Protein Acetyltransferases Mediate Bacterial Adaptation to a Diverse Environment

ABSTRACT

Protein lysine acetylation is a conserved posttranslational modification that modulates several cellular processes. Protein acetylation and its physiological implications in eukaryotes are well understood; however, its role in bacteria is emerging. Lysine acetylation in bacteria is fine-tuned by the concerted action of lysine acetyltransferases (KATs), protein deacetylases (KDACs), and metabolic intermediates, e.g., acetyl coenzyme A (Ac-CoA) and acetyl phosphate (AcP). AcP-mediated nonenzymatic acetylation is predominant in bacteria due to its high acetyl transfer potential, whereas enzymatic acetylation by bacterial KATs (bKATs) is considered less abundant. SePat, the first bKAT discovered in Salmonella enterica, regulates the activity of the central metabolic enzyme acetyl-CoA synthetase, through its acetylation. Recent studies have highlighted the role of bKATs in stress responses like pH tolerance, nutrient stress, persister cell formation, antibiotic resistance, and pathogenesis. Bacterial genomes encode many putative bKATs of unknown biological function and significance. Detailed characterization of putative and partially characterized bKATs is important to decipher acetylation-mediated regulation in bacteria. Proper synthesis of information about the diverse roles of bKATs is missing to date, which can lead to the discovery of new antimicrobial targets in future. In this review, we provide an overview of the diverse physiological roles of known bKATs and their mode of regulation in different bacteria. We also highlight existing gaps in the literature and present questions that may help clarify the regulatory mechanisms mediated by bKATs in adaptation to a diverse habitat.

OVERVIEW OF PROTEIN ACETYLATION IN BACTERIA

Posttranslational modifications (PTMs) such as phosphorylation, acetylation, and methylation expand the functional repertoire of the proteome (1–9). Protein acetylation is a common regulatory mechanism employed by all living organisms to adapt to diverse environmental conditions. Acetylation of proteins can occur via transfer of the acetyl group to the ε amino group (Nε-acetylation) of an internal lysine residue or the N-terminal amino group (Nα-acetylation) (10). Less frequently, the acetyl group can attach to the hydroxyl (OH) group of the serine/threonine residue (O-acetylation) (11). Acetylation of the Nε group of lysine neutralizes its positive charge and may impact protein stability, protein-protein interaction, and protein-nucleic acid interaction. Lysine acetylation in all living organisms is facilitated by two means: (i) enzyme-mediated acetylation by N-acetyltransferases and (ii) nonenzymatic acetylation by high-energy intermediates acetyl phosphate (AcP) or acetyl coenzyme A (Ac-CoA). In some cases, acetylation by both mechanisms can be reversed by lysine deacetylases (KDACs) (12) (Fig. 1). Eukaryotes mostly employ N-acetyltransferases to acetylate their proteins, except in mitochondria, where widespread Ac-CoA-mediated nonenzymatic acetylation is reported (13). AcP-mediated nonenzymatic acetylation is prevalent and well studied only in bacteria (14, 15). While the physiological significance of Nε-lysine acetylation is widely recognized in eukaryotes, its biological relevance in bacteria is still emerging. Recent reviews have delved into acetylation-mediated regulation of bacterial physiology and elaborated on the interplay between nonenzymatic acetylation and carbon/acetate metabolism in bacteria (16–18). Multiple studies show the physiological role and regulation of different lysine acetyltransferases (KATs) in diverse bacteria; however, a comprehensive summary of this vast knowledge is missing (19–24). In this review, we summarize the role of known bacterial KATs (bKATs) and their regulation across bacterial species, reflecting their importance in bacterial adaptation to diverse habitat. Detailed structure-function analysis of bKATs will help provide a clearer understanding of their function and pave the way for future drug development.

FIG 1.

Overview of acetylation in bacteria. Bacteria can metabolize glucose and other carbon sources (e.g., fatty acids and amino acids) to produce Ac-CoA. The phosphotransacetylase and acetate kinase pathway (PTA/AckA) constitutes the main pathway for acetate dissimilation. Ac-CoA and inorganic phosphate are reversibly converted to AcP and CoA by PTA, while AckA converts AcP and ADP to acetate and ATP in a reversible manner. AcP serves as a predominant mediator of protein acetylation in bacteria due to its high acetyl transfer potential. In the enzyme-mediated acetylation, bKATs catalyze the transfer of an acetyl group from Ac-CoA to the substrate. To date, multiple bacterial acetylome studies have reported the abundance of acetylated proteins in metabolism, translation, and transcription processes across bacterial species. However, relatively few studies have deciphered the regulatory mechanisms mediated through bacterial protein acetylation. Moreover, due to several uncharacterized and partially characterized bKATs and bKDACs, the picture of bacterial protein acetylation remains largely unclear.

NONENZYMATIC AND ENZYMATIC ACETYLATION IN BACTERIA

When Escherichia coli is grown on glycolytic substrates, the phosphotransacetylase and acetate kinase pathway (PTA/AckA) constitutes the main pathway for acetate dissimilation. Ac-CoA and inorganic phosphate are reversibly converted to AcP and CoA by PTA, while AckA converts AcP and ADP to acetate and ATP in a reversible manner (Fig. 1) (25, 26). The reversible PTA-AckA is a low-affinity acetate assimilation pathway (K_m for acetate, 7 to 10 mM) that synthesizes Ac-CoA in the presence of high levels of acetate (26). Thus, when acetate levels are high, abundant AcP is available for serving as a direct acetyl group donor to susceptible lysine residues of proteins. This is due to the high acetyl transfer potential of AcP (Gibbs free energy of hydrolysis is −40.5 kJ/mol) (14). The predominance of AcP mediated acetylation is emphasized by comparing acetylation profiles of ackA and ackA-pta mutant strains of E. coli. The ackA mutant strains show hyperacetylation due to accumulation of AcP, while the ackA pta double mutant exhibits hypoacetylation due to lack of AcP production (14, 15). Moreover, ackA and ackA-pta mutant strains exhibit severe growth defects in E. coli, which is indicative of the physiological relevance of AcP (25). In addition to its acetylating ability, AcP’s biological significance is partly attributed to its ability to serve as a direct phosphoryl group donor. Thus, as donors of acetyl and phosphoryl groups, AcP has a global regulatory impact on bacterial signaling processes through modulation of global response regulators RcsB, PhoP, and CheY (27–30).

