Integrated genome mining and molecular networking uncover the biosynthetic potential of a novel mangrove-derived strain Streptomyces sp. B1866

Abstract

Mangrove-derived actinomycetes represent a prolific reservoir of natural products, characterized by diverse chemical structures and significant pharmacological properties. Here, a mangrove-derived Streptomyces sp. B1866 was proposed to represent a novel species within the genus Streptomyces due to the low 16S rRNA gene sequence similarity (< 97%), and low overall genome relatedness indices (dDDH, 24.1%-24.7%; ANI, 80.1%-80.6%) with its closely related species. A whole genome sequence of strain B1866 comprised 42 secondary metabolite-biosynthetic gene clusters (BGCs), more than half of which exhibited low similarity (< 70%) to characterized BGCs, showing a potential for the production of compounds that are either uncharacterized or not associated with known structures. UPLC-MS/MS-based molecular networking further revealed the biosynthetic potential of strain B1866 since several nodes could not be assigned to the previously reported compounds. Eventually, investigation chemical constitutes in the fermentation of strain B1866 led to the isolation of a previously undescribed benzoxazole, designated streptoxazole A (1), along with 3 known compounds adenosine (2), diisobutyl phthalate (3), and ergosta-5,7,22-trien-3-ol (4). Streptoxazole A (1) possessed anti-inflammatory properties, inhibiting LPS-induced nitric oxide (NO) production with IC₅₀ values of 38.4 μM. This study exemplifies the discovery of novel molecules from mangrove-derived Streptomyces species.

Background Mangrove-derived microorganisms have been confirmed to produce secondary metabolites with various chemical structures and pharmaceutical activity. By integrating genome mining with UPLC-MS/MS based molecular networking approach, the metabolic potential of a novel mangrove-derived Streptomyces species B1866 was confirmed. Further investigation of the metabolite profile of strain B1866 led to the isolation of three known secondary metabolites, as well as a novel benzoxazole compound with anti-inflammatory property.

Results In this study, a Streptomyces strain B1866 isolated from mangrove sediments was identified and characterized. The whole genome of strain B1866 was sequenced to identify 21 BGCs involved in the biosynthesis of polyketides (PKS), non-ribosomal peptides (NRPS), and PKS-NRPS hybrid metabolites, highlighting its metabolic potential. Meanwhile, UPLC-MS/MS-based molecular networking revealed the presence of products with diverse structures in the crude extracts of strain B1866. Subsequently, compound 1, namely streptoxazole A (1), with a rare benzoxazole moiety was isolated. The structure of 1 was deduced from the analysis of 1D/2D NMR, HRESIMS data, and electronic circular dichroism (ECD) comparisons. The biosynthetic pathway of 1 required the shikimate pathway to generate precursor 3-amino-1,2-benzenediol and 2-(2-carboxyacetyl) benzoic acid. The isolated compound streptoxazole A (1) showed potent inhibitory effect toward NO production in LPS-induced RAW264.7 cells (IC₅₀ 38.4 μM).

Conclusions In conclusion, a novel benzoxazole, designated streptoxazole A (1), with anti-inflammatory property was isolated from a novel mangrove-derived Streptomyces sp. B1866. The integration of genome mining and UPLC-MS/MS-based molecular networking not only unveiled the metabolic potential of strain B1866 but also provided evidence for the discovery of novel chemicals from mangrove-derived Streptomyces species.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-025-11966-3.

Introduction

Mangrove forests are unique ecosystems that appear in the intertidal zones of coastlines, comprising abundant and complex biological communities, including microorganisms. Mangrove-derived microorganisms have been characterized by adaptations to salt stress, ionic stress, oxidative stress, et al., which influence their biosynthetic potential and ability to produce novel chemical entities [1, 2]. Investigation on secondary metabolites from mangrove-derived bacteria in the past were found to be quite productive and give rise to new carbon skeleton with pronounced biological activities [2]. Remarkable examples included the salinosporamide A, which is being evaluated in phase III clinical trials to treat glioblastoma [3], as well as the potent peroxiredoxin 1 inhibitor piericidin glycoside S14 [4]. Streptomyces that are major sources of antibiotics and other pharmacologically active compounds have been discovered enriched in mangrove sediments [5]. Over the last decades, more than 100 natural products with significant pharmacological activities were isolated from mangrove-derived Streptomyces spp. [6].

