Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products

Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products

Article Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoADerived Natural Products Graphical Abstract Authors J...

3MB Sizes 0 Downloads 39 Views

Article

Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoADerived Natural Products Graphical Abstract

Authors Jun-Yu Xu, Ya Xu, Zhen Xu, ..., Xiaohe Chu, Minjia Tan, Bang-Ce Ye

Correspondence [email protected] (M.T.), [email protected] (B.-C.Y.)

In Brief Xu et al. systematically investigated the general interplay between beneficial and adverse effects of cellular acyl-CoA concentrations on the biosynthesis of acyl-CoA-derived natural products in bacteria, which provided insight into the potential function of the identified lysine acylation substrates.

Highlights d

Propionyl-CoA-derived lysine propionylation affects erythromycin biosynthesis

d

Lysine acylation regulation widely exists in natural product synthesis in bacteria

d

Lysine acylation reflects imbalance between generation and consumption of acyl-CoAs

d

Beneficial and adverse effects of acyl-CoA regulate bacterial secondary metabolism

Xu et al., 2018, Cell Chemical Biology 25, 1–12 August 16, 2018 ª 2018 Elsevier Ltd. https://doi.org/10.1016/j.chembiol.2018.05.005

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Cell Chemical Biology

Article Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products Jun-Yu Xu,1,2,3,5 Ya Xu,3,5 Zhen Xu,3 Lin-Hui Zhai,2 Yang Ye,2 Yingming Zhao,2,4 Xiaohe Chu,1 Minjia Tan,2,* and Bang-Ce Ye1,3,6,* 1Collaborative Innovation Center of Yangtze River Delta Region Green Pharmaceuticals, College of Pharmaceutical Sciences, Zhejiang University of Technology, Hangzhou, Zhejiang 310014, China 2State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai 201203, PR China 3Lab of Biosystems and Microanalysis, State Key Laboratory of Bioreactor Engineering, East China University of Science and Technology, Shanghai 200237, China 4Ben May Department for Cancer Research, University of Chicago, Chicago, IL 60637, USA 5These authors contributed equally 6Lead Contact *Correspondence: [email protected] (M.T.), [email protected] (B.-C.Y.) https://doi.org/10.1016/j.chembiol.2018.05.005

SUMMARY

Coenzyme A (CoA) esters of short fatty acids (acylCoAs) function as key precursors for the biosynthesis of various natural products and the dominant donors for lysine acylation. Herein, we investigated the functional interplay between beneficial and adverse effects of acyl-CoA supplements on the production of acyl-CoA-derived natural products in microorganisms by using erythromycin-biosynthesized Saccharopolyspora erythraea as a model: accumulation of propionyl-CoA benefited erythromycin biosynthesis, but lysine propionylation inhibited the activities of important enzymes involved in biosynthetic pathways of erythromycin. The results showed that the overexpression of NAD+-dependent deacylase could circumvent the inhibitory effects of high acyl-CoA concentrations. In addition, we demonstrated the similar lysine acylation mechanism in other acyl-CoA-derived natural product biosynthesis, such as malonyl-CoA-derived alkaloid and butyryl-CoA-derived bioalcohol. These observations systematically uncovered the important role of protein acylation on interaction between the accumulation of high concentrations of acyl-CoAs and the efficiency of their use in metabolic pathways. INTRODUCTION Manipulating microorganisms for human benefit has been well studied in recent years. Traditionally, strains that naturally produced a desired product were found and then remodeled on the basis of random mutagenesis and selection (Rowlands, 1984). After that, the revolution of several biological technologies and concepts of metabolic engineering or synthetic biology has promoted the production of different chemicals in microbes,

from primary metabolites to secondary metabolites (Keasling, 2010; Nielsen and Keasling, 2011; Scherlach and Hertweck, 2009; Stephanopoulos et al., 2004). Reprogramming entire gene networks and reconstructing generic devices in engineered microbiological cell factories can produce desired chemicals with higher enhancement of substrate range, higher yield, and higher cost efficiency. In microbial chemical biosynthesis, coenzyme A (CoA) esters of short fatty acids (acyl-CoAs) are important precursors or even additional raw materials for several kinds of desired product biosynthesis, including various classes of bioactive natural products and biofuels. Polyketides are an important group of secondary metabolites with diverse pharmacological properties, such as antibiotic, anticancer, and immunosuppressive activities, and are important resources for human and veterinary medicine (Hertweck, 2009; Lackner et al., 2013; Staunton and Weissman, 2001). Using various kinds of acyl-CoAs (such as acetyl-CoA, propionyl-CoA, and malonyl-CoA) as its building block, polyketide biosynthesis is catalyzed by three classes of polyketide synthases (named type I/II/III PKSs) (Rokem et al., 2007). For example, the antibiotic macrolide, such as erythromycin, is the type I PKS-catalyzed polyketide condensed from propionyl-CoA and methylmalonyl-CoA. The aromatic polyketides, such as antibiotic tetracycline, is synthesized by type II PKS usually using malonyl-CoA as its building block. Chalcones and stilbenes, two important classes of natural polyphenols, are synthesized from coumaroyl-CoA/cinnamoyl-CoA with malonyl-CoA catalyzed by type III PKSs. In addition to polyketides, acyl-CoAs are important for biosynthesis of several classes of alkaloids, such as tropane derivative alkaloids, pyrrole derivative alkaloids, and terpene derivative alkaloids. The bacteria-derived tripyrrole alkaloid prodiginine, which has recently received attention for the promising immunosuppressive and anticancer activities, is synthesized based on acetyl-CoA and malonyl-CoA, similar to that of fatty acid synthesis (Gubbens et al., 2014; Hu et al., 2016; Williamson et al., 2006). In addition to bioactive natural products, short-chain acyl-CoAs are also important precursors for the biosynthesis of renewable and advanced biofuels (Choi and Lee, 2013; Lee et al., 2012). Most of recently attracted

Cell Chemical Biology 25, 1–12, August 16, 2018 ª 2018 Elsevier Ltd. 1

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

biofuels are short-chain alkanes, free fatty acids, fatty esters and fatty alcohols, which are involved in various acyl-CoAparticipated metabolic pathways. Thus, it is considerable that increasing corresponding acyl-CoAs is the systemic strategies for biosynthesis of desired products by using industrial microbial strains. Acyl-CoA could also function as the acyl donors for lysine acylation, an important class of protein post-translational modification (PTM). Among them, lysine acetylation, which widely occurs in both prokaryotes and eukaryotes, is closely related to kinds of physiological processes, such as DNA binding, transcription, cell signaling, and energy metabolism (Choudhary et al., 2014; Wang et al., 2010; Weinert et al., 2013a, 2013b). In addition to acetylation, several other lysine PTMs involving short-chain acyl-CoA metabolites are reported, such as malonylation, crotonylation, propionylation, butyrylation, succinylation, glutarylation, 2-hydroxyisobutyrylation, and b-hydroxybutyrylation (Hirschey and Zhao, 2015; Sabari et al., 2017). Recent studies have clearly shown that protein acylation level was dependent on cellular concentrations of various acyl-CoAs (Baeza et al., 2016; Wagner et al., 2017; Weinert et al., 2013b). Therefore, changes in acylCoA pools could directly modulate corresponding lysine acylations, further regulating signaling networks and coordinating metabolic flux (Chow et al., 2014; Colak et al., 2015; Schilling et al., 2016). We know that acyl-CoAs are the major precursors in microbial metabolite biosynthesis, as well as the key donors for lysine acylation. In many strategies of improving end-product yield, scientists modified the pathway to increase the formation of acyl-CoAs by optimizing the metabolic network or redesigning entire biological systems. Meantime, accumulation of acylCoAs pools may results in elevated protein acylation levels. Lysine acylation mechanism has been well studied in microbial physiological process. However, much less is known about the lysine acylation regulatory mechanism in the biosynthetic pathways of natural product. In the present study, we explored how lysine acylation regulated microbial metabolite biosynthesis by majorly using S. erythraea, an industrial erythromycin producing actinomycete as a model. Propionyl-CoA functions as the direct precursor of erythromycin biosynthesis and also acyl group donor of protein propionylation. Yet, the possible regulatory mechanism of lysine propionylation on erythromycin biosynthetic pathway in S. erythraea remains poorly understood. Herein, we first conducted quantitative proteomic approach to identify the difference of propionylation levels between wild-type (WT) strain and high producing E3 strain. We next used enzymatic activity assay and mutation analysis to prove that lysine propionylation could feedback regulate malonate-semialdehyde dehydrogenase (mmsA2), an important enzyme involved in the biosynthesis of endogenous propionyl-CoA. After that, we demonstrated that aldehyde-alcohol dehydrogenase (adhE) in the feeder pathway of exogenous propionyl-CoA and S-adenosylmethionine synthetase (metK) important for erythromycin biosynthesis were also inactivated by lysine propionylation, and we reversed their propionylation level by overexpression of SeSrtN to relieve the inhibition of enzymatic activity. Moreover, we expanded our study to other lysine acylations in varied species and proved that lysine acylation was a general regulatory mechanism in different kinds 2 Cell Chemical Biology 25, 1–12, August 16, 2018

of acyl-CoA-derived natural product biosynthesis in addition to macrolide polyketide, including malonyl-CoA-related alkaloid prodiginine in Streptomyces coelicolor, and butyryl-CoA-related biobutanol in Clostridium acetobutylicum. These observations showed that lysine acylation regulated natural product biosynthesis in addition to its well-known function in physiological or pathological bioprocess. Meanwhile, our study also provided insights into the relationship between lysine acylation and metabolite biosynthesis in microorganisms, which revealed a key role for lysine acylation mechanism in the microbial biosynthetic system. RESULTS Accumulation of Cellular Propionyl-CoA Resulted in High Level of Lysine Propionylation in Low-Yield S. erythraea Strain for Erythromycin Production S. erythraea, which synthesizes several bioactive secondary metabolites as a differentiated chemical arsenal for surviving in the highly competitive environments, is now regarded as the macrolide erythromycin producer in microbial fermentation industry (Cortes et al., 1990). In the process of producing erythromycin (Figure S1A), propionyl-CoA and methylmalonyl-CoA are the direct precursor metabolites, which function as the start unit and extender unit for the biosynthesis of 6-deoxyerythronolide B (erythromycin skeleton ring), respectively. Other acyl-CoAs, such as succinyl-CoA, malonyl-CoA, and acetyl-CoA, can be converted to these two acyl-CoAs and thus are indirectly utilized for erythromycin biosynthesis. As all these acyl-CoAs could also provide acyl groups for their corresponding lysine acylations, we thus explored the dynamic changes of these five protein acylation levels (no reports about lysine methylmalonylation up to now) in S. erythraea. Western blot analysis of the whole-cell lysates in wild-type (WT) and high-yield E3 strain was conducted. The results showed a significant elevation of global lysine propionylation level in WT strain in comparison with high-yield strain, whereas no difference was observed in the other four acyl modifications (Figure 1A). Several evidences have proved that cellular acyl-CoA concentrations can be directly related with protein acylation levels in cells (Chow et al., 2014; Qian et al., 2016), thus we inferred that higher level of propionylation may result from the propionyl-CoA accumulation in the S. erythraea WT strain. To prove this hypothesis, we measured intracellular propionylCoA levels in WT strain and E3 strain. The result clearly showed that intracellular propionyl-CoA level was higher in WT strain than that in E3 strain, which was likely the direct cause leading to higher protein propionylation level in WT strain (Figure 1B). We conducted quantitative proteomic analysis by using stable isotope dimethyl labeling (‘‘light’’ for WT strain and ‘‘heavy’’ for E3 strain) to explore the protein abundance of polyketide synthases involved in erythromycin biosynthesis (encoded by ery gene cluster) (Figure S1B) and the results indicated that enzymes involved in erythromycin production were obviously expressed in higher level in the E3 strain than in the WT strain (Figure 1C; Table S1). After that, we carried out comparative analysis of present proteomic data and our previously reported transcriptomic data to evaluate our data quality (Li et al., 2013). Similar to correlation analysis results reported in other species, low correlation

