Cysteinyl-tRNA synthetase governs cysteine polysulfidation and mitochondrial bioenergetics

Abstract

Cysteine hydropersulfide (CysSSH) occurs in abundant quantities in various organisms, yet little is known about its biosynthesis and physiological functions. Extensive persulfide formation is apparent in cysteine-containing proteins in Escherichia coli and mammalian cells and is believed to result from post-translational processes involving hydrogen sulfide-related chemistry. Here we demonstrate effective CysSSH synthesis from the substrate L-cysteine, a reaction catalyzed by prokaryotic and mammalian cysteinyl-tRNA synthetases (CARSs). Targeted disruption of the genes encoding mitochondrial CARSs in mice and human cells shows that CARSs have a crucial role in endogenous CysSSH production and suggests that these enzymes serve as the principal cysteine persulfide synthases in vivo. CARSs also catalyze co-translational cysteine polysulfidation and are involved in the regulation of mitochondrial biogenesis and bioenergetics. Investigating CARS-dependent persulfide production may thus clarify aberrant redox signaling in physiological and pathophysiological conditions, and suggest therapeutic targets based on oxidative stress and mitochondrial dysfunction.

Fig. 1

Formation of cysteine persulfide (CysSSH) and CysS–(S)_n–H in proteins and their biosynthesis by EcCARS. a Quantitative identification by LC-MS/MS analysis of CysS–(S)_n–H formed in recombinant ADH5 after pronase digestion of the HPE-IAM-labeled protein. b Formation of cysteine (CysSH) and CysS–(S)_n–H on tRNA (Cys-tRNA^{CysS–(S)n–H}) as identified by HPE-IAM labeling LC-MS/MS analysis, which determined the amounts of CysSH and CysS–(S)_n–H released from Cys-tRNA^Cys and Cys-tRNA^{CysS–(S)n–H} synthesized in the EcCARS enzymatic reaction after their heat or alkaline treatment. The method employed is illustrated in the upper panel. c GAPDH cysteine polysulfides are formed and incorporated into nascent polypeptides synthesized de novo in ribosomes, as identified by PUNCH-PsP (Supplementary Fig. 10; cf. Supplementary Fig. 6). d CysS–(S)_n–H formation from cysteine, catalyzed by EcCARS, as dependent on enzyme and substrate (cysteine) concentrations and reaction time (lower panel). Schematic representation of the EcCARS-catalyzed reaction (upper panel). HPE-AM, β-(4-hydroxyphenyl)ethyl acetamide; HPE-IAM, β-(4-hydroxyphenyl)ethyl iodoacetamide. Data a, b are means ± s.d. (n = 3)

Fig. 2

CysS–(S)_n–H biosynthesis catalyzed by EcCARS and its various mutant EcCARSs. a CysS–(S)_n–H (CysSSH and CysSSSH) biosynthesis from cysteine catalyzed by EcCARS as a function of reaction time and the presence or absence of PLP. CysS–(S)_n–H production was analyzed by using the HPE-IAM labeling with LC-MS/MS analysis for the reaction of recombinant EcCARS (200 μg/ml) with 100 μM cysteine in the presence or absence of 50 μM PLP. The data are means ± s.d. (n = 3). *P < 0.05. b General structure (upper panel) and conserved amino acid alignments (lower panel) of bacterial, human, and rodent CARSs. c, e Enzyme activities of EcCARS lysine (K) mutants c and cysteine (C) mutants e to form CysSSH. WT and EcCARS K and C mutants, 200 μg/ml each, reacted with 25 μM cysteine at 37 °C for 30 min. Data represent means ± s.d. (n = 3). ***P < 0.001. The enzyme activity of EcCARS Lys (d) and Cys (f) mutants was assessed by the PUREfrex assay with the cell-free translational reactions for ALDH1A1 (55 kDa), ADH5 (40 kDa), GAPDH (36 kDa), and ETHE1 (28 kDa), with protein syntheses being identified by western blotting

