OUP user menu

How microorganisms avoid phagocyte attraction

Jovanka Bestebroer, Carla J.C. de Haas, Jos A.G. van Strijp
DOI: http://dx.doi.org/10.1111/j.1574-6976.2009.00202.x 395-414 First published online: 1 May 2010


Microorganisms have developed several mechanisms to modulate the host immune system to increase their survival and propagation in the host. Their presence in the host is not only revealed by self-produced peptides but also through host-derived chemokines and active complement fragments. These so-called chemoattractants are recognized by G protein-coupled receptors (GPCRs) expressed on leukocyte cell membranes. Activation of GPCRs triggers leukocyte activation and guides their recruitment to the site of infection. Therefore, GPCRs play a central role in leukocyte trafficking leading to microbial clearance. It is therefore not surprising that microorganisms are able to sabotage this arm of the immune response. Different microorganisms have evolved a variety of tactics to modulate GPCR activation. Here, we review the mechanisms and proteins used by major human pathogens and less virulent microorganisms that affect GPCR signaling. While viruses generally produce receptor and chemoattractant mimics, parasites and bacteria such as Staphylococcus aureus, Streptococcus pyogenes, Porphyromonas gingivalis, and Bordetella pertussis secrete proteins that affect receptor signaling, directly antagonize receptors, cleave stimuli, and even prevent stimulus generation. As the large arsenal of GPCR modulators aids prolonged microbial persistence in the host, their study provides us a better understanding of microbial pathogenesis.

  • GPCR
  • immune evasion
  • chemoattractants
  • chemokines
  • complement


For successful infection of the host, invading pathogens first need to cross the mucosal surfaces or the skin. Upon subsequent entrance of host tissue by the pathogen, the acute inflammatory response of the innate immune system is initiated (Fig. 1a). Neutrophils and macrophages are the main initial effector cells of the innate immune system that can clear the invading pathogen through phagocytosis. These cells are attracted to the site of infection through sensing of chemotactic factor gradients. Chemoattractants are secreted by activated host cells and released as activated complement components upon recognition of evolutionary conserved structures presented by pathogens. The secreted chemoattractants form a gradient in the tissue and thereby direct phagocyte migration to the site of infection and inflammation (Fig. 1b). Well-known chemoattractants for phagocytes are host-derived platelet-activating factor, leukotriene B4, complement fragment C5a, and chemokines such as CXCL8 [also known as interleukin-8 (IL-8)]. Bacterial-derived products also serve as effective chemoattractants. These include formylated peptides such as N-formyl-methionyl-leucyl-phenylalanine (fMLP). All these chemoattractants activate phagocytes by binding to membrane-bound receptors that belong to the superfamily of G protein-coupled receptors (GPCRs). Activation through GPCRs not only directs phagocyte chemotaxis but also primes and activates the cells for effector functions such as phagocytosis.

Figure 1

Phagocyte extravasation. (a) For phagocytes to reach infection sites, they need to extravasate from the blood vessels into the tissue. The successful recruitment of circulating immune cells depends on the productive interaction of leukocytes with the endothelial cells lining the vessel wall and it is a multistep process. Phagocytes tether and roll on activated endothelium through transient interactions of PSGL-1 and selectins. Following stimulation of the cells by endothelium-bound chemokines, integrins are activated that mediate the firm cellular adhesion to intercellular adhesion molecule-1 (ICAM-1) and vascular adhesion molecule-1 (VCAM-1) presented by the vessel wall. The phagocytes then transmigrate through the endothelium into the underlying tissue [here junctional adhesion molecules (JAMs) play a role]. The cells are directed to the site of infection through the sensing of chemoattractant gradients by GPCRs. Here, chemokines and bacteria- and complement-derived products are important. The pathogen is cleared by phagocytosis upon recognition of the bacterium opsonized by antibodies and/or complement components. (b) At the infection site, infected cells and the surrounding tissues release chemoattractants. These secreted chemoattractants form a concentration gradient and attract leukocytes, which move through the gradient towards the higher concentrations, a process called chemotaxis. Leukocytes move through the gradient toward the higher concentrations, a process called chemotaxis. The figure was produced using Servier Medical Art.

Chemoattractants and GPCRs thus play an essential role in directing the innate immune defense against invading pathogens. It is therefore not surprising that pathogens have evolved a large variety of strategies to evade activation of phagocytes using chemoattractants. In this review, we discuss the arsenal of proteins produced by pathogens that are able to disrupt activation of the innate immune system through interference with chemoattractants and chemoattractant-induced signaling (Table 1 provides a listing of these modulators and their activities). We will focus both on products that act on GPCRs, the activating chemoattractants, or the generation and presentation of these chemoattractants. Viral mechanisms of chemoattractant inhibition primarily involve the chemokine system and have extensively been reviewed by others (Seet & McFadden, 2002; Alcami et al., 2003; Rosenkilde et al., 2008). Our review will therefore predominantly focus on bacterial evasion of activation through GPCRs. We will describe several groups of chemoattractants and their target GPCRs, followed by an overview of the mechanisms and molecules used by pathogens to combat this system.

View this table:
Table 1

GPCR modulators with their targets and mechanism of modulation

Intracellular signalingPertussis toxinB. pertussisGiαKatada & Ui (1982)
Cholera toxinV. choleraeGsαToyoshige et al. (1994)
Receptor bindingCHIPSS. aureusFPR, C5aRPostma et al. (2004)
FLIPrS. aureusFPRL-1Prat et al. (2006)
SSL5S. aureusGlycosylated GPCRsBestebroer et al. (2009)
SSL10S. aureusCXCR4Walenkamp et al. (2009)
CCL2-like moleculeM. aviumCCR5Rao et al. (2002)
Cyclophalin-I8T. gondiiCCR5Aliberti et al. (2003)
Chemokine mimicsVirusesAlcami et al. (2003)
Stimulus bindingSSL5S. aureusAll chemokinesBestebroer et al. (2009)
Evasin-1R. sanguineusCCL3, CCL4, CCL18Frauenschuh et al. (2007)
Evasin-3R. sanguineusCXCL8, CXCL1Deruaz et al. (2008)
Evasin-4R. sanguineusCCL5, CCL11Deruaz et al. (2008)
smCBPS. mansoniSmith et al. (2005)
Receptor mimicsVirusesRosenkilde et al. (2008)
Chemokine-binding proteinsVirusesSeet & McFadden (2002)
Stimulus and receptor cleavageSpyCEP, ScpCS. pyogenesCXCL8, KC, MIP2,Hidalgo-Grass et al. (2006)
CXCL1, CXCL6Sumby et al. (2008)
ScpA, ScpBS. pyogenesC5aWexler et al. (1983), Bohnsack et al. (1997)
CepIS. iniaeCXCL8Zinkernagel et al. (2008)
ElastaseP. aeruginosafMLP, C3, C5Ijiri et al. (1994), Schultz & Miller (1974)
CXCL5, CCL2, CCL5Leidal et al. (2003)
Alkaline proteaseP. aeruginosaCXCL5, CCL2, CCL5Leidal et al. (2003)
Cystein/serine proteasesP. gingivalisCXCL8, CCL2Zhang et al. (1999), Kobayashi-Sakamoto et al. (2003)
GingipainsP. gingivalisC3, C5, C5aRWingrove et al. (1992), Jagels et al. (1996)
MetalloproteaseN. americanusCCL11Culley et al. (2000)
56-kDa proteinS. marcescensC5aOda et al. (1990)
ElastaseE. faecalisC3, C3aPark et al. (2008)
Inhibition stimulus generationEcb, EfbS. aureusC3b-containing convertasesJongerius et al. (2007)
SCIN, SCIN-B, SCIN-CS. aureusC3 convertasesRooijakkers et al. (2005a)
SSL7S. aureusC5Langley et al. (2005)
  • CHIPS, Chemotaxis inhibitory protein of S. aureus; FPR, formylated peptide receptor; FLIPr, FPRL-1-inhibitory protein; FPRL-1, FPR-like receptor 1; SSL, staphylococcal superantigen-like; smCBP, S. mansonii chemokine-binding protein; SpyCEP, S. pyogenes cell envelope protease; Scp, streptococcal C5a peptidase; SCIN, staphylococcal complement inhibitor; Efb, Extracellular fibrinogen-binding protein; Ecb, extracellular complement-binding protein.

