Hydrolysis of toxic natural glucosides catalyzed by cyclodextrin dicyanohydrins

Article abstract of DOI:10.1002/ejoc.200700954

The hydrolysis of toxic 7-hydroxycoumarin glucosides and other aryl and alkyl glucosides, catalyzed by modified α- and β-cyclodextrin dicyanohydrins, was investigated using different UV, redox, or HPAEC detection assays. The catalyzed reactions all followed Michaelis-Menten kinetics, and an impressive rate increase of up to 7569 (k_cat/k_uncat) was found for the hydroxycoumarin glucoside substrate 4-MUGP. Good and moderate degrees of catalysis (k_cat/k_uncat) of up to 1259 were found for the natural glucosides phloridzin and skimmin. By using a newly developed catechol detection UV-assay, a weak degree of catalysis was also found for the toxic hydroxycoumarin esculin. A novel synthesized diaminomethyl β-cyclodextrin showed a weak catalysis of p-nitrophenyl β-D-glucopyranoside hydrolysis. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200700954

FULL PAPER
DOI: 10.1002/ejoc.200700954
Hydrolysis of Toxic Natural Glucosides Catalyzed by Cyclodextrin
Dicyanohydrins
Jeannette Bjerre,[a][‡] Erik Holm Nielsen,[a][‡] and and Mikael Bols*[a]
Keywords: Artificial enzyme / Supramolecular chemistry / Glycosidase
The hydrolysis of toxic 7-hydroxycoumarin glucosides and
other aryl and alkyl glucosides, catalyzed by modified α- and
β-cyclodextrin dicyanohydrins, was investigated using dif-
ferent UV, redox, or HPAEC detection assays. The catalyzed
reactions all followed Michaelis–Menten kinetics, and an im-
pressive rate increase of up to 7569 (kcat/kuncat) was found for
the hydroxycoumarin glucoside substrate 4-MUGP. Good
and moderate degrees of catalysis (kcat/kuncat) of up to 1259
were found for the natural glucosides phloridzin and skim-
min. By using a newly developed catechol detection UV-as-
say, a weak degree of catalysis was also found for the toxic
hydroxycoumarin esculin. A novel synthesized diamino-
methyl β-cyclodextrin showed a weak catalysis of p-ni-
trophenyl β-D-glucopyranoside hydrolysis.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
Introduction
In many parts of the world, ingestion of toxic glycosides
causes death of humans and livestock. The digitalis (heart
glycoside) family includes the potentially toxic digoxin (Fig-
ure 1), which is an inhibitor of the Na/K-ATPase, leading
to a positive inotrope effect (i.e. increases heart contraction
force).[1] Digoxin is one of the most prescribed of all cardiac
drugs, and is used for heart failure and atrial fibrillation.
However, the narrow therapeutic range of this agent in-
creases the risk of intoxication, especially in the elderly pop-
ulation.[2] Also, atractyloside (Figure 1), a toxic component
of the thistle plant Atractylis gummifera growing in North
Africa, is a glycoside responsible for intoxication and chil-
dren deaths every year.[3]
Figure 1. Toxic natural glycosides.
The glycoside esculin is the main toxic component of
Horse Chestnuts (Aesculus hippocastanum), fraxin being an-
other contributing factor.[4] Consumption of esculin can
cause diarrhoea, vomiting, bleeding and other symp-
toms,[5,6] and can be potentially fatal if ingested in larger
quantities, for humans or for livestock. However, hydrolysis
of esculin affords glucose and the virtually non-toxic agly-
con esculetin (Scheme 1).
In recent years, we have developed a number of artificial
enzymes, based on the cyclodextrin skeleton, and we imag-
ined that cyclodextrin glycosidases might be able to ef-
ficiently catalyze the hydrolysis of esculin and similar toxic
natural glycosides to non-toxic components. The size and
structure of the coumarin ring system makes it a potentially
Scheme 1. Hydrolysis of esculin affording glucose and esculetin.
[a] Department of Chemistry, University of Copenhagen,
Universitetsparken 5, 2100 Copenhagen Ø, Denmark
E-mail: bols@kemi.ku.dk
[‡] These two authors contributed equally to the paper.
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
good fit for binding inside the hydrophobic cavity of the
cyclodextrin,[7] whereafter cleavage of the glycosidic bond
could take place (see below). If this was the case, cyclodex-
Eur. J. Org. Chem. 2008, 745–752
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
745

J. Bjerre, E. H. Nielsen, M. Bols
FULL PAPER
trin glycosidases might find use as a potential starting point
for the further development of a treatment for hydroxycou-
marin glycoside poisoning.
Cyclodextrins are useful as artificial enzyme templates,
since the hydrophobic interior can bind non-polar substrate
parts, while the hydrophilic exterior makes the enzyme solu-
ble in water. Native cyclodextrins are safe to ingest orally,[8]
and are presently being used in medical agents,[9] as emulsi-
fiers and food additives, to preserve flavor in dry food prod-
ucts[10] or prolong shelf life and aromaticity of fine oils.[11]
From a green chemistry perspective, cyclodextrins are
cheaply obtained directly from starch by enzymatic diges-
tion and purification, and their use as artificial enzymes
could include industrial, medical or environmental pur-
poses. Through enzymatic catalysis of chemical processes,
one can abolish the need for employed toxic reagents, opti-
mize reaction yield and reduce by-product formation.
