Innate-immune cells can recognize several pathogen-associated molecular patterns (PAMPs) secreted or present at the bacterial cell wall via specific pathogen recognition receptors (PRRs). In mammals, these include toll-like receptors (TLRs), integrins, scavenger receptors, and C-type lectins (76). TLRs monitor the cell surface and endocytic compartments and activate cell-autonomous defenses upon detection of PAMPs, while lectins, scavenger receptors, and integrins function as phagocytic receptors that bind to particles and are able to trigger uptake even in non-phagocytic cells (76). TLRs contain ligand-binding leucine-rich repeats (LRRs) in the luminal/extracellular domains and a toll/interleukin 1 receptor (TIR) domain in the cytoplasmic tail that mediates protein–protein interactions. D. discoideum does not have TLRs, but two cytosolic proteins with TIR domains, TirA and TirB, have been identified. Depletion of TirA inhibits growth on a laboratory strain of Klebsiella used as a food bacterium, but not in media (77). How TirA is involved in sensing and/or induction of phagocytic uptake remains to be studied. The D. discoideum genome encodes >150 LRR-containing proteins, but whether they function as PRRs remains to be determined (5).

In D. discoideum, only a few phagocytic receptors have been molecularly identified so far (5). The most studied are the integrin-like Sib (similar to integrin-β) family of proteins, comprising five members (SibA-E). Like human integrins, they contain a von Willebrand factor type A and a glycine-rich transmembrane domain and can interact with the actin-binding protein talin (78). Two members of the family, SibA and SibC, are directly involved in adhesion to substrate and to particles (78). Phg1a, a member of the TM9 family, and SadA were previously identified as phagocytic receptors (79, 80); however, recent findings show that these two proteins are instead involved in regulating surface levels of SibA (81).

Lectins and scavenger receptors might also function as phagocytic receptors in D. discoideum. Three homologs of scavenger receptor class B proteins in mammals, LmpA, LmpB, and LmpC, are present in D. discoideum. LmpB is found on lipid rafts at the plasma membrane and in early phagocytic compartments and may function as a phagocytic receptor (82–84). It is thought that D. discoideum also possesses lectin-like receptors, as it was shown to be able to bind specifically to certain sugar derivatives (85, 86).

A well-studied chemoattractant and phagocytosis stimulator for D. discoideum is folate, a metabolite secreted by certain bacteria. Recently, Pan and colleagues identified fAR1, a G-protein coupled receptor for folic acid involved in signaling and initiation of uptake but not binding of bacteria (87). Presumably, the cytoskeletal rearrangements downstream of fAR1 that facilitate chemotaxis can also initiate phagocytosis when a burst of fAR1 activation occurs around a bacterium that is a concentrated source of folate. D. discoideum sensing of bacterial capsule independent of folate has also been described but not further characterized (88).

After binding of ligands to their receptors, heterotrimeric G proteins are involved in downstream signaling to initiate phagocytosis through F-actin rearrangements. In D. discoideum, the G4αGβγ complex has been proposed to be activated by the fAR1 folate receptor, was shown to be implicated in particle uptake, and is therefore the most likely candidate involved in signaling from the phagocytic receptors to drive actin polymerization (87, 89). F-actin rearrangements occur at the uptake site to drive formation of pseudopods and the phagocytic cup. F-actin polymerization at the uptake site is driven by the Arp2/3 complex, an actin nucleation factor, and its activator, the SCAR/WAVE complex, in both macrophages and D. discoideum (90, 91). In addition, several other actin-binding proteins, such as profilins, cofilins, and Abp1, are present during the formation of the phagocytic cup [(92, 93), for a more detailed review on the phagocytic process in D. discoideum, see Ref. (94)]. Regulation of actin dynamics and subsequent trafficking events by Rho GTPases is also conserved in D. discoideum, with Rac1 homologs (RacA, B, C, and G) thought to be the main regulators of phagocytic uptake. Notably, Rac1 is involved in FcγR-mediated phagocytosis in macrophages [for a comprehensive review on Rho signaling, see Ref. (95)].

