Oxidative stress is tightly regulated by cytochrome c phosphorylation and respirasome factors in mitochondria

Oxidative stress is tightly regulated by cytochrome cphosphorylation and respirasome factors in mitochondria Alejandra Guerra-Castellano a , Antonio Díaz-Quintana a , Gonzalo Pérez-Mejías a , Carlos A. Elena-Real a , Katiuska González-Arzola a , Sofía M. García-Mauriño a , Miguel A. De la Rosa a,1 , and Irene Díaz-Moreno a,1 a Instituto de Investigaciones Químicas–Centro de Investigaciones Científicas Isla de la Cartuja, Universidad de Sevilla–Consejo Superior de Investigaciones Científicas, 41092 Sevilla, Spain Edited by Alan R. Fersht, Gonville and Caius College, Cambridge, United Kingdom, and approved June 26, 2018 (received for review April 20, 2018) Respiratory cytochrome chas been found to be phosphorylated at tyrosine 97 in the postischemic brain upon neuroprotective insulin treatment, but how such posttranslational modification affects mitochondrial metabolism is unclear. Here, we report the structural features and functional behavior of a phosphomimetic cytochrome cmutant, which was generated by site-specific incorporation at position 97 of p-carboxymethyl-L-phenylalanine using the evolved tRNA synthetase method. We found that the point mutation does not alter the overall folding and heme environment of cytochrome c, but significantly affects the entire oxidative phosphorylation process. In fact, the electron donation rate of the mutant heme protein to cytochrome coxidase, or complex IV, within respiratory supercomplexes was higher than that of the wild-type species, in agreement with the observed decrease in reactive oxygen species production. Direct contact of cytochrome cwith the respiratory supercomplex factor HIGD1A (hypoxia-inducible domain family member 1A) is reported here, with the mutant heme protein exhibiting a lower affinity than the wild-type species. Interestingly, phosphomimetic cytochrome calso exhibited a lower caspase-3 activation activity. Altogether, these findings yield a better understanding of the molecular basis for mitochondrial metabolism in acute diseases, such as brain ischemia, and thus could allow the use of phosphomimetic cytochrome cas a neuroprotector with therapeutic applications. cytochrome c | electron transport chain | oxidative stress | phosphorylation | reactive oxygen species Cell redox signaling is fine-tuned by protein phosphorylation, a well-known posttranslational modification in turn modulated by the antagonistic effects of kinases and phosphatases (1, 2). Such an interplay is specifically relevant in mitochondrial metabolism, which is the main source of reactive oxygen species/reactive nitrogen species (ROS/RNS) in the cell (3). Tyrosine kinases, in particular, are key signaling mediators and their perturbations underlie many human pathologies, namely cancer and neurodegenerative diseases (4, 5). One of the main proteins that control redox signaling in mitochondrial oxidative phosphorylation (OxPhos) is cytochrome c(Cc), the functions of which are regulated by phosphorylation of several residues (6–10). Ccplays a double role in the cellular metabolism. Under homeostasis, Ccis located in the mitochondrial intermembrane space, where it acts as an electron carrier between complex III (CIII) and complex IV (CIV) in the electron transfer chain (ETC). These two respiratory complexes can be either independent or associated with each other and with complex I (CI), thus forming mitochondrial supercomplexes—the so-called respirasome—the assembly of which is strictly dependent on metabolic needs (11). Respiratory supercomplexes provide higher efficiency to the electron transport during OxPhos, thereby minimizing the generation of ROS (12). Respiratory supercomplex factors (Rcfs) are protein modulators that stabilize the membrane-embedded CIII and CIV within respiratory supercomplexes. These factors have been found in both yeasts and mammals. In yeasts, they are called respiratory supercomplex factor 1 (Rcf1, formerly Aim31) and respiratory supercomplex factor 2 (Rcf2, formerly Aim38), both being members of the conserved hypoxia-induced gene 1 (Hig1) protein family (13, 14). Rcf1 has two human orthologs, namely the HIG hypoxia-inducible domain family members 1A and 2A (HIGD1A and HIGD2A, respectively) (15). Rcf2, in contrast, is yeast-specific and necessary for oligomerization of a subclass of Ccoxidase (CcO) into supercomplexes. Under nitro-oxidative stress conditions, Ccacts as a programmed cell death (PCD) inducer (16). During early apoptosis, a Ccpopulation—tightly bound to the mitochondrial membrane— triggers peroxidation of phospholipids and, in particular, of cardiolipin (CL). CL-adducted Ccundergoes a profound tertiary conformational rearrangement that opens an entry channel for H 2 O 2 molecules, which explains how Ccenhances its peroxidase activity (17, 18). It has been proposed that Cc-CL conjugates are sufficient for formation of mitochondrial pores, allowing the release of the hemeprotein into the