Global warming and the greenhouse effect calls for renew-
able, efficient catalysts to make possible an environmentally
sustainable production in chemistry and related sciences.
We believe that cyclodextrins as artificial enzymes have
many future applications as green chemistry catalysts,[12–14]
and in the present work we direct our attention towards
their use as artificial glycosidases.
Figure 2. Mechanism of catalysis.
This mechanism is supported by the fact that when only
one cyanohydrin group is present on the primary rim, this
affords roughly half the catalysis rate of that achieved when
having two cyanohydrin groups present. The reaction fol-
lows Michaelis–Menten kinetics and requires binding of the
substrate in the cyclodextrin cavity, as proven by inhibition
experiments with cyclopentanol.
Mimicking the catalytic power of natural enzymes that
nature has spent thousands of years to perfect, is a work
still in progress.[15–19] But since the cyclodextrins are much
smaller than natural enzymes that contain several large
structures and motifs in addition to the active site, the cata-
Results and Discussion
We investigated the ability of cyclodextrin dicyanohyd-
lytic efficiency per weight unit for some of the cyclodextrin rins to catalyze the hydrolysis of various toxic natural glyco-
derivatives is now starting to approach that of some natural sides of the 7-hydroxycoumarin type. In the case of esculin,
enzymes.[20–22] By selectively modifying perpendicular we developed a quantitative catechol detection assay (de-
alcohol groups on the primary rim of the cyclodextrin,[23] scribed below), and used this for monitoring the esculinase
we have achieved artificial oxidases, peroxidases and some effect of cyclodextrin. The hydrolysis of esculin, catalyzed
of the first artificial glycosidases.[24–26] The cyclodextrin gly- by β-cyclodextrin dicyanohydrin, follows Michaelis–
cosidase activity relies on the cyanohydrin group, wherein Menten kinetics, and results can be seen in Table 1. As the
the alcohol proton is made acidic by the electron with- data show, esculin binds well to the cyclodextrin, affording
drawing effect of the nitrile group, enabling acid-catalyzed a KM value that is generally smaller than that of the pre-
hydrolysis of the glycosidic bond. The mechanism for the viously tested aryl glycoside substrates.[13] However, cataly-
cyclodextrin-mediated cleavage of various aryl glycosides is sis of esculin hydrolysis was much weaker than that of the
assumed to take place as shown in Figure 2.
aryl glycoside substrates under similar conditions (esculin
Table 1. The reactions were performed in 50 mM phosphate buffer, pH 8.0/30 mM FeIII ammonium citrate, [β-cyclodextrin dicyanohydrin]
= 1 mM. Esculin concentration was between 1–10 mM. The reactions were followed by measuring the absorption at 570 nm.
Temp. [°C]
kcat [10–6 s–1]
KM [mM]
kcat/kuncat
25
40
50
60
2.60Ϯ0.12
2.50Ϯ0.34
3.19Ϯ1.03
3.19Ϯ0.93
1.20Ϯ0.73
3.90Ϯ1.05
1.03Ϯ1.38
2.56Ϯ1.87
22
60
26
10
746
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 745–752

Cyclodextrin-Based Artificial Enzymes
kcat is 20 times smaller at 50 and 60 °C).[14] Still, the data dicyanohydrin, which correlates well with the poorer ability
clearly show some catalysis, with a kcat/kuncat value of up to of the α-cyclodextrin cavity to accommodate more bulky
60.
We also investigated as a substrate the toxic glucoside
substrates, compared to that of β-cyclodextrin.
We also tested the artificial glycoside 4-MUGP (4-meth-
skimmin,[27] found amongst others in citrus fruits.[28] As ylumbelliferyl β--glucopyranoside), which is used for de-
skimmin is not readily commercially available, we synthe- tection of various bacterial strains,[29] and diagnostically in
sized it in two steps starting from tetraacetylated α--gluco- a test for Gaucher’s disease.[30] The appearance of hydroly-
pyranosyl bromide (Scheme 2). Upon hydrolysis, skimmin sis product, 4-methylumbelliferone, was monitored directly
affords glucose and umbelliferone (7-hydroxycoumarin), the by measurement of the absorbance at 360 nm. An excellent
latter of which was detected directly by measuring the ab- degree of catalysis was found, with a kcat/kuncat value of
sorption at 365 nm.
7569 for β-cyclodextrin dicyanohydrin, and aproximately
half that value for α-cyclodextrin (Table 2). We assume that
the high activity of this substrate, compared to skimmin, is
due to the presence of the 4-methyl group. This can be ex-
plained in the following manner: skimmin, being less bulky,
may have more binding modes to the catalyst and some of
these binding modes may be unproductive. The substrate is
effectively acting as its own inhibitor. 4-MUGP on the
other hand may be limited to the productive binding modes
by its structure. Therefore it is a much better substrate.
We also tested the dihydrochalcone phloridzin, a natural
glucoside found in apple zest and pulp. Phloridzin acts as
an antidiabetic agent, competitively inhibiting the intestinal
uptake of glucose,[31] and the renal reabsorption of glucose
in the proximal convoluted tubules. Phloridzin is a recog-
Scheme 2. Synthesis of skimmin.
