Acetylation of lysine ϵ-amino groups regulates aminoacyl-tRNA synthetase activity in Escherichia coli

Previous proteomic analyses have shown that aminoacyl-tRNA synthetases in many organisms can be modified by acetylation of Lys. In this present study, leucyl-tRNA synthetase and arginyl-tRNA synthetase from Escherichia coli (EcLeuRS and EcArgRS) were overexpressed and purified and found to be acetylated on Lys residues by MS. Gln scanning mutagenesis revealed that Lys619, Lys624, and Lys809 in EcLeuRS and Lys126 and Lys408 in EcArgRS might play important roles in enzyme activity. Furthermore, we utilized a novel protein expression system to obtain enzymes harboring acetylated Lys at specific sites and investigated their catalytic activity. Acetylation of these Lys residues could affect their aminoacylation activity by influencing amino acid activation and/or the affinity for tRNA. In vitro assays showed that acetyl-phosphate nonenzymatically acetylates EcLeuRS and EcArgRS and suggested that the sirtuin class deacetylase CobB might regulate acetylation of these two enzymes. These findings imply a potential regulatory role for Lys acetylation in controlling the activity of aminoacyl-tRNA synthetases and thus protein synthesis.

Aminoacyl-tRNA synthetase (aaRS) 2 catalyzes esterification between its cognate amino acid and tRNA to produce aminoacyl-tRNA (aa-tRNA) in the initiation step of translation. A high level of accuracy is essential during aminoacylation to ensure quality control during protein synthesis. Disruption of translational fidelity can lead to mistranslation, with profound consequences for both prokaryotic and eukaryotic cells (1)(2)(3).
The 20 aaRSs can be divided into two classes, each with 10 members, based on sequence identity and characteristic structural motifs (4). Class I members have two signature peptides (HIGH and KMSK) located in the active site that form a characteristic dinucleotide binding fold (Rossmann fold, ␤-␣-␤-␣-␤). Both leucyl-and arginyl-tRNA synthetases (LeuRS and ArgRS) belong to class I aaRSs (5). Like 16 other aaRSs, LeuRS catalyzes aminoacylation of its cognate tRNA in a two-step reaction: (a) activation of the amino acid with ATP and formation of an aminoacyl adenylate and (b) transfer of the aminoacyl moiety from the aminoacyl adenylate to the cognate tRNA substrate (5,6). However, ArgRS, together with glutamyl-tRNA synthetase and glutaminyl-tRNA synthetase, requires the presence of the cognate tRNA for amino acid activation (7,8).
LeuRS consists of a Rossmann fold domain for aminoacylation, a helix bundle domain for binding the tRNA anticodon, a connective peptide 1 (CP1) domain for editing mischarged tRNA, a ZN1 module, a leucine-specific domain, and a C-terminal domain (CTD) for tRNA binding (9,10). The aminoacylation and editing mechanisms of LeuRS from various species have been thoroughly investigated (11)(12)(13). ArgRS can be divided into five domains: an N-terminal additional domain (Add1) for tRNA D-loop recognition, a catalytic Rossmann fold domain, two domains (Ins-1 and Ins-2) inserted in the active site, and a C-terminal additional domain (Add2) that participates in the binding of the tRNA anticodon (14,15). It is peculiar that the Add1 domain is conserved in ArgRS but not in other class I aaRSs. In most species, ArgRS lacks a canonical KMSK sequence, and a conserved lysine (Lys) upstream of the HIGH sequence motif in these enzymes stabilizes the transition state of the amino acid activation reaction (Arg-AMP formation) to compensate for the loss of the second Lys (K2) in the KMSK motif (16,17).
Cells are constantly faced with the challenge of changing environmental conditions, and post-translational modification (PTM) is one method of dealing with this challenge. PTM can expand the genetic lexicon by endowing proteins with diversity beyond that can be achieved with the canonical 20 proteinogenic amino acids. Acetylation of the ␣-amino group of the N-terminal amino acid (irreversible) or the ⑀-amino group of internal Lys residues (reversible) is one type of PTM. In general, acetylation of Lys refers to reversible acetylation, and it can regulate fundamental cellular processes such as transcription, translation, pathways associated with central metabolism, and stress responses (18). Although the essential regulatory role of Lys acetylation in eukaryotes is widely accepted and relatively well understood, its function in bacteria and archaea remains more obscure (19,20).
In Escherichia coli, although several putative protein acetylases are present in the genome, Gcn5-like YfiQ, which is highly similar to the acetyltransferase Pat in Salmonella enterica, is the only confirmed acetyltransferase to date (21)(22)(23), and CobB is the predominate deacetylase, which belongs the NAD ϩ -dependent sirtuin family (22)(23)(24). Recently, the serine hydrolase YcgC was identified as a Zn 2ϩ -and NAD ϩ -independent deacetylase that regulates a distinct set of substrates from CobB (25). One well studied target of protein acetylation is acetyl-CoA synthetase (Acs), which activates acetate to the high-energy intermediate acetyl coenzyme A (Ac-CoA; acetate ϩ ATP ϩ CoA 3 AMP ϩ PP i ϩ Ac-CoA). Reversible acetylation of a catalytic core Lys residue conserved in Acs enzymes from bacteria to human (Lys 609 in S. enterica Acs) could regulate enzyme activity, because it blocks ATP-dependent adenylation of acetate, preventing the formation of acetyl-AMP and the subsequent production of Ac-CoA (18,21). In S. enterica, Acs Lys 609 is regulated by a protein acetylation/deacetylation system that includes Pat and CobB (21), and this system also coordinates carbon source utilization and metabolic flux by controlling the acetylation of metabolic enzymes (26). Interestingly, acetylation was recently found to be mediated nonenzymatically in mitochondria of both prokaryotes and eukaryotes (20,23,(27)(28)(29)(30)(31). In E. coli, the majority of acetylation occurs independently of YfiQ, and the glycolysis intermediate acetyl-phosphate (AcP) is associated with a global shift in protein acetylation, whereas CobB regulates a subset of these chemical acetylation events (20,23,28). Some aaRSs are modified by phosphorylation, which influences multidrug tolerance in E. coli and the reactive oxygen species defense mechanism in mammalian cells (32,33). Despite growing knowledge, studies focusing on other forms of PTM of aaRSs are few in number. A large-scale proteomic survey demonstrated that some aaRSs from E. coli, S. enterica, Bacillus subtilis, Drosophila melanogaster, Mus musculus, Rattus norvegicus, and Homo sapiens are acetylated, and bioinformatics and network analysis of acetylation sites found aa-tRNA biosynthesis pathway enriched in some species (23, 26, 31, 34 -42). Furthermore, some of the identified acetylated Lys res-idues are conserved, and it would be intriguing to decipher the exact role of acetylation of aaRSs.
Herein, we discovered that EcLeuRS and EcArgRS are modified by acetylation of Lys residues in vivo. By utilizing an engineered Methanosarcina barkeri pyrrolysyl-tRNA synthetase (MbPylRS)/MbtRNA CUA pair (N ⑀ -acetyllysyl-tRNA synthetase/tRNA CUA pair, pAcKRS) system (43), we obtained enzymes harboring acetylation at specific sites and investigated the biochemical properties of EcLeuRS-K Ac s and EcArgRS-K Ac s. We also examined the molecular mechanism controlling regulation of the acetylation of EcLeuRS and EcArgRS and identified CobB and AcP as possible regulatory factors. Acetylation appears to be a mechanism for adjusting the activity of aaRSs and thereby controlling protein synthesis.

