Modulation of the Bacterial Cobb Sirtuin Deacylase Activity by N-Terminal Acetylation

Modulation of the bacterial CobB sirtuin deacylase activity by N-terminal acetylation

Anastacia R. Parksa and Jorge C. Escalante-Semerenaa,1

aDepartment of Microbiology, University of Georgia, Athens, GA 30606

Edited by John E. Cronan, University of Illinois at Urbana–Champaign, Urbana, IL, and approved May 20, 2020 (received for review March 23, 2020) In eukaryotic cells, the N-terminal amino moiety of many proteins acetylated in bacteria, such as secretion chaperone SecB of is modified by N-acetyltransferases (NATs). This protein modifica- Escherichia coli and the virulence factor ESAT-6 (ExsA) of several tion can alter the folding of the target protein; can affect binding Mycobacterium species. In the case of the ESAT-6 (ExsA) protein, interactions of the target protein with substrates, allosteric effec- acetylation of its N terminus abolished binding interactions with its tors, or other proteins; or can trigger protein degradation. In pro- protein partner CFP10 (ExsB) and attenuated Mycobacterium karyotes, only ribosomal proteins are known to be N-terminally marinum virulence (17–19). Even though there has been identi- acetylated, and the acetyltransferases responsible for this modifi- fication of N-terminal acetylation in prokaryotes, the acetyl- cation belong to the Rim family of proteins. Here, we report that, transferases responsible for the modifications have not been in Salmonella enterica, the sirtuin deacylase CobB long isoform identified. For example, recently published bacterial N-terminal (CobBL) is N-terminally acetylated by the YiaC protein of this bac- acetylomes of Pseudomonas aruginosa, Acinetobacter baumannii, terium. Results of in vitro acetylation assays showed that CobBL and M. tuberculosis showed that roughly ∼10% of the proteins was acetylated by YiaC; liquid chromatography-tandem mass spec- were N-terminally acetylated (8, 11, 20), but the enzymes re- trometry (LC-MS/MS) was used to confirm these results. Results of sponsible for such modifications were not identified. in vitro and in vivo experiments showed that CobBL deacetylase The Salmonella enterica genome contains ∼26 GNATs, but the activity was negatively affected when YiaC acetylated its N termi- function of about a third of them has yet to be defined. In ad- nus. We report 1) modulation of a bacterial sirtuin deacylase ac- dition, the S. enterica genome possesses one nicotinamide ade- tivity by acetylation, 2) that the Gcn5-related YiaC protein is the nine dinucleotide (NAD+)-dependent sirtuin deacylase CobB, acetyltransferase that modifies CobBL, and 3) that YiaC is an NAT. Based whose function works in concert with the protein acetyl- MICROBIOLOGY on our data, we propose the name of NatA (N-acyltransferase A) in lieu transferase (Pat) enzyme to reversibly modulate the activity of of YiaC to reflect the function of the enzyme. acetyl-CoA synthetase (Acs) (21, 22). As shown in (Scheme 1, the CobB-dependent deacetylation reaction consumes NAD+ sirtuin | CobB sirtuin deacylase | posttranslational modification | N-terminal and yields O-acetyl-ADP (adenosine diphosphate ribose) and acetylation | bacterial GNAT nicotinamide. Interestingly, this bacterium synthesizes two biologically active rotein acylation is common in prokaryotes and eukaryotes, isoforms of CobB, referred to as CobBS (CobB short isoform) Pand it is an effective and rapid means of controlling protein and CobBL (CobB long isoform), which differ in size by a function in response to diverse stimuli (1, 2). What stands out 37-residue N-terminal extension of the catalytic core (Fig. 1). about protein acylation is the diversity of organic acids used by Our group showed that both isoforms of CobB deacetylated cells to modify proteins (e.g., acetate, propionate, malonate, their bona fide protein substrate (i.e., acetylated acetyl-CoA succinate, etc.) and the large number of acyltransferases that synthetase [AcsAc]) in vivo and in vitro (23). However, the catalyze the modifications (1). Many of the acyltransferases that physiological relevance of the two CobB isoforms in S. enterica is modify proteins and small molecules belong to the protein su- unknown. Here, we report that the CobBL sirtuin deacylase perfamily PF00583, and among this family, many proteins con- isoform of this bacterium is N-terminally acetylated and that the tain the so-called GNAT (GCN5-related N-acetyltransferase) putative YiaC acetyltransferase acetylates the N terminus of domain (IPR000182). GNATs acylate free amino groups of proteins or small molecules (1, 2). For example, there is a subset Significance of well-studied GNATs that modify the e amino group (Ne)in α lysine side chains (3, 4), while other GNATs modify the amino N-terminal protein acetylation is poorly understood in bacteria. α – group (N ) of the starting residue of proteins (5 8). Herein, we report the identification of an Nα acetyltransferase To frame the work reported here, we note that many eukaryotic (NAT) that modulates the activity of a sirtuin deacylase in a hu- proteins are acetylated on their N termini and that the acetyl- man pathogen. This is significant because the alluded enzyme transferases responsible for these modifications are referred to as (named N-acyltransferase A [NatA], formerly YiaC) is a prokaryotic N-acetyltransferases (NATs). In general, NATs catalyze the non-Rim–type NAT, and N-terminal acetylation of a bacterial sir- transfer of the acetyl group from acetyl-Coenzyme A (AcCoA) to tuin has not been reported. Also significant is the fact that NatA a primary amine of a small molecule or the N-terminal amino affects the metabolism of acetate, a short-chain fatty acid known α group of a peptide or protein. In eukaryotes, N acetylation has to play an important role in pathogenesis in the human gut. been suggested to alter protein folding, protein–protein interactions, and protein degradation (9–12). In contrast, little is known Author contributions: A.R.P. and J.C.E.-S. designed research; A.R.P. performed research; about the enzymes that catalyze N-terminal protein acetylation in A.R.P. and J.C.E.-S. analyzed data; J.C.E.-S. conceptualized the project; and A.R.P. and prokaryotes and what the physiological reasons for such modifi- J.C.E.-S. wrote the paper. cation may be. Examples of N-terminal acetylation of bacterial The authors declare no competing interest. proteins, where the acetyltransferase is known, are the acetylation This article is a PNAS Direct Submission. of ribosomal proteins S18, S5, and L7/L12 by acetyltransferases Published under the PNAS license. RimI, RimJ, and RimL (13–15). These acetyltransferases were 1To whom correspondence may be addressed. Email: [email protected]. thought to have high substrate specificity until recently, when the This article contains supporting information online at https://www.pnas.org/lookup/suppl/ Mycobacterium tuberculosis RimI was shown to acetylate different doi:10.1073/pnas.2005296117/-/DCSupplemental. peptides in vitro (16). Several other proteins are known to be Nα

