Nosiheptide Biosynthesis Featuring a Unique Indole Side Ring Formation on the Characteristic Thiopeptide Framework

Abstract

Nosiheptide (NOS), belonging to the e series of thiopeptide antibiotics that exhibit potent activity against various bacterial pathogens, bears a unique indole side ring system and regiospecific hydroxyl groups on the characteristic macrocyclic core. Here, cloning, sequencing and characterization of the nos gene cluster from Streptomyces actuosus ATCC 25421 as a model for this series of thiopeptides has unveiled new insights into their biosynthesis. Bioinformatics-based sequence analysis and in vivo investigation into the gene functions show that NOS biosynthesis shares a common strategy with recently characterized b or c series thiopeptides for forming the characteristic macrocyclic core, which features a ribosomally synthesized precursor peptide with conserved posttranslational modifications. However, it apparently proceeds via a different route for tailoring the thiopeptide framework, allowing the final product to exhibit the distinct structural characteristics of e series thiopeptides, such as the indole side ring system. Chemical complementation supports the notion that the S-adenosylmethionine (AdoMet)-dependent protein NosL may play a central role in converting Trp to the key 3-methylindole moiety by an unusual carbon side chain rearrangement, most likely via a radical-initiated mechanism. Characterization of the indole side ring-opened analog of NOS from the nosN mutant strain is consistent with the proposed methyltransferase activity of its encoded protein, shedding light into the timing of the individual steps for indole side ring biosynthesis. These results also suggest the feasibility of engineering novel thiopeptides for drug discovery by manipulating the NOS biosynthetic machinery.

Thiopeptides are a growing class of sulfur-rich, highly modified heterocyclic peptides (1). Despite overall structural diversity, they share a characteristic macrocyclic core, consisting of a nitrogen-containing, 6-membered ring central to multiple thiazoles and dehydroamino acids (Figure 1). Nosiheptide-like members, classified as e series thiopeptides according to a central 2,3,5,6-tetrasubstituted pyridine domain, possess an indolic acid ring system that is appended to the side chains of the Ser/Cys and hydroxylated Glu residues of the macrocyclic core via at least two carboxylic ester linkages (e.g. O- and S- linkages for nosiheptide, and two O- linkages for nocathiacins), and in some cases, (an) attached sugar unit(s) (2, 3). There is significant clinical interest in thiopeptides due to their potent activity against various bacterial pathogens, including methicillin-resistant Staphylococcus aureus (MRSA), penicillin-resistant Streptococcus pneumoniae (PRSP), and vancomycin-resistant enterococci (VRE). For example, the nocathiacins have been investigated as drug leads for developing broad-spectrum antibiotics by chemical modifications, aiming at the generation of biologically comparable and water-soluble compounds to fight against progressively emergent bacterial resistance to traditional chemotherapies (4–8).

Figure 1.

Structures of the e series thiopeptides nosiheptide (NOS) and nocathiacin I, b series thiostrepton (TSR)/siomycin A (SIO-A), and d series thiomuracin A (TMR-A), GE2270A and thiocillin I (TCL-I).

Nosiheptide (NOS), produced by Streptomyces actuosus ATCC 25421, is one of the oldest known thiopeptides and has been widely used as a feed additive for animal growth (9, 10). The structure and stereochemistry of NOS was ultimately confirmed by X-ray crystallography (11), following extensive analysis of chemically hydrolyzed fragments by NMR spectroscopic methods (12, 13). While the total synthesis of NOS has not been achieved, NOS, as the model molecule in the e series, was one of the first thiopeptides to be investigated biosynthetically by incorporation of isotope-labeled precursors, along with thiostrepton (TSR), a representative of b series that contains a distinct quinaldic acid side ring system appended to the characteristic thiopeptide macrocyclic core (14–17). All moieties of the peptidyl backbones of NOS and TSR were shown to originate exclusively from proteinogenic amino acids, including dehydroamino acids (from the Ser or Thr residues undergoing the anti elimination of water), thiazoles (from the Cys residues with cyclodehydration followed by deoxygenation), and the central 6-membered nitrogen heterocycle (produced by cyclization between two corresponding dehydroalanine acids with incorporation of an adjacent carbonyl group). Interestingly, the Trp residue was confirmed as a common precursor for the side ring systems in both NOS and TSR biosynthesis (14, 16), in spite of the difference in structures of the resulting motifs (i.e. the indolic acid moiety for NOS and the quinalidic acid moiety for TSR) and their linkages to the thiopeptide macrocyclic core (Figure 1).

