|Home | About | Journals | Submit | Contact Us | Français|
Mitochondria play a pivotal role in apoptosis in multicellular organisms by releasing apoptogenic factors such as cytochrome c that activate the caspases effector pathway, and apoptosis-inducing factor (AIF) that is involved in a caspase-independent cell death pathway. Here we report that cell death in the single-celled organism Dictyostelium discoideum involves early disruption of mitochondrial transmembrane potential (ΔΨm) that precedes the induction of several apoptosis-like features, including exposure of the phosphatidyl residues at the external surface of the plasma membrane, an intense vacuolization, a fragmentation of DNA into large fragments, an autophagy, and the release of apoptotic corpses that are engulfed by neighboring cells. We have cloned a Dictyostelium homolog of mammalian AIF that is localized into mitochondria and is translocated from the mitochondria to the cytoplasm and the nucleus after the onset of cell death. Cytoplasmic extracts from dying Dictyostelium cells trigger the breakdown of isolated mammalian and Dictyostelium nuclei in a cell-free system, and this process is inhibited by a polyclonal antibody specific for Dictyostelium discoideum apoptosis-inducing factor (DdAIF), suggesting that DdAIF is involved in DNA degradation during Dictyostelium cell death. Our findings indicate that the cell death pathway in Dictyostelium involves mitochondria and an AIF homolog, suggesting the evolutionary conservation of at least part of the cell death pathway in unicellular and multicellular organisms.
Programmed cell death (PCD) is a genetically regulated physiological process of cell suicide that is central to the development and homeostasis of multicellular organisms (Raff, 1992 ; Steller, 1995 ; Jacobson et al., 1997 ; Vaux and Korsmeyer, 1999 ). The basic machinery that controls the onset of PCD in roundworms (Caenorhabditis elegans), insects (Drosophila melanogaster), and vertebrates (mammals) appears to be present in all cells, at all times. Crucial aspects of PCD appear to be conserved, including both the genes encoding the basic cell death machinery, and the morphological and biochemical features of apoptosis, the most frequent phenotype of PCD (Jacobson et al., 1997 ; Horvitz, 1999 ; Song and Steller, 1999 ; Vaux and Korsmeyer, 1999 ).
Mitochondria play a pivotal role in PCD in mammalian cells, in particular through the permeabilization/disruption of their outer membrane, with (or followed by) the loss of mitochondrial transmembrane potential (ΔΨm) (Kroemer et al., 1995 ; Green and Reed, 1998 ; Petit et al., 1998 ; Goldstein et al., 2000 ; Martinou et al., 2000 ), leading to the release of cytochrome c (Liu et al., 1996 ) and apoptosis-inducing factor (AIF) into the cytosol (Susin et al., 1999 ). The cytochrome c takes part in the activation of caspases, which are major effectors of PCD (Thornberry and Lazebnik, 1998 ), whereas AIF is involved in a caspase-independent cell death pathway (Susin et al., 1999 ).
Although it was initially assumed that PCD arose with multicellularity and would have been counterselected in unicellular organisms (Raff, 1992 ; Vaux et al., 1994 ; Evan et al., 1995 ; Steller, 1995 ) several recent findings indicate that a process of PCD also operates in single-celled eukaryotes (Ameisen, 1996 ). This has now been described in six species of unicellular eukaryotes, whose phylogenic origins arose 1 to 2 billion years ago. These are the free-living slime mold Dictyostelium discoideum (Cornillon et al., 1994 ); the kinetoplastid parasites Trypanosoma cruzi (Ameisen et al., 1995 ), Trypanosoma brucei rhodensiense (Welburn et al., 1996 ), and Leishmania amazonensis (Moreira et al., 1996 ); the free-living ciliate Tetrahymena thermophila (Christensen et al., 1995 ); and the dinoflagellate Peridinium gutanense (Vardi et al., 1999 ).
The cell death phenotype in unicellular eukaryotes is similar (D. discoideum) (Cornillon et al., 1994 ) or almost identical (T. cruzi) (Ameisen, 1996 ) to the PCD phenotype of multicellular animals or plants. However, nothing is currently known about the cell death machinery or the genetic control of PCD operating in these single-celled eukaryotes, apart from the fact that developmental cell death appears to be caspase-independent in D. discoideum (Olie et al., 1998 ). Thus, we cannot say how different or similar the processes are in unicellular and multicellular animals.
This report shows that cell death in Dictyostelium has several features in common with mammalian cell PCD, including a loss of mitochondrial ΔΨm followed by the exposure of cell surface phosphatidyl serine, the loss of nuclear DNA, and the engulfment of dying cells by neighboring cells. We have cloned and characterized one of the putative effectors involved in DNA degradation during cell death, a Dictyostelium homolog of mammalian AIF (DdAIF). It is released into the cytosol and targeted to the nucleus during cell death mediated by protoporphyrin IX (PPIX) and during developmental cell death induced by differentiation-inducing factor-1 (DIF-1). Cytoplasmic extracts from dying Dictyostelium cells triggered the partial degradation of isolated mammalian and Dictyostelium nuclei in a cell-free system. This process was prevented by immunodepletion with the use of a polyclonal anti-DdAIF antibody. These findings indicate that the cell death pathway of Dictyostelium amoebae involves mitochondria and an AIF homolog, and therefore may have evolved from the same ancestor as the cell death pathway of mammalian cells.
