Cell biology and molecular ecology of Francisella tularensis

Authors

  • Marina Santic,

    1. Department of Microbiology and Parasitology, Medical Faculty, Brace Branchetta 20, University of Rijeka, 51000 Rijeka, Croatia.
    Search for more papers by this author
  • Souhaila Al-Khodor,

    1. Department of Microbiology and Immunology, University of Louisville College of Medicine, 319 Abraham Flexner Way 55A, Louisville, KY 40202, USA.
    Search for more papers by this author
  • Yousef Abu Kwaik

    Corresponding author
    1. Department of Microbiology and Immunology, University of Louisville College of Medicine, 319 Abraham Flexner Way 55A, Louisville, KY 40202, USA.
    Search for more papers by this author

*E-mail abukwaik@louisville.edu; Tel. (+1) 502 852 4117; Fax (+1) 502-852-7531.

Summary

Francisella tularensis is a highly infectious intracellular bacterium that causes the fulminating disease tularemia, which can be transmitted between mammals by arthorpod vectors. Genomic studies have shown that the F. tularensis has been undergoing genomic decay with the most virulent strains having the lowest number of functional genes. Entry of F. tularensis into macrophages is mediated by looping phagocytosis and is associated with signalling through Syk tyrosine kinase. Within macrophages and arthropod-derived cells, the Francisella-containing phagosome matures transiently into an acidified late endosome-like phagosome with limited fusion to lysosomes followed by rapid bacterial escape into the cytosol within 30–60 min, and bacterial proliferation within the cytosol. The Francisella pathogenicity island, which potentially encodes a putative type VI secretion system, is essential for phagosome biogenesis and bacterial escape into the cytosol within macrophages and arthropod-derived cells. Initial sensing of F. tularensis in the cytosol triggers IRF-3-dependent IFN-β secretion, type I IFNR-dependent signalling, activation of the inflammasome mediated by caspase-1, and a pro-inflammatory response, which is suppressed by triggering of SHIP. The past few years have witnessed a quantum leap in our understanding of various aspects of this organism and this review will discuss these remarkable advances.

Francisella tularensis: causative agent of Tularemia

Francisella tularensis is a Gram-negative facultative intracellular bacterium that causes the fatal disease tularemia in humans and animals. Four closely related subspecies of F. tularensis have been identified: tularensis, holarctica, mediasiatica and novicida (Forsman et al., 1994). Subspecies tularensis and subspecies holarctica of F. tularensis cause most human illness. The subspecies tularensis has been divided into two clades A.I and A.II (Johansson et al., 2004), which differ in geographical distribution, transmission routes and manifestation of disease (Staples et al., 2006). Subspecies tularensis is confined to North America, whereas subspecies holarctica is found in many countries of the Northern Hemisphere, and subspecies novicida has a strong association with water and it is the oldest in evolutionary terms (Oyston et al., 2004; Santic et al., 2006). Because of the ease of its dissemination, its multiple routes of infection and its high infectivity, morbidity and mortality, F. tularensis has been classified as a category A select agent. There are concerns that F. tularensis could be used as a biological threat agent in bioterrorism attacks (Oyston et al., 2004; Santic et al., 2006).

The attenuated live vaccine strain (LVS) of F. tularensis is derived from ssp. holarctica in the 1940s in ex-Soviet Union by repeated passage in vitro, followed by repeated intraperitoneal inoculation of mice (Eigelsbach and Downs, 1961). Most studies on pathogenesis and virulence of F. tularensis have used the LVS and novicida strains as model systems (Oyston et al., 2004; Santic et al., 2006). Both strains are attenuated in humans but cause disease in animals similar to F. tularensis ssp. tularensis virulent strains. Fully virulent strains of F. tularensis must be handled at level 3 bio-containment (Oyston et al., 2004; Santic et al., 2006), while LVS and novicida strains are exempted from biosafety level 3 regulations and are considered suitable models to study pathogenesis of the disease.

Epidemiology and molecular ecology of F. tularensis

Tularemia is a zoonotic disease of the northern hemisphere. Human cases are typically sporadic, but outbreaks do occur such as in Martha's Vineyard (Matyas et al., 2007). Endemic areas existed during the last century and still exist in the ex-USSR and the Nordic countries. Cases of tularemia have also been reported from Japan (Ohara et al., 1998) and northern regions of China (Pang, 1987). In the USA there have been 200 cases per year from 1990 to 2000 (Feldman et al., 2001). Today, the highest incidences in the world occur in confined geographical areas of Finland and Sweden (Eliasson et al., 2002).

