A PPIX-binding probe facilitates discovery of PPIX-induced cell death modulation by peroxiredoxin

Abstract

While heme synthesis requires the formation of a potentially lethal intermediate, protoporphyrin IX (PPIX), surprisingly little is known about the mechanism of its toxicity, aside from its phototoxicity. The cellular protein interactions of PPIX might provide insight into modulators of PPIX-induced cell death. Here we report the development of PPB, a biotin-conjugated, PPIX-probe that captures proteins capable of interacting with PPIX. Quantitative proteomics in a diverse panel of mammalian cell lines reveal a high degree of concordance for PPB-interacting proteins identified for each cell line. Most differences are quantitative, despite marked differences in PPIX formation and sensitivity. Pathway and quantitative difference analysis indicate that iron and heme metabolism proteins are prominent among PPB-bound proteins in fibroblasts, which undergo PPIX-mediated death determined to occur through ferroptosis. PPB proteomic data (available at PRIDE ProteomeXchange # PXD042631) reveal that redox proteins from PRDX family of glutathione peroxidases interact with PPIX. Targeted gene knockdown of the mitochondrial PRDX3, but not PRDX1 or 2, enhance PPIX-induced death in fibroblasts, an effect blocked by the radical-trapping antioxidant, ferrostatin-1. Increased PPIX formation and death was also observed in a T-lymphoblastoid ferrochelatase-deficient leukemia cell line, suggesting that PPIX elevation might serve as a potential strategy for killing certain leukemias.

Fig. 1. ALA increases PPIX levels dose- and time-dependently without altering levels of other porphyrins.

a ALA enhances generation of PPIX in NIH3T3 cells in a dose-dependent manner following 6-h treatment with various ALA concentrations. (n = 2 biologically independent samples, representative of two independent experiments). b Overlay of untreated and 1-h 2 ALA-treated NIH3T3 cells demonstrates marked induction of PPIX but no other changes in heme or other porphyrins. c Time dependence of PPIX and heme (insert) generations in NIH3T3 cells treated with 200 µM ALA. (n = 2 biologically independent samples, representative of three independent experiments).

Fig. 2. Increased heme/porphyrin synthesis produces altered mitochondria in NIH3T3 cells.

a Electron micrographs of representative mitochondria used for scoring of morphological abnormalities incurred by ALA treatment (24 h). Arrows indicate mitochondria showing a score of 2–3 in representative slides. Scale bars indicating length are as follows, untreated, succinylacetone and ALA + succinylacetone bars indicate 600 nm, for ALA bar indicates 1 µm. b Average scores of mitochondrial morphology change after ALA treatment (illustrated by violin plot where thickness represents distribution); mitochondria viewed by electron microscopy at 6 h were scored with 0 indicating unaltered morphology and 3 indicating severe, dose-dependent impairment (number of mitochondria evaluated were: n = 71 for Control, n = 88 for ALA and n = 108 for ALA + SA). c Mitochondrial membrane potential changes with ALA treatment (200 µM). (n = 2 independent replicates representative of five independent experiments). Significance is indicated by the number of attached asterisks. *p < 0.05, **p < 0.01, ns indicates no significance. Significance in (b) was evaluated using one-way ANOVA, while unpaired Student’s t test, two-sided, was used to determine significance in (c).

Fig. 3. Development and testing of PPB, a PPIX-biotin probe.

a Model illustrating formation and movement of PPIX within the cell and molecular structures of ALA, PPIX, and heme. b Structure of PPB, with PPIX, linker and biotin domains indicated. c LC-MS spectrum of synthesized PPB indicating high level of purity. d Model indicating the dual interactions of PPB with both PPIX-binding proteins and avidin. e PPB binds albumin and quenches fluorescence comparably to PPIX. Fluorescence quenching was in the order, CPIII > Heme > UroIII > PPIX > PPB. (n = 2 independent samples, representative of three independent experiments).

Fig. 4. Confirmation of the structure of PPB by NMR.

NMR data indicating a ¹³C, ¹H-HSQC, b ¹⁵N, ¹H-HSQC, c ¹H ¹H COSY, d ¹H ¹H TOCS, and e ¹H ¹H NOESY data confirm the proposed structure of PPB.

Fig. 5. Mass spectrometry using tandem mass tag labeling indicates cellular differences in PPIX-binding proteins.

a Representative silver stain of PPB immunoprecipitation and controls from four cell lines (n = 3 or more independent experiments for each cell line). b Subcellular localization of PPIX-binding proteins identified by TMT. c Principal component analysis (PCA) of TMT proteomics data. d Unsupervised hierarchical clustering of cell lines’ quantitative TMT proteomic data from PPB immunoprecipitation assay.

Fig. 6. Analysis of PPIX-interacting proteins identifies NIH3T3-specific cluster of PPIX-binding proteins related to iron and mitochondrial metabolism.

a Weighted gene co-expression network analysis (WGCNA) of the PPB and tandem mass tag (TMT) quantitative proteomics data identified a cluster of proteins enriched in NIH3T3 cells. b Highest candidate PPIX-binding proteins from NIH-3T3 cluster 1 shown relative to other cell lines (n = 2 biologically independent samples, mean). c STRING protein-protein interaction (PPI) network analysis of cluster 1 identifies a protein module enriched for iron and oxidative phosphorylation. d Enrichr analysis of cluster 1 identifies porphyria and iron relationships. e PPIX-binding proteins identified in screen (orange) demonstrating a role in iron pathway and ferroptosis.