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Figure 1. Accumulation of Propionyl-CoA Resulted in High Level of Lysine Propionylation in Low-Yield S. erythraea Strain (A) Western blotting analysis of the four lysing acylation levels in WT and E3 strains. (B) Relative concentrations of propionyl-CoA in the WT/E3 strains. Data are represented as mean ± SEM. (C) Protein expression ratio (E3/WT) of the ery cluster in the erythromycin production pathway. (D) Correlation between mRNA expression ratio and protein expression ratio in all genes. (E) Correlation between mRNA expression ratio and protein expression ratio in genes of the ery cluster. Red dots indicate the polyketide synthases.

value (r = 0.44) was observed for whole-protein expression at transcriptomic level and proteomic level (Figure 1D). However, much higher correlation coefficient was acquired (r = 0.78) between the two omic analysis when genes in ery cluster were chosen as an independent dataset (Figure 1E), revealing the definite higher expression of polyketide synthases in E3 strain. This result indicated that in S. erythraea, the conversion efficiency of the precursor propionyl-CoA to the end-product erythromycin directly determined the cellular propionyl-CoA concentration and then the corresponding substrate propionylation level. Qualitative and Quantitative Analysis of Lysine Propionylation in S. erythraea As obviously changed propionylation level was observed in WT strain and E3 strain, we further systematically evaluated propionylated substrates and their dynamic changes in comparison with the acetylome. The light/heavy dimethyl-labeled peptides were mixed and enriched by anti-propionyllysine/anti-acetyllysine antibodies, respectively (Figure S1B). In the experiment, we totally identified 488 lysine propionylation (Kpr) sites in 271 proteins and 631 lysine acetylation (Kac) sites in 366 proteins (Tables S2 and S3). For 80 unquantified Kac sites, 57 were present in WT strain and 23 were present in E3 strain. While for 155 unquantified Kpr sites, 153 were present in WT strain and only 2 was present in E3 strain (Figure 2A). To our knowledge, this was the largest propionylome data set up in actinobacteria. We next explored the characteristics of lysine acetylation or propionylation sites in S. erythraea. The average number of pro-

pionylation in an individual protein was about 1.8 and about 38% of the proteins contained more than one site and approximately 17% contained no less than three sites (Figure S2A). We further investigated the similarities and differences among Kac/Kpr data with our previously acquired Kac and Kmal (Lys malonylation) data (Dan et al., 2015; Xu et al., 2016a). The results showed that of all the identified Kpr sites in S. erythraea, 235 (48%) sites, together with 216 (80%) proteins overlapped with Kac, and 129 (26%) sites together with 121 (45%) proteins overlapped with Kmal (Figures S2B and S2C). After that, we analyzed the proteins containing the most propionylated sites, including elongation factor G (8 sites), Fructose-bisphosphate aldolase (8 sites), elongation factor Ts (9 sites), 60 kDa chaperonin 1 (10 sites) and chaperone protein DnaK (11 sites) and found that these heavily propionylated proteins were also acetylated and malonylated (Figure S2D). Fifty-one proteins in total had at least one site shared for propionylation, acetylation and malonylation. Among them, elongation factor Ts and chaperone protein DnaK contained six and seven sites shared for propionylation, acetylation and malonylation, respectively (Figure S2E). Furthermore, we compared our data with the propionylome data from Thermus thermophilus (Okanishi et al., 2014) and found several conserved sites were propionylated in the homologous proteins of this bacterium, such as three lysine sites in ribosomerecycling factor, ATP-dependent Clp protease and glyceraldehyde-3-phosphate dehydrogenase, respectively and two lysine sites in malate synthase, which revealed evolutionary conservation of lysine propionylation (Figure S2F). After normalization to the protein expression level, the quantifiable Kpr and Kac sites were 332 and 539, respectively, and ratio distribution of the acetylated and propionylated sites were presented in Figures 2B, 2C, and S2G, respectively. The result showed that the median E3/WT ratio was 1.07 for acetylation and 0.25 for propionylation, which clearly suggested that the higher level of propionylation in WT strain than that in E3 strain and was in line with the western blot analysis. As lysine propionylation level was obviously changed in WT strain when Cell Chemical Biology 25, 1–12, August 16, 2018 3

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Figure 2. Quantitative Analysis of Lysine Acylome in S. erythraea (A) Venn diagrams showing the numbers of overlapping Kac/Kpr sites/proteins in the WT/E3 strains. (B) Scatterplot of the quantifiable acetylated sites in the WT/E3 strains. Blue dots, normalized H/L ratio less than 0.5; red dots, normalized H/L ratio between 0.5 and 2; green dots, normalized H/L ratio higher than 2. (C) Scatterplot of the quantifiable propionylated sites in WT and E3 strains. Annotation is the same as in Figure 2B.

compared with the E3 strain, we thus infer that propionylation but not acetylation was likely to regulate metabolic characteristic, especially the enzymatic activities in S. erythraea. Lysine Propionylation Impacted the Endogenous Feeder Pathway of Propionyl-CoA The biosynthesis of propionyl-CoA can originate from various sources when cultured in different media. Exogenous propanol can be converted to propionate through alcohol dehydrogenase and aldehyde dehydrogenase, along with subsequent conversion to propionyl-CoA. If there is no supplement of propanol in the course of fermentation, propionyl-CoA mainly comes from carbohydrate catabolism, lipolysis of triacylglycerol and degradation of branched-chain amino acids (Figure S3A) (Mironov et al., 2004). We then presented the Kpr ratio of all propionylated enzymes in carbohydrate catabolism and amino acid degradation (Figure S3B), and the result showed that several important enzymes involved in glycolysis and tricarboxylic acid (TCA) cycle were propionylated with elevated level in S. erythraea WT strain, which indicated that lysine propionylation might regulate carbohydrate catabolism and then affect the accumulation of propionyl-CoA from this pathway. In addition, several enzymes controlled by mmsOp1 operon, including methylmalonate-semialdehyde dehydrogenase (mmsA2), acyl-CoA dehydrogenase (sace_1457), and enoyl-CoA hydratase (echA9) were also identified to be propionylated (Figure S3B; Table S2). Previous study has proved that overexpression of mmsOp1 operon could led to the increase of erythromycin yield in S. erythraea strain (Karniar et al., 2016). Therefore, we chose mmsA2 as an example to c explore the effect of lysine propionylation on these metabolic enzymes. According to Figure 3A, mmsA2 catalyzes the decarboxylation reaction of (S)-methylmalonate semialdehyde, which is the last step for precursor propionyl-CoA generation from valine degradation pathway. This pathway was proved to be important in biosynthesis of erythromycin, as 20% elevation of erythromycin yield was observed in E3 strain when treated with valine 4 Cell Chemical Biology 25, 1–12, August 16, 2018

in the medium (Figure S3C). Thus, it was considerable that change in enzymatic activity of mmsA2 could influence the biosynthesis of propionyl-CoA, then affect the precursor supply for erythromycin production. In our proteomic data, five acylated sites were identified in mmsA2, among which K12, K94, and K100 gained higher propionylation level (ratio > 1.5) in WT strain in contrast to K100 and K174 with unchanged acetylation level (Figure 3A), suggesting that propionylation other than acetylation may regulate enzymatic activity. We carried out sequence alignment analysis and explored possible roles of the identified propionylated lysine residues on the basis of the conversation of the three propionylated sites. The result clearly showed that K12 site and K94 site were conserved from eukaryotes to prokaryotes, but K100 was unconserved (Figure 3B). We next mutated K12 site and K94 site into glutamine, which mimicked lysine propionylation with a neutral charge to investigate the possible effect of propionylation. As shown in Figures 3C and 3D, activity of mmsA2K94Q decreased more than 70% comparing with mmsA2, indicating that the K94 was the key site for the enzymatic activity of mmsA2 (Figure 3D). To further prove the propionylation regulation on this site, we nonspecifically propionylated three purified proteins with propionyl-CoA through an in vitro reaction. The result showed that enzymatic activity of the propionylated mmsA2 and propionylated mmsA2K12Q decreased as compared with their nonpropionylated forms, respectively. Meantime, the quantitative mass spectrographic result showed that propionylation of K94 in mmsA2 elevated in the in vitro propionylation reaction (Figure S3D). However, no change in activity of propionylated mmsA2K94Q was observed (Figure 3D). Together, these results demonstrated that propionylation at K94 of mmsA2 inhibited its activity, and thus decreased the supply of propionyl-CoA for erythromycin biosynthesis in S. erythraea. In addition to mmsA2, two propionylated sites were identified in sace_1457 and echA9 (the other two enzymes involved in mmsOp1 operon), respectively, and both of the two sites had higher level of propionylation in WT strain than that in E3 strain. According to the sequence alignment of the two sites, K316 in sace_1457 and K261 in echA9 were conserved (Figure S3E). Therefore, we suggested that propionylation could not only affect the activity of mmsA2, it might also inactivate sace_1457

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Figure 3. Lysine Propionylation Influenced the Endogenous Supply of Propionyl-CoA in S. erythraea (A) The acetylated/propionylated sites in mmsA2. Circles/diamonds represent the propionylated/acetylated sites, respectively. (B) Sequence alignment of methylmalonate-semialdehyde dehydrogenase (mmsA2), from Homo sapiens, Rattus norvegicus, Saccharomyces kudriavzevii, B. subtilis, S. coelicolor, and S. erythraea. (C) Purity of mmsA2 and its mutant shown by SDS-PAGE gel. (D) Activity of mmsA2 and its mutant before and after propionylation reaction.

