Glycosidase-catalyzed hydrolysis of 2-deoxyglucopyranosyl pyridinium salts: Effect of the 2-OH group on binding and catalysis

Article abstract of DOI:10.1139/v00-061

Three 2-deoxy-α-D-glucopyranosyl pyridinium tetrafluoroborates were tested for their binding affinity to a range of α-glucosidases and α-mannosidases. The α- isoquinolinium salt (11) binds approximately 275-fold more tightly to yeast α-glucosidase than does the isomeric quinolinium salt (12). In addition, compound 11 binds to the yeast enzyme approximately two-fold tighter than the corresponding glucopyranosyl isoquinolinium salt (9). The (k(cat)/k(hyd)) values for the yeast α-glucosidase-catalyzed reactions of 11 and 9 are 1.6 x 10⁵ and 2.0 x 10⁹, respectively, when compared to the spontaneous uncatalyzed reactions. Thus, the interaction of the 2-OH group in compound 9 with the yeast enzyme's active site generates a relative transition state stabilization of about 23.5 kJ mol^-1. For both compounds 11 and 12, the observed rate accelerations for the yeast α-glucosidase-catalyzed hydrolysis, relative to the spontaneous reaction in solution, (k(cat)/k(hyd) are identical within experimental error.

Full text of DOI:10.1139/v00-061

577
Glycosidase-catalyzed hydrolysis of
2-deoxyglucopyranosyl pyridinium salts: effect
of the 2-OH group on binding and catalysis
Kelly S. E. Tanaka, Jiang Zhu, Xicai Huang, Francesco Lipari,
and Andrew J. Bennet
Abstract: Three 2-deoxy-α-D-glucopyranosyl pyridinium tetrafluoroborates were tested for their binding affinity to a
range of α-glucosidases and α-mannosidases. The α-isoquinolinium salt (11) binds approximately 275–fold more
tightly to yeast α-glucosidase than does the isomeric quinolinium salt (12). In addition, compound 11 binds to the
yeast enzyme approximately two-fold tighter than the corresponding glucopyranosyl isoquinolinium salt (9). The
(kcat/khyd) values for the yeast α-glucosidase-catalyzed reactions of 11 and 9 are 1.6 × 105 and 2.0 × 109, respectively,
when compared to the spontaneous uncatalyzed reactions. Thus, the interaction of the 2-OH group in compound 9 with
the yeast enzyme’s active site generates a relative transition state stabilization of about 23.5 kJ mol–1. For both com-
pounds 11 and 12, the observed rate accelerations for the yeast α-glucosidase-catalyzed hydrolysis, relative to the spon-
taneous reaction in solution, (kcat/khyd) are identical within experimental error.
Key words: glycosidase, inhibitor, 2-deoxyglucose, pyridinium, catalysis.
Résumé : On a évalué les affinités de liaison de trois fluoroborates de 2-désoxy-α-D-glucopyranosyl pyridinium avec
un ensemble d’α-glucosidases et d’α-mannosidases. Le sel d’α-isoquinoléinium (11) se lie approximativement 275 fois
fortement à l’α-glucosidase de la levure que le sel quinoléinium isomère (12). De plus, le composé 11 se lie à
l’enzyme de la levure approximativement deux fois mieux que le sel du glucopyranosyl isoquinoléinium correspondant
(9). Les valeurs de (kcat/khyd) pour les réactions des composés 11 et 9 catalysées par l’α-glucosidase de la levure sont
égales respectivement à 1,6 × 105 et 2,0 × 109 lorsqu’on les compare aux réactions spontanées non catalysées. On en
déduit que l’interaction du groupe 2-OH du composé 9 avec le site de l’enzyme de la levure génère une stabilisation
relative de l’état de transition d’environ 23,5 kJ mol–1. Pour chacun des composés 11 et 12, les accélérations de vitesse
observées pour les hydrolyses catalysées par l’α-glucosidase de la levure, comparées à la réaction spontanée en solu-
tion (kcat/khyd) sont identiques aux erreurs expérimentales près.