Benzoxazole are important heterocyclic skeleton that feature prominently in pharmaceutical drugs, such as flunoxaprofen (anti-inflammatory), boxazomycin B (antibacterial), chloroxazone (muscle relaxant), tafamidis (transthyretin tetramers inhibitor) et al. [7–9]. The majority of bacterial origin benzoxazole-containing natural products consist of 2-oxazoline moieties, which were derived from serine, threonine, 3-hydroxyanthranilic acid (3-HAA) or 3-hydroxy-4-amino-anthranilic acid units via cyclization and oxidation [10, 11]. Although great progresses had been made in the research of benzoxazoles for the development of novel drugs in the last few decades, natural sources of benzoxazole-containing compounds were extremely rare. To date, less than 50 benzoxazoles were primarily derived from Streptomyces [12, 13]. Only a handful of benzoxazole-containing molecules biosynthetic loci (calcimycin, A33853, caboxamycin, and nataxazole) have been identified [11, 14, 15].

As vast genomic data is available in publicly accessible research databases, genome mining has been frequently employed for the discovery and characterization of genes involved in natural product biosynthesis. With the help of genome mining strategy, Streptomyces species have been found to possess a larger capability to produce diverse secondary metabolites than previously known, which was proved by the identification of a large number of cryptic biosynthetic gene clusters (BGCs) [16, 17]. To avoid the rediscovery of previously reported compounds, a combination of genome mining and metabolomics strategy has been frequently applied in discovering putative new molecules and new homologues of known natural products. Molecular networking, an approach for visualizing tandem mass spectrometry (MS/MS) data, has been introduced in metabolomics. It tends to identify similar secondary metabolites by the structural units and recognize unique metabolites, providing a comprehensive insight into the metabolite profile [18]. Recently, the integration of genome mining with molecular networking has contributed to the discovery of a new cytotoxic cyclodipeptide from Streptomyces hygrospinosus 26D9-414 [19], 15-deoxynaphthomycins with nuclear factor erythroid 2-related factor 2 (NRF2) activation function from Streptomyces sp. N50 [20], and antifungal agents antimycin analogues from Streptomyces sp. NBU3104 [21].

In the present study, mangrove-sediments derived strain B1866 was identified as a novel member of the genus Streptomyces that has received our attention due to the genome of this isolated harboring a high number of biosynthetic gene clusters (BGCs). To further investigate its biosynthetic potential, an ultra-high performance liquid chromatography-mass spectrometry (UPLC-MS/MS)-based molecular networking approach was applied for the detection and isolation of secondary metabolites. As a result, a novel benzoxazole streptoxazole A (1) was obtained. The isolation, structure elucidation, putative biosynthetic pathway, and bioactivity of this compound were discussed herein.

Results and discussion

Characterization of strain B1866

The colonies of strain 1866 exhibited a white to greyish-white coloration on modified ISP2 agar and produced abundant aerial mycelium (Fig. 1a and 1b). Alignment of 16S rRNA nucleotide sequence of strain B1866 with the type strains of the species within the genus Streptomyces revealed that it shared the highest similarity to S. malaysiensis NBRC 16446 ^T (96.4%) and lower similarity values of < 96.3% with other Streptomyces species. Genome-wide average nucleotide identity (ANI) and digital DNA-DNA hybridization (dDDH) values between strain B1866 and five closely related species of the genus Streptomyces ranged from 80.3–80.1% and 24.7–24.1%, respectively (Table S1). These values were all far below the species delineating threshold of 95–96% for ANI values and 75% for dDDH values [22, 23], confirming the genetic distinctiveness of strains B1866 in the genus Streptomyces. The taxonomy position of strain B1866 within the genus Streptomyces was further confirmed due to the formation of a distinct clade in the phylogenomic tree (Fig. 1b). Collectively, these findings demonstrated that strain B1866 isolated from mangrove sediments represented a novel species within the genus Streptomyces. The strain B1866 was deposited at the Marine Culture Collection of China and Japan Collection of Microorganisms with collection numbers of MCCC 1K04846 and JCM 36587, respectively.

Fig. 1.

