Correspondence Anetta Hartlova, Centre of Advanced Studies, Faculty of Military Health Sciences, University of Defence, Trebesska1575, 500 01 Hradec Kralove, Czech Republic. Tel.: +420 973 251 540; fax: +420 435 513 018; email: firstname.lastname@example.org
Pathogenic bacteria have developed various mechanisms to evade host immune defense systems. Invasion of pathogenic bacteria requires interaction of the pathogen with host receptors, followed by activation of signal transduction pathways and rearrangement of the cytoskeleton to facilitate bacterial entry. Numerous bacteria exploit specialized plasma membrane microdomains, commonly called membrane rafts, which are rich in cholesterol, sphingolipids and a special set of signaling molecules which allow entry to host cells and establishment of a protected niche within the host. This review focuses on the current understanding of the raft hypothesis and the means by which pathogenic bacteria subvert membrane microdomains to promote infection.
Despite the use of antibiotics and the employment of vaccination programs, infectious diseases remain a major threat to global health and security. Furthermore, the increasing ease of global travel has led to the spread of infectious diseases throughout the world. These facts indicate the importance of efforts to understand the molecular mechanisms of infection; as such efforts could lead to the identification of new drug targets and new affordable therapies.
The outcome of a host-bacterial interaction depends on the balance between the ability of the host to detect and eliminate a potentially harmful agent, and the ability of the pathogen to utilize host cell signaling. Several pathogens have in common a strategy of internalization within their host cells which enables them to avoid encountering the extracellular mechanisms of host defense. Upon microbial contact, numerous host-cell signaling pathways are simultaneously activated in order to regulate the host immune response. Internalized bacteria are delivered via endocytic vesicles to lysosomes where a combination of reactive oxygen radicals, acidic pH and hydrolytic enzymes contribute to their degradation. Pathogenic bacteria have evolved diverse strategies to co-opt host-cell signaling to promote their internalization and, therefore, their establishment of a protective niche. Most notably, bacteria modulate actin-cytoskeleton assembly and phagosome biogenesis to prevent fusion of endosomes with lysosomes (1). Bacteria can be internalized by either actin-based mechanisms, such as phagocytosis and macropinocytosis (or endocytosis). Based on the literature, endocytic pathways are classified according to the molecular mechanisms which regulate this process: clathrin-dependent endocytosis (known as receptor-mediated endocytosis), caveolae-dependent endocytosis and clathrin- and caveolae-independent pathways. Clathrin and caveolae-independent endocytosis are driven by specialized plasma membrane domains, referred to as lipid rafts. Lipid rafts are composed of specific lipids, sphingolipids, and cholesterol (2). It has been shown that several bacteria exploit lipid rafts, as well as caveolae regions, to invade host cells. This review focuses on the current understanding of the raft hypothesis and the means by which pathogenic bacteria subvert these specialized membrane microdomains to promote infection.
A NEW INSIGHT INTO THE PLASMA MEMBRANE ORGANIZATION – LIPID RAFT HYPOTHESIS
The plasma membrane of a eukaryotic cell forms a selective barrier between the outside world and the internal cellular environment. It regulates most of the information received and sent out by a cell. Therefore, its structure and properties pose key features for correctly processing information.
Over the past decades, the picture of the cell membrane has changed; it has evolved from snapshot to video, as reviewed in Edidin et al. (3). According to our current understanding, lipids are not just passive components of eukaryotic cell membranes, but actively contribute to the control of protein organization and dynamics of the membrane. It has been proposed that specific lipids, such as glycosphingolipids, and cholesterol segregate into dynamic membrane microdomains known as lipid rafts. This kind of membrane organization is based on the physical properties of particular lipids. Sphingolipids contain saturated acyl chains which allow them to pack tightly, forming a gel-like phase. In contrast, phospholipids contain unsaturated acyl chains and form a liquid-disordered phase. The presence of cholesterol within the membrane changes a gel-like phase into a laterally ordered phase with properties intermediate between those of the gel and solid phases (4). A lipid-based sorting mechanism is involved in cellular processes such as signal transduction, membrane trafficking and membrane sorting in polarized cells (5). Of particular interest is the accumulation of a specific set of proteins, including GPI-anchored proteins, and different signaling molecules in lipid rafts in response to the changing cell environment (6, 7). According to a recent definition, the term “lipid rafts” has been replaced by the term “membrane rafts” (8). This new definition takes into account the contribution of proteins to membrane raft organization. Membrane rafts are defined as small (10–200 nm), heterogeneous, highly dynamic, sterol- and sphingolipid-enriched domains which compartmentalize cellular processes. Upon certain stimuli, small rafts coalesce into larger platforms which are stabilized through protein-protein and protein-lipid interactions (8).
