Reduction potential of yeast cytochrome c peroxidase and three distal histidine mutants: dependence on pH

The pH dependence of the Fe(III) reduction potential, E⁰′, for yeast cytochrome c peroxidase (yCcP) and three distal pocket mutants, CcP(H52L), CcP(H52Q), and CcP(R48L/W51L/H52L), has been determined between pH 4 and 8. E⁰′ values at pH 7.0 for the yCcP, CcP(H52L), CcP(H52Q), and CcP(R48L/W51L/H52L) are − 189, − 170, − 224, and − 146 mV, respectively. A heme-linked ionization in the reduced enzyme affects the reduction potential for yCcP and all three mutants. Apparent pK_A values for the heme-linked ionization are 7.5 ± 0.2, 6.5 ± 0.3, 6.4 ± 0.2, and 7.0 ± 0.3 for yCcP and the H52L, H52Q, and R48L/W51L/H52L mutants, respectively. A cooperative, two-proton ionization causing a spectroscopically-detectable transition was observed in the ferrous states of yCcP, CcP(H52L) and CcP(H52Q), with apparent pK_A values of 7.7 ± 0.2, 7.4 ± 0.1 and 7.8 ± 0.1, respectively. These data indicate that: (1) the distal histidine in CcP is not the site of proton binding upon reduction of the ferric CcP, (2) the distal histidine is not one of the two groups involved in the cooperative, two-proton ionization observed in ferrous CcP, and (3) the proton-binding site is not involved in the cooperative, two-proton ionization observed in the reduced enzyme.     

Conformational change and histidine control of heme chemistry in cytochrome c peroxidase: resonance Raman evidence from Leu-52 and Gly-181 mutants of cytochrome c peroxidase.

Resonance Raman (RR) spectra are reported for Fe(III), Fe(II), and Fe(II)CO forms of site-directed mutants of the cytochrome c peroxidase variant CCP(MI), cloned in Escherichia coli. The Fe(II) form is five-coordinate (5-c) and high-spin at low pH, but it is six-coordinate (6-c) and low-spin at high pH except when the distal His-52 residue is replaced with Leu, showing the sixth ligand to be the His-52 imidazole. Although the Leu-52 mutant stays 5-c, it does undergo an alkaline transition, as revealed by upshifts and broadening of bands assigned to vinyl C = C stretching (1620 cm-1) and C beta-vinyl bending (402 cm-1). Similar changes are seen for CCP(MI) and other mutants. Thus the alkaline transition induces a conformational change that affects the vinyl groups, probably through changes in their orientation, and that permits the His-52 imidazole to bind the Fe. The RR band arising from the stretching of the proximal Fe(II)-imidazole bond contains components at ca. 235 and 245 cm-1 for CCP(MI), which are believed to reflect a double well potential for the H-bond between the proximal His-175 imidazole and the Asp-235 carboxylate group. Loss of this H-bond by mutation of Asp-235 to Asn results in the loss of these two bands and their replacement by a single band at 205 cm-1. Although the Fe(II)-imidazole stretching mode cannot be observed in the 6-c alkaline form of the enzyme, the sixth ligand in the alkaline form of CCP(MI) is photolabile, and the status of the Fe(II)-imidazole bond can be determined in the resulting 5-c-photoproduct. For CCP(MI) at alkaline pH, the conformation change induces an increase in the 235/245-cm-1 ratio, reflecting a perturbation of the H-bond potential. In the His-52----Leu mutant, a 205-cm-1 band appears along with the 235/245-cm-1 doublet at alkaline pH, indicating partial loss of the proximal H-bond due to the distal alteration. The effect of mutations that perturb the H-bonding network that extends from the distal to the proximal side of the heme is more dramatic: at alkaline pH, His-181----Gly, Arg-48----Leu, and Trp-51----Phe mutants show an Fe(II)-imidazole stretching mode at 205 cm-1 exclusively, indicating complete loss of the proximal Asp-235-His-175 H-bond.(ABSTRACT TRUNCATED AT 400 WORDS)     