MS revealed acetylation at 11 Lys residues in EcLeuRS
Previous studies demonstrated that LeuRS from E. coli, Saccharomyces cerevisiae, R. norvegicus, H. sapiens, and other species can be acetylated (38,40,41,44). To identify the acetylation sites in EcLeuRS, we overexpressed ecleuS that encodes EcLeuRS with a N-terminal His 6 tag in E. coli BL21 and purified the recombinant protein by Ni 2ϩ -NTA affinity chromatography. 11 Lys residues were detected to be acetylated in three independent MS analyses. They span the entire protein and include Lys 619 and Lys 624 in the KMSK signature sequence ( Figs. 1 and 2, A and B).
To understand the effect of acetylation on the aminoacylation activity of EcLeuRS, we separately mutated all 11 Lys residues to Gln, because this residue lacks a positive charge on the side chain and is a good mimic of acetylated Lys. We assayed the aminoacylation activity of the K-Q mutants and found that mutations at Lys 619 and Lys 624 in the amino acid activation active site and Lys 809 in the CTD displayed decreased aminoacylation activity compared with WT EcLeuRS, whereas the aminoacylation activity of mutants at all other sites was unchanged (Fig. 2C). Lys 402 is the only residue in the CP1 domain among these 11 residues, and the co-crystal structure of EcLeuRS with tRNA Leu and Leu (PDB number 4ARC) indicated that Lys 402 lies on the surface of the CP1 pocket and points away from the domain core, suggesting it is not likely to be essential for the editing function. Indeed, Ile-tRNA Leu deacylation assays showed that the post-transfer editing activity of the EcLeuRS-K402Q mutant remained unchanged compared with native EcLeuRS (data not shown).

