Total syntheses of casuarine and its 6-O-α-glucoside: Complementary inhibition towards glycoside hydrolases of the GH31 and GH37 families

Article abstract of DOI:10.1002/chem.200801578

Total synthesis of naturally occurring casuarine (1) and the first total synthesis of casuarine 6-O-α-glucoside (2) were achieved through complete stereoselective nitrone cycloaddition, Tamao-Fleming oxidation and selective α-glucosylation as key steps. B

Full text of DOI:10.1002/chem.200801578

FULL PAPER
DOI: 10.1002/chem.200801578
Total Syntheses of Casuarine and Its
ACHTUNGTRENNUNG
U
Francesca Cardona,[a] Camilla Parmeggiani,[a] Enrico Faggi,[a] Claudia Bonaccini,[b]
Paola Gratteri,[b] Lyann Sim,[c] Tracey M. Gloster,[d] Shirley Roberts,[d]
Gideon J. Davies,[d] David R. Rose,[c] and Andrea Goti*[a]
Abstract: Total synthesis of naturally
occurring casuarine (1) and the first
total synthesis of casuarine 6-O-a-glu-
coside (2) were achieved through com-
plete stereoselective nitrone cycloaddi-
tion, Tamao–Fleming oxidation and se-
lective a-glucosylation as key steps.
Biological assays of the two com-
pounds proved their strong and selec-
tive inhibitory properties towards glu-
coamylase NtMGAM and trehalase
Tre37A, respectively, which place them
among the most powerful inhibitors of
these enzymes. The structural determi-
nation of the complexes of NtMGAM
with 1 and of Tre37A with 2 revealed
interesting similarities in the catalytic
sites of these two enzymes which
belong to different families and clans.
Keywords: enzyme structures
maltase glucoamylase natural
products · total synthesis · trehalase
·
·
Introduction
lated from the bark of Casuarina equisetifolia L. (Casuarin-
AHCTUNGTREGaNNUN ceae) and from the leaves of Eugenia jambolana Lam.
Casuarine, (1R,2R,3R,6S,7S,7aR)-3-(hydroxymethyl)-1,2,6,7-
tetrahydroxypyrrolizidine (1), and its 6-O-a-glucoside, casu-
arine 6-O-a-glucopyranoside (2) (see below), have been iso-
(Myrtaceae),[1] two plants well known for their therapeutic
action against diarrhoea, dysentery and colic,[2] breast
cancer,[1] diabetes and bacterial infections.[2,3] Compounds 1
and 2 are strong inhibitors of the fungal glucoamylase from
[a] Dr. F. Cardona, C. Parmeggiani, Dr. E. Faggi, Prof. A. Goti
Department of Organic Chemistry “U. Schiff”
Laboratory of Design, Synthesis
and Study of Biologically Active Heterocycles (HeteroBioLab)
University of Florence, Via della Lastruccia, 13
50019 Sesto Fiorentino (FI) (Italy)
Fax : (+39)055-457-3531
E-mail: andrea.goti@unifi.it
[b] Dr. C. Bonaccini, Prof. P. Gratteri
Laboratory of Molecular Modeling
Cheminformatics & QSAR, Department of Pharmaceutical Sciences
Laboratory of Design, Synthesis
and Study of Biologically Active Heterocycles (HeteroBioLab)
University of Florence, Via U. Schiff 6
50019 Sesto Fiorentino (FI) (Italy)
Aspergillus niger, with very similar IC50 values (IC50 =0.7 mm
for 1, IC50 =1.1 mm for 2).[4] No inhibition data against
human a-glucosidases or glucoamylases are available to
date.
Human maltase-glucoamylase (MGAM, EC 3.2.1.20) falls
into family GH31 of the Carbohydrate Active enZyme
(CAZy) classification.[5] It is one of the two enzymes (to-
gether with sucrase-isomaltase SI, EC 3.2.1.48 and 3.2.1.10)
responsible for catalyzing the last step in starch digestion by
hydrolyzing mixtures of dextrins at the non-reducing end
into glucose with net retention of anomeric configuration.[6]
Inhibitors of enzymes involved in the starch-digestion path-
way are used to delay glucose production and thus aid in
the treatment of Type II diabetes.[7] This therapeutic poten-
[c] L. Sim, Prof. D. R. Rose
Ontario Cancer Institute and Department of Medical Biophysics
University of Toronto, 101 College St., Toronto M5G 1L7 (Canada)
[d] Dr. T. M. Gloster, S. Roberts, Prof. G. J. Davies
York Structural Biology Laboratory, Department of Chemistry
University of York, Heslington, York, YO10 5YW (UK)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801578: Experimental proce-
dures and characterization for compounds 7, 12, 13, and 14, the ace-
1
tate derivative of 9 and the diacetate derivative of 10, and H and
13C NMR spectra of all compounds.
Chem. Eur. J. 2009, 15, 1627 – 1636
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1627

tial would be encouraged by an in-depth study of their
mode of action.[8] Some of us recently reported[9] the first
crystal structure of the human N-terminal subunit of
MGAM (NtMGAM) in its apo form and in complex with
acarbose (3) (see below), a tetrasaccharide analogue cur-
rently on the market as an antidiabetic drug (Glucobay, Pre-
cose).
Trehalases, another class of glycoside hydrolase of current
interest, hydrolyze one of the two glycosidic bonds in treha-
lose (a-d-glucopyranosyl-a-d-glucopyranoside) with inver-
sion at the anomeric center.[10] Trehalases are vital for insect
flight by providing glucose from trehalose hydrolysis, and
thus their inhibition is an interesting target for novel insecti-
cides.[7] Casuarine (1) and its glucoside 2 are known to inhib-
it porcine kidney trehalase (IC50 =12 and 0.34 mm, respec-
tively).[4] Some of us recently presented the first three-di-
mensional structure of a periplasmic trehalase from Escheri-
chia coli (Tre37A), which belongs to family GH37 of the
CAZy classification, in complex with some potent inhibitors,
1-thiatrehazolin (4) and validoxylamine A (5).[11] In order to
evaluate and compare the inhibitory activity of 1 and 2 to-
wards the human NtMGAM and E. coli Tre37A we under-
took the total syntheses of both compounds.
iminosugars isolated from natural sources, thus indicating
that their occurrence in Nature is not uncommon.[14] In con-
trast, few syntheses of glycosyl iminosugars are reported in
the literature,[15,16] and they all reflect some difficulties in
targeting the final compounds with good yields and selectivi-
ties, particularly in the glucosylation step. Interestingly, their
inhibitory properties differ substantially from those of the
parent iminosugar, with the linked carbohydrate moiety usu-
ally generating higher selectivity. This can be due, in princi-
ple, to a different arrangement within the catalytic site or,
most likely, to additional stabilizing (or destabilizing) inter-
actions with other subsites.
In this paper we report a straightforward and highly selec-
tive approach to the synthesis of casuarine (1) and the first
total synthesis of casuarine-6-O-a-glucoside (2), their inhibi-
tory activity towards NtMGAM and Tre37A, and the three-
dimensional structure of NtMGAM in complex with 1 and
Tre37A in complex with 2, supported by a molecular model-
ing study. These data provide a rationale for understanding
the complementary inhibitory properties of 1 and 2 against
human NtMGAM and E. coli Tre37A. These studies also
show that although MGAM and Tre37A belong to different
families and clans,[5] have different sizes, and operate by dif-
ferent mechanisms (MGAM proceeds with retention, and
Tre37A with inversion of configuration at the anomeric
center), many similarities are found within their active sites.
Importantly, we found that casuarine (1) is able to inhibit
NtMGAM more strongly than the pseudo tetrasaccharide
acarbose (3), holding promise for the development of novel
antidiabetic drugs, and that its 6-O-a-glucoside 2 inhibits
Tre37A at levels comparable with its most powerful known
inhibitors.
Results and Discussion
Our approach to the synthesis of casuarine (1) started with
nitrone 6, available in multigram scale from 2,3,5-tri-O-ben-
zylarabinofuranose.[17] Nitrone 6 has the correct relative and
absolute stereochemistry at C1, C2 and C3 of the target
molecule. The first key step for the synthesis of the casuar-
ine skeleton was the 1,3-dipolar cycloaddition of nitrone 6
with (Z)-alkene 7. Although no close example was reported,
we were confident that (Z)-disubstituted alkene 7 could pro-
vide the correct regioselectivity in the cycloaddition step on
the basis of related nitrone cycloadditions with vinyl si-
lanes.[18] Indeed, the desired adduct 8 was obtained with
high regio- and stereoselectivity in 79% yield (Scheme 1).