www.pnas.org/cgi/doi/10.1073/pnas.2005296117 PNAS Latest Articles | 1of7 Downloaded by guest on September 27, 2021 + *AcCoA + *AcCoA + YiaC+ *AcCoA + YiaC+ *AcCoA L S L S

MM (kDa)CobB CobB CobB CobB Scheme 1. 37

CobBL CobBL. We also report in vivo and in vitro evidence that YiaC- dependent N-terminal acetylation of CobBL negatively affects its 25 deacetylase activity. CobBS

Results 20 SDS-PAGE YiaC Acetylates CobBL but Not CobBS. A search for protein substrates for the S. enterica putative GNATs led us to discover that the S. enterica YiaC protein acetylated S. enterica CobBL but not CobBS. As shown in Fig. 2, when both isoforms of S. enterica 15 CobB were incubated with [1-14C]-AcCoA as a function of YiaC, YiaC radiolabel was transferred to CobBL but not to CobBS (Fig. 2, lanes 4 and 5, respectively). Since CobBS was not acetylated, we

surmised that the sites of acetylation were located within the Ac * CobBL 37-amino acid N-terminal, arginine-rich motif of CobBL (Fig. 1).

YiaC Does Not Acetylate Ne Amino Groups of Lysine Residues. The *AcCobB PHOSPHOR IMAGE S N-terminal extension of CobBL contains two lysines (K14, K16), which we investigated as possible acetylation sites. To test this Lane # 12 3 4 5 possibility, we changed K14 and K16 to alanine, independently K14A K16A and in combination. The three variants (CobBL , CobBL , Fig. 2. YiaC acetylates CobBL but not CobBS. YiaC-dependent acetylation of 14 and CobB K14A,K16A) were overproduced, isolated, and incu- CobB isoforms was assessed after incubation of the proteins with [1- C]-AcCoA L μ bated with YiaC in the presence of [1-14C]-AcCoA. Surprisingly, (20 M) for 1 h at 37 °C. Detailed conditions of the assay are described in Ma- YiaC acetylated all three CobB variants (Fig. 3, lanes 5, 7, and terials and Methods. Controls included incubation of the CobB isoforms with [1- L 14C]-AcCoA in the absence of YiaC. Proteins were resolved by SDS-PAGE (sodium 9), suggesting that, under the assay conditions used, YiaC did not e dodecyl sulfate polyacrylamide gel electrophoresis) and visualized by Coomassie modify the N position of either K14 or K16. We note that the Brilliant Blue R stain (Upper) using Precision Plus protein (Bio-Rad) standards as K14A,K16A intensity of the signal for acetylated CobBL variant was molecular mass markers (lane 1, MM, kilodaltons). Radiolabel signal was visual- less than the single-amino acid variants and that the signal in- ized by phosphor imaging (Lower). *Radiolabeled acetyl moieties. tensity was commensurate to the amount of protein loaded on K14A,K16A the gel. The yield of CobBL variant was lower than those of the single-amino acid variants. (Fig. 5, lanes 2 and 8) but did not acetylate the CobBL peptide Ac whose first residue was L-Met (Fig. 5, lane 5). Collectively, these YiaC Modifies the N Terminus of CobBL In Vitro. To support our data showed that YiaC was an Nα protein acetyltransferase (NAT). hypothesis that CobBL is Nα acetylated, liquid chromatography- tandem mass spectrometry (LC-MS/MS) was conducted. Results of peptide fingerprinting analysis of acetylated CobB long iso- Ac form ( CobBL) unequivocally showed that the N terminus of

