A rational multi-target combination strategy for synergistic improvement of non-ribosomal peptide production

Yan, Hao; Xin, Zhenguo; Sang, Ziwei; Li, Xingwang; Xie, Jia; Wu, Jiale; Pang, Shen; Wen, Ying; Wang, Weishan

doi:10.1038/s41467-025-57073-5

Download PDF

Article
Open access
Published: 22 February 2025

A rational multi-target combination strategy for synergistic improvement of non-ribosomal peptide production

Hao Yan^1,2,3,
Zhenguo Xin²,
Ziwei Sang^1,2,
Xingwang Li¹,
Jia Xie¹,
Jiale Wu¹,
Shen Pang²,
Ying Wen¹ &
…
Weishan Wang ORCID: orcid.org/0000-0001-7827-2696^2,3,4

Nature Communications volume 16, Article number: 1883 (2025) Cite this article

8469 Accesses
6 Citations
14 Altmetric
Metrics details

Subjects

Abstract

Non-ribosomal peptides (NRPs) are pharmaceutically important natural products that include numerous clinical drugs. However, the biosynthesis of these NRPs is intricately regulated and improving production through manipulation of multiple regulatory targets remains largely empirical. We here develop a screening-based, multi-target rational combination strategy and demonstrate its effectiveness in enhancing the titers of three NRP drugs ˗ daptomycin, thaxtomin A and surfactin. Initially, we devise a reliable colorimetric analog co-expression and co-biosynthesis reporter system for screening high-yielding phenotypes. Subsequently, through coupling CRISPR interference to induce genome-wide differential expression, we identify dozens of repressors that inhibit the biosynthesis of these NRPs. To address the challenge of multi-target combination, we further developed a dual-target screen approach and introduced an interplay map based on the synergy coefficient of each pairwise interaction. Employing this strategy, we engineer the final strains with multi-target synergistic combination and achieve the titer improvement of the three NRPs. Our work provides a rational multi-target combination strategy for production improvement of NRPs.

Integrating genomics and metabolomics for scalable non-ribosomal peptide discovery

Article Open access 28 May 2021

Systematic identification of post-transcriptional regulatory modules

Article Open access 09 September 2024

Targeting Xist with compounds that disrupt RNA structure and X inactivation

Article 30 March 2022

Introduction

Microorganisms produce a diverse array of bioactive secondary metabolites (SMs), with non-ribosomal peptides (NRPs) representing a versatile class that includes antibiotics, immunosuppressants, anticancer agents, toxins, siderophores, pigments, and cytostatic^1,2. Streptomyces species are renowned for contributing approximately two-thirds of these pharmaceutical molecules³. However, in wild-type (WT) Streptomyces strains, the production of NRPs is typically constrained due to the intricate and stringent regulatory networks, involving up to 12% of the total genome dedicated to regulatory genes^4,5. This inherent regulation serves their own benefit rather than the objective of yield improvement in the pharmaceutical industry. Therefore, there is an urgent need to develop capabilities for identifying and reprogramming multiplexed synergetic regulators in order to redirect cellular resources towards NRP biosynthesis.

Construction of efficient microbial cell factories relies on globally tweaking the cell resources towards the biosynthesis of desired products, which is not always aligned with the direction of evolution that solely serves survival benefits^6,7. Therefore, systematic engineering is required for harnessing multiplex regulatory targets to reprogram genome-wide metabolism. To achieve accurate engineering of multiple regulatory targets, two key challenges must be comprehensively addressed: firstly, the rapid identification of targets that can affect production performance; secondly, determining which targets can be synergistically combined to achieve greater improvement. Given the complexity of genetic interactions in biological systems, empirical combinations of two or more regulatory targets do not always result in strains with improved performance^8,9. Henceforth, it is imperative to develop an efficient rational strategy for combining synergistic multi-target in strain improvement.

Genome-scale identification of the engineering targets involves two interconnected steps: “mutation or interference” and “selection”^10,11. Compared to existing methods for “mutation or interference”¹², such as CRISPR-based technology, which is a well-established and highly efficient mutation or interference tool used to identify functional target genes across various species^13,14, the strategy of “selection” presents greater challenges in identifying global targets¹⁵. For cases where desired products lack easily detectable characteristics such as color and growth-coupling traits, strain engineering often relies on available RNA-based riboswitches and transcription factor (TF)-based biosensors that respond to small molecules^16,17. Nevertheless, the number of well-characterized riboswitches or TFs is still insufficient to meet the diverse sensing requirements for production purpose¹⁵ — especially when it comes to detecting non-essential NRPs in Streptomyces species¹⁸.

In addition to the lack of reliable and universal selection methods, a more pressing concern lies in the selection of targets that exhibit synergetic and additive effects for combinational engineering while discarding the antagonistic targets. This is because simultaneous mutation or interference of two genes usually produces a phenotype that is unexpected in light of each loss-of-function individual effect¹⁹. Therefore, investigating the underlying functional relationship between individual favorable targets and mapping these interactions would undoubtedly provide a rational guide of multi-target combinations, enabling highly efficient successive strain improvement²⁰. However, we still face challenges in effectively combining individual favorable targets to generate synergistic effects in microbial strain improvement. Measurement of the synergy coefficient is a critical metric for evaluating the combined effects of multiple factors, such as in the synergistic screening of drugs²¹, since this parameter quantifies how the interactions between the components either enhances or diminishes their individual effects. Therefore, the determination of synergistic coefficient should be introduced in the development of a rational multi-target combination strategy for synergistic improvement of non-ribosomal peptide producing strains.

To address the aforementioned challenges, we chose daptomycin natural producer Streptomyces roseosporus, thaxtomin A heterogeneous producer Streptomyces coelicolor and surfactin producer Bacillus subtilis as a proof-of-concept to develop strategies for identifying and combining multiplex synergistic targets. Daptomycin is recognized as a clinically significant antibiotic for treating of various bacterial infections²². Thaxtomin A is acknowledged as a natural herbicide approved by the U.S. Environmental Protection Agency with high selectivity for controlling common monocot and dicot weed in rice crops and turfgrass landscapes²³. Surfactin is a cyclic lipopeptide, commonly used as an antibiotic for its capacity as a surfactant²⁴. We initially develop an analog co-expression and co-biosynthesis reporter system to establish a correlation between daptomycin titer and colorimetric indigoidine level, and evaluate the feasibility of CRISPR interference (CRISPRi) systems in S. roseosporus. Subsequently, we target genome-wide regulatory genes in S. roseosporus using a CRISPRi library and identify a total of 17 targets that inhibit daptomycin biosynthesis. Furthermore, we devise a scalable CRISPRi-based dual-screen strategy enabling massively parallel pairwise inhibition of these identified targets. By calculating synergy coefficient q of the pairwise combinations, we map the relationships between these targets and simultaneously engineer four targets with synergistic effect for high-yield construction. The resulting strain achieves a titer of 1054 mg/L in a 7.5-L fermenter, which, to the best of our knowledge, surpasses the previous highest reported yields in fermenters (812 μg/mL)²⁵. Furthermore, we demonstrate the broad applicability of this multi-target rational combination strategy by improving thaxtomins A titer to 352 mg/L in heterologous host S. coelicolor 1152 and surfactin titer to 878 mg/L in B. subtilis. Previously, the yield of thaxtomin A was increased to 728 mg/L and 504.6 mg/L in Streptomyces albidoflavus J1074²³ and Streptomyces coelicolor M1154²⁶, respectively, through combinatorial metabolic engineering and promoter engineering. To our knowledge, this work represents the first application of rational strategy for combining multiple targets in strain engineering.

Results

Overview of the screening-based rational multi-target combination strategy

The naturally occurring cellular regulation and metabolism networks were evolved to adapt to diverse environmental conditions for the survival benefit. Therefore, it is imperative to reconfigure this intricate network with the aim of achieving optimal production. In the past decade, numerous target screening strategies have been developed to identify independent regulatory targets that can enhance production²⁷. However, there is currently a lack of reliable theory or data-driven guidance on how to effectively combine these targets to achieve synergistic improvement. The existing trial-and-error combination strategy primarily relies on empirical combinations, which often result in inferior outcomes compared to single-target operations (Fig. 1 with gray background).

Compared to other microbial products, the biosynthesis of SMs in Streptomyces is subject to more intricate regulation^28,29. This necessitates the identification of genome-scale regulatory targets to inform combinational engineering of these multiple targets for synergistic performance. To achieve this objective, we developed a transferable screening-based rational multi-target combination strategy (Fig. 1 with light blue background). Following the individual target screening, we further established a method for screening the pairwise combinations of these targets. Subsequently, by determining the relationship between these targets and integrating their interactions into a network, we were able to identify multiple targets with positive synergistic effects. Finally, these synergistic targets were applied in strain improvement to achieve rational multi-target engineering.

