In order to develop a successful infectious cycle, intracellular bacterial pathogens must be able to adapt their metabolism to optimally utilize the nutrients available in the cellular compartments and tissues where they reside. Francisella tularensis, the agent of the zoonotic disease tularaemia, is a highly infectious bacterium for a large number of animal species. This bacterium replicates in its mammalian hosts mainly in the cytosol of infected macrophages. We report here the identification of a novel amino acid transporter of the major facilitator superfamily of secondary transporters that is required for bacterial intracellular multiplication and systemic dissemination. We show that inactivation of this transporter does not affect phagosomal escape but prevents multiplication in the cytosol of all cell types tested. Remarkably, the intracellular growth defect of the mutant was fully and specifically reversed by addition of asparagine or asparagine-containing dipeptides as well as by simultaneous addition of aspartic acid and ammonium. Importantly, bacterial virulence was also restored in vivo, in the mouse model, by asparagine supplementation. This work unravels thus, for the first time, the importance of asparagine for cytosolicmultiplication of Francisella. Amino acid transporters are likely to constitute underappreciated players in bacterial intracellular parasitism.
Francisella tularensis is a Gram-negative bacterium causing the zoonotic disease tularaemia in a large number of wildlife animal species and has one of the broadest host range than any other known zoonotic disease-causing organism (Sjöstedt, 2011). Rodents, lagomorphs and amoeaba are thought to be the main reservoirs for human contamination (Keim et al., 2007). This highly infectious pathogen can be transmitted to humans in numerous ways, including direct contact with sick animals, inhalation, ingestion of contaminated water or food or by insect bites. Four different subspecies (subsp.) of F. tularensis that differ in virulence and geographic distribution exist, designated subsp. tularensis, holarctica, novicida and mediasiatica respectively. F. tularensis subsp. tularensis is the most virulent subspecies causing a severe disease in humans (Sjöstedt, 2007) and is considered a potential bioterrorism agent (Oyston et al., 2004). F. tularensis subsp. novicida (or F. novicida) is considered non-pathogenic for humans but is fully virulent for mice and is widely used as a model to study highly virulent subspecies.
Francisella is able to survive and to replicate inside a variety of mammalian host cells and in particular, macrophages (Hall et al., 2008). Its virulence is tightly linked to its ability to escape rapidly from the phagosome into the host cytosol (Santic et al., 2008) where it can replicate actively (Chong and Celli, 2010).
Several recent studies, mainly based on the characterization of nucleic acid and amino acid auxotrophic mutants, support the importance of metabolism in Francisella pathogenesis. For example, utilization of uracil has been shown to be required for inhibition of the neutrophil respiratory burst and intra-macrophage survival (Schulert et al., 2009). Uracil auxotroph mutants of F. tularensis LVS showed impaired phagosomal escape and severe multiplication defects in monocyte-derived macrophages. The mutant bacteria that could reach the cytosol were unable to replicate but still triggered the secretion of IL-8 and IL-1β. Remarkably, in J774 and HepG2 cells, these uracil auxotrophs appeared to multiply normally, implying that the defects in the pyrimidine biosynthetic pathway affect Francisella virulence in a cell type-specific manner. More recently, Peng and Monack showed that tryptophan auxotrophs of F. novicida were severely affected in intracellular multiplication and were attenuated in the mouse model (Peng and Monack, 2010).
Additional metabolic pathways have been shown to contribute to Francisella pathogenesis, including gluconeogenesis and glycolysis (Meibom and Charbit, 2010a) as well as biotin biosynthesis (Napier et al., 2012).
The identification of bacterial attributes specifically required for cytosolic replication by Francisella in the host cytosol may prove technically complex and time-consuming (Celli and Zahrt, 2013). Remarkably, two genome-scale screens, reported in human macrophages and Drosophila-derived cells (Asare and Abu Kwaik, 2010; Asare et al., 2010), identified more than 90 mutants that showed deficient phagosomal escape. These included genes encoding proteins with a broad variety of predicted functions. To date the best functionally characterized genes include: (i) purine biosynthetic genes (Pechous et al., 2006; 2008), (ii) the γ-glutamyl transpeptidase gene ggt (Alkhuder et al., 2009) and (iii) several genes of still unknown biological function, such as dipA (Wehrly et al., 2009; Chong et al., 2012), FTT0989 (Brotcke et al., 2006) and ripA (Fuller et al., 2008; Mortensen et al., 2012).
Earlier genome-scale genetic screens, aimed at identifying specific Francisella intracellular survival attributes, have also repeatedly identified genes encoding membrane proteins. These included notably putative amino acid transporters (Pechous et al., 2009; Meibom and Charbit, 2010b), highlighting the potential critical role of such nutrient acquisition systems in intracellular adaptation of Francisella. However, there is currently no functional or structural information available on any of these hypothetical transporters.
The Francisella genomes encode numerous predicted transport systems, the majority of which are secondary carriers (Meibom and Charbit, 2010a). Secondary transporters encompass several major families, including the major facilitator superfamily (MFS), involved in various functions including drug efflux, sugar and amino acid uptake (Reddy et al., 2012). Of particular interest, in the pathogenic intracellular bacterium Legionella pneumophila (Sauer et al., 2005), a MFS threonine transporter, designated PhtA (for Phagosomal transporter A), was shown to be required for intraphagosomal growth and differentiation. Phylogenetic analyses revealed that the Legionella genomes encode numerous Pht paralogues (Sauer et al., 2005). Importantly, Pht family members were also found in other bacterial species (Chen et al., 2008) but exclusively among intracellular pathogenic bacteria.