and sace_1458, inhibiting simultaneously the supply of propionyl-CoA from amino acid degradation. Lysine Propionylation Impacted Supply of PropionylCoA from Assimilation of Exogenous n-propanol In the erythromycin fermentation process of industrial E3 strain, n-propanol is supplemented and then metabolized to propionylCoA for erythromycin biosynthesis (Figure S3A). To explore the relationship between n-propanol supplement and erythromycin yield, we cultured E3 strain in the increasing concentration of n-propanol from 0% to 9% with WT strain as a comparison. From Figure S4A, we found that cell growth and bacterial morphology was not obviously changed in both of the two S. erythraea strain when less than 1% n-propanol was added in the medium, which indicated that relative low n-propanol concentration (less than 1%) could not result in toxicity on the growth status of S. erythraea. In addition, a maximum yield of erythromycin was obtained at the n-propanol concentration of 0.3% for E3 strain. When the concentration of n-propanol was higher than 0.3%, the yield of erythromycin decreased gradually (Figure 4A). However, for S. erythraea WT strain, no obvious change of erythromycin yield was observed when treated with all concentrations of n-propanol (Figure 4B). Western blot analysis showed that global propionylation level of WT strain was elevated with increasing n-propanol concentrations. In comparison, protein propionylation levels of E3 strain remained constant in all n-propanol concentrations (Figure S4B). Ability of utilizing propionyl-CoA for erythromycin production of E3 strain was

higher than that of WT strain. Intracellular propionyl-CoA-derived from n-propanol supplement might be a rate-limiting step of erythromycin biosynthesis in S. erythraea E3 strain under condition of less than 0.3% n-propanol. The addition of more n-propanol than 0.3% would cause imbalance between the generation of propionyl-CoA and the efficiency of its use in various metabolic pathways, and thus result in the accumulation of high concentrations of propionyl-CoA and possible propionylation-mediated activity loss of certain important enzymes in E3 strain, similar to the case of mmsA2 as mentioned above in WT strain. We further investigated interplay among the addition of n-propanol, the accumulation of propionyl-CoA, and protein propionylation in the industrial E3 strain. Mass spectrographic assays showed that propionyl-CoA concentration greatly increased under condition of less than 0.3% n-propanol supplement, and increased only a little by elevating n-propanol to 0.66% (Figure S4C). Although no obvious concentration change was observed for cellular acetyl-CoA and succinyl-CoA as compared with propionyl-CoA, our result showed a slight decrease for acetyl-CoA concentration and a slight increase for succinyl-CoA concentration in high n-propanol supplement (Figure S4C). The observations indicated the increase in intracellular propionyl-CoA level derived from n-propanol might also promote the conversion of acetyl-CoA to succinyl-CoA through tricarboxylic cycle. Moreover, mass spectrometry-based quantitative propionylome was conducted for E3 strain cultured in 0.3% propanol (heavy labeled) and 1% propanol (light labeled). Protein propionylome data showed that the median 0.3%/1% Cell Chemical Biology 25, 1–12, August 16, 2018 5

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Figure 4. The Effect of Propionylation on the Exogenous Pathway of Erythromycin Biosynthesis in S. erythraea (A) Erythromycin production of the E3 strain in increasing concentrations of n-propanol. Data are represented as mean ± SEM. (B) Erythromycin production of the WT strain. Annotation is the same as in (A). (C) Lysine propionylation of the two important enzymes involved in the n-propanol supplemental pathway. Scatterplot of the quantifiable lysinepropionylated sites in the E3 strain cultured in 0.3% and 1% n-propanol. Blue dots, normalized H/L ratio less than 0.5; red dots, normalized H/L ratio between 0.5 and 2; green dots, normalized H/L ratio higher than 2.

ratio was 1.27, which was in accordance with the western blot analysis (Figure 4C; Tables S4 and S5). But we also found that the two enzymes involved in the assimilation of propanol pathway and synthesis of S-adenosylmethionine were more heavily propionylated in 1% propanol than in 0.3% propanol (Figure 4C). One was sace_6566 (aldehyde dehydrogenase), which catalyzed the carboxylation reaction from propylaldehyde to propionate and was directly responsible for propionyl-CoA generation. Another was sace_2103 (S-adenosylmethionine synthetase, metK), which provided methyl group for the conversion of erythromycin C to the end-product erythromycin A, and was proved to be important in the erythromycin biosynthesis (Wang et al., 2007). As shown in Figure 4C, the 0.3%/1% ratio for the K6 of sace_6566 was 0.32. Sequence alignment analysis showed that the K6 of sace_6566 was conserved in actinobacteria (Figure S4D). We next mutated site K6 into glutamine and compared the enzymatic activity of WT with that of the mutant. Photometrical activity assay showed that sace_6566K6Q remained only 40% activity, indicating that feeder pathway of propionyl-CoA biosynthesis could also be inhibited by lysine propionylation in E3 strain when excess exogenous propanol was added in the medium (Figure S4E). Sequence alignment analysis indicated that the modified K274 of metK (with 0.34 heavy-to-light [H/L] ratio shown in Figure 4C) was also conserved from prokaryotes to eukaryotes (Figure S4F). Previous study showed that the homologous site in Escherichia coli metK was identified to be vital for its enzymatic activity (Taylor and Markham, 2000), which was further validated by our high-performance liquid chromatography analysis for the lower S-adenosyl-L-methionine concentration proved in S. erythraea with 1% n-propanol addition (Figure S4G). The result suggested that propionylation could inhibit the important enzymes in pathways of precursors supply when excessive propanol was added in the medium. Taken together, these findings uncovered the function of lysine propionylation involved in propionyl-CoA-derived erythromycin biosynthesis, indicating a previously negligible PTMdriven effect have to be considered in the industrial erythromycin fermentation. 6 Cell Chemical Biology 25, 1–12, August 16, 2018

Depropionylation Relieved the Enzymatic Inhibition and Relatively Increased Erythromycin Productivity As the decrease in erythromycin yield was caused by lysine propionylation, we thus asked whether manipulation of the reversible lysine acylation system could reverse lysine propionylation and then rescue the enzymatic activity to increase erythromycin productivity. In S. erythraea, the sirtuin-type NAD+-dependent deacetylase (named as SeSrtN and encoded by sace_3798) (Figure 5A) was responsible for deacetylation of acetyl-CoA synthetase and glutamine synthetase (You et al., 2014; You et al., 2016). Its homologous enzyme in E. coil, CobB had a promiscuous deacylation ability (including deacetylation, desuccinylation, and also depropionylation), while showing no preference for both enzymatic and nonenzymatic acylated sites (Abouelfetouh et al., 2015; Castan˜ocerezo et al., 2011; Colak et al., 2013; Garrity et al., 2007). Therefore, we hypothesized that SeSrtN could also act as a bond fide depropionylase. To investigate the depropionylation activity of SeSrtN, the two propionylated peptides (_TYK(pr)LYVGGK_ and _HGGGAFSGK(pr)DPSK_) from sace_6566 and metK were synthesized and the result showed that the propionyl groups of these two peptides could be removed by SeSrtN in the in vitro depropionylation reaction (Figure 5B). Meanwhile, the comparative proteomics assay for WT and E3 strains showed that protein expression of sace_3798 in high-yield E3 strain was about two times of that in WT strain (Figure S5A), increasing its ability of depropionylation, which might be another reason for the lower propionylation level in high-yield E3 strain. After that, we compared the erythromycin production in WT and sace_3798-deletion mutant strain and found the mutant strain could only generate less than 70% of the erythromycin production in WT strain (Figure S5B). Furthermore, about 15% elevation in erythromycin production was observed in WT (O_SeSrtN) strain as compared with WT strain (Figure S5C), which would be a direct evidence that the expression level of SeSrtN was positively related with the erythromycin production through depropionylation mechanism. Based on these observations, we thought that overexpression of depropionylase SeSrtN in industrial S. erythraea E3 strain might be possible to relieve the lysine propionylation inhibition

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Figure 5. Depropionylation Relatively Reversed Propionylation Inhibition in S. erythraea (A) The catalytic mechanism of SeSrtN. (B) Depropionylation analysis of the lysine-propionylated peptides identified in the two enzymes catalyzed by SeSrtN. (C) Western blotting analysis of the protein propionylation level in the E3/E3 (O_SeSrtN) strains. The intensity representing the normalized gray level of the two lines. (D) Relative erythromycin production in E3 strain and E3 (O_SeSrtN) strain. Data are represented as mean ± SEM.

in the erythromycin fermentation process. We then constructed sace_3798 overexpression E3 (O_SeSrtN) strain. To investigate the effect of SeSrtN on global propionylation level, we cultured both E3 and E3 (O_SeSrtN) strains in high propanol concentration (0.66%). Western blotting results of whole-cell lysates with anti-propionyllysine antibody showed decreased protein propionylation level in the E3 (O_SeSrtN) strain compared with E3 strain (about 25% decrease according to the gray level difference analysis) (Figure 5C), which demonstrated that the hyperpropionylation overflow could be partly removed by overexpression of SeSrtN. In addition, in vivo mass spectrographic data showed that propionylation level of the modified sites in sace_6566 and metK decreased, further proving its ability for depropionylating the two key enzymes (Figures S5D and S5E). Furthermore, amount of erythromycin increased by more than 30% (Figure 5D) in E3 (O_SeSrtN) strain, which suggested its positive role in increasing the erythromycin production. This finding revealed that manipulating propionylation regulatory network through overexpression of the depropionylase successfully reversed lysine propionylation and thus improved the erythromycin yield. The Similar Acylation Regulation Involved in Secondary Metabolism was Observed in S. coelicolor and C. acetobutylicum Acyl-CoAs function as building blocks for several kinds of natural products, which include polyketides, alkaloids, fatty acids, bioalcohols, alkane, biodiesel, isoprenoids, and polyhydroxyalkanoates (Krivoruchko et al., 2015). Erythromycin is a typical macrolide polyketide with propionyl-CoA as its direct precursor. Our result showed that accumulation of cellular propionyl-CoA could lead to lysine propionylation-mediated inhibition of important enzymes involved in pathways of precursor supply, and then decreased the erythromycin productivity in S. erythraea. We

next explored whether lysine acylation could be a general mechanism in natural product biosynthesis. Undecylprodiginine, a red pigment, is biosynthesized by at least 18 closely linked genes (red cluster) in S. coelicolor and attracted increasing interests based on its antifungal, antibacterial, antiprotozoal, antimalarial, immunosuppressive, and anticancer activities (Papireddy et al., 2011; Williamson et al., 2007) (Figure S6A). As malonylCoA was used as its biosynthetic precursor, we thus explored whether lysine malonylation could regulate the biosynthesis of this red-pigmented prodiginine. Our previous study showed that 100 mM malonate added in S. erythraea elevated protein malonylation level to a large degree (Xu et al., 2016a). But the addition of exogenous malonate could not increase erythromycin production in S. erythraea (Figure S6B). In comparison, a 100% elevation of undecylprodiginine production was observed in S. coelicolor when malonate was added up to 100 mM (Figure S6C). Meantime, no change of lysine malonylation level was observed (Figure S6D), which indicated the additional malonate was efficiently utilized for biosynthesis of its end-product undecylprodiginine. We next increased exogenous malonate concentration up to 760 mM and found that malonate at 600 mM resulted in the maximum yield (Figure S6E). In addition, the protein malonylation level elevated gradually when the concentration of the added malonate was higher than 600 mM (Figure S6F). This result was similar to the case of n-propanol addition experiment as mentioned above, suggesting excessive exogenous malonate might inhibit important enzymes in the biosynthetic pathway of undecylprodiginine by over-malonylation of lysine. Next, we explored the possible acylation regulation in other kinds of acyl-CoA-derived natural products in addition to bioactive compounds. C. acetobutylicum is a strict anaerobic, endospore-forming bacterium, which is used for the production €tkeeversloh and Bahl, 2011). of the advanced biobutanol (Lu In the metabolic engineering of C. acetobutylicum for butanol biosynthesis, exogenous butyrate was used to improve butanol production catalyzed by a series of alcohol/aldehyde dehydrogenase (Figure 6A). We fed the strain with different concentrations of butyrate and found that the maximum butanol yield was achieved at 20 mM of butyrate (Figure 6B). Cell Chemical Biology 25, 1–12, August 16, 2018 7

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Figure 6. Lysine Butyrylation Impacted the Butanol Production in C. acetobutylicum (A) The biosynthetic pathway of butanol production in C. acetobutylicum. (B) Butanol production of C. acetobutylicum in the increasing concentrations of exogenous butyrate. Data are represented as mean ± SEM. (C) Scatterplot of the quantifiable lysine butyrylated sites in C. acetobutylicum cultured in 20 and 100 mM butyrate. Blue dots, normalized H/L ratio less than 0.5; red dots, normalized H/L ratio between 0.5 and 2; green dots, normalized H/L ratio higher than 2. (D) KEGG pathway analysis for butyrylated proteins with an obviously changed butyrylation ratio.

