Stereoselective synthesis of 2,6-trans-4-oxopiperidines using an acid-mediated 6-endo-trig cyclisation

An acid-mediated 6-endo-trig cyclisation of amine-substituted enones has been developed for the stereoselective synthesis of trans-6-alkyl-2-methyl-4-oxopiperidines. Performed under conditions that prevent removal of the Boc-protecting group or acetal formation, the key cyclisation was found to generate cleanly the 4-oxopiperidine products in high overall yields from a wide range of alkyl substituted enones. The synthetic utility of the trans-6-alkyl-2-methyl-4-oxopiperidines formed from this process was demonstrated with the total synthesis of the quinolizidine alkaloid, (+)-myrtine and the piperidine alkaloid, (-)-solenopsin A.


Introduction
Substituted piperidines are key structural units found as components of a wide range of natural products and pharmaceutically active compounds. 1 Within this structural class, compounds bearing a 2,6-trans-dialkylpiperidine core have been isolated from a wide range of sources and found to have important biological and medicinal properties. 1 These include (+)-myrtine (1), 2 a quinolizidine alkaloid from Vaccinium myrtillus (Ericaceae) and the phenylquinolizidine alkaloid (−)-lasubine I (2), 3 isolated from the leaves of Lagerstroemia subcostata (Fig. 1). Natural products bearing just the 2,6-trans-dialkylpiperidine ring include (−)-solenopsin A (3), a component of the venom of the fire ant Solenopsis invicta. 4 As well as being an inhibitor of phosphatidylinositol-3-kinase signalling and angiogenesis, 5 (−)-solenopsin A (3), has also been shown to display ceramide-like biological activity with the inhibition of the functional Akt ( protein kinase B) pathway. 6 Another example is (+)-prosopinine (4), a 2,6-trans-dialkylpiperidine alkaloid from Prosopsis africana, 7 that has antihypertensive properties and antibiotic activity against Staphyllococcus aureus. 8 As a result of their interesting and varied structures and their significant medicinal properties, a variety of strategies for the total syntheses of 2,6-trans-dialkylpiperidine natural products have been reported. 9,10 One approach in accessing the 2,6-trans-dialkylpiperidine ring system is the 6-endo-trig cyclisation 11 of amine-substituted enones. 12 In 2011, we reported a one-pot, reductive amination/ 6-endo-trig cyclisation that allowed the synthesis of 2,6-trans-4oxopipecolic acids. 13 A Zimmerman-Traxler chair-like transition state in which both the side-chain and N-substituent occupy a pseudoequatorial position was used to rationalise the stereochemical outcome of the kinetically-controlled cyclisation (Scheme 1a). Troin and co-workers also described the general synthesis of 2,6-trans-dialkylpiperidines by the intramolecular Michael-type reaction of carbamate-protected amine-substituted enones using a combination of p-tosic acid, ethylene glycol and trimethyl orthoformate (Scheme 1b). 14 This produced the ketal derivatives of the 2,6-trans-dialkyl-4- Fig. 1 Structures of 2,6-trans-disubstituted piperidine-containing natural products.
Enone 10a was then treated with an ethereal solution of 2 M hydrochloric acid in methanol at room temperature (20°C). After 1 h, TLC showed the formation of two new compounds. Characterisation by 1 H NMR spectroscopy identified these as the 2,6-trans-and 2,6-cis-4-oxopiperidines (11a and 12a), formed in a combined 58% yield and a 3 : 1 ratio, respectively ( Table 1, entry 1). Despite the acidic conditions, only trace amounts (<5%) of the deprotected amine or the oxopiperidine derived ketals could be detected by NMR spectroscopy. To explore the limits of the acid mediated 6-endo-trig cyclisation, the reaction was repeated with varying amounts of acid. Using an excess (5 equiv.), quickly led to decomposition of the starting material and the isolation of no major products (entry 2), while using 50 mol% resulted in an incomplete reaction even after 3 h and isolation of 11a and 12a in only 40% yield (entry 3). The temperature of the transformation was then investigated in attempt to improve both the diastereoselectivity and the overall yield. Although cooling the reaction to either 0 or −78°C and warming to room temperature over 2 h did result in cleaner transformations and significantly improved yields (entries 4 and 5), both reactions again gave 11a and 12a in a 3 : 1 ratio, respectively. Nevertheless, with the improved efficiency of the cyclisation, and the straightforward separation of the two cyclic products by column chromatography, this allowed the isolation of the major diastereomer 11a in 49% yield.
