Highly diastereoselective additions to polyhydroxylated pyrrolidine cyclic imines: Ready elaboration of aza-sugar scaffolds to create diverse carbohydrate-processing enzyme probes

Article abstract of DOI:10.1002/chem.200304718

Representative diastereomeric, erythritol and threitol polyhydroxylated pyrrolidine imine scaffolds have been rapidly elaborated to diversely functionalized aza-sugars through highly diastereoselective organometallic (RM) additions (R = Me, Et, allyl, hex

Full text of DOI:10.1002/chem.200304718

FULL PAPER
Highly Diastereoselective Additions to Polyhydroxylated Pyrrolidine Cyclic
Imines: Ready Elaboration of Aza-Sugar Scaffolds To Create Diverse
Carbohydrate-Processing Enzyme Probes
Timothy M. Chapman,[a] Steve Courtney,[b] Phil Hay,[b] and Benjamin G. Davis*[a]
Abstract: Representative
diastereo-
are controlled by the configuration of
the C-2 centre adjacent to the azome-
thine imine carbon and the conforma-
tion of the pyrrolidine imine. The high
potential of this method was demon-
strated by concise syntheses of 1-epi-
and 2-epi-desacetylanisomycins. In ad-
dition, the late stage addition of hydro-
phobic substituents, which this imine
addition methodology allows, enabled
the preparation of novel aza-sugars with
enhanced inhibitory potential. This was
highlighted by the screening of a repre-
sentative selection of these ™hydrophob-
ically-modified∫ aza-sugars against a
diverse panel of 12 non-mammalian
and human carbohydrate-processing en-
zymes. This identified a novel nanomo-
lar a-galactosidase inhibitor (IC50 ˆ
250nm) and a novel highly selective
glucosylceramide synthase inhibitor
(IC50 ˆ 52mm, no a-glucosidase inhibi-
tion at 1mm). Furthermore, analysis of
the structure activity relationships of
racemic series of inhibitors allowed
some validation of Fleet×s mirror-image
enzyme active site postulate.
meric, erythritol and threitol polyhy-
droxylated pyrrolidine imine scaffolds
have been rapidly elaborated to diverse-
ly functionalized aza-sugars through
highly diastereoselective organometallic
(RM) additions (R ˆ Me, Et, allyl, hex-
enyl, Ph, Bn, pMeO-Bn). The yields for
these additions have all been substan-
tially enhanced from previously opti-
mised levels (<58%) for normal addi-
tions using a reverse addition procedure
(e.g. R ˆ Ph; 44% normal mode ! 78%
reverse mode). The high diastereoselec-
tivities (>98% de for all except R ˆ
Me) are consistent with additions that
Keywords: asymmetric synthesis
¥
azasugars ¥ carbohydrates ¥ enzyme
catalysis ¥ heterocycles
carbohydrate mimic, and this approach has yielded potent
inhibitors. For example, deoxynojirimycin (DNJ), the direct
configurational counterpart of 1-deoxy-d-glucopyranose is a
potent a-glucosidase inhibitor.[5] Interestingly, however, this
type of approach does not always result in high inhibition.
Some striking exceptions exist: for example, deoxyrhamnojir-
imycin (LRJ)[6] is a poor inhibitor of a-l-rhamnosidase, and
while deoxymannojirimycin (DMJ) is a potent inhibitor of
class I a-d-mannosidases[7] it is a poor inhibitor of class II a-d-
mannosidases.[7b] In addition, in certain cases, five-membered
ring aza-sugars have been shown to give rise to higher
inhibition than their six-membered ring counterparts[8] and
subtle selectivities may be observed for five- versus six-
membered ring systems. For example, the five-membered ring
isomer of DMJ is a potent class II a-d-mannosidase inhib-
itor.[8] Therefore, it is not generally straightforward to predict
by an entirely logical design based upon configurational
analogy the structure of the best inhibitor for a given
carbohydrate-processing enzyme.
Introduction
The syntheses of aza-sugars, sugar mimics in which the ring
oxygen has been substituted by a nitrogen atom, have been
the subject of much continued interest over the last 25 years;[1]
they have proved to be highly potent enzyme inhibitors,
especially of carbohydrate-processing processing enzymes,
and have been used as invaluable probes of the nature and
mechanism of action of many enzymes.[2] For example, the
aza-sugar N-butyldeoxynojirimycin (Zavesca), a potent glu-
cosylceramide synthase inhibitor is now approved for use in
Europe for the treatment of the glycolipid storage disease,
type I Gaucher disease.[3][4] Most syntheses of aza-sugars have
focused on logical designs based on the stereochemistry of the
functional groups around the heterocyclic ring of the putative
[a] Dr. B. G. Davis, T. M. Chapman
Dyson Perrins Laboratory, Department of Chemistry
University of Oxford, South Parks Road, Oxford OX1 3QY (UK)
Fax : (‡44) 1865 275674
The need for precise probes of carbohydrate-processing
enzymes continues to build enormously. The burgeoning field
of chemical genetics[9] demands broad-ranging probes and
inhibitors of key enzymes to provide information on central
biosynthetic glycosylation processes, often where little or no
E-mail: Ben.Davis@chem.ox.ac.uk
[b] Dr. S. Courtney, Dr. P. Hay
Oxford GlycoSciences Ltd, The Forum, 86 Milton Park
Abingdon, Oxfordshire OX14 4RY (UK)
Chem. Eur. J. 2003, 9, 3397 3414
DOI: 10.1002/chem.200304718
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3397

B. G. Davis et al.
FULL PAPER
functional or structural information on putative protein
targets exists.
tions to cyclic sugar imines are scarce, and such reactions of
aza-sugar imines have been essentially limited to the iminor-
ibitol 1 (Figure 1). Early studies with 1 outlined the synthetic
potential of such an approach;[14] Furneaux et al. have since
made very valuable progress in nucleophilic additions to 1 as a
way of synthesising a wider range of potential nucleoside
hydrolase inhibitors.[15] However, in this system yields have
generally been only moderate to fair (35 64% for Grignard
reagents,[14, 15d] 44 63% for organolithium reagents[15] from
precursor amine). It has been previously suggested that such
reactions prove challenging as a result of the low inherent
electrophilicity of the azomethine carbon atom and the
tendency for competing side reactions, such as deprotonation
of tautomerisable imines.[13] Although a number of strategies
As a solution to the need for more varied potential
carbohydrate-processing probes, we have sought to design a
general method that will allow the generation and screening
of a broad range of potential inhibitors. This method retains
three key structural features that have typically generated
inhibitors: a five-membered ring nitrogen heterocycle core,
polyhydroxylation and a hydrophobic group at or near
putative aglycone binding sites. By varying hydroxyl stereo-
chemistries and functional groups within this central aza-
sugar scaffold design we sought to investigate their potential
for inhibition of a number of carbohydrate-processing en-
zymes. This required a synthetic approach that would lend
itself easily to the synthesis of diverse arrays of aza-sugars and
allow the potential introduction of a wide range of substitu-
ents to an aza-sugar scaffold.
As part of our strategy, we have generated potential
inhibitors in racemic series. This not only provides a greater
potential for hits in enzyme inhibition from either aza-sugar
enantiomer, but also allows the exploration of the concept of
™mirror-image∫ enzyme active sites. There are several exam-
ples of mirror-image enzyme active sites in nature,[10] and
Fleet has postulated that this may be a general structural
theme for many carbohydrate-processing enzymes glycosi-
dases;[11] thus, aza-sugars that are enantiomers of one another
are able to inhibit the enzymes that process the natural sugars
that are also enantiomers of one another. For example, l-
rhamnose is the enantiomer of 6-deoxy-d-mannose; Fleet
successfully predicted that l-swainsonine would be a powerful
inhibitor of l-rhamnosidase by analogy with d-swainsonine a
strong inhibitor of d-mannosidase,[11] despite the fact that no
3D structure of a rhamnosidase exists.
ˆ
for enhancing the reactivity of the C N bond of imines for the
subsequent addition of organometallics exist, for example
N-alkylation,[16] N-oxidation,[17] N-acylation[18] or N-sulfony-
lation[19] to produce more reactive iminiums, nitrones, acyli-
mines and sulfonimines, respectively, these require removal of
these activating groups, which can be difficult, and the
presence of the N-substituent may adversely affect diaster-
eoselectivities.[20] Moreover, the installation of an activating
ˆ
N-substituent may not lend itself to iterative rounds of C N
formation and nucleophile addition.
We therefore chose to realize the potential of this powerful
methodology by optimising the yields of such additions, in
particular the additions of Grignard reagents, to parent imines
bearing no N-substituent. Three model imines were chosen,
based on a five-membered ring scaffold with 2,3-dihydroxy
functionality, the threitol imine 2, and those in the epimeric
series, the erythritol imines 3 and 4 (Figure 1). These
diastereomeric imines importantly allowed the effect of
H
Previously, many syntheses of aza-sugars have relied on the
introduction of substituents to the scaffold at an early stage,
and therefore are limited in the range and number of
substituents that may be readily introduced. Aza-sugars
carrying hydrophobic substituents have shown, in some cases,
increased enzyme inhibition and increased bioavailability.[12]
Guided by these tantalizing indications we chose to inves-
tigate the introduction of such hydrophobic groups at a late
stage to allow greater flexibility in the creation of a diverse
array of potential enzyme probes.
TBSO
N
N
N
N
N
O
O TBSO
OTBS
O
O
O
O HO
OAc
OMe
1
2
3
4
5
Figure 1.
structural factors, such as hydroxyl group stereochemistries
and protecting group nature, on efficiency,[21] diastereoselec-
tivity and resulting biological activity to be probed in a
systematic way. Some initial aspects of this work have
previously appeared in a brief communication.[22] Addition-
ally, these imines allow access to scaffolds similar to those
found in natural products of known activity, such as aniso-
mycin 5, a clinically proven peptidyl-transferase inhibitor,
which exhibits strong and selective activity against pathogenic
protozoa and fungi.[23] This paper demonstrates that this
methodology may be applied to the addition of a variety of
functional groups with excellent diastereoselectivities and in
good to excellent yields beyond those previously observed for
cyclic imines. In this way, structure activity relationships for
a wide variety of carbohydrate-processing enzymes can be
readily derived from the resulting deprotected aza-sugar
products, including the identification of novel strong inhib-
itors.
The synthesis of functionalized aza-sugars by addition of
organometallic nucleophiles to cyclic imines: Our strategy
suggested the late convergent addition of substituents to an
aza-sugar scaffold. Introduction of substituents through
addition of a nucleophile to a cyclic imine is a potentially
powerful but largely unexplored methodology that by virtue
of an almost biomimetic convergence allows assembly of
probes in a manner that divides nicely into component blocks
of sugar mimic and aglycone mimic.
Due to the irreversible nature of their addition to imines,
thereby allowing kinetic control of selectivity, we focused on
the addition of organometallic reagents, and in particular
Grignard reagents, as a source of nucleophile. Whilst there are
many examples of additions of organometallic reagents to
acyclic imines reported in the literature,[13] reports of addi-
3398
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
diazabicyclo[5.4.0]undec-7-ene (DBU) ensured smooth elim-
ination to yield the desired threitol imine 2. The presence of
Results and Discussion
ˆ
Preparation of the threitol imine: Threitol imine 2 was
obtained from its iminothreitol precursor 9. This was itself
synthesized in two ways to illustrate two methods of
construction (Scheme 1): firstly from the empty, unfunction-
alized scaffold 3-pyrroline (6), and secondly via the manip-
ulation of an existing dihydroxy scaffold from the chiral
starting material, tartaric acid (11).
the imine C N bond was confirmed by characteristic IR [nÄmax
1645 cmÀ1], 1H NMR [dH (CDCl3): 7.41 ppm (pt, 1H, J ˆ
2.2 Hz)] and 13C NMR [dC (CDCl3): 169.0 ppm] resonances.
Although the imine 2 was isolated for characterisation, better
yields were obtained in subsequent reactions through in situ
imine formation: the reaction mixture, upon imine formation
as judged by TLC, 1H NMR and IR, was immediately filtered
to remove DBU ¥ HCl, then reagents added to the resulting
imine solution.
CBZ
N
Cl
N
CBZ
N
H
N
H
N
i), ii)
iii)
iv), v)
x)
Preparation of erythritol imines: Starting with l-arabinose
(14), the erythritol imines 3 and 4 were prepared as depicted
in Scheme 2. l-Arabinose was selectively 3,4-O protected as
its acetonide 15 using 2,2-dimethoxypropane according to the
procedure of Kiso and Hasegawa[26] as adapted by Thompson
et al.,[27] which was, without purification, reacted with sodium
periodate. The resulting cleavage of the 1,2-diol unit followed
by basic hydrolysis and subsequent ring closure afforded 2,3-
O-isopropylidene-l-erythrose (16) in 57% yield (a:b 1:6) in
two steps from 14.[27] Subsequent reduction with sodium
borohydride yielded the erythritol 17 in quantitative yield;
this was followed by treatment with mesyl chloride and
triethylamine in dichloromethane to afford the dimesylate
18.[28] Reaction with benzylamine led to cyclisation giving the
protected erythritol 19. Hydrogenolytic removal of N-benzyl
protection yielded the acetonide protected iminoerythritol 20.
This compound is unusually volatile, and care is required in its
isolation. As for 9, N-chlorination (86%) and subsequent
elimination with DBU sucessfully provided the desired
erythritol cyclic imine 3.
OTBS
O
7
OH TBSO
OTBS TBSO
viii), ix)
HO
9
10
6
8
xi)
Bn
Bn
N
N
N
O
O
O
O
HO2C
HO
OH
CO2H
vi)
vii)
OTBS
HO
OH
TBSO
OTBS TBSO
11
12
13
2
Scheme 1. i) C6H5CH2OCOCl, NaOH, toluene, 93%; ii) mCPBA, CH2Cl2,
46%; iii) 2m H2SO4 (aq.), Et2O, 87%; iv) TBDMSOTf, py, CH Cl2, 92%;
v) H2, Pd/C, MeOH, 92%; vi) BnNH2, xylene, reflux; vii) TBDMSCl,
imidazole, DMF, 43% over two steps from tartaric acid; viii) BH3 ¥ Me2S,
THF then MeOH, 408C, 100%; ix) H2, Pd/C, MeOH, 100%; x) NCS, Et2O,
96%; xi) DBU, Et2O.
2
Starting at 3-pyrroline 6, the iminothreitol 9 was prepared
through N-protection as its Z-carbamate then epoxidation,
followed by ring opening of the epoxide 7 in aqueous sulfuric
acid to yield the trans hydroxylated pyrrolidine 8. Subsequent
protecting group manipulation, di-O-protection using
TBDMSOTf and N-deprotection through hydrogenolysis,
afforded the iminothreitol 9, obtained in 32% overall yield
from 3-pyrroline (6) over five steps. A second route starting
from cheap, readily available racemic tartaric acid was also
used. Reaction of tartaric acid (11) with benzylamine in
refluxing xylene led to cyclisation to yield the imide 12, which
without purification was reacted with tert-butyldimethylsilyl
chloride according to the procedure of Ryu and Kim[24] to
afford the di-O-silyl ether-protected imide 13 in 43% yield
from tartaric acid (11). Reduction of imide 13 was achieved by
reaction with borane/dimethylsulfide complex yielding the
corresponding borazine as the sole product, which was
cleaved to yield N-benzylpyrrolidine by heating overnight in
methanol at 408C (100% yield from 13); this borazine-
cleavage method provides a convenient alternative to other
methods.[25] Removal of the N-benzyl group by catalytic
hydrogenation over palladium-on-carbon yielded the imino-
threitol 9 in 43% yield over four steps from tartaric acid,
identical in all respects to that synthesized from 6.
Scheme 2. i) p-TsOH, dimethoxypropane, DMF; ii) NaIO4, H2Othen
Na2CO3, 57% for 16, 57% for 23 over two steps from l-arabinose;
iii) NaBH4, MeOH, 100% for 17, 89% for 24; iv) MsCl, Et3N, CH2Cl2, 72%
for 18, 85% for 25; v) BnNH2, 658C, 76% for 19, 87% for 26; vi) H2,
Pd(OH)2/C, MeOH, 98% for 20, 97% for 27; vii) NCS, Et2O, 86% for 21,
90% for 28; viii) DBU, Et2O.
With two ready routes to 9 in hand, we focused on the
elaboration of 9 to its corresponding aldimine 2. Treatment of
iminothreitol 9 with N-chlorosuccinimide (NCS) yielded the
N-chloroamine 10 in excellent yield (96%). For the elimi-
nation of 10, a number of bases were surveyed, which caused
either decomposition, dechlorination (LDA, KHMDS) or no
reaction (Et3N, DABCO). However, treatment with 1,8-
An essentially parallel route employing 3,4-O-cyclohexyli-
dine protection gave protected iminoerythritol 27 in 36%
yield over six steps from 14. Compound 27 proved to be less
volatile than 20. Elaboration through N-chlorination (90%)
and subsequent elimination with DBU proceeded smoothly
Chem. Eur. J. 2003, 9, 3397 3414
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3399

B. G. Davis et al.
FULL PAPER
and yielded the alternatively protected erythritol cyclic
imine 4.
These routes using 14 as starting material were preferred to
alternative routes employing moderate yielding cis-dihydrox-
ylation of the Z-carbamate of 3-pyrroline (6).[29]
Screening of organometallics for addition to cyclic imines:
Having obtained access to the cyclic imines 2 4, we screened
for the addition of Grignard reagents as representative, model
nucleophiles (Table 1). Although imines 2 4 could be
isolated, better yields were obtained if the Grignard reagents
Table 1. Organometallic additions to aza-sugar pyrrolidine imines 2 4.[a]
Imine
Base
Equiv
RM
Equiv
Order of
Addition
Product
Yield[%][a]
Conditions
/
de[b]
1
2
3
4
5
6
7
8
9
threitol 2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
Et3N
DABCO9.0
LDA
DBU
DBU
DBU
DBU
PhMgBr
PhLi
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
DBU
3.0
3.0
3.0
1.0
3.0
3.0
3.0
4.0
3.0
3.0
3.0
3.0
12.0
MeMgBr
EtMgBr
5.0
5.0
5.0
1.5
5.0
5.0
5.0
5.5
5.0
5.0
5.0
5.0
normal
normal
normal
normal
normal
normal
normal
normal
normal
normal
normal
normal
29a
29b
29c
29d
29e
29f
29d
29d
29d
29d
29d
29d
58
45
26
37
48
43[c]
44
44
96
85
AllylMgBr
PhMgBr
BnMgCl
pMeOBnMgCl
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
> 98
> 98
> 98
> 98
> 98
> 98
> 98
> 98
44
40[d]
38[e]
35[f]
16[g]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
1.1
3.0
3.0
3.0
3.0
2.2
2.2[g]
1.1
1.1
1.1
1.1
1.1
1.1
1.3
1.3
1.3
1.3
1.1
1.3
1.3
1.3
3.0
3.0
3.0
3.0
3.0
1.3
PhMgBr
EtMgBr
PhLi
5.0
5.0
5.0
5.0
normal
normal
normal
normal
29d
29b
29d
29b
5[h]
10[h]
39
26[d]
0
> 98
> 98
> 98
89
2
2
2
2
Et2Zn
2
0
15
erythritol 3
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
PhMgBr
Ph2Mg
EtMgBr
BnMgCl
pMeOBnMgCl
MeMgBr
EtMgBr
PhMgBr
BnMgCl
pMeOBnMgCl
EtMgBr
3.0
2.0
3.0
3.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
normal
normal
normal
normal
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
reverse
30d
30d
30d
30d
30d
30d
30d
30d
30d
30d
30d
30b
30e
30f
> 98
> 98
> 98
3
3
3
3
3
3
3
3
3
3
3
3
3
2
2
2
2
2
4
12
25[h]
0[i]
56
> 98
> 98
> 98
> 98
> 98
> 98
> 98
> 98
> 98
> 98
66
> 98
> 98
> 98
> 98
> 98
23[j]
62[k, l]
70[k, m]
70[k, n]
68[k, o]
55[p]
76
73
76[c]
67
70
78
63
62[c]
76
29a
29b
29d
29e
29f
31b
[a] Yields quoted are determined over two-step elimination addition sequence from corresponding chloramines. [b] As determined by 1H NMR;
configuration determined by NOESYor 1D NOE NMR spectroscopy. [c] Grignard formation and addition carried out in THF. [d] Pre-treatment of the imine
solution with 0.95 equiv TMSOTf for 30 min prior to addition of the organometallic. [e] Addition of hexanes upon imine formation to encourage
precipitation of DBU ¥ HCl prior to filtration. [f] Imine formation and Grignard addition carried out in THF as solvent. [g] Addition of Grignard carried out
at À788C, then reaction mixture allowed to warm to RT before quenching. [h] Aqueous work-up performed on imine, organic extract dried over MgSO4,
filtered and dissolved in fresh dry Et2Oprior to treatment with Grignard reagent. [i] Addition of Grignard and reaction quenched at À788C. [j] No filtration
to remove DBU ¥ HCl prior to Grignard addition. [k] Reaction of imine derived from 100 mg of chloramine starting material. [l] Addition of 0.14m imine
solution in of dry Et2Oto 0.96 m Grignard reagent in dry Et2O. [m] Addition of 0.07m imine solution in of dry Et2Oto 1.96 m Grignard reagent in dry Et2O.
[n] Addition of 0.047m imine solution in of dry Et2Oto 3.0 m Grignard reagent in dry Et2O. [o] Addition of 0.093m imine solution in of dry Et2Oto 3.0 m
Grignard reagent in dry Et2O. [p] Addition of dioxane to Grignard reagent solution prior to addition of imine solution.
3400
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
were added to imine generated in situ (see above). Pleasingly,
addition of Grignard solution to a solution of threitol imine 2
(so-called ™normal∫ mode, entries 1 6) successfully yielded
the desired adducts 29a f, although the yields in all cases
were low to moderate (26 58%). The observed diastereose-
lectivities ranged from fair to excellent (de 44 > 98%); all
nucleophiles showed a strong preference for the formation of
anti diastereomers, the relative configurations of which were
assigned by NOE NMR experiments. A marked dependence
of this selectivity on nucleophile bulk was observed. Thus, for
larger nucleophiles (R ˆ Ph, Bn), including even the smaller
ethyl, de values greater than 98% were observed, whereas for
allylmagnesium bromide (85% de) and methylmagnesium
bromide (44% de) poorer preferences were observed. These
results are consistent with an addition selectivity that is
governed by steric approach control (see below).
