An aryl radical cyclisation approach to highly substituted oxindoles related to mitomycins

Aryl radical cyclisation of anilides 14 and 25 leads to highly substituted oxindoles 15 and 26 respectively in excellent yields. Oxindole 26 possesses the correct substitution pattern for the A-ring of mitomycin A.


Introduction
The mitomycins are a family of molecules that exhibit potent antibiotic and cytotoxic properties, with mitomycin C being used clinically in the treatment of adenocarcinomas of the stomach, pancreas and the colon.It is also used for the treatment of a range of other carcinomas 1 .Since the isolation of mitomycins A and B by Hata in 1956 2 , mitomycin C by Wakaki in 1958 3 (Fig. 1) and a range of structurally related mitomycins in more recent years, 4 there has been a great deal of interest in their synthesis. 5However, owing to the synthetic challenges associated with these molecules, namely the dense functionality, chemical lability, and the stereochemical problems there have been only two successful syntheses to date.The first of these by Kishi 6 involved the formation of an eight-membered ring system by an intramolecular Michael reaction followed by a mercuric chloride trans-annular cyclisation to give the B and C rings.The second by Fukuyama 7 took advantage of the fact that isomitomycins can be readily converted to the mitomycins and these were therefore his primary synthetic target.
Our own interest in the mitomycin family of molecules has centred upon the use of oxindoles as pivotal intermediates from which the C-ring can be constructed.This strategy was first proposed by Raphael 8 who attempted (unsuccessfully) to add an acetylenic anion onto the carbonyl carbon of an oxindole to form the C-ring.Our approach relies upon nucleophilic attack on the carbonyl of an oxindole.We have reported two strategies.One involved the intermolecular addition of an organolithium to the carbonyl of an oxindole followed by a Mitsunobu cyclisation to construct the C-ring 9 (Figure 2).The second involved intramolecular addition of alkyl-and vinyllithiums to the carbonyl of an oxindole 10 (Fig. 3).Both these approaches proved successful for the formation of the pyrrolo[1,2-a] indolenine ring system, the core ring system found in the mitomycins.The use of aryl radical cyclisations to generate a range of simple oxindoles has previously been reported. 11An essential requirement of this synthetic plan is the availability of oxindoles carrying the correct A-ring substitution pattern.This provides a severe test of the radical cyclisation chemistry and we now wish to disclose in detail the successful conclusion of our efforts in this area. 12The methodology relies upon the generation of a suitable aniline 1, reaction of this to form an anilide 2 and subsequent radical cyclisation to form an oxindole 3 (Scheme 1).In order to eliminate the problems inherent in handling quinone ring systems we chose to use a protected bis-phenol, which offers the opportunity for deprotection and oxidation to the required quinone at a later stage.

Results and Discussion
Following the procedure reported by Raphael, 8 the tetrasubstituted benzene 7 (scheme 2) was obtained from 2-methylresorcinol 4 in three steps.Friedel-Crafts acylation of 4 to give 5 was followed by selective benzylation of the non-hydrogen bonded hydroxyl of 5 to give 6 and the remaining hydroxyl was protected as the methyl ether in an overall yield for the three steps of 69% after recrystallisation.Catalytic hydrogenation in ethanol over palladium on carbon to remove the benzyl ether gave 8 in 99% yield, in which not only debenzylation but reduction of the benzylic ketone had occurred.However, this provided a suitable substrate to test our synthetic sequence and subsequently the final radical cyclisation.Nitration of 8 was achieved regiospecifically using concentrated nitric acid in glacial acetic acid to give 9 in 86% yield.Benzylation of 9 was simply achieved by treatment with potassium carbonate and benzyl bromide in acetone at reflux to give 10 in 93% yield.Attempts to introduce bromine into the last remaining site of the aryl ring proved problematic.Bromine in dichloromethane gave rise to benzylic bromination at the ethyl group.Bromination in acetic acid with and without iron catalysis also failed to give any of the desired product.It seemed likely that further activation of the aromatic ring would be necessary.Clearly the deactivating effect of the nitro group is sufficient to prevent the expected aromatic substitution reaction.To this end the nitro group was reduced using Bellamy's procedure 13 with tin (II) chloride in ethanol to give 11 in quantitative yield. 14reatment of 11 with bromine in acetic acid gave 12 in 54% yield, however the yield of this reaction was improved to 92% by the use of pyridinium hydrobromide perbromide 15 as the source of bromine.The conversion of the hexasubstituted aromatic 12 into the oxindole 15 proved straightforward.Aniline 12 was treated with acryloyl chloride in ether to give anilide 13 in 96% yield.In order that the molecule adopts the correct orientation for cyclisation, substitution of the amide nitrogen is essential. 16Anilide 13 was therefore alkylated on nitrogen using methyl iodide in the presence of potassium hydride in THF to give 14 in 90% yield.Cyclisation of 14 using tributyltin hydride, at a concentration of 0.012 M in toluene at reflux with AIBN as initiator gave the desired oxindole 15 in 76% yield after column chromatography.

