Synergy between Genome Mining, Metabolomics, and Bioinformatics Uncovers Antibacterial Chlorinated Carbazole Alkaloids and Their Biosynthetic Gene Cluster from Streptomyces tubbatahanensis sp. nov., a Novel Actinomycete Isolated from Sulu Sea, Philippines

Streptomyces tubbatahanensis DSD3025T is a novel actinomycete with antibiotic and anticancer activities from Tubbataha Reefs Natural Park, a United Nations Educational, Scientific and Cultural Organization World Heritage Site in Sulu Sea and considered one of the Philippines’ oldest and most-well-protected marine ecosystems. In silico genome mining tools were used to identify putative BGCs that led to the discovery of genes involved in the production of halogenated carbazole alkaloids and new natural products. By integrating bioinformatics-driven genome mining and metabolomics, we unearthed the hidden biosynthetic richness and mined the associated chemical entities from the novel Streptomyces species. ABSTRACT In this study, a novel actinomycete strain, DSD3025T, isolated from the underexplored marine sediments in Tubbataha Reefs Natural Park, Sulu Sea, Philippines, with the proposed name Streptomyces tubbatahanensis sp. nov., was described using polyphasic approaches and characterized using whole-genome sequencing. Its specialized metabolites were profiled using mass spectrometry and nuclear magnetic resonance analyses, followed by antibacterial, anticancer, and toxicity screening. The S. tubbatahanensis DSD3025T genome was comprised of 7.76 Mbp with a 72.3% G+C content. The average nucleotide identity and digital DNA-DNA hybridization values were 96.5% and 64.1%, respectively, compared with its closest related species, thus delineating the novelty of Streptomyces species. The genome encoded 29 putative biosynthetic gene clusters (BGCs), including a BGC region containing tryptophan halogenase and its associated flavin reductase, which were not found in its close Streptomyces relatives. The metabolite profiling unfolded six rare halogenated carbazole alkaloids, with chlocarbazomycin A as the major compound. A biosynthetic pathway for chlocarbazomycin A was proposed using genome mining, metabolomics, and bioinformatics platforms. Chlocarbazomycin A produced by S. tubbatahanensis DSD3025T has antibacterial activities against Staphylococcus aureus ATCC BAA-44 and Streptococcus pyogenes and showed antiproliferative activity against colon (HCT-116) and ovarian (A2780) human cancer cell lines. Chlocarbazomycin A exhibited no toxicity to liver cells but moderate and high toxicity to kidney and cardiac cell lines, respectively. IMPORTANCE Streptomyces tubbatahanensis DSD3025T is a novel actinomycete with antibiotic and anticancer activities from Tubbataha Reefs Natural Park, a United Nations Educational, Scientific and Cultural Organization World Heritage Site in Sulu Sea and considered one of the Philippines’ oldest and most-well-protected marine ecosystems. In silico genome mining tools were used to identify putative BGCs that led to the discovery of genes involved in the production of halogenated carbazole alkaloids and new natural products. By integrating bioinformatics-driven genome mining and metabolomics, we unearthed the hidden biosynthetic richness and mined the associated chemical entities from the novel Streptomyces species. The bioprospecting of novel Streptomyces species from marine sediments of underexplored ecological niches serves as an important source of antibiotic and anticancer drug leads with unique chemical scaffolds.

Actinobacteria library collection reported from the marine sediments collected from different sampling sites in the Philippine archipelago (21).
The antibiotic and anticancer potentials of S. tubbatahanensis DSD3025 T were explored via genome mining, metabolite profiling, bioinformatics, and bioassays. The S. tubbatahanensis DSD3025 T genome encodes 29 putative BGCs, including a BGC of interest that contains tryptophan halogenase and flavin reductase genes, a two-component halogenase/reductase system. The bioinformatics analysis and BGC evaluation predicted compounds belonging to PKS I, PKS II, and NRPS moieties. Metabolite profiling of S. tubbatahanensis DSD3025 T crude extract by using ultra performance liquid chromatography-electrospray ionization-time of flight mass spectrometry (UPLC-ESI-TOF-MS) and nuclear magnetic resonance (NMR) analysis identified chlocarbazomycin A compound 1, a recently discovered halogenated carbazole alkaloid, from its closely related Streptomyces species as a metabolite. Aside from compound 1, several halogenated compounds were produced by S. tubbatahanensis DSD3025 T . The extract induced cell death via membrane damage of multidrug-resistant Staphylococcus aureus ATCC BAA-44 and inhibited growth proliferation of MCF-7, HCT-116, and A2780 cancer cell lines. Moreover, compound 1 had antibacterial activity against S. aureus ATCC BAA-44 and Streptococcus pyogenes and showed antiproliferative activity against HCT-116 and A2780 cancer cell lines. Based on genome annotation data and bioinformatics analysis, we have propose the biosynthetic pathway of compound 1 in S. tubbatahanensis DSD3025 T .

RESULTS AND DISCUSSION
Phylogenetic analysis of S. tubbatahanensis DSD3025 T . The complete 16S rRNA gene sequence of S. tubbatahanensis DSD3025 T (1,531 bp) was obtained from its wholegenome sequence. S. tubbatahanensis DSD3025 T shares 16S rRNA gene sequence similarity with Streptomyces species, particularly Streptomyces diacarni LHW51701 T (99.73%) and Streptomyces reniochalinae LHW50302 T (98.32%), which are marine sponge-derived Streptomyces (24). In the maximum likelihood phylogenetic tree (Fig. 1A), S. tubbatahanensis DSD3025 T formed a monophyletic lineage with S. diacarni LHW51701 T , and their nearest neighbor was S. reniochalinae LHW50302 T . This topology was consistent with that observed in phylogenetic trees analyzed using neighbor-joining and maximum parsimony algorithms (see Fig. S1 in the supplemental material). In contrast, the multilocus sequence analysis (MLSA) of concatenated housekeeping gene sequences revealed that S. tubbatahanensis DSD3025 T formed a distinct lineage from its closely related Streptomyces species in the phylogenetic trees based on the three algorithms ( Fig. 1B; Fig. S2).
Based on MLSA, the evolutionary distance between S. tubbatahanensis DSD3025 T and its closely related Streptomyces species was 0.0107 ( Fig. 2A), which was above the cutoff value of 0.007 to delineate novel species under the genus Streptomyces (25). To further validate the species delineation of S. tubbatahanensis DSD3025 T , the digital DNA-DNA hybridization (dDDH) and orthologous average nucleotide identity (orthoANI) values were calculated to validate the taxonomic status as a novel species. The dDDH values between S. tubbatahanensis DSD3025 T and its closely related matches ranged from 22.1 to 64.1% ( Fig. 2A), which were below the threshold value of 70% (26). Notably, S. tubbatahanensis DSD3025 T and its nearest species neighbor, S. diacarni LHW51701 T , had a dDDH value of 64.1%. In addition, S. tubbatahanensis DSD3025 T showed an orthoANI value of 96.49% with S. diacarni LHW51701 T (Fig. 2B). In a recent study, a 70% dDDH corresponded to approximately 96.7% and not the 95 to 96% cutoff point in delineating Streptomyces species (27). These findings suggested that S. tubbatahanensis DSD3025 T is a novel species based on the MLSA, dDDH, and orthoANI values between S. tubbatahanensis DSD3025 T and its closest match, S. diacarni LHW51701 T . The taxonomic status of S. tubbatahanensis DSD3025 T as a novel strain was further validated using phenotypic, chemotaxonomic, and genomic analyses.