Thus, despite these extensive variations in conditions
adduct yields remained in the 40% region. It was considered
that the high basicity of Grignard reagents might lead to
unwanted side reactions such as elimination or tautomerisa-
tion, and therefore other potentially less basic and more
azophilic organometallics, phenyllithium and diethylzinc were
also investigated. Disappointingly, these yielded the respec-
tive adducts in only 39 and 26% yields, respectively (entries
18 and 19).
Corresponding variations in the addition of phenylmagne-
sium bromide to the erythritol imine 3 were similarly
disappointing with the best yield of 30d (25%), obtained by
addition of Grignard reagent at À788C and allowing the
mixture to warm to room temperature. It should be noted
however that such moderate yields were consistent with those
previously reported in the literature, which have been in the
range 35 64%.[14, 15d]
Interestingly, the major side product of the addition
reactions were the corresponding amines 9 and 20; the
formation of which is difficult to explain by a polar addition
mechanism. Reductive b-hydride transfer reactions[32] by
Grignard reagents are possible only for ethylmagnesium
bromide and yet 9 and 20 were observed for all Grignards.
Radical mechanisms have been reported in the literature for
many Grignard reactions,[33] resulting in addition, reduction
and dimerisation products, and radical mechanisms for the
Optimisation of organometallic additions: In an attempt to
improve yields of these promisingly diastereoselective reac-
tions, optimisation of the additions of phenylmagnesium
bromide to the imines 2 and 3 were investigated as model
systems. It was found, via analysis of the formation of imine 2
by 1H NMR spectroscopy, that 1.0 equivalent of DBU was not
sufficient to provide complete imine formation. However, an
increase in the amount of DBU to 3.0 equivalents while
allowing more complete imine formation only led to a small
increase in overall yield of resulting adduct (Table 1, entry 7).
A further increase to 4.0 equivalents gave no further benefit
(entry 8). Treatment of the N-chloramine with the bases
triethylamine and diazabicyclooctane (DABCO) instead of
DBU led to no imine formation, even after the adition of 12.0
and 9.0 equivalents respectively (entries 13 and 14). Similarly,
treatment with 1.1 equivalents of lithium disopropylamine
(LDA) led to no imine formation; instead the amine 20 was
formed (entry 15) presumably through halogen-metal ex-
change and quench during work-up. It has been reported that
pre-treatment of acyclic imines with the Lewis acid trime-
thylsilyltrifluoromethane sulfonate (TMSOTf), to activate the
imine to addition by formation of the TMS-iminium salt has,
in some cases, led to inceased yields of addition.[30] However,
pre-treatment of 2 with 0.95 equivalents of TMSOTf led only
to a reduced yield of adduct 29d (40%). The addition of
hexane upon imine formation, to encourage precipitation of
DBU ¥ hydrochloride, yielded 29d in 38% (entry 10). The use
of tetrahydrofuran in imine formation and subsequent
addition similarly failed to enhance efficiency and resulted
in a yield of only 35% of 29d (entry 11). Temperature effects
were also briefly investigated; addition of PhMgBr to a
solution of imine cooled to À788C followed by warming to
room temperature yielded the adduct 29d in only 16%
(entry 12). Purification of the imine by an aqueous work-up
procedure to remove excess DBU,[31] followed by addition of
Grignard led only to a much lower yield of adducts 29d, b
(5% for phenylmagnesium bromide, 10% for ethylmagnesi-
um bromide, entries 16 and 17, respectively). It is possible that
this may be attributed to the intermediate formation of a
hydrated, hemiaminal form of imine 2 on exposure to water,
which might be unreactive to subsequent Grignard addition.
ˆ
reduction of C N bonds by other reagents have also been
reported.[34] Such a radical mechanism might also account for
the formation of amines 9 and 20 from imines 2 and 3,
respectively. Consistent with this possibility, a dimerisation
product 32 was isolated as a consistent side product (8 10%),
and suggested at least a partial radical character for additions.
Indeed, the possibility of polar-radical mixed mechanisms in
additions of Grignard reagents to acyclic imines has been
noted previously.[35]
To test this possibility we used the Grignard reagent
hexenylmagnesium bromide as a probe in additions to imine
3, in which cyclisation of the hexenyl side chain during the
addition process would be indicative of a radical process.[36]
Uncyclized hexenyl adduct 30g was the sole product in 46%
yield; this inconclusive result does not preclude a radical
process, which may have a rate greater than ꢀ105 sÀ1[37] or,
alternatively, addition by any cyclized cyclopentylmethyl
radical may compete unsuccessfully with addition by uncycl-
ized hexenyl.
H
NH HN
N
TBSO
TBSO
OTBS
OTBS
O
O
32
30g
Figure 2.
Investigations into a reverse addition procedure for the
organometallic additions: Although the potential radical
nature of the additions remained unconfirmed, the formation
of dimer 32 nonetheless highlighted the propensity of imine
substrates in such homo-coupling, perhaps through Mg-
Chem. Eur. J. 2003, 9, 3397 3414
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3401

B. G. Davis et al.
FULL PAPER
chelated, pinacol reduction.[38] It was considered that addition
of the imine to the Grignard, rather than Grignard to imine,
might counter this problem. Such a reverse addition, by
presenting a high Grignard concentration to a low imine
concentration rather than vice versa, would likely curtail such
dimerization side reactions that are more dramatically
dependent on imine concentration.
addition reaction of EtMgBr with cyclohexylidene-protected
erythritol imine 4, which yielded adduct 31b in similarly good
yield (76%) and excellent de (>98%).
Diastereoselectivity in additions: Excellent diastereoselectiv-
ities were observed in almost all of the additions; for the
larger nucleophiles (R= Me, allyl) de values in excess of 98%
were observed. The erythritol imines 3,4 generated adduct de
values of greater than 98% for all nucleophiles regardless of
the mode of addition. This is consistent with the expected[40]
conformation of this imine, in which the acetonide group is
likely to enforce a rigid butterfly conformation that favours
nucleophile approach from the convex face (Figure 3). The
similarly excellent de values (>98%), similar yields and rates
for the addition of the ethyl nucleophile EtMgBr to both the
acetonide and cyclohexylidene protected imines 3 and 4,
respectively, are therefore consistent with the model shown in
Figure 3. In the case of TBDMS-protected threitol imine 2,
the bulk of the O-2-silyl group appears to be a key determin-
ing factor by limiting approach of the nucleophile to the trans
a-face of the imine (Figure 3).
Conditions for reverse addition were investigated using the
erythritol imine 3 with PhMgBr as a nucleophile. Imine
formation, with 1.1 equivalents of DBU, followed by the
addition of imine solution to phenylmagnesium bromide led
to a significant improvement in yield, and 30d was obtained in
56% yield (entry 26) compared with 15% using the previous
™normal∫ Grignard-to-imine order of addition. Reverse
addition therefore appeared to be a key factor. Other
potentially important yield determining factors were then
studied: filtration to remove DBU ¥ hydrochloride formed
during imine synthesis was also shown to be an important
factor, since the omission of filtration resulted in only a 23%
yield of adduct 30d (entry 27). As for ™normal∫ mode
additions (see above), the use of more DBU (here 1.3 equiv-
alents) resulted in a concomitant increase in yield of 30d, up
to 62%. Finally, the absolute concentrations of imine and
Grignard solutions were optimised. The use of a relatively
dilute imine solution (ꢀ0.05m) added to a concentrated
Grignard solution (ꢀ3m) gave the best yield of all achieved at
70% (entries 29 and 30), some six-fold improved over initial
yields (12%, entry 23) in this system.
R
O
N
TBSO
TBSO
TBSO
TBSO
O
N
N
R
H
H
2
3: R2C = Me2C
4: R2C = C6H10
Addition of dioxane is known to affect the form of
organomagnesium reagents in solution by altering the posi-
tion of the Schlenk equilibrium, which favours the dialkyl-
magnesium species with concomitant precipitation of
MgX2.[39] The resulting R2Mg species is less basic and might
lead to decreased eliminative side reactions. However,
addition of a solution of imine 3 to an unfiltered Ph2Mg
solution yielded 22d in only 55% yield (entry 32); no
improvement over the reaction conducted without added
dioxane.
The application of the reverse addition methodology was
extended to the other Grignard reagents ethylmagnesium
bromide and benzylmagnesium chloride, in good yields of 76
and 73% for 30b and e respectively. In view of the much
improved yields obtained for the reverse addition of the
erythritol imine 3, this strategy was also investigated for the
threitol imine 2 (entries 36 40). Gratifyingly, these yields
were also much improved. For example, using the reverse
rather than the normal mode of addition the yield of 29d from
the reaction of phenylmagnesium bromide with 2 improved
from 44% (entry 7) to 78% (entry 38). To the best of our
knowledge these represent some of the best yields of
additions of organometallic nucleophiles to cyclic imines yet
reported. It should be noted that the diastereoselectivities
were also consistently good to excellent (>98% de for all
except 66% de for 29a), and indeed the de for the least
selective addition, the addition of methylmagnesium bromide
to 2 to give 29a, was improved from 44% (entry 1) for the
normal addition to 66% (entry 36) for the reverse addition
procedure. The compatibility of the method with other
protecting groups was also briefly explored through the
Figure 3.
Deprotection of adducts 29b,d and 30b,d: Ethyl and phenyl
adducts 29b,d and 30b,d obtained from additions to both the
threitol 2 and erythritol 3 imines were deprotected to yield
their corresponding free aza-sugars 33b,d and 34b,d, respec-
tively (Scheme 3). Treatment of 29b,d with tert-butylammo-
nium fluoride (TBAF) led to a successful cleavage of the silyl
ether groups, but separation of the products from residual
TBAF proved difficult. However, in all cases hydrolysis with
aqueous trifluoroacetic acid followed by ion-exchange chro-
matography proved successful (Scheme 3).
H
N
H
N
R
R
29b: R = Et i) or ii)
29d: R = Ph
33b: R = Et
33d: R = Ph
TBSO
OTBS
R
HO
HO
OH
H
N
H
N
R
ii)
30b: R = Et
34b: R = Et
34d: R = Ph
30d: R = Ph
OH
O
O
Scheme 3. i) 50% TFA (aq.), 76% for 33b; ii) 25% TFA (aq.), THF then
Dowex H (NH4OH (aq.)), 63% for 33d, 96% for 34b, 64% for 34d.
‡
Synthesis of anisomycin analogues: The methodology of
addition of organometallics to cyclic imines offers potential
access to interesting aza-sugar derivatives, such as analogues
of the natural product anisomycin (5). Several syntheses of
anisomycin have been reported in the literature,[20a],[41] but the
3402
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
Table 3. Aza-sugar inhibition of human glycosidases.
potential activity of anisomycin and its analogues towards
glycosidases has been poorly explored.
Glycosidase
Inhibition [%]
33d
33b
33 f
34d
34b
34 f
To this end, we investigated the introduction of the
4-methoxybenzyl side chain of anisomycin through reaction
of the threitol 2 and erythritol 3 imines with 4-methoxyben-
zylmagnesium chloride (Scheme 4). These yielded the
TBDMS-protected threitol adduct, 29 f, in 43% yield for the
normal addition mode (Table 1, entry 6), and 62% for the
reverse addition mode (entry 40); both with excellent dia-
stereoselectivity (>98%). The erythritol adduct analogue 30 f
was obtained in similarly good yield (76%) and excellent
diastereoselectivity (>98% de, entry 35). Interestingly, the
use of THF solutions for the generation and use of
pMeOBnMgCl proved critical, use of Et2Oat any stage failed
to yield any adduct. Deprotection of 29 f was conducted using
essentially analogous conditions to those used for 29b and d to
yield the unnatural analogue 1-epidesacetylanisomycin (33 f,
61%; 17% overall yield from tartaric acid). This methodology
compares well in terms of the overall yields[42] reported for
syntheses of desacetylanisomycin and highlights the high
synthetic utility of the imine addition methodology. The
2-epidesacetylanisomycin diastereomer 34 f, similarly an un-
natural analogue, was synthesized in an essentially analogous
way (Scheme 4).
a-d-glucosidsase[a]
b-d-glucosidase[a]
b-d-glucosidase[b]
(non-lysosomal)
N.I.
51.5
8.7
N.I.
23.2
11.7
9.3
39.3
N.I.
3.0
38.1
11.3
20.0
13.0
57.0
a-d-galactosidase[a]
b-d-galactosidase[a]
N.I.
N.I.
N.I.
N.I.
N.I.
N.I.
N.I.
N.I.
N.I.
N.I.
N.I.
N.I.
[a] At 1mm. [b] At 1mm.
Table 4. Inhibition of human glucosylceramide synthase (GCS).
Enzyme
Inhibition [%]
33d
33b
33 f
34d
34b
34 f
glycosylceramide synthase[a]
26.5
48.1
27.6
0
0
0
[a] At 50mm.
Broad-ranging assessment of the inhibition of all of the
deprotected ™hydrophobically-modified∫ pyrrolidine aza-sug-
ars 33b, d, f and 34b, d, f towards non-mammalian glycosi-
dases revealed three broad trends: i) all the iminothreitol aza-
sugars 33, except for phenyl-substituted 33d which showed no
inhibition towards a-d-mannosidase, showed relatively potent
inhibition of a-l-rhamnosidase and a-d-mannosidase, with Ki
values in the low micromolar range (3.0 16.7mm); ii) all the
erythritol sugars 34, except for phenyl-substituted 33d which
showed no inhibition towards a-d-galactosidase, showed
relatively potent inhibition of a-l-fucosidase and a-d-galac-
tosidase, with Ki values in the sub-micromolar to low micro-
molar range (0.25 9.0mm); iii) several aza-sugars from both
the threitol (Et- and Ph-substituted, 33b, d, respectively) and
erythritol (Et- and p-MeO-Bn-substituted, 34b, f, respective-
ly) families showed moderate inhibition of b-d-glucosidase.
Additional individual moderate inhibitory activities (Ki 25.9
48.9mm) with no broad trend were also observed: a-d-
mannosidase by Et-erythritol 34b; b-d-mannosidase by Et-
threitol 33b; a-l-rhamnosidase by Et- and p-MeO-Bn-eryth-
ritol 34b, f. All inhibitors displayed competitive inhibition,
which indicates that their mode of binding and that of their
corresponding substrate are similar.
The effective inhibition of a-l-rhamnosidase by the imino-
threitol aza-sugars 33 can be dissected on the basis of
structural similarities and differences. A stereochemical basis
for the inhibition of l-rhamnosidase can be rationalized by
the configurational similarity between OH-3,4 and C-5 of l-
rhamnose (36) and OH-2,3 and C-4 of these threitol aza-
sugars (Figure 4). The best inhibitor, the 4-methoxybenzyl
derivative 33 f (Ki ˆ 3.0mm), is a slightly better inhibitor
(almost two-fold) than its ethyl counterpart 33b (Ki ˆ 5.9mm).
The phenyl side chain in 33d results in >four-fold poorer
inhibition (Ki ˆ 13.0mm). This may indicate that the CH2
group present in the hydrophobic side chain, which ethyl-
(33b) and 4-methoxybenzyl- (33 f) derivatives both contain, is
of some importance in the interactions with the active site of
rhamnosidase. This raises the possibility, which is also
consistent with the configurational mimicry outlined above,
that the hydrophobic side chains, and in particular a CH2
group within these side chains, may mimic the C-5 methyl
H
N
H
N
N
i)
O[Si]
ii)
OMe
ii)
[Si]O
HO
HO
OH
OH
[Si]O
O[Si]
OMe
OMe
2
29f
33f
H
N
H
N
N
i)
O
O
O
O
OMe
3
30f
34f
Scheme 4. i) p-MeOBnMgCl, THF, 43 62% from 10 for 29 f, 76% from
21 for 30 f; ii) 25% TFA (aq.), 61% for 33 f, 65% for 34 f.
Carbohydrate-proccessing enzyme inhibition screening: The
resulting deprotected compounds were screened for enzyme
inhibition against a range of carbohydrate-processing en-
zymes (Tables 2 4), to allow a comparison of the effects of
C-1 hydrophobic substituent and hydroxyl stereochemistry.
Table 2. Aza-sugar inhibition of non-mammalian glycosidases.[a]
Glycosidase
Inhibitor Ki [mm]
33d
33b
33 f
34d
34b
34 f
a-d-mannosidase
(Canavalia ensiformis)
b-d-mannosidase
(snail)
b-d-glucosidase
(almond)
a-d-galactosidase
(green coffee bean)
a-l-rhamnosidase
(Penicillium decumbens)
a-l-fucosidase
N.I.
15.1
16.7
N.I.
48.9
N.I.
N.I.
18.5
N.I.
13.0
N.I.
28.9
13.3
N.I.
5.9
N.I.
N.I.
N.I.
3.0
N.I.
N.I.
N.I.
N.I.
5.0
N.I.
49.4
1.43
34.0
9.0
N.I.
56.4
0.25
25.9
2.4
7.6
33.5
(bovine kidney)
[a] N.I. indicates no inhibition observed at 0.1mm.
Chem. Eur. J. 2003, 9, 3397 3414 www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3403

B. G. Davis et al.
FULL PAPER
approach: features that successfully generate inhibitory
potency in one series do so in enantiomeric series also.
The benefit of this strategy is yet more clearly emphasized
by near parallel and complementary results in the erythritol
family of aza-sugars 34. These also inhibited two ™mirror-
image∫ enzymes: a-l-fucosidase and a-d-galactosidase. In-
deed, p-methoxybenzyl erythritol 34 f with a Ki ˆ 250 nm
towards a-d-galactosidase displays the most potent inhibition
of non-mammalian carbohydrate processing enzymes deter-
mined in this study, although this is still an order magnitude
lower than the piperidine DGJ, the d-galacto analogue of
DNJ (Ki ˆ 16 nm[49]). This inhibition of the D,L galactosidase/
fucosidase enzyme pair by erythritol family 34 is, as for the
threitol family 33, partially consistent with configurational
mimicry of the cis-diol OH-3,4 unit of galactosidase/fucosi-
dase by the cis-diol OH-2,3 unit of 34 (Figure 4). Consistent
with these results, methyl erythritol aza-sugar 34a has
previously been shown to be a good inhibitor of a-l-
fucosidase (Ki 2.0mm),[22] although this activity is modest
compared with some of the most potent a-l-fucosidase aza-
sugar inhibitors known (for example, the piperidine DFJ, the
l-fuco analogue of DNJ, displays a Ki of 0.029mm towards the
same enzyme[50]). Thus, again the Fleet mirror-image postu-
late correctly predicts the observed additional inhibition of a-
galactosidase by this family of racemates. Interestingly,
however, the relative configuration of C-5 in galactose/fucose
is opposite to that of C-4 in 34. Such inhibition by aza-sugars
with apparently the opposite configuration at the carbon that
mimics C-5 has been observed previously.[6] It should also be
noted that threitols 33b,f also inhibit a-l-fucosidase perhaps
through mimicry of the C-2,C-3 unit of 37.
The inhibition pattern of threitol 33 and erythritol 34 aza-
sugars towards human glycosidases largely followed that
towards non-mammalian in that a moderate and broad
inhibition of b-d-glucosidase, both lysosomal and non-lyso-
somal, was observed. The broad nature of this inhibition for
both diastereomeric inhibitor families suggests an inhibitory
mode that does not depend on functional or configurational
mimicry. The complete lack of a- or b-d-galactosidase
inhibition observed for 33 also mirrors that observed for
non-mammalian glycosidase inhibition.
As Table 4 shows, some interesting inhibitory potency was
observed towards glucosylceramide synthase (GCS). All
three threitols 33b,d,f showed fair levels of inhibition at
50mm (26.5 48.1%). The most potent ethyl threitol 33b in
fact showed an IC50 (52mm) that is within an order of
magnitude of that for the Gaucher disease drug candidate
N-butyl deoxynojirimycin (NB-DNJ or Zavesca), which has
an in vitro IC50 ˆ 20.4mm.[3,51] This inhibitory activity may have
its origin in a number of potential modes of mimicry
(Figure 4): a) the trans OH-2,3 diol unit of 33 may mimic
the trans diol units found at OH-2,3 or OH-3,4 in the d-
glucoside unit that is transferred by GCS; or b) 33b as a whole
mimics the polar head (e.g. the 3-amino-3-deoxy-erythros-
phingenine unit 39) of the ceramide unit that is a glycosyl
acceptor in the reaction catalyzed by GCS.[3, 52] These stereo-
specific, potential modes of action are strongly supported by
the complete lack of inhibition shown by the diastereomeric
erythritols 34b, d, f.
Figure 4.
group of l-rhamnose. In this regard, the phenyl derivative,
which lacks a CH2 in its hydrophobic side chain, is a poorer
mimic of the C-5 methyl of l-rhamnose and this is consistent
with the lower inhibitory potency observed for phenyl threitol
33d, although the phenyl derivative is still a relatively good
rhamnosidase inhibitor. Interestingly, the Ki values of ethyl
threitol 33b and p-methoxybenzyl threitol 33 f are similar.
The extra aromatic functionality (replacement of the CH3
terminus in 33b by a p-methoxyphenyl terminus in 33 f)
indicates that not only is the gain from additional remote
hydrophobic or aromatic interactions in the active site small,
but also that this region of the enzyme active site is able to
accommodate the marked extra steric bulk of the 4-methox-
ybenzyl functionality.
Indeed, previous studies have established very similar
inhibitory potency for the methyl threitol 33a (Ki ˆ
5.5mm).[43] It should be noted that the results obtained here
support previous findings that pyrrolidine aza-sugars are good
inhibitors of a-l-rhamnosidase[11, 43] (for example, l-swainso-
nine and DRAM, two of the most powerful known pyrrolidine
inhibitors display Ki ˆ 0.45 and 1.0mm, respectively)[11] and
typically much better than piperidine analogues of l-rham-
nose,[6, 44] possibly by virtue of the ability of envelope
conformations of pyrrolidines to mimic boat,[45] envelope,[46]
or half chair[47] transition state conformations in the rhamno-
sidase mechanism.
The ethyl- (33b) and 4-methoxybenzyl- (33 f) threitols also
showed significant a-d-mannosidase inhibition (Ki ˆ 15.1 and
16.7mm, respectively). These inhibition levels compare well
with one of the most potent a-d-mannosidase inhibitors, l-
swainsonine (Ki ˆ 9.5mm towards same enzyme[48]). This a-d-
mannosidase inhibition may be readily rationalized since the
6-deoxy derivative of d-mannose 35 is the enantiomer of l-
rhamnose (36). Thus, our strategy of constructing and testing
racemic series of azasugars also provides the correct stereo-
chemistry to mimic d-mannose (35). This similar inhibition of
enzymes that process enantiomeric sugar substrates by a
racemate therefore provides rapid confirmation in general
terms of the applicability of Fleet×s ™mirror-image∫ enzyme
active site postulate,[11] and highlights the potential of our
3404
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
It should be noted that in the quest[53, 54] for more selective,
improved profile GCS inhibitors that display low a-glucosi-
dase inhibition, 33b with its complete lack of inhibition of
human a-glucosidase at 1mm and IC50 ˆ 52mm towards GCS
makes for a good potential lead.
exact mass measurements were performed by the Dyson Perrins mass
spectrometry service on a Walters 2790-Micromass LCT electrospray
ionization mass spectrometer, and the EPSRC mass spectrometry service at
Swansea, UK. Melting points were recorded using a Cambridge Instru-
ments Gallen III Kofler Block melting apparatus and are uncorrected.
