Synthesis of pyrrolo[1,2-a ]indole-1,8(5 H )-diones as new synthons for developing novel tricyclic compounds of pharmaceutical interest

In the course of our work aimed at developing novel heterocycles of pharmaceutical interest, a new tricycle, the tetrahydropyrrolo[1,2-a ]indole-1,8-dione, has been synthesized by an intramolecular Friedel-Crafts acylation, as a synthon suitable to be functionalized to give novel compounds with potential biological properties. Also, an unusual nucleophilic α -addition to methyl propiolate by 1,5,6,7-tetrahydro-4 H -indol-4-one was observed and discussed


Introduction
The production of heterocyclic compounds is dominating in the field of modern organic and medicinal chemistry.In particular the presence of heterocycles in drugs represents the majority of known pharmaceutical preparations.For these reasons our team has for a long time been interested in the synthesis of new heterocycles to be employed as new pharmaceutical agents.
2][3][4] More recently, our attention focused on the preparation of pyrrolo[g]indoles from which the present investigation began.The observation that the bis adduct of 4,5,6,7-tetrahydroindole-4-ketoxime with methyl propiolate (MP) 1 underwent thermal retro-Michael addition of MP to give E-2 in 22% yield 5 prompted us to find a new procedure for the preparation of 2 (Figure 1).This intermediate has been first considered a key synthon for a conversion into pyrrolo [1,2-a]indole-1,8(5H)-dione 3a which, by reduction of its 2,3-double bond, could be converted into 3b (Figure 1), both potential substrates for drug design.In fact this three-ring system possesses a peculiar feature to be purposely functionalized via the two oxo groups into possible intercalating agents 6 which are in our mind.The endeavours to obtain this skeleton are the object of this communication.

Chemistry
An overall view of the attempts made and the synthetic routes employed to prepare the targets tricycles 3a,b are presented in Schemes 1-3.Attempted preparation of 2 by carrying out a Michael addition of 4 7 on methyl propiolate (MP) in dichloromethane and in the presence of triphenylphosphine (TPP), adapting a method previously described by other authors, 8 failed.This failure was surprising since it is well known that the addition of electron poor nucleophiles to activated olefins is strongly dependent on the base.0][11] Thus, in contrast to the expected acrylate 2, we obtained regioselectively the vinyl ester 5 in 78% yield (Scheme 1).The structure of 5 was unambiguously assigned on the basis of its elemental analysis and EI-mass as well as by the data of 1 H-and 13 C-NMR experiments.Its 1 H-NMR spectrum was recorded in various organic solvents in order to evaluate a possible variation of the chemical shifts of olefinic protons.In particular, at 200 MHz in CDCl 3 it exhibited two singlets at δ H = 6.54 and 5.88 while the corresponding spectra registered in DMSO-d 6 and acetone-d 6 revealed two doublets at δ H = 6.52 and 6.12 and at δ H = 6.53 and 6.04, respectively.Both splitting patterns for the methylene protons appeared weakly coupled with a constant of 3 2).In fact, according to our prediction, no NOE interactions between H-2 and H B protons could be observed for compound 5.
In order to circumvent this unexpected behavior, we performed the reaction in DMSO using TEA as catalyst, that afforded stereoselectively 2 as E-isomer in 36% yield (Scheme 1).The chemical and spectroscopic properties of this compound were coincident with those previously reported by us.In the light of this result, it was interesting to postulate a mechanism for the formation of 5. Previous studies 8,9 described an initial attack of TPP to the β-carbon of MP to give the phosphonium enolate intermediate 6 and concomitant protonation of this 1:1 adduct to form the corresponding phosphonium salt 7.In a second step a β-addition by the nucleophile anion takes place followed by elimination of the phosphine, to be recycled as a catalyst, to afford the βacrylates.In the our case, although the nucleophilic β-addition was potentially most favourable for the major electrophilicity of the β-carbon of the adduct 7, we observed that the conjugate addition of 4 proceeded through an unusual Michael-type α-addition to the βtriphenylphosphonium acrylate counterpart.A possible explanation of this result may be due to steric hindrance of the phosphine moiety in proximity of the β-carbon of the adduct and consequently the anion 4 would attack the more free α-carbon.This assumption could be supported from the fact that compound 2 was easily obtained following other reaction conditions.Although the propenoate 2 has been obtained only in E-configuration we thought that an intramolecular cyclization of the opportune more electrophilic acid derivative could easily lead to 3a.With this in mind we saponified 13 the ester group of E-2 to obtain the acid E-8 that might undergo ring closure in the presence of polyphosphoric acid (PPA) [14][15][16] or PCl 5 17 (Scheme 2).
Unfortunately, probably due to the unfavourable stereochemistry of the precursor, this reaction failed, preventing the expected hydrogenation of the olefinic double bond to obtain the analogue 3b.This fact suggested that we must obtain 3b by another synthetic route.Then, we alkylated 4 with acrylonitrile in dioxane and in the presence of triton B 18,19 to afford 9 in 74% yield.The latter on hydrolysis 20 gave the acid 10 that by an intramolecular Friedel-Crafts cyclization in PPA at 90 °C gave the desired diketone 3b in 85% yield (Scheme 3), which was characterized by several NMR methods ( 1 H-NMR, 13 C-NMR, DEPT/APT, HETCOR).At this stage we explored the reactivity of these carbonyl functions by converting 3b into the diketoxime 11 (Scheme 4) and alternatively we investigated the possibility of a selective reduction of one of the two carbonyl of 3b into CH 2 in order to functionalize the derived monoketone.Thus, we attempted a modified Wolff-Kishner reaction 21 on 3b and found that it was regioselectively reduced to give the ketone 12 (Scheme 4).

(Figure 3 .
Figure 3. Proposed mechanism for the formation of 5.
12T spectra.This value for a methylene in the β-position of a vinyl ester was consistent with the predictions of Kalinowski et al.12that locate the CH 2 chemical shift at about 125 ppm.Besides 13 C-NMR TOTAL COUPLING further confirmed the presence of the CH 2 showing a double doublet centered at 124.8 ppm.Moreover, NOESY and NOE difference experiments showed NOE correlations between the aromatic H-2 and H A -hydrogen of the methylene in β-position.These experiments distinguished the two methylene protons (Figure Reagents and conditions: (i) LiOH•H 2 O (4 eq.), H 2 O/THF, r.t. for 1 h, then 2N HCl.(ii) a] 84% PPA, 90 °C for 12 h or b] PCl 5 , anhydr.CH 2 Cl 2 , r.t. for 12 h.