Fig. 7. ALA induces increases ROS and promotes ferroptosis.

a Total cellular ROS in ALA-treated cells, as measured by DCFDA fluorescence (n = 2 independent samples, representative of three independent experiments). b Mitochondrial superoxide levels in ALA-treated cells, as measured by MitoSOX Red fluorescence (n = 4 independent experiments, mean ± SEM). c ALA cytotoxicity is suppressed by antioxidants; ascorbic acid (500 µM), N-acetylcysteine (5 mM) or ferrostatin-1 (10 µM) (n = 3 independent samples, representative of 3 or more independent experiments for all drugs tested, mean ± SEM). d Increased lipid peroxidation with ALA treatment and attenuation by SA, as assessed by C11-Bodipy (n = 2 independent samples, representative of three independent experiments). e Oxidized/total glutathione ratio changes in ALA-treated cells or cotreated with succinylacetone (n = 3 independent experiments, mean ± SEM). f Glutathione peroxidase activity of lysate from NIH3T3 in control or 24-h ALA (200 µM) treated, following addition of cumene hydroperoxide substrate (n = 3 independent samples, representative of five independent experiments, mean ± SEM). g Specific activity of glutathione peroxidase in control and 24-h ALA-treated NIH3T3. Significance is indicated by the number of attached asterisks (n = 5 independent samples, representative of five independent experiments, mean ± SEM). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, and ns indicates no significance. Significance in this figure was evaluated using one-way ANOVA.

Fig. 8. Modulators of iron, endocytosis, and lysosome fusion alter ALA toxicity.

a Dose-dependent reduction of ALA cytotoxicity by desferrioxamine (DFO) iron chelation treatment (n = 3 independent experiments). b ALA cytotoxicity is reversed by both chloroquine (50 µM) and bafilomycin (20 nM) (n = 2 independent samples, representative of three independent experiments). c ALA cytotoxicity is reversed by ferrostatin-1 (10 µM) and liproxstatin-1 (10 µM) (n = 3 independent experiments). d Knockdown of ACSL4 in NIH3T3 or control siRNA at 48 h (n = 2 independent experiments). e Rescue of ALA toxicity of NIH3T3 by ferrostatin-1 (10 µM) or by ACSL4 knockdown 48 h prior to treatment. (n = 2 independent samples, representative of two independent experiments). f Co-treatment of cells with ALA and various concentrations of erastin increased cytotoxicity (n = 2 independent experiments). g Response surface modeling analysis of ML210 and ALA cotreatment in NIH3T3, α = −0.521, analysis of datasets from four independent experiments. h Immunoblot analysis of H-ferritin expression treated with either control siRNA or specific H-ferritin siRNA. (n = 2 biologically independent experiments). i Gene knockdown of H-Ferritin sensitizes cells to ALA cytotoxicity (n = 2 biologically independent experiments). j Immunoblot analysis of TfR expression treated with either control siRNA or specific TfR siRNA. cytotoxicity (n = 4 independent, mean ± SEM). k Gene knockdown of TfR protects cells from ALA cytotoxicity (n = 4 independent experiments, mean ± SEM). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, and ns indicates no significance.

Fig. 9. ALA mediated cell death is distinct from erastin mediated ferroptosis, is enhanced specifically by knockdown of PRDX3, and increases PPIX and toxicity in Jurkat.

a Efficient knockdown of PRDX1, 2, and 3 by siRNA treatment (n = 7 experiments). b Area-under-the-curve assessment of graphs of ALA-treated cells, with-or-without suppression of PRDX1, 2 or 3. Knockdown of PRDX3, but not that of PRDX1 or PRDX2, enhances the cell death caused by PPIX induction (n = 7 independent experiments, mean ± SEM). c Pulldown of PRDX3 from NIH3T3 lysate and silver staining (top) or immunoblot for PRDX3 (bottom). (n = 2 independent experiments). d MitoSOX Red mitochondrial superoxide indicator fluorescence increases in NIH3T3 with ALA treatment in both PRDX3 knockdown and control (n = 4 independent experiments, mean ± SEM). e ALA (156 μM) induced cytotoxicity is potentiated by siRNA PRDX3 knockdown, but reversed by ferrostatin-1, but which was reversed by the addition of ferrostatin-1 (n = 2 independent experiments). f ALA dose-dependently increases PPIX fluorescence in Jurkat cells (n = 2 independent experiments). g ALA reduced Jurkat cell viability is attenuated by ferrostatin-1 (10 µM), but not ZVAD (20 µM). (n = 2 independent experiments). h Jurkat cell viability is rescued by liproxstatin-1 under ALA treatment, 48 h. (n = 4 independent experiments, mean ± SEM). i ML210 and ALA produce additive effect on Jurkat viability. (n = 3 independent experiments, mean ± SEM). j ACSL siRNA knockdown rescues ALA toxicity in Jurkat cells. (n = 2 independent experiments). k PRDX3 overexpression in 293T cells reduces MitoSOX Red fluorescence induced by ALA or antimycin A. (n = 3 independent experiments, mean ± SEM). l PRDX3 from NIH3T3 treated with hydrogen peroxide (5 min) shows increased levels of hyperoxidized dimeric form shown in model. m PPIX treatment of NIH3T3 induces formation of hyperoxidized dimers. Total PRDX3 is shown following reduction with β-mercaptoethanol. (n = 2 biologically independent experiments) *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, and ns indicates no significance.