Elemental analyses were performed by the Durham University micro-
analysis service and the Oxford University Inorganic Chemistry Labora-
tory microanalysis service. Thin-layer chromatography (TLC) was carried
out on aluminium sheets coated with silica gel 60F254 (Merck, 1.05554).
Conclusion
Plates were developed using
a phosphomolybdic acid or potassium
permanganate dip. Flash column chromatography was performed using
silica gel (Fluorochem, 60A, 40 63 micron). Solvents were dried imme-
diately prior to use according to standard procedures: diethyl ether and
tetrahydrofuran were distilled under N2 over Na, dichloromethane was
distilled under N2 from CaH2. All solvents were removed by evaporation
under reduced pressure. All aqueous solutions were saturated unless
otherwise stated. For convenience, aza-C-glycoside imine adduct products
have been named based upon the threitol or erythritol scaffold from which
they were derived; stereochemistry at the newly derived centre is indicated
using the a/b carbohydrate convention.
Highly diastereoselective additions of Grignard reagents to
cyclic imines 2 4 has provided a ready route to novel aza-
sugars with hydrophobic substituents. Optimization has
accessed yields obtained for these additions that are some of
the best reported for additions of organometallics to cyclic
imines. This ready methodology lends itself to the easy and
general introduction of substituents to an aza-sugar scaffold,
and in this way a wide variety of substituents could be added
to yield products that are potential enzyme inhibitors. As
such, valuable information can be obtained on the mechanism
of action of enzymes that process sugars as their substrates by
investigating enzyme inhibition as a function of the substitu-
ents on an aza-sugar scaffold to create structure activity
relationships.
In this way we have generated aza-sugars as potential
inhibitors based on a designed pyrrolidine scaffold with
systematic variations in hydrophobic substituent and stereo-
chemistry and screened them against a diverse panel of 12
carbohydrate-processing enzymes. This has revealed good
inhibitors of a-l-rhamnosidase and a-d-mannosidase, with Ki
values in the range 3.0 16.7mm for the iminothreitol aza-
sugars 33, and relatively potent a-l-fucosidase and a-d-
galactosidase inhibition, with Ki values in the range 250nm
9.0mm for the iminoerythritol aza-sugars 34. It has also
identified 33b as a novel and selective inhibitor of the
Gaucher disease therapeutic target enzyme GCS. The obser-
vation of parallel patterns of inhibition in enzymes that
process enantiomeric sugar substrates also offers some con-
firmation of Fleet×s mirror-image active site postulate[11] and
highlights the value of this strategy, in the absence of other
sources of 3D information for target active sites for generating
nanomolar inhibitors such as 34 f.
1-Carboxybenzyl-3-pyrroline:[55]
Benzyl
chloroformate
(1.33 mL,
9.4 mmol, 1.3 equiv) was added to a cooled (58C) stirred mixture of
3-pyrroline (6; 97%, 0.50 g, 7.24 mmol, 1.0 equiv) in toluene (13 mL) and
3m NaOH (9 mL). After 1 h TLC (EtOAc/hexane 1:1) showed consump-
tion of SM. The layers were separated and the organic layer was washed
with distilled water (2 Â 10 mL), dried (MgSO4), filtered and concentrated
under reduced pressure to yield a yellow oil (1.59 g). This crude product
was purified by flash column chromatography on silica gel (EtOAc/hexane
3:7) to yield 1-carboxybenzyl-3-pyrroline as a pale yellow oil (1.37 g, 93%).
1H NMR (400 MHz, CDCl3): d ˆ 4.20 (m, 4H; 2  CH2N), 5.17 (s, 2H;
PhCH2), 5.79 (m, 2H; 2 Â C CH), 7.36 (m, 5H; C6H5); 13C NMR
ˆ
ˆ
(100 MHz, CDCl3): d ˆ 52.8, 53.3 (CH2N), 66.7 (PhCH2O), 112.4 ( CH),
128.3, 128.2, 127.8 (aromatic CH), 136.8 (quaternary aromatic C), 154.6
(C O); IR (film): nÄ ˆ 3064, 3032 cmÀ1 (aromatic C-H), 2952, 2862
ˆ
ˆ
ˆ
‡
(aliphatic C-H), 1707 (C O), 1623 (C C); MS (ES): m/z (%): 240 (100),
‡
226 (35) [M ‡Na], 91 (48) [C6H5CH2 ].
1-Carboxybenzyl-3,4-epoxy-pyrrolidine (7):[56] mCPBA (57 86%, 0.95 g)
in CH2Cl2 (8 mL) was added a solution of to 1-carboxybenzyl-3-pyrroline
(500 mg, 2.46 mmol) in dry CH2Cl2 (10 mL). The mixture was left to stir
under N2. After 19 h, TLC (Et2O/hexane 1:1) showed that SM still
remained. A second portion of mCPBA (0.95 g) in CH2Cl2 (6 mL) was then
added. After a total of 23 h, TLC (Et2O/hexane 1:1) showed consumption
of SM and formation of a major product (Rf 0.5). The solvent was removed
under reduced pressure, and the resultant white solid dissolved in Et2O
(30 mL) and washed with aqueous NaHCO3 (6 Â 20 mL) followed by brine
(2 Â 20 mL). The organic layer was dried (MgSO4) and concentrated under
reduced pressure. The resultant white solid (0.84 g) was purified by flash
column chromatography (EtOAc/hexane 1:1) to give 7 as a colourless oil
(249 mg, 46%). 1H and 13C NMR spectra show inequivalent CH2N
resonances that are consistent with a restricted rotation of the Cbz-N
bond. 1H NMR (400 MHz, CDCl3): d ˆ 3.38 (dd, 2J(H,H) ˆ 12.7, 3J(H,H) ˆ
6.3 Hz, 1H; 1 of CHH'N), 3.39 (dd, 2J(H,H) ˆ 12.8, 3J(H,H) ˆ 6.4 Hz, 1H; 1
2
of CHH'N), 3.68 (m, 2H; 2  OCH), 3.84 (d, J(H,H) ˆ 12.7 Hz, 1H; 1 of
Experimental Section
CHH'N), 3.89 (d, 2J(H,H) ˆ 12.8 Hz, 1H; 1 of CHH'N), 5.11 (s, 2H;
PhCH2), 7.34 (m, 5H; C6H5); 13C NMR (100 MHz, CDCl3): d ˆ 47.1, 47.3
(2 Â OCH), 54.9, 55.5 (2 Â CH2N), 66.9 (PhCH2O), 127.9, 128.0,128.4 (3 Â
General: 1H and 13C nuclear magnetic resonance (NMR) spectra were
recorded on Varian Gemini 200, Bruker DPX200, Unity 300, VXR 400,
Bruker DPX400 or Varian Inova 500 NMR spectrometers at the
frequencies indicated. Where indicated, NMR peak assignments were
made using COSY, DEPT, HETCOR or NOESY experiments; all others
are subjective. All chemical shifts are quoted on the d scale and were
referenced to residual solvent as an internal standard. The following
abbreviations are used to describe NMR multiplicities: s, singlet; d,
doublet; t, triplet; q, quartet; m, multiplet; br, broad; p, pseudo. IR spectra
were recorded on a Perkin Elmer Paragon 1000 Fourier Transform
spectrophotometer. The following abbreviations are used to describe
infrared absorption bands: br, broad; s, strong. Mass spectra were recorded
by the Durham University and Dyson Perrins mass spectrometry services
using electron impact (EI), chemical ionisation (CI), atmospheric pressure
chemical ionisation (APCI), field ionization (FI) or electrospray ionisation
(ES) techniques on Micromass LCT, Micromass GCT, Micromass Auto-
Spec-oaTof, Micromass Platform and VG Platform mass spectrometers;
ˆ
aromatic CH), 136.5 (quaternary aromatic C), 155.2 (C O); IR (film): nÄ ˆ
3062, 3034 cmÀ1 (aromatic C-H), 2945, 2875 (aliphatic C-H), 1705 (C O);
ˆ
‡
‡
MS (EI): m/z (%): 219 (91) [M ], 91 (100) [C6H5CH2 ]; HRMS (EI): m/z:
‡
calcd for C12H13NO3: 219.0895; found: 219.0898 [M ].
N-Carboxybenzyl-1,4-dideoxy-1,4-iminothreitol (8):[47] N-Carboxybenzyl-
2,3-epoxy-1,4-dideoxy-1,4-imino-threitol (7; 52 mg, 0.24 mmol) was dis-
solved in 2m aqueous H2SO4 (4 mL) and Et2O(4 mL). This resulting
solution was left to stir under argon at room temperature. After 21 h, TLC
(EtOAc) showed consumption of SM (Rf 0.7) and formation of a major
product (Rf 0.25). Et2O(4 mL) was added and the layers separated. The
organic layer was dried (MgSO4) and concentrated under reduced to give a
yellow oil (17 mg). The aqueous layer was extracted with EtOAc (6 mL),
and the organic extract dried (MgSO4) and concentrated under reduced
pressure to give a colourless oil (33 mg). Both products were shown by
1
1H NMR to be the diol 8 (50 mg in total, 88%). H and 13C NMR spectra
Chem. Eur. J. 2003, 9, 3397 3414
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3405

B. G. Davis et al.
FULL PAPER
show inequivalent CH2N resonances that are consistent with a restricted
rotation of the Cbz-N bond. 1H NMR (400 MHz, CDCl3): d ˆ 3.36 (dd,
2J(H,H) ˆ 11.6, 3J(H,H) ˆ 6.0 Hz, 2H; 2  CHH'N), 3.62 (br dd, 2J(H,H) ˆ
11.6, 3J(H,H) ˆ 2.8 Hz, 2H; 2  CHH'N), 4.06 (m, 2H; 2  CHOH), 4.24
(br s, 2H; 2 Â OH), 5.05 (s, 2H; PhCH2), 7.30 (m, 5H; C6H5); 13C NMR
(100 MHz, CDCl3): d ˆ 51.3, 51.7 (2  CH2N), 67.2 (PhCH2O), 74.4, 75.0
heated at reflux and passed through a dropping funnel packed with 4 ä
molecular sieves in order to remove water formed for 19 h. The mixture
was allowed to cool and then cooled to ice-bath temperature. The resulting
orange precipitate was recrystallised from EtOAc, then filtered and washed
with EtOAc to yield the crude hydroxyimide 12 as white crystals (10.1 g).
TBDMS-Cl (20.3 g, 0.135 mol, 3.0 equiv) was added as a solution in dry
(2 Â CHOH), 127.7, 128.0,128.4 (3 Â aromatic CH), 136.2 (quaternary
dimethylformamide (100 mL) under N2 to
a solution of 12 (10.0 g,
À1
ˆ
aromatic C), 155.6 (C O); IR (film): nÄ ˆ 3402 cm (br, O-H), 3065, 3033
0.045 mol) and imidazole (15.3 g, 0.225 mol, 5.0 equiv) in dry dimethylform-
amide (100 mL). The reaction mixture was stirred for 21 h; dimethylform-
amide was then removed by evaporation under reduced pressure to yield a
pale yellow oily solid. This residue was dissolved in CHCl3 (200 mL) and
washed with 1m HCl (3 Â 50 mL) followed by water (3 Â 50 mL). The
aqueous layers were extracted with CHCl3 (2 Â 50 mL) and the combined
organic layers dried (Na2SO4), filtered and concentrated under reduced
pressure to yield a yellow oil (20.5 g). The crude product was purified by
flash column chromatography on silica gel (EtOAc/hexane 1:10) to yield
pure 13 as a white solid (11.4 g, 43% from tartaric acid). M.p. 76.5 77.08C;
1H NMR (400 MHz, CDCl3): d ˆ 0.16, 0.22 (2s, 2  6H; 2  Si(CH3)2), 0.94
(s, 18H; 2 Â SiC(CH3)3), 4.47 (s, 2H; 2 Â CHOSi), 4.62 (s, 2H; PhCH2),
7.30 7.40 (m, 5H; C6H5); 13C NMR (100 MHz, CDCl3): d ˆ À4.5, À5.1
(2 Â Si(CH3)2), 18.2 (SiC(CH3)3), 25.6 (SiC(CH3)3), 42.3 (CH2Ph), 76.9
ˆ
(aromatic C-H), 2945, 2887 (aliphatic C-H), 1677 (s, C O); MS (EI): m/z
(%): 237 (9) [M ], 91 (100) [C6H5CH2 ]; HRMS (EI): m/z: calcd for
‡
‡
‡
C12H15NO4: 237.1001; found: 237.1006 [M ].
N-Carboxybenzyl-2,3-O-tert-butyldimethylsilyl-1,4-dideoxy-1,4-imino-
threitol: tert-Butyldimethylsilyl trifluoromethanesulphonate (4.35 mL,
5.01 g, 19.0 mmol, 3.0 equiv) was added at 08C to a solution of 8 (1.50 g,
6.32 mmol) in dry pyridine (3.70 mL, 3.00 g, 37.9 mmol, 6.0 equiv) and dry
CH2Cl2 (90 mL). The reaction mixture was stirred under N2. After 75 min
TLC (EtOAc) showed consumption of SM and TLC (EtOAc/hexane 1:4)
showed formation of a major product (Rf 0.6). The reaction mixture was
washed with distilled water (30 mL) followed by 1m aqueous HCl (30 mL)
and again by distilled water (30 mL). The aqueous extracts were back-
extracted with CH2Cl2 (30 mL) and the combined organic layers dried
(MgSO4) and concentrated under reduced pressure to yield a yellow oil
(3.50 g). The crude product was purified by flash column chromatography
on silica gel (Et2O/hexane 3:17) to yield N-carboxybenzyl-2,3-O-tert-
(CHOSi), 128.0, 128.7, 129.0 (aromatic CH), 135.3 (quaternary aromatic C),
À1
ˆ
173.0 (C O); IR (film): nÄ ˆ 2956, 2930, 2893, 2857 cm (aliphatic C-H),
‡
ˆ
1716 (C O); MS (ES): m/z (%): 472 (100) [M ‡Na]; HRMS (CI): m/z:
‡
calcd for C23H43N2O4Si2: 467.2761; found: 467.2768 [M ‡NH4]; elemental
butyldimethylsilyl-1,4-dideoxy-1,4-imino-threitol as
a pale yellow oil
(2.70 g, 92%). 1H and 13C NMR spectra show inequivalent CH2N
resonances that are consistent with a restricted rotation of the Cbz-N
bond. 1H NMR (400 MHz, CDCl3): d ˆ 0.05, 0.06 (2s, 2  6H; 2 Â
Si(CH3)2), 0.86, 0.86 (2s, 2  9H; 2  SiC(CH3)3), 3.28 (d, 2J(H,H) ˆ
11.6 Hz, 1H; CHH'N), 3.32 (d, 2J(H,H) ˆ 11.7 Hz, 1H; CHH'N), 3.57
(dd, 2J(H,H) ˆ 11.2, 3J(H,H) ˆ 3.4 Hz, 1H; CH''H'''N), 3.62 (dd, 2J(H,H) ˆ
11.2, 3J(H,H) ˆ 3.5 Hz, 1H; CH''H'''N), 5.15 (s, 2H; PhCH2), 3.98 (m, 2H;
2  CHOSi), 7.36 (m, 5H; C6H5CH2); 13C NMR (100 MHz, CDCl3): d ˆ
À4.8, À4.9 (2 Â OSi(CH3)2), 17.8, 17.9 (SiC(CH3)3), 25.6, 25.7 (2 Â
SiC(CH3)3), 52.0, 52.3 (CH2N), 66.6 (PhCH2O), 75.7, 76.4 (CHOSi), 127.6,
analysis calcd (%) for C23H39NO4Si2: C 61.42, H 8.74, N 3.11; found: C
61.26, H 8.86, N 3.05.
N-Benzyl-2,3-O-di-tert-butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol:[57a]
Borane dimethylsulfide (9.04 mL, 7.24 g, 95.3 mmol, 7.5 equiv) was added
under N2 to a solution of 13 (5.70 g, 12.7 mmol) in dry THF (90 mL). After
24 h stirring at RT the reaction was monitored by crude NMR. A little SM
appeared to remain, so a further portion of borane dimethylsulfide
(3.01 mL, 2.41 g, 31.8 mmol, 2.5 equiv) was added. After a further 6 h
stirring, crude NMR and TLC (EtOAc/toluene 5:95) showed the con-
version of SM (Rf 0.7) to product (Rf 0.75). The excess borane
dimethylsulphide was carefully quenched with MeOH until no further
effervescence was observed and the solvent removed to yield borazine
product. 1H NMR (300 MHz, CDCl3): d ˆ 0.01, 0.03 (2s, 2  3H; 2 Â
SiCH3), 0.12 (s, 6H; Si(CH3)2), 0.88, 0.93 (2s, 2 Â 9H; 2 Â SiC(CH3)3),
3.07 (dd, 2J(H,H) ˆ 12.6, 3J(H,H) ˆ 7.0 Hz, 1H; 1 of CHH'N), 3.09 (dd,
2J(H,H) ˆ 12.3, 3J(H,H) ˆ 5.0 Hz, 1H; 1 of CHH'N), 3.39 (dd, 2J(H,H) ˆ
ˆ
127.7, 128.3 (aromatic CH), 137.2 (quaternary aromatic C), 155.2 (C O); IR
(film): nÄ ˆ 2929, 2857 cmÀ1 (aliphatic C-H), 1712 (s, C O); MS (ES): m/z
ˆ
‡
‡
(%): 488 (100) [M ‡Na], 466 (10) [M ‡H]; HRMS (ES): m/z: calcd for
‡
C24H44NO4Si2: 466.2809; found: 466.2803 [M ].
2,3-O-tert-Butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol (9)[57]
Method i): Pd/C (10% wt Pd, 20 mg) was added to a solution of N-
carboxybenzyl-2,3-O-tert-butyldimethylsilyl-1,4-dideoxy-1,4-imino-threitol
(149 mg, 0.32 mmol) in dry MeOH (5 mL). The reaction mixture was
stirred under an atmosphere of H2. After 3 h, TLC (EtOAc/hexane 1:4)
showed consumption of SM. The mixture was filtered through a glass sinter
and the solvent removed under reduced pressure to yield 9 as a colourless
oil (98 mg, 92%).
12.3, 3J(H,H) ˆ 6.7 Hz, 1H;
1
of CHH'N), 3.48 (dd, 2J(H,H) ˆ 12.3,
3J(H,H) ˆ 6.5 Hz, 1H; 1 of CHH'N), 3.91 (ptd, 3J(H,H) ˆ 6.8, 3J(H,H) ˆ
4.2 Hz, 1H; 1 of CHOSi), 4.11 (s, 2H; CH2Ph), 4.32 (ptd, 3J(H,H) ˆ 6.0,
3J(H,H) ˆ 4.5 Hz, 1H; 1 of CHOSi), 7.36 7.56 (m, 5H; C6H5); 13C NMR
(75 MHz, CDCl3): d ˆ À4.9, À4.8 (SiCH3), 17.8, 17.9 (SiC(CH3)), 25.6, 25.8
(SiC(CH3)), 64.3, 64.6, 67.3 (CH2Ph, 2 Â CH2N), 78.0, 78.4 (2 Â CHOSi),
128.1, 129.0, 132.7 (aromatic C). The solvent was removed under reduced
pressure and the residue was repeatedly dissolved in MeOH followed by
removal of the solvent under reduced pressure (3 Â 50 mL). The residue
was then dissolved in MeOH (90 mL), heated to 408C and left to stir for
20 h. The reaction mixture was concentrated under reduced pressure to
yield a colourless oil (5.35 g, 100%). This product was concordant with
Method ii): Pd/C (600 mg) was added to a solution of N-benzyl-2,3-O-di-
tert-butyldimethylsilyl-1,4-dideoxy-1,4-imino-threitol (4.95 g, 11.7 mmol) in
MeOH (70 mL). The reaction mixture was stirred under an atmosphere of
H2. After 17 h the reaction was monitored by TLC (EtOAc/hexane 1:4).
Some product formation was evident but SM (Rf 0.75) still remained. A
further portion of palladium on activated carbon was added (1.00 g) and the
H2 atmosphere replenished. After a total of 20 h stirring TLC (EtOAc/
hexane 1:4) showed consumption of SM. The reaction mixture was filtered
through Celite and the solvent removed by evaporation to yield 9 as a
colourless oil (3.85 g, 99%). Compound 9 was found to be a somewhat
unstable to purification on silica gel as reported in the literature[57b] so was
used without further purification in the following reactions. 1H NMR
(400 MHz, CDCl3): d ˆ 0.03, 0.04 (2s, 2  6H; OSi(CH3)2), 0.85 (s, 18H;
2  SiC(CH3)3), 2.77 (d, 2J(H,H) ˆ 12.0 Hz, 2H; 2  CHH'N), 3.16 (dd,
2J(H,H) ˆ 11.8, 3J(H,H) ˆ 3.8 Hz, 2H; 2  CHH'N), 3.97 (m, 2H; 2 Â
CHOSi), 5.33 (br s, 1H; NH); 13C NMR (100 MHz, CDCl3): d ˆ À4.8
(OSi(CH3)2), 17.9 (SiC(CH3)3), 25.7 (SiC(CH3)3), 53.1 (CH2N), 78.5
(CHOSi); IR (film): nÄ ˆ 3304 cmÀ1 (br, N-H), 2954, 2929, 2887, 2857
pure
N-benzyl-2,3-O-di-tert-butyldimethylsilyl-1,4-dideoxy-1,4-imino-
threitol and no further purification was required. 1H NMR (400 MHz,
CDCl3): d ˆ 0.05, 0.07 (2s, 2  6H; 2  Si(CH3)2), 0.90 (s, 18H; 2 Â
SiC(CH3)3), 2.47 (dd, 2J(H,H) ˆ 9.8, 3J(H,H) ˆ 4.6 Hz, 2H; 2  CHH'N),
2.88 (dd, 2J(H,H) ˆ 9.8, 3J(H,H) ˆ 6.0 Hz, 2H; 2  CHH'N), 3.52 (d,
2J(H,H) ˆ 13.2 Hz, 1H; 1 of CHOSi), 3.74 (d, 2J(H,H) ˆ 13.2 Hz, 1H; 1
of CHOSi), 4.13 (m, 2H; PhCH2), 7.33 7.32 (m, 5H; C6H5); 13C NMR
(100 MHz, CDCl3): d ˆ À4.6, À4.7 (2  OSi(CH3)2), 18.0 (SiC(CH3)3), 25.8
(SiC(CH3)3), 60.7, 60.7 (CH2N and PhCH2), 79.8 (CHOSi), 126.8, 128.1,
128.6 (3  aromatic CH), 135.3 (quaternary aromatic C); IR (film): nÄ ˆ
3063, 3028 cmÀ1 (aromatic C-H), 2954, 2929, 2857 (aliphatic C-H); HRMS
‡
(CI): m/z: calcd for C23H44NO2Si2: 422.2910; found: 422.2911 [M ‡H].