Genome annotation and analysis. The whole genome of S. tubbatahanensis DSD3025 T , consisting of 7,760,770 bp and with a 72.3% G1C content (Table 1) Only bootstrap values above 50% (percentages of 1,000 replications) are indicated. Asterisks and crosses indicate branches that were also found using neighbor-joining and the maximum-parsimony method, respectively. of S. tubbatahanensis DSD3025 T was assembled as a single large contig (Fig. 3A). It contains 6,579 genes further divided into 4,313 operons. In the genome of S. tubbatahanensis DSD3025 T , 6,567 genes were counted as protein coding sequences (CDS), 70 as RNA genes, 6 as rRNA genes, and 69 as tRNA genes. Functional analysis by clusters of orthologous genes (COGs) revealed that 6,120 were assigned COG categories. Most of the predicted COGs had unknown functions (2,214), as shown in Fig. 3B. The other remaining 3,906 COG functional genes were largely assigned to transcription (634), carbohydrate transport and metabolism (439), and amino acid transport and metabolism (424). To some extent, the COGs with unknown functions required further classification and validation to determine their function.  Eleven genomic islands were identified in the strain's genome (Table S1), linking specialized metabolism to functional adaptation (28). Mobility genes, such as transposases and integrases, were detected (Table S2) and were associated with the movement of mobile genetic elements. Based on RAST annotation, 320 subsystems were identified where the majority of the genes were linked to amino acid and derivative metabolism (18.8%) and carbohydrate metabolism (16.1%), followed by protein metabolism (11.9%) (Fig. S3). ARTS predicted 392 core genes and 40 known resistance model hits in the S. tubbatahanensis DSD3025 T genome (Table S3). The in silico resistome analysis of S. tubbatahanensis DSD3025 T genome revealed drug resistance against 13 drug classes with six resistance mechanisms (Table S4), which was similar to closely related Streptomyces species. The resistance genes encoded in the S. tubbatahanensis DSD3025 T genome unveiled self-resistance against different antibiotic compounds, an important feature of antibiotic-producing strains to avoid suicide (29). These predicted resistance genes in DSD3025 T were not tested in vitro in this study. A total of 29 biosynthetic gene clusters associated with specialized metabolite production were predicted in the strain's genome (Fig. 3C), indicating the richness of the biosynthetic gene clusters, which are further discussed below.
Phenotypic analysis. Colonies on marine medium 1 (MM1) agar at 14 days of incubation formed well-developed aerial and substrate mycelium, which are the phenotypic features of Streptomyces. Spores were not evident in the scanning electron microscopy (SEM) image, but the intertwining of hyphal cells was observed (Fig. 4). Growth of S. tubbatahanensis DSD3025 T was abundant on MM1 agar, yeast extract-glucose-calcium carbonate (YGC) medium, nutrient agar (NA), tryptic soy agar (TSA), International Streptomyces Project 2 (ISP2) and ISP3 media, with white aerial mycelium and light to deep orange-yellow substratum mycelium (Table S5). S. tubbatahanensis DSD3025 T had moderate growth in MM11 and ISP9. Poor growth was observed in mannitol-containing agar (MM3) and ISP4. No diffusible pigment was produced by the strain in the different culture media. The API ZYM is a system used to detect selected enzymes in S. tubbatahanensis DSD3025 T . After 4 h of incubation, the results were assessed based on the standard reading table of the API ZYM The circular map of the S. tubbatahanensis DSD3025 T genome, displaying contig information in the outer circle, while the next adjacent circles show the distribution of COG genes in the forward and reverse strands, respectively. The next circle displays the rRNA and tRNA, followed by a circle for GC skew, with higher-than-average values displayed in green and lower-than-average values displayed in red. The next inner circle displays the GC ratios, with higher-than-average values in blue and lower-than-average values in yellow. (B) Clusters of Orthologous Groups (COGs) functional features. (C) Visualized locations of the 29 specialized metabolite BGC regions in the linearized genome of S. tubbatahanensis DSD3025 T , as predicted by antiSMASH v.7.0. BGC regions ranged from 7,831 to 106,511 nucleotides. system (Fig. S4). Of 19 enzymatic reactions, S. tubbatahanensis DSD3025 T showed seven positive enzymatic activities for alkaline phosphate, esterase (C 4 ), leucine arylamidase, acid phosphatase, and naphthol phosphohydrolase and weakly positive activities for esterase and lipase (C 8 ) and a-glucosidase enzymes (Table S6). The enzymatic activities expressed by different species of Streptomyces vary and can therefore be used to differentiate Streptomyces species from their closely related matches.
S. tubbatahanensis DSD3025 T had observed growth in marine medium formulated with 0 to 15% NaCl. Growth was observed at 28°C and 37°C and over a pH range of 4.0 to 10.0 ( Table 2). The closely related Streptomyces matches of S. tubbatahanensis DSD3025 T had similar observed growth in the 0 to 15% NaCl range and at 28°C and 37°C. The NaCl tolerance test demonstrated that S. tubbatahanensis DSD3025 T was not an obligate marine Streptomyces but had an adaptive capacity to grow in up to 15% (wt/vol) NaCl. Notably, S. tubbatahanensis DSD3025 T showed abundant growth at pH 4.0, in contrast with its neighboring Streptomyces species.
In silico specialized metabolite analysis of BGC in S. tubbatahanensis DSD3025 T . The 29 clusters predicted by antiSMASH (v.7.0) were associated with the production of BGCs consisting of nonribosomal peptides, indole biosynthesis, siderophore, type-1 (T1PKS) and type-2 (T2PKS) polyketides, posttranslationally modified peptides (RiPP-like), thiopeptides, and lanthipeptides (Table 4). BGCs 9, 10, and 21 have 100% similarity with known BGCs encoding desferrioxamine E, ectoine, and marineosin A and B, respectively. The remaining clusters with ,70% similarity to known specialized metabolites were identified as NRPS, indole, terpene, T1PKS, T2PKS, siderophore, lanthipeptides, betalactone, and thiopeptides. Notably, BGC regions 2, 3, 11, 15, 18, and 27 had no similarity with the reference specialized metabolites, indicating the novelty of the BGCs in S. tubbatahanensis DSD3025 T . Since S. diacarni LHW51701 T was the nearest neighbor in the phylogenetic analysis, we compared its BGCs to S. tubbatahanensis DSD3025 T . We noted similar BGCs encoded between two species for the biosynthesis of marineosin A and B, desferrioxamine E, ectoine, reductasporine, SapB, ebelactone, and hopene (31). We explored the biosynthetic potential of S. tubbatahanensis DSD3025 T by examining the putative core structures of the compounds produced by the NRPS and some PKS BGCs (using PRISM4) (Fig. S6), as well as the predicted core peptides of the RiPPs (detected using BAGEL4) (Fig. S7). These analyses highlighted the structurally diverse biosynthetic potential of S. tubbatahanensis DSD3025 T . Based on BGC analysis and comparison to known databases, the BGC region 1 of S. tubbatahanensis DSD3025 T resembled the BGC architecture of recently discovered cyclic halogenated lipoglycopeptide gausemycins A and B from Streptomyces (32). The unique structural features of the gausemycins A and B gene clusters were found in S. tubbatahanensis DSD3025 T BGC region 1, including the core gauA, gauB, gauC, and gauD NRPS gene homologs, as well as the incorporation of 4-chlorokynurenine (4-Cl-Kyn) (33) and 2-amino-4hydroxy-4-phenylbutyric acid. The halogenation and conversion of tryptophan to 4-Cl-Kyn were facilitated by a tryptophan 6-halogenase (tar14) (33), flavin reductase, tryptophan-2,3dioxygenase, and a putative a/b-hydrolase, for which the homologs were found in BGC region 1 of S. tubbatahanensis DSD3025 T .
Interestingly, an FADH 2 -dependent tryptophan halogenase and an associated flavin reductase were identified in BGC region 1 of S. tubbatahanensis DSD3025 T (Fig. 5A). This two-component halogenase/reductase system was not detected in S. diacarni LHW51701 T , thus indicating differences in the biosynthesis of halogenated compounds between the two related strains. The identified tryptophan halogenase homolog had an approximate size of 533 amino acids (Fig. 5B) that may install chlorine or bromine on tryptophan substrates with high regioselectivity and specificity.