Construction of an analog co-expression and co-biosynthesis reporter system

Initially, we established a universal reporter system that offers convenience and reliable indication of high-yield phenotypes for desired NRPs. We chose daptomycin produced by S. roseosporus as a proof-of-concept for developing this universal reporter system. Similar to other NRPs, daptomycin undergoes a complex and multifaceted regulatory process³⁰ that coordinates the initiation of daptomycin biosynthesis and its related physiological events, including expression of the daptomycin biosynthetic gene cluster (BGC), provision of cofactors and precursors, maintenance of production morphology, and other relevant pathways (Fig. 2a). These multiplex targets constitute a functional group that collectively orchestrate daptomycin biosynthesis. We therefore devised an analog co-expression and co-biosynthesis reporter system to accurately reflect the efficiency of the representative daptomycin biosynthetic machinery within this group. Specifically, we chose another NRP type indigoidine BGC containing only a single module NRP synthase (NRPS) gene (idgS) and a 4′-phosphopantetheinyl transferase (sfp) from Streptomyces lavendulae to be co-expressed with daptomycin BGC by inserting the promoterless idgS-sfp into the 3′ end of dptD (Fig. 2a and Supplementary Fig. 1a). Since both products ˗ daptomycin and indigoidine ˗ belong to NRPs, measuring colorimetric indigoidine levels can not only reflect the expression level of the daptomycin BGC and its related genes but also indicate other common targets that contribute to NRP biosynthesis. Therefore, our developed reporter system featuring analog co-expression and co-biosynthesis could improve the reliability in reporting the NRP high-yield phenotype.

**Fig. 2: Design and characterization of the reporter system.**

To assess the availability of idgS-sfp in S. roseosporus, idgS-sfp under control of the dptEp or ermE*p promoters was integrated into WT chromosome, respectively. We observed both resulting strains exhibited blue pigment on DA1 plates or in liquid medium (Supplementary Fig. 1b, c), indicating successful expression of indigoidine. Spectral scanning revealed a maximum absorption peak at 600 nm for indigoidine (Supplementary Fig. 1d), with its level being linearly and positively correlated with OD₆₀₀ value (Supplementary Fig. 1e), demonstrating the convenient use of indigoidine as a reporter for colorimetric end-point assays. We next inserted the idgS-sfp cassette after the stop codon of dptD by homologous recombination to construct the analog co-expression and co-biosynthesis reporter strain Srdi (Supplementary Fig. 2a–c). Reverse transcription PCR analysis revealed a large polycistronic transcript encompassing dptEFABCD-idgs-sfp-dptGH (Supplementary Fig. 2d), indicating this strategic placement of idgS-sfp within the daptomycin BGC ensured coordinated expression and biosynthesis of both daptomycin and indigoidine compounds. Subsequently, we assessed the impact of idgS-sfp on cell growth and daptomycin production. We did not observe significant difference in cell growth and daptomycin production between WT and Srdi fermentation broth samples (Supplementary Fig. 2e–f).

To evaluate the correlation between indigoidine production and daptomycin titer, we employed two strategies to perturb daptomycin titers in Srdi. Firstly, we individually overexpressed two activators (AtrA, BldD)^30,31 and two repressors (DasR, WblA)^32,33 known to influence daptomycin production in Srdi. Secondly, we replaced the native promoter of dptEFABCD-idgs-sfp-dptGH operon with a strong promoter kasO*p. Our finding revealed a positive correlation between indigoidine production and their respective daptomycin titer in overexpression strains Srdi^OatrA, Srdi^ObldD, Srdi^OdasR, Srdi^OwblA, Srdi^kasOdpt and control strain Srdi¹¹³⁹ (R² = 0.95662) (Fig. 2b), indicating the colored NRP indigoidine faithfully reported the level of co-expressed and co-biosynthesized daptomycin. Additionally, we also observed a positive correlation between the transcript level of the dptEFABCD-idgs-sfp-dptGH operon and the corresponding daptomycin titer (R² = 0.86465) (Fig. 2c), suggesting the expression level of daptomycin BGC could serve as an indicator for daptomycin biosynthesis to some extent. However, the indigoidine level showed a weaker correlation with the transcript abundance of daptomycin BGC compared to that with the titer of daptomycin (Fig. 2b and Supplementary Fig. 2g), suggesting that apart from genes contributing to co-transcription, other targets beneficial for NRP biosynthesis also played a noticeable role. These results indicated that our designed NRP analog co-expression and co-biosynthesis reporter system could accurately detect and quantify changes in daptomycin titer.

An exception further demonstrated the advantage of our reporter system. We observed a higher expression level of dptEFABCD-idgs-sfp-dptGH operon in Srdi^kasOdpt compared to that in Srdi^ObldD; however, both daptomycin and indigoidine yields were lower (Fig. 2d and Supplementary Fig. 2g). This discrepancy arises from the fact that Srdi^kasOdpt solely enhances daptomycin production through upregulating the expression of daptomycin BGC, while Srdi^ObldD involves coordinated regulation of a group of related genes responsible for daptomycin biosynthesis. This hypothesis was validated through comparative transcriptome analysis of these strains: only a significant upregulation in the expression levels of daptomycin BGC was observed in Srdi^kasOdpt (Fig. 2e); in contrast, multiple alterations were found in the expression levels of genes involved in amino acid metabolism, cofactor metabolism, morphology, and other related processes besides the daptomycin BGC in Srdi^ObldD (Fig. 2f). Similarly, this pleiotropic perturbation was also observed in Srdi^OatrA (Fig. 2g). Based on this finding, we concluded that our developed NRP analog co-expression and co-biosynthesis reporter system was qualified for identifying targets with pleiotropic effects on daptomycin production.

Evaluation of CRISPRi using the developed reporter system

Having developed a reliable reporter system, we proceeded to evaluate the availability of the established CRISPRi system in other Streptomyces species³⁴ for target screening in S. roseosporus. We selected bldD, encodes a global transcription factor, as target due to its role in activating daptomycin production and repressing morphological differentiation³⁰. Following the rule that single guide RNA (sgRNA) targeting the non-template strand (NT) are highly effective³⁴, we designed a sgRNA specifically targeting bldD. Our result revealed that inhibition of bldD (strain WT^bldDi) led to a phenotype similar to that of a bldD deletion mutant ΔbldD³⁰, displaying earlier and enhanced sporulation on DA1 plate (Fig. 3a). Further transcript analysis of bldD indicated a 54% reduction in WT^bldDi strain compared to WT and WT^- (containing plasmid without sgRNA) (Fig. 3b). Inhibition of bldD transcription resulted in lower expression levels of daptomycin BGC in WT^bldDi strain compared to WT, while higher than ΔbldD (Fig. 3b). Consistent with these findings, daptomycin titer exhibited the same trends (Fig. 3c). Collectively, these results demonstrated that the CRISPRi system could be employed for construction of CRISPRi library in S. roseosporus.

Fig. 3: Test of CRISPRi-based reporter system for screening in *S. roseosporus.*

To test the applicability of CRISPRi with our developed analog co-expression and co-biosynthesis reporter system, we firstly conducted a mock test using known repressor wblA³³ of daptomycin biosynthesis (Fig. 3d). We generated Srdi^wblAi strain with CRISPRi targeting wblA in Srdi. In comparison to parent Srdi and control strain Srdi^- (containing plasmid without sgRNA), the color intensity of this strain noticeably deepened in three-day culture (Fig. 3e and Supplementary Fig. 3a). The transcription level of wblA was decreased by 70%, while that of daptomycin BGC increased by 2- to 3-fold in Srdi^wblAi compared to that in Srdi^- (Supplementary Fig. 3b). Consequently, we observed a corresponding increase of around 30% in daptomycin titer in Srdi^wblAi (Fig. 3f), similar to the inhibition of wblA in WT using the same approach (a 26% increase, Fig. 3f). These results indicated that integration of the CRISPRi library with the analog co-expression and co-biosynthesis reporter system enables identification of targets influencing daptomycin production.

To further evaluate the compatibility of our analog co-expression and co-biosynthesis reporter system with the CRISPRi library for identifying unknown targets, we performed a mini-library screening using a CRISPRi approach. Specifically, we chose 11 TetR-family TFs (tetR1 through tetR11) located upstream or downstream of transporter as potential targets (Supplementary Fig. 3c). Subsequently, we constructed the library comprising sgRNA targeting the 11 tetR genes and known repressor wblA, which was then transformed into reporter strain Srdi (Fig. 3d). We randomly selected 100 transformants to measure indigoidine level with microplate reader and evaluate the bias towards targeted genes in these samples. Remarkably, all 13 genotype strains including the control strain Srdi^- were presented among the selected transformants (Fig. 3g). Furthermore, we determined the daptomycin titer for each strain and observed a strong linear relationship with the indigoidine level (R² = 0.90743) (Fig. 3h). These findings confirmed that our reporter system and CRISPRi library are highly compatible for efficient identification of unknown targets influencing daptomycin production.

We observed a noticeable darkening in the blue color of Srdi^tetR6i and Srdi^tetR7i (Fig. 3g and Supplementary Fig. 3d), indicating tetR6 and tetR7 function as negative regulators for daptomycin production. Subsequently, our transcription assay revealed a substantial increase in the transcripts of both daptomycin BGC and idgS in Srdi^tetR6i and Srdi^tetR7i compared to that in Srdi^- (Fig. 3i). Furthermore, we validated the inhibitory effect of tetR6 and tetR7 on daptomycin titer by individually deleting these genes in the WT strain (Supplementary Fig. 3e, f). Deletion mutants ΔtetR6 and ΔtetR7 exhibited an increase daptomycin titer by 35% and17% compared to WT, respectively (Fig. 3j), which correlated with enhanced indigoidine level (Fig. 3g). Taken together, these results suggested that combining CRISPRi library with the analog co-expression and co-biosynthesis reporter system could be utilized for identifying targets that inhibit daptomycin biosynthesis.

Screening targets inhibiting daptomycin production

Aiming to identify targets that directly or indirectly inhibit daptomycin production, we designed a pool of sgRNAs targeting all 610 predicted regulator genes in S. roseosporus genome, with three sgRNAs targeting different sites of each gene on the NT strand. The schematic representation of our screening workflow is depicted in Fig. 4a. Briefly, a plasmid library harboring 1830 unique sgRNA was transformed into reporter strain Srdi, and transformants exhibiting darker blue color on RM14 plates were selected for further screening through indigoidine quantitation in microplates. Transformants showing higher indigoidine level compared to Srdi were subjected to sequencing for identification of the targeted genes.

