Dictyostelium discoideum cells (Ax2) were cultured adherently in axenic HL5-C medium supplemented with 100 μg/ml Pen/Strep (Hagedorn and Soldati, 2007). Prof. Salvatore Bozzaro (University of Torino, Italy) kindly provided confirmed knockout cell lines of Nramp1, NrampB and Nramp1/B (Peracino et al., 2006, 2013). Knockout cell lines had been generated by the authors (Peracino et al., 2006, 2013) using homologous recombination to replace the respective gene with the respective coding sequence disrupted by a blasticidin resistance cassette. Ax2 cells expressing Nramp1-GFP and NrampB-GFP were obtained by transformation with the plasmids pDEX-Nramp1::GFPC and pDEX-NrampB::GFPC, respectively [provided by Prof. Bozzaro, (Peracino et al., 2006, 2013)] and grown with 10–30 μg/ml G418. Green (pKK289Km:gfp) and red (pKK289Km:mCherry) fluorescent F.n.n. wild-type and ΔiglC bacteria were cultivated in Eugon Broth (EB) shaking culture (100 rpm) at 22°C supplemented with 2 mM FeCl₃ and 15 μg/ml Kanamycin (Lampe et al., 2016). The in-frame knockout of iglC was achieved via a suicide plasmid (pDMK2) containing the fused flanking regions (~1,100 bp each) of iglC and a kanamycin resistance according to Lampe et al. (2016) and similar to other Francisella species (Lindgren et al., 2009). Results were verified by sequencing and qPCR confirmed no effects on neighboring genes iglB and iglD (Lampe et al., 2016). Iron depletion during F.n.n. cultivation was performed at 22°C using 100 μM 2,2′-dipyridyl (DP, Sigma-Aldrich) in EB as described by Brudal et al. (2013). In short, bacteria were incubated for 48 h until OD 1.5–2, diluted to OD 0.1 and grown for 24 h without FeCl₃. Subsequently, remaining iron was washed off with PBS (2x) and EB (+DP), the culture was diluted to OD 0.5 and grown for 24 h until OD 1.

The infection of Dictyostelium cells with F.n.n. was performed as described (Lampe et al., 2016). F.n.n. bacteria were grown at 22°C in Eugon Broth supplemented with 2 mM FeCl3 and 15 μg/ml Kanamycin until reaching the exponential growth phase. At least 24 h prior to infection, Dictyostelium cells were grown at 22°C in HL5-c medium without antibiotics in 25 or 75 cm² cell culture flasks. On infection day, F.n.n. equivalent to MOI 60 were centrifuged on a Dictyostelium monolayer of 80–100% confluency at 100 × g and 21°C for 30 min. After 5 min of additional time for phagocytosis, free bacteria were washed off with HL5-C without antibiotics and checked visually for remaining extracellular bacteria. Infected cells were seeded in HL5-C without antibiotics in 10 cm culture dishes at respective cell numbers (6 h post infection (hpi): 1 × 10⁷, 24 hpi: 4.5 × 10⁶, 48 hpi: 2 × 10⁶) and cultivated adherently at 22°C. At each timepoint, cells were resuspended, counted with a CASY Cell Counter and ~1 × 10⁶ cell were fixed with 4% paraformaldehyde (PFA) in Sørensen buffer for subsequent antibody staining and analysis via fluorescent microscopy or flow cytometry. For flow cytometry analysis, at least 4 × 10⁵ cells were quantified per sample using a FACSCalibur flow cytometer (Becton Dickinson) and analyzed according to Lampe et al. (2016) via FlowJo software v10. In short, dead cells and extracellular F.n.n. were excluded from analysis by gating only on living Dictyostelium cells using FSC/SSC. Infected and non-infected Dictyostelium populations were separated using SSC plotted as a function of green fluorescence (FL-1). The green fluorescence of intracellular F.n.n. was expressed as relative fluorescence units (RFU) per volume unit of cell culture (RFU/ml) as followed: the infection rate was multiplied with Dictyostelium cell number/ml to obtain infected cells/ml. This value was multiplied with the mean FL-1 value of F.n.n. only, which was caluculated as mean FL-1 of infected cells - mean FL-1 of non-infected cells. All samples were analyzed with the same fluorescence detector settings. The resulting RFU/ml correlate with F.n.n. genome equivalents/ml as shown by Lampe et al. (2016).