N-Acetyltransferases are responsible for the enzymatic acetylation of proteins in bacteria (Table 1). To uncover the physiological relevance of bacterial lysine acetylation, studies should increasingly focus on functional aspects of bKATs. Since the discovery of the first bKAT in Salmonella enterica in 2004 (31), there has been significant progress in protein acetylation research in bacteria. SePat of S. enterica, PatZ (a homolog of SePat) in E. coli, AcuA in Bacillus subtilis, and Rv0998 in Mycobacterium tuberculosis are some of the well-characterized KATs in bacteria (31–36) (Table 1). Acetylome analysis of proteomes of wild-type (WT) and bKAT deletion strains can reveal specific targets of the putative KATs. In bacteria, the high intracellular concentration of AcP mediates global acetylation of susceptible lysine residues of proteins. The AcP-mediated acetylation in bacteria is predominant and is likely to outnumber the enzyme-catalyzed events. As a consequence, the site-specific action of bKATs may remain masked. This poses challenges for elucidating the functional impact of many putative bKATs. In an exciting work by Christensen et al., a quadruple mutant (Δpta Δpat Δacs ΔcobB) strain of E. coli was constructed to identify novel KATs in the organism (20). This approach eliminated the known means of acetylation in E. coli by AcP, PatZ, and Ac-CoA synthetase (Acs) (discussed below). The overexpression of novel KATs (YjaB, PhnO, RimI, and YiaC) in the acetylation-reduced background emphasized their respective functional activity (20). Further, the deletion of the gene encoding the deacetylase CobB ensured the irreversibility of residual acetylation in the quadruple mutant (20). Similar studies can lead to the discovery of relevant bKATs across bacterial species.

TABLE 1.

bKATs

Physiological process	Organism	Protein substrate(s)	KAT(s) involved in acetylation; site(s) of modification	Impact(s) of acetylation	Reference(s)
Bacterial metabolism	S. enterica/E. coli	AMP-forming acetyl-CoA synthetase	SePat/PatZ (type I GNAT); K609	Inactivation of function by blocking the formation of acetyl-AMP intermediate	31
	Bacillus subtilis		AcuA (type IV GNAT); K549	Functional inactivation of enzyme	36
	Streptomyces lividans		SlPatB (type III GNAT); K610	Functional inactivation of enzyme	57
	Staphylococcus aureus		SaAcuA (type IV GNAT)
	Micromonospora aurantiaca		MaKAT (type III GNAT)	Functional inactivation of enzyme	69
	M. tuberculosis		Rv0998 (type III GNAT); K617	Functional inactivation of enzyme	68
	Myxococcus xanthus		MxKAT (type III GNAT); K622	Functional inactivation of enzyme	97
	Saccharopolyspora erythraea	AcsA1, AcsA2, AcsA3	SacAcuA (type IV GNAT); K620, K628, K615, K611, K237, K380, K250	Functional inactivation of enzyme	22
		Glutamine synthetases (GlnA1, GlnA4)	SacAcuA	Gain of function of GlnA1, inactivation of GlnA4	21
	M. tuberculosis	Isocitrate dehydrogenase	Rv2170 (type IV GNAT); K30, K129	Inactivation of function	19
	Mycobacterium	Mycobactin biosynthetic genes-MbtA, FadD33	Rv0998; K546 (MbtA); MsPat; K260, K511 (FadD33)	Inactivation	75, 76
Bacterial persistence	Salmonella enterica	TacA (antitoxin)	TacT (type IV-GNAT); K44	Enhances TacT acetyltransferase activity on aminoacyl-tRNA synthase	81
Bacterial transcription	M. tuberculosis	Histone-like protein HU	Eis (type V-GNAT) (32 lysines were acetylated in HU upon overexpression)	Acetylation of HU inhibited its DNA binding ability and negatively impacts the bacterial nucleoid compaction.	86
	Bacillus subtilis	Histone-like protein HBsu	YfmK (type IV); K3, K18, K14, K80	Negatively impacts nucleoid compaction	88
	E. coli	Leucine-responsive regulatory protein (LRP)	PatZ; K36	Reduced DNA binding and expression of fimbrial genes	114
	M. tuberculosis	DosR	Rv0998; K182	Transcriptional Inhibition	89
	S. enterica	HilD	SePat; K297	Transcriptional inhibition, increase in HilD stability	90
		PhoP	SePat; K201	Transcriptional Inhibition	91
	Porphyromonas gingivalis	Gingipains	PG1842, VimA; Y230 (O-acetylation), K247, K248	Gain of function in activity in the stationary phase	115
Bacterial virulence	M. tuberculosis	Mitogen activated protein kinase phosphatase 7 (MKP-7)	Enhanced intracellular survival protein (Eis); K55	Gain of function in activity	116
Prokaryotic ribonucleases	E. coli	RNase R	PatZ; K544	Proteolytic degradation	61
		RNase II	PatZ; K501	Inactivation	117

Enzyme-driven acetylation of proteins could be advantageous for bacterial adaptation in a diverse habitat, as these respond swiftly to biological stimuli. There are several partially characterized and putative bKATs, of unknown biological function and significance (37, 38). Hence, the complete characterization of relevant bKATs can unravel new modes of acetylation-mediated regulation in bacteria.

BACTERIAL N-ACETYLTRANSFERASES

Bacterial N-acetyltransferases can be grouped into Nα-acetyltransferases (bNATs) and Nε-acetyltransferases (bKATs). Nα-acetylation is catalyzed by bNATs, which transfer the acetyl group from Ac-CoA to the amino group of the first protein residue. In contrast to Nε-acetylation, Nα-acetylation is always enzyme mediated, typically irreversible, and less reported in bacteria. To date, 47, 117, and 145 proteins have been reported to be N-terminally acetylated in E. coli, Pseudomonas aeruginosa, and Acinetobacter baumannii, respectively (39, 40). RimI, RimJ, and RimL (Rim family) are well-studied bNATs that mediate acetylation of ribosomal proteins at S5, L7/12, and S18 in E. coli (41–43). A recent study in S. enterica reported the regulation of CobB (bacterial NAD⁺-dependent deacetylase [see below]) by a non-Rim family bNAT, YiaC. It negatively impacts the activity of the long isoform of CobB, CobB_L, through N-terminal acetylation (44). Further studies are likely to reveal the deeper physiological relevance of N-terminal acetylation in bacterial physiology.

bKATs.

