Genome Sequencing of Streptomyces griseus SCSIO PteL053, the Producer of 2,2′-Bipyridine and Actinomycin Analogs, and Associated Biosynthetic Gene Cluster Analysis

Abstract

Marine symbiotic actinomycetes play a key role in drug development and their ecological niches can influence a variety of natural product biosynthesis, providing potential defensive benefits. In this study, we report the whole-genome sequence analysis of marine gastropod mollusk Planaxis sp.-associated Streptomyces griseus SCSIO PteL053, which harbors 28 putative biosynthetic gene clusters (BGCs). Among them, two BGCs encoded by a hybrid non-ribosomal peptide (NRPS)/polyketide (PKS) synthetase and non-ribosomal peptide synthetase (NRPS) are responsible for the synthesis of the known therapeutic metabolites 2,2′-bipyridine and actinomycin analogs, respectively. Detailed bioinformatics analysis revealed the putative BGCs and the functions of the involved genes in the biosynthesis of the known compounds SF2738D (1), SF2738F (2), actinomycin D (3), and Actinomycin X_oβ (4). In the present study, complete-genome sequencing allowed us to rediscover known, clinically useful secondary metabolites in the newly isolated Streptomyces griseus SCSIO PteL053.

1. Introduction

Natural products (NPs) of microbial origin are the main reservoir of agricultural and therapeutic agents [1]. Over the past decade, numerous unique NPs were discovered in marine invertebrate-associated microorganisms, which suggests that actinomycetes, particularly Streptomyces sp., are a main source of novel NPs [2,3]. The understanding of interspecies interactions and their ecological contexts has created new opportunities for the discovery of natural products [4]. Actinomycetes build effective symbiotic relationships with marine invertebrates, which drives the synthesis of new classes of antibiotics and anticancer agents [5]. A substantial number of marine drugs, including trabectedin and eribulin, were isolated from microbial symbionts associated with marine animals. Recently, whole-genome sequencing analyses implied that most actinomycetes possess more than 20 biosynthetic gene clusters (BGCs). Genome-guided strategy is useful for the identification of genetic loci and their encoded gene products with new biosynthetic potential [6]. Recently, numerous NPs, including olimycin, mycemycin, and atratumycin, were yielded by a combination of gene sequencing, components, and a metabolic engineering approach by our group [7,8,9].

Bipyridine (2,2′-BP) is encoded by non-ribosomal peptide (NRPS)/polyketide synthetase (PKS), whereas actinomycin is biosynthesized by non-ribosomal peptide synthetase (NRPS). Bipyridine is a unique molecular scaffold composed of natural compounds from the actinobacteria caerulomycin and collismycin, whose 2,2′-BP core typically contains a di- or tri-alternative A ring and an unaltered B ring. The neuro-protective, anti-inflammatory, immunosuppressive, antibacterial, and antitumor activities of these compounds have been previously reported [10]. Actinomycin D and its analogs are an important family of chromopeptide lactone compounds with the same actinocin moiety but different amino acid residues on the pentapeptide lactone backbone. Many actinomycins exert strong antibacterial and antitumor action [11,12]. Among them, actinomycin D is a well-known drug and an FDA-approved chemotherapeutic agent against rare and difficult-to-treat cancers, namely gestational trophoblastic neoplasia, Wilms tumors in children, and rhabdomyosarcoma [12,13,14]. During our research on marine symbiotic actinomycetes, the Streptomyces griseus SCSIO PteL053 strain was isolated from a marine-derived gastropod mollusk Planaxis sp. The results of the bioinformatics analysis indicated the potential of the strain to produce bipyridine and actinomycin analogs. Here, we report (i) the whole-genome sequence of the marine invertebrate-associated strain SCSIO PteL053; (ii) the production of the known potential compounds bipyridine (2,2′-BP) and actinomycin (1–4) by 2,2′-Bipyridine and actinomycin BGCs; and (iii) two major known NP-encoded BGCs (termed, herein, cls and acd), which are confirmed by a bioinformatics-based approach, and the biosynthetic pathways for compounds 1–4.

2. Materials and Methods

2.1. General Methods

A 1260 infinity system (Agilent, Santa Clara, CA, USA) equipped with a Phenomenex Prodigy ODS (2) column (150 × 4.6 mm, 5 μm, Phenomenex Inc., Torrance, CA, USA) was used for HPLC-based analyses. A Primaide 1110 solvent delivery module equipped with a 1430 photodiode array detector (Hitachi, Tokyo, Japan) and a YMC-Pack ODS-A column (250 × 10 mm, 5 µm) was used for semi-preparative HPLC. A MaXis Q-TOF mass spectrometer (Bruker, Billerica, MA, USA) was used to acquire high-resolution mass spectral data. NMR spectra were recorded with an Avance-700 spectrometer (Bruker) at 700 MHz for ¹H nuclei and 125 MHz for ¹³C nuclei. Chemical shifts (δ) are provided with reference to tetramethylsilane [12]. X-ray single-crystal data were obtained with an XtaLAB PRO MM007HF X-ray diffractometer with APEX II CCD by Cu Ka radiation. Silica gel mesh 100–200 was used for column chromatography (CC). All chemicals and solvents used in this study were analytical and chromatographic grade.