Remarkably, Francisella, which resides only transiently in a phagosomal compartment and multiplies exclusively in the cytosol of infected cells, also encodes several putative Pht orthologues (four to six, depending on the e value threshold chosen for the blast analysis; Sauer et al., 2005; Chen et al., 2008). Recently, a study aimed at developing anti-Francisella vaccines (Marohn et al., 2012), showed that inactivation of several of these putative transporters had altered intracellular replication kinetics and attenuation of virulence in mice in F. tularensis LVS (including the orthologue of ansP in F. tularensis LVS, see below). However, none of them has been functionally characterized yet and their substrate specificity is currently unknown.
In the present work, we addressed the functional role of one of these transporters, hereafter designated AnsP, which has been repeatedly identified by earlier genetic screens, in different Francisella subspecies and using a variety of different screening procedures. We demonstrate that the MFS transporter AnsP is required for bacterial multiplication in the cytosolic compartment of infected macrophages and for systemic dissemination in the mouse.
AnsP, a member of the major facilitator superfamily of secondary transporters
The region of gene ansP (Fig. S1) is highly conserved in all the Francisella genomes sequenced thus far and the AnsP protein of F. novicida (FTN_1586, 437 amino acids in length) shares more than 99% identity with its orthologues in F. tularensis subsp. holartica (FTL_1645), mediasiatica (FTM_0191) and tularensis (FTT_0129) (Fig. S2). In contrast, the homologues of the Francisella AnsP protein in other bacterial species do not exceed 28% amino acid identity. Indeed, the closest homologue in L. pneumophila (PhtE) shares only 26.7% amino acid identity, suggesting that the Francisella AnsP proteins displays properties specific to Francisellae species. Secondary structure prediction predicts that, like all MFS family members, AnsP comprises 12 transmembrane helixes (not shown).
Remarkably, the ansP gene was found among the genes required for replication of F. novicida (FTN_1586) in RAW264.7 macrophages (Llewellyn et al., 2011) as well as involved in replication in both U937 macrophages and drosophila S2 cells (Asare and Abu Kwaik, 2010). The orthologue of ansP in F. tularensis subsp. tularensis Schu S4 (FTT_0129) has also been found to be required for normal bacterial replication in the hepatocytic human cell line HepG2 (Qin and Mann, 2006). These screening results imply that the transporter AnsP has specific properties that are not compensated by any other putative transporter encoded by the Francisella genome in cells as well as in vivo. In this respect, it is worth mentioning that the other predicted Pht family members of Francisella share less than 27% amino acid identity with AnsP (and in most cases, also between each other).
Role of AnsP in Francisella intracellular survival
We constructed a chromosomal deletion of the ansP gene in F. novicida (FTN_1586) by allelic replacement (Lauriano et al., 2003). We first evaluated the possible impact of the mutation in vitro (using different media and growth conditions) and then followed the capacity of the mutant to infect and multiply in different cell types as well as in vivo, in the mouse model.
AnsP is not required for growth in broth
Inactivation of ansP had no impact on bacterial growth in broth and did not prevent acquisition of any of the essential amino acid present in Chemically defined minimal medium (CDMmin; Fig. S3). We used Biolog Phenotypic Microarray GNII plates to test possible metabolic defects of the ΔansP mutant. The respiration pattern obtained with the ΔansP and wild-type strains were identical, indicating that the mutation had not led to a defect in the utilization of any of the 96 carbon sources tested (Fig. S4).
AnsP is essential for intracellular bacterial multiplication
Quantitative RT-PCR analysis demonstrated that inactivation of the ansP gene (in a ΔansP deletion strain, see below) had no impact on the transcription of the two flanking genes FTN_1585 and FTN_1587 (Fig. S1). To further confirm that inactivation of the ansP gene had no polar effect on downstream gene expression, we complemented the ΔansP mutant strain with a plasmid-encoded copy of the wild-type ansP gene under pgro promoter control (see Experimental procedures for details).
We examined the ability of wild-type F. novicida (WT), ΔansP and ΔansP-complemented strains to survive and multiply in: (i) murine macrophage-like J774 cells, (ii) primary murine bone marrow-derived macrophages (BMM) and (iii) human hepatocytic HepG-2 cells (Fig. 1). Remarkably, the ΔansP mutant showed a severe growth defect in all cell types tested. In J774 cells, the ΔansP mutant showed a 10-fold reduction of intracellular bacteria after 10 h, and almost a 100-fold reduction after 24 h, as compared to wild-type F. novicida (Fig. 1A). In BMM, the ΔansP mutant showed an eightfold reduction of intracellular bacteria after 10 h, and a 16-fold reduction after 24 h (Fig. 1B). Finally, in HepG-2 cells, the ΔansP mutant showed a 10-fold reduction of intracellular bacteria after 10 h, and a 70-fold reduction after 24 h (Fig. 1C).
In all cell types tested, the intracellular growth defect of ΔansP mutant was similar to that observed with a deletion of the entire Francisella pathogenicity island [the ΔFPI mutant (Weiss et al., 2007)]. Introduction of the complementing plasmid (pKK214-pgro-ansP) restored bacterial viability to same level as in the wild-type parent, confirming the specific involvement of gene ansP in intracellular survival.
Subcellular localization of the ΔansP mutant in infected macrophages
We used the differential solubilization process previously described (Barel et al., 2010) to follow the subcellular localization of the ΔansP mutant in infected cells (Fig. 2A and B). Briefly, the plasma membrane was selectively permeabilized with digitonin. This treatment allowed the detection of cytoplasmic bacteria and proteins. Subsequent treatment with saponin rendered intact phagosomes accessible to antibodies and allowed detection of intra-phagosomal bacteria. Intracellular localization of the bacteria or LAMP-1 (used as a specific marker of phagosomes) was analysed using specific antibodies and their colocalization was monitored at two time points (1 h and 10 h post-infection; Fig. 2A). Quantification of each colocalization was performed with the Image J software (Fig. 2B).
