Total Synthesis of Ophiorrhine A, G and Ophiorrhiside E Featuring a Bioinspired Intramolecular Diels-Alder Cycloaddition.

We report the first total synthesis of the monoterpene indole alkaloids ophiorrhine A via a late stage bioinspired intramolecular Diels-Alder cycloaddition to form the intricate bridged and spirannic polycyclic system. Several strategies were investigated to construct the indolopyridone moiety of ophiorrhiside E, the postulated biosynthetic precursor of ophiorrhine A. Eventually, the Friedel-Crafts-type coupling of N-methyl indolyl-acetamide with a secologanin-derived acid chloride delivered ophiorrhine G. Cyclodehydration of a protected form of the latter was followed by the desired spontaneous intramolecular Diels-Alder cycloaddition of protected ophiorrhiside E leading to ophiorrhine A.

Introduction

Ophiorrhine A and B (1 a, b) are monoterpene indole alkaloids isolated from Ophiorrhiza japonica and display an intricate polycyclic‐fused structure which was secured by single crystal X‐ray diffraction with a highly unusual bridged‐spirocyclic ring system (Scheme 1). These natural products exhibit in vitro immunosuppressive activity.

Scheme 1

Biosynthetic oxidative cyclization pathways from strictosidine into glycosidic monoterpene indole alkaloids.

The monoterpene indole alkaloids which comprise over 3000 natural products are biosynthetically derived from strictosidine (2) as their common precursor. Ophiorrhines A,B are among the few monoterpene indole alkaloids which are biosynthetically produced by an oxidative cyclization between the tryptamine part and one of the double bond of the terpene part without deglucosylation such as cymoside (3) or 3‐α‐dihydrocadambine (4). These pathways are in contrast with the main one which involves a deglucosylation event of strictosidine which is usually followed by the condensation of the thus released aldehyde with the secondary amine N4 to give birth to the corynanthe skeleton of 5 and eventually to a large diversity of sub‐families of monoterpene indole alkaloids.

Obviously, the biosynthesis of the azabicyclic[2.2.2]octanone central core of ophiorrhines A,B involves the intramolecular [4+2] Diels–Alder cycloaddition between the indolopyridone moiety and the C18=C19 terminal alkene of ophiorriside E (6 a) from Ophiorrhiza trichocarpon or its 6‐methoxy analogue 6 b (Scheme 2). The recently discovered ophiorrhine G and F (7 a, b) from also Ophiorrhiza japonica appeared to be hydrolyzed forms of the indolopyridone moiety of the latters (6 a, b). Of importance, several related natural products isolated from the genus Ophiorrhiza possess oxidation forms of the C‐piperidine ring of strictosidine,[ , ] the biosynthetic precursor of the monoterpene indole alkaloids.

Scheme 2

Postulated biosynthesis of ophiorrhines A,B and related monoterpene indole alkaloids from Ophiorrhiza genus.

For instance, lyaloside (8 a)[ , ] and 3,4,5,6‐tetradehydrodolichantoside (9 a)[ , ] display respectively a pyridine and a N‐methyl pyridinium, while 5‐oxodolichantoside (10 a)[ , ] and ophiorrhiside B (10 c) present a N‐methylpiperidone. The precise biosynthetic interconnections between these compounds have yet to be determined. Among others postulates, the indolopyridone of ophiorrhiside E (6 a) could be produced by oxidation of either the N‐methylpyridinium of 3,4,5,6‐tetradehydrodolichantoside (9 a) or the piperidone of 10 a or by intramolecular condensation of the N‐methyl amide of ophiorrhine G (7 a) onto its ketone.

Nevertheless, these biosynthetic hypotheses offer inspiration for organic chemists.

In the context of our recent syntheses of monoterpene indole alkaloids including cymoside by bioinspired oxidative couplings, we proposed to access ophiorrhine A (1) through a bioinspired Diels–Alder cycloaddition from ophiorrhiside E (6 a, Scheme 3). It would represent a divergent total synthesis approach from strictosidine or secologanin in complement to the syntheses of cymoside and 3‐α‐dihydrocadambine.[ , ]

Scheme 3

Retrosynthesis of ophiorrhine A (Glc=β‐d‐glucose).