Humans acquire infection by inadvertent exposure to infected arthropod vector, or by handling, ingesting, or inhaling infectious materials. F. tularensis has been isolated from over 250 animal species, including fish, birds, amphibians, rabbits, squirrels, hares, voles, ticks and flies (Oyston et al., 2004; Santic et al., 2006). Maintenance in nature is primarily associated with rodents and lagomorphs (rabbits and hares) although amoebae are a potential reservoir (Oyston et al., 2004; Santic et al., 2006). Vector-borne transmission of tularemia to mammalian hosts has an important role in pathogenesis of the disease (Petersen et al., 2009). Deer flies, horse flies, ticks and mosquitoes are common arthropod vectors of transmission between mammals (Petersen et al., 2009).

It has been suggested that holartica ssp. has a stronger association with water-borne disease (Greco et al., 1987). In vitro studies showed that F. tularensis ssp. holartica can survive and grow within Acanthamoeba castellanii (Abd et al., 2003). In addition, F. tularensis ssp. holartica was found within amoebal cysts suggesting potential for long-term survival and an important environmental reservoir for tularemia (Abd et al., 2003). The isolation of the bacterium from water ecosystem (Mironchuk Iu and Mazepa, 2002), as well as from natural spring water (Willke et al., 2009), support the hypothesis that protozoa may serve as a reservoir of F. tularensis in the nature (Morner, 1992). However, association of F. tularensis with aquatic ecosystems may also be due to interaction of F. tularensis with mosquitoes that inhabit these ecosystems for breeding grounds.

Clinical manifestation of the disease

There are several clinical forms of tularemia. The ulceroglandular form is often a result of vector-borne transmission by ticks in the USA and Central European regions and by mosquitoes in the Nordic countries (Oyston et al., 2004; Santic et al., 2006; Petersen et al., 2009). Oropharyngeal tularemia is the result of ingestion of contaminated food and water. The primary ulcer is localized in the mouth, and lymph nodes of the neck region are enlarged. Respiratory tularemia occurs after inhalation of contaminated dust and is often related to farming activities or landscaping. Pneumonia may not always be present during the respiratory form but may be present in other forms of tularemia (Santic et al., 2006; Oyston et al., 2004).

The bacteria multiply at the initial site of infection, and then spread to the regional lymph nodes, liver and spleen (Oyston et al., 2004; Santic et al., 2006). Tularemia often presents with non-specific flu-like symptoms such as headache, fever, chills, nausea, diarrhoea and pneumonia (Oyston et al., 2004; Santic et al., 2006). Most of the symptoms are non-specific and can be easily misidentified as other forms of febrile diseases.

Evolution and patho-adaptation of F. tularensis

Genomic studies on F. tularensis have been facilitated by the completion of sequencing of the genomes of many strains that represent the four subspecies (Champion et al., 2009; Vogler et al., 2009). Despite marked differences in their virulence and geographical origin, there is more than 95–98% overall identity among the four subspecies identified with an average of ∼1700 open reading frames (ORFs) (Champion et al., 2009; Vogler et al., 2009). This is a relatively small bacterial genome considering the wide variety of ecological niches and the wide host range this organism is adapted to. The relative genetic homogeneity of F. tularensis ssp. may be related to the intracellular lifestyle of the bacterium, since stable bacterial lifestyle correlates with genomic stability (Champion et al., 2009; Vogler et al., 2009).

Although a gain of virulence characteristics is usually associated with allelic diversity or gene acquisition, there are well-documented instances of specific gene loss enhancing virulence of other species and of genome decay resulting in emergence of highly virulent species (Chain et al., 2004). The evolutionary aspects of subspecies of F. tularensis are complex and apparently involve both genomic acquisition and genomic decay in the more virulent strains, suggesting a role in patho-adaptation to infect mammals (Champion et al., 2009; Vogler et al., 2009).