GPCR structure and ligand binding

GPCRs form a large family of receptors that can bind a variety of ligands such as light-sensitive compounds, odors, pheromones, hormones, and neurotransmitters. The GPCRs sensitive to chemoattractants belong to the rhodopsin subfamily of GPCRs. They are seven-transmembrane receptors with seven helical membrane-spanning regions connected by six extramembrane loops (Fig. 2a) (Palczewski et al., 2006). The N-terminus and three loops face outside the cell, while the C-terminus and the three remaining loops are localized in the cytoplasm. Post-translational modifications of the chemoattractant GPCRs include sulfation and glycosylation. The N-terminal segment and the second extracellular loop often carry sites for putative asparagine (N)-linked glycosylation. Further serine/threonine (O)-linked glycosylation sites can also be found in the N-terminus. While sulfate moieties are often important for ligand binding (Bannert et al., 2001; Farzan et al., 2001, 2002), glycans are additionally described to be essential for successful cell surface expression (Chiang et al., 2000; Ludwig et al., 2000; Bannert et al., 2001; Wenzel-Seifert & Seifert, 2003; Wang et al., 2004).

Figure 2

Structure and activation of GPCRs. (a) GPCRs are seven-transmembrane receptors, also known as serpentine receptors. In addition to the six loops that connect the seven transmembrane domains, they have an extracellular N-terminal domain and an intracellular C-terminal domain. (b) Chemoattractants use different methods to induce activation of the GPCRs. Lipid-derived chemoattractants, such as platelet-activating factor (PAF) and leukotriene B4 (LTB4), and peptides, such as formylated peptides (fMLP) and phenol-soluble modulins (PSMs) trigger activation by binding to the transmembrane (TM) regions of their target GPCRs. Chemokines and complement fragments C3a and C5a show a two-step binding model. They first interact with the N-terminus of their target GPCR and then activate it by interacting with their transmembrane regions. The figure was produced using Servier Medical Art.

The ligand-binding region of the GPCRs varies with the type of stimulus (Fig. 2b). Lipid-derived stimuli (e.g. leukotriene B4 and platelet-activating factor) and small peptides (e.g. fMLP) primarily activate target GPCRs through the transmembrane regions, and the N-terminal region or the extracellular regions do not seem to play an important role for either of these ligands (Ishii et al., 1997; Mills et al., 2000; Sabirsh et al., 2006). On the other hand, a two-step binding model has been proposed for larger protein stimuli. Chemokines and anaphylatoxins bind the N-terminus of the chemokine and anaphylatoxin receptors, whereupon they are positioned to interact with the pocket formed by the transmembrane domains, which is necessary for activation of the receptor (Siciliano et al., 1994; Allen et al., 2007).

GPCR signaling

Upon ligand binding, GPCRs are activated and the signal is transduced to enable effector functions (Fig. 3) (Hamm, 1998). As the name of the GPCRs suggests, they are coupled to heterotrimeric G proteins that are associated with their intracellular loops. Predominantly, Gαi associates with multiple chemoattractant GPCRs, but there is evidence that other Gi proteins, such as Gq or G16, may play a role. Ligand binding induces exchange of GDP for GTP by the Gα subunit, upon which Gβγ-subunit dissociates. The Gβγ-subunit in turn activates the phosphatidylinositol-3-kinase and phospholipase C, leading to the activation of the mitogen-activated protein kinase pathway and accumulation of inositol trisphosphate and diacylglycerol in the cytoplasm, respectively. These products induce mobilization of calcium from intracellular stores and activation of protein kinase C that drives downstream signaling throughout the cell. The Gα subunit also transduces the signal by interacting with adenyl cyclase, which is responsible for the synthesis of cyclic-AMP (cAMP) from ATP. The signaling pathways result in changing the cellular responses and gene expression. Generally, activation of the signaling pathways leads to actin polymerization responsible for cell shape change and chemotaxis. Effector cells are enabled to degranulate and secrete oxygen radicals crucial for killing of pathogens. Importantly, chemokine signaling induces the expression and activation of integrins on the leukocyte cell surface, allowing for firm adhesion of the leukocyte during extravasation from the blood vessel to infected tissues.

Figure 3

GPCR signaling. Upon ligand binding to GPCRs, an intracellular signaling cascade is initiated. The G protein coupled to the intracellular tail of GPCRs is essential. The signal is transduced upon dissociation of the Gα and Gβγ subunits that activate several downstream targets that regulate several processes such as actin polymerization, cell shape change, adhesion, degranulation, and protein expression. cAMP, cyclic AMP; PI3K, phosphatidylinositol-3-kinase; PKB, protein kinase B; MAPK, mitogen-activated protein kinase; PLC, phospholipase C; PIP2, phosphatidylinositol-4,5-biphosphate; IP3, inositol-1,4,5-trisphosphate; DAG, diacylglycerol; PKC, protein kinase C. The figure was produced using Servier Medical Art.

Modulation of GPCR signaling

As chemoattractant-sensing GPCRs play a central role in the immune responses initiated against invading pathogens, and they all induce common signaling molecules to transduce their signal, it is not surprising that they are often targeted by excreted bacterial toxins. The causative agent of whooping cough, Bordetella pertussis, produces a toxin that attacks the G proteins coupled to GPCRs. This well-known pertussis toxin directly binds to G proteins, modifying their action by converting them into a permanently active state.

Pertussis toxin is an oligomeric AB5 complex made up of six subunits (Tamura et al., 1982). The complex consists of a single A subunit and a pentameric B subunit. The A subunit is enzymatically active, whereas the B subunit is responsible for receptor binding. Pertussis toxin is released by the bacterium in an inactive form. When the B subunit binds to its target cell membrane-associated receptor, internal proteolytic cleavage releases the active A subunit that then passes the cell membrane and interacts with targets in the cytoplasm. The internalized A subunit binds NAD and catalyzes the ADP-ribosylation of the α subunits of the heterotrimeric Gi proteins (West et al., 1985; Locht & Antoine, 1995). Active Gα subunits typically transmit an inhibitory signal to adenyl cyclases. Because of the activity of the pertussis toxin, the inactive, ADP-bound Gα is unable to inhibit adenyl cyclase, resulting in constitutive production of cAMP (Moss et al., 1984). ADP-bound Gα is also unable to interact with its target GPCRs; thereby pertussis toxin also disturbs proper GPCR signaling through a dual mechanism. Pertussis toxin is thought to play a major role in the development of the infection by B. pertussis as it affects signaling pathways in many cells. Its inhibitory effects also involve the innate immune responses, including neutrophil and macrophage function. Pertussis toxin interferes with neutrophil influx in the lung, which is responsible for the bacterial killing (Carbonetti et al., 2003), and also inhibits early chemokine production to delay neutrophil recruitment in response to B. pertussis (Andreasen & Carbonetti, 2008). In vitro experiments have also shown that pertussis toxin inhibits neutrophil chemotaxis (Spangrude et al., 1985).

Given the dramatic effects of ADP-ribosylation of host cell components, it is not surprising that this mechanism is used by many pathogenic bacteria to disturb important processes in the infected host, such as protein synthesis and signal transduction. The factors responsible for these effects include cholera toxin and diphtheria toxin (Deng & Barbieri, 2008). Cholera toxin is responsible for ADP-ribosylation of the Gsα, also resulting in constitutive cAMP production, while diphtheria toxin interacts with elongation factor 2, which is a GTP-hydrolyzing protein important in the elongation step in protein synthesis.