The regiochemistry and relative stereochemistry of 8 were
firmly established on the basis of COSY and NOESY ex-
periments. The complete stereoselectivity of the 1,3-dipolar
cycloaddition, which allowed the selective installation of the
three new stereocenters (corresponding to C6, C7 and C7a
in the target molecule) in one step with the required config-
uration, may be ascribed to the peculiar all-trans disposition
of the benzyloxy groups in nitrone 6, which hampered any
endo or syn approaches.[17,19]
Isolation and purification of 1 and 2 from natural sources
are very difficult and expensive, and afford only minor
amounts (indeed, 2 has never been obtained in a chemically
pure form). Furthermore, very few total syntheses of casuar-
ine (1) have been reported to date,[12,13] and no total synthe-
sis of 2 has been previously described. The highly complex
and challenging structure of casuarine, as it is the most
highly hydroxylated pyrrolizidine alkaloid and possesses six
contiguous carbon stereocenters, is presumably responsible
for the limited number of synthetic approaches described.
An additional problematic issue in the synthesis of casuarine
6-O-a-glucoside lies in the selectivity of glucosylation of the
hydroxy group at C6. Carefully planned extraction proce-
dures have recently uncovered more examples of glycosyl
1628
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1627 – 1636

Total Syntheses of Casuarine
FULL PAPER
Scheme 1. Synthesis of casuarine (1). a) CH2Cl2, RT, 36 h, 79%; b) Zn,
AcOH/H2O, 60–658C, 5 h, 93%; c) HgACHTNUTRGNEUNG(CF3CO2)2, TFA, AcOH, AcOOH,
Scheme 2. Synthesis of casuarine 6-O-a-glucoside (2). a) Ac2O, Py, RT,
15 h, 100%; b) Hg
ACHTUGNRTEN(NGNU CF3CO2)2, TFA, AcOH, AcOOH, CHCl3, 82%; c)
CHCl3, 76%; d) LiAlH4, THF, reflux, 78%; e) H2, Pd/C, MeOH, HCl,
100%.
BnOC(=NH)CCl3, CF3SO3H, Et2O, RT, 3 h; d) Ambersep 900 OH,
AHCTUNGTRENNUNG
MeOH, RT, 15 h, 75% (2 steps); e) TMSOTf, Et2O, ꢀ208C, 40 min,
72%; f) LiBH4, BH3·THF, THF, 238C, 3 d, 68%; g) H2, Pd/C, MeOH,
HCl, 77%.
[17,20,21]
ꢀ
Reductive cleavage of the N O bond
with Zn in
acetic acid followed by spontaneous N-cyclization afforded
lactam 9 in 93% yield. Then, the Tamao–Fleming reaction[22]
ature the starting material was converted only marginally.
Eventually, treatment of 15 with a high excess of LiBH4 in
combination with BH3·THF at room temperature for three
days[23] afforded glucoside 16 in 68% yield. Catalytic hydro-
genation finally provided casuarine 6-O-a-glucoside (2) in
77% yield as a white foam.
ꢀ
allowed oxidation of the C Si bond with retention of config-
uration and 76% yield of diol 10. Oxidation of the C-Si
bond to C-OH with retention of configuration was con-
firmed by NOESY experiments performed on the diacetyl
derivative of 10. Finally, reduction of 10 with LiAlH4 in re-
fluxing THF (78% yield), followed by catalytic hydrogena-
tion of pyrrolizidine 11, provided quantitatively casuarine
(1) as a white solid (m.p. 180–1828C). The spectroscopic
data and physicochemical properties of the synthesized com-
pound were identical to those of natural casuarine.[1a] Syn-
thetic 1 showed a specific optical rotation ([a]2D0 =+14.4 (c
= 0.52, H2O)) in good agreement with those reported for
the natural and for the previously synthesized compound
([a]2D4 =+16.9 (c = 0.80, H2O);[1a] [a]2D7 =+10.8 (c = 1.02,
H2O)[12a]). In conclusion, our total synthesis furnished casu-
arine in five steps and 44% overall yield from nitrone 6,
which compares well with the previously reported syntheses.
For the synthesis of casuarine 6-O-a-glucoside (2), we
needed to carry out the Tamao–Fleming oxidation prior to
glycosylation as the glycosidic bond did not tolerate the
harsh reaction conditions of Tamao–Fleming oxidation. Acet-
ylation of lactam 9 and subsequent Tamao–Fleming oxida-
tion afforded 12 with 82% yield (Scheme 2). Then, a protec-
tion–deprotection sequence (75% yield over 2 steps) was re-
quired to give the intermediate 13, which bears the free hy-
droxy group at C6 suitable to be linked to the glucose
moiety. The glucosylation reaction of 13 was carried out
with the trichloroacetimidate 14 and catalytic TMSOTf in
diethyl ether and afforded selectively the a-anomer 15 in
72% yield. Only traces of the b-anomer, which could not be
isolated, were detected in the crude reaction mixture. Re-
duction of C=O double bond of 15 was not a trivial task.
LiAlH4, Red-Al, BH3·SMe2 or LiBH4 in refluxing THF af-
forded complex mixtures of products, while at room temper-
The synthesis of casuarine 6-O-a-glucoside (2) renders it
available in a chemically pure form. Indeed, the sample iso-
lated from a natural source was evaluated to be only around
1
60% pure.[1b] Only H NMR and 13C NMR data have been
reported for the isolated natural compound and they are in
good agreement with those of the compound synthesized by
us, thus confirming the assignment proposed by the au-
thors.[1b]
Inhibition studies of 1 and 2 towards NtMGAM and
Tre37A show complementary behaviour with the two en-
zymes (Table 1).
Table 1. Inhibition of NtMGAM and Tre37A by 1 and 2.
NtMGAM Ki
Tre37A Ki
casuarine (1)
casuarine 6-O-a-glucopyranoside (2)
0.45 mm
280 mm
17 mm
12 nm
NtMGAM inhibition studies revealed that 1, with a Ki of
0.45 mm, displayed over 600-fold greater inhibition against
NtMGAM compared to 2, which inhibited with a Ki of
280 mm. These results are similar to the inhibition profile of
rat intestinal maltase where 1 and 2 inhibit with IC50 values
of 0.7 and 260 mm, respectively, but are in contrast to the in-
hibition profile of fungal glucoamylases where 1 and 2 dis-
play similar and potent inhibitory activities.[4]
The similarity in inhibition profiles between NtMGAM
and the rat intestinal maltase agrees with previous kinetic
Chem. Eur. J. 2009, 15, 1627 – 1636
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1629

A. Goti et al.
studies,[9,24] which propose that the N-terminal MGAM
domain of human maltase-glucoamylase displays maltase ac-
tivity whereas its C-terminal catalytic domain displays glu-
coamylase activity.
The crystal structure of NtMGAM in complex with 1 was
solved to 2.1 ꢁ (see Experimental Section for details and
data processing and refinement statistics). The electron den-
sity clearly revealed one molecule of 1, Figure 1a, bound in
Figure 1. Stereo (divergent) ball-and-stick representation of a) casuarine
(1) in complex with NtMGAM and b) casuarine 6-O-a-glucoside (2) in
complex with Tre37A. Observed electron density for the maximum likeli-
hood weighted 2FobsꢀFcalcd map is contoured at 1s; drawn by using BOB-
SCRIPT.[25]
Figure 2. Interactions between a) 1 and NtMGAM (D443 is the catalytic
nucleophile, D542 the catalytic acid/base) and b) 2 and Tre37A (E496 is
the catalytic base, D312 the catalytic acid). Hydrogen bonds are shown
by dashed lines.
the ꢀ1 subsite and one glycerol molecule (originating from
the cryoprotection solution) bound in the +1 subsite. The
two pyrrolidine rings, named “A” and “B” (shown above),
more potent than 1 (Table 1), indicating that the glucose
moiety in the +1 subsite (see below) contributes consider-
ACHTUNGTRENaNUGN bly to binding. This fits with the fact that Tre37A hydroly-
2
both adopt an envelope configuration; ring A is in a E con-
formation, and ring B in an E6 conformation. With its heavi-
ly hydroxylated rings, 1 binds tightly to the NtMGAM
active site primarily through hydrogen-bonding interactions
with the side chains lining the ꢀ1 subsite (Figure 2a). These
include hydrogen bonds between D327 and the hydroxyl
groups at C8 and C2, H600 with the C1-OH and C2-OH,
D443 with the C7-OH and the pyrrolidine nitrogen, R526
with the C7-OH and D542 with the C6-OH. The weaker in-
hibition of 2 compared with 1 leads us to believe that the
addition of glucoside group of 2 makes unfavourable inter-
actions with the +1 subsite, or weakly competes with the
casuarine group to occupy the ꢀ1 subsite, both of which
may decrease the inhibition properties of the casuarine
group.