Immunolabeling of PFA-fixed samples for microscopy (Hagedorn et al., 2006) and flow cytometry (Lampe et al., 2016) was performed as described. Monoclonal antibody against the putative copper transporter and endosomal marker p80 (Ravanel et al., 2001) was obtained from P. Cosson (University of Geneva, Switzerland) and used 1:10 in blocking solution (0.5% FCS, 0.1% Triton in PBS). Rabbit polyclonal GFP-antibody (MBL International) was used at a dilution of 1:1,000 for microscopy and flow cytometry. Anti-rabbit and anti-mouse secondary antibodies from goat (Invitrogen) were coupled to AlexaFluor 488 and 568, respectively, and diluted 1:1,000. For quantitative microscopic analysis, a minimum of 100 bacteria was analyzed at each time-point. For statistical analysis of multiple groups to the wt control group over time, a repeated measures one-way ANOVA followed by Dunnett's post-hoc test was performed for each timepoint using Prism 7.0c software.

For live cell imaging, infected cells were resuspended in filtrated HL5-C and seeded in μ-Slide 8 well chambers (Ibidi) shortly before imaging. Microscopic analysis was performed at 21°C using a Zeiss LSM5 Live confocal microscope with a 100x Europlan apochromat oil immersion objective (N.A. 1.4) and a Diode-Laser 488, as well as a DPSS-Laser 561 [single track mode, 1 Airy unit, dual-band filter (500–545 band pass, 575 long pass)]. Image brightness and contrast were adjusted with ImageJ (Schneider et al., 2012) to whole images. A minimum of 100 bacteria were quantified at each timepoint except for F.n.n. ΔiglC at 2 hpi (50 bacteria). Imaging of PFA-fixed samples was performed with an Olympus IX81 confocal microscope equipped with an Olympus 100x UPlanSApo oil immersion objective (N.A. 1.4).

F.n.n. infected Dictyostelium cells or cultured bacteria were lysed with Trizol (Life Technologies). RNA extraction was performed using the PureLink RNA Mini Kit (Life Technologies) according to the manufacturer's instructions for Trizol extraction. For cDNA synthesis, the Maxima First Strand cDNA Synthesis Kit (Thermo Scientific) was applied according to the manufacturer's instructions. One microliter of 1:10 diluted cDNA was used for each sample as a template in a 10 μl reaction volume using LightCycler® 480 SYBR Green I Master mix (Roche) and a Rotor-Gene RG-3000 qPCR machine (Corbett Research). The thermal cycle conditions were as follows: 1 cycle at 95°C for 5 min; 40 cycles of amplification at 95°C for 30 s, 60°C for 30 s and 72°C for 30 s. A melting curve analysis was performed from 67–95°C with steps of 1°C. Samples of three infection experiments were run in duplicates. Gene-specific primer sequences of iron related genes (feoA, feoB, iucA/C) and reference genes (ftsZ, polA, fopA; Brudal et al., 2013) were designed with Primer3 software using contig data of F.n.n. (Sjodin et al., 2012) as a template and obtained from Brudal et al. (2013), respectively (Table S1). Results were analyzed using the ΔΔCt method and tested for statistical significance with a repeated measures one-way ANOVA followed by Dunnett's post test using the Prism 7.0c software.

The phagosomal proteolysis and pH of Dictostelium cell lines was quantified using fluorescently-labeled, 3 μm silica beads according to Sattler et al. (2013) with minor adaptations. In short, latex beads were labeled with 0.15 mg of either DQ Green BSA (proteolysis-sensitive) or FITC (pH-sensitive) and Alexa 568 succinimidyl ester as a reference dye. Fluorescently labeled beads were added at a bead/cell ratio of 2:1 to a 90–100% confluent monolayer of Dictyostelium cells in a 96 well cell culture microplate μClear/black (Greiner). After short centrifugation (10 s, 300 × g, 22°C), cells were left 1 min for phagocytosis. Free beads were washed off and the fluorescence emission of intracellular beads was measured every 2 min for 2 h at 535 and 590 nm using an Infinite F200 fluorescence reader (Tecan). Filter settings for excitation/emission were as followed: 485(20)/535(25) for DQ Green/FITC and 550(10)/590(20) for Alexa568. To calculate the intraphagosomal pH from the 535/590 nm emission ratio, a calibration curve was prepared for each experiment using reference pH buffers in a range of pH 3 to 7. For comparison of the acidification and proteolysis profiles, linear regression lines were calculated for the timeframes of interest (proteolysis: linear range; covering ~25 datapoints each, acidification: t = 0 min to the minimal turning point (mtp), reneutralisation: mtp to t = 120 min; Figure S4) and the slopes were tested for significance using a one-way ANOVA followed by Dunnett's multiple comparison post test using the Prism 7.0c software.