cytosol during apoptosis (19). Extramitochondrial Ccbinds to the apoptotic proteaseactivating factor 1 (Apaf1), thereby eliciting the apoptosome Significance Dysfunction of mitochondria, the powerhouses of living cells, favors the onset of human diseases, namely neurodegenerative diseases, cardiovascular pathologies, and cancer. Actually, respiratory cytochrome chas been found to be phosphorylated at tyrosine 97 during the insulin-induced neuroprotection response following a brain ischemic injury. Here, we report that the decrease in neuronal death could be directly ascribed to changes in mitochondrial metabolism—including lower production of reactive oxygen species—and cell homeostasis induced by cytochrome cphosphorylation. Our findings thus provide the basis for understanding the molecular mechanism and potential use of phosphomimetic species of cytochrome c, thereby yielding new opportunities to develop more efficient therapies against acute pathologies. Author contributions: A.G.-C., A.D.-Q., M.A.R., and I.D.-M. designed research; A.G.-C., G.P.-M., C.A.E.-R., K.G.-A., S.M.G.-M., and I.D.-M. performed research; A.G.-C., A.D.-Q., G.P.-M., C.A.E.-R., and I.D.-M. analyzed data; and A.G.-C., A.D.-Q., M.A.R., and I.D.-M. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. This open access article is distributed under Creative Commons Attribution-NonCommercial- NoDerivatives License 4.0 (CC BY-NC-ND). 1 To whom correspondence may be addressed. Email: [email protected] or idiazmoreno@ us.es. This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1806833115/-/DCSupplemental. Published online July 17, 2018. www.pnas.org/cgi/doi/10.1073/pnas.1806833115 PNAS | July 31, 2018 | vol. 115 | no. 31 | 7955–7960 BIOPHYSICS AND COMPUTATIONAL BIOLOGY platform assembly and activating the caspase cascade (20, 21). Recent data indicate that the programmed cell death network involving Ccis complex, and that Cctargets several proteins that are functionally equivalent in humans and plants (22–24). In both type of cells, for example, Ccreaches the cell nucleus upon DNA damage and sequesters human SET/TAF1βand plant NRP1 histone chaperones that remodel chromatin (25–27). In addition to regulating the functions described above, phosphorylated Ccspecies have been associated with several human pathologies. In fact, it has been proposed that Tyr97 is target of phosphorylation during the insulin-induced neuroprotection response following a brain ischemic injury (28). A possible mechanism could be that Tyr97 phosphorylation of Cc would alter the electron shuttling to CcO (7). Notably, electron transfer from wild-type (WT) Ccto CcO is positively regulated by the HIGD1A membrane protein (29). Research on tyrosine phosphorylation of Cchas normally resorted to using Tyr-to-Glu mutations (30, 31). This overcomes the difficulties of obtaining enough phosphorylated Ccfor biophysical and functional studies when the specific Cc-phos- phorylating kinase is unknown (32). However, the Tyr-to-Glu substitution leads to a substantial decrease in the residue volume and the loss of an aromatic ring. Therefore, we used the evolved tRNA synthetase technique to introduce the noncanonical amino acid p-carboxymethyl-L-phenylalanine (pCMF) at position 97, as previously reported for Tyr48 (33). This noncanonical amino acid is actually used to synthesize stable analogs of a wide array of phosphoproteins, the kinases of which are unknown or the yields of which from tissues are low. In addition, pCMF provides a dependable analog of phosphorylated tyrosine that can be used to study signal transduction pathways. In this work, the resulting Y97pCMF Ccmutant maintained its overall folding and heme environment. We found that such a posttranslational modification of Ccaffected the OxPhos process since it enhanced the electron donation rate to CcO in mitochondria isolated from yeasts grown in respiratory metabolism. In addition, we observed a decrease in the production of ROS. Thus, the CcO-driven oxidation rate of Ccis controlled by Table 1. Midpoint melting temperature (T m ) (°C) of WT and Y97pCMF Cc Ccspecies Far-UV CD, 220 nm Visible CD, 419 nm Fluorescence, 270 nm WT 88.6 ±1.9 43.9 ±2.0 (T m1 )86.3±1.4 85.1 ±2.3 (T m2 ) Y97pCMF 82.2 ±0.3 76.6 ±4.9 82.6 ±1.6 Fig. 1. Modulation of CcO activity by the HIGD1A protein and interaction of the latter with Cc.