Quantification of the immunometabolite protein modifications S-2-succinocysteine and 2,3-dicarboxypropylcysteine

Abstract

The tricarboxylic acid (TCA) cycle metabolite fumarate nonenzymatically reacts with the amino acid cysteine to form S-(2-succino)cysteine (2SC), referred to as protein succination. The immunometabolite itaconate accumulates during lipopolysaccharide (LPS) stimulation of macrophages and microglia. Itaconate nonenzymatically reacts with cysteine residues to generate 2,3-dicarboxypropylcysteine (2,3-DCP), referred to as protein dicarboxypropylation. Since fumarate and itaconate levels dynamically change in activated immune cells, the levels of both 2SC and 2,3-DCP reflect the abundance of these metabolites and their capacity to modify protein thiols. We generated ethyl esters of 2SC and 2,3-DCP from protein hydrolysates and used stable isotope dilution mass spectrometry to determine the abundance of these in LPS-stimulated Highly Aggressively Proliferating Immortalized (HAPI) microglia. To quantify the stoichiometry of the succination and dicarboxypropylation, reduced cysteines were alkylated with iodoacetic acid to form S-carboxymethylcysteine (CMC), which was then esterified. Itaconate-derived 2,3-DCP, but not fumarate-derived 2SC, increased in LPS-treated HAPI microglia. Stoichiometric measurements demonstrated that 2,3-DCP increased from 1.57% to 9.07% of total cysteines upon LPS stimulation. This methodology to simultaneously distinguish and quantify both 2SC and 2,3-DCP will have broad applications in the physiology of metabolic diseases. In addition, we find that available anti-2SC antibodies also detect the structurally similar 2,3-DCP, therefore “succinate moiety” may better describe the antigen recognized.

NEW & NOTEWORTHY Itaconate and fumarate have roles as immunometabolites modulating the macrophage response to inflammation. Both immunometabolites chemically modify protein cysteine residues to modulate the immune response. Itaconate and fumarate levels change dynamically, whereas their stable protein modifications can be quantified by mass spectrometry. This method distinguishes itaconate and fumarate-derived protein modifications and will allow researchers to quantify their contributions in isolated cell types and tissues across a range of metabolic diseases.

INTRODUCTION

The tricarboxylic acid (TCA) cycle metabolite fumarate has been shown to accumulate in several physiological and pathophysiological conditions, including adipose tissue during diabetes, select cancers, mitochondrial disease, and during inflammatory cell responses (1–8). Hyperaccumulation of fumarate drives the chemical modification of protein cysteine residues by succination, yielding the irreversible adduct S-(2-succino) cysteine (2SC) (9). This reaction is favored in an acidic microenvironment where protonated fumarate reacts with cysteine residues to modify a spectrum of mitochondrial and cytosolic proteins (10). Exogenous application of fumarate esters also increases protein succination (11) and is frequently employed experimentally to dramatically augment total cysteine succination (7, 12). The dicarboxylic acid itaconate, produced when aconitate decarboxylase (ACOD1) decarboxylates aconitate is increased during proinflammatory macrophage activation in response to lipopolysaccharide (LPS) (13, 14). Itaconate is a competitive inhibitor of succinate dehydrogenase (SDH) (15), leading to elevated succinate and hypoxia-inducible factor 1 alpha (HIF1α) activation to increase the production of interleukin 1 beta (IL-1β) (16). LPS stimulation of macrophages results in pronounced itaconate increases and the covalent modification of protein cysteine residues to generate the adduct 2,3-dicarboxypropylcysteine (2,3-DCP) (17, 18). Cysteine dicarboxypropylation has been shown to modulate macrophage inflammatory activity by suppressing NLRP3 inflammasome, gasdermin D, and IκBζ-ATF3 inflammatory axis activation (17–20). Therefore, both 2SC and 2,3-DCP are irreversible cysteine adducts that accumulate under metabolic conditions that yield increased fumarate or itaconate, respectively.

The detection of 2SC by immunological methods (anti-2SC antibody) facilitates confirmation of fumarate-derived succination in fumarate hydratase (FH)-deficient cancers and has also allowed the visualization of protein succination profiles in many cell and tissue types (1, 2, 4, 21–25). Despite the utility of the antibody, it permits only semiquantitative 2SC detection, and prior gas chromatography-mass spectrometry (GC-MS/MS) methods required substantial amounts of protein to generate trifluoroacetic acid methyl esters of free 2SC (derivatized after protein hydrolysis) (1, 9). Free 2SC is readily detected in both kidneys and serum of FH-deficient hereditary leiomyoma and renal cell carcinoma (HLRCC) patients (26), however, other metabolic conditions do not have such abundant 2SC accumulation as HLRCC (5, 7). Recently, Liu et al. reported improved liquid chromatography-mass spectrometry (LC-MS/MS)-based 2SC detection following ethyl esterification of carboxylic acids, allowing femtomolar 2SC detection (27). We employed this esterification to facilitate the dual detection of 2SC and 2,3-DCP, and further sought to address the stoichiometry of these covalent modifications by alkylating unmodified cysteines with iodoacetic acid to yield S-carboxymethylcysteine (CMC). This reaction adds an additional carboxylic acid to the cysteine, allowing its esterification and detection in the same LC-MS/MS analysis. The stoichiometric assessment of total succination has been applied only in the proteomic assessment of succinated peptides (10, 28). We employed stable isotope dilution methodology to quantify cysteine succination and cysteine dicarboxypropylation levels in a microglial cell line. Moreover, we demonstrate that immunoblotting with the anti-2SC antibody fails to discriminate when 2,3-DCP is present. Anti-2SC antibodies detect the exposed succinate moiety common to both adducts that differ only by a methylene bridge connecting the thiol and the succinate moiety. Thus, a sensitive and robust mass spectrometry method is needed to differentiate between these similar protein modifications, particularly when both metabolites are elevated.

