A universal stress protein in Mycobacterium smegmatis sequesters the cAMP-regulated lysine acyltransferase and is essential for biofilm formation

Abstract

Universal stress proteins (USPs) are present in many bacteria, and their expression is enhanced under various environmental stresses. We have previously identified a USP in Mycobacterium smegmatis that is a product of the msmeg_4207 gene and is a substrate for a cAMP-regulated protein lysine acyltransferase (KATms; MSMEG_5458). Here, we explored the role of this USP (USP⁴²⁰⁷) in M. smegmatis and found that its gene is present in an operon that also contains genes predicted to encode a putative tripartite tricarboxylate transporter (TTT). Transcription of the TTT-usp⁴²⁰⁷ operon was induced in the presence of citrate and tartrate, perhaps by the activity of a divergent histidine kinase-response regulator gene pair. A usp⁴²⁰⁷-deleted strain had rough colony morphology and reduced biofilm formation compared with the WT strain; however, both normal colony morphology and biofilm formation were restored in a Δusp⁴²⁰⁷Δkatms strain. We identified several proteins whose acetylation was lost in the Δkatms strain, and whose transcript levels increased in M. smegmatis biofilms along with that of USP⁴²⁰⁷, suggesting that USP⁴²⁰⁷ insulates KATms from its other substrates in the cell. We propose that USP⁴²⁰⁷ sequesters KATms from diverse substrates whose activities are down-regulated by acylation but are required for biofilm formation, thus providing a defined role for this USP in mycobacterial physiology and stress responses.

Introduction

Universal stress proteins (USPs)⁵ are found in bacteria, archaea, fungi, and plants (1). Their name derives from the fact that many of these proteins are induced under various stresses such as oxidative stress, salt stress, nutrient starvation, heat shock, and insult by DNA-damaging agents (2, 3). USPs are classified into two groups based on their ability to bind ATP (UspFG-like), whereas UspA-like proteins do not bind ATP (4). Although a number of these USP proteins have been characterized biochemically and structurally (4, 5), the roles of the majority of these USPs remain unknown, with many usp genes being nonessential for regular bacterial growth (6).

Mycobacteria harbor genes for a number of USP-like proteins, and many of them are conserved across both fast- and slow-growing pathogenic mycobacteria, such as Mycobacterium tuberculosis (5–7). In a manner similar to that reported in other bacteria, USP-like genes are induced under different stresses, and Rv2623 has been shown to regulate growth of M. tuberculosis and to establish latent infection (7). We have shown that Rv1636 and its ortholog MSMEG_3811 bind cAMP and ATP and could serve as a sink for cAMP in these organisms (5).

Protein acylation on lysine residues is an important post-translational modification that occurs in bacteria (8–10), and a number of proteins have been shown to be modified in proteome-wide analyses in many bacteria (11, 12). The proteins that are acylated include those involved in central carbon metabolism, transcription, fatty acid metabolism, and cell wall functions (10, 13–17). A number of bacteria harbor multiple enzymes of the GNAT family that serve to acylate proteins (15), suggesting an inbuilt redundancy. However, individual GNATs can also acylate only specific proteins (15). Bacteria also express promiscuous sirtuin-like deacetylases (9, 18–20) that allow dynamic changes in protein acylation (21), depending on the physiological requirements of the cell.

Protein lysine acetylation in mycobacteria was first identified by the biochemical characterization of a cAMP-dependent protein lysine acyltransferase (KAT; to distinguish it from N-terminal acetylation) (8). The fusion of a cAMP-binding domain to a GNAT-like acyltransferase domain in a single protein is found only in mycobacterial species (22) and could have evolved due to the abundance of genes involved in cAMP generation (adenylyl cyclases) in both pathogenic and nonpathogenic mycobacteria (23, 24). We identified the MSMEG_4207 USP (USP⁴²⁰⁷) as a substrate for the KAT found in Mycobacterium smegmatis (KATms) (8), thus providing a novel link between protein acetylation and USPs. A number of additional protein substrates were subsequently identified for KATms and KATmt, and they include those involved in metabolic pathways, fatty acid degradation, or transcription (13, 14, 17, 19).

The identification of USP⁴²⁰⁷ as a substrate for KATms utilized pulldown approaches, using GST–KATms as bait and crude cytosol from M. smegmatis (8). The presence of significant amounts of USP⁴²⁰⁷ interacting with GST–KATms indicates its relatively high expression in M. smegmatis. Nevertheless, the tight association of USP⁴²⁰⁷ with KATms is surprising, because enzyme–substrate complexes are rarely identified by pulldown approaches. We argued at that time that tight binding of USP⁴²⁰⁷ may preclude the binding of other substrates to KATms, and thus USP⁴²⁰⁷ may act as a regulator of protein acylation brought about by KATms (8).

In this study, we explore the role of USP⁴²⁰⁷ as a modulator of KATms-mediated acylation. We show that usp⁴²⁰⁷ is part of a tricarboxylic acid-regulated operon, and a strain deleted for usp⁴²⁰⁷ is defective in biofilm formation. By identifying substrates for KATms, and observing that deletion of katms in the Δusp⁴²⁰⁷ strain restores biofilm formation, we propose that USP⁴²⁰⁷ is the first proteinaceous inhibitor identified for a GNAT-like enzyme, and it has evolved specifically to regulate KATms activity in the cell.

Results

usp⁴²⁰⁷ and its gene neighborhood

The usp⁴²⁰⁷ gene is predicted to be in an operon (Fig. 1A) based on overlapping stop and start codons in the genome sequence. Using primers across the junctions of genes in the predicted operon, we obtained products across all junctions, suggesting that msmeg_4210 to msmeg_4207 could be made as a single RNA (Fig. 1B). Analysis of the sequences of these genes in NCBI predicted their protein products to be members of a tripartite tricarboxylate transporter (TTT) family (25). MSMEG_4210 may be a secreted protein that could bind the tricarboxylic acid in the periplasmic space of the cell. Indeed, bioinformatic analysis of its amino acid sequence (http://www.cbs.dtu.dk/services/SignalP/)⁶ predicted a signal peptide cleavage site after residue 24. Two integral membrane proteins are predicted to be the products of the msmeg_4209 and msmeg_4208 genes. MSMEG_4208 is predicted to have 12 transmembrane-spanning domains and along with MSMEG_4209 (with four predicted transmembrane domains) could serve as a channel to transport tricarboxylic acids into the cell. TTT operons have been characterized from other bacteria, where a similar three gene arrangement exists (26–28). Interestingly, adjacent to this operon and divergently oriented is another operon comprising a putative histidine kinase (msmeg_4211) and a response regulator (msmeg_4212) (Fig. 1A) that may function as a two-component system that could regulate the msmeg_4207 operon, with the histidine kinase serving as a sensor for citrate (29, 30). Occurrence of such a two-component system regulating the expression of TTT operons is also common (25, 31).

Figure 1.

usp⁴²⁰⁷ is in an operon with a putative tripartite TTT and adjacent to a two-component signaling system. A, schematic of the arrangement of the TTT-usp⁴²⁰⁷ operon and the two-component system. Arrows indicate primers used for amplification across the junctions of two genes as annotated in the genome. Indicated above the black arrows are suggested functions of the genes encoded in the operon, as well as the two-component system that is divergently transcribed. Indicated by an open arrow is the intergenic region that was considered as a putative promoter (∼100 bp) for the TTT-usp⁴²⁰⁷ operon, and in a reverse orientation is the promoter for the two-component system. B, reverse transcription (RT) and PCR of RNA prepared from logarithmic growing cultures of M. smegmatis, using primers across the junctions of the putative TTT-usp⁴²⁰⁷ and two-component system operons. Data shown are representative of RNA prepared from three independently grown cultures. Numbers above the lanes containing a product indicate the predicted sizes of amplicons. RNA was either subjected to RT (+) or not (−) prior to PCR. C, alignment of the USP⁴²⁰⁷ sequence across actinobacteria. Highlighted in red is the lysine residue that is acylated by KATms.