It had long been controversial whether the thiopeptides are biosynthesized via a ribosome-dependent route of maturation of short peptides to complex, highly functionalized molecules, such as lantibiotics (18), bacteriocins (19) and cyanobactins (20), or in a manner similar to peptide antibiotics vancomycin and cyclosporin, whose peptidyl backbones are assembled by non-ribosomal peptide synthetases (NRPSs) (21, 22). Very recently, we and other research groups cloned, sequenced and characterized the biosynthetic gene clusters of the bicyclic b series thiopeptides TSR and siomycin A (SIO-A), and monocyclic d series thiocillins (TCLs), GE2270A and thiomuracins (TMRs), uncovering a common paradigm for the characteristic macrocyclic core biosynthesis that features conserved posttranslational modifications on a ribosomally synthesized precursor peptide (23–26). Given the similarities in structures and precursor-labeling patterns, the biosynthesis of NOS likely shares a conserved strategy with those of above thiopeptides to form the thiopeptide macrocyclic core. However, the tailoring of the elementary framework into the e series-specific member, particularly for the indolic acid moiety formation and attachment via a route distinct from the quinaldic acid pathway in TSR biosynthesis, was unclear. To exploit the genetic basis for fulfilling the knowledge gap, we now report the localization of the nos biosynthetic gene cluster from S. actuosus ATCC 25421 by cloning the thiopeptide-specific cyclodehydratase gene nosG using our recently developed PCR approach. The sequence analysis of the entire nos gene cluster allows for assignment of functions to the deduced gene products, setting the stage to propose the NOS biosynthetic pathway. While the finding of the ribosomal origin of NOS along with conserved posttranslational modifications again validates the generality of thiopeptide biosynthesis, in vivo functional investigations of genes involved in the indole side ring formation have revealed new insights into the biosynthesis of the e series-specific thiopeptides, including a novel strategy for the carbon side chain rearrangement to convert the Trp residue into the key 3-methylindole moiety.

Results and Discussion

NOS biosynthetic gene cluster

The enzymes for the formation of the thiopeptide-characteristic framework are highly conserved in and unique to thiopeptide biosynthesis (24), as exemplified by the thiopeptide cyclodehydratases. These enzymes exhibit high sequence homology to each other, but are distinct from other proteins with similar functions, including the polythiazole synthetases in both bacteriocin and cyanobactin biosynthesis (27). We decided to localize the NOS biosynthetic gene cluster by cloning the NOS cyclodehydratase gene using a specific PCR approach developed recently (24). With the genomic DNA of S. actuosus as the template, a distinct product with the expected size of 0.7 kb was readily amplified and subsequently cloned into pMD19-T vector for sequencing. Analysis of randomly selected clones confirmed the identity of the PCR product (P1), the deduced amino acid sequence of which is highly similar to known thiopeptide cyclodehydratases. To determine the relevance of P1 to NOS production, we set out to inactivate the target allele by gene disruption in S. actuosus. As anticipated, the P1 allele (namely nosG) mutant strain SL4001 completely lost the ability to produce NOS, confirming that the cloned locus encodes NOS biosynthesis (Figure S1). Consequently, using P1 as a probe, approximately 2,000 clones from the S. actuosus genomic library were screened, resulting in the identification of 3 overlapping cosmids that span a 70 kb region on the chromosome.