Nε-Acetyltransferases (KATs) are present in all living organisms to catalyze acetylation of biomolecules. Among them, the GNAT (Gcn5-related N-acetyltransferase), MYST (MOZ, Ybf2, Sas2, and Tip60), and p300/CBP families are the best-studied examples of N-acetyltransferases in eukaryotes. Eukaryotes also have the SRC, TAFII250, TFIIIC, Rtt109, CLOCK, and Camello family of acetyltransferases (45, 46).

The bacterial GNATs (bGNATs) were first discovered in 1965 as aminoglycoside N-acetyltransferases in multiple drug-resistant E. coli strains that conferred resistance to chloramphenicol and kanamycin (47). The GNAT superfamily of enzymes have a low primary sequence homology, which makes their substrate range very versatile. These enzymes acetylate a wide range of biomolecules that includes lysine side chains, N termini of proteins, serotonin, and polyamines (48–50). The GNAT domain has a secondary structure that comprises a mix of beta sheets and alpha helices. Four distinct motifs characterize the GNAT catalytic domain in the order C, D, A, and B. Motif A is most conserved among the four motifs and is vital for Ac-CoA binding and catalysis. Despite the diversity of substrate selection, the mode of interaction of Ac-CoA is conserved across the GNAT family (50–53). BGNATs have been categorized based on the arrangement and number of GNAT domains present. Currently, there are five domain organizations of bGNAT-KATs (type I to V) (Fig. 2), as previously described (16, 17). It is important to note that bKATs mainly belong to the GNAT superfamily, and there is a minor group that belongs to the YopJ group of acetyltransferases (54) (Fig. 2).

FIG 2.

Classification of prokaryotic acetyltransferases. Bacterial acetyltransferases are broadly classified in the GCN5-type N-acetyltransferases and the YopJ family. The majority of the known lysine acetyltransferases (KATs) belong to the GNAT family and have been assigned to 5 classes based on GNAT domain arrangement. Class I KATs contain a large regulatory N-terminal domain, homologous to nucleoside diphosphate (NDP) forming CoA ligases, and a GNAT catalytic domain at the C terminus. Class II KATs are similar to type I, with their domain organization being in the reverse order. The catalytic GNAT domain is within the N terminus, and the large regulatory domain is on the C terminus. Class III KATs have an allosteric regulatory domain at the N terminus followed by the functional GNAT domain at the C terminus. Type IV KATs possess a lone GNAT domain and make up the majority of the bacterial KATs. Class V comprises two KAT domains in tandem prior to the C-terminal region. The classification scheme was adapted from references 16 and 17.

Reaction mechanism of bGNAT-KAT.

The reaction mechanism of bKATs proceeds via a ternary complex formation by the close association of the KAT, Ac-CoA, and the substrate. It comprises three steps: (i) proton removal from the target lysine of the substrate by an acidic residue of the enzyme, mostly glutamate; (ii) nucleophilic attack of the deprotonated lysine on the carbonyl carbon of bound Ac-CoA; and (iii) protonation and release of the coenzyme A, usually by a tyrosine residue in the KAT active site (16, 17, 52). GNAT-KATs have a conserved sequence motif (Q/RXXGXG/A) which serves as a hydrophobic pocket for Ac-CoA binding (50–52). This ensures that Ac-CoA is well positioned along with the enzyme and substrate to facilitate catalysis.

The target lysine in the substrate sometimes can get deprotonated by a proton wire, a water-filled channel formed by a series of residues in the enzyme to move the charge away from the active site towards the bulk solvent or protein surface (53, 55). Often the region in the vicinity of the bacterial acetylation site is rich in acidic residues, promoting lysine deprotonation (40, 55).

BACTERIAL DEACETYLASES SYNCHRONIZE REVERSIBLE LYSINE ACETYLATION

Bacterial lysine deacetylases (bKDACs) catalyze the hydrolysis of the acetyl group from specific lysine residues of proteins. These enzymes coordinate with bKATs and the nonenzymatic donor AcP to facilitate the regulation of protein function (12). Two broad families of bKDACs are NAD⁺-dependent and Zn²⁺-dependent deacetylases (56–59). The NAD⁺-dependent deacetylase named CobB (in E. coli and S. enterica) bears homology to mitochondrial sirtuins. The first and most widely studied, CobB, was discovered in S. enterica on the basis of its regulation of AMP-intermediate-forming Acs (59). CobB-dependent deacetylation reaction consumes NAD⁺ and yields O-acetyl-ADP (ADP ribose) and nicotinamide (58). Some bacteria may have more than one functional KDAC. The Streptomyces lividans genome encodes three deacetylases: two NAD⁺-dependent deacetylases (CobB and SrtA) and a zinc-dependent deacetylase (AcuC) (56). bKDACs may have evolved to inhibit high levels of acetylation that can prove detrimental to bacterial physiology. A similar observation has been made for mitochondria, where the activity of the deacetylases Sirt3 and Sirt5 suppresses deleterious nonenzymatic acetylation due to the accumulation of Ac-CoA (13). Reportedly, CobB deacetylates only 10% of the acetylation in E. coli (12, 15, 60). CobB is a predominant deacetylase in E. coli and is involved in regulating many acetylated essential proteins (60). However, there could be other unidentified bKDACs targeting specific substrates. Some acetylated proteins in E. coli could be sensitive to degradation and may not need deacetylase-mediated regulation (61). At present, it is unknown how E. coli adopts alternative strategies to regulate acetylations not reversed by CobB.

PHYSIOLOGICAL ROLE OF bKATs IN BACTERIA

Metabolism. (i) Ac-CoA synthetase.

The survival and adaptability of all living organisms are intimately linked to the regulation of metabolic processes. Proteins involved in metabolic processes show prominent acetylation in most global bacterial acetylome studies (62–66). The metabolic enzymes, especially those involved in the synthesis and utilization of Ac-CoA, show a high degree of acetylation (66).

