Investigations into the synthesis of amine-linked neodisaccharides

Article abstract of DOI:10.1016/j.tetlet.2007.10.039

Six tail-to-tail amine-linked neodisaccharides were synthesised as potential glycomimetics. Primary-primary linked compounds were synthesised using Mitsunobu chemistry with glucose-6-sulfonamides as nucleophiles and primary carbohydrate alcohols as electr

Full text of DOI:10.1016/j.tetlet.2007.10.039

Available online at www.sciencedirect.com
Tetrahedron Letters 48 (2007) 8673–8677
Investigations into the synthesis of amine-linked neodisaccharides
Tashfeen Akhtar and Ian Cumpstey*
Department of Organic Chemistry, The Arrhenius Laboratory, Stockholm University, 106 91 Stockholm, Sweden
Received 10 September 2007; revised 22 September 2007; accepted 4 October 2007
Available online 10 October 2007
Abstract—Six tail-to-tail amine-linked neodisaccharides were synthesised as potential glycomimetics. Primary–primary linked com-
pounds were synthesised using Mitsunobu chemistry with glucose-6-sulfonamides as nucleophiles and primary carbohydrate alco-
hols as electrophiles. Primary–secondary linked compounds were synthesised by epoxide ring opening with carbohydrate 6-amines.
Deprotection of all neodisaccharides was carried out using dissolving metal reduction.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Carbohydrates occurring as polysaccharides and oligo-
saccharide glycoconjugates take part in biological recog-
nition events with consequences for both usual healthy
human function and disease.1 Glycomimetics are mole-
cules that resemble a carbohydrate but that may be
modified in some way.2 Usefully, a glycomimetic may
be mistaken for, or even preferred over, the natural sub-
strate by a carbohydrate binding protein. For example,
a disaccharide with a hydrolytically stable inter-glyco-
sidic linkage may be recognised but not cleaved by a
glycosidase.3
OH
O
HO
HO
O
O
OH
OH
OH
OH
I
HO
OH
OH
O
H
N
MeO
HO
O
OH
OH
II
HO
OH
One area of research in our laboratory is the synthesis
of disaccharide mimics with unnatural linkages. One
particular area of interest is that of non-glycosidically-
linked disaccharides, comprising two sugars linked
together without using the anomeric centre. We call
these structures neodisaccharides and recently reported
the synthesis of various thioether-linked examples of this
compound class.4 Conceptually, similar work on ether-
linked5 and C-linked6 tail-to-tail disaccharides has also
been published. It is our hypothesis that such neodisac-
charides may act as glycomimetics with one sugar bind-
ing in its natural orientation to a carbohydrate binding
site in an enzyme or lectin, while the other sugar residue
assumes an unnatural orientation (Fig. 1). The pheno-
menon of different monosaccharides adopting different
orientations in a given binding site has been observed
before. For example, a crystal structure of thiodi-
Figure 1. Structures of a b(1!6)-linked disaccharide, gentobiose I and
its proposed N-linked neodisaccharide mimic II.
galactoside bound into the Galectin-1 binding pocket
shows that one galactose residue occupies the same
space as the GlcNAc residue in bound LacNAc.7
Nilsson has shown that some mannosides can mimic gal-
actose in binding to galectins, the proposed binding
mode being supported by molecular modelling.8 Jenkins
has reported that N-substituted 3-amino altrose deriva-
tives can act as glucosidase inhibitors, and has proposed
that the protonated aminosugar binds to the enzyme in
an orientation such that the amine nitrogen lies coinci-
dent with the binding position of the exocyclic oxygen
in the natural substrate.9 Adding a second carbohydrate
at the aglycon position to such a structure to give an
amine neodisaccharide (cf. II, Fig. 1) could increase
specificity for a given carbohydrate binding protein,
and possibly also binding affinity. Thus, methods to
synthesise such N-linked neodisaccharides are worth
investigating.
Keywords: Aminosugars; Disaccharides; Carbohydrates; Glyco-
mimetics.