We then conducted quantitative butyrylome analysis of C. acetobutylicum strain cultured in 20 mM (heavy labeled) and 100 mM butyrate (light labeled). The resulted showed median H/L ratio was 0.52 for lysine butyrylation (Figure 6C; Table S6), which was in accordance with the western blot analysis (Figure S7A). Lastly, KEGG pathway analysis was conducted for the modified substrates with obviously changed butyrylation ratio (ratio H/L < 0.5). The result indicated that highly butyrylatd substrates were enriched in the pyruvate metabolism and butyrate metabolism pathways (Figure 6D). Almost all enzymes in the butyrate metabolism pathway were butyrylated when treated with excess butyrate (Figure S7B). Among them, acetaldehyde dehydrogenase 1 (adhe1) and butanol dehydrogenase (bdhA), vital for catalyzing the conversion from butyryl-CoA to the endproduct butanol, bear 7 and 2 modified lysine residues with 8 Cell Chemical Biology 25, 1–12, August 16, 2018

elevated butyrylation ratio, respectively. The sequence alignment assay found that K457 of adhe1 and K43 of bdhA were conserved in microorganisms (Figure S7C). In addition, butyrylation levels of some enzymes responsible for the endogenous precursor butyryl-CoA supplement were elevated, including acetylCoA-acetyltransferase (CA_C2873), 3-hydroxybutyryl-CoA dehydrogenase (hbd), 3-hydroxybutyryl-CoA dehydratase (crt), and butyryl-CoA dehydrogenase (bcd). Three butyrylated lysine in bcd and one butyrylated lysine in crt were conserved from prokaryotes to eukaryotes (Figure S7C). Together, these observations indicated that excess addition of butyrate might lead to butyrylation-mediated activity loss of these important enzymes involved in the butanol-biosynthesized pathway and then decreased butanol production. In conclusion, these results showed that oversupply of acyl-CoA precursors could lead to acylation of enzymes in biosynthetic pathways, thereby decreasing yields of products. This phenomenon widely exists in the biosynthetic pathways for natural products with different skeleton structures, which utilize various acyl-CoAs as precursors. DISCUSSION Characterization of the acylome in many microorganisms has been promoted by recent breakthroughs in sensitive mass spectrometry-based methods, which included E. coli, Salmonella enterica, Erwinia amylovora, Rhodopseudomonas palustris, Geobacillus kaustophilus, Bacillus subtilis, S. erythraea, and Mycobacteria (Ouidir et al., 2016). Studies about lysine acylation revealed its critical functional properties in microbial physiological regulation, which mainly focused on metabolic enzymatic regulation, gene transcription, cell motility and chemotaxis, protein-protein interactions, protein-DNA interactions, protein stability, and sub-cellular distribution (Bernal et al., 2014; Hentchel

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Figure 7. The Outline of Functional Interplay between Beneficial and Adverse Effects of Acyl-CoA Supply on the Synthesis of Acyl-CoA-Derived Natural Products in Microorganisms

and Escalantesemerena, 2015). Acyl-CoAs can function as major donors for lysine acylation through enzymatic and nonenzymatic mechanism, which links protein acylation level and metabolic characteristic of bacterial cell. On the other hand, in the biosynthesis of several kinds of natural products, acyl-CoAs and its derivatives can also be supplied as precursors. Metabolic engineering and synthetic biology tools were used to raise the intracellular acyl-CoA pools, and promoting the conversion of metabolic flux into desired product. Although the two different roles of acyl-CoAs as precursors for biosynthesis and as donors for protein acylation are recognized, it remained to be investigated whether lysine acylation exerts effect on the activities of biosynthetic enzymes, and subsequently disturbs the yield of desired products. Herein, we found that the oversupply of acyl-CoAs resulted in hyperacylation of enzymes in biosynthetic pathways, thereby decreasing yields of end-products. We mainly explored the lysine propionylation-mediated inhibition on the feeder pathways of erythromycin biosynthesis. In addition, we also expanded this mechanism to biosynthesis of other kinds of natural products from different carbon skeletons, including malonylCoA-derived undecylprodiginine and butyryl-CoA-derived butanol (Figure 7). Recent studies proved that in diverse biological processes in microorganisms, over 100 cellular reactions were related with acyl-CoAs, which included glycolysis, TCA cycle, synthesis and degradation of fatty acids and amino acids, and also the biosynthesis of secondary metabolites (Hentchel and Escalantesemerena, 2015). Bulk of acyl-CoA pools and their homeostasis was affected by availability of nutrients, growth stages, diverse environmental conditions and external signal stimulation, in which the balance of acyl-CoA species would be changed by more than an order of magnitude within a short time period (Beld et al., 2014; Hentchel and Escalantesemerena, 2015). As acyl-CoA concentrations reflected cellular energy state, it was conceivable that the imbalance in acyl-CoA homeostasis resulted in intricate consequences for cellular metabolism. In this work, we investigated important roles of protein acylations on the interplay between the supply of acyl-CoAs and the efficiency of their use in biosynthetic pathways of natural product. Acyl-CoA-driven lysine acylation was a kind of carbon overflow, which resulted from the imbalance between formation and

utilization of cellular acyl-CoAs. Mitochondrial acetylation was highly related with alkaline pH and abundant acetyl-CoA in mitochondrial matrix, which was regarded as a representative inhibitory mark for several mitochondrial enzymes (Wagner and Payne, 2013). This mechanism provided negative feedback in response to acetyl-CoA overproduction and then lowered metabolic flux from carbohydrate catabolism (Baeza et al., 2016). Therefore, the acylationmediated inhibition discussed in our study reflected the oversupply of carbon precursors and was directly related with the decreased yield of products. The accumulation of high acyl-CoA concentrations had an effect on the efficiency of their use in various metabolic pathways, including the synthesis of secondary metabolites. We found that the overexpression of sirtuin-type NAD+-dependent deacylase SeSrtN in S. erythraea could reverse the inhibitory effects and promote the subsequent secondary metabolite biosynthesis. In microorganism, the sirtuin-like deacylase has a broad activity for removing different acyl-CoA groups, playing critical role in various homeostatic and stress-related cellular processes. It has been reported that mitochondrial sirtuin can improve protein function and ensure protein quality-control through repairing ‘‘acylation damage’’ on metabolic health (Finkel et al., 2009; Wagner and Hirschey, 2014). As acyl-CoA accumulation can lead to global acylation elevation in mitochondrial matrix, and mitochondrial sirtuins are responsible for optimizing activities of metabolic enzymes. Our current result is highly related with the conclusion of studies about mitochondrial acylation. In conclusion, our study revealed a regulatory mechanism underlying the interaction between biosynthetic pathways and host cell, which expanded the current knowledge of the function of these acyl-CoA-dependent lysine acylation. Moreover, as this regulatory mechanism widely exists in biosynthesis of diverse natural products and directly affects the yield of desired chemicals. It was essential to evaluate the cellular acylation level when designing biosynthetic system for biosynthesizing of desired products. SIGNIFICANCE Lysine acylation is a protein post-translational modification critical for several biological processes, such as epigenetic and metabolic regulations, spermatogenesis, hereditary metabolic diseases, and cancer. As the donors of lysine acylation, acyl-CoAs are also well-recognized precursors for biosynthesis of various natural products, including polyketides, alkaloids, and biofuels, etc. Our present study demonstrated the conceptual implication of acylation Cell Chemical Biology 25, 1–12, August 16, 2018 9

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

regulation on the production of acyl-CoA-derived natural products in microorganisms. Oversupply of propionyl-CoA led to lysine propionylation, which impacted the biosynthetic pathway of erythromycin biosynthesis. In addition, this acylation inhibitory effects widely existed in the biosynthesis of other bioactive compounds or even biofuels, which was regarded as one form of ‘‘carbon burden,’’ impacting several homeostatic and stress-related cellular processes. Overexpression of deacylase could alleviate this acylation damage and give benefit for stimulating the biosynthesis of desired products, providing potential strategies in metabolic engineering. In conclusion, this study raised the insight that lysine acylation could function in the biosynthesis of several different kinds of natural products, which expanded current knowledge of lysine acylation regulatory roles beyond the known physiological and pathological functions.

SUPPLEMENTAL INFORMATION Supplemental Information includes seven figures and seven tables and can be found with this article online at https://doi.org/10.1016/j.chembiol.2018. 05.005. ACKNOWLEDGMENTS This work was supported by grants from the National Natural Science Foundation of China (31730004, 21335003, and 21575089) (to B.-C.Y.), (91753203) (to M.T.). The Special Project on Precision Medicine under the National Key R&D Program (SQ2017YFSF090210) (to M.T.). AUTHOR CONTRIBUTIONS J.-Y.X., M.T., and B.-C.Y. designed the research. J.-Y.X. and Y.X. performed the research with the help of Z.X., L.-H.Z., X.C., and Y.Y. J.-Y.X. analyzed the data and wrote the manuscript and B.-C.Y., M.T., and Y.Z. revised it. DECLARATION OF INTERESTS

STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d

d

d d

KEY RESOURCES TABLE CONTACT FOR REAGENT AND RESOURCE SHARING EXPERIMENTAL MODEL AND SUBJECT DETAILS B S. erythraea Culture and Protein Extraction B S. coelicolor Culture and Protein Extraction B C. acetobutylicum Culture and Protein Extraction METHOD DETAILS B Western Blot Analysis B In-solution Tryptic Digestion B Stable Isotope Dimethyl Labeling B Affinity Enrichment of Peptides Containing Lysine Acylated Peptides B Nano-HPLC-MS/MS Analysis B Ms Data Processing and Analysis B In-gel Tryptic Digestion for Analysis of the Propionylated mmsA2 B Construction of an S. erythraea sace_3798-overexpressing Mutant B Cloning and Mutagenesis of the Enzymes B Expression and Purification of the Enzymes and Their Mutants B In Vitro Methylmalonate-semialdehyde Dehydrogenase Assays B In Vitro Aldehyde Dehydrogenase Assays B HPLC Assay of the Propionyl-CoA in S. erythraea WT and E3 Strain B Measurement the Intracellular Concentrations of Acyl-CoAs B Measurement of Intracellular Concentrations of S-adenosylmethionine B Erythromycin Determination by HPLC B Determination of Undecylprodiginine in S. coelicolor B Determination of Alcohol in C. acetobutylicum by Using the Colorimetric Method B The Depropionylation Activity of SeSrtN QUANTIFICATION AND STATISTICAL ANALYSIS DATA AND SOFTWARE AVAILABILITY

