Discovery of a glutathione utilization pathway in Francisella that shows functional divergence between environmental and pathogenic species

SUMMARY

Glutathione (GSH) is an abundant metabolite within eukaryotic cells that can act as a signal, a nutrient source, or serve in a redox capacity for intracellular bacterial pathogens. For Francisella, GSH is thought to be a critical in vivo source of cysteine; however, the cellular pathways permitting GSH utilization by Francisella differ between strains and have remained poorly understood. Using genetic screening, we discovered a unique pathway for GSH utilization in Francisella. Whereas prior work suggested GSH catabolism initiates in the periplasm, the pathway we define consists of a major facilitator superfamily member that transports intact GSH and a previously unrecognized bacterial cytoplasmic enzyme that catalyzes the first step of GSH degradation. Interestingly, we find that the transporter gene for this pathway is pseudogenized in pathogenic Francisella, explaining phenotypic discrepancies in GSH utilization among Francisella spp. and revealing a critical role for GSH in the environmental niche of these bacteria.

In Brief

Host glutathione (GSH) is a source of cysteine during intracellular replication of Francisella. Wang et al. find Francisella imports GSH via a previously unknown transporter and discover a cytoplasmic GSH-degrading enzyme. This pathway is functional in non-pathogenic Francisella, but pseudogenized in pathogenic lineages, suggesting its importance in an environmental niche.

Graphical Abstract

INTRODUCTION

It is increasingly appreciated that the success of bacterial pathogens relies on sophisticated strategies for scavenging nutrients from their hosts. These “nutritional virulence factors” can include mechanisms for manipulating the host to drive nutrient availability¹. For example, some intracellular pathogens hijack autophagic or proteolytic cellular machinery to release amino acids that can be exploited as carbon and energy sources^2,3. Other pathogens compete effectively with the host for nutrients that are available as a result of normal host physiology.

One metabolite present in particularly high abundance inside host cells is the tri-peptide glutathione (γ-L-glutamyl-L-cysteinyl-glycine, GSH). GSH and its oxidized counterpart GSSG play crucial roles in multiple essential processes including maintaining redox homeostasis, defense against reactive oxygen species, and protein iron-sulfur cluster synthesis⁴. Perhaps as a result of the ubiquity and high concentration of GSH in the cytosol of eukaryotic cells, certain intracellular pathogens couple GSH sensing to virulence factor induction⁵. Notable examples include Burkholderia pseudomallei, which induces type VI secretion transcription following GSH sensing by the VirAG two-component system, and Listeria monocytogenes, which senses GSH through PrfA, leading to the activation of a set of critical virulence determinants^6,7. Other pathogens, including Hemophilus influenza and Streptococcus spp. rely on co-opted host GSH to defend against oxidative stress^8,9.

In contrast to these pathogens, Gram-negative proteobacteria belonging to the genus Francisella are sulfur amino acid auxotrophs and catabolize host GSH as a source of organic sulfur. A transposon screen of Francisella tularensis subspecies holarctica LVS (F. tularensis LVS) revealed that the periplasmic enzyme γ-glutamyl transpeptidase (GGT), which cleaves 70 GSH into glutamate and cysteine–glycine (Cys–Gly), is essential for intracellular replication of this organism¹⁰. Using a similar approach, our laboratory identified an inner membrane proton-dependent oligopeptide transporter-family (POT) protein that imports Cys–Gly¹¹. We named this protein DptA and demonstrated that, consistent with its critical role in GSH catabolism, F. tularensis LVS ΔdptA is defective in intracellular replication. However, we found that inactivation of ggt or dptA neither attenuates intracellular growth nor compromises GSH catabolism in a closely related Francisella strain, F. tularensis subsp. novicida (F. novicida). Rather, we identified a predicted γ-glutamylcyclotransferase enzyme in F. novicida, ChaC, that participates in GSH catabolism and is required for robust F. novicida growth in media containing GSH as the sole organic sulfur source (GSH media).

Despite these additions to our understanding of GSH metabolism in Francisella, two lines of evidence suggested that it remained incomplete. First, if GSH breakdown by Ggt and ChaC represented the only entry points into GSH catabolic pathways in F. novicida, a strain lacking these enzymes should be unable to grow in media containing GSH as a sole cysteine source. On the contrary, we found that F. novicida Δggt ΔchaC grows in such media, albeit not at wild-type levels¹¹. Second, although ggt, dptA and chaC are present and expected to be functional in both F. novicida and F. tularensis LVS, inactivation of ggt only produces a growth defect in GSH media in F. tularensis LVS. Together, these observations led us to hypothesize that additional pathways for GSH catabolism remain to be uncovered in Francisella.

In this study, we employed Tn-seq to identify Ggt-independent pathways important for GSH utilization in Francisella. Through this analysis, we discovered that F. novicida possesses a previously unrecognized pathway for GSH utilization that consists of an outer membrane porin, an inner membrane transporter of intact GSH belonging to the major facilitator superfamily and a cytoplasmic glutamine amidotransferase family enzyme capable of initiating degradation of the molecule. We show that this pathway is mutationally inactivated in pathogenic Francisella spp., but widely conserved in members of the genus that are believed to inhabit an environmental niche. Our work thus has implications for the evolution of pathogenesis within Francisella, and provides evidence that the natural lifecycle of non-pathogenic Francisella likely includes replication within a GSH-rich habitat, such as the cytosol of unicellular eukaryotes.

RESULTS

Tn-Seq reveals genes required for GSH utilization in F. novicida U112.

Although Ggt is required for the growth of F. tularensis LVS in GSH medium, we previously found that its inactivation does not similarly impede F. novicida growth on this substrate¹¹. Moreover, using the genome of F. novicida, we were unable to identify additional characterized GSH catabolism pathways that are absent from F. tularensis LVS. This conundrum motivated us to undertake an unbiased approach for discovering GSH catabolism pathways in F. novicida. To this end, we generated transposon mutant libraries of F. novicida in the wild-type and Δggt backgrounds, and used transposon mutant sequencing to compare gene insertion frequencies for each library grown in media containing GSH versus cysteine as the sole sulfur source (Figures 1A–1C). Our decision to employ Δggt rather than ΔchaC in this experiment was motivated by our recent observation that, while ggt is not required for F. novicida growth when GSH is in excess (100 μM), the growth yield of the Δggt strain is slightly reduced when the concentration of GSH limits growth (Figure S1). Strains lacking ΔchaC grow to wild-type levels under both conditions, suggesting that Ggt has a larger role in GSH catabolism in our in vitro culturing conditions.

Figure 1. Tn-seq for discovery of F. novicida genes with synthetic phenotypes during growth on GSH.

(A) Schematic illustrating known and unidentified potential GSH catabolism pathways and their products (Glu, red circle; Cys-Gly, yellow and white circles) in the two genetic backgrounds employed in our screen. The heavy arrow depicted for the wild-type background emphasizes the primary conversion pathway (left, Ggt-mediated) while the dashed arrow indicates the residual GSH cleavage mediated by ChaC in the absence of Ggt (right). (B,C) Results of Tn-seq screen to identify genes required for growth of F. novicida on GSH medium in the wild-type (B) and Δggt backgrounds (C). Genes with the greatest difference in transposon insertion reads between growth in GSH and cysteine media in the Δggt background (purple) and other genes shown previously or in this study to participate in GSH uptake or catabolism (blue) are indicated. (D) Rank order depiction of the strength of the synthetic phenotype for genes important for growth of F. novicida Δggt in GSH medium. Rank order was calculated by dividing the ratio of transposon insertion frequency obtained for each gene during growth on GSH compared to growth on cysteine using the F. novicida Δggt background by the same ratio obtained using the wild-type background. See also Figure S1, Tables S1 and S2.

Our Tn-seq screen led to the identification of many genes important for the growth of F. novicida Δggt specifically in GSH media. Among the 40 top hits in the Δggt background – corresponding to a three-fold insertion frequency difference between cysteine and GSH as the sole sulfur source – only five were shared with wild-type (Tables S1 and S2). Interestingly, while important for growth in GSH media specifically in the Δggt background, chaC was not among the strongest hits we observed (Table S2), supporting our earlier observation that Δggt ΔchaC can propagate in GSH media. Also consistent with our prior findings, in the wild-type strain, the fitness cost of inactivating ggt or dptA in GSH media was modest (Figure 1B).