N-Chloro-2,3-O-di-tert-butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol
(10):[57a] N-Chlorosuccinimide (266 mg, 1.99 mmol, 1.1 equiv) was added to
a solution of 9 (600 mg, 1.81 mmol) in dry Et2O(30 mL). The reaction
mixture was stirred under N2. After 4.5 h stirring TLC (EtOAc/hexane 1:4)
showed the formation of a major product (Rf 0.8). The reaction mixture was
diluted with Et2O(30 mL) and washed with water (3 Â 20 mL). The
‡
(aliphatic C-H); MS (ES): m/z (%): 332 (100) [M ‡H]; HRMS (CI): m/z:
‡
calcd for C24H44NO4Si2: 332.2441; found: 332.2442 [M ‡H].
N-Benzyl-2,3-O-di-tert-butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol-1,4-
dione (13):[58b] A mixture of (Æ)-tartaric acid (9.0 g, 0.060 mol) and
benzylamine (8.52 mL, 0.078 mmol, 1.3 equiv) in p-xylene (150 mL) was
3406
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
aqueous extracts were extracted with Et2O(2 Â 20 mL) and the combined
organic extracts dried (Na2SO4), filtered and concentrated under reduced
pressure to yield 10 as a pale yellow oil (640 mg, 96%). Compound 10 was
unstable to purification on silica gel so it was used without further
purification in the following reactions. 1H NMR (400 MHz, CDCl3): d ˆ
0.05, 0.06 (2s, 2 Â 6H; 2 Â OSi(CH3)2), 0.87, 0.87 (2s, 2 Â 9H; 2 Â
SiC(CH3)3), 3.06 (dd, 2J(H,H) ˆ 10.2, 3J(H,H) ˆ 5.9 Hz, 2H; 2  CHH'N),
3.48 (dd, 2J(H,H) ˆ 10.2, 3J(H,H) ˆ 4.7 Hz, 2H; 2  CHH'N), 4.11 (m, 2H;
2  CHO[Si]); 13C NMR (100 MHz, CDCl3): d ˆ À4.7, À4.9 (2  O-
Si(CH3)2), 17.9 (SiC(CH3)3), 25.7 (SiC(CH3)3), 69.1 (CH2N), 79.0 (CHOSi);
IR (film): nÄ ˆ 2955, 2930, 2857 cmÀ1 (aliphatic C-H); HRMS (CI): m/z:
0.62 mmol) in MeOH (4 mL). The reaction mixture was stirred for 1.5 h,
whereupon TLC (EtOAc/cyclohexane 7:3) showed consumption of starting
material (Rf 0.5). Solid ammonium chloride was added to quench excess
borohydride and then the mixture was concentrated to yield a white solid.
This crude product was purified by flash column chromatography on silica
gel (EtOAc/cyclohexane 13:7 increasing to 3:1 then 17:3) to yield pure 17 as
an oily white solid (100 mg, 100%). 1H NMR (400 MHz, CDCl3): d ˆ 1.36,
1.45 (2s, 2  3H; C(CH3)2), 3.28 (t, 3J(H,H) ˆ 5.9 Hz, 2H; 2  OH), 3.77
(m, 4H; 2 Â CH2OH), 4.28 (m, 2H; 2 Â CH2CH); 13C NMR (100 MHz,
CDCl3, DEPT): d ˆ 25.0, 27.5 (2q, C(CH3)2), 60.6 (t, CH2OH), 76.9 (d,
CH), 108.4 (s, C(CH3)2); IR (film): nÄ ˆ 3398 cmÀ1 (br, O-H stretch), 2987,
2938 (aliphatic C-H stretch), 1457; HRMS (CI): m/z: calcd for C7H15O4:
‡
calcd for C16H37NO2Si2Cl: 366.2051; found: 366.2067 [M ‡H].
‡
163.0970; found: 163.0970 [M ‡H].
2,3-O-tert-Butyldimethylsilyl-1,4-dideoxy-1,4-imino-1,N-dehydrothreitol
(2): DBU (0.027 mL, 27 mg, 0.18 mmol, 1.1 equiv) was added under N2 to a
solution of 10 (58 mg, 0.16 mmol) in dry Et2O(5 mL). After 3 h the
reaction mixture was filtered under Ar to remove DBU ¥ HCl, then
2,3-O-Isopropylidene-1,4-di-O-methanesulfonylerythritol (18):[28] Metha-
nesulfonyl chloride (0.16 mL, 238 mg, 2.08 mmol, 4.0 equiv) was added
under Ar at 08C to a solution of 17 (85 mg, 0.52 mmol, 1.0 equiv) and
triethylamine (0.29 mL, 210 mg, 2.08 mmol, 4.0 equiv) in dry CH2Cl2
(3.5 mL). After 1 h TLC (Et2O) showed consumption of starting material
(Rf 0.2) and formation of a major product (Rf 0.25). The reaction mixture
was poured into ice-water (20 mL) and diluted with CH2Cl2 (30 mL). The
layers were separated and the organic phase washed with 1m HCl (10 mL)
followed by NaHCO3(aq) (10 mL). The organic layer was dried (MgSO4)
and concentrated to yield a pale yellow oil (202 mg). This crude product
was purified by flash column chromatography on silica gel (Et2O) to yield
pure 18 as a white solid (120 mg, 72%). M.p. 91 928C {lit.:[28] m.p. 92
1
concentrated to yield as an orange oil (50 mg, 96%). H NMR (400 MHz,
CDCl3): d ˆ À0.02, À0.02 (2s, 2  3H; 2  Si(CH3)), 0.04 (s, 6H; Si(CH3)2),
0.80, 0.82 (2s, 2  9H; 2  SiC(CH3)3), 3.47 (dddd, 2J(H,H) ˆ 15.8,
3J(H,H) ˆ 4.7, 2.4 Hz, 1.1 Hz, 1H; CHH'N), 3.97 (dddd, 2J(H,H) ˆ 16.0,
3J(H,H) ˆ 6.6, 2.2 Hz, 1.1 Hz, 1H; CHH'N), 4.10 (ddd, 3J(H,H) ˆ 6.5,
4.8 Hz, 3.9 Hz, 1H; NCH2CH), 4.49 (br d, 3J(H,H) ˆ 4.0 Hz, 1H;
3
N CHCH), 7.41 (pt, J(H,H) ˆ 2.2 Hz, 1H; N CH); 13C NMR (100 MHz,
CDCl3): d ˆ À4.9, À4.8, À4.7, À4.7 (4q, 4  SiCH3), 17.9, 18.0 (2s, 2 Â
SiC(CH3)3), 25.7, 25.7 (2q, 2 Â SiC(CH3)3), 66.9 (t, NCH2), 79.2 (d,
ˆ
ˆ
1
ˆ
ˆ
NCH2CH), 85.5 (d, N CHCH), 168.4 (d, N CH); IR (film): nÄ ˆ 2928,
938C}; H NMR (400 MHz, CDCl3): d ˆ 1.38, 1.49 (2s, 2  3H; C(CH3)2),
2856 cmÀ1 (aliphatic C-H), 1646 (C N); HRMS (CI): m/z: calcd for
3.09 (s, 6H; 2 Â OSO2CH3), 4.33 (m, 4H; 2 Â CH2OMs), 4.48 (m, 2H; 2 Â
CH2CH); 13C NMR (100 MHz, CDCl3, DEPT): d ˆ 25.4, 27.4 (2q,
C(CH3)2), 37.7 (q, OSO2CH3), 66.5 (t, CH2OMs), 74.1 (d, CH), 110.1 (s,
C(CH3)2); IR (KBr disc): nÄ ˆ 3023, 2979, 2942 cmÀ1 (aliphatic C-H stretch),
ˆ
‡
C16H36NO2Si2: 330.2285; found: 330.2286 [M ‡H].
2,3-O-Isopropylidene-l-erythrose
(16):[27, 59]
2,2-Dimethoxypropane
(27 mL, 22.9 g, 0.220 mmol, 3.3 equiv) was added to a solution of l-
arabinose (14) (10.0 g, 0.067 mol, 1.0 equiv) and p-toluenesulphonic acid
monohydrate (150 mg, 0.8 mmol, 0.01 equiv) in dimethylformamide
(130 mL). The resulting solution was stirred for 2 h. and then neutralized
by the addition of solid sodium carbonate and concentrated under reduced
pressure. The residue was partitioned between water (120 mL) and 40
608C petroleum ether (60 mL). To the aqueous layer was then added
sodium periodate (35.6 g, 0.166 mol, 2.5 equiv) portionwise, and the
mixture stirred for 2 h. Solid sodium carbonate was added and the slurry
was stirred for 1 h. Water was then added (80 mL) and the aqueous layer
was extracted with EtOAc (3Â 80 mL), and the combined organic extracts
concentrated under reduced pressure to yield a pale yellow oil (10.9 g). This
residue was then dissolved in CH2Cl2 (80 mL) and washed with water (2 Â
20 mL). The organic extracts were dried (MgSO4) and concentrated under
reduced pressure to yield a pale yellow oil (7.2 g). The organic extracts from
‡
1352, 1172; MS (APCI): m/z (%): 319 (100) [M ‡H]; HRMS (CI): m/z:
‡
calcd for C9H19O8S2: 319.0521; found: 319.0519 [M ‡H]; elemental
analysis calcd (%) for C9H18O8S2: C 33.95, H 5.70; found: C 33.87, H 5.73.
N-Benzyl-2,3-O-isopropylidene-1,4-dideoxy-1,4-iminoerythritol
(19):[28]
Benzylamine (3.0 mL) was added under Ar to 18 (200 mg, 0.63 mmol).
The mixture was stirred at 658C for 24 h. The reaction was monitored by
mass spectrometry (APCI), which showed consumption of SM and
‡
formation of the desired pyrrolidine (m/z 234 [M ‡H]). The reaction
mixture was diluted with EtOAc (30 mL) and washed with brine (10 mL)
followed by water (10 mL). The organic layer was dried (Na2SO4), filtered
and concentrated to yield a yellow oil. p-Xylene was then added and the
mixture concentrated (3 Â 20 mL) to remove excess benzylamine as its
azeotrope with xylene. This crude product was purified by flash column
chromatography on silica gel (EtOAc/cyclohexane 1:4) to yield pure 19 as a
colourless oil (112 mg, 76%). 1H NMR (400 MHz, CDCl3): d ˆ 1.34, 1.59
(2s, 2  3H; C(CH3)2), 2.15 (ddd, 2J(H,H) ˆ 11.6, 3J(H,H) ˆ 3.0, 3J(H,H) ˆ
1.4 Hz, 2H; 2  CHH'N), 3.05 (d, 2J(H,H) ˆ 11.5 Hz, 2H; 2  CHH'N), 3.63
(s, 2H; CH2Ph), 4.66 (m, 2H; 2 Â OCH), 7.22 7.36 (m, 5H; C6H5);
13C NMR (100 MHz, CDCl3, DEPT): d ˆ 25.1, 26.5, (2q, C(CH3)2), 59.2,
59.7 (2t, PhCH2 and CH2N), 79.6 (d, OCH), 111.2 (s, C(CH3)2), 126.9, 128.2,
128.5 (3d, 3 Â aromatic CH), 138.6 (s, quaternary aromatic C); IR (film):
nÄ ˆ 3028 cmÀ1 (aromatic C-H stretch), 2985, 2935, 2788 (aliphatic C-H
stretch), 1454, 1379; HRMS (CI): m/z: calcd for C14H20NO2: 234.1494;
1
CH2Cl2 were not found to be purer by H NMR than the original EtOAc
extracts, although the mass was significantly reduced. Therefore the
aqueous layers from the CH2Cl2 extraction were back-extracted with
EtOAc and concentrated and all the organic extracts combined to give a
yellow oil (8.2 g). This crude product was purified by flash column
chromatography on silica gel (EtOAc/cyclohexane 7:13) to yield pure 16
(6.05 g, 57% from l-arabinose) as a mixture of anomers (a:b 1:6).
25
[a]
ˆ ‡75.0 (c ˆ 0.1 in CHCl3) {lit.:[27] [a]D ˆ ‡66.3 (c ˆ 2.7 in
D; eqlbm
CHCl3); lit.:[59] [a]D ˆ ‡83.2 (c ˆ 4.36 in EtOAc)}; a-anomer: 1H NMR
(400 MHz, CDCl3): d ˆ 1.36, 1.53 (2s, 2  3H; C(CH3)2), 2.09 (brs, 1H;
OH), 3.53 (dd, 2J(H,H) ˆ 11.1, 3J(H,H) ˆ 3.7 Hz, 1H; CHH'), 3.96 (d,
‡
found: 234.1495 [M ‡H].
2,3-O-Isopropylidene-1,4-dideoxy-1,4-iminoerythritol (20):[28] Pd(OH)2/C
(20% Pd, 80 mg) was added to a solution of 19 (283 mg, 1.21 mmol) in
MeOH (5.0 mL) and the reaction mixture placed under an atmosphere of
H2. After 21 h TLC (EtOAc/cyclohexane 3:2) showed consumption of SM
(Rf 0.6). The reaction mixture was filtered through celite to remove
Pd(OH)2 on C, then the filtrate concentrated under reduced pressure to
yield 20 as a pale yellow oil (170 mg, 98%). (The product was found to be
relatively volatile so when removing the solvent the pressure was not
reduced below 60 mbar at 258C.) It was found to be unstable to purification
on silica gel, but was sufficiently pure not to require further purification for
the following reactions. 1H NMR (400 MHz, CDCl3): d ˆ 1.29, 1.43 (2s, 2 Â
3H; C(CH3)2), 2.47 (br s, 1H; NH), 2.51 (d, 2J(H,H) ˆ 13.6 Hz, 2H; 2 Â
CHH'N), 3.08 (d, 2J(H,H) ˆ 13.9 Hz, 2H; 2  CHH'N), 4.65 (m, 2H, 2 Â
OCH); 13C NMR (100 MHz, CDCl3): d ˆ 23.7, 25.9 (2q, C(CH3)2), 54.1
(CH2N), 81.4 (OCH), 110.2 (C(CH3)2); IR (film): nÄ ˆ 3395 cmÀ1 (br, N-H
stretch), 2935 (aliphatic C-H stretch), 1374; HRMS (CI): m/z: calcd for
2J(H,H) ˆ 11.6 Hz, 1H; CHH'), 4.47 (dd, 3J(H,H) ˆ 6.2, J(H,H) ˆ 3.6 Hz,
3
1H; CH2CH), 4.74 (dd, 3J(H,H) ˆ 6.2, J(H,H) ˆ 3.7 Hz, 1H; CHOHCH),
3
4.98 (s, 1H; CHOH); 13C NMR (100 MHz, CDCl3, DEPT): d ˆ 24.8, 25.9
(2q, C(CH3)2), 67.5 (t, CH2), 78.2, 79.5 (2d, 2 Â CHO-), 97.4 (d, CHOH),
1
113.4 (s, C(CH3)2); b-anomer: H NMR (400 MHz, CDCl3): d ˆ 1.30, 1.45
(2s, 3H; C(CH3)2), 3.61 (brs, 1H; OH), 3.99 (d, 2J(H,H) ˆ 10.4 Hz, 1H;
CHH'), 4.05 (dd, 2J(H,H) ˆ 10.4, 3J(H,H) ˆ 3.5 Hz, 1H; CHH'), 4.55 (d,
3J(H,H) ˆ 5.9 Hz, 1H; CHOHCH), 4.82 (dd, 3J(H,H) ˆ 5.9, 3J(H,H) ˆ
3.5 Hz, 1H; CH2CH), 5.39 (s, 1H; CHOH); 13C NMR (CDCl3,
100 MHz): d ˆ 24.6, 26.1 (2q, 2  C(CH3)), 71.8 (t, CH2), 79.9 (d, CH2CH),
85.1 (d, CHOHCH), 101.6 (d, CHOH), 112.2 (s, C(CH3)2); IR (film): nÄ ˆ
3426 cmÀ1 (br, O-H stretch), 2987, 2943, 2882 (aliphatic C-H stretch), 1460;
HRMS (CI): m/z: calcd for C7H16NO4: 178.1083; found: 178.1079
‡
[M ‡NH4].
2,3-O-Isopropylidene-erythritol (17):[28] Sodium borohydride (47 mg,
1.24 mmol, 2.0 equiv) was added under Ar to a solution of 16 (100 mg,
‡
C7H14NO2: 144.1025; found: 144.1026 [M ‡H].
Chem. Eur. J. 2003, 9, 3397 3414
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3407

B. G. Davis et al.
FULL PAPER
2J(H,H) ˆ 11.3, 3J(H,H) ˆ 4.7 Hz, H-4 a), 4.02 (d, 0.83H; 2J(H,H) ˆ
10.0 Hz, H-4' b), 4.06 (dd, 0.83H; 2J(H,H) ˆ 10.0, 3J(H,H) ˆ 3.5 Hz, H-4
b), 4.48 (dd, 0.17H; 3J(H,H) ˆ 6.2, 3J(H,H) ˆ 3.6 Hz, H-3 a), 4.56 (d,
0.83H; 3J(H,H) ˆ 6.0 Hz, H-2 b), 4.75 (dd, 0.17H; 3J(H,H) ˆ 5.4,
3J(H,H) ˆ 3.8 Hz, H-2 a), 4.82 (dd, 0.83H; 3J(H,H) ˆ 5.8, 3J(H,H) ˆ
N-Chloro-2,3-O-isopropylidene-1,4-dideoxy-1,4-iminoerythritol (21): N-
Chlorosuccinimide, NCS (513 mg, 3.84 mmol, 1.1 equiv) was added under
Ar to a solution of 20 (500 mg, 3.49 mmol) in dry Et2O(50 mL). After 5 h
TLC (EtOAc/cyclohexane 1:4) revealed consumption of SM and formation
of a major product (Rf 0.3). The mixture was diluted with Et2O(30 mL) and
washed with water (3 Â 20 mL). The organic layer was dried (Na2SO4),
filtered and concentrated under reduced pressure to afford 21 as a pale
2
3
3.5 Hz, H-3 b), 4.99 (dd, 0.17H; J(H,H) ˆ 10.5, J(H,H) ˆ 3.6 Hz, H-1 a),
5.42 (s, 0.83H; H-1 b); 13C NMR (100 MHz, CDCl3, DEPT, HMQC): d ˆ
23.6, 23.7, 24.0, 24.9, 25.0, 34.2, 34.4, 35.6, 35.9, (cyclohexylidene), 67.7 (t,
C-4 a), 72.0 (t, C-4 b), 77.8 (d, C-2 a), 79.1 (d, C-3 a), 79.4 (d, C-3 b), 84.7 (d,
C-2 b), 97.4 (d, C-1 a), 101.8 (d, C-1 b), 113.0 (s, C-5 b), 114.2 (s, C-5 a); IR
(KBr disc): nÄ ˆ 3370 (br, O-H stretch), 2934, 2860 (aliphatic C-H stretch),
1451, 1372 cmÀ1; HRMS (ES): m/z: calcd for C10H15O4: 199.0970; found:
199.0974 [MÀ À H]; elemental anlysis calcd (%) for C10H16O4: C 59.98, H,
8.05; found: C 60.01, H 8.05.
1
yellow oil (532 mg, 86%). H NMR (CDCl3, 400 MHz): d ˆ 1.30, 1.50 (2s,
2  3H; C(CH3)2), 2.84 (m, 2H; 2  CHH'N), 3.62 (d, 2J(H,H) ˆ 11.2 Hz,
2H; 2 Â CHH'N), 4.71 (m, 2H; 2 Â OCH); 13C NMR (100 MHz, CDCl3):
d ˆ 24.5, 26.0 (C(CH3)2), 67.2 (CH2N), 78.3 (OCH), 111.6 (C(CH3)2); IR
(KBr disc): nÄ ˆ 2989, 2840 cmÀ1 (aliphatic C-H stretch), 1464, 1382; HRMS
.
(FI): m/z: calcd for C7H12NO2Cl: 177.0557; found: 177.0558 [M ].
2,3-O-Isopropylidene-1,4-dideoxy-1,4-imino-1-N-dehydroerythritol
(3):
2,3-O-Cyclohexylideneerythritol (24): Sodium borohydride (39 mg,
1.02 mmol, 2.0 equiv) was added under Ar to a solution of 23 (102 mg,
0.51 mmol) in MeOH (4 mL). The reaction mixture was stirred for 30 min,
whereupon TLC (EtOAc/cyclohexane 3:1) showed consumption of starting
material (Rf 0.7) and formation of a major product (Rf 0.5). Solid
ammonium chloride was added to quench excess borohydride, and the
mixture was concentrated to yield a white solid. This crude product was
purified by flash column chromatography on silica gel (EtOAc/cyclohexane
3:2) to yield pure 24 as an oily white solid (91 mg, 89%). 1H NMR
(400 MHz, CDCl3): d ˆ 1.39 1.63 (m, 10H; cyclohexylidene), 3.09 (t, 2H;
3J(H,H) ˆ 6.1 Hz, 2  OH), 3.78 (m, 4H; H-1, H-1', H-4, H-4'), 4.28 (m, 2H;
H-2, H-3); 13C NMR (100 MHz, CDCl3, DEPT): d ˆ 23.6, 24.0, 25.0, 34.5,
37.4 (5t, cyclohexylidene), 60.8 (t, C-1, C-4), 76.5 (d, C-2, C-3), 109.0 (s,
C-5); IR (film): nÄ ˆ 3389 (br, O-H stretch), 2936, 2862 (aliphatic C-H
stretch), 1449, 1368 cmÀ1; HRMS (ES): m/z: calcd for C10H17O4: 201.1127;
found: 201.1120 [MÀ À H].