10 Cell Chemical Biology 25, 1–12, August 16, 2018

B.-C.Y. and J.X. are inventors on patent application CN201810030742.7 submitted by Zhejiang University of Technology, which covers the use of the strategies for biotechnological applications of lysine acylation regulation. The authors declare no other financial interests in this work. Received: December 11, 2017 Revised: March 25, 2018 Accepted: May 1, 2018 Published: June 7, 2018 REFERENCES Abouelfetouh, A., Kuhn, M.L., Hu, L.I., Scholle, M.D., Sorensen, D.J., Sahu, A.K., Becher, D., Antelmann, H., Mrksich, M., and Anderson, W.F. (2015). The E. coli sirtuin CobB shows no preference for enzymatic and nonenzymatic lysine acetylation substrate sites. Microbiologyopen 4, 66–83. Baeza, J., Smallegan, M.J., and Denu, J.M. (2016). Mechanisms and dynamics of protein acetylation in mitochondria. Trends Biochem. Sci. 41, 231–244. Bahadur, K., and Saroj, K.K. (2013). Study of the influence of charcoal on the acetone-butanol fermentation by different strains of Clostridium. Jpn. J. Microbiol. 4, 43–51. Beld, J., Finzel, K., and Burkart, M.D. (2014). Versatility of acyl-acyl carrier protein synthetases. Chem. Biol. 21, 1293–1299. Bernal, V., Castan˜o-Cerezo, S., Gallego-Jara, J., E´cija-Conesa, A., de Diego, T., Iborra, J.L., and Ca´novas, M. (2014). Regulation of bacterial physiology by lysine acetylation of proteins. N. Biotechnol. 31, 586–595. Boersema, P.J., Raijmakers, R., Lemeer, S., Mohammed, S., and Heck, A.J. (2009). Multiplex peptide stable isotope dimethyl labeling for quantitative proteomics. Nat. Protoc. 4, 484–494. Castan˜ocerezo, S., Bernal, V., Blancocatala´, J., Iborra, J.L., and Ca´novas, M. (2011). cAMP-CRP co-ordinates the expression of the protein acetylation pathway with central metabolism in Escherichia coli. Mol. Microbiol. 82, 1110–1128. Choi, Y.J., and Lee, S.Y. (2013). Microbial production of short-chain alkanes. Nature 502, 571–574. Choudhary, C., Weinert, B.T., Nishida, Y., Verdin, E., and Mann, M. (2014). The growing landscape of lysine acetylation links metabolism and cell signalling. Nat. Rev. Mol. Cell Biol. 15, 536–550. Chow, J.D.Y., Lawrence, R.T., Healy, M.E., Dominy, J.E., Liao, J.A., Breen, D.S., Byrne, F.L., Kenwood, B.M., Lackner, C., Okutsu, S., et al. (2014). Genetic inhibition of hepatic acetyl-CoA carboxylase activity increases liver fat and alters global protein acetylation. Mol. Metab. 3, 419–431. Colak, G., Pougovkina, O., Dai, L., Tan, M., Te, B.H., Huang, H., Cheng, Z., Park, J., Wan, X., Liu, X., et al. (2015). Proteomic and biochemical studies of lysine malonylation suggest its malonic aciduria-associated regulatory role in

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

mitochondrial function and fatty acid oxidation. Mol. Cell Proteomics 14, 3056–3071. Colak, G., Xie, Z., Zhu, A.Y., Dai, L., Lu, Z., Zhang, Y., Wan, X., Chen, Y., Cha, Y.H., and Lin, H. (2013). Identification of lysine succinylation substrates and the succinylation regulatory enzyme CobB in Escherichia coli. Mol. Cell Proteomics 12, 3509–3520. Cortes, J., Haydock, S.F., Roberts, G.A., Bevitt, D.J., and Leadlay, P.F. (1990). An unusually large multifunctional polypeptide in the erythromycin-producing polyketide synthase of Saccharopolyspora erythraea. Nature 348, 176–178. Dan, H., Li, Z.H., Di, Y., Ying, Z., and Ye, B.C. (2015). Lysine acetylproteome analysis suggests its roles in primary and secondary metabolism in Saccharopolyspora erythraea. Appl. Environ. Microbiol. 99, 1399–1413. Deutsch, J., Rapoport, S.I., and Rosenberger, T.A. (2002). Coenzyme A and short-chain acyl-CoA species in control and ischemic rat brain. Neurochem. Res. 27, 1577–1582. Finkel, T., Deng, C.X., and Mostoslavsky, R. (2009). Recent progress in the biology and physiology of sirtuins. Nature 460, 587–591. , L., Buttarelli, F.R., and Scarpa, S. (2001). Gene Fuso, A., Cavallaro, R.A., Orru silencing by S-adenosylmethionine in muscle differentiation. FEBS Lett. 508, 337–340. Garrity, J., Gardner, J.G., Hawse, W., Wolberger, C., and EscalanteSemerena, J.C. (2007). N-Lysine propionylation controls the activity of propionyl-CoA synthetase. J. Biol. Chem. 282, 30239–30245. Gubbens, J., Zhu, H., Girard, G., Song, L., Florea, B.I., Aston, P., Ichinose, K., Filippov, D.V., Choi, Y.H., Overkleeft, H.S., et al. (2014). Natural product proteomining, a quantitative proteomics platform, allows rapid discovery of biosynthetic gene clusters for different classes of natural products. Chem. Biol. 21, 707–718. Hatter, K., and Sokatch, J.R. (1988). Purification of methylmalonate-semialdehyde dehydrogenase from Pseudomonas aeruginosa PAO. Methods Enzymol. 166, 389–393. Hentchel, K.L., and Escalantesemerena, J.C. (2015). Acylation of biomolecules in prokaryotes: a widespread strategy for the control of biological function and metabolic stress. Microbiol. Mol. Biol. Rev. 79, 321–346. Hertweck, C. (2009). The biosynthetic logic of polyketide diversity. Angew. Chem. Int. Ed. 48, 4688–4716. Hirschey, M.D., and Zhao, Y. (2015). Metabolic regulation by lysine malonylation, succinylation, and glutarylation. Mol. Cell Proteomics 14, 2308–2315. Ho, K.K., and Weiner, H. (2005). Isolation and characterization of an aldehyde dehydrogenase encoded by the aldB gene of Escherichia coli. J. Bacteriol. 187, 1067–1073. Hu, D.X., Withall, D.M., Challis, G.L., and Thomson, R.J. (2016). ChemInform abstract: structure, chemical synthesis, and biosynthesis of prodiginine natural products. Chem. Rev. 47, 7818–7853. ar, K., Drobnak, I., Petek, M., Magdevska, V., Horvat, J., Vidmar, R., Karnic  Rotter, A., Jamnik, P., and Fujs, S.  (2016). Integrated omics Baebler, S., approaches provide strategies for rapid erythromycin yield increase in Saccharopolyspora erythraea. Microb. Cell Fact. 15, 1–17. Keasling, J.D. (2010). Manufacturing molecules through metabolic engineering. Science 330, 1355–1358. Krivoruchko, A., Zhang, Y., Siewers, V., Chen, Y., and Nielsen, J. (2015). Microbial acetyl-CoA metabolism and metabolic engineering. Metab. Eng. 28, 28–42. €tkeeversloh, T., and Bahl, H. (2011). Metabolic engineering of Clostridium Lu acetobutylicum: recent advances to improve butanol production. Curr. Opin. Biotechnol. 22, 634–647. Lackner, G., Bohnert, M., Wick, J., and Hoffmeister, D. (2013). Assembly of melleolide antibiotics involves a polyketide synthase with cross-coupling activity. Chem. Biol. 20, 1101–1106. Lee, J.W., Na, D., Park, J.M., Lee, J., Choi, S., and Sang, Y.L. (2012). Systems metabolic engineering of microorganisms for natural and non-natural chemicals. Nat. Chem. Biol. 8, 536–546.

Li, Y.Y., Xiao, C., Yu, W.B., Hao, L., Ye, Z.Q., Hui, Y., Liu, B.H., Yan, Z., Zhang, S.L., and Ye, B.C. (2013). Systems perspectives on erythromycin biosynthesis by comparative genomic and transcriptomic analyses of S. erythraea E3 and NRRL23338 strains. BMC Genomics 14, 523. Mironov, V.A., Sergienko, O.V., Nastasiak, I.N., and Danilenko, V.N. (2004). Biogenesis and regulation of biosynthesis of erythromycins in Saccharopolyspora erythraea: a review. Prikl. Biokhim. Mikrobiol. 40, 613–624. Nielsen, J., and Keasling, J.D. (2011). Synergies between synthetic biology and metabolic engineering. Nat. Biotechnol. 29, 693–695. Okanishi, H., Kim, K., Masui, R., and Kuramitsu, S. (2014). Lysine propionylation is a prevalent post-translational modification in Thermus thermophilus. Mol. Cell Proteomics 13, 2382–2398. Ouidir, T., Kentache, T., and Hardouin, J. (2016). Protein lysine acetylation in bacteria: current state of the art. Proteomics 16, 301–309. Papireddy, K., Smilkstein, M., Kelly, J.X., Shweta, Salem, S.M., Alhamadsheh, M., Haynes, S.W., Challis, G.L., and Reynolds, K.A. (2011). Antimalarial activity of natural and synthetic prodiginines. J. Med. Chem. 54, 5296–5306. Qian, L., Nie, L., Chen, M., Liu, P., Zhu, J., Zhai, L., Tao, S.C., Cheng, Z., Zhao, Y., and Tan, M. (2016). Global profiling of protein lysine malonylation in Escherichia coli reveals its role in energy metabolism. J. Proteome Res. 15, 2060–2071. Rokem, J.S., Lantz, A.E., and Nielsen, J. (2007). Systems biology of antibiotic production by microorganisms. Nat. Prod. Rep. 24, 1262–1287. Rowlands, R.T. (1984). Industrial strain improvement: mutagenesis and random screening procedures. Enzyme Microb. Technol. 6, 3–10. Ryu, Y.G., Butler, M.J., Chater, K.F., and Lee, K.J. (2006). Engineering of primary carbohydrate metabolism for increased production of actinorhodin in Streptomyces coelicolor. Appl. Environ. Microbiol. 72, 7132–7139. Sabari, B.R., Di, Z., Allis, C.D., and Zhao, Y. (2017). Metabolic regulation of gene expression through histone acylations. Nat. Rev. Mol. Cell Biol. 18, 90–101. Scherlach, K., and Hertweck, C. (2009). Triggering cryptic natural product biosynthesis in microorganisms. Org. Biomol. Chem. 7, 1753–1760. Schilling, B., Christensen, D., Davis, R., Sahu, A.K., Hu, L.I., WalkerPeddakotla, A., Sorensen, D.J., Zemaitaitis, B., Gibson, B.W., and Wolfe, A.J. (2016). Protein acetylation dynamics in response to carbon overflow in Escherichia coli. Mol. Microbiol. 98, 847–863. Spitzer, M., Wildenhain, J., Rappsilber, J., and Tyers, M. (2014). BoxPlotR: a web tool for generation of box plots. Nat. Methods 11, 121–122. Staunton, J., and Weissman, K.J. (2001). Polyketide biosynthesis: a millennium review. Nat. Prod. Rep. 18, 380–416. Stephanopoulos, G., Alper, H., and Moxley, J. (2004). Exploiting biological complexity for strain improvement through systems biology. Nat. Biotechnol. 22, 1261–1267. Taylor, J.C., and Markham, G.D. (2000). The bifunctional active site of S-adenosylmethionine synthetase. Roles of the basic residues. J. Biol. Chem. 275, 4060–4065. Wagner, G.R., Bhatt, D.P., O’Connell, T.M., Thompson, J.W., Dubois, L.G., Backos, D.S., Yang, H., Mitchell, G.A., Ilkayeva, O.R., and Stevens, R.D. (2017). A class of reactive acyl-CoA species reveals the non-enzymatic origins of protein acylation. Cell Metab. 25, 823–837. Wagner, G.R., and Hirschey, M.D. (2014). Nonenzymatic protein acylation as a carbon stress regulated by sirtuin deacylases. Mol. Cell 54, 5–16. Wagner, G.R., and Payne, R.M. (2013). Widespread and enzyme-independent Nε-acetylation and Nε-succinylation of proteins in the chemical conditions of the mitochondrial matrix. J. Biol. Chem. 288, 29036–29045. Wang, Q., Zhang, Y., Yang, C., Lin, Y., Yao, J., Li, H., Xie, L., Zhao, W., Yao, Y., and Ning, Z.B. (2010). Acetylation of metabolic enzymes coordinates carbon source utilization and metabolic flux. Science 327, 1004–1007. Wang, Y., Wang, Y., Chu, J., Zhuang, Y., Zhang, L., and Zhang, S. (2007). Improved production of erythromycin A by expression of a heterologous