Mechanism of activation of cytochrome c peroxidase activity by cardiolipin

In this work, the actions of bovine heart cardiolipin, synthetic tetraoleyl cardiolipin, and a nonspecific anionic detergent sodium dodecyl sulfate (SDS) on cytochrome c (Cyt c) peroxidase activity recorded by chemiluminescence in the presence of luminol and on the Fe...S(Met80) bond whose presence was estimated by a weak absorption band amplitude with peak at 695-700 nm (A(695)) were compared. A strict concurrency between Fe...S(Met80) breaking (A(695)) and cytochrome peroxidase activity enhancement was shown to exist at cardiolipin/Cyt c and SDS/Cyt c molar ratios of 0 : 1 to 50 : 1 (by chemiluminescence). Nevertheless, when A(695) completely disappeared, Cyt c peroxidase activity under the action of cardiolipin was 20 times more than that under the action of SDS, and at low ligand/protein molar ratios (=4), SDS failed to activate peroxidase activity while cardiolipin enhanced Cyt c peroxidase activity 16-20-fold. A(695) did not change on Cyt c binding with liposomes consisting of tetraoleyl cardiolipin and phosphatidylcholine (1 : 10 : 10), while peroxidase activity was enhanced by a factor of 8. Breaking of 70% of the Fe...S(Met80) bonds resulted in only threefold enhancement of peroxidase activity. Cardiolipin-activated Cyt c peroxidase activity was reduced by high ionic strength solution (1 M KCl). The aggregated data suggest that cardiolipin activating action is caused, first, by a nonspecific effect of Fe...S(Met80) breaking as the result of conformational changes in the protein globule caused by the protein surface electrostatic recharging by an anionic amphiphilic molecule, and second, by a specific acceleration of the peroxidation reaction which is most likely due to enhanced heme accessibility for H(2)O(2) as a result of the hydrophobic interaction between cardiolipin and cytochrome.     

Conversion of the proximal histidine ligand to glutamine restores activity to an inactive mutant of cytochrome c peroxidase.

Using site-directed mutagenesis, a double mutant in yeast cytochrome c peroxidase (CCP) has been constructed where the proximal ligand, His175, has been converted to glutamine and the neighboring Trp191 has been converted to phenylalanine. The refined 2.4-A crystal structure of the double mutant shows that the Gln175 side chain is within coordination distance of the heme iron atom and that Phe191 occupies the same position as Trp191 in the native enzyme with very little rearrangement outside the immediate vicinity of the mutations. Consistent with earlier work, we find that the single mutant, His175-->Gln, is fully active under steady state assay conditions and that as reported earlier (Mauro et al., 1988), the Trp191-->Phe mutant exhibits only < 0.05% activity. However, the double mutant, His175-->Gln/Phe191-->Phe, exhibits 20% wild type activity. Since it is known that the Trp191-->Phe mutant is inactive because it can no longer transfer electrons from ferrocytochrome c, changing the nature of the proximal ligand is able to restore this activity. These results raise interesting questions regarding the mechanism of interprotein electron transfer reactions.     

Redesign of cytochrome c peroxidase into a manganese peroxidase: role of tryptophans in peroxidase activity.

Trp191Phe and Trp51Phe mutations have been introduced into an engineered cytochrome c peroxidase (CcP) containing a Mn(II)-binding site reported previously (MnCcP; see Yeung, B. K.-S., et al. (1997) Chem. Biol. 5, 215-221). The goal of the present study is to elucidate the role of tryptophans in peroxidase activity since CcP contains both Trp51 and Trp191 while manganese peroxidase (MnP) contains phenylalanine residues at the corresponding positions. The presence of Trp191 in CcP allows formation of a unique high-valent intermediate containing a ferryl oxo and tryptophan radical called compound I'. The absence of a tryptophan residue at this position in MnP is the main reason for the formation of an intermediate called compound I which contains a ferryl oxo and porphyrin pi-cation radical. In this study, we showed that introduction of the Trp191Phe mutation to MnCcP did not improve MnP activity (specific activity: MnCcP, 0.750 micromol min-1 mg-1; MnCcP(W191F), 0.560 micromol min-1 mg-1. k(cat)/K(m): MnCcP, 0.0517 s-1 mM-1; MnCcP(W191F), 0.0568 s-1 mM-1) despite the fact that introduction of the same mutation to WTCcP caused the formation of a transient compound I (decay rate, 60 s-1). However, introducing both the Trp191Phe and Trp51Phe mutations not only resulted in a longer lived compound I in WTCcP (decay rate, 18 s-1), but also significantly improved MnP activity in MnCcP (MnCcP(W51F, W191F): specific activity, 8.0 micromol min-1 mg-1; k(cat)/K(m), 0. 599 s-1 mM-1). The increase in activity can be attributed to the Trp51Phe mutation since MnCcP(W51F) showed significantly increased MnP activity relative to MnCcP (specific activity, 3.2 micromol min-1 mg-1; k(cat)/K(m), 0.325 s-1 mM-1). As with MnP, the activity of MnCcP(W51F, W191F) was found to increase with decreasing pH. Our results demonstrate that, while the Trp191Phe and Trp51Phe mutations both play important roles in stabilizing compound I, only the Trp51Phe mutation contributes significantly to increasing the MnP activity because this mutation increases the reactivity of compound II, whose oxidation of Mn(II) is the rate-determining step in the reaction mechanism.     

