Cytochrome c lysine acetylation regulates cellular respiration and cell death in ischemic skeletal muscle

Article https://doi.org/10.1038/s41467-023-39820-8 Cytochrome clysine acetylation regulates cellular respiration and cell death in ischemic skeletal muscle Paul T. Morse 1,9 , Gonzalo Pérez-Mejías 2,9 ,JunmeiWan 1 ,AliceA.Turner 1,3 , Inmaculada Márquez 2 ,HasiniA.Kalpage 1 , Asmita Vaishnav 3 , Matthew P. Zurek 1,3 , Philipp P. Huettemann 1 , Katherine Kim 1 , Tasnim Arroum 1 , Miguel A. De la Rosa 2 , Dipanwita Dutta Chowdhury 3 ,IcksooLee 4 , Joseph S. Brunzelle 5 , Thomas H. Sanderson 6 ,MohH.Malek 7 , David Meierhofer 8 , Brian F. P. Edwards 3 ,IreneDíaz-Moreno 2 & Maik Hüttemann 1,3 Skeletal muscle is more resilient to ischemia-reperfusion injury than other organs. Tissue specific post-translational modifications of cytochrome c (Cytc) are involved in ischemia-reperfusion injury by regulating mitochondrial respiration and apoptosis. Here, we describe an acetylation site of Cytc, lysine 39 (K39), which was mapped in ischemic porcine skeletal muscle and removed by sirtuin5 in vitro. Using purified protein and cellular double knockout models, we show that K39 acetylation and acetylmimetic K39Q replacement increases cytochrome coxidase (COX) activity and ROS scavenging while inhibiting apoptosis via decreased binding to Apaf-1, caspase cleavage and activity, and cardiolipin peroxidase activity. These results are discussed with X-ray crystallography structuresof K39 acetylated (1.50 Å) and acetylmimetic K39Q Cytc(1.36 Å) and NMR dynamics. We propose that K39 acetylation is an adaptive response that controls electron transport chain flux, allowing skeletal muscle to meet heightened energy demand while simultaneously providing the tissue with robust resilience to ischemiareperfusion injury. Cytochrome c(Cytc) is a small, 104 aminoacid, essential protein with a covalently linked heme group that performs vital functions in both lifesustaining and cell death pathways. In the mitochondrial intermembrane space (IMS), it acts as a single electron carrier from complex III to complex IV, also known as cytochrome coxidase (COX), as part of the electron transport chain (ETC)1. Apart from respiration, Cytcplays a key role in intrinsic apoptosis, where its release from the IMS into the cytosol allows Cytcto bind apoptosis protease activating factor-1 (Apaf-1), activating the apoptosome, which in turn activates caspase-9 and the downstream caspase cascade2–4.Cytcpossesses Received: 11 October 2022 Accepted: 30 June 2023 Check for updates 1 Center for Molecular Medicine and Genetics, Wayne State University, Detroit, MI 48201, USA. 2 Instituto de Investigaciones Químicas, Universidad de Sevilla - CSIC, 41092 Sevilla, Spain. 3 Department of Biochemistry, Microbiology, and Immunology, Wayne State University, Detroit, MI 48201, USA. 4 College of Medicine, Dankook University, Cheonan-si Chungcheongnam-do 31116, Republic of Korea. 5 Life Sciences Collaborative Access Team, Northwestern University, Center for Synchrotron Research, Argonne, IL 60439, USA. 6 Department of Emergency Medicine, University of Michigan Medical School, Ann Arbor, MI 48109, USA. 7 Department of Health Care Sciences, Eugene Applebaum College of Pharmacy & Health Sciences, Wayne State University, Detroit, MI 48201, USA. 8 Max Planck Institute for Molecular Genetics, 14195 Berlin, Germany. 9 These authors contributed equally: Paul T. Morse, Gonzalo Pérez-Mejías. e-mail: i[email protected];[email protected] Nature Communications | (2023) 14:4166 1 1234567890():,; 1234567890():,; cardiolipin peroxidase activity, another function related to cell death5. Cardiolipin is a lipid that makes up about 20% of the inner mitochondrial membrane (IMM) lipid composition. During apoptosis, cardiolipin is oxidized by Cytc, which assists in outer membrane permeabilization6,7. Additionally, Cytcscavenges reactive oxygen species (ROS)8,9. Given the numerous critical processes carried out by Cytc, the protein is tightly regulated by several mechanisms: allosteric regulation by ATP, expression of tissue-specific isoforms (somatic and testes), and post-translational modifications10. Six tissue-specific phosphorylation sites and one prostate cancerspecific acetylation site have been characterized on Cytc.Fiveofthe modifications have previously been found to reduce the capacity of Cytcto perform its roles in respiration and apoptosis. The phosphorylations are tyrosine 97 in the heart11,12, tyrosine 48 in the liver13–15, threonine 28 and threonine 58 in the kidney16–18, and serine 47 in the brain18–20. These phosphorylations were generally found under basal conditions and lost during ischemia, providing a mechanism for reperfusion injury when oxygen is reintroduced into the tissue. Another phosphorylation of Cytcon threonine 49 (numbering based on mature Cytcwhich lacks the start methionine) was reported to increase in aged mouse hearts21, but, as the authors do not report knocking out endogenous Cytcprior to transfecting in recombinant, phosphomimetic Cytc, data interpretation regardingfunctional effects of this modification are limited. In prostate cancer, acetylation of lysine 53 promotes two central hallmarks of cancer: switching to Warburg metabolism and evasion of apoptosis22,23. Here,wedescribetheidentification and characterization of an acetylation site on lysine 39 (K39) of skeletal muscle Cytcpurified from porcinetibialisanterior (TA) skeletal muscle.Uniquely,K39acetylation wasfound after ischemia and was absent under basal conditions.Much like other tissue types, skeletal muscle can experience ischemiareperfusion injury. Total knee arthroplasty and other conditions that require the application of a tourniquet commonly cause ischemiareperfusion injury in skeletal muscle24,25. Other conditions, such as peripheral artery disease or compartment syndrome, also induce ischemic changes in skeletal muscle26,27. Despite this, skeletal muscle is uniquely resistant to ischemia-reperfusion injury compared to other tissues, such as the brain28,29.Usingpurified acetylated and recombinant acetylmimetic protein and a mammalian cell overexpression acetylmimetic system, we studied the effects of this modification on Cytcfunction. We report that this acetylation stimulates respiration in the ischemic muscle to meet the increased energy demand. Additionally, this modification reduces the pro-apoptoticcapabilities ofthe protein, protecting the skeletal muscle under conditions of stress. In conjunction with X-ray crystallography and nuclear magnetic resonance (NMR) studies of our Cytcvariants, we propose that K39 acetylation of Cytccontributes to the resilience of skeletal muscle to ischemia-reperfusion injury. Results Tibialis anterior muscle Cytcis acetylated on lysine 39 after ischemia Porcine tibialis anterior (TA) muscle was harvested and either immediately snap-frozen (control TA) or exposed to ischemic treatment prior to being snap-frozen (ischemic TA). Cytcwas purified from 6 control and 6 ischemic TA samples (3 male and 3 female Sus domesticus per group; 6 months old) under conditions that preserve post-translational modifications. Mass spectrometry analysis revealed an acetylation site on K39 of Cytcthat was not present in any of the control samples but was detected in 3 of the ischemic TA samples (1 male and 2 female), making this an ischemia-specific acetylation (Fig. 1A). All mass spectrometry data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org)viathe PRIDE30 partner repository, with the dataset identifier PXD040915. Note: In the raw mass spectrometry results, amino acids were numbered including the start methionine (mature Cytclacks the start methionine), therefore the numbering will be one number higher than what is reported here for the mature protein. Lysine acetylation of Cytcwas also identified on K27, K79, and K86 in some samples, however, these acetylations were present in both the control and ischemic TA samples (Supplementary Table 2). K88 acetylation was also identified, but the intensity was so low that occupancy could not be calculated. K39 acetylation was never present in the control TA samples and was gained during ischemia, making it an adaptive response to ischemia. K39 acetylation wasnot present when control TA muscle Cytcwas analyzed by mass spectrometry or western blot (Fig. 1B).Acetylationof TA muscle Cytcwas induced at 45min of ischemia and appeared to increase slightly at 60 min of ischemia (Fig. 1C). The acetylation of Cytc purified from ischemic TA muscle was removed in vitro by sirtuin5 (Fig. 1D), a mitochondria-localized deacetylase. Bacterial overexpression and protein purification of WT, acetylmimetic K39Q, and additional Cytcvariants To study the effects of K39 acetylation in vitro, a recombinant purified protein system was employed as we have previously done for lysine 53 acetylationinprostatecancer22,23.Theoverexpressed wild-typeprotein (WT) is nonacetylated. To mimic lysine