Dihydroxyacetone phosphate aldolase catalyzed synthesis of structurally diverse polyhydroxylated pyrrolidine derivatives and evaluation of their glycosidase inhibitory properties

Article abstract of DOI:10.1002/chem.200900838

The chemoenzymatic synthesis of a collection of pyrrolidine-type iminosugars generated by the aldol addition of dihydroxyacetone phosphate (DHAP) to C-α-substituted N-Cbz-2-aminoaldehydes derivatives, catalyzed by DHAP aldolases is reported. L-Fuculose-1-phosphate aldolase (FucA) and L-rhamnulose-1-phosphate aldolase (RhuA) from E. coli were used as biocatalysts to generate configurational diversity on the iminosugars. Alkyl linear substitutions at C-α were well tolerated by FucA catalyst (i.e., 40-70% conversions to aldol adduct), whereas no product was observed with C-α-alkyl branched substitutions, except for dimethyl and benzyl substitutions (20%). RhuA was the most versatile biocatalyst: C-α-alkyl linear groups gave the highest conversions to aldol adducts (60-99%), while the C-α-alkyl branched ones gave moderate to good conversions (50-80%), with the exception of dimethyl and benzyl substituents (20%). FucA was the most stereoselective biocatalyst (90-100% anti (3R,4R) adduct). RhuA was highly stereoselective with (S)-N-Cbz-2-aminoaldehydes (90-100% syn (i.e., 3R,4S) adduct), whereas those with R configuration gave mixtures of antilsyn adducts. For iPr and iBu substituents, RhuA furnished the anti adduct (i.e., FucA stereochemistry) with high stereoselectivity. Molecular models of aldol products with iPr and iBu sub-stituents and as complexes with the RhuA active site suggest that the and adducts could be kinetically preferred, while the syn adducts would be the equilibrium products. The polyhydroxylated pyrrolidines generated were tested as inhibitors against seven glycosidases. Among them, good inhibitors of a-L-fucosidase (IC₅₀ = 1-20 μM), moderate of α-L-rhamnosidase (IC₅₀=7-150 μM), and weak of α-D-mannosidase (IC₅₀ = 80-400 μM) were identified. The apparent inhibition constant values (K_i) were calculated for the most relevant inhibitors and computational docking studies were performed to understand both their binding capacity and the mode of interaction with the glycosidases.

Full text of DOI:10.1002/chem.200900838

DOI: 10.1002/chem.200900838
Dihydroxyacetone Phosphate Aldolase Catalyzed Synthesis of Structurally
Diverse Polyhydroxylated Pyrrolidine Derivatives and Evaluation of their
Glycosidase Inhibitory Properties
Jordi Calveras,[a] Meritxell Egido-Gabꢀs,[b] Livia Gꢁmez,[a] Josefina Casas,[b]
Teodor Parella,[c] Jesffls Joglar,[a] Jordi Bujons,*[a] and Pere ClapØs*[a]
7310
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

FULL PAPER
Abstract: The chemoenzymatic synthe-
sis of a collection of pyrrolidine-type
iminosugars generated by the aldol ad-
dition of dihydroxyacetone phosphate
(DHAP) to C-a-substituted N-Cbz-2-
aminoaldehydes derivatives, catalyzed
by DHAP aldolases is reported. l-Fu-
culose-1-phosphate aldolase (FucA)
and l-rhamnulose-1-phosphate aldolase
(RhuA) from E. coli were used as bio-
catalysts to generate configurational di-
versity on the iminosugars. Alkyl linear
substitutions at C-a were well tolerated
by FucA catalyst (i.e., 40–70% conver-
sions to aldol adduct), whereas no
product was observed with C-a-alkyl
branched substitutions, except for di-
methyl and benzyl substitutions (20%).
RhuA was the most versatile biocata-
lyst: C-a-alkyl linear groups gave the
highest conversions to aldol adducts
(60–99%),
while
the
C-a-alkyl
stituents and as complexes with the
RhuA active site suggest that the anti
adducts could be kinetically preferred,
while the syn adducts would be the
equilibrium products. The polyhydrox-
ylated pyrrolidines generated were
tested as inhibitors against seven glyco-
sidases. Among them, good inhibitors
of a-l-fucosidase (IC50 =1–20 mm),
moderate of a-l-rhamnosidase (IC50 =
7–150 mm), and weak of a-d-mannosi-
dase (IC50 =80–400 mm) were identified.
The apparent inhibition constant
values (Ki) were calculated for the
most relevant inhibitors and computa-
tional docking studies were performed
to understand both their binding ca-
pacity and the mode of interaction with
the glycosidases.
branched ones gave moderate to good
conversions (50–80%), with the excep-
tion of dimethyl and benzyl substitu-
ents (20%). FucA was the most stereo-
selective biocatalyst (90–100% anti
(3R,4R) adduct). RhuA was highly ste-
reoselective with (S)-N-Cbz-2-aminoal-
dehydes (90–100% syn (i.e., 3R,4S)
adduct), whereas those with R configu-
ration gave mixtures of anti/syn ad-
ducts. For iPr and iBu substituents,
RhuA furnished the anti adduct (i.e.,
FucA stereochemistry) with high ste-
reoselectivity. Molecular models of
aldol products with iPr and iBu sub-
Keywords: aldol reaction · amino
aldehydes · cyclitols · enzyme catal-
ysis · glycosidase inhibitors
Introduction
sition-state mimics they constitute valuable probes in funda-
mental biochemical studies of glycosidase mechanism.[12,13]
Many efforts have been devoted to the synthesis of both
natural and analogues of iminocyclitols with a wide variety
of structures and substituents.[11,14–19] Importantly, generation
of configurational diversity on the stereogenic centers ap-
pears to be of paramount importance to optimize their ac-
tivity and selectivity.[20–23] Consequently, novel methodolo-
gies for the preparation of these compounds with a broad
structural and configurational diversity are of increasing in-
terest.
We have recently described a novel chemoenzymatic
strategy to synthesize iminocyclitols of the piperidine and
pyrrolidine type.[24,25] The key step of the strategy devised
was the aldol addition of dihydroxyacetone phosphate
(DHAP) to N-benzyloxycarbonylaminoaldehydes catalyzed
by DHAP aldolases, such as d-fructose-1,6-diphosphate al-
dolase from rabbit muscle (RAMA), and l-rhamnulose-1-
phosphate aldolase (RhuA) and l-fuculose-1-phosphate al-
dolase (FucA), both from E. coli. This methodology allowed
us to prepare a number of iminocyclitols from N-Cbz-3-ami-
nopropanal, N-Cbz-glycinal, and both enantiomers of N-
Cbz-alaninal.
Iminocyclitols are naturally occurring polyhydroxylated al-
kaloid mimics of glycosides.[1] Many of them are potent in-
hibitors of glycosidases and glycosyltransferases.[2] Glycopro-
cessing enzymes are interesting therapeutic targets as they
are responsible for the metabolism of complex carbohydrate
structures involved in many biochemical recognition pro-
cesses. Glycosilation disorders affect to cell–cell communica-
tion, cell–matrix interaction, and immunological response;
hence, they are related to diseases such as cancer, viral in-
fections including HIV and hepatitis B and C, Gaucher and
Fabry diseases, cystic fibrosis, carbohydrate deficiency syn-
drome (CDS), rheumatoid arthritis (chronic polyarthrithis),
and IgA nephropathy (IgAN) among others.[3–6] Because of
this, iminocyclitols have attracted much interest due to their
potential therapeutic applications.[7–11] Furthermore, as tran-
[a] Dr. J. Calveras, L. Gꢁmez, Dr. J. Joglar, Dr. J. Bujons, Dr. P. Clapꢂs
Biotransformation and Bioactive Molecules group
Catalonia Institute for Advanced Chemistry-CSIC
Jordi Girona 18-26, 08034 Barcelona (Spain)
Fax : (+34)932-045-904
Therefore, it was regarded as both promising and signifi-
cant to investigate the synthetic possibilities of the afore-
mentioned strategy (Scheme 1) towards the preparation of
novel pyrrolidine derivatives with potential new inhibitory
properties against different glycosidases. Hence, in this work
we endeavored first to gain insight into both tolerance and
stereoselectivity of DHAP-dependent aldolases towards a
structural variety of N-Cbz-aminoaldehydes, for a better un-
derstanding of their substrate specificity, and second to per-
E-mail: pere.clapes@iqac.csic.es
[b] Dr. M. Egido-Gabꢃs, Dr. J. Casas
Research Unit for Bioactive Molecules (RUBAM)
Catalonia Institute for Advanced Chemistry-CSIC
Jordi Girona 18-26, 08034 Barcelona (Spain)
[c] Dr. T. Parella
Servei de Ressonꢄncia Magnꢅtica Nuclear
Universitat Autꢆnoma de Barcelona, Bellaterra (Spain)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900838.
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7311

J. Bujons, P. Clapꢂs et al.
wards the N-Cbz-aminoaldehydes was investigated as a
function of the reaction media, that is, high water content
emulsions or 1:4 dimethylformamide/water mixtures.[25] En-
zymatic aldol additions were running up to an apparent con-
stant product concentration. When using RAMA as catalyst,
no aldol adduct was detected with any of the selected N-
Cbz-aminoaldehydes in any of the reaction systems assayed.
This behavior was somehow predictable, considering the low
conversion observed in a previous work during RAMA cata-
lyzed aldolization of (S)- and (R)-N-Cbz-alaninal[25] and the
fact that Lys107 and Arg109, located at the RAMA active
site to fix its natural aldehyde acceptor, glyceraldehyde-3-
phosphate, may hamper an effective interaction with hydro-
phobic C-a substituted aldehydes.
Scheme 1. Chemoenzymatic synthesis of structural diverse polyhydroxyl-
ated pyrrolidines: a) benzyloxycarbonylsuccinimide (Cbz-OSu) in diox-
ane, b) 2-iodoxybenzoic acid (IBX), AcOEt, reflux, c) DHAP, DHAP-de-
pendent aldolase, d) acid phosphatase, e) Pd/C, H2 (50 psi).
N-Cbz-Aminoaldehydes with linear alkyl C-a substituents
(2a–c) were substrates for the FucA catalyst as well as those
with C-a benzyl (2g) and dimethyl (2h) substituents
(Figure 1). No product was detected in aldol additions with
form a preliminary evaluation of the inhibitory properties of
the products synthesized against seven glycosidases of differ-
ent types and sources.
Results and Discussion
The iminocyclitols designed consist of polyhydroxylated pyr-
rolidine derivatives with different stereochemical configura-
tions and different substituents at position C5 (Scheme 1) to
explore their influence on their glycosidase inhibitory activi-
ty. Based on bibliographic data,[26–34] improvement of poten-
cy and selectivity against glycosidases and glycosyltransfer-
ases was observed when substituents close to the nitrogen
atom were preferentially hydrophobic, aliphatic, or aromat-
ic. The configuration of the stereogenic centers is of utmost
importance for iminocyclitol inhibitors of glycoprocessing
enzymes. Hence, stereocomplementary DHAP aldolases and
the availability of pure N-Cbz-aminoaldehydes enantiomers
may formally ensure a good configurational diversity. Thus,
the selected aminoaldehydes (2) and the chemoenzymatic
strategy are depicted in Scheme 1. d-Fructose-1,6-bisphos-
phate aldolase from rabbit muscle (RAMA) and l-rhamnu-
lose-1-phosphate aldolase (RhuA) and l-fuculose-1-phos-
phate aldolase (FucA) from E. coli were the catalysts select-
ed for the aldol addition of dihydroxyacetone phosphate
(DHAP) to the N-Cbz-aminoaldehydes 2. Treatment of
aldol adducts 3 with acid phosphatase, followed by the re-
moval of the Cbz and the intramolecular reductive amina-
tion with H2 in the presence of Pd/C in a one-pot reaction
furnished the pyrrolidine type iminocyclitols 4. The configu-
ration at positions C-3 and C-4 was controlled by the
DHAP aldolase. The stereogenic center at C-2 was generat-
ed during the reductive amination and depends, in most
cases, on the stereochemistry at C-4,[24,25] whereas the stereo-
chemistry at C-5 is fixed by the starting aldehyde.
Figure 1. Molar percent conversions to products 3 from the aldol addi-
tions of DHAP to N-Cbz-aminoaldehydes ([DHAP]=50 mm; [N-Cbz-
aminoaldehydes]=85 mm, the results at higher concentrations are sum-
marized in Supporting Information) catalyzed by FucA in high water
content emulsions (black bars) and in 1:4 DMF/water (grey bars). Con-
versions with respect to the starting DHAP concentration determined by
HPLC of the crude reaction mixture using purified standards. Results in
emulsions are the mean values (max. standard deviation <5%) obtained
in the three systems, H2O/C14E4/tetradecane, H2O/C14E4/hexadecane and
H2O/C14E4/squalane 90/4/6 wt% (C14E4: non-ionic polyoxyethylene ether
surfactant, tetra(ethylene glycol)tetradecyl ether, C14H29ACHTNUTRGNE(UGN OCH2CH2)4OH,
with an average of 4 mol of ethylene oxide per surfactant molecule).
the N-Cbz-aminoaldehydes with branched alkyl C-a sub-
stituents, which appeared not to be tolerated by the FucA
catalyst. In some instances, very small peaks appeared in the
HPLC analysis (<2% conversion) after long incubation
times that were not identified, but might correspond to
some aldol adduct formation.
Investigations of the reaction media: There is a remarkable
effect of both the reaction media and aldehyde stereochem-
istry on the reaction conversion. Overall, (R)-N-Cbz-amino-
aldehydes gave higher conversions than those with S config-
uration. Moreover, the best conversions with (S)- and (R)-
N-Cbz-aminoaldehydes were achieved in 1:4 DMF/water
and emulsions, respectively. It has been observed that organ-
ic solvent molecules, for example, DMF, cluster in the active
Enzymatic aldol additions of DHAP to N-Cbz-aminoalde-
hydes: The tolerance of the selected DHAP aldolases to-
7312
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
site of enzymes like elastase modifying the substrate–
enzyme interactions in different ways.[35,36] If this may be ex-
tended to aldolases as well, DMF molecules at the FucA
active site may facilitate the interaction of sterically unfavor-
able (S)-aldehydes.
Interestingly, there were temperature effects on the stereo-
selectivity of the enzymatic reactions (vide infra), but no im-
provement on aldol conversions was observed for any of the
selected examples, indicating that this phenomenon is not of
general applicability.
RhuA was the most versatile aldolase accepting both
linear and branched C-a substituted N-Cbz-aminoaldehydes
(Figure 2).
Iminocyclitol synthesis: Aldol additions with selected ami-
noaldehydes catalyzed by RhuA and FucA were scaled up
under the reactions conditions that provided the best con-
versions to N-Cbz-aminoketotriols 3. The corresponding
aldol adducts were purified from the excess of aldehyde and
other byproducts. Special attention was paid to avoid the
elimination of possible diastereoisomers formed during the
enzymatic catalysis. Then, the isolated adducts were submit-
ted to reductive amination following the methodology de-
scribed by us.[24,25] The resulting iminocyclitols were fully
characterized by NMR spectroscopy. The results, depicted in
Table 1, show that epimeric mixtures at C-2 and C-4 were
formed in some cases, indicating partial stereoselectivity in
either the reductive amination or in the enzymatic aldol ad-
dition, respectively. Minor N-ethyliminosugars were also de-
tected that were formed during the treatment with Pd/C as
will be discussed below. Separation and purification of sev-
eral of these mixtures was accomplished successfully by
weak cation exchange chromatography on a CM-sepharose
column in ammonium form. This allowed to set up an effi-
cient purification procedure and to characterize the struc-
ture of the compounds in the mixtures unequivocally.
Stereochemistry of RhuA and FucA aldolases towards the
N-Cbz-aminoaldehydes: The stereochemistry at C-4 of the
iminocyclitols prepared accounted for the stereoselectivity
of the aldolase–DHAP complex attack (i.e., nucleophile) on
the carbonyl of the N-Cbz-aminoaldehydes.[24,25] This also
may be an indication of the ability of the aldolase to fix the
aldehyde in the right position during the donor attack in the
active site. Remarkably, FucA catalyst gave excellent stereo-
selectivities, particularly towards the (R)-aldehydes
(Figure 3), in good agreement with the previously reported
FucA reactions with (S)- and (R)-N-Cbz-alaninal.[24] The
pertinent conclusion was that the stereochemical outcome
of FucA-catalyzed aldol additions of DHAP to C-a-substi-
tuted N-Cbz-aminoaldehydes is always anti (3R,4R), regard-
less of the structure and stereochemistry of the aldehyde.
RhuA displayed excellent stereoselectivity towards the S
linear alkyl-substituted aldehydes, while significantly low for
those with R configuration (Figure 4). In this respect, the
amount of anti (3R,4R) configuration increased with the
length of the alkyl chain, the (R)-butyl giving the inverted
C-4 adduct (i.e., FucA stereochemistry) as major product
(Figure 4A).
Figure 2. Molar percent conversions to products 3 in the aldol additions
of DHAP to N-Cbz-aminoaldehydes ([DHAP]=50 mm; [N-Cbz-aminoal-
dehydes]=85 mm, the results at higher concentrations are summarized in
Supporting Information) with linear (top) and branched (bottom) C-a
substituents, catalyzed by RhuA in high water content emulsions (black
bars) and in 1:4 DMF/water solution (grey bars). Conditions as explained
in caption of Figure 1.
With the exception of the most hydrophobic butyl sub-
stituent, linear alkyl groups gave similar reaction conver-
sions regardless of stereochemistry of aldehyde and reaction
system used, while 1:4 DMF/water was the medium of
choice for the sterically more demanding branched alkyl
substituents. Furthermore, on increasing the reactant con-
centration up to [DHAP]=100 mm and [N-Cbz-aminoalde-
hyde]=170 mm (see Supporting Information) conversions
for RhuA- and FucA-catalyzed reactions were generally
higher in emulsion than those in 1:4 DMF/H2O, probably
because of the better solubilization properties of the former
reaction medium.[37]
Investigations of temperature effects: Temperature effects
were also investigated for RhuA aldolization of aldehydes
(R)-2a, (R)-2b (vide infra), (R)-2c, (R)-2 f, and (S)-2 f. To
this end, the selected reactions were conducted at 48C, be-
cause it has been reported that at this temperature the con-
version achieved on FucA-catalyzed aldol addition of
DHAP to (R)-N-Cbz-alaninal was significantly improved.[38]
The same trend was observed with C-a-branched alkyl
substituents (Figure 4B). Interestingly, (S)-2d gave the ex-
pected aldol adduct with 4S stereochemistry, whereas its en-
antiomer (R)-2d gave the opposite 4R adduct, both as single
diasteromers. Therefore, RhuA displayed an excellent ste-
reoselectivity for these reactions.
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7313