Phosphatidylinositol phosphates are important players during phagocytic uptake and maturation because they provide an identity to each compartment. PIP dynamics are well conserved between macrophages and D. discoideum. Briefly, phosphatidylinositol (4,5)-bisphosphate PI(4,5)P₂ is the predominant PIP of the plasma membrane and is involved in recruiting and activating actin-binding proteins and nucleation-promoting factors. After receptor engagement, PI(4,5)P₂ is phosphorylated into PI(3,4,5,)P₃ by phosphatidylinositol 3-kinase and hydrolyzed into diacylglycerol and inositol (1,4,5)-trisphosphate, second messengers involved in calcium release and activation of further signaling cascades, by the phospholipase C kinase. Decrease of PI(4,5)P₂ around the uptake site is necessary to then allow actin disassembly and closure of the phagocytic cup. In addition, closure of the phagocytic cup requires Dd5P4, the D. discoideum homolog of the phosphatase OCRL, which dephosphorylates PI(3,4,5)P₃ into PI(3,4)P₂ (96–98). Extensive recycling of plasma membrane components, including adhesion molecules, is a common feature shared by D. discoideum and macrophages in the early phases of phagosome formation, and in both organisms this step is regulated by the WASH complex, an Arp2/3 activator (99–101).

After closure of the phagocytic cup, the ingested particle is found in a closed compartment termed the phagosome. Extensive proteomic analyses as well as more recent live microscopy experiments have demonstrated the extraordinary plasticity of this organelle and the high degree of conservation of the phagosome maturation pathway between mammals and D. discoideum (8, 84). Phagosome maturation is a well-orchestrated series of events, which ensures killing and digestion of ingested bacteria (Figure 3). In a simplified view, Rab GTPases, notably Rab5 and Rab7, act as the masterminds of phagosome maturation by sequentially recruiting effectors involved in the various maturation steps [for an extensive review, see Ref. (102)]. In macrophages, Rab5 and its effectors are responsible for docking and fusion of endocytic compartments with the nascent phagosome and for acquisition of early phagosomal markers (103, 104). Subsequently, Rab7 ensures fusion with late-endosomal/lysosomal compartments and thus delivery of the lysosomal digestive content into the phagosome (104, 105). Rab GTPases are highly conserved at the protein sequence level between mammals and D. discoideum; indeed, this amoeba has homologs of most mammalian Rab GTPases that have been reported to be associated with phagosomes (102). D. discoideum Rab7 is recruited as early as 1–3 min after phagosome closure and regulates delivery of lysosomal proteins (106, 107). The localization and function of D. discoideum Rab5 have not been reported.

Recruited within minutes to the phagosome, the H⁺-vacuolar ATPase (V-ATPase) is the main agent of acidification and pumps protons inside the phagosome thanks to ATP hydrolysis (108). By creating a proton gradient, the V-ATPase is a crucial complex not only for the killing and digestion of bacteria but also for the progression of the maturation program. Notably, the proton gradient is necessary for proper delivery and activity of lysosomal enzymes, as well as for the function of ion transporters, involved in poisoning by or deprivation of metals (detailed below). The V-ATPase has been shown to be delivered by fusion with lysosomal vesicles or tubules in D. discoideum and macrophages, respectively (108–110). Rapid acidification of the compartment ensues, with the lowest pH reached between 10 and 30 min after phagosome formation, depending on the measuring method (93, 111). In contrast to macrophages, whose phagosomes were reported to reach a pH of 4.5–5, D. discoideum phagosomes are more acidic, with a pH as low as 3.5–4 (111–113).

Although unprocessed antigens have been shown to be regurgitated from late-endosomal compartments in dendritic cells to allow antigen uptake by other MPS cells (114, 115), in general, the acidic phagosome of MPS cells is thought of as a dead-end for ingested bacteria. In contrast, the acidic D. discoideum phagosome matures into a postlysosome. The V-ATPase and lysosomal enzymes were shown to be retrieved in several subsequent waves of recycling mediated by the WASH complex, which drives local actin polymerization (116, 117). The WASH complex is a nucleation-promoting factor necessary for the activation of the Arp2/3 complex and actin polymerization in both mammals and D. discoideum (99, 100, 117). During retrieval of the V-ATPase, the phagosome reaches its neutral pH, and progressively acquires the postlysosomal markers vacuolin A and B, homologs of the mammalian lipid raft-associated flotillins (118–120). The postlysosome is also characterized by the presence of an actin coat and the actin-binding protein coronin (118). This compartment then fuses with the plasma membrane in a mechanism akin to exocytosis to expel its non-digested materials. Interestingly, this process is reminiscent of exocytosis of secretory lysosomes in mammalian cytotoxic cells of the immune system (121). In mammals, lysosomes were shown to fuse with the plasma membrane upon increase of cytosolic Ca²⁺ concentration. In D. discoideum, Ca²⁺ was also shown to be involved in exocytosis, with mucolipin, a Ca²⁺ transporter, involved in regulating this process. Mucolipin is thought to pump Ca²⁺ inside postlysosomes, thereby inhibiting fusion with the plasma membrane by decreasing local Ca²⁺ concentration (122).