The DNA region represented by the cosmid pSL4001 was selected for sequencing, yielding a 34,713 bp contiguous sequence with the overall GC content of 71.6%, characteristic of Streptomyces DNA. Bioinformatics-based analysis of the sequenced region revealed 26 open reading frames (orfs), 15 of which, including 14 structural genes and 1 regulatory gene, were proposed to constitute the nos gene cluster according to functional assignments to the deduced products (Figure 2 and Table 1). The boundaries of the nos gene cluster were judged by identification of the immediately flanking orfs (at the both ends) that encode DNA-directed RNA polymerase subunits and various ribosomal proteins for the construction of the bacterial ribosome. Attempts to inactivate these genes were not successful, consistent with their essential nature to the survival of S. actuosus (data not shown). Considering the antibacterial mechanism of NOS, with the mode of action on the ribosome to inhibit protein synthesis by occupying a 23S ribosomal RNA domain that was known for the binding of the ribosomal protein L11 (28), the direct insertion of the nos gene cluster into the ribosomal protein encoding locus is interesting. Whether the resulting ribosome in S. actuosus is resistant to the action of NOS remains to be determined.

Figure 2.

Organization of the NOS biosynthetic genes. The deduced functions of which are labeled in color and summarized in Table 1.

Table 1.

Deduced functions of orfs in the nosiheptide biosynthetic gene cluster.

Gene	Sizea	Protein Homologb and Origin	Identity / Similarity, %	Proposed Function
orf1	324	SCO4661 (NP_628821), from S. coelicolor A3(2)	91/96	elongation factor G
orf2	156	SCO4660 (NP_628820), from S coelicolor A3(2)	99/100	30S ribosomal protein S7
orf3	123	SCO4659 (NP_628819), from S coelicolor A3(2)	99/100	30S ribosomal protein S12
nosA	151	Francci3_4114 (YP_483191), from Frankia sp. CcI3	55/67	hypothetical protein
nosB	455	Sare_3149 (YP_001537948), from Salinispora arenicola CNS-205	32/47	cytochrome P450
nosC	408	SACE_1426 (YP_001103673), From Saccharopolyspora erythraea NRRL 2338	39/55	cytochrome P450-like enzyme
nosD	345	TsrK (ACN80673), from S. laurentii ATCC 31255	32/54	dehydratase
nosE	920	TsrJ (ACN80672), from S. laurentii ATCC 31255	29/50	dehydratase
nosF	549	TsrM (ACN80675), from S. laurentii ATCC 31255	29/51	dehydrogenase
nosG	656	TsrO (ACN80677), from S. laurentii ATCC 31255	36/58	cyclodehydratase
nosH	654	TsrN (ACN80676), from S. laurentii ATCC 31255	22/37	hypothetical protein
nosI	419	SAMR0921 (CAJ88630), from S. ambofaciens ATCC 23877	29/39	AMP-dependent acyl-CoA synthetase/ligase
nosJ	79	--	--	hypothetical protein
nosK	264	BCE_2449 (NP_978758), from Bacillus cereus ATCC 10987	29/43	α/β fold family hydrolase/acyltransfer ase
nosL	400	ThiH (AAD48429), from Salmonella typhimurium	26/41	Radical S-AdoMet: Biotin and thiamine synthesis associated
nosM	50	--	--	NOS precursor peptide
nosN	395	Tlm Orf11 (ABL74954), from Streptoalloteichus hindustanus	32/48	S-AdoMet-dependent oxidase or methyl transferase
nosO	370	TsrL (ACN80674), from S. laurentii ATCC 31255	28/49	hypothetical protein
nosP	323	pSLA2-L_p071(NP_851493), from S. rochei	51/66	pathway specific regulatory protein
orf4	1299	SAV_4915 (NP_826092), from S. avermitilis MA-4680	96/98	DNA-directed RNA polymerase subunit β′
orf5	1181	SAV_4914 (NP_826091), from S. avermitilis MA-4680	95/98	DNA-directed RNA polymerase subunit β
orf6	129	SCO4653 (NP_628814), from S. coelicolor A3(2)	81/86	50S ribosomal protein L7/L12
orf7	176	SSEG_01531 (YP_002208236), from S. sviceus ATCC 29083	89/96	50S ribosomal protein L10
orf8	272	SCO4651 (NP_628812), from S. coelicolor A3(2)	65/78	lipoprotein
orf9	297	SSEG_01533 (YP_002208238), from S. sviceus ATCC 29083	55/70	lipoprotein
orf10	241	SCO4649 (NP_628810), from S. coelicolor A3(2)	92/97	50S ribosomal protein L1