J. Bujons, P. Clapꢂs et al.
Table 1. Structures of iminosugars and stereoisomeric ratio ascertained
by NMR spectroscopy from the strategy depicted in Scheme 1.
Figure 3. Stereoselectivity of FucA catalysis in the aldol additions of
DHAP to N-Cbz-aminoaldehydes. Percentage of adducts anti, that is,
(3R,4R)-3 (or (3S,4R) in the iminosugar, black bars), and syn, that is,
(3R,4S)-3 (or (3S,4S) in the iminosugar, grey bars), for each aldehyde ac-
ceptor.
Entry
Aldehyde
Aldolase
Product (ratio)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
(S)-2a
(S)-2a
(R)-2a
(R)-2a
(S)-2b
(S)-2b
(R)-2b
(R)-2b
(S)-2c
(S)-2c
(R)-2c
(R)-2c
(S)-2d
(R)-2d
(S)-2e
(S)-2 f
(R)-2 f
(R)-2 f
(S)-2g
(S)-2g
2h
RhuA
FucA
RhuA
FucA
RhuA
FucA
RhuA
FucA
RhuA
FucA
RhuA
FucA
RhuA
RhuA
RhuA
RhuA
RhuA
RhuA
RhuA
FucA
RhuA
FucA
4a (100)
4a/4b/4c (10:86:4)
4d/4e/4 f/4g (48:30:11:11)
4e (100)
4h (100)
4h/4i/4j (4:86:10)
4k/4l/4m/4n (44:34:11:10)
4l (100)
4o/4p (93:7)
4o/4p/4q (10:80:10)
4s/4r/4t/4u (28:63:5:5)
4r (100)
4v/4w (84:16)
4x (100)
Figure 4. Stereoselectivity of RhuA catalysis towards the N-Cbz-aminoal-
dehydes with linear (top) and branched (bottom) C-a substituents. Per-
centage of adducts anti, that is, (3R,4R)-3 (or (3S,4R) in the iminosugar,
black bars), and syn, that is, (3R,4S)-3 (or (3S,4S) in the iminosugar, grey
bars), for each aldehyde acceptor.
4y/4z (55:45)
4aa/4ab/4ac (71:6:23)
4ae/4ad/4af/4ag (32:43:16:9)
4ad[a] (100)
4ah (100)
4ai/4aj (62:38)
4ak/4al (83:17)
4al/4am (44:56)
was detected. This result indicates that the anti aldol adduct
was kinetically preferred and that a reduction of the temper-
ature increased the stereoselectivity, as was also observed in
other aldolases.[39]
2h
[a] Synthesis conducted at 48C.
Molecular modeling: Molecular modeling studies were car-
ried out to substantiate this hypothesis. Thus, an exhaustive
conformational analysis was carried out for the syn and anti
adducts derived from the (S)-2c, (R)-2c, (S)-2d, and (R)-2d
aldehydes, as representative examples, using a combination
of the systematic and stochastic conformational search algo-
rithms implemented in the program MOE (Chemical Com-
puting Group, Montreal), as previously described.[24,40]
Table 2 summarizes the difference in calculated energies for
Investigations of temperature effects: The effect of tempera-
ture on this stereoselectivity was also investigated as a func-
tion of the N-Cbz-aminoaldehyde. For instance, no changes
were observed with (R)-2c, whereas an increase of the anti
(3R,4R) configuration from 44 to 61% was found for (R)-2b
when cooling down from 25 to 48C, respectively. The most
significant result was obtained with (R)-2 f: at 258C it gave
a 55:45 anti:syn ratio, whereas at 48C only the anti product
7314
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
Table 2. Energy difference [kcalmolÀ1] of the global minima determined
for the syn and anti adducts derived from aldehydes 2c and 2d, in the
free state and bound in the RhuA active site. Energies determined with
the MMFF94x forcefield under continuum water solvation conditions, as
detailed in Experimental Section.
hydes (S)-2c and (S)-2d bound in the active center of
RhuA. The modeled structures show that the aldol reaction
products coordinate to the active site metal simultaneously
with the oxygen atoms of the C-2 carbonyl and the C-3 and
C-4 hydroxyl groups, thus providing a hexacoordinate envi-
ronment to the essential Zn2+, with the participation of the
histidines at positions 141, 143, and 212 of the protein. Simi-
lar geometries were obtained for the adducts derived from
(R)-2c and (R)-2d. This arrangement resembles that found
in the crystal structure of RhuA-phosphoglycolohydroxa-
mate complex (PDB 1GT7), in which a water molecule co-
ordinates the cation in addition to the two oxygen atoms
from the ligand. In a very recent paper in which high-level
computations have been used to study the enzymatic mecha-
nisms of RhuA and FucA, Jimꢂnez et al.[41] have shown that
Aldehyde
Free adduct
RhuA-adduct complex
DEsyn–anti
DEsyn–anti
(R)-2c
(S)-2c
(R)-2d
(S)-2d
À4.3
À3.1
À5.1
À0.7
18.2
12.9
23.6
34.1
the global minima determined for each pair of syn–anti ad-
ducts (DEsyn–anti) and the corresponding geometries are re-
ported in the Supporting Information. From these values it
is clear that the syn adducts are thermodynamically favored,
assuming similar entropic contributions to DG, by
À3.3 kcalmolÀ1 on average for the compounds considered.
This value correlates well with the stereochemistries ob-
served for the aldol adducts obtained from (S)-aldehydes
with RhuA catalysis. Similar results were already reported
by us for a small set of aldol adducts derived from N-pro-
tected 3-aminopropanal.[40]
À
in the transition state for the C C formation step, a negative
charge builds up on the aldehyde oxygen atom, which is sta-
bilized by interaction with the positive charge of the Zn2+
,
and that this interaction is maintained by the aldol product.
Our models for the RhuA-adduct complexes gave the
energy differences between the syn and anti products, which
are summarized in Table 2. The anti adduct complexes were
always lower in energy than the corresponding syn ones, ap-
parently due to more strained geometries for the later. De-
termination of the actual energy differences between transi-
tion states would require of accurate quantum mechanics
calculations. However, it seems plausible that since the geo-
We also looked at the complexes of the above adducts in
the active center of RhuA. Figure 5 shows the lowest energy
conformers obtained for the syn and anti adducts from alde-
À
metries of the transition states for the C C bond formation
reactions are similar to those of the aldolase adduct com-
plexes,[41] the DEsyn–anti values of Table 2 could correlate with
the energy differences for the corresponding transition
states. Therefore, these results would suggest lower activa-
tion barriers for the anti adducts and, consequently, a kinetic
preference for these, which could explain the formation of
larger proportions of anti products from aldehydes (R)-2c
or (R)-2d. If this can be extended to the other aldehydes, it
would provide an explanation to the experiments carried
out at low temperature.
Reductive amination: Conversion of the corresponding ad-
ducts to polyhydroxylated pyrrolidine derivatives 4 by Pd/C
catalyst in the presence of H2 (50 psi) gave in most cases the
expected 2-OH/4-OH syn configuration. On the other hand,
aldol adducts with 3R,4R,5S configuration gave five-mem-
bered ring iminosugars with 2-OH/4-OH anti configuration
as major product (Table 1, entries 2, 6, and 10). Exceptions
to this were the aldol adducts from the aldehydes (S)-2g
and 2h, for which limited or no stereoselectivity was ob-
served (Table 1, entries 20 and 22).
Minor amounts of N-ethylated byproducts may account
for a reductive amination of ethanal and the iminosugar.
The ethanal was formed in situ by Pd-catalyzed oxidation of
ethanol from the solvent mixture in presence of dissolved
oxygen[42] and before H2 addition. This side reaction was
avoided if the oxygen dissolved of the reactants solution was
displaced by bubbling with N2 gas prior to the addition of
Pd/C catalyst.
Figure 5. Structural models of the syn (dark grey) and anti (light grey)
adducts derived from aldehyde (S)-2c in the active center of RhuA
(top), and the syn (dark grey) and anti (light grey) adducts derived from
aldehyde (S)-2d in the active center of the same enzyme (bottom). The
corresponding color figure is available in the Supporting Information.
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7315

J. Bujons, P. Clapꢂs et al.
NMR studies: Similar to the NMR study on the iminocycli-
tols obtained in previous papers,[24,25] and as a complement
to NOE data, the shielding effects were also used as a probe
to assign the relative stereochemistry (Table 3 in the Sup-
1
porting Information). Differences on the H chemical shift
were observed depending on the protonation state of the
imino group. The protonated imino group induced 0.1–
0.5 ppm upfield shift on protons of the molecule (see Sup-
porting Information). This was useful in some cases to sepa-
rate unresolved signals. Remarkably, the NMR analysis of
4z and 4ac revealed the presence of two conformers de-
pending on the protonation of the imino group. When the
imino group existed as free base the spectra showed signals
assigned unequivocally to the product, while when the imino
group was protonated, the spectrum showed the presence of
two groups of signals corresponding to the two conformers
of the product (see Supporting Information).
Preliminary inhibitory properties against glycosidases: The
compounds, synthesized both in pure form and as mixtures,
were assayed as inhibitors of commercial glycosidases: a-d-
glucosidase from bakerꢇs yeast, b-d-glucosidase from sweet
almonds, b-d-galactosidases from bovine liver and Aspergil-
lus oryzae, a-l-fucosidase from bovine kidney, a-l-rhamno-
sidase from Penicillium decumbens, and a-d-mannosidase
from jack bean.
A cursory inspection of IC50 values
(Figure 6) indicated that among the iminocyclytols synthe-
sized, there are good inhibitors of a-l-fucosidase (IC50 =1–
20 mm), moderate inhibitors of a-l-rhamnosidase (IC50 =7–
150 mm) and poor inhibitors of a-d-mannosidase (IC50 =80–
400 mm). Furthermore, none of them had inhibitory proper-
ties against a-d-glucosidase from bakerꢇs yeast, b-d-glucosi-
dase from sweet almonds, and b-d-galactosidases from
bovine liver and A. oryzae. Next, Ki and type of inhibition
were determined for the most active compounds or mixtures
(Table 3). Mixtures showing a poor or null inhibitory activity
were not further processed.
Figure 6. IC50 values of the compounds synthesized for the inhibition of
a-l-fucosidase from bovine kidney (top), a-l-rhamnosidase from P. de-
cumbens (middle), and a-d-mannosidase from jack bean (bottom).
Empty bars: no inhibition.
Table 3. Activities of the compounds synthesized. IC50 and Ki values of the most active compounds against: a-l-fucosidase from bovine kidney, a-l-
rhamnosidase from P. decumbens, and a-d-mannosidase from jack bean.
IC50 [mm]
Ki [mm]
a-l-fucosidase
a-l-rhamnosidase
a-d-mannosidase
a-l-fucosidase
a-l-rhamnosidase
a-d-mannosidase
4a
100
0.6
1
5
200
9
100
300
40
15
300
ni
100
80
ni
ni[f]
ni
ni
400
ni
ni
450
ni
ni
500
ni
ni
nd[g]
0.30Æ0.10
0.35Æ0.15
7.6Æ0.5
nd
nd
–
–
–
4b[a]
4c[a]
4e[a]
4h
109Æ6
–
80
49Æ4
62Æ6
nd
137Æ5
150
150
100
7
80
400
20
–
–
4i[a]
4l
7.0Æ0.7
nd
nd
305Æ10
4o[b]
4p[c]
4r
nd
4.8Æ0.6
131Æ8
nd
26Æ4
55Æ5
25Æ4
–
–
8.5Æ0.8
22Æ4
nd
–
–
373Æ15
4v[d]
4y[a]
4aa[e]
–
–
–
30
20
ni
ni
[a] Purified as described in the Experimental Section. [b] Purity 93% (entry 9, Table 1). [c] Purity 80% (entry 10, Table 1). [d] Purity 84% (entry 13,
Table 1). [e] Purity 71% (entry 16, Table 1). [f] ni=no inhibition. [g] nd=not determined.
7316
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
Inhibitory properties of the compounds against a-l-fucosi-
dase from bovine kidney: The compounds given in entries 2,
3, 4, 6, 10, and 12 (Table 1) inhibited a-l-fucosidase in a
competitive manner. The iminosugars from entry 2, the mix-
ture with the highest activity, were isolated and purified,
and their inhibitory properties against a-l-fucosidase evalu-
ated. Compounds 4b and 4c, were the most potent among
those obtained in this work (Table 3), whereas 4a, which
was also obtained in entry 1, was much less active. The relat-
ed compounds 2,5-imino-1,2,5-trideoxy-l-altritol (5) and 2,5-
dideoxy-2,5-imino-l-fucitol (6), analogues of 4b and 4c, re-
spectively by which carry a 5-methyl substituent instead of
the ethyl group, were found to be more potent a-l-fucosi-
dase inhibitors with, Ki =80 and 4.9 nm, respectively.[43,44]
from animals, and of the 13 residues that constitute the
active site cavity, 11 are conserved in the human protein.
Furthermore, homology modeling suggested a high similari-
ty between the shapes of the catalytic center of both the
bacterial and human proteins, while larger differences were
found in the aglycone binding site.[54] Consequently, the
structure of a-l-fucosidase from T. maritima was used as a
reasonable model to carry computational docking studies
that help to understand the fucosidase binding capacity and
the way of interaction of the pyrrolidine inhibitors identi-
fied.
Figure 7 shows the best docked poses obtained for the a-
l-fucosidase active compounds 6 and 4b and the less active
4r, as well as the experimentally determined structure of l-
The activity of compound 6, not previously reported to the
best of our knowledge, is amongst the highest reported for
pyrrolidine and piperidine type iminosugars.[45–49] Other pyr-
rolidine derivatives with similar stereochemistry (i.e.,
3S,4R,5S) have also been reported as strong inhibitors of a-
l-fucosidase,[50–52] while some diastereomerically related iso-
mers of 5 and 6 have shown much lower activities (e.g. Ki =
165 mm for the 2R,3R,4R,5R isomer[46]), suggesting a strong
stereochemical demand of the a-l-fucosidase active site. In
this sense, looking at the compounds synthesized in this
work, modification of the stereochemistry at C-5 from S to
R, for example, compare 4c versus 4e, or at C-4 from R to
S, for example, compare 4b versus 4a, 4i versus 4h, or 4p
versus 4o, result in moderate to large decreases of the inhib-
itory activity (Table 3).
On the other hand, increasing the length of the substitu-
ent at C-5 from methyl (e.g., 5 and 6) to ethyl (e.g., 4b and
4c), propyl (e.g., 4i), or butyl (e.g., 4p) implied a progres-
sive loss of inhibitory activity. Furthermore, benzyl or di-
methyl substituents at C-5, (e.g., entries 19–22) completely
suppressed the activity of the corresponding pyrrolidine de-
rivatives (Figure 5, top). Altogether, these results suggest a
relatively small hydrophobic pocket in the active site of a-l-
fucosidase, which could be occupied by small substituents
on C-5 position, but not the larger ones. In this respect,
Laroche et al.[53] described a spirocyclopropyl pyrrolidine (7)
as a competitive inhibitor (Ki =1.6 mm). The sterically more
restricted cyclopropane group may be better accommodated
in this hydrophobic core than the dimethyl group (entries 21
and 22) and effectively interact with the enzyme active site.
Although no fucosidase structure of mammalian origin is
available up to date in the Protein Data Bank, the structure
of a-l-fucosidase from Thermotoga maritima has recently
been reported.[54] The sequence of this bacterial enzyme ex-
hibits 34–38% identity to the sequences of a-l-fucosidases
Figure 7. Top: Crystal structure of l-fucose (dark grey) bound in the
active center of a-l-fucosidase from T. maritima (PDB 1ODU) and best
docked pose obtained for 6 (light grey). Bottom: Best docked poses ob-
tained for 4b (light grey) and 4r (dark grey) in the active center of the
same enzyme. The corresponding color figure is available in the Support-
ing Information.
fucose bound into the active center of a-l-fucosidase from
T. maritima, for the sake of comparison. The pyrrolidine in-
hibitors occupy the catalytic site in a similar manner as the
crystallized ligand and establish a strong interaction between
their protonated imine group and the catalytic E266 residue.
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7317