Fig. 3

Computational modeling of EcCARS structure, and CysS–(S)_n–H biosynthesis by CARS1/2. a A molecular docking model of PLP-bound EcCARS generated by SwissDock using the crystal structure of EcCARS (PDB ID: 1LI5). Cysteinyl-tRNA is placed by superimposing the crystal structure of the EcCARS-Cysteinyl-tRNA binary complex (PDB ID: 1U0B) to the docking model. b, c PLP-dependent CysSSH and CysSSSH biosynthesis by mouse CARS1 and human CARS2. CysSSH and CysSSSH production was quantified by means of HPE-IAM labeling LC-MS/MS analysis in the reaction of recombinant mouse CARS1 and human CARS2 (200 μg/ml each) with 25 μM L-cysteine in the presence or absence of 100 μM PLP (37 °C, 2 h). The data are means ± s.d. (n = 3). *P < 0.01. d Concentration-dependent effects of PLP on CysSSH and CysSSSH production by recombinant human CARS2. Human CARS2 (200 μg/ml) reacted with 25 μM cysteine in the presence of 0, 10, 50, or 100 μM PLP at 37 °C for 30–120 min. No appreciable cysteine persulfide production was detected in the reaction mixture of cysteine and PLP alone as long as no >100 μM PLP was used. e Precisely quantitative analysis for PLP bound to human CRAS2. Human CARS2 treated with various concentrations of PLP (d) at 37 °C for 1 h was reacted with DNPH to form PLP-DNPH adduct, followed by quantification by LC-ESI-MS/MS analysis

Fig. 4

Endogenous formation of persulfides in HEK293T cells. Intracellular levels of CysSSH (a) and GSSH (b) in WT and CARS2 KO cells with CARS1 or CARS2 knocked down. Data are means ± s.d. (n = 3). *P < 0.05; **P < 0.01; N.S., not significant. c CARS1 and CARS2 Western blotting for cells used in a and b. Lane 1 and 2, duplicate determinations with each siRNA. The right panel shows the densitometric analysis for the western blot shown in the right panel. The data are means ± s.d. (n = 3). ***P < 0.001. Production of CysSSH (d) and GSSH (e) in CARS2 KO cells with WT or CARS2 C and K mutants added back. The data are means ± s.d. (n = 3). **P < 0.01; N.S., not significant vs. CARS2 KO mock. f CARS2 western blotting for WT and CARS2 KO cells with WT or CARS2 C and K mutants added back. g Western blotting for the cells in d and e with different mitochondrial proteins: MTCO1, mitochondrial cytochrome c oxidase subunit 1 (encoded by mitochondrial DNA) and SDHA, succinate dehydrogenase complex flavoprotein subunit A (encoded by genomic DNA). Supplementary Fig. 16 provides full blot images. The lower panel shows the densitometric analysis for the western blot. The data are means ± s.d. (n = 3). ***P < 0.001

Fig. 5

Generation of Cars2-deficient mice via the CRISPR/CAS9 system. a Schematic illustration of the mouse Cars2 gene structure and sequences of WT and mutant alleles around the target locus. Green and black letters indicate the first exon and intron of Cars2, respectively. The targeted locus of gRNA and protospacer-adjacent motif (PAM) sequence were indicated in the WT sequence are indicated by underlined and bold letters, respectively. A modified allele sequence obtained from the Cars2-edited mouse (line 1) is shown below. b Detection of mutations introduced by gRNA-Cas9 targeting Cars2 via PCR with genomic DNA from WT and Cars2 ^+/− mice. Cars2 ^+/−, Cars2 heterozygous KO mice, M: DNA molecular weight marker. c Western blotting of CARS2 and mitochondrial proteins, e.g., MTCO1 and SDHA, from mitochondria isolated from the liver. The lower panel shows the densitometric analysis of the western blot. Data are means ± s.d. (n = 3). ***P < 0.001. d CysSSH production in mitochondria isolated from the liver of WT and Cars2 ^+/− littermate mice. Various concentrations of isolated mitochondria were reacted with HPE-IAM for 1 h, followed by LC-MS/MS analysis (see Supplementary Methods for details). Mitochondria were obtained from line 2 Cars2 ^+/– mice (Supplementary Figs. 20 and 21). *P < 0.05, WT vs. Cars2 ^+/– mice (two-way ANOVA). e Western blotting of CARS1, CSE, CBS, and 3-MST with liver tissue obtained from WT and Cars2 ^+/– mice. Supplementary Fig. 19 provides full blot images. The right panels show the densitometric analysis of the CARS1 and CARS2 immunoblots. Data are means ± s.d. (n = 3). ***P < 0.001

Fig. 6

In vivo formation of various sulfide species in WT and Cars2 ^+/− mice. Endogenous production of CysSSH and other related polysulfide compounds was identified by means of HPE-IAM labeling LC-MS/MS analysis in the liver a and lung b obtained from WT and Cars2 ^+/− littermates (21-week-old males). The data are means ± s.d. (n = 3). *P < 0.05; **P < 0.01