(A) Bovine CcO activity with added WT (blue) or Y97pCMF (red) Ccas electron donor in the presence of the chimera MBP-HIGD1A at varying concentrations. The 1:0 bars represent the CcO activity without MBPHIGD1A. Experimental data were normalized to the Cc:HIGD1A ratio of 1:0 for WT Ccspecies. a.u.: arbitrary units. (B) Immunoprecipitation of transfected HIGD1A-mCherry with Ccin Heltog cells. (Left) Western blot shows the detection of the HIGD1A-mCherry band in the lysates, along with Cc-IP of lysates followed by probing with the HIGD1A antibody (arrow). Mouse IgG (IgG) was used as control. (Right) Detection of Ccband in lysates and Ccimmunoprecipitated lysates are also shown (arrow). Asterisks and hashtags are heavy and light chains of IgG, respectively. The black circle stands for unspecific bands of anti-Ccantibody, as reported by the manufacturer. (C) Western blots against cytosolic (tubulin), mitochondrial (Hsp60), and nuclear (H3) markers. (D) Subcellular localization of HIGD1A and Ccproteins by confocal microscopy. Heltog cells were transfected to express the chimera HIGD1A-mCherry (red; first column of Upper and Middle rows). The chimera Cc-GFP is stably expressed in the Heltog cell line (green; second columns). Nuclei were stained in blue with Hoechst (third columns) and mitochondria with MitoTracker Red CMXRos (red, first panel of Lower row). Colocalization between green Cc-GFP and red HIGD1A-mCherry in mitochondria, along with blue nuclear staining, is shown in the merge images (last columns of Upper and Middle rows). (Scale bars: 10 μm.) 7956 | www.pnas.org/cgi/doi/10.1073/pnas.1806833115 Guerra-Castellano et al. phosphorylation to maintain low levels of apoptotic-inducing ROS. Furthermore, Y97pCMF Ccwas an inefficient caspase activator. Altogether, these findings can shed light on the function of phosphorylated Ccin insulin-stimulated postischemic neuroprotection. Results and Discussion Substitution of Tyr97 with pCMF Allows for Study of CcPhosphorylation, Preserving Its Physicochemical Properties. To accurately study the phosphorylation of Ccat Tyr97, this residue was substituted with the noncanonical amino acid pCMF, which is an analog of phosphotyrosine residue. To do that, we used the evolved tRNA synthetase method as previously reported (SI Appendix, Fig. S1). As described above, this substitution simulates the charge and the volume of a phosphotyrosine residue better than a glutamate substitution (33). To confirm that the noncanonical amino acid pCMF was successfully introduced at position 97 of the heme protein, the Y97pCMF Ccmutant was purified to homogeneity, and its molecular mass was compared with that of WT Ccby MALDI-TOF (SI Appendix, Fig. S2 Aand B). The phosphomimetic mutant presented an increase of 41 Da in its molecular mass in comparison with that of WT Cc, which is consistent with the expected substitution (SI Appendix, Fig. S2A). To check whether the amino acid replacement affected the secondary structure and heme environment of Cc, circular dichroism (CD) spectroscopy measurements in the far-UV and visible region were carried out, respectively. Those data demonstrated that Y97pCMF Ccexhibited an overall secondary structure and heme properties similar to those observed for WT Cc(SI Appendix,Fig.S2Cand D). Following with the biophysical analysis of the Y97pCMF mutant, the alkaline transition phenomenon was also analyzed. This phenomenon has been related to a change in the Cccell location, which is essential for its role in PCD (34). Such transition affects the oxidized form of Ccand involves the replacement of Met80—the sixth ligand of the heme group—by Lys72, Lys73, or Lys79 as an axial ligand at high pH values (35). To analyze the Fe-Met80(S δ ) bond changes induced by pH, we therefore monitored the charge-transfer band at 699 nm, which disappears concomitantly with the alkaline transition (SI Appendix, Fig. S3A) (36). The quantitative analysis yields a difference of ∼0.8 in the pK a values for this alkaline transition in the WT and Y97pCMF Ccspecies (pK a-WT =8.9 ±0.4 and pK a-Y97pCMF =8.1 ±0.2; SI Appendix,Fig.S3B–D). On the other hand, given that most mutations in Tyr97 performed to date alter the thermal stability of Cc, we explored that possibility in Y97pCMF Cc(31, 37). In terms of secondary and overall structure, data from far-UV CD and fluorescence, respectively, showed that the midpoint melting temperature (T m ) of Y97pCMF was slightly lower than that calculated for WT Cc (Table 1). Notably, in contrast to the data recorded for the heme group environment of WT species, the visible CD data of Y97pCMF Ccfit well into a two-state model (Table 1). Such findings could indicate a slight destabilization in the N- terminal α-helix pack, which constitutes an early event in Cc folding (38–41). Altogether, the biophysical comparison of WT and Y97pCMF Ccspecies indicates that the substitution of Tyr97 with noncanonical pCMF is a good approach to studying the consequences of Ccphosphorylation without significantly altering the physicochemical and structural properties. CcO Activity Is Modulated by Y97pCMF Ccand Respiratory Supercomplex Factors. Given that phosphorylation of Cchasbeendescribedasa phenomenon that modulates the functions of this metalloprotein, we explored the behavior of Y97pCMF Ccas an electron donor. The in vitro ability of Y97pCMF Ccto reduce CcO was measured in the absence or presence of HIGD1A, which acts as a positive modulator of CcO (29). In the absence of a modulator, the CcO activities were slightly higher with the Y97pCMF mutant compared with the WT Cc(Fig. 1A). Then we wondered whether those slight changes were due to a change in the molecular recognition mode between both Ccspecies