Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2023 Mar 5.
Published in final edited form as: Nat Chem. 2022 Sep 5:10.1038/s41557-022-01020-0. doi: 10.1038/s41557-022-01020-0

Genome-based discovery and total synthesis of janustatins, potent cytotoxins from a plant-associated bacterium

Reiko Ueoka 1,3,#, Philipp Sondermann 2,4,#, Stefan Leopold-Messer 1,#, Yizhou Liu 5,6, Rei Suo 7,8, Agneya Bhushan 1, Lida Vadakumchery 1, Ute Greczmiel 1, Yoko Yashiroda 9, Hiromi Kimura 9, Shinichi Nishimura 9,10,11, Yojiro Hoshikawa 10, Minoru Yoshida 9,10,11, Annette Oxenius 1, Shigeki Matsunaga 7, R Thomas Williamson 5,12, Erick M Carreira 2,*, Jörn Piel 1,*
PMCID: PMC7613652  EMSID: EMS146508  PMID: 36064972

Abstract

Host-associated bacteria are increasingly recognized as underexplored sources of bioactive natural products with unprecedented chemical scaffolds. A recently identified example is the plant root-associated marine bacterium Gynuella sunshinyii of the chemically underexplored order Oceanospirillales. Its genome contains at least 22 biosynthetic gene clusters suggesting a rich and mostly uncharacterized specialized metabolism. In this work, in silico chemical prediction of a non-canonical polyketide synthase cluster led to the discovery of janustatins, structurally unprecedented polyketide alkaloids with potent cytotoxicity that are produced at minute quantities. A combination of MS and 2D NMR experiments, density functional theory calculations of 13C chemical shifts, and semiquantitative interpretation of T-ROESY data were conducted to determine the relative configuration, which enabled the total synthesis of both enantiomers and assignment of the absolute configuration. Janustatins feature a previously unknown pyridodihydropyranone heterocycle and an unusual biological activity consisting of delayed, synchronized cell death at subnanomolar concentrations.

graphic file with name EMS146508-f007.jpg

Open in a new tab

In the search for new chemical entities with bioactivity, genome-based methods provide powerful means of pinpointing chemodiversity hotspots among the vast domain of bacterial life. Among the resources with particular promise for chemical novelty recognized by these methods are animal- or plant-associated microbiota. Studies on these organisms have revealed various producers with large sets of biosynthetic gene clusters (BGCs)1-7, as well as enriched biosynthetic enzyme families that are underrepresented in conventionally studied bacterial groups1,8-10, suggesting major potential for molecular novelty. An example of such enzymes are the trans-acyltransferase polyketide synthases (trans-AT PKSs)11. Some of their polyketide products, such as mupirocin12,13 and the virginiamycins14-16, are well-known clinical antibiotics from Pseudomonas and Streptomyces strains, but much of the trans-AT PKS diversity resides in chemically underexplored host-associated bacteria. Compared to the more commonly known cis-AT PKS family (the “textbook” PKSs), trans-AT systems contain a much wider range of functional components that introduce unusual structural features into polyketides17-20. Trans-AT PKS-based genome mining, i.e., natural product discovery guided by genomic predictions, is therefore an attractive strategy to identify pharmacological leads with previously unknown chemical scaffolds.

A study on chemically unexplored bacterial taxa with high numbers of biosynthetic gene clusters (BGCs) had suggested Gynuella sunshinyii YC625821 as a rich source of natural products4. G. sunshinyii is a halophilic bacterium of the order Oceanospirillales isolated in South Korea from roots of the sedge Carex scabrifolia growing in a tidal flat21. Its genome contains at least 22 natural product BGCs, of which six belong to the trans-AT PKS type, the highest number reported to date for any organism4. Four of these are hybrid trans-AT PKS-nonribosomal peptide synthetase (NRPS) systems, including one (subsequently termed jan BGC) that could only be manually identified at the time of the study, as it was not detected by the automated gene cluster annotation tool AntiSMASH22 prior to version 5.0. Previous work confirmed the in silico-predicted chemical richness of G. sunshinyii by identifying compounds 1-12 from the strain YC6258 (Fig. 1)4,23-26, accounting for 8 of the 22 BGCs. Here we report the genome-guided identification and total synthesis of the jan BGC products, structurally intriguing polyketide alkaloids with exceptional and unusual bioactivity. Janustatins caused an unusual effect of delayed but synchronized cell death in cytotoxicity assays at nano- to picomolar concentrations after three days of incubation. Janustatins exhibit a novel chemical scaffold featuring a pyridodihydropyranone moiety that is to our knowledge unique among known natural products and unexpectedly rare among synthetic compounds. Modular synthetic routes were developed for both enantiomers, revealing potent bioactivity only for the natural isomer.

Fig. 1. Biosynthetic loci identified in the genome of G. sunshinyii YC6258.

Fig. 1

Open in a new tab

BGCs (I-XXII) were identified by AntiSMASH22 and manual analysis. Structures of natural products currently assigned to BGCs are shown4,23-26. Janustatin as the focus of this study is shown in the center of the circle representing the genome, previously reported compounds are shown outside. The color code for pathway types is explained in the box. RiPP: ribosomally synthesized and post-translationally modified peptide.

Results

The jan biosynthetic product is a structurally unusual polyketide alkaloid

To identify the product of the jan BGC, we attempted to predict the polyketide structure from the biosynthetic gene architecture. The cluster contains a PKS (janD) and a NRPS (janE) gene for natural product core assembly and further gene candidates encoding two predicted desaturase homologs (janA and janB) and one free-standing enoyl-reductase (ER, janC) (Fig. 2, Supplementary Table 1). Only a degraded version of a trans-AT gene was identified, perhaps due to complementation from another pathway. To obtain clues about the polyketide product, the jan cluster was subjected to a phylogenetic analysis of ketosynthase (KS) domains present in each PKS module27 and using our recently developed web application TransATor (http://transator.ethz.ch)23. For the PKS modules of the jan cluster, the prediction suggested a core polyketide structure as shown in Fig. 2 (Supplementary Table 2), which was extended by an ambiguous amino acid suggested by the NRPS module JanE. A closer analysis of domains other than KSs revealed various non-canonical features as common for trans-AT systems. For example, the first module exhibits an unusual organization that was also reported from the apratoxin28,29 and bryostatin30,31 assembly lines and comprises two methyltransferase (MT) domains flanking a pseudo-GCN5-related N-acetyltransferase (GNAT) domain32. In apratoxin, these domains were shown to install a t-butyl group in a pivaloyl starter unit33,34. The subsequent series of PKS modules extensively deviates from textbook PKS rules as well. No conclusive prediction for the amino acid incorporated by the NRPS JanE was obtained using NRPSPredictor235.

Fig. 2. The jan gene cluster in G. sunshinyii YC6258.

Fig. 2

Open in a new tab

a, Architecture of the gene cluster and domain organization of the PKS and NRPS proteins JanD and JanE. Ψ, pseudogene. Further abbreviations not defined in the main text: DH, dehydratase; KS0, non-elongating KS; OMT, O-methyltransferase; C, condensation domain; P, NRPS-para261 domain36. Small black circles represent acyl/peptide carrier proteins. b, Prediction of the putative natural product based on the computational prediction tool TransATor23 and manual refinements (Supplementary Fig. 1, Supplementary Table 2). R and R’ could not be predicted.

Assuming an acetyl starter (Fig. 2b, R = CH3) and glycine as a placeholder for the amino acid (Fig. 2b, R’ = H), the predicted structure corresponded to the molecular formula (C21H37NO6). To search for metabolites with similar mass and atom ratios, G. sunshinyii was cultivated in 100 mL test volumes, and ethyl acetate extracts were analyzed by liquid chromatography coupled to high-resolution mass spectrometry (LC-HRMS). The data revealed a candidate ion peak at m/z 448.2684 [M+H]+, which had a suggested molecular formula of C25H38NO6 (Δ -1.0 mmu) (Supplementary Fig. 2). To isolate the compound, 15 L of combined G. sunshinyii cultures were extracted with ethyl acetate and subsequently separated by several HPLC runs. MS-guided fractionation yielded 0.8 mg of janustatin A (13) and two congeners janustatin B (14, 0.4 mg) and C (15, 0.3 mg) (Fig. 3). Detailed interpretation of 2D NMR data of compound 13 (Supplementary Fig. 3-12) revealed three units I-III (Fig. 3a). While I and II could be connected based on HMBC correlations, several constitutional isomers for their connection to unit III were possible (see Supporting Information).

Fig. 3. Structure elucidation of janustatins.

Fig. 3

Open in a new tab

a, Key COSY and HMBC correlations of the partial structures (I, II and III). b, DFT calculations were performed to determine the connection of II and III by comparison of carbon chemical shifts. The structure of the constitutional isomer with the lowest mean absolute error for the prediction compared to the experimental values is shown, suggesting structure 13. Predicted shifts of other isomers and tautomers are shown in Supplementary Fig. 14. c, Relative configuration of janustatin A (13) and structures of other isolated congeners 14 and 15. The trans-configuration of red bonds is based on a large homonuclear coupling constant (3JHH = 11.5 Hz), the E-configuration of the green bond on ROESY correlations, and the relative configurations of the blue bonds on the comparison of ROESY-derived inter-proton distances <rHH>ROE with average distances calculated from DFT-optimized conformations <rHH>DFT. d, Proposed rearrangement mechanism of janustatin A (13) to janustatin A’ (13’) in the presence of MeOH.

Determining the constitution and relative configuration by DFT calculations and T-ROESY

To distinguish between possible constitutional isomers of the nitrogen-containing moiety of 13, Density Functional Theory (DFT) calculations were performed with the candidate structures. Structural suggestions were evaluated with the Spartan16 software37 using molecular mechanics minimization with the Merck molecular force field (MMFF) followed by DFT geometry optimization and chemical shift calculation using the EDF2 density functional and 6-31G* basis set. The experimental data agreed quite well with DFT predictions for structure 13 but showed large differences with the other three constitutional isomers, or their tautomers (Fig. 3b, Supplementary Fig. 14). A careful analysis of HMBC data showed a weak long-range cross-peak from H(19) to C(17), which also supported this constitution.

