Isolation of antipodal (−)-versicolamide B and notoamides L-N from a marine-derived Aspergillus sp

Abstract

Antipodal (−)-versicolamide B and notoamides L-N were isolated from a marine-derived Aspergillus sp. The possible biosynthetic pathway of enantiomeric pairs of notoamide B and versicolamide B are proposed. Notoamide L is the first metabolite containing twenty-five carbons in the related prenylated indole alkaloids. Notoamide M is potentially a precursor to the proposed azadiene species involved in the putative intramolecular Diels-Alder reaction in the biogenesis of the bicyclo[2.2.2]diazaoctane ring system.

A number of prenylated indole alkaloids containing a diketopiperazine or a bicyclo[2.2.2]diazaoctane ring were isolated from various fungi of the genera Aspergillus and Penicillium, and the study of the biosynthetic pathways to these alkaloids has recently become an area of significant interest.¹ We isolated new alkaloids, notoamides A-D (1–4),² (Figure 1) from a marine-derived Aspergillus sp. as well as the known natural product stephacidin A (5).³ Interestingly, we found that notoamide E (6) is a short-lived natural metabolite and biosynthetic precursor to notoamides C (3) and D (4).⁴ In addition, we recently isolated notoamides F-K. ⁵ Further, Williams and Gloer et al. reported the isolation of antipodal (+)-notoamide B (7) and (−)-stephacidin A (8) (Figure 2) from the terrestrial organism Aspergillus versicolor NRRL 35600 along with a new alkaloid, (+)-versicolamide B (9), which possesses a novel anti-relationship between C-21-C-22 and C-17-N-13 in the bicyclo[2.2.2]diazaoctane ring system.⁶ With regard to the generation of the antipodal stereoisomers of notoamide B and stephacidin A, they proposed that the intramolecular Diels-Alder (IMDA) reaction occurs in a face-selective manner in the marine-derived Aspergillus sp. and the terrestrial Aspergillus versicolor.⁶ Based on this presumption, the presence of (−)-versicolamide B (10) was predicted to be a metabolite in our marine-derived Aspergillus sp. (Figure 2). In the continuing search for notoamide congeners, we finally succeeded in isolating 10 along with new notoamides L-N (11–13) (Figure 1). Here we report their structures and suggest another possible biosynthetic pathway that accommodates the biogenesis of the respective antipodes of notoamide B and versicolamide B.

Figure 1.

Structures of compounds from the marine-derived Aspergillus sp.

Figure 2.

Structures of three pairs of enantiomeric alkaloids isolated from the marine-derived Aspergillus sp. and the terrestrial Aspergillus versicolor.

The fractions rich in indole alkaloids, which derived from the EtOH extract of the fungal culture,² was further purified by HPLC to afford (−)-versicolamide B (10) and notoamides L-N (11–13).

The HRFABMS of 10 showed the molecular formula C₂₆H₂₉N₃O₄. The ¹H NMR spectrum of 10 (Table S1) was superimposable on that of (+)-versicolamide B (9) and revealed that the structure and relative stereochemistry was the same as that of 9. The specific rotation of 10 (−22°) was of the opposite sign to that of 9 (+26°), and the CD spectrum of 10 (Figure S6) was also opposite to that of 9. These data conclusively reveal that 10 is the antipodal (−)-versicolamide B, and that 9 and 10 are the third pair of antipodal natural metabolites isolated from the closely-related Aspergillus fungi.

