Allylic Azide Rearrangement in Tandem with Huisgen Cycloaddition for Stereoselective Annulation: Synthesis of C-Glycosyl Iminosugars

Article abstract of DOI:10.1021/acs.orglett.5b03209

Allylic azide rearrangement is used in tandem with intramolecular azide-alkene cycloaddition to give a triazoline that when subsequently decomposed in the presence of a nucleophile gives piperidines. The tandem reaction gives two stereocenters that are generated with high control. The formation of the piperidines required the presence of innate conformational constraint. The applicability of the annulation reaction is demonstrated by the synthesis of iminosugars. A proposal is included to account for the observed stereoselectivity, which is influenced by the precursor structure.

Full text of DOI:10.1021/acs.orglett.5b03209

Letter
pubs.acs.org/OrgLett
Allylic Azide Rearrangement in Tandem with Huisgen Cycloaddition
for Stereoselective Annulation: Synthesis of C‑Glycosyl Iminosugars
Lorna Moynihan, Rekha Chadda, Patrick McArdle, and Paul V. Murphy*
School of Chemistry, National University of Ireland Galway, University Road, Galway, Ireland
S
* Supporting Information
ABSTRACT: Allylic azide rearrangement is used in tandem with
intramolecular azide−alkene cycloaddition to give a triazoline that when
subsequently decomposed in the presence of a nucleophile gives
piperidines. The tandem reaction gives two stereocenters that are
generated with high control. The formation of the piperidines required
the presence of innate conformational constraint. The applicability of
the annulation reaction is demonstrated by the synthesis of iminosugars.
A proposal is included to account for the observed stereoselectivity,
which is influenced by the precursor structure.
rganic azides act as 1,3-dipoles in cycloaddition reactions
(Huisgen cycloaddition1) and when reacted with an
alkene give a triazoline that may decompose to form an
aziridine, imine, or different products.2 The aziridine, if formed,
can undergo further reaction with nucleophiles. This sequence
has been used by our group for the synthesis of 1-
deoxynojirmycin derivatives 2 from 1 (Scheme 1).3 Here we
between 4 and stereoisomers 3 and 5 through allylic azide
rearrangement is possible, we considered that the rates of
intramolecular cycloaddition in 3 and 5 would not be equal and
that stereocontrolled ring formation could be achieved.
O
Typical syntheses of allylic azide precursors required for this
study are shown in Schemes 2 and 3. Thus, the 6-iodo
derivative of methyl α-D-mannopyranoside (7)12 was silylated
to give 8. Zinc-mediated reductive fragmentation13 converted 8
to an open-chain aldehyde, which gave ester 9 after Wittig
reaction. Reduction of 9 to the primary alcohol followed by a
Mitsunobu-type exchange of the OH for azide gave 10. One
major isomer, the primary azide 10 with the E configuration,
was observed by 1H NMR spectroscopy in CDCl3 even though
conversions to its Z isomer and secondary azides of types 3 and
5 are possible. The TES groups were next removed using TBAF
to give 11, which showed evidence for the presence of one
major stereoisomer and minor amounts of other isomers. An
acetonide group was next introduced to the 5- and 6-OH
groups of 11 in a regioselective manner by reacting it with 2,2-
dimethoxypropane in the presence of H2SO4 in dry acetone;
this ketalization reaction led to acetonide 12 with 5,6-trans
substituents, whereas the alternate acetonide would have had cis
substituents and higher steric hindrance. Acetylation of the free
4-OH to give 13 helped to verify that the acetonide had been
introduced at C-5 and C-6, as there was a significant downfield
shift (∼1 ppm) of the 1H NMR signal for the C-4 proton upon
acetylation of the adjacent OH group. In addition, the use of
2D heteronuclear multiple-bond correlation spectroscopy
(HMBC) showed the required three-bond correlation between
the carbonyl of the acetyl group and H-4 in 13.14
Scheme 1
show that allylic azide rearrangement4 can be used in tandem
with triazoline formation and that subsequent decomposition of
the triazoline leads to the formation of piperidines with
controlled formation of two new stereocenters. The intra-
molecular azide−alkyne variant was similarly successful.5−10