Ac-CoA synthetase (Acs) is a central metabolic enzyme that catalyzes the high-affinity biosynthetic pathway for Ac-CoA synthesis (K_m for acetate, 200 μM). Acetylation of Acs is a conserved regulation mechanism in bacteria and eukaryotes (31, 65–68). It activates acetate to Ac-CoA in two steps: (i) adenylation, in which acetate and AMP bind to Acs to form an acetyl-AMP intermediate (thioesterification-AMP), and (ii) replacement of thioesterification-AMP by CoA to form Ac-CoA (31). Uncontrolled Acs activity can cause excess ATP consumption and increased AMP synthesis, which is harmful to the cell (26, 66). Consequently, the cell has evolved to regulate Acs activity at various levels that include (de)acetylation-mediated regulation at the posttranslational level by KAT/KDACs. Acetylation of Acs by SePat at the conserved lysine residue K609 blocks acetyl-AMP formation by adenylation in Ac-CoA synthesis (31). CobB restores the enzyme activity by removal of the acetyl moiety. S. enterica strains lacking CobB (ΔcobB) could not grow on 10 mM acetate due to SePat-mediated inactivation of Acs. However, a double-knockout strain lacking ΔSepat/cobB could utilize the acetate and grow well due to uninhibited Acs activity (31). The precise regulation of Acs by SePat/CobB homologues is reported in other Gram-negative bacteria. SePat homologs recognize and acetylate the lysine residue within the conserved motif (PXXXXGK), which is required for the catalytic activity of AMP-forming Acs (62). Orthologs of Acs in other bacterial species are reported to be under stringent regulation by bKATs and bKDACs, though there are notable differences in their regulation mechanisms. For instance, the Acs in Mycobacterium spp. is regulated by a bKAT with an allosteric cyclic AMP (cAMP) binding domain that regulates Acs activity in response to intracellular levels of cAMP (68). Likewise, in Micromonospora aurantiaca, Acs activity is modulated by a bKAT that contains an amino acid-sensing domain which in turn is governed by cellular levels of amino acids (69). In B. subtilis and Staphylococcus aureus, Acs activity is inhibited by type IV bKAT (see Fig. 2 for types of bKATs) named AcuA. Two deacetylases remove the acetyl group from Acs in B. subtilis: the NAD⁺-independent deacetylase AcuC and the NAD⁺-dependent SrtN (56, 70). These deacetylases act in a complementary manner to remove the acetyl group from Acs under physiologically relevant conditions. To conclude, bKATs/bKDACs regulate the activity of Acs at the PTM level across bacterial species to maintain energy charge and CoA homeostasis.

(ii) Mycobacterial isocitrate dehydrogenase.

In Mycobacterium, the glyoxylate shunt supports bacterial growth from noncarbohydrate sources by utilizing acetate from fatty acids to synthesize tricarboxylic acid (TCA) cycle intermediates. Isocitrate lies at the critical branch point between the TCA cycle and glyoxylate shunt. It can be utilized in either of the processes by their respective enzymes, isocitrate dehydrogenase (TCA cycle) and isocitrate lyase (glyoxylate cycle). The isocitrate dehydrogenase has a higher substrate affinity for isocitrate than isocitrate lyase (ICL); hence, regulation of its activity is critical for determining carbon flux. A type IV bKAT, Rv2170 (Fig. 2), in Mycobacterium facilitates ICDH1 acetylation and inactivation, thus reducing carbon flux into the TCA cycle. In the presence of stearic acid as the sole carbon source, the Δicl deletion strain displayed a delayed-growth phenotype due to its inability to utilize stearic acid, whereas the Δrv2170 Δicl strain grew faster than the Δicl strain under the same growth conditions. This highlighted the significance of Rv2170-mediated regulation of carbon flux through the TCA cycle (19). This regulation is distinct from other bacteria like E. coli which employ phosphorylation of its ICDH to regulate carbon flux through the TCA cycle (71). Acetylation-mediated regulation of ICDH1 is analogous to the regulation of mammalian mitochondrial ICDH, which bears structural homology to the M. tuberculosis ICDH1 (72, 73). Parallel studies have also reported ICL1 acetylation (K322 and K392) and its role in the regulation of carbon metabolism in M. tuberculosis (74). Hence, acetylation of ICDH1 and ICL1 could be one of the strategies utilized by M. tuberculosis to remodel its metabolism to adapt to adverse conditions in the host during infection.

(iii) Mycobacterial siderophores.

M. tuberculosis has an iron chelation system comprising the siderophores mycobactin and carboxymycobactin. These siderophores allow the pathogen to combat iron starvation inside the host macrophage environment, and their regulation can determine the growth and virulence of M. tuberculosis. Mycobactin activity is modulated through Rv0998-mediated acetylation of two mycobactin biosynthesis enzymes, MbtA and FadD33. Acetylation of the siderophore biosynthesis enzymes inactivates them, which maintains energy homeostasis and also prevents excess iron accumulation in M. tuberculosis. The NAD⁺-dependent deacetylase of M. tuberculosis Rv1151c removes the acetyl moiety from the enzymes (MbtA and FadD33), thereby restoring their function. Although MbtA (de)acetylation was not directly linked to its virulence in the host, Δrv0998 and Δrv1151c mutant strains displayed distinct phenotypes while growing in low-pH and iron-deficient media. While the Δrv0998 strain showed a significant growth advantage due to uninhibited mycobactin activity, Δrv1151c displayed much slower growth in low-pH and iron-poor media (75, 76). In vivo studies of these mutant strains of mycobacteria could reveal their true impact on physiological processes.

Metabolic regulation and its link to virulence. (i) Acetate uptake from host intestine.

Protein acetylation can alter host-pathogen interaction by impacting major physiological processes like metabolism and gene expression in bacteria. SePat homologs in Gram-negative bacteria are known to acetylate the lysine residue within the conserved motif (PXXXXGK) of Acs catalytic core (62). During Vibrio cholerae infection in Drosophila, YfiQ (SePat homolog) and bKDAC-CobB regulate acetate assimilation from the host gut. Acetate accumulated in the host serves as a carbon source that promotes V. cholerae growth in the anaerobic environment of the gut. Bacterial Acs mediates acetate assimilation from the host, which causes dysregulation in acetate levels in the fly gut. The impaired short-chain fatty acid production (SCFA) from acetate in the host gut leads to fly mortality. Acs acetylation by YfiQ inhibits acetate consumption, which causes reduced V. cholerae virulence in the host. A CobB mutant of V. cholerae showed a low growth rate due to its inability to use acetate as the sole carbon source. In contrast, the ΔyfiQ ΔcobB double mutant displayed an improved growth rate due to effective acetate assimilation from uninhibited Acs activity. Thus, (de)acetylation regulates bacterial virulence by modulating metabolite exchange between the pathogen and its colonized hosts (77).