Characterization of K Ac mutants reveals that Lys acetylation reduces EcLeuRS enzyme activity
To further explore the influence of acetylation on EcLeuRS, we used a previously described system to incorporate N ⑀ -acetyl-L-Lys (AcK) at specific sites to generate EcLeuRS-K Ac s in situ (Fig. 3) (43). We transformed E. coli BL21 cells with two plasmids: pAcKRS encoding the N ⑀ -acetyllysyl-tRNA synthetase/tRNA CUA pair that activates AcK and recognizes the UAG codon and another encoding EcLeuRS in which the Lys triplet codon was substituted with TAG.
Sequence alignment showed that Lys 619 and Lys 809 of EcLeuRS are conserved in LeuRSs from various species, whereas Lys 624 is basically conserved in prokaryotic LeuRSs (Fig. 4A). Following overexpression as described above, EcLeuRS-K619 Ac , EcLeuRS-K624 Ac , and EcLeuRS-K809 Ac were purified and confirmed to be 90% homogeneous by SDS-PAGE (data not shown). Western blotting confirmed the incorporation of AcK into EcLeuRS (Fig. 4B), and comparison of CD spectra of EcLeuRS-WT and EcLeuRS Ac s confirmed that EcLeuRS Ac s were properly folded (data not shown).
Lys 619 and Lys 624 are located in or downstream of the conserved KMSK loop, which, together with the HIGH motif, is essential for the amino acid activation activity. EcLeuRS-K619 Ac completely lost its Leu activation and leucylation activities ( Fig. 4, C and D, and Tables 1 and 2). The amino acid activation and aminoacylation activities of EcLeuRS-K624 Ac were also determined. Even though EcLeuRS-K624 Ac severely lost its activation activity ( Fig. 4C and Table 1), the catalytic efficiency (k cat /K m ) in aminoacylation was not that severely damaged ( Fig. 4D and Table 2). The total effect might be because tRNA charging is the rate-limiting step. K d values between EcLeuRSs and tRNA Leu calculated by fluorescence quenching showed that the binding affinity of EcLeuRS-K619 Ac and-K624 Ac with tRNA Leu was not altered compared with that of EcLeuRS-WT (Table 3). These results suggest acetylation of these two Lys residues (especially Lys 619 ) might lead to a conformational change in the synthetic active site pocket, decreasing the Leu activation and aminoacylation activities (Fig. 4E).
The flexibly linked CTD in LeuRS makes contacts with tertiary structural base pairs and the long variable arm of tRNA Leu (10,45). The ternary complex structure of EcLeuRS, tRNA Leu UAA , and Leu in the editing conformation (PDB number 4ARC) revealed that Lys 809 is located on the edge of one ␤-sheet in the CTD. The Leu activation activity of EcLeuRS- K809 Ac was not changed compared with that of WT EcLeuRS; consistent with that, Lys 809 is distant from the activation active site core region ( Fig. 4C and Table 1). The crystal structure of EcLeuRS (PDB number 4ARC) showed that the side chain of Lys 809 lies at a distance of 3.16 Å away from the phosphate group of the U47I ribose backbone of tRNA Leu (Fig. 4F). EcLeuRS-K809 Ac had a similar k cat toward EctRNA Leu as did EcLeuRS-WT (5.0 s Ϫ1 for EcLeuRS-K809 Ac , 5.7 s Ϫ1 for WT); nevertheless, the affinity for the cognate EctRNA Leu (K m , 6.7 M for EcLeuRS-K809 Ac and 1.2 M for WT) was decreased, and the catalytic efficiency (k cat /K m ) of EcLeuRS-K809 Ac (0.7 s Ϫ1 M Ϫ1 ) was only 15% that of EcLeuRS-WT (4.8 s Ϫ1 M Ϫ1 ; Table 2). In addition, the K d of EcLeuRS-K809 Ac with tRNA Leu was 1.4-fold that of the native enzyme, implying a decrease in binding affinity between LeuRS and tRNA Leu ( Table 3). The interaction between enzyme and tRNA Leu was partially disrupted by the acetylation of Lys 809 .
These results indicate that acetylation of Lys 619 , Lys 624 , and Lys 809 could potentially inhibit either amino acid activation or tRNA-charging activities of EcLeuRS. Among these residues, Lys 619 and Lys 809 in EcLeuRS are the residues whose acetylation lead to a sharp reduction of catalytic efficiency. In addition, the aminoacylation activity of EcLeuRS Ac s was comparable with the corresponding K-Q mutants (Figs. 2C and 4D), suggesting that Gln is a suitable mimic of acetylated Lys.