and CcO. Thus, we analyzed the interaction between CcO and WT or phosphomimetic Y97pCMF Cc by isothermal titration calorimetry (ITC) (Fig. 2Aand B). Both isotherms were fitted to a model with two independent binding sites, as previously described for the cross-complex between human WT Ccand bovine CcO(42).Notably,theK D values showed that Y97pCMF Cchad a higher affinity for CcOthandidtheWT species. These results would explain the higher CcO activity shown by the phosphomimetic mutant. Recently, it has been described that the HIGD1A-dependent increase in CcO activity was higher with a Ccphosphomimetic mutant at position 48 (9). For this reason, we tested the effect of the HIGD1A in CcO activity through WT and Y97pCMF Cc.Toavoid the solubility problem of the membrane protein HIGD1A, we used a chimera with a maltose binding protein (MBP) tag in its N terminus (MBP-HIGD1A) (SI Appendix,Fig.S4). Notably, CcOactivity significantly increased upon addition of MBP-HIGD1A through Y97pCMF Cc(Fig. 1 A). To discard a possible contribution from Fig. 2. ITC assays of CcO and HIGD1A with WT and Y97pCMF Cc.(Aand B) Thermograms (Upper) and binding isotherms (Lower)ofWTCc/CcO(A) and Y97pCMF Cc/CcO(B) complexes in their reduced states. Experimental data were fitted to a 2:1 model with the same K D values. (Cand D) Thermograms (Upper) and binding isotherms (Lower)ofWTCc/MBP-HIGD1A (C) and Y97pCMF Cc/MBP-HIGD1A (D) complexes at reduced states. Data were fitted to a model with a 1:1 stoichiometry. Guerra-Castellano et al. PNAS | July 31, 2018 | vol. 115 | no. 31 | 7957 BIOPHYSICS AND COMPUTATIONAL BIOLOGY the MBP tag in CcO activity, assays were also performed with MBP protein in the presence of both Ccspecies. As shown in SI Appendix,Fig.S5A, MBP did not alter CcOactivity. Taking into account the differences observed in Fig. 1A,we would expect an eventual interaction between Ccand HIGD1A. To explore this further, we examined the interaction between MBP-HIGD1A and WT or Y97pCMF Ccby ITC (Fig. 2Cand D). These measurements revealed that both Ccspecies interacted with MBP-HIGD1A with a stoichiometry of 1:1, yielding dissociation constant (K D ) values within the micromolar range. To demonstrate that the interaction with the chimera was uniquely due to the HIGD1A module, we performed an ITC assay between WT Ccand the MBP protein (SI Appendix,Fig.S5B). These data revealed the absence of any interaction between both proteins. In addition, dilution ITC experiments showed that the observed changes in heat rate are rather due to the interaction between the Ccspecies and MBP-HIGD1A (SI Appendix,Fig.S6Aand B). The redox state of Ccdid not affect its interaction with HIGD1A (SI Appendix,Fig.S6C). Since the N-terminal domain of HIGD1A— hereafter, HIGD1A 1–26 —has been reported to retain Ccin the intermembrane mitochondrial space (IMS) and to decrease caspase activation (43), a set of ITC titrations were run. Data resulting from such titrations evinced that both Ccspecies interacted with HIGD1A 1–26 with similar stoichiometry values, although Y97pCMF Ccshowed less affinity than WT Cc(SI Appendix,Fig.S7). Thus, the lesser affinity of Y97pCMF Cctoward HIGD1A 1–26 or MBPHIGD1A may point out a higher interaction turnover of the phosphomimetic mutant in comparison with the WT species, thus facilitating CcO reduction (Fig. 1A). To test the interaction between HIGD1A and Ccin cells, coimmunoprecipitation (Co-IP) assays were performed in Heltog cells—HeLa cells constitutively expressing green fluorescent protein (GFP)-tagged Cc—transfected with HIGD1A-mCherry vector. Immunoblotting data against HIGD1A corroborated the interaction between Ccand HIGD1A (Fig. 1 B). The total lysate used in the Co-IP assays contained cytosolic and mitochondrial fractions, as shown by the detection of tubulin and Hsp60, respectively (Fig. 1C). Furthermore, we checked that both Cc and HIGD1A were in the same cellular compartment by confocal analysis of Heltog cells upon transfection with HIGD1A- mCherry vector, which codes for red fluorescence-tagged proteins. The pictures showed that Cc-GFP (green fluorescence) colocalized with HIGD1A-mCherry (red fluorescence) under homeostatic conditions (Fig. 1D,Upper and Middle). The mitochondrial localization of Cc-GFP was confirmed by codetection with the MitoTracker (Fig. 1D,Lower). Altogether, these findings revealed contacts between HIGD1A and Ccspecies. Based on these results, the HIGD1A- and Ccphosphorylationmediated regulations of CcO activity were tested in a cellular context using isolated mitochondria from yeast cells grown on a respiratory media. Under such growth conditions, Rcf1—a yeast ortholog of human HIGD1A—and Rcf2 stabilized the CIII–CIV interaction, favoring the formation of the supercomplexes (13, 14, 44). To analyze the effect of supercomplex formation, we used mitochondria obtained from WT and Rcf1/Rcf2-deficient yeast strains (Mit wt and Mit ΔΔRcf1/2 , respectively). The supercomplex formation was impaired in Mit ΔΔRcf1/2 due to the absence of Rcf1 and Rcf2 proteins (SI Appendix,Figs.S5Cand S8). In fact, CcO activity assays using Mit ΔΔRcf1/2 revealed an electron flow significantly lower in comparison with those observed using Mit WT , regardless whether WT or Y97pCMF Ccwas added (Fig. 3A). In addition, a single deletion of either Rcf1 or Rcf2 partially affected the CcO activity levels, indicating a cooperative effect when the two membrane proteins Rcf1 and Rcf2 were present (SI Appendix,Fig.S5C). Notably, the Y97pCMF Ccmutant exhibited higher CcO activity than WT Cc, both in WT (Mit WT ) and in mutant mitochondria (Mit ΔRcf1 , Mit ΔRcf2 ,andMit ΔΔRCf1/2 )(Fig.3Aand SI Appendix,Fig.S5C). This finding indicates that the Rcf-mediated ETC flow is more efficient when Y97pCMF Ccparticipates (Fig. 3A). This suggests a double regulation mechanism through Ccphosphorylation and supercomplex formation to make an accurate cellular response against hypoxia or other respiratory stresses and needs. Y97pCMF CcImpairs Mitochondrial Supercomplex-Mediated ROS Production. Changes in the ETC flow can affect the hyperpolarization of mitochondrial inner membrane and ROS production (45). For this reason, we investigated if the accelerated CcO-driven oxidation rate of phosphorylated Cccould alter mitochondrial ROS production both in the absence and in the presence of respiratory supercomplexes. ROS production in mitochondria from Saccharomyces cerevisiae occurs mainly in CIII as its mitochondria are deficient in CI (46, 47). Significant contribution of complex II (CII) to ROS generation has recently been described in mammalian mitochondria (48–50). In S. cerevisiae, however, contribution of CII to ROS generation occurs only when specific mutations are present near the proximal quinone-binding Fig. 3. Effect of the respiratory factors on CcO activity and ROS production in isolated mitochondria. (A)CcO activity was measured in isolated mitochondria with added WT Cc(blue bars) or Y97pCMF Cc(red bars). Data were normalized with respect to the activity with WT Ccin mitochondria isolated from ΔΔRcf1/2 yeast strains. Data represent the mean ±SD of three independent experiments. a.u.: arbitrary units. (B)H 2 O 2 production in isolated Mit ΔΔRcf1/2 and Mit WT . ROS generation was triggered by addition of succinate. Bar color code is the same as in A.(C) Molecular model of OxPhos and ROS production by respirasome when Ccis phosphorylated. The CcO-driven oxidation rate of Ccis facilitated by Tyr-97 phosphorylation and supercomplex formation to maintain low levels of apoptotic-inducing ROS. Blue, CI; green, dimeric CIII; and magenta, CIV. Ccmolecules at the proximal sites of one monomer of CIII or CIV are in red, whereas those at the distal sites are in yellow. HIGD1A, colored in dark blue, makes contacts with CIV and the molecule of Cclocated at its proximal site. The black arrow indicates the channeling of Cc molecules between CIII and CIV. IMM, inner mitochondrial membrane; IMS, intermembrane mitochondrial space; MM, mitochondrial matrix. 7958 | www.pnas.org/cgi/doi/10.1073/pnas.1806833115 Guerra-Castellano et al. site (Q P ) of CII (51). Since our study was focused on Cc-mediated ROS production, with Ccacting as an electron donor between CIII and CIV, no ETC inhibitors were used. Therefore, succinate was added as electron donor to both types of mitochondria, in the presence of either WT or Y97pCMF Cc, to trigger the ETC through CII. Upon addition of succinate, ETC-dependent H 2 O 2 production significantly increased, especially in Mit ΔΔRcf1/2 where the formation of supercomplexes is impaired (Fig. 3Band SI Appendix,Fig.S9). Remarkably, Y97pCMF Ccmade ROS production decrease in Mit ΔΔRcf1/2 and, more pronouncedly, in Mit WT where supercomplexes are formed (Fig. 3B). These findings suggest that the CcO-driven oxidation rate of Ccis tightly regulated by Ccphosphorylation and supercomplex formation to keep apoptotic-inducing mitochondrial ROS at a low level. Thus, we propose a respirasome OxPhos model based on the recently described extra distal Cc-binding sites in CIII and CIV, thus providing a path for diffusion of Ccmolecules (Fig. 3C) (42, 52). This way of communicating between complexes through Tyr97- phosphorylated Cccould accelerate the electron transfer flow, which, in turn, would avoid ROS-mediated cellular damage to guarantee neuronal survival during ischemia reperfusion in the brain (28, 45). ROS Scavenger Activity and Cardiolipin Binding of CcRemains Unaltered upon Phosphorylation. We wondered whether the low levels of ROS displayed by the presence of Y97pCMF Ccmight be caused by an exacerbated peroxidase activity of the heme protein. We thus measured such enzymatic activity in WT and mutant Cc, either free or bound to lipid vesicles containing CL at a 4:1 ratio [1,2-dioleoyl-sn-glycero-3-phosphocholine:1,1′2,2′- tetraoleoylcardiolipin (DOPC:TOCL)]. No differences in peroxidase activity were observed between Y97pCMF and WT Ccspecies, both when free or bound to liposomes. It should be noted that the use of liposomes rose the peroxidase activity