Abstract
Modular polyketide synthases (mPKSs) are multidomain enzymes in bacteria that synthesize a variety of pharmaceutically important compounds. mPKS genes are usually longer than 10 kb and organized in operons. To understand the transcriptional and translational characteristics of these large genes, here we split the 13-kb busA gene, encoding a 456-kDa three-module PKS for butenyl-spinosyn biosynthesis, into three smaller separately translated genes encoding one PKS module in an operon. Expression of the native and split busA genes in Streptomyces albus reveals that the majority ( >93%) of PKS mRNAs are truncated, resulting in a greater abundance of and a higher synthesis rate for the proteins encoded by genes closer to the operon promoter. Splitting the large busA gene rescues translation of truncated mRNAs into functional PKS subunits, and increases the biosynthetic efficiency of butenyl-spinosyn PKS by 13-fold. The truncated mRNA translation rescue strategy will facilitate engineering of multi-domain proteins to enhance their functions.
Similar content being viewed by others
Introduction
Modular polyketide synthases (mPKSs), the largest proteins in bacteria, are multidomain enzymes that are responsible for the biosynthesis of a variety of pharmaceutically important compounds, including antibiotics, immunosuppressants, insecticides, and anticancer drugs1,2,3,4,5. mPKS proteins are ordered by docking domains to form a megasynthase complex that achieves assembly line-like biosynthesis6. Each mPKS is composed of one or several modules, each of which contains a set of catalytic domains2,7 (Figs. 1a, 2a). Each module within an mPKS catalyses one biosynthesis step by adding one acyl-CoA building block to the polyketide chain3. Individual modules are usually larger than 150 kDa, and the molecular weight of an mPKS is determined by the number of modules and catalytic domains. Thousands of mPKSs have been identified to date8. Many efforts have been made to improve polyketide biosynthetic efficiency by optimising gene transcription9,10,11 or posttranslational modifications12,13, engineering regulatory factors14,15,16, and increasing the supply of precursors17,18. However, transcription and translation of mPKS genes are poorly understood; therefore, investigation of these characteristics will further promote the engineering of polyketide biosynthesis.
a The Modular polyketide synthases (mPKSs) for butenyl-spinosyn and salinomycin. The modules in each mPKS protein are joined with linkers. The modules between mPKS proteins are organised by docking domains to form a megasynthase complex. The operons in the butenyl-spinosyn, and salinomycin PKS gene cluster are indicated with arrows. CDDSlnA1, the C-terminal docking domain of the salinomycin PKS SlnA1. NDDSlnA2, the N-terminal docking domain of the salinomycin PKS SlnA2. CDDSlnA7, the C-terminal docking domain of the salinomycin PKS SlnA7. NDDSlnA8, the N-terminal docking domain of the salinomycin PKS SlnA8. TGA, stop codon. ATG, start codon. RBS, ribosomal binding site. b The three strategies used to split BusA. The linkers between the two modules were removed and replaced with exogenous docking domains from two adjacent salinomycin mPKSs. c Butenyl-spinosyn production by S. albus J1074 strains harbouring gene clusters containing wild-type busA (2.36 mg L−1) or split busA indicated in (b). n = 3 independent fermentation samples. Data are presented as the mean ± S.D. The p-values of the two-tailed t test are indicated. Source data are provided as a Source Data file.
a The mPKSs for avermectin and the mPKS-NRPS hybrid for epothilone. The operon in the epothilone mPKS-NRPS gene cluster is indicated with an arrow. b The two strategies used to split AveA2. c Avermectin B1a production by S. coelicolor CH999 strains harbouring gene clusters containing wild-type aveA2 or split aveA2 indicated in (b). d The strategy used to split EpoD. CDstiB, the C-terminal docking domain of the tigmatellin PKS StiB. NDDstiC, the N-terminal docking domain of the stigmatellin PKS StiC. e Epothilone A production by S. albus J1074 strains harbouring gene clusters containing wild-type epoD or split epoD indicated in (d). n = 3 independent fermentation samples. Data are presented as the mean ± S.D. The p-values of the two-tailed t test are indicated. Source data are provided as a Source Data file.
Truncated messenger RNAs (mRNAs) are produced by premature termination of transcription or degradation by RNases19,20. Because transcription and translation are coupled in bacteria, both intact and truncated mRNAs are translated into polypeptides indiscriminately21. mRNAs with intact open reading frames (ORFs) are translated into functional PKS proteins. Ribosomes translating on mRNAs with truncated ORFs that do not contain stop codons are stalled because they cannot properly terminate translation. Stalled ribosomes are then rescued by ribosome rescue systems22. Finally, mRNAs with truncated ORFs are translated into polypeptides that do not contain C-terminal catalytic domains. For mRNAs encoding full-length proteins critical for cellular processes, the translation of truncated mRNAs results in the generation of defective polypeptides that can be harmful to cells. Furthermore, the translation of truncated mRNAs also decreases the number of active ribosomes. Genes encoding mPKSs are usually larger than 10 kb and organised in operons; therefore, their mRNAs are longer than those of ordinary genes. A giant mPKS gene (51 kb) encoding an 1800 kDa protein consisting of nine modules has been found in bacteria23. The longer an mRNA is, the greater its risk of being truncated. Translation of truncated mPKS mRNAs will generate nonfunctional PKS fragments that cannot be used for polyketide biosynthesis. The effects of truncated mRNAs on mPKS translation and polyketide biosynthesis have not yet been evaluated.
In this work, to understand the transcriptional and translational characteristics of mPKS genes, we examine the expression of butenyl-spinosyn PKS genes in Streptomyces albus J107424, which is one of the most widely used hosts for the expression of PKS genes to produce polyketide natural products. Butenyl-spinosyn is a highly potent natural insecticidal agent25; however, it has not been commercialised due to low fermentation yield. We found that truncated mRNAs constitute the majority of mPKS mRNAs. The presence of truncated mRNAs results in a greater abundance and production rate of the proteins encoded by genes closer to the operon promoter. Based on these findings, we developed a strategy to rescue mPKS mRNA truncation and translation in which the multiple-module PKS protein was divided into separately translated single-module PKS subunits. The sequential organisation of the modules is maintained by adding heterologous PKS docking domains at the termini of the split PKS subunits. Applying this approach to the mPKS gene enables the translation of truncated mRNAs into functional PKS subunits and increases the concentration of PKS subunits encoded by sequences closer to the 5’ end of the mRNAs; therefore, the biosynthesis efficiency is much greater for the bacterial host harbouring split PKS genes than for that harbouring native PKS genes. These findings enhance our understanding of the transcription and translation of large genes in bacteria and will facilitate the engineering of large proteins composed of multiple modules or domains to enhance their functions.
Results
Splitting the BusA mPKS into smaller separately translated subunits improves biosynthetic efficiency of butenyl-spinosyn
To study the effects of truncated mPKS mRNAs on polyketide biosynthesis, we chose the mPKS genes for biosynthesis of butenyl-spinosyn, which is a highly potent natural insecticidal compound and the fermentation yield which needs to be improved for industrial production. The five PKS genes for butenyl-spinosyn biosynthesis, busA-E, range from 6.5 kb to 16.8 kb and each PKS comprises one to three modules25. These genes are grouped into a gene cluster composed of two large operons, consisting of busA-C and busD-E (Fig. 1a). Enhancing the concentration of the initial enzyme within a biosynthetic pathway has the potential to yield greater benefits than augmenting the downstream enzymes26,27. Furthermore, the 13.0-kb busA gene encoding the initial PKS within the buteny-spinosyn biosynthetic is closest to the operon promoter (Fig. 1a). Therefore, we selected the busA gene to investigate. Each module of the 456-kDa three-module BusA protein comprises three to five domains; therefore, we split BusA according to its module composition in three different ways: (1) two subunits, with the double-module BusA-1-1 subunit composed of the first two modules and the single-module BusA-1-2 subunit composed of the last extension module; (2) two subunits, with the single-module BusA-2-1 subunit composed of the loading module and the double-module BusA-2-2 subunit composed of the two extension modules; and (3) three subunits, with the single-module BusA-3-1 composed of the loading module, the single-module BusA-3-2 composed of the first extension module, and the single-module BusA-3-2 composed of the second extension module (Fig. 1b).
These splitting strategies necessitated removing the linkers between the two BusA modules. Therefore, to maintain module organisation in this PKS megasynthase, the C-terminal docking domain (CDD) and N-terminal docking domain (NDD) from two adjacent salinomycin PKSs (Fig. 1a) were inserted downstream of the acyl carrier protein (ACP) domain of the upstream BusA module and upstream of the ketosynthase (KS) domain of the downstream BusA module, respectively (Fig. 1b and Supplementary Fig. 1). We selected docking domains from the two adjacent salinomycin PKSs because all of the salinomycin PKS genes (slnA1-slnA9) are grouped into one operon and there is no terminator in these genes28. Therefore, the insertion of docking-domain coding sequences into the busA gene will not alter its transcription. Both the native busA gene and the split busA genes are kept in the busA-busC operon and under the control of the SA15p promoter. Furthermore, to avoid the potential confounding effect of different translation efficiencies caused by different ribosomal binding sites (RBSs) on the comparison of the biosynthesis efficiencies of recombinant BusA proteins, the RBSs between the coding sequences of the docking domains were all replaced with the RBS of the SA15p promoter, which controls transcription of the busA-busC operon. In splitting strategies 1 and 2, the CDD of the salinomycin PKS SlnA1 and the NDD of the salinomycin PKS SlnA2 were used. In splitting strategy 3, the CDD of SlnA1 and the NDD of SlnA2 were inserted between the loading module and its adjacent extension module, and the CDD of SlnA7 and the NDD of SlnA8 were inserted between the two extension modules.
