Abstract Lysine residues can be post-translationally modified by various acyl modifications in bacteria and eukarya. Here, we showed that two major acyl modifications, acetylation and succinylation, were changed in response to the carbon source in the Gram-positive model bacterium Bacillus subtilis. Acetylation was more common when the cells were grown on glucose, glycerol, or pyruvate, whereas succinylation was upregulated when the cells were grown on citrate, reflecting the metabolic states that preferentially produce acetyl-CoA and succinyl-CoA, respectively. To identify and quantify changes in acetylation and succinylation in response to the carbon source, we performed a stable isotope labeling by amino acids in cell culture (SILAC)-based quantitative proteomic analysis of cells grown on glucose or citrate. We identified 629 acetylated proteins with 1355 unique acetylation sites and 204 succinylated proteins with 327 unique succinylation sites. Acetylation targeted different metabolic pathways under the two growth conditions: branched-chain amino acid biosynthesis and purine metabolism in glucose and the citrate cycle in citrate. Succinylation preferentially targeted the citrate cycle in citrate. Acetylation and succinylation mostly targeted different lysine residues and showed a preference for different residues surrounding the modification sites, suggesting that the two modifications may depend on different factors such as characteristics of acyl-group donors, molecular environment of the lysine substrate, and/or the modifying enzymes. Changes in acetylation and succinylation were observed in proteins involved in central carbon metabolism and in components of the transcription and translation machineries, such as RNA polymerase and the ribosome. Mutations that modulate protein acylation affected B. subtilis growth. A mutation in acetate kinase (ackA) increased the global acetylation level, suggesting that acetyl phosphate-dependent acetylation is common in B. subtilis, just as it is in Escherichia coli. Our results suggest that acyl modifications play a role in the physiological adaptations to changes in carbon nutrient availability of B. subtilis. Introduction Nε-lysine acetylation is an evolutionarily conserved post-translational modification. Until recently, only a few examples of acetylated proteins had been described in bacteria [[35]1,[36]2]. However, recent proteomic studies on diverse bacteria have identified hundreds of acetylated proteins that function in various cellular processes [[37]3–[38]15]. The addition or removal of an acetyl group from a lysine residue can modify enzymatic activity, protein-protein or protein-DNA interactions, protein stability, or protein localization [[39]16–[40]20]. Lysine acetylation is widely thought to be a reversible process that is catalyzed by two types of enzymes, lysine acetyltransferases (KATs) and lysine deacetylases (KDACs) [[41]21–[42]24] However, two distinct mechanisms for acetylation have been proposed. The first mechanism is KAT-dependent; it utilizes acetyl-CoA as the acetyl group donor. The second mechanism is non-enzymatic. Recent studies have shown that acetyl-CoA serves as the acetyl donor in the mitochondria [[43]25], while acetyl phosphate (acetyl-P) serves as the acetyl donor in bacteria [[44]9,[45]12]. Acetylation in Escherichia coli commonly uses acetyl-P as the acetyl donor [[46]9,[47]12]. CobB, the only known KDAC in E. coli, can deacetylate both enzymatic and non-enzymatic lysine acetylation substrates, but not all acetylation substrates, suggesting that some of acetylations cannot be reversed [[48]26]. Recent advances in mass spectrometry have facilitated the discovery of new acyl modifications of lysine residues, including propionylation, butyrylation, succinylation, malonylation, crotonylation, and glutarylation [[49]27–[50]31]. Of these, propionylation, succinylation, and malonylation have been reported in bacteria [[51]28,[52]29,[53]32–[54]34]. It is thought that the corresponding acyl-CoA molecule is utilized in these acyl modifications. In bacteria, acetyl-CoA and succinyl-CoA are abundant and are mainly generated from glycolysis and the citrate cycle, respectively [[55]35]. Lysine succinylation and acetylation are frequent modifications in bacteria and eukarya [[56]36]. Hundreds of succinylated proteins have been reported in E. coli [[57]33,[58]36], and E. coli CobB, catalyzes not only deacetylation, but also desuccinylation [[59]33]. However, an enzyme that catalyzes lysine succinylation has not yet been identified. In Bacillus subtilis, AcuA, the only known KAT, contains a Gcn5-related acetyltransferase (GNAT) motif. AcuA is not a close homologue of YfiQ (also known as Pka or PatZ in E. coli; Pat in Salmonella), the only known KAT in E. coli. B. subtilis