The molecular formula of notoamide L (11), C₂₅H₂₉N₃O₄, was established by high-resolution FABMS, and the ¹H NMR spectrum (Table S2) showed three singlet methyl signals at δ 1.39, 1.40, and 1.64. In contrast to other prenylated indole alkaloids, one singlet methyl group is missing in 11, and the methyl signal at δ 1.64 is observed at lower field. The ¹³C NMR spectrum of 11 (Table S2) showed a ketone carbon at δ 199.1. Usually in the family of prenylated indole alkaloids, an isoprenyl unit at C-2 or C-3 cyclizes across the diketopiperazine moiety as shown in stephacidin A (5) or notoamide A (1) or exists as a substituted group at C-2 or C-3 as shown in notoamide D (4) or C (3). For compound 11, the HMBC spectrum showed the presence of an isopropenyl group at C-21 instead of an isoprenyl group; δ_H 1.64 (3H, s, H₃-23), 4.79 (1H, d, J = 1.0 Hz, H-24) and 4.83 (1H, d, J = 1.0 Hz, H-24)/δ_C 19.2 (CH₃, C-23), 116.2 (CH₂, C-24) and 144.7 (C, C-22). In addition, the HMBC correlations, δ 7.70 (H-4)/δ 199.1, δ 3.24 and 3.44 (H₂-10)/δ 199.1, showed that the ketone carbon existed between C-9 and C-10, being assignable to C-3. Thus, the gross structure of 11 was established. The 11S,17S configuration of 11 was indicated by the CD spectrum (Figure S12),⁷ and the co-occurrence of notoamides A and B (1 and 2) and (+)-stephacidin A (5) as major alkaloids indicated the 21R configuration for 11 on the basis of biogenetic considerations. In the structure of 11, a C-2 carbon derived from tryptophan is lacking. Although now more than forty members of related prenylated indole alkaloids have been reported, a metabolite containing twenty-five carbons such as 11 has not been isolated yet in this family. One possible biogenesis of 11 derived from (+)-stephacidin A (5) (see Supporting Information) is shown in Scheme S1.

Notoamide M (12) has the molecular formula C₂₆H₃₁N₃O₅, one oxygen unit more than notoamide C (3) or 3-epi-notoamide C (14).^{4, 8} The ¹H and ¹³C NMR data in CD₃OD of 12 (Table S3) revealed that a methine at C-17 in 3 or 14 was replaced with a quaternary carbon in 12, which was indicated by HMBC correlations, δ 1.68 (H-15) and 1.95 (H-16)/δ 87.6 (C-17). In the ¹³C NMR chemical shifts of three compounds,⁹ differences were observed in signals for C-3 (δ 57.1 in 12, δ 62.6 in 3 and δ 58.7 in 14) and C-11 (δ 55.3 in 12, δ 67.1 in 3 and δ 53.4 in 14), and other signals except for C-15, C-16, and C-17 match well with each other. The CD spectra of 12, 3, and 14 in MeOH (Figures S18 and S19) showed that the spectra of 3 and 14 were almost symmetrical and the spectrum of 12 was similar to that of 3. These data suggest the 3R configuration of 12. The 11S configuration is based on biogenetic considerations, but the configuration at C-17 remains to be determined.

The molecular formula of notoamide N (13) was determined as C₂₆H₂₈ClN₃O₄, indicating a replacement of a hydrogen atom with chlorine. The ¹H NMR spectrum (Table S4) showed the absence of a pair of doublet aromatic signals as shown in other notoamides and the presence of a singlet aromatic signal at δ 7.19. The ¹³C NMR spectrum of 13 was similar to that of notoamide B (2) except for C-5, δ 113.3 (C) in 13 and δ 109.6 (CH) in 2. Analysis of HMBC correlations established the structure of 13 as 5-chloro-notoamide B. The CD spectrum of 13 (Scheme S25) was similar to that of 2.² Therefore, we have assigned the 3R,11S,17S,21S configuration to 13.^{7, 10} Finally, the synthesis of 13 from synthetic notoamide B by t-butyl hypochlorite oxidation enabled the structure of 13 to be unambiguously assigned (see Supporting Information). Notoamide N (13) is the third chlorinated derivative to have been discovered in this family of prenylated indole alkaloids following the recent disclosure of malbrancheamide and malbrancheamide B. ¹¹

It is now established that a series of three pairs of enantiomeric prenylated indole alkaloids, (+)- and (−)-notoamide B (7 and 2), (+)- and (−)-stephacidin A (5 and 8), and (+)- and (−)-versicolamide B (9 and 10), are produced by distinct species of the genus Aspergillus. It is noteworthy that we have confirmed that all three pairs of these natural metabolites were isolated from their respective producing organisms in optically pure form.¹²

A biosynthetic pathway to rationalize the formation of the individual enantiomers of stephacidin A and notoamide B has been described that postulates that the respective fungi could plausibly exhibit opposite facial selectivity in a key, putative intramolecular Diels-Alder (IMDA) reaction via an achiral azadiene intermediate derived by the two-electron oxidation of notoamide E (6).⁶ In order to interrogate this proposed pathway, we carried out a feeding experiment of doubly ¹³C-labelled notoamide E in the marine-derived Aspergillus sp.⁴ The results clearly showed that 6 is a precursor for notoamides C (3) and D (4), but not for metabolites containing a bicyclo[2.2.2]diazaoctane ring, including notoamide B (2) or stephacidin A (5), although they are the major metabolites produced from this fungus.