acetylation, K39 was mutated to a glutamine residue (K39Q), which is a polar, uncharged residue like acetyl-lysine. Using glutamine as a mimetic for acetyl-lysine is an established technique for both Cytc22,23 and other proteins in general due to their similar side chains, which are both uncharged amides31–34. As another control, K39 was replaced with arginine (K39R), which retains a positive charge at the residue like unacetylated lysine. As an additional experimental control, K39 was mutated to a glutamate residue (K39E), which places a negative charge at the residue. Recombinant WT, K39R, K39Q, and K39E proteins were purified after bacterial overexpression. LabSafe GEL Blue staining of the recombinant proteins resolved on an SDS polyacrylamide gel demonstrates high purity of Cytc(Fig. 1E). The UV-Vis spectra of the reduced recombinant proteins also demonstrate purity and proper folding due to the presenceof the characteristic α,β,andγpeaks (Fig. 1F; note: the reduced K39Q spectrum overlaps with the reduced K39R spectrum). The presence of the 695 nm peak in the oxidized spectra indicates successful incorporation and correct coordination of the heme iron in all mutants (Fig. 1F, insert). Acetylmimetic K39Q CytcandinvivoacetylatedporcineTA muscle Cytcresult in increased cytochrome coxidase activity and decreased caspase-3 activity Regulatory-competent COX was purified from porcine heart under conditions that preserve post-translational modifications, and COX activity was measured with a Clark-type oxygen electrode. The maximal oxygen consumption rates using recombinant K39Q and K39E proteins were increased by 38% and 72%, respectively, compared to recombinant WT protein (Fig. 1G). Similar measurements with in vivo K39 acetylated Cytcisolated from ischemic porcine TA muscle demonstrated an oxygen consumption rate increase of 58% compared to in vivo unacetylated Cytcisolated from control porcine TA muscle (Fig. 1H). The interaction of Cytcand COX is primarily electrostatic in nature35,36. Thus, to test the hypothesis that change in charge at residue K39 drives maximal functional effects, the K39E mutant was included. Given that lysine is positively charged, glutamine is neutral but polar like acetyl-lysine, and glutamine is negatively charged, the K39Q and K39E replacements produce a change of 1 and 2 charge units compared to WT, respectively. As predicted, K39E Cytcshows a more pronounced effect to activate COX activity compared to K39Q and WT, producing the trend: WT < K39Q < K39E Cytc(Fig. 1G). Overall, these data indicate that acetylation increases oxygen consumption rate and that K39Q replacement is a good mimetic for acetylated K39. Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 2 Previous research on the structure of the apoptosome has indicatedthat K39 is a part of the Apaf-1 binding domain3. Therefore, K39Q acetylmimetic and in vivo K39 acetylated Cytccould regulate apoptosis. Interestingly, the caspase-3 activities of recombinant K39Q and K39E proteins demonstrated a significant decrease of 90% and 87%, respectively, compared to recombinant WT protein (Fig. 1I), indicating that K39 acetylation protects the cells from cell death. The caspase-3 activity of the in vivo K39 acetylated Cytcisolated from ischemic porcine TA muscle decreased caspase-3 activity by 45% compared to the in vivo unacetylated Cytcisolated from control porcine TA muscle (Fig. 1J). These data confirmthat K39Q is a good mimetic for acetylated K39andthatK39servesasanimportantregulatorforbindingApaf-1. a b ef 300 400 500 600 700 0 1 2 3 4 Wavelength (nm) Absorbance WT K39R K39Q K39E 600 650 700 750 0.10 0.15 0.20 0.25 Cytc 10- 15- 25- 35- 40- 50- 70- 100- 180- 130- WT K39R K39Q K39E TA muscle control TA muscle ischemia -15 -25 A cetyl-lysine Cytc 0 5 10 15 20 [Cytc ] ( M) WT K39R K39Q K39E p = 0.0332 p = 0.0006 p < 0.001 μ ghij 0 2 4 6 p < 0.0001 0510 15 20 25 Turnover number (s -1 ) Turnover number (s -1 ) Control TA (unacetylated) Ischemic TA (acetylated) -15 -25 -15 -15 Acetyl-lysine Cytc Sirtuin5 NAD + Ischemic TA Cytc Control TA Cytc ++ + ++ + ++ + cd Acetyl-lysine Cytc -15 -15 -10 IB: IgG 0 min(s) 15 min(s) 30 min(s) 45 min(s) 60 min(s) IP: Cytc Duration of Ischemia: WT K39R K39Q K39E 0 20 40 60 80 100 120 Caspase-3 Activity (% of WT) p = 0.0189 p < 0.001 p < 0.001 ControlTA( Unacetylated) Ischemic TA( Acetyled) 0 20 40 60 80 100 120 Caspase-3 Activity (% of Control TA) p < 0.0001 kDa: kDa: kDa: kDa: Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 3 Acetylmimetic K39Q Cytcdemonstrates an increased rate of oxidation and decreased rate of reduction In addition to respiration and apoptosis, Cytcalso functions as a ROS scavenger. Two major types of ROS that Cytcdetoxifies in the IMS are superoxide and H 2 O 2 . The effect of K39 acetylation on ROS scavenging was studied in reaction with H 2 O 2 (rate of oxidation) and superoxide (rate of reduction). To determine the rate of oxidation, recombinant ferro-(Fe2+)-Cytcproteins were reacted with 100 µMH 2 O 2 ,andthe initial reaction was monitored spectrophotometrically via the absorbance of the αpeak at 550 nm which is present in reduced, ferro-(Fe2+)- Cytcand greatly diminished in the oxidized, ferri-(Fe3+)-Cytc. The rates ofoxidationofK39QandK39E increasedby66%and63%,respectively, compared to WT (Fig. 2A). This indicates that K39Q Cytcis a better scavenger of H 2 O 2 than WT. Superoxide is a physiologically relevant source of mitochondrial ROS that is produced at complexes I and III37. For the rate of reduction, superoxide was generated by the hypoxanthine/ xanthine oxidase reaction system. To determine the rate of reduction, recombinant ferri-(Fe3+)-Cytcproteins were reacted with superoxide, and the initial reaction was monitored spectrophotometrically at 550 nm. The rates of reduction of recombinant K39Q and K39E Cytcdecreased 94% and 54%, respectively, compared to WT (Fig. 2B). Acetylmimetic K39Q Cytcdemonstrates increased redox potential The midpoint redox potential of native Cytcis between that of complex III and COX, facilitating efficient electron transfer. The literature rangefornativeCytcmidpoint redox potentials is 220 to 270 mV38.The midpoint redox potentials for the recombinant WT, K39R, K39Q, and K39E proteins were measured spectrophotometrically using the equilibration method39 and the measured values fell within the reported range (Fig. 2C). Acetylmimetic K39Q Cytcdemonstrates no change in heme stability High levels of ROS cause Cytcto lose functionality by degrading the catalytichememoiety.Thestabilitiesofthehemegroupweremeasured spectrophotometrically by tracking the dissipation of the characteristic heme Soret peak at 408 nm after the challenge with 3 mM H 2 O 2 . There was no difference in heme degradation between recombinant WT, K39R, and K39Q proteins, which decreased 86%, 86%, and 88%, respectively, in the Soret peak absorbance at 800 s (Fig. 2D). However, recombinant K39E protein demonstrated a significant reduction in heme degradation, with only a 70% decrease in the Soret peak absorbance at 800 s. Acetylmimetic K39Q Cytcdemonstrates reduced cardiolipin peroxidase activity A secondary pro-apoptotic function of Cytcis its cardiolipin peroxidase activity. During apoptosis, the peroxidase activity of Cytcis increased, and cardiolipin peroxidation can occur in the presence of H 2 O 2 catalyzed by Cytc. This process facilitates the release of Cytcinto the cytosol, committing the cell to apoptosis40. The cardiolipin peroxidase activities were measured via the fluorescence of resorufin. The Fig. 1 | Ischemic TA muscle Cytcis acetylated on lysine 39 and acetylmimetic K39Q and acetylated Cytcincreases COX activity and decreases caspase-3 activity. A Representative mass spectrum of Cytcpeptide K(acetyl)TGQAPGFSYTDANKNK identifying CytcK39 acetylation in ischemic, but not control, samples (n = 12: 6 control TA (3 male and 3 female) and 6 ischemic TA (3 male and 3 female with 1 male and 2 female samples demonstrating K39 acetylation); ProteomeXchange Consortium, PRIDE: PXD040915). BCytcis acetylated in ischemic but not control TA muscle as shown after immunoprobing for acetyl-lysine and Cytc(n =3).CRepresentative IP-Immunoblot experiment of porcine TA muscle showing acetylation state of Cytcat 0, 15, 30, 45, and 60 min of ischemia (n = 3). DRepresentative in vitro deacetylation assay showing that sirtuin5 removes acetylation from Cytc(n = 3). ERepresentative LabSafe GEL Blue stained 10% tris-tricine SDS-PAGE gel showing purity of the recombinant Cytcpurified from bacteria after overexpression (n = 3). FReduced Cytcspectra and oxidized Cytcspectra (inset) of recombinant Cytcindicate correct folding of the proteins (the reduced K39Q spectrum overlaps with the reduced K39R spectrum). Oxygen consumption rate of 26.7 nM pig heart COX was measured using an Oxygraph+ system at 25°C. GThe recombinant WT, K39R, K39Q, and K39E proteins (n = 3) or Hcontrol and ischemic purified pig TA muscle