Lysine Acetylation Regulates Alanyl-tRNA Synthetase Activity in Escherichia coli

Protein lysine acetylation is a widely conserved posttranslational modification in all three domains of life. Lysine acetylation frequently occurs in aminoacyl-tRNA synthetases (aaRSs) from many organisms. In this study, we determined the impact of the naturally occurring acetylation at lysine-73 (K73) in Escherichia coli class II alanyl-tRNA synthetase (AlaRS) on its alanylation activity. We prepared an AlaRS K73Ac variant in which Nε-acetyl-l-lysine was incorporated at position 73 using an expanded genetic code system in E. coli. The AlaRS K73Ac variant showed low activity compared to the AlaRS wild type (WT). Nicotinamide treatment or CobB-deletion in an E. coli led to elevated acetylation levels of AlaRS K73Ac and strongly reduced alanylation activities. We assumed that alanylation by AlaRS is affected by K73 acetylation, and the modification is sensitive to CobB deacetylase in vivo. We also showed that E. coli expresses two CobB isoforms (CobB-L and CobB-S) in vivo. CobB-S displayed the deacetylase activity of the AlaRS K73Ac variant in vitro. Our results imply a potential regulatory role for lysine acetylation in controlling the activity of aaRSs and protein synthesis.


Introduction
The recent development of mass spectrometry (MS)-based proteomic analysis enables the discovery of a wide variety of posttranslational modifications (PTMs) in prokaryotes [1]. Acetylation of lysine residues in proteins is distributed across all domains of life and is thought to be functionally important. Proteomic studies of acetylated protein (acetylome) indicated a large number of acetylated proteins in bacteria and archaea [2][3][4][5][6][7]. Since they also possess homologs of eukaryotic lysine acetyltransferases (KATs) and deacetylases (KDACs) [8], acetylation in bacteria and archaea are expected to be involved in the regulation of cellular processes as in eukaryotes. Recently, there has been growing interest in the role of acetylation [9][10][11][12][13][14]; however, the regulatory mechanisms and functional implications of prokaryotic acetylation are still largely unknown.
For purification of the N-terminal His-tagged CobB protein, cobB-L and cobB-S ORFs were inserted between the NdeI and BamHI site of pET15b (Novagen, Madison, WI, USA), resulting in pET15b-cobB-L and pET15b-cobB-S, respectively.

Construction of CobB-Deleted Escherichia coli Strain
To construct the CobB-deleted E. coli (∆CobB) strain, the cobB gene was removed from the E. coli DH10B genome by using the λ-red recombination system [26,27]. The kanamycin resistant gene was next removed by FLP-FRT recombination. The gene deletion was verified by PCR.
For preparation of C-terminal His-tagged AcK-incorporated AlaRS (AlaRS K73Ac), E. coli DH10B or its ∆CobB strain co-transformed with pKTS-AcKRS1-PylT and pT5C-Ec-alaS-K74amb were grown in LB medium supplemented with 25 µg/mL kanamycin and 34 µg/mL chloramphenicol at 37 • C. Twenty millimolar of nicotinamide (NAM; Sigma-Aldrich, St. Louis, MO, USA), which is an inhibitor of CobB deacetylase, was added if necessary. When the OD 600 of the culture reached 0.5, IPTG and AcK (Sigma-Aldrich) were added to a final concentration of 0.2 mM and 2 mM, respectively, and the cultivation was continued for 8 h. His-tag affinity purification was performed as described above.
For preparation of N-terminal His-tagged CobB protein used for the deacetylation assay, E. coli BL21(DE3) (Nippon Gene, Tokyo, Japan) transformed with pET15b-cobB-S was grown in LB medium supplemented with 100 µg/mL ampicillin at 37 • C. When the OD 600 of the culture reached 0.6, IPTG was added to a final concentration of 0.2 mM. The cells were continuously cultivated for 4 h and then harvested. His-tag affinity purification was performed as described above. The purified enzyme was dialyzed in 1 L of dialysis buffer (20 mM Tris-HCl pH 8.0 and 100 mM NaCl) twice and concentrated by Amicon Ultra Ultracel-10K (Merck Millipore). Finally, the enzyme was stored in 10 mM Tris-HCl pH 8.0, 50 mM NaCl, 2 mM DTT, and 50% glycerol.