Results and Discussion

The key indolopyridone moiety of 6a would be obtained through oxidation of the N‐methylpyridinium of 9 a (Scheme 3, A) which would be synthesized from the oxidative decarboxylation of the Pictet–Spengler product of L‐tryptophan (12) and secologanin (13) followed by N‐methylation of the resulting pyridine of lyaloside (8 a). Alternatively, the indolopyridone of 6 a is envisioned to be obtained via a cyclodehydration of ophiorrhine G (7 a) which would arise from the acylation of indolylacetic acid or its derivatives 15 a–c with a carboxylic acid derivative 14 of secologanin (Scheme 3, B).

In the course of our study, we prepared both racemic secologanin aglycon ethyl ether (±)‐16 and enantiopure tetra‐acetylated secologanin (−)‐17. The former was prepared according to Tietze via a hetero Diels–Alder cycloaddition between enol ether 20 and the enal derived from the condensation of formyl‐ketone 19 and aldehyde 18. In the other hand, the synthesis of (−)‐17 relied on the work of Ishikawa via an organocatalytic enantioselective Michael addition of aldehyde 22 onto 21. Oxidation of the aldehyde of (±)‐16 and (−)‐17 led to the corresponding carboxylic acids (±)‐23 and (−)‐24. Secologanin aglycone (±)‐16 could also be converted into nitrile (±)‐25 in 1 step with O‐benzoylhydroxylamine 26 in presence of camphorsulphonic acid (CSA).

For each approaches, the feasibility to access the key indolopyridone moiety was first tested with a less complex substrate in lieu of the corresponding secologanin derivative.

The indolopyridine approach (Scheme 3, A) involved the Pictet–Spengler reaction of L‐tryptophan 12 with iso‐valeraldehyde 27 or secologanin ethyl ether aglycone (±)‐16 in trifluoroacetic acid (TFA, Scheme 4). It was followed by a decarboxylation and aromatization sequence mediated by N‐chlorosuccinimide (NCS) with triethylamine to produce the pyridine ring of 28 a in 56 % yield and of lyaloside ethyl ether aglycone (±)‐28 b in 29 % yield.

Scheme 4

Attempts towards the synthesis of the indolopyridone moiety.

The corresponding N‐methylpyridinium salts 29 a and 3,4,5,6‐tetrahydrodolichantoside aglycone ethyl ether (±)‐29 b were then obtained in 93 % and 64 % yields via reaction of the pyridine of 28 a, b with excess of iodomethane at 80 °C in acetonitrile. We then pursued the pivotal oxidation of the indolopyrdinium ring into the desired indolopyridone. Known procedures for the oxidation of N‐methylpyridinium salts into pyridones were screened on 29 a such as the use of potassium ferricyanide or of Eosin Y as a photoredox catalyst under visible light irradiation or heating in air with a strong base in dimethyl sulfoxide (DMSO). Unfortunately, none of these conditions could allow to form the desired indolopyridones 30 a, b with oxidation at the 5‐position. Only pyridone 31 could be observed resulting of oxidation at the more activated 3‐position followed by a C−C bond cleavage.

Therefore, we attempted to perform the intramolecular [4+2] cycloaddition at an earlier stage (Scheme 4). The N‐methyl pyridinium of lyaloside ethyl ether aglycone (29 b) was heated in a solution of bromobenzene and water (25 : 1) at 156 °C in order to perform a Bradsher cycloaddition. However, mainly starting material could be detected.

We envisioned that the formation of a pyridine‐N‐oxide and its reaction with anhydride acetic could lead the desired indolopyridone moiety via addition of an acetate to an N‐acetoxypyridinium intermediate. Indeed, reaction of pyridines 28 a and (±)‐28 b with meta‐chloroperbenzoic acid (m‐CPBA) led respectively to pyridine N‐oxides 33 a and (±)‐33 b. Heating of 33 a in acetic anhydride (Ac2O) led to undesired α‐acetoxy pyridine (±)‐34 a. From (±)‐34 a, it was possible to effect the desired transformation via the same sequence of oxidation of the pyridine and then reflux in anhydride acetic to obtain acetoxypyridine (±)‐35 a which reaction with potassium carbonate and methyl iodide led to indolopyridone (±)‐36 a. However, this approach is not straightforward since an undesired alcohol is present on the substituent of the pyridone.

