The stress-activated fatty aldehyde dehydrogenase Hfd1 determines t-2-hex mediated toxicity and adaptive response. (A) Overview of the generation of t-2-hex within the sphingolipid degradation pathway. Stress-activated (-repressed) enzymatic functions are depicted in green (red) for yeast. The corresponding human enzymes involved in t-2-hex chemistry are shown in blue. t-2-hex stimulates the pro-apoptotic activity of Bax at human mitochondria by C126 lipidation. (B) HFD1 expression is activated by different cytotoxic stresses. Hyperosmotic (NaCl), oxidative (Menadione) and pro-apoptotic (Acetic acid) stresses were gradually applied to HFD1p-luciferase reporter containing yeast cells. The relative luciferase activity was measured in vivo (n = 3). (C) Comparison of different stress sensitivities of the HFD1 and GRE2 promoters. (D) External addition of t-2-hex (HD) activates HFD1 expression in a concentration dependent manner. Upper panel: HFD1-luciferase activation upon gradual addition of t-2-hex (n = 3). Lower panel: Comparison of HFD1 and GRE2 activation sensitivities to t-2-hex. (E) Left panel: HFD1 gene dose modulates susceptibility to t-2-hex. Growth curves of wild type, hfd1Δ and constitutively Hfd1 overexpressing strains (TDH3p-HFD1) in the absence and presence of t-2-hex (n = 3). Right panel: Quantitative comparison of growth performance (length of lag phase) of the same strains upon the indicated t-2-hex concentrations (n = 3). (F) Loss of Hfd1 function causes a hypersensitive t-2-hex response. GRE2p-luciferase reporter assay in strains with the indicated HFD1 gene dose in response to t-2-hex (n = 3). (G) HFD1 gene dose modulates the tolerance to pro-apoptotic concentrations of acetic acid (n = 3). *p<0.05 by Student's unpaired t-test.

t-2-hex stress causes profound transcriptomic remodeling. (A) Transcriptomic analysis of the cellular response to t-2-hex overload. RNA Seq experiment setup and general functional groups of up- and down-regulated genes. (B) t-2-hex activated mitochondrial functions. (C) Genomic remodeling of gene expression reveals a general proteostatic response of the cell to t-2-hex. Cytosolic chaperones are strongly activated in general, but not the Cct chaperonin or the prefoldin complex. Structural components and regulators of the proteasome are coordinately up-regulated, while the cytosolic translation machinery and cell cycle regulators are strongly repressed. Significantly up-regulated enzymes with known aldehyde dehydrogenase or unsaturated aldehyde reductase activities are shown. Colors indicate log2 fold changes in gene expression of t-2-hex treated versus mock treated cells (means of n = 3 independent biological replicates) for genes representing selected functional groups or complexes.

t-2-hex specifically induces the heat shock and proteasomal transcriptional response. (A) Significantly enriched promoter motifs were identified from the transcriptionally up- and down-regulated genes identified by our RNA seq study upon t-2-hex stress. (B) Schematic representation of the stress-specific luciferase reporters applied in the dose-response experiments upon unsaturated t-2-hex or saturated analogue t-2-hex-H2. The t-2-hex dose causing half maximal induction for each reporter is given in μM below the constructs. (C) Dose-response curves of the indicated live cell luciferase reporters upon increasing t-2-hex and t-2-hex-H2 concentrations (n = 3). All reporters were assayed in hfd1Δ cells. Initial light emission levels at time 0 were set to 1. (D) Comparison of the sensitivities of different stress type specific responses to the pro-apoptotic t-2-hex. Experimental data from (C) were analyzed by plotting the maximal reporter activation against the log[t-2-hex] concentration. (E) t-2-hex unsaturation is the cause of its severe growth inhibition. Growth of yeast wild type cells was scored upon the indicated t-2-hex and t-2-hex-H2 concentrations (upper panel) and the lag time calculated (lower panel), n = 3.

