Harnessing the Intrinsic Reactivity within the Aplysinopsin Series for the Synthesis of Intricate Dimers: Natural from Start to Finish

Abstract

Aplysinopsin monomers are considered as plausible biosynthetic precursors of the wider aplysinopsin family of marine alkaloids. The idea of harnessing their intrinsic reactivity to undertake the synthesis of dictazoles or cycloaplysinopsins logically emerged from this status. These biosynthetic considerations led us to the first total syntheses of dictazole B and other valuable cyclobutanes. When further exploiting pre-encoded reactivity, our first total synthesis of tubastrindole B originated from the ring-expansion cascade of its dictazole-type precursor. Moreover, the isolation of a transient biosynthetic intermediate combined with dimerization outcomes of a hydantoin-containing monomer allowed us to explain the formation of cycloaplysinopsins A and B.

1 Introduction

2 Easy Access to Dictazole Cyclobutanes

3 Synthesis of Cycloaplysinopsins by Ring Expansion

4 Conclusion and Future Prospects

Biographical Sketches

Adam Skiredj, born in France in 1989, studied both pharmacy and organic chemistry during a Pharm.D./M.Sc. dual degree program. He obtained his M.Sc. degree in 2013 from Paris-Sud University where he is now carrying out a Ph.D. under the guidance of Prof. Erwan Poupon. He is currently working on the biomimetic synthesis of natural products and their self-assembly via cascade reactions. His research interests also include catalysis and the total synthesis of intricate natural products in a broader sense.

Erwan Poupon is full professor of pharmacognosy and natural product chemistry at Paris-Sud University. He obtained his Pharm.D. from the University of Rennes in 1996 and his Ph.D. from Paris-Descartes University in 2000 under the guidance of Prof. Henri-Philippe Husson. After a post-doctoral period in the group of Emmanuel Theodorakis (University of California in San Diego, USA), he joined the faculty at Paris-Sud University. He is particularly interested in biomimetic strategies in total synthesis and in understanding intimate mechanisms involved in the biosynthetic pathways of specialized metabolites. Other interests include the discovery of new natural products from plants or marine invertebrates as well as natural-product-based drug design.

Delphine Joseph studied organic chemistry at the University of Metz and received her Ph.D. degree in 1995 after working on the synthesis of topoisimerase II inhibitors. She spent two years (1996–1998) as a postdoctoral fellow with Prof. Léon Ghosez at the Catholic University of Leuven (Louvain-la-Neuve, Belgium) where she studied the enantioselective addition of organozinc derivatives to imines (autoinduction of chirality). She joined the group of Prof. Jean d’Angélo as assistant professor in 1998 at the Faculty of Châtenay-Malabry working on asymmetric Michael addition. In 2005, she was promoted full professor of organic chemistry at Paris-Sud University. Her current research interests include the development of the aza-Michael reaction as source of oriented diversity, step-economical, and ecofriendly processes, the design and synthesis of pharmacological tools for studying pentameric LGICs, and the synthesis of natural products and related analogues.

Mehdi A. Beniddir graduated in pharmacy and received his M.Sc. degree from Paris-Sud University in 2009. He obtained his Ph.D. under the guidance of Dr. Françoise Guéritte and Dr. Marc Litaudon at the Institut de Chimie des Substances Naturelles (ICSN-CNRS) in 2012. He was subsequently a postdoctoral fellow with Prof. Erwan Poupon­ at Paris-Sud University, where he became­ an assistant professor of pharmacognosy in 2014. His current research aims to develop MS- and NMR-based dereplication approaches to detect and characterize natural products from complex mixtures.

Laurent Evanno received his Ph.D. degree in 2007 from the University Pierre et Marie Curie, Paris, working on total synthesis under the supervision of Dr. Bastien Nay at the National Museum of Natural History. He undertook postdoctoral research with Prof. Petri­ Pihko at Helsinki University of Technology – TKK in 2008 and Prof. Janine Cossy at ESPCI–Paris Tech in 2009–2010. Since 2010, he has been assistant professor at Paris-Sud University and his research interests include the synthesis and isolation of natural substances especially in the field of indole alkaloids.

Guillaume Bernadat received his chemical engineering diploma in 2004 from École Supérieure de Chimie Physique Électronique de Lyon and his Ph.D. in 2009 from Paris-Sud University under the supervision of Dr. Jieping Zhu and Dr. Géraldine Masson. In 2009–2010, he studied computational chemistry during a postdoctoral stay in Dr. Bogdan Iorga’s group at ICSN (CNRS, Gif-sur-Yvette). In 2011, he joined the Fluorinated Molecules and Medicinal Chemistry group headed by Prof. Sandrine Ongeri and Dr. Benoît Crousse at the Faculty of Pharmacy at Paris-Sud University, where he is now associate professor. His research activities are currently focused around applications of molecular modeling for the design and synthesis of biologically active compounds.

Introduction

Some specific marine sponges and stony corals living in shallow waters are the sources of intriguing alkaloids that share quite obvious structural links with aplysinopsin (1).[1] Hence, this rather simple alkaloid not only gives its name to the group of aplysinopsin monomers (over 20 similar compounds slightly differing from one another by N-demethylation or C6 bromination of 1 and 2, Figure [1]),[2] but it is also the eponym of the whole ‘aplysinopsin alkaloids family’ considered herein. The other congeners known to date, far more complex, define two distinct groups of dimeric compounds depending on their respective scaffolds: 1) dictazoles: remarkable spiro-fused cyclobutanes [e.g., dictazole B (3)] and 2) cycloaplysinopsins: spirocyclic tetrahydrocarbazole-type natural products [e.g., tubastrindoles B (4) and H (5) and cycloaplysinopsins A (6) and B (7)].[3] [4] Beyond the undeniable synthetic challenge represented by the access to these densely functionalized marine alkaloids (two indoles and up to eight nitrogens),[5] the puzzling biosynthetic issues raised by their study prompted us to undertake the present work. Aiming to synthesize some of these complex natural products with high simplicity, if not spontaneously, we have previously reported the first total syntheses of (±)-dictazole B (3) and (±)-tubastrindole B (4).[6] Intrigued by the reactivity of 2 and the variable stereochemical features of several related bis-hydantoin alkaloids of the family (see the structures of 5, 6, and 7 in Figure [1]), we extended our study towards this subset of natural products.