To provide some insight into the mechanism of the acidmediated 6-endo-trig cyclisation, some control experiments were performed. Initially, the cyclisation of enone 10a over a 24 h period (Fig. 2) was analysed by 1 H NMR spectroscopy. As observed in the preparative reaction, the transformation was complete after 2 h, giving 2,6-trans-11a and 2,6-cis-12a in a 76 : 24 ratio. After 2 h, a steady decrease in the amount of 11a and a concurrent increase in 12a was observed, with a final ratio of 67 : 33 in favour of 2,6-cis-product 12a (see ESI † for 1 H NMR spectra for 2-24 h). Extending the reaction time to 40 h, showed no appreciable change in ratio (69 : 31). An additional experiment was performed involving resubmission of diastereomerically pure 2,6-trans-product 11a to the acid-mediated cyclisation conditions. Over a 6 h period, 1 H NMR analysis of the reaction mixture showed increasing epimerisation of 11a with a final 68 : 32 ratio of 11a and 12a, respectively. These results taken together suggest that the 2,6-trans-4-oxopiperidine is the less stable, kinetic diastereomer and thus, is formed first during the thermodynamically controlled 6-endotrig cyclisation. On extension of the reaction time and in the presence of acid, a likely retro-conjugate addition reaction results in conversion of the 2,6-trans-4-oxopiperidine to the configurationally more stable 2,6-cis-isomer. 21,22 Therefore, to maximise selective formation of the target 2,6-trans-isomers, short reaction times should be used. While a reaction time of less than 2 h would likely lead to a more selective cyclisation, the optimisation studies have shown that a 2 h reaction time is required for full conversion. Therefore, in terms of maximising the isolation of the target 2,6-trans-isomers, a 2 h reaction is a good compromise.
Following optimisation of the acid-mediated cyclisation and with some understanding of the mechanism, the scope of the process was next investigated (Scheme 4). The 6-endo-trig cyclisation of a range of enones 10a-10f bearing n-alkyl sidechains typically present in piperidine alkaloid natural products were found to be excellent substrates for this transformation, giving a mixture of the 2,6-trans-and 2,6-cis-4-oxopiperidines (11 and 12) in high yields (73-88%). Inspection of the 1 H NMR spectra from the crude reaction mixtures showed that the 2,6trans-4-oxopiperidine 11 was the major product in each case with a diastereomeric ratio of approximately 2.5 : 1 to 3 : 1. Formation of 2,6-trans-4-oxopiperidines 11a-11f as the major diastereomers during the 2 h acid mediated 6-endo-trig cyclisa-  tion and in similar ratios irrespective of the size of the n-alkyl side-chain provides further evidence that these are the kinetic products from this process. For each case, the 2,6-trans-4-oxopiperidines 11a-11f were separated by column chromatography, allowing the isolation of these compounds in 47-55% yields. Further examination of the scope of the cyclisation with enones bearing functional groups (10g) and incorporating rings (10h and 10i), gave the corresponding 2,6-trans-4-oxopiperidines (11g-11i) in similar diastereomeric ratios and isolated yields. Limitations of the cyclisation were observed. For example, enones 10j and 10k with a branched side-chain adjacent to the alkene moiety showed no diastereoselectivity. While the cyclisation of 10j with a sec-butyl side-chain still resulted in an efficient cyclisation and the isolation of each diastereomer in 40% yield (80% overall), the cyclisation of the iso-propyl analogue 10k was impeded, generating the 2,6-transand 2,6-cis-4-oxopiperidines 11k/12k in only 54% yield. The other limitation discovered during this study is that enones with a conjugated aryl side-chain (e.g. 10l) are not cyclised under the mild conditions of this transformation.
Having developed a new approach for the synthesis of 2,6-trans-dialkyl-4-oxopiperidines, the synthetic utility of this transformation for the facile preparation of piperidine based natural products was next investigated (Scheme 5). Initially, 2,6-trans-4-oxopiperidine 11g bearing a 4-chlorobutyl sidechain was converted in two steps to the quinolizidine alkaloid, (+)-myrtine (1). While the N-Boc protecting group of 11g was stable under the anhydrous acidic conditions of the cyclisation reaction, facile removal was achieved under aqueous acidic conditions. This was followed by cyclisation of the resulting amine with the 4-chlorobutyl side-chain under basic conditions, which gave (+)-myrtine (1) in 62% yield over the two steps. 2,6-trans-4-Oxopiperidine 11f was used for a two-step synthesis of the piperidine alkaloid, (−)-solenopsin A (3). Treatment of 11f with 1,3-propanedithiol under Lewis acid conditions resulted in simultaneous formation of the cyclic dithioketal and removal of the Boc-protecting group. This gave 1,5-dithia-9-azaspiro [5.5]undecane 13 in 90% yield. Reduction of the dithioketal using RANEY® nickel then gave (−)-solenopsin A (3) in 77% yield. The spectroscopic data and optical rotations of 1 and 3 generated in this study were entirely consistent with literature data. 10b,14b