DBU (0.050 mL, 51 mg, 0.33 mmol, 1.1 equiv) was added under Ar to a
solution of 21 (54 mg, 0.30 mmol) in dry Et2O(3 mL). After 4 h TLC
(EtOAc/cyclohexane 1:4) indicated conversion of SM (Rf 0.3) to a major
product (Rf 0.1). The reaction mixture was filtered under Ar to remove
DBU ¥ HCl, and then concentrated to yield 3 as a yellow oily solid (43 mg,
100%). 1H NMR (400 MHz, CDCl3): d ˆ 1.22, 1.22 (2s, 2  3H; C(CH3)2),
3.79 (dddd, 2J(H,H) ˆ 17.3, 3J(H,H) ˆ 4.6 Hz, 2.9 Hz, 0.9 Hz, 1H; CHH'N);
3.94 (ddpt, 2J(H,H) ˆ 17.3, 3J(H,H) ˆ 2.0 Hz, 1.0 Hz, 1H; CHH'N), 4.58
(ddd, 3J(H,H) ˆ 5.5 Hz, 4.6 Hz, 1.0 Hz, 1H; NCH2CH), 4.92 (brd,
3
J(H,H) ˆ 5.5 Hz, 1H; N CHCH), 7.44 (dd, 3J(H,H) ˆ 3.0 Hz, 2.0 Hz,
ˆ
1H; N CH); 13C NMR (100 MHz, CDCl3): d ˆ 25.4, 26.7 (2q, 2  C(CH3)),
ˆ
ˆ
66.3 (t, CH2N), 76.3 (d, NCH2CH), 86.5 (d, N CCH), 111.4 (s, C(CH3)2),
À1
ˆ
165.5 (d, N CH); IR (film): nÄ ˆ 2935, 2860 cm (aliphatic C-H), 1649
ˆ
(C N); HRMS (CI): m/z: calcd for C7H12NO2: 142.0868; found: 142.0867
‡
[M ‡H]. Attempts to purify the imine by aqueous washing or by
chromatography on a short column of silica gel (30:70 Et2O/petroleum
ether (40-60) eluting with 1% Et3N) both led to extensive degradation.
1-Benzyl-3,4-O-cyclohexylidene-1,4-dideoxy-1,4-iminoerythritol
(26):
Methanesulfonyl chloride (0.11 mL, 170 mg, 1.48 mmol, 4.0 equiv) was
added under Ar at 08C to a solution of 24 (75 mg, 0.37 mmol, 1.0 equiv) and
triethylamine (0.21 mL, 150 mg, 1.48 mmol, 4.0 equiv) in dry CH2Cl2
(3.0 mL). After 45 min, TLC (Et2O) showed consumption of starting
material (Rf 0.7) and formation of a major product (Rf 0.4). The reaction
mixture was poured onto ice-water (20 mL) and diluted with CH2Cl2
(30 mL). The layers were separated and the organic layer washed with
1m HCl (aq) (10 mL), NaHCO3 (aq) (10 mL) followed by water (10 mL).
The organic layer was dried (MgSO4), filtered and concentrated to yield a
pale yellow oil (164 mg). This crude product was purified by flash column
chromatography on silica gel (Et2O) to yield pure 25 as a white solid
(113 mg, 85%). HRMS (CI): m/z: calcd for C12H23O8S2: 359.0834; found:
3,4-O-Cyclohexylidene-l-arabinose (22): Cyclohexanone dimethyl ketal
(0.61 mL, 575 mg, 3.99 mmol, 3.0 equiv) was added to a solution of l-
arabinose (14) (200 mg, 1.33 mmol, 1.0 equiv) and p-toluenesulphonic acid
monohydrate (2.5 mg, 0.0133 mmol, 0.01 equiv) in dimethylformamide
(DMF) (4 mL). After 4 h stirring, TLC (MeOH/CHCl3 1:9) showed
formation of a major product (Rf 0.3). Ion-exchange resin was added
(Dowex OHÀ form, 550A, Aldrich) to neutralize the acid. The mixture was
filtered, the residue washed with dry DMF, and the filtrate concentrated
under reduced pressure to yield a syrupy residue. This crude product was
purified by flash column chromatography on silica gel (MeOH/CHCl3 7:93
to yield 22 as a white solid (203 mg, 66%) as a mixture of anomers (a:b 1:1).
1H NMR (400 MHz, [D6]DMSO, COSY): d ˆ 1.34 1.65 (m, 10H; cyclo-
hexylidene), 3.16 (m, 0.5H; H-2 a), 3.33 (m, 0.5H; H-2 b), 3.65 (dd, 0.5H;
2J(H,H) ˆ 13.5, 3J(H,H) ˆ 2.9 Hz, H-4' a), 3.70 (d, 0.5H; 2J(H,H) ˆ
‡
359.0827 [M ‡H].
Benzylamine (3.0 mL) was added under Ar to 25 (113 mg, 0.32 mmol). The
mixture was stirred at 658C for 48 h. The reaction mixture was then allowed
to cool and diluted with EtOAc (30 mL) then washed with brine (10 mL)
followed by water (10 mL). The organic layer was dried (Na2SO4) and
concentrated to yield a yellow oil. Xylene was then added and the mixture
concentrated (3 Â 20 mL) to remove excess benzylamine as its azeotrope
with xylene. This crude product was purified by flash column chromatog-
raphy on silica gel (EtOAc/cyclohexane 3:17) to yield pure 26 as a
colourless oil (75 mg, 87%). 1H NMR (400 MHz, CDCl3, COSY): d ˆ 1.41,
1.57, 1.67, 1.84 (4m, 2H, 2H, 4H, 2H; cyclohexylidene), 2.16 (m, 2H; 2 Â
CHH'N), 3.05 (d, 2H; 2J(H,H) ˆ 11.4 Hz, 2  CHH'N), 3.63 (s, 2H,
CH2Ph), 4.65 (m, 2H; 2 Â OCH), 7.22 7.37 (m, 5H; C6H5); 13C NMR
(100 MHz, CDCl3, DEPT, HMQC): d ˆ 23.8, 24.2, 25.2, 34.5, 36.1 (5t,
cyclohexylidene). 59.3 (t, PhCH2), 59.8 (t, CH2N), 79.1 (d, OCH), 111.9 (s,
C(CH2)2), 126.9, 128.2, 128.5 (3d, 3 Â aromatic CH), 138.7 (s, quaternary
aromatic C); HRMS (CI): m/z: calcd for C17H24NO2: 274.1807; found:
3
13.5 Hz, H-4' b), 3.86 (dd, 0.5H; 3J(H,H) ˆ 6.9, J(H,H) ˆ 6.0 Hz, H-3 a),
3.94 4.01 (m, 1.5H; H-3 b, H-5 a, H-5 b), 4.08 (m, 0.5H; H-4 a), 4.14 (m,
0.5H; H-4 b), 4.19 (dd, 0.5H; 3J(H,H) ˆ 7.4, 3J(H,H) ˆ 6.6 Hz, H-1 a), 4.83
(dd, 0.5H; 3J(H,H) ˆ 4.6, 3J(H,H) ˆ 4.4 Hz, H-1 b), 4.89 (d, 0.5H;
3J(H,H) ˆ 6.8 Hz, 2-OH b), 5.12 (d, 0.5H; 3J(H,H) ˆ 5.0 Hz, 2-OH a),
6.32 (d, 0.5H; 3J(H,H) ˆ 4.7 Hz, 1-OH b), 6.60 (d, 0.5H; 3J(H,H) ˆ 6.4 Hz,
1-OH a); 13C NMR (100 MHz, [D6]DMSO, DEPT, HMQC, HMBC): d ˆ
23.4, 23.5, 23.7, 23.8, 24.6, 24.7, 35.1, 37.6, 37.7 (cyclohexylidene), 58.0 (t, C-5
b), 62.2 (t, C-5 a), 70.4 (d, C-2 b), 72.4 (d, C-4 b), 72.8 (d, C-4 a), 74.1 (d, C-2
a), 75.4 (d, C-3 b), 78.6 (d, C-3 a), 92.2 (d, C-1 b), 96.5 (d, C-1 a), 108.8 (s,
C-6 a), 113.0 (s, C-6 b). IR (KBr disc): nÄ ˆ 3382, 3272 (br, O-H stretch),
2951, 2856, (aliphatic C-H stretch), 1450 cmÀ1; HRMS (ES): m/z: calcd for
‡
C11H19O5: 231.1232; found: 231.1238 [M ‡H].
2,3-O-Cyclohexylidene-l-erythrose (23): Sodium periodate (320 mg,
1.49 mmol, 2.25 equiv) was added to a solution of 22 (151 mg, 0.66 mmol)
in water (6 mL). The reaction mixture was stirred under Ar. After 2.5 h,
solid sodium carbonate was added and the slurry stirred for 1 h. The
mixture was then diluted with water (30 mL) and extracted with EtOAc
(3 Â 20 mL) followed by CHCl3 (2 Â 20 mL). The combined organic layers
were dried (MgSO4), filtered and concentrated under reduced pressure to
yield a colourless oil (127 mg). This crude product was purified by flash
column chromatography (EtOAc/cyclohexane 35:65) to yield pure 23 as a
white solid (113 mg, 86%), as a mixture of anomers (a:b 1:5). 1H NMR
(400 MHz, CDCl3, COSY): d ˆ 1.37 1.69 (m, 10H; cyclohexylidene), 2.01
(s, 0.17H; OH a), 3.29 (d, 0.83H; 3J(H,H) ˆ 2.5 Hz, OH b), 3.54 (dd,
0.17H; 2J(H,H) ˆ 11.2, 3J(H,H) ˆ 3.8 Hz, H-4' a), 3.98 (dd, 0.17H;
‡
274.1808 [M ‡H].
2,3-O-Cyclohexylidene-1,4-dideoxy-1,4-iminoerythritol (27): Palladium hy-
droxide on carbon (20% Pd(OH)2, 25 mg) was added to a solution of 26
(75 mg, 0.27 mmol) in MeOH (3.0 mL)and the reaction mixture placed
under an atmosphere of H2. After 4 h TLC (EtOAc/cyclohexane 1:4)
showed consumption of SM (Rf 0.4). The reaction mixture was filtered
through Celite to remove Pd(OH)2 on C, then the filtrate concentrated
under reduced pressure to yield 27 as a colourless oil (48 mg, 97%).
1H NMR (400 MHz, CDCl3): d ˆ 1.36 1.67 (m, 10H; cyclohexylidene),
2.52 (m, 2H; 2 Â CHH'N), 2.64 (brs, 1H; NH), 3.12 (m, 2H; 2 Â CHH'N),
4.64 (m, 2H; 2  OCH); 13C NMR (100 MHz, CDCl3, DEPT): d ˆ 23.6,
3408
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
24.0, 25.2, 33.2, 35.7 (5t, cyclohexylidene), 54.2 (t, CH2N), 80.9 (d, OCH),
diluted with Et2O(30 mL) and washed with NH 4Cl (aq) (10 mL), water
(10 mL) followed by NaHCO3 (10 mL). The aqueous layers were basified
with 3m NaOH and back-extracted with Et2O(2 Â 10 mL) and the
combined organic layers dried (Na2SO4) and concentrated under reduced
pressure to yield a pale yellow oil (61 mg). The crude product was purified
by flash column chromatography on silica gel (EtOAc/hexane 3:7 increas-
ing to 1:1) to yield 29a (38 mg, 67%) as a mixture of diastereomers (5:1
110.9 (s, C(CH2)2); HRMS (CI): m/z: calcd for C10H18NO2: 184.1338;
‡
found: 184.1339 [M ‡H].
1-Chloro-2,3-O-cyclohexylidene-1,4-dideoxy-1,4-iminoerythritol (28): NCS
(39 mg, 0.29 mmol, 1.1 equiv) was added under Ar to a solution of 27
(48 mg, 0.26 mmol) in dry Et2O(5 mL). After 2 h stirring TLC (EtOAc/
cyclohexane 1:4) showed consumption of SM (Rf 0.0) and formation of a
major product (Rf 0.6). The reaction mixture was diluted with Et2O(30 mL)
and washed with water (3 Â 10 mL). The organic layer was dried (Na2SO4)
and concentrated to yield an oily solid (51 mg, 90%). This was consistent
with pure N-chloramine and no further purification was required. 1H NMR
(CDCl3, 400 MHz): d ˆ 1.19 1.75 (m, 10H; 5  CH2 cyclohexylidene), 2.84
(m, 2H; 2  CHH'N), 3.63 (d, 2J(H,H) ˆ 11.7 Hz, 2H; 2  CHH'N), 4.71
(m, 2H; 2  OCH); 13C NMR (100 MHz, CDCl3, DEPT): d ˆ 23.7, 24.0,
25.1, 33.9, 35.6 (5t, 5 Â CH2 cyclohexylidene), 67.3 (t, CH2N), 77.8 (d,
OCH), 112.4 (C(CH2)2); IR (KBr disc): nÄ ˆ 2935, 2853 (aliphatic C-H
1
anti:syn as determined by H NMR).
1a,b-Ethyl-2,3-O-di-tert-butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol
(29b)
i) Normal mode of addition: According to method described for 29a above
using ethyl magnesium bromide (3.0m solution in Et2O, 0.50 mL,
1.49 mmol, 5.5 equiv) was added to yield a pale yellow oil (84 mg). The
crude product was purified by flash column chromatography on silica gel
(EtOAc/hexane 1:3) to yield ethyl adducts a,b-29b (44 mg, 45%, dr 98:2 as
determined by 1H NMR). The major diastereomer was purified and shown
by NOESY NMR experiments to be the anti diastereomer a-29b. 1H NMR
(500 MHz, CDCl3, COSY, NOESY): d ˆ 0.05, 0.06, 0.06, 0.07 (4s, 4  3H;
4  Si(CH3)), 0.87, 0.87 (2s, 2  9H; 2  SiC(CH3)3), 0.98 (t, 3J(H,H) ˆ
7.5 Hz, 3H; NCH(CH2CH3)), 1.48 (m; 1H, NCH(CHH'CH3)), 1.64 (m,
1H; NCH(CHH'CH3)), 2.27 (brs, 1H; NH), 2.73 (m, 1H; NCH(Et)), 2.78
(d, 2J(H,H) ˆ 12.5 Hz, 1H; CHH'N), 2.98 (dd, 2J(H,H) ˆ 12.5, 3J(H,H) ˆ
3.5 Hz, 1H; CHH'N), 3.63 (m, 1H; NCH(Et)CHOSi), 3.90 (m, 1H;
NCH2CHOSi); 13C NMR (125.7 MHz, CDCl3, 1H-13C HETCOR):
d ˆ À4.7, À4.6, À4.6, À4.4 (OSi(CH3)2), 11.7 (NCH(CH2CH3)), 17.9
(SiC(CH3)3), 25.7, 25.8 (SiC(CH3)3), 26.8 (NCH(CH2CH3)), 53.7 (NCH2),
69.3 (NCH(CH2CH3)), 79.8 (NCH2CHOSi), 83.4 (NCH(Et)CHOSi); IR
(film): nÄ ˆ 2956, 2929, 2857 cmÀ1 (aliphatic C-H); MS (ES): m/z (%): 360
stretch), 1457, 1373 cmÀ1
217.0870; found: 217.0868 [M ].
;
HRMS (FI): m/z: calcd for C10H16NO2Cl:
.
2,3-O-Cyclohexylidene-1,4-dideoxy-1,4-imino-1,N-dehydroerythritol (4):
DBU (0.045 mL, 45 mg, 0.30 mmol, 1.3 equiv) was added under Ar to a
solution of 28 (50 mg, 0.23 mmol) in dry Et2O(3 mL). After 3 h, the
mixture was filtered under Ar to remove DBU ¥ HCl, and then concen-
trated to yield 4 as an oil (42 mg, 100%). 1H NMR (400 MHz, CDCl3): d ˆ
1.46 1.58 (m, 10H; cyclohexylidene), 3.86 (dddd, 2J(H,H) ˆ 17.2,
3J(H,H) ˆ 4.5 Hz, 2.9 Hz, 0.8 Hz, 1H; CHH'N), 4.05 (ddpt, 2J(H,H) ˆ
17.2, 3J(H,H) ˆ 2.0 Hz, 1.0 Hz, 1H; CHH'N), 4.65 (ddd, 1H, 3J(H,H) ˆ
5.6 Hz, 4.5 Hz, 0.9 Hz, 1H; NCH2CH), 5.00 (brd, 3J(H,H) ˆ 5.6 Hz, 1H;
3
ˆ
ˆ
N CHCH), 7.53 (dd, J(H,H) ˆ 2.8 Hz, 2.2 Hz, 1H; N CH).
‡
(100) [M ‡H]; HRMS (ES): m/z: calcd for C18H42NO2Si2: 360.2754;
1a,b-Methyl-2,3-O-di-tert-butyldimethylsilyl-1,4-dideoxy-1,4-dideoxy-1,4-
iminothreitol (29a)
‡
found: 360.2752 [M ‡H].
ii) Reverse mode of addition: According to method described for 29a above
using ethylmagnesium bromide (3.0m solution in Et2O, 0.27 mL,
0.82 mmol, 5.0 equiv) for adduct formation over 1.5 h to give a pale yellow
oil (57 mg). The crude product was purified by flash column chromatog-
raphy on silica gel (Et2O/40 60 petroleum ether 3:7 increasing to 1:1) to
yield 29b (41 mg, 70%) as a single diastereomer.
i) General method for normal mode of addition: DBU (0.16 mL, 164 mg,
1.08 mmol, 4.0 equiv) was added under N2 to a solution of 10 (100 mg,
0.27 mmol) in dry Et2O(6 mL). The reaction mixture was stirred and
DBU ¥ HCl observed to precipitate as a white solid. After 2.5 h the reaction
mixture was filtered through a glass sinter and the filtrate was concentrated
under reduced pressure to yield an oil. Formation of imine 2 was confirmed
1a,b-Allyl-2,3-di-O-tert-butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol
(29c)
1
by H NMR. The oil was then redissolved in dry Et2O(6 mL) and methyl
magnesium bromide (3.0m solution in Et2O, 0.50 mL, 1.49 mmol, 5.5 equiv)
was added. The reaction mixture was stirred for 3 h, diluted with Et2O
(20 mL) and washed with NaHCO3 (3 Â 10 mL). The aqueous layers were
extracted with Et2O(10 mL) and the combined organic layers dried
(Na2SO4), filtered and concentrated under reduced pressure to yield a pale
yellow oil (131 mg). The crude product was purified by flash column
chromatography on silica gel (EtOAc/hexane 3:7 then EtOAc) eluting
initially as a single diastereomer (23 mg) and then as a mixture of
diastereoisomers (32 mg) to give methyl adduct 29a (total mass 55 mg,
58%, dr 5:2 as determined from 1H NMR). The major diastereomer was
shown by NMR experiments to be the anti or a diastereomer a-29a.
1H NMR (500 MHz, CDCl3, COSY, NOESY): d ˆ À0.02 (s, 3H; SiCH3),
À0.01 (s, 3H; SiCH3), 0.00 (s, 6H; Si(CH3)2), 0.81, 0.81 (2s, 2 Â 9H; 2 Â
SiC(CH3)3), 1.14 (d, 3J(H,H) ˆ 7.0 Hz, 3H; NCH(CH3)), 2.02 (brs, 1H;
NH), 2.72 (d, 2J(H,H) ˆ 12.0 Hz, 1H; CHH'N), 2.82 (dq, 3J(H,H) ˆ 6.5,
3J(H,H) ˆ 3.0 Hz, 1H; NCH(CH3)), 2.94 (dd, 2J(H,H) ˆ 12.0, 3J(H,H) ˆ
4.0 Hz,1H; CHH'N), 3.50 (m, 1H; NCH(Me)CHOSi), 3.85 (m, 1H;
NCH2CHOSi); 13C NMR (125.7 MHz, CDCl3, 1H-13C HETCOR): d ˆ
À4.7, À4.7, À4.6, À4.5 (OSi(CH3)2), 17.9, 17.9 (SiC(CH3)3), 19.2
(NCH(CH3)), 25.8 (SiC(CH3)3), 54.1 (NCH2), 62.7 (NCH(CH3)), 80.3
(NCH2CHOSi), 85.4 (NCH(Me)CHOSi); IR (film): nÄ ˆ 2956, 2929, 2896,
i) Normal mode of addition: According to method described for 29a above
using DBU (0.12 mL, 123 mg, 0.81 mmol, 3.0 equiv) for imine formation
over 3.5 h and allylmagnesium bromide (1.0m solution in Et2O, 1.35 mL,
1.35 mmol, 5.0 equiv) over 2 h for adduct formation to yield an orange oil
(90 mg). This crude product was purified by flash column chromatography
on silica gel (EtOAc/hexane 3:17) to yield allyl adducts, major diaster-
eomer a-29c (24 mg), and a minor diastereomer b-29c (2 mg) (total
mass ˆ 26 mg, 26%, dr 12:1). The major diastereomer was shown by
NOESY NMR to be the anti diastereomer a-29c: 1H NMR (500 MHz,
CDCl3, COSY, NOESY): d ˆ 0.06 (s, 6H; Si(CH3)2), 0.07, 0.08 (2s, 2  3H;
2 Â Si(CH3)), 0.87, 0.89 (2s, 2 Â 9H; 2 Â SiC(CH3)3), 2.27 (m, 1H;
CHH'CH CH2), 2.37 (m, 1H; CHH'CH CH2), 2.80 (d, 2J(H,H) ˆ
12.0 Hz, 1H; CHH'N), 2.90 (td, 3J(H,H) ˆ 7.0, 3J(H,H) ˆ 2.5 Hz, 1H;
NCH(allyl)), 3.03 (dd, 2J(H,H) ˆ 12.0, 3J(H,H) ˆ 4.0 Hz, 1H; CHH'N),
3.70 (brs, 1H; NCH2CHOSi), 3.91 (m, 1H; NCH(allyl)CHOSi), 5.07 (brd,
ˆ
ˆ
3J(H,H)cis ˆ 10.0 Hz, 1H; CH CHH'), 5.10 (brd, 3J(H,H)trans ˆ 17.5 Hz,
ˆ
1H; CH CHH'), 5.83 (ddpt, 3J(H,H)trans ˆ 17.5, 3J(H,H)cis ˆ 10.0,
ˆ
3J(H,H) ˆ 7.0 Hz, 1 H ; CH2CH CH2); 13C NMR (125.7 MHz, CDCl3,
ˆ
DEPT, 1H-13C HSQC): d ˆ À4.7, À4.7, À4.5, À4.4 (4q, 4  SiCH3), 17.9
ˆ
(s, SiC(CH3)3), 25.8 (q, SiC(CH3)3), 38.1 (t, CH2CH CH2), 53.9 (t, NCH2),
‡
2858 cmÀ1 (aliphatic C-H); MS (ES): m/z (%): 346 (100) [M ‡H]; HRMS
66.7 (d, NCH(allyl)), 79.7 (d, NCH2CHOSi), 82.7 (d, NCH(allyl)CHOSi),
‡
ˆ
ˆ
116.9 (t, CH2CH CH2), 135.7 (d, CH2CH CH2); IR (film): nÄ ˆ 2955, 2929,
(ES): m/z: calcd for C17H40NO2Si2: 346.2597; found: 346.2602 [M ‡H].
