Overproduction of endusamycin in Streptomyces endus subsp. aureus

Abstract

Endusamycin, a naturally occurring polyether ionophore antibiotic, exhibits extensive antitumor activities. Despite its promising potential, the titer of endusamycin is significantly lower compared to widely used polyether compounds, and no reports have been published regarding its overproduction. In this study, various metabolic engineering strategies were performed to enhance endusamycin production. Notably, the deletion of competing biosynthetic gene clusters (BGCs) responsible for the biosynthesis of spore pigment and meilingmycin-like compounds based on transcriptome analysis, as well as the doubling of the endusamycin BGC, proved to be effective. These interventions resulted in a 20 % and 69 % increase in the titer of endusamycin, respectively. Furthermore, systematic optimization of fermentation medium components, including carbon source, nitrogen source, phosphorus and potassium, contributed to a further 69 % increase in the titer of endusamycin. Ultimately, the high-yielding strain YC1109 was developed through the integration of these strategies. The titer of endusamycin reached 5469 mg/L in shake-flask fermentation and 5011 mg/L in fed-batch fermentation, representing a 246 % increase compared to the original strain. This research significantly facilitates the drug development and industrialization of endusamycin. It establishes a superior chassis strain for exploring endusamycin derivatives and provides valuable insights into improving the production of polyether compounds.

1. Introduction

Natural polyether ionophore antibiotics, a subclass of polyketides, exhibit a wide range of biological activities due to their structural diversity, including antibacterial, antiparasitic, antiviral, and antitumor properties [[1], [2], [3]]. Several polyether antibiotics, such as monensin, salinomycin, narasin, maduramicin and lasalocid, have been widely used in veterinary medicine and animal husbandry, recognized as high-value natural products [4].

Endusamycin, a glycosyl polyether compound structurally similar to nanchangmycin, is produced by Streptomyces endus subsp. aureus and was first identified in 1988 [5]. However, its bioactivity only began to attract significant attention in 2018. Endusamycin and other nanchangmycin analogs exhibit potent inhibitory activity against 39 different types of cancer cells, primarily targeting the Wnt/β-catenin signaling pathway. Notably, endusamycin demonstrates greater efficacy against breast cancer stem cells compared to salinomycin [6]. Building on the endusamycin biosynthetic pathway, a novel glycosyl polyether molecule, End-16, was successfully created by combining biosynthesis strategies. End-16 exhibits superior efficacy and safety in treating bladder cancer compared to cisplatin, a frontline clinical drug for this condition [7]. Endusamycin and its derivative End-16 show promising potential as new antitumor drug candidates.

The biosynthetic gene cluster and biosynthetic pathway of endusamycin have been predicted, and the functions of its post-modification genes have been partially characterized [7] (Fig. 1). The endusamycin gene cluster comprises of 34 open reading frames (ORFs), with 11 PKS subunits (endA1-A11) encoding a single loading module and 14 extension modules, embodying the archetypal structure of Type I PKS. These subunits utilize acetyl-ACP as the starter unit and sequentially incorporate 4 malonyl-CoA and 10 methylmalonyl-CoA molecules to produce the polyether skeleton. The synthesis of endusamycin was ultimately completed through the alternating catalysis of post-PKS tailoring enzymes.

Fig. 1.

Genetic organization of the biosynthetic gene cluster (A) and the proposed biosynthetic pathway (B) of endusamycin in S. endus subsp. aureus.

Compared to the titers of other polyether antibiotics, such as salinomycin (90 g/L in fed-batch fermentation [8]) and maduramicin (7.16 g/L in shake flasks fermentation [9]), which are widely used as anti-coccidiostats in animal husbandry, the titer of endusamycin remains unsatisfactory. This significantly impedes progress in product preparation and drug development. Moreover, due to the lack of attention, no efforts have been made to improve endusamycin production.

