Synthesis of the debrominated analog of dihydroflustramine C utilizing a sulfur ylide-initiated thio-Claisen rearrangement

In investigating the scope and limitations of the sulfur ylide initiated thio-Claisen rearrangement developed in our laboratory, we have been able to efficiently synthesize highly functionalized pyrroloindoline ring systems. This functionality is present in a variety of natural and non-natural products and here we report our synthesis of the debrominated analog of dihydroflustramine C


Introduction
The presence of vicinal C(3) quaternary substitution in a variety of indoline-containing natural and non-natural products has inspired a number of groups, including ours, to develop new and improved routes to their synthesis. 1,2,3,4Our entry into this area is a result of our discovery that C(3) quaternary substituted indolines 6 can be generated from the coupling of 2-thioindoles 1 with vinyl diazoacetates 2 in the presence of Rh(II) catalysts (Scheme 1). 5 We have proposed that 6 results from a [3,3]-sigmatropic rearrangement of the charge separated ion pair 4 that comes from sulfur ylide 3 and a subsequent proton transfer. 6,7heme 1 Having a route to highly substituted indolines, we became intrigued with the possibility of applying it to the synthesis of indoline containing natural products. 8,9Among the numerous possibilities, it occurred to us that pyrroloindolines of the dihydroflustramine C and amauromine class might come from relatively straightforward manipulations of the product from the sulfur ylide reaction.We now report our initial work in this area and the synthesis of the debrominated analog of dihydroflustamine C.

Results and Discussion
Our initial plan was to generate the indoline precursor to 8 directly from the coupling chemistry of unsubstituted vinyldiazo substrate 11 with 2-thiotryptamine (Scheme 2).Although Aggarwal had previously used 11 in the generation of sulfur ylides, 10 in our hands the coupling of 2thiotryptamine 12 with 11 resulted only in the recovery of starting material.

Scheme 2
Having failed in the direct coupling to the flustramines using 11, we turned to a more circuitous path and vinyl diazoacetate 15 whose use would require the decarboxylation of the ester after coupling.To this goal, we utilized both Boc-and formamide-protected tryptamines 12 and 14, respectively, in the reaction with 15 giving vicinal quaternary substituted indolines 16 and 17, each in 83% yield.

Scheme 3
With 16 and 17 in hand, we opted to initially examine the use of a decarbonylation strategy to convert them into the requisite terminal alkene needed for the synthesis of dehydroflustramine C. 11 Reduction of 16 using i-Bu2AlH gave the corresponding allylic alcohol in 90% yield.While generally stable to a variety of conditions such as the i-Bu2AlH reduction and hydrolysis, we have found thio-imides such as 16 and 17 to be susceptible to reactions with internal nucleophiles.For example, pyrrole 18 was formed in 80% yield when the amine from 16 was treated with NaH.Interestingly, these conditions also resulted in the removal of the Boc group.Oxidation of the allylic alcohol and methyl amine formation gave decarbonylation precursor 19.Unfortunately, the attempted decarbonylation of 19 using Wilkinson's catalyst was unsuccessful in our hands.

Scheme 4
We next explored a metathesis strategy to the requisite alkene (Scheme 5).We envisioned that the incorporation of an olefin into the vinyldiazoester coupling precursor would give the desired product after ring-closing metathesis (RCM). 12With this as a goal, the Rh2(OAc)4 catalyzed coupling of 12 with vinyl diazoacetates 21 and 22 gave indolines 24 and 25, respectively.Unfortunately, all attempts at RCM using the 2 nd generation Grubbs catalyst 26 were completely unsuccessful here.

Scheme 5
Our lack of success in the decarbonylation and metathesis approaches to the terminal olefin directed our attention to an oxidative fragmentation strategy.If this were successful, a subsequent Wittig reaction would deliver the desired material.Of note was that a similar sequence had been employed by Crich during his syntheses of debromoflustramine B and pseudophyryaminol. 12With the Crich precedent in mind, we became interested in carrying out the oxidative fragmentation chemistry on pyrrolo-indoline 27 whose synthesis in a single step from the reaction of 17 with LiAlH4 is illustrated in Scheme 6.This highly efficient reaction involved the reduction of the ester and formamide, along with a subsequent cyclization and reduction of the resulting amidine.

Scheme 6
As with the decarbonylation and RCM chemistry, our attempts to cleave the alkene directly from 27 (or the corresponding N-tosyl derivative of 27) using oxidative conditions (OsO4, NaIO4 or O3) were unsuccessful.We next turned to α,β-unsaturated ester 29.The conversion of 27 into ester 28 was accomplished in a two step, single flask operation by first oxidizing 27 using MnO2 in CHCl3, concentrating the reaction mixture, dissolving the resulting residue containing the aldehyde corresponding to 27 in MeOH, and adding NaCN and additional MnO2 (Scheme 7).In our hands this was superior to the direct oxidation of 27 using the Corey procedure (MnO2, NaCN, MeOH). 13Subsequent to oxidation, the free aminal nitrogen was converted into the corresponding N-tosyl amine derivative 29 in 73% yield.

Scheme 7
While 30 proved to be amenable to oxidative cleavage using OsO4 and NaIO4, long reaction times and stoichiometric amounts of OsO4 were required.Aldehyde 31 was converted into (±)debromodihydroflustramine C in 61% overall yield following Wittig olefination and removal of the N-tosyl group (Scheme 8).

Conclusions
In summary, when coupled with reductive cyclization reactions, sulfur ylide rearrangements lead to the efficient generation of pyrroloindolines.We have shown that these reactions can be used to synthesize debromodihydroflustamine C. Further studies to utilize sulfur ylide rearrangements are ongoing.

Experimental Section
General.Di-ethyl ether (ether), THF, benzene, and toluene were distilled from sodium/benzophenone, and CH2Cl2, NEt3, DMSO, MeOH, and CH3CN were distilled from CaH2.All other reagents were used without purification unless otherwise stated.All reactions were run under an atmosphere of nitrogen.NMR spectra were recorded on the VXL-300, Unity-300, VXR-500 or Inova-500 spectrometers.Chemical shifts were reported in δ, part per million (ppm), relative to chloroform (δ = 7.24 ppm) or CH2Cl2 (δ = 5.26 ppm) as an internal standard unless otherwise stated.Mass Spectra were recorded at the Mass Spectrometry facility in the Department of Chemistry at the University of Utah.IR spectra were recorded on a Nicolet Impact 400.
To a solution of the formamide from above in CH2Cl2 (4.5 mL) at 0 ºC was added a solution of PhSCl (430 mg, 2.9 mmol) in CH2Cl2 (2 mL) dropwise via a syringe pump over 1 h.

Synthesis of ()-debromo-dihydroflustramine C 8
To a stirring solution of 31 (2.8 mg, 0.0070 mmol) in THF (0.5 mL) and NH3 (5 mL) at -78 ºC was added Na metal until the mixture turned dark blue.At this point the reaction mixture was allowed to stir for an additional 10 min.Solid NH4Cl was added to the reaction mixture, and it was warmed to RT and diluted with H2O (5 mL).The aqueous phase was extracted with ethyl acetate (3x10 mL), the extracts were dried (Na2SO4) and concentrated.Purification by flash column chromatography (6:1 CH2Cl2:MeOH) afforded 1.5 mg (85%) of 8 as a clear oil. 1 H-NMR spectra and LRMS data were in good agreement with Aiko and Wright's data.