EcLeuRS is acetylated by AcP rather than YfiQ, and CobB can deacetylate EcLeuRS Ac
Determining the enzyme responsible for acetylating EcLeuRS is of particular interest. At present, the Gcn5-like acetyltransferase YfiQ is the only known enzyme that acetylates the ⑀-NH 2 group of Lys in E. coli (23). However, purified YfiQ was unable to transfer the acetyl group of Ac-CoA to EcLeuRS in vitro (Fig. 5A). The metabolism of AcP, a high-energy intermediate between acetate and Ac-CoA, has been shown to alter global acetylation levels in vivo, and AcP acetylates proteins nonenzymatically at multiple Lys residues in vitro (23). Given that AcP is a critical regulator of acetylation in bacteria, we incubated purified EcLeuRS with AcP and detected an increase in EcLeuRS acetylation (no acetylation signal is detected on WT enzyme before AcP treatment), implying a potential role for AcP in the acetylation of LeuRS (Fig. 5A). We refer to this AcPderived form of EcLeuRS as EcLeuRS Ac .
CobB, belonging to the sirtuin class, is the main deacetylase in E. coli (22)(23)(24). To determine whether CobB is involved in deacetylation of EcLeuRS Ac , deacetylation assays were performed with EcLeuRS Ac as the substrate for CobB. The results showed that CobB could deacetylate EcLeuRS Ac , and the presence of the CobB inhibitor NAM, or the absence of its cofactor, NAD ϩ , rendered CobB inactive (Fig. 5B). We also tested the activity of CobB on purified EcLeuRS-K619 Ac and EcLeuRS-K809 Ac , and Western blotting showed that CobB effectively decreased their acetylation in a time-dependent manner (Fig.  5C). Additionally, aminoacylation assays showed that after treatment with CobB, EcLeuRS-K619 Ac and EcLeuRS-K809 Ac recovered aminoacylation activity to some extent (Fig. 5D). The above results suggest CobB deacetylates EcLeuRS in vitro.

Acetylation of EcArgRS influences its catalytic rate
To investigate whether EcArgRS is regulated by acetylation, similar experiments to those described above were performed on EcArgRS. MS (repeated three times) detected acetylation at five Lys residues in EcArgRS including Lys 126 , which is upstream of the HIGH motif in the activation site (Figs. 6 and 7) (16, 17). As described above for EcLeuRS, we performed Glnscanning mutagenesis on these five Lys residues of EcArgRS and found that K-Q mutants of Lys 126 and Lys 408 displayed a decrease in aminoacylation activity (Fig. 7C). Sequence alignment showed that Lys 126 and Lys 408 are highly conserved among prokaryotic and eukaryotic ArgRSs (Fig. 8A). As described above for EcLeuRS, we utilized the pAcKRS system to generate EcArgRS-K Ac s, and CD spectra confirmed that their secondary structures were not altered by the point mutations (data not shown).
EcArgRS-K126 Ac lost Arg activation and arginylation activities and its affinity with tRNA Arg did not change compared with WT enzyme (Fig. 8, B and C; see Table 6). Lys 126 lies in the upstream of the HIGH (HVGH in EcArgRS) sequence within the catalytic pocket of EcArgRS, where it compensates for the lack of a canonical KMSK (especially the second Lys, K2) in ArgRS in most species and thus makes an important contribution to the aminoacylation reaction (Fig. 8D) (16, 17). Acetylation of Lys 126 in EcArgRS led to the complete loss of arginyla- . Taking EcLeuRS as an example, pAcKRS and pET22b(ϩ)-ecleuS were co-transformed in E. coli BL21 (DE3), and engineered pAcKRS were induced by the addition of Ara. Subsequently, IPTG was added to induce the production of EcLeuRS in the presence of NAM, an inhibitor of CobB. With the assistance of pAcKRS, EcLeuRS was translated in full-length form with incorporation of AcK or in truncated form (terminating at the Lys codon mutation site). All other experimental details were as described previously (43).
tion, consistent with previous experiments on TtArgRS-K116G that also showed a complete loss of activation and aminoacylation activities (17).
We next focused on the Lys 408 residue in the Add2 domain that is implicated in the binding of the tRNA Arg anticodon region (15,46). EcArgRS-K408 Ac displayed a low Arg activation activity ( Fig. 8B and Table 4) and aminoacylation activity ( Fig.   8C and Table 5), and kinetic parameters revealed a weaker affinity for EctRNA Arg (K m ϭ 9.0 M) compared with WT enzyme (K m ϭ 2.7 M). The k cat (6.9 s Ϫ1 ) was also decreased to 24% that of WT (28.4 s Ϫ1 ), and the overall effect was a decrease in the catalytic efficiency (k cat /K m , 0.8 s Ϫ1 M Ϫ1 ) to only 8% that    (Fig. 8E). In yeast and E. coli, C35 is one of the major identity elements of tRNA Arg (47,48). Acetylation of Lys 408 may distort the neighboring region that interacts with the tRNA Arg anticodon and consequently decrease the affinity of EcArgRS for tRNA Arg . Given that EcArgRS requires tRNA Arg as the activator during Arg activation, the decrease in the EcAr-gRS-K408 Ac activation activity could be partly due to a loss in affinity with tRNA Arg . Overall, the in vitro data show that the acetylation of EcArgRS (Lys 126 and Lys 408 ) could severely decrease its aminoacylation activity.