With both wild-type F. novicida and the ΔansP mutant, colocalization with LAMP-1 was only around 15% after 1 h and remained in the same range throughout the infection. In contrast, for the ΔFPI mutant strain, 80% of bacteria colocalized with LAMP-1 after 1 h and colocalization remained very high after 10 h (85%). These results strongly suggest that the capacity of the ΔansP mutant to escape rapidly from the phagosomal compartment is unaffected.
To confirm these results, we performed thin section electron microscopy of J774 cells infected either with wild-type F. novicida or with the ΔansP mutant (Fig. 2C). Significant bacterial replication was observed in the cytosol of most infected cells 10 h post-infection with wild-type F. novicida. In contrast, bacterial multiplication was severely impaired in cells infected with the ΔansP mutant.
Finally, to quantify phagosomal permeabilization following F. novicida infection, we used the CCF4/β-lactamase technique originally described by Enninga and collaborators (Nothelfer et al., 2011), which we recently adapted to F. novicida (Juruj et al., 2013). Briefly, macrophages were loaded with CCF4, a fluorescence resonance energy transfer (FRET) probe retained within the host cytosol following the action of host cytosolic esterases. When cleaved by β-lactamase, the CCF4 FRET is lost leading upon excitation at 405 nm to a shift of fluorescence from green (535 nm) to blue (450 nm). F. novicida naturally express a β-lactamase able to cleave the CCF4 substrate (Broms et al., 2012). The phagosome permeabilization or F. novicida escape into the host cytosol is thus associated with a shift of the CCF4 probe from green to blue. Using this sensitive assay, we could not detect any differences in term of phagosomal permeabilization/vacuolar escape at 1 h post-infection (PI) in BMM between wild-type F. novicida and the ΔansP mutant (Fig. 2D). As expected (Juruj et al., 2013), we did not detect phagosomal permeabilization using a ΔFPI mutant.
This result confirms thus that the ΔansP mutant has normally access to the host cytosol but fails to multiply normally within this compartment.
Nutritional requirement of the ΔansP mutant
Since AnsP belongs to a family of putative amino acid transporters, we hypothesized that the intracellular multiplication defect of the ΔansP mutant might be alleviated by amino acid supplementation (Fig. 3). Therefore, we first compared the kinetics of multiplication of the ΔansP mutant in J774 macrophages, in the presence or absence of casamino acids in the cell culture medium (Fig. 3A). Addition of 0.5% casamino acids restored multiplication of the ΔansP mutant to the wild-type level.
This result suggested that AnsP might be an amino acid transporter whose function can be bypassed by other amino acids or peptide permeases.
We then reasoned that the substrate of AnsP could be deduced by further identifying the amino acid(s) responsible for this supplementation. For this, we tested the capacity of pools of amino acids to restore wild-type intracellular multiplication of the ΔansP mutant in J774 macrophages. Supplementation with a pool of six amino acids essential for growth in CDM did not improve intracellular multiplication of the ΔansP mutant (Fig. 3B). In contrast, a pool of seven amino acids (alanine, glutamate, glutamine, glycine, phenylalanine, tryptophane and asparagine), non-essential for growth in CDM (and, thus, not present in the CDM), restored wild-type intracellular multiplication to the ΔansP mutant (not shown).
As mentioned above, the ΔasnP mutant was able to grow like wild-type F. novicida in CDMmin (Fig. S3). This implies that: (i) the AnsP protein is not required for the capture of the amino acid present in the CDMmin, and (ii) that the amino acid(s) that is (are) not required for growth of the ΔasnP mutant in CDM, become(s) essential for growth in cells. Finally, each of the remaining amino acids (i.e. from the non-essential and useful but non-essential pools) was tested individually (Fig. 3C). These assays demonstrated that only asparagine was capable of supplementing the defect of the ΔansP mutant.
AnsP is an asparagine transporter required for intracellular multiplication
The multiplication defect of the ΔansP mutant was specifically and totally restored by asparagine supplementation whereas the intracellular defect of the ΔFPI mutant (Weiss et al., 2007) was not improved by asparagine supplementation (Fig. 4A).
To further assess the specificity of the AnsP transporter, we tested whether the growth defect suppression by free asparagine was dose-dependent (Fig. 4B). Supplementation with asparagine at a final concentration of 0.03 mM failed to completely restore bacterial multiplication in infected J774 cells (the initial DMEM cell culture medium is devoid of asparagine). Indeed, after 24 h, growth of the ΔansP mutant was improved 10-fold as compared to growth without asparagine, but was still 15-fold lower than that of the parental strain. Supplementation with 0.3 mM asparagine, significantly improved intracellular multiplication of the mutant but this was still fivefold lower than that of the wild-type strain after 24 h. In contrast, addition of 3 mM asparagine restored wild-type multiplication levels.
The fact that the ΔansP mutant showed no apparent defect in phagosomal escape prompted us to test the impact of supplementation with asparagine (3 mM) at later times during the infection of J774 macrophages. Therefore, we supplemented the Δansp mutant with asparagine after 1 h or after 4 h of infection (i.e. when the mutant bacteria are in the cytosolic compartment but fail to multiply normally). Remarkably, in both conditions, multiplication of the ΔansP mutant was restored to wild-type levels already at 10 h and remained similar to the wild-type strain after 24 h (Fig. 4C).
These results thus confirmed that the multiplication defect of the ΔansP mutant is due to an asparagine uptake defect when the bacteria are present in the cytosolic compartment of infected cells.