MATERIALS AND METHODS

Materials

Unless otherwise stated, chemicals were obtained from Sigma Aldrich. HPLC-grade water was purchased from Thermo Fisher Chemical. HPLC-grade ethanol, acetyl chloride, and acetonitrile were obtained from Mallinckrodt Chemicals.¹³C₃,¹⁵N-cysteine and ¹³C₄ fumarate were purchased from Cambridge Isotope Laboratories.

Synthesis of Chemical Standards

Isotopically labeled internal standards for 2SC, 2,3-DCP, and CMC were synthesized by incubating 500 µmol ¹³C₃,¹⁵N-cysteine with 550 µmol fumaric acid, itaconic acid or iodoacetic acid, respectively, for 24 h at pH 4. The yield of ¹³C₃¹⁵N-2SC,¹³C₃,¹⁵N-2,3-DCP, and ¹³C₃,¹⁵N-CMC was confirmed by mass spectrometry analysis. Unlabeled 2SC or 2,3-DCP standards were synthesized by incubating l-cysteine with maleic (cis isomer of fumaric) or itaconic acid, respectively, at a 1:5 molar ratio at pH 5 for 24 h. Standard curves were generated following serial dilution of the 2SC and 2,3-DCP standards in HPLC-grade water. The Limit of quantification (LOQ) and Limit of Detection (LOD) were calculated using the formulae LOQ = 10σ/S and LOD = 3.3σ/S, where σ is the standard error of the response and S is the slope of the standard curve. l-cysteine was dissolved in HPLC-grade water and 25 nmol was used as an l-cysteine standard.

pH-Dependent Cysteine Modification

Cysteine (100 mM) was incubated with 500 mM fumarate, maleate (fumarate isomer), itaconate, or citraconate (constitutional isomer of itaconate) for 24 h at pH 7.00 or 7.40. Triplicate individual incubations of each reaction condition were generated.

Cell Culture

Highly Aggressively Proliferating Immortalized (HAPI) microglia (29) (Millipore Sigma, SCC103) were cultured in 1 g L^-1 glucose DMEM with 100x penicillin/streptomycin and 5% fetal bovine serum (Atlanta Biologicals). HAPI microglia were treated ±100 ng mL^-1 LPS for 18–24 h upon reaching >50% confluency. Untreated HAPI cell lysates were incubated with ±10 mM itaconate after protein collection as a positive control for cysteine dicarboxypropylation.

Protein Extraction, Acid Hydrolysis, and Solid Phase Extraction

Following washing ×3 in ice-cold PBS, HAPI cell lysis was obtained following the addition of RIPA buffer [50 mM Tris Base, 150 mM NaCl, 1 mM diethylenetriamine pentaacetate (DTPA) pH 8.0, 0.1% v/v Triton X-100, 12 mM sodium deoxycholate, 3.47 mM SDS] containing protease inhibitor cocktail. The cell lysates were sonicated for 5 s using a probe sonicator before incubating with 200 mM DTT for 45 min at room temperature, with vortexing every 15 min, to facilitate disulfide bond reduction. Iodoacetic acid (250 mM) was then added for 45 min at room temperature in the dark after vortexing to facilitate cysteine alkylation. The protein (∼500 µg) was precipitated with an equal volume of 20% trichloroacetic acid (TCA) and centrifuged for 10 min at 5000 rpm at 4°C. 10 pmol ¹³C₃,¹⁵N-2SC, 10 pmol ¹³C₃,¹⁵N-2,3-DCP, and 10 pmol ¹³C₃,¹⁵N-CMC were added as internal standards to the resulting protein pellet before overnight hydrolysis in 6 M HCl at 110°C. The HCl was dried in vacuo, before resuspension in 0.1% trifluoroacetic acid (TFA). Solid Phase Extraction (SPE) was performed using 1 mL C18 SPE cartridges (Waters) with 0.1% TFA/40% methanol as the eluent. The samples were dried in vacuo before derivatization. 2SC and CMC have previously been shown to be stable following acid hydrolysis (9, 30, 31), and we find 2,3-DCP to have similar stability. The peak areas of acid hydrolyzed standards are comparable to equivalent amounts of standard that did not undergo hydrolysis. In contrast, up to 10–15% of standards are lost upon solid phase extraction. Since isotopically labeled standards are added to both standards and samples, this loss is accounted for.