2.2. Sample Collection and Strain Identification

The SCSIO PteL053 isolate was obtained from Planaxis sp., an oceanic invertebrate collected in Daya Bay, Guangdong Province, China (33′22° Latitude N; 31′15° Longitude E) using actinomycete isolation agar media, 0.01% L-asparagine, 0.2% D-mannitol, 0.01% FeSO₄·7H₂O, 0.01% MgSO₄·7H₂O, 0.05% K₂HPO₄, 3.0% CaCO₃, 3.0% sea salt, 1.8% agar, pH 7.0–7.2. The strain was further pure-cultured using the repeated streak-plate technique on a modified ISP2 agar plate (0.4% yeast extract, 0.4% glucose, 1% malt meal, 3.0% sea salt, 2% agar, pH 7.2–7.4) and kept in 15% (v/v) glycerol at −80 °C for subsequent study.

2.3. Whole-Genome Sequencing and In Silico Analysis

Complete-genome scanning and notation of strain SCSIO PteL053 were obtained using single-molecule real-time (SMRT) sequencing technology (PacBio) at Shanghai Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China. As a result, 296,471 filtered reads with a total data size of 3639599253 bp were generated. Afterward, they were assembled into the shortest contig using an ordered genome assembly operation (HGAP) [15]. The BGCs were analyzed using AntiSMASH (AntiSMASH 5.0, available at http://antismash.secondarymetabolites.org/ accessed on 22 September 2021) [15]. FramePlot (FramePlot 4.0 beta, available at http://nocardia.nih.go.jp/fp4/ accessed on 22 September 2021) was used to analyze the ORFs, whose functions were predicted using the online BLAST tool (http://blast.ncbi.nlm.nih.gov/ (accessed on 22 September 2021)).

2.4. Metabolic Profile of the Strain by LC-MS and Screening of Desired Compounds

Different production media were screened for the presence of bipyridine (2,2′-BP) and actinomycin analogs from the strain (Table S1). Initially, the strain was grown on an ISP2 agar medium. Subsequently, the 7-day-old mycelium and spore cultures were aseptically transferred into 50 mL of different media and incubated at 28 °C for 8 days with agitation at a rate of 180–200 rpm. Then, the whole cell culture was extracted with double the volume of 2-butanone, and the solvent phase was concentrated using the evaporation method. The concentrated crude extract was dissolved in methanol, and 30 µL of the sample was injected into the LC-HRESIMS analysis (a linear gradient from 10% to 100% MeCN in 30 min with a flow rate of 1.0 mL/min), with UV detection at 254 nm. The MS experiment was conducted in positive ion mode. The analytical conditions of the mass spectrometry were as follows: spray voltage, −3.50 kV; detector voltage, 1.65 kV; pressure of the TOF region, 1.6 × 10⁻⁴ Pa; pressure of the IT, 1.5 × 10⁻² Pa; rotary pump (RP) area vacuum, 70.0 Pa; drying gas pressure, 100.0 kPa; nebulizing gas (N₂) flow, 1.5 L/min; curved desolvation line (CDL) collision energy, 50%; and q = 0.251; scan range, m/z 100–2000 for MS.

2.5. Scale Up and Isolation of Active Compounds

Seed culture was prepared by inoculating the mycelium of Streptomyces griseus SCSIO PteL053 into a 250 mL conical flask consisting of 50 mL of suitable AM3 production medium with an agitation rate of 250 rpm at 28 °C for 2 days. Afterward, 10% of the culture was transferred into 200 mL of AM3 medium, and a total of 30 L was kept at 28 °C for 7 days at an agitation rate of 200 rpm. Then, the cell debris and supernatant were separated by centrifugation at 4000× g for 20 min and further solvent extraction was carried out with acetone and 2-butanone. The different organic extracts were concentrated using a rotational vacuum concentrator and the residues were combined. Further, these samples were introduced to silica column chromatography (CC; length 50 cm and diameter 7 cm) and eluted with CHCl₃-CH₃OH ratios (100:0, 98:2, 96:4, 94:6, 92:8, 90:10, 80:20, 70:30, 60:40, 50:50, (v/v) each fraction 500 mL) to provide ten fractions (Fr. A1–Fr. A10). Fractions A1 and A2 were pooled and subjected to another silica CC with petroleum ether–ethyl acetate ratios (100:0, 90:10, 80:20, 70:30, 60:40, 50:50, 40:60, 30:70, 20:80, 10:90, and 100:0 (v/v) each fraction 250 mL) to obtain the fractions Fr. B1–Fr. B11. The fractions Fr. B4–Fr. B8 were pooled and further purified by HPLC (semi-preparative) and eluted with 10% MeCN, with a flow rate of 2.5 mL/min and UV detection at 254 nm, which yielded compound 1 (4.5 mg) and compound 2 (5 mg). Similarly, fractions A3 and A4 were mixed and, subsequently, subjected to semi-preparative HPLC using the previously described method to yield compound 3 (9 mg) and compound 4 (3 mg) [11].

