GM-CSF alters Zn transporter expression in H. capsulatum infected macrophages

We hypothesized that GM-CSF alters macrophage Zn distribution by regulating transporters. In an unbiased analysis of expression of Zn importers, ZIPs (Slc39a1-14) and exporters, ZNTs (Slc30a1-10), ~60 fold and 40 fold upregulation in Slc39a2, was observed in infected activated peritoneal and bone marrow macrophages respectively, compared to resting uninfected control (Figures 2A and 2C). A significant 2–4 fold increase in Slc304 and Slc30a7 (p<0.001 for peritoneal and p<0.05 for bone marrow macrophages) was also observed upon activation and infection (Figures 2B and 2D). The expression of another importer, Slc39a8 was unchanged (Figure S2A). TNF-α treatment of did not alter Slc39a2, Slc30a4 and Slc30a7 expression (Figure 2E). These data suggest that GM-CSF activated-infected macrophages may increase Zn import and mobilize it into intracellular organelles via exporters.

Open in a separate window

Figure 2

GM-CSF alters Zn transporter and MT expression

qRT-PCR of Slc39a and Slc30a in (A, B) peritoneal; (C, D) bone marrow; (E) TNF-α treated macrophages, data are mean ± SEM, 3 independent experiments; (F, G) Mt1, Mt2 and Mt3 expression in macrophages; RP, resting peritoneal; Act, GM-CSF activated; Hc, H. capsulatum,, data are mean ± SEM, 8 independent experiments; ND, not detected. Related to Figure S2.

Proteomics analysis of SEC fractions

GM-CSF altered the Zn-binding proteome in macrophages. For proteomics analysis, five fractions identified by SEC-ICP-MS were collected with the plasma turned-off and analyzed using a bottoms-up proteomics scheme. The list of identification parameters and proteins in all fractions and Zn-binding proteins highlighted in the fractions at ~10 and ~20 min are in Table S1. The CCCH zinc finger protein, ZC3H12A, calreticulin and MT2 were selected as being potentially associated with immune response and Zn binding. MT2 was detected in the peak that scavenged a majority of Zn at 20 min. MT1, MT3 and ZIP2 were not identified in database search. ZIP2 is located on the cell membrane. The association of membrane proteins with hydrophobic lipids and their insolubility poses a challenge in proteomic analysis (Santoni et al., 2000). Peptide matching is compromised by complex covalent hetero-oligomerization of MT isoforms, likely explaining the absence of MT1/MT3 as hits in database search. We observed oligomerization at ~20 min based on a shift in retention time towards the high molecular weight region, with the peak showing maximal absorbance at 260 nm, typical of -S-S- bonds formed by MTs during oligomerization, and not 280 nm, representative of Trp and Tyr containing proteins (Figure S1C).

GM-CSF, but not TNF-α induces MT expression in macrophages

GM-CSF enhanced Zn binding to the ~20 min fraction in infected macrophages (Figures 1C and 1D). Proteomic analysis and UV-Vis spectral characteristics suggested a major contribution of MTs to this Zn signal (Figure S1C, S1D, and S1E). Therefore, we examined the effect of GM-CSF on Mt expression. GM-CSF, but not TNF-α increased Mt1, Mt2 and Mt3 expression in macrophages during infection (Figures 2F and 2G), Mt3 is predominantly expressed in the brain (Falnoga et al., 2012). To ensure that our findings are valid, Mt3 was examined in hepatocytes and thymocytes and was not upregulated (Figure S2B). Expression of the Zn finger ribonuclease, Zc3h12a was enhanced in infected macrophages but was not specific to GM-CSF (Figures S2D and S2E). GM-CSF did not alter the expression of calreticulin and of the metal response element binding transcription factor-1 (Mtf1) which regulates Mt expression (Figure S2C and S2F). Thus, GM-CSF specifically induced Mt expression.