(ii) Bacterial persistence.

Persistence is a unique form of adaptation where a metabolically inactive bacterial subpopulation shows resistance to a range of stressful conditions in bacteria. Toxin-antitoxin (TA) modules in bacterial genomes mediate bacterial quiescence. Among the six types of toxin-antitoxin modules, the type II TA module is the most studied. Recent discoveries have shown that certain type II toxins contain a catalytic GNAT fold, thus functioning as acetyltransferases. These GNAT toxins usually acetylate free amine groups on charged tRNAs, to inhibit protein translation and induce persistence. AtaT/AtaC and TacT/TacA of S. enterica, GmvT/GmvA of Shigella spp., and KacT/KacA of Klebsiella pneumoniae are some well-studied GNAT TA systems (78–81). In S. enterica, TacT toxin has two substrates: (i) the cognate antitoxin partner TacA and (ii) aminoacyl tRNAs. TacT catalyzes the acetylation of TacA at K44, which enhances the acetyltransferase activity of TacT on aminoacyl-tRNA, inhibiting bacterial translation. This acetylation was reversed by CobB (81). This study provides an interesting insight, in that KAT activity of a GNAT toxin halts protein synthesis to facilitate bacterial entry into the persister state.

Transcription. (i) Bacterial nucleoid-associated proteins.

The nucleoid-associated proteins (NAPs) predominantly drive the compaction of bacterial chromosome within the cell. These proteins are abundantly associated with bacterial chromatin and serve as global transcription factors (TFs) in bacteria. In E. coli, the major NAPs include HUPA/B, H-NS, Fis, and IHF. Global acetylome studies have shown acetylation of HU and other bacterial NAPs at multiple lysine residues (14, 15, 82). Careful analysis of the acetylome profile by Kuhn et al. revealed that some acetylation marks were absent in a Pat/YfiQ mutant of E. coli, suggesting a possible role for the bKAT-based regulation of NAPs (15).

HU is a highly conserved, basic DNA-bending protein that regulates bacterial survival, growth, and virulence (82–85). A study on acetylation of HU provided initial insights on the bacterial nucleoid modulation in M. tuberculosis. Eis, a type V bGNAT (Fig. 2), acetylated HU and inhibited its interaction with DNA, thereby negatively impacting nucleoid compaction (Fig. 3A). Mycobacterial HU contains a lysine-rich C-terminal region (PAAK/KAAK motifs). It bears a resemblance to the N-terminal domain of eukaryotic histones, which contains the most acetylatable lysines. This C-terminal extended region is unique to Actinobacteria and is not present in other HU homologs across bacteria.

FIG 3.

Examples of KAT-based regulation of virulence and transcription in M. tuberculosis. (A) Epigenetic regulation of histone-like protein by lysine acetylation in M. tuberculosis. The acetylation of the highly conserved and abundant nucleoid-associated protein HU by enhanced intracellular survival protein (Eis) reduces its DNA binding ability. It negatively impacts nucleoid compaction, leading to a relaxed form of DNA. Deacetylation by the NAD⁺-dependent deacetylase Rv1151c leads to the formation of dense bacterial chromatin. (B) Regulation of bacterial virulence in Mycobacterium. The mycobacterial KAT Rv0998 facilitates acetylation of the dormancy survival regulator (DosR) at K182, reducing its DNA binding ability and thereby inhibiting transcription of hypoxia survival genes. Further, the deacetylation of DosR by Rv1151c under hypoxic conditions enhances its DNA binding ability to promote transcription of target genes, allowing M. tuberculosis to transition to a dormant state.

In vitro acetylation of mycobacterial HU with Eis leads to acetylation of 32 lysines, and most of it is localized to the C terminus. The mycobacterial NAD⁺-dependent KDAC Rv1151c deacetylated the modified HU and restored its DNA binding ability. Eis overexpression in M. tuberculosis mediated nucleoid decompaction, while Rv1151c overexpression led to minimal change in the overall nucleoid architecture (86, 87).

The HU homolog in B. subtilis, HBsu is acetylated at seven sites in vivo (K3, K18, K37, K41, K75, K80, and K86) and plays a direct role in the regulation of bacterial nucleoid compaction. A type IV bKAT (Fig. 2), YfmK, acetylates four (K3, K18, K41, and K80) of the seven sites in vitro, highlighting its role in regulating nucleoid compaction in B. subtilis. The role of a deacetylase in the modulation of HBsu function is currently unknown (88).

(ii) Transcriptional regulation and its link to virulence.

The acetylation of transcription factors, especially at their DNA binding motifs, introduces perturbation in its interaction with DNA, thereby negatively impacting the transcription of the downstream target genes. Dormancy survival regulator (DosR) is a crucial transcription factor that promotes the adaptation of M. tuberculosis to hypoxic conditions in the host 89. Residue K182 in the DNA binding motif of DosR is considered essential for DNA binding (89). K182 of DosR is acetylated by Rv0998, which inhibits transcription of hypoxia survival genes (89). Further, the deacetylation of DosR by Rv1151c under hypoxic conditions enhances its DNA binding ability to promote transcription of target genes, allowing M. tuberculosis to transition to a dormant state (Fig. 3B) (89).

The major virulence factors and transcription regulators of S. enterica HilD and PhoP also undergo deacetylation to facilitate the expression of target genes required for bacterial adaptation during host infection (90, 91). SePat-driven acetylation of K297 located at the helix-turn-helix (HTH) motif of HilD reduced its transcriptional activity and enhanced its stability, albeit via an unknown mechanism (90). During host infection, CobB mediated deacetylation of HilD allows activation of the downstream Salmonella pathogenicity island 1 (SPI-1)-inducing genes. SPIs are defined as large gene cassettes within the Salmonella chromosome that encode virulence factors responsible for establishing specific interactions with the host. They are required for bacterial virulence in a given animal model. There are five gene cassettes of SPI, and among them, SPI-1 and SPI-2 mainly encode virulence determinants. Uncontrolled SP1-1 induction can be energy expensive for S. enterica, causing a reduced growth rate. Hence, the acetylation/deacetylation of HilD is physiologically significant, as it precisely monitors SPI induction while stabilizing its activity, thereby maintaining appropriate levels of HilD for SPI-1 induction (90).

