Ophiosphaerellins A-I, Polyketide-Derived Compounds from the Endolichenic Fungus Ophiosphaerella korrae.

Ophiosphaerellins A-I (1-9), the first example of bicyclo[4.1.0]heptenones, as well as their biosynthetic relatives ophiosphaerekorrins A-B (10-11) were isolated from the endolichenic fungus Ophiosphaerella korrae. Biosynthetically, they were derived from the polyketide pathway, and their absolute configurations were determined on the basis of the combination analysis of spectral data, circular dichroism calculations, and single-crystal X-ray diffraction measurement. Preliminary test with thin-layer chromatography bioautography found that this type of compounds showed moderate acetylcholinesterase (AChE) inhibitory effects.

Introduction

Fungal endophytes are microorganisms that reside in internal tissues of living plants without causing any immediate overt negative effects, but may turn pathogenic during host senescence.[1] Endolichenic fungi, inhabiting the thalli of lichens in a similarly asymptomatic manner, show the metabolite diversity and potential to produce pharmaceutically valuable natural products by a series of continuing reports.[2−5] During our ongoing efforts to discover bioactive natural metabolites from endolichenic fungi,[6−10] a chemical investigation on the fungus Ophiosphaerella korrae, inhabiting the lichen Physciaceae physcia collected from Ditch Muzart of Xinjiang Province in China, was carried out. Ophiosphaerellins A–I (1–9), with novel bicyclo[4.1.0]heptenones, together with two new biosynthetically oxaspiro relatives, ophiosphaerekorrins A–B (10–11), were obtained (Figure ). These two types of unprecedented scaffoldings are the first examples of natural metabolites derived from the polyketide pathway. Herein, their isolation, structure elucidation, a proposed biogenesis route, and the preliminary biotest results are presented.

Figure 1

Structures of compounds 1–11 from O. korrae.

Results and Discussion

Ophiosphaerellin A (1) was obtained as white needles from MeOH. The molecular formula of 1 was determined as C14H18O3 with five degrees of unsaturation by high-resolution electrospray ionization mass spectrometry (HRESIMS) and 13C NMR data. The IR spectrum suggested the presence of carbonyl, hydroxy, and olefinic functionalities, respectively, at 1690, 3373, and 1665 cm–1. The 13C NMR resolved 14 carbon signals, which were classified by chemical shifts and the heteronuclear single quantum coherence spectrum as 4 methyls, 2 methylenes, 2 methines, 2 quaternary carbons (1 oxygenated), with 1 double bond, and 2 ketone carbonyls. Compound 1 should be bicyclic to satisfy 5 degrees of unsaturation. Notably, the geminal coupling constant of the methylene (δH 1.95, dd, J = 8.9, 4.0 Hz and δH 1.05, dd, J = 6.8, 4.0 Hz) is 4.0 Hz (SI, Table S1), suggesting that these two signals were characteristic protons of a cyclopropyl ring.[11] The cyclopropyl ring was determined to be assembled through C-1/C-6/C-7 on the basis of the 1H–1H COSY correlation of H-6/H2-7 as well as the key heteronuclear multiple bond correlation (HMBC) correlation of H2-7/C-1. Furthermore, the cyclopropyl ring fused with a cyclohexenone was determined by the HMBC correlations from H3-8 to C-2, C-3, C-4, from OH-5 to C-4, C-5, C-6, C-9, and from H2-7 to C-2, C-5 in the HMBC spectrum (Figure ). Furthermore, one 2-methylbutanoyl group was revealed by the coupling system CH3 (4′)–CH2 (3′)–CH (2′)–CH3 (5′) in 1H–1H COSY and the linkage of which to C-1 was confirmed by the HMBC correlations of H3-5′/C-1 and H3-5′/C-1′. Thus, the structure of 1 was determined as depicted, with an unprecedented bicyclo[4.1.0]heptenone skeleton. The present correlation signal of H-6/H3-9 and the absent signal of H2a-7/H3-9 in nuclear Overhauser enhancement spectroscopy (NOESY) determined the same side orientation of H-6 and H3-9 (SI, Figure S1). Further analysis of the single-crystal X-ray diffraction with Cu Kα radiation (Figure ) (CCDC 1545859) determined the absolute configuration of 1 as (1S,5S,6R)-5-hydroxy-3,5-dimethyl-1-((S)-2-methylbutanoyl)bicyclo[4.1.0]hept-3-en-2-one. In addition, the good agreement between the calculated electronic circular dichroism (ECD) and the experimental one (SI, Figure S2a) further confirmed the structure.

Figure 2

Selected HMBC (H → C) and 1H–1H COSY (H–H) correlations of 1 and 10.

Figure 3

X-ray crystallographic structures of 1 and 5.