GM-CSF mobilizes labile Zn away from H. capsulatum

Confocal microscopy was used to visualize labile Zn with the fluorescent probe, Zinpyr-1 (Cao et al., 2001). Uninfected resting peritoneal macrophages showed labile Zn distribution in the cytoplasm and epinuclear space (Figure 3A); however, Zn staining was diffuse in infected resting peritoneal macrophages (Figure 3C). GM-CSF activation caused focal Zn accumulation (Figures 3B and 3F) and mobilized Zn away from H. capsulatum (Figures 3D and 3G) into the Golgi apparatus (Figure 3I). The reduced Zn fluorescence of yeasts observed in GM-CSF activated macrophages was not due to an impermeability of the dye (Figure S3A). Previous reports suggest that Zinpyr-1 may localize in the Golgi (Walkup et al., 2000). To ensure that labile Zn localization was a result of cytokine stimulation, we evaluated it in TNF-α treated macrophages. Zn distribution was non-focal, no or very little Zn staining was observed in the Golgi and labile Zn was observed in yeast (Figures 3E and 3H). This is consistent with lack of changes in Zn regulation in TNF-α treated macrophages (Figures 1G, 1H, 2E and 2G). Peritoneal macrophages treated with Zn and the ionophore, pyrithione, yielded bright fluorescence and very dim staining was observed in cells treated with the Zn chelator N,N,N′,N′-tetrakis-(2-pyridylmethyl)-ethylenediamine (TPEN) (Figures S3B and S3C). Labile Zn was examined in alveolar macrophages, the initial phagocyte that encounters H. capsulatum in vivo. While staining occurred in the cytosol and Golgi in resting alveolar macrophages, GM-CSF activation caused Zn localization in the Golgi (Figures S3D-S3G). Collectively, these data suggest that GM-CSF induced Zn redistribution in macrophages, which may limit its accessibility to H. capsulatum.

Open in a separate window

Figure 3

GM-CSF, but not TNF-α triggers Zn mobilization

Confocal microscopic analysis of Zn (green), nucleus (blue) and H. capsulatum (Hc) (red) in (A–E) peritoneal; (F–H) bone marrow macrophages; merge, overlay of Zn, nucleus and H. capsulatum; (I) Staining for Zn (green), nucleus (blue) and Golgi (red); merge, overlay of Zn, nucleus and Golgi; yellow, co-localization of Golgi and Zn dyes; white arrows point at the Golgi; scale bars, 20 μm; RP, resting peritoneal; Act, activated; 3 independent experiments. Related to Figure S3.

Metallothioneins sequester and deprive H. capsulatum of Zn

Since Zn regulation was similar in both types of macrophages, we utilized bone marrow macrophages for further studies. To determine if Zn sequestration in macrophages and deprivation in H. capsulatum was caused by MTs, we took two approaches. First, we silenced the expression of all four Mt genes (Figure 4A). Mt-silenced macrophages manifested reduced Zn binding to the chromatographic fraction at 20 min accompanied by an increase in the labile Zn fraction at ~25 min. (Figure 4B and zoom in). H. capsulatum recovered from Mt silenced macrophages had higher Zn content (p<0.01) (Figure 4C). Next, we examined Zn sequestration in Mt1^−/−Mt2^−/− macrophages. These cells exhibited a lack of Zn-binding to the fraction at 20 min, accompanied by an increase in labile Zn in GM-CSF treated and infected macrophages (Figure 4D and zoom in). H. capsulatum recovered from Mt1^−/−Mt2^−/− macrophages had higher Zn concentration (p<0.01) compared to wild type (WT) cells (Figure 4E). Thus, Zn deprivation within H. capsulatum resulted preponderantly from sequestration by MTs.

Open in a separate window

Figure 4

GM-CSF engages the STAT5 and STAT3 transcriptional program to regulate MT-Zn sequestration

(A) qRT-PCR of Mt genes in activated (Act) macrophages treated with scramble or Mt siRNA, % decrease in expression compared to scramble siRNA control, data are mean ± SEM, 2 independent experiments; (B, D) SEC-ICP-MS chromatograms of Mt silenced activated infected (Hc) macrophages and from WT and Mt1^−/Mt2^−/− mice and zoom in of labile Zn fraction, 2 independent experiments; (C, E) Total metal analysis of Zn in H. capsulatum (Hc) from Mt silenced and WT, Mt1^−/−Mt2^−/− macrophages, data are mean ± SD, 2 independent experiments; (F) Colony forming units (CFU) of H. capsulatum 24 h post infection from WT and Mt1^−/−Mt2^−/− activated macrophages compared to untreated control; data are mean ± SEM, 4 independent experiments; (G–J) SEC-ICP-MS chromatograms of chemically inhibited and Stat3 and Stat5 silenced macrophages, % decrease in peak area is compared to scramble siRNA treated infected macrophages; Y axis, Offset Zn-signal, 3 independent experiments; (K) p-STAT3, p-STAT5 and β-actin from silenced macrophages 24 h post activation, n=2; (L) Western blot of p-STAT3, STAT3 and β-actin from lysates 10 min post activation and infection, n=2.