The overproduction of other polyether compounds, including monensin, salinomycin and maduramicin, has been studied for a long time. Various metabolic engineering strategies, such as precursor engineering, transcriptional regulation and heterologous expression, have been successfully applied in this endeavor [[8], [9], [10], [11], [12], [13], [14], [15], [16]]. Among these strategies, precursor engineering, particularly the deletion of precursor-competing BGCs, has emerged as a preferred approach. This not only minimizes precursor loss but also reduces the interference caused by by-products [8]. That has also been employed to develop a wide range of efficient chassis strains optimized for heterologous expression [[17], [18], [19]]. However, other classic metabolic engineering strategies, such as overexpression of genes involved in rate-limiting steps, enhancement of cofactor supply, improvement of product tolerance, and doubling of BGCs, have rarely been explored for achieving overproduction of polyether compounds [20]. The doubling of BGCs can effectively circumvent metabolic imbalances caused by the differential expression of specific enzymes [21,22]. This approach has yielded remarkable results, with the titers of daptomycin [23], tautomycetin [24], 5-oxomilbemycin [25], actinorhodin [26] and acarbose [27] increased by 40 %, 14-fold, 2.9-fold, 20-fold, and 35 %, respectively. Furthermore, optimization of fermentation medium is undeniably significant, as it enhances the nutritional balance required for the synthesis of secondary metabolites, thereby fully exploiting the talent of strains. Key factors in this process include carbon sources, nitrogen sources, inorganic salts, trace elements, as well as fermentation conditions such as temperature, pH, dissolved oxygen, etc [28,29].

In this study, it is the first attempt to enhance endusamycin production. A range of metabolic engineering and fermentation medium optimization strategies were evaluated, including deletion of competing BGCs to reduce precursor losses, doubling of the endogenous endusamycin BGC to enhance product synthesis efficiency, optimization of the fermentation medium's composition and content and supplementation with the cofactor pyrroloquinoline quinone (PQQ) to stimulate primary metabolism. The integration of effective strategies led to a significant, multiple-fold increase in the titer of endusamycin. The high-yielding strain demonstrated excellent performance in both shake-flask fermentations and fed-batch fermentations.

2. Materials and methods

2.1. Strains, plasmids, primers and culture conditions

The strains, plasmids, and primers used in this study are shown in Tables S1 and S2. E. coli DH10B was used as a host for cloning. E. coli ET12567/pUZ8002 [30] was used for intergeneric conjugation between E. coli and S. endus subsp. aureus. Streptomyces and its derivatives were cultivated on soybean flour-mannitol (SFM) agar medium (2 % (w/v) soybean flour, 2 % (w/v) mannitol and 2 % (w/v) agar) at 30 °C for strain culturing and isolation. E. coli was cultured in LB medium at 37 °C with the appropriate antibiotic for selection.

2.2. Sampling and preparation of samples for transcriptomics analysis

Cells of S. endus subsp. aureus in fermentation medium were collected after incubation for 48 h. Three independent replicates were performed. The RNeasy® Plant Mini Kit (Qiagen) was employed according to the manufacturer's instructions for the extraction of total RNA. The total RNA sequencing was performed utilizing the NovaSeq 6000 platform (Illumina) by Shanghai Personal Biotechnology Co. Ltd. The gene read count value was counted using HTSeq (v0.9.1) as the original expression level of the gene. FPKM (fragments per kilobase of transcript per million fragments mapped) was used to normalize the expression.

2.3. Construction of gene knockout strains

pYH7 was used as the vector for gene knockout. The upstream and downstream homologous fragments of the knockout segment were amplified from the genome of S. endus subsp. aureus with the primer pairs pAN002-LHA-GF/GR and pAN002-RHA-GF/GR, respectively. Then, the two purified fragments were assembled with the NdeI/HindIII digested linear fragments of pYH7 using the Gibson assembly method to generate pAN002. Similarly, the upstream and downstream homologous fragments of the knockout segment were amplified from the genome of S. endus subsp. aureus using the primer pairs pAN008-LHA-GF/GR and pAN008-RHA-GF/GR, respectively. Subsequently, the two fragments were cloned into the pYH7 vector to generate pAN008.

