Functionalization of N -tosyl-2-pyridinone with silyl ketene acetal catalyzed by Lewis acid, and synthetic studies of corynantheidol

The reaction of N -tosyl-2-pyridinone with tert -butyldimethylsilyl ketene acetal, catalyzed by a Lewis acid, has been investigated. While both the C-4 Michael adduct 3 and the bicyclic compound 5 were isolated from aluminum chloride-or trimethylaluminum-catalyzed reactions, the former was afforded as the sole product from diethylaluminum chloride-or trimethylsilyl trifluoromethanesulfonate-catalyzed reactions. During investigation of a synthesis of corynantheidol, a method for construction of the key intermediate 30 was accomplished via , (1), the efficient conversion of the enamine moiety to an alkenyl bromide ( 20 (cid:198) 22 ), followed by the Sonogashira coupling reduction ( 22 (cid:198) 23 ); (2), selective reduction of the acetylene and intramolecular cyclization ( 25 (cid:198) 26 ); (3), the coupling reaction with the indole unit ( 27 (cid:198) 30 ).


Introduction
Owing to their valuable biological activities, nitrogen-containing compounds have long been of interest to organic chemists and biochemists and are present in many types of pharmaceuticals.In particular, six-membered ring systems having nitrogen atoms in their rings are found abundantly in nature as substructure(s) of alkaloids or aza-sugars. 1 Therefore, many efforts had been made over recent decades to construct such ring systems and, in particular, to develop enantio-and/or diastereo-selective synthetic methods for functionalized piperidine rings. 2 The stereoselective construction of multi-functional piperidine ring systems, the utilization of pyridinium salts or dihydropyridine derivatives has been well investigated recently, 2,3 including the Mukaiyama-Michael addition reaction, 4,5 alkylation reactions, 6,7 and reductive alkylation reactions. 8In these areas, elegant methods for the functionalization of pyridine rings using chiral 1-acylpyridinium salts as substrates have been developed by Comins' research group, 5,6 and the syntheses of many biologically active natural products were also reported. 9n comparison with the significant efforts in studies of the 4-alkoxypyridine derivatives, the researches involving the reactivity of the 2-oxypyridine (2-pyridinone) derivatives have mainly been limited to the Diels-Alder reactions 10,11 and, to the best of our knowledge, the introduction of the acetate units into oxygenated pyridine derivatives has been reported only once. 5 As part of our continuous program for the development of new methodologies for heterocyclic compounds, 12 we focused our attention on the utilization of the activated 2pyridinone derivatives.Herein, we report the regioselective functionalization of N-tosyl-2pyridinone 1 11,13 with tert-butyldimethylsilyl ketene acetal 2, activated by a Lewis acid, and its application in the synthetic study of corynantheidol 4 (Figure 1).