**Fig. 4: Screening targets that inhibit daptomycin production.**

The coverage of sgRNAs in both the CRISPRi plasmid library and Srdi transformants were determined using next generation sequencing (NGS). In the plasmid library, all 610 regulatory genes were targetable, with up to 87.4% of genes having three sgRNAs available for targeting and the total sgRNA coverage was 95.2% (Fig. 4b). Among Srdi transformants, the total sgRNA coverage was 81.6% where more than half of the genes (58.0%) were targeted with three sgRNAs and only one gene (0.2%) remained untargeted (Fig. 4c). Subsequently, over 300 darker blue colonies were visually screened from a pool of more than 20,000 transformants. (Supplementary Fig. 4a). These colonies were further subjected to an indigoidine colorimetric assay for quantification purposes (Fig. 4d and Supplementary Fig. 4b), leading to the discovery of 16 regulators that inhibit daptomycin synthesis (their genes were named as idt1 through idt16, Supplementary Fig. 4b). These regulators included one SARP-family, one LysR-family, one LacI-family, three MarR-familiy, four TetR-familiy, one XBE-family, one IclR-family, one Crp/Fnr-family, one MerR-family, one GntR-family, and one WhiB-family regulators. Transformants inhibiting these targets by cognate sgRNA exhibited a significant increase in daptomycin titer ranging from 17.4% to 37.8% compared to control strain Srdi^- (Fig. 4d). Notably, wblA and tetR6, which had been previously identified as a repressor for daptomycin production, were also included among the 16 screened genes, suggesting the reliability of our screening workflow.

We ranked these 16 targets based on their indigoidine output, which was consistent with the measurement of daptomycin production (Supplementary Table 1). To validate the identified targets, we randomly selected targets idt4, idt5, idt6, and idt7 to construct cognate mutant strains Sr4-Q57*, Sr5-Q22*, Sr6-Q94*, and Sr7-Q6*, respectively (Supplementary Fig. 4c). Shake-flask fermentation and qRT-PCR analyses revealed that loss-of-function of the four genes all led to an elevated daptomycin titer (Fig. 4e), as well as an increased expression of daptomycin BGC compared to WT (Fig. 4f). These results confirmed the inhibitory role of our identified targets in daptomycin production.

To further enhance daptomycin titer, we performed combinational inactivation by targeting two out of the four confirmed targets (idt4, idt5, idt6, and idt7) in WT, generating mutant strains Sr45, Sr46, Sr47, Sr56, Sr57, and Sr67. Surprisingly, we observed that daptomycin titer of Sr46 was lower than that of WT, and that the titer of Sr47 and Sr57 had no obvious change (Fig. 4g). These findings suggested that random combinations of these targets did not yield synergetic effects on enhancing daptomycin titer. Therefore, gaining insights into the cooperative interactions among the identified targets is crucial for further titer improvement.

Development of a dual-target screening approach

To identify the cooperative targets, we developed a construct that enables simultaneous dual-target inhibition by CRISPRi to conduct synergistic screening with the analog co-expression and co-biosynthesis reporter system (Fig. 5a). Specifically, we designed a synergistic library targeting pair of genes simultaneously to assess their effects on reporter output. Initially, we evaluated the efficiency of simultaneous dual-target inhibition through CRISPRi by targeting gfp and mCherry in S. roseosporus. Laser confocal microscopy observation (Fig. 5b) and fluorescence quantification assays (Fig. 5c) demonstrated effective parallel inhibition (>70%) of both gfp and mCherry by CRISPRi. Subsequently, we constructed a pairwise combinatorial library comprising the 17 identified target genes, and ensured coverage of all the combinations through NGS confirmation.

**Fig. 5: Development of a dual-target synergistic screening strategy.**

After transforming the pairwise combinatorial library into reporter strain Srdi, we randomly sequenced 600 colonies to quantify the synergistic effects of all the 171 combinations. Once again, we observed that the random combination of two favorable targets generated a great deal of uncertainty and does not always result in improved production performance (Fig. 5d). Subsequently, we chose five combinations with high levels (SrdiC1 through SrdiC5) and five combinations with low levels (SrdiC6 through SrdiC10) to measure daptomycin titer. As expected, their titers exhibited a consistent trend with their corresponding indigoidine level (Fig. 5e, f). These results underscored that our strategy of simultaneous dual-target screening provided a reliable guide for engineering two target combinations (Fig. 5e, f).

Synergistic multi-target engineering

To further achieve rational multi-target combination, we devised a combination strategy guided by the synergy coefficient q using the developed dual-target synergistic screen (Fig. 6a). Initially, q was calculated by comparing pairwise combinations with those containing individual targets based on indigoidine levels. Subsequently, an interaction map was constructed to identify targets within the positive synergetic. Guided by this map, we engineered the targets that exhibited positive synergetic effects with each other. Here the value of q calculated by Jin’s formula³⁵, which is a well-established approach for estimating the combined effect of two drugs on cellular pathological phenotypes²¹. Similarly, our focus was on investigating the synergistic effects between independent targets in relation to daptomycin synthesis phenotype.

We determined the q values of each pairwise combination of the identified targets (Supplementary Fig. 5a and Supplementary Table 2). These calculated q values were used to indicate antagonism (q < 0.85), additivity (0.85 ≤ q < 1.15), and synergism (q ≥ 1.15). Among these combinations, we observed antagonism in a total of 62 pairs, while additivity was found in 70 combinations, and synergy was displayed in 21 combinations (Fig. 6b). Furthermore, through mapping the interactions among these targets, we discovered that all pairwise interactions involving idt3, idt5, idt9, and idt15 demonstrated synergistic effects (Fig. 6c and Supplementary Fig. 5b).

To validate this finding, we constructed the strains inhibiting three out of these four targets in various combinations, as well as a strain simultaneously repressing all four targets. Upon comparing them to strains harboring each pairwise combination of the four targets individually, we found that all the three-target combinations exhibited significantly enhanced daptomycin titers, thereby confirming their positive synergistic effect (Fig. 6d). Notably, the strain carrying the four-target combination (denoted Sr11) achieved the highest daptomycin titer at 389 μg/mL in shake flask, representing an approximately 1.61-fold increase compared to WT (Fig. 6d). Conversely, random multi-target combinations (i.e., idt4, idt7, idt11, and idt12) exhibiting antagonism or additivity resulted in significantly lower or no significant difference in daptomycin titers (Supplementary Fig. 5c). These results highlighted the effectiveness of our synergy coefficient-guided combination strategy.

To test the production performance of the four-target combination in fed-batch condition, we conducted fermentation in a 7.5-L fermenter with sodium decanoate precursor feeding at 48 h and dextrin supplementation from 3 to 7 days (Supplementary Fig. 5d). Over a ten-day fermentation period, we found that daptomycin titer of Sr11 reached 1054 mg/L, which was two-fold higher than that of WT strain under the fed-batch condition, while both strains exhibited similar growth patterns throughout the fermentation process (Fig. 6e and Supplementary Fig. 5e).

To elucidate the underlying synergistic mechanism, we conducted comparative transcriptomics studies of the strains individually repressing these four targets (WT^idt3i, WT^idt5i, WT^idt9i, and WT^idt15i), as well as the four-target combination strain Sr11. Compared with the control strain WT^-, a total of 931, 316, 581, 328, and 2437 differentially expressed genes were identified in WT^idt3i, WT^idt5i, WT^idt9i, WT^idt15i, and Sr11, respectively (Supplementary Fig. 6a). Sr11 exhibited a broader range of genetic perturbations compared to other strains and resulted in comprehensive integration that encompassed the majority of differentially expressed genes associated with idt3, idt5, idt9, and idt15 inhibition. Functional clustering analysis revealed that the integrated differentially expressed genes in Sr11 play roles in remodeling multiplexed metabolic pathways such as biosynthesis of amino acids, biosynthesis of cofactors, quorum sensing, and nitrogen metabolism (Supplementary Fig. 6b). This integration was particularly evident in the upregulation of daptomycin BGC by approximately 15~26 fold in Sr11 (Fig. 6f). Furthermore, the time-course analysis demonstrated that individual inhibition of idt3, idt5, idt9, and idt15 led to growth stage-specific upregulation pattern of daptomycin BGC; however, the four-target combination displayed a synergistic effect throughout the entire fermentation period (Fig. 6g). These findings underscored the achievement of our developed multi-target combination strategy in enhancing daptomycin production by harnessing synergistic pleiotropic effects of multiplex regulators.

Synergistic multi-target engineering for thaxtomin A and surfactin production

To investigate the applicability of our multi-target combination strategy in other NRP-producing Streptomyces species, we applied the developed analog co-expression and co-biosynthesis reporter system along with the multi-target synergistic combination approach to enhance the titer of thaxtomin A in heterologous host S. coelicolor M1152 (Fig. 7a). We integrated the thaxtomin BGC (txt) and txt with idgS-sfp cassette inserted after the stop codon of txtB into the genome of M1152 to generate two strains: Scotxt as a control strain and Scotxti as a reporter strain (Fig. 7a). However, both strains failed to produce detectable levels of thaxtomin A when cultured in ISP4 liquid media (Fig. 7c). Transcription analyses revealed that the txt gene cluster was cryptic in both Scotxt and Scotxti. Therefore, we decided to screen for the repressors capable of inhibiting thaxtomin A production, aiming to harnessed these repressor combinations to boost the titer of thaxtomin A in S. coelicolor.