Morphology features of strain B1866 (a and b) and a phylogenomic tree generated with UBCG based on strain B1866 and 16 species from the genus Streptomyces (c). a, Scanning electron microscopy of strain B1866, scale bar presents 5 μm, 3 kV; b, Colonies of strain B1866 grown on modified ISP2 agar at 7 days. c, The numbers at the nodes indicated the Gene Support Index (GSI, maximal value is 92). Strain Kitasatospora paracochleara DSM 41656 ^T (JAMZDX000000000) was used as an outgroup. Bar, 0.05 substitutions per site

The genome properties of strain B1866

The genome of strain B1866 (accession number: JAVSCF000000000) consisted of 6,972,251 bp with 74.29% of GC content (table S2), which was consistent with the previous finding that the majority of marine-derived Streptomyces spp. possessed high GC content [24]. The strain harbored 5873 protein-coding genes, from which 73.5% were assigned to COG, 53.2% were assigned to GO, and 37.5% were assigned to KEGG (Fig. S1). Previous studies described that amino acid metabolism, carbohydrate, metabolism of cofactors and vitamins, and energy metabolisms were the most abundant metabolic functions in mangrove derived-prokaryote [25]. Similarly, 14.4, 12.7, 10.4, and 9.1% of genes respectively participated in amino acid metabolism, carbohydrate metabolism, metabolism of cofactors and vitamins, and energy metabolism. Based on COG analysis, 4319 genes were classified in 20 functional categories, of which 1490 genes were identified to be of unknown functions. In addition, most annotated genes were associated with transcription (accounted for 9.0%), amino acid transport and metabolism (accounted for 6.6%), energy production and conversion (accounted for 5.5%), signal transduction mechanism (accounted for 5.3%), and replication, recombination and repair (accounted for 5.0%). In GO categories, significant amounts of annotations were enriched in biological process and molecular function. The abundance of these genes in strain B1866 might play vital roles in environmental stress response by degrading organic matter, regulating protein synthesis and maintaining genome stability [26, 27].

Genomic and metabolomic profiles of strain B1866

The genome mining based on the antiSMASH analysis identified 42 BGCs in strain B1866, including 17 polyketide synthases (PKSs), 2 nonribosomal peptide synthases (NRPSs), and 3 PKS/NRPS hybrids (table S3, Fig. S2). Among them, two clusters showed 60% and 48–74% similarity score ranges with the biosynthetic genes involved in the biosynthesis of typical Streptomyces secondary metabolites geosmin and desferrioxamines, respectively. Marine actinomycetes are recognized as a rich source of polyphenols, which are synthesized by type II polyketide synthases and have gained significant attention due to their important pharmacological activities [28]. In the genome of strain B1866, type II PKS gene clusters located in scaffold 7, scaffold 53 and scaffold 93 displayed high amino acid identities with type II PKS genes from the BGCs of antimicrobial agents rabelomycin (70% and 87%), maduralactomycin (70% and 80%), and anthrabenzoxocinone (78% and 89%), respectively (table S3). Members of the polycyclic tetramate macrolactams (PTMs) family are structurally complex natural products with diverse biological activities. A PKS-NRPS hybrid clusters showed 70–81% identity to the gene clusters responsible for the biosynthesis of SGR PTMs in S. griseus subsp. griseus NBRC 13350 [29]. Additionally, clusters with moderate amino acid identities to well-characterized bioactive compounds such as maklamicin (46–63%), rakicidins (47–63%), salinomycin (46–57%), borrelidin (46–58%), ambruticin (46–50%), lasalocid (46–60%), hexacosalactone A (52–57%), NFAT-133 (48–76%), fostriecin (46–59%), meilingmycin (49–72%), olimycins (48–50%), and cetoniacytone A (60%) were detected, proposing that strain B1866 might generate novel derivatives of these potent compounds. These findings highlighted the potential of Streptomyces sp. B1866 to produce a wide variety of secondary metabolites and it was worthy to be investigated further.

The UPLC-MS/MS data of the crude extract from YPD medium was analyzed using the GNPS platform. By comparison of the obtained mass data with the literature data, a high number of unknown compounds were detected. In addition, integrating the identified metabolites with the genomic analysis increased the number and variety of detectable gene clusters in the genome of strain B1866 (table S4). Subsequently, scale up the fermentation of strain B1866 using YPD medium led to the discovery of a novel benzoxazole streptoxazole A (1) and three previously identified compounds (2–4). Among them, the putative biosynthetic gene cluster of adenosine (2) was also predicted in the genome of strain B1866 (table S4). The compound (1) was not detected in the crude extract derived from the modified ISP2 medium, which was used for the isolation and purification of strain B1866. However, many prediction products in UPLC-MS/MS-based molecular networking were not isolated from YPD cultures. This might be due to the low yields of these compounds and the low expression of biosynthetic genes. Thus, further investigations should be performed to optimize conditions for improving the biosynthesis of secondary metabolites in strain B1866.