^aNumbers are in amino acids.

^bNCBI accession numbers are given in parentheses.

Ribosomal origin of NOS

Within the predicted nos gene cluster, nosM encodes a 50-aa peptide that contains a 37-aa leader peptide (LP) and a 13-aa, Cys and Ser/Thr-rich structural peptide (SP) (Figure 3a). The SP (SCTTCECCCSCSS) is in complete agreement with the peptide sequence of the NOS backbone, with the exception of the C-terminal Ser residue that can be removed (except for the amino group) in the tailoring process. The LP sequence of NosM is rich in acidic amino acids, similar to those of thiopeptide precursor peptides such as TsrH for TSR and TclB for TCLs, but no overall significant homology in sequence was identified among them. This indicates that the Glu and Asp residues of LPs might be commonly important to substrate recognitions for posttranslational modifications that, however, may take place specifically on the cognate precursor peptide in each thiopeptide biosynthesis. To confirm that nosM is indispensable in NOS biosynthesis, we inactivated and complemented it in vivo. As shown in Figure 6 (II), the resulting in-frame deletion mutant strain SL4002 completely lost the ability to produce NOS, providing experimental evidence to support the origin of NOS from a ribosomally synthesized precursor peptide. Furthermore, the NOS production can be partially restored in the recombinant strain SL4003, where nosM was expressed in trans in SL4002 (Figure 6, III), implying a promising way to access the structural diversity of NOS by sequence permutations to the precursor peptide.

Figure 3.

Genetic features for the thiopeptide framework formation. (a) Peptide precursors for NOS (NosM), TSR (TsrH), SIO-A (SioH), TCLs (TclB), and GE2270A (TpdA). The SP sequences are labeled in bold. (b) Organization of the thiopeptide framework-forming genes identified from the producers of NOS (nos), TSR (tsr), SIO-A (sio), and TCLs (tcl). Their deduced functions are labeled in pattern, and homologies in sequence are indicated by dashed lines.

Figure 6.

HPLC analysis of the production of NOS or its analog, in the wild type strain S. actuosus ATCC 25421 (I), nosM mutant SL4002 (II), recombinant strain SL4003 that derives from SL4002 by expressing nosM in trans (III), nosH mutant SL4004 (IV), nosL mutant SL4005 (V), nosL mutant SL4005 complemented by chemically feeding 3-methylindole-2-carboxylic acid (8) (VI), nosN mutant SL4006 (VII), and nosP mutant SL4007 (VIII). Solid dot indicates NOS. Solid triangle indicates the NOS analog 4.