GPCR sensing of bacterial-derived chemoattractants

The host innate immune system is directed at recognizing foreign substances, often through evolutionary conserved structures. Bacterial protein synthesis is distinct in several aspects from eukaryotic synthesis, and therefore certain host receptors are sensitive to the differences in this process. During protein synthesis, translation starts with a methionine residue that can be removed during protein maturation. While eukaryotic proteins are synthesized with an unmodified methionine residue, newly synthesized bacterial proteins contain a formylated methionine at the amino terminal end due to the activity of the methionyl-tRNA-formyltransferase. The formyl group is typically removed by peptide deformylases upon protein elongation. However, the activity of the deformylases is not infallible, and consequently bacteria secrete N-formylated proteins and peptides. The percentage of secreted formylated proteins can be quite substantial. It must be noted here that bacteria are not the only source of formylated peptides; mitochondria of eukaryotic cells also use N-formylmethionine as initiators of protein synthesis.

fMLP, the prototype N-formyl-peptide, induces and potentiates chemotaxis, phagocytosis, and the generation of oxidative burst in neutrophils and monocytes. Initially, it was found as the most potent synthetic chemoattractant (Showell et al., 1976), but later it was also isolated from Escherichia coli cultures (Marasco et al., 1984). Other bacteria, such as Staphylococcus aureus and Listeria monocytogenes, secrete similar small formylated peptides inducing comparable effects (Rot et al., 1986, 1987; Rabiet et al., 2005). Formylated peptides act on receptors belonging to the GPCR family: the formyl peptide receptor (FPR) and its homologue FPR-like-1 (FPRL-1). FPR binds fMLP with high affinity, while FRPL-1 shows low affinity for the ligand. Another FPR homologue, FPRL-2, does not bind fMLP at all. However, FPRL-1 and FPRL-2 additionally recognize nonformylated chemotactic peptides from the host, for example amyloid β1–42 (Le et al., 2001a), serum amyloid A (Su et al., 1999), and prion protein PrP106–126 (Le et al., 2001c), as well as from several pathogens, for example Helicobacter pylori protein Hp(2-20) (Bylund et al., 2001), Herpes simplex virus type 2 protein gG-2p20 (Bellner et al., 2005), and several HIV-envelope proteins (Le et al., 2000). Initially, FPR expression was demonstrated in monocytes, neutrophils, microglial, and dendritic cells, but it has also been described in nonhematopoietic cells and tissues (Le et al., 2001b; Migeotte et al., 2006). FPRL-1 is also expressed in a large variety of cells and organs including monocytes, neutrophils, macrophages, and microglial cells. Neutrophils express FPR and FPRL-1 but not FPRL-2, while monocytes express all three GPCRs.

Recently, Dürr et al. (2006) demonstrated that formylated peptides are indeed the major chemoattractants in S. aureus supernatants. However, disruption of the formyl methionine transferase in this bacterium demonstrated that formyl peptides are not the only chemoattractants that are secreted. As with formylated peptides, the activity of the yet unknown chemoattractant molecules was mediated through GPCR signaling. A recent study by Wang et al. (2007) showed that several phenol-soluble modulins can indeed prime neutrophils. Phenol-soluble modulins provoke neutrophil chemotaxis and induced calcium mobilization upon stimulation, suggesting the involvement of GPCRs. Enhanced expression of phenol-soluble modulins in community-acquired methicillin-resistant S. aureus (CA-MRSA) as compared with hospital-associated MRSA is suggested to contribute to the enhanced virulence of CA-MRSA.

Evasion of bacterial-derived chemoattractants

The bacterial secretome contains formylated peptides that activate leukocytes through FPR and FPRL-1. As these bacterial-derived products are such potent stimuli, bacteria have developed strategies to affect the activation of target GPCRs. Pseudomonas aeruginosa secretes proteins that are able to abrogate bacterial-mediated immune responses (Ijiri et al., 1994). The culture supernatants of P. aeruginosa display an inhibitory effect on neutrophil chemotaxis. The potential mediator of this response was hypothesized to be elastase. Purified Pseudomonas elastase was subsequently demonstrated to be responsible for this activity when fMLP was used as a chemoattractant. Furthermore, an fMLP-induced respiratory burst was reduced by pretreatment of neutrophils with elastase. Elastase diminishes the chemoattractant capacity by hydrolyzing fMLP at the Met-Leu bond and was also found to proteolytically alter the FPR (Ijiri et al., 1994).

Two proteins have been identified from S. aureus that impair immune responses to bacterial-derived chemoattractants. Chemotaxis inhibitory protein of S. aureus (CHIPS) and FPRL-1 inhibitory protein (FLIPr) inhibit responses induced by formylated peptides by targeting both the FPR and FPRL-1, respectively. CHIPS was initially identified in a search for the protein in culture supernatants of S. aureus that was responsible for the observed inhibition of fMLP-induced activation of neutrophils (Veldkamp et al., 1997, 2000). This virulence factor impairs neutrophil chemotaxis and binding of fluorescein-labeled fMLP (de Haas et al., 2004). CHIPS is a small protein with a molecular mass of 14.1 kDa that is made up of 121 amino acids. It inhibits fMLP-induced calcium mobilization, up- and downregulation of CD11b and CD62L, and elastase release in neutrophils. CHIPS does not affect the activation of leukocytes stimulated with chemoattractants acting on the CXCR1, CXCR2, C3aR, FPRL-1, LTB4R, and PAFR. Neutrophil responses to C5a are, however, affected by CHIPS, but will be discussed in the section describing the modulation of GPCR signaling induced by complement fragments.

Epidemiological studies demonstrated that over 60% of tested clinical S. aureus isolates carry the gene encoding CHIPS; moreover, 92% of the CHIPS-positive strains produce the protein in vitro. CHIPS is human specific, and it is produced in vivo as antibodies against this protein can be detected in human sera (de Haas et al., 2004). In addition, when mice are injected with a CHIPS-positive S. aureus strain, the bacterium actively produces CHIPS in the localized infection (de Haas et al., 2004). During bacterial growth, the protein is produced in the early exponential growth phase, and it is suggested that it is synthesized in the early stage of staphylococcal infection, enabling it to modulate early host responses (Rooijakkers et al., 2006). Interestingly, the CHIPS gene is located on the β-hemolysin-integrating bacteriophage that also carries genes for other immune modulators (Van Wamel et al., 2006).

CHIPS inhibitory activity on fMLP-induced responses is mediated through direct binding to the FPR (Postma et al., 2004). The binding shows a Kd of 35.4 nM, which is in the same order as described for its natural ligand. The CHIPS-binding site on the FPR involves multiple regions in the extracellular domains of the receptor. CHIPS probably binds near or directly at the fMLP-binding pocket, because fMLP is a small peptide that would otherwise easily bypass the blockade. For CHIPS, the N-terminus is important for its activity (Haas et al., 2004). Of the N-terminal residues, the first and third phenylalanines are crucial as the removal of the phenylalanines or addition of alanine at its N-terminus reduces the FPR-blocking activity.

A search for S. aureus proteins homologous to CHIPS led to the identification of FLIPr. The gene encoding FLIPr shows 48% identity to that encoding for CHIPS (Prat et al., 2006), while the resulting protein of 105 amino acids shows 28% identity to CHIPS. Functional assays demonstrated that FLIPr has a weaker but consistent inhibition of fMLP-induced activation of neutrophils as compared with CHIPS. FLIPr also impairs neutrophil and monocyte responses to FPRL-1 agonists; it inhibits neutrophil calcium mobilization, actin polymerization, and/or chemotaxis to the synthetic peptides MMK-1 and WKYMVM, but also to endogenous proteins amyloid β1–42 and PrP106–126. In a transfection experiment, it was demonstrated that FLIPr specifically interacts with FPRL-1 and not FPR, FPRL-2, or C5aR. FLIPr thus antagonizes GPCR signaling, and its activity is mediated by binding to the FPRL-1. FLIPr is found in 59% of clinical isolates and is naturally produced by S. aureus in vivo. Together, FLIPr and CHIPS inhibit immune responses induced by bacterially secreted formylated peptides by targeting both the FPRL-1 and FPR.