Cell Chemical Biology 25, 1–12, August 16, 2018 11

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

gene encoding S-adenosylmethionine synthetase. Appl. Environ. Microbiol. 75, 837–842.

regulator, negatively controls erythromycin biosynthesis in Saccharopolyspora erythraea. J. Ind. Microbiol. Biotechnol. 41, 1159–1167.

Weinert, B., Iesmantavicius, V., Wagner, S., Scho¨lz, C., Gummesson, B., Beli, P., Nystro¨m, T., and Choudhary, C. (2013a). Acetyl-phosphate is a critical determinant of lysine acetylation in E. coli. Mol. Cell 51, 265–272.

Xu, J.Y., Xu, Z., Zhou, Y., and Ye, B.C. (2016a). Lysine malonylome may affect the central metabolism and erythromycin biosynthesis pathway in Saccharopolyspora erythraea. J. Proteome Res. 15, 1685–1701.

Weinert, B.T., Scho¨lz, C., Wagner, S.A., Iesmantavicius, V., Su, D., Daniel, J.A., and Choudhary, C. (2013b). Lysine succinylation is a frequently occurring modification in prokaryotes and eukaryotes and extensively overlaps with acetylation. Cell Rep. 4, 842–851.

Xu, Y., Liao, C.H., Yao, L.L., Ye, X., and Ye, B.C. (2016b). GlnR and PhoP directly regulate the transcription of genes encoding starch-degrading, amylolytic enzymes in Saccharopolyspora erythraea. Appl. Environ. Microbiol. 82, 6819–6830.

Williamson, N.R., Fineran, P.C., Gristwood, T., Chawrai, S.R., Leeper, F.J., and Salmond, G.P. (2007). Anticancer and immunosuppressive properties of bacterial prodiginines. Future Microbiol. 2, 605–618.

You, D., Yao, L.L., Huang, D., Escalante-Semerena, J.C., and Ye, B.C. (2014). Acetyl coenzyme A synthetase is acetylated on multiple lysine residues by a protein acetyltransferase with a single Gcn5-type N-acetyltransferase (GNAT) domain in Saccharopolyspora erythraea. J. Bacteriol. 196, 3169–3178.

Williamson, N.R., Fineran, P.C., Leeper, F.J., and Salmond, G.P.C. (2006). The biosynthesis and regulation of bacterial prodiginines. Nat. Rev. Microbiol. 4, 887–899. Wu, P., Pan, H., Zhang, C., Wu, H., Yuan, L., Huang, X., Zhou, Y., Ye, B.C., Weaver, D.T., and Zhang, L. (2014). SACE_3986, a TetR family transcriptional

12 Cell Chemical Biology 25, 1–12, August 16, 2018

You, D., Yin, B.C., Li, Z.H., Zhou, Y., Yu, W.B., Zuo, P., and Ye, B.C. (2016). Sirtuin-dependent reversible lysine acetylation of glutamine synthetases reveals an autofeedback loop in nitrogen metabolism. Proc. Natl. Acad. Sci. USA 113, 6653–6658.

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

STAR+METHODS KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Pan-anti-acetyllysine antibody

PTM Biolabs (Hangzhou, China)

Cat#PTM-101

Pan-anti-acetyllysine antibody beaded agarose

PTM Biolabs (Hangzhou, China)

Cat#PTM-103

Pan-anti-propionyllysine antibody

PTM Biolabs (Hangzhou, China)

Cat#PTM-201

Antibodies

Pan-anti-propionyllysine antibody beaded agarose

PTM Biolabs (Hangzhou, China)

Cat#PTM-202

Pan-anti-butyryllysine antibody

PTM Biolabs (Hangzhou, China)

Cat#PTM-301; RRID: AB_2687946

Pan-anti-butyryllysine antibody beaded agarose

PTM Biolabs (Hangzhou, China)

Cat#PTM-302

Pan-anti-malonyllysine antibody

PTM Biolabs (Hangzhou, China)

Cat#PTM-901; RRID: AB_2687947

Pan-anti-succinyllysine antibody

PTM Biolabs (Hangzhou, China)

Cat#PTM-401; RRID: AB_2687628

S. erythraea WT strain

(Li et al., 2013)

n/a

S. erythraea E3 strain

(Li et al., 2013)

n/a

S. erythraea WT (DSeSrtN) strain

(You et al., 2014)

n/a

S. erythraea E3 (O_SeSrtN) strain

This paper

n/a

S. erythraea WT (O_SeSrtN) strain

This paper

n/a

S. coelicolor strain

(Xu et al., 2016b)

n/a

C. acetobutylicum strain

ATCC

Cat#ATCC 824

E. coli DH5a

TransGen Biotech (Beijing, China)

Cat#CD201

E. coli BL21 (DE3)

TransGen Biotech (Beijing, China)

Cat#CD601

Peptide _TYK(pr)LYVGGK_

GL Biochem (Shanghai, China)

n/a

Peptide _HGGGAFSGK(pr)DPSK_

GL Biochem (Shanghai, China)

n/a

Recombinant Protein (PET28a-1456)

This paper

n/a

Recombinant Protein (pET28a-1456 (K12Q))

This paper

n/a

Recombinant Protein (pET28a-1456 (K94Q))

This paper

n/a

Recombinant Protein (pET28a-6566)

This paper

n/a

Recombinant Protein (pET28a-6566 (K6Q))

This paper

n/a

Recombinant Protein (pET28a-SeSrtN)

(You et al., 2014)

n/a

C18 ZipTips

Millipore, Billerica, MA

Cat#ZTC18S096

SepPak C18 cartridges

Waters, Milford, MA

Cat#WAT023590

Trypsin

Promega, Billerica, MA

n/a

Bacterial and Virus Strains

Chemicals, Peptides, and Recombinant Proteins

Protease inhibitor cocktail

Roche, Mannheim, Germany

Cat#179382

TSBY medium

Oxoid, Basingstoke, UK

n/a

chemiluminescent HRP substrate

Immobilon Western, Millipore, Germany

Cat#WBKLS0050

Other chemicals

Sigma-Aldrich

n/a

Beyotime Biotechnology (Shanghai, China)

Cat#P0011

Expression clone E. coli BL21 (DE3) with pET28a-1456

This paper

n/a

Expression clone E. coli BL21 (DE3) with pET28a-1456 (K12Q)

This paper

n/a

Expression clone E. coli BL21 (DE3) with pET28a-1456 (K94Q)

This paper

n/a

Expression clone E. coli BL21 (DE3) with pET28a-6566

This paper

n/a

Expression clone E. coli BL21 (DE3) with pET28a-6566 (K6Q)

This paper

n/a

Critical Commercial Assays BCA protein assay kit Experimental Models: Organisms/Strains

(Continued on next page)

Cell Chemical Biology 25, 1–12.e1–e6, August 16, 2018 e1

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Continued REAGENT or RESOURCE

SOURCE

IDENTIFIER

Expression clone E. coli BL21 (DE3) with pET28a-SeSrtN

(You et al., 2014)

n/a

This paper

iProx database: IPX0000970001

This paper

n/a

Deposited Data Proteomics data of this study Oligonucleotides See Table S7 for list of primers used in this study Recombinant DNA PIB139-SeSrtN

This paper

n/a

pET28a-1456

This paper

n/a

pET28a-1456 (K12Q)

This paper

n/a

pET28a-1456 (K94Q)

This paper

n/a

pET28a-6566

This paper

n/a

pET28a-6566 (K6Q)

This paper

n/a

pET28a- SeSrtN

(You et al., 2014)

n/a

MaxQuant

Version 1.4.1.2

http://www.coxdocs.org/ doku.php?id=maxquant

GraphPad Prism

Version 5.01

https://www.graphpad.com/

BoxPlotR (online)