Kinetic data were measured on Tre37A using an assay
where glucose was detected by glucose oxidase/peroxidase
linking enzymes following trehalose hydrolysis (see Experi-
mental Section for details). Ki values for 1 and 2 were deter-
mined at 378C and pH 5.5, and included a 20-minute pre-in-
cubation of the inhibitor with the enzyme to prevent any
complications in the data analysis from slow onset inhibi-
tion. Compound 2 was shown to be around a thousand fold
ses disaccharide substrates and therefore presumably has
evolved to optimise interactions in both the ꢀ1 and +1 sub-
sites.[11]
X-ray data for Tre37A in complex with 2 were collected
to 1.9 ꢁ (see Experimental Section for details and data
processing and refinement statistics). There was clear elec-
tron density in each of the four molecules of the asymmetric
unit corresponding to a molecule of 2, Figure 1b. The casu-
arine moiety of 2 is bound in the ꢀ1 subsite of Tre37A (Fig-
ure 2b). The two pyrrolidine rings are both found in an en-
2
velope conformation; ring A is in a E conformation, and
5
ring B in a E conformation. The glucose moiety of 2, in the
4
+1 subsite, is bound in a relaxed C1 conformation. The ma-
jority of the interactions between the glucose and the active
site residues of Tre37A are as described previously;[11] the
exception to this is the hydroxyl group at the C6 position,
which hydrogen bonds with a sulfate group from the crystal-
lization conditions, and appears to cause a large movement
of E511 away from the active site. The hydroxyl group at C8
of the casuarine moiety hydrogen bonds with D160 and a
water molecule, the C2-OH interacts with D160 and the
oxygen of Q207 and the C1-OH with W159 and the back-
1630
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1627 – 1636

Total Syntheses of Casuarine
FULL PAPER
bone carbonyl of G310. The nitrogen atom in the pyrrolizi-
dine ring interacts with the same water molecule as the C8-
OH. In ring B, the hydroxyl group at C7 interacts with
W447 and a different water molecule (not shown).
6-O-a-glucoside (2) and Tre37A in complex with casuarine
(1) using Glide[26] docking simulations.
Docking of 2 in NtMGAM places the casuarine moiety in
the ꢀ1 subsite, and consequently the glucose moiety in the
+1 subsite (Figure 4c). With respect to the crystallographic
structure (Figure 4a), the casuarine (aglyconic) portion of 2
is not perfectly overlapped with 1, probably due to the pres-
ence of the 1,1-linkage instead of 1,4 present in physiologi-
cal substrates. Nevertheless, all of the hydrogen bonds we
found in the casuarine complex are formed by the casuarine
portion of 2 in the modelling, with the exception of the hy-
drogen bond between C6-OH of casuarine and D542, be-
cause of the presence of the glucosidic bond in that position.
The glucose moiety which lies in the +1 subsite is able to
form a few hydrogen bonds with solvent exposed residues
but is also in close contact to bulky residues that surround
the ꢀ1 subsite entrance and which may form a sort of “link-
age selector” in order to assure 1,4-linkage selectivity for
the enzyme. All together, these observations give a rationale
to the observed difference in inhibitory activity between 1
and 2; the weaker inhibition of NtMGAM by 2 most likely
Despite having different protein sequences and architec-
tural folds, structural comparison of the two enzymes reveals
an intriguing similarity of their binding sites, as shown in
Figure 3, which was obtained by overlapping the pyrrolizi-
dine nuclei in the two enzyme–inhibitor complexes. Al-
though not perfectly superimposable, almost every residue
lining the ꢀ1 subsite of NtMGAM corresponds to a similar
and analogously oriented residue in the Tre37A active site.
The most conserved residues are those directly involved in
ligand hydrogen bonding (D327, H600 and D542 in
NtMGAM, which corresponds to D160, W159 and D312 in
Tre37A, respectively). An exception to this is the backbone
carbonyl of G310 in Tre37A, which corresponds to a water
molecule in NtMGAM. Most of the hydrophobic residues
lining the ꢀ1 subsite are also conserved and occupy approxi-
mately the same positions (Y299, W406, W441, W539 and
H600 in NtMGAM which correspond to Y157, F153, W520,
W447 and W159 in Tre37A, respectively). The two binding
sites differ in the position of the catalytic side chains (D443,
the nucleophile in the retaining NtMGAM and E496, the
base in the inverting Tre37A), by more than 3 ꢁ, reflecting
the different mechanisms utilised by the enzymes. Strikingly,
as mentioned above, the respective catalytic acids, D542 in
MGAM and D312 in Tre37A, superimpose closely. An im-
portant difference between the two binding sites is the pres-
ence of a phenylalanine (F575) in NtMGAM in close prox-
imity to the catalytic acid/base D542, while the correspond-
ing amino acid in Tre37A is an alanine (A307). In accor-
reflects the speciACTHNUTRGNEUNGficity of the enzyme which has evolved to
hydrolyse the a-1,4 linkages in starch-derived dextrins
whereas 2 is a trehalose (a,a-1,1) mimic. Interestingly, a
comparison with the crystal structure of NtMGAM in com-
plex with acarbose previously reported,[9] shows that the
acarbose (3) is stabilized by an electrostatic interaction be-
tween its positively charged nitrogen and D542, but lacks
the two hydrogen bonds with D443 (Figure 4b). Therefore,
casuarine (1) is the only one of the three inhibitors to inter-
act with both of the key residues D443 and D542, which
may be partly responsible for the two orders of magnitude
increase in inhibition with respect to acarbose (3) or 2.
ACHTUNGTRENNUNGdance with their varying functions, the +1 binding sites of
NtMGAM and Tre37A have major differences, reflecting
the specificity for a-1,4 and a,a-1,1 linkages, respectively. In
addition, this site in NtMGAM is exposed to the solvent,
while in Tre37A it is a buried cavity.
Tre37A has a buried cavity, specific both for the a,a-1,1
substrate and for a glucoside in the +1 leaving group sub-
site. The shape and dimension of 2 are thus optimal for max-
imizing transition state mimicking interactions in the ꢀ1
subsite and leaving group glycoside interactions in +1.
These are achieved through hydrogen bonds involving all of
its acceptor and donor groups (Figure 5a). A similar net-
work of interactions were observed for previously solved
complexes with 1-thiatrehazolin (4) and validoxylamine A
(5),[11] which have a similar enzyme affinity (Figure 5b and c,
respectively). As the binding of 1 leaves one of the two sub-
sites of Tre37A empty (Figure 5d and e), fewer hydrogen
bonds can be formed with the protein. The formation of
fewer interactions gives a rationalization for the lower affin-
ity of the “monosaccharide-like” inhibitor compound as op-
posed to the “diACHTNUGTRNEUNGsaccharide-like” equivalent.