Effect of alternative distal residues on the reactivity of cytochrome c peroxidase: properties of CcP mutants H52D, H52E, H52N, and H52Q

To test the effect of alternative bases at the distal histidine position, four CcP variants have been constructed that substitute the two basic residues, aspartate and glutamate, and their amides, asparagine and glutamine, for histidine-52, i.e., CcP(H52D), CcP(H52E), CcP(H52N), and CcP(H52Q). All four mutants catalyze oxidation of ferrocytochrome c by H(2)O(2) with steady-state activities that are between 250 and 7700 times slower than wild-type CcP at pH 6.0, 0.10M ionic strength, 25°C. The rate of Compound I formation is decreased between 3.5 and 5.4 orders of magnitude for the mutants compared to wild-type CcP, with the rate of the reaction between CcP(H52Q) and H(2)O(2) the slowest yet observed for any CcP mutant. A correlation between the rate of Compound I formation and the rate of HCN binding for CcP and various CcP distal pocket mutants provides strong evidence that the rate-limiting step in CcP Compound I formation is deprotonation of H(2)O(2) within the distal heme pocket under the experimental conditions employed in this study. While CcP(H52E) reacts stoichiometrically with H(2)O(2) to form Compound I, only ~36% of CcP(H52D), ~21% of CcP(H52Q) and ~8% of CcP(H52N) appear to be converted to Compound I during their respective reactions with H(2)O(2). This is partially due to the slow rate of Compound I formation and the rapid endogenous decay of Compound I for these mutants. The pathways for the endogenous decay of Compound I for the four mutants used in this study are distinct from that of wild-type CcP Compound I.     

Singlet oxygen generation from the decomposition of alpha-linolenic acid hydroperoxide by cytochrome c and lactoperoxidase

Generation of singlet oxygen is first investigated in the decomposition of polyunsaturated lipid peroxide, alpha-linolenic acid hydroperoxide (LAOOH), by heme-proteins such as cytochrome c and lactoperoxidase. Chemiluminescence and electron spin resonance methods are used to confirm the singlet oxygen generation and quantify its yield. Decomposition products of LAOOH are characterized by HPLC-ESI-MS, which suggests that singlet oxygen is produced via the decomposition of a linear tetraoxide intermediate (Russell's mechanism). Free radicals formed in the decomposition are also identified by the electron spin resonance technique, and the results show that peroxyl, alkyl, and epoxyalkyl radicals are involved. The changes of cytochrome c and lactoperoxidase in the reaction are monitored by UV-visible spectroscopy, revealing the action of a monoelectronic and two-electronic oxidation for cytochrome c and lactoperoxidase, respectively. These results suggest that cytochrome c causes a homolytic reaction of LAOOH, generating alkoxyl radical and then peroxyl radical, which in turn releases singlet oxygen following the Russell mechanism, whereas lactoperoxidase leads to a heterolytic reaction of LAOOH, and the resulting ferryl porphyryl radical of lactoperoxidase abstracts the hydrogen atom from LAOOH to give peroxyl radical and then singlet oxygen. This observation would be important for a better understanding of the damage mechanism of cell membrane or lipoprotein by singlet oxygen and various radicals generated in the peroxidation and decomposition of lipids induced by heme-proteins.     

Solution NMR study of the yeast cytochrome c peroxidase: cytochrome c interaction

Here we present a solution NMR study of the complex between yeast cytochrome c (Cc) and cytochrome c peroxidase (CcP), a paradigm for understanding the biological electron transfer. Performed for the first time, the CcP-observed heteronuclear NMR experiments were used to probe the Cc binding in solution. Combining the Cc- and CcP-detected experiments, the binding interface on both proteins was mapped out, confirming that the X-ray structure of the complex is maintained in solution. Using NMR titrations and chemical shift perturbation analysis, we show that the interaction is independent of the CcP spin-state and is only weakly affected by the Cc redox state. Based on these findings, we argue that the complex of the ferrous Cc and the cyanide-bound CcP is a good mimic of the catalytically-active Cc–CcP compound I species. Finally, no chemical shift perturbations due to the Cc binding at the low-affinity CcP site were observed at low ionic strength. We discuss possible reasons for the absence of the effects and outline future research directions.     