We queried if MT1 and MT2 participated in antimicrobial defense. We postulated that the lack of Zn sequestration in the absence of MT1 and MT2 would diminish the ability of GM-CSF to inhibit fungal growth. While H. capsulatum growth was inhibited in WT activated macrophages, Mt1^−/−Mt2^−/− activated cells showed decreased ability to limit growth of yeasts (Figure 4F).

GM-CSF engages the STAT5 and STAT3 transcriptional program to mediate Zn regulation

We sought to determine the signaling mechanism that resulted in Zn modulation. GM-CSF signals through STAT5 (Bhattacharya et al., 2001) and STAT binding sites are present on the Mt promoter (Lee et al., 1999). Inhibition of STAT5 and STAT3 reduced MT-Zn binding (Figure 4G and 4H). To confirm our observations with chemical inhibitors, we silenced Stat5 and Stat3, which decreased Zn binding to MTs in the chromatograms of activated infected macrophages (Figures 4I, 4J and 4K).

These findings led us to postulate that GM-CSF may directly activate STAT5 and STAT3 signaling. GM-CSF increased STAT3 phosphorylation 10 min post-activation and infection, proving that the cytokine directly activated STAT3 (Figure 4L). There was >2.5 and >5 fold increase in Mt1 and Mt2 expression, as early as 10 min post infection and remained high for 24 h; Mt3 was modestly altered (Figure 5A). These data indicate that Zn regulation by GM-CSF occurs via STAT5 and STAT3 signaling.

Open in a separate window

Figure 5

Early regulation of Mt genes and role of Slc39a2 in MT-Zn sequestration

(A) qRT-PCR of Mt1, Mt2, Mt3 and Slc39a2, time points are min and h post infection, data are mean ± SEM, n=3 at 0–12 h and n=10 independent experiments at 24 h; (B) qRT-PCR of Slc39a2, Mt1 and Mt2 in Slc39a2 silenced macrophages normalized to scramble siRNA control, % decrease in expression compared to respective scramble controls, data are mean ± SEM, 3 independent experiments; (C) Total metal analysis of Zn in lysates (left) and H. capsulatum (right) from scramble siRNA and Slc39a2 silenced macrophages, data are mean ± SD, 3 independent experiments; (D) SEC-ICP-MS chromatogram of scramble siRNA and Slc39a2 silenced activated macrophages, Y axis, Offset Zn- signal; right panel, zoom in of labile Zn fraction, 3 independent experiments; (E) Ratio of Zn to sulfur in the MT fraction collected from 19 to 21 min based on the areas under the chromatogram corrected by differences in sensitivity through calibration with S and Zn standards, data are mean ± SD, 3 independent experiments. Related to Figure S4.

We investigated the importance of Slc39a2 in regulating GM-CSF function. Activated infected macrophages manifested a ~10 fold increase in Slc39a2 expression 10 min post infection (Figure 5A). We questioned whether the regulation of Slc39a2 was a prerequisite in macrophages defense. Upregulation of Mt genes was partially dependent on Slc39a2, but silencing Slc39a2 did not alter Zn deprivation in H. capsulatum, ROS production or growth inhibition (Figures 5B, 5C, S4A and S4B). Slc39a2 silencing reduced Zn influx and MT-Zn signal in macrophages, but did not affect labile Zn sequestration (Figures 5C and 5D). To probe how MTs in these macrophages established a defensive state against the pathogen, we compared the amount of Zn bound per MT by determining the [Zn] to [S] ratio (Zn²⁺ to MTs) between control and Slc39a2 silenced macrophages. As expected, scramble siRNA treated macrophages had increased MT and sequestered labile Zn (Figure 5D). But, the amount of Zn bound to MTs in Slc39a2 silenced infected macrophages was ~2.4 fold greater than scramble siRNA treated control, thereby enabling efficient Zn sequestration (Figure 5E), even with lower production of MTs. Silencing Slc39a14 did not affect GM-CSF driven macrophage defense (Figure S4C–F).