Orthologs of USP are found in a number of fast-growing mycobacteria, and the lysine residue that is acetylated in USP⁴²⁰⁷ (8) is conserved (Fig. 1C).

usp⁴²⁰⁷ operon is induced by tricarboxylic acids

None of the genes in the usp⁴²⁰⁷ operon or the two-component regulatory system have been characterized earlier in any mycobacterial species, and therefore we asked whether tricarboxylic acids, such as citrate, could regulate the expression of the usp⁴²⁰⁷ operon. Citrate was indeed able to increase by 10-fold the expression of the usp⁴²⁰⁷ operon (Fig. 2A). Moreover, USP⁴²⁰⁷ protein levels were also increased in the presence of citrate (Fig. 2B). Tartrate could also increase levels of USP⁴²⁰⁷ in cells, whereas another dicarboxylic acid, succinate, failed to do so (Fig. 2B). Therefore, it appears that C4 dicarboxylates with a hydroxyl group at the C2 position serve as activators of the operon, and the presence of a third carboxyl group is not essential for operon induction. Thus, proteins encoded in this operon may serve to transport both tri- and specific dicarboxylic acids into the cell. A similar operon in Corynebacterium is induced in the presence of citrate, and its protein products serve in transporting citrate into the cell (31). However, the ability to transport tartrate was not tested.

Figure 2.

TTT-usp⁴²⁰⁷ operon is induced by tricarboxylic acids. A, reverse transcription and PCR of RNA prepared from cultures grown in the presence of 0.1 g/liter sodium citrate (low) or 2 g/liter sodium citrate (high). Primers were designed to amplify the genes indicated, and primers to the 16S gene were used for normalization in real-time PCRs. usp⁴²⁰⁷ and other genes in the operon (msmeg_4210-msmeg_4208) are induced by high levels of citrate. Experiments are representative of RNA prepared from three independent experiments. Data were analyzed by the t test. ***, p < 0.001. B, Western blot analysis using USP⁴²⁰⁷-specific antisera (8) and lysates prepared from M. smegmatis cells grown in the presence of high or low amounts of citrate and indicated concentrations of di- or tricarboxylic acids. Normalization of protein loading in the Western blotting was indicated by equivalent levels of CRP in lysates, using a CRP-specific antiserum (62). Data shown are representative of three independently grown cultures. C, ∼150-bp intergenic region between msmeg_4210 and msmeg_4211 genes was cloned in two orientations upstream of the luxAB gene, and plasmids were electroporated into M. smegmatis. Luciferase activity was measured at the indicated times in cultures grown in low (white bars) or high levels (black bars) of sodium citrate. Citrate induction was observed only when the promoter was cloned in an orientation that would drive transcription of the TTT-usp⁴²⁰⁷ operon (left panel) and not of the two-component system (right panel). Values shown are from two independent experiments, with each sample taken in duplicate, and are the mean ± S.D.

mRNA levels of katms (msmeg_5458) or two sirtuin-like genes (msmeg_4620 and msmeg_5175) were unaltered in the presence of citrate (Fig. 2A), indicating that the availability of large amounts of USP⁴²⁰⁷ could result in an increase in absolute levels of acetylated USP in the cell, when intracellular levels of citrate are high.

We cloned the intergenic region between msmeg_4210 and msmeg_4211 (∼103 bp) upstream of luciferase in two orientations and monitored luciferase activity in the presence and absence of citrate. As shown in Fig. 2C, luciferase activity was increased in the presence of citrate, in an orientation that would lie 5′ to the usp⁴²⁰⁷ operon. In a direction that would lie 5′ to the two-component system, luciferase activity remained constant in the presence and absence of citrate (Fig. 2C), suggesting constitutive expression of msmeg_4211 and msmeg_4212. This was in agreement with data shown in Fig. 2A, where RNA levels of these two genes were unchanged in the presence of citrate. Therefore, this intergenic region serves as a promoter for both the TTT-usp⁴²⁰⁷ operon and the two-component system.

Based on these findings, and as observed in Corynebacterium glutamicum (26), it is possible that citrate may bind to the periplasmic domain of the adjacent histidine kinase, MSMEG_4211. This would activate the kinase and allow phosphorylation of the response regulator MSMEG_4212, because most cognate histidine kinase-response regulator pairs are found in operons (Fig. 1B) (32). MSMEG_4212 may therefore bind to a region upstream of msmeg_4210. To test this possibility, we cloned, expressed, and purified the putative response regulator (Fig. 3A) and performed EMSAs with a radiolabeled fragment representing the promoter. As shown in Fig. 3B, increasing concentrations of the purified MSMEG_4212 showed a series of shifts in the EMSA, perhaps indicating different oligomers of the protein binding to the intergenic region, or the presence of multiple DNA-binding sites.

Figure 3.

MSMEG_4212 binds to the intergenic region between msmeg_4210 and msmeg_4211. A, putative response regulator encoded by the msmeg_4212 gene was expressed in E. coli and purified (left panel). A mutant of the response regulator where an aspartate residue was converted to an alanine was also expressed and purified. B, varying concentrations of purified proteins were used to monitor binding to the radiolabeled ∼100-bp promoter fragment by EMSA (2nd lane). Increasing amounts of the WT or D54A mutant protein (as indicated above the gels) used were 7.5, 15, 30, 50, 100, and 500 ng and 1, 2, and 4 μg. The filled arrow shows the shifted radiolabeled band representing the protein-bound fraction. The open arrow marks the mobility of the free probe. C, sequence between the start of msmeg_4210 and msmeg_4211 is shown. The sequences representing the two PCR products generated to identify the binding region of MSMEG_4212 are shown as dotted and filled lines below the sequence. Arrows represent the positioning of adjacent genes to the promoter region. Note that the two PCR products contain a 20-bp overlapping region. D, WT (300 ng) or mutant MSMEG_4212 (50 ng) were incubated with the radiolabeled intergenic region in the absence or presence of the unlabeled entire intergenic region or PCR products representing two halves of the intergenic region. The gel shown is representative of experiments performed twice with two different batches of purified protein.

It is important to note that we used MSMEG_4212 directly as purified from Escherichia coli. This may either indicate that the nonphosphorylated form of the response regulator can bind to the region upstream of msmeg_4210 or that the response regulator is phosphorylated as purified from E. coli (33). To test this, we aligned the sequence of MSMEG_4212 with OmpR from E. coli, where the aspartate residue that is phosphorylated has been identified (34). We mutated residue Asp-54 to an alanine and purified the mutant protein. Interestingly, even low concentrations of the D54A mutant protein bound to the probe and showed only a single shift (Fig. 3B). This indicated that the nonphosphorylated form of MSMEG_4212 bound to the intergenic sequence with a higher affinity than the WT protein and formed a complex comprising a single oligomeric species.