2.6. X-ray Crystallography of Compounds 1 and 2

White prismatic crystals of compounds 1 and 2 were acquired in the petroleum ether–ethyl acetate (50:50) ratio. High-quality crystal was taken and recorded using an XtaLAB PRO MM007HF X-ray diffractometer with APEXII CCD by Cu K_α radiation. The crystals were mounted at 99.99 (10) K during the data archive. Crystal structures were decoded using the ShelXT program with the dual-solution method and the structure was refined using full-matrix least-squares difference Fourier techniques [16].

3. Results and Discussion

3.1. Isolation and Identification of Strain

The strain SCSIO PteL053 (Figure S1) was obtained from the marine-derived gastropod mollusk Planaxis sp., an oceanic invertebrate in the Daya Bay in the South China Sea, using an ISP2 agar medium augment with 3% sea salt. The 16S rDNA analysis showed that the newly isolated strain has a 99.73% sequence similarity to the strain Streptomyces griseus (Figure S2). The 16S rDNA region was derived from the whole-genome sequence data.

3.2. Genome Sequencing and Elucidation of Strain SCSIO PteL053

The complete-genome sequence data set plays an essential role in the identification of novel compounds and also provides information about secondary metabolites [17,18]. The bacterial genome consists of a linear chromosome with 8,040,759 base pairs and the absence of plasmid, with average GC contents of 71.60% (Figure 1 and

Table 1. Genome features of S. griseus SCSIO PteL053.

Features	Value
Genome size (bp)	8,040,759
GC content (%)	71.60
Protein-coding genes	7077
Size of protein-coding gene (bp)	7,075,212
rRNAs	19
tRNAs	67

). The genome contains 67 tRNA, 19 rRNA, and 7077 coding genes.

The AntiSMASH [18] analysis revealed that the genome of the Streptomyces griseus SCSIO PteL053 encodes 28 BGCs for the different secondary metabolites (Figure 1 and Table S2). Most of the gene clusters were distributed in the two sub-telomeric regions of the genome [19]. The 28 BGCs spanned 0.75 Mb or 10.71% of the total genome, including two Type III PKSs, seven NRPSs, two PKS-NRPSs, four terpenes, two siderophores, two bacteriocin, three hybrid BGCs, two lanthipeptides, lasso peptides, and other types of BGCs (Table S2). Among these, NRPS cluster 6 is responsible for the synthesis of actinomycin D and PKS-NRPS cluster 19, whose products are collismycins.

3.3. Production, Isolation, and Elucidation of 2,2′-BP and Actinomycin Analogs

The LC-HRESIMS analysis of the fermentation broth culture of the isolate indicated that the desired compounds were effectively produced under standard culture conditions in an AM3 production medium (Figure S3). In order to isolate and identify the compounds, large-scale fermentation was conducted. Repeating the extractions and concentrations provided a good yield of crude extract (30 g), which was used for purification to yield compounds 1–4 (Figure 2A).

Compound 1 has the molecular formula C₁₃H₁₁N₃OS due to the HRESIMS data at m/z 258.0697 ([M + H]⁺, cald for 258.0696) and m/z 280.0515 ([M + Na]⁺, cald for 280.0515). The ¹H spectrum showed two distinct downfield-shifted methyl proton signals at δ_H 2.53 (3H, s, 4-OMe) and δ_H 4.14 (3H, s, 5-SMe), corresponding to a thiomethyl and methoxy group in the structure, which was in accordance with the ¹³C shifts at δ_C 18.6 (4-OMe) and δ_C 58.2 (5-SMe). The four olefinic protons in the spin system at δ_H 8.42 (1H, d, J = 8.0 Hz, H-2′), 7.97 (1H, m, H-3′), 7.49 (1H, dd, J = 6.9, 4.7, H-4′), and 8.68 (1H, d, J = 4.7 Hz, H-5′) showed the existence of a 2-substituted pyridine ring. The appearance of a singlet proton at δ_H 8.22 (1H, s, H-2), together with the chemical shifts of eleven carbons downfield of the ¹³C spectrum, indicated the possibility of a pyridine. The compound was determined to be SF2738D by comparing its formula and the chemical shifts of the NMR spectroscopic data (Figures S4 and S5, Table S3) with those reported previously [20], which was further verified by the X-ray analysis of the crystals of the compound (Table S4).

The HRESIMS ion peaks at m/z 244.0545 ([M + H]⁺, calcd for 244.0539) and m/z 266.0360 ([M + Na]⁺, calcd for 266.0359) supported the molecular formula of C₁₂H₉N₃OS for compound 2. The ¹H and ¹³C NMR data were similar to compound 1, except for the absence of the thiomethyl group and a non-protoned carbon at C-7, the existence of an additional sp² CH group at δ_H 9.16 (1H, s, H-7), and δ_C 158.2 (C-7). These examinations, together with the comparison of the ¹H and ¹³C NMR spectra (Figures S6 and S7, Table S3) with those reported previously, allowed the identification of the compound as SF2738F [20]. The structure of the compound was also demonstrated using X-ray analysis of the crystals of the compound, which were obtained from the petroleum ether–ethyl acetate (50:50) ratio.