Ophiosphaerellins B–D (2–4) were all diastereomers of compound 1 and simultaneously obtained with 1 by high-performance liquid chromatography (SI, Figure S3), having the same molecular formula C14H18O3 and very similar NMR data. The H-6/H3-9 NOESY correlation (SI, Figure S1) alone with the nearly identical 13C NMR data of the cyclohexenone moiety and the opposite CD at 216 nm caused by π → π* transitions[12] between 1 and 2 (Figure ) showed that both cyclohexenone moieties in 1 and 2 were symmetrically identical. As a diastereomer of 1, the agreement of circular dichroism (CD) with (1R,5R,6S,2′S)-2 led to the final stereochemical determination of 2 (SI, Figure S2b). The presence of H2a-7/H3-9 NOESY correlations (SI, Figure S1) of 3 and 4 determined the same side orientation of H2-7 and H3-9 relative to the heptenone ring. Furthermore, 3 and 4 were also symmetrically identical in the cyclohexenone moiety for their opposite CD (Figure ). The agreement between the Boltzmann-averaged ECD spectra of (1S,5R,6R,2′S)-3 and (1R,5S,6S,2′S)-4 and the experimental ones (SI, Figure S2c,d) determined the stereostructures of 3 and 4, respectively. Herein, by comparing the Cotton effects (CEs) at 210–240 nm caused by the π → π* transitions with the configuration of C-1(C-6), we can find that 1R(6S) leads to a positive CE effect, whereas 1S(6R) leads to a negative one. From the stereochemistry and 13C NMR data of 1–4, we can generalize that the chemical shift of C-5′ moves downfield when the absolute configurations of C-2′ and C-1 are inconsistent.

Figure 4

Experimental ECD spectra of 1–9.

Ophiosphaerellin E (5) was a congener of 1 with the molecular formula C14H20O4. The similar NMR data (Tables and S1) and the downfield shift of C-3′ suggested that 5 was the C-3′ hydroxylated derivative of 1. On the basis of the single-crystal X-ray diffraction with Cu Kα radiation (Figure ) (CCDC 1545863) and the consistent ECD curves with the calculated (1S,5R,6R,2′R,3′R)-5 (SI, Figure S2e), 5 was determined as (1S,5R,6R)-5-hydroxy-1-((2R,3R)-3-hydroxy-2-methylbutanoyl)-3,5-dimethylbicyclo[4.1.0]hept-3-en-2-one.

Table 1

13C NMR Data in ppm for 1–9 (100 MHz; Acetone-d6, Values in ppm)

position	1	2	3	4	5	6	7	8	9
1	42.7 s	42.0 s	42.1 s	41.5 s	42.4 s	43.7 s	43.0 s	44.8 s	40.9 s
2	194.7 s	194.7 s	195.0 s	195.0 s	196.1 s	195.2 s	195.6 s	195.4 s	195.2 s
3	130.8 s	130.8 s	133.2 s	133.1 s	133.0 s	132.6 s	130.7 s	130.6 s	132.5 s
4	147.2 d	147.1 d	144.4 d	144.4 d	145.0 s	145.1 d	147.5 d	147.5 d	144.6 s
5	67.8 s	67.8 s	66.9 s	66.9 s	66.8 s	66.6 s	67.8 s	68.0 s	67.2 s
6	37.5 d	37.4 d	40.3 d	40.3 d	40.6 d	40.9 d	36.9 d	37.4 d	36.7 d
7	20.7 t	20.5 t	21.9 t	21.8 t	22.8 t	21.6 t	21.3 t	20.2 t	19.1 t
8	16.0 q	16.0 q	16.2 q	16.2 q	16.2 q	16.1 q	15.9 q	15.9 q	16.1 q
9	31.5 q	31.5 q	28.2 q	28.2 q	27.9 q	27.4 q	31.3 q	31.5 q	28.6 q
1′	209.2 s	209.2 s	209.6 s	209.8 s	209.8 s	209.4 s	209.3 s	209.6 s	197.4 s
2′	46.1 d	45.9 d	46.0 d	46.0 d	52.4 d	52.0 d	52.3 d	53.1 d	137.5 s
3′	27.5 t	26.0 t	27.5 t	26.2 t	69.9 d	71.1 d	69.6 d	70.8 d	139.3 d
4′	11.9 q	12.1 q	11.8 q	12.1 q	21.4 q	21.5 q	21.1 q	21.6 q	14.6 q
5′	15.8 q	17.8 q	15.9 q	18.1 q	15.0 q	13.2 q	14.5 q	13.8 q	11.6 q

Ophiosphaerellins F–H (6–8), with the same molecular formula of C14H20O4 and similar NMR data, turned out to be diastereomers of 5. The same strategy was adapted for 6–8. The NOESY correlations H-6/OH-5 in 6 and H-6/H3-9 in 7 and 8 (SI, Figure S1) and the signs of the CEs (Figure ) indicated the absolute stereochemistry of the cyclohexenone moieties. On the basis of the study result that the hydroxylated methines in erythro isomers show an upfield shift compared to those in threo,[13,14] the side chain of 6–8 proved to be as threo as in 5 because none of the chemical shifts of C-3′ in 6–8 were upfield. Therefore, the (1R,5S,6S,2′R,3′R)-6 agreed with the experimental ECD spectrum (SI, Figure S2f), which further confirmed the structure of 6. Furthermore, the downfield shift of C-5′ (δC 14.5) of 7 compared with the upfield shift of C-5′ (δC 13.8) of 8 revealed that the absolute configurations of C-2′ and C-1 were inconsistent in 7 and consistent in 8. Finally, the absolute configurations of 7 and 8 were definitely determined as (1S,5S,6R,2′R,3′R) and (1R,5R,6S,2′R,3′R), respectively.