The conserved motifs of this FADH 2 -dependent halogenase (FDH) included a putative flavin-binding motif (GxGxxG) located at the amino-terminal end (34), a C-terminal motif (WxWxIP) found in the middle of the sequence (34) that prevents monooxygenase substrate binding near the flavin (35,36), and other conserved motifs identified in the halogenase genes of motifs such as GVGEAT and FVEPLESTG (37,38). The lysine and glutamate catalytic residues highly conserved in FDHs were also observed in the tryptophan halogenase homolog sequence of S. tubbatahanensis DSD3025 T . The phylogenetic analysis revealed that the tryptophan halogenase of S. tubbatahanensis DSD3025 T formed a separate clade, together with tar14 and other bacterial C-6 halogenases (sttH, ktzR, and th-Hal) (39-41) (Fig. S8), suggesting that the tryptophan halogenase of S. tubbatahanensis DSD3025 T was a tryptophan FDH that may catalyze site-selective C-6 halogenation of specialized metabolites instead of the C-7 halogenation annotated from the whole-genome sequence analysis. The flavin reductase found in the same gene cluster was essential for the enzymatic cofactor regeneration of FADH 2 needed by the tryptophan halogenase to function (42).
Metabolite profiling of S. tubbatahanensis DSD3025 T extract revealed halogenated compounds. To gain insights into the specialized metabolites biosynthesized by S. tubbatahanensis DSD3025 T , the crude extract was analyzed using ultra performance liquid chromatography-electrospray ionization-quadrupole time-of-flight mass spectrometry (UPLC-ESI-QTOF-MS). At least six mass ion peaks of S. tubbatahanensis DSD3025 T extract in the negative mode showed a 3:1 isotopic pattern of chlorine and a 1:1 isotopic pattern of bromine ( Fig. S9 and S10), suggesting that S. tubbatahanensis DSD3025 T produces halogenated specialized metabolites.
The  Fig. S14 to S17). Remarkably, 6 out of 7 compounds isolated from S. diacarni SCSIO 64983 T , including 1, 2a, and 2b, were found in S. tubbatahanensis DSD3025 T extract in the negative mode (Table  S7). The metabolite similarities observed between S. tubbatahanensis DSD3025 T and the recently reported S. diacarni SCSIO 64983 T indicated that these Streptomyces species belong to one major clade (96% bootstrap replicates), based on a reconstructed phylogenetic tree (  S20) from S. tubbatahanensis DSD3025 T that clustered with chlocarbazomycin A (56.8% similarity) not found in closely related S. diacarni strains. The extensive search of its neutral chemical formula showed no hits in MarinLit, ChemSpider, PubChem, or AntiBase, suggesting that the compound is a possible new specialized metabolite with halogenation located at C-4 of carbazole ring A.
In the MS positive mode, five mass ion peaks with chlorine isotopic patterns were identified from S. tubbatahanensis DSD3025 T extract (Table S8 and  showed no hits for known microbial natural products in MarinLit, ChemSpider, or AntiBase, suggesting new chlorinated specialized metabolites. Additionally, the MS/MS results for these unknown chlorinated compounds showed distinct fragment peaks ( Fig. S26 to S29), which indicated that they are structurally unique. However, elucidating their structure ID to confirm structural uniqueness requires validation via NMR analyses.
Isolation and structure elucidation of halogenated carbazole alkaloid. Given that S. tubbatahanensis DSD3025 T produced halogenated metabolites based on MS analysis, it was envisaged that a flavin-dependent tryptophan 6-halogenase catalyzed the halogenation of these specialized metabolites (Fig. 6). Carbazole alkaloids utilize tryptophan as a precursor in their biosynthesis (44). The indole nucleus in tryptophan contributes to rings B and C during carbazole biosynthesis; thus, the detected tryptophan 6-halogenase encoded in the S. tubbatahanensis DSD3025 T genome may facilitate the halogenation of carbazole alkaloids at the C-7 of ring C (Fig. 6). To test this hypothesis, a mass-directed purification of Streptomyces tubbatahanensis DSD3025 T crude extract was performed.
The halometabolite detected at m/z 232.0521 [M 1 H] 1 putatively identified as compound 1 was set as the target mass for isolation due to its relatively higher abundance compared to other halometabolites in S. tubbatahanensis DSD3025 T extract. The purification of 1 g S. tubbatahanensis DSD3025 T extract (Fig. S30) afforded 4.73 mg (0.5% yield) of DSD3025H1, which was later confirmed as compound 1 via high-resolution MS (HRMS), MS/MS, and NMR analyses ( Fig. S31 and S32). The purification of 1 yielded 80% purity as estimated via 1 H NMR. Notably, Streptomyces tubbatahanensis DSD3025 T is a relatively more prolific producer of halometabolites, producing 7 times more than its related strain, Streptomyces diacarni LHW51701 T (45).
Compound 1 was unambiguously identified by NMR spectroscopic data (Table S9) based on 1 H, 1 H-decoupled, 13 C, 13 C DEPTQ135, 1 H-1 H correlation spectroscopy (COSY), heteronuclear single quantum coherence (HSQC), heteronuclear multiple bond correlation (HMBC), and nuclear Overhauser effect spectroscopy (NOESY) NMR spectral analyses ( Fig. S33 to S48). The purity of the sample was approximately 80%, as determined by dividing the integral value of the 2OCH 3 peak of the compound by the total integral value of the 2OCH 3 peak of the compound and methyl peak of fatty acid impurities in the aliphatic region (2.0 to 2.15 ppm) of the 1 H NMR spectrum (Fig. S34). Data from J value, COSY, and NOESY were used to determine the position of chlorine in the elucidated compound. NOESY data showed correlations between 2NH and H1 (Fig.  S47), indicating that H1 was in proximity with 2NH and that chlorine was positioned in C-4 of carbazole ring A. In addition, NOESY data demonstrated a correlation between H2 and 2OCH 3 , suggesting that these protons were also in proximity and chlorine was adjacent to the 2OCH 3 of 1. These NOESY correlations indicated that the halogenation occurred at C-4 of carbazole ring A, indicating that compound 1 isolated from S. tubbatahanensis DSD3025 T has the structure of chlocarbazomycin A, as reported previously from its closely related Streptomyces strain, S. diacarni LHW51701 T (43,45). These NOESY correlations were verified by measuring the distances between NH and H1 (2.8 Å) and between H2 and 2OCH3 (1.9 Å) using Chemdraw 3D v21.0.0 (Fig.  S48). The correlation between H1 and H2 identified in the COSY spectrum (Fig. S41) and the calculated J value of 8.7 Hz from the 1 H NMR spectrum showed that H1 and H2 were adjacent to each other ( 3 J) and that the chlorination occurred at C-4 of carbazole ring A.
Putative biosynthesis of chlocarbazomycin A compound 1. Although compound 1 was first isolated just recently from S. diacarni strains (43,45), its biosynthetic pathway was unknown. To elucidate the biosynthetic pathway of 1, two separate libraries of compounds were created: (i) known halogenated microbial secondary metabolites from tryptophan halogenase precursors, including intermediates in the biosynthesis of rebeccamycin and pyrrolnitrin (Table S10), to deduce whether halogenation occurred before the closure of ring A to form the tricyclic carbazole alkaloid ring; and (ii) nonhalogenated carbazole alkaloids and indolocarbazoles (Table S11), to deduce whether halogenation occurred after tricyclic carbazole ring formation. Mining of the accurate mass of listed compounds and their adducts. 2 , in the mass spectrometry data showed no hits for compounds and intermediates for biosynthesis of rebeccamycin and pyrrolnitrin (Table S10) but showed 3 hits for nonhalogenated carbazole alkaloids and indolocarbazoles (Table S11) and S51), while carbazomycin H and carbazoquinocin F were identified based on the chemical formula, isotope model, and double bond equivalent ( Fig. S52 and S53). The absence of intermediates from the biosynthesis of rebeccamycin and pyrrolnitrin, including the chlorinated rebeccamycin analogs, and the detection of nonhalogenated carbazole alkaloids in S. tubbatahanensis DSD3025 T suggested that the chlorination most likely occurred after the tricyclic carbazole ring formation, i.e., after the closure of ring A, similar to the biosynthesis of tricyclic carbazoles such as neocarazostatins (44).