J. Bujons, P. Clapꢂs et al.
However, the methyl group of 6 and the ethyl group of 4b
are directed towards the inner part of the cavity, where
there is a relatively small hydrophobic pocket constituted by
residues F32, Y171, F290, and W222, while 4r shows an in-
verted orientation, with its larger butyl substituent pointing
towards the external opening of the binding pocket. The in-
teraction diagram for 6 (see Supporting Information) shows
extensive hydrogen-bonding interactions with residues H34,
E66, W67, H128, H129, D224, and E266, while the slightly
displaced conformation of 4b, required to allocate the larger
ethyl group in the hydrophobic cavity, shows hydrogen
bonding with residues W67, H129, D224, and E266. The
even larger size of the butyl group, which forces 4r to adopt
an inverted conformation, only allows for hydrogen bond in-
teractions with residues W67 and E266. This trend, that is,
smaller alkyl chains (methyl, ethyl, or propyl) located in the
hydrophobic cavity and larger chains (butyl or benzyl) ori-
ented towards the solvent exposed entry to the cavity, is ob-
served for most of the compounds considered in our docking
studies (see Supporting Information). Therefore, the deter-
mined binding modes for these compounds are in good
agreement with the observations suggesting a steric impedi-
ment for bulky substituents at the pyrrolidine C-5 position
and lower activities with increasing size of the substituents
at that position. Indeed, the scores derived from the estima-
tion of the interaction energies (MM/GBVI scores, see Ex-
perimental Section) between the inhibitors and the a-l-fuco-
sidase correlate reasonably well with the free energies of
binding calculated from the inhibition data (Table 4 and
Figure 8. Correlation between DGexptl and MM/GBVI scores from Table 4.
Table 4. Free energies of binding determined from the experimental Ki
values and docking MM/GBVI scores.
DGexptl
MM/GBVI
score
DGexptl
MM/GBVI
score
[KcalmolÀ1
]
[KcalmolÀ1
]
4a
4b
4c
4e
4i
À5.46[a]
À8.90
À8.81
À6.98
À7.03
À6.92
À25.4
À25.3
À26.5
À25.3
À23.4
À25.1
4r
4aj
4am
5
6
7
À6.35
–
–
À9.68
À11.34
À7.91
À20.7
À24.5
À19.9
À29.2
À30.7
À25.8
most potent a-l-rhamnosidase inhibitor found among the
iminosugars synthesized in this work (Ki =4.8 mm). It has
been suggested that the a-l-rhamnosidase inhibition shown
by some pyrrolidines can be rationalized in terms of stereo-
chemical similarities with a-l-rhamnose.[56,57] Considering
the structure of naringin, a natural substrate of a-l-rhamno-
sidase, it is apparent that the stereochemistry of 4o (i.e.,
2S,3S,4S,5S) matches that of the rhamnose moiety at C-3, C-
4, and C-5. This would place the 5-butyl group of 4o on the
same location as the 5-methyl group of rhamnose. A similar
reasoning was argued by Chapman et al. to explain the a-l-
rhamnosidase inhibitory activity of 8 (Ki =3.0 mm).[56] How-
ever, Kim et al. proposed a role as aglycone for hydrophobic
substituents at C-2 of pyrrolidines such as 9 (Ki =3.4 mm).[57]
4p
[a] Calculated from the IC50 value.
Figure 8). Although this correlation could be improved if
the structural differences between the a-l-fucosidase from
T. maritima and the bovine enzyme were considered, it does
reflect a plausible binding mode for pyrrolidine type inhibi-
tors that might probably be conserved for both a-l-fucosi-
dases. Better accounting for other factors like the entropic
contribution to binding or the differences in solvation–desol-
vation energies among the different compounds, recently
shown as having an enormous contribution to ligand binding
among glycosidase inhibitors,[55] would also have a large
impact to improve such correlation. Work in this sense is ac-
tually being carried out by our group.
Inhibitory properties of the compounds against a-l-rhamno-
sidase from Penicillium decumbens: Compound 4o was the
7318
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
The lack of structural data for a-l-rhamnosidase from P.
decumbens precludes a definitive answer about the binding
mode of pyrrolidine inhibitors into the active center of the
enzyme. In this sense, the crystal structure of an a-l-rham-
nosidase from Bacillus sp. GL1 (RhaB) has recently been
solved.[59] This enzyme exhibits a slight identity with the
known sequences of a-l-rhamnosidases from fungi (i.e.,
24% with RhaA and RhaB of Aspergillus aculeatus, or 20%
with RamA of Clostridium stercorarium).[60] However, the
fact that it shows activity against the common substrate nar-
ingin suggests that its catalytic site should share similar
shape and structural features with other a-l-rhamnosidases
like that of P. decumbens. Therefore, we carried out docking
experiments with the structure of RhaB from Bacillus sp.
GL1 to shed light about the binding mode of pyrrolidine 4o.
Figure 9 (top) shows the best three poses obtained for this
compound, which differ slightly but establish essentially the
same interactions with the enzyme. Unlike for the case of a-
l-fucosidase, these docking results suggest that 4o could
bind into the a-l-rhamnosidase active site orienting the
butyl substituent towards the inside of the enzyme cavity, in
a hydrophobic subsite made up of W576, I626, W629, W679,
and L681, which has been proposed as the recognition site
for the methyl group of the rhamnosyl moiety.[59] The pro-
tonated imine group of the inhibitor establishes a strong in-
teraction with the essential E572 residue, while the hydroxyl
groups at C-1, C-3, and C-4 can establish diverse hydrogen-
bond interactions with residues Q147, N849, and R858.
Naringin was also docked in the active center of a-l-
rhamnosidase (Figure 9, bottom) for comparison purposes,
since there is no structure available in the PDB of the
enzyme with a bound rhamnosyl derivative or inhibitor. The
best docked pose obtained constitutes a plausible geometry
for this substrate, with the reactive C-1 and 1,2-b-glycosidic
oxygen atoms located close to the essential E572, the rham-
nose moiety parallel to the ring of W576, fixed by hydropho-
bic stacking, and its methyl group oriented towards the hy-
drophobic residues I626, W679, and L681.[59] Comparing this
docked naringin with the docked conformations obtained
for 4o, it becomes apparent that the size of the butyl chain
forces some displacement of the pyrrolidine ring away from
the rhamnose location and towards the entrance to the
cavity. This causes a lack of coincidence of the hydroxyl
groups at C-3 and C-4 of the inhibitor with those of the
rhamnose moiety of naringin. However, it should be kept in
mind that the pyrrolidine inhibitors, containing a protonated
imino group, are thought to mimic the enzyme transition-
state complex rather than enzyme–substrate complex. None-
theless, the orientation of the alkyl chain towards the hydro-
phobic subcavity in the catalytic site, resulting in a more ef-
ficient binding for compounds with larger hydrophobic
chains, correlates with the observation that decreasing the
C-5 chain length from butyl to propyl (4h), ethyl (4a) or
methyl (10; Ki =43 mm)[25] led to a loss of inhibitory capacity.
Branched alkyl chains on C-5, for example, compounds 4v,
4y and 4aa, showed some inhibitory properties, but always
lower than those of 4o, likely reflecting additional steric re-
quirements of the biding site. In this sense, the C-5 dimethyl
substituted pyrrolidines, entries 19–22, were not inhibitors of
a-l-rhamnosidase. This is in contrast with the activities ob-
served for compounds such as 8 and 9, which are analogues
of 4ah. The lack of activity of 4ah suggests that the hydrox-
ymethyl moiety present in this pyrrolidine might have a del-
eterious effect on the activity against the enzyme. This
effect could also contribute to the lower activity observed
for 4a or 10 compared to their counterparts 11[58] or 12[56]
(Ki =5.5 and 5.9 mm, respectively).
Figure 9. Representation of the active site of a-l-rhamnosidase from Ba-
cillus sp. GL1 with the three best docked conformations obtained for pyr-
rolidine 4o (light grey; top), and the best docked pose obtained for nar-
ingin (light grey; bottom). The corresponding color figures are available
in the Supporting Information. It is apparent on both representations the
presence, at the center-left side, of a number of residues (F102, W576,
W622, I626, W629, W679, L681, W684 and W864) that constitute a rela-
tively large hydrophobic subcavity, while the center-right side is essential-
ly constituted by polar residues (Q147, D567, E572, D579, E841, N849
and R858) that can establish multiple hydrogen bonds with the substrates
and the inhibitors.
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7319

