SUMMARY
Bacterial sphingomyelinases and phospholipases are a heterogeneous group of esterases which are usually surface associated or secreted by a wide variety of Gram-positive and Gram-negative bacteria. These enzymes hydrolyze sphingomyelin and glycerophospholipids, respectively, generating products identical to the ones produced by eukaryotic enzymes which play crucial roles in distinct physiological processes, including membrane dynamics, cellular signaling, migration, growth, and death. Several bacterial sphingomyelinases and phospholipases are essential for virulence of extracellular, facultative, or obligate intracellular pathogens, as these enzymes contribute to phagosomal escape or phagosomal maturation avoidance, favoring tissue colonization, infection establishment and progression, or immune response evasion. This work presents a classification proposal for bacterial sphingomyelinases and phospholipases that considers not only their enzymatic activities but also their structural aspects. An overview of the main physiopathological activities is provided for each enzyme type, as are examples in which inactivation of a sphingomyelinase- or a phospholipase-encoding gene impairs the virulence of a pathogen. The identification of sphingomyelinases and phospholipases important for bacterial pathogenesis and the development of inhibitors for these enzymes could generate candidate vaccines and therapeutic agents, which will diminish the impacts of the associated human and animal diseases.
INTRODUCTION
Cellular membranes are dynamic structures which in the lateral dimension form lipid domains that selectively allow recruitment, clustering, and interactions of particular proteins, affecting their conformation and thus serving as signaling platforms (1). Several membrane glycerophospholipid- and sphingolipid-derived metabolites generated by cellular sphingomyelinases (SMases) and phospholipases (PLases) in response to extracellular signals play key roles in regulating lipid domain formation and intracellular vesicle trafficking (2). Some products generated by those enzymes, such as 1,2-diacylglycerol (DAG) and ceramide (Cer), change biophysical membrane properties, including charge, fluidity, and permeability (3, 4), and can recruit cytosolic proteins that induce spatial reorganization of signaling complexes, which in turn affect diverse cellular processes (5). For example, DAG, generated by cellular PLase C (PLC), plays roles in controlling cell proliferation and differentiation (6), while Cer, generated by cellular SMases, has been implicated in regulating ion transport, stress responses, cell cycle arrest, autophagy, apoptosis, and cytokine production (7). Since eukaryotic cellular membranes are interaction points with microorganisms, lipid-metabolizing enzymes produced by bacteria, such as SMases and PLases, could generate lipid-derived signaling metabolites identical to those produced by eukaryotic enzymes. Although bacterial SMases and PLases usually exert their enzymatic activities on the extracellular leaflet of the plasma membrane or on the luminal leaflet of membranes from the endolysosomal compartment, DAG and Cer could undergo spontaneous transbilayer movement, flipping to the cytosolic leaflet of those membranes and perturbing diverse cellular signaling processes (8–11).
Bacterial SMases and PLases constitute a structurally and evolutionary heterogeneous group of lipolytic esterases, usually secreted or surface associated, that are expressed by extracellular, vacuolar, and cytosolic pathogens from a variety of phylogenetic groups (Table 1). Most of the genes encoding these bacterial enzymes are chromosomally encoded, albeit a few of them are present in mobile genetic elements. Some enzyme types have orthologues in eukaryotes, whereas others are exclusively present in bacteria, although in phylogenetically distant bacterial lineages (Table 1). The patchy distribution of the genes encoding some of these enzymes across broad taxonomic boundaries suggests either an ancient origin or the occurrence of multiple horizontal gene transfer events among different phyla or even between bacteria and species from other kingdoms. According to the latter, horizontal gene transfer is recognized as playing a substantial role in the adaptive expansion of many protein families during prokaryotic genome evolution (12, 13).
Distribution of SMases and PLases which play a role in virulence among bacterial phyla
In several disease models, mutant bacterial strains lacking a SMase- or a PLase-encoding gene have impaired virulence, demonstrating conclusively the role of these enzymes in pathogenicity (Table 2). Bacterial SMases and PLases might contribute to the infection process in different hosts in several ways. Some of these enzymes hydrolyze structural membrane lipids and cause lysis of the host cells, contributing to bacterial colonization and/or dissemination and providing nutrients for pathogen survival and replication (14, 15). Others generate bioactive lipids which activate endogenous mediators of cell death without inducing cellular lysis (15–17) (Fig. 1). Bacterial SMases and PLases may act on vacuolar lipids, contributing to transform phagosomes into a replication-permissive niche or causing phagosomal membrane disruption that allows the pathogen to reach the host cell cytoplasm and survive by evading the host immune response (Fig. 1) (14, 15, 18). Bacterial PLases from enteric pathogens degrade the phospholipid-rich mucus layer overlying the gastrointestinal mucosa, whereas those from lung pathogens cleave phospholipids from pulmonary surfactant, promoting nutrient release and contributing to tissue colonization, as well as in the case of surfactant cleavage to lung dysfunction. Some bacterial SMases and PLases aid bacterial survival without inducing evident tissue damage. In some cases, these enzymes mimic their host cell counterparts and alter sphingomyelin (SM) or glycerophospholipid metabolism homeostasis, changing the balance of the different signaling molecules and thus diverting cellular processes driven by lipids to the bacterium's benefit. Products generated by bacterial SMases and PLases could hijack vesicle trafficking pathways or dysregulate signaling cascades affecting the immune response and thus promote infection establishment or progression (Fig. 1) (14–18).
SMases and PLases for which there are in vivo data supporting a role in bacterial virulence
Roles of different bacterial SMases and PLases in virulence. S. enterica serovar Typhimurium injects SseJ, an SGNH esterase with PLA1 and GCATase activities, via a T3SS into the cytoplasm of the host cell. Once in the cytosol, SseJ binds to RhoA GTPase, triggering its GCATase activity, which increases the vacuole surface. R. typhi produces two patatin-like PLA2s, Pat1 and Pat2, which are secreted during host intracellular growth and help phagosome escape. S. pyogenes SlaA is a class I-like PLA2 that enters host epithelial cells in an actin-dependent manner and plays an important role in pathogen adhesion and cytotoxicity. P. aeruginosa injects ExoU, a patatin-like PLA2, through a T3SS into the cytoplasm of the host cell. Once translocated, ExoU becomes activated and acts toward several plasma membrane substrates, leading to cytoskeletal collapse and cellular necrosis. ExoU also activates several signaling pathways, such as the arachidonic acid cascade and a PAF receptor–NF-κB pathway that leads to inflammatory mediator production. L. monocytogenes produces the PI-PLC PlcA and the Zn2+ metalloPLC PlcB, which are required for bacterial escape into the cytosol from the single-membrane primary pathogen-containing vacuole, triggering preautophagosomal structure stalling, favoring bacterial escape from the host autophagic defense. L. monocytogenes PlcB also contributes to bacterial escape into the cytoplasm. Both Bacillus spp. and S. aureus secrete a SMase C that plays an important role in virulence by increasing Cer, which changes the physical properties of the membrane and could also participate in signal transduction leading to cell death. These two bacteria also produce PI-PLC, which can facilitate signal transduction and removal of GPI-anchored proteins from plasma membranes. C. perfringens alpha-toxin, a Zn2+ metalloPLC, is required for this bacterium's escape from the macrophage phagosome in early stages of infection. At high concentrations, this toxin disrupts the plasma membrane, causing cell lysis, but at low concentrations if internalized by the target cell it cleaves PC and SM at the membranes of the endolysosomal compartment, generating DAG and Cer, which trigger signaling pathways that lead to ROS production and cell death. N. gonorrhoeae secretes a PLD named NgPLD that by both generation of PA and Akt binding drives cell surface recruitment of CR3, membrane ruffling, and bacteria engulfment, modulating host cell signaling events required for bacterial invasion.
SPHINGOMYELINASES
Sphingomyelinases are phosphoric diester hydrolases that cleave sphingomyelins, the most abundant eukaryotic membrane sphingolipids. Sphingomyelins have a phosphocholine moiety attached to Cer, a sphingosine (or sphinganine) derivative with an N-linked fatty acid of variable length and saturation. Depending on the cleavage site, there are two different types of bacterial SMases (Fig. 2): SMase Cs (EC 3.1.4.12), which hydrolyze the ester bond between Cer and phosphorylcholine, and SMase Ds (EC 3.1.4.41), which hydrolyze the phosphodiester bond between Cer-1-phosphate and choline. Three structurally different types of SMases have been described in pathogenic bacteria as playing a role in virulence. Table 3 depicts a selected example of each enzyme type.
SMase cleavage sites in sphingomyelin and cellular activities of the reaction products. SMase Cs cleave SM, generating Cer and phosphorylcholine, while SMase Ds generate Cer-1-phosphate and choline. SMase Ds also have intrinsic lysophospholipase D activities and act on lysophosphatidylcholine to generate the signaling molecule LPA (see also Fig. 5).
Types of bacterial SMases that contribute to virulencea
Sphingomyelinase CsSMase C production has been reported to occur in species of the genera Bacillus, Listeria, Staphylococcus, and Mycobacterium and also several species of Helicobacter, Chlamydia, Pseudomonas, and Leptospira. The molecular masses of bacterial SMase Cs vary from 27 to 39 kDa for most of the enzymes, except for those of Pseudomonas and Leptospira, which have a molecular mass of 58 and 63 kDa, respectively, due to the presence of additional sequences at their C termini (19, 20). The three-dimensional structures of enzymes in this group (Bacillus cereus PDB ID 2DDT, Listeria ivanovii PDB ID 1ZWX, Staphylococcus aureus β-toxin PDB ID 3I5V, and Streptomyces griseocarneus PDB ID 3WCX) (21–23) reveal that, although sharing less than 17% sequence identity with mammalian DNase I, they adopt the DNase I fold, as do the mammalian neutral SMases (23). In fact, despite a change in substrate specificity (from an endonuclease to a phosphodiesterase), both enzyme types catalyze the same chemical reaction. These similarities allowed the elucidation of the catalytic mechanism as the active site cleft, and key residues in catalysis (the general bases His150 and His285), as well as a Mg2+ ion-binding site vital for substrate binding and catalysis, are conserved (Fig. 3) (21, 24).
Structure of S. aureus beta-toxin (PDB ID 3I5V). The structure is shown as a beige cartoon, with catalytically important residues and bound diacyl glycerol shown as beige and gray sticks, respectively. Mg2+ ions from B. cereus SMase C (PDB ID 2DDT) are superimposed and shown as light blue spheres, and phosphate from L. ivanovii Smase C (PDB ID 1ZWX) is shown as sticks.
Bacterial SMase Cs bind and lyse red blood cells from different species, and thus one of their proposed roles is to aid iron acquisition from heme groups (25). Hemolysis induced by SMase Cs occurs through hydrolysis of the erythrocyte membrane SM, and susceptibility of erythrocytes from different species correlates with their SM content (25). Treatment of sheep erythrocytes with B. cereus SMase C reduces cell membrane fluidity and leads to the self-assembly of Cer-rich domains that exclude cholesterol (25). In erythrocytes, the coalescence of these domains and the formation of interfaces between rigid and fluid domains causes fragility of the plasma membrane, leading to lysis (25, 26). In the apical brush border membrane, Cer domain formation induced by SMase Cs from commensal bacteria likely modulates the epithelial barrier and affects intestinal homeostasis (27).
Bacterial SMase Cs are cytotoxic to certain cell types, but the sensitivity of nucleated cells to SMase Cs does not correlate with the SM content of their plasma membrane (28–30). However, the Cer-metabolizing speed of each cell type can influence cellular susceptibility, as it has been shown that cells deficient in Cer glucosyltransferase-1 are more susceptible to the cytotoxic effect of B. cereus SMase C, likely because of a lack of Cer clearance (31). The increase in Cer caused by bacterial SMase Cs could induce cytotoxicity by changing the physical properties of the membrane or by triggering apoptosis through Cer-dependent pathways (30, 32). However, the mechanism for the selective cytotoxicity induced by SMase Cs is still unknown. Several examples of SMase Cs which play roles in virulence are described below.