Results and Discussion
N-Tosyl-2-pyridinone, 1, was synthesized by reaction of the lithium salt of pyridinone with ptoluenesulfonyl chloride according to the reported procedure. 11The results of the reactions between 1 and the tert-butyldimethylsilyl ketene acetal 2 in the presence of various Lewis acids are summarized in Table 1.However, neither thermal conditions (toluene, 90°C, 110 h), nor treatment of 1 and 2 with zinc chloride or boron trifluoride diethyl etherate promoted any reactions, and 1 was completely recovered.Stannic chloride promoted only the migration of the tosyl group from the nitrogen-to oxygen atom to give 6 exclusively (Table 1: entry 1).However, when 1 was allowed to react with 2 in the presence of 10 mol % of aluminum chloride, the C-4 Michael adduct 3 was obtained, although the yield (22%) was disappointing (Table 1: entry 2).The yield of 3 was greatly improved by changing the catalyst to diethylaluminum chloride or trimethylaluminum: 3 was produced as the sole product by the former catalyst, and the unexpected bicyclic compound 5 was isolated in the latter case (Table 1: entries 3 and 4).The best condition tested used 10 mol % of tert-butyldimethylsilyl trifluoromethanesulfonate (TBSOTf) at -40°C, where 89% of 3 was isolated as the single product (Table 1: entry 5).The structure of 3 was established from spectra and by conversion into the δ-lactam derivatives 8 by catalytic hydrogenation.Although the bicyclic compound 5 was given as a single product, which was confirmed by the 13 C-NMR spectrum, the stereochemistry of the ethyl silyl acetal moiety was difficult to determine from the 1 H-NMR spectrum.Thus, the structure of 5 was confirmed by the NMR analysis ( 1 H, 13 C, and 1 H-COSY) of 9 which was obtained by hydrolysis of the acetal moiety (Scheme 1).When the reaction in the presence of 10 mol % of TBSOTf was terminated without treatment with H 2 O (evaporation of the solvent and the reagents), the silyl enol ether 7 was identified by 1 H-NMR.Therefore, our proposed reaction mechanisms catalyzed by TBSOTf may involve: (1) activation of the carbonyl group by the TBS group (1→10), (2) Michael addition at C4 (10→11), and (3) elimination of the TBS group to produce TBSOTf (11→7) as illustrated in Scheme 1.As the aluminum reagents for these reactions functioned as catalysts, the silyl group migration reaction (12→7) might participate in the catalytic cycle, although we do not have any evidence.On the other hand, for production of the bicyclic compound 5, it could also be that the silyl group's migration reaction was impossible for the intermediate 15, for steric reasons, and then the reaction might progress via the tandem Michael-Claisen type pathway, but not through the Diels-Alder process (Scheme 1).We have now established an efficient synthetic method for preparing 3 in large quantities, and our next aim is to synthesize natural products from 3.
Corynantheidol 4 was isolated and characterized by Shellard and Houghton in 1973. 14Total syntheses of this indole alkaloid were reported by several research groups, both as the racemate 15 and the optically active form. 16The Michael adduct 3 possesses the unique property that both C3 and C5 can be functionalized by an electrophile, although the oxidation states of the carbon atoms at the 2-and 6-positions are different from each other.Thus, we postulated that 3 should be able to convert into the D-ring system of corynantheidol 4 through two different routes according to the strategy mentioned above, namely, an alkylation reaction at the C3 position using the silyl enol ether moiety (A: 7→16→4) or at the C5 position by enamine alkylation (B: 7→17→4) (Scheme 2).We first tried the direct alkylation reaction of 7 under several reaction conditions (route A).However, both the usual alkylation of the silyl enol ether (dried TBAF and EtI in THF), and using silver acetate as an activator of the alkyl halide, were unsuccessful and only the hydrolyzed product 3 or the starting material 7 were recovered.Therefore, we decided to change the synthetic route from A to B.