Here, we wish to advance an alternative pathway for the potential unified biogenesis of the antipodal stereoisomers of notoamide B (2 and 7) and versicolamide B (9 and 10) (Scheme 1). In this biogenetic hypothesis, the common precursor is deoxybrevianamide E (15), which we have isolated from the marine-derived Aspergillus fungus.² An R-selective indole oxidase in the marine-derived Aspergillus and a corresponding S-selective oxidase in the terrestrial Aspergillus followed by pinacol rearrangement of the isoprenyl group from C-2 to C-3 affords 16 and the corresponding 3-epi diastereomer (17), respectively. Oxidation and tautomerization of 16 would yield the optically pure azadiene intermediates 18. In principle, 18 could afford four stereoisomers, 19 plus 20 and their corresponding C-21 epimers, although these epimers have not yet been detected. The preponderance of 19 over 20 might be due to the presence or absence of hydrogen-bonding, which could be observed in intermediate 26 or 27, respectively, derived from 16 (Scheme S2). In the biogenetic pathway postulated in Scheme 1, the enantio-divergence arises as a consequence of incipient R- or S-selective indole oxidase instead of opposite facial selectivities in the putative IMDA reaction; oxidation and prenylation reactions (19→2 or 20→10) to construct the pyran ring attached to the indole moiety would thus follow formation of the bicyclo[2.2.2]diazaoctane ring system. The culture of the marine-derived Aspergillus yields notoamide C (3) as a major metabolite, and 3-epi-notoamide C (14) was not detected at all (but was isolated when ¹³C-labeled notoamide E was fed to this organism).⁸ We speculate that the terrestrial Aspergillus versicolor might be expected to contain 14 as a major metabolite instead of 3. Since the biogenesis of (+)-and (−)-stephacidin A (5 and 8) is not accommodated by this pathway, further putative precursor feeding and incorporation experiments are necessary to establish the biogenesis of stephacidin A. It also remains plausible, that stephacidin A is the direct biosynthetic precursor to notoamide B by via indole oxidation and pinacol rearrangement; the laboratory transformation of which that we have experimentally established.^8b With regard to the formation of a bicyclo[2.2.2]diazaoctane ring, the isolation of notoamide M (12) is noteworthy. In order to form an azadiene intermediate (D in Scheme 2) oxidation at C-17 of a diketopiperazine ring (A→B) followed by dehydration would be necessary. Therefore, the precursor of 12 could be notoamide C (3) and the isolation of 12 provides indirect and provocative support for an oxidative manifold leading to the generation of the putative azadiene species. Alternatively, 12 could be an artifact of hydrolytic capture of the transient azadiene species (C).

Scheme 1.

Possible biogenesis of the enantiomeric pairs of notoamide B (2, 7) and versicolamide B (9, 10).

Scheme 2.

Possible formation of the bicyclo[2.2.2]diazaoctane ring system.

In conclusion, we have isolated antipodal (−)-versicolamide B (10) along with three new notoamides L-N (11–13) from a marine-derived Aspergillus sp. Namely, (+)-and (−)-versicolamide B (9 and 10) are the third enantiomeric pair, in addition to (+)- and (−)-notoamide B (7 and 2) and (+)- and (−)-stephacidin A (5 and 8), which were isolated from distinct species of Aspergillus. With regard to the generation of the enantiomeric pairs of notoamide B and versicolamide B, here we propose an alternative biogenetic pathway (Scheme 1), in which a key R- or S-selective indole oxidase would exist in the respective marine-derived or in the terrestrial Aspergillus. The identification and characterization of such enzymes in each Aspergillus are currently being pursued to shed light on this possibility. Notoamide L (11) is the first alkaloid constituted of twenty-five carbons in this family of prenylated indole alkaloids, and the biogenesis of 11 comprises an as yet undefined but alluring pathway to elucidate. Notoamide M (12) contains a hydroxy group at C-17, and its isolation provides indirect support for a potential mechanism that culminates in the construction of the unique bicyclo[2.2.2]diazaoctane ring system (Scheme 2). Notoamide N (13) is a rare chlorinated member of this family of prenylated indole alkaloids. The existence of a halogenase system in our marine-derived Aspergillus sp. is an important finding, which we are pursuing in the context of identifying the class of halogenase extant. Our laboratories are continuing studies to elucidate the origin of the fascinating enantio-divergence manifest in the natural production of the stephacidins, notoamides and versicolamides.