A stepwise solvent-promoted SNi reaction of α-d- glucopyranosyl fluoride: Mechanistic implications for retaining glycosyltransferases

Article abstract of DOI:10.1021/ja209339j

The solvolysis of α-d-glucopyranosyl fluoride in hexafluoro-2- propanol gives two products, 1,1,1,3,3,3-hexafluoropropan-2-yl α-d-glucopyranoside and 1,6-anhydro-β-d-glucopyranose. The ratio of these two products is essentially unchanged for reactions that are performed between 56 and 100 °C. The activation parameters for the solvolysis reaction are as follows: ΔH^? = 81.4 ± 1.7 kJ mol ^-1, and ΔS^? = -90.3 ± 4.6 J mol ^-1 K^-1. To characterize, by use of multiple kinetic isotope effect (KIE) measurements, the TS for the solvolysis reaction in hexafluoro-2-propanol, we synthesized a series of isotopically labeled α-d-glucopyranosyl fluorides. The measured KIEs for the C1 deuterium, C2 deuterium, C5 deuterium, anomeric carbon, ring oxygen, O6, and solvent deuterium are 1.185 ± 0.006, 1.080 ± 0.010, 0.987 ± 0.007, 1.008 ± 0.007, 0.997 ± 0.006, 1.003 ± 0.007, and 1.68 ± 0.07, respectively. The transition state for the solvolysis reaction was modeled computationally using the experimental KIE values as constraints. Taken together, the reported data are consistent with the retained solvolysis product being formed in an S_Ni (D_N^? A _Nss) reaction with a late transition state in which cleavage of the glycosidic bond is coupled to the transfer of a proton from a solvating hexafluoro-2-propanol molecule. In comparison, the inverted product, 1,6-anhydro-β-d-glucopyranose, is formed by intramolecular capture of a solvent-equilibrated glucopyranosylium ion, which results from dissociation of the solvent-separated ion pair formed in the rate-limiting ionization reaction (D_N^? + A_N). The implications that this model reaction have for the mode of action of retaining glycosyltransferases are discussed.

Full text of DOI:10.1021/ja209339j

Article
pubs.acs.org/JACS
A Stepwise Solvent-Promoted S i Reaction of α-D-Glucopyranosyl
N
Fluoride: Mechanistic Implications for Retaining Glycosyltransferases
Jefferson Chan, Ariel Tang, and Andrew J. Bennet*
Department of Chemistry, Simon Fraser University, 8888 University Drive, Burnaby, BC V5A 1S6, Canada
*
S Supporting Information
ABSTRACT: The solvolysis of α-D-glucopyranosyl fluoride in hexafluoro-
-propanol gives two products, 1,1,1,3,3,3-hexafluoropropan-2-yl α-D-
2
glucopyranoside and 1,6-anhydro-β-D-glucopyranose. The ratio of these
two products is essentially unchanged for reactions that are performed
between 56 and 100 °C. The activation parameters for the solvolysis
⧧
−1
⧧
reaction are as follows: ΔH = 81.4 ± 1.7 kJ mol , and ΔS = −90.3 ± 4.6 J
−1
−1
mol K . To characterize, by use of multiple kinetic isotope effect (KIE) measurements, the TS for the solvolysis reaction in
hexafluoro-2-propanol, we synthesized a series of isotopically labeled α-D-glucopyranosyl fluorides. The measured KIEs for the
C1 deuterium, C2 deuterium, C5 deuterium, anomeric carbon, ring oxygen, O6, and solvent deuterium are 1.185 ± 0.006, 1.080
0.010, 0.987 ± 0.007, 1.008 ± 0.007, 0.997 ± 0.006, 1.003 ± 0.007, and 1.68 ± 0.07, respectively. The transition state for the
solvolysis reaction was modeled computationally using the experimental KIE values as constraints. Taken together, the reported
data are consistent with the retained solvolysis product being formed in an S i (D *ANss) reaction with a late transition state in
which cleavage of the glycosidic bond is coupled to the transfer of a proton from a solvating hexafluoro-2-propanol molecule. In
±
⧧
N
N
comparison, the inverted product, 1,6-anhydro-β-D-glucopyranose, is formed by intramolecular capture of a solvent-equilibrated
glucopyranosylium ion, which results from dissociation of the solvent-separated ion pair formed in the rate-limiting ionization
⧧
reaction (DN + A ). The implications that this model reaction have for the mode of action of retaining glycosyltransferases are
N
discussed.