Conserved posttranslational modifications

Central to the nos gene cluster are five orfs, nosDEFGH, the deduced products of which are highly homologous to the conserved posttranslational modification enzymes for affording the thiopeptide-characteristic macrocyclic core of previously characterized thiopeptides (Figure 3b). They encode a cyclodehydratase/dehydrogenase complex (NosG and NosF, 36% and 29% identity to TsrO and TsrM, respectively), a dehydratase pair (NosE and NosD, 29% and 32% identity to TsrJ and TsrK, respectively), and NosH, a TsrN-like unknown protein (27% identity). In addition, nosO, which encodes a protein similar to TsrL (28% identity) that was proposed to be involved in the 6-membered nitrogen heterocycle formation along with TsrN, was identified in the downstream region of the gene cluster. The presence of these genes, along with nosM that encodes the ribosomally synthesized precursor peptide (whose SP is enriched by Cys and Ser/Thr residues), is in perfect agreement with the criteria for a thiopeptide biosynthetic machinery and support the newly unveiled common paradigm to furnish the family-specific framework (24). Thus, these encoded enzymes may act on NosM and catalyze 1) the polythiazole formation by nucleophilic addition of each Cys side chain to the proceeding carbonyl group followed by dehydration and dehydrogenation (with one exception of Cys45, whose –SH group is reserved for appending the indolic acid moiety); 2) multiple dehydrations of Ser and Thr residues to generate dehydroamino acids (except for Thr40, presumably due to the regiospecific activity); and 3) intramolecular cyclization to afford the 6-membered heterocyclic ring and complete the biosynthesis of the 26-membered macrocyclic system, yielding the intermediate 1 for further modifications. It should be noted that the resulting 6-membered nitrogen heterocycle may undergo an elimination of the 37-aa LP after dehydration to form the central aromatic pyridine domain, distinct from a reductive reaction required for the dehydropiperidine formation in TSR or SIO-A biosynthesis (24, 25). To support this proposed modification pathway, selected orfs were inactivated to validate their indispensability. As exemplified by the in-frame deletion of nosH, the resulting mutant strain SL4004 completely lost the ability to produce NOS (Figure 6, IV), thereby clearly confirming its involvement in NOS biosynthesis.

Unusual indole side ring system formation

Functional assignment of the remaining orfs in the nos gene cluster revealed four genes, nosIKLN, whose deduced products could serve as the candidates responsible for the biosynthesis and attachment of the indolic acid moiety to the thiopeptide macrocyclic ring system of 1 (Figures 4 and 5). NosL, containing a conserved CxxxCxxC motif essential for binding a [4Fe-4S] cluster, shows sequence similarity to various ThiH-like proteins (e.g. ThiH from Salmonella typhimurium, 26% identity) that are S-adenosylmethionine (AdoMet)-dependent and catalyze the cleavage of the Cα-Cβ of Tyr initiated by a radical Ado intermediate in thiamine biosynthesis (29, 30). Since previous isotope-labeling experiments had confirmed Trp as the precursor of the indolic acid moiety (14), this conversion should involve an unusual carbon side chain rearrangement on Trp. In a manner similar to the ThiH-catalyzed reaction, NosL may activate the substrate Trp by a radical-forming mechanism, and further carbon side chain reconstruction eventually result in the key intermediate 3-methylindolyl derivative (2) (Figure 4). To validate the above proposal, we first inactivated nosL by in-frame deletion. As anticipated, the resulting mutant strain SL4005 completely lost the ability to produce NOS (Figure 6, V). Next, feeding of the chemically synthesized 3-methylindole-2-carboxylic acid (8) into the fermentation culture of SL4005 partially restored NOS production (Figure 6, VI), strongly supporting the central role of NosL in the formation of the indolic acid moiety. Since the conversion of Trp into 2 is novel and requires multiple reactions that, to our knowledge, are not similar to those catalyzed by known enzymes, the identification of NosL in NOS biosynthesis presents an excellent opportunity to investigate its novel chemistry and enzymology.

Figure 4.

Proposed biosynthetic pathway of the indolic acid moiety of NOS.

Figure 5.

Biosynthetic hypothesis for conversion of the thiopeptide framework into NOS. Color coding indicates the posttranslational modifications on the peptide precursor NosM (orange), attachment of the indolic acid moiety to the macrocyclic core (purple), and tailoring steps including hydroxylations on Glu43 at the γ-position and central pyridine domain at the C-5 position, and final cleavage of an acrylic acid to afford NOS (blue).