Certain intracellular bacterial pathogens are known for subverting phagosome maturation to prevent the formation of an unfriendly bactericidal compartment and to enable replication within the host cell. L. pneumophila, a Gram-negative bacterium that causes Legionnaire’s disease, is able to arrest early phagosomal maturation. In fact, the V-ATPase and other early endocytic markers are not delivered to the Legionella-containing vacuole (LCV) in either D. discoideum or macrophages (123). The endoplasmic reticulum (ER) is recruited in proximity to and fuses with the LCV, which becomes enriched in ER markers such as calnexin and calreticulin. This enrichment of ER markers is a consequence of a major strategy used by L. pneumophila to proliferate inside the host cell, which is termed identity theft and consists of changing the identity of its compartment to resemble the ER by recruiting different GTPases and PIPs (124, 125). For example, thanks to several bacterial effectors secreted through its type 4 secretion system, this pathogen is able to change the PIP composition of the phagosome by notably acquiring PI(4)P, a PIP normally associated with the trans-Golgi, ER, and plasma membrane (47). Other proteins involved in PIP dynamics and phagosome maturation, such as phosphatidylinositol 3-kinase and Dd5P4 have been shown to restrict L. pneumophila growth in D. discoideum (45, 46). D. discoideum has been used successfully to isolate and characterize the proteome of LCVs (44, 126). These studies have highlighted the importance of ER/Golgi small GTPases, such as Arf1 and Rab1, in gradually modifying the identity of the LCV. Rab8, a trans-Golgi-associated Rab GTPase, has also been detected at the LCV membrane and been shown to play a role in the association of SidC, a bacterial effector that mediates ER recruitment (44). Interestingly, Hoffmann and colleagues compared the LCV proteomes purified from murine macrophages and D. discoideum and uncovered that, if only considering proteins with conserved roles in these two organisms, about 50% of LCV-associated proteins were found in both organisms (126). These include the aforementioned Arf1, Rab8, and Rab1, as well as proteins involved in lipid metabolism, suggesting that L. pneumophila uses similar mechanisms to replicate in D. discoideum and macrophages and further corroborating the case that D. discoideum is an excellent model to study L. pneumophila infection.

Like Legionella pneumophila, albeit with completely different strategies, Mycobacterium spp. manipulate the phagosome maturation pathway. In fact, Mycobacterium tuberculosis, the causative agent of tuberculosis, and Mycobacterium marinum, a closely related mycobacterium that infects frogs and fish, prevent acquisition of the V-ATPase in macrophages (127). This was confirmed in D. discoideum, where it was shown that M. marinum is able to prevent accumulation of the V-ATPase and of cathepsin D (48). Recently, it was proposed that the WASH complex plays a role in preventing association of the V-ATPase with the mycobacteria-containing vacuole (MCV) by inducing polymerization of actin around the MCV, which probably prevents fusion with acidic vesicles. This mechanism was first studied in D. discoideum, but further confirmed in human macrophages with both M. tuberculosis and M. marinum (128). Of note, the mechanism of escape of cytosolic M. marinum from the host cell, termed ejection, and of cell-to-cell spreading has been well characterized and studied using D. discoideum as a host model [(49); reviewed in Ref. (129, 130)]. Interestingly, D. discoideum has also been extensively used as a model phagocyte to screen for new mycobacterial virulence factors (131–134).

Lysozymes, glycosidases that digest the peptidoglycan layer present in the cell wall of bacteria, belong to the Aly family in D. discoideum (135). Upregulation of lysozyme expression differs depending on the bacteria used as food. D. discoideum grown on Gram-positive bacteria upregulate AlyA, AlyB, AlyC, and AlyD whereas growth on Gram-negative bacteria leads to an increase in AlyL expression (136).

Lysosomal enzymes comprise several classes of enzymes involved in hydrolysis of sugar groups, such as mannosidases, or peptide bonds, such as proteases, and have been involved extensively in resistance to certain pathogens as well as bacterial killing in MPS cells. For instance, cysteine and serine cathepsins have been implicated in resistance to and killing of Staphylococcus aureus in neutrophils and macrophages (137, 138). More recently, M. tuberculosis was shown to downregulate expression and inhibit activity of cysteine proteases in macrophages, thus ensuring its replication (139). Cathepsin D, an aspartic protease, was also involved in resistance to Listeria monocytogenes, an intracellular food-borne pathogen (140).