CobBL was acetylated by YiaC (Fig. 4). + *AcCoA + YiaC + *AcCoA YiaC + *AcCoA To confirm the LC-MS/MS data, two peptides of the first 50 + *AcCoA + YiaC + *AcCoA + *AcCoA + + *AcCoA + YiaC + K14A*AcCoA K14A K16A K16A K14A, K16A K14A, K16A L L L L L L L L

amino acids of CobBL were synthesized (Peptide 2.0, Virginia): one MM (kDa) CobB CobB CobB CobB CobB CobB CobB CobB started with unmodified L-Met, and the second one started with 37 Ac L-Met . In vitro acetylation assays were performed with the above- 25 14 mentioned peptides as substrates for YiaC. When [1- C]-AcCoA 20 SDS-PAGE was added to the reaction mixture, YiaC acetylated the CobBL peptides whose N-terminal amino group was not modified 15

*AcCobBL PHOSPHOR IMAGE Lane # 123 4 56 7 8 9 1MQSRRFHRLSRFRKNKRLLRERLRQRIFFRDRVVPEM37 Fig. 3. YiaC does not acetylate lysyl residues of CobBL. To query the site of modification, lysine variants of CobBL were purified and incubated with YiaC and [1-14C]-AcCoA. The goal was to determine whether or not YiaC was an α CobB N acetyltransferase. This experiment was conducted as described for Fig. 2 L 31 kDa (273 aa) K14A,K16A where CobBL variants (3 μM, except CobBL wasat2μM) were incubated with [1-14C]-AcCoA and either with or without YiaC (1 μM). Lanes 4 and 5 con- CobB 26 kDa (236 aa) K14A K16A S tained CobBL , lanes 6 and 7 contained CobBL , and lanes 8 and 9 con- K14A,K16A tained CobBL . Results of control experiments using CobBL plus AcCoA Fig. 1. Biologically active CobB sirtuin deacylase isoforms of S. enterica. with or without YiaC are shown in lanes 2 and 3, respectively. Lane 1 represents N-terminal amino acid sequence of CobBL. Yellow highlighted residues represent molecular mass marker (MM) in kilodaltons. SDS-PAGE, sodium dodecyl sulfate hydrophobic amino acids; red residues are arginines, and blue residues are lysines. polyacrylamide gel electrophoresis. *Radiolabeled acetyl moieties.

2of7 | www.pnas.org/cgi/doi/10.1073/pnas.2005296117 Parks and Escalante-Semerena Downloaded by guest on September 27, 2021 A 1812 100 12.20 361.169 90

80 70 AcGTMQSR - [EIC] 60

20 R e l a t i v b u n d c 10 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 T i m e (m i n) AcGTMQSR - [MS/MS]

B 2+ [MH-H2O] 352.219 y3 1.4 b3 390.301 332.021 1.2 MICROBIOLOGY 1.0

2+ 0.8 y2 y5 262.257 311.696 0.6 b4 y4 0.4 y1 521.271 b2 460.380 175.068 299.908 b5 201.140 0.2 147.092 498.347 547.275 R e l a t i v b u n d c 442.369 113.162 603.563 634.306 696.554 0.0 100 150 200 250 300 350 400 450 500 550 600 650 700 m / z

Ac Fig. 4. Mass spectrometry analysis of CobBL.CobBL (5 μM) was incubated with AcCoA (1 mM) with and without YiaC (3 μM) at 37 °C for 1 h. Reaction mixture components were resolved by SDS-PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis). CobBL was excised from the gel and sent to the University of Wisconsin-Madison Biotechnology Center Mass Spectrometry Facility for LC-MS/MS (liquid chromatography-tandem mass spectrometry) analysis. (A) Chromato- Ac Ac graphic trace for the selected ion of interest ( GTMQSR). EIC, extracted ion chromatogram. (B) Mass spectrum of CobBL. b ions are the series of fragments that extend from the N terminus; y ions are the series of fragments that extend from the C terminus. The x axis of m/z stands for mass (m) over charged number of ions (z). MASCOT software (http://www.matrixscience.com) was the online search engine used to identify peptides on the basis of their masses.

Ac Ac N-Terminally CobBL Cannot Be Deacetylated by Either Isoform of performed in vitro protein deacetylation assays with Acs* and Ac Ac CobB. To test whether N-terminally CobB could be deacety- CobBL to enhance our ability to detect deacetylation events. Ac Ac lated by either isoform of CobB, acetylation assays with YiaC, The protocols for the generation of Acs* and * CobBL are 14 Ac CobBL,and[1- C]-AcCoA were conducted. Radiolabeled, acety- described in Materials and Methods. After we had Acs* and Ac Ac lated CobB long isoform (* CobBL) samples were freed of excess * CobBL, the following experiment was performed. CobBL or Ac AcCoA and were incubated with either CobBL or CobBS,withor * CobBL was added to a reaction mixture that contained + Ac + μ without NAD . Positive controls were also added to show CobBL Acs* and NAD , and samples (25 L each) were incubated for Ac Ac and CobBS deacetylated its bona fide substrate, Acs .Asseenin 1 h. As seen in Fig. 7A, unacetylated CobBL deacetylated Acs* Fig. 6, CobBL and CobBS deacetylated radiolabeled, acetylated as indicated by the disappearance of the signal in the phosphor Ac acetyl-CoA synthetase (Acs*Ac; lanes 3 and 5) but did not to image corresponding to Acs* (molecular mass ∼72 kDa) Ac Ac deacetylate * CobBL (lanes 7 and 9). This showed that N-terminal (Fig. 7A, lane 5). In contrast, * CobBL did not deacetylate Acs*Ac as efficiently as CobB (Fig. 7A,lane3vs.lane5). acetylation of CobBL was not reversed by either CobB isoform. L Quantitative densitometry of the signals of AcsAc radiolabel dis- ∼ In Vitro and In Vivo Evidence That Nα Acetylation of CobBL Alters Its appearance showed an average of 50% reduction in CobBL Deacetylase Activity. deacetylase activity when acetylated. The results shown in Fig. 7 In vitro evidence. The N-terminal acetylation of CobBL by YiaC are representative of a set of six separate experiments. These data Ac (Fig. 2) raised questions regarding the effect of the modification showed that the enzymatic activity of * CobBL was negatively on the enzymatic activity of CobBL. To answer this question, we affected by the modification.