We then tested whether the addition of aspartic acid and/or a source of ammonium (NH4Cl) would also restore intracellular growth of the mutant. In order to avoid a possible inhibitory effect on phagosomal escape, NH4Cl was added only 4 h after infection. Strikingly, only the simultaneous addition of both aspartic acid (3 mM) and NH4Cl (1 mM) restored wild-type intracellular multiplication to the ΔansP mutant (Fig. 5A), implying either that aspartic acid and ammonium serve as substrates for the production of asparagine or, conversely, that asparagine serves as a source of aspartic acid and ammonium.
Finally, to test whether asparagine-containing dipeptides could also bypass AnsP function, we followed the kinetics of intracellular multiplication of the ΔansP mutant in J774 macrophages supplemented with either Leu–Asn or Leu–Gly dipeptides (3 mM). The asparagine-containing peptide restored wild-type multiplication of the mutant, but the Leu–Gly peptides did not (Fig. 5B). This assay confirmed thus the specific asparagine requirement of the ΔansP mutant for intracellular growth. As additional controls, we tested two dipeptides containing either glutamate (Glu–Gly) or aspartic acid (Asp–Gly). None of them compensated the multiplication defect of the ΔansP mutant (Fig. 5C).
We also constructed a chromosomal deletion mutant (ΔFTL_1645) in F. tularensis LVS and evaluated its impact on intracellular growth. Similar to our findings with F. novicida, the F. tularensis LVSΔFTL_1645 mutant had a defect in intracellular multiplication in macrophages (Fig. S5). In J774 cells, the ΔansP mutant of LVS showed a fourfold reduction of intracellular bacteria after 24 h, and an 11-fold reduction after 48 h, as compared to the parental LVS strain. Remarkably, as in F. novicida, upon supplementation of the culture media with asparagine (3 mM), the ΔFTL_1645 mutant multiplied with wild-type kinetics. These results strongly support the conserved role of AnsP among Francisella species.
AnsP transports both aspartic acid and asparagine
Since asparagine is the amide derivative of aspartic acid, we reasoned that AnsP might also be involved in the transport of aspartic acid. We therefore first compared the uptake of radiolabelled aspartic acid (14C-Asp) by wild-type F. novicida to that of the ΔansP mutant. The wild-type strain incorporated 14C-Asp more rapidly than the ΔansP mutant, at 10 μM 14C-Asp (Fig. 6A). However, aspartic acid incorporation was only partially affected in the ΔansP mutant, at all time points tested. This prompted us to evaluate the impact of ansP inactivation in a broad range of aspartic acid concentrations (Fig. 6B). At all the concentrations tested, aspartic acid incorporation was only partially affected in the ΔansP mutant (approximately 60% of the wild-type values) and the apparent Km of aspartic acid transport was similar in both strains (approximately 20 μM). These data strongly suggested the presence of additional aspartic acid transporter(s) in the ΔansP mutant.
We then measured 14C-Asp incorporation in wild-type F. novicida and in the ΔansP mutant, in the presence of increasing concentrations of competing asparagine (Fig. 6C and D). Fully supporting our first results, addition of asparagine reduced the incorporation of 14C-Asp, at all the concentrations tested in the wild-type strain. The uptake of 14C-Asp was not abrogated (40% to 60% reduction were recorded, with 100 μM cold asparagine), most likely due to the presence of additional aspartic acid transporter(s) in the strain. Interestingly, addition of asparagine had no effect on the incorporation of 14C-Asp in the ΔansP mutant, strongly suggesting that the additional aspartic acid transporters do not mediate asparagine transport in the conditions tested. Altogether, these assays strongly suggest that AnsP is involved in the transport of both asparagine and aspartic acid; and that Francisella possesses additional aspartic acid transporter(s).
Contribution of AnsP to Francisella virulence
To determine if the AnsP protein played a role for the ability of Francisella to cause disease, we performed in vivo competition assays in 6- to 8-week-old BALB/c mice. Groups of five mice were inoculated by the intraperitoneal (i.p.) route with a 1:1 mixture of the wild-type F. novicida and ΔansP mutant strains. Multiplication in the liver and spleen was monitored at day 2 (Fig. 7). The competition index (CI), calculated in both target organs, was very low (2–5 × 10−4), demonstrating that gene ansP is critical for virulence of F. novicida in the mouse model (Fig. 7).
Strikingly, upon treatment of the mice with asparagine (3 i.p. injections with 0.5 ml of a 0.5 mM solution of asparagine; see Experimental procedures), the CI in both organs raised approximately 1000-fold, in both liver and spleen. This experiment demonstrated that asparagine supplementation in vivo also restored virulence of the ΔansP mutant.
In order to survive and multiply within their hosts, intracellular bacterial pathogens must evade host innate immune surveillance pathways (Jones et al., 2012). They must also cope with a nutrient-restricted environment and have therefore evolved uptake systems adapted to the nutrient resources available in the host cell compartment where they reside (Abu Kwaik and Bumann, 2013; Zhang and Rubin, 2013). We report here the identification of a novel amino acid transporter that is required for intracellular multiplication and virulence of the facultative intracellular pathogen Francisella. We show that inactivation of this transporter does not affect phagosomal escape but prevents bacterial multiplication in the host cytosol. Importantly, this work unravels for the first time the critical role of asparagine acquisition in the virulence of this pathogen.