Mots clés : glucosidase, inhibiteur, 2-désoxyglucose, pyridinium, catalyse. Tanaka et al 582
Introduction
an endocyclic amino group, the so-called “imino-sugars” (4),
a notable example of which is 1-deoxynojirimycin (DNJ, 1)
a compound that has the same hydroxyl configuration as 1-
deoxyglucose (5). In general, imino-sugars are good inhibi-
tors of glycosidase enzymes, probably due to protonation at
physiological pH of the secondary amine of the imino-sugar,
thus generating a cation mimic of the glucosyl oxacarben-
ium ion (2) which emulates the glycosylation transition state
structure (3, 6).
Recent advances in the rapidly growing field of glyco-
biology have provided insights into the fundamental impor-
tance of glycoproteins, glycolipids, and oligosaccharides in a
myriad of biological events (1). Clearly, glycoconjugate-
processing enzymes play a crucial role in the scheme of
these various biological events (2). As a key subclass of
glycoconjugate-processing enzymes, glycosidases are the fo-
cus of ongoing research aimed at the discovery of new gly-
cosidase inhibitors (3). One class of naturally occurring
glycosidase inhibitors are carbohydrate mimics that contain
New glycosidase inhibitors isolated from natural sources
are often prototypes for the design and synthesis of ana-
logues as potential therapeutic agents. For example,
Received October 15, 1999. Published on the NRC Research Press website on May 8, 2000.
K.S.E. Tanaka,1 J. Zhu, X. Huang,2 and A.J. Bennet.3 Department of Chemistry, Simon Fraser University, Burnaby, British
Columbia, V5A 1S6, Canada.
F. Lipari.4 McGill Cancer Centre, McGill University, 3655 Drummond Street, Montréal, QC H3G 1Y6, Canada.
1Current address: Department of Biochemistry, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461,
U.S.A.
2Current address: Conjuchem, 225 President-Kennedy Avenue, Montréal, QC H2X 3Y8, Canada.
3Author to whom correspondence may be addressed. Telephone: (604) 291-3532. Fax: (604) 291-3765. e–mail: bennet@sfu.ca
4Current address: Boehringer Ingelheim (Canada) Ltd. BioMéga Research Division, 2100 Cunard Street, Laval, QC H7S 2G5,
Canada.
Can. J. Chem. 78: 577–582 (2000)
© 2000 NRC Canada

578
Can. J. Chem. Vol. 78, 2000
Scheme 1.
isofagomine (4) an analogue of the natural product fagomine
(3) isolated from buckwheat seeds (7) has recently been syn-
thesized and shown to be a potent inhibitor (Ki = 0.11 µM)
of the almond-emulsion β-glucosidase (8). Yet another re-
lated compound, 1-azafagomine (5) is a good inhibitor of
both yeast α-glucosidase (Ki = 3.9 µM) and almond-
emulsion β-glucosidase (Ki = 0.65 µM) (9).
A separate class of nitrogen-containing natural products is
the “amino-sugars” which possess an exocyclic rather than
endocyclic amino group (4). One such family of amino-
sugar glycosidase inhibitors is the validamycins. which con-
tain the critical valienamine structural motif (6). An example
of this type of glycosidase inhibitor is the tight binding natu-
ral product acarbose (10, 11). Various epimeric analogues of
valienamine, i.e., epi-valienamine (7) have also been synthe-
sized and tested as potential β-glycosidase inhibitors (12).
Using an alternative strategy, Knapp et al. (13) explored
the effectiveness of substituted 3-hydroxypyridines (8) rather
than the more customary carbohydrate-based compounds as
glycosylation TS mimics. The tightest binding of these aro-
matic compounds (8, X = H) was shown to be a weak
competitive inhibitor (Ki = 0.8 mM) of Agrobacterium β-
glucosidase (13).