Figure 3. Stereo (divergent) ball-and-stick representation of active site
overlap between NtMGAM in complex with 1 (green) and Tre37A in
complex with 2 (yellow); drawn by using BOBSCRIPT.[25]
Conclusion
In summary, a straightforward total synthesis of casuarine
(1) from nitrone 6 in five steps and 44% overall yield has
been achieved, which featured complete stereoselective cy-
cloaddition and Tamao–Fleming oxidation. This strategy al-
With the aim of rationalizing the binding data, we mod-
eled the structure of NtMGAM in complex with casuarine
Chem. Eur. J. 2009, 15, 1627 – 1636
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1631

A. Goti et al.
Figure 4. Schematic views of interactions between NtMGAM and a) casuarine (1) and b) acarbose (3) in the solved structures, and c) casuarine 6-O-a-
glucoside (2) after docking. Plots were prepared with Ligplot.[27]
Figure 5. Schematic views of interactions between Tre37A and a) casuarine 6-O-a-glucoside (2), b) 1-thiatrehazolin (4) and c) validoxylamine A (5) in
the solved structures and the two possible orientations of casuarine (1) after docking (d and e in ꢀ1 and +1 subsite, respectively). Plots were prepared
with Ligplot.[27]
1632
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1627 – 1636

Total Syntheses of Casuarine
FULL PAPER
(1R,2R,3R,6S,7S,7aS)-1,2-Bis(benzyloxy)-3-[(benzyloxy)methyl-7-
[dimethylACHTUNGTENR(UNNG phenyl)silyl]-6-hydroxyhexahydro-5H-pyrrolizin-5-one (9): A
mixture of 8 (484 mg, 0.74 mmol) and Zn dust (244 mg) in CH3COOH/
H2O 9:1 (8 mL) was heated to 60–658C for 5 h and then filtered through
cotton. The solution was cooled to 08C and, under vigorous stirring, a
lowed the total synthesis of casuarine 6-O-a-glucoside (2) to
be accomplished from nitrone 6 in nine steps and 17% over-
all yield, thus opening the way for the generation of a
number of differently glycosylated pyrrolizidine iminodisac-
charides.
AHCTUNGTERGsNNNU aturated aqueous solution of NaHCO3 (60 mL) was added until alkaline
pH was reached. The aqueous phase was extracted with AcOEt (3ꢂ
60 mL) and the combined organic phases were dried over Na2SO4. After
filtration and evaporation under reduced pressure, a viscous oil was ob-
tained. Purification on silica gel (petroleum ether/ethyl acetate 3:2) af-
forded pure 9 (420 mg, 0.69 mmol, 93%) as a colorless viscous oil. Rf =
0.4 (petroleum ether/ethyl acetate 3:2). [a]2D2 =ꢀ47.0 (c=1.87 in CHCl3);
1H NMR (400 MHz, CDCl3, 258C): d=7.55–7.53 (m, 2H; H-Ar), 7.38–
This study demonstrates that among glycoside hydrolase
enzymes that belong to different families and clans, and
which differ in mechanism of action, there is a degree of
conservation of the analogous active site residues reflecting
the convergent evolution of optimised non-covalent interac-
tions. The binding of inhibitors 1 and 2 were investigated
with NtMGAM and Tre37A by kinetic and structural meth-
ods, which provide information on the relevant interactions
in the active site and could form the basis for the synthesis
of analogues. This study shows that casuarine can be envis-
aged as a lead compound for the construction of novel di-
2
7.23 (m, 16H; H-Ar), 7.14–7.12 (m, 2H; H-Ar), 4.55 (d, J=11.7 Hz, 1H;
Bn), 4.46 (s, 2H; Bn), 4.38–4.34 (m, 2H; Bn, H3), 4.30 (d, 3J=11.1 Hz,
1H; H6), 4.22 (d, 2J=11.3 Hz, 1H; Bn), 4.12–4.09 (m, 2H; H2, Bn), 3.62
(dd, 3J=9.8, 4.1 Hz, 1H; H7a), 3.58–3.56 (m, 1H; H1), 3.47 (dd, 2J=9.6,
3J=5.8 Hz, 1H; H8a), 3.43 (dd, 2J=9.6, 3J=8.0 Hz, 1H; H8b), 2.75 (brs,
1H; OH), 1.63 (dd, 3J=11.1, 9.8 Hz, 1H; H7), 0.44 (s, 3H; SiMe),
0.39 ppm (s, 3H; SiMe); 13C NMR (50 MHz, CDCl3, 258C): d=176.6 (s;
C5), 138.0–136.3 (s, 4C; C-Ar), 134.3–127.7 (d, 20C; C-Ar), 89.0 (d; C1),
84.4 (d; C2), 73.9 (d; C6), 73.0, 72.0, 71.5 (t, 3C; Bn), 68.5 (t; C8), 63.2
(d; C7a), 59.1 (d; C3), 38.9 (d; C7), ꢀ3.6, ꢀ4.8 ppm (q, 2C; SiMe2); IR
(CDCl3): n˜ =3400, 3000, 2850, 1700 cmꢀ1; MS (70 eV): m/z (%): 516 (4),
289 (26), 181 (13), 135 (89), 91 (100); elemental analysis calcd (%) for
C37H41NO5Si (607.81): C 73.11, H 6.80, N 2.30; found: C 72.78, H 6.81, N
2.15.
ACHTUNGTRENNUNGsaccharide and polysaccharide type glycoside hydrolase in-
hibitors, the syntheses of which are currently ongoing in our
laboratories.
Experimental Section
(1R,2R,3R,6S,7S,7aR)-1,2-Bis(benzyloxy)-3-[(benzyloxy)methyl]-6,7-di-
hydroxyhexahydro-1H-pyrrolizine (10): Mercuric trifluoroacetate (64 mg,
0.15 mmol) was added to a stirred solution of 9 (47 mg, 0.079 mmol) in
CHCl3 (0.47 mL), acetic acid (0.12 mL) and trifluoroacetic acid
(0.23 mL). The solution was stirred at RT for 1 h; then, peracetic acid
(1.35 mL) was added dropwise to the mixture with ice cooling. After 1 h
stirring at room temperature the solution was cooled to ꢀ108C and a
General methods: Commercial reagents were used as received. All reac-
tions were magnetically stirred and monitored by TLC on 0.25 mm silica
gel plates (Merck F254) and column chromatography was carried out on
Silica Gel 60 (32–63 mm). Yields refer to spectroscopically and analytical-
ly pure compounds unless otherwise stated. 1H NMR spectra were re-
corded on a Varian Mercury-400. 13C NMR spectra were recorded on a
Varian Gemini-200. Infrared spectra were recorded with a Perkin-Elmer
Spectrum BX FT-IR System spectrophotometer. Mass spectra were re-
corded on a QMD 1000 Carlo Erba instrument by direct inlet; relative
percentages are shown in brackets. ESI full MS were recorded on a
Thermo LTQ instrument by direct inlet; relative percentages are shown
in brackets. Elemental analyses were performed with a Perkin-Elmer
2400 analyzer. Optical rotation measurements were performed on a
JASCO DIP-370 polarimeter.
AHCTUNGTRENNUNG
a
(30 mL), dried over anhydrous Na2SO4, filtered and evaporated under re-
duced pressure. Purification of the residue by flash column chromatogra-
phy on silica gel (petroleum ether/ethyl acetate 2:3) afforded pure 10
(29 mg, 0.06 mmol, 76%) as a colorless oil. Rf =0.2 (petroleum ether/
ethyl acetate 2:3). [a]2D0 =ꢀ41.5 (c=0.68 in CHCl3); 1H NMR (400 MHz,
CDCl3, 258C): d=7.36–7.20 (m, 15H; H-Ar), 4.71 (d, 2J=11.7 Hz, 1H;
Bn), 4.56–4.42 (m, 6H; Bn, H6), 4.26 (t, 3J=3.6 Hz, 1H; H2), 4.15–4.12
(m, 2H; H7, H3), 3.91 (t, 3J=4.7 Hz, 1H; H1), 3.66 (t, 3J=6.3 Hz, 1H;
Ethyl
[dimethyl
(8):
solution of nitrone 6[17] (1.23 g, 2.95 mmol) and
(2S,3S,3aS,4R,5R,6R)-4,5-bis(benzyloxy)-6-(benzyloxy)methyl-3-
(phenyl)silyl]-hexahydropyrrolo[1,2-b]isoxazole-2-carboxylate
(1.17 g,
G
ACHTUNGTRENNUNG
A
7
2
3
2
3
H7a), 3.58 (dd, J=9.7, J=5.8 Hz, 1H; H8a), 3.50 ppm (dd, J=9.7, J=
4.3 Hz, 1H; H8b); 13C NMR (50 MHz, CDCl3, 258C): d=172.6 (s; C5),
137.9–137.7 (s, 3C; C-Ar), 128.5–127.8 (d, 15 C; C-Ar), 86.9 (d; C1), 85.1
(d; C2), 80.5 (d; C7), 78.3 (d; C6), 73.3 (t; Bn), 72.1 (t, 2C; Bn), 68.3 (t;
C8), 66.8 (d; C7a), 59.0 ppm (d; C3); IR (CDCl3): n˜ =3372, 3011, 2866,
1703, 1454, 1214, 1102 cmꢀ1; MS (70 eV): m/z (%): 428 (11), 398 (4), 368
(3) [M+ꢀCH2OBn], 292 (20), 262 (12), 171 (90), 169 (57), 90 (100); ele-
mental analysis calcd (%) for C29H31NO6 (489.56): C 71.15, H 6.38, N,
2.86; found: C 70.90, H 6.31, N 2.80.