The aforementioned small size of membrane rafts limits their ability to be resolved by conventional light microscopy. The only morphologically identifiable raft-like domain belongs to one particular subset of rafts known as caveolae (9). Caveolae are flask-shaped invaginations of the plasma membrane and are composed of the cholesterol binding proteins caveolins. Caveolae can be defined as caveolin-containing plasma membrane microdomains. A cellular function for caveolae in membrane transport and signal transduction has been established (9, 10). For the sake of simplification, in this article the term “membrane rafts” will be used to denote both caveolae and membrane rafts, except where it is necessary to distinguish them from each other.
STRATEGIES FOR STUDYING MEMBRANE RAFTS: BIOCHEMICAL, BIOPHYSICAL APPROACHES AND MANIPULATION OF PLASMA MEMBRANE CHOLESTEROL
The predominant evidence for the functional organization of membranes has been derived from studies of membrane fragments called DRM which are insoluble in cold non-ionic detergents at low temperatures (11). This process results in a DRM fraction which is rich in cholesterol and lipids with saturated acyl chains, including sphingolipids. These lipids can form a liquid ordered phase which is distinct from the bulk liquid-disordered phase, the latter containing phospholipids with unsaturated acyl chains. Phase separation may elucidate the detergent-insolubility of rafts (12). Certain proteins accumulate in the DRM fraction, including GPI-anchored proteins; doubly acylated proteins such as tyrosine kinase of the src family; G protein coupled receptors; and some transmembrane proteins, such as the characteristic DRM-marker, flotillin (13). The specific lipid and protein composition of DRM vary depending on the identity and concentration of the detergent and the type of cells used in the assay. There are no clearly defined conditions for isolation of DRM at the present time (14). Hence, DRM isolated from cells do not precisely reflect the exact physiological composition of membrane rafts (15).
A recent study by Heerkholtz reports that the DRM method is ineffective in proving membrane raft existence and demonstrates that the detergent Triton X-100 may itself induce lipid clustering (16). To avoid this obstacle, a new method has been established in which membrane domains are isolated in the absence of detergent. It has been shown that rafts isolated from different cell types by this detergent-free method exhibit all of the typical characteristics of DRM (17).
Furthermore, a variety of sophisticated techniques such as fluorescence resonance energy transfer (18), fluorescence recovery after photobleaching (19), fluorescence correlation spectroscopy (20), single molecule microscopy (21), atomic force microscopy (22) and Laurdan 2-photon microscopy (23) have appeared. These methodologies give us information about lateral membrane organization, spatial distribution and movements of cell membrane components with a nanometer spatial resolution (Table 1). Recent studies have proposed that membrane rafts assembled by the self-aggregation of lipids must be small, unstable and most likely to reach only a few to tens of nanometers (18, 24). In agreement with Kusumi's high-resolution single particle tracking, data from Laurdan microscopy highlight the importance of the actin cytoskeleton as a regulator of the formation of a large clustered membrane platform. It is thought that non-random movement of raft-associated molecules in plasma membranes is derived from the attachment of proteins to the actin cytoskeleton (23, 25). Although data regarding the size, shape and lifetime of membrane rafts in vivo are not uniform, researchers agree on the functional significance of membrane domains in membrane sorting and signaling.