Previously, we achieved heterologous production of butenyl-spinosyn (2.36 mg L−1) by transforming the intact 86-kb butenyl-spinosyn gene cluster into the heterologous host S. albus J107429. Here, the butenyl-spinosyn gene clusters containing wild-type BusA or split BusA were transformed into S. albus J1074 and butenyl-spinosyn production was analysed by High-performance liquid chromatography-mass spectrometry (HPLC-MS). The butenyl-spinosyn production of the busA-1, busA-2, and busA-3 strains was significantly greater (by 5.3-, 5.9-, and 13.3-fold, respectively) than that of the busA strain (Fig. 1c and Supplementary Fig. 2). The impact of introducing docking domain on communication and activity of modular PKSs were quantified by replacing the CDD of BusA with that of SlnA1 and the NDD of BusB with that of SlnA2. The results showed that the exchange did not affect the biosynthesis efficiency of butenyl-spinosyn (Supplementary Fig. 3). Therefore, docking domains do not enhance the communication and activity of modular PKSs. The above results suggested that splitting a large mPKS into smaller, separately translated subunits can improve polyketide biosynthetic efficiency.
Splitting the avermectin and epothilone mPKSs also improves their biosynthetic efficiency
To examine whether this splitting approach works for other megasynthases, we chose the mPKS for the biosynthesis of the avermectin30, which is an important natural insecticidal antibiotic and the mPKS-nonribosomal peptide synthetases (NRPS) hybrid for the biosynthesis of epothilone31 which is an important anticancer drug. The four PKS genes for avermectin biosynthesis, aveA1-4, range from 11.9 kb to 18.7 kb and each PKS comprises two to four modules32 (Fig. 2a). As the 666-kDa AveA2 is the largest of the avermectin polyketide synthases, its 18.7-kb coding sequence was split as follows: the coding sequence for the linker between the second and third modules was replaced with the coding sequence for CDDSlnA1-TGA-RBS-ATG-NDDSlnA2 (TGA, stop codon; ATG, start codon) or CDDSlnA7-TGA-RBS-ATG-NDDSlnA8 (Fig. 2b and Supplementary Fig. 4). Avermectin gene clusters containing the wild-type or split aveA2 gene were transformed into S. coelicolor CH999, and avermectin production by the transformants was analysed by HPLC-MS. Avermectin B1a production by the aveA2-1 and aveA2-2 strains was 5.3 and 4.8 times greater, respectively, than that by the wild-type aveA2 strain (Fig. 2c and Supplementary Fig. 5).
The megasynthase for epothilone biosynthesis is composed of 1 single-module NRPS and 5 mPKSs consisting of one to four modules31 (Fig. 2a). The 21.8-kb coding sequence of the largest epothilone mPKS, 764-kDa EpoD, was split by replacing the coding sequence for the linker between the second and third modules with the coding sequence for CDDStiB-TGA-RBS-ATG-NDDStiC (TGA, stop codon; ATG, start codon) (Fig. 2d and Supplementary Fig. 6). NDDStiB is the N-terminal docking domain of the stigmatellin PKS StiB33. CDDStiC is the C-terminal docking domain of the stigmatellin PKS StiC33. Epothilone gene clusters containing the wild-type or split epoD genes were transformed into S. albus J1074, and epothilone production by the transformants was analysed by HPLC-MS. Epothilone A production by the epoD-1 strains was 9.8-fold higher than that by the wild-type epoD strain (Fig. 2e and Supplementary Fig. 7).
Above results suggested that splitting a large mPKS into smaller, separately translated subunits can improve biosynthetic efficiency of both mPKS and mPKS-NRPS hybrid megasynthases.
Truncated mRNAs constitute the majority of busA PKS mRNAs
Given that the mPKS splitting approach was found to increase polyketide synthesis, we next aimed to determine the underlying mechanism. Therefore, we measured the transcription and translation of wild-type busA and split busA in S. albus J1074. Precise quantification and accurate comparison of the mRNA abundance of different sites in this cluster are complicated because the PKS modules in the bus gene cluster contain many identical domains with repetitive DNA sequences, making it difficult to identify unique DNA regions in each site. Furthermore, amplification efficiency can differ by DNA region. Therefore, we adopted a reporter strategy to quantify the mRNA and protein concentrations precisely using GusA (the β-glucuronidase)34, which can tolerate large fusions at the N- or C-terminus without loss of enzyme activity35; specifically, the gusA reporter gene and a GGSGGGGGG linker coding sequence were fused to each of five constructs for the butenyl-spinosyn gene cluster: 13.0-kb intact busA (encoding three modules), 8.1-kb busA-1-1 (encoding two modules), 4.9-kb busA-1-2 (encoding a single module), 2.9-kb busA-2-1 (encoding a single module), and 10.1-kb busA-2-2 (encoding two modules) (Fig. 3a, b). All of the fusion gene constructs are in the busA-busC operon and under the control of the SA15p promoter and the same RBS (Fig. 3a, b).
a The butenyl-spinosyn PKS gene cluster. The two operons are indicated with arrows. b Schematics of the wild-type and split busA genes fused with the gusA gene. Two loci on gusA indicated with asterisks are used for qRT‒PCR-based transcription analysis. c qRT‒PCR-based transcription analysis of gusA in J1074 strains harbouring butenyl-spinosyn gene clusters containing the wild-type or split busA genes from (b) using primers annealing to locus 1. d β-Glucuronidase activity assay of J1074 strains harbouring butenyl-spinosyn gene clusters containing the wild-type or split busA genes from (b). e Parallel reaction monitoring absolute quantification of BusA, BusA-1-1, and BusA-2-1 protein levels in Del14 strains expressing the busA, busA-1-1, or busA-2-1 genes. The positions of the unique peptide fragment, A489ALVADDEPK498, used for quantifying are indicated with #. n = 3 independent biological samples. Data are presented as the mean ± S.D. The p-values of the two-tailed t test are indicated. ns, not significant. Source data are provided as a Source Data file.
The S. albus J1074 strains harbouring the above gusA fusion genes were cultivated to the stationary phase (72 h) and collected for gene transcription analysis. Two different loci (locus 1 and locus 2, Fig. 3b) in gusA were selected for quantitative real-time PCR (qRT‒PCR) analysis. The results revealed that busA-gusA, busA-1-2-gusA, and busA-2-2-gusA were transcribed at the same level (Fig. 3c). This finding suggests that BusA splitting does not affect the transcription of the busA-busC operon and the same distance from the beginning of the operon to the tested position resulted in the same mRNA abundance. Overall, busA-2-1-gusA5.0kb transcription levels were the highest, and busA-gusA14.8kb transcription levels were the lowest (Fig. 3c). More specifically, at locus 1, the transcription levels of busA-2-1-gusA5.0kb and busA-1-1-gusA10.2kb were 14.8 times greater and 6.5 times greater, respectively, than those of busA-gusA (Fig. 3c). At locus 2, the transcription levels of busA-2-1-gusA5.0kb and bus-1-1-gusA10.2kb were 15.3 times greater and 7.5 times greater, respectively than those of busA-gusA (Supplementary Fig. 8). Because all of the fusion gene constructs are in the busA-busC operon and under the control of the SA15p promoter, the above transcription analysis revealed that a considerable amount of truncated PKS mRNA was produced in S. albus J1074. For the 14.8-kb busA-gusA gene, the mRNAs with intact ORFs accounted for only 6.8% (1/14.8) (Fig. 3c) of the total transcribed mRNAs. This observation is consistent with the idea that mRNA sequences closer to the start of an operon are more abundant than those farther from the start due to mRNA truncation.
mRNA truncation increases abundance of PKS proteins encoded by genes closer to the operon promoter
This has shown that truncated mRNAs constitute the majority of mPKS mRNAs, and mRNA sequences closer to the start of an operon are more abundant than those farther from the start due to mRNA truncation. We next measured the GusA (β-glucuronidase) activity of the S. albus J1074 strains expressing above five gusA fusion gene constructs in order to determine the abundance of GusA fusion proteins. The GusA assay revealed that the BusA-GusA, BusA-1-2-GusA, and BusA-2-2-GusA proteins exhibited the same level of activity (Fig. 3d). The abundance of each of these proteins was consistent with that of the corresponding mRNA. The finding that these three genes were transcribed at the same level indicates that the insertion of RBSs upstream of the busA1-2-gusA and busA2-2-gusA genes does not improve protein production. In contrast, the β-glucuronidase activity of BusA-2-1-GusA and BusA-1-1-GusA was 3.9 times greater and 2.5 times greater than that of BusA-GusA (Fig. 3d). The abundance of these three proteins was consistent with the abundance of the corresponding mRNA: higher mRNA levels resulted in greater protein production. These findings also indicate that PKS protein production in S. albus is mainly determined by mRNA abundance.