6.16 mmol) in CH2Cl2 (3 mL) was stirred at RT for 36 h, then the solvent
was removed under reduced pressure. Purification of the residue by chro-
matography on silica gel (petroleum ether/ethyl acetate 7:1) afforded
pure 8 (1.52 g, 2.33 mmol, 79%) as a colorless oil. Rf =0.22 (petroleum
ether/ethyl acetate 7:1). [a]2D2 =ꢀ76.6 (c=0.56 in CHCl3); 1H NMR
(400 MHz, CDCl3, 258C): d=7.47–7.44 (m, 2H; H-Ar), 7.38–7.24 (m,
3
15H; H-Ar), 7.21–7.17 (m, 3H; H-Ar), 4.85 (d, J=8.8 Hz, 1H; H2), 4.58
(A part of an AB system, 2J=12.2 Hz, 1H; Bn), 4.55 (B part of an AB
2
system, 2J=12.2 Hz, 1H; Bn), 4.39 (A part of an AB system, J=11.9 Hz,
1H; Bn), 4.33 (B part of an AB system, 2J=11.9 Hz, 1H; Bn), 4.24 (A
part of an AB system, 2J=12.0 Hz, 1H; Bn), 4.16 (B part of an AB
system, 2J=12.0 Hz, 1H; Bn), 4.05–3.90 (m, 4H; H5, H3a, CH2CH3),
3.78 (dd, 2J=9.8, 3J=4.7 Hz, 1H; H8a), 3.65 (m, 1H; H4), 3.57 (dd, 2J=
9.8, 3J=8.2 Hz, 1H; H8b), 3.40 (dt, 3J=8.6, 4.7 Hz, 1H; H6), 2.40 (dd,
(1R,2R,3R,6S,7S,7aR)-1,2-Bis(benzyloxy)-3-[(benzyloxy)methyl]-6,7-di-
hydroxyhexahydro-1H-pyrrolizine (11): To a cooled (08C) solution of 10
(173 mg, 0.35 mmol) in dry THF (4.2 mL) a 1m solution of LiAlH4 in
THF (1.3 mL, 1.3 mmol) was added under nitrogen atmosphere. The mix-
ture was then refluxed for 2 h. After cooling to 08C, a saturated Na2SO4
solution (600 mL) was added dropwise. The suspension was then filtered
through Celite and washed with AcOEt. Evaporation under reduced
pressure afforded a viscous oil, that was quickly filtered over a short pad
of silica gel (ethyl acetate) obtaining pure 11 as a viscous oil (130 mg,
0.27 mmol, 78%). Rf =0.6 (ethyl acetate). [a]2D0 =ꢀ2.9 (c=0.53 in
CDCl3); 1H NMR (400 MHz, CDCl3, 258C): d=7.32–7.23 (m, 15H; H-
Ar), 4.66–4.61 (m, 2H; Bn), 4.56–4.49 (m, 4H; Bn), 4.13–4.10 (m, 2H;
H1, H6), 4.04 (t, 3J=5.6 Hz, 1H; H2), 3.96 (t, 3J=5.1 Hz, 1H; H7), 3.51
(dd, 2J=9.6, 3J=5.3 Hz, 1H; H8a), 3.46 (dd, 2J=9.6, 3J=6.3 Hz, 1H;
H8b), 3.40 (dd, 2J=11.0, 3J=5.6 Hz, 1H; H5a), 3.36 (t, 3J=5.3 Hz, 1H;
3
3J=11.1, 8.8 Hz, 1H; H3), 1.17 (t, J=7.2 Hz, 3H; CH2CH3), 0.39 (s, 3H;
SiMe), 0.38 ppm (s, 3H; SiMe); 13C NMR (50 MHz, CDCl3, 258C): d=
170.8 (s; C=O), 138.4–137.1 (s, 4C; C-Ar), 134.1–127.5 (d, 20C; C-Ar),
85.3 (d; C5), 83.9 (d; C4), 80.7 (d; C2), 73.4 (t; Bn), 72.7 (d; C6), 71.8 (d;
C3a), 71.6, 71.1 (t, 2C; Bn), 70.8 (t; C8), 61.1 (t; CH2CH3), 36.9 (d; C3),
14.0 (q; CH2CH3), ꢀ2.7, ꢀ4.2 ppm (q, 2C; SiMe2); IR (CDCl3): n˜ =3688,
3029, 3012, 2930, 1739, 1496, 1454, 1253 cmꢀ1; MS (70 eV): m/z (%): 560
(2), 400 (53), 296 (65), 188 (40), 181 (60), 105 (48), 91 (100); elemental
analysis calcd (%) for C39H45NO6Si (651.3): C 71.86, H 6.96, N 2.15;
found: C 71.58, H 7.13, N 2.08.
Chem. Eur. J. 2009, 15, 1627 – 1636
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1633

A. Goti et al.
H7a), 3.24 (dt, 3J=5.6, 5.5 Hz, 1H; H3), 2.81 ppm (dd, 2J=11.0, 3J=
5.3 Hz, 1H; H5b); 13C NMR (50 MHz, CDCl3, 258C): d=138.2–137.9 (s,
3 C; C-Ar), 128.4–127.5 (d, 15C; C-Ar), 86.8 (d; C1), 85.2 (d; C2), 81.1
(d; C7), 78.7 (d; C6), 73.6 (d; C7a), 73.3 (t; Bn), 72.5 (t; Bn), 72.0 (t;
Bn), 71. 6 (t; C8), 69.8 (d; C3), 59.9 ppm (t; C5); IR (CDCl3): n˜ =3393,
2870, 2246, 1711, 1596, 1495, 1453, 1362, 1101; MS (70 eV): m/z (%): 354
(38) [M+ꢀCH2OBn], 234 (3), 172 (18), 160 (12), 91 (100); elemental
analysis calcd (%) for C29H33NO5 (475.58): C 73.24, H 6.99, N 2.95;
found: C 72.98, H 7.23, N 2.83.
mixture was filtered and extracted with Et2O. The organic layer was
dried over Na2SO4, filtered and concentrated to dryness. The residue was
purified by flash column chromatography on silica gel (petroleum ether/
ethyl acetate 4:1) to afford pure 16 (199 mg, 0.18 mmol, 68%) as a color-
less oil. Rf =0.38 (petroleum ether/ethyl acetate 4:1); [a]2D4 =+20.3 (c=
0.60 in CDCl3); 1H NMR (400 MHz, CDCl3, 258C): d=7.29–7.14 (m,
40H; H-Ar), 4.84–4.78 (m, 3H; H1 + 2 Bn), 4.71 (d, 2J=11.7 Hz, 1H;
Bn), 4.68–4.64 (m, 2H; Bn), 4.55–4.34 (m, 12H; 11 Bn + 1H), 4.22 (d,
J=11.3 Hz, 1H), 4.16–4.09 (m, 3H; H2’+ 2H), 3.98–3.84 (m, 3H; H1’,
H5’a+ 1H), 3.76 (t, J=9.3 Hz; 1H), 3.64–3.41 ppm (m, 7H; H5’b +
6H); 13C NMR (50 MHz, CDCl3, 258C): d=138.2–137.0 (s, 8 C; C-Ar),
128.6–127.2 (d, 40 C; C-Ar), 95.1 (d; C1), 85.7 (d, C1’), 83.6 (d; C2’), 81.4
(d; 3C), 80.3, 78.9, 77.2 (d; 3C), 75.3, 74.9, 73.4, 73.0, 72.7, 72.4, 71.8, 70.9
(t, 8C; Bn), 71.4, 69.3 (d; 2C), 68.3, 67.5 (t, 2C; C6, C8’), 65.3 ppm (t;
C5’); IR (CHCl3): n˜ =3676, 3513, 3010, 2921, 2869, 2388, 1951, 1876,
1811, 1603, 1497, 1454, 1364, 1215, 1086, 1072 cmꢀ1; HRMS (ESI): m/z:
calcd for C70H74NO10 [M+H]+: 1188.5307; found: 1188.5296; elemental
analysis calcd (%) for C70H73NO10 (1088.33): C 77.25, H 6.76, N 1.29;
found: C 76.72, H 6.91, N 1.15.