Earlier studies of analyses of unidirectional genomic deletion and single nucleotide variations have indicated that the four subspecies of F. tularensis have evolved by vertical descent (Fig. 1) (Svensson et al., 2005). The genomes of the novicida ssp. strains has > 97% identity to tularensis ssp. and contain the highest percentage of intact ORFs and the least number of insertional elements (IS) compared with the virulent subspecies (Champion et al., 2009; Vogler et al., 2009). Genome-wide SNP phylogenetic studies have shown that differentiation of subspecies novicida predated differentiation of subspecies tularensis and holarctica from a common ancestor (Fig. 1) (Svensson et al., 2005).

Figure 1.

Phylogeny of four F. tularensis ssp. A phylogeny tree adopted from Champion et al. (2009), describing the evolution of both the virulent and non-virulent strains of F. tularensis from a common ancestor. The differentiation of F. tularensis ssp. novicida predated differentiation of the more virulent F. tularensis ssp. tularensis and F. tularensis ssp. holarctica subspecies. The numbers represent the branch lengths calculated using the average pathway method and are in the units of the number of changes over the whole genome sequence.

Genomes of strains of F. tularensis ssp. holarctica have a greater degree of heterogeneity than strains belonging to other F. tularensis subspecies (Champion et al., 2009), which may be partly due to the highest number of disrupted ORF as well as insertion and transposable elements incorporated in the genome. The observed increase in disrupted genes in F. tularensis ssp. holarctica lineages suggests that this subspecies has been undergoing a reductive evolution to become niche-adapted.

The presence of disrupted genes with some homology to components of type III or Type IV secretion systems is indicative of evolutionary genomic decay that may have enhanced patho-adaptation of F. tularensis (Champion et al., 2009), since effectors exported by such secretion systems could have been highly toxic and rapidly lethal to the host. In addition to genomic decay, evolution and path-adaptation of F. tularensis is further complicated by the gain of numerous genes. The FPI has a duplicate copy in subspecies tularensis and holarctica compared with the other less virulent subspecies (Nano et al., 2004). Earlier studies have shown that 41 genes are unique to tularensis and holarctica ssp. strains, but more recent and comprehensive genomic studies of 20 strains have shown that there are 14 intact genes unique to the human virulent tularensis and holarctica ssp. (Champion et al., 2009).

Moreover, a comparative study of the F. tularensis genomes revealed the presence of several regions of differences (RDs) (Salomonsson et al., 2009). Two genes within RD18 and RD19, encoding a putative type IV pilin and an outer membrane protein, have been lost in the attenuated LVS by direct repeat-mediated deletion (Salomonsson et al., 2009). Re-introduction of the two deleted genes into the LVS strain has restored virulence in mice to levels indistinguishable from the virulent F. tularensis ssp. holarctica strains (Salomonsson et al., 2009). Therefore, the loss of the two respective genes may explain, at least in part, some of the genetic bases of attenuation of the LVS strain.

Taken together, genome rearrangements due to the large number of insertion and transposable elements resulting in gradual and continuous genomic decay, in addition to acquisition of genes, have played important roles in the evolution and patho-adaptation of the human virulent subspecies of F. tularensis.

Entry of F. tularensis into macrophages

In mammals, F. tularensis invades, survives and replicates in variety of cell types including phagocytic and non-phagocytic cells of various species, as well as arthropod-derived cells (Aperis et al., 2007; Read et al., 2008; Vonkavaara et al., 2008; Santic et al., 2009). Macrophages are believed to be an important target for infection of mammals, and the pathogenesis of F.  tularensis depends on ability of the bacterium to survive and replicate within host cells (Santic et al., 2006; Oyston et al., 2004).

Francisella tularensis enters the host cell by inducing macrophages to produce unique asymmetric spacious pseudopod loops (Clemens et al., 2005). The uptake is dependent on complement receptors (CR3) and the complement factor C3 (Clemens et al., 2005; Schulert and Allen, 2006). Many intracellular pathogens exploit the complement receptor to enter macrophages, since this mode of entry blocks the oxidative burst of phagocytic cells. Thus, entry through the CR3 may have a major impact on the intracellular fate of F. tularensis, and may explain the discrepancies related to the results on phagosomal escape performed in different lab (see below). The mannose receptor (MR) and scavenger receptor class A (SRA) are also involved in uptake of F. tularensis (Balagopal et al., 2006; Pierini, 2006; Schulert and Allen, 2006). The surface nucleolin present on human monocyte-like THP-1 also contribute to uptake of F. tularensis (Barel et al., 2008). Francisella–macrophage interaction is also regulated by pulmonary surfactant proteins A and D (Beharka et al., 2002).