Skimmin assay results are listed in Table 2. With β-cyclo- nized toxic agent, albeit the toxicity is relatively low (200–
dextrin dicyanohydrin, a good rate of catalysis (kcat/kuncat
)
400 mg/kg for induction of glycosuria).[32] Hydrolysis of
of 1259 was found for the reaction. A more moderate catal- phloridzin affords glucose and the flavanoid phloretin, and
ysis rate of 162 (kcat/kuncat) is obtained with α-cyclodextrin reaction progress was monitored by HPAEC glucose detec-
Table 2. The reactions were performed at 60 °C in 50 mM phosphate buffer, pH 8.0, [cyclodextrin] = 0.34–0.91 mM. Substrate concentra-
tion was between 0.5–13.2 mM; the reactions were followed by UV (‡) or HPAEC (‡‡).
Eur. J. Org. Chem. 2008, 745–752
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
747

J. Bjerre, E. H. Nielsen, M. Bols
FULL PAPER
tion (Table 2). A moderate degree of catalysis is seen, with substrate at 60 °C (assay conditions as in Table 2), however,
β-cyclodextrin dicyanohydrin (kcat = 2ϫ10–6 s–1). Yet no catalysis was found.
strong binding does indeed occur (KM = 1.69 mM). This
The hydroxycoumarin glucoside fraxin (Figure 3) is the
suggests that much of this binding does not lead to cataly- second largest contributant to Horse Chestnut toxicity.
sis, which makes sense bearing in mind this substrate’s ex- Using HPAEC glucose detection and the same assay condi-
tended aromatic moiety.
tions as for phloridzin, we tested fraxin for β-cyclodextrin
The skimmin analogue 4-phenylskimmin (Figure 3) was dicyanohydrin-catalyzed hydrolysis, which affords fraxetin.
tested at 20 °C both with α- and β-cyclodextrin dicyanohyd- The results revealed low activity with this substrate, though
rin, but no catalysis was observed under these conditions.
the binding appeared to be quite strong (small KM value).
Using a reductive sugar detection assay (see below), we
tested as substrates the compounds n-octyl β--glucopyr-
anoside, n-pentyl β--glucopyranoside and benzyl β--gluc-
opyranoside (Figure 4). We employed as catalysts both α-
and β-cyclodextrin dicyanohydrin, testing at both 25 and
90 °C, but found no catalysis for any of these substrates.
The most likely explanation for the lack of activity is the
poorer leaving group capability of the aliphatic substituent
in these compounds compared to the aromatic substituent
in the good substrates.
Figure 4. Simple glucoside substrates.
Diaminomethyl β-Cyclodextrin
In addition to the cyclodextrin dicyanohydrins employed
in the above-mentioned assays, we also synthesized an
amine analogue hereof. Using essentially the same route of
Figure 3. Compounds that were not substrates.
Amygdalin (Figure 3) is a phenylalanine-derived cyano-
genic glucoside present in apricot kernels and bitter al-
monds (Prunus amygdalus).[33] Amygdalin was once sug-
gested to possess anticancer activity, and in spite of lack of
any evidence supporting this, it remains a popular antican-
cer agent used by people throughout the world, despite the
Scheme 3. Synthesis route for β-cyclodextrin dihydroxyamine (step
risk of cyanide poisoning.[34] We tested this compound as a 1–4 in ref.[32]).
Table 3. The assay was performed with [cyclodextrin] = 0.54 mM. Substrate concentration was between 1–26 mM. The reaction was
followed by UV (400 nm).
kcat [10–7 s–1]
2.24Ϯ0.45
KM [mM]
kcat/kuncat
2.879Ϯ1.83
8
748
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 745–752

Cyclodextrin-Based Artificial Enzymes
synthesis as for the β-cyclodextrin dicyanohydrin,[32] we esculetin have absorption maxima at 340–342 nm. The in-
achieved reduction of the nitrile group to the corresponding spiration came from the field of molecular biology, wherein
aminomethyl group by using an excess of palladium in the a ferric, bile-containing in vitro esculetin detection assay
final de-O-benzylation step (Scheme 3).
has been used for the purpose of classifying Lancefield
The second step selectively de-O-benzylates only the po- group D streptococci into enterococci and non-enterococci,
sitions 6A and 6D.[35]
the latter of which posesses esculinase activity.[38,39] The
The resulting cyclodextrin 3 was tested for catalysis of p- original ferric assay, described by Meulen,[40,41] is a non-
nitrophenyl β--glucopyranoside hydrolysis, but only little quantitative simple test with only two possible outputs, so
activity was found; however, the binding was fairly strong from this, we developed a quantitative continous catechol
(Table 3). The small rate of catalysis is not too surprising, assay.
keeping in mind the poor electron-withdrawing ability of
The vicinal hydroxy groups in esculetin form together
the aminomethyl group compared to the nitrile. This leads with the ferric ion a black-brownish complex, presumed to
to decreased acidity of the alcohol proton and thus less acid have the structure shown in Figure 5.[42] We measured the
catalysis potential.
extinction coefficient for this complex to be
1.85 mM–1 cm–1.
ε =
Reductive Sugar Assay
In order to be able to investigate the hydrolysis of glyco-
sides with non-UV detectable aglycon hydrolysis products,
an alternative non-HPAEC assay for the detection of reduc-
ing sugars was implemented, which does not require ex-
pensive equipment and can even be performed on a large
scale. For this purpose, the Bernfeld assay[36] was applied,
which relies on a 3,5-dinitrosalicylic acid reagent which
upon heating is reduced by the sugar aldehyde to 3-amino-
5-nitrosalicylic acid,[37] which can be detected at 540 nm
(Scheme 4).
