The strains used in this study were S. coelicolor M145 and S. lividans TK24 (Bentley et al., 2002). These strains were grown at 28°C in MS (ACROS, Belgium) and YEME media (ACMEC, Shanghai) for spore generation and protoplast preparation, respectively, whereas R2YE medium (BD, United States) was used for protoplasts regeneration after transformation (Kieser et al., 2000). Apramycin, and thiostrepton (HARVEYBIO, Beijing) were added in solid R2YE medium at a final concentration of 50 μg/mL in cultures of SC or SL containing the plasmids pOJ260 and pSET152 and the plasmids pDH5 and pWHM3, respectively. The following E. coli strains were used, DH5α for routine cloning, ET12567 to prepare de-methylated plasmids to transform SC and C43 for efficient protein expression. These strains were grown in Luria broth (LB) medium. Ampicillin (50 μg/mL), chloramphenicol (25 μg/mL), and kanamycin (50 μg/mL) were added to growth media when required.

In order to disrupt the gene encoding the acetyltransferase SCO0988 in S. coelicolor M145 (SC), a 370 bp DNA fragment internal to the sco0988 coding region was amplified by PCR using the primer pair DisCsco0988-F/DisCsco0988-R (Table 1) and SC chromosomal DNA as template. The resulting PCR fragments were digested by BamHI and HindIII (Takara, Japan) and ligated into the plasmid pOJ260 that carries a gene conferring resistance to apramycin cut by the same enzymes (Kieser et al., 2000). The resulting plasmid, pOJ260-sco0988^int, was transformed into E. coli ET12567 to prevent its methylation (Flett et al., 1997) and apramycin resistant (Apra^R) transformants were selected. The non-methylated plasmids extracted from the Apra^R transformants were transformed into SC and selected for Apra^R. The sco0988 disrupted mutants (SC-∆sco0988) were confirmed by PCR using the primer pair DisCsco0988-F/DisCsco0988-R (Table 1).

In order to complement the SC-∆SCO0988 mutant and to over-express sco0988 in S. lividans TK24, the coding sequence of sco0988 was amplified by PCR using the primers OEsco0988-XbaI-F and OEsco0988-HindIII-R (Table 1) and SL genomic DNA as template. The resulting PCR products were digested by XbaI and HindIII and cloned into pWHM3-ermE cut by the same enzymes in order to put the expression of sco0988 under the control of the strong ermE promoter (Bibb et al., 1985). The resulting pWHM3-ermE-sco0988 plasmid was transformed in SL and in SC-∆sco0988 and protoplasts to generate the strains SL/pWHM3-ermE-sco0988 and SC-∆sco0988/pWHM3-ermE-sco0988. The SC and SL strains harboring pWHM3-ermE empty plasmid were used as a control (see Table 2).

In order to quantify biomass yield as well as RED (undecylprodigiosin) and ACT (actinorhodin) production of native and genetically modified derivatives of SC and SL constructed in this study, 10⁷ spores of each strain were used to inoculate 200 mL of R2YE medium and grown for 72 h until stationary phase. Two hundred microliters of each culture was plated on cellophane discs deposited on the surface of 9 cm-diameter plates of R2YE solid medium in triplicates and the plates were incubated at 28°C. Half of the mycelium of each triplicate was collected every 12 h, from 24 h until 84 h of incubation, in order to establish a growth curve. Collected mycelial samples were dried at 55°C overnight and weighted. One quarter of the mycelium of each triplicate was used to assay intracellular RED and ACT concentrations and the agar medium below the last quarter was used to assay extracellular ACT concentrations as described previously but with slight modifications (Xu et al., 2010).

To assay intracellular RED, the mycelium was extracted by the addition of 1 mL of methanol then the mixture was acidified to pH 2–3 by addition of 1 mL HCl (1 M) whereas to assay intracellular ACT the mycelium was extracted by the addition of 1 mL of KOH (1 M). Samples were vortexed for 30 min at 4°C. To assay RED the optical density was measured at 530 nm against a blank constituted by methanol and HCl (1 M) (v/v:1/1) whereas to assay ACT the optical density was measured at 640 nm against a blank constituted by KOH (1 M).