considerably, although in the same way for WT and phosphomimetic Cc(Fig. 4A). This finding corroborates that the drop in ROS production upon mitochondrial supercomplex assembly is due to a more efficient CcO-driven oxidation rate by Y97pCMFthanbyWTCc. In addition, the onset of PCD is triggered by the ability of the hemeprotein to leave the mitochondria after binding to CL (53). Therefore, a plausible hypothesis is that the decrease in neuronal ischemic death may be due to a less efficient release of phosphorylated Ccto the cytosol, concomitantly with its role as a PCD inducer. Hence, we examined the binding of Ccspecies to liposomes containing CL (DOPC:TOCL) (4:1) or not (DOPC) by electrophoretic mobility shift assays (EMSA) in native agarose gels (SI Appendix, Fig. S10). The electrophoretic mobility of Y97pCMF and WT Ccspecies shifted in the same manner upon DOPC:TOCL liposome additions, unveiling that both Ccspecies were able to bind to CL (SI Appendix, Fig. S10). Likewise, EMSA data showed that none of the Ccspecies had altered the electrophoretic mobility with DOPC liposomes, corroborating that the two Ccspecies specifically interact with CL and not with other lipids of the inner mitochondrial membrane. To sum up, these data showed that there is no evidence that phosphorylated Ccat Tyr97 is preferably retained in the mitochondria. Y97pCMF CcIs an Inefficient Caspase Activator. In the context of ischemic brain damage, inhibition of both Cc–Apaf1 interaction and caspase-3 cleavage constitutes a potent neuroprotective strategy (54, 55). The binding assays involving Cc,HIGD1A, and CL suggested that both WT and Y97pCMF Ccspecies did seem to be equally retained in mitochondria. Therefore, the neuroprotective role of phosphorylated Cccannot be fully explained by a specific sequestration of Y97pCMF Ccin the IMS. Thus, we explored whether phosphorylation at Tyr97 modulated Ccability to trigger the caspase cascade in the cytosol, where the hemeprotein interacts with Apaf1 to assemble the so-called apoptosome. We analyzed the caspase-3 activation driven by WT and Y97pCMF Cc, using cytoplasmic extracts of human embryonic kidney 293 cells that lack Cc(Fig. 4B). The Y97pCMF Cc-mediated caspase-3 activity decreased by about 26% with respect to that mediated by WT Cc(Fig. 4B). These findings agree with our previous results on Y97E and Y97F Ccmutants, as well as with the nitrated mono-Tyr97 Cc, which have a caspase activation slightly lower than that of the WT species (31, 56). Interestingly, none of these Ccspecies disrupted the Cc/Apaf1 complex assembly, and this is also expected for the Y97pCMF form. However, the formation of a nonfunctional apoptosome upon posttranslational modifications of Cccannot be excluded (31, 56, 57). Altogether, our data suggest that both low oxidative stress in mitochondria and less activation of the caspase cascade in cytoplasm explain neuronal survival under ischemia damage in cells containing Tyr97-phosphorylated Cc. Materials and Methods Expression of genes in Escherichia coli and purification of Ccand HIGD1A protein constructs followed standard procedures. All experimental conditions are further available in SI Appendix, including strains, culture Fig. 4. Peroxidase and caspase activities of Y97pCMF Cc.(A)Peroxidaseactivity of WT Cc(blue) and Y97pCMF Cc(red) either free (empty bars) or in the presence of liposomes containing DOPC:TOCL (4:1) (filled bars). Data represent the mean ±SD of three independent experiments and are normalized to the peroxidase activity of free WT Cc.(B) Caspase-3 activity upon addition of exogenous WT (blue) or Y97pCMF (red) Cc. Caspase auto-activation, in the absence of added Cc, was also tested (control bar). Data represent the mean ±SD of three independent experiments and are normalized with respect to the activity determined with WT Cc.(Inset) Western blots confirmed the lack of endogenous Ccin the cytoplasmic cell extracts by immunoblotting with antitubulin (cytosolic marker) and anti-Ccantibodies. Lane 1, cytoplasmic cell extracts;lane2,purifiedCc; and lane 3, BSA as a negative control. Guerra-Castellano et al. PNAS | July 31, 2018 | vol. 115 | no. 31 | 7959 BIOPHYSICS AND COMPUTATIONAL BIOLOGY conditions, cell transfection experiments, cell-extract fraction analyses, thermal analyses, and enzymatic activity assays. All biological samples were obtained from bacteria, yeast, or human cell extracts in full compliance with University of Seville Ethical Committee bylaws. ACKNOWLEDGMENTS. We thank Prof. P. G. Schultz (Scripps Research Institute) for providing the plasmid pEVOL/pCMF/tRNA; Prof. Peter Rehling (University Medical Center Göttingen) for the antibodies αRcf1 and αRcf2; and Dr. Rosemary A. Stuart (Klingler College of Arts and Sciences, Marquette University) for the S. cerevisiae strains WT (W303-1A mata leu2 trp1 ura3 his3 ade2), RCF1::HIS3 (Δrcf1) (W303-1A mata leu2 trp1 ura3 ade2), RCF2::KAN (Δrcf2)(W303-1B Matαleu2 trp1 ura3 his3 ade2), and RCF1::HIS3 RCF2::KAN (Δrcf1 Δrcf2)(W303-1A mata leu2 trp1 ura3 ade2). This work was supported by Ministry of Economy and Competitiveness and European Regional Development Fund (MINECO/FEDER) Grant BFU2015-71017/BMC, European Union; the Ramón Areces Foundation, European Social Fund, Andalusian Government (BIO-198); Consejo Superior de Investigaciones Científicas Fellowship JaePre_2011_01248 (to A.G.