To highlight Ggt-independent pathways for GSH catabolism, we ranked F. novicida genes by the strength of their synthetic (Δggt versus wild-type) growth phenotype in GSH media. Two genes ranked substantially higher in this analysis than other hits from our screen: FTN_1011 and FTN_0435 (Figure 1D). These two genes were also those with the greatest difference in insertion frequency between growth on GSH and cysteine as sole sulfur sources in the Δggt background (46.7- and 29.6-fold difference in normalized read counts, respectively) (Figure 1C and Table S2). Neither of these genes have been characterized, nor have functions been ascribed to any close homologs. Thus, we hypothesized they could contribute to GSH catabolism through previously unknown mechanisms.

Identification and characterization of GSH transporter NgtA

The strongest synthetic phenotype during growth in GSH medium belonged to open reading frame FTN_1011 – herein named ngtA (novicida glutathione transporter A). NgtA is a member of the major facilitator superfamily (MFS) of transporters, and as is typical of these proteins, its predicted structure displays 12 transmembrane helices organized into two six-helix bundles connected by a flexible linker¹². Within the MFS, NgtA was previously classified into the Pht family¹³. Interestingly, Pht family members are found exclusively in intracellular pathogens; in Legionella pneumophila and Francisella, proteins in the family are important for intracellular replication by virtue of their role in amino acid transport or nucleoside transport ^13–17. However, the sequence of NgtA is substantially divergent from characterized Pht family members (26% sequence identity shared with PhtA, the most closely-related characterized Pht family member), and its function and substrate are unknown.

We hypothesized that NgtA could be a transporter of GSH. To test this hypothesis, we first generated in-frame deletion mutants of ngtA in the F. novicida wild-type and Δggt backgrounds. As predicted by our Tn-seq results, inactivation of ngtA in the wild-type background did not affect F. novicida growth in GSH medium (Figure 2A). However, growth of F. novicida Δggt ΔngtA was strongly impaired in GSH medium. This growth defect could be complemented by repairing the deletion of ngtA via allelic exchange, and inactivation of ngtA did not cause growth defects in media containing cysteine as a sole sulfur source in either background (Figure S2A). To determine whether NgtA contributes to GSH uptake in F. novicida, we mixed cysteine-starved strains with ³H-GSH ([Glycine-2-³H]-GSH) and measured cell-associated radiolabel following a short incubation (Figure 2B). In the wild-type strain, NgtA inactivation had no impact on ³H-GSH transport. Since Ggt and ChaC generate periplasmic Cys–Gly, which when transported into the cytoplasm by DptA would mask the potential role of NgtA in intact GSH transport, we next employed the Δggt ΔchaC background in these assays. As expected, these mutations diminished GSH transport; however, uptake of the labeled substrate dropped to levels approaching the limit of detection in F. novicida Δggt ΔchaC ΔngtA (Figure S2B). This result supports the hypothesis that in the absence of GSH cleavage in the periplasm, the intact tripeptide can be transported to the cytoplasm via NgtA.

Figure 2. NgtA is a major facilitator superfamily protein that transports intact GSH in a Ggt-independent manner.

(A) Normalized growth yield in GSH medium of the indicated F. novicida strains. (B) Quantification of the level of [Glycine-2-³H]-Glutathione (³H-GSH) uptake in the indicated strains of F. novicida after 45 min incubation. (C) Normalized growth yield of the indicated F. novicida strains after 36 hrs in defined medium containing Cys–Gly as a sole source of cysteine. (D) Neighbor-joining phylogeny of proteins from the Pht family of MFS transporters. Colored clades contain sequences identified in the original description of the family or subsequently characterized¹³. Representative proteins from the Chen et al. study or other reports are indicated by their respective clades, and transport substrate are indicated in parentheses when known. Candidate NgtA homologs are shown in purple, and the region of the phylogeny amplified in (E) is indicated (shading). (E) Neighbor-joining phylogeny of NgtA homologs in Francisella and related genera. Species names indicate the source of the protein sequences. Data in (A-C) represent mean ± s.d. Asterisks indicate statistically significant differences (unpaired two-tailed student’s t- test.; *p<0.05, ns, not significant). See also Figure S2 and Table S3.

Our data left open the formal possibility that F. novicida possesses a third mechanism to generate Cys–Gly from GSH, and that the dipeptide is the transport substrate of NgtA. Notably, our laboratory previously reported the F. novicida Cys–Gly transporter DptA¹¹. This strain exhibits only a partial growth defect in media containing Cys–Gly as a sole organic sulfur source (Cys–Gly media), suggesting that, indeed, other enzymes could support Cys–Gly transport (Figure 2C). However, inactivation of NgtA had no impact on F. novicida growth in Cys–Gly media in either the wild-type or ΔdptA backgrounds. Together, these data suggest that NgtA is a transporter with specificity for intact GSH.

In the initial report of the Pht family of MFS transporters, the NgtA homolog of F. tularensis was the only member of its cluster¹³. With many more genome sequences now available, we asked whether homologs of this protein could be found in other species. Using PSI-BLAST with the sequence of NgtA from F. novicida as the seed, we collected all publicly available sequences encoding MFS proteins from the Pht family and constructed a phylogeny. We found that while many of the clades in the phylogeny are dominated by sequences deriving from Legionella and Coxiella spp., the clade containing NgtA consists largely of sequences deriving from Francisella spp. and related Thiotrichales, with only two homologs found outside this group, in metagenome-assembled genomes of uncharacterized strains identified only as belonging to the Legionellales order and Coxiellaceae family, respectively (Figures 2D and 2E and Table S3). Our inability to identify NgtA orthologs more broadly suggests that this mechanism of transporting GSH may be an adaptation particularly exploited by organisms in this group.

A cytoplasmic glutamine amidotransferase family enzyme that initiates GSH degradation

The finding that a GSH transporter can facilitate F. novicida growth in GSH media in the absence of Ggt and ChaC implies that this organism must encode cytoplasmic proteins capable of initiating GSH catabolism. The gene with the second strongest synthetic phenotype in our transposon mutant screen, FTN_0435, encodes a predicted glutamine amidotransferase (GATase). Most GATase proteins function in biosynthetic reactions in which the amido group from glutamine is transferred to an acceptor substrate, generating glutamate and an aminated product¹⁸. However, a limited number of GATase domain-containing proteins instead function as catabolic enzymes that cleave γ-glutamyl bonds in assorted substrates, releasing glutamate. These include enzymes that hydrolyze such substrates as the folate storage and retention molecule folylpoly-γ-glutamate, the spermidine degradation intermediate γ-glutamine-γ-aminobutyrate, GSH conjugates involved in glucosinolate synthesis, and notably, GSH itself^19–23.

The latter was found to occur in yeast and is catalyzed by the enzyme Dug3p²³. Structure modeling revealed that FTN_0435, herein named CgaA (cytosolic glutathione amidotransferase A), shares an overall fold and a conserved predicted catalytic triad (C97, H184, E186) with class I GATases²⁴ (Figure 3A). This is in contrast to the GSH-targeting enzyme of yeast, a class II GATase²³. Nevertheless, we found that, as predicted by our Tn-seq results, CgaA is required for F. novicida Δggt growth in GSH medium, a phenotype that could be genetically complemented (Figure 3B,C). Substitution of the predicted catalytic cysteine with alanine (cgaA^C97A) had no impact on the protein level produced but recapitulated the growth phenotype of a cgaA deletion, supporting an enzymatic role for this protein in GSH catabolism (Figure 3B, Figure S3A). We thus asked whether CgaA encodes a cytoplasmic enzyme able to initiate GSH degradation.

Figure 3. CgaA is a cytoplasmic glutamine amidotransferase (GATase) family protein that degrades GSH.