D. discoideum cells, from the cloned Ax-2 strain (Watts and Ashworth, 1970 ), were grown in suspension in HL5 medium (Susman, 1987 ) on a gyratory shaker (150 rpm) at 22–23°C in a water-saturated atmosphere. Cultures were grown in 50-ml flasks with proper oxygenation (culture volume was 1/5 of the total). Conditioned media were prepared by starving a suspension of 4 × 107 Dictyostelium cells/ml in Soerensen buffer (100 mM Na2HPO4, 735 mM KH2PO4, 17 mM phosphate, final solution pH 6.8) for 24 h, on a gyratory shaker (150 rpm) at 22°C. A cell-free supernatant was prepared 22 h after initiation of starvation, by centrifuging the starved cell suspension at 700 × g for 5 min. These supernatants were immediately frozen and kept at −20°C. The exocytotic vesicles were prepared from cells starved for 22 h by centrifugation at 700 × g then membranes were discarded by centrifugation at 1500 × g, exocytotic vesicles were collected by centrifugation at 6,500 × g. Experiments on developmental cell death were performed in 50-ml plastic flasks (Kay et al., 1987 ). Briefly, logarithmically growing Dictyostelium cells were washed twice with Soerensen buffer (pH 6.0), incubated in the absence or presence of 3 mM cAMP (Sigma, St. Louis, MO) at 106 cells/ml for 8 h, and then treated with DIF-1 [1-(13,5-dichloro-2,6-dihydroxy-4-methoxyphenyl)-1-hexanone; Molecular Probes, Eugene, OR] (100 nM for 16 h).
The mitochondrial transmembrane potential (ΔΨm) was measured by incubating cells (5 × 105/ml) with 3,3′-dihexyloxacarbicyanine iodide [DiOC6(3); Molecular Probes; final concentration 2.5 nM (Petit et al., 1995 ) with modifications (Rottenberg and Wu, 1998 )] or 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazol carbocyanine iodide (JC-1) staining, which was performed as previously described (Kuhnel et al., 1997 ). Cells were analyzed with the use of a FACS Vantage (BD Biosciences, San José, CA), gating the forward and side scatter to exclude debris. The fluorescence was excited with an argon laser (excitation wavelength 488 nm) and collected in FL-1 [band pass 530 ± 30 nm for DiOC6(3) and JC-1 green] or FL-2 (band pass 585 ± 20 nm for JC-1 red), after suitable compensation. A minimum of 5 × 103 events was acquired in list mode and analyzed with Cellquest software (BD Biosciences).
Phosphatidylserine (PS) exposed on the outer plasma membrane was measured by staining cells with annexin V-fluorescein isothiocyanate (FITC) (1 μg/ml, 10 min, 4°C; Immunotech, Marseille, France). Relative DNA content was assessed with the use of propidium iodide (1 mg/ml) with 2% saponin.
Viability of the Dictyostelium cells was assessed with the use of TO-PRO-3 (Molecular Probes) at 1 μg/ml final in FL4 (661 ± 16 nm) to avoid any red fluorescence from the accumulated PPIX
Dictyostelium cells were fixed in 1.25% glutaraldehyde buffered with 0.1 M sodium phosphate (pH 7.4) for 24 h at 4°C, dehydrated with ethanol at 4°C, and immersed in a 1:1 mixture of propylene oxide and Epon. They were embedded in Epon by polymerization at 60°C for 48 h and examined under the electron microscope (Ryter and de Chastellier, 1977 ).
Cell suspensions were fixed for at least 24 h in a 1.25% (vol/vol) glutaraldehyde in 0.1 M sodium phosphate (pH 7.4). Aliquots were filtered through a (diameter 25 mm, 0.2 μm) Anodisc (Whatman, Maidstone, United Kingdom) and the filters were rinsed five times for 10 min in the sodium phosphate buffer. Cells were postfixed for 2 h in 1% osmium tetroxide in sodium phosphate, rinsed five times in Ultrapure water, dehydrated in a graded ethanol series (50, 70, 95, and 100% twice), soaked in isopentyl acetate, and critical point dried in a CO2 medium with an EMSDCOPE CPD 750 apparatus. The dried cells were sputter-coated with gold-palladium and examined in a JEOL JSM35CF operating at 10 kV, or an S 3000N Hitachi operating at 15 kV.
Liposomes were prepared as follows: 1-palmitoyl-2-oleoylphosphatidylcholine (POPC) was purchased from Avanti Polar Lipids (Alabaster, AL) and PPIX from Sigma. Twenty microliters of 100 μM PPIX solution was added to 320 μl of 100 mM POPC in chloroform. The mixture was dried under nitrogen and vacuum. Multilamellar vesicles were obtained by rehydration in 1 ml of phosphate-buffered saline (PBS), and sonication at 40 W for 3 min with a B-12 Sonifier (Branson, Danbury, CO). Final concentrations were 2 μM PPIX and 32 mM POPC.
