Alkylation studies of a polyhydroxylated-cyano-piperidine scaffold

Hexahydro-3-phenyl-6,7,8-trihydroxy-3 R -  3  ,5  ,6  ,7  ,8  ,8a  -5 H -oxazolo  3,2-a  pyridine-5-carbonitrile is endowed with two non equivalent reactive sites: an  -amino nitrile at the C-5 position and an  -amino ether at the C-8a position. Herein, alkylation at the C-5 position was studied. The scope and limitations of these reactions have been investigated


Figure 1
Chemists have been inspired by the numerous pharmacological activities of iminosugars, but unfortunately some of them are also toxic to human cells.Nevertheless, there is still a need for an easy access to alkaloid analogs and for the development of more potent, selective and less toxic drug candidates.In this context, we decided to consider the chiral building block 1 (Figure 1) as starting material. 25As a part of our ongoing research program on the reactivity of compound 1, we already reported a rapid access to substituted six-or seven-membered ring iminosugars via ring-expansion reactions 26 and a new asymmetric synthesis of (2S,3R,4R,5S)trihydroxypipecolic acid. 27In order to prepare indolizidine analogs such as castanospermine derivatives, we have extended, herein, our research on the alkylation study at the C-5 position of the building block 1 (Scheme 1).Compound 1 possesses two non-equivalent reactive sites on the polyhydroxylated piperidine ring system: an -amino nitrile at the C-5 position and an -amino ether at the C-8a position.

Results and Discussion
In this paper, we focus on the methylation reactions at the C-5 position.The scope and limitations of this alkylation are discussed.Our purpose is to compare the reactivity of our building block 1 with regard to the CN(R,S) synthon created by Husson's group (Scheme 2), under the same experimental conditions. 28Following the literature data, 29 we studied the methylation of 1 with iodomethane at -78 °C in THF in presence of lithium diisopropylamine.First attempts were carried out on compound 1 without success.Therefore, it was decided to use the protected form, which was obtained by benzylation of the hydroxyl functions (Scheme 2). 26n contrast with the CN(R,S) synthon, the treatment of 2 with LDA in the presence of iodomethane gave two different products: the desired C5-alkylated compound 3 and an cyanoenamine 4, in 10% and 30% yield, respectively.The major compound 4 was obtained through the elimination of the adjacent benzyloxy group.A careful analysis of 4 by 1 H-and 13 C-NMR allowed the characterization of its structure.The typical NMR signals at  = 5.29 ppm and at  = 112.5 ppm, observed in the 1 H and 13 C NMR spectra respectively, were attributed to the C-6 position.

Scheme 2
As shown above, the -elimination that took place was an obvious limitation for the access to the desired C-5-alkylated derivatives.This prompted us to change the protecting group for the hydroxyl function at the C-6 position to butane 2,3-bisacetal.This bisacetal was previously applied to various sugars [30][31][32] or shikimic derivatives [33][34][35][36] and in spite of the creation of two new stereogenic centers, the protection of vicinal diequatorial diols is described as being stereoselective.  Thusthe protection of hydroxyl functions of 1 with 2,2,3,3tetramethoxybutane, through stirring for a week in the presence of trimethylorthoformate and catalytic camphorsulfonic acid, has been performed.A 1:2 mixture (NMR determination) of the 6,7-butane-bisacetal 5 and its regioisomer 6 (Scheme 3) were obtained.

Scheme 3
Interestingly, an unexpected mixture of diastereoisomers on the butane 2,3-bisacetal protection for compounds 5 and 6, is formed.After crystallizing the diastereoisomers 6 in diethyl ether solution, this observation was confirmed by an X-Ray study.The stereochemistry of the protecting group is represented in Figure 2.

Figure 2
Structural features of interest of 6a and 6b, which crystallizes in the non-centrosymmetric P21 21 21 space group, are collected in Table 1.The protecting group of 6a favors a twist boat conformation while a chair conformation is preferred for 6b (Figure 2). 1 H NMR studies indicated that after one week of reaction, compounds 6 were present in a mixture 80/20 with prevalence for the twist boat conformation of 6a.After two weeks of reaction, compounds 6 were still a mixture 80/20, but now with prevalence for the chair conformation of 6b.This result suggests a kinetic and a thermodynamic form, with an equilibration to the more stable derivative under the acidic experimental conditions.To the best of our knowledge, no report exists for the characterization of the twist boat conformation with this stereochemistry, for the butane 2,3-bisacetal protection of vicinal diequatorial diols.
To pursue our work on the alkylation, the protection of compounds 5 and 6 on their free hydroxyl function was achieved by reacting with methoxymethyl chloride (MOMCl) in the presence of Hünig's base (Scheme 3).The structures of 7 and 8 were unambiguously deduced from their spectral data.Indeed, in the HMBC spectrum of compound 7 the cross peaks, observed between the triplet signal at  = 3.77 of H-8 proton and carbon signals at  = 91.7 and  = 95.9, were attributed to C-8a and MOM methylene C-1' positions (Figure 3).Based on the HMBC correlation between H-6 and MOM methylene carbon (Figure 3), the structure of 8 was in full agreement with the proposed skeleton.