has two KDACs, AcuC and SrtN, which are NAD^+-independent and NAD^+-dependent (sirtuin family) deacetylases, respectively. AcuA, AcuC, and SrtN control the enzymatic activity of acetyl-CoA synthetase (AcsA) through the reversible acetylation of a conserved, critical lysine residue, Lys549 [[60]37–[61]39]. A recent acetylome analysis revealed 185 acetylated proteins in B. subtilis [[62]7]. However, a comprehensive analysis of lysine succinylation in B. subtilis has not yet been reported. Increasing evidence indicates that acyl modifications contribute to the control of metabolic enzymes in bacteria and eukarya [[63]2,[64]5,[65]20,[66]21,[67]24,[68]40]. In most central pathways for carbon metabolism in B. subtilis, oscillations in enzyme concentration are insufficient to explain changes in metabolic flux, and allosteric regulation and enzyme modification have emerged as important mechanisms that control metabolic enzymes and fluxes [[69]41]. Acetylation and succinylation are closely linked to central carbon metabolism through glycolysis and the citrate cycle, respectively. Because many metabolic and regulatory networks have been studied extensively in B. subtilis [[70]41–[71]43], we thought that the organism would be suitable for examining the effect of acyl modifications on carbon flux regulation. In this study, we used a quantitative, proteomic approach based on stable isotope labeling by amino acids in cell culture (SILAC) to identify and quantify the changes in lysine acetylation and succinylation in response to the carbon source, with glucose and citrate as glycolytic and citrate cycle substrates, respectively. Our study revealed that, the two acyl modifications changed in response to the carbon source in B. subtilis. Mutations that modulate protein acylation affected B. subtilis growth in a carbon source-dependent manner. The possible role of acyl modifications in the physiological responses and adaptations to changes in carbon nutrients is discussed. Materials and Methods Bacterial strains and culture conditions B. subtilis strain 168 (trpC2), obtained from the Bacillus Genetic Stock Center (BGSC 1A1), was used as the wild type strain in this study. To construct a lysine auxotroph strain for SILAC, the codons for Arg8 (AGA) and Gln9 (CAA) in the lysA gene were replaced with nonsense mutations. To accomplish this, oligonucleotide primers were used to amplify the upstream (lysAmut1-F and lysAmut1-R) and downstream (lysAmut2-F and lysAmut2-R) regions of the lysA gene (see [72]S1 Table for the nucleotide sequences of all primers used in this study). Another PCR was performed using the primers lysAmut1-F and lysAmut2-R and the above two amplified fragments as the DNA template to generate a mutation-containing fragment. We also amplified the trpC2-hisC region by PCR using the primers trpC2hisC-F and trpC2hisC-R and the chromosomal DNA of strain 168 as the template. The resulting two PCR fragments, a lysA mutation (lysA8)-containing fragment and a trpC2-hisC fragment, were used to transform strain RIK1800 (168 hisC101) [[73]44], and histidine-autotroph transformants were selected. Tryptophan-auxotroph transformants with a trpC2 genetic background were further selected, and the resulting strain was designated strain TM61 (168 trpC2 lysA8). To construct deletion mutants of acuA, acuC, srtN, ackA, and pta, the upstream and downstream regions (approximately 0.6~1.0 kbp in length) of each target gene were amplified by PCR using the corresponding del1-F and del1-R primers and del2-F and del2-R primers, respectively (see [74]S1 Table). A 1.2 kb-erythromycin resistance gene cassette was excised from pAE41 [[75]45] through BamHI and SphI digestion and ligated with the upstream and downstream fragments of acuA, resulting in an acuA::erm fragment. A 1.3-kb neomycin resistance gene cassette was excised from pBEST501 [[76]46] through XbaI digestion and ligated with the upstream and downstream fragments of acuC, resulting in an acuC::neo fragment. A 1.4-kb spectinomycin resistance gene cassette was excised from pBEST517A [[77]47] through XbaI digestion and ligated with the upstream and downstream fragments of srtN, resulting in an srtN::spc fragment. A 0.97-kb spectinomycin resistance gene cassette was amplified by PCR using primers ackA_spc-2F and ackA_spc-2R with pBEST517A as a template; the product was connected by splicing by overlap extension (SOE)-PCR with the upstream and downstream fragments of ackA, resulting in an ackA::spc fragment. A 0.89-kb kanamycin resistance gene cassette was amplified by PCR using primers pta_kan-F and pta_kan-R with RIK1420 [[78]48] as a template; the product was connected by SOE-PCR with the upstream and downstream fragments of pta, resulting in a pta::kan fragment. The resulting acuA::erm, acuC::neo, srtN::spc, ackA::spc, and