Cytcsamples(n=5)weretitratedatconcentrationsof0,2,5, 10, 15, 20, and 25 μMorinjectedat5μM, respectively. For G, a one-way ANOVA comparing the mean of each mutant with the mean of the control mutant (WT) at the 25 μM condition with the Dunnett post-hoc test was used. For H, a student’s two-tailed t-test assuming equal variance was used. ICytosolic extracts from Cytc knockout embryonic fibroblasts were incubated with the recombinant WT, K39R, K39Q, and K39E proteins (n = 5 for WT, K39Q; n = 4 for K39R, K39E) or (J) control and ischemic purified pig TA muscle Cytcsamples (n = 4) for 2.5 h at 37°C. Rhodamine fluorescence that resulted from caspase-3 mediated cleavage of Z-DEVD- R110 was used as a measure of caspase-3 activity. For I, a one-way ANOVA comparing the mean of eachmutant with the mean of the control mutant (WT) with the Dunnett post-hoc test was used. For J, a student’s two-tailed t-test assuming equal variance was used. Data are represented as means ± standard deviation. abc de WT K39R K39Q K39E 0.0 0.1 0.2 0.3 0.4 0.5 Rate of Reduction (μM/s) p < 0.0001 p = 0.0016 WT K39R K39Q K39E 0.0 0.2 0.4 0.6 0.8 1.0 Rate of Oxidation (μM/s) p = 0.0033 p = 0.0042 WT K39R K39Q K39E 0 100 200 300 Redox Potential (mV) p = 0.0162 p = 0.0110 p < 0.0001 0 200 400 600 800 0 50 100 Time (s) Heme Degradation (% of initial Abs 408 nm) WT K39R K39Q K39E p = 0.0010 0% 20% 30% 50% 0 5 10 15 20 Cardiolipin Percentage Fluorescence/Time (au/min) WT K39R K39Q K39E p < 0.0001 p = 0.1401 p = 0.0022 p = 0.0007 p = 0.1196 p = 0.0011 p = 0.0020 p = 0.0121 p = 0.0238 p < 0.0001 p < 0.0001 p = 0.0006 Fig. 2 | Acetylmimetic K39Q Cytcdemonstrates altered rate of oxidation by H 2 O 2 , rate of reduction by superoxide, midpoint redox potential, and cardiolipin peroxidase activity. A Initial rate of oxidation of reduced recombinant Cytc proteins by 100 μMH 2 O 2 (n = 3). BInitial rate of reduction of oxidized recombinant Cytcproteins by superoxide generated with a hypoxanthine/xanthine oxidase system (n = 4). CMidpoint redox potentials of recombinant Cytcproteins were measured using the equilibration method with DCIP as the referencecompound (n =3).DHeme degradations of oxidized recombinant Cytcproteins after the addition of 3 mM H 2 O 2 (n = 3). ECardiolipin peroxidase activities of recombinant Cytc proteins with liposomes containing 0%, 20%, 30%, and 50% tetrazolyl-cardiolipin (TOCL) were measured using resorufinfluorescenceaftertheadditionof5 μMH 2 O 2 (n = 4). Data are represented as means ± standard deviation. A one-way ANOVA comparing the mean of each mutant with the meanofthe control mutant (WT)with the Dunnett post-hoc test was used. For D, the 800 s condition specifically was compared. Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 4 Cytcproteins oxidize cardiolipin, which in turn oxidizes Amplex red to resorufin. The cardiolipin peroxidase activity of recombinant K39Q Cytcwas significantly decreased compared to recombinant WT protein, ranging from 17% to 60% decreased fluorescent signal in lipid vesicles composed of 20%, 30%, and 50% cardiolipin (Fig. 2E). These data, along with the reduction in caspase-3 activity reported above for acetylated and acetylmimetic Cytc, support an anti-apoptotic role of K39 acetylation. Crystallography results Acetylmimetic K39Q and K39 acetylated Cytcproteins were crystallized and analyzed via X-ray crystallography. The K39Q structure (Supplementary Table 1: 8DZL) at 1.36 Å has continuous omit map density for residue 39 in chain-C when contoured at 1.0 root mean square deviation (RMSD) and for all four chains (chains A through D) when contoured at 0.7 RMSD (Fig. 3A–3D). The respective backbone RMSD values for the four molecules (chains A through D) relative to chain A of the native structure (5C0Z.pdb) are 0.279, 0.323, 0.362, and 0.337. The overall real space correlation coefficient (RSCC) is 0.928. The only disordered residues are Lys25-His26 in chains C and D which have broken density for their backbone atoms at 1.0 RMSD and RSCC values ranging from 0.624 to 0.767. Their objectively refined conformations match the conformations for chains A and B in the K39Q structure and chain A in the native protein. The structure of porcine K39 acetylated Cytc(Supplementary Table 1: 8VDX) at 1.5 Å has a lower overall RSCC value of 0.890, but all residues have continuous backbone density at 1.0 RMSD. The fractional occupancies for the four acetyl groups (Fig. 3E–3H) have an average value of 0.75 ± 0.26 as shown by Phenix.Refine script (Supplementary Fig. 1A). The respective backbone RMSD values for the four molecules (chains A through D) relative to chain A of the native structure (5C0Z.pdb) are 0.260, 0.307, 0.274, and 0.266. As anticipated, given the longer side chain, the K39 acetylated Cytc demonstrates greater variation at residue 39 compared to the K39Q Cytc(Supplementary Figure 1B, 1C). Interestingly, acetylmimetic mutations, K39Q (8DZL) and K53Q (7LJX)22, tend to more globally perturb the entire Cytcstructure compared to phosphomimetic mutations, T28E (5DF5)16 and S47E (6N1O)19, which show much fewer differences compared to native, rodent WT (5C0Z)19 (Supplementary Figure 1D). Specifically, both the K39Q and K53Q structures have variation in the D50-to-G60 and N70-to-E90 regions, while the T28E and S47E structures have greatly reduced variability compared to the WT in these regions and globally, highlighting that acetylation may have stronger effects on Cytcstructure and function than phosphorylation. For NMR analyses and electrostatic surface potentials, there was a switch-like effect based whether K39 was positively charged or not with recombinant WT and K39R Cytcbehaving in one manner while K39Q and K39E Cytcbehaved in a different manner. (Supplementary Fig. 2). Prior to measurement, protein purity was confirmed via Coomassie blue staining (Supplementary Fig. 3A). Proper protein folding and heme coordination were evaluated by 1D 1H NMR spectra (Supplementary Fig. 3B, C). Point mutations and monomerization state were confirmed via tryptic digestion analyses and dynamic light scattering, respectively (Supplementary Fig. 4). The chemical shift perturbations (CSP) of the amide backbone signals of the 2D 15N-1H Heteronuclear Single Quantum Correlation (HSQC) NMR spectra between the recombinant WT protein and the recombinant K39R, K39Q, and K39E proteins were calculated. Resonance assignment for each recombinant protein was confirmed via recording 3D 15N-1H Nuclear Overhauser Effect Spectroscopy (NOESY)-HSQC and 3D 15N-1H TOtal Correlation SpectroscopY (TOCSY)-HSQC spectra (Supplementary Figs. 5–7). The magnitude of the perturbation from the WT spectra induced by the mutations followed the change in charge at the residue, as K39E demonstrated greater perturbation than K39Q which demonstrated greater perturbation than K39R. Interestingly, the mutations at residue 39 not only affected the chemical environment of nearby residues, but also residues located in the M80-containing loop and the K55-to-W59 stretch (Fig. 4A, B). Mutation at K39 did not significantly alter the rotational correlationaltime (τ c ) values (WT5.75 ± 0.03 ns, K53Q 5.66 ± 0.02 ns), though these values are lower than that of the human WT Cytcprotein (6.33 ± 0.02 ns)14. However, the K39Q substitution slightly affected both relaxation rates (R 1 and R 2 ) parameters (Figs. 5A, B, Supplementary Figs. 8, 9). Comparing the differences in R 1 ,R 2 ,and15N{1H} NOE parameters revealed that the M80-containing loop of WT exhibited higher mobility in the ps-to-ns timescale, as the 15N{1H} NOE values fortheP76- to-K88 segment are smaller for the recombinant WT protein. Similar behavior is observed for the N-terminal and C-terminal αhelixes and the position of the mutation. Surprisingly, the K55-to-W59 segment, spatially close to the propionate HP7 of the heme group and to the proximity of the mutation, showed almost no difference at the 15N{1H} NOE values for the recombinant acetylmimetic K39Q protein. However, in this region, the spin-lattice (R 1 ) and spin-spin (R 2 ) relaxation rates of the recombinant WT protein revealed a possible conformational exchange in the μs-to-ms timescale. Additional 15N{1H} NOE experiments were recorded with the recombinant K39E and K39R proteins at 500 MHz (Supplementary Fig. 10). a Gln39-A Gln39-B Gln39-C Gln39-D b cd ef h g Acetyl-Lys39-A Acetyl-Lys39-B Acetyl-Lys39-C Acetyl-Lys39-D Fig. 3 | Omit density maps of K39Q and K39 acetylated Cytccrystal structures at residue 39. A–DOmit density maps of chain A through D at residue 39 of K39Q crystalstructure(8DZL.pdb,1.36Å).E–HOmitdensity mapsof chain AthroughD at residue 39 of ischemic porcine Cytccrystal structure (8DVX.pdb, 1.50 Å). The refined, fractional occupancies for the four acetyl groups are 0.37, 0.80, 0.95, and 0.89,respectivelyforchain-Athroughchain-D.Allomitdensitymapsarecontoured at 0.5 RMSD. Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 5 Cells expressing acetylmimetic K39Q Cytcdemonstrate activation of mitochondrial respiration To further study the effects of K39 acetylation, we generated cell lines stably expressing WT, K39R, K39Q, and K39E Cytc. These cell lines were created by transfecting the Cytcconstructs into Cytcdouble knockout mouse embryonic fibroblasts. This double knockout system, where both the somatic and testes isoforms of Cytcwere knocked out, was used to ensure that knockout of the somatic isoform alone would not induce expression of the testes isoform41. Additionally, a Cytc-null empty vector (EV) cell line was employed as an additional control for all cellular experiments. Clones of each Cytcvariant were selected based on similar expression of Cytc(Fig. 6A). Mitochondrial biogenesis was found to be similar between the cell lines as assessed via western blot for mitochondrial transcription factor A (TFAM), peroxisome proliferatoractivated receptor gamma coactivator 1-alpha (PGC-1α), and citrate synthase (CS) (Supplementary Fig. 11A). The cell line expressing K39Q Cytcdemonstrated a 51% increase in ATP content compared to the WT (Fig. 6B). Intact cellular respiration was measured using a Seahorse bioanalyzer. The reaction between Cytcand COX is generally considered to be the rate-limiting step of the electron transport chain (reviewed in ref. 10). Therefore, mutations to Cytcmay impact intact cellular respiration. The basal oxygen consumption rate of the K39Q cell line demonstrated a 70% increase compared to WT (Fig. 6C, D). Additionally, theK39Qcelllineshowedincreasesinnon-mitochondrialoxygencon- sumption, proton leak, ATP-coupled respiration, maximal respiration rate, and spare respiratory capacity compared to WT (Supplementary Fig. 12). Altogether, these data indicate that K39Q acetylmimetic replacement enhances electron flow through the electron transport chain resultinginanincreaseincellularrespiration. Cells expressing acetylmimetic K39Q Cytchave higher mitochondrial membrane potential which corresponds to higher ROS production Given that the in vitro protein and intact cellular data indicated an increase in respiration due to K39 acetylation or acetylmimetic replacement, we hypothesized that this would translate to an increase inthe mitochondrial membrane potential (ΔΨ m ). ΔΨ m levels of the cell lines were measured using JC-10, a ratiometric probe. Overall, the trend in ΔΨ m matchedthetrendinrespirationwithEV<WT<K39R<K39Q< K39E. ΔΨ m of the K39Q Cytccell line demonstrated a 42% increase in red/green fluorescence compared to WT, indicating higher ΔΨ m (Fig. 6E). Previous research has established a relationship between ΔΨ m and mitochondrial ROS production. Mitochondrial ROS levels of the cell lines were measured using MitoSOX. The production of mitochondrial ROS matched the trend in respiration rate and ΔΨ m with increasing ROS levels for EV < WT < K39R < K39Q < K39E. MitoSOX fluorescence of the K39Q cell line increased 50% compared to WT (Fig. 6F). Altogether, these data suggest that K39Q acetylmimetic replacement promotes respiration, which in turn increases ΔΨ m and mitochondrial ROS production. Fig. 4 | Nuclear magnetic resonance structure analysis of Cytcmutants. AChemical shift perturbation (CSP) bar plot of Cytcbackbone amide signals (ΔCSP X =CSP X -CSP WT )ofreducedCytcmutants (K39E in pink, K39Q in yellow, and K39R in purple). A scheme of the secondary structure elements is included at the top andthemutatedpositionisindicatedwith dashedlines.BSurfacemapsofNMR ΔCSP of reduced Cytcmutants. Each map shows residues colored according to ΔCSP values calculated previously. The Heme group is in green and non-assigned and proline residues are in tan. Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 6 Cells expressing acetylmimetic K39Q Cytcdemonstrate increased mitochondrial membrane potential but not ROS production after oxygen-glucose deprivation/reoxygenation After exposure to stressful conditions, such as reperfusion following ischemia, ΔΨ m increases due to ETC activation, leading to an exponential increase in mitochondrial ROS42. Because K39 acetylation was present only after ischemia, but not under normal conditions, the role of this modification was further assessed using an oxygen-glucose deprivation/reoxygenation (OGD/R) model43, which simulates ischemia-reperfusion injury in cell culture as previously described20. ΔΨ m and mitochondrial ROS levels of the cell lines were measured at normoxia (control) and after OGD/R. For ΔΨ m , each cell line showed asignificant change in JC-10 red/green fluorescence ratio after exposure to OGD/R compared to the normoxia control. Interestingly, followingOGD/R,theK39Qcelllinedemonstratedanincreaseofonly4% JC-10 red/green fluorescence compared to the WT cell line, mitigating the larger difference seen when the experiment is run under conditions of normoxia (Fig. 6G). For mitochondrial ROS, all cell lines except K39Q show a significant change in MitoSOX fluorescence after OGD/R exposure compared to normoxia control (Fig. 6H). These data show that K39Q Cytcincreases respiration, ΔΨ m , and ROS already under baseline conditions versus WT, but it does not lead to a further increase following OGD/R, potentially protecting the cells from additional ROS upon reperfusion. Cells expressing acetylmimetic K39Q Cytcare better protected from cell death after exposure to H 2 O 2 , prolonged OGD/R, or ER stress Cell death of the cell lines was analyzed via annexin V/propidium iodide (PI) staining by flow cytometry following treatment with H 2 O 2 , Fig. 5 | Difference between WT and acetylmimetic K39Q nuclear magnetic resonance parameters. A Average differences (means) in longitudinal relaxation rate R 1 (top), transversal relaxation rate R 2 (middle), and heteronuclear 15N{1H} NOE (bottom) between the experimental values at 500 MHz for the reduced forms of K39Q and WT Cytc, plotted as a function of the residue number (n = 2). Each difference was calculated from a set of relaxation parameters measured on two biologically independent samples for WT and K39Q Cytc. Raw data is represented in Supplementary Fig. 8. A scheme of the secondary structure elements is included at the top and the mutated position is indicated with dashed lines. BRibbon structure of K39Q Cytccolored according to the difference in its dynamic properties from blue (decrease) to red (increase) compared with WT protein. Undetectable or overlapping backbone resonances are in gray. Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 7 OGD/R, or thapsigargin44. Treatment with 400 μMH 2 O 2 for 16 h resulted in significantly lower total cell death of the K39Q cell line with 29% cell death compared to WT with 38% cell death (Fig. 7A). Treatmentwithoxygen-glucosedeprivationfor16hfollowedby1hof reoxygenation resulted in significantly lower total cell death of the K39Q cell line with 22% cell death compared to WT with 29% cell death (Fig. 7B). The highly glycolytic EV cell line appears unable to survive prolonged glucose deprivation, with an increased level of total cell death at 64%. Additionally, treatment with 1 mM thapsigargin for 24 h, which causes ER stress, also resulted in significantly lower total cell death of the K39Q cell line with 33% cell death compared to WT with 49% cell death (Fig. 7C). The representative flow cytometry gating strategy is available in the supplement (Supplementary Fig. 13). Overall, the decrease in total cell death for the K39Q cell line after each of the three treatments is similar to that seen with the in vitro caspase-3 activity using the recombinant K39Q Cytcprotein and K39 acetylated ischemic porcine TA muscle Cytc. Altogether, this highlights the role of K39 acetylation in reducing the apoptotic capabilities of Cytc. Cells expressing acetylmimetic K39Q Cytcshow disrupted pro-apoptotic signaling The binding of positively charged Cytcto its negatively charged binding pocket on Apaf-1 is mediated largely by electrostatic interactions3.Asdiscussedabove,K39ofCytcis known to be part of the binding site to Apaf-145. Removing the positive charge at K39 perturbs this interaction, which in turn reduces downstream apoptotic signaling. In line with this, the cell line expressing acetylmimetic K39Q Cytcshowed reduced levels of Cytcpulled down after Apaf-1 immunoprecipitation compared to the cell line expressing the WT (Fig. 7D). This also translated into reduced cleavage of caspase-9 (Fig. 7E) and caspase-3 (Fig. 7F) in the cell line expressing the acetylmimetic K39Q compared to the WT after treatment with 8 h of staurosporine, a potent inducer of intrinsic apoptosis17,46,47. Discussion In this study, we found that porcine tibialis anterior muscle Cytcis acetylated on K39 after 45min of ischemia. No other types of posttranslational modifications of Cytcwere identified beyond lysine ab c d ef gh WT K39R K39Q K39E EV Cytc GAPDH -15 -40 -35 -55 -25 -10 WT K39R K39Q K39E EV 0 200 400 600 800 1000 ATP (μmol ATP/mg protein) p = 0.0498 p = 0.0359 020406080 0 10 20 30 40 50 Time (min) OCR (pmol/min/μg