For compounds 3 and 4, the molecular formulas were identified as C₆₂H₈₆N₁₂O₁₆ and C₆₂H₈₆N₁₂O₁₇ due to the HRESIMS peaks at m/z 1255.6355 and 1269.6269. The 62 carbon signals of compounds 3 and 4 in the ¹³C NMR spectrum, combined with the UV absorptions at 256 nm and 442 nm, indicated the actinomycin skeleton of the compounds. The structures of the compounds were determined to be actinomycin D (3) and actinomycin X_Oβ (4) from the comparison of the NMR data (Figures S8–S11, Table S4) with the compounds we isolated previously [21].

3.4. Bioinformatics Analysis of the Putative 2,2′-BP BGC

For further identification of the putative BGC for the biogenesis of SF2738D (1) and SF2738F (2), all 28 BGCs were screened. Genome cluster 19, named cls, revealed the combined PKS/NRPS gene product collismycin A, which showed an 81% gene sequence similarity to the already reported clm BGC from Streptomyces sp. CS40 [22]. The putative cls BGC contained 49 open reading frames (ORFs), among which, 41 were involved in SF2738D (1) and SF2738F (2) biosynthesis. The genetic organization of the cls BGC is shown in Figure 3 in which the different-colored genes indicate their proposed function shown in

Table 2. Determined functions of ORFs in the cls BGC.

ORF	Size	Proposed Function	ID/SI	Protein Homolog and Origin
clsR1	185	Putative TetR transcriptional regulator	98/98	ClmR1, (CCC55902.1); Streptomyces sp. CS40
clsU1	118	Uncharacterized protein	50/64	CUP04914.1; Flavonifractorplautii
clsT1	581	Putative ABC transporter, ATPase, and permease	99/100	ClmT1, (CCC55903.1; Steptomyces sp. CS40
clsT6	586	ABC transporter, ATP-binding protein	99/99	MdlB, WP_032768595.1; Streptomyces sp. CNS654
clsT7	569	ABC transporter, ATP-binding protein	100/100	MdlB, WP_032768595.1; Streptomyces sp. CNS654
clsT3	322	Putative ABC transporter, solute-binding protein	96/98	ClmT3, CCC55905.1; Steptomyces sp. CS40
clsT4	327	Putative ABC transporter, permease component	99/99	ClmT4, CCC55906.1; Steptomyces sp. CS40
clsT5	261	Putative ABC transporter, ATP-binding protein	98/99	ClmT5, CCC55907.1; Steptomyces sp. CS40
clsR2	168	LuxR family transcriptional regulator	98/98	ClmR2, CCC55908.1; Streptomyces sp. CS40
clsM2	342	O-methyltransferase	96/98	ClmM2, CCC55909.1; Streptomyces sp. CS40
clsD2	532	Putative dehydrogenase	97/98	ClmD2, CCC55910.1; Streptomyces sp. CS40
clsAT	544	Putative aminotransferase	98/98	ClmAT, CCC55911.1; Streptomyces sp. CS40
clsM1	328	S-Methyltransferas	97/98	ClmM1, CCC55912.1; Streptomyces sp. CS40
clsG1	462	GriD-like dehydrogenase component	96/98	ClmG1, CCC55913.1; Streptomyces sp. CS40
clsG2	343	GriC-like dehydrogenase component	98/99	ClmG2, CCC55914.1; Streptomyces sp. CS40
clsM	397	Putative FAD-dependent oxidoreductase	97/98	ClmM, CCC55915.1; Streptomyces sp. CS40
clsAH	413	Amidohydrolase family protein	96/98	ClmAH, CCC55916.1; Streptomyces sp. CS40
clsAH1	89	Amidohydrolase family protein	99/99	WP_032768579.1; Streptomyces sp. CNS654
clsAL	562	Putative acyl CoA ligase	98/98	ClmAL, CCC55917.1; Streptomyces sp. CS40
orf-20	63	Hypothetical protein	45/54	PAZ10747.1; Streptomyces sp. SA15
clsP	86	Putative free-standing acyl carrier protein	96/97	ClmP, CCC55918.1; Streptomyces sp. CS40
clsL	458	L-Lysine 2-aminotransferase	97/98	ClmL, CCC55919.1; Streptomyces sp. CS40
clsS	393	Monomeric sarcosine oxidase	97/97	ClmS, CCC55920.1; Streptomyces sp. CS40
clsN1	2552	Non-ribosomal peptide synthetase/polyketide synthase hybrid protein	97/98	ClmN1, CCC55921.1; Streptomyces sp. CS40
clsN2	1331	Putative non-ribosomal peptide synthetase	82/84	ClmN2, CCC55922.1; Streptomyces sp. CS40
clsT	242	Putative type II thioesterase	98/99	ClmT, CCC55924.1; Streptomyces sp. CS40
clsA	375	PD-(D/E)XK nuclease family protein	50/62	NEB63020.1; Streptomyces diastaticus
orf-28	257	Hypothetical protein	97/98	WP_032768566.1; Streptomyces sp. CNS654
orf-29	302	Hypothetical protein	98/98	WP_032768565.1; Streptomyces sp. CNS654
orf-30	302	Hypothetical protein	95/96	WP_032768565.1; Streptomyces sp. CNS654
clsB	224	GNAT family N-acetyltransferase	100/100	RimI, WP_050487033.1; Streptomyces sp. CNS654
clsC	354	Alpha/beta hydrolase	100/100	AXE1, WP_100562610.1; Streptomyces sp. CB02613
clsD	167	Helix-turn-helix transcriptional winged regulator	99/99	MarR, WP_093443708.1; Streptomyces sp.Cmuel-A718b
clsU2	-	Unknown function	-	-
clsE	436	3-phytase	99/99	SCF80346.1; Streptomyces sp. Cmuel-A718b
clsF	275	Endo alpha-1,4 polygalactosaminidase	99/100	WP_032768511.1; Streptomyces sp. CNS654
orf-37	220	hypothetical protein	98/99	WP_032768508.1; Streptomyces sp. CNS654
clsG	308	NAD(P)-dependent oxidoreductase	98/99	WcaG, WP_032768507.1; Streptomyces sp. CNS654
clsH	308	NAD(P)-dependent oxidoreductase	81/82	WcaG, WP_032768507.1; Streptomyces sp. CNS654
clsI	237	Nucleotidyltransferase family protein	100/100	GCD1, WP_032768506.1; Streptomyces sp. CNS654
orf-41	333	SDR family NAD(P)-dependent oxidoreductase	99/99	WP_032768503.1; Streptomyces sp. CNS654
clsJ	468	Hypothetical protein	99/99	WP_032768501.1; Streptomyces sp. CNS654
clsU3	-	Unknown function	-	-
clsK	129	DUF3492 domain-containing protein	100/100	RfaB, WP_032768499.1; Streptomyces sp. CNS654
clsW	60	DUF3492 domain-containing protein	93/93	RfaB, WP_043252613.1; Streptomyces vinaceus
clsX	41	DUF3492 domain-containing protein	100/100	RfaB, WP_032768499.1; Streptomyces sp. CNS654
orf (+1)	524	Hypothetical protein	84/85	HSNSD, WP_032768497.1; Streptomyces sp. CNS654
orf (+2) orf(+3)	176 175	Hypothetical protein Hypothetical protein	93/94 100/100	HSNSD, WP_096629522.1; Streptomyces sp. WZ.A104, WP_032768496.1; Streptomyces sp. CNS654