Metabolite Extraction and 3-Nitrophenylhydrazine Derivatization

HAPI cells treated ± 100 ng mL⁻¹ LPS for 24 h were rapidly collected in ice-cold 80% aqueous methanol containing 50 nmol mL^{−1 13}C₄ fumarate and ¹³C₅ α-ketoglutarate. Following centrifugation to pellet the protein, the metabolites were dried in vacuo and resuspended in 50 µL methanol/water (60:40 vol/vol). The metabolites were derivatized following the addition of 25 µL each 250 mM 3-nitrophenylhydrazine (3-NPH), 150 mM 1-ethyl-3-(3-(dimethylamino)propyl) carbodiimide hydrochloride (EDC) (ThermoFisher Scientific), and 7.5% pyridine for 2 h at room temperature while rotating (32). The samples were transferred to autosampler vials for HPLC-MS/MS analysis. Standard curves were generated for succinate, itaconate, and fumarate to allow quantification.

Ethyl Ester Derivatization

The 2SC, 2,3-DCP, and CMC standards and extracted amino acids were dried in vacuo before derivatization. Ethanol/acetyl chloride (100:20 vol/vol) was mixed on ice and 25 µL was added to each tube. The tubes were vortexed thoroughly and incubated at 55°C for 2 h before being centrifuged for 10 s and dried completely in vacuo, in an adaptation of the method described by Liu et al. (27). The reaction products were resuspended in 200 µL acetonitrile/water (20:80 vol/vol) and transferred to autosampler vials for LC-MS/MS analysis. In some cases, the procedure was repeated with methanol as a substitute to generate methyl esters.

Quantification of Metabolites by LC-MS/MS

Analysis of ethyl esters.

LC separation was performed on a Thermo Vanquish Flex liquid chromatography system (ThermoFisher Scientific) using a Waters XBridge C18 Reverse phase column (Waters Corp.). The column was 2.1 mm by 100 mm with 3.5 µm particles. The column was maintained at 40°C. Following a 3 µL sample injection, a binary gradient elution was performed. Solvent A consisted of water with 0.1% formic acid and solvent B contained acetonitrile with 0.1% formic acid. The separation began with the column equilibrated at 10% B for 1.5 min then ramped to 95% B at 10 min and held there until 15 min. The system then returned to 10% B for 8 min. The column flow rate was 0.2 µl min⁻¹. Single injections of 2SC, 2,3, DCP, CMC, or cysteine (Cys) standards confirmed their separation and retention times.

Positive ion electrospray mass spectra were acquired on a Thermo Q-Exactive HF-X Quadrupole-Orbitrap performing parallel reaction monitoring (PRM). Precursor ions (150, 322, 326, 336, 340, 236, or 240 Da) were isolated by the quadrupoles using a 0.5 m z⁻¹ isolation window and fragmented in the higher-energy collisional dissociation (HCD) cell using stepped collision energy (CE) (15, 25, 35 eV); the automatic gain control (AGC) target was 1E6 and maximum ion trap (IT) was set at 150 ms. The Orbitrap resolution used for PRM was 30,000. The source capillary temperature was 275°C, other MS source settings were sheath gas flow: 45; auxiliary gas flow: 10; sweep gas flow: 2; spray voltage: 3.5 kV; funnel RF level: 40; auxiliary gas temperature: 400°C. XCalibur 4.2 software (ThermoFisher Scientific) was used to construct extracted ion chromatograms of the transitions 322 → 205 (2SC), 326 → 205 (¹³C₃,¹⁵N-2SC), 336 → 219 (2,3-DCP), 340 → 219 (¹³C₃,¹⁵N-2,3-DCP), 236 → 219 (CMC), and 240 → 222 (¹³C₃,¹⁵N-CMC) and 150 → 76 (Cys). Esterified cysteine (Cys, retention time 1.48 min) was monitored to capture any underivatized cysteine and to confirm the successful conversion of Cys to CMC in alkylated biological samples.

The area of the analyte peaks was normalized to the area of the ¹³C internal standard for the specific analyte to obtain peak area ratios. In some cases, duplicate sample injections were used to confirm peak area ratios, and the mean peak area ratio from these was used. The mass of each analyte in the biological samples was normalized to the total protein concentration as determined by the Lowry method (33). The stoichiometry of 2SC or 2,3-DCP was calculated by dividing the mass of 2SC or 2,3-DCP by the sum of the combined masses of 2SC, 2,3-DCP, CMC, and Cys. The quotient was multiplied by 100 to determine the percentage.

Analysis of 3-NPH adducts.

LC separation was performed on a Thermo Vanquish Flex liquid chromatography system (ThermoFisher Scientific) using the same column and gradient as used above for the ethyl ester analysis method. Electrospray ionization mass spectra were acquired on a Thermo Q-Exactive HF-X Quadrupole-Orbitrap performing PRM. Analyte ions were isolated by the quadrupoles using a 1.0 m z^-1 isolation window and fragmented in the high energy collison-induced dissociation (HCD) cell using stepped collision energy (CE, 20,25,40 eV); the AGC target was 2E5 and maximum IT was set at 100 ms. The orbitrap resolution was set at 30,000. Targeted metabolomics was used to monitor the tricarboxylic acid cycle intermediates, including succinate, itaconate, and fumarate. The spray voltage was 2.5 kV (all other settings same as a positive ion). XCalibur 4.2 software (ThermoFisher Scientific) was used to construct extracted ion chromatograms of the transitions 385 → 234 (fumarate), 399 → 234 (itaconate), 387 → 234 (succinate), 389 → 238 (¹³C₄ fumarate), 554 → 374 (¹³C₅ α-ketoglutarate). Peak area ratios were calculated for each metabolite following normalization to the internal standard. The metabolite mass was normalized to protein content as determined by the Lowry method (33).