The new outcomes are reported herein in addition to our previous results related to this stunning series of marine alkaloids.

Easy Access to Dictazole Cyclobutanes

In our view, from both biosynthetic and synthetic standpoints, the three different subgroups of alkaloids presented in Figure [1] should not be considered independently, but rather as a unique set of closely related congeners (Scheme [1]). Indeed, these natural products have all been isolated and, therefore, naturally produced in similar environments. A common feature of the environments from which materials have been collected attracted our attention: the shallow waters where all the relevant sponges and stony corals were collected are directly exposed to sunlight.

Guided by this unifying idea, we turned to study the photochemical dimerization of aplysinopsin-type monomers. Prepared, as previously reported, by a trivial aldolization-dehydration step on a gram scale, the desired monomers 1, 2, 8, and 9 were exposed to artificial UV-enriched light under appropriate conditions.[7] These experiments resulted in different outcomes summarized in Table [1].

Table 1 Compared [2+2] Photochemical Homodimerization Results of Aplysinopsin-type Monomers

Entry	Monomer	Additive (equiv)	Yielda (%)
1	1	–	9
2b	1	Cu(OTf) (0.5)	16c
3	8	–	6
4	8	Cu(OTf) (0.5)	12
5	2	–	5d
6	2	Cu(OTf) (0.5)	–
7	9	–	–
8	9	Cu(OTf) (0.5)	–

^a Yield of isolated product.

^b After screening of Lewis and Brønsted acids, the best results were obtained with Cu(OTf)–toluene complex (0.5 equiv).

^c Accompanied by a minor stereoisomer shown in Figure [2].

^d Notably, other stereoisomers shown in Figure [2] were also observed when 2 was exposed to light under these usual conditions.

Although common optimized conditions have been established (i.e., thin film of a 5 mM solution in DMF, hν, 14 h)[8] by iterative screening followed by a kinetic study, the recourse to additives, and their respective efficiency, differs from one monomer to another.

As mentioned in our previous reports, and in contrast with the usual lack of selectivity observed in photochemical [2+2] processes, the self-condensation of 1 by [2+2] photocycloaddition is diastereoselective and leads exclusively to the anti-dictazole-type compound 10. Moreover, the addition of copper(I) triflate significantly enhanced the yield of this transformation, even if it also induced a slight erosion of diastereoselectivity, leading to the isolation of 14, a minor syn-derivative (entries 1 and 2, Figure [2]).[9] Comparable results were observed for 8 albeit 12 is obtained in lower yields than 10 (cf. entries 2 and 4).

Whereas 2 seems very similar to 1 or 8 at first sight, the hydantoin heterocycle greatly influences its reactivity. Indeed, homodimerization of 2 is fully inhibited by copper(I) triflate. Also, the self-induced diastereoselectivity of the photodimerization is lower in the case of 2, as two other minor stereochemical isomers 15 and 16 were observed without the use of an additive for the first time (entries 5 and 6).[10]

These additional results suggests that the imino group plays a significant role in the supramolecular self-organization occurring upon dimerization of these monomers, a process probably responsible of the observed diastereoselectivity.[11] The apparent weakness of this lack of control is intriguing and it will be discussed. At this stage, monomethylated monomer 9 did not dimerize at all under the tested conditions (entries 7 and 8). Far from being sufficient, these promising [2+2] photochemical homodimerization results led us to tackle the heterodimerization issue needed to achieve the synthesis of pseudodimeric dictazole B (3) (Scheme [2]). As predicted, the equimolar heterodimerization between 1 and 8 readily provided a 1:2:1 statistic mixture of 10, pseudodimeric 17 in 8% yield, and 12, respectively. In spite of this small step forward, novel conditions had to be found to get around the initial reactivity of monomer 9 in order to couple it with 8. As mentioned in our initial report, among multiple Lewis acids, only bismuth(III) triflate significantly enhanced the reactivity of 9 and allowed us to obtain desired (±)-dictazole B (3) in 3.4% isolated yield (Scheme [2]).

Synthesis of Cycloaplysinopsins by Ring Expansion

Even though light is sufficient to achieve the synthesis of dictazole-type cyclobutanes, none of the aforementioned photochemical experiments generated any cycloaplysinopsin-type dimers.[12] An additional synthetic link was therefore required to connect this last subset of marine alkaloids to the rest of the family. Relying on a putative intramolecular rearrangement, a conceptual shift led us to consider dictazoles as both plausible biosynthetic intermediates and relevant synthetic precursors towards cycloaplysinopsins (Scheme [3]).[13] The small collection of dictazole-type compounds 10–16 synthesized during our early investigations constituted an ideal array of substrates in order to test this ring expansion hypothesis.

For greater simplicity, 10 was initially chosen for this study because of its centrosymmetry, which should lower the complexity level of our system. Moreover, 10 is readily prepared with the highest yield of the series. Inspired by prior reports, we rapidly focused our efforts on microwave experiments.[13] Under finely tuned conditions (i.e., 5% TFA–H₂O, 110 °C, 85 s under microwave irradiation), cyclobutane 10 was fully converted and (±)-tubastrindole B (4) was isolated in 40% yield (Scheme [4]). Satisfyingly, meaning that this singular intramolecular rearrangement occurs with full retention of configuration, no other stereoisomer is formed during the process, as previously reported.

Owing to the specific functionalization pattern of these compounds and the effective conditions (i.e., possible enol formation in aqueous acidic medium), an ionic formalism seemed reasonable to us.[14] In detail, a regioselective bond cleavage followed by a stereoselective Mannich-type intramolecular condensation are both needed to explain the stereochemical outcome of that cascade. This well-defined open-close sequence should lead to intermediate 18, subsequently undergoing a final tautomerism to restore the natural tetrahydrocarbazole core of (±)-tubastrindole B (4).

Even if we initially conjectured that it could be a transient intermediate rapidly generating the favored indole heterocycle of (±)-tubastrindole B (4) straight after its formation, a control experiment (i.e., identical conditions but without TFA) nonetheless provided 18 in 16% yield without any traces of 4 (Scheme [5]).

