Letter pubs.acs.org/OrgLett
A Reactive Eremophilane and Its Antibacterial 2(1H)‑Naphthalenone Rearrangement Product, Witnesses of a Microbial Chemical Warfare Séverine Amand,† Marine Vallet,† Laura Guedon,† Grégory Genta-Jouve,‡ Frank Wien,§ Stéphane Mann,† Joel̈ le Dupont,∥ Soizic Prado,*,† and Bastien Nay*,⊥,† †
Muséum National d’Histoire Naturelle, CNRS (UMR 7245 MCAM), Sorbonne Universités, 57 rue Cuvier (CP 54), 75005 Paris, France ‡ Université Paris Descartes, CNRS (UMR 8638 C-TAC), Faculté de Pharmacie de Paris, Sorbonne Paris Cité, 4 Avenue de l’Observatoire, 75006 Paris, France § Synchrotron SOLEIL, DISCO line, Gif-sur-Yvette, France ∥ Muséum National d’Histoire Naturelle, CNRS (UMR 7205 ISYEB), Sorbonne Universités, 57 rue Cuvier (CP 39), 75005 Paris, France S Supporting Information *
ABSTRACT: Two sesquiterpenes, 4-epi-microsphaeropsisin (1) and a dihydrofurano-2(1H)-naphthalenone (variabilone, 2) which represents a new skeleton, were isolated from endophytic fungus Paraconiothyrium variabile. Reactivity studies showed that eremophilane 1 is a precursor of 2 through acid-promoted methyl 1,2-migration and aromatization. An electrophilic intermediate of this transformation was intercepted by N-acetylcysteamine, a biomimetic nucleophile. Only compound 2 was antibacterial against endophytic bacterium Bacillus subtilis (coisolated with P. variabile), suggesting a role in the microbial competition in plants.
A
yew tree Cephalotaxus harringtonia.4 This last compound is representative of a new skeleton in the sesquiterpene series. Compound 1 is a reactive eremophilane and the precursor of 2 through acid-catalyzed methyl 1,2-migration. The peculiar reactivity of 1, a biologically inactive compound, raises the question of its role since it gives, once released in the fungal ecosystem, the antibacterial compound 2. The chemical reactivity and biological properties of these sesquiterpenes were investigated in the natural history context of the plant/ fungus/bacterium relationship,5 providing clues for their ecological significance.6 Compound 1 was isolated as a white solid and optically active compound ([α]20 D = −8.75 in CHCl3, c 0.8) from the ethyl acetate extract of P. variabile mycelium grown on Biomalt medium. HR-ESI-TOF mass spectrometry showed a protonated ion [M + H]+ at m/z 279.1591 compatible with the molecular formula C16H23O4 and indicating 6 degrees of unsaturation. 1H NMR spectra in DMSO-d6 (Table 1) showed three olefinic protons, two of them as doublets at δH 5.93 (H-2, d, J = 9.9 Hz) and δH 7.11 ppm (H-1, d, J = 9.9 Hz) and one as a singlet δH 6.30 (H-9). COSY correlations between H-1 and H-2 allowed defining partial structure I corresponding to a Zolefin, possibly a conjugated enone with regard to chemical shifts (Figure 2a). COSY correlations between a CH3 at δH 0.94 ppm (H-15, d, J = 6.8 Hz) and H-4 at δH 2.34 ppm (q, J = 6.8 Hz) defined second aliphatic partial structure II. Finally, last
particular trait of defense strategies in Nature involves chemical entities with latent toxicity.1 They are safely secreted by the producer and then chemically activated under environmental conditions to release the active product. This phenomenon reveals the stiff chemical warfare occurring between living organisms. A classical example from the plant kingdom is found with the phytoalexin juglone (5-hydroxynaphthoquinone), a plant growth inhibitor accumulating underneath black walnut trees, which is formed ex planta by the hydrolysis and oxidation of hydrojuglone glucoside.1a,b In microbes, this strategy has been far less documented although micro-organisms are important producers of antibiotics and toxins.2 Herein we report the isolation and properties of two new sesquiterpenes, 4-epi-microsphaeropsisin (1, Figure 1) and the dihydrofurano-2(1H)-naphthalenone 2, from P. variabile isolated in a previous investigation3 from the medicinal plum