Two main classes of lysosomal enzymes, bearing different posttranslational modifications, have been characterized in D. discoideum. The first class includes α-mannosidase, β-glucosidase, and cathepsins and is modified with mannose-6-phosphomethyldiester and/or mannose-6-sulfate, also known as common antigen-1 (141, 142). The second class of enzymes contains an N-acetylglucosamine-1-phosphate and comprises cysteine proteases (143). These different classes of enzymes reside in different vesicles at the steady state level and are recruited to phagosomes in a sequential manner, with first a wave of cysteine proteases followed by enzymes bearing the mannose-6-sulfate modification (84, 144).

Although lysosomal enzymes appear to be involved in bacterial killing in D. discoideum as in macrophages, direct evidence is lacking. Deletion of the cathepsin D gene is not sufficient to abolish growth on the food bacterium Klebsiella (142). Several D. discoideum mutants impaired in lysosomal enzyme trafficking and/or activity have been characterized including strains lacking WshA, a subunit of the WASH complex involved in lysosomal enzyme recycling (117), LvsB, a protein involved in restricting heterotypic fusion of early endosomes with postlysosomal compartments (145), and TM9 protein A, which is involved in the sorting of glycosidases, cathepsins, and lysozymes (146). Interestingly, these mutants exhibit growth defects specific to certain subsets of bacterial species (117, 145–148). These data suggest that different classes of lysosomal enzymes might play redundant roles, that they are not the sole killing strategy employed by D. discoideum, and that specific mechanisms may be used depending on the encountered bacteria.

Reactive oxygen species are key components of cell-autonomous defenses of MPS cells and function as antimicrobial effectors (149) as well as signaling molecules that regulate NF-κB (150, 151), autophagy (152), cytokine secretion (153), inflammasome activation (154), and apoptosis (155). ROS are implicated in the regulation of pH within phagosomes and the production of antigenic peptides in dendritic cells (156–158). ROS have also been implicated in the regulation of cytoskeleton dynamics and chemotaxis (159, 160). The major source of ROS in MPS cells is the NADPH oxidase (NOX) 2. Depending on its localization, NOX2 generates superoxide by transferring an electron from cytosolic NADPH to O₂ in either the extracellular space or the lumen of the phagosome. Subsequent reactions convert superoxide into additional ROS. Superoxide dismutase catalyzes its conversion to hydrogen peroxide, which in turn reacts with Fe²⁺ in the Fenton reaction to generate hydroxyl radicals or with Cl⁻ to produce hypochlorous acid via myeloperoxidase (161–163).

NOX2 is a heterodimer comprising the transmembrane proteins gp91^phox/Nox2, the catalytic subunit, and p22^phox, the regulatory subunit. NOX2 activation occurs downstream of extracellular receptors including integrins and Fc receptors and is coupled with phagocytosis. Activation requires three additional subunits, p67^phox/neutrophil cytosol factor (Ncf) 2, p40^phox/Ncf4, and p47^phox/Ncf1, which form a ternary complex in the cytosol that is recruited to membrane-localized NOX2 by the small GTPases Rac1 and 2 (161–163). Mutations in NOX2 subunits cause chronic granulomatous disease (CGD), a condition that makes patients susceptible to recurring bacterial and fungal infections and demonstrates the importance of the NOX2-generated oxidative burst in the immune response (164, 165). Mitochondrial ROS production activated downstream of TLR signaling also contributes to antimicrobial mechanisms (166, 167).

Because ROS can damage host and microbe alike, ROS production and localization are tightly regulated, and MPS cells express ROS detoxifying enzymes such as superoxide dismutases (SODs), catalases, and peroxiredoxins to prevent self-damage. Microbes that persist inside the phagosome have mechanisms to minimize oxidative stress. These include expression of robust systems to maintain internal redox homeostasis (168), secretion of SODs and catalases to detoxify their compartment (149, 169), and deployment of effectors that inhibit NOX2 activation and/or delivery to the phagosome (170–173).

The D. discoideum genome encodes three catalytic NOX subunits: NoxA and NoxB, which are homologs of gp91^phox/Nox2, and NoxC, which is a homolog of Nox5 (174–176). It also encodes one homolog of p22^phox, CybA, and NcfA, a homolog of the cytosolic activating factor p67^phox/Ncf2 (174, 175). RNAseq data indicate that NoxA, CybA, and NcfA are expressed during growth while NoxB and NoxC are mainly expressed during development (69, 174). However, long-term growth on Klebsiella can cause upregulation of NoxB in growing cells (136). Interestingly, D. discoideum expresses multiple SOD and catalase homologs (177–179) and exhibits a high resistance to oxidative stress (180), which suggests that it encounters internally and/or externally generated ROS regularly.