Hence, trifluoroacetic acid does play a crucial role in the process and not only at the final tautomerization step. Indeed, there is a substantial gap between the yields of the two experiments carried out with or without trifluoroacetic acid, all else being equal. But there is more, even if 18 was stable enough to be characterized, it rapidly underwent regiospecific hydroxylation at C8 in water after 12 hours at room temperature leading to 20. The spontaneous oxidation of 18 into 19 followed by the straightforward addition of a water molecule onto its Michael acceptor motif can be proposed for this unexpected, but not less crucial result, which will be discussed hereafter.

Then, we turned to study and analyze the reactivity of hydantoin-containing cyclobutanes 11 and 16. However, despite numerous attempts targeting the ring expansion of 11, we were unable to obtain sufficient quantities of the awaited (±)-tubastrindole H (5) even though we generated analytical quantities of a constitutional isomer of 11. Indeed, due to the utmost insolubility of 11 in water and other solvents, the addition of dimethyl sulfoxide or N,N-dimethylformamide was necessary.[15] But, whatever the power applied during microwave experiments, no transformation of 11 was noticed in the presence of dimethyl sulfoxide or N,N-dimethylformamide.[16] Nevertheless, we observed a good conversion of cyclobutane 11 into an isomeric compound when we applied similar conditions to those required for the synthesis of (±)-tubastrindole B (4) from 10 but at very low concentrations of 11 (0.2 mM vs 4.3 mM) in order to overcome its poor solubility. This parameter precluded us from isolating the rearranged product in usable quantities, but both the reaction conditions and its long-term stability in water (thereby excluding the formation of an isomer comparable to 18) suggest the probable formation of (±)-tubastrindole H (5).[17] Moreover, extrapolating our previous synthesis of 4 to minor cyclobutane stereoisomer 16, one could rationalize the formation of cycloaplysinopsins A (6) and B (7). Indeed, if 16 could undergo the aforementioned ring-expansion cascade, a classical pathway restoring the indole heterocycle could provide cycloaplysinopsin A (6) whereas the regioselective hydroxylation at C8 previously observed could lead to cycloaplysinopsin B (7) (Scheme [6]).

Conclusion and Future Prospects

This widened study reveals a remarkable synthetic network linking the intricate congeners of the aplysinopsin family of marine alkaloids (Scheme [7]). Indeed, guided by biosynthetic considerations, we harnessed the self-templating property of aplysinopsin monomers to achieve the first total synthesis of (±)-dictazole B (3) via a diastereoselective [2+2] photocycloaddition. Exploiting once again the intrinsic reactivity of the dictazole series, we came up with the first total synthesis of (±)-tubastrindole B (4) completing thereby the postulated ‘aplysinopsins’ cascade’. Despite being limited by the poor solubility of hydantoin-containing cyclobutane 11, we also took another step towards the total synthesis of (±)-tubastrindole H (5). Finally, beyond these straightforward and efficient total syntheses, the isolation of a transient intermediate and minor dictazole-type stereoisomers allowed us to explain the formation of both cycloaplysinopsins A (6) and B (7). Chiral induction and the study of dimerization under single electron transfer conditions in the aplysinopsin family of alkaloids are currently under investigation in the group and will be reported in due course.

IR spectra were recorded with a Vector 22 Bruker spectrophotometer. NMR spectra were recorded on Bruker AM-300 (300 MHz), AM-400 (400 MHz), and AV-600 (600 MHz) apparatus using CD₃OD, DMSO-d ₆, and DMF-d ₇ as solvents; solvent signals were used as references. HRMS-ESI and LC/MS were run on a Thermoquest TLM LCQ Deca ion-trap spectrometer with an XBridge analytical C₁₈ column (150 × 2.1 mm; 3.5 μm, Waters) and a preparative column (150 × 19 mm, 5 μm, Waters). Sunfire preparative C₁₈ columns (150 × 30 mm; i.d. 5 μm, Waters) were used for preparative HPLC separations using a Waters Delta Prep equipped with a binary pump (Waters 2525) and a UV-visible diode array detector (190–600 nm, Waters 2996). Microwave irradiation was performed using a CEM Discover Labmate microwave oven using 5-mL pressurized vials. Temperature measurements of microwave experiments were performed using an external infrared fiber optic probe. All other chemicals and solvents were purchased from Aldrich and SDS (France) and required no further purification.

Photodimerization; General Procedure

Monomers (0.025 mmol) were solubilized in DMF (5 mL) into an open crystallizing dish (8-cm diameter) in order to provide a thin film (1-mm thick) of solution (5 mM). Solution was allowed to concentrate to dryness (as DMF evaporation during the process is needed for efficient dimerization) with homogeneous exposure to light. The solution was exposed to light for 14 h at r.t. under a normal air atmosphere and the dried residue obtained was examined by LC/MS analysis (MeCN–H₂O, 5:95 to 100:0 + 0.1% HCO₂H, 20 min). Numerous experiments were carried out simultaneously. The batches coming from multiple irradiating runs were then pooled, centrifugated, and purified by reverse-phase HPLC. The centrifugation step is required due to the low solubility of the products prior to purification.

Cyclobutanes 10 and 14

Starting with a 5 mM solution of 1 (1.073 g) in DMF, the batch was divided into 113 individual experiments as follows: to a solution of 1 (9.5 mg, 0.025 mmol) in DMF (5 mL) was added CuOTf–toluene complex (6.5 mg, 0.0125 mmol). The mixture was then exposed to light at r.t. in an open crystallizing dish for 14 h. The dry residue obtained was then solubilized in MeOH (2 mL), pooled with 112 similar experiments and evaporated under reduced pressure. The dry residue was purified, after centrifugation, by preparative HPLC (2% to 30% MeCN–H₂O + 0.2% HCO₂H, 10 min at 42 mL/min, Sunfire) to give 10 (t _R = 5.7 min, 168 mg, 15.7%) and 14 (t _R = 6.0 min, 12 mg, 1.1%).

10

IR (neat): 3142, 1707, 1659, 1589, 1460, 1388 cm^–1.

¹H NMR (400 MHz, CD₃OD): δ = 7.43 (d, J = 7.5 Hz, 2 H), 7.37 (br s, 2 H), 7.17 (t, J = 7.5 Hz, 2 H), 7.04 (t, J = 7.5 Hz, 2 H), 6.94 (d, J = 7.5 Hz, 2 H), 5.50 (s, 2 H), 3.35 (s, 6 H), 3.33 (s, 6 H).