Figure 1. Structure of 4-epi-microsphaeropsisin 1, dihydrofurano2(1H)-naphthalenone 2, and atom numbering (following that commonly used for eremophilanes). © 2017 American Chemical Society
Received: June 13, 2017 Published: July 24, 2017 4038
DOI: 10.1021/acs.orglett.7b01788 Org. Lett. 2017, 19, 4038−4041
Letter
Organic Letters
Table 1. 1H (600 MHz) and 13C NMR (150 MHz) Data for Compounds 1 (DMSO-d6) and 2 (CD3OD) (δ in ppm and J in Hz), Including HMBC and NOESY Correlations for 1 and HMBC for 2 1 δH
no. 1 2 3 4 5 6α 6β 7 8 9 10 11 12α 12β 13 14 15 OCH3 OH a
7.11 5.93 − 2.34 − 1.56 1.93 − − 6.30 − 1.98 4.03 3.35 0.95 1.07 0.94 3.30 4.03
(d, 1H, 9.9) (d, 1H, 9.9) (q, 1H, 6.8) (dd, 1H, 14.0, 1.6 Hz) (d, 1H, 14.0)
(s, 1H) (m, 1H) (dd, 1H, 8.6, 6.7)a (dd, 1H, 8.6, 6.0) (d, 3H, 7.3) (s, 3H) (d, 3H, 6.8) (s, 3H) (s, 1H)a
2
δC 144.2 127.3 200.4 52.1 39.4 46.0 − 77.4 100.9 125.9 142.3 42.1 71.8 − 14.8 19.9 7.1 47.8 −
HMBC 3, 4, − 3, − 4, 5, − − 1, − 6, 7, 7, 7, 4, 3, 8 6,
NOESY
5, 9, 10 10 5, 6, 10, 14, 15 5, 7, 11, 14 7, 8, 10, 11, 14
5, 7, 8, 10 7, 8, 12, 13 8, 11, 13 11, 13 11, 12 5, 6, 10 4, 5 7, 8, 11
9 − − 6α, 15 − 4, 6β, 11 6α, 13, 14, 15 − − 1, OCH3 − 6α, 12α, 13 11, 13 13 11, 12 6β, 15, OH 4, 6β, 14 9, 13, OH 14, OCH3
7.54 6.07 − − − 7.39 − − − 6.77 − 3.58 4.70 4.09 1.34 1.42 1.42 − −
δH
δC
(d, 1H, 9.8) (d, 1H, 9.8)
147.7 124.2 207.6 48.6 141.7 123.2 − 137.2 160.0 110.7 129.8 37.9 79.9 − 19.5 28.6b 28.5b − −
(s, 1H)
(s, 1H) (m, 1H) (t, 1H, 8.6) (dd,1H, 8.6, 7.2) (d, 3H, 6.9) (s, 3H) (s, 3H)
HMBC 3, 4, − − − 4, − − − 1, − 7, 7, 7, 7, 3, 3, − −
5, 9 10
8, 10, 11
5, 7, 8 13 8, 13 8, 11, 13 11, 12 4, 5, 15 4, 5, 14
b
Overlapped signals. Interchangeable signals.
The relative stereochemistry of 1 was determined by NOESY experiments (Figure 2b). Strong correlations between protons H-14 and H-15/7-OH/H-6β as well as between H-13 and 8OCH3, comparatively to H-11 and H-12α or H-6α, indicated that they are on the same face. Accordingly, this compound is the 4-epimer of microsphaeropsisin which had previously been isolated from the marine sponge-associated fungus Microsphaeropsis sp.7 4-epi-Microsphaeropsisin (1) was found to be labile, which precluded any further analysis as a pure compound. Its absolute configuration was deduced from that of compound 2, as determined by circular dichroism (see below). Compound 2 was isolated as a colorless resin with an [α]20 D of +13.30 (CHCl3, c 0.15), a maximum UV absorption at 279 nm (log ε = 3.52 in MeOH) and an IR absorption band at 1653 cm−1 (CO) later found characteristic of the 1,1-dimethyl2(1H)-naphthalenone core.8 HR-ESI-TOF mass spectrometry showed a protonated ion [M + H]+ at m/z 229.1219 corresponding to the formula of C15H17O2, indicating 8 degrees of unsaturation. The 1H and 13C NMR spectra of 2 differed from those of 1 by the presence of an aromatic proton (δH 7.39 ppm, H-6, s) instead of an sp3 methylene group and by the presence of a quaternary carbon (C-4, δC 48.6 ppm) instead of a methine group. Additionally, carbons C-5 (δC 141.7), C-7 (δC 137.2), and C-8 (δC 160.0) on compound 2 were significantly shifted downfield by comparison with 1. These data showed the presence of an aromatic ring system fused to the 5-membered oxacycle, instead of the central cyclohexene moiety, and that of a gem-dimethyl on C-4. DEPT-Q spectra and the observed HSQC correlations allowed identifying hydrogenated carbons and their respective protons (Table 1). COSY correlations defined two partial structures I and II corresponding to the cyclohexenone and the methyldihydrofuran rings, respectively (Figure 3). HMBC correlations (Figure 3) confirmed these hypotheses, particularly those from H-9 to C-5, C-7, and C-8, H-6 to C-8/C-10, and those between C-8 and H-9/H-12. Additional HMBC correlations were observed between protons
Figure 2. NMR correlations observed for compound 1: (a) partial structures deduced from 1H−1H COSY correlations (bold bonds) and key 1H−13C HMBC correlations (red arrows); (b) main 1H−1H NOESY correlations (black arrows).