.

The putative cls BGC contained six core synthetic genes, clsAL, clsP, clsL, clsS, clsN1, and clsN2, which were conserved among 2,2′-BP-producing actinomycete species [22]. The biosynthesis of the compounds begins with the production of picolinic acid (PA) from l-lysine with the help of the clsL, clsS, and clsAL genes. Among them, the biosynthetic genes clsL and clsS encode l-lysine 2 aminotransferase and sarcosine oxidase enzymes, respectively.

ClsP was identical to a stand-alone acyl carrier protein (ACP) of collismycin A biosynthesis [22]. Generally, clsp is encoded within the cluster but in the case of actinomycin biosynthesis, it acts as a stand-alone process like AcmD [23]. In this strain, clsP was determined to encode ACP incriminate PA from ClsAL to ClsN1 in the biosynthetic pathway. Gene clsN1 contained seven domains, which were assembled into two modules: a PKS extender module consisting of ketosynthase (KS), acyltransferase (AT), and ACP domains and an NRPS module containing condensation or cyclization (Cy), adenylation (A), a peptidyl carrier protein (PCP), and epimerization (E) domains. Gene clsN2 encoded single-modular NRPS, which consisted of Cy-A-PCP domains. Genes clsN1 and clsN2 were very similar to the previously reported clmN1 and clmN2 in the collismycin A BGC [22]. These gene products may be used as a PA for the formation of ring B, a 2,2′-bipyridyl scaffold, for the biosynthesis of SF2738D and SF2738F. The stand-alone putative type II thioesterase encoded by the clsT gene was identical to the reported clmT of the collismycin A biosynthesis. This gene acted by releasing imperfectly added amino acids from the PCP domain of the NRPS protein or by detaching the acetyl species from the ACP domain to restore the free thiol groups [24]. Recently, a trans-acting flavoprotein-related assembly channel for 2,2′-bipyridine ring formation and differentiation was reported for dethiolated (CAE) and thiolated (COL) compounds [25]. Based on the deduced gene function correlations, we suggest that ClsT may participate in the release of the growing polyketide/non-ribosomal peptide chain from ClsN2.

The upstream region of the core biosynthetic gene begins from clsM2 to clsAH1, which is a total of nine genes. Among these, clsAH-encoded aminohydrolase was identical to the ClmAH gene from Streptomyces sp.CS40; ClmAH is mandatory for detaching the extra leucine residue incorporated during biosynthesis before the formation of collismycin C, which has a carboxylic group at the C6 position. The important role of ClmAH has been confirmed by mutation analysis [22]. ClmAH could be involved in the removal of the leucine residue before the formation of the oxime group.