Table 1. Methods characterizing the size, structure, diversity, and dynamics of membrane rafts
Characterization of membrane raft structure
FRET (Fluorescence Resonance Energy Transfer)
Characterization of spatial distribution of clusters of raft proteins
Another widely used method to study membrane raft regions involves alteration of the cholesterol content of the plasma membrane. Cholesterol is considered to be the key structural and regulatory element of membrane raft integrity (26). Depletion of membrane cholesterol using cholesterol-depleting drugs, such as MBCD or filipin, results in destabilization of the membrane raft composition and affects DRM-associated proteins, which are involved in diverse cellular processes (27). However, there has been considerable debate on this approach as such treatment may affect the general physical properties (such as permeability and fluidity) of the cellular membrane, though not the formation of membrane domains (28). Unraveling the true functional role of cholesterol could lead to a better understanding of the formation of membrane rafts.
Unfortunately, there is no similar biochemical approach with minor side effects for the extraction of sphingolipids. For this reason, a combination of morphological and biochemical approaches should be performed to study physiological processes in the raft microenvironment.
INVOLVEMENT OF MEMBRANE RAFTS IN BACTERIAL INFECTION
Despite all doubts about raft theory, we must address the physiological membrane interactions of lipids, lipid-anchored, and acylated cytoplasmic signaling proteins, interactions which have been proven by morphological or functional approaches. Over the last few years, rafts have been studied in the context of categorization of events in membrane trafficking (29), signaling (30), pathogen internalization (31) and/or other diseases (32). It is proposed that small membrane rafts are organized into larger signaling platforms in response to different stimuli, such as protein-ligand binding. Raft coalescence enables interaction of surface receptors with costimulatory molecules, thus allowing raft-dependent cellular processes (24). Assembly of a variety of signaling molecules within membrane rafts, including molecules playing essential roles in cytoskeletal reorganization, seems to be a convenient portal of entry for microorganisms. It has been shown that a wide variety of pathogens, including viruses (33), bacteria, protozoans and prions, preferentially interact with membrane rafts to invade host cells (34). Therefore, pathogen invasion may serve as a useful platform for examination of the complex physiological function of membrane rafts, or for helping to elucidate the molecular mechanisms of the pathological state. Bacteria which have been investigated in association with membrane rafts to date include Escherichia coli, Chlamydia spp., Mycobacterium spp., Shigella spp., Salmonella enteretica, Pseudomonas aeruginosa, Brucella spp., Legionella pneumophila, and Coxiella burnetii (Fig. 1).
Uropathogenic E. coli expressing FimH+ was the first bacteria studied in association with membrane rafts. E. coli is the most common causative agent of urinary tract infection in humans. It has been shown that, following binding of bacterial adhesin FimH to the GPI-linked surface molecule CD48, which is presumably present in rafts, bacteria invade mast cells via host cell caveolae. Bacterial internalization through caveolae may enable the survival of bacteria inside host cells, whereas bacteria that have been opsonized by antibodies enter host cells via clathrin-coated pits and are subsequently degraded in lysosomes. Thus, raft-dependent endocytosis provides a mechanism that allows E. coli to become an intracellular pathogen (35).
In addition, it has been shown that expression of caveolin-1 is critical to bacterial internalization within host cells. Inhibition of expression of caveolin-1 by RNA interference reduces the ability of bacteria to invade host epithelial cells. The ability of E. coli to invade bladder epithelial cells seems to be crucial both in the pathogenesis of urinary tract infections and in their recurrence. Disorganization of membrane domains by cholesterol-depleting compounds prevents E. coli uptake by bladder epithelial cells. Hence, an understanding of the role of host membrane rafts in bacterial invasion may shed light on the molecular mechanism of infection (36).