GPCR activation by chemokines

Chemokines are a large family of small proteins of 8–12 kDa that attract a variety of effector cells to inflammatory sites. They are divided in four groups according to the number and relative position of their cysteine residues (Fig. 4) (Fernandez & Lolis, 2002; Allen et al., 2007). The two large families of chemokines are the CXC- and the CC-chemokines, whereas the CX3C- and the XC-chemokines represent the two smaller families of chemokines. The CXC-chemokines are further divided into two subfamilies, based on the presence or the absence of a Glu-Leu-Arg (ELR) sequence preceding the first cysteine residue. ELR+ chemokines preferentially attract neutrophils, whereas ELR− chemokines act on lymphocytes. The CC-chemokines are involved in the attraction of various cell types, for example monocytes, eosinophils, and lymphocytes. Chemokines have a low sequence homology, but all show conserved structural homology.

Figure 4

Chemokine classes. Chemokines are classified in four distinct categories according to the number and spacing of their cysteine residues: C, CC, CXC, and CX3C. Chemokines have a glycosaminoglycan-binding domain and a GPCR-binding domain. CX3C chemokines are larger than the other chemokines and contain an additional mucin-like domain and have transmembrane and cytoplasmic domains.

Chemokines interact with GPCRs termed chemokine receptors. Interaction with the chemokine receptors results in the transduction of the signal and thus the activation of effector cells. Chemokine binding of target GPCRs varies from highly specific to highly promiscuous. To date, seven receptors have been described for CXC-chemokines (CXCR1-7), 10 receptors for CC-chemokines (CCR1-10), and one for CX3CL (CX3CR1) and XCL1 (XCR1) (IUIS/WHO Subcommittee on Chemokine Nomenclature, 2003; Balabanian et al., 2005). Chemokines not only bind to chemokine receptors, but they also interact with glycosaminoglycans through distinct sites (Witt & Lander, 1994; Kuschert et al., 1999; Ali et al., 2000). These long unbranched polysaccharides consisting of a repeating disaccharide unit allow for presentation of chemokines by endothelial cells or extracellular matrix. Thereby, they create a chemotactic gradient that directs leukocyte transmigration of the endothelial layer or movement through the tissue, respectively (Tanaka et al., 1993; Weber et al., 1999).

Chemokine evasion

As infections stimulate the production of chemokines that orchestrate the inflammatory response against the invading pathogen, it is not surprising that various pathogens have evolved strategies to block or modulate chemokine-induced responses. Viruses have evolved different strategies for chemokine evasion than those seen in bacteria. Viruses modulate chemokine signaling through downregulation of chemokine receptors in the host or the production of virally encoded chemokine mimics, chemokine receptor mimics, or chemokine-binding proteins with no homology with host chemokine receptors that directly scavenge chemokines. Bacterial proteins on the other hand are seldom host derived, and they commonly target the chemokine receptors, either for antagonism or receptor cleavage. Other bacterial proteins interact with the chemokines themselves to scavenge them from the chemokine receptors through other cell surface-expressed receptors or to cleave them, abrogating the ability to activate chemokine receptors. The reader is referred to several reviews that describe the viral chemokine evasion in detail (Alcami et al., 2003; Boomker et al., 2005; Rosenkilde et al., 2008). In this review, we will only provide a brief overview of the mechanisms used by viruses and will elaborate on chemokine-modulating mechanisms of other pathogens.

Modulation through chemokine receptor binding

Staphylococcus aureus is described to secrete two related proteins that affect chemokine signaling by binding to chemokine receptors. Staphylococcal superantigen-like (SSL) 10 specifically inhibits CXCL12-mediated cell responses by targeting CXCR4 (Walenkamp et al., 2009), while SSL5 is a broad-spectrum chemokine inhibitor that affects leukocyte activation by all classes of chemokines (Bestebroer et al., 2009). These two SSLs belong to the family of 14 staphylococcal proteins that were initially identified through sequence and structural homology to superantigens (Williams et al., 2000). Superantigens are secreted proteins that can activate large fractions of T-cell populations at very low concentrations. Their potent immunostimulatory activity results from simultaneous binding of the T-cell receptor found on T cells and the major histocompatibility complex (MHC) class II molecules on the surface of antigen-presenting cells. As superantigens, SSLs have a two-domain structure composed of an N-terminal β-barrel domain and a C-terminal β-grasp motif that are joined through an α-helix (Arcus et al., 2002; Al-Shangiti et al., 2004). Despite the high homology to superantigens, SSLs do not induce aspecific T-cell activation. Their lack of superantigenic activity is ascribed to the differences observed in the regions essential for the binding of T-cell receptors and MHC class II molecules by superantigens. Genome analysis of sequenced S. aureus strains (Fitzgerald et al., 2003) revealed that SSLs are clustered on a genomic island encoding a restriction modification system and a truncated putative transposase. This putative pathogenicity island, termed SaPI2, encodes 7–11 SSLs and is present in all S. aureus strains. Three more SSLs can be found on immune evasion cluster 2 (Jongerius et al., 2007). The sequence identity among SSLs varies between 23% and 64%, while allelic variation lies between 85% and 100% (Fitzgerald et al., 2003; Holtfreter et al., 2004). Because of their homology to superantigens and association with genomic islands, SSLs were predicted to influence immune responses. Another clue to the possible role for SSLs in immune modulation was the structural homology to CHIPS (Haas et al., 2005).

SSL10 has recently been demonstrated to inhibit cell responses mediated by the chemokine CXCL12 (or stromal cell-derived factor-1α) (Walenkamp et al., 2009). In a screen to identify bacterial inhibitors for the chemokine receptor CXCR4 using several staphylococcal proteins, SSL10 blocked the binding of an antibody directed against CXCR4. Importantly, when cells were treated with SSL10, calcium mobilization and chemotaxis in response to CXCL12 were abrogated. CXCL12-induced intracellular signaling was also affected in SSL10-pretreated cells. It must be noted that CXCL12 not only acts through CXCR4 but also targets CXCR7 for its effects. SSL10, however, does not block binding of a function-blocking antibody directed against CXCR7. SSL10 is thus specific for CXCR4-mediated responses. Indeed, in experiments with human neutrophils, SSL10 did not inhibit CXCL8- or C5a-induced calcium responses. Its antagonistic effect is exerted by interacting with CXCR4, as confirmed by transfection experiments.