Polyacrylamide Gel Electrophoresis and Immunoblotting

For sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), 15–40 μg of cell lysates were incubated with 4X Laemmli loading buffer in a 95°C water bath for 15 min. The lysates were centrifuged briefly and loaded on 12% gels. The gels were electrophoresed at 100 V for ∼55 min in running buffer (25 mM Tris + 192 mM Glycine + 0.1% SDS) and washed in transfer buffer (25 mM Tris + 192 mM Glycine + 20% Methanol) for 15 min at room temperature (RT). For proteomic identification, the gel was stained with Coomassie Brilliant blue, and bands of interest were excised. For immunoblotting, the protein was transferred to a polyvinylidene fluoride (PVDF) membrane under 250 mA for 100 min at 4°C in a transfer buffer. Equal protein loading was verified via Ponceau Red staining for 1 min at RT. Membranes were washed three times for 5 min at RT in wash buffer (pH 7.4, 20 mM Tris + 0.07% Tween-20) and blocked in 5% milk for 1 h at RT or overnight at 4°C. Membranes were incubated in primary and secondary antibodies for 1 h at RT or overnight at 4°C in 2% milk with three 5 min washes in wash buffer after each antibody. Membranes were probed with 1:6,000 Frizzell laboratory anti-2SC antibody (1); 1:2,000 CRB/Biosynth anti-2SC antibody (Biosynth Laboratories, crb2005017d); 1:6,000 anti-IL-1β antibody (Cell Signaling, D3H1Z). The secondary was goat anti-rabbit horseradish peroxidase-labeled secondary antibody (Cell Signaling, 7074S). The membranes were imaged on photographic film after 5 min at RT in an enhanced chemiluminescent (ECL) solution (Thermo Scientific, 80196). Membranes were stripped in stripping buffer (pH 6.8, 62.5 mM Tris + 2% SDS + 0.7% β-mercaptoethanol) at 65°C for 5–20 min. Stripped membranes were washed three times for 5 min in wash buffer before blocking in 5% milk for future immunodetection. Total protein detection on membranes was achieved following Coomassie staining for 5 min at RT, washed 2X for 1 h in 7% acetic acid + 10% methanol, and dried at RT.

Protein Identification by LC-MS/MS

The in-gel protein digestion method used was previously described (6, 11). Briefly, following SDS-PAGE to separate HAPI microglial protein (± 100 ng mL⁻¹ LPS), the gels were stained with Coomassie Brilliant Blue, and the bands of interest in the ∼48–50 kDa region were excised. The gel pieces were destained, washed with 50 mM ammonium bicarbonate in 50% acetonitrile, and dehydrated with 100% acetonitrile. Proteins were reduced with 10 mM dithiothreitol (DTT) and alkylated with 170 mM 4-vinylpyridine. Protein digestion was carried out overnight at 37°C in the presence of 500 ng sequencing grade modified trypsin (Promega) in 50 mM ammonium bicarbonate. After gel extraction with 5% formic acid in 50% acetonitrile, the samples were analyzed in a blinded manner on a Dionex Ultimate 3000-LC system (Thermo Scientific) coupled to a Velos Pro Orbitrap mass spectrometer (Thermo Scientific). The LC solvents were 2% acetonitrile/0.1% formic acid (Solvent A) and 80% acetonitrile/0.1% formic acid (Solvent B); the water used for these solvents was LC-MS grade. Peptides were first trapped on a 2 cm Acclaim PepMap-100 column (Thermo Scientific) with Solvent A at 3 µl min⁻¹. At 4 min, the trap column was placed in line with the analytical column, a 75 μm C18 stationary-phase LC PicoChip Nanospray column (New Objective). The peptides were eluted with a gradient from 98% A:2% B to 40% A:60% B over 30 min, followed by a 5 min ramp to 10% A:90% B that was held for 10 min. The Orbitrap was operated in data-dependent acquisition MS/MS analysis mode and excluded all ions below 200 counts. Following a survey scan (MS1), up to eight precursor ions were selected for MS/MS analysis. All spectra were obtained in the Orbitrap at 7,500 resolution. The DDA data were analyzed using Mascot 2.5.1 with a mass tolerance of 15 ppm for precursor ions and a mass tolerance of 10 ppm for fragment ions. Up to two missed cleavages were allowed following trypsin digestion. The results were filtered with a false discovery rate of 0.01. The variable modifications of methionine oxidation (MOX), cysteine pyridylethylation (CPE, 105.058), cysteine succination by fumarate (C2SC, 116.011), or cysteine dicarboxypropylation by itaconate (CDI, 130.03) were considered.

Statistical Analysis

Statistics were performed using GraphPad Prism 9. All Student’s t tests were performed using two-tailed analysis. One-way ANOVA was performed to detect differences between multiple groups. Graphs indicate mean values with standard deviation (SD).