partial structure III was assigned by the COSY correlations between the methyl protons H-13 (δH 0.95 ppm) and the methylene protons H-11 (δH 1.98 ppm), and correlations between H-11 and the oxygenated methylene protons H-12 at δH 4.03 (dd, J = 8.6 and 8.2 Hz) and 3.35 (dd, J = 8.6 and 8.2 Hz). The DEPT-Q carbon NMR spectrum (DMSO-d6) of 1 exhibited the 16 carbon atoms of the molecular formula, including five sp2 carbons distributed into one ketone at δC 200.3 ppm, three monosubstituted carbon−carbon double bonds (CH) and a quaternary one (>C). The sp3 carbons were distributed into four methyls, two methylenes including an oxygenated one, two methines, and three quaternary carbons including a monooxygenated one (δC 77.4 ppm) and a dioxygenated one (δC 100.9 ppm). Protonated carbons and their respective protons were attributed by the HSQC NMR experiment (Table 1). Connecting substructures I−III was deduced from the cross-peaks observed on the HMBC spectrum (Figure 2a). Especially, correlations of the ketone (δC 200.3 ppm) with H-1 and H-4 allowed connecting fragments I and II while HMBC correlations from H-4 to C5, C-6, and CH3-14 connected fragment II to these carbons. Finally, the 5-membered oxacycle was deduced from the HMBC correlations of acetalic carbon C-8 (δC 100.9 ppm) with olefinic proton H-9 (δH 6.30 ppm), the methoxy group (δH 3.30) ppm, and methylene protons H-12. 4039
DOI: 10.1021/acs.orglett.7b01788 Org. Lett. 2017, 19, 4038−4041
Letter
Organic Letters
Scheme 1. Biosynthetic Relationship between 1 and 2 and Proposed Mechanism for the Rearrangement of 1
Figure 3. Correlations observed for compound 2: partial structure deduced from 1H−1H COSY correlations (bold bonds) and 1H−13C HMBC correlations (red arrows).
H14 or H-15 and carbon C-4. All these data were consistent with the dihydrofurano-2(1H)-naphthalenone structure of compound 2. In addition, comparison of experimental synchrotron radiation circular dichroism (SRCD) spectra with those calculated by the time-dependent density functional theory (TDDFT) circular dichroism method, using the B3LYP at the 6-311+g(d,p) level for 15 excited states, allowed determining the absolute stereochemistry of C-11 as R (Figure 4). Three
analyzed under neutral conditions, showing that it cannot be an isolation artifact. It is striking that mild acidic conditions are compatible with not only fungal culture conditions but also soil physicochemical properties (pH 4.0−8.3)10 and some plant cellular compartments in vivo (pH 4.5−6).11 The rearrangement of 1 could thus occur once the natural product is released in the natural environment of the fungus where it could exert biological effects. Obviously, any of the structures 1−5 depicted in Scheme 1 could have strong electrophilic properties owing to the presence of enone and oxonium functional groups which are susceptible to react with biological nucleophiles. To study the mechanism of this rearrangement and with the aim of intercepting one of these hypothetic intermediates (3−5) in a biomimetic context, the reaction was performed in a weak acidic medium (pH 5) in the presence of the nucleophile Nacetylcysteamine (6). Interestingly, the addition of the thiol group was only observed at C-2, i.e. in position α of the starting enone which is usually not electrophilic, affording compound 7 (Scheme 2, Table S1). This umpolung can be explained by the
Figure 4. SRCD spectrum (blue line) of compound 2 recorded at 25 °C in methanol (c = 10 mg/mL) and TDDFT (B3LYP) calculated SRCD spectrum for the 2R configuration (red line).