We set out to investigate the viability of the tandem reaction
for the preparation C-glycosyl iminosugars, which are of
synthetic, biological, and medical interest.11 As interconversion
The 4,5-protected isopropylidene derivative 16 was also
prepared (Scheme 3). Thus, the reaction of 7 with 2,2-
Received: November 6, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.5b03209
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 2. Synthesis of Precursors
Scheme 4. Reactions from 12 and 14 and X-ray Crystal
Structure of 23α
Scheme 3. Synthesis of 16
reaction of the mixture with aqueous HCl gave iminosugar 19α
(28% from 12) after ion exchange chromatography. Catalytic
hydrogenation of 19α gave 20α. The protocol was also
investigated from 16, and in this case triazoline 22α could not
be obtained; attempts to isolate it as for 17 gave an intractable
mixture of products. However, 22α was clearly formed in a
stereoselective manner, as the heating of 16 in DMF for 6 h
followed by addition of acetic acid and further heating led to
the isolation of 21α (45%). The improved reaction from 16
compared with that from 12 is due to a lower degree of strain in
the forming piperidine ring when it is fused to a cis-dioxolane
ring rather than to the more strained trans-dioxolane ring, the
latter being obtained from 12. Removal of the protecting
groups from 21α and hydrogenation again gave 19α. Thus, a
stereoselective one-pot reaction to give C-glycosyl mannojir-
imycin derivatives occurred from 16.
In our hands it was not possible to obtain crystals of 18α−
21α for X-ray crystal structure determination. However, when
the reaction from 16 was carried out with thiophenol as the
nucleophile, 23α was obtained in crystalline form. Determi-
nation of its X-ray crystal structure confirmed its α-manno
configuration, where the vinyl group is axially oriented. The
formation of 23α is consistent with the reaction proceeding via
triazoline 22α, which in turn provides the basis for
confirmation of the structures of 18α−21α. To further support
this, we observed strong NOE cross-peaks between the
anomeric CH2 group and ring protons H-3 and H-5 of the
dimethoxypropane under acidic conditions and subsequent
silylation gave 14. Reductive fragmentation, Wittig reaction,
and reduction followed by introduction of the azide and TES
removal gave 16 via 15.
With various substrates in hand (e.g., 10−13 and 16), the
tandem allylic azide rearrangement/Huisgen cycloaddition was
investigated. A variety of conditions and solvents (e.g., MeCN,
EtOH, THF, toluene, DMF, H2O, MeOH, DMA, CHOEt3)
were explored. Reactions were achieved for acetonides but not
acyclic derivatives, indicating that innate conformational
constraint is required. Heating of acetonide 12 in DMF at
100 °C for 6 h gave triazoline 17 (50−60%) after
chromatography (Scheme 4), and a small amount of unreacted
12 (∼10%) was recovered from the reaction mixture. Given
that triazolines such as 17 are often unstable, we investigated
taking 12 and converting it directly to the piperidine in one pot.
This was most effectively achieved from 12 by heating for 6 h in
DMF at 100 °C, cooling the mixture, and then adding acetic
acid (5 equiv) with subsequent heating at 50 °C; this led to the
formation of 18α, which could not be separated from its
anomeric product 18β (∼35%, 3:1), However, subsequent
B
DOI: 10.1021/acs.orglett.5b03209
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
iminosugar ring for 20α, as would be expected. The 1H and 13
C
Scheme 6. Stereocontrolled Azide−Alkyne Cycloaddition
and X-ray Crystal Structure of 32
NMR spectroscopic data for the α-anomer 20α were in
excellent agreement with those reported for a natural product
that was isolated previously,15 and the coupling constants
derived from the 1H NMR data indicate a preference for a chair
conformer. The evidence presented herein shows that the
structure of this natural product was incorrectly assigned as 20β
(where the vinyl group is equatorial) and should in our view be
20α, assuming that the D-mannose-configured analogue is
produced naturally. Independently, Fleet and co-workers
reported the synthesis of the β-anomer 20β,16 and their 13C
NMR spectroscopic data for 20β are significantly different from
those we observed for 20α and those for the natural product.