Because GusA was fused at the C-terminus of the above native and split BusA proteins, we next measured the abundance of the peptide fragment translated from the 5’-end of mRNAs transcribed from the busA-busC operons containing the native and split busA genes. We selected the BusA, BusA-1-1, and BusA-2-1 proteins with different lengths and sharing the identical N-terminus. Furthermore, the busA, busA-1-1, and busA-2-1 genes are all under the control of the same promoter and are the first gene in the operon (Fig. 3b). The N-terminal ends of BusA, BusA-1-1, and BusA-2-1 were quantified by the parallel reaction monitoring (PRM) targeted absolute protein quantification technique36,37. To obtain a unique peptide fragment for quantifying the N-terminal ends of BusA, BusA-1-1, and BusA-2-1, the busA, busA-1-1, and busA-2-1 genes were integrated separately into the chromosome of S. albus Del1438, in which all endogenous PKS gene clusters were deleted to avoid other PKSs interfering with the accuracy of the quantification. Furthermore, other butenyl-spinosyn PKSs will also interfere with the quantification. Therefore, the busA, busA-1-1, and busA-2-1 genes were separately cloned from the constructs illustrated in Fig. 1b and integrated into the chromosome of S. albus Del14. A unique peptide fragment, A489ALVADDEPK498, in BusA, BusA-1-1, and BusA-2-1 was identified by PRM. Then, the stable isotope-labelled AALVADDEPK peptide was synthesised as the internal standard reference for absolute quantification of this fragment in S. albus Del14 expressing BusA, BusA-1-1, or BusA-2-1. The results suggested that the three S. albus Del14 strains expressed the AALVADDEPK peptide at the same concentration (Fig. 3e). Therefore, the N-terminal ends of BusA, BusA-1-1, and BusA-2-1 proteins were expressed at the same level. Abundance differences between the N-terminus (Fig. 3e) and C-terminus (Fig. 3d) of BusA, BusA-1-1, and BusA-2-1 proteins indicates that a considerable amount of truncated PKS protein was produced in S. albus. Moreover, these truncated proteins do not contain C-terminal catalytic domains or docking domains and, therefore, cannot function in polyketide biosynthesis.
mPKS gene overexpression activates the ribosome rescue system in Streptomyces
Given that a considerable amount of truncated mRNA and PKS protein was produced in mPKS gene expression, we next aimed to determine the effects of mPKS gene overexpression on the ribosome rescue system in Streptomyces. Bacteria have evolved ribosome rescue systems mediated by the tmRNA-SmpB complex39,40,41, ArfA20, or ArfB42,43 for ribosome recycling. In the tmRNA-SmpB ribosome rescue system, defective polypeptides translated from truncated mRNAs are tagged with SsrA at their C-terminus and then degraded by cellular proteases39,40,41. Translation of truncated mRNA without a stop codon traps ribosomes on mRNA. Ribosome rescue systems are activated to recycle stalled ribosomes. The expression of the tmRNA–smpB44 ribosome-rescue system, which is the major ribosome-rescue system in bacteria22, was compared across S. albus J1074 strains containing the wild-type or split busA genes. Compared with that of the wild-type S. albus J1074 strain, the transcription of the ssrA gene (encoding tmRNA) and the smpB gene increased by 5.0- and 8.3-fold, respectively, in S. albus J1074 overexpressing the 13-kb wild-type busA gene encoding a three-module PKS (Fig. 4). In contrast, when the 13-kb busA gene was split into smaller translation units (i.e., into 8.1 kb busA-1-1 and 4.9 kb busA-1-2 or into 2.9 kb busA-3-1, 5.2 kb busA-3-2, and 4.9 kb busA-3-3), transcription of ssrA and smpB genes in S. albus J1074 was restored to the level in the wild-type S. albus J1074 strain (Fig. 4). The above results indicate that large mPKS gene overexpression in S. albus J1074 produces a vast number of truncated mRNAs without stop codons, and the ribosome rescue system is activated to recycle ribosomes. Conversely, splitting the large mPKS gene into smaller, separately translated subunits by inserting stop codons eliminates the need to activate the ribosome-rescue system, even though the truncated mRNAs are still produced.
a qRT‒PCR transcription analysis of the ssrA gene (encoding the tmRNA in the tmRNA–smpB ribosome-rescue system) in J1074 strains harbouring butenyl-spinosyn gene clusters containing the wild-type or split busA genes indicated in Fig. 1b. b qRT‒PCR analysis of the smpB gene of the tmRNA–smpB ribosome rescue system in J1074 strains harbouring butenyl-spinosyn gene clusters containing the wild-type or split busA genes. n = 3 independent biological samples. Data are presented as the mean ± S.D. The p-values of the two-tailed t test are indicated. ns, not significant. Source data are provided as a Source Data file.
mRNA truncation increases production rate of PKS proteins encoded by genes closer to the operon promoter
To measure the production rates of the wild-type and split BusA proteins, the busA-gusA, busA-1-1-gusA, and busA-2-1-gusA fusion genes were cloned downstream of a cumate (p-isopropylbenzoic acid)-inducible promoter45 (Fig. 5a), and the resulting constructs were integrated into the chromosome of S. albus J1074. The S. albus J1074 strain was cultivated to the exponential growth phase, and cumate was added to induce the synthesis of GusA fusion proteins. One millilitre samples of culture were collected at 10, 20, 30, 40, 60, 80, 100, 120, 140, or 160 min of induction. The collected samples were then treated with 3 mM chloramphenicol and 1.4 mM erythromycin to stop protein synthesis, after which β-glucuronidase activity was measured.
a Schematics of the wild-type and split busA genes fused with the gusA gene under the control of the cumate-inducible promoter. b Real-time GusA (β-glucuronidase reporter) activity assay of J1074 strains containing the DNA constructs from (a). c Production rate of BusA-1-1-GusA and BusA-2-1-GusA proteins in J1074 strains after induction for 40–80 min. d Production rate of BusA-GusA, BusA-1-1-GusA, and BusA-2-1-GusA proteins in J1074 strains after induction for 80–160 min. e β-Glucuronidase activity assay of J1074 strains containing the DNA constructs from (a) after induction for 20 h. f Schematics of the wild-type and split busA genes fused with the gusA gene under the control of the cumate-inducible promoter. g Real-time GusA (the β-glucuronidase) activity of J1074 strains containing the DNA constructs from (f). h Production rate of BusA-1-2-GusA and BusA-2-2-GusA proteins in J1074 after induction for 20–30 min. i Production rate of BusA-GusA, BusA-1-2-GusA, and BusA-2-2-GusA proteins in J1074 strains after induction for 60–160 min. j β-Glucuronidase activity of J1074 strains expressing the DNA constructs from (a) after induction for 20 h. n = 3 independent biological samples. Data are presented as the mean ± S.D. The p-values of the two-tailed t test are indicated. ns, not significant. Source data are provided as a Source Data file.
Within 160 min, the concentration of BusA-2-1-GusA was greater than that of BusA-1-1-GusA, which was greater than that of BusA-GusA. Moreover, the β-glucuronidase activity of BusA-2-1-GusA and BusA-1-1-GusA was detectable at 20 min of induction, whereas that of BusA-GusA was not detectable until 60 min of induction (Fig. 5b). The protein production rate of BusA-2-1-GusA was 2.3 times faster than that of BusA-1-1-GusA between 40 and 80 min of induction (Fig. 5c). Moreover, between 80 and 160 min of induction, the protein production rates of BusA-2-1-GusA and BusA-1-1-GusA were 6.2 and 2.2 times faster, respectively, than that of BusA-GusA (Fig. 5d). After induction and incubation for 20 h, the β-glucuronidase activity assay revealed that the concentrations of BusA-2-1-GusA and BusA-1-1-GusA were 14.3 and 3.2 times greater, respectively, than that of BusA-GusA (Fig. 5e). These differences in protein production rate and concentration among BusA-GusA, BusA-1-1-GusA and BusA-2-1-GusA were consistent with the differences in their mRNA abundance; that is, the shorter proteins were expressed at faster rates.
The busA-1-2-gusA and busA-2-2-gusA fusion genes were also cloned downstream of the cumate-inducible promoter (Fig. 5f). The expression of the BusA-1-2-gusA and BusA-2-2-gusA fusion proteins was measured over a 160 min induction time course. Within 160 min, the concentration of BusA-2-2-gusA was greater than that of BusA-1-2-gusA, which was greater than that of BusA-gusA. The β-glucuronidase activities of BusA-1-2-gusA and BusA-2-2-gusA were detected beginning at 20 min of induction; however, the β-glucuronidase activity of BusA-gusA was not observed until 60 min of induction (Fig. 5g). This result indicates that insertion of the RBS and start codon upstream of gusA promotes GusA translation. The average density of ribosomes per kilobase is nearly invariant across most mRNAs and across growth conditions in bacteria46. Consequently, the protein production rate should be the same for proteins with the same mRNA abundance. Internal promoters have been found to be present in mPKS operons and regulate the transcription of operon genes47,48. The internal promoters in busA-1-1 and bus-2-1 drove the transcription of mRNAs containing RBS-ATG-busA-1-2-gusA and RBS-ATG-busA-2-2-guA. These two mRNAs, which are shorter than busA-gusA, were then translated into BusA-1-2-gusA and BusA-2-2-gusA earlier than busA-gusA was translated into BusA-gusA. At the early stage of the time course (20–30 min), the concentration of BusA-2-2-gusA was greater than that of BusA-1-2-gusA, and the protein production rate of BusA-2-2-gusA was 3.5 times greater than that of BusA-1-2-gusA (Fig. 5h); therefore, the internal promoter upstream of busA-2-2 is stronger than that upstream of busA-1-2. The protein production rates of BusA-2-2-gusA, BusA-1-2-gusA, and BusA-gusA were almost the same after induction for 60–60 min (Fig. 5i). The concentrations of BusA-2-2-gusA, BusA-1-2-gusA, and BusA-gusA were also almost the same after induction for 20 h (Fig. 5j), consistent with their similar mRNA abundances. The protein concentration and production rate are mainly determined by the mRNA concentration; therefore, the internal promoters upstream of busA-2-2-gusA and busA-1-2-gusA are weaker than the SA15p promoter, which is upstream of the large busA-busC operon.