Effects of surface amino acid replacements in cytochrome c peroxidase on complex formation with cytochrome c.

Site-directed mutagenesis was employed to examine the role played by specific surface residues in the activity of cytochrome c peroxidase. The double charge, aspartic acid to lysine, point mutations were constructed at positions 37, 79, and 217 on the surface of cytochrome c peroxidase, sites purported to be within or proximal to the recognition site for cytochrome c in an electron-transfer productive complex formed by the two proteins. The resulting mutant peroxidases were examined for catalytic activity by steady-state measurements and binding affinity by two methods, fluorescence binding titration and cytochrome c affinity chromatography. The cloned peroxidases exhibit similar UV-visible spectra to the wild-type yeast protein, indicating that there are no major structural differences between the cloned peroxidases and the wild-type enzyme. The aspartic acid to lysine mutations at positions 79 and 217 exhibited similar turnover numbers and binding affinities to that seen for the "wild type-like" cloned peroxidase. The same change at position 37 caused more than a 10-fold decrease in both turnover of and binding affinity for cytochrome c. This empirical finding localizes a primary recognition region critical to the dynamic complex. Models from the literature proposing structures for the complex between peroxidase and cytochrome c are discussed in light of these findings.     

Horseradish peroxidase catalyzed oxidation of thiocyanate by hydrogen peroxide: comparison with lactoperoxidase-catalysed oxidation and role of distal histidine

Horseradish peroxidase-catalysed oxidation of thiocyanate by hydrogen peroxide has been studied by 15N-NMR and optical spectroscopy at different concentrations of thiocyanate and hydrogen peroxide and at different pH values. The extent of the oxidation and the identity of the oxidized product of the thiocyanate has been investigated in the SCN-/H2O2/HRP system and compared with the corresponding data on the SCN-/H2O2/LPO system. The NMR studies show that (SCN)2 is the oxidation product of thiocyanate in the SCN-/H2O2/HRP system, and its formation is maximum at pH less than or equal to 4 and that the oxidation does not take place at pH greater than or equal to 6. Since thiocyanate does not bind to HRP at pH greater than or equal to 6 (Modi et al. (1989) J. Biol. Chem. 264, 19677-19684), the binding of thiocyanate to HRP is considered to be a prerequisite for the oxidation of thiocyanate. It is further observed that at [H2O2]/[SCN-] = 4, (SCN)2 decomposes very slowly back to thiocyanate. The oxidation product of thiocyanate in the SCN-/H2O2/LPO system has been shown to be HOSCN/OSCN- which shows maximum inhibition of uptake by Streptococcus cremoris 972 bacteria when hydrogen peroxide and thiocyanate are present in equimolar amounts (Modi et al. (1991) Biochemistry 30, 118-124). However, in case of HRP no inhibition of oxygen uptake by this bacteria was observed. Since thiocyanate binds to LPO at the distal histidine while to HRP near 1- and 8-CH3 heme groups, the role of distal histidine in the activity of SCN-/H2O2/(LPO, HRP) systems is indicated.     

Oxidation of cytochrome c peroxidase to compound I by peroxyacids: evidence for rate-limiting diffusion through the protein matrix