The reaction sequence was also investigated from substrates
prepared from D-galactose (Scheme 5). Hence, reaction of
Scheme 5. Reactions from Galactose-Derived 24
highly stereoselective manner, with the β-anomer being the
only isolated product (63%), as confirmed by the X-ray crystal
structure of 32. Removal of the protecting groups gave the 1-
deoxynojirimycin analogue 33 (3J1,2 = 8.3 Hz).
The stereochemical outcome of these reactions has been
rationalized as summarized in Figure 1. First, the azide−alkene
acetonide 24 gave altronojirimycin derivative 25 in a highly
stereoselective manner (45%) using acetic acid, whereas the use
of thiophenol gave 28 (40%). Removal of the protecting groups
from 25 gave 26, and subsequent reduction of the alkene gave
27. The altrose configuration (axial hydroxymethyl group at C-
5) was assigned on the basis of a comparison to the 13C data
obtained by Fleet and co-workers for the two known isomers
29 and 30.17 We found better agreement, or less variance, of
the 13C chemical shift data for the ring carbon atoms of 27 with
those of 29 than with those of 30, particularly for the signals of
C-4 to C-6. Evidence that there is a relationship between
chemical shift and stereochemistry in 13C NMR data has been
demonstrated in natural product fragments by Kishi and co-
workers.18 In addition, 1D NOE experiments with 28
supported the altrose configuration, as there was a weak
cross-peak between H-3 and one of the H-6 protons but no
cross-peak between H-3 and H-5, the latter being expected for a
galactose-configured isomer. In contrast with that observed
from mannose substrates, the anomer with the vinyl group
equatorially oriented was the major product in this case, as
Figure 1. Possible geometries generated from 16 (top left) and 27
(bottom left). Newman projection formulas viewing along C-1 to C-2
for allylic azide stereoisomers are shown at the center (corresponding
to the model on the left) and at the right (corresponding to the other
stereoisomer). In the top right and bottom right, the vinyl group and
2-substituent are close to being eclipsed, which disfavors pathways
from these stereoisomers. Minimization of allylic strain determines the
C-4 to C-5 dihedral angle in the reacting conformer.
cycloaddition of the mannose substrate ultimately gives a
product that also has a mannose configuration; this contrasted
with galactose, which gives a product with an altrose
configuration. These differing outcomes are explained by
minimization of allylic strain in the reaction transition state.19
Hence, the orientation of the alkene involved in the dipolar
cycloaddition depends on the configuration of the substituent
at C-4, and this influences the stereoselectivity observed at the
iminosugar C-5, giving a 4,5-trans product.
1
1
supported by H NMR data (3J1,2 = 9.1 Hz). The H NMR
coupling constants indicate that a chair conformer is preferred
for 27.
The tandem rearrangement/cycloaddition approach was also
extended to incorporate an alkyne as the dipolarophile (Scheme
6). Thus, acetonide 31 was heated in one pot to give 32 in a
With regard to the anomeric selectivity, substrates from
glucose and galactose give the C-glycosyl product where the
C
DOI: 10.1021/acs.orglett.5b03209
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(2) (a) Brase, S.; Banert, K. Organic Azides: Syntheses and
̈
vinyl group is equatorial, whereas substrates from mannose give
the product with the vinyl group axially oriented; this can be
summarized as giving a 1,2-trans product. The cycloaddition is
normally a concerted process20 between the azide and alkene,
and this may imply that the forming piperidine needs to adopt a
boat or twist-boat conformation to enable the required orbital
overlap in the transition state. In this situation, substituents on
the forming six-membered ring could be eclipsed or close to
eclipsed. For the reaction from mannose substrates, a reaction
of the secondary allylic azide with the S configuration would
place the vinyl group close to a C−H bond, whereas reaction of
the R isomer would place the vinyl group close to the C−O
bond. Reaction of the S isomer would be faster for steric
reasons, and this would give rise to the axially oriented anomer.