J. Bujons, P. Clapꢂs et al.
Inhibitory properties against a-d-mannosidase from jack
bean: The compounds synthesized showed weak inhibitory
properties against a-d-mannosidase. Pyrrolidine derivatives
with 2R,3S,4R,5R configuration and linear alkyl substituents
on C-5 (synthesized compounds with general structure 15,
e.g., 4e, 4l, and 4r) were the best competitive inhibitors
among the compounds synthesized. These compounds show
a cis arrangement of their hydroxyl groups at C-3 and C-4,
which parallels that of the hydroxyl groups at C-2 and C-3
of a-d-mannose (13) and other known pyrrolidine inhibitors
of a-d-mannosidase like mannostatin A (14). These cis-ar-
ranged hydroxyl groups are known to coordinate the essen-
tial Zn2+ present in the active center of the related a-d-
mannosidase from D. melanogaster,[61–64] which is a class II
a-d-mannosidase belonging to glycosyl hydrolase family 38
(Carbohydrate Active Enzymes database, http://www.cazy.
org)[65] as the a-d-mannosidase from jack bean. Both a-d-
mannosidases show the characteristic IDPFGH sequence in
their active site,[66] which is highly conserved among all clas-
s II a-d-mannosidases, and both are inhibited similarly by
compounds like mannostatin and swainsonine. Therefore,
the structure of the a-d-mannosidase from D. melanogaster
was used to model the binding of the pyrrolidines synthe-
sized in this work.
Figure 10 shows the docking results obtained for 4e and
4r, which were very similar to those found for 4l (not
shown), as well as the experimentally determined structures
of mannose and mannostatin A bound in the active center
of a-d-mannosidase from D. melanogaster, for comparison.
The docking poses found for the pyrrolidines share a similar
binding mode with mannose and mannostatin A, coordinat-
ing the Zn2+ with the hydroxyl groups at C-3 and C-4 and
hydrogen bonding to residues D92, D204, D472, and Y727.
They also establish a strong interaction between the proton-
ated imine and the negatively charged D204, similar to
those found between the charged amino group of mannosta-
tin A and the catalytic acid residues D204, D341, and
Y269.[62] However, the hydrophobic chains of 4e and 4r are
oriented towards the water-accessible opening of the cavity
and not towards the hydrophobic pocket constituted by resi-
Figure 10. Top: Best docked poses obtained for compounds 4e (light
grey) and 4r (dark grey) in the active site of a-d-mannosidase from D.
melanogaster. Bottom: Bound conformations of mannostatin A (light
grey) and mannose (dark grey) in the active center of the same enzyme
(PDB 2F7O and 3BUQ, respectively).[67] The corresponding color figures
are available in the Supporting Information.
of their hydrophobic chain to the solvent could explain the
low activity observed for pyrrolidines 4e, 4l, and 4r. Substi-
tuting the linear by branched alkyl chains, for example, 4x
and 4ad, abolish completely the inhibitory properties of the
pyrrolidine derivatives and might reflect additional steric re-
quirements of the active center of the enzyme.
Conclusions
dues F206, W415, and Y727, in which the thiomethyl group
67]
of mannostatin A is located.[62,
The interaction of this
Chemoenzymatic methods using DHAP-dependent aldolas-
es as biocatalysts for the key aldol addition step are both
convenient and advantageous for the generation of configu-
rational and functional collections of polyhydroxylated pyr-
rolidine compounds for bioactivity testing. FucA had re-
stricted substrate tolerance for C-a linear N-Cbz-2-aminoal-
dehydes, whereas RhuA accepted all types of C-a substitu-
tions selected. On the other hand, d-fructose-1,6-bisphos-
methyl group with the hydrophobic residues seems to con-
tribute importantly to the high potency of mannostatin A.
Furthermore, it allows the right orientation of the unpaired
electrons of the sulfur atom, such that they can interact with
the backbone of residue R876, which seems to be a critical
determinant of potency.[62] Therefore, the lack of these im-
portant interactions together with the unfavorable exposure
7320
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
(particle size 26 mm, mean pore diameter 17000 nm, specific surface area
BET method 2.19 m2 gÀ1) was obtained from Fluka. a-d-Glucosidase
from bakerꢇs yeast, b-d-glucosidase from sweet almonds, b-d-galactosi-
dase from bovine liver, a-d-mannosidase from jack bean, a-l-rhamnosi-
dase (naranginase) from Penicillium decumbens, and a-l-fucosidase from
bovine kidney and the corresponding substrates p-nitrophenyl-a-d-gluco-
pyranoside, p-nitrophenyl-b-d-glucopyranoside, p-nitrophenyl-b-d-galac-
topyranoside, p-nitrophenyl-a-d-mannopyranoside, p-nitrophenyl-a-l-
rhamnopyranoside, and p-nitrophenyl-a-l-fucopyranoside, respectively,
were purchased from Sigma. Deionized water was used for preparative
HPLC and Milli-Q-grade water for analytical HPLC. All other solvents
used were of analytical grade.
NMR spectroscopy: 1H (500.13 MHz) and 13C (125.76 MHz) NMR spec-
tra were recorded on an Avance 500 Bruker spectrometer equipped with
a highly sensitive, cryogenically cooled, triple-resonance TCI probehead
for samples dissolved in D2O and CD3OD. Full structural and stereo-
chemical characterization of all compounds was performed with the aid
of two dimensional COSY, NOESY, HSQC, and HMBC experiments as
well as NOE data obtained from selective 1D NOESY experiments re-
corded with a mixing time of 500 ms. The relative orientation of hydroxyl
groups was determined from the characteristic downfield/upfield effects
observed for all H-1, H-2, H-3, H-7a proton chemical shift, as recently re-
ported for other iminocyclitols.[24,25] As a general trend, strong downfield
effects are observed for the most characteristic H-1, H-2, and H-3 pro-
tons when the corresponding vicinal hydroxyl groups were found in a rel-
ative anti position. 1H and 13C chemical shifts were calibrated to a DSS
external reference.
phate aldolase from rabbit muscle (RAMA) did not catalyze
the aldol reactions with any of the selected aminoaldehydes.
FucA catalyst was the most steroselective (i.e., anti (3R,4R)
configuration of the stereogenic centers formed) catalyst,
the stereochemical outcome of the reactions being con-
trolled by the enzyme. RhuA was highly stereoselective (i.e.,
syn (3R,4S)) for the (S)-aminoaldehydes regardless of the
C-a substitution, but non-selective with the (R)-aminoalde-
hydes giving mixtures of anti/syn adducts. Molecular models
of selected RhuA–aldol adduct complexes revealed that the
anti adducts modeled were lower in energy than the corre-
sponding syn adducts. Given the similarity between the geo-
metries of the complexes and the corresponding transition
states, these models would suggest a kinetic preference that
could explain the formation of larger proportions of anti ad-
ducts from aldehydes (R)-2c or (R)-2d and could provide
an explanation to the temperature effects. Drawbacks of this
approach are thus the acceptor substrate tolerance, occa-
sionally lack of stereoselectivity, which in some instances de-
pends on acceptor structure, or an unfavorable equilibrium
position of the addition reaction. These might be overcome
with both enzyme and medium engineering, aspects that are
being currently under investigation in our lab.
Specific rotations: Measurements were made on a Perkin–Elmer Model
Examination of the inhibitory properties of the com-
pounds showed that they are good inhibitors of a-l-fucosi-
dase (IC50 =1–20 mm), moderate of a-l-rhamnosidase (IC50 =
7–150 mm) and weak of a-d-mannosidase (IC50 =80–400 mm).
The docking results obtained with a-l-fucosidase show a
good correlation with the experimental activities, and indi-
cate that small alkyl chains at C-5 position of the pyrrolidine
are preferred for high affinity. We report for the first time a
strong a-l-fucosidase inhibition for the 5-methyl substituted
pyrrolidine with 2R,3S,4R,5S configuration, more than
ꢀ30000-fold higher than for its 2R,3R,4R,5R diastereomer,
revealing the great importance of the stereochemistry for
the activity of these compounds. On the other hand, prelimi-
nary results obtained with an a-l-rhamnosidase model
would suggest that larger C-5 alkyl substituents might inter-
act more strongly with its active center. Finally, the results
obtained for a-d-mannosidase are not conclusive for the
compounds described in this work.
341 (ꢉberlingen, Germany) polarimeter.
HPLC analyses: HPLC analyses were performed on a RP-HPLC car-
tridge, 250ꢊ4 mm filled with Lichrosphere 100, RP-18, 5 mm from Merck
(Darmstadt, Germany). Samples (50 mg of the emulsion mixture or
50 mL of the reaction in DMF/water 1:4) were withdrawn from the aldol
reaction, dissolved in MeOH (1 mL) to stop the reaction and analyzed by
HPLC. The solvent system used was the following: solvent A: H2O 0.1%
(v/v) trifluoroacetic acid (TFA); solvent B: ACN/H2O 4:1 0.095% (v/v)
TFA, gradient elution from 30 to 90%
1 mLminÀ1, detection 215 nm.
B over 30 min, flow rate
General procedure for the synthesis of (S)- or (R)-N-benzyloxycarbony-
laminoaldehydes: Cbz-OSu (50 mmol) in dioxane/water 4:1 (10–50 mL)
was added dropwise to a solution of (S) or (R)-aminoalcohol (50 mmol)
in dioxane/water 4:1 (100 mL) at 258C. After stirring for 24 h, the mix-
ture was evaporated to dryness under reduced pressure. The residue was
dissolved with ethyl acetate (150 mL) and washed successively with citric
acid 5% w/v (3ꢊ50 mL), NaHCO3 10% w/v (3ꢊ50 mL), and brine (2ꢊ
50 mL). After drying over Na2SO4, the organic layer was evaporated
under reduced pressure affording white solids of (S)- and (R)-N-Cbz-ami-
noalcohols (mean yield: 85% yield; 98% purity by HPLC). NMR spec-
trum analysis of the residue showed only one product. Oxidation of (R)-
or (S)-N-Cbz-aminoalcohols (25 mmol) was efficiently carried out by a
procedure described by Ocejo et al.[69]
(S)-1-Oxobutane-2-benzylcarbamate ((S)-2a): Pale yellow oil; yield:
2.5 g, 96%; 99.9% purity by HPLC; [a]2D0 =+18.08 (c=0.9 in CH2Cl2);
1HNMR: (300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.4 (m, 5H; Ph), 5.4
(brs, 1H; NH), 5.1 (s, 2H; PhCH2O), 4.3 (dd, J=6.7, 12.8 Hz, 1H;
CHCHO), 2.0, 1.7 (m, 2H; CH3CH2), 1.0 ppm (t, J=7.9 Hz, 3H; CH3);
13CNMR: (75 MHz, CDCl3): d=199.1 (CHO), 155.0 (OCONH), 136.0 (C
ar), 128.5 (C ar), 128.2 (C ar), 127.9 (C ar), 67.0 (PhCH2O), 60.3 (CH),
22.2 (CH3CH2), 9.2 ppm (CH3CH2).
Experimental Section
Materials: d-Fructose-1,6-bisphosphate aldolase from rabbit muscle
(RAMA; EC 4.1.2.13, crystallized, lyophilized powder, 19.5 UmgÀ1) was
from Fluka (Buchs, Switzerland). Rhamnulose-1-phosphate aldolase
(RhuA; EC 4.1.2.19, suspension 100 UmLÀ1) was kindly donated by
Roche Diagnostics (Mannhein, Germany). l-Fuculose-1-phosphate aldo-
lase (FucA, EC 4.1.2.17, lyophilized 500–800 U gÀ1 or as (NH4)2SO4 sus-
pension (25 UmLÀ1) was from Departament d’Enginyeria Quꢈmica of the
Universitat Autꢆnoma de Barcelona, produced from a recombinant E.
coli (ATCC no 86984) and purified by affinity chromatography. Acid
phosphatase (PA, EC 3.1.3.2, 5.3 UmgÀ1) was from Sigma (St. Louis,
USA). The precursor of dihydroxyacetone phosphate (DHAP), the dihy-
droxyacetone phosphate diethyl ketal dimer, was synthesized in our lab
by using a procedure described by Jung et al.[68] with slight modifications.
Aluminum oxide 90 active neutral was purchased from Merck. Celite-545
(R)-1-Oxobutane-2-benzylcarbamate ((R)-2a): Pale yellow oil; yield: 2 g,
97%; 99.9% purity by HPLC; [a]2D0 =À25.08 (c=1.0 in CH2Cl2);
1HNMR: (300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.4 (m, 5H; Ph), 5.4
(brs, 1H; NH), 5.1 (s, 2H; PhCH2O), 4.3 (dd, J=6.7, 12.8 Hz, 1H;
CHCHO), 2.0, 1.7 (m, 2H; CH3CH2), 1.0 ppm (t, J=7.1 Hz, 3H; CH3);
13CNMR: (75 MHz, CDCl3): d=199.1 (CHO), 155.0 (OCONH), 136.0 (C
ar), 128.5 (C ar), 128.2 (C ar), 127.9 (C ar), 67.0 (PhCH2O), 60.3 (CH),
22.2 (CH3CH2), 9.2 ppm (CH3CH2).
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7321