protein) WT K39R K39Q K39E EV WT K39R K39Q K39E EV -5 0 5 10 15 20 25 OCR (pmol/min/μg protein) p = 0.0006 p < 0.0001 p < 0.0001 p < 0.0001 WT K39R K39Q K39E EV 0.0 0.2 0.4 0.6 0.8 1.0 Δψ m (JC-10 red/green fluorescence) p < 0.0001 p < 0.0001 p < 0.0001 WT K39R K39Q K39E EV 0 50 100 150 200 250 300 350 MitoSOX (% of WT) p = 0.0358 p < 0.0001 WT K39R K39Q K39E EV 0.0 0.5 1.0 1.5 Δψ m (JC-10 red/green fluorescence) p < 0.0001 p < 0.0001 p < 0.0001 p < 0.0001 p < 0.0001 p < 0.0001 p < 0.0001 p < 0.0001 p < 0.0001 p = 0.0003 WT K39R K39Q K39E EV 0 100 200 300 400 MitoSOX (% of WT) p = 0.0269 p < 0.0001 p = 0.007 p < 0.0001 p = 0.0024 p = 0.0284 p = 0.0207 kDa: Fig. 6 | Acetylmimetic K39Q Cytcexpressing cells show increased respiration, mitochondrial membrane potential, and ROS production under basal conditions and increased mitochondrial membrane potential but not ROS production after oxygen-glucose deprivation/reoxygenation. A Representative western blot of Cytcdouble knockout cells stably transfected with EV, WT, K39R, K39Q,and K39E constructsimmunoprobed for Cytcand loading control GAPDH (n =3).BATP levels in thestable cell lines (n= 3).CMitochondrialstresstest measured in Seahorse media supplemented with 10 mM glucose and 10 mM pyruvate was performed via sequential injections of 1 µM oligomycin, 2.5 µM carbonylcyanide-3- chlorophenylhydrazone (CCCP), and 1 µM rotenone/antimycin A (n = 4). DBasal oxygen consumption rate (OCR) from mitochondrial stress test (n = 4). EMitochondrial membrane potentials (ΔΨ m ) of cells stably expressing WT, K39R, K39Q, K39E Cytc, and EV were measured using the red/green fluorescence ratio of theJC-10probe(n=9).FMitochondrial ROS production of cells stably expressing WT, K39R, K39Q, K39E Cytc, and EV were measured using MitoSOX fluorescence and normalized to total protein content (n = 9). G,HΔΨ m and mitochondrial ROS production were measured as above at normoxia (gray bars) or after 90 min of oxygen-glucose deprivation followedby 30 minof reoxygenation(red bars) (n = 9). Data are represented as means ± standard deviation. A one-way ANOVA comparing the mean of each mutant with the mean of the control mutant (WT) with the Dunnett post-hoc test wasused. For G and H, a student’s two-tailed t-test assuming equal variance was used to compare the normoxia vs. OGD/R condition within each mutant. Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 8 acetylation. The identification of this acetylation, which is gained during ischemia, was unexpected because our previous work in other tissues revealed Cytcphosphorylation under normal conditions,which was lost during ischemia. We have shown that phosphorylation is an important regulatory modification on Cytcpresent under physiological conditions which inhibits respiration and/or apoptosis. However, becausethesephosphorylationsarelost duringischemia,weproposed that their loss causes reperfusion injury through increased ETC activity, ΔΨ m hyperpolarization, and ROS bursts eventually resulting in cell death10,48. Here, we characterized an acetylation on K39 of Cytcin skeletal muscle which is gained after ischemia and was not detected under basal conditions, reversing the pattern typically seen with Cytc phosphorylation. We here report the biological effect of K39 acetylation of Cytc with our Cytcpurified from ischemic TA muscle, which shows a 58% increase in COX activity and a 45% decrease in downstream caspase-3 activity. We further characterized K39 acetylation using a site-directed mutagenesis approach. Previous work on Cytchas found glutamine substitution to be a suitable mimic for acetyl-lysine22,23.Additionally, usingglutamineasthe primarymimicforacetyl-lysine isanestablished technique for other proteins in general due to both side chains being polar but uncharged amides31–34. While lysine is conserved at residue 39 in mammals, glutamine is evolutionarily allowed in some species of plants and yeast49. Interestingly, in some experiments the glutamate substituted Cytcproduced results comparable to those obtained with the acetylmimetic glutamine, suggesting that other modifications of this residue, including the introduction of a negative charge via the glutamate side chain, on this highly evolutionary optimized protein can cause similar functional effects. We purified the mutant proteins after bacterial overexpression and generated Cytcdouble knockout stable cell lines overexpressing the mutant proteins to characterize the effects of the modification. In most experiments, the K39R mutant behaved similarly to WT, and the acetylmimetic K39Q and K39E mutants behaved similarly, indicating a switch-like behavior between positively charged lysine or arginine versus uncharged acetyl-mimic glutamine or negatively charged glutamate. While the glutamate mutant may behave similarly to the glutamine mutant in this publication, this is not generally true for other acetylation sites, and the glutamine should therefore be considered the acetyl-mimetic. It is worth noticing that Cytcpurified from ischemic muscle exists as a mixture of acetylated and unacetylated protein, while acetylmimetic K39Q Cytc mimics 100% acetylation. This may explain the more pronounced effects seen using acetylmimetic K39Q protein vs. Cytcpurified from ischemic muscle, such as for the caspase-3 activity assay. Additionally, the side chain of the K39R mutant features a delocalized positive charge, unlike that of lysine in the WT, and is spatially close to the propionate heme groups, which may explain some of the slight functional changes for K39R compared to WT seen in some assays. For example, OCR in the cell lines increased from WT to K39R to K39Q while ATP levels were similar for the K39R and K39Q cell lines, which WT K39R K39Q K39E EV 0 10 20 30 40 50 Cell death (%) p = 0.0012 p < 0.0001 p < 0.0001 WT K39R K39Q K39E EV 0 10 20 30 40 50 60 70 Cell death (%) p = 0.0143 p = 0.0052 p = 0.0001 p < 0.0001 ab WT K39R K39Q K39E EV 0 10 20 30 40 50 60 Cell Death (%) PI+ p < 0.0001 p < 0.0001 p < 0.0001 Annexin V+ Annexin V+/PI+ c -15 -10 -180 -130 -100 -70 IP: Apaf-1 IgG WT K39R K39Q K39E Cytc A paf-1 IB: def Procaspase-9 Cleaved caspase-9 GAPDH -10 -180 -130 -100 -70 -15 -55 -40 -35 -25 WT K39R K39Q K39E EV Procaspase-3 Cleaved caspase-3 GAPDH -40 -40 -10 -180 -130 -100 -70 -15 -55 -40 -35 -25 WT K39R K39Q K39E EV kDa: kDa:kDa: Fig. 7 | Acetylmimetic K39Q Cytcexpressing cells undergo less cell death and have defective apoptotic signaling. A Annexin V/propidium iodide flow cytometry data for cells stably expressing WT, K39R, K39Q, K39E Cytc, and EV after exposure to 400 µMH 2 O 2 for 16 h (n = 3), B16 h of oxygen-glucose deprivation followed by 1 h of reoxygenation (n = 3), Cor 1 mM thapsigargin for 24 h (n = 3). DRepresentative co-IP-Immunoblot experiment of Cytcdouble knockout cells stably transfected with WT, K39R, K39Q, and K39E constructs immunoprobed for Cytcand loading control Apaf-1 (n = 3). ERepresentative western blot of Cytc double knockout cells stably transfected with EV, WT, K39R, K39Q, and K39E constructs immunoprobed for caspase-9 and loading control GAPDH after exposure to 1 µM staurosporine for 8 h (n = 3). FRepresentative western blot of Cytc double knockout cells stably transfected with EV, WT, K39R, K39Q, and K39E constructs immunoprobed for caspase-3 and loading control GAPDH after exposure to 1 µM staurosporine for 8 h (n= 3). Data are represented as means ± standard deviation. A one-way ANOVA comparing the total cell death mean of each mutant with the total cell death mean of the control mutant (WT) with the Dunnett posthoc test was used. Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 9 and III, respectively. The basal respiration was calculated using the initial OCR values prior to any chemical injections minus the nonmitochondrial oxygen consumption. The proton leak was calculated using the OCR values after the injection of 1 µM oligomycin, which inhibits ATP synthase, minus the non-mitochondrial oxygen consumption. ATP-coupled respiration was calculated using basal respiration minus the proton leak. Maximal respiration was calculated using the OCR values after the injection of 2.5 µM CCCP, which is a mitochondrial uncoupler, minus the non-mitochondrial oxygen consumption. The spare respiratory capacity was calculated using maximal respiration minus the basal respiration. OCR and ECAR were normalized to total protein content in the well after the experiment to control for variation in cell number between cell lines. For post-mitochondrial stress test normalization, cells in each well were lysed using 25 µL RIPA lysis buffer (see above) supplemented with protease inhibitor cocktail and protein concentration was determined by DC protein assay kit (#5000111, Bio- Rad). OCR is reported as pmol O 2 /minute/µgproteinandtheECARis reported as mpH/minute/µgprotein. Mitochondrial membrane potential. The relative mitochondrial membrane potentials (ΔΨ m ) of the stable cell lines expressing