(1R,2R,3R,6S,7S,7aR)-3-(Hydroxymethyl)hexahydro-1H-pyrrolizine-
1,2,6,7-tetrol (casuarine, 1): To
a stirred solution of 11 (89 mg,
0.19 mmol) in MeOH (12.5 mL), 4–5 drops of conc. HCl and 100 mg of
Pd (10% on C) were added. The suspension was stirred under hydrogen
atmosphere for 4 d, then filtered through Celite and washed with MeOH.
Evaporation under reduced pressure afforded a vitreous solid that was
transferred to a column of DOWEX 50WX8 and then washed with
MeOH (15 mL), H2O (10 mL) to remove non amine containing products
and then with 7% NH4OH (25 mL) to elute casuarine (1). Evaporation
of the solvent afforded casuarine as a white solid (39 mg, 0.19 mmol,
100%). M.p. 180–1828C; [a]2D0 =+14.4 (c=0.52 in H2O); 1H NMR
(400 MHz, D2O, 25 8C): d=4.19–4.13 (m, 2H; H6, H7), 4.15 (t, 3J=
8.2 Hz, 1H; H1), 3.80–3–75 (m, 2H; H2, H8a), 3.60 (dd, 2J=11.7, 3J=
6-a-d-Glucopyranosyl-O-casuarine (2):
A solution of 16 (196 mg,
0.18 mmol) in MeOH (25 mL) was stirred at room temperature under H2
atmosphere for 24 h in the presence of 10% Pd/C (350 mg) and 4–5
drops of conc. HCl. Filtration through Celite afforded a waxy solid that
was transferred to a column of DOWEX 50WX8 and then washed with
MeOH (15 mL), H2O (10 mL) to remove non amine containing products
and then with 7% NH4OH (25 mL) to elute pure 2 (50 mg, 0.14 mmol,
77%) as a white foam. An analytically pure sample was obtained by fil-
tration through DOWEX 50WX8–200. [a]2D3 =+91.9 (c=0.35 in H2O);
1H NMR (400 MHz, D2O, 25 8C): d=4.89 (d, 3J=3.7 Hz, 1H; H1), 4.24
(t, 3J=3.0 Hz, 1H; H7’), 4.09 (dt, 3J=4.3, 3.0 Hz, 1H; H6’), 4.00 (t, 3J=
8.2 Hz, 1H; H1’), 3.79–3.75 (m, 1H; H6a), 3.68–3.56 (m, 5H; H3, H5,
H6b, H2’, H8’a), 3.49 (dd, 2J=11.8, 3J=6.3 Hz, 1H; H8’b), 3.47 (dd, 3J=
10.0, 3.7 Hz, 1H; H2), 3.31 (t, 3J=9.4 Hz, 1H, H4), 3.13 (dd, 2J=12.9,
3J=4.3 Hz, 1H; H5’a), 3.01–2.96 ppm (m, 3H; H3’, H5’b, H7a’);
13C NMR (50 MHz, D2O, 25 8C): d=96.7 (d; C1), 83.0 (d; C6’), 77.4 (d;
C7’), 76.9 (d; C1’), 76.1 (d; C2’), 72.1 (d; C7a’), 71.9 (d; C3), 71.5 (d; C5),
70.3 (d; C2), 68.9 (d; C4), 68.6 (d; C3’), 62.0 (t; C8’), 59.8 (t; C6),
55.1 ppm (t; C5’); HRMS (ESI): m/z: calcd for C14H26NO10 [M+H]+:
368.1551; found: 368.1544; elemental analysis calcd (%) for C14H25NO10
(367.35): C 45.77, H 6.86, N 3.81; found: C 45.39, H 6.57, N 3.46.
2
3
3
6.6 Hz, 1H; H8b), 3.27 (dd, J=11.8, J=3.7 Hz, 1H; H5a), 3.10 (dd, J=
8.5, 2.5 Hz, 1H; H7a), 3.07–3.04 (m, 1H; H3), 2.94 ppm (dd, 2J=11.8,
3J=3.0 Hz, 1H; H5b); 13C NMR (50 MHz, D2O, 25 8C): d=78.9 (d; C7),
77.8 (d; C1), 77.6 (d; C6), 76.6 (d; C2), 72.7 (d; C7a), 70.3 (d; C3), 62.1
(t; C8), 58.3 ppm (t; C5); IR (KBr): n˜ =3329, 2917, 1650, 1416; MS
(70 eV): m/z (%): 205 (0.9) [M+], 188 (0.2) [M+ꢀH2O], 174 (100) [M+
ꢀCH2OH], 128 (15), 102 (17), 70 (40); elemental analysis calcd (%) for
C8H15NO5 (205.21): C 46.82, H 7.37, N 6.83; found: C 46.51, H 7.54, N
6.73.
(2,3,4,6-Tetra-O-benzyl-a-d-glucopyranosyl)-(1!6)-O-tetra(benzyloxy)-
5-oxo-casuarine (15): A solution of glucopyranosyl tricholoroacetimidate
(14, 230 mg, 0.33 mmol) and pyrrolizidine 13 (115 mg, 0.20 mmol) in dry
diethyl ether (4 mL) was stirred for 10 min at room temperature under
nitrogen atmosphere in the presence of 3 ꢁ molecular sieves (150 mg).
After cooling to ꢀ208C and addition of trimethylsilyl trifluoromethane-
sulfonate (18 mL, 0.10 mmol), stirring was continued for 40 min; during
this period the temperature was raised to RT. The mixture was washed
with a sat. Na2CO3 solution (2 mL), dried over Na2SO4, filtered and con-
centrated to dryness. The residue was purified by flash column chroma-
tography on silica gel (petroleum ether/ethyl acetate 5:1) to afford pure
15 (159 mg, 0.144 mmol, 72%) as a colorless oil. Rf =0.32 (petroleum
ether/ethyl acetate 5:1). [a]2D4 =+51.5 (c=0.87 in CHCl3); 1H NMR
(400 MHz, CDCl3, 258C): d=7.47–7.12 (m, 40H; H-Ar), 5.69 (d, 3J=
Kinetic methods: Inhibition assays of NtMGAM were carried out in 96-
well microtitre plates containing 100 mm MES buffer, pH 6.5, inhibitor
and p-nitrophenol d-glucopyranoside (pNP-glucose, Sigma) as substrate
(2.5–30 mm). The reaction was allowed to proceed for 50 minutes at 378C
before quenching with 0.5m sodium carbonate and measuring the release
of the p-nitrophenolate ion at 405 nm.
2
2
3.5 Hz, 1H; H1), 5.01 (d, J=10.8 Hz, 1H; Bn), 5.00 (d, J=11.9 Hz, 1H;
Bn), 4.84 (d, 2J=10.9 Hz, 1H; Bn), 4.77 (d, 2J=10.8 Hz, 1H; Bn), 4.75
(d, 2J=10.5 Hz, 1H; Bn), 4.73 (d, 3J=8.2 Hz, 1H; H6’), 4.64–4.32 (m,
11H; Bn), 4.31 (dd, 3J=4.1, 3.7 Hz, 1H; H2’), 4.21–4.17 (m, 2H; H3’,
Inhibition of Tre37A was determined using a stopped assay, where glu-
cose was detected using glucose oxidase/peroxidase linking enzymes
(Megazyme, Bray, Eire) in the same way as described previously.[11]
Measurements were made at trehalose concentrations between 0.05 and
6 mm and Tre37A was present at a final concentration of 0.7 nm.