*
Corresponding author. Tel.: +46 (0)8 674 7263; fax: +46 (0)8 15 49 08;
e-mail: cumpstey@organ.su.se
0040-4039/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.10.039

8674
T. Akhtar, I. Cumpstey / Tetrahedron Letters 48 (2007) 8673–8677
While C2 symmetric (6–6) N-linked neodisaccharides
have been fairly widely documented10 and reports of
one sec–sec linked compound appeared almost 30 years
ago,11 it was not until earlier this year that the first
reports of unsymmetrically substituted amine neodisac-
charides appeared: Kroutil used a method based on azi-
ridine opening by amines to give diamino compounds.12
Thiem generated primary–sec linked neodisaccharides
by reductive amination.13 In our own approach, the first
results of which are reported in this Letter, we favoured
SN2 type reactions that have the potential to dictate the
stereochemical outcome, that is, the C–N bond forma-
tion should happen stereospecifically. We focussed on
two main approaches: Mitsunobu reactions using sugar
sulfonamide nucleophiles; and epoxide opening reac-
tions using sugar amines as nucleophiles.
We then examined the potential formation of a pri-
mary–secondary linkage, i.e. by reaction between C-6
triflamide 2 and C-3 alcohol 1019 (DIAD (2 + 1 equiv),
PPh3 (2 + 1 equiv), 24 h). However, no reaction was
observed by TLC and virtually all the starting triflamide
was recovered. This difference in outcome is almost
certainly due to the increase in steric bulk in going from
a primary to a secondary alcohol. Repeating the reac-
tion in the presence of methanol gave smooth conver-
sion to the N-methylated product 11 (99%), with no
neodisaccharide formation.
O
TfMeN
O
O
BnO
O
O
BnO
BnO
OMe
O
HO
11
10
We first investigated a Mitsunobu approach.14 Mitsun-
obu chemistry can be used to form C–N bonds when
the NH proton on the nucleophile is sufficiently acidic.15
Thus, we converted a model amine 116 into its triflyl 2
and nosyl 3 derivatives (Scheme 1). Reaction between
equimolar amounts of a C-6 sulfonamide (2 or 3) and
a primary alcohol (417 or 518) under standard Mitsun-
obu conditions (i.e., DEAD or DIAD, PPh3 in THF)
gave good yields of the corresponding (6–6) N-linked
C2-symmetric (6 and 7) or unsymmetrical (8 and 9)
neodisaccharides (Scheme 2). We found that it was nec-
essary to cool the mixture to 0 ꢁC for the addition of the
reagents, but that the reaction could then be allowed to
warm to rt to complete. Attempting the reaction be-
tween 2 and 5 starting at rt failed to give any product–
both the starting materials were recovered unreacted.
O
O
O
O
O
H2N
13
We next turned our attention to an epoxide opening
reaction as a stereospecific route to amine-linked neodi-
saccharides.20 Thus, C-6 (1 and 1221) and C-3 (1322)
amines and allo 14 and manno 15 configured epoxides23
were synthesised according to the published procedures.
Reaction of the manno epoxide 15 with C-6 amines 1
and 12 in the presence of LiClO4 (2 equiv) in refluxing
acetonitrile gave the altro configured neodisaccharides
16 and 17, respectively, with (3–6) connectivity in excel-
lent regioselectivity, this being dictated by the trans
diaxial opening24 of the conformationally locked epox-
ide. The allo epoxide 14 reacted with the C-6 amines 1
and 12 under the same conditions to give the trans
diaxial (2–6) N-linked neodisaccharides 18 and 19,
respectively, once again with excellent regioselectivity
(Table 1).
RO2SHN
H2N
BnO
BnO
O
BnO
(i) OR (ii)
O
BnO
BnO
OMe
BnO
OMe
2: R=CF3
1
3: R=2-(NO2)-C6H4
However, trying to open the epoxides 14 and 15 with
C-3 amine 13 under these conditions to give a sec–sec
N-linked neodisaccharide failed: only unreacted starting
Scheme 1. Reagents and conditions: (i) Tf2O (1 equiv), Et3N (2 equiv),
DCM, 0 ꢁC, 64%; (ii) NsCl (1.2 equiv), DMAP (0.1 equiv), Et3N
(1.5 equiv), DCM, rt, 97%.