Another set of upstream-region genes includes clsG1 and clsG2, which are encoded by two dehydrogenase components that are identical to ClmG1 and ClmG2 of collismycin A biosynthesis from Streptomyces sp.CS40. Similarly, ClmG1 and ClmG2 showed a high sequence homology of GriC and GriD involved in carboxylic acid reduction reactions in grixazone biosynthesis [26,27]. Based on this gene function correlation, the homology of ClsG1 and ClsG2 indicated that they were catalyzed by similar reactions during the formation of collismycin analogs. The determined product of the clsAT gene-encoded putative aminotransferase was identical to ClmAT of the reported collismycin biosynthesis from Streptomyces sp.CS40. The ClmAT domain comes under the class III family type of protein, which is associated with the pyridoxal-phosphate-mediated decarboxylation or transamination reaction of cationic amino acid and its derivatives [22,27]. Accordingly, it is clear that ClsAT is mainly involved in oxime group formation in the biosynthesis of collismycins. The clsD2 encodes dehydrogenase homology to ClmD2 in collismycin A BGC. ClsD2 is an essential component in the oxidization of the hydroxymethyl group in position 5 of collismycin analogs to the aldehyde group, which is the complex substrate for the action of aminotransferase by ClsAT [27]. This protein is hopefully involved in the oxime formation in the crucial step of collismycin biosynthesis and its function confers the recovery of the spontaneous generation of shunt products in the main pathway [27]. Gene clsM encodes oxidoreductase with a high-level resemblance to ClmM, a putative-FAD-dependent oxidoreductase, and monooxygenase [22]. ClmM controls the UbiH domain in 2-polyprenyl-6-methoxyphenol hydroxylases and also FAD-binding oxidoreductases, which showed that clsM encodes putative monooxygenase but is not involved in the 2,2′-bipyridyl structure [28]. ClsM is important for the subsequent steps of collismycin construction, perhaps by catalyzing the formation of the oxime group. Gene clsM1 encodes S-methyltransferase and clsM2 encodes O-methyltransferase. Based on the structures of SF2738D and SF2738F, clsM2 is associated with the generation of the methoxy group in position C-4 of collismycin, and clsM1 is involved in the conversion of SF2738D from intermediate/SF2738F. The determined product of clsAT gene-encoded putative aminotransferase showed a 98% similarity with ClmAT of the reported collismycin biosynthesis from Streptomyces sp. CS40. The ClmAT domain comes under the class III family type of protein, which is associated with the pyridoxal-phosphate-mediated decarboxylation or transamination process of cationic amino acid and its derivatives [22,27]. Accordingly, it is clear that ClsAT mainly participates in oxime group formation in the biosynthesis of collismycin compounds 1 and 2. These upstream genes are involved in tailoring the enzymatic reactions for SF2738D and SF2738F biosynthesis. Gene clsAH1 encodes putative amidohydrolase. The amidohydrolase is included in the Metallo-dependent hydrolase family of proteins, which participates in the hydrolysis of the C-N bond (peptic and non-peptic) through the nucleophilic attack of a hydroxyl group [22,29]. According to the proposed pathway, the action of the enzyme is involved in the dehydration of the aldoxime intermediate from collismycin A, resulting in the formation of SF2738D and SF2738F. Nitrile groups, as part of the structure of natural products, have been studied [30,31] and a route for the nitrile group formation has been proposed [22,32].

The sets of genes that reside on the left-hand side of the cls gene cluster include two regulatory genes and six transporter genes. The detailed annotation indicated that two genes, clsR1 and clsR2, which encode the putative TetR and LuxR family transcriptional regulators, were identical to ClmR1 and ClmR2 in collismycin A biosynthesis from Streptomyces sp. CS40 [22,33]. The putative gene products also showed strong similarities to CrmR1 and CrmR2 from Actinoalloteichus cyanogriseus WH1-2216-6 [28,34]. TetR regulators act as a transcriptional repressor, which regulates drug metabolism, cellular metabolic activity, and collismycin biosynthesis through the inhibition of biosynthetic genes by clsR2 repression. Similarly, a homolog of the clmR1 protein encoded a transcriptional regulator of the TetR protein, which has been reported previously [22]. ClsR2 is involved in the positive regulation of collismycin product formation [33]. The putative transporter genes clsT1, clsT3, clsT4, and clsT5 showed strong similarity with consecutive genes from Streptomyces sp. CS40 [22]. Based on the correlation, the gene products are co-transcribed and interact to form whole ABC cassettes for a single collismycin transport system. In addition, the two genes clsT6 and clsT7 both encoded ABC transporter ATP-binding domains; their determined gene products have been previously identified as membrane-associated proteins (MdlB) in Pseudomonas putida [35]. The cls BGC also consists of additional genes that reside throughout the gene cluster and encode unrelated proteins, which are hypothetically supported in collismycin biosynthesis.