The rate of the reaction between p-nitroperoxybenzoic acid and cytochrome c peroxidase (CcP) has been investigated as a function of pH and ionic strength. The pH dependence of the reaction between CcP and peracetic acid has also been determined. The rate of the reactions are influenced by two heme-linked ionizations in the protein. The enzyme is active when His-52 (pK(a) 3.8 +/- 0.1) is unprotonated and an unknown group with a pK(a) of 9.8 +/- 0.1 is protonated. The bimolecular rate constant for the reaction between peracetic acid and CcP and between p-nitroperoxybenzoic acid and CcP are (1.8 +/- 0.1) x 10(7) and (1.6 +/- 0.2) x 10(7) M(-)(1) s(-)(1), respectively. These rates are about 60% slower than the reaction between hydrogen peroxide and CcP. A critical comparison of the pH dependence of the reactions of hydrogen peroxide, peracetic acid, and p-nitroperoxybenzoic acid with CcP provides evidence that both the neutral and anionic forms of the two peroxyacids react directly with the enzyme. The peracetate and p-nitroperoxybenzoate anions react with CcP with rates of (1.5 +/- 0.1) x 10(6) and (1.6 +/- 0.2) x 10(6) M(-)(1) s(-)(1), respectively, about 10 times slower than the neutral peroxyacids. These data indicate that CcP discriminates between the neutral peroxyacids and their negatively charged ions. However, the apparent bimolecular rate constant for reaction between p-nitroperoxybenzoate and CcP is independent of ionic strength in the range of 0.01-1.0 M, suggesting that electrostatic repulsion between the anion and CcP is not the cause of the lower reactivity for the peroxybenzoate anion. The data are consistent with the hypothesis that the rate-limiting step for the oxidation of CcP to compound I by both neutral peroxyacid and the negatively charged peroxide ion is diffusion of the reactants through the protein matrix, from the surface of the protein to the distal heme pocket.     

A novel heme and peroxide-dependent tryptophan-tyrosine cross-link in a mutant of cytochrome c peroxidase

The crystal structure of a cytochrome c peroxidase mutant where the distal catalytic His52 is converted to Tyr reveals that the tyrosine side-chain forms a covalent bond with the indole ring nitrogen atom of Trp51. We hypothesize that this novel bond results from peroxide activation by the heme iron followed by oxidation of Trp51 and Tyr52. This hypothesis has been tested by incorporation of a redox-inactive Zn-protoporphyrin into the protein, and the resulting crystal structure shows the absence of a Trp51-Tyr52 cross-link. Instead, the Tyr52 side-chain orients away from the heme active-site pocket, which requires a substantial rearrangement of residues 72-80 and 134-144. Additional experiments where heme-containing crystals of the mutant were treated with peroxide support our hypothesis that this novel Trp-Tyr cross-link is a peroxide-dependent process mediated by the heme iron.     

Replacement of the axial histidine ligand with imidazole in cytochrome c peroxidase. 1. Effects on structure

Replacement of the axial histidine ligand with exogenous imidazole has been accomplished in a number of heme protein mutants, where it often serves to complement the functional properties of the protein. In this paper, we describe the effects of pH and buffer ion on the crystal structure of the H175G mutant of cytochrome c peroxidase, in which the histidine tether between the heme and the protein backbone is replaced by bound imidazole. The structures show that imidazole can occupy the proximal H175G cavity under a number of experimental conditions, but that the details of the interaction with the protein and the coordination to the heme are markedly dependent on conditions. Replacement of the tethered histidine ligand with imidazole permits the heme to shift slightly in its pocket, allowing it to adopt either a planar or distally domed conformation. H175G crystallized from both high phosphate and imidazole concentrations exists as a novel, 5-coordinate phosphate bound state, in which the proximal imidazole is dissociated and the distal phosphate is coordinated to the iron. To accommodate this bound phosphate, the side chains of His-52 and Asn-82 alter their positions and a significant conformational change in the surrounding protein backbone occurs. In the absence of phosphate, imidazole binds to the proximal H175G cavity in a pH-dependent fashion. At pH 7, imidazole is directly coordinated to the heme (d(Fe--Im) = 2.0 A) with a nearby distal water (d(Fe--HOH) = 2.4 A). This is similar to the structure of WT CCP except that the iron lies closer in the heme plane, and the hydrogen bond between imidazole and Asp-235 (d(Im--Asp) = 3.1 A) is longer than for WT CCP (d(His--Asp) = 2.9 A). As the pH is dropped to 5, imidazole dissociates from the heme (d(Fe--Im) = 2.9 A), but remains in the proximal cavity where it is strongly hydrogen bonded to Asp-235 (d(Im--Asp) = 2.8 A). In addition, the heme is significantly domed toward the distal pocket where it may coordinate a water molecule. Finally, the structure of H175G/Im, pH 6, at low temperature (100 K) is very similar to that at room temperature, except that the water above the distal heme face is not present. This study concludes that steric restrictions imposed by the covalently tethered histidine restrain the heme and its ligand coordination from distortions that would arise in the absence of the restricted tether. Coupled with the functional and spectroscopic properties described in the following paper in this issue, these structures help to illustrate how the delicate and critical interactions between protein, ligand, and metal modulate the function of heme enzymes.     