The relative configuration of structure 13 was further determined by 3JHH and T-ROESY data38. The T-ROESY experiment largely eliminates the TOCSY effect present in conventional ROESY. With a proper choice of spinlock field strength, the effective cross-relaxation rate is approximately the average of the transverse and longitudinal cross-relaxation rates. Assuming isotropic rotational diffusion, the T-ROESY cross-peak intensity was calibrated to obtain the inter-proton distance using a known reference distance (rH11-H13 = 2.2 Å). First, the heterocyclic protons H(13) and H(14) display a large “anti” vicinal coupling of 11.5 Hz, consistent with a relative configuration (marked by an asterisk) of (13S*,14R*). Next, the configuration of the two chiral centers on unit II relative to the heterocycle was determined as (11R*,12S*,13R*,14S*), via comparison of T-ROESY-derived inter-proton distances (<rHH>ROE) with corresponding average distances calculated from DFT-optimized conformations (<rHH>DFT) for the following four diastereomers: (11S*,12S*,13S*,14R*), (11S*,12R*,13S*,14R*), (11R*,12S*,13S*,14R*), and (11R*,12R*,13S*,14R*) (Supplementary Fig. 15, Supplementary Table 5). The distance between H(11) and H(13) was 2.2 Å in all diastereomers, which was considered fixed and used as the reference for ROE calibration. Because unit I and units II & III are spatially separated by the “(E)” olefinic plane such that the configuration of unit I is not expected to significantly influence the conformation of units II & III, the configuration of unit I was arbitrarily fixed as (6S,8R) for these calculations. Finally, the configuration of unit I was also determined by comparing <rHH>ROE and <rHH>DFT through consideration of the following four diastereomers: (6S*,8S*,11S*,12R*,13S*,14R*), (6S*,8R*,11S*,12R*,13S*,14R*), (6R*,8S*,11S*,12R*,13S*,14R*), and (6R*,8R*,11S*,12R*,13S*,14R*). Only (6S*,8R*,11S*,12R*,13S*,14R*) showed excellent agreement between <rHH>ROE and <rHH>DFT whereas all other diastereomers would involve strong ROEs between far-separated protons, concluding the relative stereochemical analysis (Supplementary Fig. 15, Supplementary Table 5). The absolute configuration of C(11) and C(13) in 13 and 13' was suggested to be (11S,13S) by in silico prediction of the intermediate type accepted by the KS domains of module 5 and 7 (Supplementary Fig. S1 and Supplementary Table 2)23,27. Furthermore, the (13S) configuration is indicated by the predicted KR domain product specificity of module 723 (Supplementary Table 2). This analysis suggested (6S,8R,11S,12R,13S,14R) as the absolute configuration of janustatin A.

Isolation of janustatin A congeners

During solvent screening for NMR measurements, we realized that 13 was slowly converted to another form in CD3OD (Supplementary Fig. 16). LC-MS analysis showed the existence of a new compound 13' with the same mass as 13 (Supplementary Fig. 17), for which the 2D NMR data supported a γ-lactone instead of the δ-lactone ring present in 13 (Supplementary Fig. 18-23). A proposed mechanism of this conversion resulting in a “Janus-like” molecule is shown in Fig. 3.

We isolated two additional janustatin congeners. Janustatin B (14) had a predicted molecular formula of C25H39NO5 (Δ -1.0 mmu), which was one oxygen less and two H more than that of 13 (Supplementary Fig. 24). The 1H and HSQC spectra showed an additional methylene signal (Supplementary Fig. 25, 26). Further interpretation of 2D NMR data determined the structure of 14 (Fig. 3, Supplementary Fig. 27-30), in which the ketone group at C-5 in 13 was replaced by a methylene unit (Fig. 3). Janustatin C (15) had a molecular formula of C25H39NO7 (Δ -0.2 mmu) (Supplementary Fig. 31). 15 differs from 13 by one additional formal H2O unit. An upfield shift at C(13) compared to that of 13 indicated that the lactone ring in 15 was opened (Supplementary Fig. 32, 33). Detailed analysis of 2D NMR data confirmed the structure of 15 (Fig. 3 and Supplementary Fig. 33-37).

The structure of janustatins bears limited similarity to previously identified compounds. In addition to the unusual t-butyl-primed, polymethylated carbon chain, they feature a pyridino[3,2-b]dihydro-2H-pyran-2-one moiety that has, to our knowledge, not been reported from any natural product (Supplementary Table 9). Unexpectedly rare among synthetic compounds, the heterocyclic system is almost exclusively known in the context of pyridinocoumarins, which are of interest as drug leads39 (Supplementary Table 9). The bicyclic moiety might be a valuable addition to combinatorial synthetic compound libraries.

Total synthesis of janustatin A and its enantiomer

In view of the high number of non-protonated carbons at the bicyclic motif, we set out to confirm our initial structural hypothesis by total synthesis. This endeavour also aimed at confirming absolute as well as relative configurations of janustatin A (13) previously predicted computationally. The unprecedented pyridino-dihydropyranone motif necessitated careful retrosynthetic analysis bearing in mind the uncertain stereochemical assignment of janustatin A. Consequently, our retrosynthesis disconnected the natural product into three fragments of equal size (Fig. 4) according to the following guidelines: 1) The pyridinodihydropyranone moiety would be introduced late in the synthesis to allow variation of the respective precursor to account for the paucity of reports regarding this fragment’s reactivity. 2) All possible stereoisomers should readily be accessible. 3) All stereogenic centers should be installed with high selectivity and predictability. Special emphasis is given to the installation of stereocenters at C(6) and C(8) given the difficulties in establishing their configuration by spectroscopic methods. 4) γ- versus δ-lactone formation needs to be considered because janustatin A rearranges to the γ-lactone in methanol, which may suggest the latter is the thermodynamically favored isomer. These strategic guidelines led back to fragments 16, 17 and 18 as outlined in Fig. 5. Aldehyde 17 was envisioned to act as a linker for the consecutive addition of the two outer fragments 16 and 18. The protected secondary alcohol in 17 was to serve as key stereocontrol element inducing both -controlled addition of the alkyne fragment and the metalated pyridine.

Fig. 4. Total synthesis of janustatin A.

Fig. 4

Open in a new tab

aReagents and conditions: (I) [Ir(COD)Cl]2 (0.5 mol%), 21 (1.5 mol%), Et2MeSiH (1.2 equiv), 74%, dr = 20:1; (II) TBSOTf, 2,6-lutidine, 84%; (III) (i-Bu2AlH)2, CH2Cl2, 78%; (IV) DMP, t-BuOH, CH2Cl2, 96%; a) LDA, LiCl, 24 (1.6 equiv), THF, -78 °C to r.t., (b) LDA, H3NBH3, THF, 86% (2 steps); (c)) NMM, n-Pr4NRuO4, 4 Å molecular sieve, CH2Cl2; (d) PPh3, Zn, CBr4, 71% (2 steps), (e) n-BuLi, THF, -78 °C, then MeI, 88%; (f) Cy2BH, Et2Zn, EtZnONf, PhMe-CH2Cl2 (1:1), -20 °C, 46%; (g) Me3OBF4, 1,8-bis(dimethylamino)naphthalene, CH2Cl2, 89%; (h) AcCl, MeOH-THF, 0 °C, 83%; (i) DMP, t-BuOH, CH2Cl2, 91%; (j) t-BuLi, LaCl3·LiCl, THF, -78 °C, dr = 1:1.2, 85%; (k) DMP, t-BuOH, CH2Cl2, 89%; (l) Pd/C, AcOEt; (m) DMP, t-BuOH, CH2Cl2, 71% (2 steps); (n) 16 (4 equiv), n-BuLi (4 equiv), then 28, PhMe-Et2O, 57%; (o) HgO, HgCl2, acetone-water, 82%; (p) Me3Si(CH2)2OH; NaH, DMF, 50%; (q) NMM, n-Pr4NRuO4, 4 Å molecular sieve, CH2Cl2, 89%; (r) HCO2H-water (1:1), 50 °C, 86%; (A) i-Pr2NH (1.2 equiv), n-BuLi (1.1 equiv), THF, 29, -78 °C, DMF, 74%; (B) Z-butene-1,4-diol (3 equiv), p-TSA (0.1 equiv), C6H6 reflux, 93%; (C) 32 (5 mol%), LiEt3BH, PhMe, 100%. Both enantiomers janustatin A (13) and ent-janustatin A (36) were synthesized. The natural enantiomer is shown throughout the synthesis.

Fig. 5. Proposed biosynthesis of janustatins.

Fig. 5

Open in a new tab

The trans-AT PKS JanD assembles the polyketide portion and passes it onto the NRPS JanE. In a mechanism diverting from the classical assembly line model, a β-alanine residue (pink) selected by the NRPS adenylation (A) domain would serve as the nucleophile to attack the β-position instead of the thioester carbon. The topology is reminiscent of β-branching first reported for rhizoxin biosynthesis (inset). Finally, hydroxylation and desaturation, possibly catalyzed by the desaturase homologs JanA and JanB, generate janustatin A (13). Numbers above PKS modules refer to consecutive KS numbers. P, NRPS-para261 domain36 (unknown function); *, hypothetical trans-acting domains, possibly from other modules in JanD, speculated to complement missing domains in some PKS modules. JanC probably acts as an ER.