¹³C NMR (100 MHz, CD₃OD): δ = 174.9, 159.2, 137.6, 127.9, 125.0, 123.8, 121.2, 117.3, 113.3, 104.7, 70.1, 48.4, 30.4, 27.4.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₉N₈O₂: 509.2413; found: 509.2419.

14

IR (neat): 3142, 1707, 1659, 1589, 1460, 1388 cm^–1.

¹H NMR (400 MHz, CD₃OD): δ = 7.39 (d, J = 7.9 Hz, 2 H), 7.18–7.13 (m, 4 H), 7.04 (t, J = 7.9 Hz, 2 H), 6.98 (d, J = 7.9 Hz, 2 H), 5.25 (s, 2 H), 3.64 (s, 3 H), 3.30 (s, 3 H), 3.21 (s, 3 H), 3.13 (s, 3 H).

¹³C NMR (100 MHz, CD₃OD): δ = 176.1, 172.3, 159.6, 157.7, 137.7, 127.8, 124.9, 123.7, 121.1, 117.6, 113.2, 104.9, 71.8, 70.1, 47.7, 32.2, 27.8, 27.4, 26.7.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₉N₈O₂: 509.2411; found: 509.2408.

Cyclobutanes 11, 15, and 16

Starting with a 5 mM solution of 2 (360 mg) in DMF, the batch was divided into 45 small experiments as follow: a solution of 2 (8 mg, 0.031 mmol) was exposed to light at r.t. in an open crystallizing dish for 14 h. The dry residue obtained was then solubilized in MeOH (2 mL), pooled with 44 similar experiments and evaporated under reduced pressure. The dry residue was purified, after centrifugation, by preparative HPLC (25% to 60% MeCN–H₂O + 0.2% HCO₂H, 15 min at 42 mL/min, Sunfire) to give 11 (t _R = 13.2 min, 19 mg, 5.3%) and 15 (t _R = 13.8 min, 4 mg, 1.1%). Compound 16 co-eluted with 11 as a minor compound (see the Supporting Information for the presentation of ¹H and NOESY NMR spectra).

11

IR (neat): 3335, 1769, 1695, 1461, 1430, 1392 cm^–1.

¹H NMR (400 MHz, DMSO-d ₆): δ = 7.40 (d, J = 7.5 Hz, 2 H), 7.27 (br s, 2 H), 7.10 (t, J = 7.5 Hz, 2 H), 6.97 (t, J = 7.5 Hz, 2 H), 6.89 (d, J = 7.5 Hz, 2 H), 5.32 (s, 2 H), 3.14 (s, 6 H), 2.97 (s, 6 H).

¹³C NMR (100 MHz, DMSO-d ₆): δ = 174.5, 155.5, 135.6, 126.8, 123.2, 121.8, 119.3, 116.7, 111.9, 105.2, 66.3, 44.8, 27.2, 25.0.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₇N₆O₄: 511.2088; found: 511.2100.

15

IR (neat): 3335, 1769, 1695, 1461, 1430, 1392 cm^–1.

¹H NMR (400 MHz, DMSO-d ₆): δ = 7.39 (d, J = 7.9 Hz, 2 H), 7.15–7.04 (m, 4 H), 7.02–6.89 (m, 4 H), 6.98 (d, J = 7.9 Hz, 2 H), 5.06 (s, 2 H), 3.38 (s, 3 H), 3.04 (s, 3 H), 2.97 (s, 3 H), 2.87 (s, 3 H).

¹³C NMR (100 MHz, DMSO-d ₆): δ = 175.6, 171.8, 156.4, 154.8, 135.6, 126.5, 123.2, 121.7, 119.3, 116.8, 111.9, 104.9, 68.6, 66.0, 44.1, 29.1, 25.2, 24.8, 24.7.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₇N₆O₄: 511.2088; found: 511.2102.

16 alongside 15: ¹H NMR (300 MHz, DMSO-d ₆) see the Supporting Information for spectra of the mixture showing key signals and NOESY correlations. Notably: δ = 7.33 (d, J = 7.8 Hz, 2 H), 7.23 (br s, 2 H), [overlapped or common signals], 5.30 (s, 2 H), 3.21 (s, 6 H), 2.73 (s, 6 H).

Cyclobutane 12

To a solution of 8 (11.5 mg, 0.025 mmol) in DMF (5 mL) was added CuOTf–toluene complex (6.5 mg, 0.0125 mmol). The mixture was then exposed to light at r.t. in an open crystallizing dish for 14 h. The dry residue obtained was then solubilized in MeOH (2 mL), pooled with 13 similar experiments and evaporated under reduced pressure. Compound 12 was purified, after centrifugation by preparative HPLC (5% to 30% MeCN–H₂O + 0.2% HCO₂H, 15 min at 42 mL/min, t _R = 9.3 min) to yield an amorphous solid (18.7 mg, 11.6% isolated yield).

12

IR (neat): 3142, 1707, 1659, 1589, 1460, 1388 cm^–1.

¹H NMR (300 MHz, CD₃OD): δ = 7.58 (d, J = 1.4 Hz, 2 H), 7.34 (br s, 2 H), 7.15 (dd, J = 8.4, 1.4 Hz, 2 H), 6.84 (t, J = 8.4 Hz, 2 H), 5.41 (s, 2 H), 3.32 (s, 6 H), 3.27 (s, 6 H).

¹³C NMR (75 MHz CD₃OD): δ = 175.0, 158.9, 138.4, 127.1, 125.7, 124.2, 119.2, 117.0, 116.0, 105.9, 69.7, 48.3, 30.0, 27.0.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₇N₈O₂Br₂: 665.0624; found: 665.0622.

Cyclobutane 17

To a solution of 1 (5 mg, 0.0125 mmol) and 8 (6 mg, 0.0125 mmol, 1 equiv) in DMF (5 mL) was added CuOTf–toluene complex (6.5 mg, 0.0125 mmol). The mixture was then exposed to light at r.t. in an open crystallizing dish for 14 h. The dry residue obtained was then solubilized in MeOH (2 mL), pooled with 13 similar experiments and evaporated under reduced pressure. Compound 17 was purified after centrifugation by preparative HPLC (2% to 30% MeCN–H₂O + 0.2% HCO₂H, 15 min at 42 mL/min, t _R = 8.9 min) to yield 17 (11.2 mg, 7.6%) as an amorphous solid.