Natural substrates generally are bound much more weakly
to enzymes than the corresponding natural inhibitors (14),
and as a result of this, the design of potential reversible
glycosidase inhibitors rarely, if ever, exploit the modification
of a natural substrate structure. Most glycosidase-catalyzed
reactions are, however, monitored using nonnatural substrates
that contain an aromatic aglycon leaving group. In 1985,
Hosie and Sinnott (15) reported that yeast α-glucosidase ef-
ficiently catalyzes the hydrolysis of α -D-glucopyranosyl
pyridinium bromides with an apparent binding constant (Km)
for the isoquinolinium substrate (9) of 6.6 µM. Thus, this
man-made substrate binds to yeast α -glucosidase more
tightly than the naturally occurring inhibitor DNJ (1) (Ki =
13 µM) (16).
Given that the C-2 hydroxyl-group of glycosides is critical
for efficient glycosidase-catalyzed hydrolysis of carbohydrate
substrates (17), we decided to test whether 2-deoxyglucopy-
ranosyl pyridinium salts would retain the tight binding inter-
actions exhibited by their 2-hydroxy counterparts, while
simultaneously becoming more resistant to enzyme-catalyzed
hydrolysis. Accordingly, the 2-deoxy-α- and β-glucopyranosyl
pyridinium salts 10, 11, 12, and 13 were synthesized, and
their binding affinities and hydrolytic stabilities were evalu-
ated for various glucosidase and mannosidase enzymes.
Results and discussion
The route used for the synthesis of 12 is shown in
Scheme 1. This synthetic scheme involves a silver-promoted
SN2 reaction and is similar to the synthetic route employed
for the production of both 10 and 11 (18).
Presented in Table 1 are the inhibition constants (Ki) mea-
sured for the interaction of 10, 11, 12, and 13 with several
© 2000 NRC Canada

Tanaka et al
579
Table 1. Dissociation constants (µM) measured for the binding of 2-deoxyglucopyranosyl pyridinium salts to various glycosidases.a
Enzyme and Source
10
11
12
13
α-Glucosidase — yeastb
α-Glucosidase — riced,e
β-Glucosidase — almondsb
α-Mannosidase — almondsb
α-Mannosidase — yeastd
107 (8)
141 (52)
—
2.9 (0.5)
44 (7)
—
800c
NIf
—
—
—
—
—
7400 (800)
–
—
31000g
8000h
NIi
NIj
aData were analyzed using nonlinear least-squares regression; estimate of standard error in brackets.
bTemperature = 25°C.
cDetermined as a Km value that was estimated using five concentrations of 12 between 0.112 mM and 2.52 mM, estimated error 25%.
dTemperature = 37°C.
eSlow onset to maximal binding.
fNo inhibition detected at 1.1 mM.
gValue estimated using two concentrations of 10 (3.5 and 8.4 mM); estimated error 15%.
hValue estimated using two concentrations of 11 (1.3 and 2.5 mM); estimated error 40%.
iNo inhibition detected at 5.0 mM.
jNo inhibition detected at 1.0 mM.
Table 2. Rate constants (kcat and khyd) for the glycosidase-catalyzed and the spontaneous hydrolyses of 2-deoxyglucopyranosyl
pyridinium salts at 25°C.
9
11
12
5.5c
3.14 × 10–5 f
1.8 × 105
Yeast α-glucosidase (kcat) s–1
0.52a
2.6 × 10–10 d
2.0 × 109
7.1 × 10–3b
4.58 × 10–8 e
1.6 × 105
Spontaneous hydrolysis (khyd) s–1
Rate acceleration (kcat/khyd
)
aValue taken from ref. 15.
bCalculated from relative kcat values measured in comparison to compound 9 (15), estimated error 5%.
cCalculated from relative kcat values measured in comparison to compound 9 (15), estimated error 25%.
dExtrapolated from a pKa(B–H+) value of 5.38 and eq. [3] taken from ref. 22a.
eValue extrapolated from kinetic data measured between 55–95°C.5
fValue measured directly at 25°C.5
glycosidases. A Lineweaver–Burk plot of the kinetic data for
yeast α-glucosidase-catalyzed hydrolysis of 4-nitrophenyl α-
D-glucopyranoside in the presence of various concentrations
of 11 is illustrated in Fig. 1. When detailed kinetic measure-
ments were possible (Ki < 200 µM), only competitive inhibi-
tion of the glycosidases by the test compounds was
observed. Consequently, these compounds are binding in a
reversible manner to the enzyme active site. Table 2 presents
the measured kcat values for yeast α-glucosidase-catalyzed
hydrolysis of 9, 11, and 12. Also given in Table 2 are the
corresponding spontaneous first-order reaction rates (khyd).