2897, 2857 cmÀ1 (aliphatic C-H), 1641 (C C stretch); MS (ES): m/z (%):
ˆ
ii) Reverse mode of addition: DBU (0.074 mL, 75 mg, 0.49 mmol, 3.0 equiv)
was added under Ar to a solution of 10 (60 mg, 0.16 mmol) in dry Et2O
(5 mL). The reaction mixture was stirred and DBU ¥ HCl observed to
precipitate as a white solid. After 4 h, the reaction mixture was filtered
under Ar and the filtrate was concentrated under reduced pressure to yield
imine 2 as an oil. This was dissolved in dry Et2O(5 mL) under Ar. To a
separate flask under Ar was added methylmagnesium bromide (3.0m
solution in Et2O, 0.27 mL, 0.82 mmol, 5.0 equiv), then the imine solution
was added dropwise by syringe to the Grignard reagent. The reaction
mixture was stirred for 3 h, then quenched by the addition of NH4Cl (aq.),
‡
372 (100) [M ‡H]; HRMS (ES): m/z: calcd for C19H42NO2Si2: 372.2754;
‡
found: 372.2747 [M ‡H].
1a-Phenyl-2,3-di-O-tert-butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol
(29d)
i) Normal mode of addition: According to method described for 29a above
using DBU (0.040 mL, 41 mg, 0.27 mmol, 1.0 equiv) for imine formation
over 90 min and phenylmagnesium bromide (3.0m solution in Et2O,
0.14 mL, 0.41 mmol, 1.5 equiv) for adduct formation over 90 min to yield
a yellow oil (108 mg). The crude product was purified by flash column
Chem. Eur. J. 2003, 9, 3397 3414
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3409

B. G. Davis et al.
FULL PAPER
chromatography on silica gel (EtOAc/hexane 1:3) to yield 29d as a
colourless oil (41 mg, 37%), as a single diastereomer. This was shown by
NOESY NMR experiments to be the anti diastereomer. 1H NMR
(500 MHz, CDCl3, COSY, NOESY): d ˆ À0.19, À0.06 (2s, 2  3H; 2 Â
PhCHCHOSi(CH3)), 0.10 (s, 6H; NCH2CHOSi(CH3)2), 0.83 (s, 9H;
PhCHCHOSiC(CH3)3), 0.92 (s, 9H; NCH2CHOSiC(CH3)3), 2.65 (brs,
1H; NH), 3.00 (d, 2J(H,H) ˆ 12.0 Hz, 1H; CHH'N,), 3.17 (dd, 2J(H,H) ˆ
12.0, 3J(H,H) ˆ 4.0 Hz, 1H; CHH'N), 3.85 (d, 3J(H,H) ˆ 3.0 Hz, 1H;
CHPh), 3.98 (brs, 1H; SiOCHCHPh), 4.09 (m, 1H; SiOCHCH2N), 7.25
7.33 (m, 3H; 3 of C6H5), 7.41 (m, 2H; 2 of C6H5); 13C NMR (125.7 MHz,
CDCl3, 1H-13C HETCOR): d ˆ À4.8, À4.7, À4.7, À4.6 (OSi(CH3)2), 17.8,
18.0 (SiC(CH3)3), 25.6, 25.7, 25.8, 25.8 (SiC(CH3)3), 54.4 (CH2N), 72.2
(CHPh), 80.1 (SiOCHCH2N), 86.2 (SiOCHCHPh), 127.3, 127.7, 128.4 (3 Â
aromatic CH), 141.8 (quaternary aromatic C); IR (film): nÄ ˆ 3027 cmÀ1
(aromatic C H), 2954, 2929, 2857 (aliphatic C H); MS (ES): m/z (%): 408
water (10 mL) followed by NaHCO3 (aq) (10 mL). The aqueous layers
were extracted with CHCl3 (2 Â 10 mL) and the combined organic layers
dried (Na2SO4), filtered and concentrated under reduced pressure to yield a
yellow oily solid (277 mg). This crude product was purified by flash column
chromatography on silica gel (EtOAc/cyclohexane 1:19 then EtOAc/
cyclohexane 1:4 the EtOAc) to yield 29 f (53 mg, 43%) as a single
diastereomer. This was shown by NOESY NMR experiments to be the anti
diastereomer. 1H NMR (400 MHz, CDCl3, COSY, NOESY): d ˆ À0.16,
À0.07, 0.07, 0.10 (4s, 4 Â 3H; 4 Â SiCH3), 0.81, 0.92 (2s, 2 Â 9H; 2 Â
SiC(CH3)3), 2.14 (brs, 1H; NH), 2.78 (m, 2H; CH2C6H4), 2.82 (d,
2J(H,H) ˆ 12.0 Hz, 1H; CHH'N), 3.06 (dd, 2J(H,H) ˆ 12.0, 3J(H,H) ˆ
4.0 Hz, 1H; CHH'N), 3.08 (td, 3J(H,H) ˆ 7.8, 3J(H,H) ˆ 2.2 Hz, 1H;
NCHCH2), 3.74 (brs, 1H; NCHCHOSi), 3.79 (s, 3H; OCH3), 3.92 (m,
1H; NCH2CHOSi), 6.84 (m, 2H; 2 of C6H4), 7.14 (m, 2H; 2 of C6H4);
13C NMR (100 MHz, CDCl3, HSQC, DEPT): d ˆ À5.0, À4.7, À4.7, À4.6
(4q, 4 Â SiCH3), 17.8, 17.9 (2s, 2 Â SiC(CH3)3), 25.7, 25.9 (2q, 2 Â
SiC(CH3)3), 39.1 (t, C6H4CH2), 53.9 (t, NCH2), 55.3 (q, OCH3), 68.9 (d,
NCH), 79.6 (d, NCH2CHOSi), 82.1 (d, NCHCHOSi), 113.9, 130.0 (2d, 2 Â
aromatic CH), 131.5 (s, quaternary aromatic C), 158.0 (s, quaternary
aromatic C); IR (KBr disc): nÄ ˆ 2951, 2929, 2857 cmÀ1 (aliphatic C-H),
‡
(100) [M ‡H]; HRMS (CI): m/z: calcd for C22H42NO2Si2: 408.2754; found:
‡
408.2747 [M ‡H].
ii) Reverse mode of addition: According to method described for 29a using
phenylmagnesium bromide (3.0m solution in Et2O, 0.27 mL, 0.82 mmol,
5.0 equiv) for adduct formation over 1.5 h to yield a pale yellow oil (75 mg).
The crude product was purified by flash column chromatography on silica
gel (Et2O/40-60 petroleum ether 3:17 increasing to 3:7) to yield 29d (52 mg,
78%) as a single diastereomer.
‡
1613, 1514; MS (APCI): m/z (%): 452 (100) [M ‡H]; HRMS (CI): m/z:
‡
calcd for C24H46NO3Si2: 452.3016; found: 452.3013 [M ‡H].
ii) Reverse mode of addition: According to method described for 29a using
a THF solution of imine (6 mL) and 4-methoxybenzylmagnesium chloride
(0.25m solution in Et2O, 3.28 mL, 0.82 mmol, 5.0 equiv) for adduct
formation over 2 h to give a pale yellow oil (97 mg). The crude product
was purified by flash column chromatography on silica gel (EtOAc/
cyclohexane 1:4) to yield 29 f (46 mg, 62%) as a single diastereomer.
1a-Benzyl-2,3-O-di-tert-butyldimethylsilyl-1,4-dideoxy-1,4-iminothreitol
(29e)
i) Normal mode of addition: According to method described for 29a above
using DBU (0.040 mL, 41 mg, 0.27 mmol, 1.0 equiv) for imine formation
over 4 h and benzylmagnesium chloride (1.0m solution in Et2O, 0.41 mL,
0.41 mmol, 1.5 equiv) for adduct formation over 3 h TLC to yield an orange
oil (108 mg). The crude product was purified by flash column chromatog-
raphy on silica gel (EtOAc/hexane 1:4) whereupon 29e was obtained as an
oil (55 mg, 48%), as a single diastereomer. This was shown by NOESY
NMR experiments to be the anti diastereomer. 1H NMR (500 MHz,
CDCl3, COSY, NOESY): d ˆ À0.19, À0.08 (2s, 2  3H; 2  BnCHCHO-
Si(CH3)), 0.08, 0.11, (2s, 2 Â 3H; 2 Â NCH2CHOSi(CH3)), 0.80 (s, 9H;
BnCHCHOSiC(CH3)3), 0.93 (s, 9H; NCH2CHOSiC(CH3)3), 2.32 (brs, 1H;
NH), 2.84 (m, 3H; C6H5CH2 and CHH'N), 3.08 (brd, 2J(H,H) ˆ 10.0 Hz,
1H; CHH'N), 3.14 (m, 1H; BnCH), 3.76 (s, 1H; BnCHCHOSi), 3.93 (s,
1H; NCH2CHOSi), 7.31 7.19 (m, 5H; C6H5); 13C NMR (125.7 MHz,
CDCl3, 1H-13C HETCOR): d ˆ À5.1, À4.8, À4.7, À4.7 (Si(CH3)2), 17.9, 17.8
(SiC(CH3)3), 25.8, 25.7, 25.7 (SiC(CH3)3), 40.1 (PhCH2), 53.9 (NCH2), 68.7
(BnCH), 79.6 (NCH2CHOSi), 82.0 (BnCHCHOSi), 126.2, 128.5, 129.2
(aromatic CH), 139.4 (quaternary aromatic C); IR (film): nÄ ˆ 3027 cmÀ1
(aromatic C-H), 2953, 2928, 2857 (aliphatic C-H); MS (CI): m/z (%): 422
1
Dimer product 32: H NMR (200 MHz, CDCl3): d ˆ 0.04, 0.06, 0.08, 0.12,
0.14 (5s, 24H; 8 Â Si(CH3)), 0.87, 0.89, 0.94 (m, 36H; 12 Â SiC(CH3)), 2.59
(dd, 2J(H,H) ˆ 12.0, 3J(H,H) ˆ 9.4 Hz, 2H; 2  CHH'N), 2.71 (d,
3J(H,H) ˆ 3.0 Hz, 2H; 2  NCH), 3.48 (dd, 2J(H,H) ˆ 12.0, 3J(H,H) ˆ
8.3 Hz, 2H; 2  CHH'N), 3.97 (dd, 3J(H,H) ˆ 6.6, 3J(H,H) ˆ 3.0 Hz, 2H;
3
3
NCHCHOSi), 4.45 (ptd, J(H,H) ˆ 8.8, J(H,H) ˆ 6.6 Hz, 2H; NCH2CH);
13C NMR (100 MHz, CDCl3): d ˆ À5.0, À4.8 (SiCH3), 17.9, 18.1
(SiC(CH3)), 25.7, 25.8 (SiC(CH3)), 54.6, 56.5 (NCH2, NCH), 78.1, 86.5
(NCH2CH, NCHCH).
1b-Hex-5-enyl-2,3-O-isopropylidene-1,4-dideoxy-1,4-iminoerythritol (30g)
i) Imine formation: DBU (0.11 mL, 111 mg, 0.73 mmol, 1.3 equiv) was
added under Ar to a solution of 21 (100 mg, 0.56 mmol) in dry THF (6 mL).
The reaction mixture was stirred, whereupon the white salt DBU ¥ HCl
precipitated. After 3 h the reaction mixture was filtered under Ar to
remove DBU ¥ HCl and concentrated under reduced pressure to yield an
orange oil. This was dissolved in dry THF (5 mL) under Ar.
‡
(100) [M ‡H]; HRMS (CI): m/z: calcd for C23H44NO2Si2: 422.2910; found:
ii) Preparation of Grignard reagent: To magnesium turnings (204 mg,
8.4 mmol, 15.0 equiv) in dry THF (2 mL) under Ar with two crystals of I2
was added 6-bromo-1-hexene (0.75 mL, 913 mg, 5.6 mmol, 10.0 equiv) as a
solution in THF (6 mL) over 20 min. The activity and identity of this
Grignard stock solution was confirmed by reaction of a small portion with
benzaldehyde which yielded only 1-phenyl-hept-6-en-1-ol.
‡
422.2911 [M ‡H].
ii) Reverse mode of addition: According to method described for 29a using
benzylmagnesium chloride (1.0m solution in Et2O, 0.71 mL, 0.71 mmol,
5.0 equiv) for adduct formation over 1.5 h to give a pale yellow oil (68 mg).
The crude product was purified by flash column chromatography on silica
gel (EtOAc/cyclohexane 1:4 increasing to 1:1) to yield 29e (38 mg, 63%) as
a single diastereomer.
iii) Normal mode addition: To the stirred imine 3 as a solution in dry THF
(5 mL) under Ar was added a portion of the Grignard preparation (4 mL,
2.8 mmol, 5.0 equiv). After 3 h excess Grignard reagent was quenched by
the dropwise addition of NH4Cl (aq). The reaction mixture was concen-
1a-(4-Methoxybenzyl)-2,3-di-O-tert-butyldimethylsilyl-1,4-dideoxy-1,4-
iminothreitol (29 f)
i) Normal mode of addition: According to method described for 29a above
using DBU (0.12 mL, 123 mg, 0.81 mmol, 3.0 equiv) for imine formation
over 3.5 h. Benzylmagnesium chloride was prepared as follows. To
magnesium turnings (0.33g, 13.5 mmol, 50.0 equiv) in dry THF (5 mL)
under Ar with two crystals of I2 was added 4-methoxybenzyl chloride
(0.92 mL, 1.06 g, 6.75 mmol, 25.0 equiv) as solution in THF (10 mL) over
ꢀ30 min. The mixture turned a dark grey colour on this addition. After 1 h
a small portion was removed and quenched with D2O, then dissolved in
[D4]methanol for 1H NMR. Grignard reagent formation was confirmed by
the presence of the CH2D resonance (triplet, 1H NMR d ˆ 2.28 ppm).
4-methoxybenzylmagnesium chloride solution in THF (6 mL, 2.7 mmol,
10 equiv) was added to a solution of imine in dry THF (5 mL) under Ar.
The reaction course was followed by mass spectrometry (APCI) and after
2 h excess Grignard reagent was quenched by the dropwise addition of
NH4Cl (aq). The reaction mixture was concentrated to remove THF, the
residue dissolved in CHCl3 (30 mL) and washed with NH4Cl (aq) (10 mL),
trated and the residue dissolved in Et2O(30 mL) and washed with NH Cl
4
(aq., 10 mL), water (10 mL), NaHCO3 (aq., 10 mL). The aqueous layers
were basified with NaOH (aq., 1m) extracted with Et2O(2 Â 10 mL) and
the combined organic layers dried (Na2SO4), filtered and concentrated
under reduced pressure to yield a yellow oily solid. This crude product was
purified by flash column chromatography on silica gel (Et2O/40-60
petroleum ether 3:7 then Et2Othen EtOAc) to give 30g (58 mg, 46%) as
a single diastereomer. This was shown by NOESY NMR experiments to be
the anti b diastereomer. 1H NMR (400 MHz, CDCl3, COSY, NOESY): d ˆ
1.26 (m, 2H; 1 of (CH2)3), 1.31 (s, 3H; 1 of C(CH3)2), 1.40 (m, 4H; 2 of
ˆ
(CH2)3), 1.46 (s, 3H; 1 of C(CH3)2), 2.05 (m, 2H; CH2CH CH2), 2.21 (br s,
1H; NH), 2.82 (dd, 2J(H,H) ˆ 13.5, 3J(H,H) ˆ 4.2 Hz, 1H; CHH'N), 2.99
(d, 2J(H,H) ˆ 13.5 Hz, 1H; CHH'N), 3.08 (pt, 3J(H,H) ˆ 7.7 Hz, 1H;
NCH(hexenyl)), 4.37 (d, 3J(H,H) ˆ 5.5 Hz, 1H; NCH(hexenyl)CH), 4.67
(pt, 3J(H,H) ˆ 4.8 Hz, 1H; NCH2CH), 4.93 (ddpt, 3J(H,H)cis ˆ 10.2,
2
3
3
ˆ
J(H,H) ˆ 2.0, J(H,H) ˆ 1.1 Hz, 1H; 1 of CH CH2), 4.99 (dpq, J(H,H)trans
3410
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
3
17.1 Hz, 2/3J(H,H) ˆ 2.0 Hz, 1H; 1 of CH CH2), 5.79 (ddpt, J(H,H)trans 17.1,
2986, 2935 (aliphatic C-H stretch), 1496, 1449; HRMS (CI): m/z: calcd for
ˆ
3J(H,H)cis 10.2, J(H,H), 6.7 Hz, 1H; CH2CH CH2); 13C NMR (100 MHz,
CDCl3, HMQC, DEPT): d ˆ 23.9, 26.2 (2q, 2  (C(CH3)2), 26.3, 28.7, 30.4,
33.6 (4t, 4 Â CH2 hexenyl), 51.7 (t, NCH2), 64.8 (d, NCH(hexenyl)), 81.8 (d,
CH2CH), 86.0 (d, NCH(hexenyl)CH), 110.7 (s, C(CH3)2), 114.4 (t,
C13H18NO2: 220.1338; found: 220.1333 ([M ‡H].
1b-Benzyl-2,3-O-isopropylidene-1,4-dideoxy-1,4-iminoerythritol
3
‡
ˆ
(30e):
According to method described for 30b above using dry Et2O(6 mL) for
imine formation and dry Et2O(6 mL) with benzylmagnesium chloride
(1.0m solution in Et2O, 2.80 mL, 2.80 mmol, 5.0 equiv) over 1.5 h for adduct
formation to yield a yellow oil (168 mg). This crude product was purified by
flash column chromatography on silica gel (Et2O/40-60 petroleum ether 2:3
increasing up to Et2O) to yield 30e as a pale yellow oil (96 mg, 73%), as a
single diastereomer. This was confirmed by NOESY NMR experiments to
be the anti diastereomer. 1H NMR (400 MHz, CDCl3, COSY, NOESY):
d ˆ 1.29, 1.46 (2s, 2  3H; 2  C(CH3)), 2.49 (brs, 1H, NH), 2.61 (dd,
ˆ
ˆ
CH CH2), 138.8 (d, CH CH2); IR (film): nÄ ˆ 3315 (N-H), 2979, 2932,
2857 cmÀ1 (aliphatic C-H), 1641 (C C); HRMS (ES): m/z: calcd for
ˆ
‡
C13H24NO2: 226.1807; found: 226.1807 [M ‡H].
1b-Ethyl-2,3-O-isopropylidene-1,4-dideoxy-1,4-iminoerythritol
(30b):
DBU (0.16 mL, 167 mg, 1.10 mmol, 1.3 equiv) was added under Ar to a
solution of 21 (150 mg, 0.84 mmol) in dry Et2O(8 mL). The reaction
mixture was stirred for 3 h, then the mixture was filtered under Ar to
remove DBU hydrochloride. The filtrate was concentrated under reduced
pressure to yield an oil. This was dissolved in dry Et2O(8 mL) under Ar. To
a separate flask under Ar was added ethylmagnesium bromide (3.0m
solution in Et2O, 1.40 mL, 4.20 mmol, 5.0 equiv), and the imine solution
was added to the Grignard reagent dropwise by syringe. After 1.5 h TLC
and mass spectrometry (APCI) showed consumption of imine and
formation of the desired adduct; excess Grignard reagent was quenched
with NH4Cl (aq), then the mixture was diluted with Et2O(30 mL) and
washed with NH4Cl (aq) (10 mL), water (10 mL) and NaHCO3 (aq.,
10 mL). The aqueous layers were basified to pH 11 with 3m NaOH (aq)
3
2
2J(H,H) ˆ 14.0, J(H,H) ˆ 8.0 Hz, 1H; CHH'Ph), 2.67 (dd, J(H,H) ˆ 14.0,
3J(H,H) ˆ 8.0 Hz, 1H; CHH'Ph), 2.99 (dd, 2J(H,H) ˆ 13.3, 3J(H,H) ˆ
4.1 Hz, 1H; NCHH'), 3.07 (d, 2J(H,H) ˆ 13.3 Hz, 1H; NCHH'), 3.47 (t,
3J(H,H) ˆ 7.8 Hz, 1 H ; CHBn), 4.46 (d, 3J(H,H) ˆ 5.6 Hz, 1H; CH(Bn)CH),
4.74 (m, 1H; NCH2CH), 7.20 7.33 (m, 5H; C6H5); 13C NMR (100 MHz,
CDCl3, DEPT, HMQC): d ˆ 24.0, 26.3 (2q, 2  C(CH3)), 36.9 (t, CH2Ph),
51.8 (t, NCH2), 66.1 (d, CHBn), 81.7 (d, NCH2CH), 84.9 (d, CH(Ph)CH),
110.9 (s, CMe2), 126.3, 128.5, 129.0 (3d, aromatic CH), 138.9 (s, quaternary
aromatic C); IR (film): nÄ ˆ 3321 cmÀ1 (br, N-H stretch), 3062, 3027
(aromatic C-H stretch), 2985, 2934, 2865 (aliphatic C-H stretch), 1496,
1454; HRMS (CI): m/z: calcd for C14H20NO2: 234.1494; found: 234.1504
then extracted with Et2O(10 mL) and CHCl
(2 Â 10 mL) and the
3
‡
combined organic layers dried (Na2SO4), filtered and concentrated under
reduced pressure to yield 30b as a yellow oil as a single diastereomer
(110 mg, 76%). (Careful control of pressure was required to avoid
removing the product under vacuum, minimum p ˆ 70 mbar.). 1H NMR
(400 MHz, CDCl3, COSY, NOESY): d ˆ 0.97 (t, 3J(H,H) ˆ 7.4 Hz, 3H;
CH2CH3), 1.30 (m, 2H; CH2CH3), 1.31, 1.47 (2s, 2 Â 3H; 2 Â C(CH3)), 2.22
[M ‡H].