To confirm that the shift seen in the EMSA was specific, and also to identify a smaller region that could serve as the binding site for MSMEG_4212, we amplified two fragments, shown as dotted and filled lines in Fig. 3C. We used these purified PCR products, as well as the full-length intergenic sequence, to determine whether they could compete for binding of the radiolabeled intergenic region to MSMEG_4212 and the D54A mutant protein. As shown in Fig. 3D, the entire intergenic region could inhibit binding of the radiolabeled probe to both complexes seen with the WT protein and also the single complex seen with the mutant protein. The PCR product closer to msmeg_4211 could completely inhibit binding of MSMEG_4212 and the D54A mutant to the intergenic fragment. The product closer to msmeg_4210 showed a slight degree of competition, as seen by the enhanced intensity of the faster-moving complex with the WT protein. In the case of the D54A mutant protein, there was a marginal decrease in intensity of the shifted band. We therefore suggest that MSMEG_4212 could act as a regulator of the msmeg_4210-msmeg_4207 operon by binding to residues in the intergenic region proximal to msmeg_4211. It is unclear at present whether there are two binding sites in the intergenic region, because the PCR products contained an ∼20-bp overlapping region. However, affinity binding of mutant MSMEG_4212 suggests that in the absence of citrate, unphosphorylated MSMEG_4212 could act as a repressor of the usp⁴²⁰⁷ operon.

Deletion of msmeg_4207 alters colony morphology

We have earlier shown that USP⁴²⁰⁷ is acetylated on a single lysine residue by KATms in M. smegmatis (8). What is the role of USP⁴²⁰⁷ and its acetylation in the cell? To study this, we deleted the usp⁴²⁰⁷ gene by homologous recombination (35) and confirmed deletions using PCR (Fig. 4A). Therefore, to test the role of acetylation of USP⁴²⁰⁷, we also generated a usp⁴²⁰⁷ deletion strain in the background of a Δkatms strain. We complemented the deletion strains with either WT usp⁴²⁰⁷ or with a mutant usp⁴²⁰⁷, where Lys-104 (the site for acetylation) was mutated to Arg. WT and mutant strains grew similarly in shaking cultures (data not shown). However, as shown in Fig. 4B, colonies formed by the Δusp⁴²⁰⁷ strain were smaller and more compact, with a rougher surface than those formed by WT cells. This altered morphology was less evident when Tween 80 was included in the agar plates, suggesting that alterations in the cell wall glycopeptidolipids (GPLs) and cell–to–cell interactions may determine the morphology of the colony (36). It is possible that this was a consequence of aberrant acylation and inhibition of proteins/enzymes, which could include those involved in GPL synthesis (37, 38). Complementation with WT USP as well as the K104R mutant protein restored normal colony morphology, indicating that the presence of USP protein and not the absence of acetylated USP was responsible for normal colony morphology (Fig. 4B). Although the Δkatms strain showed no change in colony morphology, and a <10% increase in USP levels (Fig. 4C) when normalized to CRP across duplicate experiments, deletion of katms in the Δusp⁴²⁰⁷ strain restored normal colony morphology and smooth appearance (Fig. 4B). This indicated that in the absence of USP⁴²⁰⁷ and the presence of KATms, and presumably its catalytic activity, altered colony appearance.

Figure 4.

Δusp⁴²⁰⁷ strain shows altered colony morphology. A, PCR analysis of genomic DNA prepared from WT M. smegmatis and strains deleted for usp⁴²⁰⁷ in the WT strain (Δusp⁴²⁰⁷) or in a strain where katms has been deleted (Δusp⁴²⁰⁷Δkatms). Primer pairs used for the genes are indicated and confirm the specific gene deletions. B, cultures of the indicated strains were grown and resuspended to an A₆₀₀ of 0.5 and 5 μl of each culture spotted on 7H10-agar plates in the absence or presence of 0.05% Tween 80. Δusp⁴²⁰⁷ strains were transformed with a plasmid expressing either WT USP⁴²⁰⁷ or the USP⁴²⁰⁷K104R mutant protein driven by the sigA promoter to generate complemented strains (Δusp⁴²⁰⁷/usp⁴²⁰⁷ or Δusp/usp⁴²⁰⁷K104R_, respectively). Data shown are representative of at least three independently grown cultures. C, Western blot analysis to determine expression levels of USP⁴²⁰⁷ in the WT and Δkatms strains using USP⁴²⁰⁷-specific antibody. Strains were grown independently at least three times, and the blot shown is representative of lysates prepared from paired cultures.

USP⁴²⁰⁷ forms a stable complex with KATms and is the preferred substrate

USP⁴²⁰⁷ is an abundantly-expressed protein in M. smegmatis, as evidenced in our earlier experiments using GST–KATms as bait (8). We directly tested the stability of the complex formation between KATms and USP⁴²⁰⁷ in pulldown experiments. Indeed, both WT and USP⁴²⁰⁷K104R could interact with GST–KATms bound to GSH beads, indicating that the mutation in USP⁴²⁰⁷ did not alter binding of USP⁴²⁰⁷K104R to KATms (Fig. 5A). Therefore, in the cell, we hypothesized that USP⁴²⁰⁷K104R could serve to sequester KATms and regulate the extent of acetylation of other substrates, thereby providing a mechanistic explanation for the rescue by USP⁴²⁰⁷K104R of the small-colony phenotype that was seen on deletion of usp⁴²⁰⁷ (Fig. 4B).

Figure 5.

USP⁴²⁰⁷ binds to KATms and competitively inhibits acetylation of Acs. A, GST–KATms pulldown of USP⁴²⁰⁷. Lysates prepared from E. coli strains expressing either WT USP⁴²⁰⁷ or the USP⁴²⁰⁷K104R mutant were interacted with GST–KATms or GST bound to GSH beads. Protein bound to beads was analyzed either by Western blotting using the USP⁴²⁰⁷-specific antibody or staining the membrane with Coomassie to normalize for amounts of GST–KATms and GST used for the interaction. The experiment was performed at least three times with independent batches of purified proteins. Data shown are from one representative experiment. B, USP⁴²⁰⁷ acts a competitive inhibitor of KATms. Acetylation assays were performed using purified KATms (15 pmol) and Acs (75 pmol) in the absence or presence of indicated concentrations of USP⁴²⁰⁷. Samples were subjected to Western blot analysis using an acetyl-lysine–specific antibody or an antibody to hexahistidine (anti-His) found at the N terminus of the purified proteins to normalize for Acs amounts taken for assay. The graph shows quantitation of acetylated Acs as a fraction of that seen in the absence of USP. The x axis is the mole ratio of USP⁴²⁰⁷ to Acs in the reaction tube. The experiment shown is representative of two assays performed with different batches of purified protein.

To confirm that such substrate competition could occur, we monitored the acetylation of Acs, a known substrate of KATms (20), in the presence and absence of varying concentrations of USP, in vitro. As shown in Fig. 5B, acetylation of acetyl-CoA synthase (Acs) was steadily reduced in the presence of increasing concentrations of USP⁴²⁰⁷. Given the high expression levels of USP⁴²⁰⁷ in the cell, we propose that M. smegmatis can modulate acylation levels, and thereby the activity, of different substrates of KATms, depending on the levels of USP in the cell.

Deletion of usp⁴²⁰⁷ inhibits biofilm formation

We then asked what the potential substrates for KATms could be that may remain poorly acylated in the presence of USP⁴²⁰⁷. We prepared lysates from planktonic cultures of isogenic mutant strains of M. smegmatis and subjected them to Western blot analysis with antibodies to acetyl-lysine. As seen in Fig. 6A, a band of ∼80 kDa was absent in the Δkatms strain. Acetylation of this protein (or proteins) was similar in strains that harbored a WT copy of katms, and in the Δusp⁴²⁰⁷strain, equivalent to that seen in lysates prepared from WT M. smegmatis. Interestingly, acetylation of the ∼80-kDa band was markedly reduced (∼3-fold) in the Δusp⁴²⁰⁷ strain complemented with the usp⁴²⁰⁷K104R mutant. When we checked the expression levels of USP protein in different strains, it was evident that levels of USP⁴²⁰⁷K104R were much higher (∼5-fold) in the complemented strain (Fig. 6A), perhaps due to enhanced stability of the nonacetylated form of USP. Therefore, the reduced acetylation seen in the protein band at ∼80 kDa supports the hypothesis that USP⁴²⁰⁷ could indeed modulate the extent of acetylation of substrates of KATms in the cell, although its own acetylation state was dispensable. Enhanced acetylation in the Δusp⁴²⁰⁷ strain was not observed, perhaps due to the limiting amounts of KATms substrates in the cell, which remain fully acetylated even in WT cells.