In a forward sense, aldehyde 17 was prepared through Morken’s iridium-catalyzed reductive syn-aldol reaction between benzyloxyacetaldehyde 19 and methyl acrylate 20 (62%, 96% ee, dr>16:1 after purification)40,41. The newly formed alcohol was protected (TBSOTf, 2,6-lutidine, 84%) to give 22 and aldehyde 17 was obtained through a DIBAL-reduction/DMP-oxidation sequence (80%, 2 steps)42. The alkyne fragment was prepared adapting a strategy by Fürstner and coworkers43. Myers-alkylation of pseudoephedrine propionamide 24 with Roche ester-derived iodide 23 delivered 25 with both desired stereocenters of fragment 18 established. Reduction of the amide to alcohol 25 (86%, 2 steps) and Ley-Griffith oxidation delivered the corresponding aldehyde, which was subjected to the Corey-Fuchs sequence with subsequent in situ methylation to 18 (62%, 3 steps). Diastereoselective vinylation of aldehyde 17 was then achieved by regioselective hydroboration of 18 with dicyclohexylborane and subsequent transmetalation to zinc44,45. Addition of stoichiometric ethylzinc nonaflate and aldehyde 17 provided the allylic alcohol with high anti-selectivity (dr>5:1, 46%). It is noteworthy that in our hands it was necessary to use the EtZnONf as a solution in CH2Cl2/toluene, as it underwent degradation as determined by titration upon evaporation of the solvent. Subsequent methylation with sodium hydride and methyl iodide resulted in significant scrambling of the secondary silyl group, and, thus, we resorted to trimethyloxonium tetrafluoroborate to obtain 26 (89%). Transformation of the primary silyl ether was achieved through selective deprotection with HCl in THF-methanol (83%) and an oxidation-alkylation-oxidation sequence to give 27 (69%, 3 steps). While alkylation with tert-butyllithium alone delivered the tert-butyl carbinol in 52% yield, the use of stoichiometric LaCl3·2LiCl provided a significant higher yield (82%, inconsequential mixture of diastereomers, dr = 1.2:1). Deprotection of the benzyl ether and Dess-Martin oxidation (71%, 2 steps) delivered aldehyde 28 setting the stage for pyridine addition.

Development of a suitable hydroxypyridine precursor was challenging. Masking the hydroxypyridine as the less polar 2-bromo-4-chloropyridine 29 proved key. The bromide allows selective functionalization of the 2-position, while the chloride as a hydroxy surrogate facilitates handling of intermediates. Additionally, derivatization of pyridine 29 through LDA mediated deprotonation in the 3-position allows facile functionalization. Early attempts were directed at using benzyl and PMB protected 3-pyridyl carbinols obtained from reacting deprotonated 29 with BOM-Cl or p-methoxybenzyl chloromethyl ether46. However, after bromine-lithium exchange, addition of these pyridines to aldehydes suffered from low diastereoselectivity (dr anti:syn = 1.2:1 to 1.5:1 in THF), and subsequent hydrogenolysis delivered the respective deoxygenated 3-methyl pyridines. We thus changed our protecting group strategy to acetal protected pyridines prone to acidic hydrolysis. Formylation of 29 with DMF gave carbaldehyde 30, which was then converted to the corresponding acetal prepared from ethylene glycol under acidic conditions (PTSA) in refluxing benzene. Metalation of the latter and addition to aldehyde 28 suffered from product formation in low diastereomeric excess (dr anti:syn = 1.2:1 to 1.4:1 in THF, diethyl ether, toluene or diglyme). Additionally, the adduct proved reluctant to undergo acetal hydrolysis mediated by p-toluenesulfonic acid at ambient temperature and was seen to undergo decomposition upon heating. We hypothesized that the inability of the acetal to undergo hydrolysis could be related to the requirement for formation of a high energy oxocarbenium conjugated to the electron-deficient pyridine/pyridinium. To overcome this issue, we looked to implement the use of a protecting group whose cleavage would bypass this intermediate. For this purpose, formylpyridine 30 was transformed to the dihydro-1,3-dioxepine 31 with Z-butene-1,4-diol in the presence of PTSA in refluxing benzene. Isomerization with an in situ formed nickel hydride complex (from (DuPHOS)NiI2 and LiEt3BH) delivered the enol ether 1647. Following lithium-bromide exchange, the intermediate organolithium was added to aldehyde 28, delivering 33. Interestingly, the diastereoselectivity in this addition of anti:syn > 5:1 was much higher than for the previously tested lithiated pyridine nucleophiles mentioned above. The subsequent hydrolysis of enol ether 33 was achieved with mercuric dichloride and mercuric oxide48 delivering an inconsequential mixture of γ-lactol epimers 34. Next, sodium hydride-induced silyl group migration was used to obtain the six-membered lactol. Oxidation of 34 to lactone 35 (TPAP, 89%) was followed by conversion of the chloropyridine to the hydroxypyridine through hydrolysis in a warm formic acid-water mixture, which resulted in concurrent hydrolysis of the TBS ether49. Janustatin A (13) was isolated in 91% yield, and all spectral data for 13 were in agreement with the natural compound (Supplementary Table 10). Both, janustatin A (13) and ent-janustatin A (36) were synthesized and analysis by chiral HPLC indicated that the configuration of synthetic 13 matched that of the natural product (Supplementary Fig. 38).

Proposed biosynthesis of janustatins

The KS-based product prediction of the PKS protein JanD largely agrees with the open-chain portion of janustatin A (13) (Fig. 2, Supplementary Table 2), but its architecture diverges from textbook colinearity rules in almost every module (Fig. 5). The lack of a C(5) keto group in janustatin B (14) suggests that module 2 installs a reduced moiety or perhaps a mixture of the reduced and the keto species. Neither this module nor modules 3, 4, and 5 contain the complete set of domains required for the series of polyketide moieties in janustatins. Missing domains have been reported from various other trans-AT systems11, and are likely complemented by domains on other modules or by free-standing enzymes that operate in trans.

An unexpected chemical feature that we had not predicted is the unique heterobicyclic unit in janustatins. To obtain insights into its formation, we conducted feeding experiments with 15N-labeled β-alanine that we suspected as a building block incorporated by the NRPS protein JanE. Comparing extracts of labeled and unlabeled cultures by LC-MS showed a mass increase of 1 Da for janustatins after labeling in accordance with the incorporation of one 15N (Supplementary Fig. 39), thus supporting β-alanine as a building block. While this feature, together with the number of PKS modules, accounts for all atoms in the janustatin backbone, the order of the last two building blocks (β-alanine followed by a polyketide extender) appears to be reversed as compared to the architecture of the assembly line (PKS module followed by an NPRS module). Reversed building block topologies have been reported previously from the rhizoxin50 and glutarimide polyketide pathways51, in which a PKS module generates a branched polyketide intermediate (Fig. 5). In an analogous fashion, an amino acid branch installed by JanE might account for the heterocyclic unit, although such a mechanism is new and speculative at this point (Fig. 5). To unequivocally link the jan BGC to janustatin production we deleted the first KS in the PKS (Supplementary Fig. 40). In the resulting mutant G. sunshinyii ΔjanDKS1 janustatin production was abolished while other polyketides could still be detected (Supplementary Fig. 41).

Janustatins are potent cytotoxins causing delayed synchronized cell death

Janustatin A (13) was tested for bioactivity against Acinetobacter baumannii, Staphylococcus aureus, Escherichia coli, Pantoea agglomerans, Enterococcus faecalis, and Pseudomonas aeruginosa over a period of two days and against HeLa human cervical cancer cells within a regular 3-day assay period, but no significant activities were detected (Supplementary Fig. 42). Unexpectedly, while testing for antiviral activity in assays using MC57G murine fibrosarcoma cells infected with or without lymphocytic choriomeningitis virus, janustatin A (13) showed potent cytotoxicity, killing the cells at an IC50 value of 45 nM after two days (Supplementary Fig. 43). To study this effect of potent but delayed cytotoxicity, we conducted more systematic experiments testing the bioactivity of janustatins on various cell lines. Against 3Y1 murine fibroblasts and HeLa human cervical cancer cells, 13 was active at a concentration of 1.8 nM over the course of 96 hours. Until 48 hours of treatment, cells proliferated in flat shapes with protrusions, as did non-treated cells. Then, cells started to lose protrusions, round, and shrink, and stopped dividing after 72 hours of treatment, culminating in cell-death (Fig. 6, Supplementary Fig. 44, Supplementary Video 1-5). The isomer of janustatin A (13') and janustatin B (14) exhibited similar effects against 3Y1 cells (Supplementary Fig. 44, Supplementary Video 6-7). After 96 hours of treatment, 13, 13’, and 14 showed potent cytotoxicity against 3Y1 cells at IC50 values of 0.85 nM, 16.2 nM and 0.27 nM, respectively. Against HeLa cells 13 and 13’ exhibit IC50 values of 0.29 nM and 15.1 nM, respectively (Supplementary Fig. 45, Supplementary Table 11). After 4 days of treatment the synthetic ent-janustatin A (36) exhibited about 3,000-fold lower cytotoxicity against HeLa cells than the natural product (Supplementary Fig. 46).

Fig. 6. Growth of 3Y1 cancer cells treated with 1.8 nM janustatin A (13).

Fig. 6

Open in a new tab

Cells grow normally for the first 2 days. After 3 days, janustatin-treated cells stop dividing and die in a synchronized fashion. Growth of DMSO-treated cells is not affected. Treatment with 0.34 μM of the cytotoxin doxorubicin (dox) results in cell death on the first day (Supplementary Video 1-3). The experiment was performed on HeLa and 3Y1 cells in duplicates for more than ten different concentrations (see Supplementary Figure 45 a-e). For all concentrations above the IC50, a delayed cytotoxic effect was observed. Scale bar: 100 μm.

We also examined the anti-yeast activity of janustatin A using two species, the fission yeast Schizosaccharomyces pombe and the budding yeast Saccharomyces cerevisiae. The compound suppressed yeast cell growth at a wide range of concentrations. In the fission yeast, 50% growth inhibition was observed from 0.1 to 100 μg/mL (Supplementary Fig. 47, Supplementary Table 11). In the case of amphotericin B, an antifungal drug that shows acute toxicity, no cell growth was observed above 0.25 μg/mL. Janustatin A therefore likely has a fungistatic action.