EcArgRS acetylation appears to be regulated by AcP and CobB
We also attempted to identify enzymes or other molecules involved in acetylation of EcArgRS. In vitro assays showed that AcP acetylated purified EcArgRS (Fig. 9A), but YfiQ did not (data not shown) (no acetylation signal is detected on WT enzyme before AcP treatment). Furthermore, in vitro CobB deacetylated EcArgRS Ac , the product of EcArgRS following treatment with AcP (Fig. 9B). Purified EcArgRS-K126 Ac and EcArgRS-K408 Ac variants were also deacetylated by CobB, and as described above, addition of NAM or the absence of NAD ϩ caused CobB to lose its deacetylation activity (Fig. 9C). AcP and CobB therefore appear to control acetylation and deacetylation of EcArgRS, consistent with the results discovered above for EcLeuRS.

Discussion
AaRSs are found to be acetylated Some aaRSs are post-translationally modified. In E. coli, tRNA Glu -bound glutamyl-tRNA synthetase can be phosphorylated at Ser 239 in the KMSK motif by the eukaryote-like serinethreonine kinase HipA. This PTM results in a loss of aminoacylation activity, which increases uncharged tRNA Glu loading at the A site of the ribosome, triggering (p)ppGpp formation and facilitating multidrug tolerance (32).
Acetylation regulates proteins involved in transcription, amino acid, nucleotide and protein biosynthesis, protein fold- All experiments were conducted at least twice. When quantifying of the relative amount of AcK signal/His signal, the sample without NAD ϩ and NAM was defined as 100%. ing, and detoxification responses in various species. We focused on aaRSs and found that many are acetylated on Lys residues located both on the surface and within the catalytic core or the tRNA binding domain in others' papers. In the present work, we confirmed the acetylation of two class Ia aaRSs in E. coli by three independent MS experiments. The residues found to be acetylated on these two aaRSs are not exactly the same as what were found in the studies of Weinert et al. (23) and Kuhn et al. (28). This might result from the use of different E. coli strains under various growth stages. Differences of nutrients in the media used might also influence the acetylation. In addition, we purified overexpressed proteins, rather than endogenous aaRSs. However, acetylation of some crucial residues that had been identified by the above authors (like Lys 619 and Lys 624 in EcLeuRS and Lys 126 and Lys 408 in EcArgRS) were consistently identified in our studies. We screened several Lys residues at which acetylation may negatively regulate the aminoacylation activity of enzyme by using a Gln scanning mutagenesis approach. Furthermore, we used the pAcKRS sys-tem to express and purify EcLeuRS and EcArgRS acetylated at specific sites (43) and characterized the effect of acetylation on the catalytic properties.