**Fig. 7: Synergistic multi-target engineering for improvement of thaxtomin and surfactin production.**

We designed a pool of sgRNAs targeting 856 regulatory genes in S. coelicolor genome, with three sgRNAs targeting distinct sites on the NT strand of each gene. Similar to the screening conducted for inhibiting daptomycin biosynthesis, we identified 10 targets for inhibiting thaxtomin A. They are sco5803 (named as thaR1), sco5231 (thaR2), sco3226 (thaR3), sco4230 (thaR4), sco3818 (thaR5), sco5862 (thaR6), sco1596 (thaR7), sco2794 (thaR8), sco1745 (thaR9), and sco0204 (thaR10). Transformants inhibiting these targets all exhibited activation of thaxtomin A production (Fig. 7c).

To further elucidate the interaction among these targets, we constructed a pairwise combinatorial library comprising these identified target genes. The synergy coefficient q based on indigoidine levels was evaluated for all the combinations of the identified targets (Supplementary Fig. 7a). Among these combinations, antagonism was observed in a total of 25 pairs, while additivity was found in 19 combinations, and synergy was displayed in 11 combinations (Supplementary Fig. 7b). By mapping the interactions among these targets, we discovered that all pairwise interactions involving thaR4, thaR7, and thaR8 showed positive synergistic effects (Fig. 7c and Supplementary Fig. 7c).

The engineered strains carrying the combinations of these targets all demonstrated a significant increase in thaxtomin A production (Fig. 7d). Notably, thaxtomin A titer reached a very high level of 352 mg/L in shake flask for the three-target combination strain (referred to as Scotxt4, Fig. 7d). These results not only demonstrated the successful heterologous activation of thaxtomin A in S. coelicolor, but also showcased the efficacy of our multi-target combination strategy in augmenting NRPs titers across diverse Streptomyces backgrounds.

To further test the broad applicability, we employed the synergistic multi-target engineering strategy to enhance surfactin production in B. subtilis 168 (Fig. 7a). We initially activate the biosynthesis of surfactin by integrating sfp gene from Bacillus amyloliquefaciens MT45 to generate 168S1 with surfactin titer of 225 mg/L (Fig. 7e). Subsequently, we inserted idgS-sfp cassette downstream of the stop codon of srfAD (Fig. 7a) and screened the repressors of surfactin biosynthesis similar to that of daptomycin and thaxtomin A. We identified 11 repressors (bsu_03880 (srfR1), bsu_40300 (srfR2), bsu_11500 (srfR3), bsu_06830 (srfR4), bsu_16170 (srfR5), bsu_37460 (srfR6), bsu_00370 (srfR7), bsu_08730 (srfR8), bsu_03770 (srfR9), bsu_13310 (srfR10), and bsu_24600 (srfR11)), and further validated that inhibition of these targets demonstrated significant activation of surfactin production (Fig. 7e). Then we determined the interaction map of these repressors by q value (Fig. 7f and Supplementary Fig. 8a, b). The interaction map indicated that all pairwise interactions involving srfR1, srfR5, srfR7, and srfR10 displayed positive synergy (Fig. 7f and Supplementary Fig. 8c). As expected, engineered strains carrying combinations of these synergistic targets showed improvement of surfactin titer to 878 mg/L in shake flask cultures, which was nearly 4 times higher than the strain 168S1. These results again evidenced the effectiveness of our rational multi-target combination strategy in enhancing NRP titers across diverse species.

Discussion

The excellent traits of organisms are the result of synergistic interactions among multiple genes³⁶, and similarly, the high-yield traits of bacteria are not solely determined by individual genes but rather the outcome of intricate interactions among multiple targets³⁷. Progress in strain engineering has demonstrated that most enriched mutations or recombination events synergistically increase titers by hundreds or thousands of times greater than that of the original isolated strain³⁸. One impressive example is the strain improvement program for penicillin production, where a series of successive screenings and enrichment of favorable mutations has led to a titer increase from 0.06 g/L to over 100 g/L over the past 80 years³⁹. However, this process involving generating millions of mutations, and testing each mutant strain to select phenotypes with increased production is labor-intensive and inefficient⁴⁰. To achieve accurate and efficient manipulation of multiple favorable targets, we developed a screening-based rational multi-target combination strategy and significantly improved the titers of daptomycin, thaxtomin A and surfactin in S. roseosporus, S. coelicolor and B. subtilis, respectively. Our work demonstrated that understanding the complex interactions is essential for advancing the knowledge of multi-target combination and then harnessing these synergistic effects to rapidly improve microbial cell factory performance.

Strain engineering often relied on combining multiple desired targets; however, not all combinations proved effective. For instance, both redox-response regulators RshA and CosR repressed nanobody VHH production in Corynebacterium glutamicum, but double mutant strains exhibited less titer increase compared to the individual mutants⁹. Similarly, another study identified 56 genes and corresponding sgRNAs beneficial for promoting free fatty acids (FFAs) production in Escherichia coli; however, when combined the high-yielding strain (ihfA^L−-aidB⁺) with each of the beneficial sgRNAs, only 6 combinations further improved FFAs production⁸. Therefore, previous multi-target combination approaches relied on subjective trial-and-error processes that were time-consuming and often led to inferior outcomes compared to single-target operations. In this work, we propose exploring the synergistic interaction of multiplex targets involved in NRP biosynthesis, and then rationally guiding the construction of high-yielding strains through multi-target combination. To map the interplay of multi-targets from the data of dual-target screening, we introduced q value used in drug synergy screening to evaluate the synergistic effects on the yield phenotype⁴¹. Based on the q value, we obtained the interplay map of all the targets. As we expected, combinations of targets with synergistic relationships effectively improved production yield while those with antagonistic relationships had no such effect (Fig. 6d and Supplementary Fig. 5c). Moreover, we demonstrated that our selected four-target combination indeed synergistically promoted daptomycin production at the transcriptional level (Fig. 6f, g). These data demonstrated effectiveness of synergistic screening in strain improvement efforts. Thus, mapping global regulator interactions to elucidate high-yielding phenotype will hold great potential for the construction of efficient microbial cell factory.

We focused on regulators of S. roseosporus as they play a crucial role in orchestrating the global metabolism and related physiological events through diverse interactions within the transcription regulatory network³⁰. Our screen aimed to identify potential repressors, considering that SM biosynthesis is triggered under specific environmental conditions and growth stages, involving a multitude of repressors in cross-regulation⁵. In this study, we identified 17 regulators that directly or indirectly inhibit daptomycin production. Previously, identifying repressors located outside of daptomycin BGC has been challenging due to methodological limitations, and thus far only four have been reported (DepR2⁴², PhaR⁴³, WblA³³, DasR³²), of which WblA was also included in our screening results. For thaxtomin A and surfactin, we screened 10 and 11 negative regulatory proteins, respectively (Fig. 7b, e). Our study highlights the existence of numerous unknown repressors involved in regulation of NRPs diverse species, which necessitates further exploration using our developed efficient identification strategy. Furthermore, we designed a workflow to investigate the interactions between these identified repressors in relation to NRP production, and rationally combined the synergistic targets. Therefore, the workflow of identifying regulators and exploring their interactions provides a feasible rational strategy for developing high-yield strains with multi-target modification.

Numerous studies have demonstrated that genes with similar expression profiles are often functionally related and vice versa⁴⁴. Especially, the production of SMs in Streptomyces involves in the synergistic effect of multifaced metabolic and physicological regulators²⁹. Our developed reporter system, characterized by the co-expression and co-biosynthesis of the reporter indigoidine and the desired NRPs, enables genome-wide screening for these regulatory targets. Since both the reporter compound and the desired products belongs to NRP type SMs and were biosynthesized by a multi-module enzyme assembly line, indigoidine BGC could be considered as an additional BGC module of daptomycin or thaxtomin A in our reporter system design (Fig. 2a). Such design ensured the identification of targets that not only contribute to the expression of daptomycin BGC and its related genes, but also the targets that commonly benefit for such NRP compound biosynthesis (Fig. 2b–d). Previously, a reporter-guided mutant selection (RGMS) method was initially used to improve lovastatin production in Aspergillus terreus⁴⁵ and subsequently improved through double-reporter-guided mutant selection to enhance clavulanic acid production in Streptomyces clavuligerus⁴⁶. However, these studies utilized reporters that were genes conferring antibiotic resistance or enzymes catalyzing color reactions rather than achieving co-transcription and co-biosynthesis of an analog reporter compound belonging to the same class. To the best of our knowledge, our findings provided unequivocally evidence for the first time that the expression level of BGCs can serve as a representative indicator for a group of functionally related genes that collectively coordinate biosynthesis (Fig. 2f, g). We believe that this screening strategy can also be applied to the high-yield phenotype screening of other colorless NRPs.

Expression of SM BGCs in heterologous hosts amenable to industrial-scale fermentation can lead to higher yield. By implementing heterologous production of herbicide thaxtomin A in S. coelicolor, we demonstrated the efficiency of this screening-based rational multi-target combination strategy: not only were we able to activate the silent thaxtomin A BGCs in S. coelicolor, but we also increased the titer of thaxtomin A to an exceptionally high level (Fig. 7e). Given that genetic background incompatibility between a heterologous host and desired BGC is a common issue²³, our screening-based rational multi-target combination strategy provides a feasible one-stop solution for activating cryptic NRP BGCs and subsequently improving their titers.