Dictyostelium cells infected with either F.n.n. wt and ΔiglC were detached at 6 hpi and 1 ml cell culture including exocytosed bacteria was centrifuged at 500 × g for 3 min to pellet cells but not bacteria. The supernatant was centrifuged at 5,000 × g for 5 min and the bacterial pellet was dissolved in 100 μl HL5-C medium. 20 μl of undiluted and 1:50 diluted bacterial suspension was dropped on a chocolate agar plate and incubated for 1 week.

As transmembrane transporters of divalent metal ions, Nramp1 and NrampB change the ion composition of the compartment on which they are located. Thus, Nramp transporters could directly modify the environment of vacuolar bacteria. To investigate if Nramp1 and NrampB localize to the F.n.n.-phagosome (FP) in Dictyostelium cells, we infected amoeba expressing Nramp-GFP-fusion proteins with F.n.n. wild-type (wt) bacteria expressing mCherry. In addition, to monitor whether Nramp recruitment was actively manipulated by virulent F.n.n., we compared Nramp1 association with wt bacteria as well an avirulent mutant strain lacking a major component of the pathogenicity island IglC (F.n.n. ΔiglC). The proportion of Nramp-GFP associated FPs was quantified via life cell imaging in the early phase of infection (1–6 hpi). Until 1 hpi, wt and ΔiglC bacteria share similar association rates with Nramp1 at the FP (Figures 1A–C). However, whilst the association of wt F.n.n. with the iron transporter drops at 2 hpi, the avirulent mutant remains in an Nramp1-positive compartment. However, at 4 hpi, most F.n.n. ΔiglC are killed and exocytosed (Figure S1; Lampe et al., 2016) and not enough bacteria were available for reliable quantification. In contrast, wt F.n.n. were not observed in NrampB-positive compartments at any time post infection (Figure 1D).

The localization of Nramp1 on the FP and the synergistic regulation of intracellular iron homeostasis by both Nramps (Peracino et al., 2013) led us to investigate the functional impact of Nramp1 and NrampB on F.n.n. infection. We infected Dictyostelium Δnramp1, ΔnrampB and Δnramp1/B cells with GFP-expressing F.n.n. wt bacteria and compared the course of infection to Dictyostelium wt cells. The bacterial load was then monitored until 48 hpi quantifying bacteria by both fluorescence microscopy and flow cytometry.

Quantification of individual bacteria per cell by fluorescence microscopy showed a similar uptake of bacteria by wt and nramp knockout Dictyostelium cells (Figures 2A–D). However, from 24 hpi, the bacterial load per Dictyostelium cell was higher in the nramp deletion mutants. At 48 hpi, nramp mutants showed higher overall infection rates (wt: 13.6 ± 3.8%, Δnramp1: 24.5 ± 1.9%, ΔnrampB: 40.7 ± 7.8%, Δnramp1/B: 29.6 ± 3.5%) and increased proportions of highly infected cells (more than 3 bacteria: wt: 20.4 ± 12.5%, Δnramp1: 51.1 ± 8.8%, ΔnrampB: 57.4 ± 6.7%, Δnramp1/B: 41.3 ± 9.9%). The bacterial burden with more than 3 bacteria per cell over time is significantly higher in the nramp mutants indicating more bacterial growth at the single cell level.

To apply another approach to quantify F.n.n. growth we used quantitative flow cytometry and measured infection rate and bulk fluorescence of intracellular bacteria per volume over 48 h. Overall, the flow cytometry analysis recapitulated our results obtained by fluorescence microscopy. The infection rates of the Dictyostelium nramp mutants (Δnramp1: 20.84 ± 5.8%, ΔnrampB: 30.3 ± 9.4%, Δnramp1/B: 22.1 ± 5.7%) were higher in comparison to wt cells (9.9 ± 3.4%) at 48 hpi (Figure 2E). In principal, Dictyostelium cell growth can have an impact on the infection rate. However, the growth rates of the different cell lines were comparable over 48 h and could therefore not account for differing infection rates (Figure S2). The quantification of bacterial fluorescence (relative fluorescence units/ml) also showed higher bacterial loads in the mutant cell lines compared to wt cells (Figure 2F). This suggests that Nramp1 and NrampB contribute to resistance of the amoeboid host cell against F.n.n. growth.