Structure elucidation

Compound 1 (Fig. 2) was acquired in the form of a white amorphous solid, and the molecular formula C₂₄H₂₃NO₉ was established by HR-ESI–MS ion at m/z 485.1628 [M-H]⁻ (Fig. S3). Analysis of the ¹H NMR spectrum (Table 1, Fig. S4) showed a series of aromatic proton signals at δ_H 7.95 (d, J = 8.9 Hz, H-4), 7.78 (d, J = 8.9 Hz, H-6), 7.33 (t, J = 7.8 Hz, H-5''), 7.31 (t, J = 8.9 Hz, H-5), 7.24 (t, J = 7.8 Hz, H-7''), 7.05 (t, J = 7.8 Hz, H-6''), and 6.87 (d, J = 7.8 Hz, H- 8''). The ¹³C NMR and HSQC spectrum of streptoxazole A displayed 24 carbon resonances (Table 1, Fig. S5 and S6), including three CH₂, thirteen CH, one methoxy, and seven nonprotonated carbons. Of them, seven characteristic signals suggested the presence of a sugar unit (δ_C 60.9, 62.1, 80.7, 78.0, 77.3, 75.1, and 102.7), which was confirmed through HMBC (Fig. S7) correlations from H-2^' (δ_H 3.50) to C-1^' (δ_C 102.7), and C-4^' (δ_C 78.0), H-3^' (δ_H 3.48) to C-4^'' (δ_C 76.5), H-4^' (δ_H 3.59) to C-3^' (δ_C 77.3), H-6^' (δ_H 3.71, 3.87) to C-4^' (δ_C 78.0), OCH₃−7^' (δ_H 3.59) to C-5^' (δ_C 80.7), as well as ¹H − ¹H COSY correlations of H-1^'/H-2^'/H-3^'/H-4^'/H-5^'/H-6^' (Fig. 3 and Fig. S8). Connectivities between the D-glucose units and the aglycone were established by HMBC correlations from H-1'(δ_H 4.97) to C-7 (δ_C 157.5). The benzoxazole structural fragment was established from the HMBC correlations of H-2 (δ_H 9.09) with C-7a (δ_C 149.7), and C-3a (δ_C 136.1), H-4 (δ_H 7.95) with C-6 (δ_C 109.7), and C-3a (δ_C 136.1), H-5 (δ_H 7.31) with C-7 (δ_C 157.5), and C-3a (δ_C 136.1). Furthermore, HMBC correlations (Fig. S6) between H-3^'' (δ_H 2.15) and C-COOH (δ_C 182.0), C-10^'' (δ_C 132.8), and C-2^'' (δ_C 58.5), and between H-5^'' (δ_H 7.33) and C-9''(δ_C 142.7), and C-7''(δ_C 130.6), and between H-6^'' (δ_H 7.05) and C-10^'' (δ_C 132.8), and C-8^'' (δ_C 111.3), between H-7^'' (δ_H 7.24) and C-9^'' (δ_C 142.7), together with ¹H − ¹H COSY (Fig. S7) correlations located connections from H-5^''/H-6^''/H-7^''/H-8^'' reconfirmed that the existence of a tetrahydroquinoline structural fragment in streptoxazole A (1).

Fig. 2.

Chemical structures of compounds isolated from strain B1866

Table 1.

¹H and ¹³C NMR data of streptoxazole A (1) in CD₃OD

	streptoxazole A
No	δC, type	δH (J in Hz)
1	-	-
2	155.5, CH	9.09, s
3	-	-
3a	136.1, C
4	124.2, CH	7.95, d (8.9)
5	118.8, CH	7.31, t (8.9)
6	109.7, CH	7.78, d (8.9)
7	157.5, C
7a	149.7, C
1'	102.7, CH	4.97, d (7.7)
2'	75.1, CH	3.50, m
3'	77.3, CH	3.48, m
4'	78.0, CH	3.59, s
5'	80.7, CH	3.2, t (12.0)
6'	62.1, CH2	3.71, dd (12.0, 5.1), 3.87, dd (12.0, 2.2)
7'	60.9, OCH3	3.59, s
1''	-
2''	58.5, CH2	3.51, m
3''	41.3, CH2	2.15, m
4''	76.5, C	-
5''	125.2, CH	7.33, d (7.8)
6''	123.7, CH	7.05, d (7.8)
7''	130.6, CH	7.24, t (7.8)
8''	111.3, CH	6.87, t (7.8)
9''	142.7, C	-
10''	132.8, C	-
11''	182.0, C	-

Fig. 3.

