Synthesis of 1,2-cis -homoiminosugars derived from GlcNAc and GaLNAc exploiting a β-amino alcohol skeletal rearrangement

Article abstract of DOI:10.1021/ol502926f

The synthesis of 1,2-cis-homoiminosugars bearing an NHAc group at the C-2 position is described. The key step to prepare these α-d-GlcNAc and α-d-GalNAc mimics utilizes a β-amino alcohol skeletal rearrangement applied to an azepane precursor. This strateg

Full text of DOI:10.1021/ol502926f

Letter
pubs.acs.org/OrgLett
Synthesis of 1,2-cis-Homoiminosugars Derived from GlcNAc and
GalNAc Exploiting a β‑Amino Alcohol Skeletal Rearrangement
Yves Bleriot,*,† Nicolas Auberger,† Yerri Jagadeesh,† Charles Gauthier,† Giuseppe Prencipe,‡
́
Anh Tuan Tran,‡ Jerome Marrot,§ Jerome Desire,† Arisa Yamamoto,∥ Atsushi Kato,∥
́
̂
́
̂
́
́
and Matthieu Sollogoub*,‡
†Glycochemistry Group of “Organic Synthesis” Team, Universite
́ ̂
de Poitiers, UMR-CNRS 7285 IC2MP, Bat. B28, 4 rue Michel
Brunet, TSA 51106, 86073 Poitiers Cedex 9, France
‡Sorbonne Universites
Paris, France
́
, UPMC Univ Paris 06, Institut Universitaire de France, UMR-CNRS 8232, IPCM, LabEx MiChem, F-75005
§Institut Lavoisier de Versailles, UMR-CNRS 8180, Universite
́
de Versailles, 45 avenue des Etats-Unis, 78035 Versailles Cedex, France
́
∥Department of Hospital Pharmacy, University of Toyama, 2630 Sugitani, Toyama 930-0194, Japan
S
* Supporting Information
ABSTRACT: The synthesis of 1,2-cis-homoiminosugars bear-
ing an NHAc group at the C-2 position is described. The key
step to prepare these α-D-GlcNAc and α-D-GalNAc mimics
utilizes a β-amino alcohol skeletal rearrangement applied to an
azepane precursor. This strategy also allows access to naturally
occurring α-HGJ and α-HNJ. The α-D-GlcNAc-configured
iminosugar was coupled to a glucoside acceptor to yield a
novel pseudodisaccharide. Preliminary glycosidase inhibition
evaluation indicates that the α-D-GalNAc-configured homoiminosugar is a potent and selective α-N-acetylgalactosaminidase
inhibitor.
minoalditols rank among the most powerful glycosidase
inhibitors. The most well-known representative, 1-deoxy-
allowing access to α-homo-2-acetamido-1,2-dideoxynojirimycin
(α-HNJNAc, 6) and α-homo-2-acetamido-1,2-dideoxy-galacto-
nojirimycin (7, α-HGJNAc).
I
nojirimycin (1, DNJ), is found in two approved medicines.1
Nature has overcome the absence of pseudoanomeric
information in DNJ and the instability of the hemiaminal in
the related molecule nojirimycin (2, NJ)2 by producing
homonojirimycins (3, HNJ)3 that carry an extra pseudoanomeric
CH2OH group (Figure 1). This functionality not only improves
Our groups and others8 have exploited the ring isomerization
of seven-membered polyhydroxylated azacycles to generate new
or known piperidine iminosugars. This transformation, based on
a β-amino alcohol rearrangement,9 requires a free alcohol at the
β-position and proceeds via a double SN2 mechanism. We have
previously used this process to access homoglyconojirimycins10
and iminosugar C-glycosides.11 Applying this strategy to a cis β,γ-
dihydroxyazepane bearing an electron-donating group on the
endocyclic nitrogen followed by the introduction of an azido
group at C-2 could pave the way to the GlcNAc-derived 1,2-cis
homoiminosugars. To this end, a stereoselective route to a D-
arabino-like azacycloheptene was developed. Commercially
available D-arabinofuranose derivative 812 was olefinated and
the resulting free alcohol inverted using Mitsunobu conditions to
provide alcohol 9 after ester hydrolysis (66% over three steps).