PhoPQ is a highly conserved two-component signal transduction system in Gram-negative pathogens. It regulates virulence in response to various environmental conditions, such as low pH, low levels of Mg²⁺ ions, nutrient starvation, and free radicals that are likely to be encountered in the host macrophage environment. The PhoP response regulator is indispensable for S. enterica survival within the host macrophages and is subject to acetylation by enzymatic and nonenzymatic means at two different lysine residues. SePat mediates PhoP acetylation at K201, while AcP acetylates it at K102 (29, 91). When bacteria encounter hostile conditions in macrophages (low magnesium and low pH), there are decreased acetylation levels at both sites (K102 and K201). While CobB deacetylated PhoP at K201 (91), the mechanism of K102 deacetylation remained unknown (29). In both the studies, acetylation mimics of PhoP-K102Q and PhoP-K201Q in S. enterica caused significant attenuation in intestinal inflammation and systemic infection in the mouse, suggesting a crucial role of PhoP acetylation in S. enterica pathogenesis. To conclude, these instances highlight the physiological relevance of acetylation in regulating bacterial gene expression to adapt to adverse and nutrient limiting conditions in the host (29, 91).

MODES OF KAT REGULATION IN BACTERIA

Nitrogen starvation upregulates KAT activity in Actinobacteria.

Deciphering the regulatory events that modulate the bKAT activity is required to comprehend their physiological roles in bacteria. At present, there are few studies that elucidate the regulatory circuitry responsible for bKAT activity. In Saccharopolyspora erythraea, the global nitrogen regulator, GlnR senses nitrogen scarcity to mediate nitrogen and carbon assimilation through direct transcriptional activation of glutamine synthetase genes (GlnA1 and GlnA4) and acetyl-CoA synthetase genes (acs1, acs2, and acs3), respectively (21, 22) (Fig. 4A). The GlnR-mediated activation of Ac-CoA synthesis under nitrogen-shortage is imperative. It feeds into the TCA cycle, causing increased intracellular concentrations of nitrogen acceptor molecules, 2-ketoglutarate, and oxaloacetate. This leads to efficient nitrogen assimilation, which is required to synthesize amino acids, proteins, and nucleic acids. However, Ac-CoA synthesis regulation is crucial for maintaining energy homeostasis and preventing the toxic accumulation of Ac-CoA. GlnR effectively exerts transcriptional and posttranslational control over Acs synthetase genes through transcriptional activation of SacAcuA (type IV bKAT) and SacSrtN (CobB homologue) in S. erythraea (23). SacAcuA/SacSrtN modulates the activity of glutamine synthetases (GS) and Acs as per the metabolic needs of the organism. While SacAcuA mediated acetylation of GlnA4 and Acs dampens their enzymatic activity, the acetylated GlnA1 acts as a chaperone to enhance GlnR-DNA interaction. Hence, S. erythraea utilizes enzyme-mediated acetylation to coordinate carbon and nitrogen metabolism in response to nitrogen starvation in the cell (21).

FIG 4.

(A) Regulation of KAT/KDAC activity in response to nitrogen-limiting conditions in Saccharopolyspora erythraea through a positive feedback loop comprising the GlnR-KAT/KDAC-GlnA1 circuit. Under nitrogen-limited conditions, the global nitrogen regulator GlnR transcriptionally upregulates the expression of glutamine synthetase genes (glnA1 and glnA4), acetyl-CoA synthetase genes (acsA1, acsA2, and acsA3), and genes encoding enzymes of the acetylation system (acuA and srtN). The AcuA/SrtN acetyltransferase/deacetylase pair regulates the acetylation of glutamine synthetases GlnA1 and GlnA4] and genes for acetyl-CoA synthetases (AcsA1, AcsA2, and AcsA3) (not shown). Acetylation of GlnA4 downregulates its activity, whereas acetylated GlnA1 enhanced its chaperone activity, leading to conformational changes in GlnR. This increases GlnR-DNA interaction and the transcription upregulation of downstream targets. (B) cAMP-based regulation of lysine acetylation. (i) In the enteric bacteria, under carbon limitation, cAMP levels rise, and cAMP binds cAMP receptor protein (CRP) to form an active complex. cAMP-CRP directly upregulates the transcription of acetyltransferases, Pat, and acetyl-CoA synthetase (Acs). cAMP directly binds to the Acs and competes against one of its substrates, ATP. The binding of cAMP increases Acs Lys609 acetylation, causing inhibition of Acs activity. (ii) In mycobacteria, Rv0998/MtPat encodes a type III GNAT, which has a cAMP binding domain at the N terminus. The acetyltransferase is activated upon cAMP binding in the allosteric site, thereby acetylating its physiological substrate.

Mycobacterial species (Mycobacterium smegmatis and M. tuberculosis) adopt a mode of acetylation-based regulation similar to that in S. erythraea. The nitrogen-sensing regulator GlnR senses nitrogen scarcity in the host microenvironment, thus facilitating the assimilation of short-chain fatty acid (SCFA) through transcriptional activation of propionyl-CoA synthetase (MsPrpE) and Ac-CoA synthetase (MsAcs). Additionally, GlnR also activates MsPat (or Rv0998 in M. tuberculosis) to modulate acyl-CoA levels through acetylation of MsPrpE and MsAcs. Such regulation is crucial for the survival of Mycobacterium in the nutrient-deprived host microenvironment (23). To conclude, these instances provide novel insights into bKAT/bKDAC regulation in response to deficient nitrogen levels in Actinobacteria. More work is required for discerning the signaling cascades that trigger acetylation events across bacterial species.

Regulation by the catabolism repressor in Salmonella enterica.

myo-Inositol is a cyclic polyol present in the soil, and the genes required for its utilization as a carbon source are organized in a genomic island. The canonical role of the myo-inositol repressor IolR is to inhibit the expression of myo-inositol-degrading enzymes in the absence of myo-inositol (24). Intriguingly, a study in S. enterica showed its role in concomitant activation of SePat, CobB and Acs (24). This was emphasized by the inability of a ΔiolR mutant to utilize acetate as the sole carbon source in the minimal medium. Increased levels of SePat and CobB, however, compensated for the growth defect exhibited by the ΔiolR mutant (24). Though the study confirmed the role of IolR in direct transcriptional activation of SePat, its mode of action on CobB expression could not be ascertained. Such a regulatory mechanism could provide bacteria efficient means of Ac-CoA production by integrating myo-inositol and acetate metabolism (24).