WT, K39R, K39Q, K39E, and EV were measured using the JC-10 probe (#ENZ- 52305, Enzo Life Sciences). Cells were seeded at 25,000 cells/well onto 0.1% gelatin coated Costar 96-well plates and cultured overnight in 100 µL/well growth media. The next day, cells were washed with 1x PBS and incubated for 30 min at 37 °C with 3 µM JC-10 in DMEM (#31053-028, Gibco) without phenol red or FBS. After incubation, cells were washed twice with 1x PBS. JC-10 can selectively enter mitochondria where it exists as a monomer at low ΔΨ m with primarily green fluorescence (excitation/emission of 485 nm/527 nm). When ΔΨ m is high, JC-10 aggregates with primarily red fluorescence (excitation/emission of 485 nm/590 nm). Fluorescence was measured in PBS using a Synergy H1 microplate reader (BioTek Instruments Inc.; Winooski, VT, USA). ΔΨ m is reported as the ratio of red/green fluorescence. Mitochondrial ROS. The production of mitochondrial ROS of the stable cell lines expressing WT, K39R, K39Q, K39E Cytc, and EV were measured using MitoSOX (#M36008, Invitrogen). MitoSOX can selectively enter mitochondria where it fluoresces red (excitation/ emission of 510 nm/580 nm) after reaction with superoxide. Cells were seeded at 25,000 cells/well onto 0.1% gelatin coated Costar 96-well plates and cultured overnight in growth media. The next day, cells were washed with 1x PBS and incubated for 30min at 37°C with 5 µM mitoSOX in DMEM (#31053-028, Gibco) without phenol red or FBS. Afterincubation, cells were washed twice with 1x PBSand fluorescence was measured in PBS using a Synergy H1 microplate reader. Mitochondrial ROS production is reported as a percentage of change in fluorescence/µg protein (arbitrary units/µg protein) compared to WT. Oxygen-glucose deprivation/reoxygenation (OGD/R) experiments. To simulate ischemia/reperfusion injury in a cell culture model, stable cell lines expressing WT, K39R, K39Q, K39E Cytc, and EV wereexposed to transient oxygen-glucose deprivation followed by reoxygenation (OGD/R) as previously described20. Cells were seeded and cultured according to the above protocols for JC-10 or MitoSOX. After culturing overnight, the growth media was exchanged for ischemia-mimetic DMEM media (#A14430-01, Gibco) without glucose, phenol red, or FBS that had been bubbled with 95% N 2 and 5%CO 2 .Cellswereincubatedin ischemia-mimetic media at 37 °C for 90min with 1% O 2 and 5% CO 2 in a hypoxia chamber controlled by ProOx 110 Compact O 2 Controller (RRID: SCR_021129, BioSpherix; Redfield, NY, USA) and ProCO2 120 Compact CO 2 Controller (RRID: SCR_021127, BioSpherix). After ischemia, cells were reoxygenated via incubation in glucose-containing DMEM (#31053-028, Gibco) without phenol red or FBS, containing the respective probe, at 37 °C for 30 min with 5% CO 2 . The JC-10 or MitoSOX experiments were then carried out as described in the methods above. Cell death using annexin V/propidium iodide staining. Cell death levels of the stable cell lines expressing WT, K39R, K39Q, K39E Cytc, and EV were quantified after exposure to H 2 O 2 or OGD/R using annexin V and propidium iodide (PI) staining. Annexin V is impermeable to the intact plasma membrane and binds phosphatidylserine, a lipid normally present on the inner leaflet of the plasma membrane which flips to the outer leaflet early during apoptosis. Therefore, annexin V is commonly used as a marker of early apoptotic cells. PI is also impermeabletotheintactplasmamembraneandisaDNAintercalator. During necrosis, holes in the plasma membrane allow PI to enter. Therefore, PI is commonly used as a marker of necrotic cells93.Cells that stain double positive for both annexin V and PI are late apoptotic cells. Cells (1 x 106) were seeded on 10 cm cell culture dishes and cultured overnight in growth media. The next day, cells were exposed to either H 2 O 2 (400 µM for 16 h), OGD/R (16 h oxygen-glucose deprivation/1 h reoxygenation) as described above, or thapsigargin (#328570050, Thermo Fisher Scientific) (1 mM for 24 h). After 48 h from the initial seeding, cells were trypsinized, washed twice with 1x PBS, counted, and a total of 1 x 106cells were resuspended in 1 mL 1x annexin V binding buffer from the FITC annexin V apoptosis detection kit I (#556547; RRID: AB_2869082; BD Pharmingen). A total of 450 µL cell suspension was incubated for 15 min with 6 µLannexinVand6µL PI. Data were collected using a CyFlow Space flow cytometer (Sysmex America, Inc.; Lincolnshire, IL, USA). The results were analyzed using FCS Express 7 software (De Novo Software; Glendale, CA, USA). Caspase cleavage. Caspase cleavageof the cell lines stablyexpressing WT, K39R, K39Q, K39E, and EV were assessed via western blot. Cells (3x10 6) were seeded onto 15 cm cell culture dishes (#CC76823614, USA Scientific; Ocala, FL, USA) and cultured overnight. The next day, cells were treated with 1 μM staurosporine (#NC0662740, Enzo Life Science; Farmingdale, NY, USA) for 8 h, which is known to be a potent inducer of intrinsic apoptosis17,46,47. Cells were washed with 1x PBS, harvested via trypsinization, and pelleted via centrifugation. The cell pellets were resuspended and lysed in 200 µL RIPA lysis buffer supplemented with protease inhibitor cocktail, sonicated, and centrifuged at 16,900 x gfor 20 min at 4 °C to remove cell debris. Ninety μgcell lysates were run on a 10% tris-tricine SDS-PAGE gel, transferred, and blocked as described above. Membranes were incubated with either a 1:500 dilution of mouse anti-caspase-9 antibody (#9508S, Lot #7, Cell Signaling Technology) or a 1:1000 dilution of rabbit anti-caspase-3 antibody (#14220S, Lot #4, Cell Signaling Technology) in blocking reagent overnight at 4 °C. The next day, the membrane was incubated with a 1:8000 dilution of sheep anti-mouse IgG conjugated to horseradish peroxidase secondary antibody or a 1:8000 dilution of donkey anti-rabbit IgG conjugated to horseradish peroxidase secondary antibody, respectively, for 1 h at room temperature. The blots were visualized using Pierce ECL western blotting substrate. The next night, the membranes were incubated with a 1:5000 dilution of mouse anti- GAPDH antibody, as a loading control, in blocking reagent overnight at 4 °C. The next day, the membrane was incubated with a 1:8000 dilution of sheep anti-mouse IgG conjugated to horseradish peroxidase secondary antibody in blocking reagent for 1 h at room temperature. The blots were visualized using Pierce ECL western blotting substrate. Statistical analyses. Data shown represent the mean, and error bars represent the standard deviation between independent trials for protein experiments or individual wells/cell culture dishes for cellular experiments. Statistical analyses of the data were performed using Graphpad Prismv9.4.1 (Graphpad Software; San Diego, CA, USA).Most datawereanalyzed usingone-way ANOVAcomparingthemeanofeach column with the mean of the control column (WT) with the Dunnett Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 16 post-hoc test unless otherwise specified. COX activity of the recombinant purified proteins was performed using one-way ANOVA as described above on the 25 μMCytccondition. Heme degradation of the recombinant purified proteins was performed using one-way ANOVA as described above on the 800 s condition. For annexin V/PI experiments, a one-way ANOVA as described above was performed to compare total cell death (PI+ cells, annexin V+ cells, and annexin V+/PI + cells combined). A student’stwo-tailedttest assumingequalvariance was used for COX activity (to compare porcine TA muscle samples), caspase-3 activity (to compare porcine TA muscle samples), and OGD/ R (to compare normoxia and hypoxia results within each cell line). P values are indicated in the figure legends. The experiments were not randomized, and investigators were not blinded to allocation during experiments or outcome assessment. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Data availability The mass spectrometry data generated in this study have been deposited in the PRIDE repository under accession code PXD040915. The Sus domesticus proteome database UniProtKB with the ID UP000008227 and 26,104 entries (December 2016) was searched. The crystallography data generated in this study are available in the PDB repository with the identifiers 8DZL and 8DVX. The remaining data generated in this study are provided in the Supplementary Information/Source Data file. Source data are provided with this paper. Code availability The Phenix.Refine script is available in Supplementary Figure 1A. References 1. Hüttemann, M. et al. The multiple functions of cytochrome cand their regulation in life and death decisions of the mammalian cell: From respiration to apoptosis. Mitochondrion 11,369–381 (2011). 2. Galluzzi, L. et al. Molecular mechanisms of cell death: recommendations of the Nomenclature Committee on Cell Death 2018. Cell Death Differ. 25,486–541 (2018). 3. Cheng, T. C., Hong, C., Akey, I. V., Yuan, S. & Akey, C. W. A near atomic structure of the active human apoptosome. Elife 5,e17755(2016). 