Western Blotting
After SDS-PAGE, proteins were electroblotted onto a Hybond-ECL membrane (GE Healthcare, Buckinghamshire, UK), which was blocked in Tris-buffered saline containing 0.1% Tween-20 (TBS-T) and 5% skimmed milk. Proteins on the membrane were probed with antibody in blocking buffer and then the membrane was washed with TBS-T. Rabbit polyclonal anti-AcK antibody (ImmuneChem, Burnaby, BC, Canada) and Anti-His-tag mAb-HRP-DirecT (Medical & Biological Laboratories, Aichi, Japan) were used at a 1/3000 and 1/10,000 dilution respectively as primary antibodies. Anti-Rabbit IgG HRP conjugate (Promega, Madison, WI, USA) was used as the secondary antibody. Proteins on the membrane were probed with the primary antibody in blocking buffer and then washed with TBS-T. Proteins were detected by Immobilon Western Chemilum HRP substrate (Merck Millipore) and visualized on a LAS 4000 (GE Healthcare).

Alanylation Assay
Alanine tRNA (tRNA Ala ) from E. coli was transcribed in vitro by T7 RNA polymerase and purified by 12% polyacrylamide gel with 7 M urea. The alanylation assay was performed at 37 • C in a reaction mixture (50 µL) containing 50 mM HEPES-NaOH pH 7.4, 10 mM MgCl 2 , 30 mM KCl, 2 mM ATP, 10 µM [U-14 C] Ala (164 mCi/mmol; GE Healthcare), 5 µM transcribed tRNA Ala , and 5 ng/µL (50 nM) His-tagged AlaRS or its variants. At time points of 2 min, 4 min, 8 min, and 10 min, 10 µL aliquots of the reaction mixture were spotted onto Whatman 3MM filter paper (GE Healthcare) and the reaction was immediately quenched with 5% trichloroacetic acid (TCA). After washing the filter paper with 5% TCA and drying, radioactivity on the filter paper was counted by a Beckman LS 6500 Scintillation Counter (Beckman Coulter, Brea, CA, USA).

Circular Dichroism Spectrometry Analysis
The circular dichroism (CD) spectra of AlaRS and its derivatives were recorded on a J-805 CD Spectrometer (JASCO Corporation, Tokyo, Japan). AlaRS WT, K73A, and K73Ac (NAM+) were respectively prepared at a concentration of 0.1 mg/mL (1.0 µM) in a buffer containing 5 mM Tris-HCl pH 7.5 and 5 mM MgCl 2 , and scanned from 195 nm to 265 nm with a 50 nm/min speed.

Deacetylation Assay
The assay was performed at 37 • C for 12 h in a reaction mixture (30 µL) containing 25 mM Tris-HCl pH 8.0, 100 mM NaCl, 27 mM KCl, 0.5 mM NAD + (Nacalai, Kyoto, Japan), 0.5 µM AlaRS K73Ac (∆CobB), and 1.5 or 5 µM His-tagged CobB-S. The reaction was stopped by adding an equal volume of 2× sample buffer for SDS-PAGE. The resulting samples containing 0.2 pmol of AlaRS were separated by 7.5% SDS-PAGE and acetylation levels of the AlaRS K73Ac (∆CobB) were determined by Western blotting analysis using anti-AcK antibody.

Preparation of AlaRS K73Ac Variant by Expanded Genetic Code
The previously reported E. coli acetylome studies identified 15 lysine acetylation sites in AlaRS (Table A1). Of them, K73 is an essential residue in motif II of the enzyme active site (Figure 1a) [28] and interacts with the 3 -end of the cognate tRNA Ala [29,30]. In addition, this residue is well-conserved among several bacteria and eukaryotes (Figure 1b). We therefore investigated the impact of acetylation at K73 of AlaRS using the genetic incorporation of AcK into position 73 of AlaRS. The expanded genetic code system is a technology to reassign non-canonical amino acids to one of the stop codons (usually the amber codon) by pairs of engineered aaRS and its cognate suppressor tRNA [31]. One such pair, AcKRS/tRNA Pyl from M. mazei, can co-translationally introduce AcK at the amber codon, resulting in production of a protein with AcK at the specified sites [25]. Since the protein homogenously prepared by this system contains AcK, the protein with AcK is suitable to examine the impact of lysine acetylation.