Impact of metabolites and small molecules—Ac-CoA, cAMP, NADP⁺, and amino acids. (i) Ac-CoA.

Protein acetylation links the cell's metabolic status to KATs/KDACs, as they rely on the intracellular concentration of Ac-CoA and NAD⁺ for their (de)acetylation activities. Acetate activation to Ac-CoA is enabled by two means: (i) a high-affinity biosynthetic pathway via Acs and (ii) a low-affinity AckA-Pta pathway (26) (Fig. 1). Apart from acetate metabolism, bacteria can metabolize glucose and other carbon sources (e.g., fatty acids, amino acids) to produce Ac-CoA. Available Ac-CoA thus serves as a crucial metabolite taking part in multiple metabolic pathways like production of reducing equivalents in the TCA cycle and fatty acid synthesis for membrane production. In addition, Ac-CoA is an active participant in the posttranslational modification of proteins by functioning as the primary donor of acetyl groups in the KAT driven acetylation. Hence, its availability significantly regulates the activity of KATs, especially those having a higher K_m (Michaelis constant) value for Ac-CoA. For example, AcuA in Bacillus subtilis has a higher K_m value (22 mM) (36) for Ac-CoA than PatZ (K_m, 10.3 μM) (55) and exhibits a weaker KAT activity due to reduced sensitivity to Ac-CoA concentration.

(ii) Autoacetylation of bKATs by Ac-CoA.

Autoacetylation of KATs is a common and well-studied regulatory phenomenon in all eukaryotes (92, 93). A similar observation has been made in bacteria, although the mechanism and functional impact of most of the bKAT autoacetylation is unknown. In E. coli, PatZ is present as a stable homotetramer in the absence of Ac-CoA (55). The rise in cellular levels of Ac-CoA facilitates self-acetylation of PatZ tetramers at five lysine residues (K146, K149, K391, K447, and K635), mediating the formation of more stable and active PatZ octamers (55). Thus, a balancing feedback loop operates, preventing Ac-CoA accumulation and subsequent inactivation of PatZ. Here, PatZ acts as a molecular sensor for Ac-CoA, precisely regulating its functional activity in response to various metabolite levels. Acetylation induced oligomerization of PatZ was confirmed when mutation of target lysine residues to arginine (K146R, K149R, K391R, K447R, and K635R) failed to result in octamers (55). More studies on the structure and function of autoacetylated bKATs are required to reveal the impact of these modifications on KAT regulation.

(iii) cAMP.

cAMP-based tuning of KAT activity has been implicated in enteric bacteria (Fig. 4B, panel i). It coordinates the acetylation of Acs via concomitant stimulation of Acs and PatZ expression (95). In E. coli, cAMP levels increase in response to low glucose or during the stationary phase. It forms a transcriptionally active complex with the catabolite repressor protein (CRP), known as the cAMP-CRP complex, which transcriptionally activates the expression of Acs and PatZ (94, 95). However, it should be noted that only putative CRP binding sites in PatZ promoter region were identified, and there is no direct evidence of CRP participating in PatZ transcriptional activation. When Ac-CoA levels rise in the cell due to continuous Acs activity, PatZ can inhibit Acs activity through acetylation. Hence, concomitant activation of Acs and PatZ ensures that PatZ exhibits Acs regulation as per the cell's metabolic requirements. In S. enterica, under in vitro conditions, increasing amounts of cAMP bind to the ATP/AMP binding pocket of SeAcs (Acs homolog in S. enterica), thereby restraining its open conformation. Subsequently, SePat, which is activated by cAMP, mediates enhanced acetylation of SeAcs at K609 (96). Hence, Acs is regulated at transcriptional and posttranslational levels to maintain energy homeostasis.

Intracellular levels of cAMP directly tune KAT activity in M. tuberculosis. Rv0998 has an N-terminal cAMP binding domain fused with a C-terminal GNAT domain. Here, the activity of Rv0998 is allosterically stimulated in response to intracellular levels of cAMP (34, 35) (Fig. 4B, panel ii).

(iv) NADP⁺.

bKATs have acquired allosteric regulatory binding domains in their N termini in response to specific extracellular and intracellular milieus. MxKAT, from Myxococcus xanthus, negatively regulates the acetylation of Acs in response to cellular levels of NADP⁺ (97). MxKAT contains an NADP⁺-binding domain and a GNAT domain at the N terminus and C terminus, respectively. High cellular levels of NADP⁺ repress the activity of MxKAT, thereby preventing Acs acetylation and allowing Ac-CoA synthesis (97). Here, NADP⁺-mediated MxKAT modulation in M. xanthus is analogous to the regulation of acetylation events by NAD⁺-dependent CobB in other bacteria. Dehydrogenases of metabolic pathways (gluconeogenesis, TCA cycle, and pentose phosphate pathways) serve as prominent sources of NADP⁺, and their high levels are physiologically relevant to the deacetylated or active forms of enzymes. This could suggest the possibility of a regulatory circuit where active (or deacetylated) dehydrogenases in M. xanthus synthesize NADP⁺, causing negative regulation of MxKAT. As Ac-CoA levels rise due to uninhibited Acs activity, there is likely reactivation of MxKAT to repress the activity of Acs and other metabolic enzyme targets, if any.

(v) Amino acids.

Amino acid-sensing acetyltransferases (ACT-GNATs) consist of an ACT domain (aspartate kinase, chorismate mutase, and TyrA [prephenate dehydrogenase]) fused to the GNAT domain (69). These acetyltransferases have been identified in abundance in Actinomycetes and are allosterically regulated by amino acid concentration. MaKAT is the first identified ACT-GNAT in the actinomycete Micromonospora aurantiaca whose activity is allosterically stimulated by increasing concentrations of cysteine and arginine (69). Such regulation of bKATs is unique to Actinobacteria, piquing curiosity regarding its link with amino acid metabolism. Recently, ACT-GNATs have been categorized into two types based on their allosteric effectors. The asparagine-binding domain is mainly found in Streptomyces, whereas the cysteine-binding domain is prevalent in other actinobacteria (98). These ACT-GNATs may modulate enzymes involved in the metabolism of the amino acid, which signals its stimulation. However, a study to confirm this hypothesis is currently lacking.

KAT regulation by USP in M. smegmatis.