High-throughput CRISPR-based technologies facilitate rapid generation of extensive variant libraries, such as simultaneous inhibition or activation, or concurrent inhibition of one gene while activating another in pairwise fashion, thereby providing enhanced possibilities for synergistic screening⁴⁷. However, the application of CRISPR-mediated synergistic screening remains unexplored in strain improvement. In this study, we developed a synergistic CRISPRi library to simultaneously inhibit the expression of dual targets and investigated genetic interactions of identified regulators involved in inhibition of daptomycin and thaxtomin A production in different Streptomyces strains. Our findings highlight the potential utilization of widely-used CRISPR technology as a priority manipulation tool for collaborative multi-target research in Streptomyces species.

The strain Sr11, which exhibits synergistic effects on four targets, achieved daptomycin titers of 389 mg/L in shake flasks and to 1054 mg/L in a 7.5-L fermenter. These results surpass previous reported highest yields in shake flasks (350 μg/mL)⁴⁸ and fermenters (812 μg/mL)²⁵. Despite the significant increase in daptomycin yield, no metabolic burden was observed during growth (Fig. 6e). We hypothesize that this lack of burden could be due to the relatively low carbon yield of SMs, which typically accounts for less than 1% of total carbon flux. Specifically, Sr11 exhibited a carbon allocation of approximately 0.6%, while the WT strain showed about 0.2%. Given that SMs such as daptomycin are produced in much smaller quantities relative to bulk chemicals, our synergistic effect of the four targets combination may not have been sufficient to induce a noticeable growth burden. Additionally, the rich media used in fermentation could masked any potential metabolic burden, which might become more evident in selective media. To confirm this conjecture, we compared the extracellular by-products (Supplementary Fig. 5e) and intracellular concentrations of 14 amino acids (Supplementary Fig. 5f) between the final strain Sr11 and WT. The results showed no significant differences, suggesting that the strain’s robustness could be maintained to some extent.

Collectively, we developed a screening-based rational multi-target combination strategy for improving the titers of NRPs. The framework of this strategy encompasses the CRISPRi-mediated disturbance of individual target or pairwise combination, the analog co-expression and co-biosynthesis reporter system-based screening, and the mapping of cooperativity-based interaction relationships. By employing this strategy, we readily engineered multiplex targets with synergistic effect and achieved titers of daptomycin, thaxtomin A and surfactin to 1054 mg/L in 7.5-L fermenter, 352 mg/L and 878 mg/L in shake flask in S. roseosporus, S. coelicolor and B. subtilis, respectively. Our developed screening-based rational multi-target combination strategy and workflow hold potential for future strain engineering endeavors.

Methods

Strains, plasmids, primers, and growth conditions

Strains and plasmids used in this study are listed in Supplementary Table 3, and primers and sgRNAs are listed in Supplementary Table 4. S. roseosporus NRRL11379 (WT) and its derivatives were grown at 28 °C on solid DA1⁴⁹ for sporulation and phenotype observation. S. coelicolor M1152 and its derivatives were grown at 28 °C on solid MS⁵⁰ for sporulation. Solid RM14 medium was used for protoplast regeneration⁴⁹. For daptomycin titer, seed medium and fermentation medium⁴⁹ were used. For thaxtomin titer, liquid TSB and ISP4 medium were used²³. E. coli JM109, NEB 10-beta (Biomed; Beijing, China), and ET12567 (dam dcm hsdS)³⁰ were grown at 37 °C in Luria-Bertani (LB) medium and used for DNA cloning, library construction, and propagation of nonmethylated DNA for transformation into Streptomyces strains, respectively. B. subtilis 168 and its derivatives were grown at 37 °C on LB medium. Fermentation medium (6.5% sucrose, 0.8% NH₄NO₃, 0.03 M KH₂PO₄, 0.04 M Na₂HPO₄, 7 μM CaCl₂, 4 μM FeSO₄, 100 μM MgSO₄, 4 μM ethylene diamine tetra-acetic acid) was used for fermentation experiments.

Construction of S. roseosporus, S. coelicolor and B. subtilis mutant strains

To investigate idgS expression in S. roseosporus, plasmid pSET-dptEp-idgSsfp was constructed by ligation of dptEp (amplified from WT DNA with primers YH104A/YH104B) and idgS-sfp (amplified from pCIMt005⁵¹ with primers YH105A/YH105B) fragments into EcoRI-digested integrative plasmid pSET152⁵² using Gibson assembly method. Similarly, pSET-ermE*p-idgSsfp was constructed by ligation of ermE*p (amplified from pObldD³⁰ with primers YH106A/YH106B) and idgS-sfp (amplified with primers YH107A/YH107B) into pSET152. The two plasmids were then introduced into WT protoplasts, respectively, resulting in strains WT/pSET-dptEp-idgSsfp and WT/pSET-ermE*p-idgSsfp.

To construct reporter strain Srdi, two fragments flanking the dptD stop codon were amplified from WT DNA with primers YH108A/YH108B and YH110A/YH110B, and idgS-sfp fragment was amplified from pCIMt005 with primers YH109A/YH109B. dptD 5’-flanking region, idgS-sfp, and 3’-flanking region were assembled into EcoRI-digested pKC1139⁵² to generate pKC-Srdi, which was then transformed into WT protoplasts. Transformants on RM14 regeneration medium were transferred to DA1 plates for sporulation. Double-crossover mutant Srdi was screened by homologous double exchange, verified by PCR analysis with primers YH111A/YH111B and YH112A/YH112B, followed by DNA sequencing.

The mutant strains with premature introduction of a stop codon were constructed by CRISPR-cBEST system⁵³. To construct tetR6 or tetR7 deletion mutants, the corresponding sgRNA harboring the specific spacer was amplified from pKCCpf1⁵⁴ by primers YH131A/YH131B and YH136A/YH136B and then ligated into SpeI/NdeI-digested pKCCpf1. 5′- and 3′-flanking regions of tetR6 were amplified with primers YH129A/YH129B and YH130A/YH130B, and those of tetR7 were amplified with primers YH134A/YH134B and YH135A/YH135B, and then respectively ligated into HindIII-digested pKCCpf1 with corresponding sgRNA using Gibson assembly to generate pCpf1ΔtetR6 and pCpf1ΔtetR7, which were transformed into WT protoplasts. The double-crossover mutants ΔtetR6 and ΔtetR7 were selected and verified by PCR analysis using primer pairs YH132A/YH132B and YH133A/YH133B (for ΔtetR6), or YH137A/YH137B and YH138A/YH138B (for ΔtetR7), followed by DNA sequencing.

The reporter gene idgS-sfp was transferred into pSET152::txt²³ via λ-Red-mediated recombination²³, named pSET152::txt-idgSsfp. The plasmids pSET152::txt and pSET152::txt-idgSsfp were then introduced into S. coelicolor M1152 protoplasts, respectively, resulting in strains Scotxt and Scotxti.

To construct 168S1, two fragments about 1.1 kb upstream and downstream of the sfp (bsu_03570) gene was amplified from B. subtilis 168 with primers YH203A/YH203B and YH205A/YH205B, and sfp^MT45 fragment was amplified from B. amyloliquefaciens MT45 with primers YH204A/YH204B. Three fragments were assembled into EcoRI/ BamRI-digested pKSV7⁵⁵ to generate pKS-sfp^MT45, which was then transformed into B. subtilis 168 competent cells. The mutant 168S1 was obtained after two rounds of crossover and verified with primers YH206A/YH206B. To further construct reporter strain168S1R, two fragments flanking srfAD stop codon were amplified from WT DNA with primers YH207A/YH207B and YH209A/YH209B, and codon-optimized idgS-sfp fragment (codon optimization and gene synthesis by GENEWIZ Co., Ltd) with primers YH208A/YH208B. srfAD 5’-flanking region, idgS-sfp, and 3’-flanking region were assembled into EcoRI/ BamRI-digested pKSV7⁵⁵ to generate pKS-Bsudi, which was then transformed into 168S1 competent cells. Double-crossover mutant 168S1R was screened and verified by PCR analysis with primers YH210A/YH210B. B. subtilis 168 was transformed using the Spizizen method⁵⁶.

Fermentation and analysis of daptomycin titer

Spores from different S. roseosporus strains, grown on DA1 plates for 7 days, were transferred to 250 mL flasks containing 50 mL of primary seed medium and incubated at 28 °C for 48 h on a rotary shaker set at 250 rpm. The primary seed culture was then used to inoculate 50 mL of secondary seed medium at a 5% (vol/vol) inoculum, and the culture was further incubated at 28 °C with shaking for 30 h. Afterward, 5% (vol/vol) of the secondary seed culture was transferred into 50 mL of fermentation medium, and fermentation continued for 10 days. During the fermentation process, sodium decanoate was added every 12 h at a final concentration of 0.02% (wt/vol) after the first 48 hours, continuing until the fermentation was completed. The fermentation broth (1.0 mL) was centrifuged at 12,000 × g for 10 min. The resulting supernatant was filtered using a 2.2 μm membrane filter, and the filtrate was directly injected into an HPLC system (model 600; Waters, Milford, CT) equipped with a C18 column (4.6 mm × 250 mm). The mobile phase consisted of 0.1% trifluoroacetic acid in water and acetonitrile (55:45, vol/vol), with a flow rate of 1.0 mL/min. Daptomycin was detected by UV absorbance at 218 nm, and its concentration was quantified using standard curves generated with authentic daptomycin samples (Shanghai Qiao Chemical Science, China).