After plasmid sequencing verification, the constructed plasmid was introduced into Streptomyces by conjugation from E. coli ET12567 (pUZ8002). The mixed bacterial solution was cultured in SFM medium with addition of 10 mM MgCl₂ solution for 16 h at 30 °C. Then, the plates were supplemented with apramycin (15 mg/L) and trimethoprim (50 mg/L). After 7–14 d of incubation at 30 °C, the exconjugants were screened using apramycin (15 mg/L) and trimethoprim (50 mg/L). Subsequently, the exconjugants underwent iterative subculturing in antibiotic-free medium, followed by serial dilutions and plating to obtain single colonies. These single colonies were simultaneously cultured on SFM medium supplemented with apramycin (15 mg/L) and antibiotic-free to verify the loss of resistance marker. The selected candidate exconjugants were expanding cultured and further confirmed by PCR.

2.4. Construction of gene cluster-doubled strains

The pBACendBGC incorporates the intact endusamycin biosynthetic gene cluster, which was constructed in a pBeloBAC11 vector using ExoCET cloning [31]. The aac(3)IV-oriT-attP-int cassette was inserted via Redαβ recombination for intergeneric conjugation between E. coli and Streptomyces [32]. pBACendBGC was transferred into Streptomyces through conjugation. The exconjugants were selected based on their cell phenotype, which exhibited resistance to apramycin (15 mg/L) and then confirmed by PCR.

2.5. Fermentation of wild-type strains and related derivative strains

In shake flask fermentation experiment, Streptomyces preservation solution was streaked onto SFM agar plates and incubated at 30 °C for 7 d to isolate single colonies. These single colonies were then inoculated onto fresh SFM agar medium and further cultured at 30 °C for 7 d. Four well-grown colonies were picked and suspended in sterile water. After being uniformly ground using a grinder, they were inoculated into 100 mL seed medium. A seed culture was cultured in seed medium (4 % (w/v) soluble starch (Sinopharm Chemical Reagent Co., Ltd., 10021318), 1 % (w/v) soybean flour (Qingdao Haike Biotechnology Co., Ltd., 7412), 0.25 % (w/v) yeast extract (Angel Yeast Co., Ltd., FM905), 0.3 % (w/v) CaCO₃ (Sinopharm Chemical Reagent Co., Ltd., 10005760), pH 7.2) at 30 °C with shaking at 250 rpm for 30 h. Subsequently, a 5 % (v/v) inoculum was transferred to the fermentation medium, which had an identical composition to the seed medium, and incubated at 30 °C with shaking at 250 rpm for 7 d.

For fermentation experiments in bioreactors, the cultivation process of the seed culture remains consistent with the shake flask fermentation protocol. Then, a 10 % (v/v) inoculum was transferred to into 35 L fermentation medium (4 % (w/v) glucose (Hebei Jinfeng Starch Sugar Alcohol Co., Ltd.), 1 % (w/v) soybean flour (Qingdao Haike Biotechnology Co., Ltd., 7412), 0.25 % (w/v) yeast extract (Angel Yeast Co., Ltd., FM905), 0.3 % (w/v) CaCO₃ (Jiangxi Mingyuan High-tech Materials Co., Ltd.), 0.01 % (w/v) KNO₃, pH 7.2) in the 50 L bioreactor for fed-batch fermentation at 30 °C. The dissolved oxygen (DO) concentration was maintained above 40 % by adjusting the agitation speed from 200 to 550 rpm and the airflow rate from 1.6 to 2.2 m³/h. Furthermore, the glucose concentration was maintained at 1%–2% during the middle and late stages of fermentation. The bacterial broth was centrifuged to obtain bacterial precipitates for cell dry weight determination.

2.6. HPLC analysis and quantification of endusamycin

The fermentation broths of the wild type strains and related derivative strains were centrifuged at 4000 rpm for 15 min to collect the mycelia. An appropriate amount of methanol was added to the mycelia. The mixture was followed by ultrasonic extraction for 30 min and centrifuged at 4000 rpm for 15 min. The supernatant was filtered through 0.22 μm microporous membrane.