B. cereus is an opportunistic pathogen that causes endophthalmitis, pneumonia, and septicemia, as well as food poisoning (33). B. cereus SMase C induces colon epithelial cell death, kills silkworms, and contributes to the virulence of B. cereus in a Galleria mellonella larval infection model (34, 35). B. cereus clinical isolates produce significant amounts of SMase C, invade the bloodstream after intraperitoneal injection in mice, and cause death, which can be prevented by specific antibodies against this enzyme (33). B. cereus SMase C plays a crucial role in avoiding the host innate immune response during the early stages of infection, because treatment of mouse macrophages with this enzyme results in Cer-rich domain formation and a reduction in membrane fluidity that interferes with phagocytosis (33). Immunization of mice with B. cereus SMase C or administration of specific antibodies against this enzyme prevent death caused by an intraperitoneal infection with a clinical B. cereus isolate (33). Furthermore, the SMase C inhibitor SMY-540 significantly reduces the lethality of an infection with a B. cereus SMase-producing strain (36). Soil B. cereus isolates which lack the SMase C gene are not pathogenic to mice, but a transformed strain carrying the SMase C gene is fully virulent in a murine septicemia model (33). Similarly, a homologous SMase C is required for full virulence of Bacillus anthracis after intratracheal inoculation in mice (37).
L. ivanovii is a facultative intracellular pathogen of ruminants. This bacterium produces a SMase which enables it to disrupt the phagosome membrane and escape into the cytosol, thereby promoting intracellular survival and replication (38). A SMase knockout strain of L. ivanovii had an impaired capacity to proliferate intracellularly and was less virulent for mice than the wild-type strain (38).
S. aureus causes skin and soft tissue infections in humans, as well as osteomyelitis, endocarditis, pneumonia, meningitis, and sepsis (39), whereas in animals it is the most common cause of intramammary infections in lactating females (40). Although traditionally not considered an intracellular pathogen, S. aureus can survive within a wide variety of mammalian cells, including professional phagocytes (40–42). S. aureus hlb encodes a SMase C also known as beta-toxin (Fig. 3; Table 3), which is expressed in large quantities in clinical isolates and in bovine mastitis-inducing strains (39, 40). Beta-toxin is cytotoxic to human keratinocytes and could thus promote colonization on the skin (43). It also inhibits interleukin-8 (IL-8) expression by endothelial cells, which decreases neutrophil transendothelial migration (44). Furthermore, beta-toxin is cytotoxic to polymorphonuclear leukocytes, monocytes, and proliferating T lymphocytes (24, 28), contributes to the phagosomal escape of S. aureus (42), and induces biofilm formation (45). The hemolytic and lymphotoxic activities of beta-toxin are linked to its SMase C activity (24), but its capacity to induce biofilm formation is catalysis independent (45). S. aureus mutant strains lacking hlb have reduced virulence in pneumonia and murine ear skin infection models (43, 46), and a mutant strain expressing a biofilm formation-deficient beta-toxin had a diminished pathogenicity in an endocarditis model in rabbits (45). Purified beta-toxin is cytotoxic for bovine mammary epithelial cells, and non-beta-toxin-producing strains have reduced virulence in a mouse model of intramammary infection (47). Furthermore, beta-toxin is also required for S. aureus supernatants to kill silkworm larvae (48).
In pulmonary infections, SM hydrolysis by S. aureus beta-toxin and Cer accumulation in epithelial cells could contribute to pathogenesis by inhibiting the function of the cystic fibrosis transmembrane conductance regulator Cl− channel, which leads to thick mucus production that clogs the airways and fosters bacterial growth (49). Furthermore, Cer production in the membrane of epithelial cells enhances ectodomain shedding of a wide variety of cell surface proteins through stimulation of the cellular shedding machinery (50). Shedding of syndecan-1, the major heparan sulfate proteoglycan of alveolar epithelial cells, generates a chemotactic signal that potentiates neutrophil migration and inflammation in the alveoli (46).
Leptospirosis is a zoonosis of global importance caused by pathogenic members of the genus Leptospira. Leptospires migrate through skin abrasions or mucosal surfaces into the circulation and can cause severe systemic infections, leading in the severest forms to acute vasculitis, renal tubular interstitial necrosis, hepatic dysfunction, diffuse pulmonary hemorrhage, and respiratory failure (20). There are up to five SMases encoded in the genomes of pathogenic Leptospira strains which are absent in the genomes of nonpathogenic strains (20, 51). Leptospiral SMases, known to be cytotoxic to endothelial cells, lymphocytes, and macrophages (20, 29, 30), are likely involved in several aspects of leptospirosis pathogenesis (20). Among Leptospira interrogans SMases, sph2, which is likely secreted by a TolC-based type I secretion system, has the most prominent SMase C and hemolytic activities (51).
Mycobacterium tuberculosis is a facultative intracellular pathogen and the most common etiological agent of tuberculosis, a chronic infection which typically becomes active in immunosuppressed patients and destroys their lungs and sometimes other tissues (52). M. tuberculosis arrests phagosomal maturation or escapes to the macrophage cytoplasm, suppressing autophagy (53). Macrophages transport the bacterium to caseous granulomas, where it relies primarily on host lipids as energy and carbon sources (52). This bacterium encodes an outer membrane (OM)-associated SMase C (designated Rv0888) composed of a surface-exposed C-terminal sphingomyelinase domain coupled to an N-terminal integral membrane domain which putatively mediates the uptake of phosphocholine and glucose (52). This enzyme is an important hemolytic determinant in M. tuberculosis and enables it to use SM as a source for carbon, nitrogen, and phosphorous within phagosomes (52). Accordingly, an M. tuberculosis rv0888 deletion mutant replicates poorly in macrophages, indicating that this bacterium utilizes SM as a nutrient source during intracellular infection (52).
Sphingomyelinase Ds Which Adopt a TIM Barrel StructureSecretion of SMase Ds with a TIM barrel structure has been reported in Arcanobacterium (formerly Corynebacterium) hemolyticum, Corynebacterium ulcerans, and Corynebacterium pseudotuberculosis; furthermore, genes encoding those enzymes have been identified in the genomes of Streptomyces spp., Austwickia chelonae, and Burkholderia cenocepacia. These bacterial proteins are homologous to enzymes from several fungi and from spiders of the Sicariidae family (54), such as the SMase I from Loxosceles laeta (PDB ID 1XX1) (55).
These SMase Ds adopt a similar fold to that first observed in triose-phosphate isomerase, with 8 helices lying on the protein exterior and alternating with 8 parallel β-strands that form the central barrel (Fig. 4) (55). This type of SMase D lacks the HKD motif seen in phospholipase Ds (PLDs), explaining why they do not act on phosphatidylcholine (PC) and other glycerophospholipids, although sometimes they are mistakenly referred to as PLDs (56–58). Their catalytic site, identifiable by the position of the Mg2+ ion, which is essential for activity and octahedrally coordinated by strictly conserved residues, is made up of hydrophobic loops and a negatively charged surface (55). The C terminus contains a short helix (8α) and a β-strand (Hβ) motif that caps the torus of the barrel and seems to stabilize the entire internal structure. The presence of that helix and that strand, along with amino acid residues from the active site, can be considered a SMase D hallmark (55). Structural comparisons suggest that bacterial and spider SMase Ds originated from the broadly conserved glycerophosphoryl diester phosphodiesterases (55). There are unique motifs present in bacterial and spider SMase Ds but not in the ancestral family, supporting their origin through a single divergence event followed by horizontal gene transfer (59).
Molecular model of C. pseudotuberculosis Smase D. (A) Model based on the structure of the spider Loxosceles laeta Smase D (PDB ID 1XX1), shown as a blue cartoon, with the torus-capping C-terminal helix and strand in pink. The catalytically essential Mg2+ ion is shown as a green sphere, with the completely conserved coordinating aspartates and glutamates shown as sticks. A sulfate ion also coordinating the Mg2+ and likely an analogue for the sphingomyelin phosphate moiety is shown as sticks. (B) Model similar to that in panel A, but from the viewpoint of looking down into the active site, with a semitransparent electrostatic surface drawn highlighting the hydrophobic loops (white) and negatively charged center (red) that forms the active site.
SMase Ds have intrinsic Mg2+-dependent lysophospholipase D activity toward lysophosphatidylcholine (LPC), generating lysophosphatidic acid (LPA), which through the activation of G-protein-coupled receptors triggers a myriad of signaling cascades (60). Aberrant accumulation of LPA in blood causes effects similar to those induced by SMase Ds in experimental animals, including platelet aggregation, endothelial barrier dysfunction, and tissue infiltration by neutrophils (60). SMase Ds cause direct damage to the erythrocyte membrane (61) and also induce hemolysis through complement classical pathway activation (62). In addition, LPA could mediate SMase D cytotoxic activity, as transfection of LPA receptor-negative cells with the LPA(1) receptor cDNA renders them susceptible to SMase D, but only in LPC-containing media (60). Examples of SMase Ds which adopt a TIM barrel structure and play roles in virulence are those from A. haemolyticum and C. pseudotuberculosis, described below.
A. haemolyticum, which causes pharyngitis and several invasive diseases, including septic arthritis, osteomyelitis, and meningitis, produces a SMase D which likely plays a role in pathogenicity (57). A wild-type A. haemolyticum strain escapes the phagosome and causes host cell death, apparently by necrosis, whereas a mutant strain lacking the SMase D-encoding gene has a defective phagosomal escape capacity, indicating that SMase D is involved in phagosomal membrane disruption (57). The chromosomal genome of the emergent pathogen C. ulcerans, which causes among other diseases pharyngeal and pulmonary infections in humans, includes a homologous SMase D among its repertoire of virulence factors (63).
C. pseudotuberculosis causes caseous lymphadenitis in sheep and goats, as well as ulcerative lymphangitis and ventral abscesses in horses (56). It infects via skin abrasions, mucous membrane wounds, or aerosols, and it spreads into the lymph nodes, where it replicates within macrophages (56). C. pseudotuberculosis SMase D (Fig. 4) (Table 3) is highly expressed by macrophage-resident bacteria, plays a role in inducing death of infected macrophages (58), and is critical for the establishment and dissemination of the bacterium from the site of infection to lymph nodes (56), likely due to the inhibition of store-operated Ca2+ entry in T cells, which lowers the production of interleukin-2 and tumor necrosis factor alpha (TNF-α) (64). Accordingly, a C. pseudotuberculosis SMase D variant genetically engineered to be enzymatically inactive induced an immune response which protected sheep from challenge with a pathogenic C. pseudotuberculosis strain (65).
Other Sphingomyelinase DsPhotobacterium damselae subsp. damseale (formerly Vibrio damselae) is a primary pathogen for a wide range of marine animals and causes severe necrotizing fasciitis and multiple organ failure in humans (66). P. damselae subsp. damseale secretes a hemolytic SMase D, designated damselysin (Table 3), through a type II secretion system (67). This enzyme is structurally unrelated to the enzymes described in the previous sections, and it is also catalytically active against PC and phosphatidylethanolamine (PE) (67). Damselysin is encoded on a conjugative plasmid and contributes to the virulence of P. damselae in experimental infections in fish and mice (68). Homologous proteins to damselysin have been described in Aeromonas spp. (69).