Discussion

In this work, we used a combination of genomic prediction, targeted isolation, and structure elucidation enabled by computational NMR and total synthesis to identify janustatins, potent cytotoxins from the plant root-associated marine bacterium G. sunshinyii. Produced at trace amounts (most abundant congener janustatin A: ca. 50 μg/L), showing delayed bioactivity that is easily overlooked in standard assays, and encoded by a rare BGC, janustatins provide an example for the value of genomic discovery strategies in accessing elusive bioactives from non-classical bacterial sources. The biosynthetically unusual heterocyclic alkaloid structure of janustatins is only distantly related to other bioactive compounds, raising questions about the mode of action and bioactivity-conferring moieties. Some similarity exists to the piericidins, actinomycetes-derived NADH-ubiquinone oxidoreductase inhibitors52 biosynthesized by a cis-AT PKS53. Compared to the bicyclic moiety of janustatins, piericidins feature a decorated pyridine ring attached to an aliphatic chain, resembling the structure of ubiquinone54. Furthermore, the myxobacterial ajudazols share a bicyclic moiety lacking nitrogen.55 To our knowledge, delayed cytotoxicity as observed for janustatins has not been reported for piericidins or ajudazols. Delayed cell death is known for a few other compounds including 4-quinolones56,57, DNA-intercalating agents58, and nucleobase59,60 and nucleoside analogs61. For some of them, the depletion of mitochondrial DNA is discussed as an underlying mechanism56,61. However, considering the much higher potency of janustatins compared to these agents, a distinct mechanism, such as metabolic modification to an active form, cannot be excluded. With sufficient polyketide quantities accessible by total synthesis, the metabolic fate and cellular target of janustatins can now be studied. The modular total synthesis established here can be readily adapted to study structure-activity relationships or to introduce labels for target identification.

Trans-AT PKSs exhibit an extraordinary biosynthetic versatility that greatly extends the textbook PKS reaction scope shared with fatty-acid synthases11. While canonical polyketide extension is linear, two mechanisms resulting in branched chains are known that involve the attack of β-keto or conjugated olefin groups by carbon nucleophiles50,51. The topology of the unprecedented janustatin heterocycle suggests that the jan PKS uses an NRPS module to accomplish a nitrogen variant of polyketide β-branching. Studies are underway to test this hypothesis and more difficult to rationalize alternatives, such as a post-assembly line rearrangement or thioester transfer from the JanD PKS module 6 to the NRPS module and back to JanD. Utilizing the mechanism in engineered biosynthesis could provide access to compounds with a range of amino acid-derived pyridine and pyrrole moieties and thus substantially broaden the scope of polyketide biosynthesis.

The prominent role of trans-AT PKS products in symbiotic interactions has been observed in diverse host-bacterial systems1,2,4,5,8,9,62, but it remains to be clarified whether the unprecedented number of trans-AT BGCs in the janustatin producer is the basis of a symbiosis with plants. Little is known about the ecology of G. sunshinyii, which is the only species of the genus21,63. More recently, further G. sunshinyii strains were reported from a highly similar habitat (tidal flat cordgrass roots) in the United States, located more than 11,000 km away from the Korean site of the janustatin producer YC625864. This and the fact that the bacteria were isolated from surface-sterilized roots indicate that G. sunshinyii is an endophyte and might form a specific association with marine flat grasses. Endophytic colonizers of salt marsh halophytes are also known from Saccharospirillum and Marinomonas, neighbor genera of Gynuelld65,66 within the order Oceanospirillales. Thus, the rhizosphere of marine plants might be an attractive discovery resource harboring bacteria with an ecology and chemistry distinct from classical producer groups.

Methods

General

LC-ESI mass spectrometry was performed on a Thermo Scientific Q Exactive or LTQ Orbitrap XL mass spectrometer coupled to a Dionex Ultimate 3000 UPLC system operated by Xcalibur 4.1 and Chromeleon Xpress 7.2 (Thermo Scientific) respectively. NMR spectra were recorded on a Bruker Avance III spectrometer equipped with a cold probe at 500 MHz and 600 MHz for 1H NMR and 125 MHz and 150 MHz for 13C NMR at 298 K operated by TopSpin 3.5/4.1 (Bruker). Chemical shifts were referenced to the solvent peaks at δH 2.50 ppm and δC 39.51 ppm for DMSO-d6, δH 3.31 ppm and δC 49.15 ppm for CD3OD, and δH 7.26 ppm and δC 77.16 ppm for chloroform-d. Data were analyzed using Xcalibur 4.1, TopSpin 4.1, MestReNova (Mestrelab Research) and Spartan16 (Wavefunction Inc.). Purification of natural products was achieved on Agilent Infinity 1260 HPLCs operated by Open Lab CDS 2.2 (Agilent). Eukaryotic cell lines were not authenticated as they were used to test cytotoxic effects of compounds.

Prediction of the janustatin structure

For a first structural proposal, the sequences of JanD and JanE were submitted to the automated prediction tool TransATor1 (https://transator.ethz.ch/, Supplementary Fig. 1). Additionally, amino acid sequences of KS domains from PKS proteins from G. sunshinyii were aligned with 651 previously annotated KS domain sequences2,3. Alignments were performed using the MUSCLE algorithm with default settings4. A phylogenetic tree was constructed using the default settings of FastTree version: 2.1.10 +SSE3 +OpenMP (16 threads)5. The intermediate types accepted by KS domains in JanD were inferred from functionally assigned KSs in the same or phylogenetically close clades. For both approaches, the structural prediction was amended by addition of an ambiguous amino acid, suggested by the presence of the monomodular NRPS JanE. Results of the in silico analysis are shown in Supplementary Table 2.

Extraction and isolation of janustatins A-C (13-15)

Gynuella sunshinyii YC6258 was obtained from NITE Biological Resource Center (NBRC). The strain was grown in 15 L marine broth 2216 medium at 30 °C for 3 days on an orbital shaker. The culture was centrifuged, the supernatant was extracted three times with EtOAc, and the supernatant extract was dried and separated by RP-HPLC (Phenomenex Luna 5μ C18, φ 20 x 250 mm, 10.0 mL/min, 260 nm) with a gradient elution from 5% MeCN to 100% MeCN + 0.1% formic acid to afford 20 fractions (Fr.1-20). Fr.10 and Fr.11 were combined and separated by RP-HPLC (Phenomenex Luna 5μ Phenyl-Hexyl, φ 10 x 250 mm, 2.0 mL/min, 200 nm) with 40% MeCN + 0.1% formic acid. The fraction containing 13 was further separated by RP-HPLC (Phenomenex Synergi 4μ Hydro-RP, φ 10 x 250 mm, 2.0 mL/min, 200 nm) with 37% MeCN + 0.1% formic acid, and finally purified by RP-HPLC (Phenomenex Kinetex 5μ C18, φ 10 x 250 mm, 2.0 mL/min, 200 nm) with 35% MeCN + 0.1% formic acid to yield 0.8 mg of janustatin A (13). The fraction containing 15 was separated by RP-HPLC (Phenomenex Kinetex 5μ C18, φ 10 x 250 mm, 2.0 mL/min, 200 nm) with 37% MeCN + 0.1% formic acid, and further separated by RP-HPLC (Phenomenex Synergi 4μ Hydro-RP, φ 10 x 250 mm, 2.0 mL/min, 200 nm) with 42% MeCN + 0.1% formic acid to yield 0.3 mg of janustatin C (15). Fr.14 was separated by RP-HPLC (Phenomenex Luna 5μ Phenyl-Hexyl, φ 10 x 250 mm, 2.0 mL/min, 200 nm) with 50% MeCN + 0.1% formic acid, and the fraction containing 14 was further purified by RP-HPLC (Phenomenex Synergi 4μ Hydro-RP, φ 10 x 250 mm, 2.0 mL/min, 200 nm) with 65% MeCN + 0.1% formic acid to yield 0.4 mg of janustatin B (14).

Spectroscopic Analysis

Janustatin A (13)

white solid; 1H NMR (DMSO-d6) and 13C NMR (DMSO-d6) data, see Supplementary Table 3; 1H NMR (CDCl3) and 13C NMR (CDCl3) data, see Supplementary Table 4; HRESIMS m/z 448.2684 [M+H]+ (calcd. for C25H38NO6, 448.2694).

Janustatin B (14)

white solid; 1H NMR (CDCl3) and 13C NMR (CDCl3) data, see Supplementary Table 7; HRESIMS m/z 434.2891 [M+H]+ (calcd. for C25H40NO5, 434.2901).

Janustatin C (15)

white solid; 1H NMR (DMSO-d6) and 13C NMR (DMSO-d6) data, see Supplementary Table 8; HRESIMS m/z 466.2781 [M+H]+ (calcd. for C25H40NO7, 466.2799).

Structure elucidation of janustatin A (13)

13 had a predicted molecular formula of C25H38NO6, as determined by HR-ESIMS spectroscopy (Supplementary Fig. 2). 1H NMR in conjunction with HSQC data suggested three doublet methyls, three singlet aliphatic methyls, one vinylic methyl, one methoxy group, three protons connected to sp2 carbons, three oxymethines, one methylene, three methines, and two exchangeable protons (Supplementary Fig. 3-6). From the COSY spectrum, three units I-III were deduced (Fig. 3a and Supplementary Fig. 7, 8). A t-butyl group was assigned to C(1)-C(4), which connected to unit I via a ketone C(5) by HMBC correlations from H(1/2/3) to C(1/2/3), from H(1/2/3) to C(4) and C(5), from H(6), H(7), and from H(21) to C(5) (Fig. 3a and Supplementary Fig. 9, 10). Units I and II were connected by HMBC correlations from H(9) and H(23) to C(11), from H(11) to C(9) and C(23), and from H(12) to C(10). ROESY correlations between H(8) and H(23), and between H(9) and H(11) revealed a 9E configuration for the olefinic bond (Supplementary Fig. 13). The methoxy group was attached to C(11) by HMBC correlations from H(11) to C(24) and from H(24) to C(11). The terminus of unit II was elongated with a non-protonated sp2 carbon, C(15), by HMBC correlations from H(13), H(14), and OH(14) to C(15). Another non-protonated sp2 carbon C(16) was connected to C(19) in unit III via the hydroxyl sp2 carbon C(20), which was determined by HMBC correlations from H(18), H(19), and OH(20) to C(20), from H(19) and OH(20) to the non-protonated sp2 carbon of C(16), and from OH(20) to C(19). The presence of a sp2 nitrogen adjacent to C(18) was suggested by 15N-HMBC correlations from H(18) and H(19) to δN 233.8 ppm in DMSO-d6 (Supplementary Fig. 11). Finally, a non-protonated sp2 carbon at δC 161.7 ppm in CDCl3 was connected to this nitrogen by HMBC correlation from H(18). However, considering the molecular formula, a lack of one more carbon signal and ambiguous signals around δC 160 ppm in the 13C and HMBC data made it difficult to determine whether that sp2 carbon had a chemical shift identical to C(15) (Fig. 3a and Supplementary Fig. 9, 10, 12).