Cholesterol-rich lipid domains ‘lipid rafts’ together with caveolin-1 play a role in entry of F. tularensis into macrophages, and lipid rafts-associated components such as cholesterol and vaveolin-1 are incorporated into the Francisella-containing phagosome (FCP) membrane upon its biogenesis from the macrophage plasma membrane (Tamilselvam and Daefler, 2008). The recruitment of lipid rafts to the FCP may act as a platform for linking the entry process of F. tularensis at the cell membrane to the cytoskeleton and the intracellular signalling pathways (Tamilselvam and Daefler, 2008).

The uptake of F. tularensis is not affected by inhibition of the PI3K pathway, which is consistent with the findings that SHIP, an inositol phosphatase, negatively regulates F. tularensis-induced PI3K/Akt, and does not affect the uptake of F. tularensis (Fig. 2) (Parsa et al., 2006). The phagocytic pathway-related signalling component Syk is phosphorylated in F. tularensis-infected cells, its inhibition blocks uptake of F. tularensis, and its overexpression enhances bacterial uptake (Parsa et al., 2008). Interestingly, the Syk-dependent uptake is mediated via Erk but not through the PI3K/Akt pathway (Fig. 2) (Parsa et al., 2008). The Syk/Erk axis appears to be important for phagocytosis of F. tularensis, but the mechanism of signalling from Syk to Erk is not yet known (Parsa et al., 2008).

Figure 2.

Models of trafficking of F. tularensis within macrophages, its activation of the inflammasome and modulation of host cell signalling.
A. (1) The FCP matures to an early endosome that acquires the early endosomal antigen EEA1 followed by acquisition of the late endosomal markers Rab7, Lamp1 and Lamp2 for the first 30 min post infection. The FCP is rapidly and transiently acidified after acquisition of the proton vATPase pump within 15–30 min after infection. Its acidification is essential for rapid disruption of the FCP and subsequent rapid escape of F. tularensis from the phagosome into the cytosol of macrophages. Although the FCP acquires late endosomal markers for 30 min, lysosomal acid hydrolases, Cathepsin D, are excluded from the FCP. Following replication, F. tularensis re-enters the endosomal-lysosomal pathway by 24–48 h through autophagy. (2) The mglA and iglC mutants are both trafficked through the endosomal-lysosomal degradation pathway and the bacterial fail to escape into the cytosol.
B. (1) Signalling through the IFN-γ receptor upregulates SOCS3 expression, which suppresses STAT1. (2) The TLR2 and PI3k/Akt are activated, leading to NF-κB mediated production of pro-inflammatory cytokines and chemokines. (3) The activation of SHIP by F. tularensis negatively regulates the PI3k/Akt pathway.
C. (1) Upon binding to TLR2, F. tularensis induces a Myd88-dependent response that increases pro-IL-1β transcript levels. (2) Escape of F. tularensis from the phagosome into the cytosol induces activation of intracellular signalling pathway possibly by NLRs. The initial sensing of bacteria in the cytosol leads to IRF-3-dependent IFN-β secretion and type I IFNR-dependent signal. (3) The inflammasome triggers macrophage cell death by pyroptosis as well as release of pro-inflammatory cytokines IL-1β and IL-18.

Trafficking of F. tularensis within macrophages

Following entry, extracellular pathogenic or non-pathogenic bacteria are normally processed through the ‘default’ endosomal lysosomal degradation pathway, which is one of the first lines of innate defence against microbial infection (Santic et al., 2006). The nascent phagosome matures to an early endosome stage regulated by Rab5, followed by maturation into a late endosome regulated by the Rab7 GTPase (Fig. 2A). The late endosome becomes acidified upon acquisition of the proton vATPase pump that imports hydrogen protons into the phagosome. The acidified late endosome fuses to lysosomes and becomes a hydrolase-rich phagolysosome, within which most ingested bacteria are degraded. This process is very rapid and is completed within 15–30 min of formation of the phagosome (Santic et al., 2006). Therefore, many intracellular pathogens have evolved with idiosyncratic strategies to avoid a fatal fate within the phagolysosomes.