Where galactose substrates are concerned, the reaction from
the R stereoisomer is faster than that from the S stereoisomer
because of the configurational change at C-2, and thus, the
anomeric selectivity is reversed. This rationale is summarized in
Figure 1.
Applications; Wiley-VCH: Weinheim, Germany, 2009. (b) Kadaba, P.
K.; Stevenson, P. J.; P-Nnane, I.; Damani, L. A. Bioorg. Med. Chem.
1996, 4, 165.
(3) Zhou, Y.; Murphy, P. V. Org. Lett. 2008, 10, 3777.
(4) Gagneux, A.; Winstein, S.; Young, W. G. J. Am. Chem. Soc. 1960,
82, 5956.
(5) For allylic azide rearrangement used in tandem with intra-
molecular Schmidt reactions, see: Liu, R.; Gutierrez, O.; Tantillo, D. J.;
́
Aube, J. J. Am. Chem. Soc. 2012, 134, 6528.
(6) For reviews of the application of intramolecular 1,3-dipolar
cycloaddition in targeted synthesis, see: (a) Nair, V.; Suja, T. D.
Tetrahedron 2007, 63, 12247. (b) Padwa, A. In 1,3-Dipolar
Cycloaddition Chemistry; Padwa, A., Ed.; Wiley-Interscience: New
York, 1984; Vol. 2, p 316.
(7) For a review of the applications of organic azides, including
alkene−azide cycloadditions, see: Brase, S.; Gil, C.; Knepper, K.;
̈
Zimmermann, V. Angew. Chem., Int. Ed. 2005, 44, 5188.
(8) For allylic azide rearrangement followed by intramolecular azide−
́
alkyne cycloaddition, see: Vekariya, R. H.; Liu, R.; Aube, J. Org. Lett.
2014, 16, 1844−47.
In summary, we have demonstrated that the allylic azide
rearrangement when coupled to the Huisgen azide−alkene
cycloaddition (or azide−alkyne cycloaddition) can be used
productively to define a highly diastereoselective annulation in
which two stereocenters are generated in a controlled manner.
Application of the reaction has been demonstrated by the
synthesis of iminosugars (polyhydroxylated piperidines), which
are of biomedical relevance.11 Further study of this ring-
forming approach is underway, and the outcomes of that work
will be reported in due course.21
(9) For allylic azide rearrangements followed by copper-promoted
intermolecular azide−alkene cycloaddition, see: Feldman, A. K.;
Colasson, B.; Sharpless, K. B.; Fokin, V. V. J. Am. Chem. Soc. 2005,
127, 13444.
(10) For intramolecular cycloaddition of primary allylic azides and
alkynes, see: Mishra, A.; Hutait, S.; Bhowmik, S.; Rastogi, N.; Roy, R.;
Batra, S. Synthesis 2010, 2010, 2731.
(11) For recent selected applications of iminosugars, see: (a) Clark,
N.; Metcalf, M. C.; Best, D.; Fleet, G. W. J.; Garman, S. C. Proc. Natl.
Acad. Sci. U. S. A. 2012, 109, 17400. (b) Ghisaidoobe, A. T.; van den
Berg, R. J. B. H. N.; Butt, S. S.; Strijland, A.; Donker-Koopman, W. E.;
Scheij, S.; van den Nieuwendijk, A. M. C. H.; Koomen, G.-J.; van
Loevezijn, A.; Leemhuis, M.; Wennekes, T.; van der Stelt, M.; van der
Marel, G. A.; van Boeckel, C. A. A.; Aerts, J. M. F. G.; Overkleeft, H. S.