The carbazole alkaloids, such as neocarazostatin A and compound 8, produced by bacteria have a carbazole nucleus derived from tryptophan, pyruvate, and acetate (46)(47)(48). The biosynthetic pathway of 1 proposed in this study (Fig. 7) followed the initial formation of a carbazole nucleus, as observed in the biosynthesis of 8 (49). Comprehensive genome mining and bioinformatics allowed the retrieval of a candidate biosynthetic gene cluster (chlCz) in BGC region 16 spanning approximately 22.3 kB, with open reading frames (ORFs) involved in the formation of chlorinated carbazole alkaloids (Fig. 8, Table 5). The genome mining and gene cluster comparison with neocarazostatin A (44) and compound 8 (49) unveiled the key biosynthetic genes clustered in one BGC region of S. tubbatahanensis DSD3025 T (Table 5) for the formation of a tricyclic carbazole skeleton. The first step of carbazole assembly was the conversion of tryptophan into indole-3-pyruvate catalyzed by a putative aminotransferase chlCz1 (nzsD homolog in NZS biosynthesis) that shared 97.5% homology with histidinol-phosphate transaminase of S. diacarni LHW51701 T . A thiamine pyrophosphate-dependent enzyme, chlCz2 (nzsH homolog in NZS biosynthesis), with 97.3% gene similarity to a thiamine pyrophosphate-binding protein of S. diacarni LHW51701 T catalyzed the C-C bond between indole-3-pyruvate and pyruvate to form a-hydroxy-b-keto acid 9.
In contrast with the reported biosynthesis of neocarazostatin A that utilized the b-hydroxy group as its acyl substrate, the chlCz3 (nzsJ homolog in NZS biosynthesis) with 98.8% gene similarity to 3-oxoacyl-ACP synthase of S. diacarni LHW51701 T recognized only the acetyl-ACP as its non-b-hydroxyl acyl substrate (50) for carbazomycin biosynthesis. The unprecedented chlCz3 is a putative 3-oxoacyl-ACP synthase (KASIII) that possesses an acetyl-coenzyme A (CoA):ACP transacylase (ACAT) activity (51) for the catalytic formation of acetyl-ACP from acetyl-CoA and an acyl carrier protein, chlCz4 (nzsE homolog in NZS biosynthesis), with 98.8% gene similarity to an ACP in S. diacarni LHW51701 T . The acetyl-ACP may serve as a substrate for the side chain moiety of carbazomycins, since acetyl-ACP can be considered a starter unit for straight-chain fatty acid biosynthesis in Streptomyces (52). chlCz3 catalyzed the condensation of a-hydroxy-b-keto acid and acetyl-ACP. The unstable intermediate formed by the chlCz3-catalyzed reaction would undergo cyclization and ring A formation of the carbazole nucleus mediated by the putative carbazole synthase chlCz5 (nzsI homolog in NZS biosynthesis) that shared 100% gene homology to an SRPBCC family protein harbored by the S. diacarni LHW51701 T genome.
It was noted that 8 was a dihydroxyl type of carbazole metabolite, where its catechol intermediate 14 was initiated by dehydration to form 12 and nucleophilic aromatization to form 13 in a reaction catalyzed by chlCz5 (nzsI homolog in NZS biosynthesis) (53) (Fig. 7). Interestingly, 1 produced by S. tubbatahanensis DSD3025 T was plausibly catalyzed from a mono-hydroxyl type of carbazole intermediate 15, where the reduction and ring rearrangement steps were mediated by chlCz5 (  (Fig. 7). The two-step oxidation and decarboxylation of 15 catalyzed by the putative chlCz6a/chlCz6b FAD binding or FAD dependent and the chlCz7a/chlCz7b SDR family oxidoreductases were plausibly initiated to remove the methyl groups attached in C-1 and C-2 of the carbazole ring A, forming intermediates 17 and 19, respectively (Fig. 7). All the putative genes involved in the biosynthesis of 1 until intermediate 19 were found in BGC region 16 of the S. tubbatahanensis DSD3025 T genome (Fig. 8).
The hydroxyl group in C-3 of carbazole ring A in 1 was converted into a methoxy side chain catalyzed by a putative O-methyltransferase, chlCz8 (93.2% gene similarity with a methyltransferase of S. diacarni LHW51701 T ), which resulted in the formation of a tricyclic carbazole, 20, with no C-1 or C-2 methyl side chains (50). Interestingly, chlCz8 was found outside the chlCz gene cluster ( Table 5). The carbazole backbone of compounds 1 and 8 was mainly catalyzed by the putative KAS-III chlCz3 and the putative carbazole synthase chlCz5, as in the case of neocarazostatin A (50).
The putative halogenase (chlCz9) gene involved in the chlorination at C-4 of carbazole ring A was proposed to be located outside the chlCz gene cluster. The chlCz8 and chlCz9 adjacent in the genome were located approximately 2.1 Mbp away from the chlCz gene cluster of DSD3025 T and had opposite gene orientations with their corresponding gene homologs in the LHW51701 T genome (Fig. 8).
In the search for a halogenase in S. tubbatahanensis DSD3025 T genome, the tryptophan 6-halogenase was thought to be the gene responsible for the chlorination of C-4 in 1. However, the NMR analysis confirmed that the chlorine attachment in 1 was in C-4 of carbazole ring A and not in C-7 of the carbazole ring C, thus demonstrating that the tryptophan 6-halogenase was not responsible for this chlorination process. Notably, a putative single-component flavin-dependent halogenase (FDH), chlCz9, located outside the carbazomycin B gene cluster could be a bifunctional protein capable of chlorination and unusual flavin reduction. chlCz9 was annotated by antiSMASH as a NAD(P)-and/or FADdependent oxidoreductase that may act on aromatic substrates (54) with tryptophan as a precursor, but not directly with tryptophan. The gene comparison in databases revealed 95.1% similarity with a NAD(P)-and/or FAD-dependent oxidoreductase harbored in the  genome of a chlocarbazomycin producer, S. diacarni LHW51701 T , and ,50% gene similarity with halogenases. The NAD(P)-and FAD-dependent oxidoreductase as an FDH that could compensate for the chlorination of C-4 in carbazole ring A warrants further gene expression and in vitro experiments (54). More importantly, feeding 6-chlorotryptophan as a precursor for carbazomycin biosynthesis may unlock the possible role of tryptophan 6-halogenase found in the S. tubbatahanensis DSD3025 T genome in halogenating carbazole alkaloids. The putative biosynthetic gene cluster in S. tubbatahanensis DSD3025 T consisted of chlCz2/nzsH, chlCz3/nzsJ, and chlCz5/nzsI homologs for the biosynthesis of carbazole alkaloids, which were conserved in neocarazostatin producer Streptomyces sp. MA37, carquinostatin producer S. exfoliatus 2419-SVT2, carbazomycin producer S. luteoverticillatus CGMCC 15060, and chlocarbazomycin producer S. diacarni LHW51701 T (Fig. 8).
Additional biosynthetic genes were present in their respective BGCs for the modification of the carbazole nucleus and side chain moieties.