Esterification with triflic anhydride, followed by nucleophilic
displacement with allylamine, produced the corresponding
iminodiene that was protected as its tert-butyl carbamate to
give diene 10 (66% over three steps). Ring-closing metathesis
with Grubbs’ first-generation catalyst in 1,2-dichloroethane at
reflux afforded azacycloheptene 11 (85%). According to our
Figure 1. Structure of NJ, DNJ, HNJ, DNJNAc, NJNAc, and target
iminosugars α-HNJNAc and α-HGJNAc.
the glucosidase inhibition profile of the parent iminosugar but
also allows the introduction of aglycon moieties to generate
chemically stable iminoglycoconjugates with promising bio-
logical activities.4 Many synthetic routes to these highly
substituted piperidines have been reported, covering most of
the important monosaccharides.5 There is one notable exception
that has stood the test of time until very recently,6 that is the
homologues of the GlcNAc-derived iminosugars DNJNAc (4)
and NJNAc (5).7 We disclose herein a general synthetic strategy
Received: October 3, 2014
Published: October 20, 2014
© 2014 American Chemical Society
5512
dx.doi.org/10.1021/ol502926f | Org. Lett. 2014, 16, 5512−5515

Organic Letters
Letter
strategy, access to α-HNJNAc requires the preparation of the β,γ-
dihydroxyazepane. Dihydroxylation of 11 with OsO4 proceeded
with a facial diastereoselectivity (5:3 ratio) similar to that
observed for azepanes bearing a N-Cbz group.13 As the
separation of these diols proved to be difficult on a large scale,
they were converted into the separable isopropylidene acetals 12
and 13. Because of the presence of rotamers, the elucidation of
the stereochemistry of acetals 12 and 13 and corresponding diols
was difficult by NMR. This was achieved through the full
deprotection of 12 and 13 to give the known pentahydroxylated
azepanes 15 and 14, respectively.13 Removal of both the
isopropylidene and Boc groups in 12 with TFA, followed by
N-benzylation, gave the N-benzyl diol 16 (88% over two steps)
ready for ring contraction (Scheme 1).
Scheme 2. Proposed Mechanism for the Ring Contraction of
Azepane 16 Leading to Piperidine 19 and Azepane 20
displacement of the free OH at C-2 should also involve the
anchimeric assistance of the endocyclic nitrogen to generate the
transient bicycle 22 (Scheme 3). This bicycle would then be
Scheme 1. Synthesis of Protected Azepanes 12 and 13
Scheme 3. Synthesis of α-HNJNAc 6 and Pseudodisaccharide
28
We have previously exploited the superior reactivity of the β-
OH group over its γ-counterpart in syn β,γ-dihydroxyazepanes in
order to achieve its selective protection.10a As a consequence,
under β-amino alcohol rearrangement conditions, we expected
the β-OH in 16 to be selectively activated and displaced by the
endocyclic nitrogen to generate a transient fused piperidine−
aziridinium ion 18. Opening of the aziridinium ion by the acetate
at the methylene carbon should provide the corresponding
piperidine 19. The acetylated azepane 20 resulting from attack of
the aziridinium ring at the methine carbon could also be observed
(Scheme 2).
Satisfyingly, when diol 16 was treated with a small excess of
AcOH, PPh3, and DEAD in THF at 0 °C, the piperidine 19 was
isolated in 50% yield along with azepane 20 (24%). Piperidine 19
displayed a small coupling constant (J1,2 = 4.0 Hz) in agreement
with a 1,2-cis relationship for the piperidine ring. To further
confirm its structure, piperidine 19 was fully deprotected by
deacetylation, followed by hydrogenolysis, to give the naturally
occurring α-homonojirimycin (α-HNJ) 21,14 the 1H NMR data
for which agreed perfectly with that reported in the literature.
The structure of azepane 20 was supported by the production of
dihydroxyazepane 16 following its deacetylation with NaOMe in
MeOH. Protection of the pseudoanomeric alcohol in 19 as
methyl ester was exploited to investigate the azidation at the C-2
position. We relied on the hypothesis that nucleophilic
attacked at the C-2 position, providing the thermodynamically
stable six-membered ring. When piperidine 19 was treated with
diphenyl phosphoryl azide (dppa), DEAD, and PPh3 in THF, the
2-azidopiperidine 23, displaying an α-D-gluco configuration
according to the 1H NMR coupling constants (J1,2 = 6.0 Hz, J2,3
=
10.1 Hz), was obtained in 81% yield. Conversion of the azido
group into an acetamide (PPh3 in THF/water then Ac2O/py)
followed by O-deacetylation and hydrogenolysis furnished the
target α-HNJNAc 6 (73% over four steps) (Scheme 3). The
5513
dx.doi.org/10.1021/ol502926f | Org. Lett. 2014, 16, 5512−5515

Organic Letters
Letter
structure of 6 was further confirmed by X-ray crystallography of
the corresponding per-O-acetylated derivative 24 (Figure 2).