The FCP matures to an early endosome that acquires the early endosomal antigen (EEA)1 marker followed by acquisition of the late endosomal markers Lamp1, Lamp2 and the Rab7 GTPase within 15–30 min of phagosome biogenesis, but its fusion to lysosomes is very limited (Clemens et al., 2004; Santic et al., 2005a). The FCP is transiently acidified by the proton vATPase pump within 15–30 min of phagosome biogenesis (Fig. 2A), which is essential for subsequent rapid disruption of the FCP and rapid escape of F. tularensis into the host cell cytosol (Chong et al., 2008; Santic et al., 2008). Earlier studies using electron microscopy have shown that F. tularensis disrupts its phagosomal membrane 3 h after uptake (Golovliov et al., 2003; Santic et al., 2005a; Santic et al., 2006; Clemens et al., 2009). One crucial difference between the different studies from different labs is the use of fresh complement during the infection by some studies, which may alter the intracellular fate of the bacteria (Clemens et al., 2004; 2009; Chong et al., 2008; Santic et al., 2008). However, utilizing fluorescence-based phagosome integrity assays based on differential labelling of cytosolic versus vacuolar bacteria has provided an objective quantitative and sensitive approach to study the kinetics of phagosomal escape by F. tularensis, with much less potential artifacts than electron microscopy (Wehrly et al., 2009; Checroun et al., 2006; Santic et al., 2007; 2008). These fluorescence-based phagosome integrity assays are based on the ability of antibacterial antibody, loaded into the host cell cytosol after differential permeabilization of the plasma membrane, to bind cytosolic bacteria while bacteria within intact FCPs do not bind the antibody, since an intact FCP membrane is not permeable to the antibody (Checroun et al., 2006; Santic et al., 2007; 2008; Wehrly et al., 2009). The studies that did not include fresh complement during infection have shown that following acidification of the FCP by 15–30 min in human and mouse macrophages, the FCP becomes disrupted and the bacteria escape into the cytosol within 30–60 min post infection (Checroun et al., 2006; Santic et al., 2007; 2008; Wehrly et al., 2009). Importantly, inhibition of the proton ATPase pump causes a significant delay in phagosomal escape (Chong et al., 2008; Santic et al., 2008), indicating a major role of acidification of the FCP in rapid bacterial escape into the cytosol. Studies based on electron microscopy of infected cells in presence of fresh complement during infection that concluded that the FCP is not acidified are based on analyses at 3 h post infection (Clemens et al., 2009), a time point at which the bacteria have resided in the cytosol by > 2 h (Checroun et al., 2006; Santic et al., 2007; 2008; Wehrly et al., 2009). However, the use of fresh complement by Clemens et al. may result in alteration of the intracellular fate of the bacterium. It is not known whether the bacteria actively, and through a specific mechanism, limit fusion of the FCP to lysosomes or the bacteria escape upon acidification of the FCP prior to its fusion to the lysosomes.

The FCP of F. tularensis LVS excludes the NADPH oxidase subunits within polymorphonuclear leucocytes (PMN), and is devoid of superoxide and other reactive oxygen species (McCaffrey and Allen, 2006; Allen, 2007; Schulert et al., 2009); and bacterial pyrimidine biosynthesis is essential for this process (Schulert et al., 2009). Similar to macrophages, F. tularensis escapes from the phagosome within PMNs, but persists in the cytosol for at least 12 h (McCaffrey and Allen, 2006). This is a unique feature, since other pathogens that escape the phagosome in macrophages do not survive in PMNs.

Following intracellular replication of strains of F. tularensis belonging to various subspecies within bone marrow-derive mouse macrophages, the organism re-enters the endocytic compartment of murine macrophages by autophagy within 24–48 h (Checroun et al., 2006). However, it has not been reported whether autophagy is exhibited within human macrophages. Autophagy is a host cytosolic degradative pathway characterized by formation of double-membrane structures, autophagosome, to control degradation of damaged cellular vesicle and organelles (Mizushima et al., 2008). Some intracellular cytosolic pathogens have evolved strategies to avoid autophagy. Interestingly, F. tularensis downregulates several autophagy-related proteins that are required for biogenesis of autophagosomes (Butchar et al., 2008).

The Francisella pathogenicity island (FPI) and its role in phagosome biogenesis

In contrast to some intracellular pathogens, the genomes of F. tularensis strains have no functional type III or type IV secretion systems, which enable various intracellular pathogens to modulate biogenesis of their vacuoles to limit its fusion to lysosomes. The FPI-encoded proteins IglA and IglB have some homology to the Rhizobium leguminosarum proteins ImpB and ImpC that show similarity to members of type VI protein secretion system (Nano et al., 2004).