Hosie and Sinnott (15) proposed that yeast α-glucosidase-
catalyzed hydrolysis of glucopyranosyl pyridinium salts oc-
curs via the mechanism depicted in Scheme 2, a process in
which glucosylation (kgluc) is rate-limiting for kcat, and a
kinetically significant nonchemical step (k2) is rate-limiting
for kcat/Km.
The same scheme must hold for the reactions of the 2-
deoxy salts, since: (i) kcat for 12 is about 800–fold greater
than kcat for 11 (Table 2), thus kdegluc cannot be rate limiting
(i.e., kdegluc > kgluc); and (ii) kcat/Km for 11 and 12 are within
a factor of three of each other and consequently, kgluc is not
the major rate limiting step for kcat/Km. The apparent Km for
an enzyme-catalyzed reaction is given by eq. [1], where
Σ[ES] is the sum of all enzyme bound species (14). Thus,
the only two enzyme-bound species which contribute to the
apparent Km or Ki values are ES and ES′ (Scheme 2).
[E][S]
[1]
Km =
∑[ES]
Notably, the apparent dissociation constants measured for
binding of 11 (Ki = 2.9 µM) and its 2-OH analogue [Km
=
6.6 µM (15); Km = 7.7 µM (19)] to yeast α-glucosidase indi-
cate that removal of the 2-hydroxyl moderately enhances the
free energy for association by about 2 kJ mol–1 at 25°C.
The orientation of the aromatic ring is a critical variable
for binding affinity of these compounds to the yeast enzyme.
Specifically, fusion of a second aromatic ring onto the
pyridinium compound (10) gives 11 and 12, and this change
causes either an increase of 8.9 kJ mol–1 (11) or a decrease
of 5.0 kJ mol–1 (12) in the free energy of binding to the en-
zyme. An identical trend in Ki values (11 < 10 < 12) is also
apparent for inhibition of the rice α-glucosidase, despite the
attenuated potency of these compounds with this enzyme
(Table 1).
It is presumed that the origin of the tight binding between
the yeast enzyme and 11 results from strong hydrophobic in-
teractions between the active-site and the isoquinoline ring.
Lemieux (20) has stressed the important contribution to
binding free energy that occurs as ordered water molecules
5 J. Zhu and A.J. Bennet. Unpublished data.
© 2000 NRC Canada

580
Can. J. Chem. Vol. 78, 2000
Scheme 2.
Fig. 1. Lineweaver–Burk plot for the inhibition of yeast α-gluco-
sidase-catalyzed hydrolysis of 4-nitrophenyl α-D-glucopyranoside
(S) by 2-deoxy-α-D-glucopyranosyl isoquinolinium tetrafluoro-
borate; symbols represent different concentrations of 11: ᭜ 1.57 µM,
᭹ 3.77 µM, and ᭢ 6.91 µM. Kinetic data measured at concen-
trations of 11 of 0.00 and 0.63 µM have been omitted for clarity.
Scheme 3.
salts have not been reported, no definite conclusions con-
cerning the mannosidase enzymes can be made at this time.
The yeast α-glucosidase hydrolysis of compound 11 is
roughly 73 times less than for the glucosyl salt 9 (Table 2), a
rate difference that corresponds to a ∆∆G‡ (Scheme 3) of
kcat
around 10.6 kJ mol–1. A similar ∆∆G‡
of 9.7 kJ mol–1 is
kcat
observed for barley α-glucosidase-catalyzed hydrolysis of
maltose (23). This similarity is not too surprising given that
both enzymes are members of glycosidase family 31 (24).