17

IR (neat): 3010, 1653, 1587, 1458, 1397 cm^–1.

¹H NMR (400 MHz, CD₃OD): δ = 7.59 (d, J = 1.4 Hz, 1 H), 7.41 (d, J = 8.4 Hz, 1 H), 7.36 (br s, 1 H), 7.34 (br s, 1 H), 7.13–7.17 (m, 2 H), 7.03 (t, J = 7.6 Hz, 1 H), 6.93 (d, J = 8.0 Hz, 1 H), 6.84 (d, J = 8.4 Hz, 1 H), 5.46 (s, 1 H), 5.44 (s, 1 H), 3.33 (s, 6 H), 3.28 (s, 6 H).

¹³C NMR (100 MHz CD₃OD): δ = 175.2, 158.9, 138.4, 137.6, 128.1, 127.1, 125.7, 124.7, 124.2, 123.6, 121.0, 119.2, 117.6, 116.9, 116.0, 113.2, 106.0, 105.4, 69.8, 48.1, 30.04, 30.00, 27.0.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₈N₈O₂Br: 587.1519; found: 587.1514.

(±)-Dictazole B (3)

To a solution of 8 (6 mg, 0.0125 mmol) and 9 (4 mg, 0.0125 mmol, 1 equiv) in DMF (5 mL) was added Bi(OTf)₃ (8.2 mg, 0.0125 mmol). The mixture was then exposed to light at r.t. in an open crystallizing dish for 14 h. The dry residue obtained was then solubilized in MeOH (2 mL), pooled with 27 similar experiments and evaporated under reduced pressure. Compound 3 was purified after centrifugation on preparative HPLC (15% to 35% MeCN–H₂O + 0.2% HCO₂H, 30 min at 42 mL/min, t _R = 11.4 min) to yield (±)-dictazole B (3) (9.2 mg, 3.4% isolated yield) as an amorphous solid.

IR (neat): 3010, 1653, 1587, 1458, 1397 cm^–1.

¹H NMR (600 MHz, DMSO-d ₆): δ = 11.34 (br s, 1 H), 11.26 (br s, 1 H), 7.62 (br s, 1 H), 7.58 (br s, 1 H), 7.50 (br s, 1 H), 7.18 (br s, 1 H), 7.15 (br d, J = 8.5 Hz, 1 H), 7.07 (br d, J = 8.5 Hz, 1 H), 6.96 (d, J = 8.6 Hz, 1 H), 6.78 (d, J = 8.6 Hz, 1 H), 5.24 (s, 1 H), 5.04 (s, 1 H), 3.31 (s, 3 H), 3.27 (s, 3 H), 2.97 (s, 3 H).

¹³C NMR (150 MHz DMSO-d ₆): δ = 187.7, 173.6, 170.0, 156.7, 136.5, 136.4, 125.8, 125.2, 124.0, 122.3, 121.9, 119.2, 118.2, 114.6, 114.5, 114.4, 105.7, 105.2, 68.3, 68.1, 45.6, 45.4, 29.6, 28.3, 27.0.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₇H₂₅N₈O₂Br₂: 651.0467; found: 651.0464.

(±)-Tubastrindole B (4)

A solution of cyclobutane 10 (186 mg, 2.43 × 10^–1 mmol) in H₂O (56 mL) with an excess of TFA (5 v/v%), was heated under microwave (300 W) to 110 °C (divided into 43 × 1.3 mL small experiments) and maintained at that temperature for 85 s. The solvent was evaporated under reduced pressure to obtain a yellowish dry residue that was quickly purified by preparative HPLC (2% to 20% MeCN–H₂O + 0.2% HCO₂H, 10 min at 42 mL/min) to give 4 (t _R = 8.8 min, 74 mg, 40%).

IR (neat): 3318, 1769, 1690, 1589, 1460, 1388 cm^–1.

¹H NMR (400 MHz, CD₃OD): δ = 7.66 (d, J = 8 Hz, 1 H), 7.55 (d, J = 8 Hz, 1 H), 7.41 (d, J = 8 Hz, 2 H), 7.31 (t, J = 7.5 Hz, 1 H), 7.19 (t, J = 7.5 Hz, 2 H), 7.13 (t, J = 7.5 Hz, 1 H), 7.05 (s, 1 H), 4.56 (s, 1 H), 3.81 (d, J = 17.5 Hz, 1 H), 3.62 (d, J = 17.5 Hz, 1 H), 3.32 (s, 3 H), 3.12 (s, 3 H), 2.91 (s, 3 H), 2.67 (s, 3 H).

¹³C NMR (100 MHz, CD₃OD): δ = 174.4, 172.8, 161.5, 158.8, 139.6, 136.7, 128.9, 126.1, 126.6, 125.8, 124.9, 124.1, 123.9, 121.5, 121.4, 120.1, 118.2, 114.8, 113.08, 113.0, 104.1, 72.3, 72.2, 44.8, 33.1, 28.3, 27.7, 26.7, 26.5.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₉N₈O₂: 509.2413; found: 509.2415.

Biosynthetic Intermediate 18 and Compound 20

A solution of cyclobutane 10 (90 mg, 1.18 × 10^–1 mmol) in H₂O (30 mL) was heated under microwave (300 W) to 110 °C (divided into 23 × 1.3 mL small experiments) and maintained at that temperature for 85 s. The solvent was evaporated under reduced pressure to obtain a yellowish dry residue.

Compound 18 (t _R = 12 min, 14.5 mg, 16%) was obtained when the dry residue was quickly purified by preparative HPLC (1% to 20% MeCN–H₂O + 0.2% HCO₂H, 20 min at 42 mL/min, Sunfire) along with recovered 10 (t _R = 10 min, 35 mg, 39%).

When the microwaved solution remained in H₂O for 12 h, 20 (t _R = 6.7 min, 8 mg, 9%) was formed and purified by preparative HPLC (5% to 30% MeCN–H₂O + 0.2% HCO₂H, 15 min at 17 mL/min, XBridge).