Structure elucidation of isomeric janustatin A (13')

13’ had a predicted molecular formula of C25H37NO6, as determined by HR-ESIMS spectroscopy (Supplementary Fig. 17). The 1H NMR and HSQC spectra suggested three doublet methyls, three singlet aliphatic methyls, one vinylic methyl, one methoxy group, three protons connected to sp2 carbons, three oxymethines, one methylene, and three methines. (Supplementary Fig. 18, 19). COSY correlations showed three units a-c (Supplementary Fig. 20, 23). HMBC correlations from H(1/2/3) to C(1/2/3), from H(1/2/3) to C(4) and C(5), and from H(6), H(7), and H(21) to C(5) revealed the connection between a t-butyl group and unit a via a ketone C(5) (Supplementary Fig. 21, 23). Unit a and unit b were connected by HMBC correlations from H(9) and H(23) to C(11), and from H(11) to C(9), C(10), and C(23). The methoxy group was attached to C(11) in unit b, which was determined by HMBC correlations from H(11) to C(24), and from H(24) to C(11). The presence of the 4-hydroxynicotinic acid-containing unit c was suggested by HMBC correlations from H(18) to C(15) and C(20), from H(19) to C(16) and C(20), and a long-range coupling from H(19) to C(17). Finally, unit b was connected to the 4-hydroxynicotinic acid moiety by HMBC correlations from H(13) and H(14) to C(15), and from H(14) to C(16) and C(17). NOESY correlations between H(8) and H(23), and between H(9) and H(11) revealed 9E configuration (Supplementary Fig. 22).

Structure elucidation of janustatin B (14)

Janustatin B (14) had a predicted molecular formula of C25H39NO5, as determined by HR-ESIMS spectroscopy (Supplementary Fig. 24). The 1H NMR and HSQC spectra suggested three doublet methyls, three singlet aliphatic methyls, one vinylic methyl, one methoxy group, three protons connected to sp2 carbons, three oxymethines, two methylenes, and three methines. (Supplementary Fig. 25, 26). COSY correlations showed three units a-c (Supplementary Fig. 27, 30). HMBC correlations from H(1/2/3) to C(1/2/3), C(4) and C(5), from H(5) to C(1/2/3) and C(4), and from H(6) to C(4) revealed the connection between a t-butyl group and unit a. Unit a and unit b were connected by HMBC correlations from H(9) and H(23) to C(11), from H(11) to C(9), C(10), and C(23), and from H(12) to C(10) (Supplementary Fig. 28, 30). The methoxy group was attached to C(11) in unit b, which was determined by HMBC correlations from H(11) to C(24), and from H(24) to C(11). The presence of the 4-hydroxynicotinic acid-containing unit c was suggested by HMBC correlations from H-18 to C(15) and C(20), from H(19) to C(16) and C(20), and a long-range coupling from H(19) to C(17). Unit b was connected to the 4-hydroxynicotinic acid moiety by HMBC correlations from H(13) and H(14) to C(15), and from H(14) to C(16). Finally, a δ-lactone ring was identified by connecting C(13) and C(17), which was suggested by the molecular formula and the downfield shifted 13C chemical shift at C(13). NOESY correlations between H(8) and H(23), and between H(9) and H(11) revealed 9E configuration (Supplementary Fig. 29).

Structure elucidation of janustatin C (15)

Janustatin C (15) had a predicted molecular formula of C25H39NO7, as determined by HR-ESIMS spectroscopy (Supplementary Fig. 31). The 1H NMR and HSQC spectra suggested three doublet methyls, three singlet aliphatic methyls, one vinylic methyl, one methoxy group, three protons connected to sp2 carbons, three oxymethines, one methylene, and three methines. (Supplementary Fig. 32, 33). COSY correlations showed four units a-d (Supplementary Fig. 34, 37). HMBC correlations from H(1/2/3) to C(1/2/3), from H(1/2/3) to C(4) and C-5, and from H(6), H(7), and H(21) to C(5) revealed the connection between a t-butyl group and unit a via a ketone C(5) (Supplementary Fig. 35, 37). Unit a and unit b were connected by HMBC correlations from H(9) and H(23) to C(11), and from H(11) to C(9) and C(23). The methoxy group was attached to C(11) in unit b, which was determined by HMBC correlations from H(11) to C(24), and from H(24) to C(11). Unit b was connected to unit c by HMBC correlations from H(11) and H(25) to C(13), and from H(13) to C(11), C(12), and C(25). The presence of the 4-hydroxynicotinic acid-containing unit d was suggested by HMBC correlations from H(18) to C(15) and C(20), from H(19) to C(16) and C(20), and a long-range coupling from H(19) to C(17). Finally, unit c was connected to the 4-hydroxynicotinic acid moiety by HMBC correlations from H(13) and H(14) to C(15). NOESY correlations between H(8) and H(23), and between H(9) and H(11) revealed a 9E configuration (Supplementary Fig. 36).

Feeding studies of G. sunshinyii with nitrogen labeled β-alanine

G. sunshinyii was grown in 100 mL marine broth 2216 supplemented with 2 mM 15N-labeled β-alanine at 30 °C for 3 days. Ethyl acetate extracts were prepared and analysed by LC-MS (Supplementary Fig. 39).

Antibacterial assays of janustatin A

Cultures of A. baumannii (DSM 3007), S. aureus (ATTC 29213), E. coli (DSM 1103), P. agglomerans (DSM 3439), E. faecalis (DSM 2570) and P. aeruginosa (DSM 11) were grown in Mueller-Hinton broth (Sigma-Aldrich). The cultures were diluted to OD600 = 0.005 and 100 μL were transferred to each well of a 96-well plate. Janustatin A (13) was dissolved in DMSO (3 mM) and a serial dilution was performed to obtain effective concentrations of 150, 75, 37.5, 18.8, 9.4, and 4.7 μM. Sterile medium served as a blank, DMSO as a negative control, kanamycin (25 μg/mL) and ciprofloxacin (30 μg/mL) as positive controls. The optical density (OD600) was measured after 0 h, 25 h, and 48 h on a spectraMAXplus spectrometer (Molecular Devices LLC.). Results are shown in Supplementary Figure 42. After 48 h, the cultures were stamped onto a Mueller-Hinton broth agar plate. Growth on solid medium did not differ from that in liquid cultures.

Antiviral assay

The procedure followed for the lymphocytic choriomeningitis virus neutralization assay is described by Greczmiel et al.6 with a focus-forming readout established by Battegay et al.7. Instead of the normal serum or antibody dilution, a 2-fold serial dilution of janustatin A (13) in DMSO and virus, virus only or medium only. Cells were incubated for 48 h. For determination of the viability of MC57G, murine fibrosarcoma cells (ATCC CRL2295) used for the anti-viral assays, part of the cells was taken up in·phosphate buffered saline (PBS) after being detached from the 24-well plate. Cells were incubated with Near-IR (LIVE/DEAD® Fixable Near-IR Dead Cell Stain Kit, Invitrogen), diluted 1:100 in 1 x PBS. After incubation for 30 min at 4 °C, cells were washed with·PBS and resuspended in FACS buffer (1 x PBS, 5 mM EDTA, 10% FCS). Afterwards, cells were measured immediately using a LSR II FACS machine (BD) operated by FACS:FLowJov9/19 (BD) and FACSDiva (BD). Results are shown in Supplementary Figure 43. Data were analyzed with Prism 8.2 (GraphPad Software).

Bioactivity of janustatins against 3Y1, HeLa, and P388 cells

3Y1 murine fibroblasts and HeLa human cervical cancer cells (RIKEN Cell Engineering Division-CELL BANK) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, and 1% penicillin (10,000 U/mL)–streptomycin (10 mg/L) at 37 °C under an atmosphere of 5% CO2, respectively. P388 murine leukemia cells (RIKEN Cell Engineering Division-CELL BANK) were cultured in RPMI-1640 medium containing 10% fetal bovine serum, and 1% penicillin (10,000 U/mL)–streptomycin (10 mg/L) at 37 °C under an atmosphere of 5% CO2. After overnight preincubation, janustatins (13 in DMSO, 13’ in MeOH or 14 in DMSO) were added to each well of a 96-well microplate containing 200 μL of 3Y1 cell suspension and further incubated for 96 h. After the addition of 50 μL of 3’-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis(4-methoxy-6-nitro)benzenesulfonic acid hydrate (XTT) solution (1 mg/mL) containing 4% of phenazine methosulfate (PMS) solution (0.15 mg/mL) to each well, the plate was further incubated for 4 h. Finally, the absorbance at 450 nm was measured with a microplate reader (Molecular Devices SpectraMax M2). The cytotoxicities of janustatin A against HeLa cells and P388 cells were conducted in the same manner. Results are shown in Supplementary Figure 45 and Supplementary Videos 1-3.

Time-lapse imaging

3Y1 and HeLa cells were cultured as described above. After overnight preincubation, janustatins (13, 13’ or 14) were added to each well of a 96-well plate. The 96-well plate was loaded into an IncuCyteZoom (Essen Bioscience) live-cell imaging system, and the cells were imaged every 2 h for 96 h. Results are shown in Figure 6, Supplementary Figure 44 and Supplementary Video 1-8.