The molecular formula of ophiosphaerellin I (9) was established as C14H18O3 with four degrees of unsaturation on the basis of HRESIMS and NMR data. 9 was very similar to 1 except for the presence of a double bond between C-2′ and C-3′, supported by NMR data (Tables and S1). The rotating frame Overhauser enhancement spectroscopy (ROESY) correlation H3-5′/H-3′ (SI, Figure S1) demonstrated that the Δ2′(3′) olefin had the Z configuration. Combining the negative CE at 228 nm (Δε = −6.97) with the calculated CD result (SI, Figure S2g), 9 was elucidated to be (1S,5S,6R)-5-hydroxy-3,5-dimethyl-1-((Z)-2-methylbut-2-enoyl)bicyclo[4.1.0]hept-3-en-2-one.

Ophiosphaerekorrin A (10) displayed the molecular formula of C14H16O3, as determined by its 13C NMR (SI, Table S2) and HRESIMS (m/z 233.1173 [M + H] +, calcd 233.1172) data, indicating 7 degrees of unsaturation. Analysis of the one-dimensional (1D) NMR data of 10 (SI, Table S2) revealed 14 carbon signals that were attributable to 4 methyls, 2 quaternary carbons (1 oxygenated, δC 98.0), 2 ketone carbonyls, and 3 double bonds (1 oxygenated, δC 186.3), which suggested that there should be 2 rings in total. The oxaspiral framework was assembled by the key HMBC correlations from H3-13 to C-5 (δC 98.0) (Figure ), as a cyclopentenone with a 2,3-dimethyl furanone (supported by the correlations from H3-10, H3-11 to C-2 (δC 186.3) and from H3-11 to C-3, C-4). The calculated ECD spectrum for (5R,9S)-10 was consistent with its experimental data (SI, Figure S2h). Thus, compound 10 was determined to be (5R,9S)-2,3,7,9-tetramethyl-9-vinyl-1-oxaspiro[4.4]nona-2,7-diene-4,6-dione.

The molecular formula of ophiosphaerekorrin B (11) was deduced to be C14H18O3 by the 13C NMR data (SI, Table S2) and HRESIMS, suggesting one lesser degree of unsaturation than 10. The absence of the double-bond resonances was further supported by the 1H–1H COSY correlations of H3-10/H-2/H-3/H3-11. In addition, H3-10 and H3-11 were cofacial according to their NOESY correlation (SI, Figure S1). The experimental ECD spectrum was in close agreement with the calculated one of (2S,3R,5S,9S)-11 (SI, Figure S2i). Therefore, 11 was finally determined to be (2S,3R,5S,9S)-2,3,7,9-tetramethyl-9-vinyl-1-oxaspiro[4.4]non-7-ene-4,6-dione.

Very interestingly, by considering the features of the two types of isolated metabolites, a polyketide biosynthetic pathway was proposed (SI, Scheme S1). The key branch chain polyketide (d) was formed by the two intermediates a and b (biosynthesized from the propionate and acetate units).[15−17] After intramolecular 1,6-aldol reaction, the cyclopropane ring would be readily constructed by the attack from C-7 to C-1 carbocation.[18,19] After a series of reduction, dehydration, and oxygenation, 1–9 were formed. Meanwhile, the spiro-derivatives of 10 and 11 were biosynthesized by decarboxylation, followed by radical induced intramolecular cyclization.

Most compounds (1–6 and 9–11) were evaluated preliminarily for AChE inhibitory activity using a thin layer chromatography-bioautographic assay.[20] The minimum quantity of 3 required for AChE inhibition was 1.25 μg, compared to galanthamine (positive control, 0.006 μg) (SI, Figure S4). Moreover, compounds 9–10 exhibited weak activity, with minimal inhibitory quantities from 30 and 10 μg.

Conclusions

In summary, compounds 1–11 are two types of unprecedented scaffoldings with bicyclo[4.1.0]heptenone (1–9) and oxaspiro[4.4]nonenone (10–11) substructures, which are first isolated from the endolichenic fungus O. korrae. These isolated metabolites are optical diastereomers, which means that they are formed through the enzyme-catalyzed pathway. Furthermore, the rule of the CEs at 210–240 nm caused by π → π* transitions is a useful indicator to determine the stereochemistry of C-1 of the bicyclo[4.1.0]heptenone derivatives (SI, Table S3). In addition, by comparing the chemical shift of C-5′ in 13C NMR, it can be decided whether the absolute configurations of C-1 and C-2′ are consistent or not. On the basis of the steric effect of 5-OH on the chemical shifts of its “neighbors”, a way of determining the relative configurations of the cyclohexenone moiety is of value as well.[21] In addition, these two scaffoldings exhibited inhibitory activity against AChE and provided a possibility of finding AChE inhibitors by further chemical modification.