Whether ROS contribute to cell-autonomous defenses during the growth phase of D. discoideum is not clear. Mutants lacking noxA or both noxA and noxB exhibit no growth or killing defects when grown on Klebsiella (148, 174). This lack of a defect might be a consequence of redundant killing mechanisms or of challenge with a bacterium that is easily killed. Intriguingly, a D. discoideum mutant lacking the Xpf nuclease, a component of DNA damage repair machinery, accumulates more mutations when grown on a range of bacteria including Klebsiella than when grown axenically in media (181). One possible explanation is that Xpf is required to repair DNA damaged by ROS generated in response to bacteria.

Excessive ROS production in growing cells due to a deletion of a surface-localized SOD causes defects in chemotaxis and cell motility via sustained Ras activation (182, 183). The authors hypothesize that the chemotaxis defect prevents the inclusion of the cell in the multicellular stage and thus prevents propagation of the potentially mutagenized genome. It is tempting to speculate that excessive ROS production in response to a resistant microbe causes the same effect and inhibits the inclusion of an infected cell in the multicellular stages. Although the context is different, this “self-sacrifice” might be analogous to ROS-induced apoptosis of infected macrophages (155, 184).

Reactive oxygen species have functions during development. Extracellular ROS scavengers inhibit aggregation (185), and mutations in the individual NOX subunit genes, noxA, noxB, noxC, or cybA, or in the development stage-specific catalase gene, catB, cause defects in fruiting body formation when developed under axenic conditions (174, 186). These results indicate a signaling role for ROS. When developed after feeding on bacteria, a noxABC⁻ triple mutant exhibits increased bacterial contamination of fruiting bodies compared with wild type (187). Thus, ROS also have an immunity function.

One possible cause of the immunity defect in the noxABC⁻ mutant is the abrogation of DNA extracellular trap (ET) formation (Figure 1). The slug stage can persist for multiple days, during which it migrates through a dangerous melange of infectious bacteria and fungi that could decrease and/or prevent spore production. Phagocytic flux is limited in non-feeding developed cells (188). Protection from infection and intoxication appears to be mediated in part by a subpopulation of cells (<1%) that retain the capacity for phagocytosis (77). These so-called sentinel cells (S cells) are motile within the slug and phagocytose bacteria and toxins until they are eventually shed. Compared with the other cell types in slugs, S cells are enriched for tirA mRNA, and S cells from tirA⁻ mutants have a decreased capacity to kill bacteria (77). In response to bacteria or LPS, S cells secrete mitochondrially derived ETs via a mechanism that requires both TirA and NOX-generated ROS, and increased contamination of fruiting bodies correlates with decreased ET production (187). Binding of the TLR2 TIR domain with Nox2 during M. tuberculosis infection of macrophages has been reported (189). How NOX and TirA fit into the pathway and whether they interact awaits further examination, as do the questions of whether S cells kill intracellular bacteria and, if so, whether TirA and NOX are involved. ET formation during the growth phase of the D. discoideum life cycle has not been observed.

First discovered in neutrophils and named neutrophil extracellular traps (NETs), ETs have been observed in numerous immune cell types including macrophages (190–193). ETs comprise antimicrobial peptides, proteases, and signaling molecules bound to a meshwork of DNA released from the nucleus or mitochondria [(190); reviewed in Ref. (194)]. The mechanisms by which ETs kill extracellular bacteria and/or prevent their dissemination are not well understood. Neutrophils from CGD patients fail to generate NETs (195), although NOX-independent mechanisms have also been described (196–199), some of which utilize ROS from mitochondria (200). ET production by S cells further illustrates the conservation of cell-autonomous defense mechanisms in D. discoideum, a precursor to the specialized phagocytes of the vertebrate immune system (201).

Maintaining the concentration of divalent trace metals such as iron, manganese, zinc, and copper is essential for every living organism to preserve metabolism and cell growth. Metalloproteins with trace metals as cofactors play a role in many important cellular functions such as signaling, respiration, transcription, translation, and cell division. Tight regulation of divalent metals is necessary: low levels of iron and manganese have detrimental metabolic effects, whereas zinc and copper are toxic at high concentrations.