Figure 5. Structure of FeIII-esculetin complex.
The development of the assay proceeded in a systematic
order: First, we confirmed that dark coloration was a prod-
uct only of esculetin in combination with FeIII, and no dif-
ferent combination of these and other assay components
gave rise to color change. Then, we performed a series of
UV-measurements under a range of different conditions
and concentrations, in order to determine the appropriate
wavelength range for detection of the FeIII-esculetin com-
plex. From this, we saw that esculin and esculetin domi-
nated the whole 200–400 nm range, but the 450–600 nm
area showed some absorption of the ferric complex, unre-
lated to other components. In order to determine the pre-
cise optimal ferric complex detection wavelength, we per-
formed an assay with β-glucosidase and esculin, with mea-
surement of the increase of absorption at a number of dif-
ferent wavelengths in the 400–600 nm range (50 mM phos-
phate buffer, pH 8.0, 25 °C). From this, 570 nm proved to
be the optimal detection wavelength. In the same assay, we
experimented using different concentrations of FeIII ion
present, and found 30 mM to yield the best results.
We then validated the assay with β-glucosidase and escu-
lin as a substrate. The resulting Hanes plot (Figure 6) is of
good correlation and thus confirms the general validity of
the assay.
We furthermore wanted to check the validity of the assay
using cyclodextrin dicyanohydrins as enzymes. For this pur-
pose, we synthesized the substrate ortho-hydroxyphenyl β-
-glucopyranoside 5 in two steps from β--glucose penta-
acetate (Scheme 5). 5 is presumably cleaved by the cyclodex-
Scheme 4. Bernfeld reducing sugar assay.
Using, for conversion purposes, a prepared standard
curve of glucose concentration vs. absorption, the Bernfeld
assay was used to measure the extent of hydrolysis of simple
glycosides (see Exp. Sect. for details). As this assay is of
general nature and easily conducted, the range of substrates
that can be studied is very broad, allowing for various fu-
ture experiments on a diverse selection of substrates.
Development of Quantitative Catechol Detection Assay
One of the most straightforward and convenient ways to trin in the same fashion as the previously tested nitrophenyl
monitor the kinetics of enzymatic reactions, is by con- glucopyranosides[14]
tinuous UV-measurements in a standard spectrophotome- Using similar conditions as for the β-glucosidase assay
.
ter. We wanted to create an assay methodology applicable (see Exp. Sect.), we found that the catechol assay was useful
for following the hydrolysis of esculin, a reaction which can for monitoring the β-cyclodextrin dicyanohydrin-catalyzed
not be monitored directly by UV since both esculin and hydrolysis of 5, thus confirming the validity of the assay.
Eur. J. Org. Chem. 2008, 745–752
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
749

J. Bjerre, E. H. Nielsen, M. Bols
FULL PAPER
cal modification of the cyclodextrins, we hope in the future
to accomplish improved rates of catalysis and to learn more
about the factors governing catalysis.
Experimental Section
General: Solvents were distilled under anhydrous conditions. All
reagents were used as purchased without further purification.
Evaporation of organic solvents was performed on a rotary evapo-
rator under reduced pressure and with the temperature kept below
40 °C. Columns for flash chromatography were packed with silica
gel 60 (230–400 mesh) as the stationary phase. TLC silica plates
(Merck 60, F254) were visualized by addition of CMOL-visualizer
[cerium sulfate (1%) and molybdic acid (1.5%) in 10% sulfuric
acid] and heating until visible spots appeared, and by UV light
when applicable. 1H and 13C NMR spectroscopic data were re-
corded on a Varian Mercury 400 MHz NMR spectrometer. Chemi-
cal shifts are given in ppm and referred to internal SiMe4 (δH, δC
= 0.00). Optical rotations were recorded at room temperature on a
Perkin–Elmer 241 polarimeter. IR spectra were recorded on a Per-
kin–Elmer FT-IR Paragon 1000. Enzymatic UV-assays were per-
formed using a Spectronic Genesys 5 spectrophotometer.
Figure 6. Hanes plot of β-glucosidase-catalyzed esculin hydrolysis,
in 50 mM phosphate buffer at pH 8.0 and 25 °C, in the presence
of 30 mM FeIII ammonium citrate (R2 = 0.98).
Procedure for Determining the Rate of Hydrolysis by Direct UV
Measurement: Each assay was performed on 50 mM phosphate
buffer samples (1 mL), prepared from aqueous solutions of the ap-
propriate aryl glycoside at different concentrations (0.5 mL) mixed
with phosphate buffer solution (0.5 mL) containing either cyclo-
dextrin catalyst or nothing as control. Concentration of cyclodex-
trin dicyanohydrin catalyst was between 0.34–0.91 mM, and con-
centration of β-cyclodextrin diaminomethyl was 0.54 mM. The re-
actions were followed continuously at 60 °C using UV absorption
at 365 nm for skimmin (ε = 15.02 mM–1 cm–1), 360 nm for 4-
MUGP (ε = 8.92 mM–1 cm–1) and 400 nm for p-nitrophenyl β--
glucopyranoside (ε = 17.59 mM–1 cm–1). The reactions were moni-
tored for 1–3 h. Velocities were determined as the slope of the pro-
gress curve of each reaction. Uncatalyzed velocities were obtained
directly from the control samples. Catalyzed velocities were calcu-
lated by subtracting the uncatalyzed velocity from the velocity of
the appropriate cyclodextrin-containing sample. The catalyzed ve-
locities V were used to determine KM and Vmax from non-linear
regression of V vs. [S] using the program Dataplot; kcat was calcu-
lated as Vmax/[cyclodextrin], kuncat was determined as the slope
from a plot of Vuncat vs. [S].