The bioinformatics analysis of the biosynthetic mechanism for the production of 1 highlighted the importance of exploring the different detection strictness and extra features predicted by antiSMASH. The careful analysis of proposed antiSMASH BGC boundaries could be a valuable aspect in mining biosynthetic gene clusters, rather than merely relying on the default settings. The putative gene cluster for the biosynthesis of 8 encoded in the S. tubbatahanensis DSD3025 T genome was identified through a detailed analysis of the BGC proposed by antiSMASH. The identified gene cluster was part of a larger BGC region 16 annotated as "T2PKS" due to its close proximity to signature genes used by antiSMASH as anchor points for the T2PKS gene cluster. The putative carbazomycin B gene cluster in the S. tubbatahanensis DSD3025 T genome was found using "relaxed" detection strictness, while the BGC for carbazomycin B in the S. luteoverticillatus CGMCC 15060 and S. diacarni LHW51701 T genomes was predicted only by antiSMASH in "loose" detection strictness.
S. tubbatahanensis DSD3025 T extract induced cell death via membrane damage against the multidrug-resistant S. aureus ATCC BAA-44. The crude extract of S. tubbatahanensis DSD3025 T had an oily consistency with dark brown coloration due to the high pigmented melanin content (55,56). Initial screening using an absorbancebased assay revealed that the nature of S. tubbatahanensis DSD3025 T extract caused false-positive results and low reproducibility. Thus, a fluorescence-based assay was utilized in this study to remove the bias due to the nature of the extract.
The live/dead cell differentiation in a cell population can be easily distinguished using the calcein-propidium iodide (PI) staining (57). Calcein acetoxymethyl ester (calcein-AM) is a permeable nonfluorescent molecule that is rapidly hydrolyzed by intracellular esterases in viable cells to green fluorescent calcein, which is subsequently retained intracellularly (58). Alternatively, the nuclei-staining dye PI is a nonpermeable molecule that can only enter a membrane-damaged cell and intercalates with the DNA (59), producing a red fluorescence. Therefore, the effect of the extract on the cell viability and membrane integrity of multidrug-resistant S. aureus ATCC BAA-44 was determined via calcein-PI dual staining.
After 6 h of treatment exposure, the live/dead cells were qualitatively identified via fluorescence microscopy. DMSO-treated cells displayed green fluorescence (Fig. 9A), indicating viable cells. On the other hand, cells treated with 70% ethanol and S. tubbatahanensis DSD3025 T extract (5 mg/mL) showed red fluorescence, indicating nonviable cells. The physiological status of the cell population of each treatment was then analyzed quantitatively by flow cytometric (FCM) analysis.
The cell population density in the flow cytometry profile in Fig. 9B was clustered into two regions based on their physiological status: R1, with PI-labeled cells (dead cells with damaged membranes), and R2, calcein-labeled cells (viable cells with an intact membrane) (23,60). The DMSO-treated cells showed a high cell population (94.6%) in R2, which was considered live, viable cells. Results showed that S. tubbatahanensis DSD3025 T extract-treated cells exhibited a high cell population (99.8%) in R1, indicating nonviable cells. The bactericidal activity of S. tubbatahanensis DSD3025 T extract was significantly comparable to that in 70% ethanol, which killed 98.5% of the cell population. These cells were considered dead with damaged membranes, as propidium iodide can only penetrate bacterial cells with compromised membrane permeability or integrity. In this study, 70% ethanol was used as a positive model for cell membrane damage. Ethyl alcohol, at 60 to 80% concentration, is a known disinfecting agent that denatures proteins and disrupts the cell membrane. The loss of cellular membrane integrity results in increased permeability and uncontrolled transport of solutes, decreasing the proton flux across the membrane and cytoplasmic leakage (61-63) leading to cell death. The bactericidal activity of S. tubbatahanensis DSD3025 T extract at 5 mg/mL was further validated by standard total plate counts to confirm that the membrane-damaged cell population observed in flow cytometry was dead cells. An antibiotic agent is considered "bactericidal" if it can completely prevent bacterial growth or results in a $99.9% decrease in the initial inoculum (64). Interestingly, both ethanol-treated and S. tubbatahanensis DSD3025 T extract-treated cells showed no visible growth after 24 h of incubation. Compared to DMSO-treated cells, which showed bacterial growth with 2.2 Â 10 8 CFU/mL (Fig. 9C). Collectively, these results confirmed that the S. tubbatahanensis DSD3025 T extract induced cell death of S. aureus ATCC BAA-44 via membrane damage.
Compound 1 produced by DSD3025 T was tested against ESKAPE pathogens. The results showed that 1 exhibited antibacterial activity against S. aureus ATCC BAA-44 (98% growth inhibition) and S. pyogenes (95% growth inhibition), with MIC 90 values of 128 mg/mL and 64 mg/mL, respectively (Table 6). Minimal growth inhibition was observed against E. faecium, with only 87% growth inhibition at 128 mg/mL, while none of the Gram-negative pathogens was inhibited. Carbazole derivatives have attracted attention due to their wide range of pharmacological applications (65)(66)(67)(68)(69), as the carbazole moiety is known to be useful in research and clinical studies (69)(70)(71)(72)(73). The compound 1 produced by S. diacarni LHW51701 T has a MIC of .128 mg/mL against Bacillus mycoides, methicillin-resistant S. aureus (MRSA), Mycobacterium smegmatis, and Candida albicans (45). In addition, compound 1 produced by coral reef sand-derived S. diacarni SCSIO 64983, when tested using the diskdiffusion method at 10 mg, showed no antibacterial activity against Gram-positive or Gramnegative pathogens. Halo-and thio-carbazomycin analogs produced by S. diacarni SCSIO 64983 against the mentioned test pathogens also showed no antibiotic properties (43). The observed antibacterial activities of 1 produced by Streptomyces tubbatahanensis DSD3025 T in this study against disease-causing pathogens S. aureus and S. pyogenes showed a propitious result that fills in the gap for antibiotic drug discovery.
S. tubbatahanensis DSD3025 T extract exhibited dose-dependent antiproliferative activities and toxicities against cancer and nontumor cells. The 3-(4,5-dimethyl-2-thiazolyl)-2,5-2H-tetrazolium bromide (MTT) assay is one of the most exploited in vitro cytotoxicity assays used in cancer research for measuring cell viability (74) and in the determination of the anticancer potential of a new compound (75,76). The antiproliferative activities of S. tubbatahanensis DSD3025 T extract as well as that of 1 were determined against MCF-7, HCT-116, and A2780 cancer cell lines in MTT assays. S. tubbatahanensis DSD3025 T extract demonstrated antiproliferative activity in a dose-dependent manner, exhibiting a complete  inhibitory activity against all cancer cells at 2 mg/mL (Fig. 10A). The S. tubbatahanensis DSD3025 T extract had 50% inhibitory concentration (IC 50 ) values of 457.4 mg/mL for MCF-7, 302.00 mg/mL for HCT-116, and 277.50 mg/mL for A2780 cells. Regardless of the IC 50 value of the extract, its inhibitory activity still showed an anticancer potential, suggesting that further purification of the extract may increase its anticancer activity. Compound 1 demonstrated anticancer activity above the 50% threshold against two cancer cell lines. At 125 mg/mL, 1 showed 55.4% and 50.2% growth inhibition against HCT-116 and A2780 cells, respectively (Fig. 10C). Carbazole derivatives have been gaining interest for their wide range of biological activities (64)(65)(66). However, 1 isolated from the sponge-associated bacterium Streptomyces diacarni LHW51701 T demonstrated no anticancer activity against the human lung adenocarcinoma cell line SPCA-1 (45). In this study, the antiproliferative activities of 1 produced by Streptomyces tubbatahanensis DSD3025 T against HCT-116 and A2780 indicated that the compound had a broader spectrum for targeted cancer cell lines.
The S. tubbatahanensis DSD3025 T extract showed decreasing toxicity in a dose-dependent manner against kidney, liver, and cardiac cell lines (Fig. 10B). Moreover, 1 showed moderate nephrotoxicity, high cardiotoxicity, and nonhepatotoxicity at 125 mg/mL (Fig. 10D). A low percent toxicity relates to safe levels of a compound suitable for administration toward the target organ. On the other hand, absence of compound toxicity signifies viability of the cells, since elevated LDH levels are associated with cell injury and death.