DdAIF was identified by a search for homology with mammalian AIF in the Cellular Slime Molds cDNA with the use of TBLASTN (National Center for Biotechnology Information, Bethesda, MD). A cDNA clone named SLB348 was obtained. The 5′ end of DdAIF was cloned by polymerase chain reaction (PCR) from a D. discoideum cDNA library (kindly provided by D. Fuller, Loomis lab, University of California at San Diego, La Jolla, CA) with the use of gene-specific primers and anchor primers. The PCR product obtained was cloned into a pGEM-T vector (Promega, Madison, WI), sequenced, and fused to SLB348 to obtain full-length DdAIF cDNA. The accession number of DdAIF is AJ272500 (EMBL Nucleotide Sequence Database).
Polyclonal anti-DdAIF sera were obtained from rabbits immunized with a mixture of two DdAIF peptides (amino acids 221–234 and 465–478, coupled to keyhole limpet hemocyanin). Cytosol and heavy membranes from Dictyostelium cells were separated on a 4/20% polyacrylamide gel (Bio-Rad, Hercules, CA) and then transferred to polyvinylidene difluoride (Bio-Rad). The membrane was immunoblotted with rabbit anti-DdAIF (1/1000) and visualized with horseradish peroxidase-conjugated sheep anti-rabbit IgG F(ab′)2 fragment (Amersham Pharmacia Biotech UK, Little Chalfont, Buckinghamshire, United Kingdom), followed by enhanced chemiluminescence (Amersham Pharmacia Biotech UK). For immunofluorescence microscopy, cells were fixed in methanol (5 min at −15°C) and permeabilized in 0.1% Triton X-100. They were then incubated with anti-DdAIF antibodies (1/100 in PBS, 1% bovine serum albumin) or with anti-heat shock protein (Hsp)-60 antibodies (Stressgen, Victoria, BC, Canada; 1/200 in PBS, 1% BSA) for 2 h at room temperature. Bound antibodies were vizualized with FITC-labeled antirabbit antibodies (Sigma) or tetramethylrhodamine B isothiocyanate-labeled anti-mouse antibodies (Sigma).
DNA was extracted from mammalian cells or Dictyostelium cells as previously described (Martin et al., 1995 ). The pulsed gel electrophoresis was performed with the use of 1.1% agorose gels in 0.5× Tris borate-EDTA buffer on a Gene Navigator (Amersham Pharmacia Biotech UK) for 17 h at 226 V with phase A pulse 0.5 s/4 h, phase B 1.0 s/9 h, and phase C pulse 2.0 s/4 h. Gels were stained for 15 min in ethidium bromide with agitation.
Cytoplasmic extracts of Dictyostelium cells and Jurkat cells, nuclei from CEM cells, and the reconstituted cell-free extracts were prepared as described previously (Martin et al., 1995 ). Cytoplasmic extracts were prepared as follows: cells were washed twice in PBS and incubated on ice for 20 min with cell extract buffer (50 mM piperazine-N,N′-bis(2-ethanesulfonic acid) [PIPES], pH 7.4, 50 mM KCl, 5 mM EGTA, 2 mM MgCl2, 1 mM dithiothreitol [DTT], 10 μM cytochalasin B, and 1 mM phenylmethylsulfonyl fluoride [PMSF]). They were lysed by homogenization with a B-type pestle. Lysis was monitored under a phase contrast microscope. Cell lysate was first centrifugated for 5 min at 800 × g to eliminate nuclei and unbroken cells and then centrifuged at 4°C for 15 min at 17,000 × g. The clear cytosol (corresponding to light fraction) was carefully removed removed and stored at −80°C. CEM nuclei were prepared as follows; CEM cells were washed twice in PBS and once with nuclei isolation buffer (NB: 10 mM PIPES, pH 7.4, 10 mM KCl, 2 mM MgCl2, 1 mM DTT, 10 μM cytochalasin B, and 1 mM PMSF); they were suspended in this buffer, allowed to swell on ice for 20 min, and gently lysed with a Dounce homogenizer. Liberated nuclei were then layered >30% sucrose in NB and centrifuged at 800 × g for 10 min, followed by washing in NB and suspension in nucleus storage buffer (10 mM PIPES, pH 7.4, 80 mM KCl, 20 mM NaCl, 250 mM sucrose, 5 mM EGTA, 1 mM DTT, 0.5 mM spermidine, 0.2 mM spermine, 1 mM PMSF, and 50% glycerol) at 2 × 108 nuclei/ml. Nuclei were stored at −80°C.
Dictyostelium nuclei were prepared as described by Charlesworth and Parish (1975) with a modified lysis medium composed of 10 mM PIPES pH 7.9, 200 mM NaCl, 300 mM NaF, 1 mM DTT, 1 mM ATP, 0.5 μM leupeptin, 2.5 μM pepstatin, 0.53 g of β-glycerophosphate, 200 μl of saturated solution of vanadate, 0.1 mM PMSF for 100 ml. Cells were lysed, nuclei were purified on sucrose gradients and then stored in nucleus storage buffer.
Cell-free reactions (25 μl) were performed with the use of 20 μl of cytoplasmic extract (20–30 mg/ml protein), 1 μl (2 × 105) of nuclei, and 4 μl of extract dilution buffer (10 mM HEPES, 50 mM NaCl, 2 mM MgCl2, 5 mM EGTA, 1 mM DTT, 2 mM ATP, 10 mM phosphocreatine, and 50 μg/ml creatine kinase). For flow cytometric analysis of the nuclei, propidium iodide (1 μg/ml) was added for 30 min and then the nuclei analyzed with the use of the FL3 LP pass filter (FACScalibur 4C; BD Biosciences).