J. Med. Chem. 2014, 57, 9096. (c) Lopez, O.; Qing, F.-L.; Pedersen, C.
M.; Bols, M. Bioorg. Med. Chem. 2013, 21, 4755. (d) Sanchez-
Fernandez, E. M.; Risquez-Cuadro, R.; Ortiz Mellet, C.; Garcia
Fernandez, J. M.; Nieto, P. M.; Angulo, J. Chem. - Eur. J. 2012, 18,
8527. (e) Decroocq, C.; Stauffert, F.; Pamlard, O.; Oulaidi, F.;
Gallienne, E.; Martin, O. R.; Guillou, C.; Compain, P. Bioorg. Med.
Chem. Lett. 2015, 25, 830−833.
(12) Skaanderup, P. R.; Poulsen, C. S.; Hyldtoft, L.; Jørgensen, M. R.;
Madsen, R. Synthesis 2002, 2002, 1721.
(13) (a) Bernet, B.; Vasella, A. Helv. Chim. Acta 1979, 62, 1990.
(b) Skaanderup, P. R.; Hyldtoft, L.; Madsen, R. Monatsh. Chem. 2002,
133, 467.
(14) HMBC in conjunction with COSY and HSQC experiments was
generally used to establish the connectivity in the molecules prepared.
(15) Kato, A.; Kato, N.; Adachi, I.; Hollinshead, J.; Fleet, G. W. J.;
Kuriyama, C.; Ikeda, K.; Asano, N.; Nash, R. J. J. Nat. Prod. 2007, 70,
993.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.5b03209.
Crystallographic data for 32 (CIF)
Crystallographic data for 23α (CIF)
NMR spectra (PDF)
Experimental procedures and analytical data for com-
pounds, schemes and experimental section for the
preparation of 24 and 31, tables of NMR data, and
plots of chemical shift differences (PDF)
AUTHOR INFORMATION
Corresponding Author
■
(16) Fleet, G. W. J.; Namgoong, S. K.; Barker, C.; Baines, S.; Jacob,
G. S.; Winchester, B. Tetrahedron Lett. 1989, 30, 4439.
(17) Shilvock, J. P.; Wheatley, J. R.; Nash, R. J.; Watson, A. A.;
*paul.v.murphy@nuigalway.ie
Notes
The authors declare no competing financial interest.
Griffiths, R. C.; Butters, T. D.; Muller, M.; Watkin, D. J.; Winkler, D.
̈
A.; Fleet, G. W. J. J. Chem. Soc., Perkin Trans. 1 1999, 2735.
(18) Kobayashi, Y.; Lee, J.; Tezuka, K.; Kishi, Y. Org. Lett. 1999, 1,
2177−2180.
ACKNOWLEDGMENTS
■
This research was supported by the Irish Research Council
(Ph.D. scholarship to L.M.) and NUI Galway (College of
Science Ph.D. scholarship to R.C.). This publication has also
emanated in part from research supported by Science
Foundation Ireland (SFI, 07/IN.1/B966) that was cofunded
under the European Regional Development Fund.
(19) Hoffmann, R. W. Chem. Rev. 1989, 89, 1841.
(20) Huisgen, R. J. Org. Chem. 1968, 33, 2291.
(21) For related previous work from our laboratory, see:
(a) McDonnell, C.; Cronin, L.; O’ Brien, J. L.; Murphy, P. V. J. Org.
Chem. 2004, 69, 3565. (b) Cronin, L.; Murphy, P. V. Org. Lett. 2005,
7, 2691. (c) O’Reilly, C.; O’Brien, C.; Murphy, P. V. Tetrahedron Lett.
2009, 50, 4427.
REFERENCES
■
(1) (a) Huisgen, R. Angew. Chem. 1963, 75, 604. (b) Huisgen, R.
Angew. Chem., Int. Ed. Engl. 1963, 2, 633.
D
DOI: 10.1021/acs.orglett.5b03209
Org. Lett. XXXX, XXX, XXX−XXX