Parks and Escalante-Semerena PNAS Latest Articles | 3of7 Downloaded by guest on September 27, 2021 circles in Fig. 7B). These results were consistent with our in vitro data, which showed that N-terminal acetylation of CobBL by YiaC had a negative effect on the deacetylase activity of CobBL (Fig. 7A). Again, under the growth conditions used, a reduction in deacetylase activity would prevent the deacetylation (hence reactivation) of Acs, blocking the conversion of acetate into AcCoA with the concomitant negative effect on cell growth. As predicted by the above results, the specific activity of Acs was different between the ΔcobB strain that ectopically synthesized CobBL and YiaC and the ΔcobB strain that synthesized CobBL but did not ectopically synthesized YiaC (Fig. 8A). In the above-mentioned strains, we measured a statistically significant reduction of Acs activity when YiaC was overproduced. These results were consistent with the data showing that YiaC- dependent acetylation of the CobBL N terminus reduced its activity, maintaining Acs acetylated (hence inactive) and thus,

Fig. 5. YiaC acetylates the N-terminal methionine of CobBL. SDS-PAGE and arresting growth due to reduced levels of AcCoA. phosphor imaging analysis of unacetylated or N-terminally acetylated syn- To verify that the concentration of Acs was the same in both thetic peptides of amino acids spanning the first 50 residues of CobBL that strains, quantitative western blots using rabbit polyclonal anti- 14 were incubated with YiaC without [1- C]-AcCoA (lanes 3 and 6), with [1- Acs antibodies were performed to determine the amount of Acs 14 14 C]-AcCoA but no YiaC (lanes 4 and 7), or with YiaC and [1- C]-AcCoA protein present in lysates used to assay for Acs activity. Indeed, (lanes 2, 5, and 8). Lanes 1 and 9 are the Precision Plus protein standard + molecular mass markers (MM) (kilodaltons). SDS-PAGE, sodium dodecyl sul- lysates of strains that either overexpressed yiaC or carried the fate polyacrylamide gel electrophoresis. *Radiolabeled acetyl moieties. empty vector contained the same amount of Acs (Fig. 8B and SI Appendix, Fig. S2).

Acetylation of CobB Does Not Cause CobB Degradation In Vivo. In In vivo evidence. To verify that acetylation of the N terminus of L L eukaryotes, some N-terminally acetylated proteins are targeted CobB affected its enzymatic activity, we performed in vivo ex- L for degradation (27). Other examples suggest that N-terminal periments to assess the activity of CobB as a function of YiaC L acetylation stabilizes acetylated proteins and prevents degrada- during growth on minimal medium containing a low concentration of acetate (10 mM). It is known that, under such conditions, tion (28, 29). To assess whether acetylation of CobBL altered CobB function is required to maintain Acs deacetylated, hence CobBL protein levels in vivo, we used rabbit polyclonal anti- CobB antibodies to quantify levels of CobB in cell-free extracts active (3, 22, 26). For this purpose, we moved two plasmids into + + + an S. enterica ΔcobB strain. One plasmid encoded either of three strains: yiaC /vector, yiaC::kan /vector, and yiaC::kan / (M37A,M38A) (M1A) pYiaC. Strains were grown in minimal medium supplemented CobBL or CobBS ; the second plasmid encoded μ YiaC or the vector control. Genes cobB and yiaC were under the with acetate (10 mM) and with L-(+)-arabinose (500 M) or in control of the L-(+)-arabinose–inducible promoter ParaBAD (24). To get only CobBL and no CobBS protein, the plasmid encoding + + CobBL contained the natural starting methionine (M1) for + +