Figure 6.

Identification of substrates of KATms and biofilm formation in the Δusp⁴²⁰⁷ strain. A, samples prepared from planktonic cultures of the indicated strains were subjected to Western blot analysis using either an antibody to acetyl-lysine, USP⁴²⁰⁷, or CRP (to normalize for protein loading). Blots were repeated three times with independently-grown cultures. USP⁴²⁰⁷ expressed in the complemented strains contained a hexahistidine N-terminal tag, so the proteins migrate at a higher molecular weight than USP⁴²⁰⁷ in the WT strain. The lower molecular weight band in the complemented strains could represent protein in which the hexahistidine tag had been cleaved within the bacteria. Parallel gels were run, and 0.5-cm pieces corresponding to a region of ∼80 kDa was cut from the lanes containing lysates of WT or Δkatms strains. The gel piece was analyzed by MS to identify acetylated peptides found in the WT strain but absent in the Δkatms strain. Peptides thus identified are shown in the table below with the acetylated lysine residue (K) in bold and italicized. The proteins in which these peptides are present are indicated. B, strains as indicated were grown in 7H9 media containing 0.2% glycerol, in the absence of Tween 80, and biofilm was allowed to develop over 3 days. Pictures are shown from the top (left-hand side) and from the side of the individual dishes (right-hand side). Biofilm was formed in all but the Δusp⁴²⁰⁷ strains. The experiment was performed four times with independently grown cultures. C, lipids were extracted from cultures after formation of the biofilm or from flocculent cells from the Δusp⁴²⁰⁷ strain. Lipids were spotted after normalizing for protein content present in cells taken for lipid extraction. GPLs were resolved by performing one-dimensional TLC using chloroform/methanol (100:7; v/v) as the solvent system and detected by spraying the plate with 0.1% orcinol in 40% sulfuric acid.

The prominent acetylated bands we observed here, which were absent in a Δkatms strain, were identified in an earlier report utilizing similar biochemical approaches (19), and included propionyl-CoA synthetase (PrpE) and Acs. PrpE and Acs were also shown to be acetylated in a global acylome analysis of M. smegmatis (39).

We cut out a 0.5-cm piece of the gel corresponding to the region that contained proteins that were acetylated in planktonic cultures and subjected the gel piece to tryptic digestion and MS. We specifically looked for additional peptides whose acetylation status changed in the Δkatms strain. We were able to identify the same peptides in Acs and PrPE as reported earlier, but we also identified a peptide from MSMEG_0401 (product of mps) that is part of the cluster of genes involved in mycobacterial GPL synthesis (40–42). Deletion of genes in this cluster results in a rough colony phenotype (40, 41, 43), similar to that seen in the Δusp⁴²⁰⁷ strain.

In a proteomic analysis of the biofilm pellicle in M. smegmatis, propionyl-CoA synthetase levels were reported to be increased (44). Moreover, alteration in GPL biosynthesis has been shown to affect biofilm formation (38, 43). It is therefore possible that the synthesis of GPLs is modulated by acylation during biofilm formation and consequently are dependent on the presence of USP and KATms in the biofilm. Indeed, biofilm formation was dramatically lost in the Δusp⁴²⁰⁷ strain, and cells became flocculent and dispersed (Fig. 6B). Biofilm formation was restored and marginally enhanced in the usp⁴²⁰⁷ complemented strains. Interestingly, deletion of katms in the background of Δusp⁴²⁰⁷ restored biofilm formation, whereas deletion of katms alone had no effect on biofilm formation. Moreover, complementation with the K104R mutant of usp⁴²⁰⁷ was able to restore the biofilm formation, emphasizing again that association of USP with KATms, and not acylation of USP, is sufficient for preventing acylation by KATms of its other substrates.

To confirm that deletion of usp⁴²⁰⁷ altered the levels of GPLs in the biofilm, we prepared lipids from strains that formed a biofilm and from flocculent cells seen in the Δusp⁴²⁰⁷strain (Fig. 6C). GPLs were dramatically reduced in the usp⁴²⁰⁷ strain, but restored on deletion of katms, concomitant with biofilm formation. Therefore, it appears that GPLs play a critical role in biofilm formation, and their production is regulated by the presence of KATms and levels of USP.

We monitored transcript levels of prpE, katms, mps, the sirtuin genes, and usp⁴²⁰⁷ in the pellicle and planktonic cultures. Levels of transcripts of usp⁴²⁰⁷, prpE, mps, and msmeg_4620 were found to be higher (∼3–50-fold) in the pellicle than in planktonic cells. A modest increase in katms or sirtuin gene expression (∼1.5-fold) was observed (Fig. 7A). Comparing the qPCR efficiency of all primer pairs for individual genes, and generating a standard curve using known amounts of genomic DNA (data not shown), we could estimate the amount of gene-specific cDNA in samples taken for qPCR (Fig. 7B). Based on this analysis, mRNA levels of usp⁴²⁰⁷ were higher than other transcripts, which could reflect a higher concentration of protein, assuming half-lives of protein and mRNA are similar across the genes tested. We could estimate levels of USP⁴²⁰⁷ by performing quantitative Western blot analysis using anti-USP⁴²⁰⁷ antibodies and defined concentrations of USP. Indeed, USP⁴²⁰⁷ protein levels in the pellicle reached ∼10 μg/mg protein, whereas levels were ∼1 μg/mg protein in planktonic cultures (Fig. 7C).

Figure 7.

USP⁴²⁰⁷ is up-regulated in the biofilm and regulates protein acetylation in the pellicle. A, RNA was prepared from planktonic cultures of M. smegmatis and from the biofilm pellicle. Levels of mRNA transcripts were monitored following reverse transcription and real-time PCR, with 16S rRNA levels used for normalization. Cultures were grown independently. Data shown are from three independently grown cultures. *, p < 0.05; **, p < 0.01; ***, p < 0.001. B, gene-specific primer efficiency was determined using varying amounts of DNA and was used to estimate the amount of specific cDNA in planktonic and pellicle fractions. Data are shown in mean ± S.D. of three independent experiments. *, p < 0.05; **, p < 0.01; ***, p < 0.001. C, Western blot analysis was performed on lysates prepared from planktonic cultures and pellicles of biofilm formed by WT M. smegmatis. CRP was used to check whether its expression changed in the pellicle or planktonic cultures. The blot shown is representative of two independent experiments. D, samples were prepared from the pellicle (or flocculent cells in the case of the Δusp⁴²⁰⁷). Samples were subjected to Western blot analysis with antibodies against acetyl-lysine or CRP. Note the absence of a band of ∼80 kDa from pellicles of WT and Δkatms strains. E, M. smegmatis cells were harvested from planktonic cultures and at the times indicated following pellicle formation. cAMP was measured by radioimmunoassay. Cultures were grown twice, and extracts were prepared independently for measurement. (**, p < 0.01; Student's t test).

Our results therefore strongly suggest that proteins/enzymes that may be required for biofilm formation (such as PrpE and MSMEG_0401) should remain nonacylated in order to remain active. The high concentrations of USP⁴²⁰⁷ could serve to sequester KATms from these and other substrates. In the absence of USP in a KATms-expressing strain (Δusp⁴²⁰⁷), increased acylation of these proteins required for biofilm formation would occur, thus inhibiting their activity and affecting biofilm formation.