PHOSPHOLIPASES
Phospholipases are lipolytic esterases which cleave glycerophospholipids, the major structural lipids in eukaryotic membranes. Glycerophospholipids contain two fatty acids esterified to the sn-1 and sn-2 positions of a glycerol backbone and a polar head group (the most common being choline, ethanolamine, serine, and inositol) attached to the sn-3 position, which defines the phospholipid class (PC, PE, phosphatidylserine [PS], or phosphatidylinositol [PI]). Each class, however, includes several molecular species, due to the diversity of fatty acids of various chain lengths and degrees of saturation. According to the cleavage site on their substrate (Fig. 5), PLases are classified into several groups: carboxyl ester acyl hydrolases, PLCs, and PLDs. Carboxyl ester acyl hydrolases include PLase As (PLAs), PLase Bs (PLBs), and lysophospholipase As (LPLAs). PLAs release a fatty acid from the glycerol backbone, leaving a lysophospholipid. LPLAs cleave the single fatty acid linked to the glycerol backbone, thereby detoxifying lysophospholipids. According to their site of ester bond hydrolysis, PLAs are classified as PLA1 (EC 3.1.1.32), which hydrolyzes the fatty acid at the sn-1 position of the glycerol moiety, or PLA2 (3.1.1.4), which removes the fatty acid at the sn-2 position. PLBs (EC 3.1.1.5) can hydrolyze both acyl groups at the sn-1 or sn-2 position of the glycerophospholipid and also exhibit LPLA activity. PLCs (EC 3.1.4.3) and PLDs (EC 3.1.4.4) are phosphoric diester hydrolases which cleave either the glycerol-oriented or the alcohol-oriented phosphodiester bond, respectively. PLCs release the phosphorylated head group (e.g., inositol triphosphate [IP3] or choline phosphate) and DAG, while PLDs cleave the terminal phosphodiester bond, releasing the head group (e.g., choline or inositol) and a phosphatidic acid (PA) (Fig. 5).
PLase cleavage sites in phosphatidylcholine and cellular activities of the reaction products. PLA1s remove the fatty acid at the sn-1 position of the glycerol moiety, while PLA2s remove it at the sn-2 position, generating cellular mediators like lysophosphatidylcholine and arachidonic acid. PLBs hydrolyze both acyl groups from the glycerophospholipid and also exhibit LPLA activity. PLCs hydrolyze the glycerol-oriented phosphodiester bond, releasing DAG and a phosphorylated head group [e.g., IP3, acting on PI(4,5)P2, or choline phosphate acting on PC]. PLDs cleave the alcohol-oriented phosphodiester bond, releasing the head group (e.g., choline or inositol) and generating LPA. (Based on data from reference 82.)
Acyl Hydrolases: Phospholipases A and BBacterial PLAs, PLBs, and LPLAs usually are membrane-associated or secreted carboxyl ester hydrolases, some of which are translocated to the host cell cytosol by specialized secretion systems. Six structurally different types of acyl hydrolases have been described in pathogenic bacteria to play a role in virulence, and Table 4 depicts a selected example of each enzyme type.
Bacterial acyl hydrolases that contribute to virulence
Outer membrane acyl hydrolases which adopt an antiparallel β-barrel structure.The OM acyl hydrolases, referred to as OMPLAs, have been identified in many species from the Enterobacteriaceae family, as well as in the genera Helicobacter, Campylobacter, and Neisseria. These enzymes regulate the OM lipid composition to maintain its permeability barrier and confer fitness for survival in certain environments (70). They are serine hydrolases with PLA, PLB, and/or LPLA activities and are structurally related to OM proteins involved in lipolysis and fatty acid uptake (70), which in Echerichia coli contributes to maintenance of lipid homeostasis and cell envelope integrity (71).
OMPLAs are integral membrane proteins with a membrane-spanning region consisting of a 12-stranded antiparallel β-barrel with the strands connected by long extracellular loops and short periplasmic turns (Fig. 6) (72). These enzymes form dimers with the interface formed at the flat side of the membrane-spanning barrel (72). The active site triad (Ser144, His142, and Asn156 in the E. coli OMPLA) is contained in a single monomer, but dimerization is required for maximal catalytic activity because the substrate-binding cleft, which allows the accommodation of a range of different aliphatic tail lengths, is formed at the dimer interface (72). OMPLAs have been found to be involved in the virulence of Yersinia pseudotuberculosis and Helicobacter pylori, as described below.
Molecular model of the Helicobacter pylori PldA1 dimer. (A) The model based on E. coli OMPLA (PDB ID 1QD6), shown as a pink and green cartoon. The inhibitor hexadecanoylsulphoic acid is shown as sticks at the dimer interface, and the Ser-His-Gln catalytic triad is shown as sticks and labeled. The Ca2+ ion that enhances activity by influencing active-site electrostatics is shown as a blue sphere, with its coordinating main-chain carbonyl ions drawn as sticks. (B) Increased magnification image of the binding pocket lined with hydrophobic residues at the dimer interface. Colors are as for panel A, but with all residues drawn as sticks and with a semitransparent molecular surface to highlight the large pocket that can accommodate diverse lipid tails.
Y. pseudotuberculosis, which causes human gastroenteritis and mesenteric lymphadenitis, encodes an OMPLA named PldA which hydrolyzes PC and SM (73). A Y. pseudotuberculosis mutant strain lacking pldA is 200 times less lethal than the wild-type strain in a murine yersiniosis model, indicating that PldA is an important virulence determinant (73).
H. pylori, which is the primary cause of peptic ulcer disease and is associated with several gastric disorders, produces an OMPLA named PldA1 (Fig. 6) (Table 4) (74). The pldA1 gene could have been horizontally acquired and is highly conserved due to purifying selection at the vast majority of residues (75). The human gastric epithelium protects itself against noxious factor attacks by secreting a lipid-rich hydrophobic mucus which is colonized by H. pylori before it attaches to epithelial cells (74). H. pylori PldA1 likely degrades phospholipids in the mucus layer, reducing the hydrophobicity of the gastric lining and allowing long-term colonization of the mucosa (74). When H. pylori grows under acidic conditions, PldA1 expression increases, leading to a change in its membrane lipid composition (from less than 2% to more than 50% lysophospholipids), increased adhesiveness, and the release of other virulence factors advantageous for survival at low pH (76). Accordingly, an H. pylori mutant strain lacking pldA1 is unable to colonize the murine gastric mucosa (77), and strains isolated from patients suffering peptic ulcer display high lysophospholipid content (78). However, pldA1expression does not play a role in enhancing IL-8 production or the acute inflammatory response to H. pylori (79).
Surface-associated phospholipase A2s.The surface-associated phospholipase A2 group includes the PlaB produced by different Legionella pneumophila strains as well as homologues, largely still uncharacterized, encoded in the genomes of other Legionella species, Pseudomonas aeruginosa, and other water-associated bacteria (80).
L. pneumophila is an intracellular parasite of diverse species of free-living freshwater amoebae and ciliated protozoa (81). This bacterium is ubiquitous to the environment, and in humans it is an opportunistic pathogen that causes Legionnaires' disease, a severe pneumonia frequently accompanied by acute respiratory distress syndrome (81). L. pneumophila possesses a large variety of about 20 established and potential PLAs (15 enzymes), PLCs (3 enzymes), and PLDs (1 enzyme) (82, 83). L. pneumophila PLase activity destroys phospholipids from lung surfactant, which covers small airways, bronchioles, and the alveolar surface and reduces the surface tension, and thereby its activity may contribute to lung disease (84). The most prominent PLA activity of L. pneumophila is PlaB (Table 4) (83).
PlaB hydrolyzes PC and phosphatidyl glycerol (PG), localizes at the bacterial surface, and has hemolytic activity, and its gene is transcribed mainly in the exponential growth phase (85, 86). Amino acid residues Ser85, Asp203, and His251 are required for catalysis and likely constitute the catalytic triad, whereas Ser129 is important for PC substrate specificity (80). A Ser129 mutant displaying an ∼90% reduction in its enzymatic activity against PC, but almost unchanged activity against PG, was not hemolytic to human erythrocytes (80).
An L. pneumophila mutant strain lacking plaB is less hemolytic and propagates less efficiently in RAW 246.7 macrophages (85, 87); it also has impaired capacities to colonize and replicate in lungs and to disseminate to the spleen in a guinea pig pneumonia model (86).
Acyl hydrolases from the SGNH esterase family.The SGNH esterase family acyl hydrolases are found through all kingdoms of life and are widely distributed among Gammaproteobacteria (88); they are also present in Burkholderia spp. and Streptomyces spp. Despite having poor sequence similarity, they have one domain that adopts the fold of the GDSL hydrolase superfamily, which is a three-layered α/β/α structure with a conserved core of five β-strands and at least four α-helices and a conserved SGNH catalytic motif, making up the S-D-H triad required for catalysis (88). Some of the single-domain enzymes of this group are secreted and display PLA and LPLA activities, though a few also have glycerophospholipid cholesterol acyl transferase (GCATase; EC 2.3.1.43) activity, which cleaves phospholipids and transfers the acyl chain onto sterols (88). Some acyl esterases present in the OM of several species of Gammaproteobacteria have an additional C-terminal β-barrel domain that functions as an autotransporter (89). Both single- and two-domain acyl hydrolases of the SGNH esterase family have a flexible substrate-binding pocket and are active on a wide range of substrates (89). Acyl hydrolases from the SGNH esterase family have been shown to be virulence factors in Moraxella bovis, Vibrio harveyi, Salmonella enterica, and Yersinia enterocolitica, as described below.
M. bovis is the causative agent of infectious bovine keratoconjunctivitis, the most important ocular disease affecting cattle worldwide, which can result in corneal ulceration and permanent blindness (90). The M. bovis plb gene encodes a PLB of the SGNH family which consists of two domains and is active against PC and LPC. A molecular model of M. bovis PLB (Fig. 7), which was built based on the structure of the homologous acyl hydrolase EstA from P. aeruginosa (91), shows conserved blocks of residues close in space, thus highlighting the likely active site region. M. bovis PLB contributes to host cell lysis after bacterial adhesion to cattle corneal cells and is a candidate vaccine for infectious bovine keratoconjunctivitis (90). P. aeruginosa EstA is required for cell motility and biofilm formation, but its role in pathogenesis has not been yet determined (92).
Molecular model of M. bovis PLB. The model is based on P. aeruginosa EstA (PDB ID 3KVN), and the active transporter is drawn as a gold cartoon, with the conserved blocks of residues associated with acyl hydrolase activity shown in red and the residues of the catalytic motif shown as sticks.
Vibrio species that are pathogenic for fish and crustaceans secrete acyl hydrolases from the SGNH esterase family that likely play a role in pathogenicity (93). The V. harveyi hemolysin VHH is a PLB active against PC, hemolytic for fish erythrocytes, cytotoxic for fish cells, and upon intraperitoneal injection in flounder induces hemorrhage and death (93–95).
S. enterica causes severe gastroenteritis that can lead to death in infants, the elderly, and immunosuppressed patients (96). This bacterium has the ability to invade, reside, and replicate within a variety of cells, which allows it to avoid clearance by neutrophils (96). Salmonella strains have evolved with two type III secretion systems (T3SS), encoded on Salmonella pathogenicity island 1 (SPI1) and Salmonella pathogenicity island 2 (SPI2), to deliver sophisticated virulence factors into host cells, which reprogram cell functions for the pathogen's benefit (96). Salmonella T3SS effectors are essential for invasion, phagosomal maturation arrest, and the establishment of the bacterium in its intracellular niche, which is in a phagosome that undergoes extensive membrane remodeling (96). SseJ from S. enterica subsp. enterica serovar Typhimurium (Table 4) is an SGNH hydrolase with PLA1, deacylase, and GCATase activities that is injected from the bacterium-containing vacuole into the host cell cytoplasm through the SPI2 secretion system (97–99). The role of SseJ in modulating the phagosomal membrane was recently reviewed (100). SseJ destabilizes the phagosomal membrane when the stabilizing factor SifA is not present (96). SseJ is recruited to the cytoplasmic face of endosomal membranes, where binding to the RhoA GTPase triggers its GCATase activity (101, 102), demonstrating a sophisticated evolution of a virulence factor upon intimate interaction with a host factor. Thus, RhoA regulation of sseJ constitutes a mechanism to tightly regulate the enzyme and to determine the localization of the activated protein to exert its pathogenic function. SseJ generates lysophospholipids and esterifies cholesterol, modifying the Salmonella-containing vacuole membrane and affecting its curvature, which leads to morphological changes, including the formation of filamentous microtubule-dependent structures that increase vacuole surface (103). Whether this changes function to provide nutrients or to stabilize the integrity of the vacuole is yet to be determined. The SseJ residues S151, D247, and H384 are necessary for enzymatic activity (104), and mutant strains lacking a functional sseJ display attenuated intracellular replication in peritoneal macrophages and diminished virulence in mice after intraperitoneal inoculation (98, 102, 104, 105).