Dictyostelium cytoplasmic extracts (CEs) were immunodepleted by diluting anti-DdAIF serum and nonimmune serum to 1/200 in 0.1 M Tris-HCl pH 9.6 buffer and coating 96-well plates (Maxisorp; Nunc, Wiesbaden, Germany) by incubation overnight at 4°C. The plates were washed four times with PBS. Dictyostelium cytoplasmic extracts were incubated for 5× 1 h in the wells at 4°C.
Exocytosis, a potential mechanism of detoxification, occurs in Dictyostelium cultures and leads to the shedding of vesicles (Tatischeff et al., 1998 ). Vesicles are also released during starvation of suspended Dictyostelium cells that subsequently die. Living Dictyostelium cells harvested from exponentially growing cultures were incubated for 45 h with the supernatant from a suspension of Dictyostelium cells starved for 22 h and containing a high concentration of exocytotic vesicles. Epifluorescence microscopy and flow cytometric analysis demonstrated that >85% of the Dictyostelium cells were dead at 43 h (Figure (Figure1a)1a) and 99% at 45 h (Figure (Figure1b),1b), whereas incubation for 45 h in starvation medium alone caused the spontaneous death of <10% of cells (data not shown).
Plasma membrane integrity is a feature that distinguishes apoptotic from necrotic cell death. Necrotic cells versus apoptotic cells were assessed with TO-PRO-3 staining. This stain only enters into necrotic cells, and as shown in Figure Figure1a1a only 4% of the cells were stained by TO-PRO-3 at 43 h, whereas 85% exposed their PS residues as shown by annexin V staining. Living Dictyostelium cells are round, and there was no significant increase in the cell granulometry (side scatter) of dying Dictyostelium cells (Figure (Figure1b).1b). Cell death was associated with several other apoptosis-like features as well as cell shrinkage (transmission light microscopy, Figure Figure1a),1a), reduced forward scatter and side scatter in flow cytometry (Figure (Figure1b),1b), and loss of mitochondrial ΔΨm measured with tetramethylrhodamine ethyl ester (epifluorescence microscopy; Figure Figure1a),1a), or with JC-1 probes (flow cytometric analysis; Figure Figure1b)1b) (Vayssière et al., 1994 ; Petit et al., 1995 ; Zamzami et al., 1995 ). Almost all (85.7%) of the dying cells had a low mitochondrial ΔΨm, compared with <2.5% of control cells. Control experiments showed that carbamoyl cyanide m-chlorophenylhydrazone (mClCCP), an uncoupler of oxidative phosphorylation, abolished the JC-1 dye uptake, demonstrating that this is driven by the ΔΨm (Figure (Figure11b).
Cell death was also associated with an externalization of PS (Figure (Figure1,1, c and d) at the plasma membrane in 94% of dying cells at 45 h compared with 8.5% of cells at 35 h (Figure (Figure1d).1d). A kinetic analysis (Figure (Figure1c)1c) revealed that PS was exposed after the loss of ΔΨm, as in mammalian cells (Kroemer et al., 1995 ; Petit et al., 1995 ; Zamzami et al., 1995 ). Dictyostelium cells began to lose ΔΨm 31 h after incubation with the supernatant from 22-h Dictyostelium culture containing exocytotic vesicles. Most of the cells (97%) had lost their ΔΨm after 37 h, whereas the PS exposure had just begun. Dictyostelium cells (58.8%) had exposed their PS at 40 h, by at which time they had all lost their ΔΨm (Figure (Figure1c).1c). At 45 h most of the cells had externalized PS.
We also observed that Dictyostelium cell death was associated with a loss of nuclear DNA (59.3 vs. 6% in control cells) (Figure (Figure1e).1e). This loss of nuclear DNA occur slightly later than the exposure of phosphatidyl residues on the external surface of the plasma membrane. The increased vacuolization that is evidenced in Figure Figure2a2a also appeared to increase simultaneously with the dramatic mitochondrial drop as soon as the 50% mitochondrial decrease is reached at 37 h (Figure (Figure2c)2c) but could be detected as early as 34 h (Figure (Figure1c,1c, arrow). The cellular condensation was almost parallel to the loss of mitochondrial membrane potential (Figure (Figure1c,1c, stars).
Transmission and scanning electron microscopy of the dying Dictyostelium cells showed that they had a phenotype presenting most of the characteristic features of apoptosis (Figure (Figure2a),2a), including a progressive cell shrinkage (Figure (Figure2,2, a and c), intense autophagic vacuolization (Figure (Figure2,2, a and b), and blebbing of the cell surface (Figure (Figure2c).2c). The development of autophagic vacuolization consisted of vacuoles containing damaged mitochondria with an electron-dense matrix and an altered outer mitochondrial membrane (Figure (Figure2b),2b), which was either fully digested or released with the extracellular vesicles.