COSY, key HMBC, and Key NOESY correlations of streptoxazole A (1)

The relative configuration of streptoxazole A (1) was determined according to NOESY correlations and coupling constants (Fig. S9). Analysis of the large diaxial coupling constant (J = 7.7 Hz) for the anomeric proton H-1^′ indicated a β-configuration for the sugar moiety (J_1'-J_2' > 7 Hz) [30]. A NOESY (Fig. 3) signal from H-1′ to H-5′ indicated that the sugar unit of 1 was D-glucose [31]. The absolute configuration of (4''S)-streptoxazole A was further determined by a comparative analysis of its experimental electronic circular dichroism (ECD) spectrum (Fig. 4) with the calculated spectrum. Additionally, the other compounds (Fig. 2) were identified as adenosine (2) [32], diisobutyl phthalate (3) [33], and ergosta-5,7,22-trien-3-ol (4) [34], based on comprehensively comparing the spectroscopic data with literature (Fig. S10-S15).

Fig. 4.

Comparison between experimental and calculated ECD spectra of streptoxazole A (1)

Proposed biosynthetic pathway of streptoxazole A (1).

Several benzoxazole-containing natural products within the bacterial kingdom featured the same benzoxazole moiety, which was derived from 3-hydroxyanthranilic acid (3-HAA), such as nataxazole, UK-1, A33853, caboxamycin, and calcimycin/cezomycin (Fig. 5a) [11, 35–37]. Based on the reported assembly logic, the 2-substituted benzoxazole moiety in these molecules were formed through the dimerization of 3-HAA by ester formation and subsequent heterocyclization. However, the structures of streptoxazole A (1) and the known compound sarubicinol A were exceptions because of their benzoxazole systems without substitute groups at position-2 [13]. In contrast to the well-studied biosynthetic loci of 2-subsitituted benzoxazole derivatives in bacteria, benzoxazole derivatives with different substitution sites have not been received considerable attention due to their limited amount. Gene clusters for published known pathways of benzoxazole-containing compounds have not been identified in strain B1866. Anthranilic acid served as a precursor of L-tryptophan and is a product of chorismite in the shikimate pathway [38]. In view of this, we presumed that streptoxazole A (1) is derived from the precursor anthranilic acid, followed by enzymatic esterification and cyclization (Fig. 5b). In the genome of strain B1866, a putative anthranilate phosphoribosyltransferase gene (RKE29_14475) might catalyze the conversion of chorismite to anthranilate that was identified in the scaffold 91 (table S5). Located in the scaffold 91 gave three cytochrome c oxidase genes (RKE29_14515, RKE29_14520, and RKE29_14525) that might participate in oxidation–reduction reaction. However, genes coding for glycosyl hydrolase and glycosyltransferase were not identified in scaffold 91. To further determine whether the identified genes are responsible for the biosynthesis of streptoxazole A (1) and elucidate the transformation of glycosyl moiety in this pathway, gene knockout and heterologous expression need to be carried out in the further study.

Fig. 5.

The assembly logic of benzoxazole-containing natural products (a) and proposed biosynthetic pathway for streptoxazole A (1, b)

Bioactivity assays and molecular docking study

Several benzoxazole derivatives have been reported to exhibit anti-inflammatory properties, acting through multiple mechanisms [39, 40]. The streptoxazole A (1) displayed significant inhibitory effects on nitric oxide (NO) production induced by lipopolysaccharide (LPS) with IC₅₀ values of 38.4 μM, which were comparable to positive control dexamethasone (6.5 μM), while no obvious cytotoxicity was observed (table S6). The molecular docking revealed that the two lowest binding energy were determined to be −8.589 and −6.766 for the sodium/glucose cotransporter and squalene synthase (Fig. S16), indicating the greatest binding affinity of this protein for streptoxazole A (1). Additionally, streptoxazole A (1) was found to have a binding energy of −6.415, −6.342, −6.253, −5.732, −5.72, −5.514, −5.384, −5.324, −5.293, −5.209, and −5.174 kcal/mol, with leucine-tRNA ligase, DNA topoisomerase 1, neprilysin, fibroblast growth factor receptor 1, splicing factor 3B subunit 3, tyrosine-protein kinase, endothelin-converting enzyme, hexokinase-1, valine–tRNA ligase, tyrosine-protein kinase, and hexokinase-2, respectively (table S7). Considering the complexity of the inflammatory process and functions of predicted targets, further experiments are required for identifying key vital targets of streptoxazole A (1).