Few factors that regulate phagosome escape and intracellular survival of F. tularensis have been identified. Many of these factors are encoded by the FPI, which is composed of 17 ORFs (Nano et al., 2004). The iglC gene is essential for evasion of lysosomal fusion and bacterial escape into the cytosol (Santic et al., 2005a; Bonquist et al., 2008). However, IglD is required for bacterial proliferation within the cytosol and does not play a role in evasion of lysosomal fusion (Bonquist et al., 2008; Santic et al., 2005a). Persistent colocalization of the FCP harbouring the iglD mutant with the late endosomal marker Lamp has been reported (Bonquist et al., 2008), but escape of the iglD mutant has not been directly examined in that study. Reliance on colocalization of the FCP with Lamps as an indicator of phagosomal escape is likely to be unreliable, since cytosolic bacterial may be surrounded by endocytic vesicles that are LAMP positive, which would show as an artifact of colocalization of these vesicles with the FCP when examined by confocal microscopy (R. Asare et al., unpubl. data). The FPI-encoded PdpA is required for phagosome biogenesis, since the FCP harbouring the pdpA mutant persistently colocalizes with the late endosomal marker LAMP-1 and the mutant is defective in intracellular proliferation (Schmerk et al., 2009). In addition to the FPI genes, the four acid phosphatases of F. tularensis ssp. novicida are required for bacterial escape into the cytosol (Mohapatra et al., 2008).

The iglA, iglCD and the two FPI genes pdpA and pdpD are regulated by MglA (Lauriano et al., 2004; Guina et al., 2007), which is a global regulator (Guina et al., 2007). The FCP harbouring the iglC and mglA mutants fuses to lysosomes and the mutant fails to escape into the cytosol (Santic et al., 2005a; Bonquist et al., 2008). In addition to MglA, the FPI is partially regulated by an orphan two-component system response regulator, PmrA (Mohapatra et al., 2007). In addition, FerR has been shown to regulate FPI gene expression (Brotcke and Monack, 2008).

Transcriptional profiling of intracellular F. tularensis Schu S4 strain at various stages of phagosomal and cytosolic location within bone murine marrow-derived macrophages has revealed that the general oxidative and stress response pathways of F. tularensis are upregulated during both early phagosome and late endosomal stages of the intracellular infection (Wehrly et al., 2009), which is likely to be due, at least in part, to early acidification of the FCP within the first 15–30 min of phagosome biogenesis (Chong et al., 2008; Santic et al., 2008). The metabolic genes involved in amino acid catabolism are upregulated intracellularly during the cytosolic replication stage (Wehrly et al., 2009). Interestingly, despite the crucial role of the FPI in early stages of phagosome biogenesis and bacterial escape into the cytosol, maximal expression of the FPI genes is triggered at the end of cytosolic replication. This may suggest a role for the FPI genes during late stages of the infection or that this is an adaptive mechanism to have the bacterial ready and ‘armed’ to invade a future host cell after lysis of the infected cell (Wehrly et al., 2009).

Trafficking of F. tularensis within arthropod-derived cells

Arthropods, such as ticks, deer flies and horseflies, and mosquitoes, transmit F. tularensis to humans, and this is the major form of transmission in certain parts of Europe. It is likely that interaction of F. tularensis with arthropods has played a factor in its ecology and patho-adaptation to infect mammals, since there are numerous evolutionary conserved pathways between the two species, such as innate immune responses and intracellular trafficking events. Recently, arthropod models to study the infection of the arthropod host by F. tularensis, such as mosquitoes (Aperis et al., 2007; Read et al., 2008) and D. melanogaster (Vonkavaara et al., 2008; Santic et al., 2009) have been explored, and many FPI genes are required for intracellular proliferation within arthropod species-derived cells, similar to mammalian cells. In addition, F. tularensis is trafficked within D. melanogaster-derived cells similar to mammalian cells as the FCP transiently matures to a late endosome stage but has limited fusion to the lysosomes and followed by rapid bacterial escape into the host cell cytosol (Santic et al., 2009). In addition, F. tularensis proliferates in D. melanogaster adult flies, while FPI-defective iglC and iglD mutants, as well as the mglA mutant, are defective (Vonkavaara et al., 2008; Santic et al., 2009), recapitulating the mouse model of infection. Therefore, some F. tularensis factors required for virulence in mammals are also required for virulence in the arthropod model system.