Bioactivity tests of ent-janustatin A (36) against HeLa cells

HeLa cells (ATCC CCL-2) were cultivated at 37 °C, 5% (v/v) CO2 for 3-4 days. Cells were washed twice with PBS buffer (Sigma D8537), 2 mL of 0.05 % trypsin-EDTA solution (Thermo 25300-054) were added and the plate was incubated for 5 min at 37 °C. The cells were resuspended in 10 mL medium (DMEM-GlutaMAX) supplemented with 10% FCS (Eurobio CVFSVF00-01), 1% non-essential amino acids (Thermo 11140-035), and 50 μg/mL gentamicin and centrifuged for 5 min at room temperature. The supernatant was discarded and the pellet resuspended in 10 mL medium. After counting the cells under a ZEISS Axiovert 25 microscope using a Neubauer hemocytometer a 20,000 cells/mL suspension was prepared and 200 μL were transferred into each well of a 96-well plate. After one day of cultivation 2 μL of the ent-janustatin A (36) were added and a 5-fold serial dilution was performed. After four days of cultivation, 50 μL of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, 1 mg/mL in sterile H2O) were added and the cells incubated for 3 h at 37 °C. The supernatant was discarded and 150 μL of DMSO were added to the wells. The absorbance at 570 nm was measured on a spectraMAXplus spectrometer (Molecular Devices LLC). Results are shown in Supplementary Figure 46. Data were analyzed with Prism 8.2 (GraphPad Software).

Bioactivity tests of janustatin A against yeast

Effects of janustatin A (13) on cell growth of fission yeast wild-type cells (Schizosaccharomyces pombe JY1, h-) were examined in YES liquid culture (0.5% yeast extract, 2% glucose, and 225 mg/mL of supplements including adenine, uracil, leucine, histidine and lysine). Mid-log phase inocula were diluted to 0.0033 OD595 and the cells were exposed to various concentrations of janustatin A (13) or amphotericin B at 27 °C for 24 h. After incubation, the turbidity was measured at OD620 using a microplate reader MULTISKAN FC (Thermo Scientific). For further tests with the budding yeast Saccharomyces cerevisiae BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0), overnight culture in YPGal (1% yeast extract, 2% peptone, and 2% galactose) was diluted to 0.01 OD600, aliquots of 198 μL of the diluted cultures were pipetted into each well of 96-well plates, and 2 μL of stock solution of janustatin A (3.2 mg/mL in DMSO), dilutions thereof or DMSO was added directly to the wells containing yeast culture. After incubating at 30 °C for 24 h, the OD600 of each well was measured using a microplate reader SpectraMax M2 (Molecular Device). Results are shown in Supplementary Figure 47. Data were analyzed with Prism 8.2 (GraphPad Software).

Deletion of janustatin BGC

To delete the first ketosynthase domain in janD, 500 base pairs upstream and downstream were amplified from G. sunshinyii with the primer pairs (1_deltaKS_H1_fwd/2_deltaKS_H1_rev and 3_deltaKS_H2_fwd/ 4_deltaKS_H2_rev). The backbone of the plasmid pSW81978 was amplified with the primer pair (5_deltaKS_backbone_fwd/6_deltaKS_backbone_rev; Supplementary Table 12). The three fragments were designed using Geneious 7.1.9 (Biomatter Limited) so that each had 20 base pair overlap to one another. After Gibson assembly plasmids were electroporated into E. coli DH5α pir and grown on Luria-Bertani (LB) broth containing kanamycin (50 μg/mL) and 1% (w/V) glucose. The plasmid pSW8197_deltaKS was isolated with a NucleoSpin Plasmid kit (Macherey-Nagel AG) and sequenced (Microsynth AG).

G. sunshinyii cultures were grown at 30 °C in marine broth 2216 to an OD600=0.8 and washed three times with ice cold sucrose solution (300 mM). Cells were transformed with 900 ng plasmid by electroporation in 2 mm cuvettes at 2.5 kV in an electroporator (Biorad MicroPulser). After 5 hours of recovery in one mL marine broth at 30 °C and 700 rpm, cultures were centrifuged and plated onto marine agar plates containing kanamycin (5 μg/mL) and 1% (w/V) glucose. After four days colonies were picked into marine broth containing kanamycin (12.5 μg/mL) and 1% (w/V) glucose. Two days later, integration of pSW8197_deltaKS into the G. sunshinyii genome by homologous recombination was verified with the two primer pairs (7_pSW8197_fwd/10_deltaKS_downstream and 8_pSW8197_rev/9_deltaKS_upstream). One primer of each pair binds in the genome up-/downstream of the recombination site, the other one in the plasmid. PCR positive colonies were then grown in marine broth without additives for 2 days, and subsequently selected on marine broth containing 0.25% (w/V) arabinose. The primer pair (9_deltaKS_upstream/10_deltaKS_downstream) was used to confirm deletion of the first ketosynthase (Supplementary Figure 40).

Three independent 20 mL cultures of G. sunshinyii wild type and G. sunshinyii ΔjanDKS1 were grown in marine broth, 30 °C 150 rpm for 3 days. Supernatants were extracted with 20 mL ethyl acetate and concentrated. The residues were dissolved in acetonitrile and subjected to UPLC HRESIMS (Supplementary Figure 41).

Supplementary Material

Inventory of supporting information
Supplementary data set 1
Supplementary Information
Supplementary Video 1
Download video file (10.7MB, mp4)
Supplementary Video 2
Download video file (8.8MB, mp4)
Supplementary Video 3
Download video file (2.1MB, mp4)
Supplementary Video 4
Download video file (13.9MB, mp4)
Supplementary Video 5
Download video file (3MB, mp4)
Supplementary Video 6
Download video file (3.5MB, mp4)
Supplementary Video 7
Download video file (3.5MB, mp4)
Supplementary Video 8
Download video file (1.8MB, mp4)

Acknowledgments

We thank Young Ryun Chung and Dmitri Mavrodi for insightful discussions and Didier Mazel for sharing plasmid pSW8197 with us. J.P. acknowledges funding by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement No 742739), the Gordon and Betty Moore Foundation (#9204, https://doi.org/10.37807/GBMF9204), and the Swiss National Science Foundation (NRP72 “Antimicrobial Resistance”, 407240_167051). Y.Y. is supported by a Grant-in-Aid for Scientific Research on Innovative Areas (17H06411), the Japan Society for the Promotion of Science (JSPS). M.Y. and S.N. are supported in part by a JSPS Grant-in-Aid for Scientific Research (S) (19H05640).

Footnotes

Author contributions

All authors devised experiments. R.U. and S.L.-M. isolated and characterized natural products. P.S. prepared and characterized synthetic compounds. Y.L. and R.T.W. analyzed NMR data and conducted calculations. S.L.-M., R.S., Y.Y, H.K., A.B., U.G., S.N., Y.H., M.Y., A.O. and S.M. performed biological assays. S.L.-M., A.B. and J.P. performed bioinformatic analysis. S.L.-M. and L.V. generated the mutant strain. R.U., S.L.-M., P.S., E.M.C., and J.P. wrote the paper with contributions of all authors.

Competing interests

The authors declare no competing interests.

Data availability

The data supporting the findings of this study are available in this article and the Supplementary Information. The genome sequence of G. sunshinyii YC6258 is accessible in GenBank under accession number NZ_CP007142.1, the janustatin gene cluster is located at locus YC6258_05439 (AJQ97469.1) to YC6258_05446 (AJQ97476.1). The closest homologs to these proteins can be found under the accession numbers: WP_158657926.1, WP_086931657.1, WP_038924948.1, RKZ46405.1, WP_087684108.1 and WP_086931660.1. Candidate proteins for the freestanding KR are: WP_044620326.1; WP_044617647.1; WP_044617493.1; WP_044616081.1; WP_044617507.1; WP_044618569.1; WP_044619377.1; WP_044617759.1; WP_144407613.1 and WP_052830250.1. The janustatin BGC was deposited in MIBiG under accession number BGC0002136.