RESULTS AND DISCUSSION

Derivatization and Resolution of 2SC, 2,3-DCP, and CMC

Increases in both fumarate and itaconate have been described in primary bone marrow-derived macrophages following stimulation by LPS ex vivo (7, 15, 17). In addition, biological tissues have heterogeneous cell populations that may variably contribute to metabolite concentrations in response to a metabolic stressor. We reasoned that the irreversible cysteine modifications 2SC and 2,3-DCP correlate with the reactivity of these metabolites and represent a stable measure of metabolite abundance. Given their structural similarity, they can be quantified in parallel by LC-MS/MS following esterification. In addition, we wished to address the stoichiometry of these modifications following the immediate reduction and alkylation of remaining cysteines to generate S-carboxymethylcysteine (CMC). We used a rat microglial cell line as a model system, these cells demonstrate robust increases in IL-1β following 100 ng mL⁻¹ LPS stimulation. Following protein precipitation with trichloroacetic acid, we hydrolyzed total proteins in 6 M HCl at 110°C for ∼18 h (Fig. 1A) (1). This yields free 2SC, 2,3-DCP, and CMC that are converted to ethyl esters (indicated in red, Fig. 1A) after recovery from solid phase extraction (SPE). Cellular protein extracts were subjected to LC separation and positive ion mode mass spectrometry in parallel with both unlabeled and isotopically labeled standards. Similar to Liu et al. (27), we observed successful esterification of all three carboxyl groups of 2SC (mass 322, Fig. 1B) with 100:20 ethanol:acetyl chloride, and found that increasing the derivatization reaction temperature to 55°C for 2 h yielded products with a high-intensity MS signal. We also detected the esterification of three carboxyl groups for 2,3-DCP (mass 336) and 2 carboxyl groups of CMC (mass 236) as the prevalent reaction products (Fig. 1, C and D). The alternative use of methanol to generate methyl esters was also tested, and while we detected esterification of all carboxyl groups and chromatographic separation of 2SC, 2,3-DCP, and CMC, we observed peak splitting at the resolution used (Supplemental Fig. S1). Since the ethyl esters yielded well-resolved peaks and consistent retention times of ∼6.79 min for 2SC (Fig. 1B), ∼6.90 min for 2,3 DCP (Fig. 1C), and ∼4.90 min for CMC (Fig. 1D), we employed ethanol for all further esterification in the stable isotope dilution studies. Isotopically labeled standards exhibited the same retention times and facilitated the calculation of peak area ratios following parallel reaction monitoring (PRM) of robust fragment ions from each precursor ion, as shown in the representative extracted ion chromatograms and standard curves (Fig. 1, B–D). The standard curves for serially diluted 2SC, 2,3-DCP, and CMC were linear in the 10 fmol to 10 pmol range. The limit of quantification (LOQ) and limit of detection (LOD) for each analyte are specified in Table 1. The LOQ and LOD improved for each analyte following SPE before derivatization, with an LOD of 0.3 pmol for 2SC and 2,3-DCP. SPE significantly improved the detection of all analytes in the subsequent cellular analyses and is recommended for the detection of these analytes in more complicated matrices.

Figure 1.

A: workflow for the quantification of 2SC, 2,3-DCP, and CMC in HAPI microglia. HAPI microglial cells were stimulated ± 100 ng mL⁻¹ lipopolysaccharide (LPS) for 18 h. Following cell lysis, total microglial cell protein containing 2SC or 2,3-DCP are immediately reduced with DTT and alkylated with IA to generate CMC as a measure of total cysteines. Following protein precipitation samples are hydrolyzed in 6 M HCl at 110 °C and dried, followed by SPE of modified amino acids. Esterification is performed in 100:20 ethanol:acetyl chloride at 55°C followed by LC-MS/MS analysis of esterified amino acids. B: structure of esterified 2SC and representative extracted ion chromatogram for 2SC and ¹³C₃,¹⁵N-2SC, retention time 6.77 min. The linearity of the standard curve to 10 pmol is shown. C: structure of esterified 2,3-DCP and representative extracted ion chromatogram for 2,3-DCP and ¹³C₃,¹⁵N-2,3-DCP standards, retention time 6.89 min. The linearity of the standard curve to 10 pmol is shown. D: structure of esterified CMC and representative extracted ion chromatogram for CMC and ¹³C₃,¹⁵N-CMC, retention time 4.90 min. The linearity of the standard curve to 10 pmol is shown. CMC, carboxymethylcysteine; DCP, dicarboxypropylcysteine; DTT, dithiothreitol; HAPI, highly aggressively proliferating immortalized; IA, iodoacetic acid; LC-MS/MS, liquid chromatography-mass spectrometry; SPE, solid phase extraction. Created with BioRender.com.

Table 1.

LOQ and LOD for 2SC, 2,3-DCP, and CMC ± SPE

	No SPE	SPE
2SC
LOQ	1.404 pmol	1.136 pmol
LOD	0.444 pmol	0.359 pmol
2,3-DCP
LOQ	3.884 pmol	1.050 pmol
LOD	1.228 pmol	0.332 pmol
CMC
LOQ	3.653 pmol	1.745 pmol
LOD	1.155 pmol	0.552 pmol

CMC, carboxymethylcysteine; DCP, dicarboxypropylcysteine; LOD, limit of detection; LOQ, limit of quantification; SPE, solid phase extraction; 2SC, 2-succinocysteine.