Cotton effects were experimentally observed by SRCDweak negative at 341 nm (n → π* transition), strong positive at 278 nm (π → π*) overlapped with a positive effect at 240 nm, and negative at 205 nm (π → σ*)9also clearly visible on the calculated spectrum for the two higher wavelengths. Compound 2, for which we suggest the name of variabilone, is the first member of a new sesquiterpene skeleton (variabilane, by analogy) characterized by the presence of a gem-dimethylated carbone at position 4. Although the carbon skeletons of 1 and 2 are different, their structural similarity suggests a common biosynthetic origin. In fact, 2 could result from acid-promoted 1,2-shift of the methyl at position 5 onto position 4 (Wagner−Meerwein rearrangement) through highly conjugated and electrophilic oxonium species 3 which may result from the elimination of the methoxy group of 1, followed by enolization into 4 and mesomery with carbocation 4′ (Scheme 1). Methyl migration would release carbocation 5 which would be quenched through dienol isomerization. Final aromatization of the central ring would deliver 2 through a highly favored dehydration. This hypothesis was confirmed by the transformation of 1 into 2 under various pH conditions. The rearrangement of 1 was observed under mild acidic conditions, with a rate increasing at lower pH. The reaction was complete within 2 h at pH = 3 or 48 h at pH = 5 (see Supporting Information), while at neutral pH = 7, compound 2 was only observed in minute amounts after 48 h. The natural character of rearrangement product 2 was confirmed by its presence in crude extracts obtained and
Scheme 2. Reaction of 1 in the Presence of a Thiol Nucleophile, Involving Cationic Intermediate 3
existence of an electrophilic intermediate like the oxonium species 3. Furthermore, the conservation of the relative stereochemistry at C-4 during this reaction is also indicative that intermediate 3 is persistent, ketone enolization being ratelimiting for the rearrangement 1 → 2. To conclude this experiment, the observation of sulfide adduct 7 (yet still labile) combined with the absence of rearranged product 2 under these conditions demonstrates that a reactive intermediate like 3 may have a life span compatible with the alkylation of biological nucleophiles, in vivo. Eremophilanes have been described as host selective phytotoxins, 6c,12 entomotoxins, 13 or phytohormones. 14 Although fungal eremophilanes have been less studied than plant ones, many were described with a high degree of 4040
DOI: 10.1021/acs.orglett.7b01788 Org. Lett. 2017, 19, 4038−4041
Organic Letters
■
unsaturations on their decalin moiety while one of them, periconianone A,15 displayed an extra ring within a complex 6/ 6/6 carbocyclic architecture. Fungal eremophilanes may be regarded as defensive compounds against microbial competitors. Owing to its reactivity, 4-epi-microsphaeropsisin (1) turns out to behave as a molecular trigger under weakly acidic environments, leading to profound structure reorganization. In this context, the search for biologically and ecologically relevant activities led to the discovery that variabilone 2 is significantly antibacterial against an endophytic Bacillus subtilis strain previously isolated from the same host tree as P. variabile.3 While compound 1 did not show any significant antibacterial activity during these essays (IC50 > 100 μg/mL), compound 2 behaved as a potent growth inhibitor of B. subtilis at an IC50 of 2.13 μg/mL after 24 h (0.36 μg/mL for kanamycin). In conclusion, we show that the endophytic fungus P. variable is able to secrete an inactive eremophilane (1) whose acid-catalyzed rearrangement leads to antibacterial compound 2 called variabilone, inhibiting the growth of the endophytic bacterial competitor B. subtilis. Variabilone (2) holds a new natural product skeleton and appears to be an end-product of the eremophilane oxidative metabolism. 4-epi-Microsphaeropsisin (1) was shown to be reactive in an acid-dependent manner compatible with in vivo physicochemical conditions. We demonstrated that its rearrangement may involve a reactive oxonium intermediate (3) which was intercepted by a biomimetic nucleophile, showing its potential capability to interact with biological nucleophiles. More generally, this study brings mechanistic insight into the role of small chemical entities in the plant-fungus mutualism6,16 and in microbial competitions.17