To provide support for this, we performed Western blot analysis of lysates prepared from the pellicle of WT, Δusp⁴²⁰⁷, Δkatms and complemented strains. As shown in Fig. 7D, the major acetylated band at ∼80-kDa band was absent in pellicle fractions of all strains that formed the biofilm. However, fractions prepared from flocculent cells formed by the Δusp⁴²⁰⁷ strain showed a prominent band at ∼80 kDa, corresponding to the band seen in planktonic cultures (Fig. 6A). Therefore, reduced acetylation of proteins such as PrpE and Acs is critical for biofilm formation. Interestingly, transcript levels of the sirtuin genes increased in the biofilm (Fig. 7A), suggesting that acylation of proteins in general may be detrimental to biofilm formation.

The blot also revealed changes in KATms-independent protein acetylation in different strains. A band of 140 kDa is absent in the Δusp⁴²⁰⁷ strain but present in the Δkatms strain. 55-, 65-, and 110-kDa proteins remain acetylated. Therefore, protein acetylation seems to be exquisitely regulated during biofilm formation.

Attempts to normalize protein loading using CRP antiserum revealed that, intriguingly, CRP levels were markedly reduced in the Δusp⁴²⁰⁷ strain (Fig. 7D). Nevertheless, the increased intensity of the band at ∼80 kDa and equivalent intensity of other bands allowed us to conclude that KATms-mediated acetylation is indeed regulated by high levels of USP⁴²⁰⁷ during biofilm formation.

cAMP levels increase during biofilm formation

Our hypothesis related to the role of USP⁴²⁰⁷ in regulating KATms activity arises from genetic analysis that showed that the presence of KATms in a Δusp⁴²⁰⁷ strain was inhibitory to biofilm formation. Because KATms activity can be enhanced in the presence of cAMP, we asked whether levels of cAMP were altered in the pellicle. We found that, in fact, cAMP levels were increased during biofilm formation (Fig. 7E). RNA levels of usp⁴²⁰⁷ were increased in the pellicle (Fig. 7A), suggesting transcriptional up-regulation of the operon, perhaps by elevation of citrate and other tricarboxylates during biofilm formation.

In summary, cells have devised a means to regulate KATms activity in the presence of cAMP by increasing USP⁴²⁰⁷ levels, rather than by altering KATms levels. cAMP may be required during biofilm formation to modulate the activities of other cNMP-binding proteins, including CRP. The presence of USP⁴²⁰⁷, serving as a substrate and “inhibitor” of KATms, could bring in a high degree of specificity in terms of regulating the activity of a specific cNMP-binding protein (i.e. KATms) during biofilm formation, following the elevation of cAMP.

Discussion

In this study, we propose a novel mechanism of regulation of protein acylation, that is by cells providing a proteinaceous competitive inhibitor of an acyltransferase. The inhibitor in this case is a universal stress protein, which also serves as a substrate for a cAMP-regulated acyltransferase. As shown in Fig. 8A, usp⁴²⁰⁷ is part of an operon containing genes that probably encode a TTT, because expression of genes in that operon are up-regulated in the presence of citrate. We also propose that the divergent two-component, histidine kinase-response regulator gene pair encodes a signaling system to transcriptionally activate the TTT-usp⁴²⁰⁷ operon in the presence of tricarboxylic acids. The putative response regulator can bind in a nonphosphorylated state to a region present between the TTT-usp⁴²⁰⁷ operon and the genes of the two-component signaling system (Fig. 3B), suggesting that it could serve as a repressor, or allow only weak transcription, under conditions when citrate levels are low. The multiple concentration-dependent shifts seen with the WT protein (Fig. 3B) could suggest that phosphorylation of the WT protein may stabilize oligomers that bind to the DNA, without altering DNA binding affinity or specificity (45), and/or represent binding of the WT protein to two sites in the intergenic region. The single shift seen with the D54A mutant protein suggests that this protein forms a stable oligomer following mutation or adopts a DNA-binding conformation more readily (46). In a manner similar to that proposed for CpxRA from Legionella pneumophila (47), MSMEG_4212 may adopt different oligomeric states that bind to different sequences contained in the intergenic region to either activate or repress transcription of the usp⁴²⁰⁷operon. Mutational analysis of this intergenic region could determine the recognition motif for this newly identified response regulator.

Figure 8.

Proposed role of USP⁴²⁰⁷ in regulating KATms-mediated acylation. A, predicted regulation of the TTT-usp⁴²⁰⁷ operon by the divergent histidine kinase-response regulator gene pair. The TTT-usp⁴²⁰⁷ operon is regulated by tricarboxylic acids such as citrate. Citrate can bind to the secreted protein (encoded by the msmeg_4210 gene represented by a brown arrow) or directly to the histidine kinase (MSMEG_4211), resulting in its phosphorylation. Phosphotransfer to the response regulator may prevent binding of the response regulator to the promoter of the TTT-usp⁴²⁰⁷ operon, thereby increasing transcription of the three genes that encode the tripartite transporter (two membrane proteins encoded by genes represented in purple (msmeg_4208) and red (msmeg_4209) and the secreted protein). usp⁴²⁰⁷ (green) is the last gene in the operon. B, distribution of the usp⁴²⁰⁷ gene, kat orthologs, and presence of the TTT-usp⁴²⁰⁷ operon across actinobacterial members, including slow- and fast-growing mycobacteria. Green boxes represent the presence of a particular gene(s), and purple boxes represent the absence of the gene(s). To the right of the boxes is a graphical representation of the organization of the TTT-usp⁴²⁰⁷ operon in organisms where it is present. The two-component gene pair is shown in brown arrows, and the usp gene in a red arrow.

Proteins containing a cAMP-binding domain fused to a GNAT-like acyltransferase domain are found only in mycobacteria (8). We explored the distribution of USP⁴²⁰⁷-like proteins and the TTT operon in various genera of Actinobacteria, and we found that in mycobacteria, the presence of the TTT operon in fast-growing species is always seen along with genes encoding USP⁴²⁰⁷-like and KATms-like proteins (Fig. 8B). Therefore, we predict that in these mycobacterial species, the level of acylation of substrates of KATms-like proteins is modulated by USP⁴²⁰⁷ orthologs, as we report here in M. smegmatis. Slow-growing mycobacteria do not contain either a USP⁴²⁰⁷-like protein or a TTT operon. The acyltransferase activity of KATms is substantial even in the absence of cAMP (8). However, the KAT protein from M. tuberculosis (KATmt) obligatorily requires cAMP for its activation, and the structural basis for these differences has been elucidated (8, 13, 22, 48, 49). We therefore speculate that in slow-growing mycobacteria, where the KAT protein has been exquisitely tuned to respond to levels of cAMP, there is no requirement for additional levels of regulation, such as the presence of proteinaceous competitive inhibitors. Indeed, in preliminary pulldown experiments using GST–KATmt as a bait and lysates prepared from M. tuberculosis, no major protein could be identified that interacts with KATmt (data not shown), in contrast to our results with KATms (8). The levels of KAT, cAMP, and protein acylation during biofilm formation in slow-growing mycobacteria, such as in M. tuberculosis which lacks the USP⁴²⁰⁷ ortholog, remain to be explored.