Acidic pH Increases the Yield of 2SC and 2,3-DCP

Protonated fumarate reacts with cysteines to yield protein succination, emphasizing the role of an acidic microenvironment (10, 34). We examined the impact of a physiological pH shift on the formation of both 2SC and 2,3 DCP. In vitro, cysteine incubations at pH 7.4 versus 7.0 were compared with reflect physiological pH fluctuations that might be observed during an acute metabolic crisis. We contrasted the reactivity of fumarate with its cis isomer maleate and itaconate with its constitutional isomer citraconate (Supplemental Fig. S2A). Compared to control l-cysteine incubations, incubation with fumarate at pH 7.4 yielded a 19-fold increase in 2SC, whereas pH 7.0 contributed to a 104-fold increase in 2SC (P < 0.0001, Supplemental Fig. S2B). As expected, the more reactive cis isomer maleate yielded a 178-fold increase in 2SC at pH 7.4 and a 208-fold increase at pH 7.0. (P < 0.0001, Supplemental Fig. S2B). l-cysteine incubation with itaconate at pH 7.4 and 7.0 yielded 126- and 1518-fold increases in 2,3-DCP, respectively (P < 0.0001, Supplemental Fig. S2C). Incubation with citraconate at pH 7.4 yielded only a 1.7-fold increase in 2,3-DCP, and pH 7.0 yielded a 11-fold increase, relative to the l-cysteine control (Supplemental Fig. S2C). These results confirm that lower pH enhances the formation of these covalent cysteine adducts. As expected, the cis isomer maleate formed more 2SC than fumarate due to the increased electrophilicity of the cis bond; however, citraconate had reduced reactivity compared to itaconate. Chen et al. (35), previously reported an eightfold increase in citraconate-modified glutathione versus itaconate-modified glutathione in incubations performed at pH 6.5, as detected by LC-MS/MS. Therefore, a lower pH than 7.00 may be required to facilitate enhanced citraconate reactivity. Overall, our data confirm previous reports that acidity enhances fumarate reactivity and that pathophysiological pH alterations can drive enhanced thiol modification.

Exogenous Maleate Treatment Increases Intracellular Succination

Maleate has been proposed as a nontoxic alternative to dimethyl fumarate as an exogenous treatment to increase intracellular succination (34). HAPI microglial cells were treated ± 5 mM maleate for 24 h and total levels of 2SC were quantified using the stable isotope dilution method detailed above. A significant 23-fold increase in 2SC was detected with 444 pmol mg⁻¹ protein in maleate-treated cells versus 19.4 pmol mg^-1 protein in control untreated cells (P < 0.0001, Fig. 2A). No 2,3-DCP was detected by LC-MS/MS in the maleate-treated cells. We employed the detection of CMC as a readout of total reduced thiol content in cells and used this measurement to estimate the stoichiometry of 2SC in the maleate-treated cells. Detection of free cysteine (Cys) before conversion to CMC was considered to represent a large but incomplete fraction of Cys in biological samples subjected to hydrolysis, as we sometimes detected esterified cystine or esterified oxidized cysteine products. Therefore, derivatization to CMC was considered a rapid and stable way to capture reduced cysteine thiols. Since CMC can be endogenously generated in cells when reactive carbonyl species such as glyoxal or glycolaldehyde are present (30, 36), we first measured total CMC levels in protein from control HAPI cells where no reduction or alkylation with iodoacetic acid was performed. We detected endogenous CMC levels of 0.052 nmol mg^-1 protein, and this increased to 1.15 nmol mg⁻¹ protein following the synthesis of CMC in untreated control HAPI cells (P < 0.001, Supplemental Fig. S3A). Considering all cysteine-derived metabolites detected in our PRM assay (2SC, 2,3-DCP, CMC, Cys), the endogenous CMC accounted for 4.7% of total cysteine, which increased to 87.6% of cysteines following reduction and alkylation to synthesize CMC (P < 0.0001, Supplemental Fig. S3B). Supporting this, we detected 8.06% free Cys in the alkylated samples, versus 90.6% free Cys in the unalkylated samples (P < 0.0001, Supplemental Fig. S3C). Although we account for 2SC, 2,3-DCP, CMC, and reduced Cys in our analyses of total cysteine, we recognize that other cysteine modifications such as lipidation, nitration, and oxidative modifications exist and are not captured in this stoichiometric assessment. However, the use of alkylation to capture total free cysteine is routinely employed for quantitative purposes in proteomics, e.g., the use of chemical reactivity probes such as iodoacetamide-alkyne (IA) in isotopic Tandem Orthogonal Proteolysis - Activity-Based Protein Profiling (isoTOPP ABBP) (10, 37). Applying the detection of derivatized CMC to the control and maleate-treated cells, we find that 1.6% of cysteines are succinated in control conditions, increasing to 28.4% in maleate-treated cells (P < 0.0001, Fig. 2B). In parallel, we used our anti-2SC antibody (1) to detect the profile of succinated proteins in the maleate-treated HAPI microglia. A wide spectrum of proteins is intensely succinated in the maleate-treated cells versus controls (Fig. 2C), confirming our quantitative analyses.

Figure 2.