Conclusions. Advancements in bioinformatics and genomics have improved the ability to analyze key features and core biosynthetic gene clusters of novel genomes. These major developments in natural products research have facilitated the discovery of novel compounds through metabolomics. The discovery of the tryptophan halogenase gene encoded in the novel S. tubbatahanensis DSD3025 T genome enabled the identification of halogenated compounds by high-resolution mass spectrometry. The mass spectrometry metabolomics demonstrated that S. tubbatahanensis DSD3025 T produces halogenated metabolites, which could be investigated for their anticancer and antibacterial activity. The halogenated metabolites produced by S. tubbatahanensis DSD3025 T were less likely attributed to the original tryptophan halogenase harbored in the genome, and this finding led to the focused search for halogenated compounds using metabolomics. The two-component halogenase-reductase system was not involved in the halogenation of carbazole metabolites produced by S. tubbatahanensis DSD3025 T , but the chlorination step was rather compensated by a putative FDH gene that facilitated the C-4 chlorination of carbazole ring A. The tryptophan halogenase and its associated flavin reductase encoded in the S. tubbatahanensis DSD3025 T genome may influence the halogen moiety of other tryptophan substrates, and this needs further investigation for the identification of novel specialized metabolites. Future purification and NMR analysis of S. tubbatahanensis DSD3025 T extract warrant the profiling, structure elucidation, and identification of the halogenated specialized metabolites as antibacterial and anticancer candidates in the drug discovery pipeline.
The integration of bioinformatics-driven genomics and metabolomics unearths the hidden biosynthetic gene machineries that will further accelerate the natural drug discovery in novel Streptomyces. Overall, the bioprospecting of novel Streptomyces species from marine sediments of underexplored ecological niches serves as an important source of drug leads with hidden biosynthetic potential and unique chemical scaffolds.
It is a Gram stain-positive and aerobic actinomycete that forms well-developed, branched substrate mycelia with intertwining hyphal cells in MM1 agar and grows well in YGC, NA, TSA, ISP2, and ISP3 media. S. tubbatahanensis DSD3025 T has moderate growth in MM11 and ISP9 and poor growth in MM3 and ISP4. Optimum growth is at 28°C, pH 4.0 to 10.0, and 0 to 15% (wt/vol) NaCl. In the API ZYM, it is positive for alkaline phosphate, esterase, leucine arylamidase, acid phosphatase, and naphthol phosphohydrolase and weakly positive for esterase-lipase and a-glucosidase enzymes.
The type strain, DSD3025 T (=DSM 33792 T ), was isolated from the marine sediment samples collected in Tubbataha Reefs Natural Park, Cagayancillo, Palawan, in the middle of Sulu Sea, Philippines. The complete genome of S. tubbatahanensis DSD3025 T is composed of one linear chromosome 7,760,770 bp long, with a G1C content of 72.3%, a total of 6,579 predicted genes, and 29 biosynthetic gene clusters. The whole-genome sequence of S. tubbatahanensis DSD3025 T has been deposited in the GenBank database under the accession number CP093846.

MATERIALS AND METHODS
Strain isolation and maintenance. S. tubbatahanensis DSD3025 T was isolated from a marine sediment sample at a collection site (latitude 8.74000000, longitude 119.81916667) in Tubbataha Reefs Natural Park in the middle of Sulu Sea, Philippines, in April 2018. S. tubbatahanensis DSD3025 T was recovered from marine sediment in the subsurface layer (26 to 50 cm depth below the seafloor) sampled using a 110-cm core sampler (22). Marine sediment samples were air-dried completely and inoculated in a trehalose-containing marine agar using the dry stamp method. Pure culture of S. tubbatahanensis DSD3025 T was obtained and maintained on MM1 (77) agar at 28°C and stocked in glycerol stock suspensions (20% [vol/vol]) at 280°C.
Genomic and phylogenetic analyses. The genomic DNA extraction of S. tubbatahanensis DSD3025 T was carried out by using DNeasy blood and tissue kits (Qiagen) (21). Whole-genome sequencing of S. tubbatahanensis DSD3025 T was performed by applying a long-read PacBio single-molecule real-time (SMRT) sequencing system (78). The SMRTbell template library was based on the procedure for 10-kb library preparation according to the instructions from PacBio. The size selection for BluePippin was performed according to the manufacturer's recommendations using the 0.75% DF Marker S1 High-Pass 6-kb to 10-kb v3 run protocol and S1 marker (79). The SMRTbell library was annealed based on the SMRT link setup and then sequenced using Sequel II. For whole-genome analyses, S. tubbatahanensis DSD3025 T was sequenced with its genome assembled and deposited at NCBI GenBank under the accession number CP093846.
The complete 16S rRNA gene sequence of S. tubbatahanensis DSD3025 T (1,531 bp) was retrieved from the draft genome sequence data and initially analyzed using the EzBioCloud server (80). For phylogenetic analysis, the 16S rRNA gene sequences of the type strains of closely related Streptomyces species were obtained from the NCBI GenBank database. The 16S rRNA gene sequences were aligned using the Clustal W algorithm in MEGA 11.0 (81). The phylogenetic trees were constructed using neighbor-joining (82), maximum parsimony (83), and maximum likelihood (84) methods with bootstrap analysis based on 1,000 replications. The five housekeeping genes, atpD (ATP synthase F1 beta subunit), gyrB (DNA gyrase B subunit), rpoB (RNA polymerase beta subunit), recA (recombinase A), and trpB (tryptophan synthetase beta subunit) were retrieved from the complete genome sequence of S. tubbatahanensis DSD3025 T , and the related gene sequences of the type strains were obtained from the GenBank database and concatenated head-to-tail, inframe. The phylogenetic tree of the concatenated protein-coding sequence (atpD-gyrB-rpoB-recA-trpB) of S. tubbatahanensis DSD3025 T and its closely related strains in the GenBank database was reconstructed using neighbor-joining (82), maximum parsimony (83), and maximum-likelihood algorithms (84) in MEGA 11.0 (81). The multilocus sequence analysis (MLSA) evolutionary distances were calculated using Kimura's two-parameter model (85). The in silico digital DNA-DNA hybridization (dDDH) values, based on formula 2, were calculated using the Genome-To-Genome Distance Calculator (GGDC v3.0) (86) at https://ggdc.dsmz.de/ggdc .php#. Calculation of orthoANI values and generation of an ANI heat map were performed by using OAT software v0.93.1 available at https://www.ezbiocloud.net/tools/orthoaniu (87).
Cultural and phenotypic properties. The cultural characteristics of S. tubbatahanensis DSD3025 T were determined following the growth in MM1 agar, mannitol-containing agar (MM3), glucose-containing (MM11) agar, YGC (88) agar, TSA, NA, and ISP2, ISP3, ISP4, and ISP9 at 28°C for 7 days. The morphological characteristics of S. tubbatahanensis DSD3025 T were observed in terms of aerial spore color, vegetative mycelium, and diffusible pigmentation produced in MM1 agar after 7 days of incubation at 28°C and with scanning electron microscopy (JEOL JSM 5510LV). Growth at different pH, temperature, and salinity was observed in MM1 agar. S. tubbatahanensis DSD3025 T was grown in MM1 agar with different pH levels ranging from 4.0 and 10.0 at an interval of 1 pH unit. The temperatures used to determine the growth of S. tubbatahanensis DSD3025 T were 4, 28, and 37°C. Tolerance to NaCl was carried out by growing S. tubbatahanensis DSD3025 T in MM1 agar with different NaCl concentrations (0, 2.6, 3, 5, 7, 10, 12, 15% [wt/vol]). Enzymatic and biochemical characteristics of S. tubbatahanensis DSD3025 T were analyzed using a commercially available API ZYM kit (bioMérieux, Marcy-l' Etoile, France) (89)(90)(91). The API strips were inoculated according to the manufacturer's instructions. The reference strain, Pseudomonas aeruginosa ATCC 27853, was used as a standard control for the assay. Briefly, S. tubbatahanensis DSD3025 T and P. aeruginosa bacterial suspensions were prepared in an API suspension medium with a turbidity of McFarland standard 5. A volume of 65 mL of bacterial suspension was inoculated into each cupule of the API strips. The strips were then incubated for 4 h at 37°C. One drop of ZYM A and ZYM B reagents was added into each cupule for the enzyme and biochemical analyses.
