Total synthesis of (-)-N-methylwelwitindolinone B isothiocyanate via a chlorinative oxabicycle ring-opening strategy.

The first total synthesis of N-methylwelwitindolinone B isothiocyanate is reported. The route features several key steps, including a regio- and diastereoselective chlorinative oxabicycle ring-opening reaction to introduce the challenging alkyl chloride motif.

The total synthesis of indole alkaloids continues to be a fruitful area of scientific pursuit. One particular class of molecules that has provided an exciting arena of chemical discovery is the welwitindolinone natural products, wherein the majority of congeners contain a [4.3.1]-bicyclic core (e.g., 2–4, Figure 1).[1,2] Since Moore’s first isolation report in 1994,[3] roughly 25 manuscripts describing efforts toward these complex structures have appeared from many research groups worldwide.[2b,4] The majority of initially reported studies established a variety of methods for assembling the [4.3.1]-bicyclic core, and subsequent efforts have focused on completing the total syntheses. Toward this latter end, the most recent publications describe formal[5] as well as total syntheses of 3, 4, and C3-oxidized variants of 3.[6,7]

Figure 1

Welwitindolinones 1–4.

Although [4.3.1]-bicyclic welwitindolinones in the C- and D- “series” have been synthesized (e.g., 3 and 4), compounds in the B- “series” (e.g., 2) have yet to be prepared by total synthesis.[4u,4y] Structurally, 2 is quite similar to 3 with the key difference being a variation in the oxidation state at C13.[8,9] However, this seemingly simple change is deceptive, as the alkyl chloride resides on the more congested face of the [4.3.1]-bicycle, adjacent to a quaternary center,[10] and thus presents a formidable challenge with regard to synthesis. In addition to this subtle feature, the alkyl chloride in these systems is prone to undergo a variety of undesirable side reactions under basic reaction conditions.[4y,11] Herein, we describe our efforts toward (−)-2 and the first total synthesis of this elusive natural product.

In our initial efforts, we considered several approaches to 2 that ultimately proved unsuccessful (Scheme 1). In what can be considered the most direct assault, we envisioned 2 as arising from C13–14-reduction of 5 or a related derivative. However, attempts to realize this strategy were thwarted by the facile reduction of the terminal olefin.[12] We also pursued a strategy wherein the alkyl chloride would derive from alcohol 6 by activation and chlorination with stereochemical inversion. Similar to the observations made by Rawal,[4u] we found that the proximal vinyl group underwent formal migration to C13 upon activation of the alcohol.[13] Even in the absence of the vinyl group, the chlorination is known to be difficult and only proceeds under specialized conditions,[4u] as the necessary approach of a chloride nucleophile is somewhat hampered by the steric congestion of the bicyclic scaffold.[14]

Scheme 1

Failed Approaches toward Installation of the Alkyl Chloride

After numerous failed attempts to advance late-stage intermediates from our previous synthesis, we devised the alternative retrosynthetic plan highlighted in Scheme 2. In this revised approach, it was envisioned that 2 would arise from oxazolidinone 7 by late-stage cleavage of the carbamate and further manipulation, all in the presence of the sensitive alkyl chloride. In turn, oxazolidinone 7 would derive from nitrene insertion of carbamate 8.[15] We have previously studied related insertion reactions for C11 functionalization of welwitindolinone scaffolds, but in all prior cases the substrates possessed the opposite stereochemical configuration at C10.[7] Thus, the attempted nitrene insertion of 8 would serve as an opportunity to probe the generality of this method for C11 functionalization. In a critical transformation, we sought to introduce the alkyl chloride of 8 by performing a regio- and diastereoselective chlorinative ring opening of an oxabicycle-containing intermediate (see transition structure 9). This transformation, largely inspired by Shea’s seminal studies,[4y] could provide a solution to the challenge faced earlier. Namely, the necessary approach of the chloride appeared favorable, owing to the restricted conformation of the oxabicycle unit.[16] Importantly, the oxabicycle was envisioned to be readily available from indole 10, which is accessible from enantioenriched carvone derivative 11 and indole 12 in three steps using our previously established procedure involving an indolyne cyclization.[7a,17]

Scheme 2

Modified Retrosynthetic Plan for the Total Synthesis of (−)-2

To implement the plan illustrated in Scheme 2, we first targeted construction of oxabicycle 15 (Scheme 3). To this end, ketone 10 was elaborated to mesylate 13 in two steps involving reduction with LiAlH4 followed by sulfonylation. Upon treatment of 13 with Bu4NF in THF at 80 °C, desilylation readily occurred with concomitant cyclization to afford oxabicycle 14 in 84% yield. Subsequently, a one-pot oxidation/hydrolysis protocol was used to elaborate 14 to the corresponding oxindole 15, which was formed as a single diastereomer.