(Spitzer et al., 2014)

http://boxplot.tyerslab.com/

Software and Algorithms

CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Bang-Ce Ye ([email protected]). EXPERIMENTAL MODEL AND SUBJECT DETAILS S. erythraea Culture and Protein Extraction S. erythraea NRRL23338 WT and E3 strain was grown in TSBY medium (30 g of tryptone soya broth and 5 g of yeast extract per L of distilled H2O) at 30 C and cultured for 24 h. We then add 1 ml cultured cells to new medium for further cultivation and each for three biological replicates. During the late stationary phase, cells were collected and washed with cold PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KHPO4) twice before lysis. The pellet was resuspended in the lysis buffer (8 M Urea in PBS buffer, containing protease inhibitor cocktail and 20 mM nicotinamide) and sonicated for 5 min. After being incubated on ice for 20 min, cell debris was removed by centrifugation at 21000 g for 30 min. Protein concentration was monitored by the BCA method using lysis buffer as control. The protein was combined from the three replicates. In the n-propanol addition experiment, S. erythraea WT and E3 strain were cultured in TSBY medium at 30 C for 24 h. 1 ml cultured cells were added into new medium for 24 h further cultivation. After that, different amounts of n-propanol were added in the medium as controlling the final n-propanol concentration at 0%, 0.03%, 0.1%, 0.3%, 0.66%, 1%, 3% and 9%. During the late stationary phase, cells were collected and proteins were extracted as mentioned above. Three replicates were performed. Similarly, S. erythraea WT (DSeSrtN) strain, S. erythraea WT (O_SeSrtN), and S. erythraea E3 (O_SeSrtN) were cultured in TSBY medium with or without treatment of n-propanol. Bacterial cells were collected and their proteins were extracted as mentioned above. S. coelicolor Culture and Protein Extraction The strain of S. coelicolor was cultured in tryptic soy broth (TSB) for 25 h at 30 C. Then, 1 ml cultured cells was added to 50 ml modified Evans medium (per liter of distilled water was supplemented with 1% glucose, 7.5 mM (NH4)2SO4, 8 mM NaH2PO4, 10 mM KCl, 1.25 mM MgCl2, 2 mM Na2SO4, 2 mM citric acid, 0.25 mM CaCl2, 1 mM NaMoO4, 25 mM TES [N-tris (hydroxymethyl) methyl-2-aminoethanesul-phonic acid] and 0.5% Evans trace elements, pH 7.2). When the strain entered into the stationary phase, malonate with concentrations of 0 mM, 6.6 mM, 20 mM, 60 mM, 100 mM, 300 mM, 600 mM, 700 mM and 760 mM was added to the medium and each for three biological replicates. Protein extraction process was same to that described above. C. acetobutylicum Culture and Protein Extraction The medium for C. acetobutylicum strain was containing (grams per liter): 2 g (NH4)2SO4, 1 g K2HPO4, 0.5 g KH2PO4, 0.1 g MgSO4,7H2O, 0.015 g FeSO4,7H20, 0.01 g CaCl2, 0.01 g MnSO4,H20, 0.002 g CoCl2, 0.002 g ZnSO4,7H20, 0.00025 g Na2SeO3, 2 g tryptone, 1 g yeast extract, 50 g glucose. The cells were cultured in the anaerobic Hungate-type tube with screw cap and butyl e2 Cell Chemical Biology 25, 1–12.e1–e6, August 16, 2018

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

rubber septum at 37 C for 24 h. The volume of the culture was scaled up to 100 ml through transfer. When the strain entered into stationary phase, butyrate with concentrations of 0 mM, 6.6 mM, 20 mM, 60 mM and 180 mM was added to the medium and each for three biological replicates. Protein extraction process was same to that described above. METHOD DETAILS Western Blot Analysis Protein lysate was separated by 12% SDS-PAGE and then transferred to the nitrocellulose membrane for 90 min at 100 V. After blocking with 3% BSA in PBST buffer (phosphate buffered saline containing 0.1% Tween-80) at room temperature for 1 h, the membrane was incubated with different pan anti-acyllysine antibodies (anti-acetyllysine/ anti-propionyllysine/ anti-butyryllysine/ anti-malonyllysine/ anti-succinyllysine) at 4 C overnight. After the membrane was washed with PBST buffer four times for 8 min each, it was incubated with diluted HRP conjugated secondary antibody for 40 min at room temperature with gentle shaking. The membrane was then washed with PBST buffer four times for 8 min each, the blot was performed via ImageQuant LAS 4000 system (GE Healthcare, UK) after chemiluminescent HRP substrate treatment. In-solution Tryptic Digestion 6 mg of Protein extracts were reduced in 5 mM dithiothreitol at 56 C for 30 min, followed by incubation with 15 mM iodoacetamide at room temperature in the darkness for 30 min. The alkylation reaction was quenched with 30 mM cysteine at room temperature for 30 min. Trypsin (an enzyme-to-substrate ratio of 1:50) was added to the protein solution at 37 C for 16 h after 4-fold dilution. In the next day, additional trypsin (an enzyme-to-substrate ratio of 1:100) was added for a 4 h complete digestion. The tryptic peptides were desalted through SepPak C18 cartridges and dried by Speed Vac. Stable Isotope Dimethyl Labeling The stable isotope dimethyl labeling method was based on the protocol previously reported (Boersema et al., 2009). In general, 6 mg of peptides from each group were reconstituted in 600 ml of triethylammonium bicarbonate buffer (100 mM). Then, 30 ml of 20% CH2O or CD2O was added to sample to be light-labeled or heavy-labeled, and 30 ml of 3 M NaBH3CN was added to each samples, respectively. The reaction solution was incubated for 60 min at room temperature and the labeling efficiency were confirmed. We then quenched the reaction with 30 mL of ammonia (20%) and acided it with TFA for solid-phase extraction. A gradient elution (9%, 15%, 20%, 25%, 30%, and 80%) was used for separating the sample. Three fractions were acquired for anti-acetyllysine or anti-propionyllysine antibodies enrichment and two fractions were acquired for anti-butyryllysine antibody enrichment. Affinity Enrichment of Peptides Containing Lysine Acylated Peptides The acetyllysine/ propionyllysine/ butyryllysine (Kac/ Kpr/ Kbu) antibody beaded agarose were used for enrichment. Briefly, peptides from each HPLC fraction were resolubilized in NETN buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, 0.5% NP-40, pH 8.0). Samples were centrifuged at 21000 g for 10 min to remove insoluble particles. The peptides of each fraction were incubated with 20 ml of agarose beads conjugated with anti-acetyllysine, anti-propionylated or anti-butyrylated antibody beaded agarose at 4 C overnight with gentle rotation. The beads were then washed three times with NETN buffer, once with ETN buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, pH 8.0), and twice with purified water. Enriched Kac, Kpr or Kbu peptides were eluted from the beads after washed three times with 0.1% trifluoroacetic acid. The eluted peptides were desalted with C18 ZipTips and then dried in SpeedVac. Nano-HPLC-MS/MS Analysis Modified peptides were dissolved in solvent A (0.1% (v/v) formic acid and 2% acetonitrile in water) and then were injected onto a manually packed reverse-phase C18 column (10 cm length 3 75 mm inner diameter; C18 resin with 3 mm particle size; 90 A˚ pore diameter; Dikma Technologies Inc., Lake Forest, CA) coupled to an EASY-nLC 1000 system (Thermo Fisher Scientific, Waltham, MA) and eluted by 90 min gradient with 8%-35% solvent B (0.1% formic acid and 10% water in acetonitrile) for 54 min, 35%-45% solvent B for 10 min, 45%-80% solvent B for 3 min, and 80% solvent B for 3 min at a flow rate of 300 nl/min. The HPLC elute was directly electrosprayed into an orbitrap Fusion mass spectrometer (Thermo Fisher Scientific) using a nanospray source. Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific) was used for analysis. Full MS spectra with an m/z range of 300 to 1400 with a mass resolution of 120000 at m/z 200 were acquired. The electrospray voltage was 2.2 kV. Automatic gain control was used by setting the ion count at 5E5 for orbitrap. The isolation window (m/z) was set as 1, and precursor ions with intensity greater than 5000 are used for MS2 analysis. MS/MS acquisition was performed in top-speed mode with a 3s cycle time. Ions with charge states 2+ to 6+ were fragmented in the ion trap by higher energy collisional dissociation (HCD) with normalized collision energy (NCE) of 32%. The dynamic exclusion duration was set as 30 s. Two technical replicates were conducted for all quantitative analysis of lysine acylomes. Peptides used for protein level normalization was also analyzed by Orbitrap Fusion mass spectrometer. Peptides were also dissolved in solvent A and eluted by a 60 min gradient with 8%-32% solvent B for 51 min, 32%-48% solvent B for 5 min, 48%-80% solvent B for 1 min, and 80% B for 3 min at a flow rate of 300 nl/min. The electrospray voltage was controlled at 2.2 kV. The resolution was set at 120000 at m/z 200. The m/z scan range was 350 to 1300 m/z. When the intensity of precursor ions was greater than 5000, they were used for MS2 analysis. MS/MS acquisition was conducted in top-speed mode with a cycle time of 3s. Ions with 2+ to 6+ charge states were fragmented in HCD model with normalized collision energy (NCE) of 32%. 60 s was set as the dynamic exclusion Cell Chemical Biology 25, 1–12.e1–e6, August 16, 2018 e3