■
REFERENCES
(1) (a) Daglish, C. Biochem. J. 1950, 47, 458−462. (b) Duroux, L.; Delmotte, F. M.; Lancelin, J.-M.; Kéravis, G.; Jay-Lallemand, C. Biochem. J. 1998, 333, 275−283. (c) Saunders, M.; Kohn, L. M. Appl. Environ. Microbiol. 2008, 74, 136−142. (d) Klessig, D. F.; Malamy, J. Plant Mol. Biol. 1994, 26, 1439−1458. (2) Fischbach, M. A. Curr. Opin. Microbiol. 2009, 12, 520−527. (3) Langenfeld, A.; Prado, S.; Nay, B.; Porcher, E.; Cruaud, C.; Lacoste, S.; Bury, E.; Hachette, F.; Hosoya, T.; Dupont, J. Fungal Biol. 2013, 117, 124−136. (4) C. harringtonia has been the focus of numerous phytochemical studies, especially from our group: (a) Abdelkafi, H.; Nay, B. Nat. Prod. Rep. 2012, 29, 845−869. (b) Evanno, L.; Jossang, A.; NguyenPouplin, J. Planta Med. 2008, 74, 870−872. (5) The antagonism between P. variabile and B. subtilis has been deeply studied in our group; see also: Vallet, M.; Vanbellingen, Q. P.; Fu, T.; Le Caer, J.-P.; Della-Negra, S.; Touboul, D.; Duncan, K. R.; Nay, B.; Brunelle, A.; Prado, S. J. Nat. Prod., paper under revision (np2016-011858.R1). (6) (a) Prado, S., Li, Y., Nay, B. Diversity and ecological significance of fungal endophyte natural products. In Studies in Natural Product Chemistry, Vol. 36; Atta-Uhr-Raman, Ed.; Elsevier Science: Amsterdam, 2012; pp 249−296. For other fungal sesquiterpenes with ecological significance, see: (b) Amand, S.; Langenfeld, A.; Blond, A.; Dupont, J.; Nay, B.; Prado, S. J. Nat. Prod. 2012, 75, 798−801. (c) Del Valle, P.; Figueroa, M.; Mata, R. J. Nat. Prod. 2015, 78, 339−342. (7) Höller, U.; König, G. M.; Wright, A. D. J. Nat. Prod. 1999, 62, 114−118. Comparison of 1H NMR data of 1 in CDCl3 with those of microsphaeropsisin showed significant differences (not shown). However, marked instability of 1 in CDCl3 precluded complete analysis in this NMR solvent. (8) Wenkert, E.; Youssefyeh, R. D.; Lewis, R. G. J. Am. Chem. Soc. 1960, 82, 4675−4680. (9) Crabbé, P. ORD and CD in Chemistry and Biochemistry; Academic Press: New York, 1972. (b) Berova, N.; Nakanishi, K.; Woody, R. W. Circular Dichroism − Principles and Applications, 2nd ed.; WileyBlackwell: 2000. (10) Rousk, J.; Båat̊ h, E.; Brookes, P. C.; Lauber, C. L.; Lozupone, C.; Caporaso, J. G.; Knight, R.; Fierer, N. ISME J. 2010, 4, 1340−1351. (11) Gout, E.; Bligny, R.; Douce, R. J. Biol. Chem. 1992, 267, 13903− 13909. (12) Sugawara, F.; Strobel, G.; Fisher, L. E.; van Duyne, G. D.; Clardy, J. Proc. Natl. Acad. Sci. U. S. A. 1985, 82, 8291−8294. (13) García, M.; Sosa, M. E.; Donadel, O. J.; Giordano, O. S.; Tonn, C. E. J. Chem. Ecol. 2003, 29, 175−187. (14) Sugawara, F.; Hallock, Y. F.; Bunkers, G. D.; Kenfield, D. S.; Strobel, G.; Yoshida, S. Biosci., Biotechnol., Biochem. 1993, 57, 236− 239. (15) Zhang, D.; Ge, H.; Zou, J.-h.; Tao, X.; Chen, R.; Dai, J. Org. Lett. 2014, 16, 1410−1413. (16) Tian, Y.; Amand, S.; Buisson, D.; Kunz, C.; Hachette, F.; Dupont, J.; Nay, B.; Prado, S. Phytochemistry 2014, 108, 95−101. (17) We showed in a previous study that P. variabile also competes with a fungal phytopathogen: Combès, A.; Ndoye, I.; Bance, C.; Bruzaud, J.; Djedjat, C.; Dupont, J.; Nay, B.; Prado, S. PLoS One 2012, 7, e47313.
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.7b01788. Experimental details and copies of 1D and 2D NMR spectra (PDF)
■
Letter
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected] (B.N.). *E-mail:
[email protected] (S.P.). ORCID
Soizic Prado: 0000-0002-8071-9642 Bastien Nay: 0000-0002-1209-1830 Present Address ⊥
Ecole Polytechnique, Lab. de Synthèse Organique (UMR 7652 CNRS), Route de Saclay, 91128 Palaiseau, France. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank the MNHN (especially the ATM “Microorganismes”), the French Ministry of Research and CNRS for funding. NMR was funded by Région Ile-de-France, MNHN and CNRS. SRCD measurements on DISCO beamline at SOLEIL Synchrotron were collected under Proposal No. 20151217. 4041
DOI: 10.1021/acs.orglett.7b01788 Org. Lett. 2017, 19, 4038−4041