CO dissociation in cytochrome c peroxidase: site-directed mutagenesis shows that distal Arg 48 influences CO dissociation rates

To investigate the molecular basis for the 100-fold slower rate of CO dissociation in ferrous peroxidases relative to myoglobin, CO dissociation rates were measured as a function of pH in the cloned cytochrome c peroxidase from yeast [CCP(MI)] and in several mutants in the heme binding pocket prepared by site-directed mutagenesis. The mutants included Asp 235----Asn; Arg 48----Lys, Leu; and His 181----Gly. Changes in the absorption spectrum with pH are consistent with conversion of the CO-ferrous CCP(MI) complex from acidic to alkaline forms by a two-proton cooperative ionization, with an apparent pKa = 7.6, analogous to that described for CCP from bakers' yeast [Iizuka, T., Makino, R., Ishimura, Y., & Yonetani, T. (1985) J. Biol. Chem. 260, 1407-1412]. The rate of CO dissociation (koff) was increased 11-fold (from 0.7 x 10(-4) to 8.0 x 10(-4) s-1) by conversion of the acidic to the alkaline form. Analogous acidic and alkaline forms of the CO complex were also observed in the mutants of CCP(MI) examined here. In the acidic form, koff was increased 5- and 20-fold when Arg 48 was replaced with Lys and Leu, respectively, while in the acidic form of mutants that possess Arg 48, koff was similar to that observed in CCP(MI). Conversion of the CO complex from the acidic to alkaline form increased koff in all the mutants, and the pH-dependent increase in koff correlated with a two-proton cooperative ionization, except in the case of His 181----Gly. In this mutant, pH-dependent increase in koff correlated with a single-proton ionization, implicating His 181 as one of the two residues that is deprotonated in the conversion of CO-ferrous CCP(MI) from acidic to alkaline forms. Only a 2.5-fold variation was observed for koff between the alkaline form of CCP(MI) and the Arg 48----Leu mutant, suggesting that the influence of Arg 48 on the rate of CO dissociation is decreased in the alkaline form by a conformational change.(ABSTRACT TRUNCATED AT 400 WORDS)     

Oxidation of ABTS by silicate-immobilized cytochrome c in nonaqueous solutions

Cytochrome c can be readily adsorbed onto mesoporous silicates at high loadings of up to 10 mmol g(-)(1) of silicate. The adsorbed protein retains its peroxidative activity, with no diffusional limitations being observed. The protein can be adsorbed onto the external surface of the silicate or, provided that the pore diameter is sufficiently large, into the channels. In aqueous buffer, the catalytic activity of the adsorbed protein (for the oxidation of ABTS) decreased with increasing temperature, with the decrease being less marked for cytochrome c held within the silicate channels. Similar results were obtained in 95% methanol. Analysis of kinetic data showed that significant increases in k(cat)/K(M) occurred in methanol, ethanol, and formamide, with slight decreases occurring in 1-methoxy-2-propanol. The observed increases were primarily a result of substantial increases in k(cat), while the results in 1-methoxy-2-propanol can be ascribed to increases in K(M). Resonance Raman spectroscopy indicated that the structure of the heme environment of the adsorbed protein was essentially unchanged, in aqueous buffer and in the nonaqueous solvents, methanol, 1-methoxy-2-propanol, and ethanol. In addition, Raman spectra of the lyophilized protein indicated that there were no apparent changes in the heme structure.     

The role of cytochrome c in caspase activation in Drosophila melanogaster cells

The release of cytochrome c from mitochondria is necessary for the formation of the Apaf-1 apoptosome and subsequent activation of caspase-9 in mammalian cells. However, the role of cytochrome c in caspase activation in Drosophila cells is not well understood. We demonstrate here that cytochrome c remains associated with mitochondria during apoptosis of Drosophila cells and that the initiator caspase DRONC and effector caspase DRICE are activated after various death stimuli without any significant release of cytochrome c in the cytosol. Ectopic expression of the proapoptotic Bcl-2 protein, DEBCL, also fails to show any cytochrome c release from mitochondria. A significant proportion of cellular DRONC and DRICE appears to localize near mitochondria, suggesting that an apoptosome may form in the vicinity of mitochondria in the absence of cytochrome c release. In vitro, DRONC was recruited to a >700-kD complex, similar to the mammalian apoptosome in cell extracts supplemented with cytochrome c and dATP. These results suggest that caspase activation in insects follows a more primitive mechanism that may be the precursor to the caspase activation pathways in mammals.