Preparation of AlaRS K73Ac Variant by Expanded Genetic Code
The previously reported E. coli acetylome studies identified 15 lysine acetylation sites in AlaRS (Table A1). Of them, K73 is an essential residue in motif II of the enzyme active site (Figure 1a) [28] and interacts with the 3′-end of the cognate tRNA Ala [29,30]. In addition, this residue is well-conserved among several bacteria and eukaryotes (Figure 1b). We therefore investigated the impact of acetylation at K73 of AlaRS using the genetic incorporation of AcK into position 73 of AlaRS. The expanded genetic code system is a technology to reassign non-canonical amino acids to one of the stop codons (usually the amber codon) by pairs of engineered aaRS and its cognate suppressor tRNA [31]. One such pair, AcKRS/tRNA Pyl from M. mazei, can co-translationally introduce AcK at the amber codon, resulting in production of a protein with AcK at the specified sites [25]. Since the protein homogenously prepared by this system contains AcK, the protein with AcK is suitable to examine the impact of lysine acetylation. We prepared the AlaRS K73Ac variant using the AcKRS/tRNA Pyl system and verified AcKincorporation in AlaRS by Western blotting with the anti-AcK antibody (Figure 2a). Specific incorporation of AcK into position 73 of AlaRS was also verified by MALDI-TOF/TOF analysis ( Figure A2). It has been reported that CobB is responsible for deacetylation against many proteins in E. coli [18]. In order to know whether acetylation of K73 was reversed by CobB, we examined the acetylation levels of AlaRS K73Ac purified from NAM-treated (NAM+) and untreated (NAM−) E. coli We prepared the AlaRS K73Ac variant using the AcKRS/tRNA Pyl system and verified AcKincorporation in AlaRS by Western blotting with the anti-AcK antibody (Figure 2a). Specific incorporation of AcK into position 73 of AlaRS was also verified by MALDI-TOF/TOF analysis ( Figure A2). It has been reported that CobB is responsible for deacetylation against many proteins in E. coli [18]. In order to know whether acetylation of K73 was reversed by CobB, we examined the acetylation levels of AlaRS K73Ac purified from NAM-treated (NAM+) and untreated (NAM−) E. coli DH10B. NAM is an inhibitor of sirtuin-type lysine deacetylases including CobB. AlaRS K73Ac (NAM+) showed higher acetylation signals than AlaRS K73Ac (NAM−), although AlaRS WT and its K73A mutant as negative controls did not give a signal of acetylation. The observation that acetylation of the AlaRS WT and K73A mutant could not be detected by Western blotting is thought to be due to the amount of the overexpressed enzymes being excess to a capability of endogenous lysine acetylation, resulting in reducing populations of acetylated AlaRS. We next examined the acetylation level of K73Ac purified from the ∆CobB strain. The acetylation level of K73Ac (∆CobB) was almost the same as that of K73Ac (NAM+). Quantification of the band intensity showed that the acetylation levels of K73Ac (NAM+) and K73Ac (∆CobB) were 3.3-fold and 4.0-fold higher than that of K73Ac (NAM−), respectively ( Figure 2b). These results indicated that NAM treatment and CobB deletion enhanced the acetylation of AlaRS, suggesting that CobB can deacetylate K73Ac in vivo. It should be noted that lysine acetylation was not detected in the AlaRS WT and K73A mutant purified from the ∆CobB strain ( Figure 2c). Taken together, these results support that the acetylation signal of the Western blot represented the K73 acetylation and CobB was responsible for the decreased acetylation of K73Ac.
Genes 2018, 9, x FOR PEER REVIEW 6 of 16 DH10B. NAM is an inhibitor of sirtuin-type lysine deacetylases including CobB. AlaRS K73Ac (NAM+) showed higher acetylation signals than AlaRS K73Ac (NAM−), although AlaRS WT and its K73A mutant as negative controls did not give a signal of acetylation. The observation that acetylation of the AlaRS WT and K73A mutant could not be detected by Western blotting is thought to be due to the amount of the overexpressed enzymes being excess to a capability of endogenous lysine acetylation, resulting in reducing populations of acetylated AlaRS. We next examined the acetylation level of K73Ac purified from the ∆CobB strain. The acetylation level of K73Ac (∆CobB) was almost the same as that of K73Ac (NAM+). Quantification of the band intensity showed that the acetylation levels of K73Ac (NAM+) and K73Ac (∆CobB) were 3.3-fold and 4.0-fold higher than that of K73Ac (NAM−), respectively (Figure 2b). These results indicated that NAM treatment and CobB deletion enhanced the acetylation of AlaRS, suggesting that CobB can deacetylate K73Ac in vivo. It should be noted that lysine acetylation was not detected in the AlaRS WT and K73A mutant purified from the ∆CobB strain ( Figure 2c). Taken together, these results support that the acetylation signal of the Western blot represented the K73 acetylation and CobB was responsible for the decreased acetylation of K73Ac.