J. Bujons, P. Clapꢂs et al.
(S)-1-Oxopentane-2-benzylcarbamate ((S)-2b): Oil; yield: 2.7 g, 98%;
99.9% purity by HPLC; [a]2D0 =+138 (c=0.8 in CH2Cl2); 1HNMR:
(300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.3 (m, 5H; Ph), 5.4 (brs, 1H;
NH), 5.1 (s, 2H; PhCH2O), 4.3 (dd, J=6.7, 12.8 Hz, 1H; NHCH), 1.8, 1.6
(m, 2H; CHCH2), 1.4 (m, 2H; CH2CH3), 0.9 ppm (t, J=7.5 Hz, 3H;
CH2CH3); 13CNMR: (75 MHz, CDCl3): d=199.4 (CHO), 156.0
(OCONH), 136.0 (C ar), 128.5 (C ar), 128.2 (C ar), 128.0 (C ar), 67.0
(PhCH2O), 60.0 (CH), 31.1 (CHCH2), 18.3 (CH2CH3), 13.7 ppm (CH3).
(R)-4-Methyl-1-oxopentane-2-benzylcarbamate ((R)-2 f): Pale yellow oil;
yield: 4.1 g, 97%; 99.9% purity by HPLC; [a]2D0 =+30.88 (c=1.1 in
MeOH); HNMR: (300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.4 (m, 5H;
1
Ph), 5.2 (brs 1H; NH), 5.1 (s, 2H; PhCH2O), 4.3 (m, 1H; CHCHO), 1.7
(m, 2H; CH2), 1.4 (m, 1H; CHACHTNUGRTNEUNG
(CH3)2), 1.0 ppm (m, 6H; CH3); 13CNMR:
(75 MHz, CDCl3): d=199.6 (CHO), 158.3 (OCONH), 166.7 (C ar), 128.5
(C ar), 128.2 (C ar), 128.0 (C ar), 67.1 (PhCH2O), 58.7 (CHCHO), 38.0
(CH2), 23.0 (CHACHTNUTRGNE(UGN CH3)2), 24.5, 22.0 ppm (CH3).
(S)-1-Oxo-3-phenylpropan-2-benzylcarbamate ((S)-2g): Oil; yield: 3.4 g,
97%; 99.9% purity by HPLC; [a]2D0 =À53.08 (c=0.6 in DMF); 1HNMR:
(300 MHz, CDCl3): d=9.7 (s, 1H; CHO), 7.4 (m, 10H; Ph), 5.4 (d, J=
6.1 Hz, 1H; NH), 5.1 (s, 1H; PhCH2O), 4.5 (dd, J=6.8, 13.5 Hz, 1H;
PhCH2CH), 3.2 ppm (d, J=6.5 Hz, 2H; PhCH2CH); 13CNMR: (75 MHz,
CDCl3): d=199.2 (CHO), 155.2 (OCONH), 136.3 (C ar), 135.7 (C ar),
129.6 (C ar), 129.1 (C ar), 128.9 (C ar), 128.8 (C ar), 128.8 (C ar), 128.8
(C ar), 67.4 (CH), 61.4 (PhCH2O), 35.6 ppm (PhCH2CH).
(R)-1-Oxopentane-2-benzylcarbamate ((R)-2b): Oil; yield: 2.6 g, 97%;
99.9% purity by HPLC; [a]2D0 =À15.08 (c=1.0 in CH2Cl2); 1HNMR:
(300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.3 (m, 5H; Ph), 5.4 (brs, 1H;
NH), 5.1 (s, 2H; PhCH2O), 4.3 (dd, J=6.7, 12.8 Hz, 1H; NHCH), 1.8, 1.6
(m, 2H; CHCH2), 1.4 (m, 2H; CH2CH3), 0.9 ppm (t, J=7.5 Hz, 3H;
CH2CH3); 13CNMR: (75 MHz, CDCl3): d=199.4 (CHO), 156.0
(OCONH), 136.0 (C ar), 128.5 (C ar), 128.2 (C ar), 128.0 (C ar), 67.0
(PhCH2O), 60.0 (CH), 31.1 (CHCH2), 18.3 (CH2CH3), 13.7 ppm (CH3).
2-Methyl-1-oxopropane-2-benzylcarbamate (2h): Oil; yield: 3.6 g, 94%;
99.9% purity by HPLC; 1HNMR: (300 MHz, CDCl3): d=9.4 (s, 1H;
CHO), 7.3 (m, 5H; Ph), 5.4 (brs 1H; NH), 1.4 ppm (s, 6H; CH3);
13CNMR: (75 MHz, CDCl3): d=200.3 (CHO), 171.1 (OCONH), 155.1 (C
ar), 128.5 (C ar), 128.2 (C ar), 128.1 (C ar), 66.8 (PhCH2O), 59.4
(CCHO), 21.7 ppm (CH3).
(S)-1-Oxohexane-2-benzylcarbamate ((S)-2c): Oil; yield: 3.1 g, 97%;
99.9% purity by HPLC; [a]2D0 =À8.78 (c=0.5 in MeOH); 1HNMR:
(300 MHz, CDCl3): d=9.5 (s, 1H; CHO), 7.3 (m, 5H; Ph), 5.6 (brs, 1H;
NH), 5.1 (s, 2H; PhCH2O), 4.2 (m, 1H; CH), 1.8 (m, 1H; CHCH2), 1.6
(m, 1H; CHCH2), 1.3 (m, 4H; CHCH2CH2CH3), 0.9 ppm (t, J=6.9 Hz,
3H; CH3); 13CNMR: (75 MHz, CDCl3): d=199.7 (CHO), 156.1
(OCONH), 136.1 (C ar), 128.4 (C ar), 128.1 (C ar), 128.0 (C ar), 66.9
(PhCH2O), 60.1 (CHCHO), 28.6 (CHCH2), 27.1 (CHCH2CH2), 22.3
(CH3CH2), 13.7 ppm (CH3).
Enzymatic aldol condensation: Analytical scale reactions (2.5 mL total
volume) were conducted in 10 mL test tubes with screw caps at
[DHAP]=50–100 mm, [N-Cbz-aminoaldehyde]=85–170 mm depending
on the results obtained at analytical scale. Reactions at preparative scale
(15–20 mL total volume) were performed in 50 mL test tubes with screw
caps. The other variables were scaled up in proportion to the final
volume. As example, the procedure followed is described below.
(R)-1-Oxohexane-2-benzylcarbamate ((R)-2c): Oil; yield: 3.1 g, 97%;
99.9% purity by HPLC; [a]2D0 =+8.28 (c=0.7 in MeOH); 1HNMR:
(300 MHz, CDCl3): d=9.5 (s, 1H; CHO), 7.3 (m, 5H; Ph), 5.6 (brs, 1H;
NH), 5.1 (s, 2H; PhCH2O), 4.2 (m, 1H; CH), 1.8 (m, 1H; CHCH2), 1.6
(m, 1H; CHCH2), 1.3 (m, 4H; CHCH2CH2CH3), 0.9 ppm (t, J=6.9 Hz,
3H; CH3); 13CNMR: (75 MHz, CDCl3): d=199.7 (CHO), 156.1
(OCONH), 136.1 (C ar), 128.4 (C ar), 128.1 (C ar), 128.0 (C ar), 66.9
(PhCH2O), 60.1 (CHCHO), 28.6 (CHCH2), 27.1 (CHCH2CH2), 22.3
(CH3CH2), 13.7 ppm (CH3).
Enzymatic aldol condensations in emulsions: (S)- or (R)-N-Cbz-aminoal-
dehyde (0.13–4.8 mmol, 1.7 equiv per mol DHAP), the oil (6% w/w),
and the surfactant (4% w/w) were mixed with a vortex mixer. Then, the
DHAP solution (volume corresponding to 90% w/w of the mixture,
0.075–3 mmol) at pH 6.9, freshly prepared as described by Effenberger
et al.,[70] was added dropwise while stirring at 258C or 48C depending on
the experiment with a vortex mixer. The reactions were started by
adding RAMA (8 UmLÀ1 reaction mixture), RhuA (2 UmLÀ1 reaction
mixture), or FucA (8 UmLÀ1 reaction mixture) and mixed again. The test
tubes were placed on a horizontal shaking bath (100 rpm) at constant
temperature (4 or 258C depending on the experiment). The reactions
were followed by HPLC, as indicated above, until the peak of the prod-
uct reached a maximum.
(S)-3-Methyl-1-oxobutane-2-benzylcarbamate ((S)-2d): Pale yellow oil;
yield: 4.2 g, 95%; 99.9% purity by HPLC; [a]2D0 =+8.28 (c=1.7 in
1
MeOH); HNMR: (300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.4 (m, 5H;
Ph), 5.4 (brs 1H; NH), 5.1 (s, 2H; PhCH2O), 4.3 (m, 1H; CHCHO), 2.3
(m, 1H; CHACHTUNGTRENNUNG(CH3)2), 1.0 (d, J=6.9 Hz, 3H; CH3), 0.9 ppm (d, J=7.0 Hz,
3H; CH3); 13CNMR: (75 MHz, CDCl3): d=199.7 (CHO), 156.3
(OCONH), 136.1 (C ar), 128.5 (C ar), 128.2 (C ar), 128.0 (C ar), 67.0
(PhCH2O), 64.9 (CHCHO), 29.0 (CHACHTUNTRGNEUNG(CH3)2), 18.9, 17.4 ppm ((CH3)2).
Enzymatic aldol condensations in mixtures water/dimethylformamide 4:1:
(S)- or (R)-N-Cbz-aminoaldehyde (0.13–4.8 mmol, 1.7 equiv per mol
DHAP) was dissolved in DMF (the amount corresponding to 20% v/v of
the total). Then, the DHAP solution (volume corresponding to 80% v/v
of the total, 0.075–3 mmol) at pH 6.9, freshly prepared as described
above was added dropwise while stirring at 25 or 48C, depending on the
experiment, with a vortex mixer. Finally, RAMA (8 UmLÀ1 reaction mix-
ture), RhuA (2 UmLÀ1 reaction mixture), or FucA (8 UmLÀ1 reaction
mixture) was added and mixed again. The rest of the experimental proce-
dure was identical to that described for the reaction in emulsions.
(R)-3-Methyl-1-oxobutane-2-benzylcarbamate ((R)-2d): Pale yellow oil;
yield: 4 g, 96%; 99.9% purity by HPLC; [a]2D0 =À8.58 (c=1.3 in MeOH);
1HNMR: (300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.4 (m, 5H; Ph), 5.4
(brs 1H; NH), 5.1 (s, 2H; PhCH2O), 4.3 (m, 1H; CHCHO), 2.3 (m, 1H;
CHACHTUNGTRENNUNG(CH3)2), 1.0 (d, J=6.9 Hz, 3H; CH3), 0.9 ppm (d, J=7.0 Hz, 3H;
CH3); 13CNMR: (75 MHz, CDCl3): d=199.7 (CHO), 156.3 (OCONH),
136.1 (C ar), 128.5 (C ar), 128.2 (C ar), 128.0 (C ar), 67.0 (PhCH2O), 64.9
(CHCHO), 29.0 (CHACHTUNGTRENNUNG(CH3)2), 18.9, 17.4 ppm ((CH3)2).
ACHTUNGTRENNUNG(2S,3S)-3-Methyl-1-oxopentane-2-benzylcarbamate ((S)-2e): Pale yellow
oil; yield: 2.4 g, 95%; 99.9% purity by HPLC; [a]2D0 =+5.88 (c=1.2 in
The enzymatic reactions were stopped by addition of MeOH
(1.5xreaction volume). Then, the methanol was evaporated and the aque-
ous solution washed with ethyl acetate to remove the unreacted N-pro-
tected aminoaldehyde. The aqueous layer was collected, the remaining
ethyl acetate removed under reduced pressure, and lyophilized. The solid
obtained was dissolved in plain water (ca. 10–20 mL) and the pH was ad-
justed to 5.5 either with TFA or with HCl. To this solution, acid phospha-
tase (5.3 UmmolÀ1 phosphated adduct) was added. The reaction was fol-
lowed by HPLC until no starting material was detected. Further phos-
phatase units were added, if necessary, to direct the reaction to comple-
tion. Then, the reaction mixture was filtered through a 0.45 mm cellulose
membrane filter. The filtrate was loaded onto a Perkin–Elmer semipre-
parative 250ꢊ25 mm column, filled with C18 (10 mm) and eluted with a
gradient of CH3CN (0 to 48% over 30 min) in plain water. Pure fractions
were pooled and lyophilized.
1
MeOH); HNMR: (300 MHz, CDCl3): d=9.7 (s, 1H; CHO), 7.4 (m, 5H;
Ph), 5.4 (brd, J=6.8 Hz, 1H; NH), 4.4 (m, 1H; CHCHO), 1.5, 1.3 (m,
2H; CH2CH3), 1.2 (m, 1H; CHCH3), 1.0 ppm (m, 6H; CH3); 13CNMR:
(75 MHz, CDCl3): d=200.0 (CHO), 156.2 (OCONH), 136.1 (C ar), 128.5
(C ar), 128.2 (C ar), 128.1 (C ar), 67.0 (PhCH2O), 64.6 (CHCHO), 36.3
(CHCH3), 25.2 (CH2CH3), 15.5 (CHCH3), 11.9 ppm (CH2CH3).
(S)-4-Methyl-1-oxopentane-2-benzylcarbamate ((S)-2 f): Pale yellow oil;
yield: 4.1 g, 97%; 99.9% purity by HPLC; [a]2D0 =À33.58 (c=1.1 in
1
MeOH); HNMR: (300 MHz, CDCl3): d=9.6 (s, 1H; CHO), 7.4 (m, 5H;
Ph), 5.2 (brs 1H; NH), 5.1 (s, 2H; PhCH2O), 4.3 (m, 1H; CHCHO), 1.7
(m, 2H; CH2), 1.4 (m, 1H; CHACHTNUGRTNEUNG
(CH3)2), 1.0 ppm (m, 6H; CH3); 13CNMR:
(75 MHz, CDCl3): d=199.6 (CHO), 158.3 (OCONH), 166.7 (C ar), 128.5
(C ar), 128.2 (C ar), 128.0 (C ar), 67.1 (PhCH2O), 58.7 (CHCHO), 38.0
(CH2), 23.0 (CHACHTUNGTRENNUNG(CH3)2), 24.5, 22.0 ppm (CH3).
7322
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
Removal of Cbz group and reductive amination: The aldol adduct ob-
tained (0.011–0.24 mmol) was dissolved in ethanol (5–10 mL) followed
with the addition of plain water (20–45 mL). Pd/C (200 mg) was added to
this solution. The reaction mixture was shaken under hydrogen gas
(50 psi) overnight at room temperature. After removal of the catalyst by
filtration through neutralized and deactivated aluminum, the pH of the
filtrate was adjusted to pH 5.5 with HCl, and the solvent was evaporated
under reduced pressure and then lyophilized. Characterization of the
products from the lyophilisates was accomplished by NMR spectroscopy.
Physical, 1H and 13C NMR data of the compounds synthesized: For the
sake of clarity in the systematic nomenclature, we always assign the hy-
droxymethyl group at position C-2. For detailed assignments of H and C
in the spectra see Supporting Information. Entry numbers correspond to
the entries in Table 1.
1H), 0.97 ppm (m, 3H overlapped); 13C NMR (101 MHz, D2O): d=75.1
(C4), 72.0 (C3), 72.4 (C5), 61.0 (C2), 60.1 (C6), 52.9 (C9,9’), 24.9 (C7),
11.5 (C10), 11.3 ppm (C8). ESI-TOF: m/z calcd for [M+H]+ C9H20NO3:
190.1437; found: 190.1433.
Entry 4—(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-ethyl-3,4-diol
(4e): Yield: 0.137 mmol, 27 mg, 20%; [a]2D0 =+47.68 (c=0.8 in MeOH);
1H NMR (500 MHz, D2O): d=4.28 (t, J=3.5 Hz, 1H), 4.09 (dd, 3J-
ACHUTNGRENNUG CAHTUNGTRENNUNG
(H,H)=9.3, 3.7 Hz, 1H), 3.96 (A of AB system, 3J
1H), 3.85 (A of AB system, 3J
N
3
3.42 (dt, J
N
ACTHNUTRGNEUNG
(t, 3J(H,H)=7.5 Hz, 3H); 13C NMR (101 MHz, D2O): d=77.8 (C4), 72.9
(C3), 64.8 (C2), 64.1 (C5), 60.4 (C6), 26.1 (C7), 12.8 ppm (C8); ESI-TOF:
m/z calcd for [M+H]+ C7H16NO3: 162.1130; found: 162.1126.
Entry 5—(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(4h): Yield: 0.708 mmol, 150 mg, 67%; [a]2D0 =À40.68 (c=1.0 in MeOH);
Entry 1—(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-ethyl-3,4-diol
(4a): Yield: 0.222 mmol, 44 mg, 63%; [a]2D0 =À36.78 (c=1 in MeOH);
1H NMR (400 MHz, D2O): d=4.01 (t, 3J
ACTHNUTRGNE(UNG H,H)=7.2 Hz, 1H), 3.91 (m,
3
1H NMR (500.13 MHz, D2O): d=4.03 (t, J
G
1H), 3.86 (A of AB system, 1H), 3.81 (B of AB system, 3J
5. 9 Hz, 1H), 3.51 (m, 1H), 3.38 (m, 1H), 1.82 (m, 1H), 1.70 (m, 1H),
1.41 (m, 2H), 0.90 ppm (t, 3J(H,H)=7.3 Hz, 3H); 13C NMR (101 MHz,
ACHTUNGTRENUN(NG H,H)=12.7,
1H), 3.90 (A of AB system, 3J
N
3
system, J
A
AHCTUNGTRENNUNG
1H), 1.75 (m, 1H), 1.03 ppm (t, 3J
ACHTUNGTRENNUNG
D2O): d=80.8 (C4), 76.9 (C3), 64.8 (C5), 64.0 (C2), 60.6 (C6), 34.9 (C7),
21.4 (C8), 15.6 ppm (C9); ESI-TOF: m/z calcd for [M+H]+ C8H18NO3:
176.1287; found: 176.1281.
(101 MHz, D2O): d=80.6 (C4), 76.9 (C3), 65.7 (C5), 64.8 (C2), 60.6 (C6),
26.2 (C7), 12.4 ppm (C8); ESI-TOF: m/z calcd for [M+H]+ C7H16NO3:
162.1130; found: 162.1125.
Entry 6—(2S,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(4i; 86%), (2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(4h; 4%), and (2R,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-
diol (4j; 10%): Overall yield: 0.103 mmol, 22 mg, 21%; [a]2D0 =À28.38
(c=0.3 in MeOH).
Entry 2—(2R,3S,4R,5S)-2-(Hydroxymethyl)pyrrolidine-5-ethyl-3,4-diol
(4b; 86%), (2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-ethyl-3,4-diol
(4a; 10%), and (2R,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-ethyl-3,4-
diol (4c; 4%): Overall yield: 0.101 mmol, 20 mg, 28%; [a]2D0 =+14.38
(c=0.9 in MeOH).
Data for 4i: 1H NMR (500 MHz, D2O): d=4.28 (m, 1H), 4.27 (s, 1H),
Data for 4b: 1H NMR (500 MHz, D2O): d=4.23 (m, 1H), 4.22 (m, 1H),
3.96 (A of AB system, 3J
ACHTUNGTRENNUNG
3.91 (A of AB system, 3J
G
system, 3J
G
ACHTUNGTRENNUNG
3
3
system, J
G
3
AHCTUNGTREG(NNNU H,H)=7.2 Hz, 3H).
1.5 Hz, 1H), 1.96–1.66 (m, 2H), 0.95 ppm (t, 3J
ACHTUNGTRENNUNG
13C NMR (101 MHz, D2O): d=74.1 (C4), 73.3 (C3), 64.2 (C5), 64.2 (C2),
61.1 (C6), 30.8 (C7), 21.4 (C8), 15.7 ppm (C9). ESI-TOF: m/z calcd for
[M+H]+ C8H18NO3: 176.1287; found: 176.1281.
13C NMR (101 MHz, D2O): d=74.1 (C4), 73.1 (C3), 65.9 (C5), 64.2(C2),
61.0 (C6), 22.2 (C7), 12.4 ppm (C8). ESI-TOF: m/z calcd for [M+H]+
C7H16NO3: 162.1130; found: 162.1125.
Data for 4h: See the main 1H and 13C NMR chemical shifts listed for
entry 5.
1
Data for 4a: See H and 13C NMR of entry 1.
Data for 4c: The main 1H and 13C NMR chemical shifts are similar to
those of entry 6 (4j). ESI-TOF: m/z calcd for [M+H]+ C7H16NO3:
162.1130; found: 162.1127.
Data for 4j: 1H NMR (500 MHz, D2O d ppm 4.53 (dd, 3J
1H), 4.33 (t, 3J(H,H)=4.4 Hz, 1H), 3.94 (A of AB system, 3J
12.1, 4.2 Hz, 1H), 3.85 (B of AB system, 3J
(H,H)=12.1, 8.5 Hz, 1H),
3.78 (m, 1H), 3.54 (m, 1H), 1.81 (m, 1H), 1.71 (m, 1H), 1.41 (m, 1H),
0.95 ppm (t, 3J(H,H)=7.4 Hz, 1H); 13C NMR (101 MHz, D2O): d=72.7
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
Entry 3—(2S,3S,4S,5R)-2-(hydroxymethyl)pyrrolidine-5-ethyl-3,4-diol
(4d; 48%), (2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-ethyl-3,4-diol
(4e; 30%), N-ethyl-(2S,3S,4S,5R)-2-(hydroxymethyl)pyrrolidine-5-ethyl-
3,4-diol (4 f; 11%), and N-ethyl-(2R,3S,4R,5R)-2-(hydroxymethyl)pyrro-
lidine-5-ethyl-3,4-diol (4g; 11%): Overall yield: 0.152 mmol, 30 mg,
61%; [a]2D0 =À12.48 (c=1.0 MeOH).
AHCTUNGTRENNUNG
(C4), 72.7 (C3), 62.9 (C5), 62.9 (C2), 60.2 (C6), 30.1 (C7), 21.1 (C8),
15.3 ppm (C9). ESI-TOF: m/z calcd for [M+H]+ C8H18NO3: 176.1287;
found: 176.1284.
Entry 7—(2S,3S,4S,5R)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(4k;44%), (2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(4l; 34%), N-ethyl-(2S,3S,4S,5R)-2-(hydroxymethyl)pyrrolidine-5-propyl-
3,4-diol (4m; 11%), and N-ethyl-(2R,3S,4R,5R)-2-(hydroxymethyl)pyrro-
lidine-5-propyl-3,4-diol (4n; 10%): Overall yield: 0.180 mmol, 38 mg,
78%; [a]2D0 =+15.78 (c=1.0 in MeOH).
Data for 4d: 1H NMR (500 MHz, D2O): d=4.15 (m, 1H), 4.04 (m, 1H),
3.84 (A of AB system, 3J
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
(C3), 77.6 (C4), 70.2 (C2), 67.1 (C5), 62.4 (C6), 21.2 (C7), 12.7 ppm (C8).
ESI-TOF: m/z calcd for [M+H]+ C7H16NO3: 162.1130; found: 162.1125.
Data for 4k: 1H NMR (500 MHz, D2O): d=4.14 (brd, 1H), 4.06 (m,
Data for 4e: The main 1H and 13C NMR chemical shifts are listed in
entry 4. ESI-TOF: m/z calcd for [M+H]+ C7H16NO3: 162.1130; found:
162.1125.
1H), 3.97 (A of AB system, 3J
ACHTUNGTRENNUNG
system, 3J(H,H)=7.3, 2.8 Hz, 1H), 3.78 (dd, 3J
ACTHNUGTRENNUGN ACHTUNGTRENNNUG
3.55 (m, 1H), 1.75 (m, 2H), 1.43 (m, 2H), 0.94 ppm (m, 3H). 13C NMR
(101 MHz, D2O): d=78.8 (C3), 78.0 (C4), 72.2 (C2), 65.3 (C5), 63.0 (C6),
29.7 (C7), 21.8 (C8), 15.8 ppm (C9).
Data for 4 f: 1H NMR (500 MHz, D2O): d=4.30 (d, 3J
1H), 4.19 (s, 1H), 3.96 (d, 3J
(H,H)=1.2 Hz, 1H), 3.95 (s, 1H), 3.70 (m,
1H), 3.61 (t, 3J(H,H)=6.5 Hz, 1H), 3.53 (A of AB system, 3J
(H,H)=
14.6, 7.4 Hz, 1H), 3.31 (B of AB system, 3J
(H,H)=14.5, 7.3 Hz, 1H),
1.87 (m, 1H), 1.35 (t, 3J(H,H)=7.3 Hz, 1H), 1.04 ppm (t, 3J
(H,H)=
ACHTUNGTREN(NUNG H,H)=2.8 Hz,
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
Data for 4l: See the 1H and 13C NMR chemical shifts listed in entry 8.
Data for 4m: The main 1H and 13C NMR chemical shifts are coincident
AHCTUNGTRENNUNG
A
ACHTUNGTRENNUNG
7.5 Hz, 1H); 13C NMR (101 MHz, D2O): d=77.4 (C3), 75.9 (C4), 76.0
(C2), 72.0 (C5), 61.7 (C6), 50.7 (C9,9’), 19.3 (C7), 12.3 (C8), 10.3 ppm
(C10). ESI-TOF: m/z calcd for [M+H]+ C9H20NO3: 190.1437; found:
190.1432.
to those for 4 f.
Data for 4n: The main H and 13C NMR chemical shifts are coincident to
1
those for 4g.
Entry 8—(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(4l): Yield: 0.136 mmol, 29 mg, 28%; [a]2D0 =+62.38 (c=1.0 in MeOH);
1H NMR (500 MHz, D2O): d=4.28 (t, J=3.4 Hz, 1H), 4.08 (dd, 3J-
1
Data for 4g (free base): H NMR (500 MHz, D2O): d=4.21 (m, 1H), 3.95
(m, 1H), 3.85 (A of an AB system, 3J
ACHTUNGTRENNUNG
ACHUTNGRENNUG CAHTUNGTRENNUNG
(H,H)=9.3, 3.8 Hz, 1H), 3.96 (A of AB system, 3J
N
1H), 3.85 (B of AB system, 3J
ACHTUNGTRENNUNG
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7323