Zn sequestration enhances oxidative burst

GM-CSF enhances ROS production (Weisbart et al., 1987). Therefore, we asked if Zn sequestration affected oxidative burst during infection. GM-CSF increased ROS in macrophages. Zn chelation by TPEN in serum-free medium augmented ROS production by ~40%, which was significantly higher (p<0.05) compared to activated infected macrophages (Figure 6A). This effect was reversed upon addition of ZnSO₄ (p<0.001). To ensure that this was not due to chelation of other metals by TPEN, we specifically depleted Zn from serum-free and serum-containing media. Low-Zn serum containing media did not show a significant increase in ROS over activated infected macrophages (Figure 6A), possibly due to the presence of residual Zn contributed by serum or serum-LPS that may have an inhibitory effect on NADPH oxidase (Nox) activity. However, macrophages exposed to Zn-depleted serum-free media manifested ~36% increase in ROS, significantly higher (p<0.01) than activated infected macrophages (Figure 6A). These data emphasize that a labile Zn-deprived state induced by GM-CSF facilitated increased oxidative burst, which was entirely reversible upon addition of Zn.

Open in a separate window

Figure 6

Zn sequestration enhances ROS production

(A) % Change in ROS MFI in cells cultured in Zn replete, low Zn serum⁻, low Zn serum⁺, 10 μM TPEN serum⁻, 20 μM TPEN serum⁺, and 20 μM TPEN with 100 μM ZnSO4 treated serum⁺ media, serum⁻ media was included as serum containing media has considerably higher Zn. Where serum⁺ media was used, a higher concentration of TPEN was added (20μM), last group is significantly different (p<0.01) from all groups; ROS histograms demonstrating the effect of Zn sequestration; Act, activated, Hc, H. capsulatum, 2–3 independent experiments; (B) % change in ROS MFI in Ncf1^−/− activated macrophages, compared to WT activated uninfected cells, 3 independent experiments, histogram from Ncf1^−/− macrophages; (C) % inhibition of yeast growth in apocynin treated or Ncf1^−/− activated over WT macrophages, 3 independent experiments; (D) CFU of H. capsulatum in normal or low Zn media with increasing concentration of XO, with or without addition of Zn, 3 independent experiments; (E) % change in ROS MFI in WT and Mt1^−/−Mt2^−/− macrophages compared to WT uninfected cells, ROS histogram of WT and Mt1^−/−Mt2^−/− macrophages, 3 independent experiments; (F) Hvcn1 expression normalized to untreated macrophages, 3 independent experiments; (G) % change in ROS MFI in WT activated macrophages in response to infection, Zn and CCCP, 2 independent experiments, and Hvcn1^−/− activated cells compared to WT, 3 independent experiments; (H) % inhibition of H. capsulatum growth in infected and activated WT vs Hvcn1^−/− macrophages over WT control, 3 independent experiments; all data are mean ± SEM. Related to Figure S5.

The sources of cellular ROS include mitochondrial respiration (Drose and Brandt, 2012) and ROS produced by Nox (Minakami and Sumimotoa, 2006). Nox contributes to phagosomal ROS production that targets pathogens contained within these vacuoles. Activated macrophages from Ncf1^−/− mice failed to generate an increased ROS response upon H. capsulatum challenge or Zn depletion (Figure 6B), indicating that reduction in labile Zn specifically enhanced ROS generation via Nox. We postulated that ROS production would be compromised in Mt1^−/−Mt2^−/− macrophages as a consequence of reduced Zn sequestration. While WT activated infected macrophages effectively induced an oxidative burst, this was dampened in Mt1^−/−Mt2^−/− macrophages (Figure 6E).