18

IR (neat): 3142, 1769, 1707, 1670, 1589, 1460, 1388 cm^–1.

¹H NMR (400 MHz, CD₃OD): δ = 7.49 (d, J = 7.5 Hz, 1 H), 7.44 (d, J = 7.5 Hz, 1 H), 7.32 (d, J = 7.5 Hz, 1 H), 7.23 (t, J = 7.5 Hz, 1 H), 7.19 (t, J = 7.5 Hz, 1 H), 7.13 (t, J = 7.5 Hz, 1 H), 6.99 (s, 1 H), 6.85 (t, J = 7.5 Hz, 1 H), 6.77 (d, J = 7.5 Hz, 1 H), 5.94 (d, J = 3.5 Hz, 1 H), 5.18 (d, J = 3.5 Hz, 1 H), 4.50 (s, 1 H), 3.26 (s, 3 H), 3.20 (s, 3 H), 3.03 (s, 3 H), 2.82 (s, 3 H).

¹³C NMR (100 MHz, CD₃OD): δ = 175.6, 172.4, 159.2, 158.5, 155.4, 146.9, 137.4, 132.9, 128.3, 126.9, 123.9, 123.4, 123.1, 121.3, 121.1, 117.5, 113.4, 113.1, 107.7, 104.9, 74.7, 72.0, 63.5, 45.4, 31.4, 27.1, 26.8, 26.5.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₉N₈O₂: 509.2411; found: 509.2408.

20

IR (neat): 3318, 1769, 1707, 1670, 1589, 1460, 1388 cm^–1.

¹H NMR (300 MHz, CD₃OD): δ = 7.78 (d, J = 7.7 Hz, 1 H), 7.57 (d, J = 7.7 Hz, 1 H), 7.45 (d, J = 8 Hz, 1 H), 7.39 (d, J = 8 Hz, 1 H), 7.34 (t, J = 7.5 Hz, 1 H), 7.23 (t, J = 7.5 Hz, 1 H), 7.19 (t, J = 7.5 Hz, 1 H), 7.14 (t, J = 7.5 Hz, 1 H), 7.00 (s, 1 H), 5.50 (s, 1 H), 5.03 (s, 1 H), 3.18 (br s, 6 H), 2.84 (s, 3 H), 2.77 (s, 3 H).

¹³C NMR (75 MHz, CD₃OD): δ = 171.3, 170.2, 159.3, 156.6, 137.9, 135.1, 127.6, 126.1, 124.7, 124.1, 123.9, 122.0, 120.3, 119.7, 118.6, 116.9, 111.6, 111.3, 103.6, 73.0, 69.8, 63.6, 38.5, 31.0, 26.5, 25.1, 24.6.

HRMS (ESI-TOF): m/z [M + H]⁺ calcd for C₂₈H₂₉N₈O₃: 525.2363; found: 525.2359.

Acknowledgment

We thank Prof. Cyrille Kouklovsky, Dr. Guillaume Vincent, and Terry Tomakinian (ICMMO, Université Paris-Sud, France) for their help in the realization of microwave experiments. We also thank Jean-Christophe Jullian and Jean-François Gallard (ICSN-CNRS, Gif-sur-Yvette, France) for NMR assistance. We gratefully acknowledge Karine Leblanc­ for HPLC analysis. Finally, LabEx LERMIT (ANR-10-LABX-0033-LERMIT) is acknowledged for funding (grant to A.S.).

• References

• 1a Bialonska D, Zjawiony J. Mar. Drugs 2009; 7: 166 ; http://www.mdpi.com/journal/marinedrugs
  CrossrefSearch in Google Scholar
• 1b Boyd EM, Sperry J. Chem. N. Z. 2010; 74: 109
  Search in Google Scholar
• 2a Kazlauskas R, Murphy PT, Quinn RJ, Wells RJ. Tetrahedron Lett. 1977; 18: 61
  CrossrefSearch in Google Scholar
• 2b Cachet N, Loffredo L, Vicente OO, Thomas OP. Phytochem. Lett. 2013; 6: 205
  CrossrefSearch in Google Scholar
• 3 Dai J, Jiménez JI, Kelly M, Williams PG. J. Org. Chem. 2010; 75: 2399
  CrossrefSearch in Google Scholar
• 4a Mancini I, Guella G, Zibrowius H, Pietra F. Tetrahedron 2003; 59: 8757
  CrossrefSearch in Google Scholar
• 4b Meyer M, Delberghe F, Liron F, Guillaume M, Valentin A, Guyot M. Nat. Prod. Res. 2009; 23: 178
  CrossrefSearch in Google Scholar
• 4c Iwagawa T, Miyazaki M, Okamura H, Nakatani M, Doe M, Takemura K. Tetrahedron Lett. 2003; 44: 2533
  CrossrefSearch in Google Scholar
• 4d Iwagawa T, Miyazaki M, Yokogawa Y, Okamura H, Nakatani M, Doe M, Morimoto Y, Takemura K. Heterocycles 2008; 75: 2023
  CrossrefSearch in Google Scholar
• 4e Balansa W, Islam R, Gilbert DF, Fontaine F, Xiao X, Zhang H, Piggott AM, Lynch JW, Capon RJ. Bioorg. Med. Chem. 2013; 21: 4420
  CrossrefSearch in Google Scholar
• 4f Dai J, Jiménez JI, Kelly M, Barnes S, Lorenzo P, Williams P. J. Nat. Prod. 2008; 71: 1287
  CrossrefSearch in Google Scholar
• 5a Gutekunst WR, Baran PS. J. Org. Chem. 2014; 79: 2430
  CrossrefSearch in Google Scholar
• 5b Snyder SA. Nature (London) 2010; 465: 560
  CrossrefSearch in Google Scholar
• 6a Skiredj A, Beniddir MA, Joseph D, Leblanc K, Bernadat G, Evanno L, Poupon E. Angew. Chem. Int. Ed. 2014; 53: 6419
  CrossrefSearch in Google Scholar
• 6b Skiredj A, Beniddir MA, Joseph D, Leblanc K, Bernadat G, Evanno L, Poupon E. Org. Lett. 2014; 16: 4980
  CrossrefSearch in Google Scholar
• 7 Procedures adapted among others from ref. 4e and previously reported in ref. 6a. Compound 2 was obtained in 63% yield (providing 4.5 g of hydantoin-containing monomer 2) from 3-formylindole and 1,3-dimethylhydantoin.
• 8 Photodimerization conditions: a thin film of a 5 mM monomer solution in DMF is exposed to artificial UV-enriched light under air atmosphere for 14 h with solvent evaporation. A kinetic study monitored over 20 h by ¹H NMR showed the maximum formation of dictazole-type cyclobutanes after 14 h of exposure followed by substantial decomposition for longer experiments.
• 9 Notable decomposition of the monomers into the corresponding formylindoles was observed when adding Cu(OTf) to the reaction medium. Retro-aldol and/or fragmentation of a transient 1,2-dioxetane heterocycle may account for this outcome. We thank Dr Michaël de Paolis (IRCOF, Rouen, France) for suggesting this latter explanation.
• 10 Structures of 14 and 15 were fully ascertained by 2D NMR analysis. Structure of isomer 16 was attributed thanks to its ¹H NMR aromatic pattern: identical to 10 combined with its only two methyl signals and their NOESY correlations (see the Supporting Information for detailed spectra analysis). Nevertheless, we were not able to obtain a pure sample of 16, which was always isolated as the minor constituent of an inseparable mixture with 10.
• 11a Johnson J, Canseco D, Dolliver D, Schetz J, Fronczek F. J. Chem. Crystallogr. 2009; 39: 329
  Search in Google Scholar
• 11b Cohen MD, Schmidt GM. J. J. Chem. Soc. 1964; 1996
  Crossref
• 11c Cohen MD, Schmidt GM. J, Sonntag FI. J. Chem. Soc. 1964; 2000
  Crossref