High-performance liquid chromatography (HPLC, Thermo Fisher Scientific, USA) analysis was performed with an ChromCore Polar C18 column (250 mm × 4.6 mm, 5 μm) using acetonitrile-H₂O (0.1 % formic acid) system at a flow rate of 1 mL/min. The gradient elution program was as follows: an initial 70 % acetonitrile was linearly increased to 99 % over 14 min, increased to 100 % for 5 min, then maintained at 100 % for 1 min. Subsequently, the acetonitrile concentration was gradually reduced from 100 % to 70 % over 3 min and maintained at 70 % for an additional 2 min. The absorbance of endusamycin was detected at 232 nm throughout the process.

3. Results

3.1. Metabolic competition in the omics perspectives

S. endus subsp. aureus, the natural producer of endusamycin, harbors a genome containing 9431 predicted CDSs. In silico analysis of the draft genome using antiSMASH [33] allowed the identification of a total of 44 BGCs, associated with polyketides (PKS), nonribosomal peptides (NRPS), saccharides, ribosomally synthesized and post-translationally modified peptides (RiPPs), hybrid NRPS-PKS, terpenes and others. Fourteen BGCs were classified as PKS or PKS-NRPS types (Table 1). Region 1, a Type I PKS with 90 % similarity to the nanchangmycin BGC, has been confirmed to be responsible for the synthesis of endusamycin [7].

Table 1.

Putative PKS/NRPS-PKS gene clusters in S. endus subsp. aureus.

Region	Type	Length (bp)	Most similar known cluster	Type	Similarity %
1	T1-PKS	133,252	nanchangmycin	Polyketide	90
2	PKS-like	40,659	acarviostatin	Saccharide	18
3	T2-PKS,NRPS	161.209	cinnapeptin	NRP	78
4	transAT-PKS, NRPS	107,076	oxazolepoxidomycin A	NRP + Polyketide	93
5	T1-PKS, NRPS-like	46,514	borrelidin	Polyketide	11
6	T2-PKS	72,497	spore pigment	Polyketide	83
7	NRPS, T1-PKS	55,492	landepoxcin	NRP + Polyketide	22
8	T1-PKS	53,928	foxicin	NRP + Polyketide	12
9	NRPS, PKS-like	70,439	misaugamycin A	NRP	5
10	T1-PKS, NRPS	573,939	meilingmycin	Polyketide	64
11	T1-PKS, NRPS-like	48,273	NFAT-133	Polyketide	31
12	T1-PKS	46,269	gausemycin A	NRP + Saccharide	2
13	NRPS, T1-PKS	112,704	hexacosalactone A	Other	13
14	NRPS, T1-PKS	64,567	guadinomine	Saccharide	15

In general, only actively transcribed BGCs are supposed to compete with the endusamycin synthesis for the starter units and extender units. Therefore, transcriptome analysis was performed during the rapid accumulation period of endusamycin production utilizing the NovaSeq 6000 platform (Illumina). The FPKM values of the genes in the fourteen BGCs were calculated and visualized in a violin plot. The violin plot reveals that, except for the BGC in region 5, the transcription levels of the other twelve BGCs, along with the endusamycin BGC, were comparable and fell within the moderate range compared to the overall transcription level of the genome (Fig. 2). This finding suggests that these BGCs engage in intense precursor competition during secondary metabolites synthesis.

Fig. 2.

Violin plot of the transcription levels of putative PKS/NRPS-PKS gene clusters in S. endus subsp. aureus.

3.2. Elimination of precursor competition pathway

To minimize precursor loss, the deletion of competing BGCs was prioritized. To avoid unintended effects caused by unknown functions of unknown products, gene clusters with high similarity to known product BGCs and clearer elucidated biosynthetic pathway were preferentially selected for editing. Region 6, corresponding to the spore pigment BGC with 83 % similarity, and Region 10, corresponding to the meilingmycin BGC with 64 % similarity, were targeted. The spore pigment, classified as a type II PKS, needs one molecule of acetyl-CoA as the starter unit and eleven molecules of malonyl-CoA as extender units for biosynthesis in S. coelicolor [34]. Meilingmycin, a type I PKS, demands one molecule of acetyl-CoA, seven molecules of malonyl-CoA and five molecules of methylmalonyl-CoA for biosynthesis in S. nanchangensis NS3226 [35]. Similarly, the biosynthesis of endusamycin involves one molecule of acetyl-CoA, four molecules of malonyl-CoA and ten molecules of methylmalonyl-CoA (Fig. 1B). Acetyl-CoA, malonyl-CoA and methylmalonyl-CoA serve as common precursors.