J. Bujons, P. Clapꢂs et al.
3
3.50 (dt, J
N
ACTHNGUTERNNUG
1H), 1.04 ppm (dd, 3J(H,H)=12.9, 6.8 Hz, 6H); 13C NMR (101 MHz,
D2O): d=79.3 (C4), 77.1 (C3), 69.6 (C5), 64.7 (C2), 60.1 (C6), 32.2 (C7),
21.1 (C8), 20.6 ppm (C8’). ESI-TOF: m/z calcd for [M+H]+ C9H20NO3:
190.1443; found: 190.1437.
2H), 0.93 ppm (t, 3J
78.0 (C4), 72.9 (C3), 64.1 (C2), 63.0 (C5), 60.3 (C6), 34.9 (C7), 21.8 (C8),
15.6 ppm (C9); ESI-TOF: calcd for [M+H]+ C8H18NO3: 176.1287;
found: 176.1282.
A
Data for 4w: The main 1H and 13C NMR chemical shifts similar to those
Entry 9—(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-butyl-3,4-diol
(4o; 93%) and (2S,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-butyl-3,4-
diol (4p; 7%): Overall yield: 0.155 mmol, 35 mg, 62%; [a]2D0 =À27.1 (c=
0.9 in MeOH).
for the two conformations of 4ac (entry 16).
Entry 14—(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-isopropyl-3,4-
diol (4x): Yield: 0.093 mmol, 21 mg, 50%; [a]2D0 =+38.0 (c=0.7 in
MeOH); 1H NMR (500 MHz,D2O): d=4.26 (m, 1H), 4.24 (m, 1H), 3.95
Data for 4o: 1H NMR (500 MHz, D2O): d=4.02 (t, 3J
1H), 3.93 (m, 1H), 3.90 (A of AB system, 3J
(H,H)=12.6, 3.8 Hz, 1H),
3.83 (B of AB system, 3J
(H,H)=12.6, 6.1 Hz, 1H), 3.53 (m, 1H), 3.38 (m,
1H), 1.86 (m, 1H), 1.72 (m, 1H), 1.35 (m, 4H), 0.86 ppm (t, 3J
(H,H)=
ACHTUNGTREN(NUNG H,H)=6.9 Hz,
ACTHNUTRGNEUNG
(A of AB system, 3J(H,H)=12.1, 5.3 Hz, 1H), 3.85 (B of AB system, 3J-
(H,H)=12.1, 8.1 Hz, 1H), 3.70 (m, 1H), 3.27 (t, 3J
2.02 (m, 1H), 1.03 ppm (t, 3J(H,H)=7.2 Hz, 6H). 13C NMR (101 MHz,
ACHTUNGTREN(NUGN H,H)=8.5 Hz, 1H),
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
D2O): d=76.2 (C4), 73.4 (C3), 68.7 (C5), 64.1 (C2), 60.2 (C6), 32.0 (C7),
21.4 (C8), 20.5 ppm (C8’); ESI-TOF: m/z calcd for [M+H]+ C8H18NO3:
176.1287; found: 176.1280.
7.2 Hz, 3H); 13C NMR (101 MHz, D2O): d=80.8 (C4), 77.0 (C5), 64.8
(C2), 64.2 (C5), 60.6 (C6), 32.5 (C7), 30.0 (C8), 24.3 (C9), 15.7 ppm
(C10); ESI-TOF: m/z calcd for [M+H]+ C9H20NO3: 190.1443; found:
190.1437.
Entry 15—(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-((S)-sec-butyl)-
3,4-diol (4y; 55%) and N-ethyl-(2S,3S,4S,5S)-2-(hydroxymethyl)pyrroli-
dine-5-((S)-sec-butyl)-3,4-diol (4z; 45%): Overall yield: 0.151 mmol,
34 mg, 50%; [a]2D0 =À22.98 (c=1.0 in MeOH).
Data for 4p: The 1H and 13C NMR chemical shifts listed in entry 10.
Entry 10—(2S,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-butyl-3,4-diol
(4p; 80%), (2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-butyl-3,4-diol
(4o; 10%), and (2R,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-butyl-3,4-
diol (4q; 10%): Overall yield: 0.088 mmol, 20 mg, 18%; [a]2D0 =À28.38
(c=0.3 in MeOH).
Data for 4y: 1H NMR (500 MHz, D2O): d=4.08 (t, 3J
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGERTG(NNUN H,H)=7.9 Hz, 1H), 1.86 (m, 1H), 1.54 (m, 1H), 1.23 (m, 1H), 1.02 (m,
3H), 0.90 ppm (m, 3H). 13C NMR (101 MHz, D2O): d=79.0 (C4), 77.1
(C3), 68.0 (C5), 64.7 (C2), 60.0 (C6), 38.5 (C7), 28.1 (C8), 16.6 (C10),
12.7 ppm (C9); ESI-TOF: m/z calcd for [M+H]+ C9H20NO3: 190.1443;
found: 190.1434.
Data for 4p: 1H NMR (500 MHz, D2O): d=4.24 (m, 1H), 4.22 (m, 1H),
3.92 (A of AB system, 3J
G
3
system, J
G
1H), 1.71 (m, 1H), 1.33 (m, 4H), 0.86 ppm (t, 3J
ACTHNUTRGNE(NUG H,H)=7.1 Hz, 3H).
13C NMR (101 MHz, D2O): d=74.1 (C3), 73.4 (C4), 64.4 (C5), 64.2 (C2),
Data for 4z: Two conformations were observed—major conformation I:
1H NMR (500 MHz, D2O): d=4.24 (dd, 3J
ACHTUNGTRENNUNG
61.0 (C6), 30.0 (C8), 28.4 (C7), 24.5 (C9), 15.7 ppm (C10).
Data for 4o: The 1H and 13C NMR chemical shifts listed in entry 9.
(t, 3J
1H), 3.52 (t, JACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
3
1
Data for 4q: The main H and 13C NMR chemical shifts are coincident to
AHCTUNGTRENNUNG
those for entry 6 (4j).
A
ACHTUNGTRENNUNG
Entry 11—(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(4r; 63%), (2S,3S,4S,5R)-2-(hydroxymethyl)pyrrolidine-5-propyl-3,4-diol
(101 MHz, D2O): d=77.6 (C4), 77.6 (C3), 77.3 (C5), 71.0 (C2), 58.9 (C6),
49.6 (C11), 37.6 (C7), 29.3 (C8), 16.1 (C10), 13.8 (C9), 12.7 ppm (C12);
minor conformation II: selected signals 1H NMR (500 MHz, D2O): d=
4.29 (brs, 1H), 4.09 (brs, 1H), 3.96–3.88 (m, 2H), 3.58 (m, 1H), 3.56 (m,
1H), 3.38–3.27 (m, 1H), 2.08 (m, 1H), 1.38 (m, 3H), 1.37 (m, 1H), 1.33
(m, 1H), 1.10 ppm (m, 6H); ESI-TOF: m/z calcd for [M+H]+
C11H24NO3: 218.1756, found 218.1752.
(4s;
28%),
N-ethyl-(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-
propyl-3,4-diol (4t; 5%), and N-ethyl-(2S,3S,4S,5R)-2-(hydroxymethyl)-
pyrrolidine-5-propyl-3,4-diol (4u; 5%): Overall yield: 0.181 mmol, 41 mg,
67%; [a]2D0 =+19.68 (c=0.9 in MeOH).
Data for 4r: 1H NMR (500 MHz, D2O): d=4.28 (t, J=3.5 Hz, 1H), 4.07
(dd, 3J(H,H)=9.2, 3.9 Hz, 1H), 3.95 (A of AB system, 3J
ACHTUNGTRENNUNG ACHTUNGTRENNUNG
Entry 16—(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-isobutyl-3,4-diol
(4aa; 71%), (2S,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-isobutyl-3,4-
diol (4ab; 6%), and N-ethyl-(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-
5-isobutyl-3,4-diol (4ac; 23%): Overall yield: 0.129 mmol, 29 mg, 12%;
[a]2D0 =À39.3 (c=1.0 in MeOH).
5.1 Hz, 1H), 3.85 (B of AB system, 3J
1H), 3.47 (dt, 3J
0.86 ppm (t, 3J
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
(C4), 72.9 (C3), 64.1 (C2), 63.3 (C5), 60.3 (C6), 32.5 (C7), 30.5 (C8), 24.4
(C9), 15.7 ppm (C10).
Data for 4aa: 1H NMR (500 MHz, D2O): d=4.03 (t, 3J
ACHTUNGTRENNUNG
Data for 4s: 1H NMR (500 MHz, D2O): d=4.13 (brd, 1H), 4.04 (brs,
1H), 3.92 (m, 1H), 3.89 (A of AB system, 3J
3.83 (B of AB system, 3J
ACHTUNGTRENNUNG
1H), 3.91 (A of AB system, 3J
N
AHCTUNGTRENNUNG
3
system, J
A
1H), 1.74 (m, 1H), 1.67 (m, 1H), 1.63 (m, 1H), 0.92 ppm (m, 6H);
13C NMR (101 MHz, D2O): d=81.8 (C4), 76.9 (C3), 64.8 (C2), 62.4 (C5),
60.6 (C6), 42.0 (C7), 27.1 (C8), 24.7 (C9), 23.6 ppm (C9’). ESI-TOF: m/z
calcd for [M+H]+ C9H20NO3: 190.1443; found: 190.1444.
2H), 1.34 (m, 4H), 0.86 ppm (t, 3J
ACHTUNGTRENNUNG
(101 MHz, D2O): d=78.8 (C3), 77.9 (C4), 70.2 (C2), 65.6 (C5), 62.4 (C6),
27.4 (C7), 24.5 (C8), 24.5 (C9), 15.7 ppm (C10).
1
Data for 4t: The main H and 13C NMR chemical shifts are coincident to
1
Data for 4ab: H NMR (500 MHz, D2O): d=4.29 (m, 1H), 4.26 (m, 1H),
those for 4g.
3.96 (A of AB system, 3J
ACHTUNGTRENNUNG
Data for 4u: The main H and 13C NMR chemical shifts are coincident to
1
system, 3J
(ddd, 3J
N
ACHTUNGTRENNUNG
those for 4 f.
AHCTUNGTRENNUNG
Entry 12—(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-butyl-3,4-diol
(4r): Yield: 0.087 mmol, 20 mg, 19%; [a]2D0 =+53.28 (c=1.0 in MeOH);
AHCTUNGTRENNUNG
d=74.2 (C4), 73.5 (C3), 64.1 (C2), 62.7 (C5), 61.0 (C6), 37.5 (C7), 27.1
(C8), 24.3 (C9’), 24.2 ppm (C9). ESI-TOF: m/z calcd for [M+H]+
C9H20NO3: 190.1443; found: 190.1439.
1
the H and 13C NMR chemical shifts are coincident to those for entry 11;
ESI-TOF: calcd for [M+H]+ C9H20NO3: 190.1443, found 190.1439.
Entry 13—(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-isopropyl-3,4-
diol (4v; 84%), N-ethyl-(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-
isopropyl-3,4-diol (4w; 16%): Overall yield: 0.217 mmol, 46 mg, 45%;
[a]2D0 =À36.3 (c=1.0 in MeOH).
Data for 4ac: Two conformations were observed: major conformation I:
1H NMR (500 MHz, D2O): d=4.25 (brs, 1H), 4.15 (brs, 1H), 4.01 (m,
2H), 3.76 (m, 1H), 3.56 (m, 1H), 3.44 (m, 1H), 3.35 (m, 1H), 1.83 (m,
1H), 1.78 (m, 1H), 1.69 (m, 1H), 1.36 (t, 3J
ACHTUNGTRENNUNG
1
Data for 4v: H NMR (500 MHz, D2O): d=4.03 (m, 2H), 3.88 (A of AB
1.02 ppm (t, 3J
ACHTUNGTRENNUNG
system, 3J
G
N
(C4), 79.0 (C3), 73.6 (C2), 69.2 (C5), 60.3 (C6), 47.2 (C10), 35.2 (C7),
26.9 (C8), 25.1 (C9’), 23.2 (C9), 12.2 ppm (C11); minor conformation II:
12.7, 5.8 Hz, 1H), 3.49 (m, 1H), 3.18 (t, 3J
A
7324
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
d=ppm 4.33 (brs, 1H), 4.10 (brs, 1H), 4.08–4.03 (m, 2H), 3.83 (brs,
Data for 4ak: 1H NMR (500 MHz, D2O): d=4.06 (dd, 3J
5.8 Hz, 1H), 3.89 (A of AB system, 3J
(H,H)=16.4, 3.9 Hz, 1H), 3.89 (s,
1H), 3.81 (B of AB system, 3J(H,H)=12.6, 6.7 Hz, 1H), 3.55 (dt, 3J-
(H,H)=7.1, 3.9 Hz, 1H), 1.48 (s, 3H), 1.36 ppm (s, 3H); 13C NMR
(101 MHz, D2O): d=83.2 (C4), 77.3 (C3), 67.5 (C5), 64.6 (C2), 61.1 (C6),
26.2 (C7), 22.6 ppm (C7’); ESI-TOF: m/z calcd for [M+H]+ C7H16NO3:
162.1130; found: 162.1120.
ACHTUNGTRENUN(NG H,H)=7.5,
1H), 3.52 (m, 1H), 3.44 (m, 2H), 1.82 (m, 1H), 1.78 (m, 1H), 1.69 (m,
ACHTUNGTRENNUNG
1H), 1.41 (t, 3J
N
G
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
Entry 17—(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-isobutyl-3,4-
diol (4ad; 43%), (2S,3S,4S,5R)-2-(hydroxymethyl)pyrrolidine-5-isobutyl-
3,4-diol (4ae; 32%), N-ethyl-(2R,3S,4R,5R)-2-(hydroxymethyl)pyrroli-
dine-5-isobutyl-3,4-diol (4af; 9%), and N-ethyl-(2S,3S,4S,5R)-2-
Data for 4al: 1H NMR (500 MHz, D2O): d=4.04 (dd, 3J
ACHTUNGTRENNUNG
1.9 Hz, 1H), 3.95 (m, 1H), 3.94–3.85 (m, 2H), 3.47 (brq, 3J
AHCTUNGTRENNUNG
5.3 Hz, 1H), 1.41 (s, 3H), 1.38 ppm (s, 3H); 13C NMR (101 MHz, D2O):
d=83.2 (C4), 78.5 (C3), 75.5 (C2), 67.5 (C5), 60.6 (C6), 21.0 (C7),
20.7 ppm (C7’).
(hydroxyACHTUNGTRENNUNGmethyl)pyrrolidine-5-isobutyl-3,4-diol (4ag; 16%): Reaction at
258C; overall yield: 0.067 mmol, 15 mg, 50%.
Entry 22—(2R,3S,4R)-2-(hydroxymethyl)pyrrolidine-5-dimetil-3,4-diol
(4am) and (2S,3S,4R)-2-(hydroxymethyl)pyrrolidine-5-dimetil-3,4-diol
(4al): Yield: 0.136 mmol, 27 mg, 12%; [a]2D0 =+58 (c=1 in MeOH).
Data for 4ad: 1H NMR (500 MHz, D2O): d=4.27 (t, 3J
ACTHNUGRTENUNG(H,H)=3.5 Hz,
ACHTUNGTRENNUNG
1H), 4.05 (dd, 3J(H,H)=9.2, 3.8 Hz, 1H), 3.95 (A of AB system, 3J-
ACHTUNGTRENNUNG ACHTUNGTERN(NUGN H,H)=12.1, 8.3 Hz,
(H,H)=12.1, 5.1 Hz, 1H), 3.85 (B of AB system, 3J
Data for 4am: 1H NMR (500 MHz, D2O): d=4.39 (t, 3J
N
3
1H), 4.33 (dd, 3J
(H,H)=4.5, 8.4 Hz, 1H), 3.93 (d, J
G
1H), 3.74 (m, 1H), 3.55 (m, 1H), 1.65 (m, 3H), 0.92 ppm (m, 6H);
13C NMR (101 MHz, D2O): d=78.3 (C4), 72.7 (C3), 64.2 (C2), 61.4 (C5),
60.3 (C6), 42.0 (C7), 27.6 (C8), 24.8 (C9), 23.5 ppm (C9’).
Data for 4ae: 1H NMR (500 MHz, D2O): d=4.15 (s, 1H), 4.07 (s, 1H),
3.90–3.78 (m, 2H), 3.80 (m, 1H), 3.54 (m, 1H), 1.65 (m, 3H), 0.94 ppm
(t, J=6.4 Hz, 6H); 13C NMR (101 MHz, D2O): d=78.6 (C3), 78.0 (C4),
61.6 (C2), 63.8 (C5), 60.1 (C6), 42.1 (C7), 27.5 (C8), 24.5 (C9), 24.5 ppm
(C9’).
3.82 (A of AB system, 1H), 3.73 (B of AB system, 1H), 3.70 (m, 1H),
1.32 (s, 3H), 1.29 ppm (s, 3H); 13C NMR (101 MHz, D2O): d=79.31
(C4), 72.9 (C3), 67.5 (C5), 62.2 (C2), 60.3 (C6), 26.8 (C7), 23.6 ppm
(C7’); ESI-TOF: m/z calcd for [M+H]+ C7H16NO3: 162.1130; found:
162.1124.
Data for 4al: The chemical shifts differ slightly from those of entry 21:
1H NMR (500 MHz, D2O): d=4.33 (dd, J=4.5, 8.9 Hz, 1H), 3.84 (m,
1H), 3.78–3.70 (m, 2H), 3.50 (m, 1H), 1.33 (s, 3H), 1.28 ppm (s, 3H);
13C NMR (101 MHz, D2O): d=78.7 (C4), 72.9 (C3), 69.2 (C5), 65.2 (C2),
60.7 (C6), 26.1 (C7), 22.8 ppm (C7’).
Data for 4af: The 1H and 13C NMR chemical shifts are coincident to
1
those for 4g; selected signals: H NMR: d=4.32 (brs, 1H), 3.46 (m, 1H),
3
3.26 (m, 1H), 1.28 ppm (t, J
N
Data for 4ag: The 1H and 13C NMR chemical shifts are coincident to
Purification by ion exchange chromatography: The mixtures of entries 3,
6, 15, and 16 (10–20 mg, Table 2) were separated by ion exchange chro-
matography on a FPLC system following a method described by Asano
1
those for 4 f; selected signals: H NMR: d=4.21 (brs, 1H), 3.40 (m, 1H),
3
3.20 (m, 1H), 1.30 ppm (t, J
N
+
et al.[71] CM-Sepharose CL-6B (Amersham Pharmacia) in NH4 form sta-
Entry 18—(2R,3S,4R,5R)-2-(hydroxymethyl)pyrrolidine-5-isobutyl-3,4-
diol (4ad): Reaction at 48C, yield: 0.265 mmol, 60 mg, 16%; [a]2D0 =À57.3
tionary phase was packed into a glass column (160ꢊ20 mm) to give a
final bed volume of 50 mL. The flow rate was 0.9 mLminÀ1. The CM-
1
(c=0.8 in MeOH); the H and 13C NMR chemical shifts are coincident to
+
those for entry 17 (4ad); ESI-TOF: m/z calcd for [M+H]+ C9H20NO3:
Shepharose-NH4 was washed initially with H2O. Then, an aqueous solu-
tion of the crude material at pH 7 was loaded onto the column. Minor
colored impurities were washed away with H2O (150 mL, 3 bed volumes).
The retained compounds were eluted with aqueous NH4OH (0.01m):
entry 2 (20 mg) compounds: 4b (elution volume 116 mL, 5 mg); 4c (elu-
tion volume 156 mL, 2 mg); entry 3 (15 mg) compounds: 4d (elution
volume 136 mL, 4 mg), 4e (elution volume 162 mL, 2 mg), 4 f (elution
volume 104 mL, 2 mg), 4g (elution volume 143 mL, 1 mg); entry 6
(20 mg) compounds: 4i (elution volume 204 mL, 4 mg); 4j (elution
volume 224 mL, 1 mg); entry 15 (15 mg) compounds: 4y (elution volume
91 mL, 9 mg), 4z (elution volume 104 mL, 3 mg); entry 16 (18 mg) com-
pounds: 4aa (elution volume 84 mL, 7 mg), 4ab (elution volume 195 mL,
3 mg), 4ac (elution volume 97 mL, 2 mg). In each case, the operation was
repeated until the whole crude was consumed. Pure fractions were
190.1443; found: 190.1444.
Entry 19—(2S,3S,4S,5S)-2-(hydroxymethyl)pyrrolidine-5-benzyl-3,4-diol
trifluoroacetate salt (4ah): Yield: 0.5 mmol, 170 mg, 10%; [a]2D0 =À45.08
(c=1.0 in MeOH); 1H NMR (500 MHz, CD3OD): d=7.33 (m, 5H), 4.02
(t, 3J
system, 3J
11.8, 7.2 Hz, 1H), 3.70 (dt, J
(A of AB system, 3J
(H,H)=14.2, 7.6 Hz, 1H), 3.02 ppm (A of AB
system, 3J(H,H)=14.2, 8.0 Hz, 1H); 13C NMR (101 MHz, CD3OD): d=
ACHTUNGTRENNUNG ACHTUNGTRENNUNG
(H,H)=4.8 Hz, 1H), 3.95 (t, 3J
G
ACHTUNGTRENNUNG
3
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
136.1 (C8), 129.1 (C10), 128.8 (C9), 127.2 (C11), 78.2 (C4), 76.0 (C3),
65.9 (C2), 65.3 (C5), 58.8 (C6), 35.9 ppm (C7); ESI-TOF: m/z calcd for
[M+H]+ C12H18NO3: 224.1287; found: 224.1282.
Entry 20—(2S,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-benzyl-3,4-diol
(4ai;62%) and (2R,3S,4R,5S)-2-(hydroxymethyl)pyrrolidine-5-benzyl-3,4-
diol (4aj; 38%): Overall yield: 0.298 mmol, 77 mg, 6%; [a]2D0 =À14.08
(c=1.0 in MeOH).
1
pooled and lyophilized. Physical and NMR data are listed above. H and
13C NMR spectra are given in Supporting Information.
Enzymatic inhibition assays: Commercial glycosidase solutions were pre-
pared with the appropriate buffer and incubated in 96-well plates at
378C without (control) or with inhibitor (1.6–4.2 nm) for 3 min for a-d-
glucosidase, b-d-glucosidase, a-d-mannosidase, a-l-rhamnosidase, and a-
l-fucosidase, and 5 min for b-d-galactosidase. After addition of the corre-
sponding substrate solution, incubations were prolonged for different
time periods: 10 min for a-d-glucosidase, 3 min for b-d-glucosidase,
6 min for a-d-mannosidase, 5 min for a-l-rhamnosidase, 7 min for a-l-fu-
cosidase, and 16 min for b-d-galactosidase and stopped by addition of
Tris solution (50 mL, 1m) or glycine buffer (180 mL, 100 mm, pH 10), de-
pending on the enzymatic inhibition assay. The amount of p-nitrophenol
formed was determined at 405 nm with UV/VIS Spectramax Plus (Molec-
ular Devices Corporation) spectrophotometer. For a-d-glucosidase from
rice, the activity was determined with p-nitrophenyl-a-d-glucopyranoside
(1 mm) in sodium acetate buffer (50 mm, pH 5.0). b-d-Glucosidase activi-
ty was determined with p-nitrophenyl-b-d-glucopyranoside (1 mm) in
sodium acetate buffer (100 mm, pH 5.0). b-d-Galactosidase activity was
determined with p-nitrophenyl-b-d-galactopyranoside (1 mm) in sodium
phosphate buffer (100 mm, 0.1 mm MgCl2, pH 7.2). a-d-Mannosidase ac-
Data for 4ai: 1H NMR (500 MHz, CD3OD): d=7.35 (m, 5H), 4.15 (dd,
3J
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
(101 MHz, CD3OD): d=136.4 (C8), 129.1 (C10), 128.7 (C9), 127.1 (C11),
72.1 (C4), 70.5 (C3), 63.2 (C2), 62.3 (C5), 58.5 (C6), 35.4 ppm (C7). ESI-
TOF: m/z calcd for [M+H]+ C12H18NO3: 224.1287; found: 224.1282
Data for 4aj: 1H NMR (500 MHz, CD3OD): d=7.26 (m, 5H), 4.41 (dd,
3J(H,H)=8.1, 4.5 Hz, 1H), 4.09 (t, 3J
ACHTUNGTRENNUNG ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
d=138.0 (C8), 129.0 (C10), 128.7 (C9), 127.1 (C11), 72.1 (C4), 70.3 (C3),
63.0 (C2), 62.2 (C5), 58.7 (C6), 32.1 ppm (C7).
Entry 21—(2S,3S,4S)-2-(hydroxymethyl)pyrrolidine-5-dimetil-3,4-diol
(4ak; 83%) and (2R,3S,4R)-2-(hydroxymethyl)pyrrolidine-5-dimetil-3,4-
diol (4al; 17%): Overall yield: 0.520 mmol, 103 mg, 21%; [a]2D0 =À38.78
(c=1.3 in MeOH).
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7325