Alanylation Activity of AlaRS K73Ac Variants
Given that the K73 in AlaRS was important for the alanylation activity [29,30], we inferred that acetylation of the residue should influence the activity and consequently downregulate protein translation. To examine the activity of a series of AlaRS K73 variants, we performed alanylation assays in vitro ( Figure 3). We confirmed that the K73A mutation abolished the alanylation activity of AlaRS, as reported previously [30]. AlaRS K73Ac (NAM−) showed decreased alanylation activity compared to the AlaRS WT. The alanylation activities of K73Ac (∆CobB) and K73Ac (NAM+) variants, which showed higher levels of acetylation than K73Ac (NAM−), were more impaired. Considering that the acetylation signals reflect only K73 acetylation, these results indicated that K73 acetylation inhibits the alanylation activity of AlaRS.

Alanylation Activity of AlaRS K73Ac Variants
Given that the K73 in AlaRS was important for the alanylation activity [29,30], we inferred that acetylation of the residue should influence the activity and consequently downregulate protein translation. To examine the activity of a series of AlaRS K73 variants, we performed alanylation assays in vitro (Figure 3). We confirmed that the K73A mutation abolished the alanylation activity of AlaRS, as reported previously [30]. AlaRS K73Ac (NAM−) showed decreased alanylation activity compared to the AlaRS WT. The alanylation activities of K73Ac (∆CobB) and K73Ac (NAM+) variants, which showed higher levels of acetylation than K73Ac (NAM−), were more impaired. Considering that the acetylation signals reflect only K73 acetylation, these results indicated that K73 acetylation inhibits the alanylation activity of AlaRS.

Circular Dichroism Spectrum of AlaRS K73Ac Variants
To exclude the possibility of a structural defect caused by genetic incorporation of AcK, we compared CD spectra of AlaRS WT, K73A and K73Ac (NAM+). Almost the same spectral curves were observed among the three proteins ( Figure 4). The proportion of secondary structures contained in those proteins was calculated by the CAPITO web server [34]. WT, K73A and K73Ac (NAM+) contained 21%, 13%, and 22% α-helix, and 22%, 24%, and 29% β-sheet, respectively. This suggested that incorporation of AcK did not impair the protein structure of AlaRS.

Escherichia coli Expresses Two CobB Isoforms with Each Promoter
It has been reported that there are long and short isoforms of CobB (CobB-L and CobB-S, respectively) in Salmonella enterica, both of which are biologically active in the cell [35]. The existence of the two CobB isoforms has not been experimentally confirmed in E. coli yet. The cobB ORF in the public E. coli database encodes a 279-amino acid isoform (CobB-L), while a shorter 242-amino acid isoform (CobB-S) can be translated from the 112 nt-downstream second Met codon in the ORF. CobB-L contains a 37-amino acid N-terminal extension comprised of 15 basic and 13 hydrophobic residues, similar to CobB-L of S. enterica (Figure 5a). We expressed a C-terminal His-tagged CobB ORF with different

Circular Dichroism Spectrum of AlaRS K73Ac Variants
To exclude the possibility of a structural defect caused by genetic incorporation of AcK, we compared CD spectra of AlaRS WT, K73A and K73Ac (NAM+). Almost the same spectral curves were observed among the three proteins ( Figure 4). The proportion of secondary structures contained in those proteins was calculated by the CAPITO web server [34]. WT, K73A and K73Ac (NAM+) contained 21%, 13%, and 22% α-helix, and 22%, 24%, and 29% β-sheet, respectively. This suggested that incorporation of AcK did not impair the protein structure of AlaRS.