(A) Alignment of the predicted structure of CgaA (orange) and the crystal structure of a characterized class I GATases, P. aeruginosa SpuA (blue, PDB: 7D4R, only one subunit of the SpuA homodimer is shown). The conserved catalytic triad is indicated (numbers correspond to amino acid positions in F. novicida CgaA). (B,C) Normalized 36 hrs growth yields of the indicated strains of F. novicida. (C) Coomassie stained SDS-PAGE analysis of purified CgaA and CgaA^C97A. (D) Glutamate released following 60 min incubation of purified CgaA or CgaA^C97A (1 μM protein) with GSH or Gln (10 mM substrate). Data in (B) (C) and (E) represent means ± s.d. Asterisks represent statistically significant differences (unpaired two-tailed student’s t- test.; *p<0.05, ns, not significant). See also Figure S3.

To determine the substrate specificity of CgaA, we used established in vitro assays to measure the activity of CgaA and CgaA^C97A purified from E. coli. Biosynthetic GATase proteins exhibit glutaminase activity, generating glutamate and ammonia in the absence of their respective amido group-accepting substrates. However, we detected only a low level of glutamate accumulation following incubation of CgaA with glutamine. On the contrary, we readily detected glutamate released from GSH by CgaA, and this product was not detected above background levels in reactions with CgaA^C97A (Figures 3D and 3E). In total, these data support the hypothesis CgaA is a GATase that acts downstream of NgtA to initiate the degradation of GSH via cleavage into Glu and Cys–Gly.

Although our genetic and biochemical data strongly suggest that GSH is a physiological substrate of CgaA, we noted the rate of glutamate release from the purified enzyme is low. In yeast, the GATase enzyme Dug3p acts in concert with two other proteins, Dug1p and Dug2p ²³. Purified Dug3p is inactive in vitro unless bound to Dug2p, which allosterically activates the enzyme. Dug1p is a Cys–Gly specific peptidase that does not physically associate with the Dug2p–Dug3p complex. CgaA is encoded by the third gene in a predicted five gene operon. Examination of our Tn-seq results suggested that the genes encoded upstream of cgaA within this operon may also be important for growth on GSH in the F. novicida Δggt background (Figures S3B-D, Tables S1 and S2); however, we also considered that insertions within these genes may lead to polar effects on cgaA. To distinguish between these possibilities, we generated a conservative in-frame deletion in the first gene in the operon in F. novicida Δggt and measured the growth of this strain relative to F. novicida Δggt ΔngtA in GSH media. This strain exhibited robust growth in GSH media (Figure S3E), strongly suggesting that polar effects underlie the apparent depletion of genes upstream of cgaA in our Tn-seq study, and moreover that CgaA does not require adjacently encoded proteins for its activity.

Parallel pathways for GSH catabolism contribute to F. novicida intramacrophage growth

Previous studies indicate that ggt mutants of F. tularensis SCHU S4 and LVS are attenuated in virulence ^{10,11,25–27}. This has led to the consensus in the field that GSH serves as an important source of organic sulfur for these bacteria during infection^1,5,28. To our knowledge, the role of host GSH catabolism during F. novicida infection has not been examined. Unlike F. tularensis SCHU S4 and LVS, our results suggest that F. novicida may be capable of utilizing multiple pathways for GSH scavenging in vivo. To explore this possibility, we measured the growth of F. novicida strains lacking the function of one or both GSH uptake pathways in bone marrow-derived murine macrophages (BMMs). Interestingly, we found that only F. novicida strains in which both pathways are inactivated display a detectable intracellular growth defect (Figure 4A). We next examined the importance of the two GSH catabolism pathways in a more complex model of infection, a murine intranasal model²⁹. At 48 hrs post-infection in the intranasal model, we observed a modest decrease in recovery of F. novicida Δggt from lung samples. However, in contrast to our macrophage infection study, no further decrease was detected when Δggt was combined with ΔngtA or ΔcgaA (Figure 4B).

Figure 4. NgtA contributes to intramacrophage replication of F. novicida but is mutationally inactivated in pathogenic Francisella strains.

(A) Normalized intracellular growth of the indicated strains of F. novicida in bone marrow-derived murine macrophages (24 hrs post-infection. (B) Bacterial burden in mouse lungs at 48 hrs post intranasal infection with ~100 CFU of the indicated strains of F. novicida. (C) Structural model of F. novicida NgtA highlighting differences with NgtA in other Francisella species. The terminal residues resulting from truncating frameshift mutations (FS) indicated in parentheses. FS-1, location of truncation resulting from frameshift in F. tularensis subsp. holarctica; FS-2, location of truncation resulting from frameshift in F. noatunensis; loop, poorly conserved region with many differences between species; IGS, three-residue deletion found in F. tularensis subsp. tularensis. The first six and the last nine residues in the NgtA structure are trimmed. (D-F) Normalized growth yields in GSH medium of the indicated strains of F. novicida (D,E) or F. tularensis LVS (F) . (G) Schematized phylogeny of Francisella species indicating predicted functionality of NgtA (functional, green; inactivated, solid grey; absent, dashed grey), animal association when known (mammal association indicated by rabbit schematic) and mutations present in ngtA (colors correspond to panel C). The ngtA sequence of P. persica contains many mutations (***) and ngtA appears to have been lost completely from F. endociliophora. Phylogenetic relationships derived from Vallesi et al.⁴². Data shown in (A) (B) and (D-F) represent as means ± s.d. Data points in (A-B) indicate technical replicates from 3 (A) or 4 (B) biological replicates conducted. Asterisks indicate statistically significant differences (A and B, one-way ANOVA with Dunnett’s multiple comparison test comparing mutant strains to wild-type; D-F, unpaired two-tailed student’s t- test; *p<0.05, ns, not significant.) See also Figure S4.

The results of our murine infection model study suggest that in the context of an animal infection, Ggt-mediated cleavage of GSH may be the primary mechanism by which F. novicida acquires organic sulfur. This is consistent with the observation that ggt mutants of F. tularensis SCHU S4 and F. tularensis LVS have significant virulence defects; however, it remained unclear why ggt inactivation alone is sufficient to inhibit in vitro growth in GSH medium in these other subspecies, but not in F. novicida. Furthermore, the magnitude of virulence defect for the Δggt background of F. novicida is qualitatively lower than that of F. tularensis SCHU S4 and LVS^{10,11,25–27}. While investigating explanations for this difference, we found that the ngtA genes of SCHU S4 (ngtA^SCHU), LVS (ngtA^LVS), as well as those of other F. tularensis subsp. tularensis and subsp. holarctica strains encode proteins with a three amino acid deletion relative to NgtA^NOV. In the predicted structure of NgtA^NOV, these amino acids (I113-G114-S115) reside within a transmembrane helix located in the core of the protein (IGS, Figure 4C). F. holarctica ngtA genes further contain a small 5′ in-frame deletion and a premature stop codon that removes the last two predicted transmembrane helices (FS-1, Figure 4C). Taken together with our current findings, these observations led us to hypothesize that NgtA, and thus GSH transport, is compromised in these pathogenic strains of Francisella. Indeed, we found that F. novicida Δggt carrying ngtA^SCHU or ngtA^LVS in place of ngtA^NOV demonstrated growth behavior matching F. novicida Δggt ΔngtA in GSH medium (Figure 4D). Furthermore, F. novicida Δggt carrying ngtA^NOV engineered to contain only the three-residue deletion found in ngtA alleles from human pathogenic strains (ngtA^ΔIGS) was similarly unable to grow in GSH medium (Figure 4E). All NgtA variants bearing ΔIGS were undetectable despite efforts at enrichment by immunoprecipitation, suggesting that this deletion is sufficient to destabilize NgtA (Figure S4). We also performed the converse experiment in F. tularensis LVS by over-expressing ngtA^NOV in the Δggt background. The expression of ngtA^NOV resulted in a small, but reproducible restoration of growth in GSH medium (Figure 4F). We speculate that the limited degree to which NgtA^NOV expression restores F. tularensis LVS GSH autotrophy could be the result of pseudogenization or regulatory alteration of elements downstream of NgtA that are important for efficient GSH catabolism.