Reinforcing the transcription of downstream mPKS genes further increases butenyl-spinosyn production
The above results suggest that splitting the busA gene into smaller separate translation units enables their translation into functional PKS subunits and enhances biosynthetic efficiency. Therefore, we hypothesised that adopting a similar approach for the C-terminal PKS subunits and downstream PKS proteins (BusB-E) could further improve the biosynthetic efficiency of the PKS megasynthase assembly line. To test this, the strong constitutive KasOp* promoter49 from Streptomyces was inserted upstream of busA-1-2 in the busA-1 gene cluster, busA-2-2 in the busA-2 gene cluster, and busA-3-2, busB, busC, busD, and busE in the busA-3 gene cluster (Fig. 6a). These recombinant gene clusters were subsequently transformed into S. albus J1074 and integrated into the chromosome through phiC31 site-specific recombination; then, butenyl-spinosyn production was analysed in the transformed strains containing the above recombinant gene clusters.
a Schematic of the kasOp* promoter insertion. b–d Butenyl-spinosyn production by S. albus J1074 strains harbouring gene clusters from (a). n = 3 independent fermentation samples. e qRT‒PCR analysis of the transcription of bus genes in J1074 strains harbouring the busA-3 butenyl-spinosyn gene cluster with the kasOp* promoter inserted upstream of busB, busC, busD, or busE. n = 3 independent biological samples. Data are presented as the mean ± S.D. The p-values of the two-tailed t test are indicated. Source data are provided as a Source Data file.
S. albus expressing the kasOp-busA-1-2 gene cluster exhibited a 2.6-fold increase in butenyl-spinosyn production compared with S. albus containing the busA-1 gene cluster (Fig. 6b). Similarly, the kasOp-busA-2-2 gene cluster resulted in a 1.9-fold increase in butenyl-spinosyn production compared with the busA-2 gene cluster (Fig. 6c), and the kasOp-busA-3-2, kasOp-busB, kasOp-busC, kasOp-busD, and kasOp-busE gene clusters resulted in 1.4-, 2.0-, 2.2-, 1.8-, and 1.6-fold enhanced butenyl-spinosyn production, respectively, compared with the busA-3 gene cluster (Fig. 6d). Ultimately, butenyl-spinosyn production by S. albus containing the kasOp-busC gene cluster was 30 times greater than that by S. albus containing the native unsplit busA gene (Fig. 6d).
The transcript levels of busB, busC, busD, and busE from the kasOp-busB, kasOp-busC, kasOp-busD, and kasOp-busE gene clusters were compared with those from the busA-3 gene cluster in S. albus J1074. The results revealed that insertion of the kasOp* promoter upstream of busB, busC, busD, and busE increased gene transcription by 62.7-, 188.8-, 130.3-, and 145.8-fold, respectively (Fig. 6e). Therefore, transcriptional reinforcement of downstream PKS genes further increased butenyl-spinosyn production.
Discussion
mPKSs synthesise numerous compounds that are important sources of pharmaceuticals. mPKS genes are usually larger than 10 kb. Understanding the transcription and translation characteristics of these large genes is important for the engineering of biosynthetic pathways. In this study, we compared the transcription and translation of native and split 13-kb busA genes and investigated the consequences of different constructs for the biosynthesis of butenyl-spinosyn in Streptomyces. Based on the results obtained in this study, we propose a model for the transcription and translation of the native and split PKS genes, as illustrated in Fig. 7. The transcription of large PKS operons in Streptomyces produces both intact and truncated mRNAs. Truncated mRNAs constitute the majority of transcribed mRNAs (Fig. 7a). Both intact and truncated mRNAs are translated into polypeptides. mRNAs with intact ORFs are translated into functional PKS proteins. Ribosomes translated on mRNAs with truncated ORFs that do not contain stop codons are stalled because they cannot properly terminate translation. Stalled ribosomes are then rescued by ribosome rescue systems (Fig. 7a). Finally, mRNAs with truncated ORFs are translated into polypeptides that do not contain C-terminal catalytic domains or docking domains; therefore, these defective truncated PKS fragments are nonfunctional and cannot be used for biosynthesis. When tmRNA–SmpB, the major ribosome rescue system in bacterial cells22, is used to rescue stalled ribosomes, the SsrA tag is added to the C-termini of truncated PKS fragments. These SsrA-tagged polypeptides are then degraded by cellular proteases (Fig. 7a). Therefore, the transcription of truncated mRNAs and translation of truncated nonfunctional polypeptides waste cellular resources.
a Transcription and translation of native PKS genes. b Transcription and translation of the split PKS genes.
Splitting large PKS genes by inserting the CDD-TGA-RBS-ATG-NDD coding sequence between module coding sequences rescues the translation of truncated mRNAs and allows the generation of functional PKS subunits (Fig. 7b). Because truncated mRNAs constitute the majority of transcribed mRNAs, many functional PKS subunits encoded by sequences closer to the 5’ end of mRNAs were generated with this approach. Splitting PKS genes does not affect operon transcription; therefore, the concentration of the C-terminal subunit produced by the split PKS gene is the same as that produced by the native unsplit full-length PKS gene (Fig. 7b). PKS subunits produced by split genes form functional PKS complexes through protein-protein interactions mediated by terminal docking domains. Splitting of PKS genes caused a significant increase in the concentration of N-terminal PKS subunits; therefore, the biosynthetic efficiency of the PKS megasynthase assembly line was improved significantly. Currently, we do not know if BusA, AveA2, or EpoD investigated in this study are the rate-limiting enzymes in their biosynthetic pathways. But, if such a rate-limiting PKS can be identified, improving its activity might yield the most benefits.
In addition to the protein concentration and production rate measured in this study, the smaller PKS subunits may be more soluble and stable. Large multidomain or multimodule proteins have more complex folding processes than small proteins and are, therefore, more prone to misfolding50. In protein structural studies, large multidomain proteins are often truncated into smaller subunits for better expression, solubility and stability; these factors may also contribute to the enhanced polyketide biosynthesis efficiency observed here.
The average density of translating ribosomes is nearly constant on mRNAs46. The number of stalled ribosomes on the unsplit busA truncated mRNAs are much larger than those on the split busA truncated mRNAs (Fig. 7). Each stalled ribosome needs one tmRNA-smpB complex to rescue; therefore, the strain expressing the unsplit busA needs more tmRNA-smpB complexes compared with the strain expressing the split busA gene. This is consistent with the transcription analysis results of the tmRNA–smpB ribosome-rescue system in S. albus J1074 expressing the unsplit or split busA genes (Fig. 4). Both long unsplit BusA truncated proteins and short split BusA truncated proteins should be programmed for rapid degradation. Amino acids regained from the degradation of long unsplit BusA truncated proteins are much more than those of short split BusA truncated proteins. However, amino acids consumed for the synthesis of long unsplit BusA truncated proteins are much more than those of short split BusA truncated proteins. Overall, we estimated that the cost of unsplit busA gene expression is higher.
In bacteria, the transcript abundance of genes is regulated at both transcriptional and post-transcriptional levels. The type of exponentially distributed truncated busA mRNA that we observe may arise from low transcriptional processivity of the endogenous RNA polymerases in S. albus J1074. Transcribing target mPKS genes using exogenous orthogonal RNA polymerases with higher processivity than the endogenous bacterial RNA polymerase could reduce the abundance of truncated mRNA. T7 RNA polymerase (T7RNAP) exhibits a very rapid elongation rate and a tendency not to pause until it encounters terminator sequences51. The elongation rate of T7RNAP is nearly 5-fold higher than that of the E. coli RNA polymerase52. T7RNAP has been used in vitro to produce RNA longer than 27 kb, and the majority of transcripts are of full length53. Furthermore, the gene expression level of the T7RNAP system in Corynebacterium glutamicum was 4-fold higher than that of the endogenous RNA polymerase54. The T7RNAP-based expression system has been established in Streptomyces55,56, and it is worthwhile to try employing this system to regulate the transcription of mPKS genes for higher expression levels.
Endonuclease cleavage within the intergenic region of a polycistronic mRNA can dissect a transcript into two separate ones with different stability57. Truncated busA mRNAs may also arise from 3′-5′ exonuclease degradation of the unstable busA mRNA fragment when there is an endonuclease cleavage site downstream of the busA mRNA. If the intergenic endonucleolytic cleavage sites can be identified, they can be swapped into strong transcriptional terminators to increase the half-life of the busA mRNA. The transcription unit architecture of S. lividans TK24 has been systematically determined through high-throughput sequencing techniques at a nucleotide resolution to identify genetic regulatory elements including intergenic endonucleolytic cleavage sites58. The regulatory elements for transcription termination and post-transcriptional processing have also been elucidated in the S. clavuligerus genome59. However, intergenic endonucleolytic cleavage sites for mRNAs of large mPKS operons have not been determined in their work. Systematically elucidation of intergenic endonucleolytic cleavage sites of long mRNAs encoded by large mPKS operons will broaden our understanding of their regulatory mechanisms and facilitate engineering of them for better expression.
In summary, we split large BusA, AveA2, or EpoD mPKSs into smaller proteins and improved the biosynthetic efficiency of corresponding natural products. The truncated PKS mRNA translation rescue strategy developed in this study enriches the toolbox for PKS pathway engineering. This approach can be combined with known engineering strategies to improve the biosynthesis efficiency of polyketide or polyketide-nonribosomal peptide hybrid natural products. In addition, this approach could be applied to other large proteins composed of multiple modules or domains to enhance their functions.