CobBL but had two mutations that change the starting methio- + NAD + NAD + NAD + NAD L L S S + L S S nine and neighboring methionine for CobBS (i.e., M37 and M38) L M37A,M38A to alanines. The cobB allele encoding CobB effectively + CobB + CobB + CobB + CobB + CobB + NAD L L L L L Ac * + CobBAc * + CobBAc * + CobBAc * blocks the synthesis of CobBS (23). Conversely, for the cell to CobB CobB CobB CobB CobB MM (kDa)Acs Acs Acs Acs Ac Ac Ac Ac Ac * * * * * synthesize only CobBS, the CobBS-encoding plasmid has a cobB 100 allele with the first methionine changed to encode alanine so 75 AcsAc* that CobBL protein cannot be made from M1, resulting in the 50 exclusive synthesis of CobBS starting at position M38 (23). The growth behaviors of strains ΔcobB and cobB+ harboring empty cloning vectors were used as controls. An additional control used 37 Ac + + SDS-PAGE * CobB a cobB strain harboring a plasmid carrying yiaC (open circles L in Fig. 7B). All strains were grown on minimal medium with 25 CobBS acetate (10 mM) as the sole carbon and energy source. AcsAc* Data presented in Fig. 7B show the growth behavior of strains of interest in the presence of inducer [L-(+)-arabinose, 100 μM]. The following observations were made. 1) As expected, the

ΔcobB strain failed to grow on 10 mM acetate (black triangles in IMAGE Ac PHOSPHOR * CobB Fig. 7B) because Acs remained acetylated, hence inactive. 2) The L Δ phenotype of the cobB strain was corrected by ectopic ex- Ac * CobBS pression of cobB alleles that directed the synthesis of functional Lane # 123 4 5 6 7 8 9 10 CobBS (open squares in Fig. 7B) or CobBL (black squares in α 14 Ac Fig. 7B). 3) Synthesis of YiaC did not affect the growth of the Fig. 6. N acetylated proteins are not substrates for CobB. [1- C]- CobBL + was incubated with either CobBL or CobBS with (lanes 7 and 9) and without cobB strain (open circles vs. gray squares in Fig. 7B). 4) Syn- + thesis of CobB and YiaC by a strain with a genomic deletion of NAD (lanes 6 and 8). Positive controls of unlabeled CobBL and CobBS were S also tested for their ability to deacetylate [1-14C]-AcsAc (lanes 2 to 5). Lane 1 cobB resulted in a growth behavior that was very similar to that + represents molecular mass standards (MM) reported in kilodaltons. Samples of the cobB strain that synthesized YiaC (open triangles vs. gray were resolved by SDS-PAGE (sodium dodecyl sulfate polyacrylamide gel 14 squares and open circles in Fig. 7B). 5) Synthesis of CobBL and electrophoresis), and transfer of [1- C] label was revealed by phosphor YiaC by the ΔcobB strain prematurely arrested growth (gray image analysis. *Radiolabeled acetyl moieties.

4of7 | www.pnas.org/cgi/doi/10.1073/pnas.2005296117 Parks and Escalante-Semerena Downloaded by guest on September 27, 2021 + +

+ NAD

Ac * Ac * + NAD Ac * Ac *

+ Acs + Acs Acs A L L + Acs + B L L *CobB *CobB MM (kDa)Ac Ac CobB CobB cobB+ / vectors 75 0.25 cobB+/ pYiaC cobB / pCobB 50 S cobB / pCobB / pYiaC 37 0.15 S SDS-PAGE cobB / pCobBL cobB / pCobB / pYiaC AcsAc* 0.1 L cobB

Optical Density (630 nm) / vectors 0102030 Time (h) IMAGE Ac PHOSPHOR * CobBL Lane # 12345

Ac Fig. 7. Acetylation of the N terminus of CobBL negatively affects its deacetylase activity in vitro and in vivo. (A) To assess the enzymatic activity of CobBL in vitro, Acs 14 Ac Ac + protein radiolabeled with [1- C]-AcCoA (Acs *) was incubated with either CobBL (lane 5) or * CobBL (lane 3) and NAD to visualize the removal of the radiolabeled 14 Ac Ac acetyl group from Acs. CobBL was acetylated with YiaC using [1- C]-AcCoA as substrate. * CobBL helped visualize the mobility of CobBL on the phosphor image. Negative controls included reactions listed above except no NAD+ was added (lanes 2 and 4). Lane 1 shows molecular mass marker (MM) in kilodaltons. SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis. (B) All strains were grown on NCE (no-carbon essential) minimal medium supplemented with acetate (10 mM) + as the sole source of carbon and energy. The strains used were ΔcobB/pCV1/pCV3 (black triangles; negative control), ΔcobB/pcobB (CobBS)/pCV1 (open squares; + + + + complementation), ΔcobB/pcobB (CobBL)/pCV1 (black squares; complementation), cobB /pCV1/pCV3 (gray squares; wild-type control), cobB /pyiaC /pCV3 (open circles), + + + + ΔcobB/pcobB (CobBS)/pyiaC (open triangles), and ΔcobB/pcobB (CobBL)/pyiaC (gray circles). Cloning vectors pCV1 and pCV3 contain the same arabinose-inducible promoter (24, 25). Maps of cloning vectors pCV1 and pCV3 can be found in ref. 25. The concentration of arabinose used was 100 μM. Each strain was grown in biological and technical triplicate, and analyses were conducted three independent times. Error bars represent SD of technical triplicates. *Radiolabeled acetyl moieties.