Chemotaxonomic properties. The biomass and lyophilized cells of S. tubbatahanensis DSD3025 T were obtained by growing in MM1 broth at 28°C for 7 days in a shaking incubator (15, 21-23, 92, 93). The analyses of fatty acids (92,93), polar lipids (94), whole sugar (95), and respiratory quinones were carried out by the Identification Service, Leibniz-Institut DSMZ-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany. The cellular fatty acids were analyzed according to the instructions of the Microbial Identification System (MIDI; microbial ID). Polar lipids were extracted from lyophilized cells using chloroform-methanol-0.3% aqueous NaCl mixture (94) and identified by twodimensional silica gel thin-layer chromatography. Diagnostic sugars in whole-cell hydrolysates (96) were analyzed by thin-layer chromatography on cellulose plates for identification. Respiratory quinones were extracted from freeze-dried cells using hexane, purified further by a silica-based solid-phase extraction, and identified by HPLC-DAD based on confirmed spectrum and retention time in MS. S. tubbatahanensis DSD3025 T was deposited in DSMZ under the accession number DSM 33792 T .
Extract preparation. A 7-day-old broth culture of S. tubbatahanensis DSD3025 T was inoculated onto MM1 agar and grown for 14 days at 28°C. Harvested biomass was extracted with ethyl acetate and concentrated in vacuo until dried extract was obtained (15,(21)(22)(23). Solid-phase extraction (SPE) was performed to remove methanol (MeOH)-insoluble melanin and impurities in the extract. The dried extract was reconstituted in 100% MeOH (HPLC grade) to create a to 40-mg/mL suspension and then semipurified using Sep-Pak Plus Short C 18 cartridges (Waters, Ireland) eluted with 100% MeOH. The S. tubbatahanensis DSD3025 T SPE extract was concentrated in vacuo for metabolite profiling, antibiotic testing, and anticancer screening.
UPLC-QTOF of S. tubbatahanensis DSD3025 T extract. The chemical profile of S. tubbatahanensis DSD3025 T extract was analyzed using MS in UPLC-QTOF-MS. A 0.2-mg/mL solution of S. tubbatahanensis DSD3025 T extract in MeOH (MS grade) was prepared and placed in an autosampler at 4°C. Five microliters of the solution was injected into the UPLC system. Separation of compounds was performed using an Acquity UPLC BEH C 18 column (130 Å, 1.7 mm, 2.1 mm by 50 mm) held at 40°C. Water (H 2 O; solvent A) and acetonitrile (MeCN; solvent B) containing 0.1% (vol/vol) formic acid (HCOOH) were used as mobile phase. The mobile phase was pumped at 0.3 mL/min as follows: 20% B (0 to 0.55 min), 20% to 100% B (0.55 to 9.92 min), 100% B (9.92 to 11.58 min), 100% to 20% B (11.58 to 12.68 min), and 20% B (12.68 to 13.78 min). Compounds eluted at different retention times were analyzed using a Waters Synapt XS Q-ToF mass spectrometer equipped with an ESI source. The mass spectrometer was calibrated in both positive and negative ions in resolution mode using sodium iodide (NaI). A lockspray mass correction was performed using leucine-enkephalin (m/z 556.2771 [M 1 H] 1 , m/z 554.2615 [M 2 H] 2 ). The MS E data (low energy, 0 eV; high energy, ramp 25 to 75 eV) was acquired in the centroid mode over a mass ion range of 100 to 2,100 Da and a scan time set at 0.15 s. The analyte was subjected to a capillary voltage of 3 kV (positive) or 1.0 kV (negative), with a 100°C source temperature, 500 liters/h desolvation gas (N 2 ) flow, and 300°C desolvation temperature. Acquired MS E data were processed using MassLynx software version 4.2 (Waters Corporation, Milford, MA, USA). Accurate mass, isotope pattern, chemical formula, and double bond equivalent were used for extensive library search in AntiBase v2017, ChemSpider, and MarinLit.
Molecular networking analysis. The acquired Waters MS E data (.raw) was converted to Analysis Base File (.abf) format using the Reifycs Analysis base file converter (Tokyo, Japan). The converted files were then processed using MS-DIAL with optimized parameters (Table S12).
A library of 12,879 unique compounds in the negative mode was downloaded from http://prime.psc .riken.jp/compms/msdial/main.html for the identification of compounds. Molecular networking analysis was performed using the Data Visualization Navigator on a target m/z value, and the MS/MS tolerance was set at 0.05 Da, 40% cutoff similarity, and retention time tolerance to 1.5 min. The molecular network was then visualized using the Mozilla Firefox HTML viewer.
Isolation and structural elucidation of chlocarbazomycin A compound 1 produced by S. tubbatahanensis DSD3025 T . A solid-phase extraction was performed with 1 g dried S. tubbatahanensis DSD3025 T crude extract via reversed-phase flash column chromatography (Biotage Isolera, Biotage Sfar, C 18  equipped with a make-up pump (Waters 515) and a splitter that split the column outlet flow to the single-quadrupole mass detector (Waters Acquity QDa detector) and fraction collector (WFC III) at 1:1,000 (vol/vol). Mass ion scans were set at a range of 150 to 700 m/z in the positive polarity. The fraction collection was triggered at m/z 232 [M 1 H] 1 , .1 Â 10 3 intensity. The collected fraction was dried in vacuo to afford 4.73 mg of 1 at ;80% purity as estimated via 1 H NMR.
The structure of 1 was described and verified using NMR data. Briefly, the one-dimensional (1D) and 2D NMR spectra of 1 were recorded in CDCl 3 (99.8% D) at 298 K with CDCl 3 peaks as a reference on a Bruker Avance cryoprobe (5 mm TCI 1 H& 19 F/ 13  Antibacterial activity. (i) Flow cytometry assay by calcein-PI dual staining. The effect of S. tubbatahanensis DSD3025 T on bacterial cell viability and membrane integrity was assessed using a calcein-PI dual staining assay, followed by fluorescence microscopy and flow cytometry. The antibiotic activity of S. tubbatahanensis DSD3025 T extract was evaluated against a multidrug-resistant pathogen, Staphylococcus aureus ATCC BAA-44, an Iberian MRSA clone with multidrug resistance against 18 antibiotics (52). Briefly, S. aureus ATCC BAA-44 cells were grown for 3 h at 37°C (200 rpm) to obtain mid-logarithmic-phase cells. The cells were harvested by centrifugation (4,000 rpm for 5 min) and resuspended in phosphate-buffered saline (PBS) to a final optical density at 600 nm (OD 600 ) of 0.5. In a 96-well plate, the bacterial suspension in PBS was treated with S. tubbatahanensis DSD3025 T extract at 5 mg/mL. DMSO-and 70% ethanol-treated cells were used as the negative and positive controls, respectively. Treatment plates were incubated for 6 h at 37°C (80 rpm). A premixed dye solution was prepared using a Live/Dead cell double staining kit (Sigma-Aldrich, MO, USA) according to the manufacturer's protocol with modifications. The dye solution was prepared by adding 5 mL of calcein-AM (CA) and 20 mL of PI to 24 mL PBS. After treatment exposure, cells were dualstained by adding 10 mL CA-PI dye solution. Single-stained cells were also prepared for fluorescence compensation in the flow cytometer. The plate was incubated for 30 min in a dark room at 37°C.