Modulation of host cell signalling by F. tularensis

Two families of recognition receptors, the membrane-bound Toll-like receptors (TLRs) and cytosolic nuclear oligomerization domain (NOD)-like receptors (NLRs), recognize conserved bacterial molecules, such as lipopolysaccharide (LPS), peptidoglycan and flagellin (Manicassamy and Pulendran, 2009). When activated, TLRs recruit adaptor molecules to activate signalling through MyD88, ultimately leading to induction of genes involved in the inflammatory response (Fig. 2) (Manicassamy and Pulendran, 2009). Therefore, it is not surprising that F. tularensis can dampen or subvert host immune response by negative regulation of TLR signalling and this process is closely related to the virulence of F. tularensis strains.

Upon contact to the host cell, F. tularensis elicits signalling through TLR2 and triggers TLR2 expression (Katz et al., 2006; Malik et al., 2006; Cole et al., 2007). Within 1 h of infection of thioglycolate-elicited mouse peritoneal macrophages the bacteria colocalize with TLR2 and MyD88, but whether this colocalization is with the cytosolic organisms or with the FCP is not known (Cole et al., 2007). F. tularensis possesses a unique LPS with weak endotoxic activity and does not activate TLR4 (Hajjar et al., 2006). However, TLR4 triggering by F.  tularensis has been demonstrated and may contribute to signalling by other TLRs, in addition to TLR2 (Lembo et al., 2008).

During infection of human monocytes with F. tularensis TLR1, 4, 5, 6, 7 and 8 are downregulated while TLR2 is upregulated (Butchar et al., 2008). The downregulation of TLR signalling pathways results in dampening of the monocyte response to infection (Butchar et al., 2008). Interference with IFN-γ signalling by F. tularensis is one other strategy to evade the immune system to proliferate and cause disease. The mechanism of IFN-γ-mediated suppression does not require live bacteria and immunosuppressive effect is mediated by a heat-stable and constitutively expressed bacterial factor (Butchar et al., 2008). The F. tularensis-mediated suppression of IFN-γ signalling is associated with upregulation of SOCS3, a negative regulator of IFN-γ signalling (Fig. 2B) (Parsa et al., 2008). Upregulation of SOCS3 suppresses phosphorylation of STAT1 (Fig. 2B) (Parsa et al., 2008). Suppression of IFN-γ signalling is likely to be a major component of evasion of the immune system by F. tularensis, since IFN-γ triggers iNOS induction (Parsa et al., 2008), fusion of the FCP to lysosomes and failure of the bacteria to escape into the cytosol and proliferate (Santic et al., 2005b).

Although the PI3K/Akt pathway is activated and plays a critical role in the NF-κB-mediated production of inflammatory cytokines, there is a marked downregulation of this pathway accompanied by suppression of cytokine production upon escape of F. tularensis into the cytosol (Fig. 2B) (Parsa et al., 2006; Rajaram et al., 2006). The molecular mechanism of downregulation of PI3K/Akt is not known, but is dependent on proteins whose expression is regulated by MglA and are involved in phagosomal escape (Rajaram et al., 2006). The SH2 domain-containing inositol phosphatese SHIP is phosphorylated upon infection by F. tularensis, and the activation of SHIP negatively regulates the PI3K/Akt pathway and its role in production of pro-inflammatory cytokines (Fig. 2B). Therefore, it is thought that activation of SHIP by F. tularensis antagonizes the PI3K/Akt pathway and suppresses NF-κB-mediated production of pro-inflammatory cytokines (Fig. 2B) (Parsa et al., 2006). It has been recently shown that microRNA-155 (MiR-155) targets SHIP directly (O'Connell et al., 2009), and it would interesting to decipher whether SHIP is modulated by F. tularensis through an effect on expression of MiR-155.