rich medium (Lysogeny Broth, LB) containing L-(+)-arabinose (1 activity. Our data also show that YiaC modifies the long isoform MICROBIOLOGY mM). In both cases, growth was monitored at 600 nm. Data pre- of the CobB sirtuin deacylase (CobBL)ofthisbacteriumbutnot sented in Fig. 9 show that during growth in rich medium or in the short isoform of CobB (Fig. 2). These results placed the sites minimal medium with a low concentration of acetate, the con- of acetylation within the 37-amino acid, N-terminal extension of e centration of CobBL did not vary in any of the strains. Pure protein CobBL (Fig. 1). Second, YiaC does not modify N amino groups controls of CobBL and CobBS were added as positive controls, and of lysine side chains of the CobBL protein (Figs. 3 and 4). Whether an anti-DnaK blot was included to ensure that all samples were YiaC can modify lysyl residues of other proteins remains to be loaded equally. Statistical analysis of pixel density of the anti-CobB determined. Third, YiaC appears to have somewhat broad speci- α western blots shows no difference in CobBL concentration in any of ficity for its target since it acetylated the N amino group of the glycyl residue that remained fused to the protein after protease the strains tested, as well as no difference in CobBS or the internal control protein DnaK (Fig. 9C and SI Appendix,Fig.S5). treatment to remove the MBP (maltose binding protein)-H6 tag (Figs. 2 and 4). YiaC also acetylated the Nα amino group of the Discussion N-terminal methionine of the C-terminally H6-tagged CobBL,and In S. enterica, the YiaC protein is an Nα acetyltransferase that the N-terminal methionine of a synthetic peptide comprised the controls the activity of the long isoform of the CobB sirtuin deacylase. first 50 amino acids of CobBL (Fig. 5 and SI Appendix,Fig.S4). We note that repeated attempts to isolate N-terminally Data reported here support several conclusions regarding the function Ac of the S. enterica YiaC protein. First, YiaC has Nα acetyltransferase CobBL from cells were unsuccessful, despite the fact that several different proteases were used in combination prior to mass spectrometry analysis. We posit that the amino acid composition AB of the N terminus of CobBL makes this analysis difficult because it is so rich in arginines. However, when we removed the N-terminal tag from MBP-H6-CobBL, the resulting protein had two additional residues on its N terminus, namely Gly-Thr (GT-CobBL). YiaC acetylated GT-CobBL in vitro (Fig. 4 and SI Appendix,Fig.S4)but did not acetylate a protein that had an N-terminal hexahistidine tag (H6-CobBL protein) (SI Appendix,Fig.S4). Collectively, our results show that YiaC is an Nα acetyltransferase (NAT) and that the CobBL isoform is a substrate of it. To the best of our knowledge, YiaC is a bacterial NAT that Fig. 8. Acs activity is decreased when YiaC is overproduced in S. enterica.(A) does not belong to the Rim family of proteins. Based on our M37A,M38A Strains of S. enterica with cobB deleted and with only pCobBL in trans data, we propose to change the name of YiaC to NatA, to reflect with or without pYiaC were grown to midlogarithmic phase with 10 mM acetate the fact that it is an Nα acetyltransferase. as the sole carbon and energy source and were lysed and tested for Acs-specific activity from 4 μg of lysate. Specific activity (micromoles AMP [adenosine mono- Can YiaC Acetylate the e Amino Group of Lysine Side Chains? Re- −1 −1 phosphate] minute milligram ) was calculated using a continuous spectropho- cently, Christensen et al. (30) showed that overexpression of yiaC tometric assay described in Materials and Methods. The activity of Acs decreased in in an E. coli pta patZ acs cobB deletion strain displayed increased lysates of the strain with pYiaC overexpressed. The experiment was performed in α biological triplicate with nine technical replicates each. *P = 0.03. (B)Acsprotein protein acetylation measured by western blot analysis using -AcK concentration is the same in both strains tested for Acs activity based on quanti- antibodies. These results are interesting because these investigators tative anti-Acs western blot analysis of lysates used in the experiment in A.Error conducted AcK enrichment and mass spectrometry to identify pu- bars represent unpaired t test with equal SD; not significant (ns) P value = 0.5118. tative YiaC protein lysine targets. CobB was not identified from the

Parks and Escalante-Semerena PNAS Latest Articles | 5of7 Downloaded by guest on September 27, 2021 Rich medium

+ / vector + / pYiaC + / vector + / pYiaC + / vector + / pYiaC

A + / vector + / vector + / vector ::kan

yiaC yiaC::kan yiaC::kan yiaC yiaC::kan yiaC::kan yiaC yiaC yiaC::kan Pure CobB OD600 nm Control 0.2 0.2 0.2 0.5 0.5 0.5 0.7 0.7 0.7

CobBL -CobB

CobBS

-DnaK

Low acetate medium B C 1500000 P=0.9747 + / vector + / pYiaC

+ / vector

Pure CobB signal 1000000 L Control yiaC yiaC::kan yiaC::kan

-CobB 500000 -DnaK Intensity of CobB 0

+ / vector+ / vector / pYiaC

yiaC

yiaC::kan

Fig. 9. YiaC does not alter CobB protein levels in vivo. The concentration of CobB isoform protein in S. enterica strains where yiaC was deleted or com- + plemented by pYiaC was compared with the parent strain (yiaC /vector) by western blotting. (A) Western blots of cells grown on LB medium with 1 mM L-