Fermentation in a 7.5-L fermenter (New Brunswick, BioFlo®/CelliGen®115) was performed at 28 °C with an initial culture volume of 3.0 L. Inoculation was 5% (v/v) and the aeration rate was controlled at 3.0 vvm. Dissolved oxygen (DO) was maintained above 20% of the saturated dissolved oxygen concentration by adjusting the agitation speed above 200 rpm. Aqueous solution containing 2% (w/v) sodium decanoate precursor was added at the rate of 40 μL/min after 48 h, and dextrin was supplemented at a flow rate of 25 mg/min during day 3 (when reducing sugar and total sugar contents dropped to 3 g/L and 20 g/L) to day 7. During the fermentation process, biomass was determined by dry cell weight from 80 °C for 24 h, and pH value was maintained at 6.5 with 25% NH₄OH.

Fermentation and analysis of thaxtomin A titer

For thaxtomins production, S. coelicolor spores were inoculated in liquid TSB and incubated for 48 h as a seed culture, and then 2.5 mL of each seed culture was transferred into a shake flask containing 50 mL ISP4 with 1% cellobiose²³. HPLC analysis of thaxtomin A titer was performed with a SGIMADZU LC-20A HPLC system and a ZORBAX SB-C18 column (5 μm pore size; 4.6 by 250 mm).

Fermentation and analysis of surfactin titer

For surfactin production, B. subtilis 168 and its derivatives were precultured in LB medium and then incubated at 37 °C with shaking at 200 rpm for 16 h as seed cultures. Five milliliters of seed culture was inoculated into 50 mL of fermentation medium in a 250 mL flask and cultured at 37 °C with shaking at 200 rpm for 48 h. To determine surfactin, the cell-free culture broth was adjusted to pH 2.0 with 6 M HCl, then centrifuged at 10,000 × g for 30 min. The supernatant was discarded, and the pellet was re-dissolved in 100% methanol to obtain crude surfactin. The surfactin quantification was carried out using a Waters UPLC H-class system, which included a binary solvent delivery system and an auto-sampler. Chromatographic separation was achieved on a Waters Acquity C18 column (50 mm × 2.1 mm, 1.7 μm particle size), with a 5 μL injection volume. The mobile phase consisted of solvent A (HPLC-grade water with 0.1% formic acid) and solvent B (HPLC-grade methanol). A linear gradient elution from 85% to 100% solvent B over 6 min was applied at a flow rate of 0.3 mL/min. The elution profile was monitored by measuring absorbance at 215 nm. The surfactin concentration was calculated and quantified using Empower Version 3.0 software⁵⁷. The total surfactin concentration was the sum of the concentrations of four homologs. Surfactin standards were purchased from Sigma (St. Louis, MO, USA). All experiments were conducted in triplicate.

GC-MS analysis

GC-MS was used to identified relative level of intracellular amino acids analysis. The GC–MS system (Agilent) employed for the analysis consisted of a 7890 GC, a 5975 MS, and a 7683 auto-sampler, with a HP-5 MS capillary column (30 m × 250 μm × 0.25 μm, J&W Scientific). For metabolomic profiling, 1 μL of the sample was injected with a 1:10 split ratio. The temperature program started at 70 °C for 1 minute, then ramped at 5 °C/min to 230 °C, followed by an increase of 10 °C/min to 290 °C, and was held at 290 °C for 6 min. Ions were produced using the 70 eV electron beam at a 40 μA ionization current, and the mass spectrum was recorded in the range of 50–650 m/z. For amino acid analysis, 1 μL of the sample was injected with a 1:30 split ratio. The chromatographic conditions were as follows: 70 °C for 2 minutes, then a temperature increase to 290 °C at 5 °C/min, which was held for 3 minutes. Helium was used as the carrier gas at a flow rate of 1.0 mL/min. Both the injector and detector temperatures were maintained at 290 °C.

Next-generation sequencing (NGS)

The DNA samples were amplified using primers YH139A/YH139B with plasmid library and transformants library as templates respectively. The NGS was carried out and analyzed by GENEWIZ Co., Ltd. on illumina Miseq.

Determination of sugar content

The concentration of reducing sugars and total sugars (acid hydrolysis) was determined by the 3,5-dinitrosalicylic acid (DNS) method⁵⁸.

Real-time RT-PCR (qRT-PCR) assay

S. roseosporus cells grown in liquid seed or fermentation medium and S. coelicolor cells grown in ISP4 liquid medium for 2 days were harvested for RNA extraction using the TRIzol method. Each RNA sample (4 μg) was treated with DNase I (TaKaRa; Japan) to digest genomic DNA, and then subjected to reverse transcription with M-MLV reverse transcriptase (Promega; USA). The cDNA obtained was used as template for qRT-PCR analysis (using primers listed in Supplementary Table 4). Relative expression level of tested genes was normalized with housekeeping gene 16S rRNA as the internal control using comparative Ct method. Expression value for each gene was determined in triplicate.

Quantitative assay of indigoidine production

S. roseosporus and S. coelicolor strains carrying idgS-sfp genes were cultured in liquid seed medium and ISP4 liquid medium for 3 days, and indigoidine production was measured by recording OD₆₀₀ of the culture supernatants using multifunctional plate reader (Molecular Devices; USA). The supernatant of WT or plasmid control strain was used as a blank control.

Gene inhibition and base editing in S. roseosporus

Plasmid pSETdCas9Rg-2 for CRISPRi was constructed based on pSET-dCas9⁵⁹ and pKCcas9dO⁵⁹. sgRNA expression element (J23119p-sgRNA-terminator) was amplified from pKCcas9dO with primers YH201A/YH201B and then assembled into EcoRI-digested pSET-dCas9 by Gibson assembly to generate pSETdCas9Rg. pSETdCas9Rg-2 backbone was amplified from pSETdCas9Rg with primers YH202A/YH202B to introduce two BsaI cleavage sites between J23119p and sgRNA core scaffold. For gene inhibition, the oligo 5´-TAGT(N20)G-3´ and the reverse complement oligo 5´-AAAAC(N20)-3´were synthesized, annealed into double strands, and then ligated into BsaI-digested pSETdCas9Rg-2 to generate CRISPRi plasmids.

CRISPR-cBEST plasmids for base editing were constructed based on pCRISPR-cBEST (Addgene #125689)⁵³. The 20-nt spacer (N20) of sgRNAs was designed by software CRISPy-web (https://crispy.secondarymetabolites.org)³⁴. For base editing, 5´-CGGTTGGTAGGATCGACGGC(N20)GTTTTAGAGCTAGAAATAG-3´ and the reverse complement strand were synthesized, annealed, and assembled into NcoI-digested pCRISPR-cBEST by Gibson assembly to generate CRISPR-cBEST plasmids. The demethylated CRISPRi and CRISPR-cBEST plasmids from E. coli ET12567 were transformed into S. roseosporus protoplasts. 5–8 transformants were selected for sequencing to screen the correct constructs.

To obtain plasmid-free strains, the base-edited S. roseosporus strains containing CRISPR-cBEST plasmid were cultured at 37 °C on DA1 plates for plasmid curing. After incubation for 5 days, spores were spread on DA1 plates without apramycin and cultured at 28 °C for 5–7 days. Then, 15–20 colonies were separately inoculated onto apramycin-containing DA1 plates and those could not grow were plasmid-free.

Construction and analysis of the CRISPRi library and synergistic library

The sgRNA pool was used for targeted inhibition of 610 predicted regulatory genes in S. roseosporus and 856 regulator genes in S. coelicolor genome. All sgRNAs were designed to target the NT strand and 3 target sites were selected for each gene. The sgRNAs were synthesized as 108-nt oligonucleotides (ATCGGATCCTTGACAGCTAGCTCAGTCCTAGGTATAATACTAGT(N20)TTTTAGAGCTAGAAATAGCAAGTTAAAATAAGGCTAGTCCGTTA), each consisting of the 20-nt sgRNA spacer sequence and 44-nt flanking region homolog to pSETdCas9Rg-2 backbone in S. roseosporus and pSETdCas9Rg-3 backbone in S. coelicolor. The sgRNA pool was designed and synthesized in batches by GenScript (Nanjing, China). Oligonucleotides were amplified by PCR with primers YH139A/YH139B. To minimize bias in the library, 15 cycles of amplification was used. The pSETdCas9Rg-2 backbone was amplified by PCR and then was digested with DpnI. Amplified sgRNA oligonucleotides were assembled into pSETdCas9Rg-2 or pSETdCas9Rg-3 backbone by Gibson assembly to construct CRISPRi plasmid library. To improve the coverage of the plasmid library, the library was independently constructed 20 times. The assembled product was sequentially transformed into E. coli NEB 10-beta, ET12567 and S. roseosporus Srdi or Scotxti. The resulting transformants were subjected to visual screening and sequencing analysis to determine the targeted genes.

For constructing the synergistic library, the CRISPRi plasmids of screened genes that inhibit daptomycin or thaxtomin production were separately constructed. The obtained CRISPRi plasmids and control plasmid pSETdCas9Rg-2/ pSETdCas9Rg-3 generated SlA library. The plasmid backbones and sgRNAs were both amplified from SlA library, and sgRNAs were then assembled into the other site of the SlA plasmid backbones i.e., between the resistance gene and the origin of replication, to generate synergistic library SlB (Fig. 5a). In addition to control plasmid, each SlB plasmid contained 1 or 2 sgRNAs. The SlB library was transformed into Srdi for synergistic screening based on indigoidine production. The SlA library was also transformed into reporter strain for calculating the synergy coefficient.