The dying Dictyostelium cells released large vesicles (~500 nm in diameter) that resembled apoptotic bodies (Figure (Figure2c).2c). These bodies sometimes enclosed part of the nucleus and contained large amounts of DNA (Tatischeff et al., 1998 ). Pseudoapoptotic bodies and/or dying cells were also ingested by neighboring cells, indicating that Dictyostelium, a unicellular organism, can perform this function (Figure (Figure2d2d left), similar to multicellular animals. The engulfed cells or fragments of cells were found in digestive vacuoles, where they were degraded (Figure (Figure2d,2d, left). This is the first time that such engulfment has been reported in unicellular eukaryotic cell death. These observations suggest the existence of Dictyostelium homologs of the engulfment machinery as described during apoptosis in multicellular animals.
A nucleus characteristic of a living Dictyostelium cell is shown in Figure Figure2a2a (t = 0). Nucleoli, with highly condensed chromatin, are linked to the nuclear membrane and the nuclear chromatin is uniformly distributed within the nucleus. We observed that spots of condensed nuclear chromatin appeared during Dictyostelium cell death, and nucleoli were damaged, shown by the fragmented state of their condensed chromatin (Figure (Figure2b,2b, damaged nuclei). However, the chromatin was not fully fragmented (Figure (Figure6c),6c), and there was no oligonucleosomal DNA degradation as assessed on agarose gel electrophoresis. The mitochondria of Dictyostelium underwent metabolic changes during death, as indicated by the drop in mitochondrial membrane potential (Figure (Figure1,1, a and b). When mitochondria were not found in autophagic vacuoles (Figure (Figure2b)2b) and were present in the cytoplasm, they appeared to be condensed with their outer membrane partly disrupted (Figure (Figure2e),2e), suggesting that intermembrane space proteins might have been released into the cytosol during cell death. The “apoptosis-like ” phenotype was not specific of the cell death induced by exocytotic vesicles, because actinomycin D, another potent inducer of apoptosis, produced similar changes (unpublished data).
We purified exocytotic vesicles (Figure (Figure3a)3a) or Dictyostelium exosomes by differential centrifugation. The vesicles were mainly spherical (30–100 nm in diameter) (Figure (Figure3a)3a) and were enriched in a 17-kDa protein that was identified as ponticulin, as assessed by Western blotting (unpublished data), an integral membrane glycoprotein that binds to F-actin and the nuclear actin assembly (Hitt et al., 1994 ). High-performance liquid chromatography (HPLC) and spectral analysis of shed vesicles revealed (Figure (Figure3,3, b and c) that they contained PPIX. This type of porphyrin has previously been reported to cause apoptosis in mammalian cells by a process that requires photoactivation (Noodt et al., 1996 ). Exogenous PPIX (that enters the cells) has also been found mainly in membranes, including mitochondrial membranes, and it triggers the release of apoptogenic factors such as cytochrome c and AIF (Marchetti et al., 1996 ). PPIX has a red fluorescence, with a major emission at 637 nm and a minor emission peak at 705 nm (Figure (Figure3c).3c). We used this property to assess the intracellular accumulation of PPIX during Dictyostelium cell death. The exposure of PS by Dictyostelium cells incubated with purified exocytotic vesicles was accompanied by an increase in the red fluorescence intensity, indicating an accumulation of PPIX (Figure (Figure3d).3d). Most of the cells (93.4%) accumulated PPIX and exposed PS at 45 h, whereas 80.4% did so at 40 h and only 8.5% did so at 35 h. We then incubated Dictyostelium cells with either purified exocytotic vesicles or with liposomes loaded with purified PPIX. More than 90% of the cells were dying or dead in both cases after incubation for 6 h with exposure to light (15 min at 3 J/cm2), as assessed by flow cytometry (Figure (Figure3e),3e), whereas exposure to light in the absence of purified exocytotic vesicles or liposome loaded with PPIX induced death in <5% of the cells (Figure (Figure3e).3e). Again, cell death was accompanied by cell shrinkage, loss of ΔΨm (Figure (Figure3e),3e), exposure of PS, and loss of nuclear DNA (unpublished data). Incubation of Dictyostelium cells with pure PPIX alone (5 μM) for a similar time and exposure to light also caused cell death (unpublished data). The finding that purified exocytosis vesicles and PPIX required light to cause Dictyostelium apoptosis, whereas the 22-h Dictyostelium supernatant containing exocytosis vesicles caused death without such exposure indicates that the supernatant contains other unidentified light-independent inducers of Dictyostelium cell death.
DNA degradation during apoptosis can be due either to caspases activation (Nagata, 2000 ) or to caspase-independent effectors, such as AIF (Susin et al., 1999 ). One major difference is that caspases are involved in stage II of nuclear apoptosis, i.e., oligonucleosomal DNA fragmentation (180 bp) and that AIF is involved in stage I of nuclear apoptosis, i.e., perinuclear chromatin condensation and large scale DNA fragmentation (several kbp) (Susin et al., 2000 ).
We observed a DNA degradation during Dictyostelium cell death (Figure (Figure1e).1e). Incubation of Dictyostelium cells with peptide inhibitors of caspases (zVAD-fmk, ac-DEVD-CHO, YVAD-fmk), with broad-spectrum cysteine-proteinases/calpain inhibitors (leupeptin, E64), or with a cathepsin inhibitor (FA-fmk) did not prevent the DNA degradation during Dictyostelium cell death, suggesting that this process is caspase and cysteine proteinase independent. Moreover, we did not observe oligonucleosomal DNA fragmentation.