INTRODUCTION
documented for substitutions at silicon; however, at carbon
■
centers, the distinction between a classically defined S i
Glycoconjugates are carbohydrate-containing structures that are
involved in numerous biological processes, including many
complex interactions such as the mediation and modulation of
N
8
(
D *D*A ) such as for the breakdown of chlorosulfites
N
N
with retention of configuration and a concerted mechanism
1
(A D ) has become blurred in the literature.
cell adhesion and cell signaling events. As a consequence,
N
N
The pioneering investigation by Sinnott and Jencks of the
solvolysis products for the reactions of glucopyranosyl
derivatives in mixtures of ethanol and trifluoroethanol (TFE)
showed clearly that for substitution reactions on carbohydrates9
the product stereochemistry is influenced by the leaving group.
In particular, solvolysis of α-D-glucopyranosyl fluoride (α-GluF)
gave trifluoroethyl products in which the retained diastereomer
predominated, an observation from which Sinnott and Jencks
concluded that generation of trifluoroethyl α-D-glucopyranoside
could be regarded as a type of “internal return” involving the
research in this field has involved studying the main classes of
enzymes responsible for either the addition or removal of sugar
residues, namely, glycosyltransferases or glycosidases, respec-
tively. The mechanisms by which the corresponding non-
enzymatic reactions of glycosides occur can be used as a
valuable cornerstone for the understanding of how enzymes
2,3
accelerate these important biological reactions. Central to a
mechanistic discussion of these acetal substitutions is whether
the reaction at the anomeric center occurs via an oxacarbenium
4
ion intermediate. In the context of the Winstein ion pair
−
9
complex of the leaving group and solvent (F ···HOCH CF ).
mechanism for solvolyses, the reaction of glucopyranosyl
2
3
With regard to carbohydrate-processing enzymes, the SNi
derivatives can occur through either a direct S 2 reaction to
N
mechanism is enjoying something of a renaissance. For
give an inverted product or an S 1 pathway in which product
N
example, it recently was concluded on the basis of a mechanistic
study involving a transition state mimicry analysis, linear free
energy relationships, and KIEs that for the retaining
glycosyltransferase trehalose-6-phosphate synthase the cata-
formation occurs at the stage of intimate, solvent-separated, or
5
,6
solvent-equilibrated ion pairs (Scheme 1).
Notably, the
stereochemical outcome of these reactions can be influenced by
the nature of the leaving group when oxacarbenium ion capture
occurs on solvent-separated ion pairs. Recently, a substitution
10
lyzed reaction involves a “front-face” mechanism. In this
study, the authors concluded that their results were consistent
with the presence of either a metastable intermediate (SNi,
mechanism, which was originally labeled as an “intramolecular
7
nucleophilic change (S i)” by Ingold and co-workers, in which
N
the in vogue transition state structure that is drawn has partial
bonding to both the nucleophile and the leaving group has
received renewed attention. Such reactions have been well
Received: October 4, 2011
Published: December 8, 2011
©
2011 American Chemical Society
1212
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
a
Scheme 1. General Mechanistic Scheme for the Reactions of α-D-Glycopyranosyl Derivatives
a
The green circles represent solvent molecules.
D *ANss) in a stepwise mechanism or a frontside concerted
metal, and dichloromethane was dried and distilled over calcium
hydride. For anhydrous reactions, all glassware was flame-dried and
cooled under a nitrogen atmosphere immediately prior to use. NMR
spectra were recorded on a Bruker 500 or 600 MHz spectrometer.
Chemical shifts (δ) are listed in parts per million downfield from TMS.
N
pathway with an exploded transition structure geometry
1
0
(
A D ). Other proposed mechanisms for retaining glycosyl-
N N
transferases include standard double displacements in which the
1
1,12
enzymatic nucleophile is either a carboxylate
or a side chain
1
1
1
13
All NMR peak assignments are based on H− H COSY and H− C
HMQC experiments; coupling constants are reported in hertz. Melting
points were measured on a Gallenkamp melting point apparatus and
1
3
amide group. The mechanism by which glycosyltransferases
3
,14
operate has recently been reviewed.
18
To probe the intrinsic kinetic signatures for frontside
substitution reactions on glucopyranosyl derivatives, we
decided to examine the mechanism of solvolysis of α-GluF in
the mildly acidic and non-nucleophilic solvent hexafluoro-2-
propanol (HFIP). HFIP was selected because in comparison to
are uncorrected. The synthesis of α-D-(6- O)glucopyranosyl fluoride
1
8
(1e) requires 1,2,3,4,6-penta-O-acetyl-α,β-D-(6- O)glucopyranose
(3), which was accessed in four steps beginning from 1,2-O-
isopropylidene-α-D-glucofuranose in 43.2% yield overall as detailed
below.