Y. enterocolitica causes a gastrointestinal disease characterized by enteritis, ileitis, and mesenteric lymphadenitis that in severe cases can lead to septicemia and death (106). Y. enterocolitica penetrates intestinal M cells and is delivered to Peyer's patches, where it survives and replicates within macrophages (106). Y. enterocolitica produces YspM which, similar to its SseJ homologue, is translocated into the host cells via a T3SS, but its role in pathogenesis remains to be determined (107).
Acyl hydrolases that adopt the α/β hydrolase fold.Genes encoding acyl hydrolases with an α/β hydrolase fold are present in the genomes of many species of Mycobacterium, Brucella, and several Gammaproteobacteria species. These enzymes adopt the α/β hydrolase fold, common to a diverse and widespread group of enzymes with no obvious sequence similarities that includes lipases, esterases, and cutinases, among others (108). This fold consists of eight β-strands forming a left-handed, superhelically twisted β-sheet partly connected to and surrounded on both sides by α-helices (108). There is a conserved core of secondary structural elements and nearly always a catalytic triad of Ser/Cys-His-Asp at the active site (108). The α/β hydrolase fold is able to incorporate numerous inserted loops and secondary structural elements that allows it to recognize a vast range of substrates (108). These enzymes show interfacial activation, whereby a loop of residues constitutes a “lid” which moves on contact with membrane to expose the hydrophobic substrate-binding pocket (108). Examples of acyl hydrolases which adopt the α/β hydrolase fold that play roles in virulence are those from M. tuberculosis, Brucella melitensis, Y. enterocolitica, and Vibrio cholerae, as described below.
The M. tuberculosis genome contains seven genes encoding homologous proteins predicted to adopt the α/β hydrolase fold; these were originally annotated as cutinases (109). Although none of these enzymes has cutinase activity (110), at least two have PLA2 activity, Cut4 (also known as Rv3452 or Culp4) and Cut6 (also known as Rv3802c or Culp6) (111, 112). M. tuberculosis Cut4 is secreted (111), is cytotoxic to macrophages, and is likely important for virulence, as genes encoding homologous proteins to Cut4 are also found in genomes of several virulent Mycobacterium species (112). Interestingly, the Cut4 homologue from Mycobacterium smegmatis is also active against SM (109). M. tuberculosis Cut6, which is cell wall associated (110), also has thioesterase activity and plays a critical role in mycolic acid synthesis, which likely explains why Cut6 is conserved among mycobacteria (113, 114). Interestingly, Cut6 induces an immune response in mice which protects them from aerosol challenge with a virulent M. tuberculosis strain (115). M. tuberculosis Culp1 (also known as Cfp21 or Rv1984c) is secreted, exhibits strong acyl esterase activity (110), and induces production of transforming growth factor β and reactive oxygen species (ROS) by alveolar epithelial cells, which leads to apoptotic cell death (116). It has therefore been suggested that Culp1 contributes to disruption of the host alveolar barrier, facilitating mycobacterial dissemination (116). Accordingly, M. tuberculosis Culp1 enhances the M. bovis bacillus Calmette-Guérin vaccine-induced immunity that results in lower bacterial loads in the lungs and reduced lung pathology in murine models of M. tuberculosis infection (117, 118).
B. melitensis, a facultative intracellular bacterial pathogen that causes abortion in goats and sheep and Malta fever in humans, produces a PLA1 active against PE which is required for survival and replication of this bacterium in macrophages, as well as for persistent infection in mice (119). The bveA gene, which encodes this enzyme, is broadly conserved among the genus Brucella and has paralogues in the genus Xanthomonas, many species of Alphaproteobacteria, and Burkholderia (119).
Y. enterocolitica produces a PLA2 which adopts the α/β hydrolase fold, and this protein is named YplA (Table 4); it is secreted by the flagellar secretion apparatus and by two contact-dependent T3SSs (120). A Y. enterocolitica mutant strain lacking yplA exhibits lower colonization levels in the Peyer's patches and mesenteric lymph nodes and induces reduced inflammatory infiltration and bowel tissue necrosis, in comparison to the wild-type strain in a mouse intragastric model of infection (106).
Another heterogeneous group of acyl hydrolases which bear the GxSxG motif and the conserved Ser-Asp-His catalytic triad used by enzymes which adopt the α/β hydrolase fold is encoded in the genomes of several classes of proteobacteria; these acyl hydrolases are referred to as type VI lipase effectors (Tle) (121). These enzymes alter the host membrane architecture and thus play a role in antagonistic intra- and interspecies bacterial interactions after being injected by a type VI secretion system into the periplasmic space of adjacent target cells (121). The Tle effector referred to as Tle2VC/VC1418/TseL is produced by Vibrio cholerae, the etiological agent of the diarrheal disease cholera, has PLA1 activity (121), contributes to colonization of the upper intestinal mucosa during the infection in rabbits, likely via the displacement of competing bacteria (122), and is also required for V. cholerae to escape predation by amoeba (123). Two of the P. aeruginosa Tle effectors, referred to as Tle1PA and Tle4PA, have been structurally characterized and found to adopt the α/β hydrolase fold (124, 125). Tle1PA (96.5 kDa; PDB ID 4O5P) has PLA2 activity (121) and consists of a canonical α/β hydrolase fold domain with an additional putative membrane-anchoring module different from known structures (124). Tle4PA (63.3 kDa; PDB ID 4R1D) consists of a canonical α/β hydrolase fold domain and a cap domain, which is observed in many homologous lipases (125). The antibacterial activities of Tle1PA and Tle4PA could help P. aeruginosa prevent invasion by other bacteria in the lungs of cystic fibrosis patients, which would explain why a small number of P. aeruginosa clonal lineages persist for years during pulmonary infections in these patients (126). However, the role of these enzymes in pathogenesis has not been determined.
Another PLA1 which adopts the α/β hydrolase fold is found in a heterogeneous family of large toxins designated multifunctional autoprocessing repeats-in-toxin (MARTX) toxins, which are secreted through a type I secretion system by several Vibrio species and other Gram-negative bacteria of the genera Aeromonas, Proteus, Photorhabdus, and Xenorhabdus (127). MARTX toxins are composed of conserved repeat regions and an autoprocessing protease domain that functions as a delivery platform to transfer up to five cytopathic effectors in eukaryotic cells (128). They form pores in the membrane of the target eukaryotic cell which allow the entrance of the effectors that disrupt diverse cellular processes for bacterial benefit (128). MARTX toxins are cytopathic for a wide range of eukaryotic cell types, including erythrocytes, epithelial cells, and phagocytic cells from human, mouse, fish, and eel origins (128). The best-characterized MARTX toxin is the one secreted by nearly all clinical and environmental isolates of V. cholerae (129). This toxin, which promotes colonization of the small intestine in mice, includes as one of its effectors a PI 3-phosphate (PIP3)-specific PLA1 that reduces the intracellular PIP3 levels on endosomes and preautophagosomal structures, thus blocking endosomal and autophagic pathways, preventing bacterial clearance, and facilitating bacterial survival (129). The MARTX toxin of the opportunistic pathogen Vibrio vulnificus, which plays a role in defense against predation by amoeba and is required for full virulence in mammals, also encompasses a domain predicted to adopt the α/β hydrolase fold (127), although its exact role in pathogenicity has not been clarified.
Secreted patatin-like acyl hydrolases.PLAs homologous to the potato tuber storage protein patatin are encoded in the genomes of a wide range of Gram-negative and Gram-positive bacteria, including P. aeruginosa, L. pneumophila, Rickettsia spp., and M. tuberculosis (130). Patatin-like PLAs are also present in fungi, plants, and animals, harbor a domain characterized by a three-layer α/β/α architecture, and contain the conserved serine lipase motif Gly-x-Ser-x-Gly and the catalytic Ser-Asp dyad (130).
P. aeruginosa is a free-living and ubiquitous extracellular bacterium which forms part of natural microbial communities inhabiting diverse aquatic and terrestrial environments (131). P. aeruginosa is capable of causing opportunistic infections in a wide variety of hosts, including plants, worms, insects, and mammals (131). In humans, P. aeruginosa causes keratitis, ventilator-associated pneumonia, severe chronic infections in cystic fibrosis patients, and acute hospital-acquired infections in postoperative and burn patients (132). ExoU (Table 4) is a patatin-like protein translocated by P. aeruginosa into the host cell cytosol by a T3SS (133–135). Inside the bacterium, ExoU is held inactive by the SpcU chaperone, which facilitates its interaction with the secretion apparatus and protects the bacterium membrane (136, 137). Once translocated into the host cell, ExoU is activated by ubiquitylation (138) or by contact with monoubiquitin, ubiquitin polymers, or ubiquitylated proteins, such as SOD1 (139, 140) and PI 4,5-biphosphate [PI(4,5)P2] (141).
The exoU/spcU locus constitutes a bicistronic operon within a genomic island that likely originated from the chromosomal integration of an ancestral plasmid acquired through horizontal gene transfer from Betaproteobacteria or Gammaproteobacteria (142). exoU/spcU-carrying strains show many signs of being clonally related, and even though the ancestral plasmid underwent deletions, full-length exoU and spcU are retained, suggesting a clonal expansion of the original plasmid recipient strain and the existence of environmental pressures selecting for the ability to secrete functional ExoU (142). It is speculated that the presence of ExoU-producing strains in environmental reservoirs might be maintained because ExoU kills amoebae, P. aeruginosa's natural predators (143–145), and thus provides a public aid for ExoU-negative bacteria (146).
The crystal structure of ExoU in complex with SpcU (PDB ID 3TU3) shows the four ExoU domains (colored differently in Fig. 8): the chaperone-binding domain (residues 55 to 101 and 472 to 502), the PLA2 domain (residues 107 to 471), and two helical bundle domains, domain 3 (residue 503 to 603) and domain 4 (residues 604 to 687), which constitute the membrane-targeting and ubiquitin-binding regions (136, 137). Although the catalytic Ser142 and the oxyanion hole are ordered, part of the active site is disordered and not visible in an electron density map, in particular, the catalytic Asp344 (136, 137). Whereas the catalytic site appears to be inaccessible to the substrate, the ubiquitinylated residue Lys178 is solvent exposed and accessible to ubiquitin ligase (136, 137).
Structures of P. aeruginosa ExoU and its chaperone, SpcU (PDB ID 3TU3), shown as a cartoon. SpcU is shown in white, the ExoU chaperone binding domain is in gold, the PLA2 domain is green, and the membrane localization domain is shown in pink and purple. Catalytically important residues are shown as yellow sticks, and the ubiquitylable Lys178 is shown as a space-filling representation.
Distinct amino acid residues are critical for ExoU activation by ubiquitin and PI(4,5)P2, indicating that these factors activate ExoU by different mechanisms (141). The eukaryotic specificity of these cofactors ensures that ExoU acts selectively on host cell phospholipids without affecting bacterial membranes. Even though activation by ubiquitin and PI(4,5)P2 binding are conserved by patatin-like PLA2s from diverse bacterial species, enzymatic and toxicity profiles vary between different enzymes (147, 148).