During Dictyostelium cell death, the cells show nuclear features of apoptosis similar to those of stage I (Figure (Figure2b),2b), and the DNA is degraded on a large scale (Figure (Figure6c)6c) in a caspase-independent manner, suggesting the involvement of an AIF as nuclear effector.
A TBLASTN search disclosed a homologous partial cDNA clone (clone SLB348, kindly provided by T. Morio, University of Tsukuba, Tsukuba, Japan) in the Dictyostelium cDNA database of Japan (University of Tsukuba); which is very similar to mammalian AIF and confirming that AIF seems to be conserved across several phyla (e.g., Schizosaccharomyces pombe, Drosophila melanogaster, Caenorhabditis elegans, Arabidopsis thaliana, mammals) (Lorenzo et al., 1999 ). We obtained the full-length cDNA by cloning the 5′ end by PCR from a Dictyostelium cDNA library, with the use of gene-specific and vector-specific primers. The final 1.7-kb cDNA encoded a protein of ~60-kDa (Figure (Figure4a).4a). We generated anti-DdAIF polyclonal antibodies by immunizing rabbits with a mixture of two DdAIF peptides. The molecular weight of the mature form was ~53 kDa, as assessed by Western blotting (Figure (Figure4b).4b). DdAIF shares >30% identity and 60% similarity with human AIF (Figure (Figure4c).4c). Like mammalian AIF, DdAIF contains a mitochondria localization site (MLS) (Claros and Vincens, 1996 ) and a putative nuclear localization site (NLS) (Boulikas, 1993 ). DdAIF also contains a putative helix-turn-helix DNA binding motif (Dodd and Egan, 1990 ) at its C-terminal end (Figure (Figure4a),4a), suggesting that DdAIF may bind to DNA. All the amino acids believed to interact with the prosthetic groups FAD and NAD are present in DdAIF, as in human AIF (Lorenzo et al., 1999 ) (Figure (Figure4c).4c).
In mammalian cells, AIF has a mitochondrial localization and upon cell death induction, AIF is released to the cytosol and then targeted to the nucleus (Susin et al., 1999 ). Immunofluorescence microscopy showed that in living Dictyostelium cells, DdAIF, like Hsp 60 (a protein of the mitochondrial matrix), was associated with the mitochondria. On triggering cell death, DdAIF was released into the cytosol and secondarily targeted to the nucleus after the triggering of apoptosis (Figure (Figure5a),5a), whereas Hsp 60 remained localized in the mitochondria. We used cell fractionation to confirm that DdAIF was present in heavy membranes (including mitochondria) in living Dictyostelium cells like Hsp 60 (Figure (Figure5b).5b). By cell fractionation, we also confirmed that DdAIF was translocated from the heavy membranes to the cytosol after treatment with PPIX or actinomycin, whereas Hsp 60 was still present in the heavy membranes (Figure (Figure5c).5c). These data show that DdAIF, like mammalian AIF, is translocated from mitochondria to the cytosol and the nucleus during cell death.
D. discoideum can undergo, upon adverse environmental conditions, a complex developmental process, depending on DIF-1, cAMP secretion, and cAMP responsive gene expression, resulting in the emergence of a multicellular aggregated body (the fruiting body) made up of 20% dead stalk cells forming the stalk wall that supports the 80% viable spores (Soderbom and Loomis, 1998 ; Brown and Firtel, 1999 ; Thomason et al., 1999 ).
Dictyostelium cells treated in the presence of DIF alone, or cAMP plus DIF, exhibit a decrease of their forward and side scatter and show a markedly decreased mitochondrial membrane potential in up to 74 and 83% of cells, respectively (Figure (Figure6a).6a). The loss of mitochondrial ΔΨm was associated with an externalization of PS. Indeed, only 3% of the control living cells bound annexin-V, whereas 35 and 45% of Dictyostelium cells treated with DIF-1 alone or cAMP plus DIF-1 were respectively labeled. Only 12 and 18% of the Dictyostelium cells showed plasma membrane damage during developmental cell death versus 1.5% in the control living cells as assessed with TO-PRO-3 staining. These data confirm that DIF-1 and cAMP plus DIF-1 cause mainly an “apoptotic-like” cell death in Dictyostelium cells as described by Cornillon et al. (1994) .
Developmental cell death, as cell death induced by PPIX, was also associated with DdAIF translocation from mitochondria to the cytosol (Figure (Figure6b)6b) and with large-scale DNA fragmentation, as assessed with pulse field gel electrophoresis (Figure (Figure6c).6c). DdAIF is released from mitochondria during developmental cell death, but the amounts are smaller than those released by Dictyostelium cells treated with PPIX or actinomycin D (Figure (Figure66b).