References

  • 1.Wilson MC, et al. An environmental bacterial taxon with a large and distinct metabolic repertoire. Nature. 2014;506:58–62. doi: 10.1038/nature12959. [DOI] [PubMed] [Google Scholar]
  • 2.Elshahawi SI, et al. Boronated tartrolon antibiotic produced by symbiotic cellulose-degrading bacteria in shipworm gills. Proc Natl Acad Sci USA. 2013;110:e295–304. doi: 10.1073/pnas.1213892110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Sit CS, et al. Variable genetic architectures produce virtually identical molecules in bacterial symbionts of fungus-growing ants. Proc Natl Acad Sci USA. 2015;112:13150–13154. doi: 10.1073/pnas.1515348112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Ueoka R, et al. Genome-Based Identification of a Plant-Associated Marine Bacterium as a Rich Natural Product Source. Angew Chem Int Ed. 2018;130:14727–14731. doi: 10.1002/anie.201805673. [DOI] [PubMed] [Google Scholar]
  • 5.Waterworth SC, et al. Horizontal Gene Transfer to a Defensive Symbiont with a Reduced Genome in a Multipartite Beetle Microbiome. mBio. 2020;11:e02430–02419. doi: 10.1128/mBio.02430-19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Zan J, et al. A microbial factory for defensive kahalalides in a tripartite marine symbiosis. Science. 2019;364:eaaw6732. doi: 10.1126/science.aaw6732. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Tobias NJ, et al. Natural product diversity associated with the nematode symbionts Photorhabdus and Xenorhabdus. Nat Microbiol. 2017;2:1676–1685. doi: 10.1038/s41564-017-0039-9. [DOI] [PubMed] [Google Scholar]
  • 8.Rust M, et al. A multiproducer microbiome generates chemical diversity in the marine sponge Mycale hentscheli. Proc Natl Acad Sci USA. 2020;117:9508–9518. doi: 10.1073/pnas.1919245117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Storey MA, et al. Metagenomic Exploration of the Marine Sponge Mycale hentscheli Uncovers Multiple Polyketide-Producing Bacterial Symbionts. mBio. 2020;11:e02997–02919. doi: 10.1128/mBio.02997-19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Freeman MF, et al. Metagenome Mining Reveals Polytheonamides as Posttranslationally Modified Ribosomal Peptides. Science. 2012;338:387–390. doi: 10.1126/science.1226121. [DOI] [PubMed] [Google Scholar]
  • 11.Helfrich EJN, Piel J. Biosynthesis of polyketides by trans-AT polyketide synthases. Nat Prod Rep. 2016;33:231–316. doi: 10.1039/c5np00125k. [DOI] [PubMed] [Google Scholar]
  • 12.Fuller AT, et al. Pseudomonic Acid: an Antibiotic produced by Pseudomonas fluorescens. Nature. 1971;234:416–417. doi: 10.1038/234416a0. [DOI] [PubMed] [Google Scholar]
  • 13.El-Sayed AK, et al. Characterization of the Mupirocin Biosynthesis Gene Cluster from Pseudomonas fluorescens NCIMB 10586. Chem Biol. 2003;10:419–430. doi: 10.1016/s1074-5521(03)00091-7. [DOI] [PubMed] [Google Scholar]
  • 14.Pulsawat N, Kitani S, Nihira T. Characterization of biosynthetic gene cluster for the production of virginiamycin M, a streptogramin type A antibiotic, in Streptomyces virginiae. Gene. 2007;393:31–42. doi: 10.1016/j.gene.2006.12.035. [DOI] [PubMed] [Google Scholar]
  • 15.Mast Y, et al. Characterization of the ‘pristinamycin supercluster’ of Streptomyces pristinaespiralis. Microb Biotechnol. 2011;4:192–206. doi: 10.1111/j.1751-7915.2010.00213.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Cocito CG. Antibiotics of the Virginiamycin Family, Inhibitors Which Contain Synergistic Components. Microbiol Rev. 1979;43:145–198. doi: 10.1128/mr.43.2.145-192.1979. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Pöplau P, Frank S, Morinaka BI, Piel J. An Enzymatic Domain for the Formation of Cyclic Ethers in Complex Polyketides. Angew Chem Int Ed. 2013;52:13215–13218. doi: 10.1002/anie.201307406. [DOI] [PubMed] [Google Scholar]
  • 18.Ueoka R, Bortfeld-Miller M, Morinaka BI, Vorholt JA, Piel J. Toblerols: Cyclopropanol-Containing Polyketide Modulators of Antibiosis in Methylobacteria. Angew Chem Int Ed. 2018;130:989–993. doi: 10.1002/anie.201709056. [DOI] [PubMed] [Google Scholar]
  • 19.Meoded RA, et al. A Polyketide Synthase Component for Oxygen Insertion into Polyketide Backbones. Angew Chem Int Ed. 2018;57:11644–11648. doi: 10.1002/anie.201805363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Ma M, Lohman JR, Liu T, Shen B. C-S bond cleavage by a polyketide synthase domain. Proc Natl Acad Sci USA. 2015;112:10359–11364. doi: 10.1073/pnas.1508437112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Chung EJ, Park JA, Jeon CO, Chung YR. Gynuella sunshinyii gen. nov., sp. nov., an antifungal rhizobacterium isolated from a halophyte, Carex scabrifolia Steud. Int J Syst Evol Microbiol. 2015;65:1038–1043. doi: 10.1099/ijs.0.000060. [DOI] [PubMed] [Google Scholar]
  • 22.Blin K, et al. AntiSMASH 5.0: updates to the secondary metabolite genome mining pipeline. Nucleic Acids Res. 2019;47:W81–W87. doi: 10.1093/nar/gkz310. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Helfrich EJN, et al. Automated structure prediction of trans-acyltransferase polyketide synthase products. Nat Chem Biol. 2019;15:813–821. doi: 10.1038/s41589-019-0313-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Ueoka R, et al. Genome Mining of Oxidation Modules in trans-Acyltransferase Polyketide Synthases Reveals a Culturable Source for Lobatamides. Angew Chem Int Ed. 2020;59:1–6. doi: 10.1002/anie.201916005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Chung YR, Khan H. Novel nitrilase and method for preparing sangivamycin by using same. 2014 [Google Scholar]
  • 26.Helfrich EJN, et al. Evolution of combinatorial diversity in trans-acyltransferase polyketide synthase assembly lines across bacteria. Nat Commun. 2021;12:1422. doi: 10.1038/s41467-021-21163-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Nguyen T, et al. Exploiting the mosaic structure of trans-acyltransferase polyketide synthases for natural product discovery and pathway dissection. Nat Biotechnol. 2008;26:225–233. doi: 10.1038/nbt1379. [DOI] [PubMed] [Google Scholar]
  • 28.Grindberg RV, et al. Single Cell Genome Amplification Accelerates Identification of the Apratoxin Biosynthetic Pathway from a Complex Microbial Assemblage. PLoS One. 2011;6:e18565. doi: 10.1371/journal.pone.0018565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Luesch H, Yoshida WY, Moore RE, Paul VJ, Corbett TH. Total Structure Determination of Apratoxin A, a Potent Novel Cytotoxin from the Marine Cyanobacterium Lyngbya majuscula. J Am Chem Soc. 2001;123:5418–5423. doi: 10.1021/ja010453j. [DOI] [PubMed] [Google Scholar]
  • 30.Buchholz TJ, et al. Polyketide ß-Branching in Bryostatin Biosynthesis: Identification of Surrogate Acetyl-ACP Donors for BryR, an HMG-ACP Synthase. Chem Biol. 2010;17:1092–1100. doi: 10.1016/j.chembiol.2010.08.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Davidson SK, Allen SW, Lim GE, Anderson CM, Haygood MG. Evidence for the Biosynthesis of Bryostatins by the Bacterial Symbiont “Candidatus Endobugula sertula” of the Bryozoan Bugula neritina. Appl Environ Microbiol. 2001;67:4531–4537. doi: 10.1128/AEM.67.10.4531-4537.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Skiba MA, et al. Repurposing the GNAT Fold in the Initiation of Polyketide Biosynthesis. Structure. 2020;28:63–74. doi: 10.1016/j.str.2019.11.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Skiba MA, et al. A Mononuclear Iron-Dependent Methyltransferase Catalyzes Initial Steps in Assembly of the Apratoxin A Polyketide Starter Unit. ACS Chem Biol. 2017;12:3039–3048. doi: 10.1021/acschembio.7b00746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Skiba MA, et al. Biosynthesis of t-Butyl in Apratoxin A: Functional Analysis and Architecture of a PKS Loading Module. ACS Chem Biol. 2018;13:1640–1650. doi: 10.1021/acschembio.8b00252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Röttig M, et al. NRPSpredictor2—a web server for predicting NRPS adenylation domain specificity. Nucleic Acids Res. 2011;39:W362–W367. doi: 10.1093/nar/gkr323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Lu S, et al. CDD/SPARCLE: the conserved domain database in 2020. Nucleic Acids Res. 2020;48:D265–D268. doi: 10.1093/nar/gkz991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Hehre W, et al. Efficient Protocol for Accurately Calculating 13C Chemical Shifts of Conformationally Flexible Natural Products: Scope, Assessment, and Limitations. J Nat Prod. 2019;82:2299–2306. doi: 10.1021/acs.jnatprod.9b00603. [DOI] [PubMed] [Google Scholar]
  • 38.Hwang T-L, Shaka AJ. Cross Relaxation without TOCSY: Transverse Rotating-Frame Overhauser Effect Spectroscopy. J Am Chem Soc. 1992;114:3157–3159. [Google Scholar]
  • 39.Nunez-Vergara LJ, et al. Chromenopyridines: Promising Scaffolds for Medicinal and Biological Chemistry. Curr Med Chem. 2011;18:4761–4785. doi: 10.2174/092986711797535272. [DOI] [PubMed] [Google Scholar]
  • 40.Duffey MO, LeTiran A, Morken JP. Enantioselective Total Synthesis of Borrelidin. J Am Chem Soc. 2003;125:1458–1459. doi: 10.1021/ja028941u. [DOI] [PubMed] [Google Scholar]
  • 41.Zhao C-X, Duffey MO, Taylor SJ, Morken JP. Enantio- and Diastereoselective Reductive Aldol Reactions with Iridium-Pybox Catalysts. Org Lett. 2001;3:1829–1831. doi: 10.1021/ol015859f. [DOI] [PubMed] [Google Scholar]
  • 42.Evans DA, Ratz AM, Huff BE, Sheppard GS. Total Synthesis of the Polyether Antibiotic Lonomycin A (Emericid) J Am Chem Soc. 1995;117:3448–3467. [Google Scholar]
  • 43.Ahlers A, De Haro T, Gabor B, Fürstner A. Concise Total Synthesis of Enigmazole A. Angew Chem Int Ed. 2016;55:1406–1411. doi: 10.1002/anie.201510026. [DOI] [PubMed] [Google Scholar]
  • 44.Stanton GR, Koz G, Walsh PJ. Highly Diastereoselective Chelation-Controlled Additions to α-Silyloxy Ketones. J Am Chem Soc. 2011;133:7969–7776. doi: 10.1021/ja201629d. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Stanton GR, Kauffman MC, Walsh PJ. Diastereoselective Chelation-Controlled Additions to β -Silyloxy Aldehydes. Org Lett. 2012;14:3368–3371. doi: 10.1021/ol301354w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Benneche T, Strande P, Undheim K. A New Synthesis of Chloromethyl Benzyl Ethers. Synthesis. 1983:762–763. [Google Scholar]
  • 47.Frauenrath H, Brethauer D, Reim S, Maurer M, Raabe G. Highly Enantioselective Isomerization of 4,7-Dihydro-1,3-dioxepins Catalyzed by Me-DuPHOS-Modified Dihalogenonickel Complexes and Determination of the Absolute Configuration of the Isomerization Products. Angew Chem Int Ed. 2001;40:177–179. doi: 10.1002/1521-3773(20010105)40:1<177::AID-ANIE177>3.0.CO;2-X. [DOI] [PubMed] [Google Scholar]
  • 48.Gigg R, Warren CD. The Allyl Ether as a Protecting Group in Carbohydrate Chemistry. Part II. J Chem Soc C. 1968:1903–1911. doi: 10.1016/s0040-4039(00)90699-3. [DOI] [PubMed] [Google Scholar]
  • 49.Seixas RSGR, Silva AMS, Alkorta I, Elguero J. An experimental NMR and computational study of 4-quinolones and related compounds. Monatsh Chem. 2011;142:731–742. [Google Scholar]
  • 50.Bretschneider T, et al. Vinylogous chain branching catalysed by a dedicated polyketide synthase module. Nature. 2013;502:124–128. doi: 10.1038/nature12588. [DOI] [PubMed] [Google Scholar]
  • 51.Lim S-K, et al. Iso-Migrastatin, Migrastatin, and Dorrigocin Production in Streptomyces platensis NRRL 18993 is Governed by a Single Biosynthetic Machinery Featuring an Acyltransferase-less Type I Polyketide Synthase. J Biol Chem. 2009;284:29746–29756. doi: 10.1074/jbc.M109.046805. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Hall C, Wu M, Crane FL. Piericidin, a new inhibitor of mitochondrial electron transport. Biochem Biophys Res Commun. 1966;25:373–377. doi: 10.1016/0006-291x(66)90214-2. [DOI] [PubMed] [Google Scholar]
  • 53.Liu Q, et al. Elucidation of Piericidin A1 Biosynthetic Locus Revealed a Thioesterase-Dependent Mechanism of α-Pyridone Ring Formation. Chem Biol. 2012;19:243–253. doi: 10.1016/j.chembiol.2011.12.018. [DOI] [PubMed] [Google Scholar]
  • 54.Zhou X, Fenical W. The unique chemistry and biology of the piericidins. J Antibiot. 2016;69:582–593. doi: 10.1038/ja.2016.71. [DOI] [PubMed] [Google Scholar]
  • 55.Kunze B, Jansen R, Höfle G, Reichenbach H. Ajudazols, new inhibitors of the mitochondrial electron transport from Chondromyces crocatus. Production, antimicrobial activity and mechanism of action. J Antibiot. 2004;57:151–155. doi: 10.7164/antibiotics.57.151. [DOI] [PubMed] [Google Scholar]
  • 56.Lawrence JW, Darkin-Rattray S, Xie F, Neims AH, Rowe TC. 4-Quinolones Cause a Selective Loss of Mitochondrial DNA From Mouse L1210 Leukemia Cells. J Cell Biochem. 1993;51:165–174. doi: 10.1002/jcb.240510208. [DOI] [PubMed] [Google Scholar]
  • 57.Lawrence JW, Claire DC, Weissig V, Rowe TC. Delayed Cytotoxicity and Cleavage of Mitochondrial DNA in Ciprofloxacin-Treated Mammalian Cells. Mol Pharmacol. 1996;50:1178–1188. [PubMed] [Google Scholar]
  • 58.Esnault C, Roques BP, Jacquemin-Sablon A, Le Pecq JB. Effects of New Antitumor Bifunctional Intercalators Derived from 7H-Pyridocarbazole on Sensitive and Resistant L1210 Cells. Cancer Res. 1984;44:4355–4360. [PubMed] [Google Scholar]
  • 59.Tidd DM, Paterson ARP. A Biochemical Mechanism for the Delayed Cytotoxic Reaction of 6-Mercaptopurine. Cancer Res. 1974;34:738–746. [PubMed] [Google Scholar]
  • 60.Wotring LL, Roti Roti JL. Thioguanine-induced S and G2 Blocks and Their Significance to the Mechanism of Cytotoxicity. Cancer Res. 1980;40:1458–1462. [PubMed] [Google Scholar]
  • 61.Chen C-H, Cheng Y-C. Delayed Cytotoxicity and Selective Loss of Mitochondrial DNA in Cells Treated with the Anti-human Immunodeficiency Virus Compound 2’,3’-Dideoxycytidine. J Biol Chem. 1989;264:11934–11937. [PubMed] [Google Scholar]
  • 62.Wakimoto T, Egami Y, Abe I. Calyculin: Nature’s way of making the sponge-derived cytotoxin. Nat Prod Rep. 2016;33:751–760. doi: 10.1039/c5np00123d. [DOI] [PubMed] [Google Scholar]
  • 63.Kevbrin V, et al. Natronospirillum operosum gen. nov., sp. nov., a haloalkaliphilic satellite isolated from decaying biomass of a laboratory culture of cyanobacterium Geitlerinema sp. and proposal of Natronospirillaceae fam. nov., Saccharospirillaceae fam. nov. and Gynuellaceae fam. nov. Int J Syst Evol Microbiol. 2020;70:511–521. doi: 10.1099/ijsem.0.003781. [DOI] [PubMed] [Google Scholar]
  • 64.Mavrodi OV, et al. Rhizosphere Microbial Communities of Spartina alterniflora and Juncus roemerianus From Restored and Natural Tidal Marshes on Deer Island, Mississippi. Front Microbiol. 2018;9:1–13. doi: 10.3389/fmicb.2018.03049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65.Lucena T, et al. Marinomonas spartinae sp. nov., a novel species with plant-beneficial properties. Int J Syst Evol Microbiol. 2016;66:1686–1691. doi: 10.1099/ijsem.0.000929. [DOI] [PubMed] [Google Scholar]
  • 66.Fidalgo C, et al. Saccharospirillum correiae sp. nov., an endophytic bacterium isolated from the halophyte Halimioneportulacoides. Int J Syst Evol Microbiol. 2017;67:2026–2030. doi: 10.1099/ijsem.0.001914. [DOI] [PubMed] [Google Scholar]