The host cytosol, initially considered as a safe and nutrient-replete haven (Ray et al., 2009), is now recognized as a potentially life-threatening and nutrient-deprived environment for invading bacteria. Intracellular bacterial pathogens have therefore learned to adapt their metabolism and are equipped with efficient uptake systems, to capture limiting nutrients (amino acids, carbohydrates, ions, …) from the host. Furthermore, intracellular bacterial have also developed a variety of strategies to modulate the nutritional content of the host cell compartment where they reside.
To fulfil their needs in amino acids, intracellular bacteria may take advantage of natural degradative pathways, such as proteasomal degradation and autophagy, like Legionella and Anaplasma, respectively (Price et al., 2011; Niu et al., 2012), to access to plentiful sources of amino acids. Indeed, L. pneumophila, which resides in a vacuolar compartment that evades lysosomal fusion and is remodelled by the endoplasmic reticulum, injects the eukaryotic-like F box protein effector AnkB into the infected host cells (Al-Quadan et al., 2012). After lipidation by the host farnesylation machinery, AnkB anchors to the vacuolar membrane and serves then as a platform for the assembly of Lys48-linked poly-ubiquitinated proteins. Their degradation by the host proteasome generates elevated levels of amino acids, which can be imported into the vacuole (Price et al., 2011). The obligate intracellular bacterium Anaplasma phagocytophilum, which replicates inside a membrane-bound vacuole that resembles an early autophagosome but that does not fuse to lysosomes (Niu et al., 2008), was recently shown to actively induce autophagy, by secreting the effector Ats-1. After translocation to the cytoplasm of infected cells, a portion of Ats-1 interacts with the host autophagosome initiation complex. Ats-1 autophagosomes are then recruited to the vacuole containing bacteria. Upon fusion of the two outer membranes, the autophagic content is released in the resulting vacuolar compartment (Niu et al., 2012).
Of note, F. tularensis has been reported to induce the formation of a multi-membranous, autophagosome-like structure in primary murine macrophages, by Celli and co-workers (Checroun et al., 2006), initially suggesting that the bacterium might use autophagosome as an alternative intracellular survival niche. However, more recent studies have shown that only replication-deficient and chloramphenicol-treated F. tularensis were degraded via canonical autophagy (Chong et al., 2012), implying that wild-type F. tularensis avoids classical autophagy. Of particular interest, Steel et al. very recently reported (Steele et al., 2013) that F. tularensis induced autophagy in infected mouse embryonic fibroblasts and primary human monocytes. The authors showed that F. tularensis not only survived autophagy, but required autophagy for optimal intracellular growth. F. tularensis was shown to be able to import amino acids derived from host proteins. Furthermore, impaired intracellular growth upon autophagy inhibition could be rescued by supplying excess non-essential amino acids or pyruvate. Remarkably, F. tularensis intracellular replication was not affected in ATG5−/− macrophages or ATG5−/− MEF cells, indicating that the canonical autophagy pathway was not involved. Altogether, these data suggest that F. tularensis induces an ATG5-independent autophagy process in infected cells to obtain carbon and energy sources to support their intracellular replication.
Supporting the notion that Legionella and Anaplasma use these natural degradative pathways to acquire host nutrients for their growth, inhibition of these pathways (proteasomal degradation and autophagy respectively) block bacterial multiplication. Remarkably, in both cases, the resulting growth defects can be totally bypassed upon supplementation with excess amino acids (Price et al., 2011; Niu et al., 2012).
Alternatively, intracellular bacteria may also stop growing to restrict their need as much as possible upon amino acid starvation (Chlamydia); or be self-sufficient and constitutively produce all their amino acids (Mycobacterium tuberculosis) (Zhang and Rubin, 2013). In the case of Francisella, we have previously shown that the bacterium used the cysteine-containing tripeptide glutathione (GSH) as a source of cysteine, to replicate in infected cells (Alkhuder et al., 2009), thus suggesting that Francisella has evolved by exploiting the natural abundance of GSH in the host cytosol to compensate its natural auxotrophy for cysteine. We have also recently shown that Francisella infection simultaneously upregulated the surface expression of the eukaryotic amino acid transporter SLC1A5 and the downregulation of SLC7A5, in infected monocytes (Barel et al., 2012). Since both SLC7A5 and SLC1A5 work in concert to equilibrate the cytoplasmic amino acid pool (Fuchs and Bode, 2005), the differential effect of Francisella infection on their expression could be profitable to the bacterium for increasing the intracellular concentration of essential amino acids. These two examples suggest that Francisella is capable of triggering optimal intracellular nutrient growth conditions as well as taking advantage of abundant nutrients already present.
The capacity of the host cell to adapt its metabolism for starving the invading bacteria of essential nutrients has led to the now widely accepted paradigm of ‘nutritional immunity’ (Weinberg, 1975). In turn, the terms ‘nutritional virulence’ have been very recently proposed to designate the specific virulence properties of intracellular pathogens (Abu Kwaik, 2013; Abu Kwaik and Bumann, 2013). We found here that inactivation of the ansP gene (FTN_1586) severely impaired Francisella intracellular multiplication in all cell types studied and successfully identified asparagine as the unique amino acid residue able to restore normal intracellular multiplication of the ΔansP mutant. Altogether, these data presented support the following model that is depicted in Fig. 8: AnsP is an asparagine transporter whose function is, not required in broth, but essential inside host cells. Remarkably, in infected cells, the absence of AnsP can be bypassed by alternative (probably less efficient) amino acid transporters (Paa) when the level of asparagine is high; and/or by peptide transporters (Pdipep) (Meibom and Charbit, 2010a) supplying asparagine-containing peptides. The fact that asparagine supplementation of the ΔansP mutant, even after 1 h or 4 h of infection, completely alleviated the growth defect, establishes that asparagine is specifically required for bacterial multiplication in the cytosolic compartment of the host. We found that AnsP was also involved in aspartic acid transport. Remarkably, intracellular growth of the ΔansP mutant was only restored when both aspartic acid and ammonium were added, strongly suggesting that it is primarily the role of asparagine transporter that is critical for bacterial cytosolic multiplication.