To assay extracellular ACT, a quarter of each of the three R2YE agar plates was smashed and subjected to diffusion in 10 mL water at 4°C for at least 2 h. The mixture was then centrifuged and 10 mL of KOH (1N) was added to the collected supernatant. The solution was mixed by inversion, then 10 mL of HCL (3N) was added. The resulting mixture was incubated on ice for 10 min and the precipitated ACT was then collected by centrifugation. The supernatant was discarded and the ACT pellet was re-suspended in 1 mL KOH (1N). The optical density of the solution was measured at 640 nm against a blank constituted by KOH (1 M) in order to determine ACT concentration. In all cases a spectrophotometer (Molecular Devices, United States) was used to determine optical density.

In order to delete bldKB/sco5113 in S. coelicolor M145 (SC), two 1.5-kb DNA fragments flanking the sco5113 coding region were amplified by PCR using primer pairs DelUpsco5113-HindIII-F/DelUpsco5113-XbaI-R and DelDownsco5113-XbaI-F/DelDownsco5113-KpnI-R (Table 1) and SC genomic DNA as template. The resulting two PCR fragments were individually digested with the restriction enzymes present in the primers (Takara, Japan) and cloned into the pDH5 plasmid that carries thiostrepton resistance determinant (Kieser et al., 2000), using a triple ligation strategy. The resulting plasmid pDH5-Delsco5113 suitable for sco5113 deletion was transformed into SC protoplasts. The transformants were grown for three generations on plates of R2YE medium in absence of any selective pressure. Subsequently, spores resulting from the transformants were diluted and spread to yield isolated colonies. The thiostrepton resistance or sensitivity of the resulting sporulated colonies was determined by their replicate plating on R2YE and on R2YE with thiosptrepton at 50 μg/mL. The chromosomal structure of the transformants, that had lost the delivery plasmid and were only able to grown on R2YE, was assessed by PCR using the primer pair VerifDelsco5113-F and VerifDelsco5113-R (Table 1) and genomic DNA originating from these transformants. The mutants in which the successful in-frame deletion of sco5113 was confirmed were called SC-ΔbldKB.

In order to generate a mutagenized version of BldKB/SCO5113, a plasmid containing BldKB/SCO5113 ORF together with its native promoter was first constructed. To do so, a PCR fragment encompassing this region was amplified using the primer pair CPsco5113-EcoRI-F/CPsco5113-XbaI-R (Table 1) and SC chromosomal DNA as template. The resulting PCR product cut by EcoRI and XbaI was ligated into pUC19 to generate pUC19-bldKB-wt (AAG).

Subsequently, in order to generate mutagenized BldKB, the primer pairs K425Rsco5113-F/K425Rsco5113-R and K425Qsco5113-F/K425Qsco5113-R (Table 1), in which the 425th Lys codon (AAG) was replaced by an Arg codon (AGG) or by a Gln codon (CAG), were used to amplify the mutated fragment by PCR using pUC19-bldKB-wt as a template. The resulting PCR products were treated with DpnI to digest the methylated parental DNA template, and subsequently transformed into E. coli DH5α competent cells. Plasmids isolated from various colonies were sequenced and the plasmids harboring the desired mutation (AAG/K to AGG/R) and (AAG/K to CAG/Q) were selected and named pUC19-bldKB-K425R and pUC19-bldKB-K425Q, respectively.

bldKB-wt (AAG), bldKB-K425R (AGG) and bldKB-K425Q (CAG) together with their native promoters were amplified by PCR using primer pair CPsco5113-EcoRI-F/CPsco5113-XbaI-R (Table 1) from pUC19-bldKB-wt (AAG), pUC19-bldKB-K425R (AGG) and pUC19-bldKB-K425Q (CAG), respectively. The resulting DNA fragments were ligated into the EcoRI and XbaI sites of pSET152 to generate pSET152-bldKB-wt, pSET152-bldKB-K425R and pSET152-bldKB-K425Q, respectively. These three plasmids were transformed into SC-ΔbldKB protoplasts to generate SC-ΔbldKB/pSET152-bldKB-wt, SC-ΔbldKB/pSET152-bldKB-K425R and SC-ΔbldKB/pSET152-bldKB-K425Q and the phenotype of the transformants was assessed for complementation.