NosN shares high sequence similarity to various putative methyltransferases with a radical S-AdoMet-dependent manner, including Tlm-Orf11 (32% identity) in tallysomycin biosynthesis (31). To determine its involvement and corresponding function in NOS biosynthesis, we inactivated nosN by in-frame deletion. As shown in Figure 6, the resulting mutant strain SL4006 indeed lost the ability of NOS production (Figure 6, VII) and, intriguingly, accumulated a new compound with ultraviolet (UV) absorptions at λ_max 254 nm, 280 nm and 350 nm, characteristic for all thiopeptides. HR-ESI-MS analysis showed its positive [M + Na]⁺ ion at m/z = 1270.1528 (± 0.003), establishing the molecular formula as C₅₃H₄₅N₁₃O₁₂S₆Na (1270.1257 calculated). Although it is too labile for complete structural elucidation by NMR spectroscopic methods, this compound, upon extensive tandem MS spectrometry analysis, was deduced to be the intermediate 4 (Figure 7), a side ring-opened NOS analog that contains the 3-methylindolyl moiety appended to the –SH group of Cys45 on the thiopeptide framework via a single thioester linkage. These findings not only support that NosN functions as a methyltransferase and acts on the 3-methylindolyl group at the C-4 position for furnishing the 3,4-dimethylindolyl moiety, but also shed light into the timing of the individual steps for indole side ring biosynthesis. As shown in Figure 5, the intermediate 4 can be resulted from the thioesterification between 2 and Cys45 of 1 (pathway I). NosK, containing a putative acyltransferase domain at the C-terminus, may contribute to this enzymatic process. The NosN-catalyzed methylation may occur on 4 to yield the intermediate 5, and the subsequent hydroxylation of the resulting C-4 methyl group could give the intermediate 6. The priority of the S- linkage formation to the methyl group hydroxylation in NOS biosynthesis is consistent with previous studies that 8 and 3,4-dimethylindole-2-carboxylic acid 9, but not 4-(hydroxylmethyl)-3-methylindole-2-carboxylic acid, can be activated as the adenylate derivatives and incorporated into NOS (14). However, it does not exclude another possibility of the substrate specificity of NosN. It may act on 2 (instead of the 3-dimethylindolyl moiety-attached compound 4) to generate the 3,4-dimethylindolyl derivative 3 (Figure 4), which could then be appended directly onto the thiopeptide framework via the NosK-catalyzed thioesterification to give the intermediate 6 (pathway II, Figure 5). Finally, 6 may undergo an intramolecular esterification, presumably catalyzed by NosI (belonging to an AMP-dependent CoA ligase family), to form the O- linkage between the indole moiety and Glu43 on the macrocyclic core, close the 19-membered side ring system and give the product 7 for further tailoring (Figure 5).

Figure 7.

MS/MS spectrum of the [M + Na]⁺ ion at m/z = 1269.74 corresponding to the NOS analog 4. The obtained fragments clearly indicate a permutation and combination pattern (labeled in color) for cleavage on the free carboxylic group(s) of the C-terminal dehydroalanine acid and/or side chain of Glu43, and/or on the linkage of the 3-methylindole moiety (2) to Cys45 of the macrocyclic core.

Tailoring to NOS and regulation

To eventually produce NOS, post-modifications on 7 are postulated to proceed with a set of tailoring enzymes encoded by nosA, nosB and nosC as outlined in Figure 5. NosB and NocC, belonging to a P450 hydroxylase family, may oxidatively act on Glu43 at the γ-position and the central pyridine domain at the C-5 position, giving the hydroxylation pattern characteristic to all known e series thiopeptides. The final formation of NOS could be resulted from the removal of an acrylic acid group to form the C-terminal amide moiety, and NosA serves as a candidate to catalyze this reaction. This proposed mechanism is distinct from that in TSR biosynthesis, which employs an asparagine synthetase-like protein TsrC for the terminal amidotransfer (25).