t-2-hex leads to cytosolic protein aggregation and inhibition of proteasomal function. (A) GFP-tagged Hsp104 was used to visualize intracellular protein aggregation in wild type and proteasome deficient rpn4Δ cells. Specifically, unsaturated t-2-hex caused a slowly increasing protein aggregation. Cells were treated with 100 μM of the bioactive lipids for the indicated times. Lower panels: Quantification of t-2-hex induced protein aggregates across cell populations. Number of analyzed cells: wt DMSO n = 601, wt t-2-hex n = 584, wt t-2-hex-H2 n = 558; rpn4 DMSO n = 551, rpn4 t-2-hex n = 549, rpn4 t-2-hex-H2 n = 329. (B) t-2-hex activated protein aggregation was no longer observed after inhibition of protein synthesis with cycloheximide (CHX). Lower panel: Quantification of t-2-hex induced protein aggregates across cell populations. Number of analyzed cells: t-2-hex n = 584, CHX + t-2-hex n = 134.(C) Effect of pro-apoptotic t-2-hex on proteasomal activity. Proteasomal activity was quantified in whole cell extracts before (Cnt) or after treatment with 200μM t-2-hex, t-2-hex-H2 or vehicle (DMSO) for 3h in wild type or rpn4Δ cells (n = 3). Activity of untreated wild type cells was set to 100%. (D) Response of proteasomal subunit Rpn8 expression upon t-2-hex exposure. Rpn8 was expressed as a Tap fusion from its chromosomal locus and cells were treated or not with 200μM t-2-hex for the indicated times. Rpn8 protein abundance was quantified by anti-Tap western blot (upper panel) and quantified relative to uninduced levels (n = 2). (E) Proteasomal deficiency causes t-2-hex sensitivity. Growth of yeast wild type and rpn4Δ cells was quantified upon the indicated t-2-hex concentrations (left panel) and the doubling time calculated (right panel), n = 6. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0005 by Student's unpaired t-test.

t-2-hex overload induces mitochondrial pre-protein accumulation and aggregation of de novo synthesized proteins. (A) The appearance of unimported mitochondrial precursor proteins (*) was induced by the uncoupler CCCP (20μM) or t-2-hex (200μM) for the indicated times. Aim17, Cox5a, Ilv6 and Sdh4 were visualized in chromosomally Tap-tagged wild type yeast strains by anti-Tap western blot. (B) Mitochondrial import block depends on t-2-hex unsaturation. The same strains as in (A) were treated with DMSO (control), CCCP (20μM), t-2-hex (200μM) or t-2-hex-H2 (200μM) for 40min and mitochondrial pre-protein accumulation visualized by anti-Tap western blot. (C) Mitochondrial fragmentation depends on t-2-hex unsaturation. Yeast wild type cells expressing mt-GFP were treated with vehicle (control), 200μM t-2-hex or t-2-hex-H2 for 1h. (D) t-2-hex induces the formation of aberrant mitochondrial proteins. Yeast wild type strains expressing Aim17-, Cyc7- or Mpc3-Tap tagged fusion proteins from their chromosomal locus were treated with 200μM t-2-hex for the indicated times. Fusion proteins were detected by anti-Tap western blot. (E) Inhibition of de novo protein synthesis abolishes the formation of aberrant mitochondrial proteins. Experimental conditions as in (D), but including a pretreatment with cycloheximide (CHX, 250μg/ml) where indicated. (F) t-2-hex induces aberrant forms of highly expressed proteins. Yeast wild type strains expressing Cis1- or Tma10-Tap tagged fusion proteins from their chromosomal locus were treated or not with 200μM t-2-hex for the indicated times. Fusion proteins were detected by anti-Tap western blot. (G) Hfd1 gene dose determines the extent of t-2-hex induced pre-protein accumulation. Aim17 pre-protein accumulation in response to different t-2-hex concentrations was quantified in wild type, hfd1Δ and HFD1 overexpressing cells. Cells were either untreated, or treated with the indicated lipid concentrations or DMSO for 20 min. The percentage of pre-protein relative to total Aim17 protein was calculated in the right panel (n = 4) (H) t-2-hex causes rapid protein ubiquitination. Yeast wild type cells were treated for the indicated times with t-2-hex (200μM) and protein ubiquitination visualized by anti-Ub western blot (upper panel) and quantified relative to the Pgk1 loading control (lower panel) (n = 2).

Functional genomics screen SATAY for the identification of pro- and anti-apoptotic functions upon t-2-hex stress. (A) Schematic outline of the SATAY experiment. hfd1Δ cells harboring the galactose inducible transposase (TPase) and a transposon (TN) disrupting the ADE2 gene were grown in galactose-containing medium to induce transposition. TN generated mutant cells were inoculated in synthetic glucose medium lacking adenine in the presence or absence of 50μM t-2-hex, grown for several generations and harvested for DNA extraction and sequencing of transposon insertion sites (TNs). TNs are mapped to the genome to identify genes that become required for proliferation under t-2-hex stress conditions (red) or genes whose mutation is beneficial for t-2-hex tolerance (green). (B) Identification of anti- and pro-apoptotic gene functions upon t-2-hex overload. Volcano plot showing the fold change of number of transposon insertions (TN) per gene of libraries grown in t-2-hex excess versus control conditions. TN under-enriched (anti-apoptotic) genes were analyzed with a log2 ratio below -0.75 and are summarized in Supplemental Table 3. TN enriched (pro-apoptotic) genes were analyzed with a log2 ratio > 1.5 and are available in Supplemental Table 4.