Methods
Strains and reagents
The RedEx seamless mutagenesis experiments were performed using Escherichia coli strains GBred-gyrA46229,60 and GB200529. GBred-gyrA462 carries the GyrAArg462Cys mutation to confer ccdB resistance and a recombineering operon composed of the redγ, redβ, redα, and recA genes under the PBAD promoter in the chromosome. GB2005 is a derivative of the DH10B strain which carries the wild-type GyrA and is sensitive to the ccdB toxin. Recombineering was performed using Escherichia coli strain GB05-red61 which carries the recombineering operon composed of the redγ, redβ, redα, and recA genes under the PBAD promoter in the chromosome. Expression of butenyl-spinosyn and epothilone gene clusters was performed using the S. albus J1074 strain. Expression of avermectin gene clusters was performed using S. coelicolor CH999 strain.
E. coli strains were grown in LB medium at 37 °C. Antibiotic selection for E. coli strains was performed using 15 µg mL−1 of chloramphenicol (A100230, Sangon Biotech), 100 µg mL−1 of ampicillin (A010, MDBio), or 20 µg mL−1 of apramycin (A600090, Sangon Biotech). Induction of expression of the redγ-redβ-redα-recA in E. coli GBred-gyrA462 and GB05-red strains was performed using 2.5 mg mL−1 L-arabinose (E003256, Sigma-Aldrich). Streptomyces strains were grown on mannitol soya flour agar plates or brain heart infusion agar plates. Antibiotic selection for Streptomyces strains was performed using 50 µg mL−1 of apramycin (A600090, Sangon Biotech) and 25 µg mL−1 of nalidixic acid (N4382, Sigma-Aldrich).
Seamless mutagenesis of PKS genes by the RedEx method
The HAL-CDD-TGA-RBS-ATG-NDD-HAR20-PacI-ampccdB-PacI-HAR cassette was prepared by overlap extension PCR. The overlapping HAL-CDD-TGA-RBS fragment and ATG-NDD-HAR20 fragment were amplified via PCR and then joined together via overlap extension PCR to generate the HAL-CDD-TGA-RBS-ATG-NDD-HAR20 fragment. The PacI-ampccdB-PacI-HAR fragment carrying an overlapping sequence with the HAL-CDD-TGA-RBS-ATG-NDD-HAR20 fragment was amplified via PCR. Finally, the HAL-CDD-TGA-RBS-ATG-NDD-HAR20 fragment and the PacI-ampccdB-PacI-HAR fragment were joined together by overlap extension PCR to generate the HAL-CDD-TGA-RBS-ATG-NDD-HAR20-PacI-ampccdB-PacI-HAR cassette. The templates and oligonucleotides used for PCR are listed in Supplementary Data 1. The full example DNA sequence of the HAL-CDD-TGA-RBS-ATG-NDD-HAR20-PacI-ampccdB-PacI-HAR cassette for splitting busA into busA-1 is provided in Supplementary Data 2. The amino acid sequences of the CDD of SlnA1, NDD of SlnA2, CDD of SlnA7, NDD of SlnA8, CDD of StiB, and NDD of StiC are provided in Supplementary Table 1.
Five hundred nanograms of the HAL-CDD-TGA-RBS-ATG-NDD-HAR20-PacI-ampccdB-PacI-HAR cassette was electroporated into E. coli GBred-gyrA462 containing pBAC11-phiC31-bus29 (or pBAC-phiC31-avm) and expressing Redγ, Redβ, Redα, and RecA. The HAL-CDD-TGA-RBS-ATG-NDD-HAR20-PacI-ampccdB-PacI-HAR cassette was inserted into the busA gene (or the aveA2 gene) through recombineering. E. coli GBred-gyrA462 colonies containing pBAC-phiC31-bus-ampccdB (or pBAC-phiC31-ave-ampccdB) were selected on LB plates supplemented with ampicillin (100 µg mL−1). Correct clones were identified by restriction analysis, and the DNA sequence of the HAL-CDD-TGA-RBS-ATG-NDD-HAR20-PacI-ampccdB-PacI-HAR cassette was confirmed by sequencing.
A 200-ng sample of PacI-digested pBAC-phiC31-bus-ampccdB (or pBAC-phiC31-ave-ampccdB) was reacted in 1 × ABclonal MultiF Seamless Assembly Mix (ABclonal, cat. no. RK21020) in a 20 µL reaction system, which was incubated at 50 °C for 1 h and then held at 4 °C in a thermocycler. The reaction mixture was transferred to a Millipore membrane filter (Merck-Millipore, cat. no. VSWP01300) and dialysed against ddH2O at room temperature for 30 min. Then, E. coli GB2005 cells were electroporated with 12 µL of dialysed reaction mixture. E. coli GB2005 colonies containing recombinant BACs carrying the desired sequence were selected on LB plates supplemented with chloramphenicol (15 µg mL−1). Twelve colonies were picked to identify correct clones by restriction analysis, and the accuracy was greater than 80%. Correct clones were then transformed into S. albus J1074 or S. coelicolor CH999 through conjugation to integrate the gene cluster into the phiC31 attB site in the chromosome via site-specific recombination mediated by the phiC31 integrase.
Direct cloning of the avermectin gene cluster
The 80-kb avermectin gene cluster30 was directly cloned from the chromosome of Streptomyces avermitilis DSM46492 using the strategy illustrated in Supplementary Fig. 9. Briefly, the chromosome of S. avermitilis DSM46492 was isolated and digested with BglII (New England BioLabs, Beijing) according to the protocol described for the RecET direct cloning method62 to release a 73-kb fragment containing the aveR-aveA4 gene or with SnaBI + EcoRV to release a 15-kb fragment containing the aveA4-aveBVIII gene. The 73-kb fragment and the 15-kb fragment were cloned and inserted into the pBeloBAC11 vector and the pBR322 vector, respectively, using the ExoCet method63. Finally, the 73-kb fragment and the 15-kb fragment, which have 2.7 kb of overlap, were stitched by recombineering64 to obtain the intact avermectin gene cluster in the pBeloBAC11 vector (pBAC-ave; Supplementary Fig. 10). The PCR oligonucleotides and templates used to prepare the pBeloBAC11 and pBR322 vectors for ExoCET cloning and to prepare the cassettes for stitching are listed in Supplementary Table 2. The apra-oriT-attP-phiC31int cassette62 (apra, the apramycin resistance gene; oriT, the origin for conjugal transfer; phiC31int, the phiC31 integrase gene; attP, the phiC31 attachment site) was then inserted into pBAC-ave using recombineering to form the E. coli-Streptomyces shuttle vector pBAC-phiC31-ave (Supplementary Fig. 11). Correct recombinant BACs and plasmids were identified via PstI restriction analysis. pBAC-phiC31-ave was transformed into S. coelicolor CH99965 through conjugation and then integrated into the phiC31 attB site in the chromosome via site-specific recombination mediated by the phiC31 integrase. The recombinant strains were selected on mannitol soya flour agar plates supplemented with apramycin (50 µg mL−1) and nalidixic acid (25 µg mL−1). Correct recombinants were confirmed by PCR.
Construction of the epothilone expression plasmids
The p15A-epo-IR-Tps-bsd-oriT-IR-kan plasmid carries the epothilone operon for heterologous expression in Myxococcus xanthus66. The SA15p promoter54 from S. albus J1074 was inserted upstream of the epothilone operon in the p15A-epo-IR-Tps-bsd-oriT-IR-kan plasmid using the RedEx method. The apra-oriT-attP-phiC31int cassette62 was then inserted into the plasmid using recombineering to form the E. coli-Streptomyces shuttle vector p15A-phiC31-epo which was transformed into S. albus J1074 through conjugation and then integrated into the phiC31 attB site in the chromosome via site-specific recombination mediated by the phiC31 integrase. The recombinant strains were selected on mannitol soya flour agar plates supplemented with apramycin (50 µg mL−1) and nalidixic acid (25 µg mL−1). Correct recombinants were confirmed by PCR.
Fermentation of Streptomyces strains
The seed culture of each S. albus J1074 strain harbouring the butenyl-spinosyn or epothilone gene clusters was prepared by inoculating colonies from a plate into a 250 mL flask containing 30 mL of tryptic soy broth, followed by incubation at 30 °C with shaking at 220 rpm for 72 h. Individual 250−mL flasks containing 30 mL of fermentation broth (1% glycerol, 1.5% soytone, 4% glucose, 3% soluble starch, 0.65% peptone, 1% beef extract, 0.1% magnesium sulfate, 0.05% yeast extract, 0.24% CaCO3, and 0.2% NaCl) were inoculated with 1 mL of the seed culture and incubated for 7 days. Then, 600 µL (2%, v/v) of Amberlite XAD-16 adsorber resin was added, and the culture was incubated for another 2 days.
The seed culture of each S. coelicolor CH999 strain harbouring the avermectin gene clusters was prepared by inoculating colonies from a plate into a 250 mL flask containing 30 mL of seed medium A (1.5% soybean meal, 2.5% glucose, 0.3% CaCO3), followed by incubation at 30 °C with shaking at 220 rpm for 24 h. Then, 1 mL of culture from seed medium A was inoculated into 30 mL of seed medium B (1.5% soybean meal, 4% glucose, 1.5% skim powdered milk, 0.5% yeast extract) and incubated at 30 °C with shaking at 220 rpm for 72 h. Subsequently, 1 mL of the culture from seed medium B was inoculated into 30 mL of fermentation broth (3% glucose, 0.2% (NH4)2SO4, 0.5% CaCO3, 0.3% NaCl, 0.005% FeSO4, 0.005% MnSO4, 0.01% MgSO4) and incubated for 6 days. Finally, 600 µL (2%, v/v) of the Amberlite XAD-16 adsorber resin was added, and the culture was incubated for another 2 days.