The only putative regulatory gene identified within the nos gene cluster was nosP (Figure 2). It encodes a protein that contains an effector domain of a response regulator at the N-terminus and a transcriptional activator domain at the C-terminus. Inactivation of nosP by in-frame deletion completely abolished NOS production (Figure 6, VIII), validating that NosP functions as a pathway specific regulator to positively control the NOS biosynthesis in S. actuosus.

In summary, we have uncovered the biosynthetic machinery of nosiheptide by cloning, sequencing and characterizing the nos gene cluster in S. actuosus ATCC 25421. Our findings extend the current paradigm for thiopeptide biosynthesis into the e series members with distinct features. Thus, consistent with the similarities in structures to the b or c series thiopeptides whose biosynthetic pathways were established, NOS biosynthesis shares the common strategy for forming the characteristic thiopeptide macrocyclic core that features a ribosomally synthesized precursor peptide and conserved posttranslational modifications. However, the NOS biosynthetic machinery is unique, apparently proceeding via a different route for tailoring the intermediate into the final product by installing an unusual indole side ring system and two regiospecific hydroxyl groups, characteristics of e series thiopeptides. Although Trp serves as a common precursor for the indolic acid moiety of NOS and the quinaldic acid moiety of TSR, the distinct pathways for biosynthesis and attachment onto the thiopeptide macrocyclic cores lead to major differences in the side ring systems. The conversion of Trp into the key 3-methylindolyl moiety (2) requires an unusual carbon side chain rearrangement, in which the S-AdoMet-dependent protein NosL may play a central role with a radical-initiated mechanism. NOS biosynthesis involves many novel enzymes and reactions, and the characterization of these will contribute new chemistry and enzymology to thiopeptide biosynthesis. Since all e series members contain these structurally similar features (e.g. the indole side ring system and the hydroxylation pattern) (1), their biosyntheses may employ similar approaches in affording the NOS-like scaffold, but with additional oxidoreduction(s) and glycosylation(s) for tailoring into the other individual members, such as nocathiacins. Finally, the success in identification of the novel NOS analog 4 by gene inactivation continuously inspires attempts to apply this knowledge to metabolic engineering for structural derivation in the pharmaceutical development of thiopeptide-like antibiotics.

Methods

Bacterial Strains, Plasmids, and Reagents

Bacterial strains and plasmids used in this study are summarized in Table S1. Biochemicals and media were purchased from Sinopharm Chemical Reagent Co., Ltd. and Oxoid Ltd. unless otherwise stated. Restriction enzymes were purchased from TaKaRa Biotechnology Co., Ltd.

DNA Isolation, Manipulation, and Sequencing

DNA isolation and manipulation in Escherichia coli and Streptomyces were performed according to standard methods (32, 33). PCR amplifications were carried out in an Eppendorf AG 22331 thermal cycler using LA Taq DNA polymerase (TaKaRa). Primer synthesis and DNA sequencing were performed at Shanghai Invitrogen Biotech Co., Ltd. and the Chinese National Human Genome Center. For cloning the cyclodehydratase gene from S. actuosus, PCR amplification was carried out by using the degenerate primers ThioF (5′-TAC GAG ACC TCC AAY GGN TGY GCN-3′) and ThioR (5′-GTG GCC RAA SGT CAT NGG-3′) described previously (24). Gene inactivations and complementation in this study are described in the Supporting Information.

Construction of the Genomic Library

The genomic library of S. actuosus ATCC 25421 was constructed in pOJ446. Briefly, the total DNA of S. actuosus was randomly sheared to approximately 30–40 kb fragments (judged by electrophoresis at 4 °C), which were blunt-ended by the treatment with the End-Repair Enzyme Mix (EPICENTRE Biotechnologies). The resultant DNA fragments were then ligated into the EcoRV site of pOJ446. DNA packaging and transfection were carried out by using the Packagene Lambda DNA Packaging System (Promega) and E. coli LE392 according to the manufacture's instructions.