Is there additional information on possible links between protein acylation, colony morphology, and biofilm formation that would validate our hypothesis on the interplay of USP⁴²⁰⁷ and KATms? In a study performed to characterize the acetylome of M. tuberculosis, a sirtuin-like gene (rv1151c) was deleted, which should result in hyperacetylation of proteins (11). It was observed that the Δrv1151c strain showed a rough colony morphology and did not form biofilm. These data agree with our findings in M. smegmatis where the deletion of usp⁴²⁰⁷, which in our model would result in high activity of KATms and therefore increased acylation of its substrates (as should be observed in a sirtuin-deletion strain), also altered colony morphology and biofilm formation (Figs. 4B and 6B).

The promiscuous nature of sirtuins poses a difficulty in deacetylating a specific subset of acetylated proteins (13, 18, 50, 51). In M. smegmatis, there exists 25 putative acyltransferases and 3 lysine deacetylases (15, 19). In M. smegmatis, we show that a specific USP⁴²⁰⁷ could regulate acylation of substrates by directly interacting with its cognate acyltransferase, KATms, thereby overcoming the lack of specificity of sirtuins to regulate the KATms-specific acylome.

The M. smegmatis gene cluster msmeg_0400-msmeg_0402 encodes a nonribosomal peptide synthase that is involved in synthesizing the tetrapeptides in GPLs that are found in nontuberculosis-causing mycobacteria as well as certain pathogenic mycobacteria (37, 38, 40, 43, 52). MSMEG_0401 is a member of the FadD enzyme family, some of which we have shown earlier are acetylated by KATmt (13). MSMEG_0401 harbors a lysine residue (residue 432) in the first condensation domain (https://pfam.xfam.org/)⁷ that catalyzes the formation of a peptide bond in nonribosomal peptide synthesis that could be the site for acetylation by KAT.

In Mycobacterium abscessus, rough, cording, and nonbiofilm phenotypes are associated with low GPL levels (53). Smooth biofilm-competent strains show increased GPL biosynthesis and higher expression of mps1. A KAT-like enzyme is present in this species (Fig. 8B) but resembles the KATmt protein that does not contain the PP-loop that contributes to the reduced requirement for cAMP binding to activate the acyltransferase domain of KATms (48). This may account for the absence of the USP⁴²⁰⁷-like protein-encoding gene in M. abscessus (Fig. 8B). In summary, we propose that in nontuberculosis-causing mycobacteria GPL biosynthesis is regulated by protein lysine acylation and, in turn, through the interplay between KATms, USP⁴²⁰⁷, and intracellular cAMP levels.

No TTT-like transporter has been characterized in mycobacteria, and citrate is usually added to media along with iron in culturing mycobacteria where it aids in the uptake of iron by the bacteria (54, 55). Our findings here have identified a putative transporter, based on the fact that genes in the operon are up-regulated by citrate, as is seen in other organisms where such TTT transporters have been characterized (25). However, in other bacteria, such as Salmonella typhimurium, the tricarboxylate-binding protein can interact with a variety of citrate analogs and organic acids (56). Therefore, in M. smegmatis, it is possible that the TTT operon is regulated by a number of substances that may be found in environments where M. smegmatis thrives. The presence of usp⁴²⁰⁷ in the TTT operon, and therefore linking protein acylation to cellular metabolism, is worth investigating in future.

USPs are found in many bacteria, and some have been shown to be important in regulating biofilm formation. For example, deletion of the uspA gene in Porphyromonas gingivitis resulted in poor biofilm formation, and the levels of UspA were increased during biofilm formation (57). A uspA gene in Staphylococcus epidermidis is up-regulated in planktonic aggregates that show biofilm-like behavior (58). Finally, deletion of a usp gene in Micrococcus luteus regulated the expression of a number of genes, including those involved in central carbon metabolism (59). Therefore, because the expression of various USPs is increased in a number of stress conditions, they may play a direct role in regulating genes involved in overcoming stress such as during infection (2), perhaps independently of a role in regulating protein acylation.

In conclusion, we have provided new information on the role of protein acylation in M. smegmatis, a link to colony morphology and biofilm formation, and the involvement of a universal stress protein in acting as a rheostat to tune levels of protein acylation in the cell. This complex interplay of small molecules like cAMP and citrate, unique enzymes such as KATms that can regulate the activities of enzymes as diverse as those involved in nonribosomal peptide synthesis and metabolism, and a widely-expressed small protein like USP⁴²⁰⁷ are depicted in Fig. 9. What remains unknown is how cAMP regulates biofilm formation and whether citrate levels are increased during biofilm formation. Importantly, it appears that usp⁴²⁰⁷-like genes are found exclusively in an operon with TTT–transporter-like genes in Actinobacteria, even in those genera that do not have a KAT-like enzyme (Fig. 8B). This raises the interesting possibility that USP⁴²⁰⁷ could regulate the activity of other GNAT-like enzymes present in bacteria that do not contain KAT-like enzymes, because of structural similarity among acyltransferase enzymes (60).

Figure 9.

Complex interplay between protein acylation, cAMP, and citrate in M. smegmatis biofilm formation. cAMP binds to MSMEG_3811, which is a ATP–cAMP-binding USP (5) and also to KATms. cAMP can regulate transcription via CRP (62). Given the abundance of putative cAMP-binding proteins in mycobacteria, cAMP is likely to have a number of additional roles, which are not known at present (indicated by a question mark). KATms can acylate USP⁴²⁰⁷ and also PrpE, Acs, and MSMEG_0401. Transcript levels of prpE and msmeg_0401 increase in biofilm pellicles, suggesting their requirement in biofilm formation. Acylated PrpE (19) and MSMEG_0401 are inactive. During biofilm formation, both cAMP and USP⁴²⁰⁷ levels are increased (Fig. 7). To prevent the unwanted acylation of PrpE, MSMEG_0401 (and perhaps other substrates of KATms), USP⁴²⁰⁷ sequesters KATms from its substrates. The thickness of the blue arrows represent the extent of enzymatic activity of the indicated enzymes, KATms and sirtuins, toward their substrates.

The action of USP⁴²⁰⁷ as a competitive inhibitor of additional protein substrates of GNAT-like enzymes is distinct from the inhibitory actions of small motifs that inhibit acetyl-CoA binding to histone acetyltransferases (61), for example. Indeed, our findings have more general implications when considering the activity of an enzyme against multiple substrates. The relative concentrations of different protein substrates and their varying affinities for the enzyme will determine which protein becomes modified by enzymatic activity in the cell. Therefore, promiscuity in enzyme–substrate interactions may actually provide an additional level of regulation within the cell, as we have shown here.

Experimental procedures

Mycobacterial strains, culture conditions and preparation of cell lysates

M. smegmatis mc² 155 (ATCC 700084) was grown at 37 °C in Middlebrook 7H9 broth (BD Biosciences) supplemented with 0.2% glycerol and 0.05% Tween 80 with shaking at 200 rpm or in Middlebrook 7H10 agar (BD Biosciences) supplemented with 0.5% glycerol. Wherever necessary, cells were grown in Sauton's medium containing 0.2% glycerol and 0.05% Tween 20 and further supplemented with 2 g/liter of citrate, succinate, or tartrate as required.

For the preparation of cell lysates, stationary-phase cells were harvested and washed with TBST containing 10 mm Tris-Cl (pH 8.2), 0.9% NaCl, and 0.1% Tween 80. Cells were then lysed by sonication in buffer containing 50 mm Tris-Cl (pH 8.2), 100 mm NaCl, 10% glycerol, 5 mm 2-mercaptoethanol (2-ME), 2 mm phenylmethylsulfonyl fluoride (PMSF), 1 mm benzamidine hydrochloride, 3 mm nicotinamide, and 1 μm trichostatin A, followed by centrifugation at 30,000 × g for 30 min at 4 °C. The supernatant was collected, and protein estimation was performed by the Bradford method.