The CRISPRi library and synergistic library were constructed based plasmid pLCg6-dcpf1⁶⁰ (dCpf1 expression plasmid) and pcra3⁶⁰ (sgRNA expression plasmid) in B. subtilis 168. The yield of surfactin in 168S1R^- (with empty plasmid pLCg6-dcpf1 and pcra3) was comparable to that of 168S1R. The pLCg6-dcpf1 was first transformed into 168S1R and inserted into lacA gene. To construct the CRISPRi library, all sgRNAs were designed to target the NT strand and 3 target sites were selected for each gene. The sgRNAs were synthesized as 27 oligonucleotides AGAT(N23) and AATT(N23(Reverse complementary sequence)). The oligonucleotides were annealed and ligated into Eco31I-digested pcra3. The synergistic library and multi-gene targeting plasmids were constructed by golden gate assembly⁶⁰. All 66 combinations in the synergistic library (55 dual-target combinations and 11 single-targeting) were independently constructed.

Quantitative measurement of fluorescence intensity

S. roseosporus strains carrying corresponding plasmids were cultured in liquid seed medium. 10 mL culture was centrifuged, and resuspended in 1 mL Tris-HCl (20 mM, pH 7.5). The cell suspension was sonicated on ice for 5 min (pulse 3 s, stop 5 s), and the protein concentration in the supernatant was quantified with Bradford protein stain (Bio-rad, Quick Start™ Bradford 1× Dye Reagent). 200 μL protein solution with the same OD₅₉₅ was used for fluorescence measurement by SpectraMax M3 microplate reader (Molecular Devices; USA). GFP abundance was quantified at an excitation wavelength of 488 ± 10 nm and an emission wavelength of 511 ± 10 nm. mCherry abundance was quantified at an excitation wavelength of 588 ± 10 nm and an emission wavelength of 644 ± 10 nm.

Fluorescence microscopy

5 μL of 3-day S. roseosporus culture in liquid seed medium was spotted onto a glass slide and a coverslip was put on the top. Mycelia were observed by confocal laser scanning microscopy (GFP excitation at 488 nm; mCherry excitation at 588 nm) (model TCS SP8, Leica Microsystems; Germany) with an oil-immersion objective.

Synergistic effect

The indigoidine level of each strain on day 3 was determined by multifunctional plate reader, with WT^- as zero and Srdi^- as blank control.

Synergistic effect was calculated by the Jin’s formula:

$$q=\frac{E(A+B)}{{EA}+{EB}-E{{{\rm{control}}}}}$$

Where E(A + B) and EA, EB are the average effects of simultaneous targeting of two genes vs. targeting two genes individually. In this method, q < 0.85 indicates antagonism, 0.85 ≤ q < 1.15 indicates additivity, and q ≥ 1.15 indicates synergism³⁵.

Statistics and reproducibility

Statistical analyses and data graphing were performed using GraphPad Prism (version 8.4.0) and Microsoft Excel (version 16.45). All statistical analyses performed reflect comparisons between distinct samples, rather than repeated measures. Statistical tests used in each experiment are indicated in their respective figure legends. Data are presented as means ± standard deviation (SD), and statistical significance between groups is assessed by p-values, denoted by asterisks (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, and n.s. no significant difference). No statistical method was used to predetermine sample sizes.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data associated with this study are present in the paper or the Supplementary Information. The NGS for library coverage generated in this study have been deposited in the NCBI Sequence Read Archive database under accession code PRJNA1220263. Source data are provided with this paper.