Subsequently, the spore pigment and meilingmycin-like BGCs were deleted as planned. Homologous proteins in region 6 were aligned using interactive Blastp, referencing the spore pigment BGC from S. sahachiroi ATCC 33158 (GenBank: JQ951949.1). GM005673, GM005672, GM005671, GM005674, GM005670, GM005669, GM005676, and GM005675 share 74 %, 71 %, 53 %, 75 %, 71 %, 64 %, 69 %, and 56 % sequence identities, to SahA, SahB, SahC, SahD, SahE, SahF, SahG, and SahH, respectively (Table S3). Thus, the gene GM005669-GM005676 was knocked out to generate strain YC1102 (Fig. 3AB), which resulted in a 13 % increase in the titer of endusamycin (Fig. 3C, Fig. S1). Similarly, homologous proteins in region 10 were aligned with reference to the meilingmycin BGC from S. nanchangensis NS3226 (GenBank: FJ952082.1) and each gene exhibited >85 % identity (Table S4). GM007821, homologous to MeiA1, the PKS responsible for the loading module and two extension modules in meilingmycin biosynthesis [35], was selected for knockout. The resulting strain YC1108, derived from YC1102 (Fig. 3AB), exhibited a 20 % increase in the titer of endusamycin compared to S. endus subsp. aureus, reaching 1888 mg/L (Fig. 3C, Fig. S1). These results clearly demonstrate that the individual or tandem deletion of competing BGCs can significantly enhance endusamycin synthesis.

Fig. 3.

In-frame deletion for spore pigment and meilingmycin-like gene clusters in S. endus subsp. aureus. (A) Schematic diagram of the in-frame deletion. (B) Confirmation of mutant strains by PCR. PCR was carried out using the primers listed in Table S2. The arrows indicate the expected size of the PCR fragments in the wild-type and mutant strains. (C) HPLC analysis of endusamycin production from wild-type and mutant strains. Statistical analysis between the experimental group and the CK group was conducted using a two-tailed Student's t-test (∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001).

3.3. Doubling of the endusamycin biosynthetic gene cluster

During endusamycin synthesis, the post-PKS tailoring enzymes exhibit broad substrate specificity, meaning that enhancing the expression of a single enzyme may lead to the preferential accumulation of certain by-products. Therefore, the doubling of the complete BGC is undoubtedly a more effective strategy.

The pBACendBGC plasmid, containing the complete endusamycin BGC, is 135.8 kb in length (Fig. 4A). Initial attempts using traditional gene manipulation systems failed to obtain conjugants. Therefore, the conjugation system was systematically optimized to improve efficiency, including adjustments to culture mediums, conjugation methods (biparental conjugation based on E. coli ET12567/pUZ8002 and triparental conjugation based on E. coli ET12567/pUB307), morphology of Actinomycete, the mixing ratio of recipient to donor cells, and the incubation time on plates. Finally, an optimal conjugation system was established (Table 2). It was observed that the mixing ratio of E. coli to Actinomycete had a more pronounced effect on the conjugation efficiency. The low number of conjugants highlights the inherent difficulty in conjugation of large plasmids. The optimized gene manipulation method was employed to introduce pBACendBGC into S. endus subsp. aureus to generate strain YC1107 (Fig. 4B). The endusamycin titer of YC1107 reached 2678 mg/L, 69 % higher than that of S. endus subsp. aureus (Fig. 4C, Fig. S1).

Fig. 4.