Even during our early attempts of non-photochemical dimerization with Ledwith–Weitz salt, no dimeric entity was observed. For the use of Ledwith–Weitz salt in the total synthesis of kingianins via RCDA see:
• 12a Drew SL, Lawrence AL, Sherburn MS. Angew. Chem. Int. Ed. 2013; 52: 4221
  CrossrefSearch in Google Scholar
• 12b Drew SL, Lawrence AL, Sherburn MS. Chem. Sci. 2015; 6: 3886
  CrossrefSearch in Google Scholar
• 12c Lim HN, Parker KA. Org. Lett. 2013; 15: 398
  CrossrefSearch in Google Scholar
• 12d Lim HN, Parker KA. J. Org. Chem. 2014; 79: 919
  CrossrefSearch in Google Scholar

Also, preliminary results of photoredox catalysis experiments carried out on monomers 1, 2, 8, and 9 with [Ru(bpy)₃](PF₆)₂ under compatible conditions only led to E/Z isomerization of the starting materials without formation of any dimeric compound, but further investigations are still ongoing to test a questioned single electron transfer (SET) mechanism:
• 12e Prier CK, Rankic DA, MacMillan DW. C. Chem. Rev. 2013; 113: 5322
  CrossrefPubMedSearch in Google Scholar
• 12f Narayanam JM. R, Stephenson CR. J. Chem. Soc. Rev. 2011; 40: 102
  Search in Google Scholar
• 12g Tucker JW, Stephenson CR. J. J. Org. Chem. 2012; 77: 1617
  CrossrefSearch in Google Scholar
• 12h Du J, Yoon TP. J. Am. Chem. Soc. 2009; 131: 14604
  CrossrefSearch in Google Scholar
• 12i Ischay MA, Ament MS, Yoon TP. Chem. Sci. 2012; 3: 2807
  CrossrefSearch in Google Scholar
• 12j Riener M, Nicewicz DA. Chem. Sci. 2013; 4: 2625
  CrossrefSearch in Google Scholar
• 12k Yoon TP. ACS Catal. 2013; 3: 895
  CrossrefSearch in Google Scholar

Inspired by previously reported ring expansion cascade of Baran and Wenkert respectively:
• 13a Baran PS, O’Malley DP, Zografos AL. Angew. Chem. Int. Ed. 2004; 43: 2674
  CrossrefPubMedSearch in Google Scholar
• 13b Wenkert E, Moeller PD. R, Piettre SR, McPhail AT. J. Org. Chem. 1987; 52: 3404
  CrossrefSearch in Google Scholar
• 14 A SET mediated pathway is, however, also conceivable and is currently under investigation in our laboratory.
• 15 Saturation of a solution of 11 in water is visible from less than 0.1 mg/mL (i.e., 0.2 mM).
• 16 Experiments carried out with various proportions of water and DMSO or DMF as cosolvent from 10% to 100% of cosolvent.
• 17 See the Supporting Information for an LC/MS chromatogram showing the formation of an isomeric compound [M + H]⁺ = 511.2092 from 10 in highly diluted medium.