pta::kan fragments were used to transform strain 168. The transformants that harbored a deletion mutation in the target gene as a result of a double-crossover event were selected for resistance to erythromycin (0.5 μg/ml), spectinomycin (50 μg/ml), neomycin (7.5 μg/ml), or kanamycin (5 μg/ml) at 37°C. Proper disruption of each target gene was confirmed by PCR amplification and DNA sequencing. Finally, strain SS38 (acuC::neo srtN::spc), SS51 (ackA::spc), SS52 (pta::kan), SS53 (ackA::spc pta::kan), SS110 (acuA::erm), and SS111 (ackA::spc pta::kan acuA::erm) were obtained, as listed in [79]Table 1. Table 1. Bacillus subtilis strains used in this study. Strain Relevant characteristic Source/reference 168 trpC2 BGSG 1A1 TM61 168 lysA8 This study SS38 168 acuC::neo srtN::spc This study SS51 168 ackA::spc This study SS52 168 pta::kan This study SS53 168 ackA::spc pta::kan This study SS110 168 acuA::erm This study SS111 168 ackA::spc pta::kan acuA::erm This study [80]Open in a new tab For western blotting, wild type strain 168 cells were grown in modified Spizizen minimal medium (0.6 g/l KH[2]PO[4], 1.4 g/l K[2]HPO[4], 0.2 g/l (NH[4])[2]SO[4], 1 mM MgSO[4], 1 mM CaCl[2], 10 μM MnCl[2], and 1 μM FeSO[4]) supplemented with 50 μg/ml tryptophan, 0.01% yeast extract, and a carbon source (30 mM). They were harvested at OD[660] = 0.5–0.7. For SILAC labeling, strain TM61 colonies grown on a minimal medium plate supplemented with 30 mM glucose, a 50 μg/ml amino acid mixture (without lysine), and 50 μg/ml ^13C[6]-lysine (heavy lysine) were inoculated into fresh minimal medium supplemented with glucose and heavy lysine at OD[660] = 0.1. They were grown until the OD[660] reached 0.5 for pre-cultivation. For glucose-heavy labeling (exp. 1), the cultures were grown in fresh minimal medium supplemented with glucose and heavy lysine or in minimal medium supplemented with citrate and light lysine. For citrate-heavy labeling (exp. 2), the cultures were grown in minimal medium supplemented with glucose and light lysine or in minimal medium supplemented with citrate and heavy lysine. The cultivation was started at OD[660] = 0.1, and the cells were harvested at OD[660] = 0.5. Preparation of a pan-anti-succinyllysine antibody A pan-anti-succinyllysine antibody against succinylated bovine serum albumin (BSA) was generated as previously described [[81]49]. BSA was chemically succinylated with succinic anhydride, and the degree of chemical succinylation of primary amines was confirmed by MALDI-TOF MS (76% of the amino groups were succinylated). The rabbit polyclonal anti-succinyllysine antiserum was generated by the Research Resource Center (RRC) of RIKEN-BSI. Specificity of the homemade pan anti-succinyl lysine antibody was confirmed by dot blot analyses ([82]S1 Fig). Preparation of protein lysates and western blot analysis Cells were lysed in NET buffer (50 mM Tris-HCl, pH 7.6, 150 mM NaCl, and 1 mM EDTA) supplemented with 1 mM DTT, 1 mM PMSF, 10 μg/ml DNase, 10 μg/ml RNase, 10 mM sodium butyrate (class-I, II KDAC inhibitor), and 20 mM nicotinamide (class-III or sirtuin KDAC inhibitor) by exposure to high pressure using an EmulsiFlex-B15 (Avestin). After removing the cell debris by centrifugation, the cleared lysates were concentrated with a Vivaspin 20 (Sartorius). The protein concentration was measured with the Quick start protein assay (Bio-Rad). Lysate aliquots containing 20 μg of protein were separated by 10% SDS-PAGE and then transferred to an Immobilon-P membrane (Millipore) using a semidry apparatus. The blot was blocked with 3% (w/v) skim milk in TBST, and then incubated with a mixture of rabbit polyclonal antibodies (1:1000 each in 3% [w/v] milk-TBST) as the primary antibody at 4°C overnight. The blot was then incubated with an horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (1:5000 in 3% milk-TBST; Sigma) for 1 h at room temperature. Antigens on the membranes were detected with an LAS4000 image analyzer (GE Healthcare). Preparation of protein lysates, tryptic digestion, and enrichment of lysine-acetylated and-succinylated peptides Equal amounts of protein (2.5 mg of protein from the glucose and citrate conditions) were mixed, precipitated with acetone, and dissolved in 0.1 M NH[4]HCO[3]. Proteins were reduced with 20 mM DTT at 56°C for 30 min, and subsequently alkylated with 30 mM iodoacetamide at 37°C for 30 min. Samples were incubated overnight at 37°C with sequencing grade trypsin (Promega) at a 1:100 enzyme:substrate ratio (w/w). Proteolytic peptides were concentrated by vacuum centrifugation, and then resuspended in NETN buffer (50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1 mM EDTA, and 0.1% NP-40). A mixture of polyclonal anti-acetyllysine antibodies (Cell Signaling and Rockland) or the anti-succinyllysine