Conclusion

In this study, a novel mangrove-derived Streptomyces sp. B1866 was identified. The biosynthetic potential of strain B1866 was assessed by the prediction of 42 BGCs and genes encoding putative adenosine, 4-homoectoine, indolelactic acid, apigenin, canthaxanthin, and 6’-hydroxymethyl simvastatin synthases. The biosynthetic capability of this strain was highlighted by the isolation of adenosine (2), diisobutyl phthalate (3), and ergosta-5,7,22-trien-3-ol (4). Although the metabolite profile of strain B1866 was complicated due to the presence of plenty of unknown metabolites in the metabolomic molecular networking analysis, a novel benzoxazole named streptoxazole A (1), exhibiting anti-inflammatory activity, was obtained here. This compound represented the first unusual benzoxazole derivatives reported from a mangrove-derived Streptomyces species and its putative biosynthesis pathway was also proposed.

Materials and methods

General experimental procedures

NMR spectra were recorded on a Bruker Avance 600 MHz spectrometer and chemical shifts are given in δ values (ppm). IR spectra were measured on a Bruker Tensor 27 FT-IR spectrometer. HRESI-MS were recorded on a Bruker maXis TOF-Q mass spectrometer. Column chromatographies were performed on Silica gel (100–300 mesh, China), Sephadex LH-20 (GE, Sweden), and YMC ODS-A, methanol system equipped with a ODS column (YMC-5 μm, 12 nm, ODS-A). Fractions were analyzed by TLC using GF254 silica gel (China).

Isolation of strain B1866

Mangrove sediments were collected in Beibu Gulf, Beihai, P. R. China (N21°53′15″, E108°31′25″). For isolation, 1 g sediment samples were suspended in 100 mL and 1000 mL distilled water. The obtained dilution (10^–2 and 10^–3 g mL⁻¹) liquid was plated on modified ISP2 (yeast extract 0.2% w/v, maltol 0.2% w/v, glucose 0.2% w/v, agar 1.5% w/v, water 1 L, pH 7.2–7.6) agar plates. After 7 days of aerobic incubation at 28 °C, a greyish white colony was transferred and purified on modified ISP2 agar. The purified strain was preserved at −80 °C in modified ISP2 liquid medium supplemented with 20% glycerol (v/v).

16S rRNA gene and genome sequencing

The genomic DNA extraction and PCR-mediated amplification of the 16S rRNA gene sequence were performed according to the method described by Hu et al. [41]. The amplified 16S rRNA gene was obtained from PCR product using a DNA Gel Extraction Kit (QIAGEN), ligated to pMD19-T vector (Takara,), transformed into Escherichia coli DH 5α competent cells followed by sequencing (Tsingke Biotechnology Co., Ltd., Beijing). To search the closely related type strains, the almost complete 16S rRNA gene sequence was compared with the published type species in NCBI database. The genome of strain B1866 was sequenced by using Hiseq 2500-PE150 platform at Majorbio Bio-pharm Technology Co., Ltd (Shanghai, PR China). The paired-end reads were assembled using SOAP denovo software (v2.04) [42]. Then, the quality control of sequence reads was conducted using SeqPrep (https://github.com/jstjohn/SeqPrep) and Sickle (https://github.com/najoshi/ -sickle) softwares. The annotation of gene-encoded proteins was performed using BLAST searches (E-value ≤ 1e-5) in Gene Ontology (GO) database, Cluster of Orthologous Groups (COG) of proteins, and Kyoto Encyclopedia of Genes and Genomes (KEGG) database [43–45]. The biosynthetic gene clusters of secondary metabolites were annotated by antiSMASH bacteria version 7 [46]. The similarity score was obtained based on the comparative analysis between the identified gene clusters and published BGCs in MIBiG database [47]. The putative function of all amino acid sequences was predicted by the comparison of available protein sequences in NCBI database using a BLAST-p search.