3
H7’), 3.99 (dd, J=9.4, 9.2 Hz, 1H; H4), 3.99–3.94 (m, 2H; H5, H1’), 3.74
(dd, 3J=6.6, 6.4 Hz, 1H; H7a’), 3.68 (dd, 3J=9.6, 9.2 Hz, 1H; H3), 3.64
(dd, 3J=9.6, 3.5 Hz, 1H; H2), 3.61–3.53 (m, 3H; H6a, H8’a, H8’b),
3.45 ppm (dd, 2J=10.7, 3J=1.7 Hz, 1H; H6b); 13C NMR (50 MHz,
CDCl3, 258C): d=169.4 (s; C5’), 138.9–137.2 (s, 8 C; C-Ar), 128.5–127.5
(d, 40 C; C-Ar), 95.2 (d; C1), 87.2 (d; C1’), 86.1 (d; C2’), 84.4 (d; C7’),
81.6 (d; C4), 78.6 (d; C2), 77.9 (d; C6’), 77.3 (d; C3), 75.7, 74.9 (t, 2 C;
Bn), 73.3 (t, 2C; Bn), 72.3 (t; Bn), 72.0 (t, 2 C; Bn), 71.7 (t; Bn), 70.4 (d;
C5), 69.2 (t; C8’), 68.1 (t; C6), 65.8 (d; C7a’), 58.9 ppm (d; C3’); IR
(CHCl3): n˜ =3066, 3032, 3010, 2867, 1711, 1454, 1098 cmꢀ1; HRMS (ESI):
m/z: calcd for C70H71NO11Na [M+Na]+: 1124.4919; found: 1124.4918; el-
emental analysis calcd (%) for C70H71NO11 (1102.31): C 76.27, H 6.49, N
1.27; found: C 76.53, H 6.76, N 1.38.
The Ki values of 1 and 2 against NtMGAM and Tre37A were calculated
by determining the reaction rates of the enzymes in the absence and
presence of inhibitor. Concentrations of 0.2–0.6 mm 1 or 0.4–1.2 mm 2
were used to inhibit NtMGAM and concentrations of 10–40 mm 1 or 10–
20 nm 2 were used to inhibit Tre37A. Rates were determined and the
data fitted to the Michaelis–Menten equation in GRAFIT (Erithacus
Software Ltd., Horley, UK) to obtain a KM (in absence of inhibitor) or
apparent KM (KMapp) (in presence of inhibitor). Ki values were deter-
mined using the equation KMapp =KM (1+[I]/Ki).
Structural methods: The crystallization of NtMGAM was previously re-
ported by Sim et al.[9] The complex of NtMGAM with 1 was obtained by
soaking NtMGAM crystals for ꢁ12 h in mother liquor supplemented
with 200 mm 1. X-ray diffraction data were collected on an ADSC Quan-
tum-4 CCD detector at beamline F1 at the Cornell High Energy Syn-
chrotron Source (CHESS) and were processed with HKL2000.[28] Since
the crystal of NtMGAM in complex with 1 was isomorphous to the crys-
tal of NtMGAM in complex with acarbose,[9] this structure was used as
(2,3,4,6-Tetra-O-benzyl-a-d-glucopyranosyl)-(1!6)-O-tetra(benzyloxy)-
casuarine (16): To a solution of 15 (300 mg, 0.27 mmol) in dry THF
(8.0 mL) a 2m solution of LiBH4 in THF (2.01 mL, 4.05 mmol) and 1m
BH3 in THF (4.05 mL, 4.05 mmol) were added dropwise at 08C. The re-
action mixture was stirred at 238C for 3 d, then, after cooling to ꢀ158C,
H2O (3 mL) was added dropwise. The THF was then evaporated and the
1634
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1627 – 1636

Total Syntheses of Casuarine
FULL PAPER
an initial model to calculate Fo-Fc maps. These difference maps clearly re-
vealed the position of casuarine in the NtMGAM active site. Inhibitor
topologies and restraints were generated using the PRODRG server
(http://davapc1.bioch.dundee.ac.uk/programs/prodrg/) and the model was
subsequently modified and refined using COOT[29] and REFMAC.[30]
Data processing and refinement statistics are shown in Table 2.
NMR spectrometer. Brunella Innocenti and Maurizio Passaponti (Di-
ACHUTNGRENUpNG artiAHCUTNGTRENmNUNG ento di Chimica Organica “U. Schiff”) are acknowledged for tech-
nical assistance. CINMPIS is gratefully acknowledged for a grant to C.P.
L.S. holds a scholarship from NSERC Canada and D.R.R. grants from
CIHR (MOP79400) and Heart and Stroke Canada (NA-6305). The York
laboratory would like to thank the BBSRC for funding. G.J.D. is a recipi-
ent of a Royal Society/Wolfson Research Merit award. T.M.G. is a Sir
Henry Wellcome postdoctoral fellow. The Cornell High Energy Synchro-
tron Source (CHESS), is supported by the National Science Foundation
under award DMR-0225180, and the Macromolecular Diffraction at
CHESS (MacCHESS) facility, by award RR-01646 from the National In-
stitutes of Health, through its National Center for Research Resources.
Table 2. Data collection and refinement statistics.
Crystal
NtMGAM in complex
Tre37A in complex
with 1
with 2
space group
unit cell dimensions
a [ꢁ]
P212121
P212121
86.7
92.5
[1] a) R. J. Nash, P. I. Thomas, R. D. Waigh, G. W. J. Fleet, M. R. Wor-
mald, P. M. De Q. Lilley, D. J. Watkin, Tetrahedron Lett. 1994, 35,
7849–7852; b) M. R. Wormald, R. J. Nash, A. A. Watson, B. K. Bha-
doria, R. Langford, A. Sims, G. W. J. Fleet, Carbohydr. Lett. 1996, 2,
169–174.
b [ꢁ]
c [ꢁ]
109.1
109.4
117.9
203.5
data collection[a]
beamline
CHESS F1
0.9175
ESRF ID29
0.9700
l [ꢁ]
[2] a) R. N. Chopra, S. L. Nayar, I. C. Chopra, Glossary of Indian Me-
dicinal Plants, Council of Scientific and Industrial Research (India),
New Dehli, 1956, p. 55; b) R. B. Nair, G. Santhakumari, Ancient Sci-
ence of Life: Journal of International Institute of Ayurveda 1986, 6,
80–84.
[3] a) J. K. Grover, S. Yadav, V. Vats, J. Ethnopharmacol. 2002, 81, 81–
100; b) S. R. Mentreddy, J. Sci. Food Agric. 2007, 87, 743–750.
[4] A. Kato, E. Kano, I. Adachi, R. J. Molyneux, A. A. Watson, R. J.
Nash, G. W. J. Fleet, M. R. Wormald, H. Kizu, K. Ikeda, N. Asano,
Tetrahedron: Asymmetry 2003, 14, 325–331.
[5] P. M. Coutinho, B. Henrissat in Recent Advances in Carbohydrate
Bioengineering (Eds.: H. J. Gilbert, G. J. Davies, B. Henrissat, B.
Svensson), Royal Society of Chemistry, Cambridge, 1999, pp. 3–12.
[6] a) B. L. Nichols, J. Eldering, S. Avery, D. Hahn, A. Quaroni, E. E.
Sterchi, J. Biol. Chem. 1998, 273, 3076–3081; b) B. L. Nichols, S.
Avery, P. Sen, D. M. Swallow, D. Hahn, E. E. Sterchi, Proc. Natl.
Acad. Sci. USA 2003, 100, 1432–1437.
resolution range [ꢁ]
unique reflections
redundancy
I/sI
completeness [%]
Rsym
refinement statistics
Rcryst [%]
Rfree [%]
16–2.1
57768
6.3 (5.5)
11.9 (5.0)
95.6 (97.4)
0.118 (0.432)
20–1.9
174266
7.1 (5.9)
18.9 (2.8)
99.6 (96.2)
0.098 (0.524)
22.5
27.8
0.014
1.53
3CTT
17.7
22.4
0.012
1.38
2JJB
R.M.S.D. bonds [ꢁ]
R.M.S.D. angles [8]
PDB code
[a] Values in parentheses refer to the highest-resolution shell; R.M.S.D.,
root mean square deviation.
Gene expression and purification of Tre37A was carried out as described
previously.[11] Protein, at a concentration of 10–12 mg/mL, was co-crystal-
lized with 5 mm 2 from 1.7m ammonium sulfate and 0.1m citric acid
buffer, pH 3.5. Crystals were cryoprotected in the mother liquor supple-
mented with 25% ethylene glycol and flash frozen. Data were collected
on beamline ID29 at the European Synchrotron Radiation Facility
(ESRF), and were processed with the HKL2000 suite.[28] As the cell di-
mensions differed from previous complexes of Tre37A, structure solution
required molecular replacement in PHASER[31] using a monomer of
2 JF4 as the search model. The model was refined with manual building
in COOT[29] interspersed with refinement of geometric restraints in
REFMAC.[30] Data processing and refinement statistics are shown in
Table 2.