In order to carry out in vitro acetylation of BldKB, recombinant His-tag fused BldKB was first expressed and purified in E. coli C to facilitate protein expression and purification (Miroux and Walker, 1996). To do so the whole bldKB coding region, except its termination codon, was amplified by PCR using primer pairs PEsco5113-EcoRI-F and PEsco5113-HindIII-R (Table 1). The resulting PCR fragments digested by EcoRI and HindIII (Takara, Japan) were cloned into the plasmid pET28a under the control of a LacI-like IPTG inducible promoter (Schifferli, 1995). In this plasmid, that carries a gene conferring resistance to kanamycin (Kan^r), the cloned gene is fused to a DNA sequence translated as a 6-His-tag. The recombinant plasmid pET28a-bldKB was transformed into E. coli C43. E. coli C43/pET28a-bldKB was cultured in LB medium containing 25 μg/mL kanamycin at 37°C. When OD₆₀₀ reached 0.4–0.5, IPTG was added at a final concentration of 0.5 mM and induction was allowed for 5–6 h. The cells were collected by centrifugation, washed in PBS (137 mM NaCl, 3 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH = 7.4) and sonicated to obtain homogenous cell crude extracts. His-tag fused BldKB was purified to near homogeneity using Ni-NTA column according to the manufacturer’s instructions (Sangon Biotech, Shanghai Co., Ltd.).

However, since the acetyltransferase SCO0988 is a membrane protein, that could not be purified in E. coli, crude extracts of SL/pWHM3-ermE and SL/pWHM3-ermE-sco0988 had to be prepared as described in You et al. (2019) in order to carry out the acetylation assays. To do so, these two strains were cultivated in R2YE liquid medium until stationary phase and their mycelium was collected by centrifugation, washed in PBS, re-suspended in sonication buffer (50 mM Tris-HCl pH 8.0, 10% glycerol, 0.1 mM EDTA, 1 mM dithiothreitol) containing a deacetylase inhibitor cocktail (Beyotime, Shanghai) and sonicated in order to obtain a homogeneous solution that was centrifuged to remove cell debris. The protein concentration of each supernatant was determined using the Bradford reagent (Catalogue No. 500-0006; Bio-Rad) and adjusted to 500 μg/mL for further uses. In the acetylation assay, 5 μg of purified six-histidine-tagged BldKB (His₆-BldKB) and 1 μg of the crude extract of either SL/pWHM3-ermE or SL/pWHM3-ermE-sco0988 were mixed together with 20 μM of acetyl-CoA in HAT buffer (50 mM Tris-HCl pH 8.0, 10% glycerol, 0.1 mM EDTA, 1 mM dithiothreitol) and incubated at 30°C for 3 h. Control was carried out in parallel with the reaction mix containing BldKB alone in presence of AcetylCoA.

At the end of the incubation period, His₆-BldKB present in the reaction mixture was purified using Ni-NTA column following the manufacturers’ instructions. The collected fractions were separated on 12% SDS-PAGE and transferred on a polyvinylidene fluoride (PVDF) membrane that was incubated with antibodies against acetylated lysine (Cat. No. ICP0380, Runzekang, Beijing). Visualization of the acetylated protein was achieved by the addition of secondary antibody (HRP Goat Anti-Rabbit IgG) and a solution containing 3,3′-Diaminobenzidine (DAB) (Cat. No. PK10005, Proteintech) the substrate of HRP as described in Kim et al. (2013).

SCO0988 belongs to the Gcn5-related N-acetyltransferase (GNAT) (Pfam00583) superfamily whose carboxyl terminal end catalyzes the transfer of the acetyl moiety of acetyl-CoA to the ε-amino group of a lysine residue (Vetting et al., 2005). The Kyte and Doolite hydropathic profile (Kyte and Doolittle, 1982) of this protein suggested that it is a membrane protein with multiple trans membrane segments (Supplementary Figure S1). The level of expression of sco0988 was shown to increase with time indicating that it might regulate late developmental processes such as morphological differentiation and secondary/specialized metabolism (Supplementary Figure S2) (Zhang et al., 2020).