The flow cytometry analysis was performed to quantify the live or dead cells in the cell population. Data were acquired using Amnis FlowSight imaging flow cytometer (Luminex, Austin, TX, USA) with 505 to 560 nm (channel 2) and 642 to 745 nm (channel 5) for CA and PI fluorescence detection, respectively. Data acquisition was set to 10,000 events for each sample. The experiment was performed in three trials with triplicates. Data were analyzed using AMNIS IDEAS software v6.3.23. The PI fluorescing cells (R1) were gated as dead cells with damaged cell membranes, and CA fluorescing cells (R2) were gated as live cells with intact membranes. Treated cells were also observed in an IX83 inverted fluorescence microscope (Olympus, USA) to morphologically identify the physiological status of the cells through fluorescence emission.
(ii) Total plate count. A total plate count was performed using the spread plate method to confirm the bactericidal activity. Cell treatment was performed using the same protocol as described above. The treated cells were serially diluted up to 1 Â 10 6 dilution in PBS. The cell suspension (50 mL) was then inoculated onto TSA plates and incubated for 24 h at 37°C. Three trials with triplicates were done. CFU were calculated using the following equation: CFU per milliliter = [(number of colonies)(dilution factor)]/(volume of culture plate).
(iii) MIC of chlocarbazomycin A compound 1. The MIC 90 of 1 was determined using a microbroth susceptibility assay with 2-fold serial dilution. The compound was tested starting from an initial test concentration of 128 to 0.25 mg/mL against multidrug-resistant pathogens Enterococcus faecium ATCC 70021, Staphylococcus aureus ATCC BAA-44, Klebsiella pneumoniae ATCC BAA-1705, Acinetobacter baumannii ATCC BAA-1605, and Streptococcus pyogenes ATCC 12384. The positive controls were vancomycin for E. faecium, S. aureus, and S. pyogenes, meropenem for A. baumannii, imipenem for K. pneumoniae, and tetracycline for all test pathogens. DMSO was used as the negative control. The bacterial suspension (195 mL) with an OD 600 of 1 Â 10 6 CFU/mL was dispensed into the wells of a 96-well plate followed by the addition of the treatment (5 mL), and then incubated for 18 to 24 h at 37°C. The OD 600 was determined using an absorbance microplate reader (BioTekTM ELx808, Biotek, Winooski, VT, USA). The last concentration showing .90% growth inhibition was considered the MIC 90 of the extract (53). The assay was performed in triplicates and three trials except for S. pyogenes, which was tested in two trials only, due to sample limitation. The percent growth inhibition was calculated using the following equation: percent growth inhibition = [(OD negative control 2 OD treatment )/(OD negative control )] Â 100. Anticancer activity. (i) Cell line maintenance. Human ovarian carcinoma (A2780; ECACC 93112519), human colon carcinoma (HCT-116; ATCC CCL-247), and human breast Caucasian adenocarcinoma (MCF-7; ECACC 86012803) were used for MTT testing. The human renal proximal tubule epithelial cell line (HK-2; ATCC CRL-2190), human hepatocellular carcinoma (Hep G2; ECACC 85011430), and rat cardiomyocytes (H9c2 ; ECACC 88092904) were utilized for toxicity testing. The A2780 cells and HCT-116 cells were grown in RPMI 1640 (D6456). The MCF-7 and Hep G2 cells were grown in minimum essential medium (MEM; 51411C) with 1% nonessential amino acid solution (NEAA; M7145). The MCF-7 growth medium had an added supplement of 1 mM sodium pyruvate solution (S8636). The HK-2 and H9c2 [2-1] cell lines were grown in Dulbecco's minimum essential medium (DMEM; D6546). All maintenance media contained 10% fetal bovine serum (F2442), 2 mM alanyl-glutamine (A8185) supplement, and 1% penicillin-streptomycin (P4333), with the exception of MCF-7 and HCT-116 growth media, which did not contain 2 mM alanyl-glutamine supplement.
The cell lines used in this study were cultured in accordance with the medium components and concentrations publicly listed in the European Collection of Authenticated Cell Cultures (ECACC) and American Type Culture Collection (ATCC). The growth media and supplements used for cell line maintenance were purchased from Merck (Sigma-Aldrich), Philippines.
(ii) MTT assay. Human cancer cells for the antiproliferation assay of S. tubbatahanensis DSD3025 T extracts and 1 were seeded in 96-well plates using an automated liquid workstation Biomek i5 (Beckman Coulter, Germany) at 25,000, 6,000, and 8,000 cells/well for A2780, HCT-116, and MCF-7, respectively. Cells were incubated for 24 h at 37°C with 5% CO 2 and 80 to 90% relative humidity. The first batch of cells was then treated with 2-fold dilutions of S. tubbatahanensis DSD3025 T extract from 2 to 0.007 mg/mL, while another batch was treated with 2-fold dilutions of 1 at 125 to 0.49 mg/mL for 24 h. The negative control was DMSO, while cisplatin (PHR1624), 5-fluorouracil (F6627), and tamoxifen (85256) were used as positive controls for A2780, HCT-116, and MCF-7 cancer cells, respectively. Ten microliters of a 5-mg/mL MTT solution was then added to each well and incubated for 4 h. The medium was then gently aspirated and replaced with 100 mL DMSO to dissolve the formazan crystals. Absorbance was measured at 570 nm using a CLARIOstar multimode microplate reader (BMG Labtech GmbH, Germany). The percent growth inhibition was calculated using the following equation: percent inhibition = {[(negative control absorbance) 2 (experimental absorbance)]/(negative control absorbance)} Â 100.
The IC 50 values of S. tubbatahanensis DSD3025 T extract were analyzed using GraphPad Prism v9.5.0 for Windows (GraphPad Software, San Diego, CA USA). The inhibitory activity of 1 tested at a final concentration of 125 mg/mL was compared with that of the positive controls using the same software.
(iii) Lactate dehydrogenase (LDH) assay for toxicity profiling with nontumor cells. Cell lines for the LDH assay were seeded at 5,000 cells/well in a 96-well plate using an automated liquid handling workstation (Biomek i5; Beckman-Coulter, Germany). The experimental plates were incubated for 2 to 4 h before the addition of treatment for the cells to attach at 37°C, 5% CO 2 with a relative humidity of 80 to 90%. After incubation, the 2-fold dilutions of S. tubbatahanensis DSD3025 T SPE extract and 1 with decreasing concentrations starting from 1 mg/mL to 0.063 mg/mL, and 125 mg/mL were added, respectively.
The negative control used was 0.1% DMSO. Doxorubicin hydrochloride (PHR1789) was used as the positive control for HK-2 and H9c2 [2-1] cell lines, while tamoxifen (85256) was used as the positive control for the Hep G2 cell line. The experimental plates were incubated for 18 h followed by LDH toxicity testing using a cytotoxicity detection kit (4744934001; Roche Diagnostics). The stop solution used was 1 N HCl, and the absorbance was measured at 490 nm using a CLARIOstar multimode microplate reader (BMG Labtech GmbH, Germany). The percent toxicity was calculated using the following formula which was based on the assay kit protocol: percent toxicity = {(experimental value 2 experimental blank) 2 (DMSO control 2 DMSO control blank)]/[(high control value 2 high control blank) 2 (low control value 2 low control blank)]} Â 100. The results were illustrated using GraphPad Prism v9.5.0 for Windows (GraphPad Software, San Diego, CA, USA).
The threshold value for analyzing the toxicity in this study was set to 10%, to obtain the lowest minimum compound efficacy against the tested cell line. Any value exceeding the baseline of 10% was considered toxic.
Data availability. The whole genome of S. tubbatahanensis DSD3025 T , consisting of 7,760,770 bp and with 72.3% G1C content (Table 1), was deposited in the GenBank database under accession number CP093846.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 6.5 MB.