antibody (this study) was added at a 1:100 antibody:peptide ratio (w/w) to collect the acetylated or succinylated peptides, respectively. The acetylated and succinylated peptides captured by the respective antibodies were precipitated with protein-G beads (Veritas). The beads were washed three times in NETN buffer and twice in NET buffer. The enriched acetylated peptides were then eluted with 1% trifluoroacetic acid:50% acetonitrile. The acetylated peptide samples were cleaned with ZipTip-scx (Millipore) according to the manufacturer’s instructions, and then subjected to HPLC-MS/MS analysis. To compare the relative abundance of proteins in the glucose and citrate conditions, equal amounts of protein from the two growth conditions were mixed, reduced, alkylated, digested with trypsin, and subjected to MS analysis. The labeling efficiency was calculated as the amount of light peptides among the total peptides in labeled samples without affinity enrichment. Mass spectrometry analysis, peptide identification, and SILAC quantification Samples were analyzed by reverse-phase nano HPLC-MS/MS using an EASY-nLC 1000 HPLC system (Thermo Fisher Scientific) connected to a Q Exactive mass spectrometer (Thermo Fisher Scientific). Peptides were trapped on an Acclaim PepMap pre-column (100 μm × 2 cm; Thermo Fisher Scientific) and then eluted onto a Nano-HPLC capillary C18 column (0.075 × 150 mm; Nikkyo Technos). They were then separated with a 120-min gradient from 0% to 65% solvent B (0.1% formic acid in acetonitrile; solvent A, 0.1% formic acid) at a flow rate of 300 nl/min. Full scan MS spectra (350–1,800 m/z) were acquired in the Orbitrap with a target value of 3.00E+06 at a resolution of 70,000 at 200 m/z. The 10 most intense ions were selected for higher-energy-collisional-dissociation (HCD) fragmentation in the HCD cell with a target value of 1.00E+05 and normalized collision energy of 28%. Mass spectrometric data were processed using Proteome Discoverer (ver. 1.4; Thermo Fisher Scientific). Data were searched against B. subtilis sequences (4,185 entries) in the SwissProt database (2013_11) using with the MASCOT search engine (ver. 2.4.1). The search parameters in MASCOT included trypsin digestion with two missed cleavages allowed. Variable modifications included ^13C[6]-lysine, lysine acetylation, lysine succinylation, and methionine oxidation; cysteine carbamidomethylation was set as a fixed modification. Precursor ion and fragment ion mass tolerances were set to 6 ppm and 20 mmu, respectively. For both identification and quantification, only spectra with expectation values less than 1% of the false discovery rate (FDR) were accepted. Identified peptides were validated using the Percolator algorithm with a q-value threshold of 0.01. The identified peptides with Mascot ion scores below 20 were removed to ensure high quality peptide identification. The event detector and precursor ion quantifier algorithms of Proteome Discoverer were used for quantification with a 2 ppm mass variability and a 0.2 min retention time tolerance on precursor ion pairs. Quantification was based on the ratio of the peak area for each peptide in the glucose and citrate conditions. The peptide ratios were calculated using the same number of isotopes. Protein ratios were calculated using the total trypsinized peptides without affinity enrichment based on a previously reported method [[83]50]. The mass spectrometry proteomics data have been deposited with the ProteomeXchange Consortium ([84]http://proteomecentral.proteomexchange.org) via the PRIDE partner repository [[85]51] (dataset PXD001615). The change in acyl modification was determined from the R-value, which was calculated from the ratio of the peptide peak areas (Heavy/Light [pep]) normalized to the ratio of the protein areas (Heavy/Light [pro]). When multiple peptides with a single lysine modification were detected, the peptide with the highest ion score was chosen. Comprehensive lists of the unique acetylated and succinylated peptides identified are provided with the R-values in [86]S2 Table (for each experiment) and [87]S3 Table (merged). Bioinformatics analysis For protein function annotation, we used the Kyoto Encyclopedia of Genes and Genome (KEGG) pathway database with BRITE functional hierarchies. For functional enrichment analysis, we used the Database for Annotation, Visualization, and Integrated Discovery (DAVID) [[88]52]. A Bonferroni cutoff of 0.05 was used to determine the statistical significance. For motif analysis, the 10 amino acid residues (-10 to +10) on either side of a modification site were selected, and a consensus logo was generated using the iceLogo webserver [[89]53]. We also analyzed the same sets of data using the motif-X webserver [[90]54] to confirm the sequence preferences ([91]S4