Our finding that ngtA is inactivated in multiple F. tularensis subspecies prompted us to examine the nature and prevalence of mutations in ngtA amongst Francisella spp. more broadly. Interestingly, we found evidence supporting pseudogenization of ngtA in two additional lineages of animal-associated Francisella: the tick endosymbiont F. persica and the fish pathogens F. noatunensis and F. orientalis (Figures 4C and 4G). On the contrary, the ngtA sequences of Francisella without a known animal association bore mutations primarily restricted to a hypervariable cytoplasmic loop that are not expected to inactivate the transporter (Figures 4C and 4G). Given that we find evidence for repeated ngtA pseudogenization events limited to Francisella lineages adapted to animal hosts, our results suggest that intact GSH uptake is most beneficial for bacteria in this genus in the environment, perhaps during replication within unicellular eukaryotes.

FupA is a porin that mediates GSH uptake.

We were surprised to find fupA as a gene with highly differential transposon insertion frequency during growth on GSH versus cysteine as sole sulfur sources in both the wild-type and Δggt backgrounds of F. novicida (Figures 1B, 1C, Tables S1 and S2). FupA is a member of a family of paralogous predicted outer membrane proteins unique to Fransicella species, several of which, including FupA, are widely thought to mediate high affinity uptake of ferrous iron^30–32. Despite this dogma, F. tularensis SCHU S4 ΔfupA exhibits a general growth defect in minimal media regardless of iron source or type, and proteoliposome assays using purified FupA provide evidence that it promotes membrane permeability^30,33. We thus hypothesized that FupA may contribute to F. novicida growth in GSH media by facilitating GSH passage through the outer membrane.

To directly examine the role of FupA in GSH catabolism, we generated an in-frame deletion of fupA in the wild-type and Δggt backgrounds of F. novicida. Consistent with our Tn-seq results, deletion of fupA in both backgrounds resulted in a strong growth defect specifically in GSH media (Figure 5A). We then evaluated the role of FupA in GSH transport by measuring the impact of ΔfupA on cellular uptake of ³H-GSH by F. novicida. We found that in the absence of FupA, ³H-GSH uptake was reduced below levels observed in F. novicida Δggt (Figures 2B and 5B). Furthermore, inactivation of ggt in the ΔfupA background did not further reduce ³H-GSH uptake. These data support our hypothesis and further suggested that FupA could act as a general porin of F. novicida. Indeed, a prior analysis of predicted β-barrel proteins in Francisella did not identify clear homologs of previously characterized general porins³⁴. In addition to serving as a conduit for the uptake of nutrients, a common feature of porins is that they present a vulnerability by providing entry to harmful molecules such as antibiotics and hydrogen peroxide ^35,36. We found that F. novicida ΔfupA is significantly more resistant to hydrogen peroxide than the wild-type, further supporting its functional assignment as a porin of F. novicida (Figure 5C). These findings show that GSH accesses the periplasm of F. novicida via FupA, thus providing an explanation for the insertion frequency in fupA observed in our screen. In total, our genetic, biochemical and phenotypic data allow us to assemble a new, complete model for GSH transport and catabolism in Francisella (Figure 6).

Figure 5. FupA is a porin required for GSH uptake in F. novicida.

(A) Normalized growth in GSH medium of the indicated strains of F. novicida. (B) Quantification of the level of ³H-GSH uptake in the indicated strains of F. novicida after 45 min incubation. (C) Survival of the indicated strains of F. novicida after incubation of mid-log phase cultures with 1.5 mM H₂O₂ for 30 min or 60 min. Data in (A-C) represent mean ± s.d. Asterisks indicate statistically significant differences (unpaired two-tailed student’s t- test.; *p<0.05, ns, not significant).

Figure 6. Comprehensive model of GSH transport and catabolism in Francisella.

Both the pathways for import and cytosolic catabolism of GSH discovered in this study (left, green shading) and for periplasmic degradation of GSH and subsequent fate of imported Cys-Gly (right, grey shading) are indicated.

DISCUSSION

In this study, we report the finding that Francisella spp. encode a previously uncharacterized, Ggt-independent pathway for GSH uptake that has been lost in each established animal-colonizing lineage of the genus. On one hand, the correlation between inactivation of the GSH transporter encoding gene ngtA and adaptation to animal association is counterintuitive, as GSH is only present at sufficient concentrations to be useful as a source of sulfur in host-associated environments. However, we found evidence that in F. novicida, a species without a known physiological animal host, NgtA and the intracellular GSH-degrading enzyme CgaA work in concert with Ggt to support intramacrophage replication. This was not observed in a mouse model of infection, where a range of cell types are infected³⁷. The mechanisms macrophages employ to kill bacteria share many features in common with those utilized by predatory protozoa^38–40. While the environmental niches colonized by non-pathogenic Francisella species remain largely uncharacterized, several species have been isolated from bacterivorous ciliates, including the deeply branching species F. adeliensis, which encodes intact ngtA^41,42. Accordingly, we speculate that functional NgtA is maintained in environmental lineages due to its utility during colonization of a macrophage-like intracellular habitat within eukaryotic microbes. In support of an intracellular environment representing the natural niche of diverse Francisella species, species that encode functional NgtA also encode the host cell-targeting type VI secretion system associated with the Francisella pathogenicity island^43–45.

Unlike the GSH uptake mechanisms characterized in other bacterial pathogens, which consist of ABC transporters, intact GSH import in Francisella is mediated by an MFS transporter. The consequences of this are unclear; however, one difference between the transporter types is the steepness of the concentration gradient of GSH that each can overcome. ABC transporters rely on ATP and can achieve transport across gradients much steeper than those achievable with MFS transporters, which can only overcome concentration gradients equivalent to those of the coupling ions⁴⁶. Interestingly, the bacteria in which ABC transporters for GSH have been identified, including S. pneumoniae and H. influenzae, reside in extracellular host-associated niches, where the concentration of GSH is much lower than the intracellular habitat of F. novicida^8,9,47. Thus, differences in the GSH concentration encountered in the different primary habitats these organisms colonize appears to correlate with the GSH uptake mechanism employed, in a manner consistent with the energetics of uptake by each route.

While our data clearly demonstrate that NgtA is capable of transporting GSH and suggests it does not play a role in Cys–Gly import, we have not defined the extent of its physiologically relevant substrates. To our knowledge, the only other MFS protein previously shown to transport GSH is Gex1 of yeast⁴⁸. While Gex1 can export GSH, its primary function appears to be related to cadmium detoxification via the extrusion of GSH-cadmium conjugates. This raises the possibility that NgtA could transport substrates beyond GSH. Several other members of the Pht family of transporters to which NgtA belongs facilitate uptake of amino acids that are limiting during intracellular growth of Francisella and Legionella^15–17. Candidate additional substrates for NgtA could include other γ-Glu amide bonded molecules, or a broader range of oligopeptides, such as those transported by members of the proton-dependent oligopeptide transporter class of MFS proteins⁴⁹.

During both growth in GSH medium and intramacrophage replication, we find that either Ggt or NgtA and CgaA-mediated degradation of GSH are sufficient to support growth of F. novicida, raising the question as to why the two pathways are maintained in parallel in many strains. GSH plays several important roles beyond serving as a source of nutrients, including redox buffering, combating oxidative stress and detoxifying metals and xenobiotics⁵⁰. Import of intact GSH via NgtA could thus provide a source of GSH under conditions where de novo biosynthesis may provide an insufficient supply of the intact molecule to counteract particular stresses. For the non-pathogenic Francisella species in which we find intact ngtA, one such condition may be encountered during replication within a protozoan host. These organisms employ many of the same mechanisms for killing phagocytosed bacteria as macrophages, including generation of a reactive oxygen burst³⁹. Interestingly, while Francisella have yet to be isolated from amoeba in a natural setting, laboratory studies employing model amoeba strains suggest that in these hosts, Francisella species replicate within the vacuole where the oxidative burst is delivered, rather than escaping to the cytosol as in mammalian cell infections^51–54. Additionally, virulent strains of Francisella can limit the oxidative burst within cells they infect, by mechanisms that are not yet fully characterized^55–57; it remains to be determined if these are conserved in other non-pathogenic species. We speculate that NgtA may provide a means of rapidly acquiring GSH for Francisella species that must contend with acute episodes of oxidative stress.