Scheme 5. Synthesis of substrate for catechol assay validation.
Conclusions
The breakdown and neutralization of toxic glycosides is
an important task that could be performed enzymatically
with great economical and environmental benefits. We have
developed and implemented novel assay procedures and
used these for performing hydrolysis assays of various toxic
substrates with cyclodextrin dicyanohydrins as catalysts. As
expected, little catalysis of p-nitrophenyl β--glucopyrano-
side hydrolysis was found for the analogue diaminomethyl
β-cyclodextrin. Using a newly developed continuous cate-
chol detection assay, a weak degree of catalysis was found
for the toxic natural glycoside esculin with β-cyclodextrin
dicyanohydrin. An equipment-light reductive sugar assay
was used to asses the catalysis of simple benzyl or alkyl
Procedure for Determining the Rate of Hydrolysis by Catechol De-
tection Assay: Each assay was performed as a direct UV-measure-
glucoside substrates, but none of the compounds were ment assay, as described above. 30 mM FeIII ammonium citrate was
present in all cuvettes. UV-absorption was measured continuously
found active. A number of toxic hydroxycoumarin and re-
lated compounds were tested using HPAEC or UV detec-
tion. For skimmin and phloridzin, good to moderate de-
grees of catalysis were found with cyclodextrin dicyanohyd-
rins, with kcat/kuncat values up to 1259. For the artificial
glycoside 4-MUGP, β-cyclodextrin dicyanohydrin was able
at 570 nm for FeIII-esculetin complex (ε = 1.85 mM–1 cm–1).
For assay performed with esculin, substrate concentration was be-
tween 1–10 mM and β-cyclodextrin dicyanohydrin concentration
was 1.0 mM.
For catechol-assay validation with glucoside 5, the substrate con-
centration was between 3–15 mM and the β-cyclodextrin dicya-
nohydrin concentration was 2.5 mM. A linear correlation between
the measured β-cyclodextrin dicyanohydrin-related increase in ab-
sorption and the ortho-hydroxyphenyl β--glucopyranoside con-
centration was found, as expected.
to efficiently catalyze the hydrolysis with an impressive kcat
/
kuncat value of up to 7569. Making artificial enzymes from
scratch and using these successfully to achieve significant
catalysis of chemical and biological reactions is challenging,
but in several cases we observed good enzymatic ability of
cyclodextrin dicyanohydrins. The results are encouraging,
and open up the possibility for alternative uses of modified
Procedure for Determining the Rate of Hydrolysis by HPAEC
(High-Performance Anion-Exchange Chromatography): HPEAC
cyclodextrins as artificial enzymes. Through further chemi- was performed on a Dionex ICS 3000 system equipped with an
750 Eur. J. Org. Chem. 2008, 745–752
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Cyclodextrin-Based Artificial Enzymes
Electrochemical Detector with a gold working electrode and a sil-
ver/silver chloride reference electrode. The column used was a Di-
onex CarboPac PA1 (4ϫ250 mm analytical column).
was added sodium methoxide (25 w/w% in 6.00 mL MeOH,
26.2 mmol) and the mixture was stirred at 5 °C for 2 h. The reac-
tion was quenched by addition of dry ice (until pH 6–7). Methanol
was removed under reduced pressure and the residue was subjected
to ion exchange (IR-120 H). The resulting aqueous solution was
concentrated in vacuo and subsequently dried using lyophilization,
which afforded the desired product (175 mg, 89%). 1H NMR
(400 MHz, [D6]DMSO): δ = 8.00 (d, J = 9.6 Hz, 1 H), 7.65 (d, J
= 8.6 Hz, 1 H), 7.05 (d, J = 2.3 Hz, 1 H), 7.01 (dd, J = 2.4, J =
8.6 Hz, 1 H), 6.33 (d, J = 9.6 Hz, 1 H), 5.38 (d, J = 4.8 Hz, 1 H,
OH), 5.12 (d, J = 4.6 Hz, 1 H, OH), 5.05 (d, J = 5.3 Hz, 1 H, OH),
5.03 (d, J = 7.3 Hz, 1 H), 4.58 (t, J = 5.6 Hz, 1 H, OH), 3.71 (dd,
J = 5.6, J = 10.4 Hz, 1 H), 3.44 (m, 2 H), 3.28 (m, 2 H), 3.17 (td,
J = 7.1, J = 14.1 Hz, 1 H) ppm. 13C NMR (100 MHz, [D6]DMSO):
δ = 160.1, 154.9, 144.1, 129.3, 113.5, 113.1, 113.0, 103.0, 99.8, 77.0,
76.3, 73.0, 69.5, 60.5 ppm. HR-MS (ES): m/z calcd. for
C15H16NaO8: 347.0743, found: 347.0750.