Phylogenomic analyses

Cell morphological characteristics of strain B1866 were observed under scanning electron microscopy (HITACHI SU8010). Genome relatedness referred by values of ANI and dDDH were determined by OrthoANI online service (https://www.ezBioCloud.net/tools/ANI) [48] and the Genome-to-Genome Distance calculator platform (http://ggdc.dsmz.de/ggdc.php) [23], respectively. A phylogenomic tree was constructed with 92 universal bacterial core gene sequences that generated by UBCG (Up-to-date Bacterial Core Gene) using genome sequences of strain B1866 and 16 closely related strains [49]. Gene Support Index (GSI) was calculated and represented the number of genes supporting the branch in the UBCG tree. Values of GSI exhibited a positive correlation with the robust of branch [49].

UPLC-MS/MS Analysis

A 1L culture of Streptomyces sp. B1866 was grown in YPD medium (1% yeast extract, 2% peptone, 2% glucose, seawater, pH 7.8) for 7 days at 28 °C, 160 rpm. Then, the biomass and supernatant were separated by centrifugation (8000 rpm, 10 min). The suspension was extracted with equal volume of ethyl acetate for three times. The organic phases were collected and dried to afford 4.9 mg of crude extracts. Ultra-performance liquid chromatography (UPLC) analysis was performed using a Waters Acquity Ultra Performance LC system. Chromatographic separation was carried out on an ACQUITY UPLC HSS T3 column (100 mm × 2.1 mm, 1.7 urn) at a column temperature of 35℃. The mobile phase consisted of solvent A (0.1% formic acid in water, v/v) and solvent B (0.1% formic acid in acetonitrile), with gradient polarity (A:B) of 100:0 to 0:100. The flow rate was set at 0.3 mL/min. The column and auto-sampler were maintained at 35 °C and 10 °C, respectively. The injection volume was 5 μL. The extract was resuspended in methanol and the mass spectrometry analysis was performed on a Xevo G2-XS QTof mass spectrometer (Waters MS Technologies, Milford, CT, USA). Electrospray ionization was adopted. The scan range was from 50 to 1000 m/z. The capillary was set at 3 kV, and positive electron spray mode was adopted. The desolvation gas was set to 800 L/h at a temperature of 400 °C, the cone gas was set to 30 L/h, and the source temperature was set to 100 °C.

Molecular networking

The obtained MS data was imported into MS-Convert software to convert the files to mzXML format. A molecular network was generated using the Feature-Based Molecular Networking (FBMN) workflow on the GNPS platform [50] with the following parameters: a minimum cosine score of 0.65, a MS/MS fragment ion tolerance of 0.02 Da, and the precursor ion mass tolerance of 0.02 Da. The molecular formular were predicted based on the comparison of MS/MS data (NPAtlas and StreptomeDB). Annotations, including SMILES strings for identified compounds, were retrieved from GNPS library matches. To calculate the structural similarity scores, the raw LC–MS/MS data files from strain B1866 were directly processed using MS2DeepScore 2.0 [51]. Query spectra were compared against the GNPS MS/MS library ‘ALL_GNPS_210409_positive’ (or specify the exact model/library version used with MS2DeepScore). Putative structural annotations (SMILES strings) corresponding to high-scoring matches were retained.

To link BGCs to their potential products, candidate metabolites identified in the crude extracts of strain B1866 via GNPS and MS2DeepScore were considered. For each BGC, annotations from antiSMASH (specifically, the"most similar known cluster"field referencing MIBiG entries) were used to retrieve SMILES strings of known related compounds from public databases (e.g., PubChem). These BGCs associated SMILES strings were then compared against the SMILES strings derived from the GNPS and MS2DeepScore annotations of observed metabolites. Molecular similarity represented by Tanimoto similarity coefficients were calculated using the ChemmineR package in R [52]. The observed metabolite with similarity score (Tanimoto similarity score) > 0.5 was considered the most likely product of the related BGC.

Fermentation, extraction, and isolation of compounds

A slant culture of the strain B1866 was inoculated into 500 mL Erlenmeyer flasks containing 100 mL of YPD medium and maintained at 28℃. After culturing for 48 h at 180 rpm, 100 mL of the seed culture was inoculated into liquid medium (30% artificial seawater with 5.0% malt extract, 2.0% yeast powder, and 2.0% glucose, pH 7.8) and incubated at 28℃ for 7 days on a rotary shaker at 180 rpm. Large-scale fermentation was carried out. The completed fermentation broth (60 L) was filtered through cheesecloth to separate the culture filtrate and the mycelia. The filtrate and the mycelia were extracted three times with ethyl acetate (EtOAc). The combined EtOAc extracts were concentrated under reduced pressure to yield 4.6 g of crude extract (7.6%, g:L). The extract (4.6 g) was subjected to open silica gel (100–200 mesh) column chromatography (CC) with a CH₂Cl₂-MeOH solvent system (from 50:1 to 20:1 and finally 1:1) to separate into five fractions based on TLC properties. Fraction 2 was further separated on a silica gel CC to produce 5 subfractions (Fr. 2.1-Fr. 2.5). Fr. 2.5 was subjected to preparative thin layer chromatography (TLC) to produce ergosta-5,7,22-trien-3-ol (4) (9.0 mg).