If the primary role of NgtA is to enable import of intact GSH for non-nutritional uses, the question then arises as to why Francisella species additionally encode an intracellular enzyme for GSH degradation, CgaA. In eukaryotic cells, constitutive degradation of GSH by intracellular enzymes contributes to GSH homeostasis. In yeast, this is mediated by the Dug complex, which shares the same predicted enzymatic function as CgaA, whereas in mammalian cells, constitutive turnover of GSH appears to be mediated by the γ-glutamyl cycotransferase enzyme ChaC2^23,58,59. The ChaC2 homologs that have been characterized to date exhibit a very slow rate of GSH turnover, which has been suggested to be important for preventing unchecked depletion of intracellular GSH levels⁵⁹. We similarly observed a low rate of GSH turnover by purified CgaA. While this observed rate of turnover may be the result of our in vitro assay conditions, it is consistent with CgaA playing a role in GSH homeostasis. Of note, Francisella spp. also encode a homolog of ChaC2, but this protein localizes to the periplasm. Additionally, strains lacking ChaC exhibit pleitropic phenotypes¹¹, and the corresponding gene was not a strong hit in our Tn-seq screen for genes important in GSH utilization, suggesting its role in GSH catabolism is likely a minor part of its overall function in Francisella.

A previously missing component of the GSH utilization pathway in Francisella is the means by which the tripeptide crosses the outer membrane. Here, we provide evidence that FupA provides this function by acting as a porin. Our findings challenge the prior assertion that FupA serves as a high affinity ferrous iron transporter³². Upon reexamination, two pieces of published data support our conclusion that FupA functions as a general porin: i) unlike typical high affinity transport mechanisms, FupA expression is not induced by limiting iron, and ii) growth of F. tularensis ΔfupA is reduced in minimal media regardless of the concentration or type of iron supplied^30,32. Additionally, we note that in other Gram-negative species, porins related to OmpC or OmpF, which are absent in Francisella spp., allow passive entry of Fe²⁺ that is then imported across the inner membrane by high affinity transporters⁶⁰. Together with our discovery of the NgtA and CgaA-mediated pathway for GSH uptake and degradation, our identification of the role of FupA in GSH import allows us to construct a substantially revised model for GSH catabolism in Fransicella that highlights the central importance of this molecule for this diverse group of organisms.

STAR METHODS

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Joseph Mougous (mougous@uw.edu).

Materials Availability

Plasmids and bacterial strains generated in this study are available upon request from the Lead Contact.

Data and Code Availability

• Sequence data associated with this study has been deposited to the NCBI Sequence Read Archive: PRJNA967744.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Bacterial strains and growth conditions

Bacterial strains used in this study include Francisella tularensis subspecies novicida U112 (F. novicida) and F. tularensis subsp. novicida MFN245 (F. novicida MFN245, both are gifts from Colin Manoil, University of Washington, Seattle, WA), F. tularensis subsp. holarctica LVS (F. tularensis LVS, provided by Karen Elkins, Food and Drug Administration, Rockville, MD), Escherichia coli strain DH5α (E. coli DH5α, Thermo Fisher Scientific), E. coli strain BL21 (DE3) (E. coli BL21, EMD Millipore). F. novicida strains were routinely grown aerobically at 37°C in tryptic soy broth or agar supplemented with 0.1% (w/v) cysteine (TSBC or TSAC). F. tularensis LVS was grown aerobically at 37 °C in either liquid Mueller-Hinton broth (Difco) supplemented with glucose (0.1%), ferric pyrophosphate (0.025%), and Isovitalex (2%) (MHB) or on cystine heart agar (Difco) supplemented with 1% hemoglobin (CHAH). For selection, antibiotics were used at the following concentrations: kanamycin at 5 μg/mL (LVS), 15 μg/mL (U112) or 50 μg/mL (E. coli), carbenicillin at 150 μg/mL, and hygromycin at 200 μg/mL. F. novicida strains were stored at −80°C in TSBC supplemented with 20% (v/v) glycerol. E. coli strains were stored at −80°C in LB supplemented with 15% (v/v) glycerol.

Murine bone marrow-derived macrophage generation

Murine bone marrow-derived macrophages (BMMs) were differentiated from bone marrow of female, 6–12 -weeks-old C57BL/6J mice (Jackson Laboratory) for 5 days in non-tissue culture-treated Petri dishes at 37°C under 10% CO2 in Dulbecco’s Modified Eagle’s Medium, containing 1g/L glucose, L-glutamine and sodium pyruvate (DMEM) supplemented with 10 % fetal bovine serum (FBS) and 20% L-929 mouse-fibroblast conditioned medium (L-CSF). 5 days post-plating, non-adherent cells were washed out with ice-cold phosphate buffered saline (PBS) and the differentiated BMMS were incubated for 10 min in ice-cold cation-free PBS (Corning) supplemented with 1 g/L glucose, detached by pipetting and harvested by centrifugation for 7 min at 200xg/ 25°C. Pelleted cells were resuspended in BMM complete medium (DMEM, 10% FBS, 10% L-CSF) and plated at a density of 5×10⁴ cells/ well in 24-well, tissue culture-treated plates followed by incubation for 48 h at 37°C under 10% CO2 with replenishment of BMM complete medium at 24 hrs post-plating.

Mice

C57BL/6J mice used in this study were purchased from Jackson Labs. Mice were maintained under SPF conditions ensured through the Rodent Health Monitoring Program overseen by the Department of Comparative Medicine at the University of Washington. All experiments involving mice were performed in compliance with guidelines set by the American Association for Laboratory Animal Science (AALAS) and were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Washington.

METHOD DETAILS

Strain and plasmid construction

Deletion mutations, in cis gene complementation strains , and in trans cgaA complementation strains of F. novicida were generated via allelic exchange as described previously⁶¹. Briefly, sequences containing ~1000 bp flanking the site of deletion or the insertion of the complementation allele (the intergenic site between FTN_0485 and FTN_0486 was used for cgaA) were amplified by PCR and cloned into the BamHI and PstI sites of the vector pEX18-pheS-km using Gibson assembly⁴⁴. . Naturally competent F. novicida was prepared by back-diluting overnight cultures 1:100 in 2 mL TSBC, growing for 3 hrs at 37°C with shaking, harvesting by centrifugation, and resuspending in 1 mL Francisella transformation buffer (per liter; L-arginine, 0.4 g; L-aspartic acid 0.4 g; L-histidine, 0.2 g, DL-methionine, 0.4 g; spermine phosphate, 0.04 g; sodium chloride, 15.8 g; calcium chloride, 2.94 g; tris(hydroxymethyl) aminomethane 6.05 g)⁶¹. Approximately 1 μg of pEX18-pheS-km-based deletion or complementation plasmid was added to freshly prepared competent cells. Bacterial suspensions were then incubated at 37°C with shaking for 30 min, followed by addition of 2 mL TSBC and an additional 3 hrs of incubation. Transformants were selected by plating on TSAC with kanamycin. The resulting merodiploids were grown overnight in non-selective TSBC, diluted 1:100 into Chamberlain’s defined medium (CDM)⁶² containing 0.1% p-chlorophenylalanine (w/v) and allowed to grow to stationary phase. Cultures were then streaked onto TSAC, colonies were patched onto TSAC with and without kanamycin to test for kanamycin sensitivity, and kanamycin sensitive colonies were screened for mutations by colony PCR.