O
O
TfN
NsN
BnO
OMe
OMe
HO
BnO
BnO
BnO
BnO
BnO
BnO
BnO
O
(i)
(ii)
BnO
O
BnO
OBn
OBn
+ 2
3 + 4
BnO
O
OMe
4
6
7
BnO
MeO
BnO
MeO
BnO
BnO
O
O
TfN
NsN
OBn
O
OMe
OMe
HO
BnO
BnO
BnO
BnO
BnO
BnO
(iii)
(iv)
BnO
O
BnO
+ 2
3 + 5
OBn
OBn
O
OMe
5
9
8
BnO
MeO
BnO
MeO
Scheme 2. Reagents and conditions: (i) DIAD (2 + 2 equiv), Ph3P (2 + 2 equiv), THF, 0 ꢁC!rt, 7 h, 78%; (ii) DIAD (2 + 0.5 equiv), Ph3P
(2 + 0.5 equiv), THF, 0 ꢁC!rt, 26 h, 78%; (iii) DIAD (3 equiv), Ph3P (3 equiv), THF, 0 ꢁC!rt, 3 h, 87%; (iv) DEAD (2 + 2 equiv), Ph3P
(2 + 2 equiv), THF, 0 ꢁC!rt, 48 h, 78%.

T. Akhtar, I. Cumpstey / Tetrahedron Letters 48 (2007) 8673–8677
8675
Table 1. Reactions carried out with epoxide (1 equiv), amine (1.1 equiv) and LiClO4 (2 equiv) in refluxing acetonitrile
Entry
Epoxide
Amine
Time (h)
Product
Isolated yield (%)
HO
O
OMe
O
O
OMe
OMe
H2N
O
O
O
BnO
BnO
O
HN
1a
1b
45
84
62 (14% epoxide recovered)
79
Ph
Ph
Ph
O
BnO
O
OMe
BnO
O
O
1
15
15
14
16
BnO
BnO
OMe
HO
O
OMe
OBn
OBn
O
H2N
O
O
O
BnO
BnO
O
HN
2
46
58 (17% epoxide recovered)
Ph
O
O
O
BnO
17 BnO
OMe
12
OMe
Ph
O
O
OMe
OMe
H2N
O
O
O
H
N
BnO
BnO
3
72
81
O
Ph
Ph
BnO
O
OMe
HO
MeO
OMe
O
O
BnO
1
BnO
OBn
18
Ph
OBn
O
H2N
O
O
O
BnO
BnO
H
4
24
89
BnO
N
O
O
O
OMe
OMe
19
HO
MeO
12
14
BnO
OBn
materials could be seen on TLC. For the reaction of allo
epoxide 14 with C-3 amine 13, carrying out the reaction
for one week in a sealed tube at 90 ꢁC (2 equiv LiClO4)
gave no detectable product formation. Unfortunately, it
seems that once again, the increase in steric crowding at
a potential reaction transition state was enough to kill
off any reaction and further investigation is necessary
to achieve general formation of N-neodisaccharides
including sec–sec linked examples by this method.
50% conversion). No partially deprotected compounds
were ever observed.
The fully protected altrose derivatives appear to exist
4
predominantly in a C1 chair conformation, as judged
by the NMR 1H–1H 3J coupling constants (e.g.,
for (2–6) N-linked compound 17, J1,2 <1 Hz; J2,3
2.8 Hz; J3,4 2.8 Hz; J4,5 9.6 Hz; and for (3–6) N-linked
compound 16, J1,2 <1 Hz; J2,3 1.4 Hz; J3,4 4.1 Hz; J4,5
9.8 Hz). When the conformationally rigid benzylidene
protection is removed, however, both the 3-substituted
and the 2-substituted altrose residues lose this confor-
mational rigidity in solution, adopting either a single
other or multiple conformations, as indicated by the
NMR coupling constants (e.g., for (2–6) N-linked com-
pound 24, J1,2 4.6 Hz; J2,3 8.2 Hz; J3,4 3.6 Hz; and for
(3–6) N-linked compound 22, J1,2 3.3 Hz; J2,3 5.7 Hz;
J3,4 and J4,5 4.7 and 7.8 Hz).26
Attempted deprotection of some of the compounds (e.g.,
18) by catalytic hydrogenolysis was not satisfactory, and
while product formation was seen, the reactions
proceeded very slowly even when acid was added. How-
ever, all the neodisaccharides 7–8 and 16–19 could be
smoothly deprotected in one step (i.e., removal of benzyl
ethers, benzylidene acetals and sulfonamide protection)
using dissolving metal conditions25 to give the free
neodisaccharides 20–25, all as their bis-methyl
glycosides (Scheme 3). A reverse addition procedure,
where sodium was first added to the condensed ammo-
nia, followed by slow addition of the sugar solution in
THF was found to work best, especially on larger scales.