For intracellular bacterial pathogens, trace metals are an important micronutrient resource with a role in many metabolic processes and are, as a consequence, essential for intracellular growth. Host cells such as macrophages have developed strategies to control growth of intracellular bacteria by sequestering metals such as iron and manganese (i.e., metal deprivation or nutritional immunity) or by pumping toxic metals inside the pathogen-containing compartment [i.e., metal poisoning; reviewed in Ref. (202–204)]. Bacteria have established ways to counteract metal deprivation or poisoning by expressing siderophores (i.e., small molecules with high affinity for the relevant metal) and uptake systems or by upregulating efflux systems such as P-type ATPases (203).

Consequently, the phagosomal concentration of essential trace metals varies during phagocytosis and infection with bacterial pathogens. Wagner et al. elegantly measured the metal concentration in the phagosome upon macrophage activation with inflammatory cytokines and upon infection with M. tuberculosis or Mycobacterium avium [(205); reviewed in Ref. (206)]. At 1-h post infection (hpi), the early phagosome was shown to be enriched in sulfur and chloride and depleted of calcium and potassium. At a later stage (24 hpi), the MCV harbored zinc, iron, and calcium at high concentrations. In addition, activation of infected macrophages with cytokines leads to a large increase in zinc and copper and a depletion of iron and chloride (205). Importantly, the metal concentration in the phagosome is very likely coupled to the proton gradient. Acidification of the phagosome is achieved by the combined actions of the V-ATPase and the Hv1 H⁺-channel (207). To counter-balance the electrogenic H⁺-gradient across the phagosomal membrane, Cl⁻ is imported into the phagosome by transporters of the CFTR (208) and CLC family [(209); reviewed in Ref. (210)], respectively [reviewed in Ref. (206)]. Lysosomal acidification is also facilitated by the efflux of cations (211).

The galectins compose a family of 15 mammalian lectins with affinity for β-galactoside sugars that share a characteristic carbohydrate recognition domain (285). They have been classified into three groups according to their overall structure: the “prototype,” the “tandem-repeat,” and the “chimera-type.” Galectins are present in the cytosol (286) and nucleus (287) of cells but are also secreted extracellularly via an unconventional secretion mechanism that has remained elusive for decades (288). Galectins play a role in multiple biological processes, including angiogenesis (289), tumor growth (289), and inflammation (290), due to their ubiquitous localization as well as the high diversity of self and non-self glycoconjugates that they recognize [for a review, see Ref. (291)]. Importantly, extracellular lectins bind to the complex glycocalyx coat at the cell surface, which enables the formation of specific microdomains known as the galectin lattice (292). The galectin lattice restricts the mobility of glycoconjugates in the plasma membrane and has important functions in signaling and endocytosis (293). Some galectins recognize the surface of various pathogens (294–297) and have been proposed to have a direct antibacterial effect (298, 299). These interactions involve galectin binding to specific bacterial species or clades according to the carbohydrates displayed on their cell wall or capsules. In addition to their roles in the extracellular milieu, galectins have recently emerged as general innate-immune factors against a wide range of intracellular infections. This newly described function appears not to involve the binding of cytosolic galectins directly to the surface of intracellular pathogens but rather to result from the recognition of self glycans present on the luminal leaflet of pathogen-containing vacuoles. These glycans become accessible to the cytosolic lectins when membrane damage occurs, as shown during infection with S. enterica, L. pneumophila, and Yersinia pseudotuberculosis (300, 301). This process leads to the recruitment of the autophagy machinery (300) or of GBPs (301) to the compartment.

Although D. discoideum lacks galectin homologs, discoidins share molecular and biological characteristics with galectins. They form a family of four β-galactoside-binding and β-N-acetylgalactosamine-binding lectins with two carbohydrate binding domains (302, 303), and recent genome analyses identified three potential discoidin-like proteins in D. discoideum (http://dictybase.org). Discoidins are expressed throughout the D. discoideum life cycle (69, 70). They are highly abundant in the cytosol and are also secreted despite their lack of a signal peptide (304). Early reports suggested a possible role in self-recognition during the multicellular cycle of D. discoideum (305), but it was later observed that mutants with much reduced expression of discoidins (1–2% of wild type) were able to form apparently normal fruiting bodies (306). In addition, subsequent studies were unable to confirm surface localization of discoidins (307, 308). Consequently, the role of discoidins in self-recognition and adhesion during D. discoideum development has remained controversial for several decades. Whether discoidins bind bacteria and/or have a role in cell-autonomous defenses similar to that of galectins is currently under investigation.