ROS was essential for H. capsulatum growth inhibition; macrophages in which Nox was inhibited by apocynin and Ncf1^−/− macrophages had decreased ability to inhibit H. capsulatum growth (Figure 6C). To discern whether the defensive action of GM-CSF was a direct result of Zn limitation or enhanced ROS caused by Zn deprivation, we subjected H. capsulatum to ROS in vitro under Zn limitation. Though xanthine oxidase (XO) at a higher concentration reduced growth in normal media, Zn limitation greatly enhanced the inhibitory effect of ROS on H. capsulatum growth and this was reversible upon addition of Zn (Figure 6D). To determine the mechanism by which Zn sequestration facilitated increased Nox activity, we examined the regulation of hydrogen voltage gated proton channel (Hvcn1). Macrophages enhanced Hvcn1 expression in response to GM-CSF in a STAT3 and STAT5 dependent manner (Figures 6F and S5A). Proton flux by Hvcn1 accelerates ROS generation via Nox and Zn inhibits Hvcn1 function (DeCoursey et al., 2003). Addition of the protonophore, carbonyl cyanide m-chlorophenyl hydrazone (CCCP) reversed the inhibitory effect of Zn on ROS generation (Figure 6G). ROS production was attenuated in Hvcn1 silenced and Hvcn1^−/− macrophages upon GM-CSF activation and infection. Hvcn1^−/− activated macrophages showed compromised ability to control H. capsulatum growth compared to WT control (Figures 6G, 6H, S5B, S5C). These data signify the impact of Zn sequestration in enhancing phagocyte defense via ROS production.

GM-CSF drives Zn binding to MT in vivo

GM-CSF has a profound influence on the fate of infection in vivo (Deepe et al., 1999). We hypothesized that GM-CSF signaling was required for Zn sequestration in macrophages in vivo. In a pulmonary model of H. capsulatum infection, F4/80⁺ CD11b⁺ GFP⁺ and F4/80⁺ CD11b⁺ GFP⁻ macrophages were sorted from lung leukocytes of green fluorescent protein (GFP) H. capsulatum infected mice (Schematic, Figure 7A). F4/80⁺ CD11b⁺ GFP⁺ macrophages showed enhanced Zn binding to MTs compared to F4/80⁺ CD11b⁺ GFP⁻ cells (Figure 7B). To evaluate the importance of GM-CSF in Zn binding to MTs in vivo, it was neutralized and Zn profiles of macrophages were evaluated 7 days post infection (p.i). There was a sharp reduction in Zn bound to MTs in F4/80⁺ CD11b⁺ GFP⁺ macrophages in GM-CSF neutralized mice making it comparable to F4/80⁺ CD11b⁺ GFP⁻ macrophages (Figure 7D). No alteration was seen in Rat IgG controls (Figure 7C). We asked whether GM-CSF neutralization enhanced fungal burden in macrophages. The percent and MFI of GFP⁺ cells in F4/80⁺ CD11b⁺ macrophages were higher in anti-GMCSF treated mice (Figure 7F). To further examine the importance of GM-CSF signaling in Zn regulation, WT macrophages were adoptively transferred into Csf2ra^−/− mice lacking the GM-CSF receptor, infected with GFP H. capsulatum and WT vs. Csf2ra ^−/− macrophages were sorted 7 days post infection. Mt1, Mt2 and Slc39a2 expression was enhanced in infected WT macrophages, while Csf2ra^−/− macrophages showed over 60, 93 and 97% decrease in Mt1, Mt2 and Slc39a2 expression respectively compared to infected WT macrophages (Figures 7G and 7H). Moreover, the percent infectivity of F4/80⁺ CD11b⁺ GFP⁺ cells in Csf2ra^−/− mice was higher (16.7%) as compared to infected WT CD45.1 macrophages (1.26%). These data highlight the importance of GM-CSF in regulating Zn during infection in vivo.