Figure 3
Initial attempts towards the stereoselective alkylation at the C-5 position of compounds 7 (mixture 50/50 of 7a and 7b confirmed by 1 H NMR), under classic conditions (LDA at -78 °C), were unsuccessful and only starting material was recovered.In contrast, when HMPA was added to the reaction and the temperature was gradually raised up to -15 °C, lactams 9 were obtained in 60% yield as confirmed by NMR and MS spectroscopy (δ 162.7 (CO); MS [M + H] + 424) and represented in Scheme 4. According to the literature, 38 the lactam formation can be explained by in situ addition of oxygen from the solvent to an extremely oxygen sensitive anion.

Scheme 4
This result showed that the anion formation occurred and that the 6,7-butane-bisacetal protection prevented the -elimination.However, the alkylation reaction probably failed because of steric hindrance.In order to confirm this last hypothesis, we investigated the alkylation study on the less hindered 7,8-butane-bisacetals 8 (Scheme 5).Alkylation of 8 with iodomethane (LDA, THF, -78 °C) led to the formation of the unexpected -methylated--cyanoenamines 10 (Scheme 5).These compounds were isolated in 77% yield after flash chromatography.

Scheme 5
These results suggested that the alkylation at -aminonitrile position was strongly dependant on the adjacent substituent.In the presence of a leaving group, the enamine formation is favored, whereas the presence of a sterically bulky group leads to the lactam derivative.Finally, we chose to use the non protected compounds 6 as a starting material, as seen in Scheme 6.

Scheme 6
Under similar conditions, compounds 6 are transformed into the corresponding alkylated compound 11a and the lactams 12 in equivalent yield, based on recovered starting material.Encouraged by this result, we decided to investigate the reactivity of the two diastereisomers separately.In this context, we first studied reactivity of compound 6a.The alkylation reaction, following the same procedure, led to a mixture of 11a and 12a in 60% and 40% yield respectively.We then turned our attention to the second diasteroisomer 6b.Reaction of 6b furnished the lactam 12b in quantitative yield.This result showed in our case the substratedependent reactivity concerning the butane 2,3-bisacetal protection group.The above observations led us to consider this method as a new and alternative access to alkylated or lactam compound, with increased yields.

Conclusions
In conclusion, we have demonstrated that compound 6a is a useful building block for alkylation reaction and that its diastereoisomer 6b allowed a selective access to the lactam scaffold.Considering the high potential of so-called "azasugars" for drug discovery, we want to extend this methodology to a new approach for the synthesis of indolizidine scaffold, through alkylation with suitable side chains and subsequent cyclization on the nitrogen after elimination of the chiral appendage.Efforts in this direction are currently being pursued in our laboratory.

Experimental Section
General.All reactions were carried out in dried glassware under an argon atmosphere.All solvents were purchased with an analytical grade from SDS. Acetone and chloroform were dried over molecular sieves.Reactions were monitored using thin-layer chromatography (TLC) carried out on 0.25mm E. Merk silica gel plates (60F-254) using UV light as visualizing agent and sulfuric vanillin heated as developing agent.Flash chromatography was performed with silica gel CHROMATOGEL 60 (particle size 20-45 µm or 35-70 µm) supplied by SDS.Yield refers to chromatography and spectroscopically pure compounds, unless otherwise noted. 1 H NMR spectra were recorded on a BRUCKER AC 300 MHz spectrometer or a BRUCKER AVANCE 400 MHz spectrometer.Chemical shifts are reported in ppm and coupling constants (J) in Hz. 13   C NMR spectra were recorded on a BRUCKER AC 300 spectrometer at 75 MHz.The assignments are based upon 1D and 2D spectra recorded using the following pulse sequences from the Bruker standard pulse program library: DEPT, COSY, HSQC, HMBC and NOESY.Mass spectra were measured with a ZQ 2000 Waters mass spectrometer (ESI).High resolution mass spectra were obtained on a Q-ToF1 ESI mass spectrometer (Waters).Infra-red spectra were recorded with a PerkinElmer spectrum 65 FT-IR spectrometer and wavelengths (ν) are given in cm -1 .Specific rotations were measured using a PerkinElmer model 341 polarimeter in a chloroform solution at 20 °C.