Scheme 2
Having established that the radical cyclisation of an anilide bearing a highly functionalised aromatic ring was feasible and proceeded in good yield, attention was turned to the synthesis of an anilide system more closely resembling that required for the A ring of the mitomycins.Once again the quinone functionality was protected with the intention of unmasking it at a later stage in the synthesis.Owing to their ease of removal, methoxymethyl (MOM) ethers were the first choice of protecting group.Careful hydrogenolysis of 7 with continuous monitoring enabled us to isolate 16 in 84% yield after two hours with no trace of reduction of the ketone.This was nitrated using concentrated nitric acid in glacial acetic acid to give 17 in 74% yield after recrystallisation from ethanol.Baeyer-Villiger oxidation of 17 in dichloromethane with mCPBA gave the dihydroquinone 18 8 in 78% yield which was converted to its bismethoxymethyl ether derivative 19 8 in 78% yield by treatment with N,N-diisopropylethylamine and chloromethyl methyl ether in dichloromethane.Reduction of the nitro group was then attempted.Unfortunately using the conditions previously used for 10, namely stannous chloride in ethanol resulted in cleavage of the MOM ethers to give 18 in near quantitative yield.However, the desired transformation to give 20 was achieved quantitatively by catalytic hydrogenation in ethyl acetate using palladium on carbon as catalyst.Bromination was attempted using pyridinium hydrobromide perbromide in acetic acid as for 12 but no identifiable products were isolated.Clearly, acidic conditions had to be avoided and the reaction was repeated in dichloromethane containing N,N-diisopropylethylamine. Once again none of the desired product was isolated.In order to alleviate the potential problem of oxidative decomposition of the aromatic ring due to its high nucleophilicity, bromination of the nitro aromatic 18 was investigated.Once again under a variety of conditions none of the desired product was obtained.
These results forced a reconsideration of the protection protocol, and the use of methyl ether protecting groups was chosen since oxidation to the quinone required for the A ring of the mitomycins has been reported previously. 17Protection of 18 was achieved in nearly quantitative yield using potassium carbonate and dimethylsulfate in acetone to give 21. 14he nitro group of 21 was reduced by catalytic hydrogenation in ethyl acetate using palladium on carbon catalyst to give 22 in 93% yield. 14Bromination was straightforward using pyridinium hydrobromide perbromide in dichloromethane and pyridine to give 23 in 70% yield.Acylation with acryloyl chloride gave 24 in 95% yield followed by Nmethylation using sodium hydride and methyl iodide to give 25 in 95% yield.The radical cyclisation of 25 proceeded smoothly to give the oxindole 26 in 97% yield after chromatography.
In summary, we have demonstrated the feasibility of constructing an oxindole containing all the functionality required for ring-A of the mitomycins using an aryl radical cyclisation.General Procedures.NMR spectra were recorded on a Bruker AM360 or AM250 spectrometer at 360 MHz and 250 MHz respectively using CDCl3 as solvent with SiMe4 as an internal standard, unless otherwise stated.J-values are given in Hz.IR spectra were recorded on a Perkin-Elmer 983G infrared spectrometer, using nujol mulls or carbon tetrachloride solutions unless otherwise stated.Mass spectral data were recorded on a Jeol AX505W with complement data system.Samples were ionised electronically at 70eV with typical accelerating voltage of 6 kV.Melting points were determined using a Kofler hot plate apparatus and are uncorrected.All column chromatography was carried out using the flash chromatography technique, using Merck 60 (230-400 mesh) silica gel.Analytical TLC was carried out on Merck plastic backed TLC plates, coated with silica gel 60 F-254.Plates were visualised using ultraviolet light, unless otherwise stated.Eluting solvent systems are stated where appropriate.All dry reactions were performed in an inert argon atmosphere using a vacuum-argon manifold for the exclusion of water.Stirring was by internal magnetic bead.All syringes, needles and glassware were pre-dried at 110°C and cooled in an anhydrous atmosphere before use.Diethyl ether, tetrahydrofuran (THF), and toluene were pre-dried over sodium wire and refluxed over sodium under argon with benzophenone as an indicator in the reaction vessel.Dichloromethane was refluxed under argon, over CaH2 and distilled directly into the reaction vessel.

4-Ethyl-3-methoxy-2-methyl-6-nitrophenol (9).
A solution of phenol 8 (32.0 g, 192 mmol) in acetic acid (100 ml) was treated with a solution of concentrated nitric acid (12 ml) in acetic acid (25 ml).Initially 5 ml of this solution was added and the reaction mixture was warmed to 40 °C.After 15 minutes the remaining nitric acid solution was slowly added.The reaction was stirred for a further 1 hour, then poured into water (500 ml), extracted with diethyl ether (3 x 300 ml), washed with brine (3 x 20 ml) and dried with magnesium sulfate.The ethereal solution was then treated with charcoal, filtered and the solvents removed at reduced pressure to leave a brown crystalline solid.This was recrystallised from hot ethanol and the mother liquor concentrated and purified by chromatography on silica using petroleum spirit (40 -60 °C) -ethyl acetate (10 : 1) as eluent to give nitrophenol 9 (34.