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

duration. Two technical replicates were conducted for protein profiling of S. erythraea WT and E3 strain. The results only used for protein level normalization were performed once. Ms Data Processing and Analysis MaxQuant software (version 1.4.1.2) was used for identifying and quantifying protein and the acylated peptides. The isotope dimethyl labeling pair search was conducted, in which the peptide N-terminal and lysine dimethylation (+28 Da; 2 CH2) are set for light and (+32 Da; 2 CD2) are set for heavy. S. erythraea database from Uniprot (7154 sequences; release date, September 2015) or C. acetobutylicum database from Uniprot (3847 sequences; release date, November 2015) was used for data processing. Trypsin/P was specified as the cleavage enzyme, and the maximum number of missed cleavage was set at 2. Mass error was set to 20 ppm for precursor ions and 0.5 Da for fragmentations. Carbamidomethyl (C) was specified as the fixed modification and variable modifications were oxidation (M), acetylation (Protein N-term) and acetylation/propionylation/butyrylation (K). False discovery rates (FDR) at protein, peptide and modification level were all set as 0.01. In the quantification, Lys acylation site identifications with localization lower than 0.75, or from reverse or contaminant protein sequences were removed. The ratios of the acylated peptides were normalized to the ratios of their corresponding proteins level. Normalized ratios of the peptides were used for further analysis. In-gel Tryptic Digestion for Analysis of the Propionylated mmsA2 For determining the propionylation level of K94 in mmsA2. The gel bands in SDS-PAGE were sliced and destained in 50% ethanol treatment. After fully dehydrated in 100% ACN, the samples were reduced by 10 mM DTT at 56 C for 40 min and then alkylated by 15 mM iodoacetamide (IAA) in darkness for another 40 min. After that, the gels were washed in washing buffer (50% ACN/50% mM NH4HCO3 (v/v)) and proteins were digested by trypsin at an enzyme to substrate ratio of 1:40 for 16 hours. The tryptic peptides were extracted in 50% ACN/ 5% TFA, 75%/0.1% TFA and 100% TFA in sequence. The samples were then dissolved in solvent A and analyzed by the orbitrap Fusion mass spectrometer in two technical replicates. The raw data was converted to mgf files and then analyzed by Mascot search engine (v 2.3.01, Matrix Science). The search parameters were enzymes, trypsin/P; missed cleavage, 2; fixed modification, Carbamidomethyl (C); variable modifications, propionylation (K), oxidation (M) and acetyl (protein N-terminal); peptide mass tolerance, 10 ppm; fragment mass tolerance, 0.5 Da. +2, +3 and +4 were used as selected charge states. Areas under the curves (AUC) of the precursor ion’s peak were used for evaluating the peptide intensity. Unmodified peptide (VHDLVGVGVEEGAELLVDGR) was used for protein level normalization. Construction of an S. erythraea sace_3798-overexpressing Mutant The reconstructed plasmid PIB139-SeSrtN carrying 780 bp fragment of the sace_3798 open reading frame inserted between NdeI and EcoRV restriction sites was transformed to the S. erythraea E3 or S. erythraea WT cell. After that, the constructed plasmid was integrated into the S. erythraea genome using the phiC31 integrase. Apramycin resistance was used for selection. The selected mutants were verified by PCR and real-time quantitative PCR (RT-qPCR). Cloning and Mutagenesis of the Enzymes The list of primers used for PCR and mutagenesis were provided in Table S7. The sace_1456 (mmsA2), and sace_6566 (aldehyde dehydrogenase) were amplified by using primers 1456F, 1456R, 6566F and 6566R from the genomic DNA of S. erythraea. After restriction digest, the genes coding for sace_1456 and sace_6566 were cloned into pET28a plasmid to generate pET28a-1456/ pET28a-6566 constructed plasmids in E. coli DH5a strain. The positive clone was then confirmed by sequencing. Point mutations in the two reconstructed plasmids (sace_1456 and sace_6566) were generated by site-directed mutagenesis. The mutation genes were obtained through overlap extension PCR and transformed to pET28a for sequencing. Then the mutant plasmids (pET28a1456 (K12Q), pET28a-1456 (K94Q) and pET28a-6566 (K6Q)) were generated. Expression and Purification of the Enzymes and Their Mutants The proteins were expressed using the E. coli BL21 (DE3) strain. A single colony of each strain (E. coli BL21 (DE3) with pET28a-1456/ pET28a-1456 (K12Q)/ pET28a-1456 (K94Q)/ pET28a-6566/ pET28a-6566 (K6Q)/ pET28a-SeSrtN) was selected to start a 5 ml overnight culture, which was then transferred to a 100 ml broth (LB) medium supplemented with 1& kanamycin, respectively. The samples were cultured at 37 C and then induced with 0.7 mM isopropyl-b-D-thiogalactoside at 20 C overnight. Cells were harvested by centrifugation and resuspended in PBS buffer. The cells were then disrupted by sonication, and cell debris was removed by centrifugation at 12000 g for 15 min. The resulting supernatant of each protein was loaded onto a 2 ml Ni-nitrilotriacetic acid (NTA)-agarose column that was pre-equilibrated with the binding buffer (50 mM NaH2PO4, 300 mM NaCl, and 20 mM imidazole, pH 8.0). After that, the column was washed with 20 ml of binding buffer, and bound proteins were eluted using the elution buffer (250 mM imidazole in 50 mM NaH2PO4 and 300 mM NaCl, pH 8.0). The purified proteins were pooled and dialyzed against PBS buffer. The protein concentration was determined by BCA protein assay kit using buffer R as the control, and the same amounts of each protein after concentration determination were analyzed by SDS-PAGE. In Vitro Methylmalonate-semialdehyde Dehydrogenase Assays The activity of methylmalonate-semialdehyde dehydrogenases (recombinant protein (PET28a-1456), recombinant protein (pET28a1456 (K12Q)) and recombinant protein (pET28a-1456 (K94Q))) were assessed by using a well-established method (Hatter and e4 Cell Chemical Biology 25, 1–12.e1–e6, August 16, 2018

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Sokatch, 1988). Briefly speaking, the reaction buffer contained 10 ml Tris buffer (1 M, pH 9.2), 10 ml 1 M 2-mercaptoethanol, 1 ml 0.1 M NAD+ and same concentrations of enzyme to give an initial absorbance change at 340 nm of 0.03 to 0.2 against a PBS buffer blank. The volume was made to 95 ml and the reaction was started with 5 ml of 0.5 M propionaldehyde. The dynamic absorbance was determined at 340 nm by using a microplate reader (BioTek Instruments). One unit was 1 mmol NADH formed per milligram protein and specific activity was units per milligram protein. Enzymatic activities of mmsA2, mmsA2K12Q, mmsA2K94Q with or without propionyl-CoA incubation were conducted. Triplicate for each protein and mutant were performed and relative enzyme activity was then determined. Three replicates were conducted. The data was presented by using the box plot (Spitzer et al., 2014). In Vitro Aldehyde Dehydrogenase Assays Enzymatic activity of aldehyde dehydrogenases (recombinant protein (pET28a-6566) and recombinant protein (pET28a-6566 (K6Q))) were determined spectrophotometrically at 340 nm by conversion of NAD+ to NADH at room temperature in 0.1 ml total volume (Ho and Weiner, 2005). The assay buffer included 50 mM sodium pyrophosphate (pH 7.5), 0.5 mM NAD+ and 1 mM propionaldehyde. The enzyme reaction was initiated by adding the purified protein to the mixture. Absorbance at 340 nm was recorded by microplate reader. Triplicate for the protein and its mutant were conducted, and relative enzyme activity was calculated. The data was presented by using the box plot. HPLC Assay of the Propionyl-CoA in S. erythraea WT and E3 Strain Intracellular level of propionyl-CoA was analyzed by using reserve-phase high-performance liquid chromatography (HPLC) (Deutsch et al., 2002). 10 ml ice-cold 6% perchloric acid was added to same amounts of cultured cells and vortexed for 10 s. After that, the reaction buffer was sonicated for 1 min and then followed by ice-bath cooling for 4 min. After that, the mixture was centrifuged at 21000 g for 10 min and was adjusted to pH 3 by dropwise addition of 3 M K2CO3. The supematant was filtered through a SepPak C18 cartridges (Waters, Milford, MA), eluted by 40% acetonitrile and dried by Speed Vac. The samples were analyzed by using a HPLC system (Agilent 1100, Agilent TechnologiesInc., Folsom, CA ) equipped with an Xbridge C18 column (4.63250mm, Waters Corp., Milford, MA). The two mobile-phase solvents used were buffer Ahplc (0.2 M sodium phosphate, pH 5.5) and buffer Bhplc (800 ml of 0.25 M sodium phosphate pH 5.0 mixed with 200 ml of acetonitrile). Analysis was performed using a linear gradient system of 48% buffer Bhplc for 50 min. Initial conditions were 90% buffer Ahplc and 10% buffer Bhplc; a constant flow rate of 0.7 mL/min was used during the separation. At 10 min, buffer Bhplc was increased to 15%, then increased to 23% over 10 min. At 40 min, percent buffer Bhplc was increased to 48% and held at that composition for 10 min. The column was equilibrated at 10% buffer Bhplc for 20 min between injections. The UV detector was set at 254 nm. Peak area analysis was performed on the propionyl-CoA for each of the samples. Two replicates were conducted. Measurement the Intracellular Concentrations of Acyl-CoAs The cellular metabolites were extracted as mentioned above. In the desalting process, samples were eluted by 60% ACN and dried in Speed Vac. Orbitrap Fusion mass spectrometer was used for metabolite qualitative and quantitative analysis. Samples were dissolved in solvent A and analyzed by a 30 min nano-HPLC gradient with 7%-27% solvent B for 24 min, 27%-80% solvent B for 3 min, and 80% solvent B for 3 min at a flow rate of 300 nl/min. The mass resolution was set as 240000 at m/z 200. The maximum injection time was set as 50 ms for Orbitrap and 35 ms for ion trap. The electrospray voltage was set as 2 kV. The standard of acetyl-CoA, propionyl-CoA and succinyl-CoA was used for localization in the spectrogram. Two replicates were conducted. Areas under the curves (AUC) of the precursor ion’s peak were used for evaluating the intensity. Measurement of Intracellular Concentrations of S-adenosylmethionine The method reported was used for the measurement of cellular SAM through HPLC analysis (Fuso et al., 2001). In brief, same amounts of cell cultures were washed twice with PBS buffer and were scraped into 1 ml of water. After that, the samples were precipitated by using 1.5 M perchloric acid at 4 C for 1 h. After the complete lysis, the mixture was neutralized with KOH to pH 4. The extraction buffer was centrifuged for 15 min at 21000 g and later subjected for HPLC analysis. The experiments were conducted in duplicate, and SAM concentration was evaluated by the peak area. Two replicates were conducted. Erythromycin Determination by HPLC The fermentation samples of each strain were frozen and erythromycin was extracted by acetonitrile and analyzed by high-performance liquid chromatography at 210 nm absorbance (Wu et al., 2014). The mobile phase was a mixture of acetonitrile-50 mM K2HPO4 pH 8 (50:50) at rate (1 ml/min), the model of HPLC system and chromatographic column were mentioned above. The experiments were performed in duplicate and the standard curve was constructed for the direct determination of the erythromycin production for each sample. Two replicates were conducted. Determination of Undecylprodiginine in S. coelicolor We measured the undecylprodiginine in S. coelicolor according to the previous studies (Ryu et al., 2006). Briefly speaking, the same amounts of cell pellets cultured in different concentration of malonate were mixed with 800 ml methanol (pH 1.0) and were vortexed thoroughly for 2 min. Cell debris was then removed by centrifugation at 21000 g for 30 min. Undecylprodiginine concentration was evaluated at absorbance 530 nm recorded by microplate reader. Cell Chemical Biology 25, 1–12.e1–e6, August 16, 2018 e5

Please cite this article in press as: Xu et al., Protein Acylation is a General Regulatory Mechanism in Biosynthetic Pathway of Acyl-CoA-Derived Natural Products, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.05.005

Determination of Alcohol in C. acetobutylicum by Using the Colorimetric Method We determine the production of alcohol by using the potassium dichromate colorimetric method according to the previous study (Bahadur and Saroj, 2013). Briefly speaking, 1 ml 2% Cr2K2O7 solution and 0.5 ml 98% H2SO4 were added to the glass curette with 0.5 ml diluted culture medium. After mixed, the reaction buffer was in the darkness for 10 min. Alcohol concentration was determined at absorbance 610 nm recorded by microplate reader and two replicates were conducted. The Depropionylation Activity of SeSrtN The recombinant SeSrtN (pET28a-sace_3798) was purified according to our previous studies (You et al., 2014). The depropionylation activity of SeSrtN was conducted by HPLC-MS/MS analysis in the reaction buffer, including 4 mg SeSrtN, 0.3 mM synthesized peptides (TYK (pr) LYVGGK or HGGGAFSGK (pr) DPSK), 20 mM Tris-HCL (PH 7.4), 1 mM NAD+, 1 mM DTT, at 37 C for 2 h. After desalted by C18 ZipTips, the samples were analyzed and the peptides (propionylated and its unpropionylated form) were normalized to the same height. QUANTIFICATION AND STATISTICAL ANALYSIS FDR was controlled at protein, peptide and modification level for the proteomics data in the MaxQuant software. GraphPad Prism software or online BoxPlotR tool was used to calculate the statistics. Replicates are indicated in the figure legends. Error bars represent the standard deviation of the mean in all presented data. DATA AND SOFTWARE AVAILABILITY The proteomics data reported in this paper have been deposited to iProx database (URL: http://www.iprox.org/page/PDV014.html? projectId=IPX0000970000) and are available under accession number IPX0000970001.

e6 Cell Chemical Biology 25, 1–12.e1–e6, August 16, 2018