1b-(4-Methoxybenzyl)-2,3-O-isopropylidene-1,4-dideoxy-1,4-iminoerythri-
tol (30 f): According to method described for 30b above using dry Et2O
(6 mL) for imine formation and dry THF (6 mL) with 4-methoxybenzyl-
magnesium chloride (0.25m solution in THF, 11.2 mL, 2.80 mmol, 5.0 equiv,
prepared as described above) over 1.5 h for adduct formation to yield a
yellow oil (421 mg). This crude product was purified by flash column
chromatography on silica gel (Et2O/40-60 petroleum ether 4:1, increasing
to Et2Othen MeOH/Et 2O1:9) to yield 30 f as a pale yellow oil (108 mg,
73%), as a single diastereomer. This was confirmed by NOESY NMR
experiments to be the anti diastereomer. 1H NMR (400 MHz, CDCl3,
COSY, NOESY): d ˆ 1.28, 1.45 (2s, 2  3H; 2  C(CH3)), 2.45 (brs, 1H;
NH), 2.54 (dd, 1H, 2J(H,H) ˆ 14.1, 3J(H,H) ˆ 8.0 Hz, CHH'Ar), 2.60 (dd,
(brs, 1H; NH), 2.83 (dd, 2J(H,H) ˆ 13.8, J(H,H) ˆ 4.1 Hz, 1H; NCHH'),
3
2.99 (d, 2J(H,H) ˆ 13.5 Hz, 1H; NCHH'), 3.01 (d, 3J(H,H) ˆ 7.7 Hz, 1H;
CHEt), 4.39 (d, 3J(H,H) ˆ 5.2 Hz, 1H; CH(Ph)CH), 4.68 (pt, 3J(H,H) ˆ
5.1 Hz, 1H, NCH2CH); 13C NMR (100 MHz, CDCl3, DEPT, HMQC): d ˆ
11.4 (q, CH2CH3), 23.6 (t, CH2CH3), 23.9, 26.3 (2q, 2 Â C(CH3)), 51.7 (t,
NCH2), 66.6 (d, CHEt), 81.8 (d, NCH2CH), 85.6 (d, CH(Et)CH), 110.7 (s,
C(CH3)2); IR (film): nÄ ˆ 3337 cmÀ1 (br, N-H stretch), 2962, 2934, 2874
(aliphatic C-H stretch); HRMS (CI): m/z: calcd for C9H18NO2: 172.1338;
3
2
2J(H,H) ˆ 14.0, J(H,H) ˆ 7.8 Hz, 1H; CHH'Ar), 2.96 (dd, J(H,H) ˆ 13.4,
3J(H,H) ˆ 4.0 Hz, 1H; NCHH'), 3.05 (d, 2J(H,H) ˆ 13.3 Hz, 1H; NCHH'),
3.41 (t, 3J(H,H) ˆ 8.0 Hz, 1H; CH(CH2Ar), 3.78 (s, 3H; OCH3), 4.43 (d,
3J(H,H) ˆ 5.5 Hz, 1H; CH(CH2Ar)CH), 4.72 (m, 1H; NCH2CH), 6.83
6.87 (m, 2H; 2 of C6H4), 7.09 7.13 (m, 2H, 2 of C6H4); 13C NMR (100 MHz,
CDCl3, DEPT, HMQC): d ˆ 24.0, 26.3 (2q, 2  C(CH3)), 36.0 (t, CH2Ar),
51.8 (t, NCH2), 55.2 (q, OCH3), 66.3 (d, CHCH2Ar), 81.7 (d, NCH2CH),
84.9 (d, CH(Ph)CH)), 110.9 (s, CMe2), 114.0, 129.9 (2d, 2 Â aromatic CH),
138.9, 158.1 (2s, 2  quaternary aromatic C); IR (film): nÄ ˆ 3325 cmÀ1 (br,
N-H stretch), 3028 (aromatic C-H stretch), 2987, 2934, 2836 (aliphatic C-H
stretch), 1612; HRMS (CI): m/z: calcd for C15H22NO3: 264.1600; found:
‡
found: 172.1337 [M ‡H].
1b-Phenyl-2,3-O-isopropylidene-1,4-dideoxy-1,4-iminoerythritol (30d)
i) Normal mode of addition: DBU (0.11 mL, 111 mg, 0.73 mmol, 1.3 equiv)
was added under Ar to a solution of 21 (100 mg, 0.56 mmol) in dry Et2O
(6 mL). The reaction mixture was stirred for 4 h, then the mixture was
filtered under Ar to remove DBU hydrochloride. The filtrate was
concentrated under reduced pressure to yield an oil. This was dissolved
in dry Et2O(6 mL) under Ar, and phenylmagnesium bromide (3.0
m
solution in Et2O, 0.56 mL, 1.68 mmol, 3.0 equiv) was added. After 1.5 h
TLC showed consumption of imine; excess Grignard reagent was quenched
with NH4Cl (aq) then the mixture was diluted with Et2O(30 mL) and
washed with NH4Cl (aq) (10 mL), water (10 mL) followed by NaHCO3 (aq)
(10 mL). The aqueous layers basified and back-extracted with Et2O(2 Â
10 mL) and the combined organic layers dried (Na2SO4) and concentrated
under reduced pressure to yield a yellow oil (198 mg). This crude product
was purified by flash column chromatography on silica gel (Et2O/40-60
petroleum ether 2:3) to yield 30d as a pale yellow oil (19 mg, 15%), as a
single diastereomer.
‡
‡
264.1595 [M ‡H]
1b-Ethyl-2,3-O-Cyclohexylidene-1,4-dideoxy-1,4-iminoerythritol
(31b):
DBU (0.045 mL, 45 mg, 0.30 mmol, 1.3 equiv) was added under Ar to a
solution of 28 (50 mg, 0.23 mmol) in dry Et2O(3 mL). The reaction mixture
was stirred for 3 h, then the mixture was filtered under Ar to remove DBU
hydrochloride. The filtrate was concentrated under reduced pressure to
yield an oil. 1H NMR showed this to be consistent with imine formation
1
ˆ
( H NMR (200 MHz, CDCl3): d ˆ 7.54 (pt, J(H,H) ˆ 2.4 Hz, 1H; N CH)).
ii) Reverse mode of addition: According to method described for 30b above
using dry Et2O(6 mL) for imine formation and phenylmagnesium bromide
(3.0m solution in Et2O, 0.93 mL, 2.80 mmol, 5.0 equiv) over 2 h for adduct
formation to yield a yellow oil (146 mg). This crude product was purified by
flash column chromatography on gel (Et2O/40-60 petroleum ether 2:3) to
yield 30d as a pale yellow oil (86 mg, 70%), as a single diastereomer. This
was confirmed by NOESY NMR experiments to be the anti diastereomer.
1H NMR (400 MHz, CDCl3, COSY, NOESY): d ˆ 1.36, 1.55 (2s, 2  3H;
2  C(CH3)), 2.45 (brs, 1H; NH), 2.95 (dd, 2J(H,H) ˆ 13.4, 3J(H,H) ˆ
This was dissolved in dry Et2O(4 mL) under Ar. To a separate flask under
Ar was added ethylmagnesium bromide (3.0m solution in Et2O, 0.38 mL,
1.15 mmol, 5.0 equiv), and the imine solution was added to the Grignard
reagent dropwise by syringe. After 2 h, TLC and APCI mass spectrometry
showed consumption of imine and formation of the desired adduct; excess
Grignard reagent was quenched with NH4Cl (aq), then the mixture was
diluted with Et2O(30 mL) and washed with NH 4Cl (aq) (10 mL), water
(10 mL) followed by NaHCO3 (aq) (10 mL). The aqueous layers were
basified to pH 11 with 3m NaOH (aq) then back-extracted with Et2O
(10 mL) followed by CHCl3 (2 Â 10 mL) and the combined organic layers
dried (Na2SO4) and concentrated under reduced pressure to yield pure 31b
as a yellow oil (37 mg, 76%). No further purification was required, and
1H NMR showed that the product was a single diastereomer. 1H NMR
(400 MHz, CDCl3, COSY, NOESY): d ˆ 0.97 (t, 3J(H,H) ˆ 7.4 Hz, 3H;
CH2CH3), 1.25 1.68 (m, 10H, cyclohexylidene), 1.32 (m, 2H; CH2CH3),
2.85 (dd, 2J(H,H) ˆ 13.7, 3J(H,H) ˆ 4.2 Hz, 1H; NCHH'), 2.94 (brs, 1H;
2
4.4 Hz, 1H; NCHH'), 3.13 (d, J(H,H) ˆ 13.4 Hz, 1H; NCHH'), 4.38 (brs,
1H; CHPh), 4.71 (m, 1H; NCH2CH), 4.85 (dd, 3J(H,H) ˆ 5.7, 3J(H,H) ˆ
0.8 Hz, 1H; CH(Ph)CH), 7.23 7.44 (m, 5H; C6H5); 13C NMR (100 MHz,
CDCl3, DEPT, HMQC): d ˆ 24.1, 26.4 (2q, 2  C(CH3)), 52.7 (t, NCH2),
67.6 (d, CHPh), 82.2 (d, NCH2CH), 88.1 (d, CH(Ph)CH), 111.1 (s, CMe2),
126.7, 126.9, 128.5 (3d, aromatic CH), 139.5 (s, quaternary aromatic C); IR
(film): nÄ ˆ 3337 cmÀ1 (br, N-H stretch), 3061, 3027 (aromatic C-H stretch),
Chem. Eur. J. 2003, 9, 3397 3414
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3411

B. G. Davis et al.
FULL PAPER
NH), 3.03 (d, 2J(H,H) ˆ 13.7 Hz, 1H; NCHH'), 3.05 (app t, 3J(H,H) ˆ
7.8 Hz, 1H; CHEt), 4.39 (d, 3J(H,H) ˆ 5.5 Hz, 1H; NCH(Et)CH), 4.67
(app t, 3J(H,H) ˆ 7.8 Hz, 1H; NCH2CH); 13C NMR (100 MHz, CDCl3,
DEPT, HMQC): d ˆ 11.4 (q, CH2CH3), 23.6, 24.0, 25.2, 29.7, 33.5, 36.1 (6t,
5 Â CH2 cyclohexylidene, 1 Â CH2CH3)), 51.8 (t, NCH2), 66.7 (d, CHEt),
81.2 (d, NCH2CH), 85.1 (d, CH(Et)CH), 111.5 (s, C(CH2)2); IR (film): nÄ ˆ
3323 cmÀ1 (br, N-H stretch), 2934, 2854 (aliphatic C-H stretch), 1449, 1369;
HRMS (CI): m/z: calcd for C12H22NO2: 212.1651; found: 212.1655
1b-Ethyl-1,4-dideoxy-1,4-iminoerythritol (34b): THF (1 mL) and a solution
of TFA (aq., 25% v/v, 2 mL) was added to 30b (110 mg, 0.64 mmol). The
mixture was stirred vigorously under Ar. After 60 h stirring, TLC (EtOAc)
showed consumption of SM (Rf 0.2). The mixture was concentrated under
vacuum. Water was added and the solvent removed under high vacuum
(3 Â 20 mL), then toluene was added and the solvent evaporated (2 Â
20 mL) to yield an oily solid. This was purified by ion-exchange
‡
chromatography (Dowex, 50X2-200, H form, Acros, 10% v/v solution of
‡
[M ‡H].
™880∫ ammonia) the flash column chromatography on silica gel (CHCl3/
MeOH/™880∫ ammonia solution 30:15:4) to yield pure 34b (81 mg, 96%).
1H NMR (400 MHz, D2O, COSY): d ˆ 0.94 (t, 3J(H,H) ˆ 7.5 Hz, 3H;
CH2CH3), 1.61, 1.79 (2m, 2  1H; CH2CH3), 3.20 (dd, 2J(H,H) ˆ 13.1,
1a-Ethyl-1,4-dideoxy-1,4-imino-threitol (33b): A solution of TFA (aq.,
50% v/v, 2 mL) and THF (2 mL) was added to 29b (53 mg, 0.15 mmol). The
reaction mixture was stirred under N2. After 65 h TLC (EtOAc/hexane 1:1)
showed consumption of SM (Rf 0.3). The reaction was then concentrated
under reduced pressure, water was added to the residue and removed under
reduced pressure (2 Â 20 mL), toluene was added and removed under
reduced pressure (2 Â 20 mL) to yield a yellow oil (38 mg). This crude
3
3
3J(H,H) ˆ 2.2 Hz, 1H; NCHH'), 3.31 (td, J(H,H) ˆ 5.4, J(H,H) ˆ 8.9 Hz,
2
3
1H; NCH(Et)), 3.41 (dd, J(H,H) ˆ 13.1, J(H,H) ˆ 4.5 Hz, 1H; NCHH'),
3.95 (dd, 3J(H,H) ˆ 8.6, J(H,H) ˆ 4.2 Hz, 1H; CH(Et)CH), 4.26 (m, 1H;
3
NCH2CH); 13C NMR (100 MHz, D2O, APT, HMQC): d ˆ 10.4 (q,
CH(CH2CH3)), 23.4 (t, CH(CH2CH3), 49.2 (t, NCH2), 62.6 (s, CH(Et)),
69.7 (d, NCH2CH), 75.0 (d, CH(Et)CH); IR (film): nÄ ˆ 3355 cmÀ1 (br, O-H,
N-H stretches), 2969 (aliphatic C-H stretch); HRMS (CI): m/z: calcd for
‡
product was purified by ion-exchange chromatography (Dowex H form,
50X2-200, Acros; water then 0.1m aqueous ammonia solution) to yield 33b
(15 mg, 76%). 1H NMR (500 MHz, D2O): d ˆ 0.83 (t, 3J(H,H) ˆ 7.5 Hz 3 H ;
CH2CH3), 1.37 (m, 1H; CHH'CH3), 1.56 (m, 1H; CHH'CH3), 2.62 (m, 1H;
‡
C10H14NO2: 132.1025; found: 132.1018 [M ‡H].
1b-Phenyl-3,4-dihydroxy-1,4-dideoxy-1,4-iminoerythritol (34d): THF
(1 mL) and a solution of TFA (aq., 25% v/v solution, 2 mL) was added
to 30d (75 mg, 0.34 mmol). The mixture was vigorously stirred under Ar.
After 48 h TLC (50:50 EtOAc/cyclohexane) showed consumption of SM
(Rf 0.2). The solution was concentrated under vacuum, water was added
and the solvent removed under vacuum (3 Â 20 mL) then toluene was
added and the solvent evaporated (2 Â 20 mL) to yield an oily solid. This
2
3
NCH(Et)), 2.72 (dd, J(H,H) ˆ 12.5, J(H,H) ˆ 3.5 Hz, 1H; CHH'N), 2.94
(dd, 2J(H,H) ˆ 12.5, 3J(H,H) ˆ 5.5 Hz, 1H; CHH'N), 3.57 (dd, 3J(H,H) ˆ
5.5, 3J(H,H) ˆ 3.5 Hz, 1H; NCH2CHOH), 3.98 (m, 1H; NCH(Et)CHOH);
13C NMR (125.7 MHz, D2O, DEPT): d ˆ 10.7 (q, NCH(CH2CH3), 25.9 (t,
NCH(CH2CH3), 50.5 (t, CH2N), 64.1 (d, NCH(Et)), 77.6 (d, NCH2CHOH),
82.1 (d, CH(Et)CHOH); IR (KBr disc): nÄ ˆ 3352 cmÀ1 (br, O-H stretch),
2966, 2932, 2880 (aliphatic C-H stretch); MS (ES): m/z (%): 132 (100)
‡
was purified by ion-exchange chromatography (Dowex, 50X2-200, H
‡
[M ‡H]; HRMS (ES): m/z: calcd for C6H14NO2: 132.1025; found:
form, Acros; 10% v/v solution of ™880∫ ammonia) then flash column
chromatography on silica gel (CHCl3/MeOH/™880∫ ammonia solution
30:15:4) to yield 34d (39 mg, 64%). 1H NMR (400 MHz, D2O, COSY):d ˆ
3.41 (d, 2J(H,H) ˆ 13.3 Hz, 1H; NCHH'), 3.73 (d, 2J(H,H) ˆ 13.3,
3J(H,H) ˆ 4.3 Hz, 1H; NCHH'), 4.48 (brt, 3J(H,H) ˆ 3.6 Hz, 1H;
NCH2CH), 4.55 (d, 3J(H,H) ˆ 10.1 Hz, 1H; NCH(Ph)), 4.62 (dd,
3J(H,H) ˆ 10.1, 3J(H,H) ˆ 3.9 Hz, 1H; NCH(Ph)CH), 7.46 7.50 (m, 5H,
C6H5); 13C NMR (100 MHz, D2O, APT, HMQC): d ˆ 50.4 (t, NCH2), 63.0
(d, CH(Ph)), 69.6 (d, NCH2CH), 76.3 (d, CH(Ph)CH), 128.7, 130.0, 130.6
(3d, 3  aromatic CH), 132.2 (s, quaternary aromatic C); IR (film): nÄ ˆ
3336 cmÀ1 (br, O-H, N-H stretches), 3032 (aromatic C-H stretch), 2933
(aliphatic C-H stretch); HRMS (CI): m/z: calcd for C10H14NO2: 180.1025;
‡
132.1025 [M ‡H].
1a-Phenyl-1,4-dideoxy-1,4-iminothreitol (33d): A 25% solution of TFA in
water (4 mL) and THF (4 mL) to 29d (74 mg, 0.18 mmol). The reaction
mixture was stirred under N2. After 48 h TLC (EtOAc/hexane 3:2) showed
the consumption of SM (Rf 0.5). The reaction was concentrated under
reduced pressure. Water was added to the residue and removed under
reduced pressure (2 Â 20 mL), then toluene was added and removed under
reduced pressure (2 Â 20 mL) to yield a yellow oil resulted (38 mg). This
crude product was purified by flash column chromatography on silica gel
(MeOH/CHCl3 3:17) to yield pure 33d, as its trifluoroacetate salt, a pale
yellow oil (33 mg, 63%). 1H NMR (500 MHz, D2O): d ˆ 3.18 (dd,
2J(H,H) ˆ 13.0, 3J(H,H) ˆ 3.0 Hz 1H; CHH'N), 3.45 (dd, 2J(H,H) ˆ 12.5,
‡
found: 180.1025 [M ‡H].
3
3J(H,H) ˆ 5.5 Hz, 1H; CHH'N), 4.19 (d, J(H,H) ˆ 6.5 Hz, 1H; NCHPh),
1b-(4-Methoxybenzyl)-2,3-dihydroxy-1,4-dideoxy-1,4-iminoerythritol
(34 f): THF (1 mL) and trifluoroacetic acid (TFA) (25% aqueous solution,
2 mL) was added to 30 f (96 mg, 0.36 mmol). The mixture was vigorously
stirred under Ar. After 50 h stirring, TLC (EtOAc) showed consumption of
SM (Rf 0.2). The solution was concentrated under vacuum, water was added
and the solvent removed under vacuum (3 Â 20 mL), then toluene was
added and the solvent evaporated (2 Â 20 mL) to yield an oily solid. This
4.25 (m, 2H; 2 Â CHOH), 7.32 7.37 (m, 5H; C6H5); 13C NMR (125.7 MHz,
D2O, DEPT): d ˆ 50.1 (CH2N), 67.5 (NCH(Ph)), 74.9 (NCH2CHOH), 80.6
(CH(Ph)CHOH), 116.5 (q, CF3COOÀ, 1J(C,F) ˆ 290 Hz), 128.1, 129.4,
129.7, 134.0 (4  aromatic C), 163.2 (q, CF3COOÀ, 2J(C,F) ˆ 35 Hz); IR
(KBr disc): nÄ ˆ 3357 cmÀ1 (br, O-H), 2946 (aliphatic C-H), 1679 (CF3C O);
ˆ
‡
MS (ES): m/z (%): 180 (100) [M ‡H]; HRMS (ES): m/z: calcd for
‡
C10H14NO2: 180.1025; found: 180.1020 [M ‡H].
‡
was purified by ion-exchange chromatography (Dowex, 50X2-200, H
form, Acros; 10% solution of ™880∫ ammonia) then flash column
chromatography on silica gel (CHCl3/MeOH/™880∫ ammonia solution,
30:15:4) to yield 34 f (52 mg, 65%) as a yellow oil. 1H NMR (400 MHz,
D2O, COSY): d ˆ 2.63 (dd, 2J(H,H) ˆ 14.3, 3J(H,H) ˆ 9.3 Hz, 1H;
CHH'C6H4), 2.81 (dd, 2J(H,H) ˆ 12.9, 3J(H,H) ˆ 3.0 Hz, 1H; CHH'N),
2.95 (dd, 2J(H,H) ˆ 14.3, 3J(H,H) ˆ 5.0 Hz, 1H; CHH'C6H4), 3.23 (dd,
2J(H,H) ˆ 12.8, 3J(H,H) ˆ 5.4 Hz, 1H; CHH'N), 3.27 (td, 3J(H,H) ˆ 9.0,
3J(H,H) ˆ 5.0 Hz 1H; NCH(CH2Ar), 3.69 (s, 3H; OCH3), 3.78 (dd,
3J(H,H) ˆ 8.3, 3J(H,H) ˆ 5.0 Hz, 1H; NCHCHOH), 4.13 (td, 3J(H,H) ˆ
5.0, 3J(H,H) ˆ 3.0 Hz, 1H; NCH2CHOH), 6.86 (m, 2H; 2 of C6H4), 7.14
(m, 2H; 2 of C6H4); 13C NMR (100 MHz, D2O, APT, HMQC): d ˆ 36.3 (t,
CH2C6H4), 49.9 (t, NCH2), 55.7 (q, OCH3), 62.5 (d, NCH(CH2Ar)), 70.2 (d,
NCH2CH), 75.4 (d, NCH(CH2Ar)CH), 114.6 (d, aromatic CH), 130.4 (s,
quaternary aromatic C), 130.5 (d, aromatic CH), 158.0 (s, quaternary
aromatic C); IR (film): nÄ ˆ 3328 cmÀ1 (br, O-H, N-H stretches), 3030
(aromatic C-H stretch), 2934 (aliphatic C-H stretch), 1612; HRMS (CI):
1a-(4-Methoxybenzyl)-1,4-dideoxy-1,4-iminothreitol (33 f): A solution of
TFA (aq., 25% v/v, 4 mL) and THF (2 mL) was added to 29 f (77 mg,
0.17 mmol). The reaction mixture was stirred under Ar. After 65 h TLC
(EtOAc) revealed showed consumption of SM (Rf 0.2). The reaction was
concentrated under reduced pressure, water was added to the residue and
removed under reduced pressure (2 Â 20 mL), then toluene was added and
removed under reduced pressure (2 Â 20 mL) to yield a yellow oil (60 mg).