In addition to SSL10, SSL5 has been recently described to inhibit chemokine-induced leukocyte activation (Bestebroer et al., 2009). SSL5 was the first SSL with identified crystal structure (Arcus et al., 2002) and was initially demonstrated to exhibit immune-modulating properties by targeting P-selectin glycoprotein ligand-1 (PSGL-1) (Bestebroer et al., 2007). This sialomucin is an important adhesion receptor that binds P-selectin encountered on activated endothelial cells and platelets. It is essential in leukocyte recruitment to inflamed or infected sites (Cummings et al., 1999). By binding PSGL-1, SSL5 blocks the interaction with its natural ligand P-selectin. More importantly, SSL5 effectively abrogates neutrophil rolling on endothelial cells. SSL5 binding to PSGL-1 is glycan dependent. Subsequent data on SSL5 revealed its structure in complex with sialyl Lewis x (sLex) (Baker et al., 2007), the predominant sugar moiety on PSGL-1. Similar glycan-dependent effects have been recently described for SSL11 (Chung et al., 2007), suggesting that more homologous SSLs target glycoproteins. We have recently demonstrated that SSL5 inhibits leukocyte responses to CXC-, CC-, and CX3C-chemokines (Bestebroer et al., 2009). SSL5 also inhibits leukocyte responses to the complement fragments C3a and C5a. SSL5 not only affects the calcium mobilization induced by all tested chemokines but also abrogates chemokine-driven actin polymerization and chemotaxis. Importantly, these antagonistic effects were also demonstrated to be glycan dependent. Competition experiments with antibodies directed against several chemokine receptors demonstrated the interaction of SSL5 with this class of GPCRs. Subsequent transfection experiments showed that SSL5 directly binds the N-termini of chemokines receptors. Its binding was also dependent on the presence of glycans. As not only chemokine receptors but all GPCRs are known to be glycosylated, the interaction of SSL5 with the N-termini of GPCRs responding to other chemoattractants was investigated. Indeed, SSL5 also binds to the FPR, FPRL-1, leukotriene B4 receptor, platelet-activating factor receptor, and the nucleotide receptor P2Y2. However, SSL5 does not inhibit the stimulation of these receptors by fMLP, MMK-1, leukotriene B4, platelet-activating factor, and ATP, respectively. Inhibition of a subset of GPCR ligands by SSL5 thus appears to be inherent in the ligand. The lipid-, bacterial-, and nucleotide-derived stimuli primarily interact with the transmembrane regions of their target GPCR, while the N-terminal domain does not seem to play an important role for either of these ligands. Chemokines on the other hand bind the N-terminus of their receptors, whereupon they are positioned to interact with the pocket formed by the transmembrane domains, which is necessary for activation of the receptor. SSL5 thus targets the glycosylated N-termini of all GPCRs and thereby only inhibits stimuli of protein nature that require the receptor N-terminus for activation.

Rather than binding chemokine receptors for antagonism, some pathogens secrete proteins that modulate GPCR responses through activation of the receptors. Mycobacterium avium produces such a molecule with chemotactic activity (Rao et al., 2002). Mycobacterial supernatants are described to induce chemotaxis of macrophages, and they contain a CCL2-like molecule that mediates the response. This mycobacterial protein of about 34 kDa is recognized by antibodies directed against human CCL2, indicating that mycobacterial and human CCL2 share common epitopes. Toxoplasma gondii also produces a protein that targets chemokine receptors for activation. It has a major role in driving CCR5-dependent IL-12 production that is beneficiary for the protozoan chronic infection (Aliberti et al., 2003). This cyclophilin-I8 shows high affinity for CCR5 and upon binding triggers cell signaling, as measured by calcium mobilization and IL-12 production, and chemotactic migratory responses. Cyclophilin-I8 has no obvious sequence homology with known CC-chemokines and lacks their characteristic β-sheet and α-helical domains.

Viruses can encode chemokine homologues that function either as receptor agonists or their antagonists. The human herpesvirus 8 (HHV8) encodes the broad-spectrum chemokine antagonist vMIP2 (Kledal et al., 1997; Crump et al., 2001). It binds to the receptors for chemokines from all four classes, blocking their activity, without inducing activation of the chemokine receptors itself. A similar mechanism is used by the molluscum contagiosum virus (MCV) by means of MC148 (Damon et al., 1998). This protein inhibits several CC- and CXC-chemokines, and it is suggested to bind CCR8 (Luttichau et al., 2000). The human cytomegalovirus on the other hand expresses vCXCL1 that binds CXCR2 and enables chemotaxis of targeted cells (Penfold et al., 1999). Also, the HHV8 encodes viral chemokines that induce migration as exemplified by vMIP1 that acts on CCR4 and CCR8.

Chemokine binding and scavenging

In addition to GPCR binding, SSL5 displays another mechanism to inhibit chemokine-induced leukocyte activation (Bestebroer et al., 2009). When the effect of SSL5 on chemokine binding was examined, a surprising increase in the binding of three tested CXC- and CC-chemokines was observed, while the stimulation of cells with these chemokines was inhibited by SSL5. The presence of glycans at the cell surface was demonstrated to be crucial and the effect was independent of the expression of chemokine receptors. Moreover, SSL5 was not able to directly interact with the chemokines, but required cell surface-associated receptors. Thus, SSL5 scavenges chemokines from the chemokine receptors through other glycoproteins. The question of how SSL5 interacts with different classes of chemokines was resolved using heparan sulfate-loaded chemokines. Chemokines are not only structurally similar, but all contain a highly homologous glycosaminoglycan-binding site, which allows their presentation by endothelial cells or extracellular matrix. When the glycosaminoglycan-binding site of chemokines is occupied by the glycosaminoglycan heparan sulfate, SSL5 no longer induces chemokine binding to cells. SSL5 thus targets the glycosaminoglycan-binding site common among all chemokines.

Although not a pathogen, the common brown dog tick Rhipicephalus sanguineus secretes several proteins that directly bind chemokines and neutralize their activity (Hajnicka et al., 2001). These so-called Evasins are found in the saliva of this blood-sucking parasite and have stringent chemokine selectivity. Evasin-1 binds to CCL3, CC4, and CCL18 (Frauenschuh et al., 2007), while Evasin-3 binds to CXCL8 and CXCL1 and Evasin-4 binds to CCL5 and CCL11 (Deruaz et al., 2008). Through chemokine binding, Evasins block their interaction with target chemokine receptors. The structure of Evasin-1 and Evasin-3 shows them to be distinct proteins. Both proteins, however, show anti-inflammatory activities in animal models. The helminthic parasite Schistosoma mansoni also secretes a protein into host tissues that binds certain chemokines. This so-called S. mansoni chemokine-binding protein inhibits chemokine interaction with host chemokine receptors and thereby impairs their biological activity (Smith et al., 2005). Schistosoma mansoni chemokine-binding protein is unrelated to host chemokine receptors and is only found in schistosome egg secretions but not in other life cycle stages. The protein can be found in glycosylated forms, although glycosylation is not important for its activity, and it suppresses inflammation in several disease models.

Viruses show two strategies to modulate chemokine responses by means of chemokine binding and scavenging. They express virally encoded chemokine receptors and chemokine-binding proteins not homologous to any host chemokine receptor. Virally encoded chemokine receptors manipulate chemokine responses by scavenging host chemokines or directly affect host cell responses to chemokines. HHV8 ORF74 encodes a viral CXC-chemokine receptor that is constitutively active. Thereby, activation of the infected cells is ligand independent and is linked to induce cell proliferation (Arvanitakis et al., 1997). The herpesvirus saimiri homologue ECRF3 is described to show constitutive as well as a ligand-regulated signaling (Rosenkilde et al., 2004). This receptor induces calcium mobilization of infected cells in response to CXCL1 and CXCL8 (Ahuja & Murphy, 1993). Human cytomegalovirus carries UL33, US27, and US28 that are homologous to host CC-chemokine receptors (Chee et al., 1990; Margulies et al., 1996). Only US28 has been shown to be a functional receptor (Neote et al., 1993; Gao & Murphy, 1994). US28 constitutively traffics to and from the cell surface to drive chemokine uptake (Bodaghi et al., 1998; Fraile-Ramos et al., 2001); thereby it sequesters CC-chemokines, making them unavailable for host receptors.