4. Elena-Real, C. A. et al. Cytochrome cspeeds up caspase cascade activation by blocking 14-3-3epsilon-dependent Apaf-1 inhibition. Cell Death Dis. 9, 365 (2018). 5. Kalpage, H. A. et al. Tissue-specific regulation of cytochrome cby post-translational modifications: respiration, the mitochondrial membrane potential, ROS, and apoptosis. FASEB J. 33,1540–1553 (2019). 6. Belikova, N. A. et al. Peroxidase activity and structural transitions of cytochrome cbound to cardiolipin-containing membranes. Biochemistry 45, 4998–5009 (2006). 7. Diaz-Quintana, A., Perez-Mejias, G., Guerra-Castellano, A., De la Rosa, M. A. & Diaz-Moreno, I. Wheel and deal in the mitochondrial inner membranes: the tale of cytochrome c and cardiolipin. Oxid. Med Cell Longev. 2020, 6813405 (2020). 8. Wang, Z. B., Li, M., Zhao, Y. & Xu, J. X. Cytochrome cis a hydrogen peroxide scavenger in mitochondria. Protein Pept. Lett. 10, 247–253 (2003). 9. Pereverzev, M. O., Vygodina, T. V., Konstantinov, A. A. & Skulachev, V. P. Cytochrome c, an ideal antioxidant. Biochem Soc. Trans. 31, 1312–1315 (2003). 10. Kalpage, H. A. et al. Cytochrome cphosphorylation: Control of mitochondrial electron transport chain flux and apoptosis. Int J. Biochem Cell Biol. 121, 105704 (2020). 11. Lee, I. et al. New prospects for an old enzyme: mammalian cytochrome cis tyrosine-phosphorylated in vivo. Biochemistry 45, 9121–9128 (2006). 12. Guerra-Castellano, A. et al. Oxidative stress is tightly regulated by cytochrome cphosphorylation and respirasome factors in mitochondria. Proc.NatlAcad.Sci.USA115,7955–7960 (2018). 13. Yu, H., Lee, I., Salomon, A. R., Yu, K. & Hüttemann, M. Mammalian liver cytochrome cis tyrosine-48 phosphorylated in vivo, inhibiting mitochondrial respiration. Biochim Biophys. Acta 1777,1066–1071 (2008). 14. Moreno-Beltran, B. et al. Structural basis of mitochondrial dysfunction in response to cytochrome cphosphorylation at tyrosine 48. Proc.NatlAcad.Sci.USA114,E3041–E3050 (2017). 15. Guerra-Castellano, A. et al. Mimicking tyrosine phosphorylation in human cytochrome c by the evolved trna synthetase technique. Chemistry 21,15004–15012 (2015). 16. Mahapatra, G. et al. Phosphorylation of cytochrome c threonine 28 regulates electron transport chain activity in kidney: implications for amp kinase. J. Biol. Chem. 292,64–79 (2017). 17. Wan, J. et al. Regulation of respiration and apoptosis by cytochrome c threonine 58 phosphorylation. Sci. Rep. 9,15815(2019). 18. Guerra-Castellano, A., Diaz-Moreno, I., Velazquez-Campoy, A., De la Rosa, M. A. & Diaz-Quintana, A. Structural and functional characterization of phosphomimetic mutants of cytochrome cat threonine 28 and serine 47. Biochim Biophys. Acta 1857,387–395 (2016). 19. Kalpage, H. A. et al. Serine-47 phosphorylation of cytochrome cin the mammalian brain regulates cytochrome coxidase and caspase- 3activity.FASEB J. 33,13503–13514 (2019). 20. Kalpage, H. A., Wan, J., Morse, P. T., Lee, I. & Hüttemann, M. Brainspecific serine-47 modification of cytochrome c regulates cytochrome c oxidase activity attenuating ros production and cell death: implications for ischemia/reperfusion injury and akt signaling. Cells 9,1843(2020). 21. Li, F. et al. Increased cytochrome C threonine 50 phosphorylation in aging heart as a novel defensive signaling against hypoxia/reoxygenation induced apoptosis. Aging (Albany NY) 14,5699–5709 (2022). 22. Bazylianska, V. et al. Lysine 53 acetylation of cytochrome c in prostate cancer: warburg metabolism and evasion of apoptosis. Cells 10, 802 (2021). 23. Marquez, I. et al. Structural and functional insights into lysine acetylation of cytochrome cusing mimetic point mutants. FEBS Open Bio 11, 3304–3323 (2021). 24. Arthur, J. R. & Spangehl, M. J. Tourniquet use in total knee arthroplasty. J. Knee Surg. 32,719–729 (2019). 25. Tran, T. P. et al. Tourniquet-induced acute ischemia-reperfusion injury in mouse skeletal muscles: Involvement of superoxide. Eur. J. Pharm. 650,328–334 (2011). 26. Simon, F. et al. Pathophysiology of chronic limb ischemia. Gefasschirurgie 23,13–18 (2018). 27. Blaisdell, F. W. The pathophysiology of skeletal muscle ischemia and the reperfusion syndrome: a review. Cardiovasc Surg. 10, 620–630 (2002). 28. Eckert, P. & Schnackerz, K. Ischemic tolerance of human skeletal muscle. Ann. Plast. Surg. 26,77–84 (1991). 29. Gillani, S., Cao, J., Suzuki, T. & Hak, D. J. The effect of ischemia reperfusion injury on skeletal muscle. Injury 43,670–675 (2012). 30. Martens, L. et al. PRIDE: the proteomics identifications database. Proteomics 5,3537–3545 (2005). 31. Kamieniarz, K. & Schneider, R. Tools to tackle protein acetylation. Chem. Biol. 16,1027–1029 (2009). 32. Yang, A., Cho, K. & Park, H. S. Chemical biology approaches for studying posttranslational modifications. RNA Biol. 15,427–440 (2018). Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 17 33. Wang, X. & Hayes, J. J. Acetylation mimics within individual core histone tail domains indicate distinct roles in regulating the stability of higher-order chromatin structure. Mol. Cell Biol. 28,227–236 (2008). 34. Okada, A. K. et al. Lysine acetylation regulates the interaction between proteins and membranes. Nat. Commun. 12, 6466 (2021). 35. Roberts, V. A. & Pique, M. E. Definition of the interaction domain for cytochrome c on cytochrome c oxidase. III. Prediction of the docked complex by a complete, systematic search. J. Biol. Chem. 274,38051–38060 (1999). 36. Schmidt, T. R. et al. Rapid electrostatic evolution at the binding site for cytochrome c on cytochrome c oxidase in anthropoid primates. Proc.NatlAcad.Sci.USA102,6379–6384 (2005). 37. Wong, H. S., Dighe, P. A., Mezera, V., Monternier, P. A. & Brand, M. D. Production of superoxide and hydrogen peroxide from specific mitochondrial sites under different bioenergetic conditions. J. Biol. Chem. 292,16804–16809 (2017). 38. Nicholls, D.G. & Ferguson, S.J. Bioenergetics 2,(AcademicPress Limited, San Diego, CA, USA, 1992). 39. Cammack, R. Redox States and Potentials. in Bioenergetics: a Practical Approach (eds.Brown,G.C.&Cooper,C.E.)85-109(Oxford,1995). 40. Kagan,V.E.etal.Cytochromecactsasacardiolipinoxygenase required for release of proapoptotic factors. Nat. Chem. Biol. 1, 223–232 (2005). 41. Vempati, U. D. et al. Role of cytochrome cin apoptosis: increased sensitivity to tumor necrosis factor alpha is associated with respiratory defects but not with lack of cytochrome crelease. Mol. Cell Biol. 27,1771–1783 (2007). 42. Kalogeris, T., Bao, Y. & Korthuis, R. J. Mitochondrial reactive oxygen species: a double edged sword in ischemia/reperfusion vs preconditioning. Redox Biol. 2,702–714 (2014). 43. Ryou, M. G. & Mallet, R. T. An in vitro oxygen-glucose deprivation model for studying ischemia-reperfusion injury of neuronal cells. Methods Mol. Biol. 1717, 229–235 (2018). 44. Crowley, L.C., Marfell, B.J., Scott, A.P. & Waterhouse, N.J. Quantitation of Apoptosis and Necrosis by Annexin V Binding, Propidium Iodide Uptake, and Flow Cytometry. Cold Spring Harb Protoc 2016,953–957 (2016). 45. Zhou, M. et al. Atomic structure of the apoptosome: mechanism of cytochrome c- and dATP-mediated activation of Apaf-1. Genes Dev. 29,2349–2361 (2015). 46. Johansson, A. C., Steen, H., Ollinger, K. & Roberg, K. Cathepsin D mediates cytochrome c release and caspase activation in human fibroblast apoptosis induced by staurosporine. Cell Death Differ. 10, 1253–1259 (2003). 47. Nakazono-Kusaba, A., Takahashi-Yanaga, F., Morimoto, S., Furue, M. & Sasaguri, T. Staurosporine-induced cleavage of alpha-smooth muscle actin during myofibroblast apoptosis. J. Invest Dermatol 119, 1008–1013 (2002). 48. Hüttemann, M. et al. Regulation of mitochondrial respiration and apoptosis through cell signaling: cytochrome coxidase and cytochrome cin ischemia/reperfusion injury and inflammation. Biochim Biophys. Acta 1817,598–609 (2012). 49. Zaidi, S., Hassan, M. I., Islam, A. & Ahmad, F. The role of key residues in structure, function, and stability of cytochrome-c.Cell Mol. Life Sci. 71, 229–255 (2014). 50. Sato, W. et al. Energetic mechanism of cytochrome c-cytochrome c oxidase electron transfer complex formation under turnover conditions revealed by mutational effects and docking simulation. J. Biol. Chem. 291,15320–15331 (2016). 51. Kaim, G. & Dimroth, P. ATP synthesis by F-type ATP synthase is obligatorily dependent on the transmembrane voltage. EMBO J. 18, 4118–4127 (1999). 52. Liu,S.S.Cooperationofa“reactive oxygen cycle”with the Q cycle and the proton cycle in the respiratory chain–superoxide generating and cycling mechanisms in mitochondria. J. Bioenerg. Biomembr. 31,367–376 (1999). 53. Korshunov, S. S., Skulachev, V. P. & Starkov, A. A. High protonic potential actuates a mechanism of production of reactive oxygen species in mitochondria. FEBS Lett. 416,15–18 (1997). 54. Davies,K.J.,Quintanilha,A.T.,Brooks,G.A.&Packer,L.Free radicals and tissue damage produced by exercise. Biochem Biophys. Res Commun. 