Scheme 4. Synthesis of α-HGJNAc 7
Figure 2. X-ray crystallography of iminosugar 24 (CCDC 1023880).
The main synthetic advantage of this new class of inhibitors
over previously reported iminosugars that mimic GlcNAc is their
ability to incorporate aglycons at the pseudoanomeric position to
form chemically stable iminoglycoconjugates. We examined this
capacity through the synthesis of a pseudodisaccharide. The
anchimeric assistance of the endocyclic amine associated with the
sensitivity of the acetamide to basic conditions required a specific
coupling strategy. It should involve an aglycon partner bearing a
leaving group and a nucleophilic homominosugar derivative
bearing an azido group as a masked NHAc functionality.
Piperidine 23 was O-deacetylated with sodium methoxide and
treated with the known activated sugar 2515 to furnish the
protected pseudodisaccharide 26 (71% over three steps).
Subsequent generation of the acetamide group as described
above (70% yield) provided compound 27 that was further
hydrogenolyzed to afford the α-1,6-pseudodisaccharide 28
(Scheme 3).
Encouraged by these results, we used our strategy to access the
α-homo-2-acetamido-1,2-dideoxygalactonojirimycin (α-
HGJNAc), a GalNAc analogue that could interfere with human
α-N-acetyl-D-galactosaminidase (α-GalNAcase) involved in
Shindler−Kanzaki disease.16 Starting from L-ribose (Carbo-
synth), glycosylation in methanol followed by perbenzylation
and selective anomeric deprotection furnished furanose 29 (66%
over three steps). Wittig olefination under classical conditions
proved problematic17 and needed optimization. While BuLi gave
poor yields, the use of t-BuOK (2.5 equiv) in toluene at reflux
provided alkene 30 in 78% yield. Introduction of allylamine as
described above produced the corresponding diene that was
protected as its benzyl carbamate to afford compound 31 (52%
over three steps). Ring-closing metathesis gave the azacyclo-
heptene 32 (78% yield) that furnished the required dihydrox-
yazepane 33 (67% over three steps) upon dihydroxylation
(OsO4, NMO) and protecting group interconversion. The
stereochemistry of 33 was firmly established through its full
deprotection to yield the known pentahydroxylated azepane 3418
(Scheme 4).
group removal afforded the target α-HGJNAc (7) (Scheme 4).
Both α-HNJNAc (6) and α-HGJNAc (7) adopt a 4C1
conformation in CD3OD as illustrated by the key coupling
constants (J1,2 = 5.6 Hz, J2,3 = 10.5 Hz, J3,4 = 8.3 Hz, J4,5 = 9.3 Hz
for α-HNJNAc (6) and J1,2 = 5.9 Hz, J2,3 = 10.9 Hz, J3,4 = 2.1 Hz
for α-HGJNAc (7)). The α-HNJNAc (6), α-HGJNAc (7), and
pseudodisaccharide 28 were assayed on a panel of β-N-
acetylhexosaminidases from human placenta, bovine kidney,
HL-60, Jack bean, and Aspergillus oryzae and α-N-acetylgalacto-
saminidase from chicken liver (Table 1).
While pseudodisaccharide 28 is a poor inhibitor of these
enzymes, α-HNJNAc (6) is a moderate inhibitor of β-N-
acetylhexosaminidases, displaying two digit IC50. The α-
Application of the optimized ring-contraction conditions to
diol 33 generated the desired piperidine 35 (57% yield). As the
1H NMR spectrum of 35 contained broad signals that could not
be easily assigned, the structure of 35 was established by
achieving its full deprotection. It cleanly provided the known α-
homogalactonojirimycin (α-HGJ, 36), the spectroscopic data for
which agreed perfectly with that reported in the literature.19 The
parallel introduction of an azido group at the C-2 position of
piperidine 35 with dppa, DEAD, and PPh3 furnished the
azidopiperidine 37 in 61% yield while also retaining an α-D-
galacto configuration as confirmed by the NOE effects observed
between H-3 and the pseudoanomeric methylene group, H-5 and
the pseudoanomeric methylene group, and H-1 and H-6,
respectively. Final acetamide installation at C-2 and protecting
Table 1. Concentration (in μM) of Iminosugars 6, 7, and 28
Giving 50% Inhibition of Various Glycosidases (IC50)
enzyme
6
7
28
β-N-acetylhexosaminidase
human placenta
56
67
46
36
269
310
416
917
NI
bovine kidney
HL-60
265
48
52
Jack bean
22
Aspergillus oryzae
α-N-acetylgalactosaminidase chicken liver
NI
NI
ND
1.1
567
NI: no inhibition (less than 50% inhibition at 1 mM). ND: Not
determined.