The fupA complementation strain (ΔfupA Tn7:Pbfr-fupA) was constructed using the mini-Tn7 system⁶³. Briefly, the fupA gene was amplified from F. novicida by PCR and cloned using Gibson assembly into the pMP749 plasmid along with the sequence encoding the bacerioferritin promoter (Pbfr) for high constitutive expression^63,64. Using natural transformation as described above, the resulting plasmid, pMP749-Pbfr -fupA, was transformed into plasmid compatible strain F. novicida MFN245 carrying the helper plasmid encoding the transposase for Tn7 integration, pMP720. Tn7 integrants were selected on kanamycin, and colonies were screened for the presence of the inserted transposon at the glmS locus using PCR. To transfer the Tn7:Pbfr-fupA insertion from F. novicida MFN245 to F. novicida U112, genomic DNA was prepared from Tn7:Pbfr-fupA MFN245 strains and 10 ng was used to transform competent U112, prepared as described above. For expressing NgtA^NOV in F. tularensis LVS, the gene was amplified from F. novicida using PCR and cloned into the expression plasmid pF behind the constitutive groEL promoter ⁶⁵. Empty pF plasmid or pF-ngtA^NOV were electroporated into wild-type or Δggt F. tularensis LVS as previously described⁶⁶, and transformants were selected by plating on CHAH with kanamycin.

For constructing protein expression plasmids pET-28b(+)-6xHis-CgaA or pET-28b(+)-6xHis-CgaA^C97A, cgaA and cgaA^C97A were amplified by PCR and cloned into the NdeI and BamHI sites of the vector pET-28b(+) using Gibson assembly

Transposon mutant library generation

Transposon mutant libraries containing 100,000 to 300,000 Mariner transposon insertions in F. novicida wild-type and Δggt were constructed using delivery plasmid pKL91¹¹. The plasmid was delivered via natural transformation as described above, cells were allowed to recover 2 hrs and then plated on TSAC with kanamycin. Plates were incubated for 20 hrs at 37°C and the resulting kanamycin-resistant colonies were scraped and resuspended in CDM broth lacking a cysteine source (CDM-Cys). Each library was washed 4x times with CDM-Cys broth prior to freezing of aliquots containing ~10⁶ CFU each in CDM-Cys with 20% (v/v) glycerol at −80°C.

Tn-seq screen

For each genetic background, two aliquots of the respective transposon libraries were thawed and used as the inocula for 50 mL CDM-Cys media. Cultures were placed at 37°C with shaking for 2 hrs prior to addition of 100 μM cysteine or GSH, followed by incubation for 20 hrs at 37°C with shaking. Cells were then collected via centrifugation and genomic DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen). Sequencing libraries were generated essentially as described⁶⁷. In brief, 3 μg DNA from each sample was sheared to ~300 bp on a Covaris LE220 Focused-Ultrasonicator followed by DNA-end repair, terminal C-tailing and of amplification of the transposon-genome junctions by two rounds of PCR. The first round employed the tranposon-specific primer Fn_TnSeq_1_F and olj376, and second round employed the transposon-specific primer Fn_TnSeq_2_F and distal primer Fn-TnSeq-R2-N701 or Fn-TnSeq-R2-N702 to add unique de-multiplexing barcodes per sample. The libraries were pooled and sequenced using custom sequencing primer Fn_TnSeq-CustomSeq and Read1_SEQ primer by single-end 150 bp sequencing with a single index read on an Illumina MiSeq at 11 pM density with 15% PhiX spike-in.

Tn-seq data processing

Custom scripts^67,68 (https://github.com/lg9/Tn-seq) were used to process the Illumina sequencing reads and map sites of transposon insertion. First, reads from each sample were filtered for those displaying transposon end sequence (the sequencing primer was designed to anneal five bases from the end of the transposon). The filtered reads were mapped to the genome after removing the transposon end sequence. Reads per unique mapping position and orientation were tallied and read counts per gene were calculated by summing reads from all unique sites within each gene’s ORF except those within the 5’ 5% and 3’ 10% (insertions at gene termini may not be fully inactivating). Gene counts per sample were normalized based on a comparison between all samples of the median reads per gene per gene length for genes with insertions in all samples, as described⁶⁸.

Bacterial growth assays

To examine the proliferation of F. novicida on different cysteine sources, strains were first grown overnight in CDM with 57 μM cysteine at 37°C with shaking. Cells were washed three times and resuspended in CDM lacking cysteine (CDM-cys) followed by cysteine starvation for 2 hrs at 37°C with shaking. Cultures were then diluted to an OD₆₀₀= 0.01 in CDM-Cys supplemented with either GSH, cysteine, or Cys-Gly at 100 μM. Cultures were transferred into a 96-well plate and incubated in a plate reader at 37°C. OD₆₀₀ measurements were taken following a 2 s shake every 10 min. Final growth yields reported represent the OD₆₀₀ obtained after 36 hrs of growth. For F. tularensis LVS, strains containing either empty pF plasmid or pF-ngtA^NOV were first grown overnight in CDM with 1 μM cysteine and then washed and back-diluted to starting OD₆₀₀ = 0.1 in CDM-Cys supplemented with GSH at 100 μM concentration. Cultures were then incubated at 37°C with shaking, and the growth yield was determined by measuring OD₆₀₀ at 16 hrs.

GSH uptake assays

The indicated strains of F. novicida were grown to mid-log phase at 37°C in CDM-Cys liquid medium supplemented with 57 μM cysteine. Cells were spun down, washed three times in uptake buffer (25 mM Tris pH 7.5, 150 mM NaCl, 5 mM glucose), then concentrated 20-fold. The OD₆₀₀ was measured and normalized to OD₆₀₀ = 10. Reactions were established containing 20 μL cells, 5 μL 100 μM GSH, 0.5 μCi ³H-GSH ([Glycine-2-³H]-Glutathione,>97%, 50μCi, PerkinElmer), and 25 μL uptake buffer and incubated for 45 min at 37°C followed by quenching with 1 mL ice-cold uptake buffer. Cells were then pelleted by centrifugation, washed three times in 1 mL cold uptake buffer, then resuspended in 50 μL uptake buffer. Samples were then added to scintillation cocktail (National Diagnostics, Ecoscint Ultra) and counts were measured over 1 minute on a scintillation counter (Beckman, LS6500).

Protein expression and purification

For protein expression, overnight cultures of E. coli BL21 carrying pET-28b(+)-6xHis-CgaA or pET-28b(+)-6xHis-CgaA^C97A were back diluted 1:500 in 2xYT broth and grown at 37°C until the OD₆₀₀ reached 0.4 ~ 0.6. Protein expression was then induced by the addition of IPTG (IPTG), and cultures were then incubated with shaking at 18°C for 18 hrs. Following this incubation, cells were collected by centrifugation and resuspended in buffer containing 500 mM NaCl, 50 mM Tris-HCl pH 7.5, 10% glycerol, 5 mM imidazole, 0.5 mg/mL lysosome, 1 mM AEBSF, 10 mM leupeptin, 1 mM pepstatin, 1 mU benzonase, and 5 mM β-mercaptoethanol (BME). Cells were disrupted by sonication and cellular debris was removed by centrifugation at 45,000 x g for 40 min. Lysates were run over a 1 mL HisTrap HP column on an AKTA FPLC purification system to purify the His-tagged proteins. The bound proteins were eluted using a linear imidazole gradient from 5 mM to 500 mM. The purity of each protein sample was assessed by SDS- PAGE and Coomassie brilliant blue staining, and fractions with high purity were concentrated using a 10 kDa cutoff filter. Protein samples were further purified by running over a HiLoad™ 16/600 Superdex™ 200 pg column equilibrated in sizing buffer (300 mM NaCl, 50 mM Tris-HCl pH 7.5, and 1 mM TCEP). Again, the purity of each fraction was assessed by SDS-PAGE and Coomassie brilliant blue staining. Fractions of the highest purity were pooled, concentrated, and utilized in biochemical assays.

In vitro glutaminase assays

Glutamine amidotransferase activity of purified CgaA and CgaA^C97A toward different substrates was assayed in vitro using a glutamate detection kit. 1 μM purified protein was mixed with GSH, glutamine or buffer alone (300 mM NaCl, 50 mM Tris-Cl (pH 8.5)) in 50 μl reactions mixes and incubated 1 hr at at 37 °C. The reactions were stopped by heating at 95 °C for 5 min to inactivate the enzyme. After inactivation, 45 μL of reaction mixtures were added to 100 μL of glutamate detection reaction mix (Abcam Glutamate Assay Kit, ab83389) in a 96-well plate. The reaction mixture was incubated for 5 min at RT followed by 5 min at 37°C. The color change is proportional to the glutamate generated and was measured at A450 in a plate reader. Reported glutamate concentrations were calculated by subtracting background absorbance readings from a no enzyme control.