Viral chemokine-binding proteins, without known host receptor homology, target host chemokines and neutralize their ability to activate effector cells. Myxomaviruses, poxvirus pathogen of rabbits, produce M-T7 that interacts with CC-, CXC-, and XC-chemokines through interaction with their glycosaminoglycan-binding domain (Lalani et al., 1997). As a consequence, chemokines are released from glycosaminoglycans, which disrupts their presentation to leukocytes. The myxomavirus protein M-T1 selectively binds to various CC- chemokines with a high affinity. Thereby, it prevents the chemokine receptor activation and influences influx of leukocytes to sites of viral infection (Graham et al., 1997; Lalani et al., 1999). For the chemokine CCL2, it is described that M-T1 targets its receptor-binding domain and thereby prevents the binding of the chemokine to its cellular receptors (Lalani et al., 1998). Finally, some viral chemokine-binding proteins apply both strategies. The HHV8-encoded M3 protein (van Berkel et al., 2000) is secreted by infected cells and binds to various chemokines from all four subclasses. The crystal structure of M3 in complex with CCL2 and XCL1 has been determined (Alexander et al., 2002; Alexander-Brett & Fremont, 2007). The C-terminal domain of M3 targets the GPCR-binding domain of chemokines, whereas its N-terminal domain interacts with the glycosaminoglycan-binding domain of chemokines. M3 thus mimics both the sites important for chemokine presentation and activation. M3 is also described to bind CXCL8 and a similar mechanism of chemokine inhibition was observed for this CXC-chemokine (Webb et al., 2003).

Chemokine cleavage

Some secreted bacterial proteins not only bind chemokines to prevent their receptor interactions but also neutralize them through cleavage. Group A Streptococcus shows protease activity toward the neutrophil-recruiting chemokine CXCL8 (Hidalgo-Grass et al., 2004). This CXCL8-cleaving activity was identified as belonging to Streptococcus pyogenes cell envelop protease (SpyCEP), also known as streptococcal chemokine protease (SpcC) (Edwards et al., 2005; Hidalgo-Grass et al., 2006). SpyCEP not only degrades human CXCL8 but is also capable of cleaving the mouse orthologues CXCL1 and CXCL2, while cleavage of stimuli such as tumor necrosis factor α, IL-1β, interferon γ, IL-6, or complement factor C5a remain unaffected. The degradation of CXCL8 is the result of a single specific cleavage between Gln-59 and Arg-60 within the CXCL8 C-terminal α-helix (Edwards et al., 2005). Because of its protease activity, SpyCEP prevents CXCL8-induced shedding of l-selectin, cell adhesion molecule, by neutrophils. SpyCEP activity in vivo was related to the observation that necrotizing fasciitis caused by isolates of invasive group A streptococci is correlated to the striking paucity of infiltrating neutrophils. Subsequent studies by Sumby et al. (2008) demonstrated that SpyCEP not only degrades CXCL8 and its orthologues but also other CXC-chemokines, namely CXCL1 and CXCL6. SpyCEP is described to inhibit neutrophil priming by these chemokines, as was measured by monitoring CD11b expression. SpyCEP does not induce neutrophil priming itself. In addition to its presence in culture media, SpyCEP is also shown to be surface expressed. SpyCEP expression is negatively regulated by the CovR/S two-component regulatory system. Group A streptococci often carry mutations in CovR/S, and these mutants upregulate transcription of SpyCEP >25-fold (Sumby et al., 2008). Surprisingly, in a mouse soft tissue model, the SpyCEP-deficient mutant strain of group A Streptococcus generates significantly larger skin lesions than parental strains. Subsequent study by Sjolinder et al. (2008) also investigated the effect of an SpyCEP-deficient mutant strain. Here, the SpyCEP-deficient strain had the ability to recruit immune cells during soft tissue infections in mice, in contrast to the wild-type strain. The SpyCEP mutant induced more severe sepsis with higher bacteremia and mortality rates. Thus, the SpyCEP mutant triggers high neutrophil infiltration, but not a lethal outcome, after soft tissue infection, whereas intravenous infection leads to highly aggressive systemic disease. These data were very recently confirmed by Zinkernagel et al. (2008), who demonstrated that the knockout mutant lacking SpyCEP was attenuated for virulence in murine infection models. SpyCEP decreased CXCL8-dependent neutrophil endothelial transmigration and bacterial killing, the latter by reducing neutrophil extracellular trap formation. They also showed that the zoonotic pathogen Streptococcus iniae possesses a functional homologue of SpyCEP, termed CepI, which cleaves CXCL8, promotes neutrophil resistance, and contributes to virulence. In conclusion, the SpyCEP protease impairs neutrophil clearance mechanisms, contributing to the pathogenesis of invasive streptococcal infection.

Similar effects on chemokine degradation have been described for Porphyromonas gingivalis (Zhang et al., 1999). Here, after infection with this periodontal pathogen, accumulation of CXCL8 produced by gingival and oral epithelial cells in response to the pathogen could not be detected. The infected epithelial cells were still capable of expressing CXCL8, but incubation of CXCL8 with P. gingivalis resulted in a rapid loss of the chemokine. In the presence of a protease inhibitor, the loss was significantly retarded. Together, the results indicated that CXCL8 is degraded by extracellular proteases. Activity of a gingivalis cysteine and/or serine protease was also demonstrated for CCL2 (Kobayashi-Sakamoto et al., 2003). It is suggested that gingipains are responsible for the protease activity (Mikolajczyk-Pawlinska et al., 1998).

Supernatants of P. aeruginosa also display chemokine-degrading properties (Leidal et al., 2001, 2003). Metalloproteases elastase and alkaline protease cleave CXCL5, CCL2, and CCL5, while CXCL8 is considerably more resistant to proteolysis. Degradation of chemokines is time, temperature, and concentration dependent, and results in the loss of chemotactic activity for CXCL5, CCL2, and CCL5. The basis for the sensitivity of some chemokines, but not others, to bacterial proteases is uncertain.

In addition to bacteria, the helminthic parasite Necator americanus secretes a protease that can cleave chemokines (Culley et al., 2000). Initial infection studies demonstrated that secreted products of N. americanus inhibit eosinophil recruitment in vivo in response to the chemokine CCL11, but not the chemoattractant leukotriene B4. Treatment with protease inhibitors EDTA and phenanthroline demonstrated that the active enzyme is a metalloprotease. The yet unidentified protein specifically cleaves CCL11 and not CXCL8 or its close homologue CCL24.

Glycosaminoglycan–chemokine interaction-modulating proteins

Chemokines are displayed through glycosaminoglycans on proteins of the extracellular matrix or endothelial tissues to subsequently activate target cells (Petersen et al., 1998; Kuschert et al., 1999). Bacteria and viruses have possibly evolved mechanisms to disrupt the chemokine–glycosaminoglycan interaction and thereby abrogate the chemokine gradient formed. Staphylococcal SSL5 and virally encoded chemokine-binding proteins are described to interact with the glycosaminoglycan-binding domain of chemokines (Lalani et al., 1997; Bestebroer et al., 2009). SSL5 uses the mechanism to scavenge the chemokines through glycoproteins and thereby fix them on cell surfaces, rather than disrupt the chemotactic gradient (Bestebroer et al., 2009). Several bacteria, however, express enzymes that can cleave the glycosaminoglycan–protein linkage region and thereby generate soluble glycosaminoglycans (Tsuda et al., 1999). Such chondroitinases and heparinases are described for Proteus vulgaris, Flavobacterium heparinum, and Arthrobacter aurescens. Liston & McColl (2003) have previously proposed that generation of soluble glycosaminoglycans may disrupt the chemokine gradient and abrogate chemokine-directed movement and activation.