For members of other glycosidase families, the removal of a
2-OH group from the substrate can lead to larger differences
surrounding nonpolar surfaces in solution are liberated into
bulk solution when a hydrophobic moiety binds to a biologi-
cal receptor. Carbohydrate–protein interactions of this nature
have been discussed by Wong (21) in terms of providing a
favourable entropic contribution towards the free energy for
binding of hydrophobic groups.
in ∆∆G‡
.
For example, the β-glucosidase from
kcat
Agrobacterium gives a value for ∆∆G‡ of 22 kJ mol–1 for
kcat
the hydrolysis of 4-nitrophenyl β-D-glucopyranoside when
compared to its 2-deoxy derivative (25).
Another potential source of the tight binding of these
compounds may be that their ground state conformations are
At the present time, a more detailed analysis of the mag-
4
1S3 skew-boats rather than the usual C1 chair (18, 22), and
nitude of ∆∆G‡
is unwarranted since X-ray crystallo-
kcat
thus, if these α-glucosidase enzymes bind substrates in non-
chair conformations, it is to be expected that conformat-
ionally biased compounds such as 9, 10, and 11 will display
lower binding constants (Kms) than the natural substrates.
The β-epimer of 11, 2-deoxy-β-D-glucopyranosyl isoquino-
linium bromide (13), is a mediocre inhibitor of almond-emul-
sion β-glucosidase (Ki = 7.4 mM). Therefore, it is clear that
the attachment of an extended aromatic ring onto the
anomeric carbon of a carbohydrate analogue will not guaran-
tee an increase in binding affinity of glucosidase inhibitors.
Both mannosidases investigated, namely the retaining en-
zyme from almond and the inverting enzyme from yeast,
showed either weak or no affinity for the 2-deoxyglucopy-
ranosyl pyridinium salts, suggesting that the 2-OH group is
important for binding to α-mannosidases. However, since ki-
netic studies using the parent α-mannopyranosyl pyridinium
graphic 3-D structure analysis has not been reported for any
member of the α-glucosidase family 31.
The spontaneous hydrolysis of 11 occurs approximately
176-fold more rapidly than that of the corresponding
glucoside (Table 2) (18, 22a). From this, it can be concluded
that removal of the 2-OH group reduces ∆∆G‡ by approx-
stab
imately 23.5 kJ mol–1 (i.e., 176 × 79–fold) for the yeast en-
zyme-catalyzed process. Furthermore, yeast α-glucosidase
accelerates the hydrolysis of both 11 and 12 (the best and
worst inhibitors, respectively) by similar amounts (Table 2).
Therefore, the differential hydrophobic interactions that oc-
cur in the enzyme-substrate complexes for 11 and 12 are
maintained at their respective α-glucosidase-catalyzed hy-
drolysis TSs, and the free energy differences associated with
the measured kcat/khyd rate accelerations (∆G‡
– ∆G‡
,
kcat
Scheme 3) are indistinguishable. In other words, the nonshpyed-
© 2000 NRC Canada

Tanaka et al
581
cific, hydrophobic interactions of the quinoline (and
isoquinoline) moiety with the yeast enzyme’s active site are
of a similar magnitude in both the enzyme–substrate com-
plex and the rate limiting glucosidation transition state, and
as a consequence, the observed rate acceleration must be
caused by an increase in binding of the 2-deoxyglucose unit
to the glucosidase-catalyzed transition state relative to the
E–S complex.
for 15 min, the resulting solution was neutralized by the ad-
dition of I-120 Amberlite resin (H+-form) (2.5 g). After an
additional 3 min, a solution of HBF4 in ether (35% w/v,
1 mL) was added and the resultant acidic solution was fil-
tered directly into cold anhydrous ether (–78°C, 250 mL).