Functional analysis of pro- and anti-apoptotic genes upon t-2-hex excess. (A) The pleiotropic drug response activator Pdr1 and the mitochondrial protein Fmp52 are necessary for t-2-hex tolerance. Quantitative growth assay of wild type, pdr1Δ and fmp52Δ cells upon the indicated concentrations of t-2-hex. The relative lag time was quantified as an estimator of growth inhibition (n=3). (B) Deletion of ribosomal protein subunits causes t-2-hex resistance. As in (A), but using the yeast deletion strains rps30bΔ, rpl6aΔ, rps28aΔ and rpl40aΔ. Growth curves are depicted only for the t-2-hex tolerant strains rps30bΔ and rpl6aΔ (n=3). (C) Inhibition of ribosomal biogenesis is beneficial for t-2-hex tolerance. Yeast wild type cells were grown upon the indicated t-2-hex concentrations in the presence or absence of the inhibitor diazaborine (DAB, 20μg/ml) (n=3). (D) Galactose growth improves t-2-hex tolerance. Growth of wild type and hfd1Δ cells upon the indicated concentrations of t-2-hex on synthetic glucose or galactose media (n=3). **p<0.01, ***p<0.001 by Student's unpaired t-test. (E) Summary of anti-apoptotic genes identified by SATAY with significant transcriptional up-regulation (red) and pro-apoptotic genes identified by SATAY with significant transcriptional down-regulation (green) according to our RNA-seq study.

t-2-hex targets the mitochondrial TOM complex. (A) Hfd1 co-purifies with Tom70. Constitutively Hfd1-HA expressing cells were used in co-immunoprecipitation experiments from yeast whole cell extracts in the presence or not of endogenously expressed Tom22- or Tom70-Tap. Lower panel: Quantification of Hfd1 co-purification with respect to Tom20 and Tom70; n = 6. (B) A chemoproteomic screen with t-2-hex alkyne identifies the TOM and Tim23 complexes as lipidation targets. Purified mitochondrial fractions from yeast wild type cells were treated or not with the clickable analogue t-2-hex-Alkyne at high and low doses (100μM or 10μM). After addition of biotin to the modified proteins and purification with Streptavidin agarose pull-down, the protein identities of the t-2-hex targets were determined by mass spectrometry. Cysteine containing subunits of the TOM and Tim23 complexes are depicted as direct lipidation targets. The tables show the mean spectral reads for selected subunits in the chemoproteomic analysis for mock treated (Ctrl) and t-2-hex treated mitochondria (n = 2), as well as the log2 fold enrichment and adjusted p-value. (C) Tom40 is lipidated in vitro by t-2-hex. Upper panel: t-2-hex-Alkyne was used in the indicated concentrations to lipidate proteins in enriched mitochondrial preparations from Tom40-HA expressing cells or control cells. After t-2-hex addition, input samples were generated directly, while pull-down samples were treated with click chemistry for covalent biotin linkage and subsequent Streptavidin purification. Tom40-HA was detected in all samples by anti-HA western blot. Lower panel: Competition of Tom40 t-2-hex lipidation by free thiol groups. t-2-hex-Alkyne was used at 100μM to lipidate Tom40-HA in mitochondrial preparations in the presence or absence of 2mM DTT. (D) Mutations of TOM accessory subunits 20 or 70 cause t-2-hex tolerance. t-2-hex toxicity assays comparing wild type cells with the cysteine free Tom40 mutant (Tom40 CFREE) and deletion mutants tom20 or tom70. Lower panel: Quantitative comparison of growth performance (length of lag phase) of the tom20 and tom70 strains upon the indicated t-2-hex concentrations (n = 3). **p<0.01, ***p<0.001 by Student's unpaired t-test.

Model of the pro-apoptotic function of t-2-hex and the anti-apoptotic function of Hfd1 at the mitochondrial Tom complex. Upper panel: Under normal conditions, the Hfd1 lipid aldehyde dehydrogenase located at the Tom complex safeguards mitochondrial protein import by t-2-hex degradation. Lower panel: Upon severe stress conditions or in the absence of Hfd1 function, an excess of t-2-hex directly lipidates Tom subunits such as Tom40 and thus inhibits mitochondrial pre-protein transport across the outer mitochondrial membrane. The resulting proteostatic imbalance in the cytosol, if not sufficiently repaired or counteracted by the heat shock response, proteasomal clearance or diminished de novo protein synthesis, can induce apoptotic cell death.

Tom20 function determines the extent of t-2-hex induced pre-protein accumulation. Upper panels: Aim17 pre-protein accumulation in response to different t-2-hex concentrations was quantified in wild type and tom20Δ cells. Cells were either untreated, or treated with the indicated lipid concentrations or DMSO for 20 min. The percentage of pre-protein relative to total Aim17 protein was calculated in the lower panel (n = 3). *p<0.05, **p<0.01 by Student's unpaired t-test, ns = not significant.