(+)-arabinose using anti-CobB or anti-DnaK antibodies. Cells were harvested at three different optical densities (OD600 nanometers, nm) of 0.2, 0.5, and 0.7. (B) Western blots of CobB protein levels or DnaK protein levels from cells grown with 10 mM acetate minimal medium supplemented with 500 μM L- (+)-arabinose. Anti-DnaK western blotting was used as positive controls to show that all lanes were loaded equally and to use as a standard for densitometry

calculations. (C) The density of pixels for the CobBL-only sample of the anti-CobB western blot from B showing the concentration of CobBL in all strains are the same. Calculations were conducted using one-way ANOVA, and differences of CobBL protein concentrations are not significant with a P value of 0.9747.

aforementioned proteome because the cobB gene was deleted in the Materials and Methods strains used, and N-terminal acetylation of proteins was not reported. Detailed protocols used in this study are presented in SI Appendix. SI Ap-

As mentioned above, at present we cannot rule out the possibility pendix contains protocols for the purification of CobBS, CobBL, and YiaC that YiaC acetylates lysyl residues as suggested by Christensen et al. proteins; size exclusion chromatography; in vitro acetylation and deacety- (30). Additional work is needed to determine whether YiaC can lation assays; a protocol for the in vitro determination of Acs activity; perform Nα and Ne protein acetylation. The identified activity of quantitative western blot analysis; lists of strains, plasmids, and primers; YiaC as an N-terminal acetyltransferase raises questions about the methods for the construction of strains and plasmids; culture media, growth role of N-terminal acetylation in prokaryotic cell physiology. conditions, and growth behavior analysis; and mass spectrometry analysis of the acetylation state of proteins of interest. SI Appendix also contains tables Concluding Remarks. We have shown that N-terminal acetylation which contain information about strains, plasmids, and primers, and YiaC of CobBL occurs both in vitro and in vivo, as well as that YiaC- and CobB homologs in other enteric bacteria, respectively (SI Appendix, Tables S1–S4). In addition, SI Appendix contains figures (SI Appendix, Figs. dependent acetylation of CobBL negatively impacts its deacetylase activity, which then negatively affects growth on acetate. How the S1–S5) that present results of control experiments, data regarding the oligomeric state of YiaC in solution, and quantification of western blots of addition of an acetyl group to the N terminus of CobBL impacts its activity is a question of interest, and ongoing studies in our lab- CobB isoforms or DnaK (control) used in this study. oratory are focused on answering this question. YiaC is an NAT that does not belong to the Rim protein family of Nα amino group Data Availability. All data of this work are reported in the paper. acetyltransferases. Also, we report acylation of a prokaryotic sir- ACKNOWLEDGMENTS. We thank Grzegorz Sabat from the Biotechnology tuin deacylase. As pointed above, based on the evidence reported Center of The University of Wisconsin–Madison for the performance of the here, we propose to change the name of YiaC to NatA (for Nα LC-MS/MS analysis. Rachel Burckhardt first observed CobB acetylation. This acetyltransferase) to reflect the biochemical activity of the enzyme. work was supported by NIH Grant R35 GM130399 (to J.C.E.-S.).

1. K. L. Hentchel, J. C. Escalante-Semerena, Acylation of biomolecules in prokaryotes: A 4. J. G. Gardner, J. C. Escalante-Semerena, Biochemical and mutational analyses of AcuA, widespread strategy for the control of biological function and metabolic stress. Mi- the acetyltransferase enzyme that controls the activity of the acetyl coenzyme a crobiol. Mol. Biol. Rev. 79, 321–346 (2015). synthetase (AcsA) in Bacillus subtilis. J. Bacteriol. 190, 5132–5136 (2008). 2. C. M. VanDrisse, J. C. Escalante-Semerena, Protein acetylation in bacteria. Annu. Rev. 5. H. Takakura, S. Tsunasawa, M. Miyagi, J. R. Warner, NH2-terminal acetylation of ri- Microbiol. 73, 111–132 (2019). bosomal proteins of Saccharomyces cerevisiae. J. Biol. Chem. 267, 5442–5445 (1992). 3. V. J. Starai, J. C. Escalante-Semerena, Identification of the protein acetyltransferase 6. A. Yoshikawa, S. Isono, A. Sheback, K. Isono, Cloning and nucleotide sequencing of (Pat) enzyme that acetylates acetyl-CoA synthetase in Salmonella enterica. J. Mol. the genes rimI and rimJ which encode enzymes acetylating ribosomal proteins S18 Biol. 340, 1005–1012 (2004). and S5 of Escherichia coli K12. Mol. Gen. Genet. 209, 481–488 (1987).