The ensuing cloudy solution was kept in a freezer (–16°C)
for 24 h. After decanting the solvent, addition of acetonitrile
(5 mL) to the solid residue gave a clear solution. Subsequent
addition of anhydrous ether (250 mL) gave a cloudy solution
and this suspension was placed into a freezer (–16°C). After
48 h, the resultant colourless, hygroscopic solid was filtered
Materials and methods
and dried to give an analytically pure sample of 12 (0.65 g,
All reagents and procedures used for the measurement of
enzymatic binding constants with compounds 10, 11, and 13
were identical to those described previously (26, 27).
Quinoline was purchased from Aldrich and purified by
recrystallization of its hydrogen sulfate salt from HOAc–
Et2O, followed by neutralization and fractional distillation at
atmospheric pressure.
All new compounds were fully characterized using NMR
spectroscopy. All observed resonances in the NMR spectra
were fully assigned using 1H–homonuclear (28) and 1H–13C-
heteronuclear (29) chemical shift correlated NMR spectros-
copy techniques. All coupling constants (J) are listed in
hertz (Hz). Synthesis of the substrate α-D-glucopyranosyl 4′-
bromoisoquinolinium bromide (19) and the 2-deoxy com-
pounds 10 (18), 11 (18), and 13 (30) were accomplished ac-
cording to published procedures.
1
71%). H NMR (400 MHz, D2O) δ: 2.63 (ddd, 1 H, J2a,1
5.3, J2a,2e = 15.4, J2a,3 = 9.9, H-2a), 2.92 (dt, 1 H, J2e,1 + J2e,3
=
=
8.3, H-2e), 3.41 (ddd, 1 H, J5,6a = 2.3, J5,6b = 5.7, J5,4 = 8.6,
H-5), 3.67 (dd, 1 H, J6a,6b = 12.5, H-6a), 3.71 (t, 1 H, J4,3
+
J4,5 = 16.0, H-4), 3.82 (dd, 1 H, H-6b), 4.25 (ddd, 1 H, J3,2e
= 4.3, J3,4 = 7.6, H-3). 7.16 (brt, 1 H, J2e,1 = 4.0, H-1), 7.99
(brt, 1 H, J5′,6′ + J6′,7′ = 15.4, ArH-6′), 8.07 (dd, 1 H, J2′,3′
=
6.2, J3′,4′ = 8.1, ArH-3′), 8.22 (brt, 1 H, J6′,7′ + J7′,8′ = 16.2,
ArH-7′), 8.35 (brd, 1 H, J5′,6′ = 8.2, ArH-5′), 8.55 (brd,
J7′,8′ = 9.2, ArH-8′), 9.16 (brd, 1 H, ArH-4′), 9.43 (brd, 1 H,
ArH-2′). Anal. calcd. for C15H18BF4NO4: C 49.62, H 5.00,
N 3.86; found: C 49.46, H 5.19, N 4.09.
Measurement of kcat values
The kcat value for yeast α-glucosidase-catalyzed hydroly-
sis of 11, relative to that for 9, was measured under satura-
tion conditions ([11] = 303 µM and [9] = 110 µM) in 50 mM
phosphate buffer (pH 6.8, 0.1% w/v BSA) at 25°C. The rela-
tive rate of formation of isoquinoline from the two substrates
was monitored at 337 nm. Yeast α-glucosidase-catalyzed hy-
drolysis of 12 was monitored (using an identical enzyme
concentration to that from the experiments with both 9 and
11) at 320 nm in 50 mM phosphate buffer (pH 6.8, 0.1%
w/v BSA) at 25°C. The kinetic parameters Vmax and Km were
calculated using a nonlinear least-squares fit of the initial
rate vs. concentration data. The relative kcat value for α-
glucosidase-catalyzed hydrolysis of 12 was calculated from
the observed Vmax values and the measured ∆(ε) values for
11 and 12 at the respective wavelengths used to monitor
their hydrolysis reactions (11, ∆(ε) = 2010 at 337 nm; 12,
∆(ε) = 2770 at 320 nm).