Conclusions
In summary, an acid-mediated 6-endo-trig cyclisation of aminesubstituted enones has been optimised for the stereoselective synthesis of 2,6-trans-6-alkyl-2-methyl-4-oxopiperidines. Mechanistic studies have shown that the 2,6-trans-4-oxopiperidine is the kinetic product, that is then converted over time via a possible retro-conjugate addition reaction to the configurationally more stable 2,6-cis-isomer. Therefore, using a short reaction time, the scope of the process was explored with a range of amine-substituted enones bearing alkyl side-chains, resulting in the preparation of a small library of 2,6-trans-6alkyl-2-methyl-4-oxopiperidines. Despite the acidic conditions, neither Boc-group removal or acetal formation was observed, with the Boc-protected 4-oxopiperidines isolated as the major products in 54-95% yield. The potential of the 2,6-trans-4-oxopiperidines as building blocks was demonstrated with the twostep preparation of the quinolizidine alkaloid, (+)-myrtine (1) and the piperidine alkaloid, (−)-solenopsin A (3). Work is currently underway to investigate further synthetic applications of this acid-mediated 6-endo-trig cyclisation reaction.

Experimental
All reagents and starting materials were obtained from commercial sources and used as received. All dry solvents were purified using a solvent purification system. All reactions were performed in oven-dried glassware under an atmosphere of argon unless otherwise stated. Brine refers to a saturated solution of sodium chloride. Flash column chromatography was performed using silica gel 60 (40-63 μm). Aluminium-backed plates pre-coated with silica gel 60F 254 were used for thin layer chromatography and were visualised with a UV lamp or by staining with potassium permanganate. 1 H NMR spectra were recorded on a NMR spectrometer at either 400 or 500 MHz and data are reported as follows: chemical shift in ppm relative to tetramethylsilane or the solvent as the internal standard (CDCl 3 , δ 7.26 ppm), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet or overlap of nonequivalent resonances, integration). 13 C NMR spectra were recorded on a NMR spectrometer at either 101 or 126 MHz and data are reported as follows: chemical shift in ppm relative to tetramethylsilane or the solvent as internal standard (CDCl 3 , δ 77.0 ppm), multiplicity with respect to hydrogen (deduced from DEPT experiments, C, CH, CH 2 or CH 3 ). IR spectra were recorded on a FTIR spectrometer; wavenumbers are indicated in cm −1 . Mass spectra were recorded using electrospray or electron impact techniques. HRMS spectra were recorded using a dual-focusing magnetic analyser mass spectrometer. Melting points are uncorrected. Optical rotations were determined as solutions irradiating with the sodium D line (λ = 589 nm) using a polarimeter.

Conflicts of interest
There are no conflicts to declare.