• References

• 1a Bialonska D, Zjawiony J. Mar. Drugs 2009; 7: 166 ; http://www.mdpi.com/journal/marinedrugs
  CrossrefSearch in Google Scholar
• 1b Boyd EM, Sperry J. Chem. N. Z. 2010; 74: 109
  Search in Google Scholar
• 2a Kazlauskas R, Murphy PT, Quinn RJ, Wells RJ. Tetrahedron Lett. 1977; 18: 61
  CrossrefSearch in Google Scholar
• 2b Cachet N, Loffredo L, Vicente OO, Thomas OP. Phytochem. Lett. 2013; 6: 205
  CrossrefSearch in Google Scholar
• 3 Dai J, Jiménez JI, Kelly M, Williams PG. J. Org. Chem. 2010; 75: 2399
  CrossrefSearch in Google Scholar
• 4a Mancini I, Guella G, Zibrowius H, Pietra F. Tetrahedron 2003; 59: 8757
  CrossrefSearch in Google Scholar
• 4b Meyer M, Delberghe F, Liron F, Guillaume M, Valentin A, Guyot M. Nat. Prod. Res. 2009; 23: 178
  CrossrefSearch in Google Scholar
• 4c Iwagawa T, Miyazaki M, Okamura H, Nakatani M, Doe M, Takemura K. Tetrahedron Lett. 2003; 44: 2533
  CrossrefSearch in Google Scholar
• 4d Iwagawa T, Miyazaki M, Yokogawa Y, Okamura H, Nakatani M, Doe M, Morimoto Y, Takemura K. Heterocycles 2008; 75: 2023
  CrossrefSearch in Google Scholar
• 4e Balansa W, Islam R, Gilbert DF, Fontaine F, Xiao X, Zhang H, Piggott AM, Lynch JW, Capon RJ. Bioorg. Med. Chem. 2013; 21: 4420
  CrossrefSearch in Google Scholar
• 4f Dai J, Jiménez JI, Kelly M, Barnes S, Lorenzo P, Williams P. J. Nat. Prod. 2008; 71: 1287
  CrossrefSearch in Google Scholar
• 5a Gutekunst WR, Baran PS. J. Org. Chem. 2014; 79: 2430
  CrossrefSearch in Google Scholar
• 5b Snyder SA. Nature (London) 2010; 465: 560
  CrossrefSearch in Google Scholar
• 6a Skiredj A, Beniddir MA, Joseph D, Leblanc K, Bernadat G, Evanno L, Poupon E. Angew. Chem. Int. Ed. 2014; 53: 6419
  CrossrefSearch in Google Scholar
• 6b Skiredj A, Beniddir MA, Joseph D, Leblanc K, Bernadat G, Evanno L, Poupon E. Org. Lett. 2014; 16: 4980
  CrossrefSearch in Google Scholar
• 7 Procedures adapted among others from ref. 4e and previously reported in ref. 6a. Compound 2 was obtained in 63% yield (providing 4.5 g of hydantoin-containing monomer 2) from 3-formylindole and 1,3-dimethylhydantoin.
• 8 Photodimerization conditions: a thin film of a 5 mM monomer solution in DMF is exposed to artificial UV-enriched light under air atmosphere for 14 h with solvent evaporation. A kinetic study monitored over 20 h by ¹H NMR showed the maximum formation of dictazole-type cyclobutanes after 14 h of exposure followed by substantial decomposition for longer experiments.
• 9 Notable decomposition of the monomers into the corresponding formylindoles was observed when adding Cu(OTf) to the reaction medium. Retro-aldol and/or fragmentation of a transient 1,2-dioxetane heterocycle may account for this outcome. We thank Dr Michaël de Paolis (IRCOF, Rouen, France) for suggesting this latter explanation.
• 10 Structures of 14 and 15 were fully ascertained by 2D NMR analysis. Structure of isomer 16 was attributed thanks to its ¹H NMR aromatic pattern: identical to 10 combined with its only two methyl signals and their NOESY correlations (see the Supporting Information for detailed spectra analysis). Nevertheless, we were not able to obtain a pure sample of 16, which was always isolated as the minor constituent of an inseparable mixture with 10.
• 11a Johnson J, Canseco D, Dolliver D, Schetz J, Fronczek F. J. Chem. Crystallogr. 2009; 39: 329
  Search in Google Scholar
• 11b Cohen MD, Schmidt GM. J. J. Chem. Soc. 1964; 1996
  Crossref
• 11c Cohen MD, Schmidt GM. J, Sonntag FI. J. Chem. Soc. 1964; 2000
  Crossref

Even during our early attempts of non-photochemical dimerization with Ledwith–Weitz salt, no dimeric entity was observed. For the use of Ledwith–Weitz salt in the total synthesis of kingianins via RCDA see:
• 12a Drew SL, Lawrence AL, Sherburn MS. Angew. Chem. Int. Ed. 2013; 52: 4221
  CrossrefSearch in Google Scholar
• 12b Drew SL, Lawrence AL, Sherburn MS. Chem. Sci. 2015; 6: 3886
  CrossrefSearch in Google Scholar
• 12c Lim HN, Parker KA. Org. Lett. 2013; 15: 398
  CrossrefSearch in Google Scholar
• 12d Lim HN, Parker KA. J. Org. Chem. 2014; 79: 919
  CrossrefSearch in Google Scholar

Also, preliminary results of photoredox catalysis experiments carried out on monomers 1, 2, 8, and 9 with [Ru(bpy)₃](PF₆)₂ under compatible conditions only led to E/Z isomerization of the starting materials without formation of any dimeric compound, but further investigations are still ongoing to test a questioned single electron transfer (SET) mechanism:
• 12e Prier CK, Rankic DA, MacMillan DW. C. Chem. Rev. 2013; 113: 5322
  CrossrefPubMedSearch in Google Scholar
• 12f Narayanam JM. R, Stephenson CR. J. Chem. Soc. Rev. 2011; 40: 102
  Search in Google Scholar
• 12g Tucker JW, Stephenson CR. J. J. Org. Chem. 2012; 77: 1617
  CrossrefSearch in Google Scholar
• 12h Du J, Yoon TP. J. Am. Chem. Soc. 2009; 131: 14604
  CrossrefSearch in Google Scholar
• 12i Ischay MA, Ament MS, Yoon TP. Chem. Sci. 2012; 3: 2807
  CrossrefSearch in Google Scholar
• 12j Riener M, Nicewicz DA. Chem. Sci. 2013; 4: 2625
  CrossrefSearch in Google Scholar
• 12k Yoon TP. ACS Catal. 2013; 3: 895
  CrossrefSearch in Google Scholar

Inspired by previously reported ring expansion cascade of Baran and Wenkert respectively:
• 13a Baran PS, O’Malley DP, Zografos AL. Angew. Chem. Int. Ed. 2004; 43: 2674
  CrossrefPubMedSearch in Google Scholar
• 13b Wenkert E, Moeller PD. R, Piettre SR, McPhail AT. J. Org. Chem. 1987; 52: 3404
  CrossrefSearch in Google Scholar
• 14 A SET mediated pathway is, however, also conceivable and is currently under investigation in our laboratory.
• 15 Saturation of a solution of 11 in water is visible from less than 0.1 mg/mL (i.e., 0.2 mM).
• 16 Experiments carried out with various proportions of water and DMSO or DMF as cosolvent from 10% to 100% of cosolvent.
• 17 See the Supporting Information for an LC/MS chromatogram showing the formation of an isomeric compound [M + H]⁺ = 511.2092 from 10 in highly diluted medium.

• Supplementary Material