The Streptomyces culture described above was centrifuged to pellet the cells and Amberlite XAD-16; then, 40 mL of methanol was added to each pellet to resuspend the cells, and Amberlite XAD-16 was added. The suspension was sonicated for 20 min, incubated at 30 °C with shaking at 220 rpm for 2 h, and evaporated. The residue was dissolved in 1 mL of methanol and then used for high-performance liquid chromatography−mass spectrometry (HPLC-MS) analysis.
HPLC‒MS analysis
High-resolution mass spectrometric analysis of butenyl-spinosyn A, epothilone A, and avermectin B1a prepared above was performed on an Impact HD micro TFF-Q III mass spectrometer (Bruker Daltonics, Bremen, Germany) using a standard ESI source operating in centroid mode (100 to 1500 m/z) with positive ionisation mode and automatic MS2 fragmentation. HPLC analysis was performed on an Ultimate 3000 UHPLC-DAD system (Thermo Fisher Scientific) equipped with an Acclaim RSLC 120 C18 column (2.2 m, 2.1 × 100 mm, Thermo Scientific) at a flow rate of 0.3 mL min−1 and at a UV wavelength of 200–600 nm. The mobile phase solvent A was 0.1% (v/v) formic acid in H2O. The mobile phase solvent B was 0.1% (v/v) formic acid in acetonitrile. The HPLC gradient programme for butenyl-spinosyn was set as follows: 0‒5 min, 5–50% solvent B with a linear gradient; 5‒20 min, 50% solvent B; 20‒25 min, 50–95% solvent B with a linear gradient; 25‒30 min, 95% solvent B; and 30‒35 min, 5% solvent B. The HPLC gradient programme for avermectins was set as follows: 0‒5 min, 40% solvent B; 5‒6 min, 40–70% solvent B with a linear gradient; 6‒18 min, 70% solvent B; 18‒24 min, 70–90% solvent B with a linear gradient; 24‒28 min, 90% solvent B; and 28‒31 min, 40% solvent B. The HPLC gradient programme for epothilones was set as follows: 0‒5 min, 5% solvent B; 5‒25 min, 5–95% solvent B with a linear gradient; 25‒30 min, 95% solvent B; and 30‒35 min, 5% solvent B.
Fusion of the gusA gene
The gusA-amp cassette containing the gusA gene, the ampicillin resistance gene (amp), and the recombineering homology arms for insertion at the 3’-end of the busA, busA-1-1, busA-1-2, busA-2-1, or busA-2-2 gene was prepared by overlap extension PCR. The gusA and amp genes with overlapping sequences were amplified via PCR, and these two fragments were subsequently combined via overlap extension PCR to generate a PCR fragment containing the gusA-amp cassette. The templates and oligonucleotides used for PCR are listed in Supplementary Table 3. E. coli GB05-red cells containing pBAC11-phiC31-bus, pBAC11-phiC31-busA-1, or pBAC11-phiC31-busA-2 were induced with L-arabinose to express Redγ, Redβ, Redα, and RecA. The gusA-amp PCR fragment was electroporated into the above strains. The gusA-amp cassette was then inserted at the 3’-end of the busA, busA-1-1, busA-1-2, busA-2-1, or busA-2-2 genes by recombineering. E. coli GB05-red cells containing the correct recombinant BACs were selected on LB plates supplemented with ampicillin. Correct clones were identified by restriction analysis, and the DNA sequences of the homology arms, the linker and the gusA gene were confirmed by sequencing. Correct clones were then transformed into S. albus J1074 through conjugation to integrate the gene cluster onto the phiC31 attB site in the chromosome via site-specific recombination mediated by the phiC31 integrase.
Quantitative real-time PCR analysis of gene transcription
Colonies of S. albus J1074 were inoculated into a 250 mL flask containing 30 mL of tryptic soy broth and incubated at 30 °C with shaking at 220 rpm for 72 h. Then, 500 μL of the culture was centrifuged to collect the cells. Total RNA was purified using an RNAprep Pure Kit (Tiangen, cat. no. DP430). Reverse transcription was performed with a PrimeScript RT Kit (Takara, cat. no. RR047A). Quantitative real-time PCR (qRT‒PCR) was performed with SYBR Premix Ex Taq GC (Takara, cat. no. RR071A) on a StepOnePlus Real-Time PCR System (Applied Biosystems). The hrdB gene, which encodes the sigma factor, was used as the reference gene. The primers used for qRT‒PCR are listed in Supplementary Table 4.
Glucuronidase activity assay
Colonies of S. albus J1074 strains containing the gusA gene were inoculated into a 250 mL flask containing 30 mL of tryptic soy broth and incubated at 30 °C with shaking at 220 rpm for 72 h. The optical density at 600 nm (OD600) of the cultures was measured, and the cultures were diluted to an OD of 1.0. Culture samples (2 mL) were centrifuged at 12,000 × g for 1 min to pellet the cells. Nine hundred microlitres of lysis buffer (50 mM phosphate buffer (pH 7.0), 0.1% Triton X-100, 5 mM dithiothreitol, 1 mg mL−1 lysozyme) was added to the resuspended cells. Cell lysis was performed at 37 °C for 20 min, and 900 µL of dilution buffer (50 mM phosphate buffer (pH 7.0), 0.1% Triton X-100, 5 mM dithiothreitol) was added to the lysate. The diluted lysate was centrifuged at 16,400 × g for 10 min at 4 °C. Then, 100 mL of lysate was added to a 96-well plate and mixed with 100 μL of dilution buffer supplemented with 2 mM p-nitrophenyl-β-D-glucuronide. A fully automatic microplate reader was used to measure the optical density at 415 nm (OD415) of the reaction mixture every minute for a total of 90 min of incubation at room temperature. One unit of glucuronidase activity indicates the amount of enzyme that catalyses the production of 1 μmol of p-nitrophenol in one minute. Glucuronidase activity was calculated in units per gram of dry biomass. As a reference, 100 μL of lysate was mixed with 100 μL of dilution buffer.
Integrating genes into S. albus Del14
The p15A-apra-phiC31 vectors carrying recombineering homology arms for cloning the busA, busA-1-1, or busA-2-1 genes from pBAC11-phiC31-bus, pBAC11-phiC31-busA-1, or pBAC11-phiC31-busA-2 were prepared by PCR. The templates and oligonucleotides used for PCR are listed in Supplementary Table 5. E. coli GB05-red cells containing pBAC11-phiC31-bus, pBAC11-phiC31-busA-1, or pBAC11-phiC31-busA-2 were induced with L-arabinose to express Redγ, Redβ, Redα, and RecA, respectively. The p15A-apra-phiC31 PCR fragment was electroporated into the above strains. The busA, busA-1-1, and busA-2-1 genes were cloned and inserted individually into the p15A-apra-phiC31 vector by recombineering. E. coli GB05-red cells containing the correct recombinant BACs were selected on LB plates supplemented with apramycin (20 µg mL−1). Correct clones were identified by restriction analysis, and the DNA sequences of the homology arms, the oriT site, the phiC31 integrase gene, and the attP site were confirmed by sequencing. p15A-apra-phiC31-busA, p15A-apra-phiC31-busA-1-1, and p15A-apra-phiC31-busA-2-1 were transformed into S. albus J1074 Del14 through conjugation and then integrated into the phiC31 attB site in the chromosome via site-specific recombination mediated by the phiC31 integrase.
Absolute quantification of targeted proteins by parallel reaction monitoring
Colonies of the S. albus J1074 Del14 strain expressing the BusA, BusA-1-1, or BusA-2-1 protein were inoculated into a 250 mL flask containing 30 mL of tryptic soy broth and incubated at 30 °C with shaking at 220 rpm for 72 h. Then, 2 mL of the culture was centrifuged at 8000 × g for 10 min to pellet the cells. The cell pellets were then sent to Beijing Biomarker Technologies Co., Ltd., for PRM analysis. The quality of the samples, including the protein amount (Supplementary Table 6) and the SDS-PAGE picture (Supplementary Fig. 12), was evaluated to be sufficient for parallel reaction monitoring (PRM). The AALVADDEPK stable 13C-isotope-labelled peptide was synthesised for use as an internal standard and added to each sample.
The mobile phase solvent A was 0.1% (v/v) formic acid in H2O. The mobile phase solvent B was 0.1% (v/v) formic acid in acetonitrile. The lyophilised powder of tryptic peptides was dissolved in 10 μL of the mobile phase solvent A. After centrifugation at 14,000 × g for 20 min at 4 °C, 1 μL of the supernatant was used for liquid chromatography−mass spectrometry (LC‒MS) analysis. LC was performed on an Easy 1200 (Thermo Scientific), and the gradient programme was set as follows: 0‒8 min, 6% solvent B; 8‒13 min, 6–12% solvent B; 13‒46 min, 12–30% solvent B; 46‒53 min, 30–40% solvent B; 53‒54 min, 40–95% solvent B; 54‒64 min, 95% solvent B; and 64‒65 min, 6% solvent B. MS was performed on an Orbitrap Fusion Lumos mass spectrometer (Thermo Scientific) using a Nanospray Flex™ (NSI) ion source operating in data-dependent acquisition mode (m/z 350‒1500). The ion spray voltage was 2.0 kV, and the temperature of the ion transfer tube was 320 °C. The MS1 resolution was 120,000, the automatic gain control (AGC) target value was 4 × 105, and the C-trap maximum ion injection time was 50 ms. The parent ions were fragmented by the high-energy collision dissociation (HCD) method for MS2 detection. The MS2 resolution was 15,000, the automatic gain control (AGC) target value was 5 × 104, and the C-trap maximum ion injection time was 22 ms. The collision energy for peptide fragmentation was 33%.
