Efficiency is the major concern of contemporary organic synthesis [1]. Tandem/cascade/domino reaction [2] is an effective strategy to achieve high efficiency. (−)-Chaetominine (1 in Fig. 1) is a hexacyclic fungal tripeptidal alkaloid isolated from endophytes of different origins. Following the isolation of chaetominine [3], its homologues and diastereomers (−)-isochaetominines A–C (2–4) and (+)-14-epi-isochaetominine C(5) were isolated from the solid-substrate culture of an Aspergillus sp. Fungus [4]. In 2016, Lin and coworkers reported that during a screening of their marine microorganism library, the gorgonian (Pseudopterogorgia sp.)-derived fungus Aspergillus versicolor LZD-14-1 was found to possess inhibitory effects against thioredoxin reductase (TrxR), a potential target for the treatment of cancer, AIDS and autoimmune diseases. The investigation led to the isolation of eleven new fumiquinazoline-type alkaloids named versiquinazolines A–K [5]. Among them, (−)-versiquinazoline H(6, the structure displayed in the original paper [5]) also belongs to the chaetominine family. Its relative stereochemistry was determined by 2D NMR techniques, and the absolute configuration was determined to be 14S, 16S, 17S, 26R (referring to 2S, 3S, 11R, 14S according to the numbering system used for compounds 1–4) based on the comparison with the known compound (–)-isochaetominine C(4) [6], and on a comparison of the experimental ECD data with those of calculated for 6 and its enantiomer. On the other hand, the configuration of the isoleucine (Ile) unit was determined to be l by analysis of the degradation product.

The enantioselective total synthesis of (−)-chaetominine has been reported by several groups [7], and the diastereodivergent synthesis of this family of alkaloids has been achieved by the Huang group [6, 8]. We have developed three enantioselective approaches to the chaetominine-family alkaloids. The first-generation approach is the four-step total synthesis of (−)-chaetominine from D-tryptophan (Trp) [7c, f, i, 8a]. The second approach involves the first total synthesis of (−)-chaetominine from L-Trp [7g]. The third-generation method resides in using amino acid benzyl esters as components that allows accessing C₂/C₁₄ cis-stereochemistry of the isochaetominine series of alkaloids [8b]. On the basis of these precedents, we report herein the fourth-generation total synthesis. Through this improved approach, we have achieved the first total synthesis of (−)-versiquinazoline H (7) and the proposed structure (6) as well as their diastereomers. The puzzle about the stereochemistry of (−)-versiquinazoline H has been clarified.

We opted for the total synthesis of the displayed structure (2S, 3S, 11R, 12S, 14S)-versiquinazoline H (6, Fig. 1) as the first objective of this investigation. For this purpose, Huang's third generation approach was adopted [8b]. Thus, the synthesis commenced with the known N-aroyl-L-tryptophan 8 [7g], prepared by aroylation of L-tryptophan (L-Trp) with o-nitrobenzoyl chloride (Scheme 1). The coupling of the mixed anhydride, generated in situ from 8 and i-BuOCOCl/N-methylmorpholine (NMM), with benzyl D-allo-isoleucinate p-toluenesulfonic acid salt (THF, −20 ℃, 12 h) afforded 9 in 92% yield. The low-valent Ti-reagent mediated-quinazolinone formation [9] (Zn/TiCl₄ and trimethyl orthoformate, THF) [7c] proceeded smoothly at 0 ℃ to give 10 in 95% yield. Exposure of 10 to DMDO [8, 10] at −78 ℃ followed by treating the resultant intermediates with an aqueous solution of Na₂SO₃ [8] at 45 ℃ for 2 h afforded, in one-pot, the double cyclization product 11 along with monocyclization product 12 in 32% and 33% yield, respectively. O-Debenzylation of 12 under catalytic hydrogenolytic conditions yielded the corresponding acid derivative 13, which, without isolation, was subjected to lactamization conditions. However, under the previously used lactamization conditions [(COCl)₂ (1.5 equiv.), DIPEA (2.0 equiv.), cat. DMF, DCM (0.04 mol/L), −10 ℃, 45 min], the desired product 6 was obtained in only 29% yield. Increasing (COCl)₂ from 1.5 equiv. to 3.0 equiv. led to an even lower yield (20%). Employing the Ye coupling reagent (DEPBT) [(2.0 equiv.) [11], DIPEA (2.0 equiv.), DCM (0.04 mol/L), 25 ℃, 8 h] resulted in a higher yield of 54%. Finally, it was found that another coupling method employing HOAt (6.0 equiv.) and EDCI (3.0 equiv.) in dichloromethane (0.05 mol/L) (30 ℃, 12 h) produced the proposed structure of versiquinazoline H (6) in 95% yield over two steps. The sense of optical rotation {[α]_D²⁰ 69.6 (c 0.1, CH₃OH)} and spectral (¹H and ¹³C NMR) data of our synthetic compound are different from those reported for the natural (−)-versiquinazoline H {[α]_D²⁰ −100 (c 0.1, MeOH) [5]}, suggesting that the originally proposed stereochemistry (6) for (−)-versiquinazoline H was incorrect.