J. Bujons, P. Clapꢂs et al.
tivity was determined with p-nitrophenyl-a-d-mannopyranoside (1 mm)
in sodium acetate buffer (50 mm, pH 5.0). a-l-Rhamnosidase activity was
determined with p-nitrophenyl-a-d-rhamnopyranoside (1 mm) in sodium
acetate buffer (50 mm, pH 5.0). a-l-Fucosidase activity was determined
with p-nitrophenyl-a-d-fucopyranoside (0.15 mm) in sodium acetate
buffer (50 mm, pH 5.0). The commercial glycosidase solutions were pre-
pared as follows: a-d-glucosidase (NH4)2SO4 suspension (100 mL) in
buffer (5 mL); b-d-glucosidase: (0.1 mgmLÀ1 buffer), b-d-galactosidase
from Aspergillus oryzae (0.5 mgmLÀ1 buffer), a-l-rhamnosidase (naringi-
nase) (0.3 mgmLÀ1 buffer); a-d-mannosidase (NH4)2SO4 suspension
(25 mL) in buffer (10 mL); b-d-galactosidase from bovine liver
(0.1 mgmLÀ1 buffer), and a-l-fucosidase (NH4)2SO4 suspension (33 mL)
in buffer (10 mL).
mology Modeling module contained in MOE. The Amber99 force field
and the default parameters implemented in the program were used for
the energetic calculations. Thus, ten models were generated that were
minimized and scored. The Ramachandran plots were used to asses the
quality of the models. The best scored model was used to carry out the
subsequent docking experiments. The protonation state of every residue
in the proteins was generally determined with the Protonate 3D applica-
tion implemented in MOE.[78] Protonate 3D solves the macromolecular
protonation state assignment problem by selecting a protonation state for
each chemical group that minimizes the total free energy of the system.
In the particular case of the a-l-fucosidase, which contains three residues
of histidine in the active center (H34, H128 and H129), the eight possible
protonation states derived from considering each His residue as neutral
or charged were manually generated and the docking experiments were
performed with the eight structures for the protein. All docking experi-
ments were carried out allowing full flexibility for the ligands and treat-
ing the proteins as rigid. To ensure an appropriate search of the confor-
mational space of the pyrrolidine ring, a conformational search centered
on the ring was carried out previous to the docking. In this way, several
ring conformers were generated for each compound and all of them were
used in independent docking runs. The docking protocol implied several
stages:
Kinetics of inhibition: The nature of the inhibition against enzymes and
the Ki values were determined from the Lineweaver–Burk or Dixon
plots.
Computational methods: All molecular simulations were conducted with
the program MOE (v. 2007.09 and 2008.10, Chemical Computing Group,
Montreal). The implemented MMFF94x force field, a modified version
of the MMFF94s force field,[72,73] was used in all protein–ligand and
ligand alone calculations. The electrostatic interactions were approximat-
ed using the generalized Born/Volume Integral (GB/VI) methodology,[74]
without cutoffs when no protein was implied or with a smoothed cutoff
between 14–15 ꢋ when there was a protein in the system. The coordi-
nates of the complex E. coli rhamnulose-1-phosphate aldolase with phos-
phoglycolohydroxamic acid (PGH), Thermotoga maritima a-l-fucosi-
dase-fucosyl fluoride complex, a-l-rhamnosidase from Bacillus sp. GL1
and a-d-mannosidase from Drosophila melanogaster were obtained from
the Protein Data Bank[75] at Brookhaven National Laboratory (en-
tries 1GT7, 1L9, 2OKX, and 1HXK respectively). The protocols used to
determine the lowest energy conformations for the aldol adducts in free
state and as complexes with the RhuA active center were the same as
those previously reported.[24,40] For the free adducts, starting from an ini-
tially optimized structure, a systematic conformational search was run in
which every non-terminal single bond was rotated in 60–1208 steps. The
conformations generated for each compound were then minimized and
ranked according to their energy. To confirm the nature of the lowest
energy conformer determined, a subsequent stochastic conformational
search was run in which the conformational space of the molecules was
explored by random rotation of bonds and simultaneous Cartesian per-
turbation. The conformations thus generated were minimized and
checked to determine if they were duplicates, within a RMS tolerance
(0.1 ꢋ), of previously generated conformations. The process was finished
when the number of failures to find new conformations exceeded a large
enough number (1000) of attempts. The adducts bound in the active
center of RhuA were first built by modifying the structure of the PGH
molecule included in PDB 1GT7. After a preliminary minimization, the
conformational space of the ligands was explored by running a stochastic
conformational search, as described above, to find the lowest energy
minima. During this conformational search, a restraint was imposed to
keep the coordination of the DHAP moiety to the Zn2+. All the confor-
mations obtained were further minimized without restraints and keeping
the protein rigid.
1. Conformer generation: Only the dihedral angles of non-ring bonds
were varied and up to 5000 conformations/ligand were generated.
2. Placement: Up to 1000 poses were generated for each ligand using
the alpha PMI method, in which ligands are placed by aligning ran-
domly their principal moments of inertia in the receptor site.
3. Scoring: The London dG scoring function implemented in MOE was
used as a crude estimation of the free energy of binding of the ligand
from a given pose. In this step up to 50 poses, after removal of dupli-
cates, were obtained for each ligand.
4. Refinement: Energy minimization of the system was carried out using
the MMFF94x forcefield and keeping the protein rigid.
5. Re-scoring: Poses were re-scored by determining the MM/GBVI non-
bonded interaction energy, which comprises van der Waals, Coulomb
and generalized Born implicit solvent interaction energies, between
the receptor and the ligand.
Interaction diagrams were built with MOE for the best docked poses ob-
tained for some of the inhibitors (see Supporting Information).
Acknowledgements
This work was supported by the Spanish MEC CTQ2006-01345/BQU, La
Maratꢁ de TV3 foundation (Ref.: 050931), Generalitat de Catalunya
(DURSI 2005-SGR-00698 and Grant 2005SGR05-01063), and ESF proj-
ects COST D25/0006/05 and CM0701. J. Calveras acknowledges the
CSIC for the I3P predoctoral scholarship.
The docking calculations to determine the binding modes of the pyrroli-
dine inhibitors to the active site of a-l-fucosidase, a-l-rhamnosidase, and
a-d-mannosidase were performed with the docking application included
in MOE. Previous to these docking simulations, the structures of the pro-
teins were prepared by removing all the solvent molecules and ligands,
adding hydrogen atoms, and minimizing the energy using the Amber99
force field[76,77] with implicit water solvation conditions (GB/VI), to
remove any steric clashes. Furthermore, the structure of T. maritima a-l-
fucosidase was modified in the following way. From the two chains con-
tained in the PDB structure, chain A, which contains a molecule of fuco-
syl fluoride in the active center of the enzyme, was used as starting point
to build the complete sequence of the protein. For that purpose, the coor-
dinates of the inhibitor at the active site were removed, the lacking resi-
dues 1–6, 47–55, 267–274, 297–300, and 449 were introduced in the pro-
tein sequence, and their structure was modeled and refined using the Ho-
[1] N. Asano, R. J. Nash, R. J. Molyneux, G. W. J. Fleet, Tetrahedron:
Asymmetry 2000, 11, 1645–1680.
[2] P. Compain, O. R. Martin, Bioorg. Med. Chem. 2001, 9, 3077–3092.
[3] C. R. Bertozzi, L. L. Kiessling, Science 2001, 291, 2357–2364.
[4] A. Varki, Cell 2006, 126, 841–845.
[5] K. M. Koeller, C. H. Wong, Nat. Biotechnol. 2000, 18, 835–841.
[6] R. K. Yu, M. Yanagisawa, J. Neurochem. 2007, 103, 39–46.
[7] N. Asano, J. Enzyme Inhib. 2000, 15, 215–234.
[8] A. A. Watson, G. W. J. Fleet, N. Asano, R. J. Molyneux, R. J. Nash,
Phytochemistry 2001, 56, 265–295.
[9] T. D. Butters, R. A. Dwek, F. M. Platt, Curr. Top. Med. Chem. 2003,
3, 561–574.
[10] P. Compain, O. R. Martin, From Synthesis Iminosugars to Therapeu-
tic Applications, Wiley, New York, 2007, p. 467.
7326
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328