We used a cell-free system (Martin et al., 1995 ) to explore the role of DdAIF as a nuclear effector. Mammalian nuclei (isolated from CEM cells) were incubated for 4–5 h with cytoplasmic extracts of Dictyostelium. Cytoplasmic extracts from dying cells (cell death mediated by PPIX) caused the condensation and partial degradation of CEM nuclei (86,4% of nuclei with low DNA content), whereas cytosol from living untreated cells did not (20.4% of nuclei with low DNA content) (Figure (Figure7a).7a). Nuclear damage was not prevented by treating the cytoplasmic extracts with various proteinase inhibitors (zVAD-fmk, ac-DEVD-CHO, YVAD-fmk, leupeptin, E64, and FA-fmk) (unpublished data), suggesting that these early chromatin changes caused by cytoplasmic extracts from dying Dictyostelium cells did not depend on caspase/proteinase activation (whether in the cytoplasmic extracts or in the nuclei). Concordant with these findings, nuclear chromatin was not fully fragmented by cytoplasmic extracts from dying Dictyostelium cells, in contrast to the full fragmentation caused by cytoplasmic extracts from human Jurkat cells treated with an agonistic Fas antibody (Figure (Figure7b).7b). The chromatin changes caused in the cell-free system by the cytoplasmic extract from dying Dictyostelium cells seemed to depend mostly on DdAIF, because the condensation and partial degradation of CEM nuclei were prevented by immunodepletion of the cytoplasmic extracts with the polyclonal anti-DdAIF antibody (27% of nuclei with low DNA content), but not with nonimmune serum (81.6% of nuclei with low DNA content) (Figure (Figure7a).7a). However, DdAIF may not be the sole nuclear effector(s) in the cytoplasmic extracts extract from dying Dictyostelium cells, because late (>6 h after incubation) nuclei damage was observed even after immunodepleting the cytosol with the anti-DdAIF antibody. Pulsed field gel electrophoresis demonstrated that cytoplasmic extracts from dying Dictyostelium cells caused large-scale DNA fragmentation (50–24 kbp) of mammalian cell nuclei (Figure (Figure7c,7c, lane 2) and that the fragmentation was prevented after immunodepletion with the DdAIF antibody (lane 4) but not with the nonimmune serum (lane 3). This fragmentation did not occur with cytoplasmic extracts from control cells (lane 1). z-VAD.fmk (100 μM) did not inhibit the DNA fragmentation caused by cytoplasmic extracts from dying cells confirming that the large DNA fragmentation was a caspase-independent process (lane 5).
We also incubated isolated Dictyostelium nuclei with cytoplasmic extract from dying Dictyostelium cells. Epifluorescence microscopy (Figure (Figure7d)7d) showed that cytoplasmic extracts from dying Dictyostelium cells caused the partial degradation of the nuclei with the dissipation of nucleoli, whereas cytoplasmic extracts from living untreated cells did not. Once again, DNA damage was not prevented by treating the cytoplasmic extracts with various proteinases inhibitors (unpublished data). The changes in nuclei caused by the Dictyostelium cytoplasmic extracts seemed to depend mostly on DdAIF in the cell-free system, because the partial degradation of Dictyostelium nuclei was prevented by immunodepleting the cytoplasmic extracts with an anti-DdAIF polyclonal antibody but not with nonimmune serum (Figure (Figure7,7, d and e).
Together, our data suggest that DdAIF has a conserved nuclear apoptosis function and is one of the effectors needed for degradation of the nucleus during Dictyostelium cell death.
In mammalian cells, PCD or apoptosis is a cell suicide process that depends on two major executionary pathways, one involving proteolytic activation of effector caspase proteases, the other involving mitochondria outer membrane permeabilization, leading to the release in the cytosol of intermembrane space proteins such as cytochrome c and AIF (Green and Reed, 1998 ; Martinou et al., 2000 ). Although both pathways usually operate together and amplify each other (cytochrome c inducing caspase activation, and activated caspases inducing mitochondria permeabilization [Green and Reed, 1998 ; Marzo et al., 1998 ; Thornberry and Lazebnik, 1998 ]), caspase activity, in several instances, is not required for the execution of cell death (Borner and Monney, 1999 ), and AIF has been proposed to be one of the candidates involved in the caspase-independent cell death pathway (Susin et al., 1999 ).
Our findings indicate that Dictyostelium cells can undergo cell death that shares essential features with mammalian cell apoptosis. This involves a loss of ΔΨm, resulting in the release of the Dictyostelium homolog of AIF from the mitochondria. The loss of ΔΨm precedes nuclear shrinkage and extranucleolar chromatin condensation, the dispersion and partial fragmentation of nuclear chromatin, the exposure of phosphatidylserine at the cell surface, and engulfment of dying cells by healthy or dying neighboring cells. Neither Dictyostelium cell death nor its apoptotic features was blocked by the broad caspase inhibitor zVAD-fmk (unpublished data), confirming a previous report that Dictyostelium cell death appears to be caspase independent (Olie et al., 1998 ). The involvement of DdAIF in at least some of the apoptotic features of dying Dictyostelium cells, e.g., the nuclear chromatin condensation and partial DNA fragmentation, is suggested by our observation that in a cell-free system, cytoplasmic extracts from dying Dictyostelium cells induced similar changes in isolated nuclei, and these were prevented when DdAIF was immunodepleted from the extracts with the use of an anti-DdAIF antibody. This is, to our knowledge, the first evidence of the existence of phylogenic conservation between the basic cell death pathway (DNA degradation) that operates in a single-celled eukaryotic organism and part at least of the cell death pathway which operates in cells of metazoan organisms, particularly from mammals.