Chlamydiae are obligate intracellular parasites which are believed to exploit membrane rafts in both phagocytic and non-phagocytic cells. Chlamydiae are major causative agents of sexually transmitted disease in the western hemisphere. Each species is subdivided into a trachoma biovariant (biovars) that can be further divided into serological variants (serovars). Like entry of E. coli, Chlamydiae uptake is also sensitive to raft-disruption agents. However, not all biovars and serovars of particular Chlamydiae spp. exploit rafts in order to enter host cells (37).
Differences in the route of infection depend on the composition of host surface molecules that serve as receptors for particular serovars. Additionally, the route of entry is affected by cell culture conditions and the growth phase of the bacteria. Chlamydial entry is caveolin-1-independent. All strains have been shown to be able to enter caveolin-1 negative Fischer rat thyroid cells, regardless of whether or not cholesterol-depleting drug treatment had been used (38). Unlike E. coli, strains of chlamydia are believed to actively modify their compartments in order to avoid lysosomal fusion. Chlamydiae-containing vesicles traffic to the Golgi apparatus and actively sequester sphingolipid- and cholesterol-containing vesicles. Thus, Chlamydiae-containing vesicles mimic vesicles derived from the biosynthetic compartment of host cells (39).
Another group of intracellular pathogens causing severe disease in humans are the Mycobacteria. When these bacteria are internalized by macrophages, they successfully avoid fusion with lysosomal compartments of the host cells. Moreover, Mycobacteria are able to survive inside cells due to active recruitment of the tryptophan-aspartate containing coat protein (TACO/Coronin1) into the phagosomal membrane. This protein prevents fusion of a bacterium-containing phagosome with a lysosome. It has been shown that, as well as being required for internalization of these bacteria, cholesterol is required for association of the TACO protein with the phagosomal membrane. Depletion of cholesterol by pharmacological agents inhibits bacterial internalization. In addition, disruptors prevent integration of the TACO protein into the phagosomal membrane. Furthermore, it has been shown that cholesterol accumulates at the site of bacterial entry. The cell walls of Mycobacteria are particularly rich in glycolipids, and it is apparent that some membrane molecules directly interact with host cell cholesterol (40).
Further, CR3 has been demonstrated to be one of the main receptors that these bacteria use for raft-dependent uptake. However, removal of cholesterol from macrophage cell walls does not influence CR3 function. Therefore, it has been proposed that cholesterol is not required for CR3 function, but is necessary for stable interaction of Mycobacteria with host plasma membranes (41).
Shigellae are a genus of Gram-negative bacteria that are known to be etiological agents of human bacillary dysentery. Bacterial uptake requires a functional T3SS which injects effecter proteins into the cytoplasm of host cells in order to activate crucial cellular processes, such as cytoskeleton remodeling or signaling. T3SS is constitutively expressed on the bacterial membrane independent of any interaction with host cells. Upon bacterial contact with host cells, T3SS is activated and the bacterial effectors are delivered into the host cytoplasm (42). Membrane rafts are implicated in the early stages of infection with Shigella flexneri, namely in binding and entering. The bacterial invasive protein IpaB binds to the host raft associated proteins CD44 or β1 integrin. In addition, disruption of membrane rafts markedly inhibits bacterial binding and internalization (43).
In the van der Goot study, it was shown that lipids, not proteins, are essential for binding of Shigellae to the cell surface. Protein-free artificial liposomes, which are composed of raft-like lipids, can trigger contact-mediated secretion of T3SS. Thus, it has been suggested that a special assembly of lipids affects the state of the secretion system. Nevertheless, the possible role of host proteins in the regulation of secretion processes has not been excluded. Furthermore, it has been suggested that membrane rafts may also promote escape of Shigellae from phagosomes into the host cell cytoplasm by influencing host cell processes (44).
Like Shigellae, S. enteritica utilizes a T3SS to invade host cells. Interestingly, S. enteritica employs two T3SS. It has been shown that cholesterol accumulates at the site of entry of S. enteritica (45). Once inside host cells, S. enteritica survives and replicates within cholesterol rich vacuoles that possess the features of late endosomes. It is proposed that an increase in cholesterol in Salmonella-containing vacuoles and recruitment of the GPI-linked protein CD55 relate to a role of membrane rafts in this compartment (46). Therefore, intracellular rafts, Salmonella-containing vacuoles on the membrane, may be implicated in bacterial survival within host cells. Furthermore, pathogen subversion of intracellular rafts seems to be connected with the association of PipB and PipB2, the bacterial effectors of the Salmonella pathogenicity island-2 encoded secretion system, with membrane microdomains (47).