Circular Dichroism Spectrum of AlaRS K73Ac Variants
To exclude the possibility of a structural defect caused by genetic incorporation of AcK, we compared CD spectra of AlaRS WT, K73A and K73Ac (NAM+). Almost the same spectral curves were observed among the three proteins ( Figure 4). The proportion of secondary structures contained in those proteins was calculated by the CAPITO web server [34]. WT, K73A and K73Ac (NAM+) contained 21%, 13%, and 22% α-helix, and 22%, 24%, and 29% β-sheet, respectively. This suggested that incorporation of AcK did not impair the protein structure of AlaRS.

Escherichia coli Expresses Two CobB Isoforms with Each Promoter
It has been reported that there are long and short isoforms of CobB (CobB-L and CobB-S, respectively) in Salmonella enterica, both of which are biologically active in the cell [35]. The existence of the two CobB isoforms has not been experimentally confirmed in E. coli yet. The cobB ORF in the public E. coli database encodes a 279-amino acid isoform (CobB-L), while a shorter 242-amino acid isoform (CobB-S) can be translated from the 112 nt-downstream second Met codon in the ORF. CobB-L contains a 37-amino acid N-terminal extension comprised of 15 basic and 13 hydrophobic residues, similar to CobB-L of S. enterica (Figure 5a). We expressed a C-terminal His-tagged CobB ORF with different

Escherichia coli Expresses Two CobB Isoforms with Each Promoter
It has been reported that there are long and short isoforms of CobB (CobB-L and CobB-S, respectively) in Salmonella enterica, both of which are biologically active in the cell [35]. The existence of the two CobB isoforms has not been experimentally confirmed in E. coli yet. The cobB ORF in the public E. coli database encodes a 279-amino acid isoform (CobB-L), while a shorter 242-amino acid isoform (CobB-S) can be translated from the 112 nt-downstream second Met codon in the ORF. CobB-L contains a 37-amino acid N-terminal extension comprised of 15 basic and 13 hydrophobic residues, similar to CobB-L of S. enterica (Figure 5a). We expressed a C-terminal His-tagged CobB ORF with different upstream regions from pACYC184P plasmids in E. coli DH10B (Figure 5b). We detected two His-tagged protein bands by Western blotting, which corresponded to CobB-L and CobB-S from pACYC184P-1218-cobB and pACYC184P-417-cobB plasmids (Figure 5c). Meanwhile, CobB-S was only expressed from pACYC184P-30-cobB and pACYC184P-cobB-L. Neither form of CobB was detected with pACYC184P-cobB-S. These results indicated that the promoter for CobB-L expression existed between −417 and −30 within the nagK ORF and that for CobB-S expression existed between +1 and +111 within the cobB ORF. Using the online BPROM web tool (Softberry, Inc., Mount Kisco, NY, USA), we detected the candidates of the −10 and −35 consensus sequences between −347 and −319 for CobB-L and between +51 and +83 for CobB-S expression, respectively (Figure 5d). These results indicated that two isoforms of CobB-L and CobB-S exist in E. coli and they are directed from each promoter. upstream regions from pACYC184P plasmids in E. coli DH10B (Figure 5b). We detected two His-tagged protein bands by Western blotting, which corresponded to CobB-L and CobB-S from pACYC184P-1218-cobB and pACYC184P-417-cobB plasmids (Figure 5c). Meanwhile, CobB-S was only expressed from pACYC184P-30-cobB and pACYC184P-cobB-L. Neither form of CobB was detected with pACYC184P-cobB-S. These results indicated that the promoter for CobB-L expression existed between −417 and −30 within the nagK ORF and that for CobB-S expression existed between +1 and +111 within the cobB ORF.
Using the online BPROM web tool (Softberry, Inc., Mount Kisco, NY, USA), we detected the candidates of the −10 and −35 consensus sequences between −347 and −319 for CobB-L and between +51 and +83 for CobB-S expression, respectively (Figure 5d). These results indicated that two isoforms of CobB-L and CobB-S exist in E. coli and they are directed from each promoter.