ExoU has PLA2 and LPLA activities and targets a broad range of substrates, including PC, PE, PA, and PI(4,5)P2 (149, 150). ExoU PLA2 activity toward PI(4,5)P2 and neighboring phospholipids leads to focal adhesion disruption and triggers actin depolymerization, cytoskeletal collapse, and cell detachment from extracellular matrix, causing cell rounding and membrane blebbing, which later triggers plasma membrane disruption leading to host cell necrosis (150). Due to its high affinity for PI(4,5)P2, this lipid seems to be the scaffold for ExoU membrane engagement (150). Upon translocation into the host cell, ExoU rapidly associates with PI(4,5)P2 and orients the catalytic site toward phospholipid substrates in the plasma membrane (148). In the lungs, ExoU targets epithelial cells and endothelial cells, as well as resident alveolar macrophages and recruited neutrophils (149, 151, 152).
ExoU is cytotoxic to corneal epithelial cells, is required for P. aeruginosa traversal of the corneal epithelium, and plays a substantial role in the ocular colonization and pathogenesis of corneal disease in an experimental keratitis model in mice (153, 154).
ExoU contributes to promote the host inflammatory response and oxidative stress through different mechanisms (151, 155–158). ExoU releases arachidonic acid from cell membranes, leading to a subsequent increase in eicosanoid production by endothelial cells and airway epithelial cells (156). ExoU induces c-fos and c-jun activation, leading to changes in cytokine expression (157). ExoU inhibits caspase-1 activation and thereby the secretion of interleukin-1β and interleukin-18 (151). ExoU induces NF-κB activation through the canonical pathway by platelet activating factor (PAF) receptor signaling, leading to IL-8 production and neutrophil recruitment (158). The ExoU-mediated activation of NF-κB in turn favors PAF receptor expression, which amplifies its effects (159). ExoU leads epithelial cells to overexpress both the inducible and endothelial nitric oxide synthases, which leads to a redox imbalance that results in increased concentrations of lipid hydroperoxides and decreased levels of reduced glutathione (155).
ExoU is induced early during acute P. aeruginosa pneumonia, and the amount of ExoU present in the lungs increases over time (160). ExoU expression plays a significant role in P. aeruginosa dissemination from lung tissue into the bloodstream (160), and mutant strains lacking a functional ExoU are more easily cleared than wild-type bacteria and therefore have a dramatically reduced pathogenicity in experimental acute pneumonia models (152, 161, 162). T3SS-negative P. aeruginosa benefits from the public good provided by ExoU-mediated killing of neutrophils during an experimental pulmonary infection (146). Furthermore, ExoU constitutes a virulence marker, since ExoU-expressing strains are significantly associated with bacteremia, more severe disease, higher complication rates, and greater mortality in pneumonia patients (132, 163).
L. pneumophila encodes 11 patatin-like PLases, designated VipD/PatA, PatB, VpdA/PatC, PatD, PatE, VpdC/PatF, VpdB/PatG, and PatH to PatK (83). VipD, VpdA, VdpB, and VpdC are translocated by the Dot/Icm type IVB secretion system into the host cell (164–166). Once in the cytosol, VipD binds Rab5 and Rab22 (167, 168), two key regulators of endosomal trafficking from early endosomes, which induces a structural rearrangement that triggers VipD's PLA1 activity, resulting in inhibition of phagosomal maturation by catalyzing PIP3 depletion and altering the protein composition of endosomes (169). Furthermore, VipD hydrolyzes PE and PC on the mitochondrial membrane, contributing to cytochrome c release, caspase 3 activation, and apoptosis induction (166). Although single or quadruple knockout mutants for the genes that encode VipD, VpdA, VpdB, and VpdC grow normally in macrophages (164), the role of these patatin-like PLases for intracellular survival in these mutants could be masked by the functional redundancy of the additional homologous enzymes.
Rickettsiae are obligate intracellular pathogens which enter host cells by phagocytosis, escape from the phagosome, replicate within the host cell cytoplasm, and exit the host cell by actin-mediated motility (e.g., Spotted Fever group rickettsiae) (170) or host cell lysis (e.g., Typhus group rickettsiae) (171). R. prowazekii, the etiological agent of epidemic typhus in humans, produces two patatin-like proteins, RP534 and RP602. RP534 has PLA2, PLA1, and LPLA activities that can hydrolyze PC in the absence of any eukaryotic cofactor (170). R. typhi causes murine typhus, a mild to severe flu-like human illness which can be fatal if untreated (171). R. typhi produces two patatin-like proteins: Pat1, with PLA2 activity, and Pat2, with PLA2, PLA1, and LPLA activities (171). Both enzymes are surface exposed and translocated into the host cell during intracellular growth (171). Pretreatment of R. typhi with anti-Pat1 or anti-Pat2 specific antibodies decreases Rickettsia species' capacity to invade cells and escape from the phagosome (171).
Another type of patatin-like acyl hydrolases is present in the genomes of some pathogenic, commensal, and environmental bacteria of the phyla Proteobacteria, Bacteroidetes, Fusobacteria, and Chlorobi. This enzyme type is secreted fused to a transporter 16-stranded β-barrel domain that is similar to those of TpsB transporters of type Vb secretion systems (172, 173). The archetype of this acyl hydrolases type is PlpD from P. aeruginosa (PDB ID 5FYA), which acts as a PLA1 on phosphatidylinositols, PS, phosphatidic acid, and phosphatidylglycerol (173), although its role in pathogenesis has not been determined.
Class XIB phospholipase A2s.A prophage-encoded class XIB PLA2 denominated SlaA (Table 4) is present in the genomes of the group A streptococcus Streptococcus pyogenes serotypes M1, M2, M3, M4, M6, M22, M28, and M75 (17, 174). Streptococcus equi, the causative agent of equine strangles, also encodes a SlaA homologue suggested to be acquired by genetic exchange with S. pyogenes (175). Furthermore, analysis of diverse genomes has shown that genes encoding this PLA2 class are also present in Clostridia, Bacilli, some Alphaproteobacteria, and in plants (176).
SlaA contains five Cys residues, four of which form disulfide bonds conserved with plant class XIB PLA2 (176), with which it shares just 17% sequence identity over its whole length. The structures of PLA2s from this group show a conserved set of three α-helices and a two-stranded β-sheet (the β-wing) as well as a conserved Ca2+-binding loop (177). Although sequence alignments show that the catalytic His and the Ca2+-coordinating residues are present in SlaA, it contains a large inserted region at the N terminus, so the production of a meaningful structural model of this protein was not possible.
S. pyogenes causes pharyngitis, skin or deep tissue infections and, even though it is primarily considered an extracellular bacterium, it can persist within macrophages (174). SlaA requires intracellular localization to exert cytotoxicity, and it enters host epithelial cells in an actin-dependent manner, suggesting it utilizes a membrane receptor (178). An S. pyogenes mutant strain lacking slaA has reduced adhesion and cytotoxicity to human epithelial cells (178). Mice infected subcutaneously or intraperitoneally with a slaA+ strain die faster than mice infected with the isogenic strain lacking slaA (178). In addition, this slaA-lacking mutant strain has a reduced capacity to colonize the upper respiratory tract in a macaque pharyngitis model (178). Furthermore, immunization of mice with SlaA protects from systemic infection by a slaA+ S. pyogenes strain (178).
Phospholipase CsDifferences in the phospho head groups or the acyl chains of the glycerophospholipid substrate determine its susceptibility to a PLC's catalytic activity. Thus, based on substrate preference, PLCs are classified as phosphatidyl inositol specific (PI specific), PC preferring, or nonspecific PLCs with or without SMase activity. Four structurally different types of PLCs have been described in pathogenic bacteria as playing a role in virulence, and Table 5 depicts a selected example of each enzyme type.
Bacterial PLCs that contribute to virulence
PI-specific phospholipase Cs.PI-specific PLCs are produced by several Gram-positive pathogens, including Bacillus spp., Listeria monocytogenes, S. aureus, Clostridium spp., Streptomyces spp., and Leptospira interrogans (179). They hydrolyze PIP3 into IP3 and DAG, and they can also cleave the glycosyl PI (GPI) from membrane-anchored proteins (179, 180).
B. cereus (PDB ID 1GYM) (181) and L. monocytogenes (PDB ID 2PLC) (182) PI-specific PLCs adopt a modified TIM barrel structure that is open between strands V and VI, and it lacks two α-helices between strands IV and V and strands V and VI, respectively (181, 182). The active site is located at the C-terminal end of the barrel in a deep cleft lined by polar and charged amino acids. The PI head group, myo-inositol, has been observed bound to both enzymes in an edge-on orientation and with a number of hydrogen bonds to side chains in the active site. Stereochemical and site-directed mutagenesis studies indicated that the catalytic site in B. cereus PI-PLC (Fig. 9) consists primarily of three residues: His32 (general base), His82 (general acid), and Arg69, acting as a phosphate-activating residue (183). The higher capacity of the B. cereus PI-PLC to release GPI-anchored proteins from cellular membranes relative to the L. monocytogenes enzyme seems to be the result of the Vb strand insertion that extends the binding pocket and allows considerable additional interactions with glycan substrates (184).
Structure of L. monocytogenes PI-PLC (PDB ID 2PLC), shown as a cartoon. Bound myo-inositol is shown as gray sticks. General base His45, general acid His93, and phosphate-interacting residue Arg84, as well as residues interacting with the PI head group. are shown as sticks and labeled.
L. monocytogenes causes severe foodborne infections and survives within macrophages, hepatocytes, and endothelial cells (185). Once internalized by host cells, L. monocytogenes escapes from phagosomes and multiplies within the cytoplasm, from where it spreads to neighboring cells by an actin-based motility mechanism without leaving the intracellular environment (185). After such cell-to-cell spread, the bacterium resides in a secondary vacuole surrounded by a double membrane, disseminating in host tissues sheltered from the humoral immune response (185). L. monocytogenes PI-specific PLC, designated PlcA (Table 5), is required for bacterial escape into the cytosol from the single-membrane primary pathogen-containing vacuole (186–189). L. monocytogenes PlcA contributes to alter the host inflammatory response in several ways: it modulates phagocyte NADPH oxidase activation (190, 191), induces a persistent NF-κB activation, and upregulates the expression of adhesion molecules in endothelial cells (192). Furthermore, L. monocytogenes PlcA reduces the PIP3 level on preautophagosomal structures, leading to their stalling and prevention of autophagic flux, thus favoring the escape of cytosolic bacterial from the host autophagic defense (193). Accordingly, L. monocytogenes mutant strains lacking plcA are defective for growth in mouse macrophages and show reduced virulence in a murine infection model (187, 194).
PI-specific PLCs from different species dysregulate the immune response in different ways: the L. interrogans enzyme induces death of infected macrophages (195), the S. aureus PI-specific PLC is implicated in the survival of S. aureus in human neutrophils (196), whereas the enzyme from B. anthracis downmodulates dendritic cell function and T cell responses, possibly by cleaving GPI-anchored proteins important for Toll-like receptor-mediated activation (197).
Zn2+ metallophospholipase Cs.PLCs from the group of Zn2+ metallophospholipase Cs (Zn2+ metalloPLCs) are produced by species of Bacillus, Listeria, and Clostridium, as well as by P. aeruginosa. They contain three conserved Zn2+ ions in the substrate-binding pocket, have one or two domains, and although they vary in their substrate specificity (198, 199), their catalytic activity is readily inhibited by the competitive inhibitor D609 (200). Zn2+ metallophospholipase Cs have been shown to play a role in the pathogenesis of diseases caused by B. anthracis, L. monocytogenes, and Clostridium perfringens, as described below.
B. anthracis, the anthrax causative agent, accesses the body through the skin or the gastrointestinal or respiratory tract (37). After inhalation, B. anthracis endospores are engulfed by alveolar phagocytes and germinate, resulting in bacilli outgrowth in the circulatory and lymphatic systems, which leads to septicemia, toxemia, and often death (37). During infections, B. anthracis expresses a PC-preferring Zn2+ metalloPLC and a PI-specific PLC together with the SMase C (37). Simultaneous deletion of the genes encoding these three enzymes decreases B. anthracis virulence after intratracheal inoculation in mice (37).