Amino acids represent the major sources of carbon and energy for Legionella and Francisella. Both species have lost several of their amino acid biosynthetic pathways [arginine, serine, valine, threonine, methionine, cysteine, for Legionella (Ristroph et al., 1981)]; arginine, histidine, lysine, tyrosine, methionine, cysteine, for Francisella (Larsson et al., 2005). Hence, they both rely on the host to obtain these essential amino acids for their intracellular multiplication. However, mammalian cells are auxotrophic for several of these amino acids (Young, 1994). Their intracellular concentration is thus likely to be particularly critical for multiplication of these pathogens.
Asparagine is a non-essential amino acid for Francisella as well as for mammalian cells. Its physiological intracellular concentration is estimated in the μM range in mammalian cells (Rinehart and Canellakis, 1985; Wang et al., 1998). The severe intracellular growth defect of the ansP mutant suggests that this concentration of asparagine is limiting for the multiplication of Francisella in infected cells, in the absence of the AnsP transporter. In vivo, the mammalian asparagine synthase (ASNS) enzyme catalyses the conversion of l-aspartic acid and l-glutamine into l-asparagine and l-glutamate. However, although ubiquitously expressed, the levels of ASNS expression in adult animals may vary considerably among tissues (Balasubramanian et al., 2013). Thus, Francisella must cope with variable pools of asparagine during in vivo dissemination.
Our data indicate that AnsP also transports aspartic acid, an amino acid at the crossroad of numerous important metabolic pathways, including the biogenesis of several amino acids (asparagine, methionine, threonine, isoleucine, alanine and lysine) and which can also serve as a precursor for pyrimidine biogenesis. The intracellular multiplication defect of the ΔansP mutant might thus be the consequence of the reduced capacity of the mutant to acquire both aspartic acid and asparagine.
Of note, in silico analysis of the predicted proteome of F. novicida revealed a broad heterogeneity in asparagine residues content per protein (ranging from 17.6% to less than 1% per coding sequence), with an average of 5.9% (i.e. close to 1/20). This rules out the possibility that asparagine acquisition might be needed due to an over-representation of this amino acid residue in Francisella proteins.
AnsP belongs to the Phagosomal transporter (Pht) subclass of MFS that is exclusively found among intracellular pathogenic bacteria. As recalled in the Introduction, L. pneumophila encodes the PhtA threonine transporter (Sauer et al., 2005) that is required for intra-phagosomal bacterial growth and differentiation. Threonine is an essential amino acid for Legionella that is likely to be present in limiting concentration in the host (human cells are notably auxotroph for threonine). The growth defect of phtA mutant could be abrogated by supplementation with excess threonine or threonine-containing peptide, supporting the notion that the Legionella-containing phagosomes indeed do not contain sufficient peptides or threonine to bypass the phtA defect.
Our study confirms that, at least some, Pht members are involved in amino acid uptake but indicates that these transporters are not restricted to bacteria residing exclusively in a phagosomal compartment. This subfamily of MFS transporter might therefore be renamed ICAATs (for Intra Cellular Amino Acid Transporters). The present study unravels for the first time the role of an MFS transporter as a key player in Francisella intracellular nutrition, further strengthening the link between nutrition and virulence for intracellular pathogens. It is likely that a number of other Francisella MFS members are specifically involved in acquisition of other limiting intracellular substrates, still to be discovered. More generally, it is reasonable to assume that transporter family members contribute to the intracellular nutrition of many (if not all) intracellular bacteria and could thus constitute attractive targets for the rational design of molecules to manage spread of these pathogens.
Bacterial strains and plasmids
Francisella tularensis subsp. tularensis strain U112 (kindly provided by A. Sjöstedt) and its derivatives were grown: (i) in liquid, on Schaedler broth (BioMérieux, Marcy l'Etoile, France), Tryptic Soya broth supplemented with cysteine (Becton, Dickinson and company) or Chamberlain chemically defined medium (Chamberlain, 1965), and (ii) in solid, on pre-made chocolate agar PolyViteX (BioMerieux SA Marcy l'Etoile, France) or chocolate plates prepared from GC medium base, IsoVitalex vitamins and haemoglobin (BD Biosciences, San Jose, CA, USA), at 37°C. Escherichia coli was grown in LB (Luria–Bertani, Difco) at 37°C. Ampicillin was used at a final concentration of 100 μg ml−1 to select recombinant E. coli carrying pGEM and its derivatives. All bacterial strains, plasmids and primers used in this study are listed in Table S1.
Metabolic profiling was tested, using Biolog Phenotypic Microarray GNII plates. A positive respiration pattern (visualized by the appearance of a dark colour in the well) reflects the capacity of the bacterium to utilize the carbon sources tested. Each plate contained 96 different carbon sources, each present in non-limiting amounts. The assay was performed according to the Biolog guidelines (available at the internet site http://www.biolog.com).
Secondary structure predictions were performed using the program HMMTM available at the internet site http://www.cbs.dtu.dk).