Complement activation of GPCRs

Small complement fragments C3a and C5a act on receptors that belong to the GPCR family. These stimuli, also commonly called anaphylatoxins, are generated during the activation of the complement cascade. The complement system can be activated via three separate pathways that differ in their mode of recognition (Fig. 5) (Walport et al., 2001; Gasque et al., 2004). Antibody-mediated complement activation is called the classical pathway. The lectin pathway is activated upon recognition of conserved microbial sugars by mannan-binding lectin and ficolins. The alternative pathway acts as an amplification loop after C3b is formed. However, it can also be activated spontaneously by interacting with hydrolyzed C3. All the three pathways converge by the formation of C3 convertases (Sim & Laich, 2000). These complexes cleave C3, which results in the release of the small chemoattractant C3a and the deposition of C3b molecules on the microbial surface. C3b deposition is crucial for the eradication of microorganisms. The C3 convertases can also bind an additional C3b molecule to form C5 convertases. The C5 convertase cleaves C5, resulting in the release of the chemoattractant C5a and the sequential deposition of C5b-C9, which leads to the formation of the so-called membrane attack complex (Pangburn & Rawal, 2002). Formation of the membrane attack complex is responsible for direct killing of Gram-negative bacteria through lysis, while Gram-positive bacteria are rather resistant to this response due to their thick peptidoglycan cell wall (Muller-Eberhard et al., 1986). The formation of the small cleavage products C3a and C5a plays an important role in the attraction of phagocytes to the site of infection and priming and activation of phagocytes for bacterial uptake. Generation of C3a and C5a is therefore of utmost importance in the clearance of Gram-positive bacteria (Easmon & Glynn, 1976; Mullaly & Kubes, 2006). The immune responses initiated by C3a and C5a are mediated through binding to the C3a receptor (C3aR) and the C5a receptor (C5aR), both belonging to the GPCR family.

Figure 5

Complement activation. Complement activation can be initiated via three different pathways. The classical pathway is triggered by antibody-bound bacteria, while the lectin pathway specifically recognizes microbial sugar moieties with mannose-binding lectin (MBL) or ficolins. The alternative pathway is activated spontaneously or amplifies complement on bacteria previously opsonized via the classical and lectin pathway. Upon initiation, signaling pathways are activated. All pathways converge at the level of C3 by the formation of C3 convertases. The C3 convertases are responsible for the generation of C3b and iC3b, two components that mediate bacterial uptake by phagocytes via complement receptors, and the chemoattractant C3a. Incorporation of another C3b molecule into existing C3 convertases generates C5 convertases. C5 convertases cleave C5, resulting in the release of C5a and the sequential deposition of C5b-C9, which leads to the formation of the so-called membrane attack complex (MAC). The small active fragments C3a and C5a are chemoattractants. The figure was produced using Servier Medical Art.

Complement evasion

Several pathogens have developed several strategies to abrogate the generation of or the responses to the anaphylatoxins. They can abrogate C3a- and C5a-induced signaling by inhibiting the C3 and C5 convertases responsible for their generation by preventing the cleavage of C3 or C5, by degrading the generated anaphylatoxins, or by antagonizing the target receptors C3aR and C5aR. Many pathogens have also evolved mechanisms that interfere with complement activation and regulation by other means. Here, only mechanisms directly related to C3 and C5 will be discussed. The general complement escape is described elsewhere (Kraiczy & Wurzner, 2006; Lambris et al., 2008).

Inhibition of C3a and C5a generation

Staphylococcus aureus secretes a number of proteins that effectively impede the complement system (as reviewed in Rooijakkers et al., 2005b; Lambris et al., 2008). Briefly, several of these proteins accomplish complement inhibition by blocking the activity of convertases. Staphylococcal complement inhibitor (SCIN) was the first described convertase inhibitor (Rooijakkers et al., 2005a). SCIN is an excreted protein of 9.8 kDa that interferes with all complement activation pathways by blocking C3 convertases. By binding C3 convertases, SCIN directly blocks their enzymatic activity. Furthermore, SCIN stabilizes the C3 convertases and thereby prevents the formation of new convertases. Through this dual activity, SCIN effectively inhibits the opsonization, phagocytosis, and killing of S. aureus by neutrophils and abrogates the generation of C5a and associated neutrophil chemotaxis. SCIN is produced in 92% of S. aureus strains. In addition to SCIN, S. aureus excretes two homologues termed SCIN-B and SCIN-C with the same mechanism of action (Jongerius et al., 2007). Staphylococcal extracellular fibrinogen-binding protein (Efb) and extracellular complement-binding protein (Ecb or Ehp) also block the conversion of C3. Efb and Ecb bind the thioester-containing the C3d domain within C3 and thereby these proteins block C3b-containing convertases, i.e. the C3 convertases of the alternative pathway and the C5 convertases of all complement pathways. Thereby, Ecb and Efb inhibit C3b deposition via the alternative pathway and more potently block C5a formation via all pathways. Inhibition of the C5a generation was demonstrated in vivo using a mouse model in which Ecb completely inhibited C5a-dependent neutrophil influxes. Although SCIN, Efb, and their homologues display different mechanisms of complement inhibition, they show a remarkably high structural homology. The complement inhibitors are composed of three parallel α-helices (Hammel et al., 2007; Rooijakkers et al., 2007).

In addition to convertase inhibitors, generation of complement fragments can be blocked by binding to C3 or C5 directly and thereby prevent their conversion. Staphylococcal protein SSL7 is such a protein, and it specifically binds C5 (Langley et al., 2005). SSL7 also interacts with immunoglobulin A and prevents its binding to the FcαRI. Through C5 binding, SSL7 was shown to inhibit complement-mediated cell lysis of erythrocytes and E. coli, presumably by blocking the C5b-9 formation. Concomitantly, SSL7 also prevents the generation of C5a.

Stimulus degradation

Streptococci have developed mechanisms to neutralize generated anaphylatoxins. Virulent strains of S. pyogenes express a cell wall-anchored C5a peptidase termed ScpA (O'Connor & Cleary, 1986; Stafslien & Cleary, 2000). This peptidase inhibits the chemotactic activity of neutrophils towards serum pre-exposed to zymosan. ScpA activity is highly specific for the anaphylatoxin C5a and does not affect its precursor C5. C5a is cleaved by ScpA between residues His-67 and Lys-68 (Cleary et al., 1992), which is at the receptor-binding site on neutrophils. ScpA is demonstrated to be a virulence factor for group A streptococci as in animal models ScpA-deficient strains are cleared more efficiently than the wild-type strains. ScpA affects normal host inflammatory responses by delaying the accumulation of neutrophils at streptococcal infection sites (O'Connor & Cleary, 1987; Ji et al., 1996). Recently, it was demonstrated that ScpA-mediated cleavage of C5a by group A streptococci requires the expression of a plasmin receptor, also known as surface dehydrogenase or glyceraldehyde-3-phosphate dehydrogenase (Terao et al., 2006). While ScpA weakly binds C5a, plasmin receptor binds C5a with a high affinity. Plasmin receptor, however, does not have protease activity itself. It thus seems that the plasmin receptor is responsible for the recruitment of C5a to the cell surface, while ScpA mediates the cleavage, resulting in an effective C5a degradation. Soluble plasmin receptor is also capable of binding C5a and can inhibit C5a-induced neutrophil chemotaxis and oxidative burst. C5a peptidases similar to ScpA are also expressed by group B (ScpB) (Hill et al., 1988; Bohnsack et al., 1991) and group G (Cleary et al., 1991) streptococci. The C5a peptidase of group B streptococci also affects the rapid neutrophil recruitment to sites of infection (Bohnsack et al., 1997).

In addition to streptococci, there is evidence that proteases from Serratia marcescens and P. aeruginosa also degrade C5a. In serratia, a 56-kDa protease results in the degradation of C5a into an inactive molecule (Oda et al., 1990). Pseudomonas elastase, on the other hand, is highly destructive for fluid-phase and cell-bound C3 and fluid-phase C5 (Schultz & Miller, 1974). The elastase generates and then inactivates a chemotactic factor from human C5 but not from C3. Gingipains of the oral anaerobic bacterium P. gingivalis can also cleave C3 and C5 (Wingrove et al., 1992; Discipio et al., 1996; Popadiak et al., 2007). Instead of degrading the complement factors, these proteases release a chemotactically active C5a molecule. It is proposed that through this mechanism the bacterium attracts neutrophils and exacerbates periodontal disease.