Interestingly, Akt is also involved in phagosome biogenesis through activation of Rab5, and a dominant negative mutant of Akt blocks endosomal fusion, but this block is reversed by constitutively active Rab5 (Barbieri et al., 1998). In addition, Akt promotes the assembly of NADPH oxidase through phosphorylation of phox proteins leading to triggering of the oxidative burst (Chen et al., 2003; Hoyal et al., 2003). Therefore, SHIP-mediated downregulation of the PI3K/Akt signalling pathway by F. tularensis may serve multiple purposes that include down-modulation of the pro-inflammatory response, limiting biogenesis of the FCP through the endosomal-lysosomal pathway, and downregulating the oxidative burst, all of which would facilitate bacterial survival, fitness and subsequent proliferation within the host cell cytosol.

Activation of the inflammasome by F. tularensis

Few years ago, it was thought that by escaping into the host cell cytosol, bacteria reached a safe haven to hide from the immune system and are not exposed to microbicidal mechanisms. Recent evidence clearly show that bacterial escape into the cytosol triggers several innate immune pathways such as the inflammasome. This cystosolic signalling leads to pyropoptosis of the infected macrophage and a paracrine and systemic response that leads to the production of pro-inflammatory cytokines.

Escape of F. tularensis form the phagosome into the cytosol is essential for replication as well as for the cytosolic innate immune recognition system including activation of the inflammasome, which is a multiprotein complex that contain intracellular receptors, NLRs, an adaptor protein and the effector molecule caspase-1 (Martinon et al., 2002). In macrophages, F. tularensis activates the cytosolic surveillance system resulting in expression of genes regulated by IFN-β (Henry and Monack, 2007). Furthermore, the initial sensing of bacteria in the cytosol leads to IRF-3-dependent IFN-β secretion and type I IFNR-dependent signalling (Fig. 2C). Macrophages deficient for the type I IFN receptor (IFNRA−/−) fail to activate the inflammasome in response to infection by F. tularensis (Henry and Monack, 2007). In F. tularensis-infected macrophages, activation of the inflammasome triggers caspase-1, subsequent release of pro-inflammatory cytokines, such as interleukin-1β (IL-1β) and IL-18, and eventually cell death by pyropoptosis (Fig. 2C) (Henry and Monack, 2007). F. tularensis modulate activation of the inflammasome by at least the two proteins FTT0748 and FTT0584, but their functions are not known (Weiss et al., 2007). Thus, extracellular TRL-mediated signalling as well as signalling within the cytosol through type I IFN cooperates to control the intracellular infection by F. tularensis.

Concluding remarks

It is evident that genomic decay as well as gene acquisition has played roles in patho-adaptation of F. tularensis. It is possible that evolution of this organism within arthropods have facilitated this patho-adaptation, since intracellular trafficking of F. tularensis within arthropod and mammalian cells is very similar and the FPI genes are required for phagosome biogenesis and bacterial escape into the cytosol within both evolutionarily distant hosts. This may be similar to patho-adaptation of Legionella pneumophila to infect amoeba that has played a major role in its subsequent invasion of mammalian cells (Molmeret et al., 2005). It is most likely that studies on interaction of F. tularensis with the arthropod hosts will shed light on pathogenic evolution, patho-adaptation and transmission to mammals. While phagosome biogenesis and escape into the cytosol by F. tularensis seem to be very similar to Listeria, the mechanism of phagosomal escape of F. tularensis is still not known. In silico genomic analyses, as well as large scale mutant screens, have not identified a potential factor responsible for degradation of the phagosomal membrane. The relevance of preferential triggering of TLR2 instead of TLR4 by F. tularensis is still to be identified, but that may be a common theme among intracellular pathogens, such as L. pneumophila that also triggers TLR2. Downregulation of the PI3K/Akt signalling through activation of SHIP by F. tularensis may serve multiple purposes that include, but not limited to, subversion of the pro-inflammatory response, limiting fusion of the FCP to lysosomes, and blocking the oxidative burst, all of which would facilitate bacterial survival, fitness and subsequent proliferation within the host cell cytosol. It is not known whether activation of the inflammasome by F. tularensis serves fitness of the organism within the host cell cytosol or that it enables the host cell to eliminate the invading bacteria. The pace of our rapid understanding of F. tularensis in the past few years has been a remarkable progress that will hopefully continue to unravel the various mysteries that continue to surround various aspects of this organism.

Acknowledgements

Y.A.K. is supported by Public Health Service Awards R01AI43965 and R01AI069321 from NIAID and by the commonwealth of Kentucky Research Challenge Trust Fund and by Ministry of Science Education and Sports Republic of Croatia (Grant No. 062-0621273-0950).

Ancillary