The catalytic activity of aaRSs is regulated by acetylation
Acetylation of Lys 619 and Lys 809 greatly impacted the aminoacylation activity of EcLeuRS, and both residues are highly conserved among LeuRS in various species (Fig. 4A). Lys 622 is the second Lys in the KMSK motif (K2), which is the key residue that directly interacts with Leu-AMP and believed to stabilize the negatively charged transition state of the first reaction step in class I aaRSs. Mutation of this residue led to a severe loss of enzymatic activity (49 -51). Acetylation of the adjacent residues, Lys 619 and Lys 624 caused significant inhibition of the first step of aminoacylation. In Bacillus stearothermophilus tyrosyl-tRNA synthetase (TyrRS), another class I aaRS, 230 KFGK 233 corresponds to the signature sequence KMSK. Previous data showed that the BsTyrRS-K230N mutant lost its activation activity, indicating that the first Lys in the KMSK is crucial to  the activation of tyrosine (49). Lys 619 is the first Lys in the signature sequence of EcLeuRS, and Lys 624 is adjacent to the signature sequence. Acetylation of the two Lys residues should inhibit the active site of amino acid activation and thus influence the activation and aminoacylation activities of EcLeuRS.
Hsu et al. (52) utilized Ala scanning mutagenesis to identify specific sites in the CTD that may be important for RNA-protein interactions and found that mutation of Lys 809 had a negligible effect on aminoacylation catalytic efficiency. By contrast,   nm EcArgRS-K408 Ac 9.0 Ϯ 0.6 6.9 Ϯ 0.5 0. 8 8 our results suggest acetylation of Lys 809 has a marked negative effect on the affinity for EctRNA Leu . Previously, relatively less attention has been paid to EcArgRS. Lys 126 , the residue upstream the signature sequence HIGH (HVGH), makes up for the absence of K2 in the KMSK motif in EcArgRS (17), and acetylation of key residues in the activation pocket leads to the complete loss of activation activity and consequent aminoacylation activity. Lys 408 (PhArgRS Lys 455 ) is in the vicinity of the tRNA Arg anticodon region that harbors the identity element C35 (PDB number 2ZUE), and acetylation of Lys 408 inhibited amino acid activation and transfer of the arginyl group from Arg-AMP to the 3Јend of tRNA Arg . These results suggest for the first time that acetylation regulates the amino acid activation and aminoacylation activities of EcArgRS and EcLeuRS.
Acetylation of EcArgRS and EcLeuRS was first reported in previous MS studies, including acetylation of Lys residues surrounding or within the conserved HIGH and KMSK motifs (23,40,42). We noticed that the conserved Lys residue upstream of the HIGH motif was also acetylated in various ArgRSs (Lys 156 in the yeast cytosol, Lys 139 in R. norvegicus mitochondria, and Lys 205 in the H. sapiens, R. norvegicus, and M. musculus cytosol; Fig. 10A) (35,38,39,54). These results indicate that Lys acetylation is a conserved mechanism for regulating the catalytic activity of aaRSs in particular conditions.
Recently, TyrRS has also been shown to be highly acetylated in response to oxidative stress. This aaRS is primarily acetylated on Lys 244 near the nuclear localization signal, and the acetyla-   (23). Pta, AckA, and Acs are crucial enzymes involved in the interconversion of Ac-CoA and acetate.

Acetylation regulates aminoacyl-tRNA synthetases
tion inhibits the aminoacylation activity of TyrRS. Acetylation, which is regulated by PCAF and sirtuin 1, also promotes TyrRS translocation from cytoplasm to the nucleus and protects against DNA damage caused by oxidative stress in mammalian cells and zebrafish. This study provided us with other perspectives about the biological role of aaRS acetylation (55).

AcP and CobB appear to regulate acetylation of EcArgRS and EcLeuRS
Our results showed that AcP and CobB may regulate acetylation of EcLeuRS and EcArgRS in vitro. We also tried to explore the acetylation state of these two endogenous aaRSs, but haven't yet found a physiological state that leads to an obvious increase in the acetylation of aaRSs. In a previous study (23), the global acetylation state of E. coli was found to be elevated in growth-arrested (GE) cells compared with those in the exponential phase (EP). This accumulation required AcP, and most acetylation was independent of YfiQ. Additionally, CobB can suppress chemical acetylation by AcP in growing and GE cells (23), and proteins functioning in translation, transcription and central metabolism are acetylated, with a considerable number of the acetylated sites regulated by AcP (28).
The supplementary data in the study of Weinert et al. (23) showed that acetylation of Lys 126 in EcArgRS can differ in E. coli cells cultured in M9 minimal medium on different genetic background and under different metabolic states. Acetylation of EcArgRS Lys 126 in stationary phase BL21 cells was ϳ10-fold higher compared with EP cells, indicating elevation of acetylation during this stage. Furthermore, acetylation of this site in EP ⌬ackA BW25113 cells was increased ϳ10-fold compared with EP control cells, suggesting that elevation of AcP levels (⌬ackA) by genetic manipulation could stimulate acetylation at specific sites (Fig. 10B). Acetylation of EcArgRS Lys 126 in EP ⌬pta BW25113 cells was 62.5% that of the level measured in EP WT cells, indicating that a decrease in AcP in ⌬pta cells reduces the acetylation of Lys 126 (Fig. 10B). In addition, deletion of the yfiQ gene had no direct effect on the acetylation of Lys 126 in EP ⌬yfiQ MG1655 cells compared with EP control MG1655 cells, consistent with our in vitro assay results that similarly indicated that YfiQ is not involved in the acetylation of the equivalent residue in EcArgRS. Further studies should focus on investigating the ability of CobB to deacetylate EcLeuRS Ac and EcAr-gRS Ac in vivo, as well as the physiological significance of acetylation. In addition, questions about whether there are some undiscovered acetyltransferases that can take AcP as a substrate and catalyze acetylation should be addressed.