Fraction 3 was subjected to Sephadex LH-20 chromatography (MeOH) to produce three fractions (Fr. 3.1-Fr. 3.3). Fr.3.2 was separated by ODS CC (MeOH: H₂O, 30:70) to yield three subfractions (Fr. 3.2.1-Fr. 3.2.3). Fr.3.2.2 was further fractionated by preparative TLC using CH₂Cl₂-acetone (8:1) to give diisobutyl phthalate (3) (8.0 mg). Fr. 3.2.3 was further separated by high performance liquid chromatography (MeOH: H₂O, 70:30) to obtain streptoxazole A (1) (3.1 mg) (0.06%). Fraction 4 (1.2 g) was separated by Sephadex LH-20 (MeOH) to yield five fractions (Fr. 4.1-Fr. 4.5). Fr.4.4 (96.5 mg) was further isolated through ODS CC (MeOH: H₂O, 60:40) to yield adenosine (2) (5 mg).

Streptoxazole A (1): White powder, [α]25 D-18 (c 0.2, MeOH), HRESIMS m/z: 485.1628 [M-H]⁻ (calcd. for C₂₄H₂₅N₂O₉, 485.1625, Fig. S2). ¹H NMR (CD₃OD, 600 MHz, Fig. S3) and ¹³C NMR (CD₃OD, 150 MHz, Fig. S4), refer to Table 1.

Cytotoxicity assay

The HCT-8, A2780, HeLa, and RAW 264.7 cells were cultured in DMEM medium at 37 ◦C under 5% CO₂ atmosphere. The cytotoxicity of new compound against four cell lines, were evaluated by the MTT method described in the literature [53].

Assessment of biological activity

The activity of compound 1 was examined by inhibiting NO production in LPS-stimulated RAW 264.7. The detailed process of the assay is described in the previously published paper [54]. Dexamethasone was served as positive control.

Potential target analysis of streptoxazole A (1)

The potential target of streptoxazole A (1) was predicted using SwissTargetPrediction (http://www.wisstargetprediction.ch). Then, potential streptoxazole A (1)-related targets were retrieved from the UniProt (https://www.uniprot.org) database with status set as “reviewed” and species set as “Homo sapiens”. The binding active pockets and binding energy between proteins and the prepared ligand were calculated using AutoDock (4.2.6) [55].

Abbreviations

ANI

Average Nucleotide Identity

dDDH

Digital DNA-DNA hybridization

YPD

Yeast extract peptone dextrose

MeOH

Methanol

UPLC

Ultra Performance Liquid Chromatography

NMR

Nuclear Magnetic Resonance

HSQC

Heteronuclear Single Quantum Correlation

COSY

Correlation Spectroscopy

HMBC

Heteronuclear Multiple Bond Correlation

NOESY

Nuclear Overhouser Effect Spectroscopy

Authors’ contributions

J.L.Y. conducted the strain and natural products isolation and structure elucidation. H.W.J. performed the bioinformatic and metabolic profiling analyses. T.B.R., Z.D. and C.S.W. performed the strain cultivation and activity assay. Q.J.L. prepared the Fig. 1. L.H. and P.X.L. designed the experiment and wrote the manuscript.

Funding

This work was supported by the National natural science foundation of China (82260689), the Guangxi natural science foundation (2025GXNSFAA069978 and 2025GXNSFAA069643), department of Science and Technology of Sichuan Province (2022NSFSC0109), Sichuan Science and Technology Program (2022YFS0624), and Sichuan Traditional Chinese Medicine Administration (2023zd008).

Data availability

Sequence data that support the finding of the study have been deposited in NCBI GenBank database.The 16S rRNA gene sequence and genome sequence of strain B1866 were deposited in the NCBI GenBank database under accession numbers of OR437945 and JAVSCF000000000, respectively. The feature-based molecular networking analysis (derived from MZmine-processed data) is publicly accessible on GNPS (Job ID: `39942b6c1c2f41b194a2ef6a2a42ca26`).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.