Each assay was performed on 50 mM phosphate buffer samples
(3 mL), prepared from aqueous solutions of the appropriate aryl
glycoside at 4 different concentrations (1.5 mL) mixed with phos-
phate buffer (1.5 mL) containing either cyclodextrin or nothing as
control. The reactions were conducted at 60 °C and samples
(200 µL) were taken out at specified times and cooled with ice. The
sample was injected onto the HPAEC and then eluted at 25 °C at
a flowrate of 1 mL/min. The following gradient system was used:
gradient A (150 mM NaOH), gradient B (500 mM NaOAc in
150 mM NaOH). 0–5 min 100% A, 5.1–11.5 min 100% B and 12–
15 min 100% A. The amount of glucose was determined from a
standard curve which was produced from triple determinations of
6 different concentrations of glucose using the Dionex Chromeleon
software.
(6AR,6DR)-6A,6D-Di-C-Aminomethyl-β-cyclodextrin (3): The prod-
uct was obtained from de-O-benzylation of the corresponding per-
benzylated dicyanohydrin β-cyclodextrin [(6AR,6DR)-6A,6D-Di-C-
cyano-2A–G,3A–G,6B,C,E–G-nonadecakis-O-benzyl-β-cyclodextrin][43]
(1.0 g, 0.3 mmol), which was dissolved in MeOH/EtOAc (1:1,
150 mL). Then Pd/C (5%, 1.5 g) and TFA (7 drops) were added,
and the reaction was left to stir at room temperature under hydro-
gen atmosphere for 2 days. The reaction progress was monitored
by MALDI-TOF-MS. Filtration through a bed of Celite and evap-
oration of the solvent afforded 3 in quantitative yield. 1H NMR
(400 MHz, CDCl3): δ = 5.08–4.76 (m, 7 H, 1-H), 4.31–4.12, (m, 1
H), 4.12–3.90 (m, 1 H), 3.90–3.18 (m, 40 H), 3.18–2.96 (m, 2 H, 2-
H) ppm. 13C NMR (100 MHz, D2O): δ = 101.9, 82.1, 81.4, 81.1,
80.7, 73.2, 72.8, 72.3, 71.8, 71.1, 68.9, 65.2, 65.1, 63.4, 60.7, 60.3,
42.8, 39.5 ppm. MALDI-TOF-MS m/z calcd. for C44H76N2NaO35:
1215.4121, found: 1215.6175.
Procedure for Determining the Rate of Hydrolysis by the Reductive
Sugar Assay: Experiments were performed in 50 mM phosphate
buffer at pH 8.0, with [cyclodextrin] = 0.42 mM. Each of the glyco-
side substrates was tested for hydrolysis at 3 different substrate con-
centrations, with each concentration tested for hydrolysis under 3
different conditions: in the presence of α-cyclodextrin dicyanohyd-
rin, β-cyclodextrin dicyanohydrin and with buffer only, respectively.
All of the above-mentioned measurements were furthermore per-
formed at two different temperatures: 25 and 90 °C. For each ex-
periment, a blank (water-containing) sample was included. A repre-
sentative selection of the experiments were reconducted under iden-
tical conditions to assess validity and reproducability of the data.
Reaction mixtures (with differering amounts of substrate and
cyclodextrin) of 3 mL volume were left stirring in sealed glass vials
in a heating block or at room temperature. At regular time intervals
(hours) a small amount (0.35 mL) of each reaction mixture was
added an equal volume (0.35 mL) of 3,5-dinitrosalicylic acid test
solution. The assay mixture was then incubated at 100 °C for 5 min.
The developed assay mixture was cooled to room temperature, di-
luted with water (3.5 mL), and the absorbance at 540 nm was mea-
sured. By comparing to a standardcurve of absorbance vs. glucose
concentration, the hydrolysis progress was monitored.
ortho-Hydroxyphenyl
2,3,4,6-Tetra-O-acetyl-β-D-glucopyranoside
(4): β--Glucose pentaacetate (2.00 g, 5.12 mmol) and pyrocatechol
(621 mg, 5.64 mmol, 1.1 equiv.) were dissolved in dry dichloro-
methane (10 mL), in the presence of molecular sieves (4 Å, 5.0 g).
Boron trifluoride–diethyl ether (0.64 mL, 5.12 mmol, 1 equiv.) was
added dropwise. The reaction was stirred overnight at room tem-
perature under nitrogen atmosphere. Reaction progress was moni-
tored by TLC (silica, eluent diethyl ether/pentane 1:1). The reaction
mixture was filtered and to the filtrate was added dichloromethane
(100 mL). The organic layer was washed with sat. aq. NaHCO3
(3ϫ40 mL), 0.5  NaOH (3ϫ33 mL) and water, dried (MgSO4),
filtered and concentrated in vacuo, affording the crude product as
a colorless foam. Partial purification of the crude product was per-
formed by flash chromatography (eluent gradient, EtOAc/pentane
1:3 Ǟ EtOAc), affording primarely the desired product as a color-
less solid (340 mg, 0.77 mmol), upon which the next step of the
synthesis was based.