18
9
1,2-O-Isopropylidene-6- O-α-D-glucofuranose (2). A flame-dried
TFE, used in the study by Sinnott and Jencks, it is (i) more
flask was charged with 1,2-O-isopropylidene-α-D-glucofuranose (1.0 g,
acidic, (ii) a stronger hydrogen bond donor, (iii) a more
18
20
4
.54 mmol), ( O )benzoic acid (544 mg, 4.38 mmol), PPh (2.38 g,
2 3
15
ionizing solvent, and (iv) less nucleophilic. These solvent
properties are expected to favor a greater selectivity for the
formation of retained products. A panel of isotopically labeled
α-D-glucopyranosyl fluorides (1a−f) was synthesized for the
measurement of kinetic isotope effects on this solvolysis
reaction. In addition, computational chemistry was employed to
locate possible transition state structures using the experimental
KIEs as restraints during modeling.
9
.07 mmol), and anhydrous THF (75 mL). The resultant solution was
cooled to 0 °C with an ice bath and treated with a solution of diethyl
azodicarboxylate in toluene [40% (w/v), 5.0 mL, 10.46 mmol]. After
the solution had been stirred at room temperature for 48 h, the
volatiles were removed under reduced pressure and the crude residue
was purified via flash chromatography [5:95 (v/v) MeOH/CH Cl ] to
afford the intermediate as a white solid. This material was treated with
2
2
a solution of sodium methoxide in MeOH (∼0.5 M, 25 mL) and was
stirred until it was shown by TLC analysis [R = 0.28; 5:95 (v/v)
f
MeOH/CH Cl ] that reaction was complete. Following neutralization
2
2
+
of the solution with Amberlite (H ) resin, it was filtered and
concentrated to afford the desired product as a white solid (520 mg,
2
1
5
(
4% yield over two steps): mp 160−161 °C (lit. 161 °C); HRMS
+ 18
M + Na ) C H O O requires m/z 245.0882, found m/z 245.0894.
9 16 5 1
1
8
1
,2,3,4,6-Penta-O-acetyl-α,β-D-(6- O)glucopyranose (3). A sol-
ution of 2 (500 mg, 2.25 mmol) in water (15 mL) and acetone (15
+
mL) was treated with Amberlite (H ) resin (1 g). The mixture was
MATERIALS AND METHODS
stirred while being heated to reflux for 3 h. After cooling, the resin was
removed via filtration and rinsed with water (∼20 mL). The filtrate
was concentrated under reduced pressure, and the resultant syrupy
residue was dissolved in pyridine (15 mL) and treated with acetic
anhydride (5 mL). After being stirred at room temperature for 12 h,
■
Materials. 1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP) and
1
,1,1,3,3,3-hexafluoro-2-propan(ol-d) (HFIP-d), which were purified
16
by using a standard procedure, were purchased from Matrix
18
Scientific and Aldrich, respectively. Water ( O) was purchased from
the solution was diluted with CH Cl2 (50 mL) and washed
Olinax Inc. (98 atom % 18O, batch 090215A4). 1,6-Anhydro-β-D-
2
sequentially with water (50 mL), cold 10% H SO (50 mL), saturated
13
2
4
glucopyranose was purchased from TCI America. D-Glucose-1- C and
NaHCO (2 × 25 mL), and brine (25 mL). The organic layer was
2
3
D-glucose-2- H were purchased from Omicron Biochemicals, Inc. D-
dried (Na SO ) and concentrated to give a solid residue that was
2
2
18
2
4
Glucose-1- H, D-glucose-5- H, and D-glucose-5- O were prepared
purified via column chromatography [1:1 (v/v) EtOAc/hexanes] to
afford the title compound as a white foam (705 mg, 80% yield over
1
7,18
according to known procedures.
α-D-Glucopyranosyl fluoride and
its singly labeled isotopologues were prepared from the correspond-
+
18
9
two steps): HRMS (M + Na ) C H O O requires m/z 415.1097,
16
22
1
ingly labeled 1,2,3,4,6-penta-O-acetyl-α,β-D-glucopyranose in two steps
found m/z 415.1109.
19
using a known protocol. All other reagents were purchased from
Aldrich and used without purification. Thin-layer chromatography
Product Isolation, Purification, and Identification. A 20 mL
ampule was charged with 1 (220 mg, 1.21 mmol) and HFIP (10 mL).