L. monocytogenes secretes PlcB, a single-domain PC-preferring Zn2+ metalloPLC which can also hydrolyze PE, PS, and even SM, though only one-fourth as rapidly as PC (201). PlcB is secreted at acidic pH, which restricts its action to the lumen of L. monocytogenes-containing phagosomes (202). The dual enzymatic activity of PlcB as a PLase/SMase allows membrane fusion, which could be important for L. monocytogenes cell-to-cell spreading (203). Both the lecithinase and the SMase C activities of this enzyme are important for bacterial spread in cultured cells (189, 204). PlcB catalytic activity is regulated by vacuolar pH, and its compartmentalization to the spreading vacuole is critical for intracellular survival in neutrophils (205). Accordingly, an L. monocytogenes strain carrying a mutated plcB gene encoding an enzyme lacking control of its catalytic activity has 100-fold-attenuated virulence in mice (205, 206). Furthermore, an L. monocytogenes mutant strain lacking plcB has defects in phagosomal escape and cell-to-cell propagation and consequently has a 10- to 20-fold increased 50% lethal dose (LD50) (187) and reduced virulence in a murine cerebral listeriosis model (207). Accordingly, PlcB plays a critical role in the protective immunity elicited by recombinant L. monocytogenes strains used as vaccines (208).
P. aeruginosa secretes PlcB, a Zn2+ metalloPLC essential for directed twitching motility up a phospholipid gradient of either PE or PC (209). Lung surfactant PC serves as a primary carbon and energy source during P. aeruginosa lung infections (210, 211), and therefore, the ability of P. aeruginosa to travel along spatio-chemical gradients by chemotaxis could be critical for fitness and virulence. Hence, PlcB could play a role in pathogenicity during lung infections, although this remains to be assessed.
Several Clostridium species produce two-domain Zn2+ metalloPLCs, possessing an α-helical N-terminal catalytic domain homologous to the B. cereus PLC and an additional Ca2+-binding C-terminal β-sandwich domain similar to mammalian C2 domains (212). The best-characterized clostridial PLC is C. perfringens alpha-toxin (Fig. 10B; Table 5) (213), which preferentially cleaves PC and SM but also hydrolyzes PE, PI, and PG (214). Recent results utilizing a Langmuir monolayer technique showed the requirement for unsaturated carbon-carbon bonds in the acyl tails of glycerophospholipids for alpha-toxin activity (215). C. perfringens alpha-toxin shows significant lecithinase and SMase C activity in the pH range of 6.5 to 4.5, with its lecithinase activity being 5-fold higher than the SMase activity (216). Relative to the B. cereus PLC, C. perfringens alpha-toxin has lost a pair of helices in the N-terminal domain (199) (Fig. 10A and B), resulting in a planar surface able to interact directly with the target membrane. C. perfringens alpha-toxin likely anchors to membranes via Ca2+ ion-binding sites in the C-terminal domain, and thus the active site is oriented to interact directly with membrane phospholipids (Fig. 10C). The absent B. cereus helix hairpin is replaced with a tryptophan residue which, together with a number of hydrophobic residues in the C-terminal domain (Phe334, Tyr 331), is well placed to interact with the hydrophobic membrane interior. In addition, the C-terminal domain, in a manner analogous to C2 domains, binds Ca2+ ions but does not complete their coordination spheres and has a number of positively charged residues (such as Lys300) placed to interact with the phosphate at the membrane surface (199) (Fig. 10C). These adaptations for membrane interaction explain the much higher enzymatic activity and toxicity of C. perfringens alpha-toxin compared to its B. cereus homologue.
Structures of Zn2+ metalloPLCs from B. cereus and C. perfringens, similarly oriented. (A) B. cereus PLC (PDB ID 1P6D) is shown as a purple cartoon, with inserted hairpin in white. Active-site Zn2+ ions are shown as lighter-colored spheres. The bound PC analogue is shown by white sticks. (B) C. perfringens alpha-toxin (PDB ID 1CA1) with the N terminal (homologous to B. cereus PLC) in green and membrane-binding C terminus in pink. Zn2+ (violet) and Ca2+ (blue) ions are shown as spheres, and coordinating residues are shown as sticks. (C) The alpha-toxin shown in panel B, modeled bound to mixed PC/cholesterol membrane, highlighting how the B. cereus hairpin would impinge into the membrane.
C. perfringens is the most common cause of gas gangrene, an acute infection associated with traumatic or surgical wounds and characterized by intravascular leukocyte accumulation, thrombosis, and severe myonecrosis (217). A C. perfringens mutant strain with an inactivated plc is unable to produce gas gangrene (218), and immunization with the alpha-toxin C-terminal domain protects mice against experimental gas gangrene induced by C. perfringens (219). Similarly, immunization of guinea pigs with the Clostridium haemolyticum PLC elicits a protective immunity from an intramuscular challenge with 50 LD50s of this bacterium (220). Alpha-toxin is required for C. perfringens to escape from the macrophage phagosome in early stages of the infection (221), and once the bacteria are established, it causes extensive myonecrosis, leading to creatine kinase release to the circulation (222). Furthermore, C. perfringens alpha-toxin suppresses myocardial contractility, induces hypotension, and plays multiple roles in the pathogenesis of gas gangrene (Fig. 11) (reviewed in reference 217). C. perfringens alpha-toxin-induced myotoxicity and lethality are at least partially mediated by ROS, as myonecrosis and mortality in a murine model of gas gangrene are reduced by free radical scavengers (222).
General scheme of the events triggered by C. perfringens alpha-toxin during gas gangrene. The pleiotropic effects of C. perfringens alpha-toxin in diverse cell types contribute to reducing vascular perfusion and causing tissue damage, which enhance the conditions for bacterial growth and lead to shock and multiple organ failure.
C. perfringens alpha-toxin induces cell death in various cell types, particularly those with low ganglioside content in their plasma membrane (223). Muscle has the lowest concentration of complex gangliosides among mammalian tissues, offering a possible explanation for the high susceptibility of muscle fibers to alpha-toxin (223). Other clostridial PLCs with a structure similar to C. perfringens alpha-toxin display lower toxicities due to different amino acid substitutions (212, 224).
In artificial membranes, measurements of membrane dipoles show that following exposure to alpha-toxin, the membranes swell as DAG is accommodated in their interior, increasing the curvature of the local membrane to sites of alpha-toxin binding (215). In rabbit neutrophils, C. perfringens alpha-toxin induces DAG production and protein kinase C (PKC) activation, which leads to stimulation of fibrinogen and fibronectin adhesion (225, 226). In endothelial cells, alpha-toxin stimulates prostacyclin and PAF production, which likely contributes to increased vascular permeability and neutrophil adhesion to endothelial cells (227). In platelets, alpha-toxin causes fibrinogen receptor gpIIbIIIa translocation to the membrane and triggers a conformational change in the surface-expressed receptors, inducing aggregation and leading to vascular occlusion (228, 229). Thus, alpha-toxin activates transduction pathways in neutrophils, endothelial cells, and platelets, resulting in uncontrolled expression of adhesion molecules and contributing to intravascular leukostasis and thrombosis, which promotes vascular injury and enhances the conditions for anaerobic bacterial growth (230, 231).
C. perfringens alpha-toxin induces the respiratory burst in neutrophils by inducing the activation of PKC θ via two separate pathways: elevation of DAG generated by activation of an endogenous PLC through a pertussis toxin-sensitive GTP-binding protein, and activation of a tropomyosin-related kinase receptor (TrkA receptor) that leads to PDK1 phosphorylation (225, 226). In epithelial cells, alpha-toxin binding to GM1a ganglioside induces TNF-α and IL-8 production through the activation of the TrkA receptor and the p38 mitogen-activated protein kinase (MAPK) pathway, as well as a PLC-γ1, ERK1/2–NF-κB-dependent pathway (232–234). TNF-α plays an important role in the toxic effect of alpha-toxin, since the administration of specific antibodies against TNF-α protects mice from its lethal effect (235).
C. perfringens alpha-toxin induces PKC activation in rabbit erythrocytes, rabbit neutrophils, MDCK cells, and ganglioside-deficient cells (8, 226, 236). At sublytic concentrations, alpha-toxin is internalized, inducing ERK1/2 activation and triggering NF-κB in a MEK/ERK-dependent manner (236, 237). PKC, MEK/ERK, and NF-κB signaling induced by C. perfringens alpha-toxin results in ROS production and cell death, and inhibition of either of these signaling pathways significantly reduces cytotoxicity in cultured cells and myotoxicity in vivo (222, 236). C. perfringens alpha-toxin is internalized in a dynamin-dependent manner, reaches early and late endosomes, and causes lysosomal damage (237). Furthermore, dynamin inhibition, which impairs alpha-toxin internalization, also inhibits its cytotoxic effect (237). Τherefore, C. perfringens alpha-toxin, previously considered to act only locally on the plasma membrane, has a more complex mode of action.
In poultry, C. perfringens causes necrotic enteritis characterized by inflammation, intestinal necrosis, and dramatic flock morbidity and mortality, which lead to significant productivity losses (238). C. perfringens alpha-toxin affects the jejunal mucosa of laying hens and could contribute to necrotic enteritis pathogenesis (239). Accordingly, vaccination with alpha-toxin or its C-terminal domain reduces intestinal pathology and growth depression in chickens challenged with C. perfringens (240–242).
Phospholipase Cs from the acid phosphatase superfamily.Enzymes from the acid phosphatase superfamily include those homologous to AcpA from Francisella tularensis and the well-characterized PLCs produced by Pseudomonas spp., Burkholderia, and Mycobacterium, as well as hypothetical phosphoesterases from several alpha-, beta-, and gammaproteobacteria and Actinobacteria (243–245). The catalytic activity of these enzymes is Zn2+ independent and, unlike Zn2+ metalloPLCs, it is not abolished by the inhibitor D609 (243).
P. aeruginosa secretes two PLCs from the acid phosphatase superfamily, PlcN and PlcH/PlcS (Table 5), via the twin-arginine translocation (Tat) and type II Xcp-dependent systems (246). PlcN hydrolyzes PC and PS, but not SM, and is nonhemolytic (247), whereas PlcH hydrolyzes PC, SM, and lyso-PC at equal rates (248) and also acts on cardiolipin, PE, and PG (249). P. aeruginosa PlcH, which is expressed during lung infections in humans (250), is hemolytic to human erythrocytes (26) and selectively cytotoxic to endothelial cells (248). Furthermore, it suppresses the neutrophil respiratory burst and induces a strong chemokine response in mice and human granulocytes (248, 251). In vivo, PlcH induces recruitment and activation of platelets at the endothelium, leading to thrombotic lesions similar to those observed in P. aeruginosa sepsis (248). P. aeruginosa mutants lacking plcH have reduced virulence and result in decreased mortality in systemic experimental infections in mice (252, 253). Interestingly, adjacent to plcH the P. aeruginosa genome encodes a neutral ceramidase which is secreted with PlcH/PlcS and enhances its hemolytic activity (254). Remarkably, PlcH/PlcS is also required for P. aeruginosa to elicit disease in Arabidopsis thaliana (131, 255) and G. melonella (256) and has a role in killing Candida albicans filaments, suggesting that the antagonism with this fungus could have contributed to the evolution and maintenance of this virulence factor (257).
Pulmonary surfactant PC and PG hydrolysis by P. aeruginosa PlcH contributes to bacterial dispersion and alveolar dysfunction, because a P. aeruginosa mutant strain lacking plcH produces more localized damage and a less severe lung disease (258). Catabolism of the PlcH-released phosphorylcholine provides nutrients and activates biofilm formation and anaerobic metabolism, thus contributing to bacterial fitness and survival in the lungs (210, 211, 259). Furthermore, SM hydrolysis by PlcH and Cer accumulation in epithelial cells could contribute to pathogenesis of P. aeruginosa lung infections by inhibiting the function of the cystic fibrosis transmembrane conductance regulator Cl− channel, which leads to thick mucus production that clogs the airways, fostering bacterial growth (49).