RNA isolation and reverse transcription

Cells were harvested, resuspended in Tri Reagent (Sigma), and lysed by bead beating (BioSpec Products). The lysate was then heated at 65 °C for 5 min and centrifuged at 16,000 × g for 10 min at 4 °C. The supernatant was collected and mixed with chloroform followed by centrifugation at 16,000 × g for 15 min at 4 °C. The upper aqueous phase was collected, and RNA was precipitated with isopropyl alcohol. The RNA pellet was washed with 75% ethanol, dissolved in RNase-free Milli-Q water, and treated with RNase-free DNase (Thermo Fisher Scientific). 2 μg of RNA was used for reverse transcription using 200 units of reverse transcriptase (Thermo Fisher Scientific). Sequences of primers used to study the transcript level of different genes are shown in Table 1. Real-time PCR was performed using SYBR Premix Ex Taq (Tli RNase H Plus) on a CFX96 touch real-time PCR detection system (Bio-Rad). The housekeeping gene 16s was used as internal normalization control.

Table 1.

Primers used in this study

Primer sequences are shown 5′ to 3′, and a brief description of their use is provided.

Western blotting

Protein samples were electrophoresed on an SDS-polyacrylamide gel prior to transfer to a polyvinylidene difluoride membrane (Immobilon-P, Millipore). Both MSMEG_4207 and CRP (Rv3676) polyclonal antibodies were generated in the laboratory and were used at a dilution of 1:5000. Anti-acetyl-lysine antibody (Cell Signaling Technology, Inc.) was used at a dilution of 1:1000 for whole-cell lysate and 1:2500 for recombinant protein. Horseradish peroxidase-conjugated secondary antibody (GE Healthcare) was used at a dilution of 1:50,000 and detected by enhanced chemiluminescence (Luminata Crescendo, Millipore) following the manufacturer's protocol. A standard curve was generated by measuring the intensity of bands obtained on Western blot analysis using known concentrations of USP, from which levels of USP could be measured in planktonic cultures and the pellicle.

Luciferase reporter assay

The intergenic region (103 bp) present upstream of msmeg_4210/4211 was amplified by PCR using primers MSMEG_4210–4211 INT FWD SpeI and MSMEG_4210–4211 INT REV BamHI. The PCR amplicon was cloned into pBluescript II KS vector, and the sequence of the insert was verified (Macrogen, South Korea). The BamHI–SpeI-digested insert from the above pBluescript II KS construct was ligated separately with XbaI–BamHI-digested pMV-no prom-LuxAB and BamHI–SpeI-digested pMV-no prom-LuxAB, respectively. WT M. smegmatis Δkatms strains were electroporated with the plasmids. Cultures from individual single colonies were grown until OD₆₀₀ of ∼1.0 and diluted into fresh medium to an OD₆₀₀ of 0.02. Aliquots of cultures at indicated time points were taken, and luciferase counts were measured as described previously (62).

To generate shorter fragments for gel-shift assays, PCR was performed on plasmid harboring the entire intergenic region using primers USP-prom_short_SpeI fwd and MSMEG_4210–4211INT REV BamHI (to generate a PCR product containing 56 bp of the intergenic region proximal to msmeg_4210) and primers MSMEG_4210–4211 INT REV BamHI and MSMEG_4210–4211 INT FWD SpeI (to generate a PCR product containing 60 bp of the intergenic region proximal to msmeg_4211). PCR products were purified and used in EMSA.

Cloning, expression, and purification of proteins

PCR was carried out on genomic DNA of M. smegmatis mc² 155 using MSMEG_4212 FWD and MSMEG_4212 REV primers (Table 1). The amplified product was digested with EcoRI–NotI and cloned into similarly digested pPROEx-HTa to generate plasmid pPRO-MS4212. The clone was confirmed by sequencing (Macrogen, South Korea).

The mutant MSMEG_4212 (D54A) was generated by site-directed mutagenesis (63) of the pPRO-MS4212 plasmid. The mutation was confirmed by sequencing. Hexahistidine-tagged WT and mutant MSMEG_4212 were purified by nickel-nitrilotriacetic acid affinity chromatography as described earlier (8).

Cloning, expression, and purification strategies for KATms, USP⁴²⁰⁷, GST–KATms, and Rv3667 (Acs) and generation of the USP⁴²⁰⁷K104R mutant have been described earlier (8, 13).

EMSA

A double-stranded oligonucleotide harboring the intergenic region (104 bp) present upstream of msmeg_4210/4211 was end-labeled using [γ-³²P]ATP by T4 polynucleotide kinase. The labeled oligonucleotide was purified, and ∼50 fmol (∼10,000 cpm) was incubated with varying concentrations of either WT or D54A mutant MSMEG_4212 in buffer containing 25 mm Tris-Cl (pH 7.8), 5 mm MgCl₂, 50 mm KCl, 10 μm EDTA, 10% glycerol, 1 μg of poly(dI/dC), and 50 μg/ml BSA in a 20-μl reaction volume (62) at 37 °C for 20 min. Gels were electrophoresed at 50 V for 1 h at 4 °C prior to loading the samples on 6% polyacrylamide gels in Tris borate/EDTA (45 mm Tris borate and 1 mm EDTA) buffer at 100 V for 1.5 h at 4 °C for the separation of bound complex from the free probe. The gels were dried and scanned using a phosphorimager (Azure Biosystems).

For competition with unlabeled DNA fragments, WT MSMEG_4212 (300 ng; 12.6 pmol) or the D54A mutant protein (30 ng; 1.26 pmol) was incubated at 37 °C for 20 min in the absence or presence of the unlabeled double-stranded oligonucleotide (104-bp intergenic region or two halves as indicated in Fig. 3C; 20 pmol each), along with the radiolabeled probe.

Generation of Δusp⁴²⁰⁷ strain

A strain with deletion of usp⁴²⁰⁷ was generated using the homologous recombination-based suicide vector approach (35). The region 5′ to usp⁴²⁰⁷consisting of ∼988 bp upstream was amplified by PCR using primers up 4207FWD and up 4207REV, and a fragment of ∼1035 bp downstream of usp⁴²⁰⁷was amplified using down 4207FWD and down 4207REV primers (Table 1). The 5′-amplicon was cloned into pBluescript II KS vector using PstI–HindIII sites, and the 3′-amplicon was cloned into pBluescript II KS vector using HindIII–NotI sites to generate plasmids pBKS-MS4207–5′KO and pBKS-MS4207–3′KO, respectively. The clones were confirmed by sequencing (Macrogen, South Korea). The PstI–HindIII-digested pBKS-MS4207–5′KO and HindIII-NotI digested pBKS-MS4207–3′KO inserts were ligated into PstI–NotI-digested p2NIL plasmid to generate p2NILMS4207–5′3′KO plasmid. The PacI fragment from pGOAL19 containing the marker gene cassette (Ag85p-lacZ, hyg^r, and hsp60p-sacB) was cloned into p2NIL-MS4207–5′3′KO to generate plasmid p2NIL-MS4207–5′3′KO-PacI. 5 μg of the plasmid p2NIL-MS4207–5′3′KO-PacI was electroporated into M. smegmatis mc² 155, and single crossovers and double crossovers were obtained essentially as described in the original method (35). Double crossovers were further tested by genomic PCR and Western blotting to obtain the Δusp⁴²⁰⁷ strain. The strategy for generation of Δkatms has been described earlier (8). Similar strategy was taken to generate Δusp⁴²⁰⁷ Δkatms strain.