5514
dx.doi.org/10.1021/ol502926f | Org. Lett. 2014, 16, 5512−5515

Organic Letters
Letter
Huang, S.-H.; Chang, Y.-F.; Cheng, T.-J. R.; Cheng, W.-C.; Wong, C.-H.
J. Org. Chem. 2014, 79, 8629−8637.
HGJNAc (7) is a potent and rather selective α-N-acetylgalacto-
saminidase inhibitor (IC50 1.1 μM).
(7) (a) Fleet, G. W. J.; Smith, P. W.; Nash, R. J.; Fellows, L. E.; Parekh,
R. B.; Rademacher, T. W. Chem. Lett. 1986, 1051−1054. (b) Fleet, G. W.
J.; Fellows, L. E.; Smith, P. W. Tetrahedron 1987, 43, 979−990.
(c) Kappes, E.; Legler, G. J. Carbohydr. Chem. 1989, 8, 371−378.
In summary, the synthesis of 1,2-cis homoiminosugars derived
from GlcNAc and GalNAc bearing an α configuration for the
pseudoanomeric CH2OH group has been achieved by exploiting
a β-aminoalcohol rearrangement applied to a seven-membered
iminosugar. This strategy also allows access to naturally
occurring α-HGJ and α-HNJ. These derivatives can be coupled
to various aglycon moieties and a pseudodisaccharide incorpo-
rating the α-HNJNAc has been obtained. These molecules
increase the array of homoiminosugars available for biologists
that could be used as probes to decipher the conformational
itineraries and mechanisms of glycosidases.
(d) Gradnig, G.; Legler, G.; Stutz, A. E. Carbohydr. Res. 1996, 287, 49−
̈
57. (e) Furneaux, R. H.; Gainsford, G. J.; Lynch, G. P.; Yorke, S. C.
Tetrahedron 1993, 49, 9605−9612. (f) Best, D.; Chairatana, P.; Glawar,
A. F. G.; Crabtree, E.; Butters, T. D.; Wilson, F. X.; Yu, C.-Y.; Wang, W.-
B.; Jia, Y.-M.; Adachi, I.; Kato, A.; Fleet, G. W. J. Tetrahedron Lett. 2010,
51, 2222−2224. (g) de la Fuente, A.; Martin, R.; Mena-Barragan, T.;
Verdaguer, X.; Garcia Fernandez, J. M.; Ortiz Mellet, C.; Riera, A. Org.
Lett. 2013, 15, 3638−3641.
(8) (a) Poitout, L.; Le Merrer, Y.; Depezay, J.-C. Tetrahedron Lett.
1996, 37, 1613−1616. (b) Bagal, S. K.; Davies, S. G.; Lee, J. A.; Roberts,
P. M.; Russell, A. J.; Scott, P. M.; Thompson, J. E. Org. Lett. 2010, 12,
136−139.
ASSOCIATED CONTENT
■
S
* Supporting Information
(9) Met
2010, 39, 89−102.
(10) (a) Liu, T.; Zhang, Y.; Bler
(b) Lammerts van Bueren, A.; Fayers-Kerr, J.; Luo, B.; Zhang, Y.;
Sollogoub, M.; Bleriot, Y.; Rovire, C.; Davies, G. J. J. Am. Chem. Soc.
2010, 132, 1804−1806.
(11) Mondon, M.; Fontelle, N.; Des
Marrot, J.; Bleriot, Y. Org. Lett. 2012, 14, 870−873.
(12) Purchased from Sigma-Aldrich.
(13) Li, H.; Bleriot, Y.; Chantereau, C.; Mallet, J.-M.; Sollogoub, M.;
́
ro, X.; Duthion, B.; Gomez Pardo, D.; Cossy, J. Chem. Soc. Rev.
Experimental details, spectra, and X-ray crystallographic data.
This material is available free of charge via the Internet at http://
pubs.acs.org.
́
iot, Y. Synlett 2007, 905−907.
́
AUTHOR INFORMATION
■
́
́ ́
ire, J.; Lecornue, F.; Guillard, J.;
Corresponding Authors
́
*E-mail: yves.bleriot@univ-poitiers.fr.
*E-mail: matthieu.sollogoub@upmc.fr.
Notes
́
Zhang, Y.; Rodriguez-Garcia, E.; Vogel, P.; Jimenez-Barbero, J.; Sinay, P.
̈
Org. Biomol. Chem. 2004, 2, 1492−1499.
The authors declare no competing financial interest.