Methods only References

  • 1.Helfrich EJN, et al. Automated structure prediction of trans-acyltransferase polyketide synthase products. Nat Chem Biol. 2019;15:813–821. doi: 10.1038/s41589-019-0313-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Helfrich EJN, Piel J. Biosynthesis of polyketides by trans-AT polyketide synthases. Nat Prod Rep. 2016;33:231–316. doi: 10.1039/c5np00125k. [DOI] [PubMed] [Google Scholar]
  • 3.Nguyen T, et al. Exploiting the mosaic structure of trans-acyltransferase polyketide synthases for natural product discovery and pathway dissection. Nat Biotechnol. 2008;26:225–233. doi: 10.1038/nbt1379. [DOI] [PubMed] [Google Scholar]
  • 4.Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32:1792–1797. doi: 10.1093/nar/gkh340. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Price MN, Dehal PS, Arkin AP. FastTree 2--Approximately Maximum-Likelihood Trees for Large Alignments. PLoS One. 2010;5:e9490. doi: 10.1371/journal.pone.0009490. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Greczmiel U, et al. Sustained T follicular helper cell response is essential for control of chronic viral infection. Sci Immunol. 2017;2:1–12. doi: 10.1126/sciimmunol.aam8686. [DOI] [PubMed] [Google Scholar]
  • 7.Battegay M, et al. Quantification of lymphocytic choriomeningitis virus with an immunological focus assay in 24- or 96-well plates. J Virol Methods. 1991;33:191–198. doi: 10.1016/0166-0934(91)90018-u. [DOI] [PubMed] [Google Scholar]
  • 8.Le Roux F, Binesse J, Saulnier D, Mazel D. Construction of a Vibrio splendidus Mutant Lacking the Metalloprotease Gene vsm by Use of a Novel Counterselectable Suicide Vector. Appl Environ Microbiol. 2007;73:777–784. doi: 10.1128/AEM.02147-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Sayers EW, et al. GenBank. Nucleic Acids Res. 2019;47:D94–D99. doi: 10.1093/nar/gky989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Reaxys. Elsevier; Amsterdam, NL; Oxford, UK; Cambridge MA, USA: [accessed: June 18. 2020]. https://www.reaxys.com/ [Google Scholar]
  • 11.PubChem. National Library of Medicine: National Center for Biotechnology Information U.S; Bethesda, MD, USA: [accessed: June 18. 2020]. https://pubchem.ncbi.nlm.nih.gov/ [Google Scholar]
  • 12.SciFinder-n. American Chemical Society; Columbus, OH, USA: [accessed: June 18. 2020]. scifindern.cas.org/ [Google Scholar]
  • 13.CHEMnetBASE Dictionary of Natural Products. CRC Press, Taylor&Francis Group; Abingdon, UK: [accessed: June 18. 2020]. http://dnp.chemnetbase.com/ [Google Scholar]
  • 14.van Santen JA, et al. The Natural Products Atlas: An Open Access Knowledge Base for Microbial Natural Products Discovery. ACS Cent Sci. 2019;5:1824–1833. doi: 10.1021/acscentsci.9b00806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.MarinLit: a database of the marine natural products literature. Royal Society of Chemistry: Department Chem. Univ. Canterbury; Canterbury, NZ: [accessed: June 18. 2020]. http://pubs.rsc.org/marinlit/ [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Inventory of supporting information
EMS146508-supplement-Inventory_of_supporting_information.docx (68.6KB, docx)
Supplementary data set 1
EMS146508-supplement-Supplementary_data_set_1.xlsx (2.5MB, xlsx)
Supplementary Information
EMS146508-supplement-Supplementary_Information.pdf (6MB, pdf)
Supplementary Video 1
Download video file (10.7MB, mp4)
Supplementary Video 2
Download video file (8.8MB, mp4)
Supplementary Video 3
Download video file (2.1MB, mp4)
Supplementary Video 4
Download video file (13.9MB, mp4)
Supplementary Video 5
Download video file (3MB, mp4)
Supplementary Video 6
Download video file (3.5MB, mp4)
Supplementary Video 7
Download video file (3.5MB, mp4)
Supplementary Video 8
Download video file (1.8MB, mp4)

Data Availability Statement

The data supporting the findings of this study are available in this article and the Supplementary Information. The genome sequence of G. sunshinyii YC6258 is accessible in GenBank under accession number NZ_CP007142.1, the janustatin gene cluster is located at locus YC6258_05439 (AJQ97469.1) to YC6258_05446 (AJQ97476.1). The closest homologs to these proteins can be found under the accession numbers: WP_158657926.1, WP_086931657.1, WP_038924948.1, RKZ46405.1, WP_087684108.1 and WP_086931660.1. Candidate proteins for the freestanding KR are: WP_044620326.1; WP_044617647.1; WP_044617493.1; WP_044616081.1; WP_044617507.1; WP_044618569.1; WP_044619377.1; WP_044617759.1; WP_144407613.1 and WP_052830250.1. The janustatin BGC was deposited in MIBiG under accession number BGC0002136.

RESOURCES