7-[(2,3,4,6-Tetra-O-acetyl-β-D-galactopyranosyl)oxy]-2H-1-benzo-
pyran-2-one (1): To a solution of 2,3,4,6-tetra-O-acetyl-α--gluco-
pyranosyl bromide (2.00 g, 4.86 mmol) and 7-hydroxycoumarin
(790 mg, 4.87 mmol) in acetonitrile (15 mL) was added Ag2O
(1.35 g, 5.83 mmol) and the mixture was stirred overnight at room
temperature in the absence of light. The mixture was then diluted
with EtOAc (20 mL), filtered through a bed of Celite and washed
with EtOAc (3ϫ20 mL). The filtrate was washed with aq. KOH
(0.25 , 2ϫ20 mL), brine (20 mL), and water (20 mL), dried
(MgSO4) and concentrated under reduced pressure. The crude
product was recrystallized from methanol which afforded the de-
sired product as white crystals (626 mg, 26%). 1H NMR (400 MHz,
CDCl3): δ = 7.64 (d, J = 9.6 Hz, 1 H), 7.39 (d, J = 8.6 Hz, 1 H),
6.95 (d, 1 H), 6.90 (dd, J = 2.3, J = 8.5 Hz, 1 H), 6.31 (d, J =
9.5 Hz, 1 H), 5.36–5.26 (m, 2 H), 5.19–5.13 (m, 2 H), 4.28 (dd, J
= 5.8, J = 2.3 Hz, 1 H), 4.18 (d, J = 12.3 Hz, 1 H), 3.96–3.88 (m,
1 H), 2.11 (s, 3 H), 2.06 (s, 6 H), 2.03 (s, 3 H) ppm. 13C NMR
(100 MHz, CDCl3): δ = 170.3, 169.9, 169.3, 169.1, 160.4, 159.2,
155.2, 143.0, 128.8, 114.3, 114.2, 114.0, 103.8, 98.1, 72.4, 72.1, 70.8,
68.0, 61.7, 20.5, 20.4 ppm. HR-MS (ES): m/z calcd. for
C23H24NaO12: 515.1166, found: 515.1166.
ortho-Hydroxyphenyl β-D-Glucopyranoside (5): ortho-Hydroxy-
phenyl 2,3,4,6-tetra-O-acetyl β--glucopyranoside (4) (340 mg,
0.772 mmol) was dissolved in dry methanol (8 mL), sodium meth-
oxide (3 mL, freshly made from sodium and methanol) was added
to the solution. The reaction mixture was stirred at room tempera-
ture and the reaction progress was monitored by TLC (silica, eluent
EtOAc/pentane 1:3). After 3.5 h the reaction mixture was made
slightly acidic by addition of acetic acid to the mixture, after which
all solvent was removed in vacuo. The product was purified by flash
chromatography (eluent EtOAc/MeOH 5:1), which afforded the
pure and desired product (141 mg, 0.518 mmol, 67%) as a colorless
7-(β-D-Glucopyranosyloxy)-2H-1-benzopyran-2-one (Skimmin) (2):
To a solution of 1 (0.30 g, 0.61 mmol) in dry methanol (40 mL)
powder. IR (film): ν = 3419, 2925, 1600, 1501, 1465, 1375, 1268,
˜
Eur. J. Org. Chem. 2008, 745–752
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
751

J. Bjerre, E. H. Nielsen, M. Bols
FULL PAPER
1204, 1109, 1069, 1015, 791, 731 cm–1. 1H NMR (400 MHz,
CDCl3): δ = 7.09 (dd, 3J = 8.0, 4J = 1.6 Hz, 1 H, 3Ј-H/6Ј-H), 6.96–
6.82 (m, 3 H, 3Ј-H/6Ј-H, 4Ј-H, 5Ј-H), 4.97 (d, J1,2 = 7.6 Hz, 1 H,
1-H), 3.80 (dd, J5,6a = 2.4, J6a,6b = 12.4 Hz, 1 H, 6a-H), 3.65 (dd,
J5,6b = 5.2, J6a,6b = 12.4 Hz, 1 H, 6b-H), 3.54–3.38 (m, 4 H, 2-H,
3-H, 4-H, 5-H) ppm. 13C NMR (100 MHz, CDCl3): δ = 145.7,
144.7 (C-1Ј, C-2Ј), 124.3, 121.3, 117.0, 116.7 (C-3Ј, C-4Ј, C-5Ј, C-
6Ј), 101.3 (C-1), 76.3, 75.6, 73.0, 69.5 (C-2, C-3, C-4, C-5), 60.6 (C-
6) ppm. HR-MS (ES): m/z calcd. for C12H16NaO7: 295.0794,
found: 295.0802.
[18]
[19]
P. Travascio, D. Sen, A. J. Bennet, Can. J. Chem. 2006, 84, 613–
619.
K. Kikuchi, R. B. Hannak, M. J. Guo, A. J. Kirby, D. Hilvert,
Bioorg. Med. Chem. 2006, 14, 6189–6196.
Z. L. Fang, R. Breslow, Org. Lett. 2006, 8, 251–254.
N. M. Milovic, J. D. Badjic, N. M. Kostic, J. Am. Chem. Soc.
2004, 126, 696–697.
[20]
[21]
[22]
[23]
L. G. Marinescu, M. Bols, Angew. Chem. Int. Ed. 2006, 45,
4590–4593.
For a new method of modifying the secondary rim, see: M.
Fukudome, A. Matsushima, D. Q. Yuan, K. Fujita, Tetrahe-
dron Lett. 2006, 47, 6599–6602.
C. Rousseau, F. Ortega-Caballero, L. U. Nordstrøm, B. Chris-
tensen, T. E. Petersen, M. Bols, Chem. Eur. J. 2005, 11, 5094–
5101.