It seems most likely that Nramp1 and NrampB control bacterial replication by affecting the phagosomal stage of F.n.n. infection. Like F. tularensis, F.n.n. escapes the p80-positive phagosome in the late phagosomal stage in Dictyostelium (Lampe et al., 2016). Therefore, the endosomal protein p80, a putative copper transporter (Ravanel et al., 2001), was applied as a marker for intraphagosomal vs. cytosolic bacteria phagosomal membranes (Figures 3A,B).

At 1 hpi, more than 80% of F.n.n. were localized in phagosomes of wt and nramp knockout cells (Figure 3C). However, starting at 2 hpi, significantly less bacteria were located in p80-positive compartments in the nramp deletion mutants with a comparable total load of bacteria per cell between wild-type and mutant cells. This suggests that more bacteria translocate into the cytosol and reach the replicative, cytosolic phase in the nramp mutants, which results in a higher bacterial load per cell in the mutant strains at the late infection phase. Even though not statistically significant, in comparison to the ΔnrampB strain at 2 and 4 hpi, bacteria seem to escape more efficiently in the Δnramp1 cells. This transient delay of escape might highlight a rather indirect impact of NrampB, which is located at the contractile vacuole, in contrast to a direct impact of Nramp1 which is present at the F.n.n. vacuole.

In principle, intracellular bacteria respond to limiting iron concentrations in their environment by upregulating the transcription of genes encoding for iron transporters (Rodriguez et al., 2002; Deng et al., 2006; Ledala et al., 2010). F.n.n. possesses two iron accumulation systems: the Feo system for uptake of soluble Fe²⁺, represented by FeoA and FeoB, and the siderophor synthetase IucA/C for transport of insoluble Fe³⁺. Members of both uptake systems (feoA, iucA/C) showed highly increased mRNA levels after depletion of iron during in vitro growth of F.n.n. (Figure S3).

We monitored mRNA levels of the iron accumulation genes feoA, feoB and iucA/C during infection of wt cells and normalized it to cultured bacteria (dotted line; Figures 4A–C). At 6 hpi, we observed an induction of all three genes, when most of the bacteria are entering the cytosolic growth phase. At later timepoints, mRNA levels of the entire gene set were still elevated in comparison to in vitro cultured bacteria but gradually decreased until 48 hpi. These observations suggest that F.n.n. relies on the uptake of external iron during the cytosolic growth phase in Dictyostelium cells.

Peracino et al. (2013) observed that nramp mutants contain lower levels of intracellular, bioavailable iron. Accordingly, F.n.n. showed an increased induction of all iron related genes in the late infection phase of the nramp mutants compared to wt cells. Together, these results suggest that cytosolic F.n.n. experience an enhanced iron limitation in the absence of Nramp1 and NrampB, to which the bacteria respond by an upregulation of iron transporters.

Nramp1 is known to play a role in the maturation of phagosome-derived, pathogen-containing compartments during infection with bacteria such as Salmonella and Mycobacteria (Hackam et al., 1998; Govoni et al., 1999; Frehel et al., 2002). In Dictyostelium, phagosomal maturation is characterized by rapid lowering of the pH and delivery of proteolytic enzymes followed by the reneutralization of the compartment after 1–2 h via recycling of the V-ATPase from the phagosomal membrane (Clarke et al., 2002, 2010). To investigate an impact of Nramp1 and NrampB on phagosomal maturation in our system, we measured the phagosomal proteolysis and pH in wt and nramp mutant cell lines over 2 h using DQ Green BSA- (proteolysis-sensor) or FITC-labeled (pH-sensor) latex beads (Sattler et al., 2013).

After phagocytosis of DQ Green BSA-labeled beads, we observe a significantly faster and stronger bulk proteolysis in nramp single mutants compared to Dictyostelium wt cells (Figure 5A, Figure S4). The double mutant shows an intermediate phenotype between wt and single mutant cell lines. All cell lines show an immediate decrease of the pH in the FITC-labeled bead containing phagosomes, which is significantly stronger in Δnramp1 compared to wt cells (Figure 5B, Figure S4). However, nramp mutants differ slightly in starting point and duration of acidification represented by a pH below 5 (indicated by the elevated line). An earlier (Δnramp1) or shorter (ΔnrampB, Δnramp1/B) acidification below pH 5 coincided with a significantly faster reneutralization of the latex bead-containing phagosome compared to wt cells. Together, these results indicate an impact of the Nramp transporters on phagosomal maturation of inert particles.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.