Deacetylation of K73 by CobB-S In Vitro
We attempted to produce N-terminal His-tagged CobB isoform proteins for deacetylation assays in vitro, but CobB-L could not be successfully obtained as soluble protein after several attempts. CobB-S contains the conserved Sir2 catalytic core and has been shown to have the deacetylase activity for a histone H4 acetylated peptide [36]. We therefore performed an in vitro deacetylation assay with CobB-S. After incubation of the His-tagged CobB-S and AlaRS K73Ac (∆CobB) for 12 h, the residual acetylation was determined by Western blotting (Figure 6a). An NAD + -dependent decrease in acetylation was observed and it was enhanced by increasing concentrations of CobB-S; 27% and 39% decreases in acetylation were observed when 1.5 µM or 5.0 µM of CobB-S were used, respectively (Figure 6b). The deacetylase activity of CobB-S for AlaRS K73Ac seemed to be moderate, suggesting that CobB may require additional factors for full activity.

Deacetylation of K73 by CobB-S In Vitro
We attempted to produce N-terminal His-tagged CobB isoform proteins for deacetylation assays in vitro, but CobB-L could not be successfully obtained as soluble protein after several attempts. CobB-S contains the conserved Sir2 catalytic core and has been shown to have the deacetylase activity for a histone H4 acetylated peptide [36]. We therefore performed an in vitro deacetylation assay with CobB-S. After incubation of the His-tagged CobB-S and AlaRS K73Ac (∆CobB) for 12 h, the residual acetylation was determined by Western blotting (Figure 6a). An NAD + -dependent decrease in acetylation was observed and it was enhanced by increasing concentrations of CobB-S; 27% and 39% decreases in acetylation were observed when 1.5 µM or 5.0 µM of CobB-S were used, respectively (Figure 6b). The deacetylase activity of CobB-S for AlaRS K73Ac seemed to be moderate, suggesting that CobB may require additional factors for full activity.

Discussion
Serine phosphorylation in glutamyl-tRNA synthetase [37], lysine succinylation in methionyl-tRNA synthetase (MetRS) [38], and lysine acetylation in LeuRS, ArgRS [22] and TyrRS [23] have been reported as meaningful PTMs toward the aaRS function in E. coli. These modifications, with the exception of lysine succinylation in MetRS, downregulate their tRNA synthetase activities. The lysine succinylation of MetRS decreases the discrimination of cognate tRNA under stress conditions [38]. In this study, we newly found K73 acetylation that downregulates the activity of E. coli AlaRS which belongs to class II aaRSs.
K73 in AlaRS is known to be an essential residue that interacts with the 3′-end of tRNA Ala [29]. Substitution of this residue to glutamine, asparagine, alanine or glutamate significantly reduces the alanylation activity but not the alanine activation, suggesting that the positive charge of the lysine plays an important role in the interaction [30]. The recently described crystal structure of AlaRS with the cognate tRNA Ala from Archaeoglobus fulgidus elucidated the recognition mechanism of tRNA Ala by AlaRS [39]. Although A. fulgidus AlaRS lacks the corresponding lysine residue in the primary sequence, the cytidine-75 in the tRNA Ala of the complex is located in the vicinity of K73 of the E. coli enzyme when its structure is superposed on the complex structure. Based on the above-mentioned evidence, our results suggest that downregulation of AlaRS activity by K73 acetylation is due to the