(14) Dondoni, A.; Nuzzi, A. J. Org. Chem. 2006, 71, 7574−7582.
(15) Haines, A. L. Org. Biomol. Chem. 2004, 2, 2352−2358.
(16) Clark, N. E.; Metcalf, M. C.; Best, D.; Fleet, G. W. J.; Garman, S.
C. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 17400−17405.
(17) Hekking, K. F. W.; Moelands, M. A. H.; van Delft, F. L.; Rutjes, F.
P. J. T. J. Org. Chem. 2006, 71, 6444−6450.
ACKNOWLEDGMENTS
■
Support for this research was provided by the Sanfilippo
Foundation Switzerland, Dorphan, and “Vaincre les Maladies
Lysosomales“. We thank Dr. Matthew Young, University of
Oxford, for proofreading this manuscript.
(18) Li, H.; Bler
́
iot, Y.; Mallet, J.-M.; Zhang, Y.; Rodriguez-Garcia, E.;
Vogel, P.; Mari, S.; Jimenez-Barbero, J.; Sinay, P. Heterocycles 2004, 64,
65−74.
̈
(19) Glawar, A. F. G.; Jenkinson, S. F.; Newberry, S. J.; Thompson, A.
L.; Nakagawa, S.; Yoshihara, A.; Akimitsu, K.; Izumori, K.; Butters, T. D.;
Kato, A.; Fleet, G. W. J. Org. Biomol. Chem. 2013, 11, 6886−6899 and
references cited therein.
REFERENCES
■
(1) (a) Jacob, G. S. Curr. Opin. Struct. Biol. 1995, 5, 605−611.
(b) McCormack, P. L.; Goa, K. L. Drugs 2003, 63, 2427−2434.
(2) (a) Niwa, T.; Tsuruoka, T.; Goi, H.; Kodama, Y.; Itoh, J.; Inouye,
S.; Yamada, Y.; Niida, T.; Nobe, M.; Ogawa, Y. J. Antibiot. 1984, 37,
1579−1586.
(3) For α-HNJ see: (a) Kite, G. C.; Horn, J. M.; Romeo, J. T.; Fellows,
L. E.; Lees, D. C.; Scofield, A. M.; Smith, N. G. Phytochemistry 1990, 29,
103−105. (b) Kite, G. C.; Fellows, L. E.; Fleet, G. W. J.; Liu, P. S.;
Scofield, A. M.; Smith, N. G. Tetrahedron Lett. 1988, 29, 6483−6486.
For β-HNJ see: (c) Asano, N.; Kato, A.; Miyauchi, M.; Kizu, H.;
Kameda, Y.; Watson, A. A.; Nash, R. J.; Fleet, G. W. J. J. Nat. Prod. 1998,
61, 625−628. (d) Asano, N.; Nishida, M.; Kizu, H.; Matsui, K.; Watson,
A. A.; Nash, R. J. J. Nat. Prod. 1997, 60, 98−101.
(4) (a) Liu, P. S. J. Org. Chem. 1987, 52, 4717−4721. (b) Boshagen, H.;
Geiger, W.; Junge, B. Angew. Chem., Int. Ed. Engl. 1981, 20, 806−807.
(c) Liotta, L. J.; Lee, J.; Ganem, B. Tetrahedron 1991, 47, 2433−2447.
(d) Liu, P. S.; Rogers, R. S.; Kang, M. S.; Sunkara, P. S. Tetrahedron Lett.
1991, 32, 5853−5856. (e) Johnson, C. R.; Miller, M. W.; Golebiowski,
A.; Sundram, H.; Ksebati, M. B. Tetrahedron Lett. 1994, 35, 8991−8994.
(f) Baudat, A.; Vogel, P. Tetrahedron Lett. 1996, 37, 483−484.
(g) Martin, O. R.; Liu, L.; Yang, F. Tetrahedron Lett. 1996, 37, 1991−
1994. (h) Martin, O. R.; Saavedra, O. M. Tetrahedron Lett. 1995, 36,
799−802.
(5) (a) Compain, P.; Chaignault, V.; Martin, O. R. Tetrahedron:
Asymmetry 2009, 20, 672−711. (b) Dhavale, D. D.; Matin, M. M.
ARKIVOC 2005, iii, 110−132.
(6) During the revision of this manuscript, the Wong group reported
the synthesis of iminosugar C-glycoside analogues of a-D-GlcNAc-1-
phosphate; see: Hsu, C.-H.; Schelwies, M.; Enck, S.; Huang, L.-Y.;
5515
dx.doi.org/10.1021/ol502926f | Org. Lett. 2014, 16, 5512−5515