Recently, complement evasive properties for Enterococcus faecalis have been described. Elastase of this bacterium cleaves C3 into C3b-like molecules, which leads to consumption of C3 from serum and degrades iC3b molecules deposited on antigen surfaces (Park et al., 2008). These proteolytic activities contribute to reduced phagocytosis of E. faecalis by neutrophils. Importantly, elastase also effectively degrades C3a (Park et al., 2007).

Receptor cleavage and antagonism

Some pathogens modulate GPCR activation by C5a by targeting the C5aR. Lys-gingipains from P. gingivalis can cleave the C5aR (Jagels et al., 1996). This mechanism was identified after diminished binding of an antibody directed against the N-terminus of C5aR was observed upon exposure of neutrophils to gingipains. Evidence of cleavage rather than receptor antagonism was demonstrated through mass spectral analysis of gingipain-treated C5aR N-terminal peptides. Lys-gingipains cleave the C5aR N-terminal peptide at Lys-17 and Lys-28. In addition to Lys-gingipain, gingivalis vesicles carry other proteases that carry C5aR-cleaving capacity. While Lys-gingipain activity seems limited to the N-terminus of C5aR, the other proteases also include other ligand-binding sites and/or signal transduction elements. Both types of proteases can, however, inhibit C5a-mediated neutrophil responses such as enzyme release and calcium mobilization.

Two staphylococcal proteins are described to target the C5aR and/or C3aR and prevent their activation. Rather than cleaving the receptors, they exert their effect through antagonism. As described above, CHIPS inhibits fMLP-induced responses through its binding to the FPR. Additionally, CHIPS binds the C5aR and inhibits C5a-induced cell responses such as calcium mobilization, up- and downregulation of CD11b and CD62L, and elastase release in neutrophils (de Haas et al., 2004; Postma et al., 2004). CHIPS exerts its effect through binding to amino acids 10–18 in the N-terminus of the C5aR (Postma et al., 2005). While the N-terminus of CHIPS was demonstrated to be essential for its activity on the FPR, deletion of the first 30 amino acids did not affect its C5aR-inhibitory activity (Haas et al., 2004), suggesting two different active sites. Recently, the 3D structure of the C5aR-blocking domain of CHIPS (CHIPS31–121) was resolved (Haas et al., 2005). CHIPS31–121 is composed of an α-helix packed onto a four-stranded antiparallel β-sheet. This domain is structurally homologous to the C-terminal domain of SSL5 (Arcus et al., 2002) and SSL7 (Al-Shangiti et al., 2004), suggesting that some of the SSLs may carry GPCR antagonistic effects.

Indeed, as described above, it was demonstrated that SSL10 displays CXCR4-inhibitory activity, while SSL5 is a broad-chemokine receptor inhibitor. In addition to its inhibition of chemokine-induced leukocyte activation, SSL5 also potently antagonizes C3a- and C5a-induced responses (Bestebroer et al., 2009). SSL5 effectively blocks calcium mobilization, actin polymerization, and chemotaxis towards C3a and C5a. SSL5 binds the C5aR as SSL5 blocked antibody binding recognizing the N-terminus of the C5aR. Furthermore, SSL5 binds to cells transfected with the N-termini of the C3aR and C5aR. As for the chemokine receptors and other GPCRs, SSL5 recognizes glycan epitopes in these receptors. Additional determinants are suggested to be important in its binding to GPCRs, as SSL11, another glycan-recognizing SSL, does not show the effects observed for SSL5.

Concluding remarks

Chemoattractants are important in the recruitment of leukocytes to sites of inflammation and infection (Fig. 6). Several classes of chemoattractants acting through GPCRs are generated in response to invading pathogens. These receptors are thus crucial and central for our immune system. Pathogens have evolved a diverse array of mechanisms to antagonize the effector functions directed by the chemoattractants. Understanding their mechanism of action not only contributes to our knowledge on microbial pathogenesis, but in addition informs us how GPCRs function and provides interesting approaches for the development of novel therapeutics. Many autoimmune and chronic inflammatory diseases such as stroke, reperfusion/ischemia, transplant rejection, rheumatoid arthritis, and tumor progression are mediated through chemokines and anaphylatoxins. Cancer metastasis is also driven by different chemoattractants. In addition to providing potential anti-inflammatory compounds, current efforts to resolve mechanisms modulating GPCR signaling may also contribute to the identification of new antimicrobial targets. This is especially valuable in cases where high antibiotic resistance prevents effective eradication of the infective agent.

Figure 6

Leukocyte extravasation and modulation of GPCR signaling. Leukocyte extravasation from blood into tissue and migration to the site of infection require multiple sequential steps. Phagocytes such as monocytes and neutrophils are most important here. Initially, phagocytes tether and roll on activated endothelium. Following stimulation of phagocytes by chemokines, they firmly adhere to the vessel wall. The cells then transmigrate through the endothelium into the underlying tissue and migrate to the site of infection through the sensing of various chemoattractants by GPCRs. Chemoattractants are either pathogen-derived products or are produced by the host, such as chemokines and complement fragments. Several pathogens produce proteins that affect GPCR signaling through a variety of mechanisms: mimicry, cleavage, antagonism or scavenging. PSGL-1, P-selectin glycoprotein-1; ICAM, intercellular adhesion molecule-1; VCAM, vascular adhesion molecule-1; JAM, junctional adhesion molecule; C3 conv., C3 convertase; C5 conv., C5 convertase; MAC, membrane-attack complex; FPR, formylated peptide receptor; FPRL-1, FPR-like 1; SSL, staphylococcal superantigen-like; SCIN, staphylococcal complement inhibitor; Efb, extracellular fibrinogen-binding protein; Ecb, extracellular complement-binding protein; Scp, streptococcal C5a peptidase; fMLP, formylated peptides; PSM, phenol-soluble modulins; CHIPS, chemotaxis inhibitory protein of Staphylococcus aureus; FLIPr, FPRL-1-inhibitory protein; vCBP, viral chemokine-binding protein; SpyCEP, Streptococcus pyogenes cell envelope protease; vGPCR, virally encoded GPCR; vC, virally encoded chemokine. The figure was produced using Servier Medical Art.

Certainly, the importance of GPCR modulators for microbial pathogenesis needs to be verified. For some proteins, an association has been found between the presence of the modulators and virulence of the pathogen. In the case of S. pyogenes, a mutant lacking SpyCEP, which cleaves the chemokine CXCL8, was attenuated for virulence in mouse infection models. Pertussis and cholera toxin are well-established virulence factors; however, for other modulators, further research needs to be performed. Proteins such as staphylococcal Ecb and several of the evasins from R. sanguineus show anti-inflammatory properties in animal models of disease. However, animal infection experiments using bacterial knockout strains are difficult due to functional redundancy with other proteins. However, these experiments are now ongoing in several labs. Another problem arises when proteins are human specific, for example staphylococcal CHIPS and SCIN. In such cases, the use of knockout S. aureus strains in animal infection models is not informative with regard to their virulence in humans. The solution in the near future is the use of humanized mice in which the mouse receptor is knocked out and replaced with a copy of the human receptor.

In this review, we have provided an overview of proteins secreted by major human pathogens and by less virulent microorganisms that affect GPCR signaling through various mechanisms. Undoubtedly, microorganisms will be shown to produce a much larger arsenal of GPCR inhibitors, especially because GPCRs are central in the pathophysiology of infectious diseases. Further research is thus warranted to identify novel GPCR modulators that will increase insight into microbial pathogenisis as well as into the molecular basis of our own defense system.


This work was supported by the Technology Foundation STW (#UKG-6609), the Netherlands Organisation for Scientific Research NWO-TOP (#9120.6020), and EU-STREP AMIS (#FP6-512093). The authors declare no competing financial interests.


  • Editor: Neil Fairweather


View Abstract