P. aeruginosa infection is one of the most frequent and severe infections occurring in patients with cystic fibrosis and other immunocompromising conditions. Consequently, there is clinical interest in defining the molecular mechanism of P. aeruginosa infection (48). It has been demonstrated that P. aeruginosa infection of both respiratory and corneal epithelial cells triggers fusion of small membrane rafts into more stable ceramide-enriched membrane platforms at the site of bacterial entry (49). Bacterial stimulation through receptors, such as CD95 and CFTR, results in membrane reorganization by translocation of the acid ASM onto the extracellular leaflet of the cell membrane. Translocation of ASM is connected with enzyme activation and conversion of sphingomyelin into ceramide, which may result in formation of a large ceramide-enriched domain. CFTR and CD95 both cluster in ceramide-rich membrane platforms in the presence of P. aeruginosa infection (50). It has been revealed that ceramide-enriched membrane platforms are involved in P. aeruginosa uptake, induction of apoptosis in host cells, and even in regulation of cytokine release from the infected cells. The central role of ceramide-enriched membrane rafts in P. aeruginosa pathogenesis has been tested using cholesterol-depleting drugs and also by genetic deficiency of the ASM in vitro and in vivo. Both conditions lead to disruption of membrane rafts, which manifests as inhibition of both bacterial uptake and induction of apoptosis. The hypothesis that ceramide-enriched membrane rafts play a central role in P. aeruginosa pathogenesis has been confirmed by the finding that mice lacking CFTR show a great delay in induction of apoptosis, some such mice being unable to induce apoptosis at all and succumbing to systemic infection within a few days (51). Therefore, it seems that ceramide-enriched platforms play a crucial role in effective control of host defense against P. aeruginosa.
Brucella spp. are Gram-negative, facultative, intracellular pathogens that cause a highly infectious zoonotic disease known as undulant fever. As a causative agent of human infection following aerosol exposure, Brucellae have been classified as a potential agent for biological warfare (52). Brucellae are able to enter and replicate in both phagocytic and non-phagocytic cells. The ability of Brucellae to invade macrophages and adapt to their intracellular environment is essential for their bacterial pathogenesis. To evade the host immune system, Brucellae regulate their own intracellular trafficking in order to avoid maturation of the phagosome-lysosome compartment. Brucellae enter macrophages via membrane ruffling and macropinosomes. During interaction with macrophages via lipopolysaccharides (smooth O chain) and/or surface Hsp60, they recruit characteristic membrane raft molecules, such as the GM1 ganglioside and GPI proteins, at the site of bacterial entry (53). Moreover, disruption of membrane raft assembly by the cholesterol-sequestering agent MBCD has been shown to result in elimination of all bacteria (54).
Additionally, it has been documented that cyclic β-1, 2-glucan is implicated in the biogenesis of endoplasmic reticulum-derived Brucella-containing vacuoles. It is proposed that cyclic glucans sequester cholesterol from the phagosomal membrane and disrupt the rafts present on this membrane. Consequently, signaling pathways that drive phagosome maturation are modulated. Nevertheless, more detailed investigation of the mechanism of action of cyclic glucans is required (55).
Legionella spp. and Coxiella burnetii
Legionella pneumophila, a gram negative intracellular pathogen, causes a severe human pneumonia known as Legionnaire's disease. It has been demonstrated that membrane rafts participate in the uptake of Legionellae into host cells (56). Cholesterol depletion of host cells by cyclodextrin results in inhibition of Legionella invasion. Moreover, endoplasmic reticulum-derived vacuoles containing Legionellae are rich in cholesterol, GM1 gangliosides and GPI-proteins. It has been reported that the presence of these membrane raft components in Legionellae-containing vacuoles is responsible for recruitment of the autophagic pathway (57).