6of7 | www.pnas.org/cgi/doi/10.1073/pnas.2005296117 Parks and Escalante-Semerena Downloaded by guest on September 27, 2021 7. C. J. Walsh, “Protein N-acetylation” in Posttranslational Modification of Proteins: 20. T. Ouidir, F. Jarnier, P. Cosette, T. Jouenne, J. Hardouin, Characterization of Expanding Nature’s Inventory, (Roberts and Company Publishers, Greenwood Village, N-terminal protein modifications in Pseudomonas aeruginosa PA14. J. Proteomics – CO, 2006), pp. 151 170. 114, 214–225 (2015). 8. T. Kentache, T. Jouenne, E. Dé, J. Hardouin, Proteomic characterization of Nα- and Ne- 21. V. J. Starai, I. Celic, R. N. Cole, J. D. Boeke, J. C. Escalante-Semerena, Sir2-dependent – acetylation in Acinetobacter baumannii. J. Proteomics 144, 148 158 (2016). activation of acetyl-CoA synthetase by deacetylation of active lysine. Science 298, 9. K. T. Nguyen et al., N-terminal acetylation and the N-end rule pathway control 2390–2392 (2002). degradation of the lipid droplet protein PLIN2. J. Biol. Chem. 294, 379–388 (2019). 22. V. J. Starai, H. Takahashi, J. D. Boeke, J. C. Escalante-Semerena, Short-chain fatty acid 10. K. T. Nguyen, S. H. Mun, C. S. Lee, C. S. Hwang, Control of protein degradation by activation by acyl-coenzyme A synthetases requires SIR2 protein function in Salmo- N-terminal acetylation and the N-end rule pathway. Exp. Mol. Med. 50,1–8 (2018). – 11. C. R. Thompson, M. M. Champion, P. A. Champion, Quantitative N-terminal foot- nella enterica and Saccharomyces cerevisiae. Genetics 163, 545 555 (2003). printing of pathogenic mycobacteria reveals differential protein acetylation. 23. A. C. Tucker, J. C. Escalante-Semerena, Biologically active isoforms of CobB sirtuin J. Proteome Res. 17, 3246–3258 (2018). deacetylase in Salmonella enterica and Erwinia amylovora. J. Bacteriol. 192, 12. R. Ree, S. Varland, T. Arnesen, Spotlight on protein N-terminal acetylation. Exp. Mol. 6200–6208 (2010). Med. 50,1–13 (2018). 24. L. M. Guzman, D. Belin, M. J. Carson, J. Beckwith, Tight regulation, modulation, and 13. M. Yaguchi et al., Cooperative control of translational fidelity by ribosomal proteins high-level expression by vectors containing the arabinose PBAD promoter. in Escherichia coli. II. Localization of amino acid replacements in proteins S5 and S12 J. Bacteriol. 177, 4121–4130 (1995). altered in double mutants resistant to neamine. Mol. Gen. Genet. 142,35–43 (1975). 25. C. M. VanDrisse, J. C. Escalante-Semerena, New high-cloning-efficiency vectors for 14. B. Wittmann-Liebold, B. Greuer, The primary structure of protein S5 from the small complementation studies and recombinant protein overproduction in Escherichia coli – subunit of the Escherichia coli ribosome. FEBS Lett. 95,91 98 (1978). and Salmonella enterica. Plasmid 86,1–6 (2016). 15. C. Terhorst, B. Wittmann-Liebold, W. Möller, 50-S ribosomal proteins. Peptide studies 26. C. H. Chan, J. Garrity, H. A. Crosby, J. C. Escalante-Semerena, In Salmonella enterica, on two acidic proteins, A 1 and A 2, isolated from 50-S ribosomes of Escherichia coli. the sirtuin-dependent protein acylation/deacylation system (SDPADS) maintains en- Eur. J. Biochem. 25,13–19 (1972). ergy homeostasis during growth on low concentrations of acetate. Mol. Microbiol. 80, 16. D. Pathak, A. H. Bhat, V. Sapehia, J. Rai, A. Rao, Biochemical evidence for relaxed 168–183 (2011). substrate specificity of Nα-acetyltransferase (Rv3420c/rimI) of Mycobacterium tuber- 27. C. S. Hwang, A. Shemorry, A. Varshavsky, N-terminal acetylation of cellular proteins culosis. Sci. Rep. 6, 28892 (2016). – 17. V. F. Smith, B. L. Schwartz, L. L. Randall, R. D. Smith, Electrospray mass spectrometric creates specific degradation signals. Science 327, 973 977 (2010). investigation of the chaperone SecB. Protein Sci. 5, 488–494 (1996). 28. H. Jörnvall, Acetylation of protein N-terminal amino groups structural observations – 18. L. M. Okkels et al., CFP10 discriminates between nonacetylated and acetylated ESAT-6 of on alpha-amino acetylated proteins. J. Theor. Biol. 55,1 12 (1975). Mycobacterium tuberculosis by differential interaction. Proteomics 4, 2954–2960 (2004). 29. B. Persson, C. Flinta, G. von Heijne, H. Jörnvall, Structures of N-terminally acetylated 19. A. Mediero, M. Perez-Aso, B. N. Cronstein, Activation of EPAC1/2 is essential for os- proteins. Eur. J. Biochem. 152, 523–527 (1985). teoclast formation by modulating NFκB nuclear translocation and actin cytoskeleton 30. D. G. Christensen et al., Identification of novel protein lysine acetyltransferases in rearrangements. FASEB J. 28, 4901–4913 (2014). Escherichia coli. MBio 9, e01905–e01918 (2018). MICROBIOLOGY

Parks and Escalante-Semerena PNAS Latest Articles | 7of7 Downloaded by guest on September 27, 2021