H₂O₂ sensitivity assays

To monitor H₂O₂ tolerance levels, strains of F. novicida were grown at 37°C in CDM medium to mid-log (OD₆₀₀= 0.4–0.6). Cultures were diluted to an OD₆₀₀= 0.1 and cell viability was assayed via plating for CFU enumeration. 1.5 mM of H₂O₂ was then added and the cultures were placed at 37°C with shaking. After 30 mins and 60 min, samples were collected for CFU enumeration. Survival rates were calculated by comparing the CFU numbers pre and post-H₂O₂ exposure.

Macrophage infection assays

48 hrs post-plating, BMMs were infected with mid-log phase F. novicida at a multiplicity of infection (MOI) of 1 in pre-chilled BMM complete medium. Bacterial uptake was synchronized by centrifugation for 10 min at 400xg/ 4°C after which the plates were immediately placed in a water tray pre-warmed to 37°C and incubated for 30 min at 37°C under 10% CO2. Following incubation, extracellular bacteria were removed by 4 washes with plain DMEM medium pre-warmed to 37°C, the complete BMM medium was replenished, and the plates were placed back at 37°C under 10% CO₂. At 2 hrs and 24 hrs post-infection (p.i.) the BMMs were rinsed 3 times with sterile PBS and lysed in PBS/0.1% sodium deoxycholate (Sigma), followed by serial dilution in sterile PBS and plating on TSAC plates for CFU enumeration. Bacterial growth at 24 h p.i. was normalized to the CFU counts obtained at 2 h p.i.

F. novicida inoculum preparation and intranasal infection

F. novicida inoculum was prepared as described previously⁶¹. Briefly, 3 mL TSBC was inoculated with each F. novicida strain and incubated aerobically for 18 h at 37°C with shaking. After overnight growth, cultures were adjusted to OD₆₀₀ =1 in TSBC, diluted 1:1 with 40% glycerol in TSBC (20% v/v final glycerol concentration), aliquoted and stored at −80°C. The post-freeze titer of each stock was determined by culturing on TSAC. Just prior to infection, an aliquot of each strain was quickly thawed at 37°C and diluted in sterile 1X PBS to ~100 CFU in 30 μL (~3.3 × 10³ CFU/mL).

Mice were infected with indicated F. novicida strains by intranasal instillation (30 μL total) under light isoflurane anesthesia. Mice were weighed just prior to and 48 hrs post infection. After 48 hrs mice were euthanized with CO₂. The lungs and spleens were harvested in 5 mL lysis buffer (0.1% IGEPAL, in 1x PBS sterile filtered) and homogenized using a Tissue Teaeror Homogenizer (BioSpec Products, Cat# 985370–14). Organ homogenate was serially diluted into 1X PBS and dilutions plated on TSAC. Plates were incubated at 37°C overnight aerobically. Colonies were counted and CFU per organ calculated.

Protein expression level analysis

To analyze the expression of CgaA and CgaA^C97A by western blot, F. novicida Δggt, Δggt cgaA-VSV-G and Δggt cgaA^C97A-VSVG were grown overnight in CDM with 100 μM cysteine at 37°C with shaking and the equivalent of 1 ml culture at OD₆₀₀ was collected for each strain and centrifuged to pellet cells. Cell pellets were resuspended in 50 μL 1x Laemmli buffer⁶⁹, boiled at 95°C for 5 min, and proteins in 5 μL of each sample were separated by SDS-PAGE. Proteins were then transferred to nitrocellulose membranes, and membranes were blocked in TBST (10 mM Tris-HCl pH 7.5, 150 mM NaCl₂, and 0.05% v/v Tween-20) with 5% (w/v) bovine serum albumin (BSA) for 1 hr at room temperature, followed by incubation with primary antibodies (α-VSV-G with 1:5,000 dilution or α-SodB with 1:20,000 dilution) diluted in TBST with 5% (w/v) BSA for 1 hr at room temperature. Blots were then washed 3x times with TBST, followed by incubation with secondary antibody (Goat α-Rabbit, HRP conjugated, 1:5,000 dilution) diluted in TBST for 1 hr at room temperature. Finally, blots were washed 3x times with TBST, developed using ECL substrate (BIO-RAD), and visualized using the iBright FL1500 Imaging System (Thermo Fisher).

To analyze the expression of NgtA and NgtA variants, F. novicida strains Δggt, Δggt ngtA-3xFLAG, Δggt ngtA^ΔIGS-3xFLAG, Δggt ngtA^SCHU-3xFLAG, and Δggt ngtA^LVS-3xFLAG were grown to OD ~ 2 in CDM with 100 μM cysteine at 37°C with shaking, pelleted, and resuspended in lysis buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 2% (v/v) Glycerol, 1% (v/v) Triton X-100, 1 mM β-mercaptoethanol, 25 U/mL benzonase, 0.25 mg/mL Lysozyme, and protease inhibitor (cOmplete Protease Inhibitor Cocktail, EDTA-Free, Sigma)). Cells were lysed by sonication (5 rounds, 15 s each) and the resulting cell lysates were clarified by centrifugation at 17,000 rcf for 30 min at 4°C. To concentrate low-abundant NgtA proteins, clarified supernatants were normalized by BCA assay (Thermo Fisher) and equal amounts of protein (5 mg total each sample) were incubated with 40 μL anti-FLAG Affinity Resin (Thermo Scientific) for 4 hrs at 4°C. After incubation, resins were washed 3x times with 1 mL wash buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5, 2% (v/v) Glycerol, 0.1% (v/v) Triton X-100, 1 mM β-mercaptoethanol) and proteins were eluted by incubating in 2x Laemmli buffer for 30 min at room temperature. Eluted proteins were subjected to western blot analysis as described above with α-FLAG primary antibody (1:800 dilution) and Goat α-Rabbit HRP conjugated secondary antibody (1:5,000 dilution).

NgtA sequence and phylogenetic analysis

To generate a phylogeny, NgtA homologs were identified by collecting the top 5,000 hits from Psi-BLAST, then aligned using the Clustal Omega plug in of Geneious Prime (Dotmatics). Positions with gaps present in at least 30% of sequences were masked in the alignment, and then a neighbor-joining phylogeny was constructed using the Geneious Tree Builder. This phylogeny included both Pht family members and a number of clades of related MFS family proteins from other subfamilies. Non-Pht family clades were eliminated by performing additional BLASTp searches with representatives from each clade that did not contain a previously characterized Pht family member; clades were eliminated when these sequences had higher percent identity matches with other MFS transporter families than with the closest Pht family member. This yielded a set of 1,043 sequences that were re-aligned and masked as described above, and used to construct a new neighbor joining phylogeny.

Inactivating mutations in ngtA coding sequences were identified by performing a tBLASTn search with NgtA, limited to the Thiotrichales. All protein sequences obtained were filtered to remove those sharing <50% identity with NgtA of F. novicida, as these were found to represent other Pht family members. Remaining protein sequences were then aligned. In cases where a premature stop codon had been introduced into the coding sequence of ngtA, our tBLASTn search retrieved multiple hits, which manifested as truncated sequences in the protein sequence alignment. For F. endociliophora, the absence of ngtA was confirmed by performing a BLASTp search with NgtA against the complete genome, and by examining the conserved genomic location where ngtA is encoded in other strains.

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical significance in bacterial growth assays, GSH uptake assays, in vitro glutaminase assays, and H₂O₂ sensitivity assays was assessed by unpaired two-tailed student’s t-test between relevant samples. Statistical significance in macrophage infection assays and mouse infection assays was assessed by one-way ANOVA with Dunnett’s multiple comparison test comparing mutant strains to wild-type. Details of statistical significance is provided in the figure legends.