Nevertheless, switching acetic anhydride by triflic anhydride (Tf2O) allowed to convert selectively pyridine N‐oxide 33 a at room temperature to the desired α‐triflyloxylpyridine 37 a in 74 % yield. Methylation into 38 a and treatment with lithium hydroxide led with delight to N‐methylpyridone 39 a. However, we were disappointed that this sequence could not be implemented to the more complex secologanin‐derived pyridine N‐oxide (±)‐33 b since its treatment with triflic anhydride led to a complex mixture with less than 10 % of (±)‐37 b.

We then turned our attention towards the reaction of indolylacetic acid derivatives 15 a–c with carbonyl reagents according to the alternative retrosynthesis (Scheme 3, B).

Without surprised the acid‐mediated reaction between iso‐valeraldehyde aldehyde 27 and indolylacetamide 15 a did not lead to indolopiperidone 41 but to the double addition of the indole of 15 a to the carbonyl of 27 (Scheme 4).

Very interestingly, the palladium catalyzed reaction of indolylocarboxylic acid 15 c with butyronitrile 42 led to indolopyridone 43 a with bipyridine as ligand (bpy). Unfortunately, the same reaction with secologanin‐derived nitrile (±)‐34 b did not succeeded (Scheme 4).

Eventually direct acylation of 15 a, b with a carboxylic acid into a 2‐acylindole was considered (Scheme 5). As a test substrate, we selected 5‐hexenoic acid 44 which leads to a suitable substrate model for the intramolecular Diels–Alder cycloaddition.

Scheme 5

Total synthesis of ophiorrhine A via a cyclodehydration/Diels–Alder cycloaddition from protected ophiorrhine G.

After optimization, we were able to perform the Friedel–Crafts‐type coupling of indolyl acetic acid methyl ester 15 b with 5‐hexenoic acid chloride 45 in presence of a catalytic amount of tin tetrachloride to furnish 46 b in 80 % yield. The same acylation reaction from indolylacetamide 15 a required a stoichiometric amount of tin tetrachloride and delivered the 2‐acylated indole 46 a in 52 % yield.

Both 46 a and 46 b could be converted efficiently into the expected indolopyridone 47, either via reaction of methyl ester 46 b with excess of an ethanolic solution of methylamine and ammonium acetate at 95 °C in toluene or by cyclodehydration of keto‐amide 46 a with triethylamine in reflux of acetic acid.[ , ] In the literature, only one example of a Diels–Alder cycloaddition of indolopyridone is reported with electron‐poor dienophiles while our present approach involves an electron‐rich dienophile. The aromatic character of the 2‐pyridone moiety part via tautomeric equilibrium necessitates to supply a high activation energy such as a high temperature to accomplish the [4+2] cycloaddition. Nevertheless, it has been shown that the presence of a N‐alkyl substituent could enable the cycloaddition of 2‐pyridones. In addition, in these harsh reaction conditions, avoiding the retro Diels–Alder reaction of the cycloadduct with extrusion of methyl isocyanate is another challenge to overcome.[ , ] Rewardingly, the key intramolecular Diels–Alder cycloaddition was achieved via heating indolopyridone 47 in bromobenzene at 156 °C leading to azabicyclo[2.2.2]octenone skeleton of (±)‐48 as one diastereoisomer which relative stereochemistry could not be determined.