The low solubility of the protected disaccharides in
ammonia/THF mixtures at ꢀ78 ꢁC meant that in the
alternative procedure in which ammonia was condensed
into a THF solution of disaccharide and then sodium
added, some starting material was often recovered from
the reaction mixture as well as the product (typically
In conclusion, we have investigated two new approaches
to the synthesis of amine-linked neodisaccharides, that is,
sulfonamide-based Mitsunobu chemistry, and the open-
ing of epoxides by carbohydrate amines, and demon-
strated their utility by the formation of some primary–
primary and primary–secondary N-linked structures.
We are currently investigating potential extensions of
these methods to form more challenging primary–sec-
ondary and sec–sec N-linkages, and the results of our
investigations along with the biological results from

8676
T. Akhtar, I. Cumpstey / Tetrahedron Letters 48 (2007) 8673–8677
HO
O
HN
HO
OMe
HO
HO
(i)
OH
8
86%
O
20
HO
MeO
O
HN
OMe
OH
HO
HO
(i)
HO
O
OH
7
62%
21
HO
MeO
HO
HO
O
HO
H
N
O
OMe
(i)
HO
HO
O
(i)
18
HN
HO
HO
OMe
OH
88%
16
17
80%
O
HO
OH
MeO
HO
HO
24
22
OH
OMe
HO
O
HO
HO
OMe
H
N
O
HO
HO
(i)
HO
OH
O
(i)
HN
OH
O
19
OMe
74%
72%
HO
HO
HO
23
MeO
25
OMe
Scheme 3. Reagents and condition: (i) Na, NH3(l), THF, ꢀ78 ꢁC.
testing the glycosidase inhibitory properties of our neodi-
saccharides will be presented elsewhere in due course.
fied by flash column chromatography (pentane/ethyl
acetate).
Typical procedure for the Mitsunobu reaction: Sulfon-
amide (1.0 equiv) and alcohol (1.0 equiv) were dissolved
in THF (3 mL) under N2 with stirring, and the mixture
cooled to 0 ꢁC. Triphenylphosphine (2–3 equiv) was
added, and after 10 min, DIAD (2–3 equiv) was added
slowly. The yellow solution turned into a milky suspen-
sion within 10 min. The ice bath was removed after
30 min, and stirring was continued at room temperature.
After 2 h, the reaction mixture became a clear solution. If
necessary, as judged by TLC, the mixture was recooled to
0 ꢁC and further PPh3 and DIAD added. After TLC
showed the formation of a major product, the reaction
mixture was concentrated and the residue purified by
flash column chromatography (pentane/ethyl acetate).
Typical procedure for deprotection: Ammonia (15–
20 mL) was condensed into a flask at ꢀ78 ꢁC. Sodium
metal (required quantity) was added and the solution
turned deep blue immediately. A solution of the
protected neodisaccharide in THF (2 mL) was added
dropwise followed by MeOH (3–4 drops). The reaction
mixture (a deep blue solution) was stirred at this temper-
ature for 4–5 h, under a N2 atmosphere, then quenched
by the addition of NH4Cl (solid, ca. 400 mg). Ammonia
was allowed to evaporate at rt, then the mixture was
concentrated in vacuo. The crude product was dissolved
in CHCl3/EtOH (1:1) and filtered. This procedure was
repeated twice to get rid of most of the inorganic salts.
The product was then purified by flash column chromato-
graphy (CHCl3/MeOH/AcOH/H2O; 60:30:3:5) to give
the deprotected neodisaccharides as their acetic acid
salts.
Typical procedure for epoxide opening: Amine (1.1 equiv),
epoxide (1.0 equiv) and lithium perchlorate (2.0 equiv)
were dissolved in acetonitrile (3.0 mL), and the reaction
mixture stirred under reflux. After TLC showed
consumption of the epoxide and the formation of one
major product (45–65 h), the reaction mixture was
poured into water (50 mL) and extracted with ethyl ace-
tate (2 · 75 mL). The combined organic extracts were
dried (MgSO4) and concentrated, and the residue puri-
Acknowledgements
The authors would like to thank the Swedish Research
˚
Council (Vetenskapsradet) for financial support. TA is
grateful to the Higher Education Commission of Paki-

T. Akhtar, I. Cumpstey / Tetrahedron Letters 48 (2007) 8673–8677
8677
11. (a) Coxon, B. Carbohydr. Res. 1979, 73, 47–57; (b) Coxon,
B. Carbohydr. Res. 2007, 342, 1044–1054.
12. Kroutil, J.; Budesinsky, M. Carbohydr. Res. 2007, 342,
147–153.
stan for a fellowship under the international research
support initiative program.