A: HAPI microglia were treated ±5 mM MAL for 24 h to promote protein succination. 2SC levels were detected in total protein hydrolysates and B: quantified as a percentage of total cysteine following alkylation of free thiols to CMC. C: anti-2SC antibody immunoblotting of protein lysates from HAPI cells ±5 mM MAL was used to confirm protein succination in parallel with the LC-MS/MS. Quantification of 2SC (D) and 2,3-DCP (F) in HAPI microglia treated ±100 ng mL⁻¹ LPS for 18 h. 2SC (E) and 2,3-DCP (G) levels were expressed as a percentage of total detected cysteine content in cells following alkylation of cysteines to CMC. H: detection of anti-2SC or the succinate moiety in HAPI cells ±100 ng mL⁻¹ LPS indicates increased modification of select proteins. N = 6/group. **P < 0.01, ****P < 0.0001. sCMC, S-carboxymethylcysteine; HAPI, highly aggressively proliferating immortalized; LPS, lipopolysaccharide; MAL, maleate. Created with BioRender.com.

2,3-Dicarboxypropylcysteine Increases in LPS-Stimulated HAPI Microglia

LPS stimulation of HAPI cells induces several phenotypic changes indicative of pro-inflammatory activation, including nitric oxide generation, cytokine production, and release (29, 38, 39). LPS challenge induces ACOD1 to increase the production of itaconate from cis-aconitate. We confirmed increased production of itaconate in LPS-stimulated HAPI microglia, detecting 89.8 nmol mg⁻¹ protein, a ∼15-fold increase over the variable trace levels detected in unstimulated control cells (Supplemental Fig. S4A). Fumarate levels tended to decrease after 24 h LPS-stimulation (Supplemental Fig. S4B), whereas succinate levels increased significantly from 9.70 nmol mg⁻¹ protein in control cells to 38.4 nmol mg⁻¹ protein in LPS-stimulated cells (P < 0.05, Supplemental Fig. S4C). The expected increase in succinate in parallel with itaconate is due to inhibition of succinate dehydrogenase (15). Quantification of the 2SC content in proteins from control cells following acid-hydrolysis demonstrated levels around 25 pmol mg⁻¹ protein, with no significant change in LPS-stimulated cells (Fig. 2D). Stoichiometric quantification demonstrated that 2SC accounted for ∼1.5% of total cysteines detected, with a small but significant decrease to ∼1% total cysteines in LPS-stimulated cells (P < 0.01, Fig. 2E). The quantification of 2SC mirrors the decrease in fumarate detected in LPS-stimulated HAPI cells versus control (Supplemental Fig. S4B), confirming that the succination is not increased in HAPI microglia in response to LPS.

We next quantified the covalent dicarboxypropylation of protein cysteine thiols by itaconate to yield 2,3-dicarboxypropylcysteine (2,3-DCP). Total levels of 2,3-DCP were significantly increased 10-fold, with 27.5 pmol mg⁻¹ in control cells versus 292 pmol mg⁻¹ in LPS-stimulated HAPI microglia (P < 0.01, Fig. 2F). Stoichiometric quantification 2,3-DCP demonstrated an increase from 1.57% in control cells to 9.07% in LPS-stimulated cells (P < 0.001, Fig. 2G). Therefore, although both 2SC and 2,3-DCP account for ∼1–2% of total cysteine content detected in the control cells, only 2,3-DCP is significantly increased upon LPS stimulation. Our quantitative data support observations in bone marrow-derived macrophages where dicarboxypropylation of proteins regulating the immune response have been described (17–19). Despite the lack of fumarate-mediated succination in LPS-activated HAPI cells, it is evident that itaconate-derived dicarboxypropylation drives immunometabolic responses in these microglia. The LC-MS/MS-based quantification for distinguishing 2SC and 2,3-DCP will also be relevant for longitudinal immunometabolism studies in primary macrophages where dynamic fluctuations in both fumarate and itaconate occur (7, 17).

We also examined the succination profile in the control and LPS-stimulated cells and were surprised to note increases in select bands given that neither fumarate nor 2SC increased upon LPS stimulation (Fig. 2H). A lighter exposure focusing on the pronounced band at ∼50 kDa showed that this reflects increases in two discrete bands (Supplemental Fig. S5A, 2SC/Succinate moiety). The increase in the detected modification occurred in parallel with the LPS activation of the cells, as indicated by increased uncleaved IL-1β levels (Supplemental Fig. S5A, IL-1β panel). Since only 2,3-DCP increased following LPS stimulation, we hypothesized that the structurally similar succinate moieties present on both 2SC and 2,3-DCP might be detected by anti-2SC antibodies. To test this, we generated a strong positive control for itaconate-modified proteins by incubating control HAPI microglia protein extracts with exogenous itaconate, yielding high levels of 2,3-DCP. The “anti-2SC” antibody detects these dicarboxypropylated positive control samples (CON + ITA lanes, Supplemental Fig. S5B), therefore we now propose that “succinate moiety” better describes the antigen that the current anti-2SC antibody detects, since this is common to both 2SC and 2,3-DCP.