[7] N. Asano, Glycobiology 2003, 13, 93R–104R.
[8] For reviews on glycosidase mechanisms, see: a) T. D. Heightman,
A. T. Vasella, Angew. Chem. 1999, 111, 794–815; Angew. Chem. Int.
Ed. 1999, 38, 750–770; b) D. L. Zechel, S. G. Withers, Acc. Chem.
Res. 2000, 33, 11–18; c) C. S. Rye, S. G. Withers, Curr. Opin. Chem.
Biol. 2000, 4, 573–580.
[9] L. Sim, R. Quezada-Calvillo, E. E. Sterchi, B. L. Nichols, D. R.
Rose, J. Mol. Biol. 2008, 375, 782–792.
[10] J. Defaye, H. Driguez, B. Henrissat, Carbohydr. Res. 1983, 124, 265–
273.
[11] R. P. Gibson, T. M. Gloster, S. Roberts, R. A. J. Warren, I. S. de Gra-
cia, A. Garcꢃa, J. L. Chiara, G. J. Davies, Angew. Chem. 2007, 119,
4193–4197; Angew. Chem. Int. Ed. 2007, 46, 4115–4119.
[12] a) S. E. Denmark, A. R. Hurd, Org. Lett. 1999, 1, 1311–1314; b) S.
Denmark, A. R. Hurd, J. Org. Chem. 2000, 65, 2875–2886; c) I. Iz-
quierdo, M. T. Plaza, J. A. Tamayo, Tetrahedron 2005, 61, 6527–
6533; d) R. J. Nash, G. W. J. Fleet, J. Van Ameijde, G. Horne, PCT
Int. Appl., WO2006008493, 2006.
[13] Casuarine has also been obtained as a side-product during the syn-
thesis of targeted 3-epi-casuarine: a) J. Van Ameijde, G. Horne,
M. R. Wormald, R. A. Dwek, R. J. Nash, P. W. Jones, E. L. Evinson,
G. W. J. Fleet, Tetrahedron: Asymmetry 2006, 17, 2702–2712. For
Computational methods: All calculations were performed on 2ꢂAMD
Opteron285 Dual Core processors running Linux RedHat OS. The struc-
tures of the complexes between NtMGAM and casuarine (1) and Tre37A
and casuarine 6-O-a-glucoside (2) were superimposed by the pyrrolizi-
dine nuclei. After the removal of all non-protein molecules, the crystallo-
graphic structures were prepared according to the Glide protein prepara-
tion procedure[26] used to assign missing hydrogens and to reduce struc-
tural problems. Docking calculations were performed with the Glide pro-
gram using default input parameters (no scaling factor for the vdW radii
of non polar protein atoms, 0.8 scaling factor for non polar ligand atoms)
and the SP scoring function was used to evaluate the resulting poses. The
best docked conformations were chosen using the GlideScore, a modified
and extended version of the empirically based ChemScore function.[32]
AHCTUNGERTGrNNUN elated syntheses of casuarine diastereoisomers, see also: b) ref.
[12d]; c) I. Izquierdo, M. T. Plaza, J. A. Tamayo, J. Carbohydr.
Chem. 2006, 25, 281–295; d) T. Ritthiwigrom, S. G. Pyne, Org. Lett.
2008, 10, 2769–2771.
[14] N. Asano, T. Yamauchi, K. Kagamifuchi, N. Shimizu, A. Takahashi,
H. Takatsuka, K. Ikeda, H. Kizu, W. Chuakul, A. Kettawan, T. Oka-
moto, J. Nat. Prod. 2005, 68, 1238–1242.
Acknowledgements
[15] Chemical syntheses of glycosyl iminosugars: a) P. B. Anzeveno, L. J.
Creemer, J. K. Daniel, C.-H. R. King, P. S. Liu, J. Org. Chem. 1989,
54, 2539–2542; b) P. S. Liu, J. K. Daniel, B. L. Rhinehart, Eur. Pat.
Appl. EP297534, 1989 [Chem. Abstr. 1989, 111, 7659r]; c) H.
We thank MiUR and Ente Cassa di Risparmio di Firenze, Italy, for finan-
cial support. Ente CRF is also acknowledged for granting a 400 MHz
Chem. Eur. J. 2009, 15, 1627 – 1636
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1635

A. Goti et al.
Ardron, T. D. Butters, F. M. Platt, M. R. Wormald, R. A. Dwek,
G. W. J. Fleet, G. S. Jacob, Tetrahedron: Asymmetry 1993, 4, 2011–
2024; d) F.-X. Felpin, K. Boubekeur, J. Lebreton, J. Org. Chem.
2004, 69, 1497–1503.
Parker, H. E. Plaut, P. E. J. Sanderson, J. Chem. Soc. Perkin Trans. 1
1995, 317–337.
[23] S. Nukui, M. Sodeoka, H. Sasai, M. Shibasaki, J. Org. Chem. 1995,
60, 398–404.
[24] R. Quezada-Calvillo, L. Sim, Z. Ao, B. R. Hamaker, A. Quadroni,
G. D. Brayer, E. E. Sterchi, C. C. Robayo-Torres, D. R. Rose, B. L.
Nichols, J. Nutr. 2008, 138, 685–692.
[25] R. M. Esnouf, J. Mol. Graph. Model 1997, 15, 132–134.
[26] a) R. A Friesner, J. L. Banks, R. B. Murphy, T. A. Halgren, J. Klicic,
D. T. Mainz, M. P. Repasky, E. H. Knoll, M. Shelley, J. K. Perry, D.
Shaw, P. Francis, P. S. Shenkin, J. Med. Chem. 2004, 47, 1739–1749;
b) Schrçdinger, LLC, New York (http://www.schrodinger.com).
[27] A. C. Wallace, R. A. Laskowski, J. M. Thornton, Prot. Eng., 1995, 8,
127–134.
[28] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276, 307–326.
[29] P. Emsley, K. Cowtan, Acta Crystallogr. Sect. A 2004, 60, 2126–2132.
[30] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. Sect.
A 1997, 53, 240–255.
[31] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn,
L. C. Storoni, R. J. Read, J. Appl. Crystallogr. 2007, 40, 658–674.
[32] M. D. Eldridge, C. W. Murray, T. R. Auton, G. V. Paolini, R. P. Mee,
J. Comput. Aided Mol. Des. 1997, 11, 425–445.
[16] Enzymatic syntheses of glycosyl iminosugars: a) N. Asano, K. Oseki,
E. Kaneko, K. Matsui, Carbohydr. Res. 1994, 258, 255–266; b) N.
Asano, A. Kato, H. Kizu, K. Matsui, R. C. Griffiths, M. G. Jones,
A. A. Watson, R. J. Nash, Carbohydr. Res. 1997, 304, 173–178.
[17] F. Cardona, E. Faggi, F. Liguori, M. Cacciarini, A. Goti, Tetrahedron
Lett. 2003, 44, 2315–2318.
[18] P. DeShong, J. M. Leginus, S. W. Lander Jr, J. Org. Chem. 1986, 51,
574–576.
[19] J. Revuelta, S. Cicchi, A. Goti, A. Brandi, Synthesis 2007, 4, 485–
504.
[20] A. Goti, S. Cicchi, M. Cacciarini, F. Cardona, V. Fedi, A. Brandi,
Eur. J. Org. Chem. 2000, 3633–3645.
[21] For our related work on the synthesis of necine bases, see: a) A.
Goti, V. Fedi, L. Nannelli, F. De Sarlo, A. Brandi, Synlett 1997, 577–
579; b) A. Goti, M. Cacciarini, F. Cardona, F. M. Cordero, A.
Brandi, Org. Lett. 2001, 3, 1367–1369.
[22] a) K. Tamao, N. Ishida, T. Tanaka, M. Kumada, Organometallics
1983, 2, 1694–1696; b) K. Tamao, N. Ishida, J. Organomet. Chem.
1984, 269, C37-C39; c) I. Fleming, R. Henning, H. E. Plaut, J. Chem.
Soc. Chem. Commun. 1984, 29–31; d) I. Fleming, R. Henning, D. C.
Received: August 1, 2008
Published online: January 2, 2009
1636
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1627 – 1636