The raw data were analysed with Skyline (MacCoss Lab, University of Washington) to derive the peak area of each AALVADDEPK target peptide segment (Supplementary Figs. 13–15). The signal intensities for the AALVADDEPK peptide of the BusA, BusA-1-1, and BusA-2-1 proteins were quantified relative to each sample and normalised to the internal standard reference.
Cloning of the gusA fusion genes under control of the cumate-inducible promoter
The zeo-Pcum cassette containing the zeocin resistance gene (zeo), the cumate-inducible promoter (Pcum)67, and the homology arms for recombineering insertion upstream of the busA-gusA, busA-1-1-gusA, busA-1-2-gusA, busA-2-1-gusA, or busA-2-2-gusA fusion gene was prepared by overlap extension PCR. The zeocin resistance gene (zeo) and the cumate-inducible promoter (Pcum) with overlapping sequences were amplified via PCR and then joined together via overlap extension PCR to generate a PCR fragment containing the zeo-Pcum cassette. The templates and oligonucleotides used for PCR are listed in Supplementary Table 7. E. coli GB05-red cells containing pBAC11-phiC31-bus-gusA, pBAC11-phiC31-busA-1-1-gusA, pBAC11-phiC31-busA-1-2-gusA, pBAC11-phiC31-busA-2-1-gusA, or pBAC11-phiC31-busA-2-2-gusA were induced with L-arabinose to express Redγ, Redβ, Redα, and RecA. The gusA-amp PCR fragment was electroporated into the above strains. The zeo-Pcum cassette was then inserted upstream of the busA-gusA, busA-1-1-gusA, busA-1-2-gusA, busA-2-1-gusA, or busA-2-2-gusA fusion genes by recombineering. E. coli GB05-red cells containing the correct recombinant BACs were selected on LB plates supplemented with zeocin (20 µg mL−1). Correct clones were identified by restriction analysis, and the DNA sequences of the homology arms and the cumate-inducible promoter were confirmed by sequencing. Correct clones were then transformed into S. albus J1074 through conjugation to integrate the gene cluster onto the phiC31 attB site in the chromosome via site-specific recombination mediated by the phiC31 integrase.
Insertion of the kasOp* promoter upstream of the bus genes
The amp-kasOp cassette containing the ampicillin resistance gene (amp), the kasOp* promoter, and homology arms for recombineering insertion upstream of the busA-1-2, busA-2-2, busA-3-2, busB, busC, busD, or busE genes was prepared by overlap extension PCR. The templates and oligonucleotides used for PCR are listed in Supplementary Table 8. E. coli GB05-red61 cells containing pBAC11-phiC31-busA-1, pBAC11-phiC31-busA-2, or pBAC11-phiC31-busA-3 were induced with L-arabinose to express Redγ, Redβ, Redα, and RecA. The amp-kasOp PCR fragment was electroporated into the above strains. E. coli GB05-red cells containing the correct recombinant BACs were selected on LB plates supplemented with ampicillin. Correct clones were identified by restriction analysis and DNA sequencing of the homology arms, and the kasOp* promoter was confirmed by sequencing.
Statistical analysis
Statistical analysis comparing two groups was performed using a two-tailed unpaired t test. All statistical analyses were performed using GraphPad Prism 8.0.1 software. The p-values less than 0.05 were considered significant. Data are presented as the means ± s.d.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Data supporting the findings of this work are available within the paper and its Supplementary Information files. A reporting summary for this Article is available as a Supplementary Information file. Source data are provided in this paper.
References
Bagde, S. R., Mathews, I. I., Fromme, J. C. & Kim, C. Y. Modular polyketide synthase contains two reaction chambers that operate asynchronously. Science 374, 723–729 (2021).
Dutta, S. et al. Structure of a modular polyketide synthase. Nature 510, 512–517 (2014).
Grininger, M. Enzymology of assembly line synthesis by modular polyketide synthases. Nat. Chem. Biol. 19, 401–415 (2023).
Winn, M., Fyans, J. K., Zhuo, Y. & Micklefield, J. Recent advances in engineering nonribosomal peptide assembly lines. Nat. Prod. Rep. 33, 317–347 (2016).
Atanasov, A. G., Zotchev, S. B. & Dirsch, V. M. International Natural Product Sciences, T. & Supuran, C.T. Natural products in drug discovery: advances and opportunities. Nat. Rev. Drug Discov. 20, 200–216 (2021).
Cai, W. & Zhang, W. Engineering modular polyketide synthases for production of biofuels and industrial chemicals. Curr. Opin. Biotechnol. 50, 32–38 (2018).
Weissman, K. J. The structural biology of biosynthetic megaenzymes. Nat. Chem. Biol. 11, 660–670 (2015).
Nivina, A., Yuet, K. P., Hsu, J. & Khosla, C. Evolution and diversity of assembly-line polyketide synthases. Chem. Rev. 119, 12524–12547 (2019).
Bilyk, O. & Luzhetskyy, A. Metabolic engineering of natural product biosynthesis in actinobacteria. Curr. Opin. Biotechnol. 42, 98–107 (2016).
Tan, G. Y. & Liu, T. Rational synthetic pathway refactoring of natural products biosynthesis in actinobacteria. Metab. Eng. 39, 228–236 (2017).
Liu, R., Deng, Z. & Liu, T. Streptomyces species: Ideal chassis for natural product discovery and overproduction. Metab. Eng. 50, 74–84 (2018).
Fu, Y. et al. A meet-up of acetyl phosphate and c-di-GMP modulates BldD activity for development and antibiotic production. Nucleic Acids Res. 51, 6870–6882 (2023).
Xu, J. Y. et al. Protein acylation is a general regulatory mechanism in biosynthetic pathway of Acyl-CoA-derived natural products. Cell Chem. Biol. 25, 984–995 (2018).
Zhuo, Y. et al. Reverse biological engineering of hrdB to enhance the production of avermectins in an industrial strain of Streptomyces avermitilis. Proc. Natl Acad. Sci. USA 107, 11250–11254 (2010).
Xu, Y., You, D., Yao, L. L., Chu, X. & Ye, B. C. Phosphate regulator PhoP directly and indirectly controls transcription of the erythromycin biosynthesis genes in Saccharopolyspora erythraea. Microb. Cell Fact. 18, 206 (2019).
Wang, B., Guo, F., Dong, S. H. & Zhao, H. Activation of silent biosynthetic gene clusters using transcription factor decoys. Nat. Chem. Biol. 15, 111–114 (2019).
Wang, W. et al. Harnessing the intracellular triacylglycerols for titer improvement of polyketides in Streptomyces. Nat. Biotechnol. 38, 76–83 (2020).
Li, J. et al. A non-carboxylative route for the efficient synthesis of central metabolite malonyl-CoA and its derived products. Nat. Catal. 7, 361–374 (2024).
Herzel, L., Stanley, J. A., Yao, C. C. & Li, G. W. Ubiquitous mRNA decay fragments in E. coli redefine the functional transcriptome. Nucleic Acids Res. 50, 5029–5046 (2022).
Huter, P. et al. Structural basis for ArfA-RF2-mediated translation termination on mRNAs lacking stop codons. Nature 541, 546–549 (2017).
Wang, C. et al. Structural basis of transcription-translation coupling. Science 369, 1359–1365 (2020).
Keiler, K. C. Mechanisms of ribosome rescue in bacteria. Nat. Rev. Microbiol. 13, 285–297 (2015).
Stinear, T. P. et al. Giant plasmid-encoded polyketide synthases produce the macrolide toxin of Mycobacterium ulcerans. Proc. Natl. Acad. Sci. USA 101, 1345–1349 (2004).
Zaburannyi, N., Rabyk, M., Ostash, B., Fedorenko, V. & Luzhetskyy, A. Insights into naturally minimised Streptomyces albus J1074 genome. BMC Genomics 15, 97 (2014).
Hahn, D. R. et al. Butenyl-spinosyns, a natural example of genetic engineering of antibiotic biosynthetic genes. J. Ind. Microbiol. Biotechnol. 33, 94–104 (2006).
Li, J. et al. Chloroplastic metabolic engineering coupled with isoprenoid pool enhancement for committed taxanes biosynthesis in Nicotiana benthamiana. Nat. Commun. 10, 4850 (2019).
An, Z. H. et al. Increasing the heterologous production of spinosad in Streptomyces albus J1074 by regulating biosynthesis of its polyketide skeleton. Synth. Syst. Biotechnol. 6, 292–301 (2021).
Zhu, Z. H. et al. SlnR is a positive pathway-specific regulator for salinomycin biosynthesis in Streptomyces albus. Appl. Microbiol. Biotechnol. 101, 1547–1557 (2017).
Song, C. et al. RedEx: a method for seamless DNA insertion and deletion in large multimodular polyketide synthase gene clusters. Nucleic Acids Res. 48, e130 (2020).
Zhuo, Y. et al. Synthetic biology of avermectin for production improvement and structure diversification. Biotechnol. J. 9, 316–325 (2014).
Molnar, I. et al. The biosynthetic gene cluster for the microtubule-stabilizing agents epothilones A and B from Sorangium cellulosum So ce90. Chem. Biol. 7, 97–109 (2000).
Ikeda, H. & Omura, S. Avermectin biosynthesis. Chem. Rev. 97, 2591–2610 (1997).