To construct the Δusp⁴²⁰⁷ strain complemented with WT usp⁴²⁰⁷driven by the sigA promoter, the KpnI–BamHI fragment from pBKS-sigA and BamHI–HindIII fragment from pPRO-MS4207 was cloned into KpnI–HindIII-digested pMV 10–25 plasmid to generate pMV-MS4207Comp. Identical cloning strategy was taken to generate pMV-MS4207_K104RComp. The pMV-MS4207Comp or pMV-MS4207_K104RComp plasmids were electroporated into the Δusp⁴²⁰⁷strain. The positive transformants carrying the required plasmid were screened by colony PCR and further validated by RT-PCR and Western blotting.

Study of colony morphology

Mid-log–phase cells grown in Middlebrook 7H9 medium supplemented with 0.2% glycerol and 0.05% Tween 80 were harvested and resuspended in fresh 7H9 medium to an OD₆₀₀ of 0.5. 5 μl of neat and serially diluted (10⁻¹ and 10⁻² dilution) cultures were spotted on the 7H10 agar plates, with or without 0.05% Tween 80, and incubated at 37 °C for 48–72 h until visible growth was observed.

GST pulldown assays

GST–KATms was expressed in the E. coli SP850 strain upon induction by 0.5 mm isopropyl 1-thio-β-d-galactopyranoside. Following purification, GST or GST fusion proteins bound to GSH beads interacted with 0.5 mg of cytosolic proteins from the Δusp⁴²⁰⁷ strain complemented either with the WT or mutant (usp⁴²⁰⁷K104R) form of Δusp⁴²⁰⁷at 4 °C for 1 h. The beads were then washed five times with 1 ml of buffer containing 100 mm Tris-Cl (pH 8.2), 150 mm NaCl, 10% glycerol, 5 mm 2-ME, 2 mm PMSF, and 1 mm benzamidine hydrochloride. The bound proteins were analyzed on a 13.5% SDS-polyacrylamide gel.

Mass spectrometric analysis of acetylated proteins

Lysates were subjected to SDS gel electrophoresis analyzed by Western blotting using anti-acetyl-lysine antibodies. The blot was aligned with the gel and a region containing acetylated proteins from the WT strain, and the corresponding regions from the Δkatms strain were excised. Pieces were directly shipped to the Taplin Mass Spectrometric Facility at Harvard University, Boston, MA, where they were reduced with 10 mm DTT, followed by alkylation with iodoacetamide. Samples were then washed, dried, and subjected to tryptic digestion at 37 °C for 16 h. Peptides were collected and sequenced using an Orbitrap Mass Spectrometer from Thermo Fisher Scientific. Peptides, and the proteins in which they were presented, were identified based on the predicted proteome of M. smegmatis. Further analysis was performed to detect acetylated peptides differing by a mass of 42 Da, representing the addition of an acetyl group to a lysine residue in the sequence. It was ensured that the same peptides were identified in a nonacetylated form in the Δkatms strain. Here, we report only peptides whose acetylation was completely lost in the Δkatms strain.

Pellicle formation assay

Pellicle formation was monitored by growing static cultures of mycobacteria with an initial OD₆₀₀ of 0.1 in Middlebrook 7H9 medium (without Tween 80) at 37 °C for 72–96 h in either polystyrene 3.5-cm dishes or T-25 tissue culture flasks. The formation of pellicle was identified visually. To study the expression profile of different genes during the biofilm formation, the pellicle formed at the air–liquid interface was collected, followed by RNA and/or protein isolation as required.

To prepare whole-cell lysates for anti-acetyl-lysine Western blottings, cells, planktonic cultures at OD ∼2, or those harvested from strains after pellicle formation (flocculent cells seen in the Δusp strain) were washed with TBS containing 10 mm Tris-Cl (pH 8.2) and 0.9% NaCl. Cells were lysed by bead beating using 0.5-mm diameter glass beads (BioSpec Products) in buffer containing 50 mm Tris-Cl (pH 8.2), 100 mm NaCl, 10% glycerol, 5 mm 2-ME, 2 mm PMSF, 1 mm benzamidine hydrochloride, and 3 mm nicotinamide. Samples were centrifuged at 1000 × g for 5 min at 4 °C to separate the glass beads. The supernatant was centrifuged again at 3000 × g for 5 min at 4 °C to separate the unlysed cells, and the supernatant was collected, and protein estimation was performed by Bradford method. Protein (50 μg) was loaded for Western blot analysis.

Analysis of GPLs in pellicle

Cells from biofilm or flocculent cells (in the case of Δusp⁴²⁰⁷) were harvested in PBS, and the suspension was sonicated at 25 °C for 60 min to disrupt clumps. A fraction of the cell suspension was used to prepare lysates for protein estimation. The remainder was centrifuged at 3000 × g for 10 min; the supernatant was discarded, and the biomass was dried at 65 °C in a glass tube. To the dried biomass, 5 ml of chloroform/methanol (2:1, v/v) was added, vortexed, and sonicated at 25 °C for 45 min. The tubes were then incubated overnight in an end-over rocker. The suspension was centrifuged at 4000 rpm for 20 min, and the supernatant was collected. 0.2 volume of 0.9% NaCl was added to the supernatant, and the tube was vortexed and mixed on a rotator for 20 min. The mixture was centrifuged for 5 min at 4000 rpm allowing two phases to form. The lower organic phase containing GPLs was collected, transferred to a separate tube, and evaporated to dryness at 65 °C. Dried GPLs were resuspended in different volumes of solvent chloroform/methanol (2:1, v/v), according to the protein content in the sample (final concentration of 64 mg/ml protein). 5 μl from each sample was spotted on a silica gel 60 F₂₅₄ plate (Merck) and resolved using chloroform/methanol (100:7, v/v) as the solvent system. The plate was dried for 3 h followed by spraying with 0.1% orcinol in 40% sulfuric acid. The plate was incubated at 95 °C for 45 min until bands were revealed.

In vitro acetylation assay

Assays were carried out in a 20-μl total reaction volume containing 25 mm Tris-Cl (pH 8.2), 100 mm NaCl, 5 mm EDTA, 30 μm acetyl-CoA, 10 μmf cAMP, KATms (15 pmol), Acs (Rv3667; 75 pmol), and varying concentrations of USP⁴²⁰⁷. Reactions were incubated at 25 °C for 20 min, stopped by heating in SDS sample buffer, and analyzed by Western blotting using anti-acetyl-lysine antibody, quantitated by densitometric analysis of the bands, and normalized for protein content in a lane by blotting with anti-His antibody. To determine the extent of acetylation of Acs, the intensity of the band in each lane of the blot probed with acetyl-lysine antibodies was estimated and normalized to the amount of Acs in that lane based on the intensity of bands seen in the blot with anti-hexahistidine antibodies. The value obtained was then expressed as a fraction of the extent of acetylation seen in the absence of USP⁴²⁰⁷.

Measurement of cAMP

Both the planktonic and pellicle cells were harvested by centrifugation; cell pellets were resuspended in 0.1 n HCl and heated for 10 min at 95 °C. Samples were stored at −70 °C until further use. Aliquots were taken for estimation of cAMP by radioimmunoassay (64). All data shown was analyzed using GraphPad Prism 8 using the two-tailed t test.

Author contributions

S. S., P. B., A. Bose, N. S., and S. N. investigation; S. S., P. B., A. Banerjee, A. Bose, N. S., and S. N. methodology; S. S., P. B., and S. S. V. writing-original draft; A. Banerjee, S. N., D. K. S., and S. S. V. conceptualization; A. Banerjee, S. N., and S. S. V. formal analysis; N. S. and S. S. V. validation; S. N. and S. S. V. supervision; S. S. V. resources; S. S. V. data curation; S. S. V. project administration; S. S. V. writing-review and editing.