Scheme 3

Elaboration of 10 to Oxabicycle 15

With rapid access to oxabicycle 15, we were poised to attempt the key chlorinative ring-opening reaction (Scheme 4).[18] We surveyed several conditions that have previously been used for related transformations such as ZnCl2 and acetyl chloride,[19] ethanolic HCl,[20] and TiCl4.[21] Although the use of most reaction conditions led to the recovery of starting material or decomposition, treatment of 15 with BCl3[22] led to consumption of the substrate with opening of the oxabicycle. Unfortunately, the two products obtained were 16, which had undergone formal vinyl migration, and 17, an unproductive constitutional isomer of the desired product, which forms as a result of undesired chloride attack at C10 (rather than C13). In hope of avoiding the vinyl migration, and to perturb the electronic environment at C13,[4u] alkene 15 was exposed to modified oxidative cleavage conditions,[23] which furnished aldehyde 18. To our delight, treatment of 18 with BCl3 in CH2Cl2 at 50 °C delivered the desired chlorinated product 19 in 64% yield. Of note, 19 was obtained as a single diastereomer and the analogous undesired regioisomer was not observed.[24]

Scheme 4

Chlorinative Oxabicycle Opening Studies

Having introduced the alkyl chloride, we turned our attention to installing the C11 nitrogen substituent via the key nitrene insertion reaction (Scheme 5). The requisite substrate for this transformation (8) was accessed from 19 in four steps that began with conversion to silyl ether 20 using a protection/olefination sequence.[25] Deprotection of 20 followed by carbamoylation delivered the nitrene insertion substrate 8 in quantitative yield over two steps. As mentioned above, our previous studies of related nitrene insertion reactions were performed on substrates epimeric at C10.[7] Although these prior attempts routinely delivered the desired C11-functionalized products, Ag-[26] or Rh-promoted[27] nitrene insertion reactions of 8 were regrettably found to predominantly furnish 21, the product of nitrene insertion into the C9–H bond.[28]

Scheme 5

Attempted Nitrene Insertion of Substrate 8

To test if the formation of 21 was strictly an artifact of the stereochemical configuration, we prepared the corresponding C10 epimer of nitrene insertion substrate 8 (Scheme 6). To that end, oxidation of alcohol 19, followed by Wittig olefination, afforded ketone 22. Subsequent reduction of 22 with LiAlD4[29] occurred with complete diastereoselectivity to furnish an alcohol intermediate, which was carbamoylated to provide 23. Fortunately, carbamate 23 proved to be a viable substrate for the desired nitrene insertion reaction; upon treatment of 23 with AgOTf, PhI(OAc)2, and bathophenanthroline in CH3CN at 50 °C, we obtained the C11 functionalized product 24 in 55% yield with 10% recovered 22. The dichotomy regarding the nitrene products derived from substrates 8 and 23 underscores the subtleties often seen in late-stage manipulations in total synthesis. Moreover, the successful formation of 24 is noteworthy in that He’s Ag-based nitrene insertion conditions[26] tolerate the sensitive alkyl chloride unit.

Scheme 6

Nitrene Insertion, Oxazolidinone Cleavage, and Completion of (−)-2

From insertion product 24, all that remained to complete the total synthesis of 2 was cleavage of the carbamate, followed by oxidation and N-functionalization. Despite previously having success with carbamate hydrolysis on related compounds, we found that treatment of 24 with Ba(OH)2 led to decomposition of the alkyl chloride. This led us to develop a milder means for cleaving the carbamate. Prompted by Snieckus’ recent report of cleaving N,N-dialkylcarbamate derivatives of phenols,[30] cyclic carbamate 24 was exposed to Schwartz’ reagent in THF (Scheme 6). Gratifyingly, the carbamate was cleaved selectively to give an amidoalcohol intermediate, where C23 of 24 had conveniently been retained as a formyl group on the bridgehead nitrogen. Oxidation of the alcohol intermediate delivered 25. With the chloride still intact, dehydration with Burgess reagent and sulfurization[31] afforded (−)-N-methylwelwitindolinone B isothiocyanate (2). Analytical data for (−)-2 were found to be identical to those of the natural material in all respects.

In summary, we have completed the first total synthesis of (−)-N-methylwelwitindolinone B isothiocyanate (2) in 15 steps from indolyne cyclization product 10. Critical to the success of our enantiospecific route is the use of a regio- and diastereoselective chlorinative oxabicycle ring-opening reaction to introduce the challenging alkyl chloride. To complete the synthesis, a number of steps were taken, including substrate-specific installation of the C11 nitrogen substituent and oxazolidinone cleavage, all of which proceeded in the presence of the alkyl chloride motif. With our completed synthesis of (−)-2, all structural classes of the welwitindolinones are now accessible by synthetic chemistry.