-C.); a Fundación Cámara PhD Fellowship (to C.A.E.-R.); Biointeractomics Platform (cicCartuja, Seville); and TA Instruments. The INSTRUCT platform of Baculovirus expression (expression screening@OPPF) was used for cloning the HIGD1A into the pOPINM vector. 1. Monteiro HP, Stern A (1996) Redox modulation of tyrosine phosphorylationdependent signal transduction pathways. Free Radic Biol Med 21:323–333. 2. Corcoran A, Cotter TG (2013) Redox regulation of protein kinases. FEBS J 280: 1944–1965. 3. Turrens JF (2003) Mitochondrial formation of reactive oxygen species. J Physiol 552: 335–344. 4. Sefton BM, Hunter T, Beemon K, Eckhart W (1980) Evidence that the phosphorylation of tyrosine is essential for cellular transformation by Rous sarcoma virus. Cell 20:807–816. 5. Ingley E (2012) Functions of the Lyn tyrosine kinase in health and disease. Cell Commun Signal 10:21. 6. Hüttemann M, Lee I, Grossman LI, Doan JW, Sanderson TH (2012) Phosphorylation of mammalian cytochrome cand cytochrome coxidase in the regulation of cell destiny: Respiration, apoptosis, and human disease. Adv Exp Med Biol 748:237–264. 7. Lee I, et al. (2006) New prospects for an old enzyme: Mammalian cytochrome cis tyrosine-phosphorylated in vivo.Biochemistry 45:9121–9128. 8. Yu H, Lee I, Salomon AR, Yu K, Hüttemann M (2008) Mammalian liver cytochrome c is tyrosine-48 phosphorylated in vivo, inhibiting mitochondrial respiration. Biochim Biophys Acta 1777:1066–1071. 9. Moreno-Beltrán B, et al. (2017) Structural basis of mitochondrial dysfunction in response to cytochrome cphosphorylation at tyrosine 48. Proc Natl Acad Sci USA 114: E3041–E3050. 10. Guerra-Castellano A, Díaz-Moreno I, Velázquez-Campoy A, De la Rosa MA, Díaz- Quintana A (2016) Structural and functional characterization of phosphomimetic mutants of cytochrome c at threonine 28 and serine 47. Biochim Biophys Acta 1857: 387–395. 11. Ramírez-Aguilar SJ, et al. (2011) The composition of plant mitochondrial supercomplexes changes with oxygen availability. J Biol Chem 286:43045–43053. 12. Genova ML, Lenaz G (2014) Functional role of mitochondrial respiratory supercomplexes. Biochim Biophys Acta 1837:427–443. 13. Strogolova V, Furness A, Robb-McGrath M, Garlich J, Stuart RA (2012) Rcf1 and Rcf2, members of the hypoxia-induced gene 1 protein family, are critical components of the mitochondrial cytochrome bc 1 -cytochrome coxidase supercomplex. Mol Cell Biol 32:1363–1373. 14. Vukotic M, et al. (2012) Rcf1 mediates cytochrome oxidase assembly and respirasome formation, revealing heterogeneity of the enzyme complex. Cell Metab 15:336–347. 15. Shoubridge EA (2012) Supersizing the mitochondrial respiratory chain. Cell Metab 15: 271–272. 16. Ow YP, Green DR, Hao Z, Mak TW (2008) Cytochrome c: Functions beyond respiration. Nat Rev Mol Cell Biol 9:532–542. 17. Mohammadyani D, et al. (2018) Structural characterization of cardiolipin-driven activation of cytochrome c into a peroxidase and membrane perturbation. Biochim Biophys Acta 1860:1057–1068. 18. Vladimirov YA, et al. (2006) Cardiolipin activates cytochrome cperoxidase activity since it facilitates H( 2 )O( 2 ) access to heme. Biochemistry (Mosc) 71:998–1005. 19. Bergstrom CL, Beales PA, Lv Y, Vanderlick TK, Groves JT (2013) Cytochrome ccauses pore formation in cardiolipin-containing membranes. Proc Natl Acad Sci USA 110: 6269–6274. 20. Zou H, Li Y, Liu X, Wang X (1999) An APAF-1.cytochrome c multimeric complex is a functional apoptosome that activates procaspase-9. J Biol Chem 274:11549–11556. 21. Desagher S, Martinou JC (2000) Mitochondria as the central control point of apoptosis. Trends Cell Biol 10:369–377. 22. Martínez-Fábregas J, et al. (2013) New Arabidopsis thaliana cytochrome cpartners: A look into the elusive role of cytochrome cin programmed cell death in plants. Mol Cell Proteomics 12:3666–3676. 23. Martínez-Fábregas J, et al. (2014) Structural and functional analysis of novel human cytochrome Ctargets in apoptosis. Mol Cell Proteomics 13:1439–1456. 24. Martínez-Fábregas J, Díaz-Moreno I, González-Arzola K, Díaz-Quintana A, De la Rosa MA (2014) A common signalosome for programmed cell death in humans and plants. Cell Death Dis 5:e1314. 25. González-Arzola K, et al. (2015) Structural basis for inhibition of the histone chaperone activity of SET/TAF-Iβby cytochrome c.Proc Natl Acad Sci USA 112:9908–9913. 26. González-Arzola K, et al. (2017) Histone chaperone activity of Arabidopsis thaliana NRP1 is blocked by cytochrome c.Nucleic Acids Res 45:2150–2165. 27. Díaz-Moreno I, Velázquez-Cruz A, Curran-French S, Díaz-Quintana A, De la Rosa MA (2018) Nuclear cytochrome c: A mitochondrial visitor regulating damaged chromatin dynamics. FEBS Lett 592:172–178. 