In order to determine whether this acetyltransferase has an impact on these processes in the strong antibiotic producer, SC, a sco0988 disruptive mutant of SC, SC-Δsco0988 was constructed. The phenotypes of the mutant, and of the wt strain are shown in Figure 1. The sco0988 deletion mutant had a slightly slower growth rate than the original strain (Supplementary Figure S3) and was characterized by a strong inhibition of both antibiotic production and sporulation (Figure 1A). Quantitative analysis demonstrated a 2/3 reduction of RED production and a complete abolition of ACT biosynthesis in SC-Δsco0988 compared to the original strain (p < 0.05) (Figure 1B). The complementation of the SC-Δsco0988 by sco0998 carried by the high copy number plasmid pWHM3 (Bibb et al., 1985) restored growth (Supplementary Figure S3) as well as RED and ACT production to the level of the original strain whereas sporulation was enhanced compared to that of the wt strain (Figures 1A,B).

In order to determine whether this acetyltransferase has an impact on these processes in the weak antibiotic producer, SL, sco0988 carried by pWHM3 was over-expressed in this strain. This over-expression led to a two-fold increase of RED production and the originally undetectable ACT production became now detectable (Figure 1C). The over-expression of sco0988 had no impact on the growth rate of the strain that was similar to that of the strain containing the empty plasmid (data not shown). These results clearly demonstrated that SCO0988 has a positive impact on both antibiotic production and morphological differentiation in these two model Streptomyces species.

In order to gain an insight into the biological processes affected by the over-expression of sco0988 in S. lividans TK24, a comparative analysis of the acetylome of the control strain (SL/pWHM3-ermE) and of the strain overexpressing sco0988 (SL/pWHM3-ermE-sco0988) was carried out. Acetylated lysine (acK) peptides were purified by immunoaffinity-based acetyl-lysine peptide enrichment, identified and quantified in both strains by high resolution mass spectrometry. An overview of the experimental procedures used is provided in Figure 2A.

A total of 1,399 acetylated lysine sites were discovered in 740 proteins (Figure 2B) that represented approximately 10% of S. lividans TK24 proteome (Rückert et al., 2015). 59% of these proteins, contained a single acetylated site, 21% had two acetylated sites and 9% contained three acetylated sites (Figure 2C). The average frequency of lysine acetylation was of 1.45 acetylated residue in 100 amino acids (Figure 2D). Detailed information concerning all acetylated peptides and the corresponding proteins is provided in Supplementary Table S1. A total of 643 statistically differentially acetylated peptides (p < 0.05) at 24 h and 36 h is shown in Supplementary Table S2. Considering that 54 of them were present at the two time points, the amount of the acetylated peptides is thus 589 (643-54). Among these 589 acetylated peptides, 177 (30.05%) were more acetylated while 358 (60.78%) were less acetylated in the strain overexpressing sco0988 than in the control strain. These observations suggested that in SL the pool of acetylCoA is limited and that the over-expression of sco0988 consumes acetylCoA and thus limits the acetylation of the specific targets of other acetyltransferases. Interestingly, 92 acetylated peptides were found exclusively present in the strain over-expressing sco0988 whereas 159 acetylated peptides were only present in the control strain (Supplementary Tables S1, S2). The proteins over-acetylated or exclusively acetylated in SL/pWHM3-ermE-sco0988 can thus be considered as potential specific SCO0988 targets.

Gene Ontology (GO) terms and KEGG pathway enrichment analysis of differentially acetylated proteins (DAPs) was performed with Applied Protein Technology Co., Ltd (Shanghai, China). The ontological classifications, determined by Blast2Go (see text footnote 2) and Interpro scan,⁴ are shown in Figure 3. A large proportion of DAPs (>60%) were classified as involved in nitrogen (19.6%) and carbon (14.93%) metabolisms, transcriptional and translational processes (18.97%) as well as transport (12.44%).

In order to identify a possible consensus acetylation motif of SCO0988, the frequency of amino acids present from position −5 to +5 around the 92 lysines exclusively acetylated in the strain over-expressing SCO0988 were determined using the MEME program, and similarly, the frequency of amino acids presents from position −5 to +5 around 159 lysines exclusively acetylated in the control strain were determined using the MEME program. The results shown in Figure 4 demonstrated that, as expected, the consensus of the sites acetylated by SCO0988 was different and also far less diverse than the consensus of the other acetylated sites that are likely to be acetylated by multiple acetyltransferases other than SCO0988. In the latter the central pair KK is highly conserved but is replaced by KR in the SCO0988 consensus. K and R are both basic amino acids. In the SCO0988 consensus the positions 2, 3, 5 and 6 are highly conserved and are P, D, K and R, respectively and in the other positions only 2 alternative amino acids are present.