Sequence Analysis

ORFs were deduced from the sequence with the assistance of FramePlot 4.0beta program (http://nocardia.nih.go.jp/fp4/). The corresponding deduced proteins were compared with other known proteins in the databases by available BLAST methods (http://www.ncbi.nlm.nih.gov/blast/). Amino acid sequence alignments were performed with CLUSTALW from BIOLOGYWORKBENCH 3.2 software (http://workbench.sdsc.edu).

Fermentation, Isolation and Analysis of NOS

S. actuosus wild-type and recombinant strains were grown on ISP2 agar plates at 30 °C for sporulation. To produce NOS, 50 μl of S. actuosus or recombinant strains spores were inoculated into a 250-ml flask containing 50 ml of seed medium (sucrose 2.0%, corn steep liquor 3.0%, peptone 0.5% and CaCO₃ 0.5%) and incubated at 28 °C and 250 rpm for 24 hrs. Ten ml of seed culture was then transferred into 100 ml of fermentation medium (Pharmamedia cotton meal 1.0%, NaCl 0.3%, glucose 3.0%, 2 × trace elements solution 0.5% (32) and CaCO₃ 0.3%, pH 7.0) in a 500-ml flask and incubated at 28 °C for 4 days. For chemical feeding of the synthetic 3-methylindole-2-caboxylic acid (8) to restore NOS production in the nosL mutant SL4005, 8 was added into the fermentation culture of SL4005 on day 2 at a final concentration of 20 μg ml^-1. For NOS isolation, each 100 ml of culture broth was centrifuged for 10 minutes at 6,000 rpm. After removing the supernatant, the precipitate was soaked with 100 ml of ethyl acetate for 6 hrs. The extract was then concentrated in vacuum and resolved in 1 ml of methanol for further analysis. HPLC analysis was carried out on an Agilent ZORBAX SB-C18 column (4.6×250 mm, part number 880975-902, S/N USCL024998), which was equilibrated with 85% solvent A (H₂O containing 0.1% TFA) and 15% B (CH₃CN containing 0.1% TFA), and developed with the following program: 0 to 3 min, constant 85% A/15% B; 3 to 6 min, a linear gradient from 85% A/15% B to 60% A/40% B; 6 to 12 min, constant 60% A/40% B; 12 to 19 min, a linear gradient from 60% A/40% B to 45% A/55% B; 19 to 22 min, a linear gradient from 45% A/55% B to 15% A/85% B; 22 to 28 min, constant 15% A/85% B; and 28 to 32 min, a linear gradient from 15% A/85% B to 85% A/15% B. This was carried out at a flow rate of 1 ml min^-1 and UV detection at 254 nm using an Agilent 1100 HPLC system (Agilent Technologies). The identity of NOS was confirmed by HPLC-ESI-MS analysis (for HPLC, Agilent 1100 used; and for ESI-MS, Thermo Fisher LTQ Fleet used) performed under the same conditions. NOS showed an [M+H]⁺ ion at m/z 1222.36 consistent with the molecular formula C₅₁H₄₃N₁₃O₁₂S₆ (1221.15 calculated).

Tandem Mass Spectrometry Analysis of the NOS analog 4

Production and isolation of 4 was carried out according to the method described previously (24). Tandem Mass Spectrometry analysis of 4 was performed on a LCQ Fleet Ion Trap MSn tandem mass spectrometer (Thermo Fisher Scientific) (Figure 7).

Synthesis of 3-Methylindole-2-Carboxylic Acid (8)

The synthesis of 8 was achieved by following the previously described method (34). ¹HNMR assignments in CDCl₃ at 300 MHz are as follows (s means singlet, d means doublet, and t means triplet): δ8.85 (s, 1H), δ7.69 (d, 1H, J = 7.8 Hz), δ7.40 (d, 1H, J = 9.3 Hz), δ7.35 (t, 1H, J = 8.7 Hz), δ7.16 (t, 1H, J = 7.2 Hz), δ2.67 (s, 3H). ESI-MS m/z = 174.0 (M - H⁺).