The significance of ubiquitination in the regulation of various aspects of mammalian immunity has been increasingly recognized in recent years. Ubiquitination is an omnipresent posttranslational modification in which the 76-amino acid ubiquitin is covalently linked to lysine (K) residues of substrate proteins. The stepwise enzymatic cascade of ubiquitination involves following three proteins: ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2), and ubiquitin ligase (E3). Their activity results in the attachment of one ubiquitin to the substrate protein linked by an isopeptide bond between the ubiquitin C-terminus and the NH₂ group of the substrate K residue. This is referred to as monoubiquitination. Repeated ubiquitination leads to the generation of a polyubiquitin chain, known also as polyubiquitination. Ubiquitin contains seven K residues (K6, K11, K27, K29, K33, K48, and K63). Typically, the attachment of K48-linked polyubiquitin chains to substrate proteins serves as a signal for their degradation by the proteasome. However, the other linkages and the C- or N-terminal linear linkage of ubiquitin moieties play roles in almost all aspects of plant and animal biology, such as growth, development, stress responses, and immunity. In mammals, K63-linked polyubiquitination has been associated with a broad range of immunity-related processes and particularly the activation of the NF-κB pathway, xenophagy and apoptosis (309). D. discoideum has 13 genes encoding ubiquitin, and the TRAFs and TRIMs responsible for K63-linked polyubiquitination are conserved from D. discoideum to mammals.

TNF receptor-associated factors are a family of proteins primarily involved in the regulation of inflammation, antiviral responses, and apoptosis (309, 310). Currently, seven TRAF proteins have been characterized in humans: TRAF1–7. Typically, the TRAF proteins comprise an N-terminal RING domain that mediates the interaction between an E2 ligase and the substrate, followed by a zinc-finger domain, which may play a role in DNA, RNA, protein, and/or lipid binding, and a C-terminal TRAF homology (MATH) domain. The TRAF/MATH domain has an N-terminal TRAF region that mediates homo- and hetero-oligomerization between TRAF members and a C-terminal region that is important for interactions with receptors and adaptor proteins (310).

TNF receptor-associated factor 6, perhaps the most ancient mammalian TRAF, is a RING-type E3-ligase that ubiquitinates via the K63-linkage (311–313). It is required for mTORC1 translocation to the lysosome, and TRAF6-catalyzed K63 polyubiquitination modulates mTORC1 amino acid sensing capacity (314). Moreover, in macrophages, TRAF6 is responsible for the decoration of pathogens and pathogen-containing vacuoles with polyubiquitin chains, which serve as a cue for GBP recruitment and as a recognition signal for the autophagy cargo receptor p62 (313, 315).

BLAST analyses predict that TRAF-like proteins are also present in several social amoeba species. More than 40 TRAF-like proteins in D. discoideum are predicted, of which 16 contain the RING, zinc-finger and TRAF domains present in mammalian TRAF2, TRAF3, TRAF5, and TRAF6 (Figure 9; Table 2). Despite the fact that human and D. discoideum TRAF proteins show significant similarities with respect to their major domains, their evolutionary divergence precludes ortholog assignment. Determining whether D. discoideum TRAFs regulate nutrient sensing and/or ubiquitinate pathogens will expand our understanding of their roles in infection.

The TRIM superfamily is remarkably conserved among metazoans and, perhaps as a result of an expansion during vertebrate evolution, is represented by more than 80 members in humans. TRIMs typically comprise an N-terminal RING domain, a B-box domain containing several zinc-binding motifs, a coiled-coil domain, and a considerably diverse C-terminal domain important for substrate binding (316).

Tripartite motif-containing proteins are important for many aspects of immunity resistance to pathogens. Recent studies in mouse macrophages demonstrate that various TRIM proteins are induced upon infection with influenza virus or activation of TLRs in a type-I-interferon (IFN)-dependent manner (317, 318). TRIMs are involved in restriction of HIV replication and activation of NF-κB downstream of TLRs (318). Moreover, they play a dual role as receptors and regulators of autophagy. As regulators, TRIMs serve as platforms for the assembly of the core autophagy initiators ULK1 (Atg1 in yeast and D. discoideum) and Beclin1 (Atg6 in yeast and D. discoideum) (319). In macrophages, autophagy cargo receptors recognize and bind K63-linked polyubiquitin chains and galectins, which serve as “eat-me” signals and mediate the binding of the cargo to phagophore-conjugated LC3. As receptors, TRIMs are able to recognize endogenous and exogenous (e.g., bacteria) cargo intended for autoloysosomes via binding of their diverse C-terminal domains to the cargo in a ubiquitin-independent manner and mediate delivery to the phagophore by also binding LC3 (320, 321).