F. tularensis causes tularemia, a systemic disease with multiple clinical manifestations, including ulceroglandular, glandular, oropharyngeal, oculoglandular, respiratory, and typhoidal forms (244). The bacterium subverts the oxidative burst in neutrophils, escapes from the phagosome, and replicates mainly in the macrophage cytosol, although it can infect a wide variety of cells (244). The acid phosphatase AcpA from F. tularensis has an amino acid sequence that is 23% identical to that of P. aeruginosa PlcH, and it also inhibits the neutrophil respiratory burst (244, 260). The fold adopted by ApcA has been observed in enzymes with a range of catalytic functions, including an arylsulphatase (PDB ID 1AUK (261) and phosphoglycerate mutase (PDB ID 1EJJ) (262). ApcA is a dimer, and each monomer consists of a core domain comprised of a twisted 8-stranded β-sheet flanked on each side by three α-helices and two smaller domains located above the C-terminal edge of the β-sheet (244). A molecular model of P. aeruginosa PlcH built based on the ApcA structure (Fig. 12) shows that the residues interacting with the metal ion found at the active site are conserved in PlcH. In addition, several of the phosphate analogue-interacting residues are also conserved, suggesting that the phosphate from phospholipid would bind in a similar location. The likely nucleophile identified in ApcA, Ser175, has a Thr at an equivalent position in P. aeruginosa PlcH; thus, this residue could function as a nucleophile. Since the essential hydroxyl nucleophile from the PLCs of this superfamily is conserved in AcpA, it has been suggested that it could facilitate phagosomal escape by hydrolyzing lipids from the phagosomal membrane (244). Accordingly, a mutant F. tularensis strain lacking acpA has reduced virulence in mice due to a defect in phagosomal escape (260). However, whether AcpA or homologous phosphatases display PLC and/or SMase C activities remains to be assessed.
Molecular model of P. aeruginosa PlcH. The model is based on P. aeruginosa ApcA (PDB ID 2D1G) and is shown as a gold cartoon, with active-site residues conserved between the two enzymes shown as sticks. The metal ion found in the ApcA structure is shown as a blue sphere, and the phosphate analogue is shown as white sticks. The majority of the conserved residues interact with either the metal ion or the phosphate analogue, with the exception of the likely nucleophile, Thr178 (labeled).
B. pseudomallei is a facultative intercellular pathogen capable of survival in human phagocytic cells and causes melioidosis, a febrile illness with disease states ranging from chronic abscesses to acute pneumonia and septicemia (263). B. pseudomallei encodes three PLCs, PLC-1/PlcN1, PLC-2/PlcN2, and PLC-3 (263), with the first two being secreted in a GspD-dependent manner (264). Functional analysis of plc1 and plc2 mutant strains showed that these PLCs have a role in nutrient acquisition, plaque formation, and cytotoxicity (265). PLC-3 is upregulated in vivo and was shown to be required for full virulence in a hamster model of acute melioidosis (266).
M. tuberculosis enters the host by inhalation, and once in the lungs is phagocytosed by macrophages, which may lead to its elimination or to a persistent infection with the formation of granulomas having foamy macrophages containing bacteria (267). M. tuberculosis accumulates inclusions of host cell membrane-derived lipids and remains dormant, or under host immune depression can become active, replicate, and spread into the lung and other tissues (267). The genome of most M. tuberculosis strains encodes various lipases and up to four PLCs, encoded by plcA/rv2351c, plcB/rv2350c, plcC/rv2349c, and plcD/rv1755c, two of which have SMase C activity (267, 268). These PLCs are strongly upregulated during the first 24 h of macrophage infection and are cytotoxic to mouse macrophages (268, 269). PlcA and PlcB have been experimentally shown to be transferred through the cytoplasmic membrane by means of the Tat system (270). Phospholipid hydrolysis by M. tuberculosis PLCs releases DAG, which may be hydrolyzed by lipases, releasing fatty acids that serve either as a carbon source or as building blocks for cell wall lipid synthesis in chronically infected lungs (267). Although a quadruple mutant, M. tuberculosis plcABCD, was reported to be attenuated in its growth kinetics during the late phase of pulmonary infection in mice (268), in a recent study no differences were found between wild-type and mutant strains lacking PLCs in their phagosomal escape, growth abilities within macrophages, or in a mouse infection model (271). Nevertheless, PLCs from M. tuberculosis might help in phosphate acquisition during phagosomal containment, because the M. tuberculosis PLC-encoding genes are strongly upregulated under phosphate starvation, and mutant strains lacking them have decreased survival when PC is the only phosphate source, compared to survival of the parental strains (271).
The lung and mucocutaneous pathogen Mycobacterium abscessus encodes a PLC that is homologous to M. tuberculosis PLCs (39% amino acid sequence identity), is highly cytotoxic to macrophages, and is involved in survival of amoeba (272). This enzyme enhances M. abscessus virulence in a murine model of pulmonary infection (272), and when used as a vaccine, it induces an immune response that allows mice to rapidly clear the pulmonary infection (273).
Other phospholipase Cs.L. pneumophila PLCs constitute a separate Zn2+-dependent PLC class, together with the Pseudomonas fluorescens PC-preferring PLC and not-yet-characterized homologous enzymes from fungi (274).
L. pneumophila produces three PLCs, designated PlcA, PlcB, and PlcC/CegC1 (Table 5), which are induced during host cell infection (275, 276). plcA is found in L. pneumophila genomes, and plcB is also present in some nonpneumophila strains, whereas plcC/cegC1 is found in all Legionella genomes sequenced so far (274). PlcC/CegC1 is injected into the host cell by a Dot/Icm type IVB secretion system, whereas PlcA and likely PlcB are secreted via an Lsp type II secretion system (274, 277–279). In addition, Tat-dependent transport of PlcA was suggested in experiments with the 130b strain of L. pneumophila. PLC-related secreted activity was reduced by ∼30% in a tatB mutant, whereas activity of a plcA knockout was reduced ∼70%, indicating either concurrent involvement of the Sec system or a selective contribution of Sec under conditions when the Tat pathway is not functional (277). However, the absence of a twin-arginine motif in PlcA proteins of other L. pneumophila strains, such as Philadelphia-1 or Corby, indicates that this is not a general pathway for PlcA export and that the Sec system may be preferred. The three L. pneumophila PLCs require Zn2+ ions for catalysis; PlcA and PlcB hydrolyze PG but not PC, whereas PlcC has a broad substrate specificity that includes PC, PG, SM, and PI (274). Fifteen conserved amino acids essential for enzymatic activity and potential Zn2+ binding were identified via mutations of plcC (274). Knockout of the three genes plcA, plcB, and plcC resulted in reduced killing of G. mellonella larvae, showing their contribution to insecticidal activity in this model (274). Similarly, the P. fluorescens PLC also contributes to insecticidal activity against Plutella xilostella larvae (280).
Phospholipase DsPLD-encoding genes are present in several intracellular and extracellular pathogens, including Neisseria gonorrhoeae, Acinetobacter baumannii, Rickettsia prowazekii, Legionella spp., Yersinia spp., Chlamydia spp., H. pylori, P. aeruginosa, and Klebsiella pneumoniae (281).
Bacterial PLDs are either secreted into the extracellular milieu or directly injected into the host cell cytosol (281). They display the HxK(x)4D(x)6GSxN motif present in eukaryotic PLDs of the PLD protein superfamily, together with endonucleases, cardiolipin, and glycerophospholipid synthases (281). Bacterial PLDs fold into two homologous domains, each with a PLD motif, and thus form a pseudodimer. Each domain consists of a β-sandwich fold, with two 8-stranded β-sheets sandwiched between 18 α-helices (Fig. 13) (282). The active site is at the interface between the domains, so both PLD motifs are involved in the interaction with substrate (282, 283). Two flexible loops extend over the entrance to the active site and are thought to modulate interfacial lipid interactions and substrate specificity (282). Structural studies of Streptomyces PLD-substrate complexes revealed that the reaction proceeds via a phosphohistidine intermediate, with the phospholipid phosphate covalently bound to a PLD motif histidine in the second domain (283).
Molecular model of Acinetobacter baumanii PLD1. The model is based on the structure of the Streptomyces PLD (PDB ID 1F0I) and is shown as a cartoon. The N-terminal domain is shaded green, and the C terminal is pink. Residues at the active site and part of the PLD motif are shown as sticks, and the covalently bound reaction intermediate is shown as white sticks at the interface between the domains. Note the symmetrically positioned motif residues in each homologous domain.
N. gonorrhoeae, the causative agent of human gonorrhea, replicates within cervical and urethral epithelial cells (284). This bacterium is endocytosed by a clathrin-dependent mechanism or by complement receptor type 3 (CR3)-mediated endocytosis (284). N. gonorrhoeae produces a PLD, designated NgPLD, which is necessary for efficient epithelial cell invasion (284). NgPLD-produced PA drives cell surface recruitment of CR3, membrane ruffling, and bacteria engulfment (284). Furthermore, NgPLD also modulates host cell signaling by binding directly to the PH domain of the Akt kinase, which results in its activation as well as in CR3 receptor translocation to the cell surface, promoting further N. gonorrhoeae uptake (285). Accordingly, an NgPLD-lacking mutant strain exhibited a decreased ability to invade primary cervical cells due to its impaired capacity to recruit CR3 to the cell surface and elicit membrane ruffles (284).
A. baumannii is an opportunistic pathogen that causes a wide range of severe human infections, including ventilator-associated pneumonia, soft tissue and bone infections, meningitis, urinary tract infections, and bacteremia (286). The A. baumannii genome encodes up to three PLDs, designated pld1, pld2, and pld3 (287). A molecular model of A. baumannii PLD1 was built based on the structure of the Streptomyces PLD (PDB ID 1F0I) (Fig. 13). pld3 was apparently acquired first, whereas pld1 and pld2 evolved more recently through gene duplication and divergence, as pld2 encodes a membrane-anchored enzyme (287). The three A. baumannii PLDs act in a concerted manner to promote bacterial invasion of host cells (287), but inactivation of pld2 only reduces virulence significantly after intranasal inoculation of mice (286).
R. prowazekii produces a dimeric PLD which is cytotoxic and associated with phagosomal escape (288–290). This bacterium causes fever as well as severe and prolonged body weight loss when injected intraperitoneally into guinea pigs, whereas an R. prowazekii pld-deficient mutant strain showed attenuated virulence in this disease model (289).
L. pneumophila produces a PLD, designated LpdA, which is injected into host cells through the Dot/Icm type IVB secretion system (291, 292). This enzyme is posttranslationally modified at its C terminus by S-palmitoylation and thus targeted to the plasma membrane as well as to Rab4- and Rab14-containing vesicles (292). LpdA hydrolyzes PIP3, PIP4, and PG to PA and triggers disruption of the Golgi apparatus in cultured cells (292). An L. pneumophila mutant strain lacking lpdA shows reduced replication in murine lungs, indicating that LpdA plays a role in virulence (292).
Yersinia pestis, the agent of bubonic plague, harbors a plasmid which encodes an intracellular PLD, named Y. pestis murine toxin (Ymt), which has β-adrenergic blocking activity and is highly toxic to mice and rats when released upon bacterial lysis in vivo (293). Although Ymt is not required for virulence in mice (294), it is required for bacterial survival in the midgut of its principal vector, the rat flea Xenopsylla cheopis, and thereby for the arthropod-borne transmission (295).
Chlamydophila pneumoniae is an obligate intracellular pathogen associated with a wide range of acute and chronic diseases in humans and animals (296, 297). This bacterium is a major cause of severe respiratory diseases in humans, but it also causes conjunctival and genitourinary infections and is implicated in atherosclerosis and acute coronary syndromes (296, 297). C. pneumoniae infects and multiplies in endothelial, smooth muscle, and blood mononuclear cells (296, 297). It produces a chromosomally encoded PLD which induces a specific antibody response in acute coronary syndrome patients (297). C. pneumoniae PLD is a Toll-like receptor 4 (TLR4) agonist able to elicit a Th17 immune response within atherosclerotic plaques in an experimental mouse model of atherosclerosis, as it activates endothelial cells and macrophages to express and secrete a wide variety of chemokines and cytokines (298).
