into apoenzyme KatA (apo-KatA) is O₂ dependent, KatA only becomes functional when exogenous heme is provided. In the absence of heme, KatA remains a nonfunctional apoenzyme. As is typical for most catalases, KatA is inhibited by azide and cyanide (Whittenbury, 1964). The katA gene and its corresponding protein have been isolated and characterized for E. faecalis (Frankenberg, Brugna, & Hederstedt, 2002; Baureder, Reimann, & Hederstedt, 2012).

NADH Peroxidase

E. faecalis NADH peroxidase (Npr, EF1211), like KatA, catalyzes the decomposition of H₂O₂ to molecular oxygen and water. A survey of available genomic sequences indicates that genes that encode proteins with amino acid identities of 34% (E. casseliflavus, E. gallinarum) to 47% (E. faecium) to EF1211 are present in other enterococcal species. This activity was originally considered to be due to a “pseudocatalase,” because the responsible protein did not contain heme. The nonheme nature of Npr renders it insensitive to azide and cyanide, and permits continued enzymatic activity, despite exposure to high concentrations of H₂O₂ (Whittenbury, 1964). Npr is a flavoprotein disulfide reductase that is similar to lipoamide dehydrogenase, glutathione reductase, and thioredoxin reductase (Claiborne, et al., 1999). A peroxidase with similar reactivity has been described in E. hirae (Miller, Poole, & Claiborne, 1990), although this particular species does not encode an npr gene with sequence homology to that of E. faecalis or other enterococci. In the absence of functional KatA, Npr is absolutely required for aerobic growth on glycerol and for maximal growth on lactose, galactose, or ribose (La Carbona, et al., 2007). In addition, Npr contributes to the survival of enterococci against exogenous H₂O₂ (La Carbona, et al., 2007) generated, for example, by other facultative or anaerobic bacteria in the intestine (Gordon, Holman, & McLeod, 1953). Sequence analysis of npr reveals highly conserved segments that correspond to three of the four structural domains for glutathione reductase (Ross & Claiborne, 1992). A putative OxyR protein in E. faecalis that binds to an upstream region of npr potentially assists with coordinate regulation of genes following oxidative stresses (Ross & Claiborne, 1997).

Peroxiredoxins

A serious consequence of oxidative stresses is the formation of organic hydroperoxides that subsequently initiate free-radical chain reactions, which lead to DNA damage. Peroxiredoxins use cysteine thiols to detoxify such harmful peroxides and thereby protect cells from mutation. This is especially important for E. faecalis, as it generates endogenous ROS through multiple mechanisms. The first peroxiredoxin system was characterized in Salmonella typhimurium (Aph) (Jacobson, Morgan, Christman, & Ames, 1989). The catalytic site for reductase activity in this bacterium (AhpC, EF2739) contains cysteines that react with peroxides to yield corresponding alcohols. AhpCs reduce a broad range of peroxides from H₂O₂ to complex organic hydroperoxides. The activities of these enzymes depend on small FAD-containing thioredoxin reductases (AhpF, EF2738) that use NADPH to recycle the reductase. The E. faecalis ahpCF operon is expressed following oxidative stress and controlled by the hydrogen peroxide regulator (HypR–EF2958) that also regulates genes for SodA, KatA, glutathione reductase, and thiol peroxidase (Verneuil, et al., 2004). Notably, HypR does not contain cysteines that are commonly used to sense oxidative stress (Verneuil, et al., 2004). This transcriptional regulator for oxidative stress is therefore functionally unrelated to OxyR. Finally, AhpC and AhpF represent important defenses against oxidative stresses, as shown for E. faecalis when grown on glycerol (La Carbona, et al., 2007). Corresponding homologous operons are also present in other sequenced enterococcal species.

Thiol peroxidases (Tpx; EF2932) are members of another peroxidase class that confers broad substrate specificity. The E. faecalis Tpx is typical of members in this class and contains two cysteine residues that align with catalytically important residues found in the E. coli Tpx (La Carbona, et al., 2007). For E. faecalis, this antioxidant enzyme appears more important than Npr or AhpC for survival in macrophages and in promoting virulence using a murine peritonitis model. Homologous enzymes with approximately 60% amino acid identity are present in E. casseliflavus and E. gallinarum genome databases, but are lacking for E. faecium.

Organic hydroperoxide resistance (Ohr) comprises a third class of peroxiredoxins that, similar to AhpC, contain invariant cysteine residues that help reduce peroxides. Ohr genes occur in only a few eubacteria, including E. faecalis (Ochsner, Hassett, & Vasil, 2001). In the genome database for E. faecalis V583, two genes (EF0453 and EF3201) encode Ohr family proteins. EF0453 appears to be a general stress response protein, termed Gsp65, with significant homology to Ohr proteins from other organisms (Rincé, Giard, Pichereau, Flahaut, & Auffray, 2001). Knockout of gsp65 is deleterious to E. faecalis, with mutants showing increased susceptibility to tert-butylhydroperoxide and ethanol. Proteins with approximately 50% amino acid identity to EF0453 are present in other sequenced enterococci.

Relationship between oxidative stresses and antibiotic tolerance

The most effective antimicrobials function to rapidly kill pathogens, not just inhibit them, and thereby limit the severity of infections and emergence of resistance. Enterococci, however, are intrinsically tolerant to many bactericidal drugs, including β-lactams and glycopeptides. Mechanisms that confer protection against the lethal effects of these drugs have recently been investigated (Bizzini A. , Zhao, Auffray, & Hartke, 2009; Ladjouzi, et al., 2013). Singly- and multiply-deficient mutants of the E. faecalis strain JH2-2 affected in oxidative stress defense activities or DNA repair were screened for loss of tolerance to bactericidal antibiotics. The only mutant that was efficiently killed by vancomycin or penicillin (but not by bacteriostatic drugs) was deficient in manganese SOD (Bizzini A. , Zhao, Auffray, & Hartke, 2009). This dependence of tolerance on active SOD was confirmed for another E. faecalis strain (OG1RF) and the manganese SOD (SodA) of E. faecium also appeared to be key for tolerance to bactericidal antibiotics (Ladjouzi et al. 2013). The combined results implied that these drugs increase intracellular superoxides. Although SodA efficiently detoxifies O₂^- to H₂O₂, this latter ROS must be less problematic, even in peroxidase- or catalase-deficient strains, since mutants with deficiencies in these enzymes were still tolerant to these antibiotics. In sodA mutants, however, O₂^- presumably accumulates to toxic levels that damage cellular targets and cause cell death. The mechanisms by which bactericidal antibiotics promote intracellular superoxide production remain to be determined. Finally, E. faecalis can also generate extracellular superoxide through the leakage of electrons from membrane-associated demethylmenaquinone (Huycke M. M., et al., 2001). However, this site of O₂^- production does not seem to play a role in the bactericidal activity of antibiotics, since mutants that lacked demethylmenaquinone (i.e., extracellular O₂^- production) along with SodA were no less sensitive to these antibiotics than the ΔsodA mutant alone (Bizzini A. , Zhao, Auffray, & Hartke, 2009).