Based on the available information and determined functions of genes in the cls BGC, we propose putative biosynthetic pathways for SF2738D (1) and SF2738F (2) (Figure 4). Three different stages are involved in the assembly of SF2738D (1) and SF2738F (2). In the first phase, skeleton assembly is initiated by biosynthesis and the activation of PA from a precursor, l-lysine, with the assistance of three enzymes ClsL, ClsS, and ClsAL. Afterward, the activated picolinic CoA is supplied to the hybrid PKS/NRPS system by a stand-alone ClsP. Therefore, collismycin’s first pyridine ring is formed. In the second phase, the hybrid PKS/NRPS system (ClsN1 and ClsN2) is associated with the incorporation of a malonate unit, cysteine, and extra leucine residue to provide 2,2′-BP-l-leucine. The N-acetyl group of cysteine is ultimately attached to PA via the carbonyl group to make C2-N1 bonding, and its intermediate undergoes a further intramolecular cyclization reaction to provide a seven-membered sulfur-containing heterocycle ring. Later, a six-membered intermediate is formed by the release of the SH group via a subsequent rearrangement process for the biosynthesis of the skeleton structure. A similar type of reaction has been reported for the biosynthesis of caerulomycins and collismycins [28]. In the end phase, the leucine residue may be eliminated by ClsAH, and then the reduction of the carboxylic group to an aldehyde group by ClsG1 and ClsG2. The oxime group formation is catalyzed by ClsAT and the final methylation of the hydroxyl group at the C4 by ClsM2 to yield collismycin A [22,27]. Collismycin A is catalyzed by ClsAH1 and ClsM1 to obtain the final products SF2738D (1) and SF2738F (2).

3.5. Bioinformatics Analysis of the Putative Actinomycin BGC

The complete genome of S. griseus SCSIO PteL053 was used to locate the entire BGC coding for potent antitumor actinomycin biosynthesis by computational analysis. Genomic DNA with a size of 50 kb was predicated as a putative actinomycin BGC (termed, herein, acd). Among these data, 20 open reading frames (ORFs) were proposed to be involved in the biosynthesis of actinomycin. The genetic organization of the strain was compared with known actinomycin producers [11,16,36,37] (Figure 5) and the determined gene product functions were noted (

Table 3. Determined functions of ORFs in the acd BGC.

ORF	Size	Proposed Function	ID/SI	Protein Homologue and Origin
acdU3	210	Hypothetical protein	30/49	GLYR1; Drosophila pseudoobscura (Q29NG1.2)
acdU4	187	Hypothetical protein	27/36	SWS; Drosophila mojavensis (B4L535.1)
acdD	66	MbtH protein	63/77	MbtH; Mycobacterium tuberculosis variant bovisAF2122/97(P59965.1)
acdE	78	4-MHA carrier protein	25/54	AflC; Aspergillus parasiticus SU-1 (Q12053.1)
acdN1	410	Acyl–CoA ligase	44/60	Schizosaccharomyces pombe 972h- (O74976.1)
acdN2	2537	Non-ribosomal peptide synthetase	44/60	DhbF; Bacillus subtilis subsp. subtilis str. 168 (P45745.4)
acdN3	4250	Non-ribosomal peptide synthetase	29/45	Metarhiziumrobertsii ARSEF 23 (E9FCP4.2)
acdF	210	Hypothetical protein(DA)	29/45	Schizosaccharomyces pombe 972h (O13799.1)
acdG	299	Arylformamidase	34/50	Afmid; Danio rerio (Q566U4.2)
acdH	285	Tryptophan 2,3-dioxygenase	43/58	KynA; Polaromonas sp. JS666 (Q126P7.1)
acdL	420	Kynureninase	44/62	KynU; Pseudomonas fluorescens (P83788.1)
acdM	347	Methyltransferase	54/70	AcmL; Streptomyces lavendulae NRRL 11,002 (ABI22137.1)
acdP	435	Cytochrome P450	49/67	CypA; Saccharopolyspora erythraea NRRL 2338 (P33271.1)
adnU	63	Ferredoxin	45/60	SoyB; Streptomyces griseus (P26910.1)
acdO	216	LmbU-like protein	99/99	LmbU: Streptomyces parvus (WP_167533479)
acdR	281	TetR family transcriptional regulator	35/53	TetR; Vibrio anguillarum (P51560.1)
acdQ	289	Siderophore-interacting protein	33/48	Mb2919c; Mycobacterium tuberculosis variant bovis AF2122/97 (P65050.1)
acdT1	327	ABC transporter, ATP-binding protein	46/60	DrrA; Streptomyces peucetius(P32010.1)
acdT2	255	ABC-2-type transporter	27/44	DrrB; Streptomyces peucetius (P32011.1)
acdT3	753	ATP-binding protein	53/70	UvrA; Methanothermobacterthermautotrophicus str. Delta H (O26543.1)

). They showed a strong similarity to homologs of previously reported actinomycin BGC (acd) from Streptomyces chrysomallus [16].