The P. aeruginosa genome encodes two PLDs, designated PldA/Tle5PA and PldB, which are translocated into other bacteria through the H2- and H3-type VI secretion systems, respectively, and were likely acquired by horizontal gene transfer through independent events (121, 299, 300). Homologous proteins to these enzymes, referred to as Tle5 type VI effectors, are present in many other genera of Betaproteobacteria and Gammaproteobacteria (121). PldB is a trans-kingdom effector which displays antibacterial activity and promotes intracellular invasion of nonphagocytic cells by activation of the AKT/PI3 pathways (299). PldA/Tle5PA, which preferentially targets PE, causes lysis when delivered to competing bacteria and is required for P. aeruginosa to persist in a chronic pulmonary infection model in rats (121, 300, 301).
K. pneumoniae is responsible for blood, urinary, and respiratory tract infections in humans (302). Its genome contains within a type VI secretion system locus the pld1 gene, which encodes an enzyme homologous to the P. aeruginosa PLDs. PLD1/Tle5KP is involved in the control of bacterial membrane composition (302), and a K. pneumoniae pld1-deficient strain lacks virulence when injected intranasally in a murine pneumonia model (302).
EVOLUTIONARY CONSIDERATIONS
The traditional view of virulence evolution considers the interaction between pathogenic bacteria and their multicellular hosts as the primary driving force to develop virulence traits that confer adaptive advantages for transmission or growth during infection (303). However, bacteria have existed for more than 3 billion years and adapted to a huge diversity of environments long before multicellular organisms appeared, about 600 million years ago (304, 305). Thus, for bacterial pathogens that have a free-living state, selection of traits associated with virulence may have occurred in settings outside infection if they provide an advantage for survival in their natural environment, and virulence to multicellular organisms could be just coincidental (303–306).
Growth in environments with phosphorous limitation or dramatic variations in water availability could have favored the evolution of some bacterial PLCs associated with virulence. In P. aeruginosa, PlcH and the phosphorylcholine phosphatase PchP might have a role in phosphate scavenging and osmoprotection (307). Under phosphate limitation, in niches where eukaryotic membrane debris are present, PlcH would be induced to generate phosphate monoesters that would be subsequently hydrolyzed by PchP to phosphate and choline (308–311). Choline could be used as an energy source or as a precursor for the synthesis of glycine betaine, an important protectant for survival in environments where osmolarity is not constant (307). Homologous enzymes to P. aeruginosa PlcH and PchP are present in several species of Actinobacteria and Proteobacteria (311). Interestingly, in the soil resident nonpathogenic alphaproteobacterium Sinorhizobium meliloti during phosphorous limitation, a PLC also plays a function in environmental fitness by scavenging for phosphate in its own membrane glycerophospholipids (312). Under phosphate-rich conditions, the membrane of S. meliloti is composed of PG, PE, monomethyl-PE, cardiolipin, and PC (312). Phosphate limitation induces the S. meliloti PLC to degrade PE, monomethyl-PE, and PC to DAG and the corresponding phosphoalcohols and unidentified phosphatases to release inorganic phosphate from the phosphoalcohols for the synthesis of essential phosphorous-containing biomolecules (312). The released DAG serves as a substrate for the formation of the phosphate-lacking lipids betaine DAG-(N,N,N,-trimethyl)homoserine and sulfoquinovosyl-DAG, which together with ornithine lipids are the main constituents of the bacterial membrane under phosphorus-limiting conditions (312). Membrane lipid remodeling is a widespread response used by heterotrophic bacteria to reduce their requirement for phosphorus in oligotrophic habitats (313, 314); however, it is not known whether this strategy is also used by opportunistic pathogens in their natural environments.
Membrane targeting by PLases seems to have evolved independently on several occasions in some bacterial taxa as an antibacterial strategy to compete for space and resources (315). Bacterial PLases with antagonistic activity could play indirect roles in pathogenesis. These enzymes could help pathogens to overcome the colonization barrier constituted by the established microbiota or could provide a competitive advantage over other potential colonizers when a commensal colonization barrier is compromised. Following colonization, those PLases could enable pathogens to defend their niche by resisting invasion by incoming competitors, of the same or other species. However, bacterial PLases exerting deleterious effects on a broad array of competitors likely evolved as mediators of interbacterial interactions in their natural environments outside infection settings (315).
Bacterial predation by heterotrophic protists could have served also as a selection force during the evolution of bacterial SMases and PLases associated with virulence in pathogens of environmental origin. Bacterial survival in diverse ecosystems has been constrained for more than 800 million years by the pressure of predation by heterotrophic protists, and this is now considered a major force in the selection of traits serving as antipredator adaptations (316, 317). Those traits might have evolved before the emergence of multicellularity, played a role in bacterial persistence, and only by chance provided enhanced fitness for pathogenicity in metazoans (304, 306). Phagocytosis evolved in an ancestral unicellular eukaryote as a way to acquire food by predation of other microorganisms and was then maintained as a core function of the innate immune response in metazoans (317–320). Phagocytosed bacteria are killed by multiple mechanisms activated after extensive remodeling of the phagosomal membrane and a complex phagosome maturation process which requires a sequence of fusion events with multiple components of the endocytic pathway (321). Different mechanisms to resist phagocytosis by free-living amoebae arose during evolution, including biofilm or endospore formation, development of strategies to exit the phagosome after internalization, or to avoid killing, and even establishment of a replicative intracellular niche (144, 145, 316, 320, 322, 323). Due to the striking similarities between the phagocytic and microbicide molecular machineries in amoebae and macrophages, these mechanisms are also effective for escape or survival within macrophages (324). Thus, virulence traits that help bacteria thrive within macrophages, likely emerged as adaptations for intra-amoebal survival and multiplication (305, 306, 320, 325). Therefore, obligate and facultative intracellular bacterial pathogens of animals and humans could have evolved after surviving phagocytosis by free-living amoebae and adapting in them to an intracellular lifestyle (305, 306, 320, 326). A large number of bacterial species from different phylogenetic groups can withstand predation by protozoa and persist as intracellular parasites, suggesting that the capability to thrive within these microorganisms evolved several times independently (320, 325). The human bacterial pathogens known to be amoeba-resistant include species from the genera Chlamydia, Bacillus, Staphylococcus, Listeria, Mycobacterium, Helicobacter, Rickettsia, Pseudomonas, Yersinia, Legionella, Salmonella, Francisella, Acinetobacter, Vibrio, Aeromonas, and Burkholderia (320, 327). Furthermore, fungi and viruses also survive within different free-living amoeba species (320, 327). Thus, free-living amoebae constitute a niche that brings together diverse microorganisms allowing genetic exchanges among themselves and with the genomic content of the amoebal host itself (328). Since the metabolism of phagosomal membrane lipids controls a remarkable number of events during phagocytosis, the acquisition of bacterial SMases and PLases which contribute to phagosomal escape, arrest, or diversion of phagosomal maturation could have been favored by the selective pressure of fighting predator amoebae. Genes encoding SMases and PLases which contribute to intracellular survival could have been horizontally transferred among diverse microorganisms with a sympatric lifestyle within free-living amoebae; this constitutes a plausible explanation for the patchy distribution of some of the genes encoding these enzymes among intracellular pathogens from phylogenetically divergent bacterial phyla.
CONCLUDING REMARKS
Bacterial SMases and PLases are diverse and widely expressed esterases that play important roles in the pathogenesis of several diseases caused by a variety of Gram-positive and Gram-negative bacteria. Some of these enzymes help pathogens in a variety of ways to colonize tissues and/or to obtain substrates from the host, thus promoting bacterial survival or invasion. SMases and PLases may be crucial to allow bacterial escape from the phagocytic vacuole, thus promoting intracellular replication, as in the case of L. monocytogenes. PLases could contribute also to the establishment of a specialized intracellular niche, as in the case of the gastrointestinal pathogen S. enterica serovar Typhimurium, and could stall autophagosome formation, promoting escape from autophagic defenses, as with L. monocytogenes. Some pathogens express several PLases which confer redundancy for lipid metabolism and membrane modulation, highlighting the importance of these enzymes for virulence.
SMases and PLases may be good targets for vaccination against bacterial infections, as observed for B. cereus, C. pseudotuberculosis, S. pyogenes, C. perfringens, C. haemolyticum, and M. abcesus, whereas inhibitors of some of these enzymes, such as SMY-540, could be used as new antibacterial therapeutic agents for adjunctive therapy.
The use of site-directed mutagenesis along with experimental work on artificial membranes, molecular dynamics simulations, and fluorescence correlation spectroscopy contributes to the comprehension of the molecular basis of the interaction between SMases and PLases and host cell membranes, as well as the substrate specificities of these enzymes. Experiments with mutant strains lacking specific PLases and SMases have allowed elucidation of the roles of several of these enzymes in pathogenicity. However, the potential roles as virulence factors of other enzymes recently identified in bacterial genomes await further investigations, and in bacteria such as L. pneumophila and P. aeruginosa, which express several PLases, the potential overlapping or synergistic effects between the distinct enzymes remain to be evaluated.
The paradigm regarding the function of bacterial PLases discussed in this work is biased toward their roles in pathogenesis in animal and human hosts. However, these enzymes also contribute to other aspects of the lifestyles of bacteria, including membrane remodeling and competition with other microorganisms for survival in their natural environments and, thus, their multifunctional nature reflects the remarkable adaptability conferred by PLases to some bacteria.
Finally, although important contributions have been made during the last 15 years in the field of bacterial SMases and PLases, the effects of many of these enzymes in host cells are not fully understood at the molecular level, and the roles of several of them in disease pathogenesis are not clear. New tools and established techniques will aid future studies to further elucidate the roles of SMases and PLases as virulence factors.
ACKNOWLEDGMENTS
This work was supported by Vicerrectoría de Investigación Universidad de Costa Rica (grants 741-B1-601, 741-B1-603, and 741-A9-503), FEES CONARE (FR 6441), MICIT (FR 1471), the Robert Koch-Institute, and the German Research Foundation (grants DFG FL359/4-3, DFG FL359/6-1, 6-2, and 7.1).
We thank José María Gutierrez, Inger Florin, and César Rodríguez for critical reading of the manuscript. We thank Dorothea Eitze, Andrés Hernández, and Patrick Lane for support in preparation of some figures.
- Copyright © 2016, American Society for Microbiology. All Rights Reserved.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 129.↵
- 130.↵
- 131.↵
- 132.↵
- 133.↵
- 134.↵
- 135.↵
- 136.↵
- 137.↵
- 138.↵
- 139.↵
- 140.↵
- 141.↵
- 142.↵
- 143.↵
- 144.↵
- 145.↵
- 146.↵
- 147.↵
- 148.↵
- 149.↵
- 150.↵
- 151.↵
- 152.↵
- 153.↵
- 154.↵
- 155.↵
- 156.↵
- 157.↵
- 158.↵
- 159.↵
- 160.↵
- 161.↵
- 162.↵
- 163.↵
- 164.↵
- 165.↵
- 166.↵
- 167.↵
- 168.↵
- 169.↵
- 170.↵
- 171.↵
- 172.↵
- 173.↵
- 174.↵
- 175.↵
- 176.↵
- 177.↵
- 178.↵
- 179.↵
- 180.↵
- 181.↵
- 182.↵
- 183.↵
- 184.↵
- 185.↵
- 186.↵
- 187.↵
- 188.↵
- 189.↵
- 190.↵
- 191.↵
- 192.↵
- 193.↵
- 194.↵
- 195.↵
- 196.↵
- 197.↵
- 198.↵