Doubling of the endusamycin biosynthetic gene cluster in S. endus subsp. aureus. (A) pBACendBGC with whole endusamycin gene cluster and cassette of int-attP-oriT-aac (3) IV from pSET152. (B) Confirmation of mutant strains by PCR. PCR was carried out using the primers listed in Table S2. The arrows indicate the expected size of the PCR fragments in the wild-type and mutant strains. (C) HPLC analysis of endusamycin production from wild-type and mutant strains. Statistical analysis between the experimental group and the CK group was conducted using a two-tailed Student's t-test (∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001).

Table 2.

Exploration of the conjugantion conditions of E. coli to S. endus subsp. aureus.

Condition	Set value	Optimal value	Number of exconjugant (Cultivate to 7 d)
Culture medium	SFM, ABB13	SFM	4∼10
Conjugation method	biparental, triparental	biparental
Morphology of Actinomycetes	spores, mycelium	spores
Recipient (CFU)	103, 2 × 103, 4 × 103, 8 × 103	103 or 2 × 103
Donor (CFU)	103, 5 × 103, 104	104
The incubation time (h)	14, 16, 21	16

3.4. Optimization of fermentation medium composition

Substrate composition plays a crucial role in antibiotic production in Actinomycetes [36]. Here, the contents of the carbon source, nitrogen source, phosphorus and potassium in the fermentation medium were systematically optimized to promote the synthesis of endusamycin in S. endus subsp. aureus.

For carbon source optimization, 4 % soluble starch was used as the control. The effects of slow-release carbon sources (corn starch, soybean oil) and a rapid-release carbon source (glucose) were compared, including 4 % corn starch, 4 % glucose, 2 % corn starch + 2 % glucose, and 2 % corn starch + 2 % soybean oil. The addition of soybean oil significantly reduced the titer of endusamycin. The group with 2 % corn starch + 2 % glucose outperformed the 4 % corn starch group, while the addition of 4 % glucose led to a notable 25 % increase in endusamycin titer, reaching 1912 mg/L. These results demonstrate that the strain prefers rapid-release carbon sources over slow-release ones (Fig. 5A).

Fig. 5.

Optimization of fermentation medium components and addition of PQQ. (A) The effect of carbon source selection on endusamycin production. (B) The effect of nitrogen source, phosphorus and potassium selection on endusamycin production. (C) The effect of addition of PQQ on endusamycin production. Statistical analysis between the experimental group and the CK group was conducted using a two-tailed Student's t-test (∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001).

Next, using 4 % glucose as the carbon source, the existing organic nitrogen sources (yeast extract, soybean flour) in the fermentation medium and commonly used inorganic salts ((NH₄)₂SO₄, K₂HPO₄, (NH₄)₂HPO₄, KNO₃) were evaluated with varying concentrations (Fig. 5B). The impact of organic nitrogen sources content on endusamycin production followed a similar trend. The titer of endusamycin decreased slightly when the content of soybean flour and yeast extract was reduced, while the titer of endusamycin decreased significantly after additional supplementation. The biomass of the strain exhibited a consistent trend with production. When (NH₄)₂SO₄ or (NH₄)₂HPO₄ was added, both the biomass and endusamycin production decreased significantly, with further declines observed at higher concentrations. In contrast, the addition of K₂HPO₄ notably reduced the titer of endusamycin but increased biomass. This phenomenon is consistent with previous studies and was attributed to the coordinated regulation of primary and secondary metabolism of the bacteria for environmental adaptation [37]. When KNO₃ was added at a concentration of 0.01 %, the titer of endusamycin increased by 20 %—2597 mg/L, without significantly affecting biomass. Previous studies have similarly reported that the supplementation of nitrate into the fermentation medium can significantly enhance the titers of several antibiotics, including rifamycin B, lincomycin A, azalomycin B, and lividomycin [38,39]. However, the titer of endusamycin decreased as the KNO₃ concentration increased further.

Consequently, the optimal fermentation medium formulation is considered to replace the original carbon source with 4 % glucose and add extra 0.01 % KNO₃.