In addition to the antibody that we have previously generated (1), a commercially available “anti-2SC” antibody is available from Biosynth Laboratories (formerly Cambridge Research Biochemicals). We compared the itaconate-modified positive control (CON +ITA) to the control microglia, as well as LPS-stimulated positive controls. Both antibodies detected similar itaconate-modified proteins at a low exposure (Supplemental Fig. S5C), confirming that the current commercial “anti-2SC” antibody also detects dicarboxypropylated proteins when itaconate is elevated, likely by also recognizing the succinate moiety. The recent identification of fumarate-mediated succination in LPS-activated macrophages was confirmed in part by the anti-2SC antibody (7). However, it is equally likely that the antibody was detecting the itaconate adduct, especially since the increase in fumarate was comparatively low versus the increase of itaconate in these cells. Importantly, independent measures of increased free 2SC were also used to confirm distinct succination in this work (7). We expect the current methodology to have distinct utility in cells such as bone marrow-derived macrophages where both metabolites have been shown to increase in parallel. In addition, anti-2SC immunodetection is predominantly used to differentiate fumarate hydratase deficient cells in cases of hereditary leiomyoma and renal carcinoma (25, 40, 41). Given that elevated fumarate and 2SC accumulation is a marked feature of these cells, it is likely that 2SC is the primary antigen. However, it may be useful for clinical pathologists to note any disperse 2SC-positive cells that are not part of the tumor, particularly if immune cells are present that may contain 2,3-DCP.

We sought to identify the proteins that showed more intense modification around the 48–50 kDa region and performed proteomic analyses on bands excised from parallel Coomassie gels. Following in-gel trypsin digestion, we identified cysteine dicarboxypropylation on β-tubulin isoforms (predominantly TBB5 and TBB4B) (Supplemental Fig. S5D). The rat β-tubulin identified is modified by itaconate on either one of Cys¹²⁹β or Cys¹³¹β (as the specific site of dicarboxypropylation could not be distinguished). We have previously detected succination by fumarate esters on this same peptide of porcine β-tubulin (42). Both Cys¹²⁹β and Cys¹³¹β are known to be solvent-accessible surface cysteines but were previously found to be unreactive with iodoacetamide as their electrostatic environment is inhibitory due to the presence of carboxylates (Cys¹²⁹β is adjacent to Glu³β and Cys¹³¹β is positioned near Asp¹³⁰β, Glu⁹⁷α and Asp⁹⁸α) (43). These tubulin cysteines are not typically considered reactive, however, given that an acidic microenvironment facilitates protonation and reactivity of itaconate, these carboxylate neighbors greatly increase the likelihood of dicarboxypropylation at Cys¹²⁹β and Cys¹³¹β. Future studies will be required to determine the functional importance of this dicarboxypropylation site in activated microglia. In previous work, we found that modestly succinated tubulin was still capable of polymerization, but that protein–protein interactions between microtubules and other proteins might be impacted. Dicarboxypropylation of Cys¹²⁹β or Cys¹³¹β might also impact interactions with other proteins as Cys¹³¹β is proposed to be thiol-disulfide exchange site for Tau, a protein that stabilizes microtubules. Modification at this site may prevent this interaction with Tau, destabilizing the microtubule. LPS-activated microglia undergo microtubule reorganization to form organized radial arrays that allow cytokine transport and release (44, 45), and dynamic microtubule reorganization is critical for this process. It is interesting to speculate that itaconate-driven tubulin dicarboxypropylation is a contributor to this process in microglia.

Conclusion

The expanding roles of itaconate and fumarate as immunomodulatory metabolites demand a robust method to discern these immunometabolite-derived cysteine modifications. We demonstrate a sensitive LC-MS/MS method for the dual detection of the fumarate and itaconate-derived protein modifications 2SC and 2,3-DCP, respectively. Importantly, this method allows the quantification of both modifications alongside an estimation of their cysteine stoichiometry based on the detection of reduced thiols following conversion to CMC. This method will allow researchers to determine the effects of itaconate and fumarate in distinct cell types isolated from tissues across a range of diseases where these metabolites have been implicated. We propose the “succinate moiety” nomenclature for previously described anti-2SC antibodies that also detect the structurally similar antigen 2,3-DCP.

SUPPLEMENTAL DATA

All Supplemental Figures are available at https://doi.org/10.6084/m9.figshare.25137482.

GRANTS

This work was supported by the National Institutes of Health R01NS126851, R56NS116174, R01NS092938 (to N.F.) and R21058586 (to M.S. and N.F.), the National Science Foundation 1828059 (to N.F. and M.D.W.) and an ASPIRE II grant from the University of South Carolina (to N.F. and M.S.). ChemDraw Professional 22.2.0 was used to generate chemical structures.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.H.C., R.S.M., and N.F. conceived and designed research; J.H.C., R.S.M., E.T., S.S., H.H.S., M.D.W., and W.E.C. performed experiments; J.H.C., R.S.M., E.T., S.S., H.H.S., M.D.W., W.E.C., and N.F. analyzed data; J.H.C., R.S.M., E.T., S.S., H.H.S., M.D.W., W.E.C., and N.F. interpreted results of experiments; J.H.C., R.S.M., E.T., S.S., H.H.S., M.D.W., W.E.C., and N.F. prepared figures; J.H.C., R.S.M., and N.F. drafted manuscript; J.H.C., R.S.M., G.G.P., M.S., and N.F. edited and revised manuscript; J.H.C., R.S.M., E.T., S.S., H.H.S., G.G.P., M.S., M.D.W., W.E.C., and N.F. approved final version of manuscript.