BldKB/SCO5113 is the extracellular solute-binding protein of the ABC transporter BldKABCDE (SCO5112-SCO5116) also constituted of BldKA and BldKC (SCO5112 and SCO5114) two integral membrane proteins, BldKD/SCO5115/intracellular ATPase subunit and BldKE/SCO5116/ATP-binding protein. This ABC transporter is involved in the up-take of the BLD261 oligopeptide that is a signaling molecule of the quorum sensing pathway (Willey et al., 1993). The up-take of this signaling molecule by the BldK transporter triggers the expression of genes of the “bald” signaling cascade that controls positively antibiotic production as well as aerial mycelium formation and sporulation in SC (Nodwell et al., 1999). Since, 4 of the 7 acetylated lysine (acK) peptides detected in BldKB/SCO5113 showed a great increase in their acetylation level upon sco0988 overexpression, with the Lys425 showing the largest increase (Supplementary Table S2), BldKB was selected for further analysis.

In order to determine whether SCO0988 was able to acetylate BldKB in vitro, recombinant His₆-BldKB was expressed in E. coli and purified (Figure 5A). Since the membrane protein SCO0988 formed inclusion bodies when E. coli/pET expression system was used (data not shown), cell crude extracts from S. lividans over-expressing sco0988 were prepared for the acetylation assay. Western blot probed with an antibody against acetylated lysine clearly demonstrated an intensified band corresponding to acetylated BldKB in the reaction carried out in the presence of crude extract of SL/pWHM3-ermE-sco0988 (Figure 5B) but not in that carried out in the presence of crude extract of SL/pHWM3-ermE (Figure 5B). No acetylation was observed in the control assays with BldKB alone in presence of AcetylCoA (data not shown). These data thus demonstrate the acetylation of BldKB by SCO0988.

According to NCBI’s Conserved Domain Database (CDD), four structural domains could be identified in BldKB: a periplasmic component DdpA-like (57–592), an oligopeptide binding domain OppA2-like (75–582), an extracellular solute-binding domain SBP_bac_5-like (120–516) and a substrate-binding domain PRK09755 (356–599) (Figure 6A).⁵ All four domains are critical for the BldKB function. Seven acetylated lysine sites were detected in BldKB and among them four (K/248, K/337, K/341 and K/425) were over-acetylated upon sco0988 over-expression (Supplementary Tables S1, S2 and Figure 6). Interestingly K₄₂₅, the most sensitively acetylated Lys upon SCO0988 over-expression, is present in the four domains detected in BldKB (Figure 6). In order to determine whether K₄₂₅ was important the function and/or the acetylation of BldKB, a SC strain disrupted for bldKB, SC-ΔbldKB, was first constructed. SC-ΔbldKB had a similar growth pattern as the original strain (Supplementary Figure S4) but its RED and ACT production levels were reduced, being 32.03 and 34.53% lower, respectively, than that of the original strain, at 72 h (Figures 7A,B). In contrast, morphological differentiation and sporulation seemed totally abolished in SC-ΔbldKB. These observations are consistent with those reported in previous studies (Nodwell et al., 1996; Nodwell and Losick, 1998; Park et al., 2005). Subsequently, variants of BldKB in which Lys⁴²⁵ was substituted by Arg/R or Gln/Q were constructed. Arg/R was chosen since it cannot be acetylated and Gln/Q was chosen since many reports in the literature mentioned that it mimics acetylated Lys (Kamieniarz and Schneider, 2009; Okada et al., 2021; Morse et al., 2023). When bldKB (K425R) and bldKB (K425Q) were introduced into SC-ΔbldKB, complementation did not occur since ACT and RED production as well as sporulation were not restored to wild type level (Figures 7A,B). In contrast, these features were restored to wild type level when the native bldKB gene was introduced into SC-ΔbldKB. These results indicated that Lys⁴²⁵ is crucial for BldKB function and that the replacement of Lys⁴²⁵ by Arg or Gln greatly alters BldKB function. Unfortunately, these results were inconclusive concerning the impact of acetylation on BldKB function (see discussion).