Supplementary Material

3

Table S1. Normalized transposon insertion frequency from a library constructed in wild-type F. novicida grown in GSH or cysteine media. Related to Figure 1.

4

Table S2. Normalized transposon insertion frequency from a library constructed in F. novicida Δggt grown in GSH or cysteine media. Related to Figure 1.

5

Table S4. Oligonucleotides and linear DNA fragments used in this study. Related to the STAR Methods section.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-VSV-G	Sigma-Aldrich	Cat#V4888; RRID:AB_261872
Rabbit polyclonal anti-FLAG	Sigma-Aldrich	Cat#F7425; RRID:AB_439687
Rabbit anti-SodB	Gift from Dr. Karsten Hazlett	N/A
Goat anti-Rabit HRP conjugated	Sigma-Aldrich	Cat#A6154; RRID:AB_258284
Anti-FLAG affinity resin	Thermo Scientific	Cat#A36803
Bacterial and virus strains
Francisella novicida U112	Gift from Dr. Colin Manoil	N/A
F. novicida Δggt	Ramsey et al.11	N/A
F. novicida ΔchaC	Ramsey et al.11	N/A
F. novicida ΔngtA	This paper	N/A
F. novicida Δggt ΔngtA	This paper	N/A
F. novicida Δggt ΔngtA::ngtA	This paper	N/A
F. novicida ΔdptA	This paper	N/A
F. novicida ΔdptA ΔngtA	This paper	N/A
F. novicida Δggt ΔchaC	Ramsey et al.11	N/A
F. novicida Δggt ΔchaC ΔngtA	This paper	N/A
F. novicida ΔcgaA	This paper	N/A
F. novicida Δggt ΔcgaA	This paper	N/A
F. novicida cgaA C97A	This paper	N/A
F. novicida Δggt cgaAC97A	This paper	N/A
F. novicida Δggt ΔcgaA + Tn7:Pnat-cgaA	This paper	N/A
F. novicida Δggt cgaA-VSV-G	This paper	N/A
F. novicida Δggt cgaAC97A-VSV-G	This paper	N/A
F. novicida Δggt ΔFTN_0433	This paper	N/A
F. novicida ΔdotU	Eshraghi et al.44	N/A
F. novicida Δggt ΔngtA::ngtALVS	This paper	N/A
F. novicida Δggt ΔngtA::ngtASCHU	This paper	N/A
F. novicida ngtA ΔIGS	This paper	N/A
F. novicida Δggt ngtAΔIGS	This paper	N/A
F. novicida Δggt ngtA-3xFLAG	This paper	N/A
F. novicida Δggt ngtAΔIGS-3xFLAG	This paper	N/A
F. novicida Δggt ΔngtA::ngtASCHU-3xFLAG	This paper	N/A
F. novicida Δggt ΔngtA::ngtALVS-3xFLAG	This paper	N/A
F. novicida ΔfupA	This paper	N/A
F. novicida ΔfupA + Tn7:Pbfr-fupA	This paper	N/A
F. novicida ΔfupA Δggt	This paper	N/A
F. novicida MFN245	Gift from Dr. Colin Manoil	N/A
F. novicida MFN245 Tn7:Pbfr-fupA	This paper	N/A
F. tularensis LVS	Gift from Dr. Karen Elkins	N/A
F. tularensis LVS Δggt	Ramsey et al.11
Escherichia coli DH5α	Thermo Fisher Scientific	Cat#18258012
E. coli BL21 (DE3)	EMD Millipore	Cat#69450
Chemicals, peptides, and recombinant proteins
[Glycine-2-3H]-Glutathione	PerkinElmer	Cat#NET282050UC
Ecoscint Ultra	National Diagnostics	Cat# LS-270
Hydrogen peroxide solution	Sigma	Cat#H1009–5ML
Dulbecco’s Modified Eagle’s Medium	Corning	Cat#10–014-CM
Fetal Bovine Serum	Atlanta Biologicals	Cat#S10350H
L-929 mouse-fibroblast conditioned medium	This paper
Phosphate buffered saline	Corning	Cat#21–030-CV
Cation-free PBS	Corning	Cat#21–040-CV
Phosphate buffered saline	Life Technologies	Cat#10010049
Sodium deoxycholate	Sigma	Cat#D6750
IGEPAL	MP Biomedicals	Cat#198596
Critical commercial assays
Glutamate Assay Kit	Abcam	Cat#ab83389
DNeasy Blood & Tissue Kit	Qiagen	Cat#69506
Deposited data
Transposon insertion sequencing data	This paper	NCBI Sequence Read Archive: PRJNA967744
Experimental models: Organisms/strains
C57BL/6J mice	The Jackson Laboratory	Cat#000664
Murine bone marrow-derived macrophages (BMMs) differentiated from bone marrow of female, 6–12 -weeksold C57BL/6J mice	The Jackson Laboratory	Cat#000664
Oligonucleotides
Primers are listed in Table S4
Recombinant DNA
pEX18-pheS-km	Eshraghi et al.44	N/A
pEX18-pheS-km-Δggt	Ramsey et al.11	N/A
pEX18-pheS-km-ΔchaC	Ramsey et al.11	N/A
pEX18-pheS-km-ΔngtA	This paper	N/A
pEX18-pheS-km-ΔngtA::ngtA	This paper	N/A
pEX18-pheS-km-ΔdptA	This paper	N/A
pEX18-pheS-km-ΔcgaA	This paper	N/A
pEX18-pheS-km-cgaAC97A	This paper	N/A
pEX18-pheS-km-Tn7:Pnat-cgaA	This paper	N/A
pEX18-pheS-km-cgaA-VSV-G	This paper	N/A
pEX18-pheS-km-cgaAC97A-VSV-G	This paper	N/A
pEX18-pheS-km-ΔFTN_0433	This paper	N/A
pEX18-pheS-km-ΔngtA::ngtALVS	This paper	N/A
pEX18-pheS-km-ΔngtA::ngtASCHU	This paper	N/A
pEX18-pheS-km-ngtAΔIGS	This paper	N/A
pEX18-pheS-km-ngtA-3xFLAG	This paper	N/A
pEX18-pheS-km-ΔngtA::ngtALVS-3xFLAG	This paper	N/A
pEX18-pheS-km-ΔngtA::ngtASCHU-3xFLAG	This paper	N/A
pEX18-pheS-km-ngtAΔIGS-3xFLAG	This paper	N/A
pEX18-pheS-km-ΔfupA	This paper	N/A
pMP720	LoVullo et al.63	N/A
pMP749	LoVullo et al.63	N/A
pMP749-Pbfr-fupA	This paper	N/A
pF	Charity et al.65	N/A
pF-ngtANOV	This paper	N/A
pET-28b(+)-6xHis-CgaA	This paper	N/A
pET-28b(+)-6xHis-CgaAC97A	This paper	N/A
pKL91	Ramsey et al.11	N/A
Software and algorithms
Geneious Prime 2023.1.2	Geneious, Software, Newark, New Jersey, USA	https://www.geneious.com; RRID:SCR_010519
Prism 9 for macOS	GraphPad, Software, La Jolla, California, USA	https://www.graphpad.com; RRID:SCR_022798
Adobe Illustrator 27.3.1	Adobe Systems Incorporated, San Jose, California, USA	https://www.adobe.com/products/illustrator; RRID:SCR_010279
Chimera version 1.16	UCSF, Software, San Francisco, California, USA	www.rbvi.ucsf.edu/chimera; RRID:SCR_004097
Other
1 mL HisTrap HP column	Cytiva	Cat#17–5247-01
HiLoad 16/600 Superdex 200 pg	Cytiva	Cat#28989335

Highlights.

• Tn-seq identifies a previously unrecognized GSH utilization pathway in F. novicida.

• A major facilitator transporter pseudogenized in pathogenic Francisella imports GSH.

• Francisella uses a glutamine amidotransferase to degrade cytoplasmic GSH.

• A single porin mediates the bulk of outer membrane GSH transport in Francisella.