Construction of a chromosomal deletion mutants
We generated a chromosomal deletion of gene ansP in wild-type F. novicida strain U112 (ΔFTN_1586), by allelic replacement of the wild-type region with a mutated region deleted of the entire gene (from the start codon to the TAA stop codon), substituted by the kanamycin (Kan) resistance gene npt placed under the control of the pgro promoter. First, the two regions (c. 600 bp each) flanking gene ansP (designated FTN_1586up and FTN_1586down respectively), and the npt gene [1161 bp, amplified from plasmid pFNLTP16H3 (Maier et al., 2006)] were amplified by PCR using the following pairs of primers: (i) FTN_1586up, p3/p4; FTN_1586down, p5/p6, (ii) npt, p1/p2. The region FTN_1586up-npt (c. 1700 bp) was then amplified by overlap PCR, using the FTN_1586-up and npt products, and cloned into pGEM–T easy vector to yield plasmid pGEM FTN_1586-up/npt. The fragment FTN_1586-down was finally digested with PstI and SalI (New England Biolabs) and subcloned into the corresponding sites of plasmid pGEM-FTN_1586up-npt (immediately downstream of the npt gene) to yield plasmid pGEM-FTN_1586up-npt-FTN_1586down. This plasmid was used as a template for the amplification of a 2260 bp fragment comprising FTN_1586up-npt-FTN_1586down, using primers p9/p10. This PCR product was gel purified (using the QIAquick Gel extraction kit, QIAgen) and directly used to transform wild-type U112. Recombinant bacteria, resulting from allelic replacement of the wild-type region with the mutated FTN_1586up/npt/FTN_1586down region, were selected on Kan-containing plates (10 μg ml−1). The mutant strain, designated ΔansP was checked for loss of the wild-type ansP gene, using specific primers, and by PCR sequencing (GATC Biotech).
We also generated a chromosomal deletion of the orthologous gene FTL_1645 in F. tularensis subsp. holarctica strain LVS. For this, we used the counter selectable plasmid pMP812 (LoVullo et al., 2009). The recombinant plasmid pMP812-ΔFTL_1645 was constructed by overlap PCR. Primers p15/p16 amplified the 994 bp region upstream of position +1 of the FTL_1645 coding sequence, and primers p17/p18 amplified the 1012 bp region immediately downstream of the FTL_1645 stop codon (Table S1). Primers p16/p17 have an overlapping sequence of 23 nucleotides, resulting in complete deletion of the FTL_1645 coding sequence after cross-over PCR. PCR reactions with primers p15/p16 and p17/p18 were performed with exTaq polymerase (Fermentas). The products were purified using the QIAquick PCR purification kit (QIAgen, CA). 200 μM of each was used as a template for PCR with primers p15/p18 and treated with 30 cycles of PCR (94°C for 30 s, 54°C for 30 s and 72°C for 120 s). The gel-purified 1996 bp fragment was digested with BamHI and NotI (New England Biolabs) and cloned into BamHI–NotI digested pMP812 (LoVullo et al., 2009). The plasmid is introduced into F. tularensis LVS by electroporation. F. tularensis LVS was grown to OD600 0.3–0.6 in Schaedler-K3 broth; bacteria were collected and washed twice with 0.5 M sucrose. Bacteria were suspended in 0.6 ml of 0.5 M sucrose and 200 μl was used immediately for electroporation in a 0.2 cm cuvette (2.5 kV, 25 mF, 600 W). After electroporation, bacteria were mixed with 1 ml of Schaedler-K3 broth and incubated at 37°C for 6 h before selection on chocolate agar (Bio-Rad, Hercules, CA, USA) and 5 μg ml−1 Kan. Colonies appeared after 3 days of incubation at 37°C and were subsequently passed once on plates with selection, followed by a passage in liquid medium without selection (to allow recombination to occur). Next, bacteria were passed once on agar plates containing 5% sucrose. Isolated colonies were checked for loss of the wild-type FTL_1645 gene by size analysis of the fragment obtained after PCR using primers combination p15/p18, p33/p18 and p15/p34. One colony harbouring a FTL_1645 deletion, as determined by PCR analysis, was used for further studies. Genomic DNA was isolated and used as the template in a PCR with primers p33/p34. The PCR product was directly sequenced using primers p33/p34 to verify the complete deletion of the FTL_1645 gene.
The plasmid used for complementation of the ΔansP mutant, pKK214-pgro-ansP, was constructed by overlap PCR. Primers p13/p14 amplified the wild-type ansP gene (1454 bp), and the primers p11/p12 amplified the 333 bp of the pgro promoter. Primers p12/p13 have an overlapping sequence of 20 nucleotides. PCR reactions with primer pairs p11/p12 and p13/p14, were performed with exTaq polymerase (Fermentas), and the products were purified using the QIAquick PCR purification kit (QIAgen, CA). 200 μM of each amplification product was used as a template for PCR with primers p11/p13 and treated with 30 cycles of PCR (94°C for 30 s, 54°C for 30 s and 72°C for 120 s). The gel-purified 1787 bp fragment was digested with SmaI and PstI (New England Biolabs) and cloned into SmaI–PstI-digested pKK214. The plasmids pKK214 (empty plasmid) and pKK214-pgro-ansP (complementing plasmid) were introduced into U112 and the ΔansP mutant by electroporation.
Growth kinetics in broth
Stationary-phase bacterial cultures of wild-type U112 and U112ΔansP mutant strains were diluted at a final OD600 of 0.1 in tryptic soya broth (TSB). Every hour, the OD600 of the culture wads measured, during a 9 h period.