A similar intracellular phenomenon has been observed for Coxiella burnetii. This intracellular pathogen is also internalized into host cells through membrane rafts and resides within an autophagosome-like compartment that subsequently fuses with lysosomes (58).
Autophagy is a conserved pathway that regulates cellular turnover of long-lived proteins and organelles and also forms an important protective barrier against infection (59). However, it has been proposed that bacteria can subvert autophagy to avoid fusion with lysosomes (60). This is probably true of Legionellae, which deliver virulence factors into the host cell in order to delay maturation of their autophagosomal vacuoles (56).
The raft hypothesis was formulated more than ten years ago and gives insight into the functional organization of cell membranes, where lipids are much more than the silent partners of proteins. The membrane raft concept proposed that sphingolipids and cholesterol have different biophysical properties and form liquid-ordered phases in plasma membranes (61). Subsequently, this model was supported by studies in model membranes showing phase separation. Tightly-packed sphingolipid-cholesterol clusters were found to be insoluble to the detergent Triton X-100 at 4°C and to give rise to DRM (62). Furthermore, it was found that signaling molecules are concentrated in membrane rafts (63). Thus, two main directions of inquiry concerning rafts emerged: biophysical analyses focused on the lateral organization of the plasma membrane, and biochemical and cell biology studies which explored the functionality of membrane rafts (64).
Using simpler model membranes, biophysical in vivo studies proved the lateral heterogeneity of lipids in cell membranes, but the nature and effect of this membrane organization is still uncertain. The size and distribution of lipids in the membrane and their role in the mechanism of membrane formation remain to be elucidated. Despite great advances in biophysical techniques, development of new microscopy approaches with better spatial and temporal capabilities is required in order to image membrane dynamics in living cells (65). As in the saying “Seeing is believing”, observation of nanodomains clustering into a large-scale membrane platform would lead to an understanding of the fundamental principles of functional domains in living cells. In contrast, biology-based experiments have confirmed the implication of membrane rafts in endocytic pathways and intracellular signaling.
In recent years, new membrane transport mechanisms have been recognized. Endocytosis, exocytosis and recycling pathways have been found to be the main processes by which cells are able to communicate with their environment. It should not be forgotten that these processes are extensively regulated by cell signaling and vice versa. Cell signaling is controlled by compartmentalization of signal transduction domains in space and time. Therefore, cell membrane trafficking, together with signaling, forms one complex system (66). Both processes are essential for bacterial infection. Most bacteria exploit endogenous host pathways in order to survive inside host cells. Bacterial pathogens preferentially target host proteins involved in the regulation of phagocytosis and bacterial transport into lysosomes. This strategy makes these proteins possible targets of microbial virulence factors. In raft theory, it is proposed that small unstable membrane rafts cluster together and that the resultant raft-clusters function in such a way as to address this issue. It has been shown that many bacteria enter host cells via raft-dependent machinery. It is also proposed that bacteria exploit membrane rafts to avoid fusion with degradative compartments. Receptor-ligand engagement during bacterial internalization may influence the fate of bacteria inside host cells, possibly by altering the host immune response. For instance, it has been shown that raft components are involved in the establishment of the intracellular niche of Salmonella enteretica (46). In addition, Brucellae create special vacuole-containing proteins from the endoplasmic reticulum in order to be masked from the host immune system (55). Determination of the function of membrane rafts as signaling compartments could lead to an understanding of the host mechanism that governs vesicular trafficking. This discovery could then help to elucidate the molecular mechanisms of infection. In order to resolve these issues, further studies that combine different disciplines, including cell biology, biophysics, biochemistry and microbiology, are required.
Anetta Hartlova is supported by projects of the Ministry of Defense, Czech Republic, (MO0FVZ0000501 and OVUOFVZ200808) and of the Czech Science Foundation (GACR 310/07/0226).