This crude product was purified by ion-exchange chromatography (Dowex,
‡
H
form, 50X2-200, Acros; water then 0.1m aqueous ammonia solution) to
yield 33 f (23 mg, 61%) as a yellow oil. 1H NMR (400 MHz, CD3OD,
COSY): d ˆ 2.74 (dd, 2J(H,H) ˆ 13.8, 3J(H,H) ˆ 8.2 Hz, 1H; CHH'C6H4),
2.88 (dd, 2J(H,H) ˆ 12.0, 3J(H,H) ˆ 2.6 Hz, 1H; CHH'N), 2.99 (dd,
2J(H,H) ˆ 13.8, 3J(H,H) ˆ 6.2 Hz, 1H; CHH'C6H4), 3.07 (dd, 2J(H,H) ˆ
12.0, 3J(H,H) ˆ 5.3 Hz, 1H; CHH'N), 3.08 (m, 1H, CH-pMeOBn), 3.75
3
(dd, J(H,H) ˆ 3.2, 5.2, 1H; NCHCHOH), 3.78 (s, 3H; OCH3), 4.04 (pdt,
3J(H,H) ˆ 2.5, 2.5, 5.2, 1H; NCH2CHOH), 6.88 (m, 2H; 2 of C6H4), 7.19 (m,
2H; 2 of C6H4); 13C NMR (100 MHz, CD3OD, DEPT): d ˆ 39.9 (t,
CH2C6H4), 53.3 (t, NCH2), 56.1 (q, OCH3), 68.7 (d, NCH), 79.4 (d, CHOH),
83.1 (d, CHOH), 115.4, 131.5 (2d, 2 Â aromatic CH), 132.4 (s, quaternary
aromatic C), 160.2 (s, quaternary aromatic C); IR (KBr disc): nÄ ˆ 3273 cmÀ1
(brs, O-H stretch), 2928 (aliphatic C-H stretch), 1613, 1515; HRMS (CI):
‡
m/z: calcd for C12H18NO3: 224.1287; found: 224.1284 [M ‡H].
Non-mammalian glycosidase inhibition assays: p-Nitrophenyl glycosides
were purchased from Sigma-Aldrich Co. Ltd. Enzymes were purchased
from Sigma-Aldrich: a-mannosidase (Canavalia ensiformis, jack beans,
M7257), b-mannosidase (snail acetone powder, M9400), b-glucosidase
(almonds, G0395), a-galactosidase (green coffee beans, G8507), a-rham-
‡
m/z: calcd for C12H18NO3: 224.1287; found: 224.1285 [M ‡H].
3412
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414

Aza-Sugar Scaffolds
3397 3414
nosidase (Penicillium decumbens, naringinase, N1385), a-l-fucosidase
(bovine kidney, F5884). Recordings were made using a Molecular Devices
SPECTRAmax Microplate Spectrophotometer.
[4] R. A. Dwek, T. D. Butters, F. M. Platt, N. Zitzmann, Nature Rev. Drug
Discovery 2002, 1, 65.
[5] S. Inouye, T. Tsurouka, T. Ito, T. Niida, Tetrahedron 1968, 24, 2125.
[6] B. G. Davis, A. Hull, C. Smith, R. J. Nash, A. A. Watson, D. A.
Winkler, R. C. Griffiths, G. W. J. Fleet, Tetrahedron: Asymmetry 1998,
9, 2947.
[7] a) U. Fuhrmann, E. Bause, G. Legler, H. Ploegh, Nature 1984, 307, 755;
b) J. Bischoff, R. Kornfeld, Biochem. Biophys. Res. Commun. 1984,
125, 324.
[8] B. Winchester, S. Al Daher, N. C. Carpenter, I. Cenci di Bello, S. S.
Choi, A. J. Fairbanks, G. W. J. Fleet, Biochem. J. 1993, 290, 743.
[9] a) D. S. Tan, Nat. Biotechnol. 2002, 20, 561; b) K. M. Specht, K. M.
Shokat, Curr. Opin. Cell Biol. 2002, 14, 155.
[10] a) D. L. Ollis, E. Cheah, M. Cygler, B. Dijkstra, F. Frolow, S. M.
Franken, M. Harel, S. J. Remington, I. Silman, J. Schrag, J. L. Sussman,
K. H. G. Vershueren, A. Goldman, Protein Eng. 1992, 5, 197; b) M. S.
Kimber, E. F. Pai, EMBO J. 2000, 19, 1407; c) A. Mattevi, M. A.
Vanoni, F. Todone, M. Rizzi, A. Teplyakov, A. Coda, M. Bolognesi, B.
Curti, Proc. Natl. Acad. Sci. USA 1996, 93, 7496; d) M. C. Fitzgerald, I.
Chernushevich, K. G. Standing, S. B. H. Kent, C. P. Whitman, J. Am.
Chem. Soc. 1995, 117, 11075.
3.5 mm stock solution of p-nitrophenyl a-l-rhamnopyranoside in 0.1m
pH 7.0 orthophosphate buffer and 5.0 mm stock solutions of all other p-
nitrophenyl glycosides in 0.1m pH 7.0 orthophosphate buffer were prepared
and diluted to 3.0, 2.5, 2.0, 1.5, 1.0, 0.5, 0.25 mm and 4.0, 3.0, 2.0, 1.5, 1.0, 0.5,
0.25 mm, respectively. Enzyme solutions were prepared in 0.1m pH 7.0
orthophosphate buffer to a stock concentration of 2.0 UmLÀ1 and diluted
to give a final assay concentration of 0.1 UmLÀ1. Stock inhibitor solutions
were prepared in deionised water to a concentration of 40 mm and diluted
to give a final assay concentration range of 0.1, 0.05, 0.01, 0.005, 0.001 mm.
Substrate solutions (185 mL of each) were pipetted into a multiwell plate
and incubated for ten minutes at 378C prior to use. Enzyme solution
(10 mL) was incubated for 5 minutes at 378C with appropriate inhibitor
solution (5 mL) prior to use and pipetted into substrate solution. Release of
p-nitrophenol from substrate was recorded at 405 nm for 5 minutes, reading
every 10 s with agitation between each reading. Inhibition data was
compiled and analysed using Excel2000 (Microsoft), and Grafit4.0
(Erithacus Software Ltd.), using Dixon plot analysis.
Human lysosomal glycosidase inhibition assays: Enzymes were extracted
from an MCF7 cell-line harvest: cells were harvested from ꢀ100 mL of
standard culture, washed in phosphate buffered saline solution (PBS) and
sonicated (3 Â 10 s) in water (1 mL). Extract (5 mL) and inhibitor solution
(0.04, 0.4 or 4 mm diluted using water from 100mm stock in DMSO, 5 mL)
were diluted with the appropriate enzyme assay solution (see below, 10 mL)
and incubated for the appropriate length of time (see below). The course of
the assay was stopped by addition of glycine-carbonate buffer solution
(0.17m, pH 9.8, 150 mL) and absorbance (405 nm) or fluorescence (excita-
tion 460 nm, emission 355 nm) recorded as appropriate. Assay solutions
and incubation times: a-d-glucosidase [1.25mm p-nitrophenyl a-d-gluco-
pyranoside in 0.2m citrate/phosphate buffer, pH 4.4, 378C, 16 h]; b-d-
glucosidase [5mm 4-methylumbelliferyl b-d-glucopyranoside in 0.2m
citrate/phosphate buffer, pH 5.8, 378C, 3 h]; a-d-galactosidase [20mm p-
nitrophenyl a-d-galactopyranoside, 180mm N-acetyl-d-glucosamine in
0.2mm citrate/phosphate buffer, pH 4.4, 378C, 4 h]; b-d-galactosidase
[5mm p-nitrophenyl b-d-galactopyranoside in 0.2m citrate/phosphate
buffer, pH 4.3, 378C, 2 h]. IC50 values were determined from Â2 serial
dilutions as appropriate of inhibitor concentrations from 40mm. All assays
were recorded in duplicate.
[11] B. Davis, A. A. Bell, R. J. Nash, A. A. Watson, R. C. Griffiths, M. G.
Jones, C. Smith, G. W. J. Fleet, Tetrahedron Lett. 1996, 37, 8565.
[12] a) G. W. J. Fleet, S. K. Namgoong, C. Barker, S. Baines, G. S. Jacob, B.
Winchester, Tetrahedron Lett. 1989, 30, 4439; b) F. M. Platt, G. R.
Neises, G. B. Karlsson, R. A. Dwek, T. D. Butters, J. Biol. Chem. 1994,
¬
269, 27108; c) Y. Bleriot, C. R. Veighey, K. H. Smelt, J. Cadefau, W.
Stalmans, K. Biggadike, A. L. Lane, M. M¸ller, D. J. Watkin, G. W. J.
Fleet, Tetrahedron: Asymmetry 1996, 7, 2761; d) T. D. Butters,
L. A. G. M. van den Broek, G. W. J. Fleet, T. M. Krulle, M. R. Wor-
mald, R. A. Dwek, F. M. Platt, Tetrahedron: Asymmetry 2000, 11, 113.
[13] a) R. Bloch, Chem. Rev. 1998, 98, 1407; b) S. Kobayashi, H. Ishitani,
Chem. Rev. 1999, 99, 1069; c) R. A. Volkmann in Comprehensive
Organic Synthesis: Additions to C-X p-Bonds, Part 1, Vol. 1, (Ed.: S. L.
Schreiber), Pergamon, Oxford, 1991, p. 355.
[14] B. A. Horenstein, R. F. Zabinski, V. L. Schramm, Tetrahedron Lett.
1993, 34, 7213.
[15] a) G. B. Evans, R. H. Furneaux, T. L. Hutchison, H. S. Kezar, P. E.
Morris, V. L. Schramm, P. C. Tyler, J. Org. Chem. 2001, 66, 5723;
b) G. B. Evans, R. H. Furneaux, G. J. Gainsford, V. L. Schramm, P. C.
Tyler, Tetrahedron 2000, 56, 3053; c) R. H. Furneaux, V. L. Schramm,
P. C. Tyler, Bioorg. Med. Chem. 1999, 7, 2599; d) R. H. Furneaux, G.
Limberg, P. C. Tyler, V. L. Schramm, Tetrahedron 1997, 53, 2915.
[16] H. Bˆhme, H. G. Viehe in Advances in Organic Chemistry; Methods
and Results, Vol. 9 (Eds.: H. Bˆhme, H. G. Viehe), Wiley, New York,
1976, p. 183.
Non-lysosomal b-glucosidase inhibition assay: Conducted as for human
lysosomal b-glucosidase inhibition assays except prior to assay, MCF7
extract (450 mL) was incubated for 30 min with the irreversible lysosomal b-
glucosidase inhibitor conduritol b-epoxide (25mm diluted from a 250mm
stock in DMSO, 4.5 mL).
Glucosylceramide synthase (GCS) assay: GCS [UDP-glucose N-acyl-
sphingosine glucosyltransferase (EC 2.4.1.80)] assay was conducted using
HL-60 cell microsomes as described in ref. [60].
[17] a) J. Hamer, A. Macaluso, Chem. Rev. 1964, 64, 473; b) G. Tennant in
Comprehensive Organic Chemistry, Vol. 1 (Eds.: D. H. R. Barton,
W. D. Ollis), Pergamon, Oxford, 1979, p. 504; c) R. Giovannini, E.
Marcantoni, M. Petrini, J. Org. Chem. 1995, 60, 5706; d) F. Schieweck,
H.-J. Altenbach, J. Chem. Soc. Perkin Trans. 1 2001, 3409; e) M.
Lombardo, S. Fabbroni, C. Trombini, J. Org. Chem. 2001, 66, 1264.
[18] a) H. Hellmann, Angew. Chem. 1957, 69, 463; b) H. E. Zaugg,
Synthesis 1970, 49; c) H. E. Zaugg, Synthesis 1984, 85; d) W. N.
Speckamp, H. Hiemstra, Tetrahedron 1985, 41, 4367.
[19] a) P. Perlmutter, C. C. Teo, Tetrahedron Lett. 1984, 25, 5951; b) F. A.
Davis, J. Wei, A. C. Sheppard, S. Gubernick, Tetrahedron Lett. 1987,
28, 5115; c) J. Sisko, S. M. Weinreb, J. Org. Chem. 1990, 55, 393; d) J.-
C. Adelbrecht, D. Craig, B. W. Dymock, S. Thorimbert, Synlett 2000, 4,
467; e) J.-C. Adelbrecht, D. Craig, S. Thorimbert, Tetrahedron Lett.
2001, 42, 8369.
Acknowledgements
This work was supported by the UK Engineering and Physical Sciences
Research Council (EPSRC) and by Oxford Glycosciences Ltd. We thank
the EPSRC for access to the Mass Spectrometry Service at Swansea & the
Chemical Database Service at Daresbury, Dr. D. Scopes for helpful advice
and Mr. M. Robinson for technical assistance.
[20] a) R. Ballini, E. Marcantoni, M. Petrini, J. Org. Chem. 1992, 57, 1316;
b) R. Giovannini, E. Marcantoni, M. Petrini, J. Org. Chem. 1995, 60,
5706; c) M. Lombardo, S. Fabbroni, C. Trombini, J. Org. Chem. 2001,
66, 1264; d) P. J. Dransfield, P. M. Gore, M. Shipman, A. M. Z. Slawin,
Chem. Commun. 2002, 150.
[1] a) B. Winchester, G. W. J. Fleet, Glycobiology 1992, 2, 199; b) D.
O×Hagan, Nat. Prod. Rep. 1997, 14, 637; c) A. E. St¸tz, Iminosugars as
Glycosidase Inhibitors, Wiley, Weinheim, 1999; d) N. Asano, R. J.
Nash, R. J. Molyneaux, G. W. J. Fleet, Tetrahedron: Asymmetry 2000,
11, 1645.
[21] Imines 3 and 4 contain the same core substructure as ribitol imine 1
(see ref. [14]), which prior to this work had been the substrate for one
of the highest yielding examples of addition of an organometallic
nucleophile to a cyclic imine reported in the literature. We therefore
wondered if the increased rigidity of this bicyclic [3.3.0] geometry
could be an important factor.
[2] a) M. L. Sinnott, Chem. Rev. 1990, 90, 1171; b) G. Legler, Adv.
Carbohydr. Chem. Biochem. 1990, 48, 319; c) A. E. St¸tz, Angew.
Chem. 1996, 108, 2054; Angew. Chem. Int. Ed. Engl. 1996, 35, 1926;
d) T. D. Heightman, A. T. Vasella, Angew. Chem. 1999, 111, 794;
Angew. Chem. Int. Ed. Engl. 1999, 38, 750.
[3] T. D. Butters, R. A. Dwek, F. M. Platt, Chem. Rev. 2000, 100, 4683.
Chem. Eur. J. 2003, 9, 3397 3414
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3413

B. G. Davis et al.
FULL PAPER
[22] B. G. Davis, M. A. T. Maughan, T. M. Chapman, R. Villard, S.
Courtney, Org. Lett. 2002, 4, 103.
[23] a) A. Jiminez, D. Vasquez in Antibiotics (Ed.: F. E. Hahn), Springer,
Berlin, 1979, pp. 1 19; b) A. P. Grollman, J. Biol. Chem. 1967, 242,
3226.
[43] C.-H. Wong, L. Provencher, J. A. Porco, S.-H. Jung, Yi.-F. Wang, L.
Chen, R. Wang, D. H. Steensma, J. Org. Chem. 1995, 60, 1492.
[44] A. J. Fairbanks, N. Carpenter, G. W. J. Fleet, N. G. Ramsden, I.
Cenci de Bello, B. G. Winchester, S. S. Al-Daher, G. Nagahashi,
Tetrahedron 1992, 48, 3365.
[24] Y. Ryu, G. Kim, J. Org. Chem. 1995, 60, 103.
[25] M. Corturier, J. L. Tucker, B. M. Andresen, P. Dube, J. T. Negri, Org.
Lett. 2001, 3, 465.
[26] M. Kiso, A. Hasegawa, Carbohydr. Res. 1976, 52, 95.
[27] D. K. Thompson, C. N. Hubert, R. H. Wightman, Tetrahedron 1993,
49, 3827.
[28] A. E. McCaig, K. P. Meldrum, R. H. Wightman, Tetrahedron 1998, 54,
9429.
[29] S. H. Rosenberg, K. W. Woods, H. L. Sham, H. D. Kleinert, D. L.
Martin, H. Stein, J. Cohen, D. A. Egan, B. Bopp, I. Merits, K. W.
Garren, D. J. Hoffman, J. J. Plattner, J. Med. Chem. 1990, 33, 1962.
[30] M. A. Brook, Jahangir, Synth. Commun. 1988, 18, 893.
[31] Y. Arakawa, S. Yoshifuji, Chem. Pharm. Bull. 1991, 39, 2219.
[32] H. S. Mosher, E. L. Combe, J. Am. Chem. Soc. 1950, 72, 3994.
[33] See for example: a) M. Okubo, Bull. Chem. Soc. Jpn. 1985, 58, 3108;
b) K. Maruyama, T. Katagiri, Chem. Lett. 1987, 735; c) K. Maruyama,
T. Katagiri, J. Am. Chem. Soc. 1986, 108, 6263; and references therein.
[34] R. Bianchini, C. Forte, G. Musumarra, C. Pinzino, C. Sergi, Tetrahe-
dron 1997, 53, 6907.
[35] a) G. Alvaro, G. Martelli, D. Savoia, J. Chem. Soc. Perkin Trans. 1
1998, 775; b) C. Berti, L. Greci, L. Marchetti, J. Chem. Soc. Perkin
Trans. 2 1977, 1032; c) L. Eberson, L. Greci, J. Org. Chem. 1984, 49,
2135.
[36] J. F. Garst, F. E. Barton, J. Am. Chem. Soc. 1974, 96, 523.
[37] D. Lal, D. Griller, S. Husband, K. U. Ingold, J. Am. Chem. Soc. 1974,
96, 6355.
[38] H. Thies, H. Schoenenberger, Chem. Ber. 1956, 89, 1918.
[39] W. Schlenk, W. Schlenk Jr., Ber. 1929, 62B, 920.
[40] For useful examples of an acetonide-protected [3.3.0] dihydroxypyr-
rolidine 3D structures containing an sp2 centre see Cambridge
Structural Database Refcode WAQNUX, WAQPAF: M. Godskesen,
I. Lundt, U. Sotofte, Tetrahedron: Asymmetry 2000, 11, 567.
[41] See for example: a) D. P. Schumacher, S. S. Hall, J. Am. Chem. Soc.
1982, 104, 6076; b) J. G. Buchanan, K. A. MacLean, H. Paulsen, R. H.
Wightman, J. Chem. Soc. Perkin Trans. 1 1985, 1463; c) H. Iida, N.
Yamazaki, C. Kibayashi, J. Org. Chem. 1986, 51, 1069; d) O. Schwardt,
U. Veith, C. Gaspard, V. J‰ger, Synthesis 1999, 1473.
[45] a) V. M.-A. Ducros, D. L. Zechel, G. N. Murshudov, H. J. Gilbert, L.
Szabo, D. Stoll, S. G. Withers, G. J. Davies, Angew. Chem. 2002, 114,
2948; Angew. Chem. Int. Ed. 2002, 41, 2824; b) V. Notenboom, S. J.
Williams, R. Hoos, S. G. Withers, D. R. Rose, Biochemistry 2000, 39,
11553.
[46] S. J. Williams, V. Notenboom, J. Wicki, D. R. Rose, S. G. Withers, J.
Am. Chem. Soc. 2000, 122, 4229.
[47] H. Huang, C.-H. Wong, J. Org. Chem. 1995, 60, 3100.
[48] G. W. J. Fleet, P. W. Smith, S. V. Evans, L. E. Fellows, J. Chem. Soc.
Chem. Commun. 1984, 1240.
[49] B. Ganem, Acc. Chem. Res. 1996, 29, 340.
[50] D. P. Dumas, T. Kajimoto, K. K.-C. Liu, C.-H. Wong, D. B. Berkowitz,
S. J. Danishefsky, Bioorg. Med. Chem. Lett. 1992, 2, 33.
[51] F. M. Platt, G. R. Neises, R. A. Dwek, T. D. Butters, J. Biol. Chem.
1994, 269, 8362.
[52] T. D. Butters, L. A. G. M. van den Broek, G. W. J. Fleet, T. M. Kr¸lle,
M. R. Wormald, R. A. Dwek, F. M. Platt, Tetrahedron: Asymmetry
2000, 11, 1113.
[53] U. Andersson, T. D. Butters, R. A. Dwek, F. M. Platt, Biochem.
Pharmacol. 2000, 59, 821.
[54] F. M. Platt, G. R. Neises, G. B. Karlsson, R. A. Dwek, T. D. Butters, J.
Biol. Chem. 1994, 269, 27108.
[55] a) S. H. Rosenberg, K. W. Woods, H. L. Sham, H. D. Kleinert, D. L.
Martin, H. Stein, J. Cohen, D. A. Egan, B. Bopp, I. Merits, K. W.
Garren, D. J. Hoffmann, J. J. Plattner, J. Med. Chem. 1990, 33, 1962;
b) X. Wu, S. Toppet, F. Campernolle, G. J. Hoorneart, Tetrahedron
2000, 56, 6279.
[56] Q. Li, W. Wang, K. B. Berst, A. Claiborne, L. Hasvold, K. Raye, M.
Tufano, A. Nilius, L. L. Shen, R. Flamm, J. Alder, K. Marsh, D.
Crowell, D. T. W. Chu, J. J. Plattner, Bioorg. Med. Chem. Lett. 1998, 8,
1953.
[57] a) Y. Arakawa, S. Yoshifuji, Chem. Pharm. Bull. 1991, 39, 2219; b) A.
Goti, F. Cardona, A. Brandi, S. Picasso, P. Vogel, Tetrahedron:
Asymmetry 1996, 7, 1659.
[58] a) H. Yoda, H. Kitayama, W. Yamada, K. Katagiri, K. Takabe,
Tetrahedron: Asymmetry 1993, 4, 1451; b) Y. Ryu, G. Kim, J. Org.
Chem. 1995, 60, 103.
[42] From tartaric acid a) 14%: P. Hutin, M. Haddad, M. Larcheveque,
Tetrahedron: Asymmetry 2000, 11, 2547; b) 12%: R. Ballini, E.
Marcantoni, M. Petrini, J. Org. Chem. 1992, 57, 1316; c) 5%: C. M.
Wong, J. A. Buccini, I. N. Chang, R. Te, J. Arend, R. J. Schwenk,
Can. J. Chem. 1969, 47, 2421.
[59] L. M. Lerner, Carbohydr. Res. 1969, 9, 1.
[60] F. M. Platt, G. R. Neises, G. B. Karlsson, R. A. Dwek, T. D. Butters, J.
Biol. Chem. 1994, 269, 8362.
Received: January 8, 2003 [F4718]
3414
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3397 3414