28. Sanderson TH, et al. (2013) Cytochrome cis tyrosine 97 phosphorylated by neuroprotective insulin treatment. PLoS One 8:e78627. 29. Hayashi T, et al. (2015) Higd1a is a positive regulator of cytochrome coxidase. Proc Natl Acad Sci USA 112:1553–1558. 30. Pecina P, et al. (2010) Phosphomimetic substitution of cytochrome Ctyrosine 48 decreases respiration and binding to cardiolipin and abolishes ability to trigger downstream caspase activation. Biochemistry 49:6705–6714. 31. García-Heredia JM, et al. (2011) Tyrosine phosphorylation turns alkaline transition into a biologically relevant process and makes human cytochrome cbehave as an antiapoptotic switch. J Biol Inorg Chem 16:1155–1168. 32. Kadenbach B, Urban PF (1968) Application of a quantitative chromatographic method of purification in the study of the biosynthesis of cytochrome c.Fresenius Z Anal Chem 243:542–554. 33. Guerra-Castellano A, et al. (2015) Mimicking tyrosine phosphorylation in human cytochrome cby the evolved tRNA synthetase technique. Chemistry 21:15004–15012. 34. Godoy LC, et al. (2009) Disruption of the M80-Fe ligation stimulates the translocation of cytochrome c to the cytoplasm and nucleus in nonapoptotic cells. Proc Natl Acad Sci USA 106:2653–2658. 35. Ying T, et al. (2009) Evolutionary alkaline transition in human cytochrome c. J Bioenerg Biomembr 41:251–257. 36. Schejter A, George P (1964) The 695-mμband of ferricytochrome cand its relationship to protein conformation. Biochemistry 3:1045–1049. 37. Díaz-Moreno I, García-Heredia JM, Díaz-Quintana A, Teixeira M, De la Rosa MA (2011) Nitration of tyrosines 46 and 48 induces the specific degradation of cytochrome c upon change of the heme iron state to high-spin. Biochim Biophys Acta 1807:1616–1623. 38. Zaidi S, Hassan MI, Islam A, Ahmad F (2014) The role of key residues in structure, function, and stability of cytochrome-c.Cell Mol Life Sci 71:229–255. 39. Chothia C, Levitt M, Richardson D (1981) Helix to helix packing in proteins. J Mol Biol 145:215–250. 40. Pielak GJ, et al. (1995) Protein thermal denaturation, side-chain models, and evolution: Amino acid substitutions at a conserved helix-helix interface. Biochemistry 34: 3268–3276. 41. Ptitsyn OB (1998) Protein folding and protein evolution: Common folding nucleus in different subfamilies of c-type cytochromes? J Mol Biol 278:655–666. 42. Moreno-Beltrán B, et al. (2015) Respiratory complexes III and IV can each bind two molecules of cytochrome cat low ionic strength. FEBS Lett 589:476–483. 43. An H-J, et al. (2011) The survival effect of mitochondrial Higd-1a is associated with suppression of cytochrome C release and prevention of caspase activation. Biochim Biophys Acta 1813:2088–2098. 44. Chen YC, et al. (2012) Identification of a protein mediating respiratory supercomplex stability. Cell Metab 15:348–360. 45. Starkov AA, Fiskum G (2003) Regulation of brain mitochondrial H 2 O 2 production by membrane potential and NAD(P)H redox state. J Neurochem 86:1101–1107. 46. de Vries S, Grivell LA (1988) Purification and characterization of a rotenone-insensitive NADH:Q6 oxidoreductase from mitochondria of Saccharomyces cerevisiae.Eur J Biochem 176:377–384. 47. Herrero E, Ros J, Bellí G, Cabiscol E (2008) Redox control and oxidative stress in yeast cells. Biochim Biophys Acta 1780:1217–1235. 48. Quinlan CL, et al. (2012) Mitochondrial complex II can generate reactive oxygen species at high rates in both the forward and reverse reactions. J Biol Chem 287: 27255–27264. 49. Dröse S (2013) Differential effects of complex II on mitochondrial ROS production and their relation to cardioprotective pre- and postconditioning. Biochim Biophys Acta 1827:578–587. 50. Grivennikova VG, Kozlovsky VS, Vinogradov AD (2017) Respiratory complex II: ROS production and the kinetics of ubiquinone reduction. Biochim Biophys Acta 1858: 109–117. 51. Guo J, Lemire BD (2003) The ubiquinone-binding site of the Saccharomyces cerevisiae succinate-ubiquinone oxidoreductase is a source of superoxide. J Biol Chem 278: 47629–47635. 52. Moreno-Beltrán B, et al. (2014) Cytochrome c1 exhibits two binding sites for cytochrome c in plants. Biochim Biophys Acta 1837:1717–1729. 53. Iverson SL, Orrenius S (2004) The cardiolipin-cytochrome cinteraction and the mitochondrial regulation of apoptosis. Arch Biochem Biophys 423:37–46. 54. Namura S, et al. (1998) Activation and cleavage of caspase-3 in apoptosis induced by experimental cerebral ischemia. J Neurosci 18:3659–3668. 55. Cao G, et al. (2004) Cloning of a novel Apaf-1-interacting protein: A potent suppressor of apoptosis and ischemic neuronal cell death. J Neurosci 24:6189–6201. 56. García-Heredia JM, et al. (2010) Nitration of tyrosine 74 prevents human cytochrome c to play a key role in apoptosis signaling by blocking caspase-9 activation. Biochim Biophys Acta 1797:981–993. 57. García-Heredia JM, et al. (2012) Specific nitration of tyrosines 46 and 48 makes cytochrome cassemble a non-functional apoptosome. FEBS Lett 586:154–158. 7960 | www.pnas.org/cgi/doi/10.1073/pnas.1806833115 Guerra-Castellano et al.