Fig. 7

Endogenous protein polysulfidation in vivo and in HEK293T cells. The amounts of CysSSH formed in whole cell protein recovered from the mouse livers of WT and Cars2 ^+/− (line 1) 21-week-old male littermates (a) and from WT and CARS2 KO HEK293T cells (b) were quantified by using HPE-IAM labeling LC-MS/MS analysis. Data are means ± s.d. (n = 3). *P < 0.05; **P < 0.01. c Schematic drawing of the mechanism of the extramitochondrial release of CysSSH into the cytosol, which may regulate whole cell protein polysulfidation

Fig. 8

CARS2-dependent mitochondrial morphogenesis and bioenergetics. a Mitochondrial morphological analyses with MitoTracker Red fluorescent mitochondrial stain: morphometric analysis of mitochondrial length in HEK293T cells (WT and CARS2 KO; CARS2 WT and mutants added back). AU, arbitrary unit. The data are means ± s.d. (n = 3). **P < 0.01. b Transmission electron microscope (TEM) images of the cells in a. Scale bars, 1 μm. c Identification of Drp1 activity in HEK293T cells (WT and CARS2 KO; CARS2 WT and mutants added back). The GTP-agarose pulldown assay was performed. The data are means ± s.d. (n = 3). **P < 0.01. d Drp1 expressed in extensively polysulfidated (biotin-PEG-MAL capture method) HEK293T cells. Drp1 was markedly suppressed and nullified by CARS2 KO and CARS1/2 knockdown. Lanes 1 and 2 show duplicate determinations with each siRNA. Supplementary Fig. 23 provides full blot images. e A schematic drawing of Drp1 activity as regulated by protein polysulfidation and depolysulfidation, as affected by polysulfides vs. electrophiles and Trx/TrxR system. f Changes in membrane potential as assessed by using JC-1 staining of HEK293T cells (WT and CARS2 KO; CARS2 WT and mutants added back). The data are means ± s.d. (n = 3). *P < 0.05; **P < 0.01. g Assessment of mitochondrial electron flow in HEK293T CARS2 KO cells with or without adding back WT and C78/257D, K124/127A, and K317/320A mutants, as analyzed by measuring OCR using an extracellular flux analyzer. Time dependence of oxygen consumption before/after inhibition of mitochondrial respiration at complexes I and III by rotenone/antimycin A, and its statistical summary; the data are means ± s.d. (n = 3). ***P < 0.001

Fig. 9

Mitochondrial ETC-mediated reduction of CysSSH. a, b Sulfide metabolite profiling for the reaction of the recombinant human CARS2 in vitro (a) and of CARS2 expressed in HEK293T cells (b). c–h Changes in amounts of CysSSH (ΔCysSSH) and HS⁻ (ΔHS⁻) induced by complex III inhibition by antimycin A c–e or by mitochondrial DNA (mtDNA) elimination induced by ethidium bromide (f–h) in WT and CARS2 KO HEK 293 T cells. The values of CysSSH and HS⁻ shown in b–h indicate the quantity of each compound produced in the cells in a manner dependent on CARS2 expression, which was determined by subtracting each amount in CARS2 KO HEK293T cells from that in the WT cells, after quantification of each metabolite via HPE-IAM labeling LC-MS/MS analysis. e, h Stoichiometric alterations (conversion) between CysSSH and HS⁻ in cells by the ETC inhibition. i Schematic diagram of ETC-mediated CysSSH reduction to form HS⁻ and possible further conversion to S₂O₃ ²⁻. The data are means ± s.d. (n = 3). *P < 0.05; **P < 0.01; N.S., not significant

Fig. 10

CARS-mediated protein polysulfidation and mitochondrial functions. a The physiological relevance of co-translational protein polysulfidation that is reversibly regulated by various post-translational modifications, including depolysulfidation. b A CysS–(S)_n–H regulation mechanism for mitochondrial functions with regard to mitochondrial biogenesis and bioenergetics. CysSSH is reductively metabolized to CysSH and HS⁻, which may be oxidized by sulfide:quinone reductase (SQR) and other enzymes, e.g., sulfur dioxygenase (SD) and sulfur transferase (ST), in a manner linked to ETC in mitochondria. The CysS–(S)_n–H-dependent HS⁻ metabolism may be coupled with formation of the iron-sulfur clusters, as being controlled by the mitochondrial ETC. I, II, III, and IV: complexes I, II, III, and IV; TCA tricarboxylic acid (Krebs) cycle