3.5. Addition of cofactor PQQ

PQQ acts as a redox cofactor in alcohol or glucose dehydrogenase reactions in many Gram-negative bacteria. The addition of PQQ or the introduction of the pqq gene cluster has been shown to effectively enhance polyketides production [40]. Subsequently, an attempt was performed to add PQQ at various concentrations (0.04 μM, 28 μM, and 280 μM) at the early stage of fermentation (20 h). The addition of 0.04 μM PQQ resulted in a 33 % increase in the titer of endusamycin, with no further improvement observed at higher concentrations. It is noteworthy that simultaneous addition of 0.01 % KNO₃ and 0.04 μM PQQ did not lead to an additional increase in endusamycin production (Fig. 5C).

3.6. Integration of effective strategies

Strategies that positively impacted endusamycin production were integrated, including the deletion of spore pigment and meilingmycin-like BGCs, doubling of the endusamycin BGC and fermentation medium optimization. Specifically, the pBACendBGC plasmid containing the endusamycin BGC was introduced into the strain YC1108, which had tandem deletions of competing BGCs, generating strain YC1109. Shake-flask fermentations of the wild-type strain and YC1109 were performed using the optimized fermentation medium. The endusamycin titer of YC1109 was increased by 98 %—5469 mg/L, compared with S. endus subsp. aureus (2761 mg/L) (Fig. 6A, Fig. S2).

Fig. 6.

Endusamycin production in shake-flask fermentations and fed-batch fermentations. (A) HPLC analysis of endusamycin production from S. endus subsp. aureus and YC1109. Statistical analysis between the experimental group and the CK group was conducted using a two-tailed Student's t-test (∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001). (B) Fed-batch fermentation of S. endus subsp. aureus and YC1109 in a 50 L bioreactor.

To evaluate the ability of S. endus subsp. aureus and YC1109 to synthesize endusamycin in a fermenter, fed-batch fermentation was conducted in a 50-L bioreactor with 35 L working volume, scaling up from shake-flask fermentations. Key parameters, including dissolved oxygen (DO), pH, glucose concentration, dry cell weight (DCW) and the titer of endusamycin, were monitored throughout the process. DO was maintained above 40 % to ensure sufficient oxygen supply. However, when following the exact shake-flask fermentation formula, glucose was rapidly consumed within 0 h–80 h. As glucose neared depletion, significant apoptosis occurred, leading to a stagnation in productivity (data not shown). Notably, maintaining the glucose concentration at 1%–2% after 80 h preserved strain viability (Fig. 6B). That indicates a higher glucose demand in fed-batch fermentation compared to shake-flask fermentation. YC1109 exhibited higher rates of oxygen and glucose consumption than S. endus subsp. aureus within 20 h–80 h, accompanied by an increase in biomass. Despite similar rates of product accumulation occurring during the early stages of fermentation (1 h–40 h), the yield of YC1109 was higher and remained consistent at 40 h–120 h. The titer of endusamycin for S. endus subsp. aureus reached a maximum of 2285 mg/L at 120 h, whereas the titer of endusamycin for YC1109 continued to accumulate and ultimately achieved a peak titer of 5011.39 mg/L at 173 h (Fig. 6B). This improvement could be attributed to greater biomass accumulation in the earlier stages and enhanced metabolic capabilities.

4. Discussion

The poultry industry incurs costs of approximately £7.7 to £13.0 billion (based on 2016 prices) across seven countries for the prophylaxis, treatment, and production losses caused by avian coccidiosis [41]. Polyether compounds as ionophores have been widely used against coccidiosis in chickens since 1939. The development of new polyether compounds can help mitigate the risk of drug resistance resulting from the long-term use of existing compounds [42], and endusamycin might be a candidate. Moreover, endusamycin and other analogs of nanchangmycin exhibit extensive antitumor activities, offering significant potential for drug development. While substantial progress has been made in the metabolic engineering of polyketides [20], there have been limited efforts to enhance the production of polyether compounds. In this study, we evaluated the effectiveness of metabolic engineering strategies to increase the titer of endusamycin.