Open in a separate window

Figure 7

GM-CSF triggers Zn binding to MTs in vivo and in human macrophages

(A) Schematic, mice were left untreated or treated with rat IgG or anti-GM-CSF and infected i.n. with 2 × 10⁶ GFP⁺ yeasts (Hc); lungs were harvested 7 days p.i. and F4/80⁺CD11b⁺GFP⁺ and F4/80⁺CD11b⁺GFP⁻ cells were sorted. SEC-ICP-MS profiles of (B) F4/80⁺CD11b⁺GFP⁺ and F4/80⁺ CD11b⁺GFP⁻ cells isolated from lungs; (C) rat IgG treated and (D) anti-GM-CSF treated mice, 2 independent experiments; (E) SEC-ICP-MS profiles of lysates from human resting (h R) and GM-CSF activated (h Act) macrophages; Y axis, offset Zn signal; inset, zoom in of 15–25 min fraction, 2 independent experiments; (F) Dot plot of lung leukocytes gated on F4/80⁺CD11b⁺ macrophages; histogram of F4/80⁺CD11b⁺GFP⁺ macrophages; bar graphs are quantification of %, absolute numbers and MFI of F4/80⁺CD11b⁺GFP⁺ macrophages in rat IgG vs anti-GM-CSF treated mice, data are mean ± SEM, n=4; (G) Schematic representing transfer of CD45.1 WT macrophages into CD45.2 Csf2ra^−/− mice, infection and sorting of WT vs Csf2ra^−/− macrophages from lungs for gene expression analysis; (H) qRT-PCR of Mt1, Mt2 and Slc39a2 from sorted CD45.1⁺ (WT) and CD45.1⁻ (Csf2ra^−/−) GFP⁻ and GFP⁺ macrophages normalized to CD45.1⁺GFP⁻ macrophages, % values are decrease in gene expression in Csf2ra^−/− CD45.1⁻ GFP⁺ macrophages compared to WT CD45.1⁺ GFP⁺ macrophages, data are mean ± SEM, data are from 1 experiment. Related to Figure S6.

Mice

C57BL/6, CBA/J, Mt1^tm1BriMt2^tm1Bri (Mt1^−/−Mt2^−/−), 129S1/SvImJ WT (control for Mt1^−/−Mt2^−/−) and B6.SJL-Ptprc^a Pepc^b/BoyJ (CD45.1) mice were obtained from Jackson Laboratory. We thank Dr. Bruce Trapnell, Cincinnati Children’s Hospital Medical Center for the Csf2ra^−/− mice; Dr. John Engelhardt, Dr. Yulong Zhang (University of Iowa), Dr. Brahm Segal, Dr. Nazmul Khan (Roswell Park Cancer Institute) and Dr. Long-Jun Wu and Dr. Jiyun Peng (Rutgers University) for providing bone marrow from Ncf1^−/− and Hvcn1^−/− mice. Animal experiments were in accordance with Animal Welfare Act guidelines of National Institutes of Health.

Activation, STAT inhibition and infection

Bone marrow macrophages were activated with cytokines for 24 h before infection. Peritoneal macrophages were rested for 24 h, followed by activation for 24 h before infection and were activated again during infection. Where indicated, macrophages were cultured throughout with 100μM STAT3 or STAT5 inhibitors, S3I-201 and N′-((4-Oxo-4H-chromen-3-yl)methylene) nicotinohydrazide (Calbiochem) respectively. For infection, G217B or G186R H. capsulatum were cultured in Ham’s F12 (Zn concentration, 250 ± 18 ng/ml or 4 μM) and washed before use.

Microscopy

Macrophages were infected with PKH-26 stained H. capsulatum for 24 h followed by staining with nuclear, Golgi and Zn dyes where indicated for 30 min. Images were acquired on a Zeiss LSM710 confocal and analyzed using ZEN 2011 software.

Silencing

Genes were silenced using TransIT-TKO transfection reagent (Mirus Bio LLC) and 100nM Mt (Santa Cruz), Slc39a2, Slc39a14, Stat3 or Stat5 (50nM each of Stat5a and Stat5b), Hvcn1 or scrambled siRNA (Dharmacon) as per manufacturer’s instructions.

ROS

Macrophages were activated for 24h and then incubated with 5μM Dihydroethidium (DHE) (Invitrogen) in fresh media for 15 min in dark, followed by GM-CSF and infected for 2h. Where indicated, 100μM ZnSO₄ and TPEN or Zn depleted media was added 20 min prior to DHE. CCCP was added 30 min before analysis by flow cytometry (Accuri C6). Data were analyzed with FCS Express 3 De Novo Software (CA) or FlowJo.

In vivo

Eight week old C57BL/6 mice were infected with 2 × 10⁶ GFP⁺ H. capsulatum yeasts i.n. Mice were injected ip. with 0.5mg control mAb Rat IgG2a (BioXcell) or MP1–22E9 mAb Rat IgG2a anti-GM-CSF a day before and on the day of infection. After 7 days, lungs were homogenized with gentleMACS Dissociator (Miltenyi Biotec, CA), collagenase treated, filtered through 60 μM nylon mesh and washed. Leukocytes were isolated using Lympholyte M (Cedarlane Laboratories, Canada) and stained with APC F4/80 (AbD Serotec) and PerCP CD11b mAbs (BD Biosciences) and CD16/CD32 blocking mAbs for 30 min at 4°C and washed; gated on F4/80⁺CD11b⁺ and sorted into GFP⁺ (containing phagocytosed H. capsulatum) and GFP⁻ (not containing H. capsulatum) using 5-laser FACS Aria II (BD Biosciences) in a BSLII facility at Cincinnati Children’s Hospital; collected at 4°C in culture media and washed prior to lysis. For adoptive transfer, 2×10⁶ bone marrow macrophages from WT (CD45.1) were transferred intratracheally into 8 week old Csfr2a^−/− (CD45.2) mice 1 day prior to H. capsulatum infection. On day 7 post infection, lung leukocytes were sorted into 4 populations: F4/80⁺CD11b⁺CD45.1⁺GFP⁻ or GFP⁺ and F4/80⁺CD11b⁺CD45.1⁻GFP⁻ or GFP⁺ macrophages for analysis of gene expression.

Human macrophages

PBMCs were isolated from human blood (Hoxworth Blood Center, University of Cincinnati) using Ficoll-Paque. Monocytes were purified with Monocyte Isolation Kit II on a VarioMACS separator (Miltenyi Biotec) and cultured in RPMI with 12.5% human male AB serum (Sigma-Aldrich) for 7 days, and macrophages were exposed to 10 ng/ml GM-CSF (Peprotech) or vehicle for 24 h before infection, re-exposed to GM-CSF at the time of infection and lysates were prepared 24 h later.

Cell lysates for ICP-MS

Macrophages were washed and lysed with 0.1% sodium dodecyl sulfate (SDS) in HPLC water for 20 min on ice. Lysates were passed through a 0.22 μm filter or centrifuged at 13,000rpm for 5 min for analysis. For total metal analysis of H. capsulatum, lysates were centrifuged at 13,000rpm for 5 min to pellet the yeasts and the pellet was washed twice with HBSS.

Total elemental analysis and SEC-ICP-MS screening

ICP-MS analysis was performed on Agilent 7700x ICP-MS instrument (Agilent Technologies). A conventional Meinhard nebulizer, Peltier-cooled spray chamber, and shield torch constituted the sample introduction system under standard conditions. Sc was used as internal standard and SRM (DORM-3 or NIST1745) were analyzed for every digestion.

For SEC-ICP-MS, Agilent 1100 series HPLC system equipped with a binary pump, vacuum membrane degasser, thermostated auto sampler, column oven, and diode array detector (DAD), with a semi-micro flow UV-Vis cell was coupled to the ICP-MS through a 0.17mm internal diameter short PEEK tube. The system was controlled with Chemstation software. A TSK Gel 3000SW 7.5 × 300 mm column was used. For proteomics, samples were concentrated before injection using MWCO filter 3 kDa (Millipore). For analysis of MT saturation, the MTs-SEC signal (20 min) was collected and analyzed for S and Zn to determine the MT:Zn ratio in a Superdex peptide SEC-ICP-MS using O₂ reaction mode.

Proteomics

The SEC-ICP-MS fractions were analyzed using an Agilent 1200 nanoHPLC, nanoChip ESI and 6300 series MSD Ion Trap XCT Ultra system. Peptide and corresponding protein identification were conducted using MASCOT server (Matrix Science).

We thank Agilent Technologies for instrumentation, ICP-MS and nanoHPLC-ESI-IT-MS and Drs. B. Klein, W. Nauseef, and D. Hildeman for their valuable feedback. This work was supported by grants from the NIH, AI-094971, AI-106269 and a Merit Review from Veterans Affairs.