References

Li, J. W. & Vederas, J. C. Drug discovery and natural products: end of an era or an endless frontier? Science 325, 161–165 (2009).
Article ADS PubMed MATH Google Scholar
Harvey, A. L., Edrada-Ebel, R. & Quinn, R. J. The re-emergence of natural products for drug discovery in the genomics era. Nat. Rev. Drug Discov. 14, 111–129 (2015).
Article CAS PubMed MATH Google Scholar
Lacey, H. J. & Rutledge, P. J. Recently discovered secondary metabolites from Streptomyces species. Molecules 27, 887 (2022).
Urem, M., Swiatek-Polatynska, M. A., Rigali, S. & van Wezel, G. P. Intertwining nutrient-sensory networks and the control of antibiotic production in Streptomyces. Mol. Microbiol. 102, 183–195 (2016).
Article CAS PubMed Google Scholar
Liu, G., Chater, K. F., Chandra, G., Niu, G. & Tan, H. Molecular regulation of antibiotic biosynthesis in Streptomyces. Microbiol. Mol. Biol. Rev. 77, 112–143 (2013).
Article CAS PubMed PubMed Central Google Scholar
Cho, J. S., Kim, G. B., Eun, H., Moon, C. W. & Lee, S. Y. Designing microbial cell factories for the production of chemicals. JACS Au 2, 1781–1799 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Volk, M. J. et al. Metabolic engineering: methodologies and applications. Chem. Rev. 123, 5521–5570 (2023).
Fang, L. et al. Genome-scale target identification in Escherichia coli for high-titer production of free fatty acids. Nat. Commun. 12, 4976 (2021).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Yu, X. et al. CRISPRi-microfluidics screening enables genome-scale target identification for high-titer protein production and secretion. Metab. Eng. 75, 192–204 (2023).
Article CAS PubMed MATH Google Scholar
Esvelt, K. M. & Wang, H. H. Genome-scale engineering for systems and synthetic biology. Mol. Syst. Biol. 9, 641 (2013).
Article PubMed PubMed Central MATH Google Scholar
Si, T., Xiao, H. & Zhao, H. Rapid prototyping of microbial cell factories via genome-scale engineering. Biotechnol. Adv. 33, 1420–1432 (2015).
Article CAS PubMed MATH Google Scholar
Hanna, R. E. & Doench, J. G. Design and analysis of CRISPR-Cas experiments. Nat. Biotechnol. 38, 813–823 (2020).
Article CAS PubMed MATH Google Scholar
Klann, T. S. et al. CRISPR-Cas9 epigenome editing enables high-throughput screening for functional regulatory elements in the human genome. Nat. Biotechnol. 35, 561–568 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Schuster, A. et al. RNAi/CRISPR screens: from a pool to a valid hit. Trends Biotechnol. 37, 38–55 (2019).
Article CAS PubMed MATH Google Scholar
Sarnaik, A., Liu, A., Nielsen, D. & Varman, A. M. High-throughput screening for efficient microbial biotechnology. Curr. Opin. Biotechnol. 64, 141–150 (2020).
Article CAS PubMed Google Scholar
Tucker, A. T. et al. Discovery of next-generation antimicrobials through bacterial self-screening of surface-displayed peptide libraries. Cell 172, 618–628.e613 (2018).
Article CAS PubMed PubMed Central MATH Google Scholar
Lee, S. W. & Oh, M. K. A synthetic suicide riboswitch for the high-throughput screening of metabolite production in Saccharomyces cerevisiae. Metab. Eng. 28, 143–150 (2015).
Article CAS PubMed MATH Google Scholar
Ruiz, B. et al. Production of microbial secondary metabolites: regulation by the carbon source. Crit. Rev. Microbiol. 36, 146–167 (2010).
Article CAS PubMed MATH Google Scholar
Mani, R. et al. Defining genetic interaction. Proc. Natl Acad. Sci. USA 105, 3461–3466 (2008).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Boone, C., Bussey, H. & Andrews, B. J. Exploring genetic interactions and networks with yeast. Nat. Rev. Genet. 8, 437–449 (2007).
Article CAS PubMed MATH Google Scholar
Cao, J. Y. et al. Prognostic significance and therapeutic implications of centromere protein F expression in human nasopharyngeal carcinoma. Mol. Cancer 9, 237 (2010).
Article PubMed PubMed Central MATH Google Scholar
Gaudelli, N. M., Long, D. H. & Townsend, C. A. β-Lactam formation by a non-ribosomal peptide synthetase during antibiotic biosynthesis. Nature 520, 383–387 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Li, Z. et al. Stepwise increase of thaxtomins production in Streptomyces albidoflavus J1074 through combinatorial metabolic engineering. Metab. Eng. 68, 187–198 (2021).
Article CAS PubMed MATH Google Scholar
Koglin, A. et al. Structural basis for the selectivity of the external thioesterase of the surfactin synthetase. Nature 454, 907–911 (2008).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Ng, I. S., Ye, C., Zhang, Z., Lu, Y. & Jing, K. Daptomycin antibiotic production processes in fed-batch fermentation by Streptomyces roseosporus NRRL11379 with precursor effect and medium optimization. Bioprocess Biosyst. Eng. 37, 415–423 (2014).
Article CAS PubMed Google Scholar
Zhao, X., Zong, Y., Wei, W. & Lou, C. Multiplexed promoter engineering for improving thaxtomin A production in heterologous Streptomyces hosts. Life 12, 689 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Wang, X. R. et al. Elucidation of genes enhancing natural product biosynthesis through co-evolution analysis. Nat. Metab. 6, 933–946 (2024).
Article CAS PubMed MATH Google Scholar
Kirby, R. Chromosome diversity and similarity within the Actinomycetales. FEMS Microbiol. Lett. 319, 1–10 (2011).
Article CAS PubMed MATH Google Scholar
Yan, H., Li, S. & Wang, W. Reprogramming naturally evolved switches for Streptomyces chassis development. Trends Biotechnol. 43, 12–15 (2024).
Yan, H. et al. BldD, a master developmental repressor, activates antibiotic production in two Streptomyces species. Mol. Microbiol. 113, 123–142 (2020).
Article CAS PubMed MATH Google Scholar
Mao, X. M. et al. Transcriptional regulation of the daptomycin gene cluster in Streptomyces roseosporus by an autoregulator. AtrA. J. Biol. Chem. 290, 7992–8001 (2015).
Article CAS PubMed MATH Google Scholar
Chen, Q., Zhu, J., Li, X. & Wen, Y. Transcriptional regulator DasR represses daptomycin production through both direct and cascade mechanisms in Streptomyces roseosporus. Antibiotics 11, 1065 (2022).
Huang, X. et al. wblA, a pleiotropic regulatory gene modulating morphogenesis and daptomycin production in Streptomyces roseosporus. J. Appl. Microbiol. 123, 669–677 (2017).
Article CAS PubMed MATH Google Scholar
Tong, Y. et al. CRISPR-Cas9, CRISPRi and CRISPR-BEST-mediated genetic manipulation in streptomycetes. Nat. Protoc. 15, 2470–2502 (2020).
Article CAS PubMed MATH Google Scholar
Jin, Z. J. About the evaluation of drug combination. Acta Pharmacol. Sin. 25, 146–147 (2004).
CAS PubMed MATH Google Scholar
Pérez-Pérez, J. M., Candela, H. & Micol, J. L. Understanding synergy in genetic interactions. Trends Genet. 25, 368–376 (2009).
Article PubMed MATH Google Scholar
Cai, P. et al. Recombination machinery engineering facilitates metabolic engineering of the industrial yeast Pichia pastoris. Nucleic Acids Res. 49, 7791–7805 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Adrio, J. L. & Demain, A. L. Genetic improvement of processes yielding microbial products. FEMS Microbiol. Rev. 30, 187–214 (2006).
Article CAS PubMed Google Scholar
Barreiro, C. & Garcia-Estrada, C. Proteomics and Penicillium chrysogenum: Unveiling the secrets behind penicillin production. J. Proteom. 198, 119–131 (2019).
Article CAS MATH Google Scholar
Zeng, W., Guo, L., Xu, S., Chen, J. & Zhou, J. High-throughput screening technology in industrial biotechnology. Trends Biotechnol. 38, 888–906 (2020).
Article CAS PubMed MATH Google Scholar
Han, K. et al. Synergistic drug combinations for cancer identified in a CRISPR screen for pairwise genetic interactions. Nat. Biotechnol. 35, 463–474 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Mao, X. M., Luo, S. & Li, Y. Q. Negative regulation of daptomycin production by DepR2, an ArsR-family transcriptional factor. J. Ind. Microbiol. Biotechnol. 44, 1653–1658 (2017).
Article CAS PubMed Google Scholar
Luo, S., Chen, X. A., Mao, X. M. & Li, Y. Q. Transposon-based identification of a negative regulator for the antibiotic hyper-production in Streptomyces. Appl. Microbiol. Biotechnol. 102, 6581–6592 (2018).
Article CAS PubMed MATH Google Scholar
Sastry, A. V. et al. The Escherichia coli transcriptome mostly consists of independently regulated modules. Nat. Commun. 10, 5536 (2019).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Askenazi, M. et al. Integrating transcriptional and metabolite profiles to direct the engineering of lovastatin-producing fungal strains. Nat. Biotechnol. 21, 150–156 (2003).
Article CAS PubMed MATH Google Scholar
Xiang, S. H. et al. Application of a double-reporter-guided mutant selection method to improve clavulanic acid production in Streptomyces clavuligerus. Metab. Eng. 11, 310–318 (2009).
Article CAS PubMed MATH Google Scholar
Wang, J. Y. & Doudna, J. A. CRISPR technology: A decade of genome editing is only the beginning. Science 379, 251 (2023).
Article MATH Google Scholar
Li, X., Sang, Z., Zhao, X. & Wen, Y. Metabolic engineering of Streptomyces roseosporus for increased production of clinically important antibiotic daptomycin. Microb. Biotechnol. 17, e70038 (2024).
Article CAS PubMed PubMed Central MATH Google Scholar
Zhang, Q. et al. A MarR family transcriptional regulator, DptR3, activates daptomycin biosynthesis and morphological differentiation in Streptomyces roseosporus. Appl. Environ. Microbiol. 81, 3753–3765 (2015).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Gomez-Escribano, J. P. & Bibb, M. J. Engineering Streptomyces coelicolor for heterologous expression of secondary metabolite gene clusters. Microb. Biotechnol. 4, 207–215 (2011).
Article CAS PubMed PubMed Central Google Scholar
Li, P. et al. An efficient blue-white screening based gene inactivation system for Streptomyces. Appl. Microbiol. Biotechnol. 99, 1923–1933 (2015).
Article CAS PubMed MATH Google Scholar
Bierman, M. et al. Plasmid cloning vectors for the conjugal transfer of DNA from Escherichia coli to Streptomyces spp. Gene 116, 43–49 (1992).
Article CAS PubMed MATH Google Scholar
Tong, Y. et al. Highly efficient DSB-free base editing for streptomycetes with CRISPR-BEST. Proc. Natl Acad. Sci. USA 116, 20366–20375 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Li, L. et al. CRISPR-Cpf1-assisted multiplex genome editing and transcriptional repression in Streptomyces. Appl. Environ. Microbiol. 84, e00827–18 (2018).
Smith, K. & Youngman, P. Use of a new integrational vector to investigate compartment-specific expression of the Bacillus subtilis spoIIM gene. Biochimie 74, 705–711 (1992).
Article CAS PubMed Google Scholar
Spizizen, J. Transformation of biochemically deficient strains of Bacillus subtilis by deoxyribonucleate. Proc. Natl Acad. Sci. USA 44, 1072–1078 (1958).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Wu, Q., Zhi, Y. & Xu, Y. Systematically engineering the biosynthesis of a green biosurfactant surfactin by Bacillus subtilis 168. Metab. Eng. 52, 87–97 (2019).
Article CAS PubMed MATH Google Scholar
Miller, G. L. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31, 426–428 (1959).
Article CAS MATH Google Scholar
Zhao, Y. et al. CRISPR/dCas9-mediated multiplex gene repression in Streptomyces. Biotechnol. J. 13, e1800121 (2018).
Article PubMed Google Scholar
Wu, Y. et al. CAMERS-B: CRISPR/Cpf1 assisted multiple-genes editing and regulation system for Bacillus subtilis. Biotechnol. Bioeng. 117, 1817–1825 (2020).
Article CAS PubMed MATH Google Scholar

Download references

Acknowledgements

This study was supported by the National Natural Science Foundation of China (Grant No. 32170081 to Y.W., 32100066 to H.Y., and 32170095 to W.W.) and the National Key Research and Development Program of China (Grant No. 2021YFC2100602 to W.W.).

Author information

Authors and Affiliations

State Key Laboratory of Animal Biotech Breeding and College of Biological Sciences, China Agricultural University, Beijing, China
Hao Yan, Ziwei Sang, Xingwang Li, Jia Xie, Jiale Wu & Ying Wen
State Key Laboratory of Microbial Resources, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
Hao Yan, Zhenguo Xin, Ziwei Sang, Shen Pang & Weishan Wang
Beijing Key Laboratory of Genetic Element Biosourcing & Intelligent Design for Biomanufacturing, Beijing, China
Hao Yan & Weishan Wang
University of Chinese Academy of Sciences, Beijing, China
Weishan Wang

Authors

Hao Yan
View author publications
Search author on:PubMed Google Scholar
Zhenguo Xin
View author publications
Search author on:PubMed Google Scholar
Ziwei Sang
View author publications
Search author on:PubMed Google Scholar
Xingwang Li
View author publications
Search author on:PubMed Google Scholar
Jia Xie
View author publications
Search author on:PubMed Google Scholar
Jiale Wu
View author publications
Search author on:PubMed Google Scholar
Shen Pang
View author publications
Search author on:PubMed Google Scholar
Ying Wen
View author publications
Search author on:PubMed Google Scholar
Weishan Wang
View author publications
Search author on:PubMed Google Scholar

Contributions

W.W., Y.W., and H.Y. designed the research. H.Y. conducted most of the experiments. Z.S., Z.X., X.L., J.X., J.W. and S.P. conducted part of the experiments. W.W., W.W., and H.Y. contributed reagents/materials/analysis tools. W.W., Y.W., and H.Y. wrote the paper with contributions from all other authors.

Corresponding authors

Correspondence to Ying Wen or Weishan Wang.

Ethics declarations

Competing interests

W.W., Y.W., H.Y., Z.X. and Z.S. have filed provisional patents for this work to the China National Intellectual Property Administration (CNIPA). The patent application numbers are CN202210143368.8, CN202410571683.X, CN202510175729.0 and CN202510175726.7. The remaining authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers 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.

Supplementary information

Supplementary Information

Reporting Summary

Transparent Peer Review file

Source data

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/.

Reprints and permissions

About this article

Cite this article

Yan, H., Xin, Z., Sang, Z. et al. A rational multi-target combination strategy for synergistic improvement of non-ribosomal peptide production. Nat Commun 16, 1883 (2025). https://doi.org/10.1038/s41467-025-57073-5

Download citation

Received: 24 September 2024
Accepted: 07 February 2025
Published: 22 February 2025
DOI: https://doi.org/10.1038/s41467-025-57073-5

This article is cited by

Scalable secondary metabolite production in Streptomyces using a plug-and-play system
- Bowen Yang
- Zilong Li
- Weishan Wang
Nature Biotechnology (2025)