Scheme 2
The selective reduction of the lactam carbonyl group of 3 by sodium borohydride in the presence of cerous chloride heptahydrate at -78 to -20°C produced the acetal 18 as inseparable diastereomeric mixtures.Formation of the acetal 18 under standard conditions (triethyl orthoformate, PPTS, EtOH, r.t.) gave two separable diastereomers 19 and 20 in 14 and 70% yield, respectively (Scheme 3).The stereochemistry of the major isomer 20 was determined by 1 H-NMR analysis.Between the possible two conformations, the conformer 20a in which the axial position is occupied by an ethoxy group might be more favored, owing to the anomeric effect.The signal of the C3-βH appeared at 0.73 ppm as a triple doublet (J = 12.7 and 2.4 Hz), which showed the typical coupling pattern for the axial proton.Also, the unusually high field shift of this proton in comparison with the usual methylene signals, which is the result of shielding by the lone pair of the sp 2 nitrogen atom, might also be evidence for this conformation (Figure 2).Although both 19 and 20 seemed to be used for the synthesis of corynantheidol 4, the major isomer 20 was used for the following sequences.Because our efforts at alkylation at the C5 position were also all unsuccessful, we next planned to introduce the C2-unit by palladium-complex-catalyzed coupling between an alkenyl halide and acetylene.In order to introduce halogen atom at C5, 20 was reacted with NBS in THF-H 2 O (2:1) to afford the desired bromohydrin 21.The dehydration of 21 was successful using thionyl chloride and piperidine in DMF, giving the alkenyl bromide 22, and setting the stage for the coupling reaction (Scheme 3).Unfortunately, the typical reaction conditions for the Sonogashira coupling reaction 17 (trimethylsilyl)acetylene, PdCl 2 (Ph 3 P) 2 , Et 3 N, CuI), and also those using other palladium complexes [e.g., Pd(OAc) 2 , Pd 2 (dba) 2 , PdCl 2 (MeCN) 2 ] were found to be ineffective.However, when tetraphenylphosphonium chloride was added to the reaction mixture, the yield of 23 was greatly improved. 18Finally, the best yield (89%) was recorded when PdCl 2 (MeCN) 2 , N,Ndiisopropylamine, and tetraphenylphosphonium chloride were used as the catalyst, the base, and the additive, respectively, as shown in Scheme 4 (22→23).The trimethylsilyl group was removed by tetrabutylammonium fluoride in THF (93%), followed by reduction of the ester group with lithium aluminum hydride, to afford the alcohol 25 in a reasonable overall yield.Next, we planned the reduction of both the ethynyl group and internal enamine moiety at the same time.Thus, the alcohol 25 was subjected to catalytic hydrogenation (Pd-C in chloroform) to afford the bicyclic compound 26 in 79% yield.The production of the bicyclic compound 26 was rather convenient for us, because we expected that the reducing reagents would attack the enamine moiety from the convex (α−face) to afford the cis stereochemistry between C-4 and C-5 which corresponds to the D-ring system of corynantheidol 4. Surprisingly, the internal olefin resisted catalytic hydrogenation using Pd-C, Pd(OH) 2 , or PtO 2 as the catalyst, even when the hydrogen pressure was increased to 5 Kg/cm 2 .Therefore, we changed the route with the removal of the tosyl group first, then reduction of the double bond or coupling with the indole part.The tosyl group of 26 was easily removed by stirring with magnesium in methanol at 30°C. 19The enamine 27 was found to be unstable, so it was subjected to catalytic hydrogenation or sodium borohydride reduction without purification.However, we also dropped this route due to only decomposition of the enamine was observed under these reaction conditions.Finally, the crude enamine 27 was reacted with 3-indoleacetyl chloride 29 20

Conclusions
The efficient method for functionalizing N-tosyl-2-pyridinone 1 was established by the reaction with tert-butyldimethylsilyl ketene acetal 2 catalyzed by several kinds of Lewis acids.Due to the reactivities of the C-4 Michael adduct 3 or 7 were quite different from the usual enamines or silyl enol ethers, we could not perform direct syntheses from these compounds.However, the introduction of the C2 unit at the desired position was carried out by using a palladium catalyzed reaction.Although we have not finished our synthesis of corynantheidol 4, we are now able to produce the key intermediate for total synthesis by 11 steps from 1.The total synthesis of corynantheidol 4 and related natural products are now under way in our laboratory.

Experimental Section
General Procedures.All melting points were determined with Yazawa Micro Melting Point BY-2 and are uncorrected. 1H NMR spectra (400 or 500 MHz) and 13 C-NMR spectra (100 or 125 MHz) were recorded on JEOL JMN AL-400 or JEOL GX-500 spectrometers, respectively. 1H-NMR spectra (300 MHz) were measured with a Varian Gemini 2000.Chemical shifts (δ) are given from TMS (0 ppm) as internal standard for 1 H-NMR, and 13 CDCl 3 (77.0ppm) for 13 C-NMR.Mass spectra and high resolution mass spectra were measured on JEOL JMS-DX303 and MS-AX500 instruments, respectively.IR spectra were recorded on a Shimadzu FTIR-8400."RT" denotes room temperature.
and pyridine to afford the key intermediate 30 for the corynantheidol synthesis.The total synthesis of corynantheidol and the chiral version of the Michael addition reaction are now in progress in our laboratory.