There are various strategies to increase the precursor supply for polyketides, including enhancing the conversion efficiency of coenzyme A, deleting competing gene clusters, strengthening primary metabolism and so on [22,43]. Among these, the deletion of competing gene clusters has proven consistently effective. However, given the inherent challenges and the time-consuming process of genetic manipulation in Streptomyces, eliminating all non-target BGCs is impractical. Targeting highly competitive BGCs by eliminating trace or silent gene clusters based on omics analysis is a more optimal choice. Interestingly, 12 of the 13 PKS/PKS-NRPS BGCs exhibited transcription levels comparable to that of the endusamycin BGC, and BGC in Region 5 showed a higher transcription level (Fig. 2). However, Region 5 shares only 11 % similarity with known gene clusters (Table 1), suggesting a strong potential for producing novel compounds with unexpected functions. Therefore, BGCs like this were excluded from deletion. Instead, two well-studied gene clusters, responsible for spore pigmentation and meilingmycin-like production, were targeted. The resulting mutant strain YC1108, with serial deletions, achieved a 20 % increase in endusamycin titer.

The synthesis of type I polyketide backbone relies on the cascade catalysis of multi-module PKS, resulting in BGCs often exceeding 50 kb in size [31]. Consequently, doubling of polyketide BGCs in Streptomyces is challenging, primarily due to the complexities of gene cluster acquisition and genetic manipulations. However, recent advances in highly efficient gene cluster capture technologies have enabled the rapid and convenient acquisition of large gene cluster fragments [31,44,45]. The pBACendBGC plasmid containing the endusamycin gene cluster was obtained using ExoCET. Systematic optimization of the genetic manipulation method allowed the doubling of the endusamycin BGC in the native strain, resulting in a 69 % increase in endusamycin titer. This highlights the significant effectiveness of this strategy. This is also a rare case of doubling such a large Type I PKS gene cluster in a native strain.

Given the complexity of microbial metabolism, optimizing the fermentation medium is essential to ensure a balanced and diverse nutrient supply. In shake flask fermentation, replacing 4 % soluble starch in the initial medium with 4 % glucose led to a 25 % increase in endusamycin titer. In fed-batch fermentation, strain growth and product accumulation showed a clear dependency on glucose. These observations highlight the strain's preference for glucose as a nutrient source. Interestingly, S. albus DSM41398, the producer of the polyether salinomycin, demonstrates a preference for oil-rich media. The titer of endusamycin in S. endus subsp. aureus and its derivatives remained consistent for shake flask and fed-batch fermentation, whereas the titer of salinomycin in fed-batch fermentation exhibited a significant "scale-up" effect [8]. Based on these different habits of strains, endusamycin could potentially be heterologously expressed in the salinomycin-producing strain S. albus DSM41398 or the high-yielding variant S. albus BK3-25 in the future. However, the risk of metabolic incompatibility must be carefully considered, as it may lead to substantial yield losses, as observed when salinomycin was expressed in three exogenous hosts [15].

In summary, the deletion of competing BGCs and doubling of the endogenous BGC significantly enhanced endusamycin production. Through effective metabolic engineering and fermentation medium optimization, the titer of endusamycin was increased to 5469 mg/L, 3.46 times higher than the initial titer (1580 mg/L). This study not only provides a high-fielding strain for the drug development and industrialization of endusamycin but also offers a superior chassis cell for exploring endusamycin derivatives, such as YP-9 or those with further structural modifications. Furthermore, the proven strategies in this study can be extended to enhance the production of other polyether compounds.

CRediT authorship contribution statement

Yingying Chang: Writing – original draft, Methodology, Investigation. Zhen Liu: Investigation. Zixin Deng: Supervision, Conceptualization. Tiangang Liu: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interests

The author Zixin Deng is the Founding Editor for Synthetic and Systems Biotechnology, the author Tiangang Liu is an Editorial Board Member for Synthetic and Systems Biotechnology and they were not involved in the editorial review or the decision to publish this article. The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Zhen Liu is currently employed by Wuhan Hesheng Technology Co., Ltd.; the research project is funded by Wuhan Hesheng Technology Co., Ltd.

Appendix B. Supplementary data

The following is the Supplementary data to this article: