Stereoselective synthesis of (+)-1-deoxyaltronojirimycin

Abstract A stereocontrolled, facile and high-yield approach for producing (+)-altroDNJ, has been developed starting from the inexpensive commercial cis 2-butene-1,4-diol. Sharpless epoxidation and a subsequent dihydroxylation were used for the introduction of all stereocentres; finally, the ring closure under basic conditions afforded the piperidine heterocycle.


Introduction
Iminosugars, carbohydrate analogues in which the endocyclic oxygen is replaced by a nitrogen atom, are nowadays the most attractive class of sugar mimics because of their high glycosidase and glycosyltransferase inhibitor activity and hence their therapeutic potential in a vast array of diseases, such as diabetes, glycosphingolipid storage disorders and viral infections (e.g. HIV, hepatitis B and C) (Compain & Martin 2007).
Usually the main problem is the lack of selectivity towards the target enzyme, causing several adverse effects. In order to increase the binding affinity to a particular substrate, many studies have been carried out and several different substitution patterns have been designed and tested. The understanding of the complicated metabolism of glycoconjugates is a challenge for modern medicine and will allow to design iminosugars specific for a particular molecular target.

Results and discussions
Some of the already reported syntheses of deoxynojirimycin and its various isomers (Compain & Martin 2007) employ chiral pool as starting material: carbohydrates as D-glucose derived (Dhavale et al. 2004), amino acids as the Garner aldehyde (Karjalainen & Koskinen 2011;Singh et al. 2014) or pyroglutamic acid (Ikota et al. 1997).
Our approach, based on both our experience on stereo and regiocontrolled opening of oxirane ring (Antonioletti et al. 2000;Righi et al. 2011) and the results we recently obtained on dihydroxylation reaction of optically active trans α,β-unsaturated epoxy esters (Righi et al. 2012), starts from a suitable optically active 2,3-epoxy alcohol 2, easily obtained from the commercially available cis 2-butene-1,4-diol (Roush et al. 1991).
The oxidation to aldehyde of epoxy alcohol 2 followed by a Horner-Emmons reaction afforded the corresponding trans α,β-unsaturated epoxy ester 4, substrate of choice for the dihydroxylation reaction (Scheme 1).
Osmium-catalysed dihydroxylation of 4 provided a diastereomeric mixture (60:40 dr) of chromatographically inseparable epoxy diols 5A and 5B in nearly quantitative yield (VanRheenen et al. 1976, Scheme 2). The ratio has been calculated by integration of the signals of the CHOH in α of the ester moiety on the 1 H NMR spectra of the crude mixture. The correct stereochemistry was assigned comparing the data of the final iminosugar with those reported in literature (Singh & Han 2003;van den Nieuwendijk et al. 2012).
In order to protect the diolic moieties and to achieve a simpler separation, the mixture of 5A and 5B was transformed in the corresponding acetonides (Hermitage et al.1998). Unfortunately, also 6A and 6B proved to be extremely difficult to separate and consequently the subsequent opening with azide of the epoxide ring was performed on the diastereomeric mixture. The regioselectivity of the attack is hardly predictable, as the epoxide ring is functionalised with two ether moieties. However, the use of the system NaN 3 /NH 4 Cl in methanol at 70 °C (Behrens et al. 1985) led exclusively to the attack on C-5 (Scheme 3).
After this reaction, the major diastereomer 7A was isolated via chromatographic purification, while 7B was obtained only as mixture with 7A. Consequently, from this step the synthesis was carried out only from 7A.
Hydrogenation of azido derivative 7A in the presence of ditert-buthyldicarbonate afforded carbamate 8. The free alcoholic moiety was then protected as tert-buthyldimethylsilyl ether in 80% yield from 7A, followed by the nearly quantitative reduction of the ester to alcohol 10.

Conclusions
In summary, through stereocontrolled, facile and high-yield reactions, we have developed a new approach for producing (+)-altroDNJ hydrochloride starting from the inexpensive commercial cis 2-butene-1,4-diol. Moreover, even if the number of steps is comparable to that of some already reported syntheses, in our sequence only few chromatographic purifications were required, making the route accessible and efficient.
It is noteworthy that, the regio and stereoselective control of the key steps allows achievement of the enantiomer (−)-1-deoxyaltronojirimicin simply using (−)-DET instead of (+)-DET in the Sharpless asymmetric epoxidation, following the same synthetic pathway.