3,4,6-Tri-O-acetyl-2-deoxy-α-D-arabino-hexopyranosyl
quinolinium tetrafluoroborate (14)
Silver tetrafluoroborate (0.63 g, 3.1 mmol) was added to a
solution of the 2-thiono-1,3-dioxa-2-phosphorinane 15 (18)
(1.50 g, 3.1 mmol) in quinoline (2 mL) and dichloromethane
(1.5 mL). After the solution had stirred at rt for 2 h, the sil-
ver salts were precipitated by the addition of methanol
(250 mL). Following filtration, the solvent was removed un-
der reduced pressure. The resulting residue was dissolved in
a minimum volume of methanol and subsequent addition of
diethyl ether (500 mL) resulted in precipitation of the prod-
uct. This procedure was repeated and the final colourless
powder was crystallized from methanol–ether to give 14
(0.83 g, 53%), mp 111–113°C. 1H NMR (400 MHz, D2O) δ:
2.00, 2.17, 2.26 (s, 9 H, 3 × CH3), 2.86 (ddd, 1 H, J2a,1
4.2, J2a,2e = 15.4, J2a,3 = 7.3, H-2a), 3.02 (ddd, 1 H, J2e,1
6.2, J2e,3 = 4.2, H-2e), 4.23 (dd, 1 H, J6a,5 = 3.0, J6a,6b
=
=
=
Acknowledgments
12.6, H-6a), 4.29 (ddd, 1 H, J5,4 = 6.5, J5,6b = 6.7, H-5), 4.66
(dd, 1 H, H-6b), 5.21 (t, 1 H, J4,3 + J4,5 = 12.4, H-4), 5.53
(ddd, 1 H, J3,4 = 5.6, H-3). 7.28 (dd, 1 H, H-1), 8.07 (ddd, 1
H, J5 ′,6 ′ = 7.9, J6′,7 ′ = 7.0, J6′,8 ′ = 0.7, ArH-6′), 8.19 (dd, 1 H,
J3′,2′ = 6.2, J3′,4′ = 8.2, ArH-3′), 8.30 (ddd, 1 H, J5′,7′ = 1.4,
J7 ′,8 ′ = 8.7, ArH-7′), 8.43 (dd, 1 H, J2′,4′ = 1.4, ArH-4′), 8.50
(brd, 1 H, ArH-5′), 9.26 (brd, 1 H, ArH-8′), 9.52 (dd, 1 H,
ArH-2′). Anal. calcd. for C21H24BF4NO7: C 51.52, H 4.94,
N 2.86; found: C 51.50, H 4.87, N 2.72.
The authors are grateful to the Natural Sciences and Engi-
neering Research Council of Canada (NSERC) and the Brit-
ish Columbia Health Research Foundation for financial
support. In addition, the authors would like to thank NIH for
a grant (GM 31265, to Dr. A. Herscovics) that supports re-
search on the yeast α -mannosidase enzyme. Finally, we
would like to thank Dr. T. E. Kitos for editorial assistance
with this manuscript.
2-Deoxy-α-D-arabino-hexopyranosyl quinolinium
tetrafluoroborate (12)
A solution of sodium methoxide (4 equiv) in methanol
(10 mL) was added in one portion, with stirring, to an ice-
cold solution of 14 (0.12 g, 2.5 mmol) in methanol (15 mL)
that was maintained under an inert atmosphere. After stirring
References
1. R.A. Dwek. Chem. Rev. 96, 683 (1996).
2. P. Sears and C.-H. Wong. Angew. Chem. Int. Ed. Eng. 38,
2300 (1999).
© 2000 NRC Canada

582
3. (a) G. Legler. Adv. Carbohydr. Chem. Biochem. 48, 319
Can. J. Chem. Vol. 78, 2000
15. L. Hosie and M.L. Sinnott. Biochem. J. 226, 437 (1985).
16. G. Hanozet, H.-P. Pircher, P. Vanni, B. Oesch, and G.
Semenza. J. Biol. Chem. 256, 3703 (1981).