Enzyme Catalysis
FULL PAPER
[11] T. M. Wrodnigg, A. J. Steiner, B. J. Ueberbacher, Anti-Cancer
Agents Med. Chem. 2008, 8, 77–85.
[12] T. D. Heightman, A. T. Vasella, Angew. Chem. 1999, 111, 794–815;
Angew. Chem. Int. Ed. 1999, 38, 750–770.
[13] A. Vasella, G. J. Davies, M. Bohm, Curr. Opin. Chem. Biol. 2002, 6,
619–629.
[14] L. Cipolla, B. La Ferla, F. Nicotra, Curr. Top. Med. Chem. 2003, 3,
485–511.
[15] T. Ayad, Y. Genisson, M. Baltas, Curr. Org. Chem. 2004, 8, 1211–
1233.
[16] M. Sugiyama, W. Greenberg, C. H. Wong, J. Synth. Org. Chem. Jpn.
2008, 66, 605–615.
[17] J.-B. Behr, R. Plantier-Royon, Recent Res. Dev. Org. Chem. 2006, 10,
23–52.
[18] L. Cipolla, B. La Ferla, M. Gregori, Comb. Chem. High Throughput
Screening 2006, 9, 571–582.
[19] P. Greimel, J. Spreitz, F. K. Sprenger, A. E. Stutz, T. M. Wrodnigg,
Org. Chem. Sugars 2006, 383–424.
[20] A. Kato, N. Kato, E. Kano, I. Adachi, K. Ikeda, L. Yu, T. Okamoto,
Y. Banba, H. Ouchi, H. Takahata, N. Asano, J. Med. Chem. 2005,
48, 2036–2044.
[21] H. Ouchi, Y. Mihara, H. Watanabe, H. Takahata, Tetrahedron Lett.
2004, 45, 7053–7056.
[22] N. Asano, K. Ikeda, L. Yu, A. Kato, K. Takebayashi, I. Adachi, I.
Kato, H. Ouchi, H. Takahata, G. W. J. Fleet, Tetrahedron: Asymme-
try 2005, 16, 223–229.
[23] H. Takahata, Y. Banba, H. Ouchi, H. Nemoto, A. Kato, I. Adachi, J.
Org. Chem. 2003, 68, 3603–3607.
[24] L. Espelt, J. Bujons, T. Parella, J. Calveras, J. Joglar, A. Delgado, P.
Clapꢂs, Chem. Eur. J. 2005, 11, 1392–1401.
[25] L. Espelt, T. Parella, J. Bujons, C. Solans, J. Joglar, A. Delgado, P.
Clapꢂs, Chem. Eur. J. 2003, 9, 4887–4899.
[26] C. Saotome, C. H. Wong, O. Kanie, Chem. Biol. 2001, 8, 1061–1070.
[27] C. Y. Wu, C. F. Chang, J. S. Y. Chen, C. H. Wong, C. H. Lin, Angew.
Chem. 2003, 115, 4809–4812; Angew. Chem. Int. Ed. 2003, 42, 4661–
4664.
[43] A full study on the inhibition activity against a-l-fucosidase from
bovine kidney of 5 and 6 and its diasteromerically related pyrroli-
dine isomers will be published elsewhere.
[44] In our hands, the inhibitory activity of 5 was higher than the value
previously reported by Wang et al.[48]
[45] M. Joubert, A. Defoin, C. Tarnus, J. Streith, Synlett 2000, 1366–
1368.
[46] N. Asano, K. Yasuda, H. Kizu, A. Kato, J.-Q. Fan, R. J. Nash,
G. W. J. Fleet, R. J. Molyneux, Eur. J. Biochem. 2001, 268, 35–41.
[47] M. Dubernet, A. Defoin, C. Tarnus, Bioorg. Med. Chem. Lett. 2006,
16, 1172–1174.
[48] Y. F. Wang, D. P. Dumas, C. H. Wong, Tetrahedron Lett. 1993, 34,
403–406.
[49] In a control experiment the known a-l-fucosidase inhibitor, 1-
deoxy-l-fuconojirimycin, was assayed giving a value of Ki =4.9 nm in
good agreement with that reported in the literature.[45]
[50] E. Moreno-Clavijo, A. T. Carmona, Y. Vera-Ayoso, A. J. Moreno-
Vargas, C. Bello, P. Vogel, I. Robina, Org. Biomol. Chem. 2009, 7,
1192–1202.
[51] A. J. Moreno-Vargas, A. T. Carmona, F. Mora, P. Vogel, I. Robina,
Chem. Commun. 2005, 4949–4951.
[52] C. Chevrier, D. LeNouen, M. Neuburger, A. Defoin, C. Tarnus, Tet-
rahedron Lett. 2004, 45, 5363–5366.
[53] C. Laroche, J. B. Behr, J. Szymoniak, P. Bertus, C. Schutz, P. Vogel,
R. Plantier-Royon, Bioorg. Med. Chem. 2006, 14, 4047–4054.
[54] G. Sulzenbacher, C. Bignon, T. Nishimura, C. A. Tarling, S. G. With-
ers, B. Henrissat, Y. Bourne, J. Biol. Chem. 2003, 279, 13119–13128.
[55] T. M. Gloster, P. Meloncelli, R. V. Stick, D. Zechel, A. Vasella, G. J.
Davies, J. Am. Chem. Soc. 2007, 129, 2345–2354.
[56] T. M. Chapman, S. Courtney, P. Hay, B. G. Davis, Chem. Eur. J.
2003, 9, 3397–3414.
[57] J. H. Kim, M. J. Curtis-Long, W. D. Seo, J. H. Lee, B. W. Lee, Y. J.
Yoon, K. Y. Kang, K. H. Park, Bioorg. Med. Chem. Lett. 2005, 15,
4282–4285.
[58] L. Provencher, D. H. Steensma, C.-H. Wong, Bioorg. Med. Chem.
1994, 2, 1179–1188.
[59] Z. Cui, Y. Maruyama, B. Mikami, W. Hashimoto, K. Murata, J. Mol.
Biol. 2007, 374, 384–398.
[60] W. Hashimoto, H. Nankai, B. Mikami, K. Murata, J. Biol. Chem.
2003, 278, 7663–7673.
[28] I. Robina, A. J. Moreno-Vargas, J. G. Fernandez-Bolanos, J. Fuentes,
R. Demange, P. Vogel, Bioorg. Med. Chem. Lett. 2001, 11, 2555–
2559.
[29] A. T. Carmona, J. Fuentes, I. Robina, E. R. Garcia, R. Demange, P.
Vogel, A. L. Winters, J. Org. Chem. 2003, 68, 3874–3883.
[30] A. T. Carmona, F. Popowycz, S. Gerber-Lemaire, E. Rodriguez-
Garcia, C. Schutz, P. Vogel, I. Robina, Bioorg. Med. Chem. 2003, 11,
4897–4911.
[31] F. Popowycz, S. Gerber-Lemaire, C. Schutz, P. Vogel, Helv. Chim.
Acta 2004, 87, 800–810.
[32] T. D. Butters, L. A. G. M. Van den Broek, G. W. J. Fleet, T. M.
Krulle, M. R. Wormald, R. A. Dwek, F. M. Platt, Tetrahedron:
Asymmetry 2000, 11, 113–124.
[61] S. P. Kawatkar, D. A. Kuntz, R. J. Woods, D. R. Rose, G. J. Boons, J.
Am. Chem. Soc. 2006, 128, 8310–8319.
[62] D. A. Kuntz, W. Z. J. Guo, D. R. Rose, G.-J. Boons, ChemBioChem
2009, 10, 268–277.
[63] D. A. Kuntz, C. A. Tarling, S. G. Withers, D. R. Rose, Biochemistry
2008, 47, 10058–10068.
[64] W. Zhong, D. A. Kuntz, B. Ember, H. Singh, K. W. Moremen, D. R.
Rose, G. J. Boons, J. Am. Chem. Soc. 2008, 130, 8975–8983.
[65] B. L. Cantarel, P. M. Coutinho, C. Rancurel, T. Bernard, V. Lom-
bard, B. Henrissat, Nucleic Acids Res. 2009, 37, D233–D238.
[66] S. Howard, S. He, S. G. Withers, J. Biol. Chem. 1998, 273, 2067–
2072.
[33] A. J. Moreno-Vargas, I. Robina, R. Demange, P. Vogel, Helv. Chim.
Acta 2003, 86, 1894–1913.
[67] Although the bound conformation of mannostatin A in structure
2F7O shows the thiomethyl group oriented away from the hydro-
phobic pocket constituted by residues F206, W415, and Y727 and to-
wards the hydrophilic environment, a recent paper[62] from the
group that originally deposited the structure points out that this was
actually due to an experimental artifact and that a later structural
determination under different conditions showed that the actual ori-
entation of the thiomethyl group is towards the above mentioned
hydrophobic pocket.
[68] S.-H. Jung, J.-H. Jeong, P. Miller, C.-H. Wong, J. Org. Chem. 1994,
59, 7182–7184.
[69] M. Ocejo, J. L. Vicario, D. Badia, L. Carrillo, E. Reyes, Synlett 2005,
2110–2112.
[34] T. M. Wrodnigg, S. G. Withers, A. E. Stutz, Bioorg. Med. Chem.
Lett. 2001, 11, 1063–1064.
[35] C. Mattos, D. Ringe, Curr. Opin. Struct. Biol. 2001, 11, 761–764.
[36] C. Mattos, C. R. Bellamacina, E. Peisach, A. Pereira, D. Vitkup,
G. A. Petsko, D. Ringe, J. Mol. Biol. 2006, 357, 1471–1482.
[37] L. Espelt, P. Clapꢂs, J. Esquena, A. Manich, C. Solans, Langmuir
2003, 19, 1337–1346.
[38] T. Suau, G. Alvaro, M. D. Benaiges, J. Lopez-Santꢈn, Biotechnol.
Bioeng. 2006, 93, 48–55.
[39] M. L. Gutiꢂrrez, X. Garrabou, E. Agosta, S. Servi, T. Parella, J.
Joglar, P. Clapꢂs, Chem. Eur. J. 2008, 14, 4647–4656.
[40] J. Calveras, J. Bujons, T. Parella, R. Crehuet, L. Espelt, J. Joglar, P.
Clapꢂs, Tetrahedron 2006, 62, 2648–2656.
[70] F. Effenberger, A. Straub, Tetrahedron Lett. 1987, 28, 1641–1644.
[71] N. Asano, H. Kuroi, K. Ikeda, H. Kizu, Y. Kameda, A. Kato, I.
Adachi, A. A. Watson, R. J. Nash, G. W. J. Fleet, Tetrahedron:
Asymmetry 2000, 11, 1–8.
[41] A. Jimꢂnez, P. Clapꢂs, R. Crehuet, Chem. Eur. J. 2009, 15, 1422–
1428.
[42] M. Hayashi, H. Kawabata, Adv. Chem. Res. 2006, 1, 45–62.
Chem. Eur. J. 2009, 15, 7310 – 7328
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7327

J. Bujons, P. Clapꢂs et al.
[72] T. A. Halgren, J. Comput. Chem. 1999, 20, 730–748.
[73] T. A. Halgren, J. Comput. Chem. 1999, 20, 720–729.
[74] P. Labute, J. Comput. Chem. 2008, 29, 1693–1698.
[75] H. M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T. N. Bhat, H.
Weissig, I. N. Shindyalov, P. E. Bourne, Nucleic Acids Res. 2000, 28,
235–242.
[76] W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz,
D. M. Ferguson, D. C. Spellmeyer, T. Fox, J. W. Caldwell, P. A. Koll-
man, J. Am. Chem. Soc. 1995, 117, 5179–5197.
[77] J. Wang, P. Cieplak, P. A. Kollman, J. Comput. Chem. 2000, 21,
1049–1074.
[78] P. Labute, Proteins 2009, 75, 187–205.
Received: March 31, 2009
Published online: July 3, 2009
7328
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 7310 – 7328