Gaitatzis, N. et al. The biosynthesis of the aromatic myxobacterial electron transport inhibitor stigmatellin is directed by a novel type of modular polyketide synthase. J. Biol. Chem. 277, 13082–13090 (2002).
Myronovskyi, M., Welle, E., Fedorenko, V. & Luzhetskyy, A. β-Glucuronidase as a sensitive and versatile reporter in actinomycetes. Appl. Environ. Microbiol. 77, 5370–5383 (2011).
Schmitz, U. K., Lonsdale, D. M. & Jefferson, R. A. Application of the beta-glucuronidase gene fusion system to Saccharomyces cerevisiae. Curr. Genet. 17, 261–264 (1990).
Peterson, A. C., Russell, J. D., Bailey, D. J., Westphall, M. S. & Coon, J. J. Parallel reaction monitoring for high resolution and high mass accuracy quantitative, targeted proteomics. Mol. Cell Proteom. 11, 1475–1488 (2012).
Zhang, G. et al. Integrated analysis of transcriptomic, miRNA and proteomic changes of a novel hybrid yellow catfish uncovers key roles for miRNAs in heterosis. Mol. Cell Proteom. 18, 1437–1453 (2019).
Myronovskyi, M. et al. Generation of a cluster-free Streptomyces albus chassis strains for improved heterologous expression of secondary metabolite clusters. Metab. Eng. 49, 316–324 (2018).
Keiler, K. C., Waller, P. R. & Sauer, R. T. Role of a peptide tagging system in degradation of proteins synthesized from damaged messenger RNA. Science 271, 990–993 (1996).
Fritze, J., Zhang, M., Luo, Q. & Lu, X. An overview of the bacterial SsrA system modulating intracellular protein levels and activities. Appl. Microbiol. Biotechnol. 104, 5229–5241 (2020).
Himeno, H., Kurita, D. & Muto, A. tmRNA-mediated trans-translation as the major ribosome rescue system in a bacterial cell. Front. Genet. 5, 66 (2014).
Handa, Y., Inaho, N. & Nameki, N. YaeJ is a novel ribosome-associated protein in Escherichia coli that can hydrolyze peptidyl-tRNA on stalled ribosomes. Nucleic Acids Res. 39, 1739–1748 (2011).
Chadani, Y., Ono, K., Kutsukake, K. & Abo, T. Escherichia coli YaeJ protein mediates a novel ribosome-rescue pathway distinct from SsrA- and ArfA-mediated pathways. Mol. Microbiol. 80, 772–785 (2011).
Braud, S., Lavire, C., Bellier, A. & Mazodier, P. Effect of SsrA (tmRNA) tagging system on translational regulation in Streptomyces. Arch. Microbiol. 184, 343–352 (2006).
Horbal, L., Fedorenko, V. & Luzhetskyy, A. Novel and tightly regulated resorcinol and cumate-inducible expression systems for Streptomyces and other actinobacteria. Appl. Microbiol. Biotechnol. 98, 8641–8655 (2014).
Balakrishnan, R. et al. Principles of gene regulation quantitatively connect DNA to RNA and proteins in bacteria. Science 378, eabk2066 (2022).
Jiang, C. et al. Establishing an efficient salinomycin biosynthetic pathway in three heterologous Streptomyces hosts by constructing a 106-kb multioperon artificial gene cluster. Biotechnol. Bioeng. 118, 4668–4677 (2021).
Wang, Y. et al. Internal promoters and their effects on the transcription of operon genes for epothilone production in Myxococcus xanthus. Front. Bioeng. Biotechnol. 9, 758561 (2021).
Wang, W. et al. An engineered strong promoter for streptomycetes. Appl. Environ. Microbiol. 79, 4484–4492 (2013).
Lyu, X., Yang, Q., Zhao, F. Z. & Liu, Y. Codon usage and protein length-dependent feedback from translation elongation regulates translation initiation and elongation speed. Nucleic Acids Res. 49, 9404–9423 (2021).
Pavco, P. A. & Steege, D. A. Characterization of elongating T7 and SP6 RNA polymerases and their response to a roadblock generated by a site-specific DNA binding protein. Nucleic Acids Res. 19, 4639–4646 (1991).
Borkotoky, S. & Murali, A. The highly efficient T7 RNA polymerase: A wonder macromolecule in biological realm. Int. J. Biol. Macromol. 118, 49–56 (2018).
Thiel, V., Herold, J., Schelle, B. & Siddell, S. G. Infectious RNA transcribed in vitro from a cDNA copy of the human coronavirus genome cloned in vaccinia virus. J. Gen. Virol. 82, 1273–1281 (2001).
Kortmann, M., Kuhl, V., Klaffl, S. & Bott, M. A chromosomally encoded T7 RNA polymerase-dependent gene expression system for Corynebacterium glutamicum: construction and comparative evaluation at the single-cell level. Microb. Biotechnol. 8, 253–265 (2015).
Wei, J. et al. Development and application of a T7 RNA polymerase-dependent expression system for antibiotic production improvement in Streptomyces. Biotechnol. Lett. 39, 857–864 (2017).
Lussier, F. X., Denis, F. & Shareck, F. Adaptation of the highly productive T7 expression system to Streptomyces lividans. Appl. Environ. Microbiol. 76, 967–970 (2010).
DeLoughery, A., Lalanne, J. B., Losick, R. & Li, G. W. Maturation of polycistronic mRNAs by the endoribonuclease RNase Y and its associated Y-complex in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 115, E5585–E5594 (2018).
Lee, Y. et al. The transcription unit architecture of streptomyces lividans TK24. Front. Microbiol. 10, 2074 (2019).
Hwang, S. et al. Elucidating the regulatory elements for transcription termination and oosttranscriptional porocessing in the streptomyces clavuligerus genome. mSystems 6, https://doi.org/10.1128/msystems.01013-20 (2021).
Wang, H. et al. Improved seamless mutagenesis by recombineering using ccdB for counterselection. Nucleic Acids Res. 42, e37 (2014).
Song, C. et al. Enhanced heterologous spinosad production from a 79-kb synthetic multi-operon assembly. ACS Synth. Biol. 8, 137–147 (2019).
Wang, H. et al. RecET direct cloning and Redab recombineering of biosynthetic gene clusters, large operons or single genes for heterologous expression. Nat. Protoc. 11, 1175–1190 (2016).
Wang, H. et al. ExoCET: exonuclease in vitro assembly combined with RecET recombination for highly efficient direct DNA cloning from complex genomes. Nucleic Acids Res. 46, e28 (2018).
Zhang, Y., Buchholz, F., Muyrers, J. P. & Stewart, A. F. A new logic for DNA engineering using recombination in Escherichia coli. Nat. Genet. 20, 123–128 (1998).
McDaniel, R., Ebert-Khosla, S., Hopwood, D. A. & Khosla, C. Engineered biosynthesis of novel polyketides. Science 262, 1546–1550 (1993).
Fu, J. et al. Efficient transfer of two large secondary metabolite pathway gene clusters into heterologous hosts by transposition. Nucleic Acids Res. 36, e113 (2008).
Li, X. et al. Improving spinosad production by tuning expressions of the forosamine methyltransferase and the forosaminyl transferase to reduce undesired less active byproducts in the heterologous host Streptomyces albus J1074. Microb. Cell Fact. 22, 15 (2023).
Acknowledgements
This work was supported by the National Natural Science Foundation of China (32122049 to H.W.); Natural Science Foundation of Shandong Province (ZR2022JQ11 and ZR2019ZD22 to H.W.); the Fundamental Research Funds of Shandong University (2023QNTD001 to H.W.); National Key Research & Development Programme of China (2021YFC2101000 to J.L.); the Fund for Distinguished Young Scholars of SDU (H.W.); the SKLMT Frontiers and Challenges Project (SKLMTFCP-2023-05 to H.W.); and the 111 Project (B16030 to H.W.). The authors acknowledge Prof. Andriy Luzhetskyy from Universität des Saarlandes for providing the Streptomyces albus Del14 strain; Jingyao Qu, Jing Zhu and Zhifeng Li of the Core Facilities for Life and Environmental Sciences, State Key Laboratory of Microbial Technology of Shandong University for their help and guidance in LC-MS.
Author information
Authors and Affiliations
Contributions
H.W., J.L. and Y.L. conceived and designed the research. H.W., Y.L., C.S. and C.J. designed the PKS splitting strategy. Y.L. designed and performed experiments on butenyl-spinosyn and avermectin PKS splitting, promoter exchange, fermentation, HPLC-MS, GusA fusion, qRT-PCR, glucuronidase activity assay, and PRM analysis. Q.C. designed and performed experiments on epothilone PKS splitting, fermentation, and HPLC-MS analysis. H.S. and Z.L. performed cloning of the avermectin gene cluster. Y.L., R.G. and R.H. performed transformation and restriction analysis. H.W. and J.L. supervise the project. H.W., Y.L. and J.L. discussed the data and wrote the manuscript.
Corresponding author
Ethics declarations
Competing interests
H.W., Y.L. and J.L. are inventors of the pending patent application submitted by Shandong University that covers the PKS module splitting method described in this study (CN2024118692176). The remaining authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Måns Ehrenberg, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Liu, Y., Song, C., Cui, Q. et al. Improving polyketide biosynthesis by rescuing the translation of truncated mRNAs into functional polyketide synthase subunits. Nat Commun 16, 774 (2025). https://doi.org/10.1038/s41467-025-55973-0
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41467-025-55973-0
This article is cited by
-
Reply to: Docking domains from modular polyketide synthases and their use in engineering
Nature Communications (2025)
-
Docking domains from modular polyketide synthases and their use in engineering
Nature Communications (2025)