17. (a) B.W. Sigurskjold, B. Duus, and K. Bock. Acta. Chem.
Scand. 45, 1032 (1991); (b) J.D. McCarter, M.J. Adam, and
S.G. Withers. Biochem. J. 286, 721 (1992).
18. J. Zhu and A.J. Bennet. J. Am. Chem. Soc. 120, 3887 (1998).
19. X. Huang, K.S.E. Tanaka, and A.J. Bennet. J. Am. Chem. Soc.
119, 11147 (1997).
20. R.U. Lemieux. Acc. Chem. Res. 29, 373 (1996).
21. C.-H. Wong. Acc. Chem. Res. 32, 376 (1999).
22. (a) L. Hosie, P.J. Marshall, and M.L. Sinnott. J. Chem. Soc.
Perkin Trans. 2, 1121 (1984); (b) R.U. Lemieux and A.R.
Morgan. Can. J. Chem. 43, 2205 (1965).
23. T.P. Frandsen and B. Svensson. Plant Mol. Biol. 37, 1 (1998).
24. (a) B. Henrissat. Biochem. J. 280, 309 (1991); (b) B.
Henrissat. Biochem. J. 293, 781 (1993); (c) B. Henrissat and
A. Bairoch. Biochem. J. 316, 695 (1996).
25. M.N. Namchuk and S.G. Withers. Biochemistry, 34, 16194
(1995).
26. K.S.E. Tanaka and A.J. Bennet. Can. J. Chem. 76, 431 (1998).
27. F. Lipari and A. Herscovics. Biochemistry, 38, 1111 (1999).
28. A.E. Derome and M.P. Williamson. J. Magn. Reson. 88, 177
(1990).
29. A. Bax and S. Subramanian. J. Magn. Reson. 67, 565 (1986).
30. X. Huang, C. Surry, T. Hiebert, and A.J. Bennet. J. Am. Chem.
Soc. 117, 10614 (1995).
(1990); (b) M.L. Sinnott. Chem. Rev. 90, 1171 (1990); (c) J.D.
McCarter and S.G. Withers. Curr. Opin. Struct. Biol. 4, 885
(1994).
4. A.D. McNaught. Adv. Carbohydr. Chem. Biochem. 52, 43 (1997).
5. (a) S. Inouye, T. Tsuruoka, and T. Niida. J. Antibiot. 19, 288
(1966); (b) A.B. Hughes and A.J. Rudge. Nat. Prod. Rep. 135
(1994).
6. D.E. Koshland. Biol. Rev. Cambridge Philos. Soc. 28, 416
(1953).
7. M. Koyama and S. Sakamura. Agric. Biol. Chem. 38, 1111
(1974).
8. T.M. Jespersen, W. Dong, M.R. Sierks, T. Skrydstrup, I. Lundt,
and M. Bols. Angew. Chem. Int. Ed. Eng. 33, 1778 (1994).
9. M. Bols, R.G. Hazell, and I.B. Thomsen. Chem. Eur. J. 3, 940
(1997).
10. (a) Y. Kameda and S. Horii. J. Chem. Soc. Chem. Commun.
746 (1972); (b) Y. Kameda and S. Horii. J. Chem. Soc. Chem.
Commun. 747 (1972).
11. B. Svensson and M.R. Sierks. Carbohydr. Res. 227, 29 (1992).
12. (a) J.C. McAuliffe, R.V. Stick, D.M.G. Tilbrook, and A.G.
Watts. Aust. J. Chem. 51, 91 (1998); (b) J.C. McAuliffe and
R.V. Stick. Aust. J. Chem. 50, 193 (1997); (c) S. Ogawa, C.
Uchida, and Y. Shibata. Carbohydr. Res. 223, 279 (1992).
13. S. Knapp, Y. Dong, K. Ruptiz, and S.G. Withers. Bioorg. Med.
Chem. Lett. 5, 763 (1995).
14. A. Fersht. Enzyme structure and mechanism. 2nd. ed. W.H.
Freeman and Co., New York. 1985.
© 2000 NRC Canada