Discussion
Serine phosphorylation in glutamyl-tRNA synthetase [37], lysine succinylation in methionyl-tRNA synthetase (MetRS) [38], and lysine acetylation in LeuRS, ArgRS [22] and TyrRS [23] have been reported as meaningful PTMs toward the aaRS function in E. coli. These modifications, with the exception of lysine succinylation in MetRS, downregulate their tRNA synthetase activities. The lysine succinylation of MetRS decreases the discrimination of cognate tRNA under stress conditions [38]. In this study, we newly found K73 acetylation that downregulates the activity of E. coli AlaRS which belongs to class II aaRSs.
K73 in AlaRS is known to be an essential residue that interacts with the 3 -end of tRNA Ala [29]. Substitution of this residue to glutamine, asparagine, alanine or glutamate significantly reduces the alanylation activity but not the alanine activation, suggesting that the positive charge of the lysine plays an important role in the interaction [30]. The recently described crystal structure of AlaRS with the cognate tRNA Ala from Archaeoglobus fulgidus elucidated the recognition mechanism of tRNA Ala by AlaRS [39]. Although A. fulgidus AlaRS lacks the corresponding lysine residue in the primary sequence, the cytidine-75 in the tRNA Ala of the complex is located in the vicinity of K73 of the E. coli enzyme when its structure is superposed on the complex structure. Based on the above-mentioned evidence, our results suggest that downregulation of AlaRS activity by K73 acetylation is due to the loss of the positive charge in the lysine residue, resulting in interference of the interaction between AlaRS and tRNA Ala .
Our in vivo and in vitro studies indicate that the acetylation of K73 is reversed by CobB deacetylase. Systematic analysis to search for substrates of CobB has shown that CobB does not recognize a specific sequence, but tends to prefer aspartate (D), glutamate (E), alanine (A), glycine (G) and tyrosine (Y) on β-sheet or loop structures adjacent to an acetylated lysine [18]. The K73 exists in the AGGKHND sequence of the β-turn, which contains the favored amino acids of CobB (shown with underline). This observation supports our result that the acetylated K73 is a substrate of CobB. Our in vitro study showed the deacetylase activity of CobB-S for AlaRS K73Ac, but it seemed to be moderate. We consider the possibility that CobB-L may prefer to remove this acetylation or another factor like a chaperone might be required for the reaction.
We clarified the CobB-dependent deacetylation mechanism of K73 in AlaRS in vivo, while the mechanism of K73 acetylation remains elusive. The K73 acetylation has been detected at the stationary phase of growth without glucose [7,16] and at the early stationary phase when glucose is present [17]. In both cases, the intracellular level of acetyl-phosphate is likely increased by carbon overflow in the central metabolism pathway [16,40]. Thus, the acetylation of K73 is thought to be introduced by a non-enzymatic mechanism with acetyl-phosphate. This K73 acetylation has also been detected in Vibrio parahaemolyticus, mice, and humans [41,42]. The conserved K73 acetylation among different organisms may be of physiological significance, although it remains to be experimentally determined in other organisms.
Promoter analysis for CobB expression in E. coli was previously performed using an enhanced green fluorescent protein (EGFP)-based reporter system and real time quantitative PCR (RT-qPCR), showing that promoter activity for a nagK-cobB transcription was detected in the 5 -upstream region (−300 to +50) of nagK but no significant activity was detected in the upstream region (−400 to +151) of cobB [43]. Contrary to a previous report, we could find two promoter activities within the region of −417 to +111 of cobB: the one within nagK directs CobB-L expression and the other within cobB directs CobB-S expression (Figure 5d). These findings are similar to the case of S. enterica, in which the upstream and downstream promoters correspond to P2 and P3 promoters, respectively [35]. Since CobB-S contains the conserved catalytic core which is sufficient for the deacetylase activity [35,36], the extended N-terminal region composed of positively charged and hydrophobic amino acids may add an alternative function to CobB-L. It was proposed in S. enterica that the extended N-terminal region may be involved in interaction with other molecules such as nucleic acids and acetylated protein substrates, and cellular localization [35]. Our failure to purify N-terminal His-tagged CobB-L might be due to masking of the N-terminal His-tag by unknown interacting partners. Elucidation of the physiological significance of CobB-L and CobB-S is an intriguing challenge for the future studies.
As aaRSs are essential enzymes, severe inhibition of their activities causes growth retardation and cell death [44]. Therefore, it is thought that downregulation of AlaRS activity by K73 acetylation is likely transient and AlaRS should be rapidly reactivated by deacetylation when growth conditions are improved. Hence, our findings will provide insight into a new regulatory mechanism of translation by post-translational lysine acetylation of aaRSs. Future studies will shed light on the acetylation introduced into other aaRSs for a more comprehensive understanding of translation regulation.
Author Contributions: T.U. conceived and designed research; T.U. and S.K. performed research; T.U., S.K., D.S. and K.T. analyzed data; T.U. and S.K. wrote the paper; D.S. and K.T. supervised the work.   Table A3. List of plasmids in this study. nts, nucleotides.