Mycobacterium smegmatis adopts a unique mechanism to regulate cAMP-dependent KAT activity (Rv0998). A recent study in M. smegmatis reported the sequestration of MsKAT by universal stress protein (USP), which inhibited its acetylating action on a range of substrates involved in biofilm formation (99). A Δusp strain of M. smegmatis showed rough colony morphology and reduced biofilm formation, which was restored in the Δusp Δkatms mutant. The Δusp strain also showed a phenotype similar to that of the hyperacetylated M. tuberculosis strain, devoid of its cognate NAD⁺-dependent deacetylase (Δrv1151c). These results collectively highlight a novel mechanism of KAT regulation in M. smegmatis, where a proteinaceous competitive inhibitor restricts cAMP-regulated KAT activity (99).

CONCLUDING REMARKS

Protein acetylation in bacteria was previously considered either insignificant or absent. The significance of bacterial acetylation began to emerge in the late 1990s, when studies implicated its role in bacterial chemotaxis (acetylation of CheY at K92 and K109) (30). Later, the discovery of enzyme-driven regulation of Acs by SePat/CobB in S. enterica set the stage for protein acetylation studies in bacteria. Subsequently, the roles of many bacterial acetyltransferases with the GNAT functional domain have come to light. Recent developments in high-resolution mass spectrometry and high-affinity purification technology for lysine-acetylated peptides have led to the availability of a plethora of information on the bacterial acetylome. Initial proteomic studies were focused on model organisms—E. coli and S. enterica—followed by pathogens like M. tuberculosis, V. cholerae, Borrelia burgdorferi, Streptococcus pneumoniae, and Francisella spp. (9, 63–65, 100–103). Now, acetylproteome studies have extended to marine microbes like Shewanella baltica, (a food spoilage organism of aquatic products) and Vibrio alginolyticus (a marine pathogen). This has opened up new avenues of research and has provided an insight into the physiological relevance of acetylation in diverse environmental contexts.

Bacterial acetylation profiles are unique to their environmental conditions. For instance, a recent study compared the lysine acetylome profiles of human gut microbiota in patients of Crohn's disease with those of healthy individuals (104). The differential levels of protein acetylation in disease versus control samples suggested the relevance of lysine acetylation in the gut microbiome of patients who have Crohn's disease (104). Hence, to examine the physiological relevance of the acetylation, lysine acetylomes of clinical isolates (for pathogenic strains) or environmental isolates should be studied to assess the role of these modifications pertinent to their distinct phenotypes (105, 106).

In eukaryotes, KATs catalyze high stoichiometric acetylation of their endogenous substrates (105, 106). This may be true for KAT-mediated acetylation in bacteria as well. Enzymes provide a favorable environment for catalysis; hence, a higher proportion of lysines may be modified by KAT-mediated acetylation. In E. coli, proteins involved in metabolism, transcription, and translation showed the highest acetylation stoichiometry (66). Numerous acetylome studies in bacteria have reported the acetylation of the transcription factors RpoD (housekeeping sigma factor) and HU at their respective conserved lysines; however, their physiological significance and acetylation mechanism remain unknown (9, 14, 15, 63, 64). Interestingly, in a recent study in Streptomyces venezuelae, the RpoD homolog HrdB was reported to be constitutively acetylated at K259 throughout bacterial growth. The modification led to an increased association of HrdB with the RNA polymerase (RNAP) core complex, which caused enhanced promoter binding by RNAP holoenzyme (107). The nucleoid protein HUPA/B undergoes acetylation at residues K3 and K13. K3 acetylation could significantly alter its interaction with DNA, as it facilitates HU dimerization by forming internal salt bridges with acidic residues (108). In Acinetobacter baumannii, the acetylated variant of K13 of HU impacted its thermal stability and DNA-binding kinetics (109). The housekeeping sigma factor and HU serve to regulate bacterial transcription globally, and we speculate that bKAT/bKDACs precisely regulate their acetylation levels. Moreover, these global TFs reportedly undergo other modifications like phosphorylation and methylation (110, 111). Whether these modifications undergo cross talk with acetylation to spatiotemporally modulate global gene expression is of interest. This could reveal the coordinated role of these bacterial PTMs in regulating global gene expression in bacteria.

To attain clarity on bacterial protein acetylation, future work should emphasize (i) studying the biological significance of putative or partially characterized bKATs, especially those which attribute to phenotypes under stress or infection conditions and (ii) functional characterization of consistent/stable acetylation of conserved proteins observed across different bacterial acetylome studies. The eukaryotic KATs and KDACs have been successfully explored as drug targets, since they are well known to regulate broad cellular processes like enzymatic activity, transcription, metabolism, and protein interactions (112, 113). Our expanding knowledge on the roles of putative bGNAT-KATs in bacterial physiological processes like metabolism, gene expression, and host-pathogen interactions can pave the way for designing KAT inhibitors for use as antimicrobial agents.

Biographies

Aiswarya Dash received her master of science degree in biotechnology (gold medalist) from Siksha O Anusandhan University, India (2012), and is pursuing her doctoral research with Rahul Modak at Kalinga Institute of Industrial Technology, Bhubaneswar, India. In 2014, she was awarded the DST-INSPIRE fellowship from the Department of Science and Technology, Government of India, to pursue her doctoral studies. She has identified a novel lysine acetyltransferase in Salmonella enterica and is studying its role in bacterial physiology and pathogenesis. On completion of her Ph.D., she wishes to explore a postdoctoral position in microbiology or scientific writing and related fields and is currently scouting for employment opportunities.

Rahul Modak is an assistant professor at the School of Biotechnology, Kalinga Institute of Industrial Technology, Bhubaneswar, India. His research group is interested in deciphering role of reversible lysine acetylation in regulation of host pathogen interactions. They use cell line and mouse model systems to study gut bacterium-host interactions. His group has identified novel acetyltransferases in Salmonella and a deacetylase in Vibrio spp. They have developed in vitro nonradioactive assays to study enzymatic acetylation and deacetylation. His earlier work showed that E. coli and S. aureus infection induced alteration of specific histone acetylation and methylation marks in a mouse model of bovine mastitis. He discovered embelin, a specific natural small-molecule inhibitor of the mammalian PCAF enzyme. He received his master of science degree in biotechnology from Indian Institute of Technology Bombay, Mumbai, India, and a Ph.D. from Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR), Bangalore, India. His postdoctoral training was at JNCASR and the Institute of Molecular Biology, Mainz, Germany.