Having demonstrated the proof of principle with a simplified carboxylic acid, we then aimed to implement our strategy to the secologanin template (Scheme 5). Based on the results on the model substrate, indolyl acetic acid methyl ester 15 b seemed to be a more suitable substrate for the acylation than the corresponding amide 15 a. Thus, the acid of secologanin aglycon ethyl ether (±)‐23 a was converted into the corresponding acid chloride and the Friedel–Crafts‐type reaction with 15 b in presence of a catalytic amount of tin tetrachloride followed by treatment with methylamine allowed to obtain ophiorrhiside E aglycone ethyl ether (±)‐51 in 14 % over two steps. Heating of the latter in bromobenzene at 156 °C delivered stereoselectively ophiorrhine A aglycon ethyl ether (±)‐53 a in 60 %. Unfortunately, the same sequence could not be applied to protected secologanin acid (−)‐24. While the acylation of 15 b furnished (−)‐50 b in 46 %, the reaction of the latter with methyl amine could not deliver protected ophiorrhiside E (−)‐52, only decomposition or partial deacetylation of protected sugar were observed. Therefore, we believed that starting the sequence from a substrate already containing the acetamide functionality would prevent this problem. Rewardingly, reaction of 15 a with the acid chloride of protected secologanin acid (−)‐24 with a stoichiometric amount of tin tetrachloride allowed us to obtain, in 44 % yield, (−)‐50 a, which deacetylation with potassium carbonate in methanol could yield quantitatively ophiorrhine G 7 a.

Formation of the key indolopyridone motif was sought by cyclodehydration of 2‐(2‐acyl‐3‐indolyl)‐acetamide (−)‐50 a with triethylamine in acetic acid according to the conditions developed on the model substrate 46 a. Thus, (−)‐50 a was submitted to heating at 125 °C with a large excess of triethylamine in acetic acid for 6 h. Surprisingly, only traces of indolopyridone 52 were observed in these conditions. To our delight, the successful cyclodehydration was followed by the spontaneous bioinspired diastereoselective intramolecular Diels–Alder cycloaddition of transient 52 to deliver the desired spirocyclic azabicyclic[2.2.2]octanone (−)‐54 in 58 % yield. Finally, the total synthesis of ophiorrhine A (−)‐1 a was achieved via quantitative methanolysis of the four acetate of the glucose moiety of (−)‐54.

Intriguingly, during the dehydrative intramolecular condensation/Diels–Alder sequence, carbazole (+)‐55 was also observed which could be methanolysed into (+)‐56 in 29 % over 2 steps. We demonstrated that carbazole (+)‐55 could be formed from azabicyclic[2.2.2]octanone (−)‐54 in reflux of acetic acid and triethylamine (Scheme 6). The benzamide part of (+)‐55 probably raised from the ring opening of the azabicyclic[2.2.2]octanone via cleavage of the N4−C3 bond followed by aromatizing C18−C19 dehydrogenation of 57. This remarkable bond reorganization delivers an original strictosidine‐derived skeleton which is not a natural product or yet to be discovered from natural sources. This mechanism is in contrast with the anticipated retro Diels–Alder reaction from (−)‐54 with extrusion of methyl isocyanate which was not observed in our case.[ , ]

Scheme 6

Formation of carbazole (+)‐55 via aromatizing ring opening of the bicyclo[2.2.2]template of ophiorrhine A.

In order to form indolopyridone 52 and prevent the spontaneous Diels–Alder cycloaddition, the cyclodehydration had to be performed at a lower temperature (80 °C) after which, removal of the four acetates with methanol and potassium carbonate led to ophiorrhiside E 6 a in 30 % over two steps (Scheme 5). At 80 °C, the conversion of (−)‐50 a into 52 is modest since 57 % of ophiorrhine G 7 a was also isolated. Eventually, ophiorrhiside E 6 a could also be obtained in 20 % via cyclodehydration at 80 °C of ophiorrhine G 7 a.

Conclusion

In conclusion, we performed the first total synthesis of ophiorrhines A and G, as well as ophiorrhiside E. Ophiorrhine A and its spirocyclic ring systems were accessed through a bioinspired intramolecular Diels–Alder cycloaddition between the pyridone moiety and terminal alkene of protected ophiorrhiside E. The latter was obtained through acylation of N‐methyl indolylacetamide by a secologanin derivative followed by the cyclodehydration of the resulting protected ophiorrhine G.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.