Acylation of aaRSs might regulate the metabolic state of cells
In E. coli and yeast, acetylation is much less abundant than phosphorylation, and succinylation of aaRSs (Lys 619 and Lys 624 on EcLeuRS) has recently been reported (56). Nonenzymatic protein acylation has been linked to negative regulation of protein function because carbon stress and deacylases are evolved to reverse this form of PTM in both prokaryotes and eukaryotes (29). Our results, together with those of previous studies, suggest that acetylation and other forms of acylation may inhibit the activity of aaRSs in response to environmental stresses. Under normal conditions, aaRSs may endure basal-level acetylation. When there is a stimulus that requires cells to reduce their growth rates, cells could utilize an economical way to slow down protein synthesis. Environmental stresses may cause AcP-mediated acetylation of aaRSs, which inhibits their aminoacylation activities, and the deacetylase CobB may reverse this PTM to recover aa-tRNA biosynthesis when conditions improve (Fig. 11). It is very interesting to understand exactly how acetylation specificity is achieved. To our understanding, first, even though AcP can acetylate peptides nonenzymatically at high concentrations, with some salt and Mg 2ϩ , selectivity and specificity can increase (28). Second, it is possible that there might be an/some undiscovered acetyltransferase(s) that can utilize AcP as a substrate and catalyze acetylation (23); lastly, deacetylase could preferentially remove acetylation on some specific sites. Moreover, cross-talk between different types of acylation may be important and should be investigated. Whether acetylation could also affect other functions of aaRSs beyond translation, as demonstrated by TryRS, is a fascinating to question to be answered (55).

Concluding remarks
Herein, we confirmed acetylation of EcLeuRS and EcArgRS in vivo and identified the Lys residues involved. To investigate the significance of this form of PTM, we engineered K-Q mutants to identify residues that may affect the aminoacylation activity and employed a novel site-directed AcK incorporation Figure 11. Proposed acetylation mechanism for aaRSs. In this model, AcP nonenzymatically acetylates aaRSs, which negatively regulates their aminoacylation activities, and this PTM is removed by CobB to recover aaRS function and maintain cellular homeostasis (29). system to prepare EcLeuRS and EcArgRS acetylated at specific sites. Characterization of the amino acid activation and tRNAcharging activities of these EcLeuRS-K Ac and EcArgRS-K Ac variants confirmed that acetylation of several Lys residues negatively regulates their catalytic activities. Subsequent in vitro assays suggest that AcP might be the source of the nonenzymatic acetylation of EcLeuRS and EcArgRS, and CobB is likely to be responsible for deacetylation. This work extends our understanding of acetylation of aaRSs. Whether this type of PTM is prevalent and physiologically important remains to be elucidated. Qiagen. Superdex 75 resin and PVDF membranes were purchased from GE Healthcare. dNTP mixtures, arabinose (Ara), Tween 20, Triton X-100, BSA, Na 4 PPi, KH 2 PO 4 , and K 2 HPO 4 were purchased from Sangon (Shanghai, China). Oligonucleotide primers were synthesized by Invitrogen. A DNA fragment rapid purification kit and a plasmid extraction kit were obtained from Yuanpinghao Biotech (Tianjin, China). Protein standard markers, T4 ligase, restriction endonucleases and Zeba spin desalting columns were obtained from Thermo Scientific (Waltham, MA). The KOD-plus mutagenesis kit and KOD-plus Neo enzyme were purchased from TOYOBO (Osaka, Japan), and DNA sequencing was performed by Biosune (Shanghai, China). The pAcKRS system was gift from Prof. Jiang-yun Wang. Competent E. coli Top10 and BL21 (DE3) cells were prepared in our laboratory.

Gene cloning, mutagenesis, protein expression, and purification
Plasmid pET30a(ϩ)-ecleuS encoding EcLeuRS with a N-terminal His 6 tag was constructed previously in our lab (57). Mutation of ecleuS was performed by PCR as reported (57). EcLeuRS and its K-Q mutants were purified by Ni-NTA affinity chromatography as reported (58). To obtain site-directed EcLeuRS-K Ac , we constructed pET22b(ϩ)-ecleuS encoding EcLeuRS with a C-terminal His 6 tag, in which target Lys codons were separately mutated to TAG. We co-transformed E. coli BL21 (DE3) cells with pAcKRS and pET22b(ϩ)-ecleuS with a C-terminal His 6 tag. As described previously (43), EcLeuRS-K Ac s were overexpressed in E. coli BL21 (DE3) following the addition of Ara and IPTG, in the presence of ampicillin, chloramphenicol, 10 mM NAM, and AcK (Fig. 3). EcLeuRS-K Ac was purified by Ni-NTA affinity chromatography, followed by gel-filtration chromatography with a Superdex 75. The purity of the preparations was assessed by SDS-PAGE, and protein concentration was determined using the Bradford protein assay kit. EcLeuRS enzymes were stored in buffer containing 20 mM potassium phosphate (pH 6.8) and 1 mM DTT (58).
The gene encoding EcArgRS was amplified from plasmid pUC18-ecargS and inserted into pET28a(ϩ) (59). WT EcArgRS and K-Q mutants were purified by Ni-NTA affinity chromatography as described above for EcLeuRS. Because EcArgRS with a C-terminal His 6 tag displayed no activity (data not shown), EcArgRS-K Ac enzymes were also expressed using the pET28a(ϩ) vector. EcArgRS-K Ac enzymes were purified by two-step chromatography as described above for EcLeuRS and stored in buffer containing 20 mM potassium phosphate (pH 7.5) and 1 mM DTT (59). Cloning of genes encoding YfiQ, CobB, and YcgC and purification of the recombinant enzymes was performed as reported previously (25,26).