Cells were grown in Chamberlain medium to mid-exponential phase and then harvested by centrifugation and washed twice with Chamberlain without amino acid. The cells were suspended at a final OD600 of 0.5 in the same medium containing 50 μg ml−1 of chloramphenicol. After 15 min of pre-incubation at 25°C, uptake was started by the addition of l-[U-14C] aspartic acid (Perkin Elmer), at various concentrations (14C-Asp ranging from 1 to 100 μM); with or without competitor asparagine, at various concentrations (Asn ranging from 10 μM to 1 mM). The radiolabelled 14C-Asp was at a specific activity of 7.4 GBq mmol−1. Samples (100 μl of bacterial suspension) were removed at regular intervals and collected by vacuum filtration on membrane filters (Millipore type HA, 25 mm, 0.22 μm) and rapidly washed with Chamberlain without amino acid (2 × 5 ml). At the end of each experiment, the filters were transferred to scintillation vials and counted in a Hidex 300 scintillation counter. The counts per minute (cpm) were converted to picomoles of amino acid taken up per sample, using a standard derived by counting a known quantity of the same isotope under similar conditions.
Multiplication in macrophages
J774 (ATCC TIB67), BMM (ATCC TIB-202) and HepG-2 (ATCC HB 8065) cells were propagated in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum. Cells were seeded at a concentration of 2 × 105 cells per well in 12-well cells tissue plates, and monolayers were used at 24 h after seeding. J774, BMM and HepG2 macrophages monolayers were incubated for 60 min at 37°C with the bacterial suspensions (multiplicity of infection of 100) to allow the bacteria to enter. After washing (time zero of the kinetic analysis), the cells were incubated in fresh culture medium containing gentamicin (10 μg ml−1) to kill extracellular bacteria.
At several time points, cells were washed three times in DMEM, macrophages were lysed by addition of water and the titre of viable bacteria released from the cells was determined by spreading preparations on chocolate agar plates. For each strain and time in an experiment, the assay was performed in triplicate. Each experiment was independently repeated at least three times, and the data presented originate from one typical experiment. For ΔansP suppression experiments, amino acids or peptides were added to 3 mM at the time of infection and maintained throughout the infection.
J774 cells were infected with wild-type F. novicida and mutant ΔansP bacteria. Samples for electron microscopy were prepared using the thin-sectioning procedure as previously described (Alkhuder et al., 2009).
J774 cells were infected with wild-type F. novicida, ΔansP or ΔFPI, strains for 1 h, 4 h and 10 h at 37°C, and were washed in KHM (110 mM potassium acetate, 20 mM Hepes, 2 mM MgCl2). Cells were incubated for 1 min with digitonin (50 μg ml−1) to permeabilize membranes. Then cells were incubated for 15 min at 37°C with primary anti-F. novicida mouse monoclonal antibody (1/500 final dilution). After washing, cells were incubated for 15 min at 37°C with secondary antibody (Ab) (Alexa Fluor 488-labelled GAM, 1/400 final dilution) in the dark. After washing, cells were fixed with PFA 4% for 15 min at room temperature (RT) and incubated for 10 min at RT with 50 mM NH4Cl to quench residual aldehydes. After washing with PBS, cells were incubated for 30 min at RT with primary anti-LAMP1 Ab (1/100 final dilution) in a mix with PBS, 0.1% saponine and 5% goat serum. After washing with PBS, cells were incubated for 30 min at RT with secondary anti-rabbit Ab (Alexa 546-labelled, 1/400 dilution). DAPI was added (1/5000 final dilution) for 1 min. After washing, the glass coverslips were mounted in Mowiol. Cells were examined using an X63 oil-immersion objective on a LeicaTSP SP5 confocal microscope. Colocalization tests were performed by using Image J software; and mean numbers were calculated on more than 500 cells for each condition. Confocal microscopy analyses were performed at the Cell Imaging Facility (Faculté de Médecine Necker Enfants-Malades).
Phagosome permeabilization assay
Quantification of phagosome permeabilization/vacuolar escape using the β-lactamase-CCF4 assay (Life technologies) was performed following manufacturer's instructions. Briefly, macrophages were infected as previously described for 1 h, washed and incubated in CCF4 for 1 h at room temperature in the presence of 2.5 mM probenicid (Sigma). Cells were collected by gentle scraping and analysed by flow cytometry on a canto 2 cytometer (BD Bioscience) without any fixation step. Propidium iodide-negative cells were considered for the quantification of cells containing cytosolic F. novicida using excitation at 405 nm and detection at 450 nm (cleaved CCF4) or 535 nm (intact CCF4).
All experimental procedures involving animals were conducted in accordance with guidelines established by the French and European regulations for the care and use of laboratory animals (Décrets 87-848, 2001-464, 2001-486 and 2001-131 and European Directive 2010/63/UE) and approved by the INSERM Ethics Committee (Authorization Number: 75-906).
Wild-type F. novicida and mutant strains were grown in TSB to exponential growth phase and diluted to the appropriate concentrations. Six- to 8-week-old female BALB/c mice (Janvier, Le Genest St Isle, France) were intraperitoneally (i.p.) inoculated with 200 μl of bacterial suspension. The actual number of viable bacteria in the inoculum was determined by plating dilutions of the bacterial suspension on chocolate plates. For competitive infections, wild-type F. novicida and mutant bacteria were mixed in 1:1 ratio and a total of 100 bacteria were used for infection of each of five mice. After 2 days, mice were sacrificed. Homogenized spleen and liver tissue from the five mice in one experiment were mixed, diluted and spread on to chocolate agar plates. Kanamycin selection to distinguish wild-type and mutant bacteria were performed. For ΔansP supplementation experiments, asparagine was injected i.p. into mice (500 μl of a 0.5 mM solution) 14 h before infection and during infection, at days 1 and 2.
We thank Dr A. Sjöstedt for providing the strain LVS. These studies were supported by INSERM, CNRS and Université Paris Descartes Paris Cité Sorbonne. Gael Gesbert was funded by a fellowship from the ‘Délégation Générale à l'Armement’ (DGA) and Elodie Ramond by a fellowship from the ‘Région Ile de France’.