13. Neumann, J.; Weingarten, S.; Thiem, J. Eur. J. Org.
Chem. 2007, 1130–1144.
14. (a) Mitsunobu, O.; Yamada, M. Bull. Chem. Soc. J. 1967,
40, 2380–2382; (b) Mitsunobu, O. Synthesis 1981,
1–28.
15. (a) Henry, J. R.; Marcin, L. R.; McIntosh, M. C.; Scola, P.
M.; Harris, G. D.; Weinreb, S. M. Tetrahedron Lett. 1989,
30, 5709–5712; (b) Fukuyama, T.; Jow, C.-K.; Cheung, M.
Tetrahedron Lett. 1995, 36, 6373–6374.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tetlet.
2007.10.039.
16. Tagmose, T. M.; Bols, M. Chem. Eur. J. 1997, 3, 453–
462.
References and notes
17. Garegg, P. J.; Iversen, T.; Oscarson, S. Carbohydr. Res.
1976, 50, C12–C14.
1. Varki, A. Glycobiology 1993, 3, 97–130.
18. Boren, H. B.; Eklind, K.; Garegg, P. J.; Lindberg, B.;
Pilotti, A. Acta Chem. Scand. 1972, 26, 4143–4146.
19. Austin, G. N.; Baird, P. D.; Fleet, G. W. J.; Peach, J. M.;
Smith, P. W.; Watkin, D. J. Tetrahedron 1987, 43, 3095–
3108.
20. Chini, M.; Crotti, P.; Macchia, F. J. Org. Chem. 1991, 56,
5939–5942.
21. Prosperi, D.; Ronchi, S.; Lay, L.; Rencurosi, A.; Russo,
G. Eur. J. Org. Chem. 2004, 395–405.
2. For the state of the art, see: Carbohydr. Res. 2007, 342,
1537–1982 (special issue on Glycomimetics).
3. For example, see: Shing, T. K. M.; Kwong, C. S. K.;
Cheung, A. W. C.; Kok, S. H.-L.; Yu, Z.; Li, J.; Cheng, C.
H. K. J. Am. Chem. Soc. 2004, 126, 15990–15992.
4. Cumpstey, I. Synlett 2006, 1711–1714.
5. Haines, A. H. Org. Biomol. Chem. 2004, 2, 2352–2358.
6. Harding, M.; Hodgson, R.; Majid, T.; McDowall, K. J.;
Nelson, A. Org. Biomol. Chem. 2003, 338–349.
22. Guo, J.; Frost, J. W. J. Am. Chem. Soc. 2002, 124, 10642–
10643.
23. Robertson, G. J.; Griffith, C. F. J. Chem. Soc. 1935, 1193–
1197.
7. (a) Bianchet, M. A.; Ahmed, H.; Vasta, G. R.; Amzel, L.
M. Proteins 2002, 40, 378–388; (b) Leffler, H.; Barondes,
S. J. Biol. Chem. 1986, 261, 10119–10126.
8. Tejler, J.; Skogman, F.; Leffler, H.; Nilsson, U. J.
Carbohydr. Res. 2007, 342, 1537–1982.
24. Furst, A.; Plattner, P. A. Helv. Chim. Acta 1949, 32,
¨
275.
9. Maxwell, V. L.; Evinson, E. L.; Emmerson, D. P. G.;
Jenkins, P. R. Org. Biomol. Chem. 2006, 2724–2732.
10. For a recent example, see: Neimert-Andersson, K.; Sauer,
S.; Panknin, O.; Borg, T.; So¨derlind, E.; Somfai, P. J. Org.
Chem. 2006, 71, 3623.
25. Iserloh, U.; Dudkin, V.; Wang, Z. G.; Danishefsky, S. J.
Tetrahedron Lett. 2002, 43, 7027–7030.
26. Lichtenthaler, F. W.; Mondel, S. Carbohydr. Res. 1997,
303, 293–302.