To address the enantioselective total synthesis of the natural (−)-versiquinazoline H, we commenced with the identification of the stereoisomers of the amino acids present in (−)-versiquinazoline H.

A closer inspection of the original paper [5] showed that the evidences provided and conclusion obtained about the stereochemistry of (−)-versiquinazoline H are contradictory. First, the authors stated: "the absolute configuration was determined to be 14S, 16S, 17S, 26R based on the closely similar ECD and specific rotation data of (−)-versiquinazoline H and (−)-isochaetominine C (4)" [6]. However, as can be seen from Fig. 1, in terms of stereochemistry, the ring system of (−)-isochaetominine C (4) appears to be an "enantiomer" of the proposed structure of "(−)-versiquinazoline H" (6) for the on-ring stereogenic centers. Second, they described: "the configuration of the isoleucine unit has been determined to be l by the advanced Marfey's method". However, the proposed structure of "(−)-versiquinazoline H" (6) contains a D-allo-Ile instead of an L-Ile residue. This analysis allowed us to suggest stereoisomer 7 as the actual structure of the natural (−)-versiquinazoline H.

To further confirm the structure of the natural (−)-versiquinazoline H, we proceeded to synthesize the stereoisomer 7 containing D-Trp and L-Ile residues. Under the standard conditions [7f], the coupling of the known (R)-8 with benzyl L-isoleucinate p-toluenesulfonic acid salt yielded compound 14 in 93% yield. The latter was converted to 15 in 95% yield. The DMDO oxidation of 15 followed by treating the resulting mixture with a saturated aqueous solution of Na₂SO₃ at 45 ℃ for 2 h produced the double cyclization product 16 and monocyclization product 17 in 32% and 35% yield, respectively. O-Debenzylation of 17 under catalytic hydrogenolytic conditions yielded the corresponding amino acid, which, without separation, was subjected to our newly established lactamization conditions [HOAt (6.0 equiv.), EDCI (3.0 equiv.), DCM (0.05 mol/L), 30 ℃, 12 h], which produced compound 7 as a white solid in 97% yield over two steps. The sense of optical rotation and spectral (¹H and ¹³C NMR) data of our synthetic compound fully matched those reported for the natural versiquinazoline H, but a difference exists for the values of specific rotation {synthetic 7: [α]_D²⁰ −74.6 (c 0.1, MeOH); natural product (colorless oil): [α]_D²⁰ −100 (c 0.1, MeOH) [5]}. The results allowed us to conclude that the stereochemistry of the natural (−)-versiquinazoline H to be 2R, 3R, 11S, 12S, 14R as represented by 7. It is worth noting that, H14, H19, H25, C13, C14, C17, C18, and C23 of compounds 11 and 16 are very broad, which are similar to the observations noted by Tan [3a], Snider [7a], and Huang for chaetominine [3a, 7a, 8b] (Scheme 2).

In summary, through evolution of our previously developed strategy, we have completed the first total synthesis of both the proposed and the revised structures of versiquinazoline H both in six steps with an overall yield of 24.4% and 27.3%, respectively. Through this fourth-generation strategy, the stereochemistry of this natural product has been revised to 2R, 3R, 11S, 12S, 14R. Work is in progress in our laboratories to further extend this efficient strategy [12], and results will be reported in due course.

This paper is dedicated to Professor Henry N.C. Wong for his significant contribution to the development of organic chemistry in China. The authors are grateful for the financial support from the National Natural Science Foundation of China (Nos. 21672176 and 21332007) and the National Key R & D Program of China (No. 2017YFA0207302). We thank Ms. Yan-Jiao Gao for assistance in the preparation of this manuscript.

Supplementary material related to this article can befound, in the online version, at doi:https://doi.org/10.1016/j.cclet.2019.06.043