Tripartite motif-containing protein homologs are found in multiple social amoeba species, and a single TRIM has been identified in D. discoideum, DdTRIM, which is an ortholog of human TRIM37 (Figure 10; Table 2) (http://dictybase.org). According to an accumulating amount of evidence, human TRIMs are emerging as critical regulators of cell-autonomous defenses. Particularly, human TRIM37 has been shown to restrict HIV-1 replication (322). At present, the role of DdTRIM remains unknown, and the presence of a single member of the TRIM superfamily early in evolution makes D. discoideum an interesting model to explore the primordial role of TRIM proteins before their expansion.

The GBP proteins are IFN-gamma-inducible, immunity-related GTPases. Generally, the GBP proteins comprise a globular N-terminal domain and a C-terminal alpha-helical baculovirus inhibitor of apoptosis repeat (BIR) domain. Both the globular N-terminal domain, which confers GTPase activity, and the C-terminal BIR domain mediate protein–protein and protein–lipid interactions and contribute to nucleotide-dependent homotypic and heterotypic GBP protein assembly (323). In addition, BIR domains have been reported to act as caspase regulators and mediators of homotypic interactions (324). GTP binding to the GTPase activity domain allows dimer formation, and its hydrolysis enables conformational changes resulting in GBP tetramer formation. In vertebrates, the GBPs proteins have been linked to a multitude of innate immunity-related responses such as inflammasome activation and xenophagy (325). Essential for their function is their ability to oligomerize and to bind target endomembranes (323, 326, 327). In mouse macrophages, GBP2 recruitment to Chlamydia trachomatis- and Toxoplasma gondii-containing vacuoles correlates with their host-mediated lysis, underlying the importance of these large GTPases in the successful immune response against intracellular invaders (313, 328).

Multiple social amoeba species have homologs of GBPs, and a single GBP homolog has been identified in D. discoideum, DdGBP (Figure 11; Table 2). Its role remains to be elucidated. As a single GBP representative, it may allow a better understanding of the primordial and conserved role that GBPs play in MPS cells.

In humans, there are seven STATs, which have a unique N-terminus important for nuclear translocation and protein–protein interactions. This region is followed by a coiled-coil domain involved in nuclear export and regulation of tyrosine phosphorylation, a DNA-binding domain that mediates the recognition of sequences related to TTCN3–4GAA in the promoters of responsive genes, and an Src-homology (SH2) domain, which allows for specific recognition and docking to phosphotyrosines on cytokine receptors, Janus kinases (JAKs), and other STAT molecules. The C-terminus contains a divergent transactivator domain, which mediates STAT transcription factor transactivation via various cofactors (329).

Signal transducers and activators of transcription proteins are activated mainly by cytokines and growth factors. Binding of these signaling molecules to their receptors triggers receptor dimerization, allowing transphosphorylation and activation of receptor-associated JAKs. The JAKs also phosphorylate the cytoplasmic tails of the receptors, which promotes recruitment of the STAT proteins through their SH2 domain. The subsequent tyrosine phosphorylation of STATs results in the formation of homodimers and/or heterodimers and nuclear translocation, whereupon they bind to the promoter regions and initiate transcription of various immunity-related genes such as galectins, GBPs, NOXs, SQSTM1, TRIMs, and NDP52 (330). As part of their functional cycle, STATs shuttle between the cytosol and the nucleus (329). Mammalian STAT proteins act in a tissue-specific manner to regulate growth and development, various immunity-related processes, and cellular stress responses (331, 332). Gain- or loss-of-function mutations in components of the JAK–STAT signaling pathway have been associated with a broad range of innate-immune deficiencies and autoimmune diseases (333).

Multiple social amoeba species have STAT homologs, and four D. discoideum homologs of STATs have been identified, DstA, DstB, DstC, and DstD [reviewed in Ref. (334)] (Figure 12; Table 2). These proteins have a predicted SH2 domain (335), and their activity is regulated in part by tyrosine kinases (336, 337) and phosphatases (338, 339). Like mammalian STATs, D. discoideum STATs regulate growth and development in response to extracellular signaling molecules (335, 340–342) and are activated by cellular stresses (339, 343). Whether they also contribute to antimicrobial responses in D. discoideum is unknown, and addressing this question will provide insight into their possible roles in cell-autonomous defenses.