CD spectroscopy
The secondary structures of EcLeuRS, EcArgRS, and their variants were determined by CD spectroscopy as described previously (9).

Acquisition of tRNAs
E. coli tRNA Leu CAG (EctRNA Leu ) and E. coli tRNA Arg ICG (EctRNA Arg ) were isolated from the corresponding overexpression strains as previously described (60,61). The charging level of EctRNA Leu and EctRNA Arg was measured with 1 M corresponding EcLeuRS and EcArgRS for more than 20 min. Both tRNAs harbored ϳ1400 pmol/A 260 units of accepting activity.

Mass spectrometry
E. coli BL21 (DE3) cells harboring the corresponding plasmids containing ecleuS or ecargS were inoculated from overnight culture in a ratio of 1:100 to 2ϫ YT medium. When the A 600 reached 0.6 -0.8, ecleuS and ecargS genes were induced by IPTG in the presence of 10 mM NAM in 22°C for 6 h. The cells pellets were lysed by sonication in the presence of PMSF. After centrifugation, the supernatants were applied to the Ni-NTA column. Then His 6 -taggged EcLeuRS and EcArgRS were purified by the affinity chromatography. After SDS-PAGE, the target bands were separated, excised, and sent to Shanghai Applied Protein Technology (Shanghai, China). Protein bands were in-gel digested, subjected to nanoLC to separate the resultant peptides, and identified by mass spectrometry (Thermo Finnigan, Silicon Valley, CA). Data processing and analysis of raw files were conducted using Proteomics Tools 3.1.6 and Mascot 2.2.

Amino acid activation, aminoacylation, misaminoacylation, and deacetylation assays
Assays of EcLeuRS were performed as previously described (62). The amino acid activation of EcLeuRS and its mutants was assayed by monitoring ATP-PP i exchange reactions at 37°C in reaction mixture containing 100 mM HEPES (pH 7.8), 10 mM Determination of the tRNA Leu dissociation constant (K d ) by fluorescence quenching assays 0.1 M proteins in 400 l of equilibrium titration buffer containing 100 mM Tris-HCl (pH 8.2), 12 mM MgCl 2 , and 0.5 mM DTT were excited at 280 nm in a quartz cuvette, and the appropriate emission wavelength was monitored at room temperature as described previously (64). Maximum emission of EcLeuRS and EcArgRS was observed at 340 nm, and this was used to measure the fluorescence intensity of enzymes titrated with their cognate tRNAs. The final concentration of tRNAs ranged from 0.06 to 4.21 M, and the total volume of titrations containing tRNA was less than 20 l (1/20 of the original volume). We calculated the K d values by plotting the fluorescence intensity change (a.u.) against final tRNA concentration (M) according to the "one site-specific binding" option in GraphPad Prism software. BSA was used as a control.
In vitro deacetylation reactions were carried out as described previously with purified EcLeuRS or EcArgRS and purified CobB in reaction mixtures containing 40 mM HEPES (pH 7.0), 6 mM MgCl 2 , 1 mM NAD ϩ , 1 mM DTT, and 10% glycerol at 37°C (26). To investigate the effect of deacetylation by CobB on the aminoacylation activities of EcLeuRS-K Ac , enzymes were incu-bated with or without CobB in deacetylation buffer for 1 h at 37°C and subsequently diluted to 5 nM to initiate the aminoacylation reaction. For deacetylation of EcLeuRS Ac s and EcAr-gRS Ac s (corresponding enzymes preincubated with AcP), proteins were pretreated (desalted) using spin desalting columns.

Western blotting
For the detection of AcK, PVDF membranes were blocked in buffer containing 50 mM Tris-HCl (pH 7.5), 1% peptone, and 10% (v/v) Tween 20 and subsequently incubated with diluted primary antibody in 50 mM Tris-HCl and 0.1% peptone (53). All other Western blotting assays were conducted using conventional methods.