F. Ortega-Caballero, J. Bjerre, L. S. Laustsen, M. Bols, J. Org.
Chem. 2005, 70, 7217–7226.
J. Bjerre, T. H. Fenger, L. Marinescu, M. Bols, Eur. J. Org.
Supporting Information (see also the footnote on the first page of
this article): NMR spectrum of compound 3.
[24]
Acknowledgments
[25]
[26]
This work has been supported by the Lundbeck Foundation and
The Danish National Science Research Council.
Chem. 2007, 704–710.
[27]
[28]
K. Taeufel, Nahrung 1968, 12, 275–282.
M. Runkel, M. Bourian, W. Legrum, Fruit Processing 1997, 7,
213–216.
[1] E. Weinhouse, J. Kaplanski, A. Danon, D. B. Nudel, Life Sci.
1989, 44, 441–450.
[2] C. G. Hanratty, P. McGlinchey, D. Johnston, A. P. Passmore,
Drug. Aging 2000, 17, 353–362.
[3] C. Hamouda, A. Hedhili, N. Ben Salah, M. Zhioua, M. Ama-
mou, Vet. Hum. Toxicol. 2004, 46, 144–146.
[4] G. Stanic´, B. Jurisˇic´, D. Brkic´, Croat. Chem. Acta 1999, 72,
827–834.
[29]
[30]
S. Oh, D. Kang, Appl. Environ. Microbiol. 2004, 70, 5692–5694.
L. B. Daniels, P. J. Coyle, R. H. Glew, N. S. Radin, R. S. La-
bow, Archives of Neurology 1982, 39, 550–556.
V. Crespy, O. Aprikian, C. Morand, C. Besson, C. Manach, C.
Demigné, C. Remésy, J. Nutr. 2001, 3227–3230.
V. L. Singleton, F. H. J. Kratzer, Agr. Food. Chem. 1969, 17,
497–512.
Y. P. D avo l , Indian J. Biochem. 1967, 4, 54–55.
G. W. Newton, E. S. Schmidt, J. P. Lewis, E. Conn, R. Law-
rence, Western J. Med. 1981, 134, 97–103.
A. J. Pearce, P. Sinaÿ, Angew. Chem. Int. Ed. 2000, 39, 3610–
3612.
P. Bernfeld, Adv. Enzymol. Relat. Areas Mol. Biol. 1951, 12,
379–428.
L. D. Frost, Chem. Educ. 2004, 9, 239–241.
S. M. H. Qadri, M. I. DeSilva, S. Zubairi, J. Clin. Microbiol.
1980, 12, 472–474.
[31]
[32]
[5] Z. Stefanova, H. Neychev, N. Ivanovska, I. Kostova, J. Ethno-
pharmacol. 1995, 46, 101–106.
[33]
[34]
[6] M. J. Martín, C. Alarcón, V. Motilva, Ann. Pharm. Fr. 1990,
48, 306–311.
[7] M. V. Rekharsky, Y. Inoue, Chem. Rev. 1998, 98, 1875–1917.
[8] J. Szejtli, Comprehensive Supramolecular Chem. 1996, 3, 5–40.
[9] V. J. Stella, R. A. Rajewski, Pharm. Res. 1997, 14, 556–567.
[10] J. Szejtli, Chemical Intelligencer 1999, 5, 38–45.
[11] H. J. Buschmann, D. Knittel, C. Jonas, E. Schollmeyer, Lebens-
mittelchemie 2001, 55, 54–56.
[12] B. Sueur, L. Leclercq, M. Sauthier, Y. Castanet, A. Mortreux,
H. Bricout, S. Tilloy, E. Monflier, Chem. Eur. J. 2005, 11, 6228–
6236.
[13] D. Kirschner, T. Green, F. Hapiot, S. Tilloy, L. Leclercq, H.
Bricout, E. Monflier, Adv. Synth. Catal. 2006, 348, 379–386.
[14] A. Cassez, A. Ponchel, F. Hapiot, E. Monflier, Org. Lett. 2006,
8, 4823–4826.
[15] R. Breslow, J. Phys. Org. Chem. 2006, 19, 813–822.
[16] E. Kassianidis, D. Philp, Angew. Chem. Int. Ed. 2006, 45, 6344–
6348.
[35]
[36]
[37]
[38]
[39]
[40]
S. C. Edberg, R. W. Trepeta, C. M. Kontnick, A. R. Torres, J.
Clin. Microbiol. 1985, 21, 363–365.
F. C. Harrison, J. van der Leck, Zentralbl. Bakteriol. Para-
sitenkd. Infektionskr. Hyg. 1909, 22, 547–551.
A. Miskin, S. C. Edberg, J. Clin. Microbiol. 1978, 7, 251–254.
B. D. Jain, H. B. Singh, J. Indian Chem. Soc. 1963, 40, 263–
264.
F. Ortega-Caballero, C. Rousseau, B. Christensen, T. E. Pe-
tersen, M. Bols, J. Am. Chem. Soc. 2005, 127, 3238–3239.
Received: October 9, 2007
[41]
[42]
[43]
[17] W. B. Motherwell, C. E. Atkinson, A. E. Aliev, Y. F. Stephanie,
B. H. Wong, D. Warrington, Angew. Chem. Int. Ed. 2004, 43,
1225–1228.
Published Online: December 3, 2007
752
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 745–752