The acd cluster contained five consecutive NRPS genes (acdD, acdE, acdN1, acdN2, and acdN3), which are the homologs of acnR, acnD, acnA, acnB, and acnC in S. chrysomallus responsible for the synthesis of the 4-MHA-pentapeptide halves for actinomycin. This pentapeptide was synthesized from the precursor unit, a 4-methyl-3-hydroxy-anthranilic acid (4-MHA), along with five different amino acids (l-threonine, d-valine, l-proline, sarcosine, and l-methylated valine). Similarly, the action of five consecutive NRPS genes has been reported in the biological synthesis of actinomycin D from mangrove-derived Streptomyces costaricanus SCSIO ZS0073 [11]. The AcdD protein is homologous to the MbtH protein, which is similar to the protein from the Mycobacterium tuberculosis variant bovis [38]. It is found in many NRPS gene clusters and is involved in adenylation reactions [39,40,41]. The acdN1 gene encoded the adenylation protein and the acdE gene encoded the 4-MHA carrier protein. The acdN2 and acdN3 genes encoded the non-ribosomal peptide synthetase enzyme, which was indicated in the peptide chain assembly and release process from the NRPS system. In the downstream of the NRPS, there were four other consecutive genes, acdG, acdH, acdL, and acdM, which had high similarity with enzymes involved in the biosynthesis of 4-MHA. These gene homologs have been shown to play a role in the 4-MHA biosynthesis process in the actinomycin-producing strain S. antibioticus IMRU3720 [16]. The acdF gene was predicted to be involved in the dimerization of actinomycin halves during actinomycin synthesis according to recent reports [11].

A group of genes, including acdO, acdR, acdQ, acdT1, acdT2, and acdT3, located at the end of the right-arm cluster, had regulatory and protective roles. Within the acd BGC, the acdO gene encoded the LmbU regulatory protein. The LmbU gene was previously shown to be involved in lincomycin biosynthesis in S. lincolnensis 78-11 [42]. In addition, the inactivation of the homologous acnO gene in a mutant strain of S. costaricanus SCSIO ZS0073 showed a complete shutdown of actinomycin D production, which demonstrates the importance of the positive regulatory role of the gene [11]. The acdR gene encoded the TetR family of proteins, which play a transcriptional activator and repressor role to regulate the production of secondary metabolites [43,44]. As previously indicated, the trdK gene was found to be highly similar to the TetR protein associated with tirandamycin synthesis in Streptomyces sp. SCSIO 1666 and its inactivation led to an increase in tirandamycin titers [45]. Apart from this, the essential role of acnR in actinomycin D biosynthesis in S. costaricanus SCSIO ZS0073 has been validated by gene inactivation studies [11]. Apart from these genes in the acd cluster, we found three putative transporter genes, namely acdT1, acdT2, and acdT3. Based on the analysis, the acdT1-encoded ATP-binding cassette (ABC) transporter was found to be similar to the drrA gene from daunorubicin-producing Streptomyces peuceticus [46]. Likewise, acnT2, which encoded an ABC-type 2 transporter permease, was found to be similar to the drrB gene from Streptomyces peuceticus [46,47]. acnT3 encoded a 753 amino acid DNA-binding domain that was similar to the UvrA-like protein found in E. coli [48]. The inactivation of three genes, acnT1, acnT2, and acnT3, in the acn BGC resulted in no significant impact on actinomycin D biosynthesis in S. costaricanus SCSIO ZS0073 [11]. Accordingly, we suggest that these genes are mainly involved in the transport of actinomycin analogs to the environment but not in actinomycin biosynthesis. The gene acnQ located upstream of the transporter genes was predicted to encode siderophore-interacting proteins to regulate the intracellular concentrations of metabolites. Hence, the inactivation of acnQ had no impact on actinomycin D production, which has recently been reported in S. costaricanus SCSIO ZS0073 [11]. Based on the available data, we suggest a putative biosynthetic pathway leading to the production of actinomycin D (3) and actinomycin X_Oβ (4) in S. griseus SCSIO PteL053 (Figure 6). The acd BGC also consisted of a few additional genes throughout the gene cluster that encoded unrelated proteins, which are either directly or hypothetically involved in biosynthesis.

4. Conclusions

In the present study, we attained the whole genome sequence of S. griseus SCSIO PteL053, whose average GC content was 71.6%. The marine invertebrate-associated strain harbors 28 putative BGCs. Here, we characterized two putative BGCs, which were predicted to generate pharmaceutically important 2,2′-BP and actinomycin analogs. The study also provided good evidence to show that the isolated strain has a close association with marine invertebrates and can protect these soft-bodied organisms from bacterial infection through the production of prevalent antibiosis metabolites. These interactions are driven by the biosynthesis of specialized secondary metabolites, which exhibit biological activity and contribute to the chemical diversity in highly diverse environments [5,49]. Understanding these mutualistic associations and the ecology of NPs can provide more insights into the future direction of NP discovery. Furthermore, this work highlights the importance of genome mining techniques in identifying BGCs and potential lead metabolites for drug discovery from wild strains.