Dictyostelium is a single-celled organism that can undergo a complex developmental process under adverse environmental conditions that depends on a factor of differentiation (DIF-1), on cAMP secretion, and on cAMP responsive gene expression. These changes involve cell chemotaxis and aggregation and differentiation into two main cell populations, the stalk cells and the spore cells. This results in the emergence of a multicellular aggregated body (the fruiting body) made up of ~20% dead stalk cells that support the 80% of viable spores (Soderbom and Loomis, 1998 ; Brown and Firtel, 1999 ; Thomason et al., 1999 ). Terminal differentiation into stalk cells has been reported to be a caspase-independent form of PCD (Cornillon et al., 1994 ; Olie et al., 1998 ), with nuclear chromatin condensation, cytoplasmic vacuolization, and the formation of a rigid cellulose cell wall. The resulting stalk cell corpses retain their structural integrity and are not engulfed by neighboring cells. This process is similar to the developmental cell death in Myxobacterium and to the terminal differentiation of bark cells in multicellular plants (Ameisen, 1998 ). This developmentally regulated PCD has been the only cell death process studied in Dictyostelium until now. Our findings show that Dictyostelium can undergo a form of “developmental default ” cell death, a cell death under conditions that do not lead to either cell aggregation or to stalk cell differentiation. Both processes of cell death share crucial features with metazoan cell apoptosis, in that they involves permeabilization/disruption of the mitochondrial membrane, the release of DdAIF, and DNA fragmentation in large scale. However, cell death induced by PPIX leads to the engulfment of the dying cell by neighboring cells but not during developmental cell death.
DdAIF is very similar to mammalian AIF, in particular in its phylogenetically conserved oxidoreductase domain. This domain appears to be required for the nuclear apoptogenic function of AIF (Lorenzo et al., 1999 ; Susin et al., 1999 ). DdAIF also contains an MLS and putative NLSs, as does mammalian AIF. The nuclear apoptosis function of DdAIF is apparently conserved because cytoplasmic extracts from dying Dictyostelium cells caused chromatin changes and large-scale DNA fragmentation in a caspase-independent manner in isolated human nuclei in a cell-free system. These changes were similar to those induced by recombinant murine AIF (Susin et al., 1999 ) and were prevented by immunodepletion of DdAIF from the Dictyostelium cytoplasmic extracts.
Our findings are thus consistent with the idea that mitochondria play an evolutionary conserved role in the control of cell suicide (Ameisen et al., 1995 ; Green and Reed, 1998 ) and that AIF may be an ancestral and phylogenetically conserved mitochondrial effector of nuclear degradation that could have been recruited to the executionary pathway earlier than caspases (Lorenzo et al., 1999 ). Such a possibility is also supported by the finding that homologs of caspases (or of proteases that cleave substrates at the same sites as caspases) appear to take part in cell differentiation processes, rather than cell death in Dictyostelium (Olie et al., 1998 ).
Dictyostelium may however represent a particular case in the evolution of apoptosis in single-celled eukaryotes. There are indeed several examples suggesting that molecular effectors recruited to the cell suicide machinery have undergone phylogenetic variation in both metazoan and single-celled eukaryote lineages. For example, although caspases are not needed for cell death in mammals under certain circumstances, caspase activity appears to be crucial in D. melanogaster and C. elegans (Horvitz, 1999 ; Song and Steller, 1999 ; White, 2000 ), indicating that the recruitment of caspases and mitochondrial effectors may have varied during vertebrate and invertebrate evolution. Apoptosis involving full chromatin and oligonucleosomal DNA fragmentation has also been described in single-celled eukaryotes such as the kinetoplastides (Ameisen, 1996 ), whose phylogenetic divergence predates that of Dictyostelium by several hundred million years, suggesting that these organisms may have effectors of nuclear apoptosis other than AIF.
In summary, although the phylogeny of the eukaryote cell death machinery remains an open question that will require the investigation of cell suicide in other single-celled eukaryotes, our findings provide evidence for the existence of shared ancestry, rather than evolutionary convergence, in some of the molecular mechanisms of apoptosis that operate in mammalian cells and in the single celled eukayote D. discoideum. We further believe that D. discoideum is a simple, valuable model system for further studies of the function of AIF and mitochondria in cell death.
We thank Dr. T. Morio (University of Tsukoba, Tsukoba, Japan) and the Dictyostelium cDNA project in Japan (supported by Japan Society for the Promotion of Science [RFTF96L00105] and Ministry of Education, Science, Sports, and Culture of Japan ). We also thank O. Seksek (Laboratoire de Physico Chimie Biomoléculaire et Cellulaire, Paris) for the liposomes containing PPIX, Marie-France Szajnert (Institut National de la Sante et de la Recherche Medicale U129) for many interesting discussions, Frédéric Petit for technical assistance (EMI 9922 Bichat), and Dr. D. Fuller (Loomis lab, University of California at San Diego, La Jolla, CA) for Dictyostelium cDNA library. The English text was edited by Dr. Owen Parkes. This study was supported by Institut National de la Sante et de la Recherche Medicale, Centre National de la Recherche Scientifique, and the Association pour la Recherche contre le Cancer (to P.X.P.), Institut National de la Sante et de la Recherche Medicale, Paris VII University, Agence Nationale de Recherches sur le SIDA, and Sidaction (to J.C.A.), and by a grant from the Délégation Générale de l'Armement (to D.A.).