107, 1198–1205 (1982). 55. O'Neill,C.A.,Stebbins,C.L.,Bonigut,S.,Halliwell,B.&Longhurst,J. C. Production of hydroxyl radicals in contracting skeletal muscle of cats. J. Appl Physiol. (1985) 81, 1197–1206 (1996). 56. Pearson, T. et al. Skeletal muscle contractions induce acute changes in cytosolic superoxide, but slower responses in mitochondrial superoxide and cellular hydrogen peroxide. PLoS One 9, e96378 (2014). 57. Wagner, G. R. & Payne, R. M. Widespread and enzyme-independent Nepsilon-acetylation and Nepsilon-succinylation of proteins in the chemical conditions of the mitochondrial matrix. J. Biol. Chem. 288, 29036–29045 (2013). 58. Konig,A.C.,Hartl,M.,Boersema,P.J.,Mann,M.&Finkemeier,I.The mitochondrial lysine acetylome of Arabidopsis.Mitochondrion 19 Pt B,252–260 (2014). 59. Manning, J. R. et al. Cardiac-specific deletion of GCN5L1 restricts recovery from ischemia-reperfusion injury. J. Mol. Cell Cardiol. 129, 69–78 (2019). 60. Scott, I., Webster, B. R., Li, J. H. & Sack, M. N. Identification of a molecular component of the mitochondrial acetyltransferase programme: a novel role for GCN5L1. Biochem J. 443,655–661 (2012). 61. Wan, C. et al. Panorama of ancient metazoan macromolecular complexes. Nature 525,339–344 (2015). 62. Nakamura, Y., Ogura, M., Tanaka, D. & Inagaki, N. Localization of mouse mitochondrial SIRT proteins: shift of SIRT3 to nucleus by coexpression with SIRT5. Biochem Biophys. Res Commun. 366, 174–179 (2008). 63. Schlicker, C. et al. Substrates and regulation mechanisms for the human mitochondrial sirtuins Sirt3 and Sirt5. J. Mol. Biol. 382, 790–801 (2008). 64. Lim,J.H.etal.Sirtuin1modulatescellularresponsestohypoxiaby deacetylating hypoxia-inducible factor 1alpha. Mol. Cell 38, 864–878 (2010). 65. Olsen, J. V. & Mann, M. Status of large-scale analysis of posttranslational modifications by mass spectrometry. Mol. Cell Proteom. 12, 3444–3452 (2013). 66. Hansen, B. K. et al. Analysis of human acetylation stoichiometry defines mechanistic constraints on protein regulation. Nat. Commun. 10,1055(2019). 67. Prus, G., Hoegl, A., Weinert, B. T. & Choudhary, C. Analysis and interpretation of protein post-translational modification site stoichiometry. Trends Biochem Sci. 44,943–960 (2019). 68. Weinert,B.T.etal.Acetylation dynamics and stoichiometry in Saccharomyces cerevisiae. Mol. Syst. Biol. 10, 716 (2014). 69. Zhou,T.,Chung,Y.H.,Chen,J.&Chen,Y.Site-specificidentifica- tion of lysine acetylation stoichiometries in mammalian cells. J. Proteome Res 15, 1103–1113 (2016). 70. Choudhary, C., Weinert, B. T., Nishida, Y., Verdin, E. & Mann, M. The growing landscape of lysine acetylation links metabolism and cell signalling. Nat. Rev. Mol. Cell Biol. 15,536–550 (2014). 71. Gonzalez-Alonso, J., Richardson, R. S. & Saltin, B. Exercising skeletal muscle blood flow in humans responds to reduction in arterial oxyhaemoglobin, but not to altered free oxygen. J. Physiol. 530, 331–341 (2001). 72. Heinonen, I. H. et al. Regulation of human skeletal muscle perfusion and its heterogeneity during exercise in moderate hypoxia. Am. J. Physiol. Regul. Integr. Comp. Physiol. 299,R72–R79 (2010). Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 18 73. Lee, I. et al. Isolation of regulatory-competent, phosphorylated cytochrome coxidase. Methods Enzymol. 457, 193–210 (2009). 74. Gielisch, I. & Meierhofer, D. Metabolome and proteome profiling of complex I deficiency induced by rotenone. J. Proteome Res 14, 224–235 (2015). 75. UniProt, C. UniProt: the universal protein knowledgebase in 2023. Nucleic Acids Res 51,D523–D531 (2023). 76. Trelle, M. B. & Jensen, O. N. Utility of immonium ions for assignment of epsilon-N-acetyllysine-containing peptides by tandem mass spectrometry. Anal. Chem. 80,3422–3430 (2008). 77. Pecina, P. et al. Phosphomimetic substitution of cytochrome ctyrosine 48 decreases respiration and binding to cardiolipin and abolishes ability to trigger downstream caspase activation. Biochemistry 49,6705–6714 (2010). 78. Liu, Z. et al. Remarkably high activities of testicular cytochrome cin destroying reactive oxygen species and in triggering apoptosis. Proc.NatlAcad.Sci.USA103, 8965–8970 (2006). 79. Fridovich, I. Quantitative aspects of the production of superoxide anion radical by milk xanthine oxidase. J. Biol. Chem. 245, 4053–4057 (1970). 80. Quick,K.L.,Hardt,J.I.&Dugan,L.L.Rapidmicroplateassayfor superoxide scavenging efficiency. J. Neurosci. Methods 97, 139–144 (2000). 81. Vonrhein, C. et al. Data processing and analysis with the autoPROC toolbox. Acta Crystallogr D. Biol. Crystallogr 67,293–302 (2011). 82. Liebschner, D. et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr D. Struct. Biol. 75,861–877 (2019). 83. Joosten, R. P., Long, F., Murshudov, G. N. & Perrakis, A. The PDB_REDO server for macromolecular structure model optimization. IUCrJ 1,213–220 (2014). 84. Krieger, E. & Vriend, G. New ways to boost molecular dynamics simulations. J. Comput Chem. 36,996–1007 (2015). 85. Molecular Operating Environment (MOE). 2022.02 edn (Chemical Computering Group ULC, 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2023). 86. Pettersen,E.F.etal.UCSFChimera–a visualization system for exploratory research and analysis. J. Comput Chem. 25, 1605–1612 (2004). 87. Autenrieth, F., Tajkhorshid, E., Baudry, J. & Luthey-Schulten, Z. Classical force fieldparametersforthehemeprostheticgroupof cytochrome c.J. Comput Chem. 25,1613–1622 (2004). 88. Case, D.A. et al. Amber 2021: Reference Manual,(Universityof California, San Francisco, California, 2021). 89. Lee, W., Tonelli, M. & Markley, J. L. NMRFAM-SPARKY: enhanced software for biomolecular NMR spectroscopy. Bioinformatics 31, 1325–1327 (2015). 90. Delaglio,F.etal.NMRPipe:amultidimensionalspectralprocessing system based on UNIX pipes. J. Biomol. NMR 6,277–293 (1995). 91. Dosset, P., Hus, J. C., Blackledge, M. & Marion, D. Efficient analysis of macromolecular rotational diffusion from heteronuclear relaxation data. J. Biomol. NMR 16,23–28 (2000). 92. Berlin, K., Longhini, A., Dayie, T. K. & Fushman, D. Deriving quantitative dynamics information for proteins and RNAs using ROTDIF with a graphical user interface. J. Biomol. NMR 57, 333–352 (2013). 93. Lecoeur, H. Nuclear apoptosis detection by flow cytometry: influence of endogenous endonucleases. Exp. Cell Res 277,1–14 (2002). Acknowledgements This work was supported by the National Institutes of Health grants R01 GM116807 (M.H.) and R01 NS120322 (M.H.), a Competitive Graduate Research Assistant Award from the Wayne State University Graduate School (P.T.M), and research grants from the Spanish Ministry of Science and Innovation (PGC2018-096049-B-I00 and PID2021-126663NB-I00; I.D-M.), European Regional Development Fund (FEDER; I.D-M.), Andalusian Government (BIO-198, US-1254317, US-1257019, P18-FR-3487, and P18HO-4091, US/JUNTA/FEDER, UE; I.D-M), University of Seville (VI PPIT) and the Ramón Areces Foundation (I.D-M.). G.P-M. was awarded a Ph.D. fellowship from the Spanish Ministry of Education, Culture, and Sport (FPU17/04604). This research used resources from the Advanced Photon Source, a U.S. Department of Energy Office of Science User Facility operated for the Department of Energy Office of Science by Argonne National Laboratory under Contract DE-AC02- 06CH11357. Use of the Life Sciences Collaborative Access Team (LS-CAT) Sector 21 was supported by the Michigan Economic Development Corp., the Michigan Technology Tri-Corridor (Grant 085P1000817; M.H.), and Wayne State University’sOffice of the Vice-President for Research (M.H.). We thank Dr. Carlos Moraes at the University of Miami for providing the Cytc double knockout cell line, Beata Lukaszewska-McGreal for proteome sample preparation, and Dr. Pavel Afonine for an alternate script for Phenix. refine, Philippe Archambault of the Chemical Computing Group for his technical advice, the support of the Max Planck Society, and the staff from the NMR facility at CITIUS (University of Seville). Author contributions P.T.M., G.P-M., J.W., A.A.T., I.M., H.A.K., A.V., M.P.Z., P.P.H., K.K., T.A., M.A.R., D.D.C., I.L., M.H.M, T.H.S., B.F.P.E., I.D-M., and M.H. designed and performed experiments. D.M. conducted mass spectrometry experiments. A.V., J.S.B., and B.F.P.E. determined the crystal structures. P.T.M., A.A.T., G.P-M., B.F.P.E., I.D-M., and M.H. wrote the manuscript. All authors have read and agreed to the final version of the manuscript. Competing interests The authors declare no competing interests. Additional information Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-023-39820-8. Correspondence and requests for materials should be addressed to Irene Díaz-Moreno or Maik Hüttemann. Peer review information Nature Communications thanks Sharma Shivani, Eva de Alba and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available. Reprints and permissions information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2023 Article https://doi.org/10.1038/s41467-023-39820-8 Nature Communications | (2023) 14:4166 19