After being sealed, the reaction vessel was heated for 2 h in a steam
bath. After the solution had cooled to room temperature, the volatiles
were removed under reduced pressure and the crude residue was
treated with pyridine (15 mL) and acetic anhydride (5 mL). After
being stirred for 16 h at room temperature, the solution was diluted
with CH Cl (50 mL) and washed sequentially with cold 10% H SO
(
TLC) was performed on aluminum-backed TLC plates precoated
with Merck silica gel 60 F254. Compounds were visualized with UV
light and/or staining with a p-anisaldehyde solution. Flash
chromatography was performed using silica gel 60 (230−400 mesh).
Solvents used in anhydrous reactions were dried and distilled
immediately prior to use. THF was dried and distilled over sodium
2
2
2
4
1
213
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
(
50 mL), water (50 mL), saturated NaHCO (50 mL), and brine (50
the internal standard 1,4-difluorobenzene (−124.27 ppm), 1 (−153.02
ppm), and 1a (−153.74 ppm) were integrated. At various points
throughout the course of the reaction, the same NMR tube was heated
in a steam bath, cooled, and analyzed by 19F NMR spectroscopy as
3
mL). The organic fraction was dried (Na SO ) and concentrated to
afford a syrupy residue. The protected HFIP solvolysis product (R =
0
2
4
f
.29) and 2,3,4-tri-O-acetyl-1,6-anhydro-β-D-glucopyranose (R = 0.13)
f
were purified via column chromatography [3:7 (v/v) EtOAc/hexanes].
Each product was treated with freshly prepared NaOMe in MeOH
described above. The resultant fraction of reaction (F ) and individual
1
peak areas were analyzed using a nonlinear least-squares fit to eq 1. Of
note, the intense 19F signal corresponding to HFIP (−78.66 ppm)
does not overlap or interfere with the peaks of interest. All other KIEs
(
∼0.5 M, 10 mL). The reaction mixtures were neutralized with
+
Amberlite IR-120+ resin (H form), filtered, and concentrated to
2
afford 4 (α-GluHFIP) and 5 (1,6-anhydro-β-D-glucopyranose).
were determined relative to the α- H KIE using eq 2
1
,1,1,3,3,3-Hexafluoropropan-2-yl α-D-glucopyranoside (4): mp
2
0
1
R
R
0
1
25−126 °C; [α] = +89.8 (c 0.017, H O); H NMR (600 MHz,
(1/KIE−1)
D
2
= (1 − F1)
D O) δ 5.20 (d, J = 3.3, 1 H, H-1), 5.10 [m, 1 H, OCH(CHF ) ], 3.76
(1)
2
3
2
(
3
m, 1 H, H-6), 3.74−3.64 (m, 3 H, H-3, H-6′, H-5), 3.57 (dd, J = 10.0,
.8, 1 H, H-2), 3.40 (t, J = 9.5, 1 H, H-4); C NMR (151 MHz, D O)
13
1D
2
KIE
= k
H
δ 121.84 (q, JCF = 62.6, −CF ), 119.97 (q, J = 65.5, −CF ), 100.31
KIE = Rel
3
CF
3
k
(2)
(
6
3
C-1), 72.54, 72.00 (septet, JCF = 32.4, −CHCF ) 71.67, 70.34 (C-2),
KIE
x
3
+
8.55, 59.70 (C-6); HRMS (M + Na ) C H F O requires m/z
9 12 6 6
where 1DKIE is the α- H isotope effect (k
k ) is the relative isotope effect measured in the competitive
experiment between 1a and another labeled α-GluF. In these
calculations, 1a is treated as the “heavy” isotopologue such that the
ratio of 1DKIE and KIE will give the KIE for x (i.e., x = 1b). All
calculations of standard deviations on KIE values computed using eq 2
involved standard propagation of the errors associated with the
2
/k ) and RelKIE (kx/
unlabeled 1D
53.0430, found m/z 353.0443.
,6-Anhydro-β-D-glucopyranose (5): [α]
20
1D
1
= −67.6 (c 0.017,
D
22
1
H O) (lit. = −66.0); H NMR (600 MHz, D O) δ 5.38 (s, 1 H, H-
2
2
1
), 4.56 (d, J = 5.7, 1 H, H-5), 4.02 (d, J = 7.7, 1 H, H-6), 3.69 (m, 1
Rel
13
H, H-6′), 3.62−3.59 (m, 2 H, H-3, H-4), 3.45 (s, 1 H, H-2); C NMR
151 MHz, D O) δ 100.96 (C-1), 75.78 (C-5), 72.01 (C-3), 70.32 (C-
), 69.69 (C-2), 64.71 (C-6); HRMS (M + Na ) C H O requires m/
z 185.0420, found m/z 185.0420.
(
4
2
+
6 10 5
1D
Rel
23
measured values of KIE and KIE.
Simultaneous Measurement of Three KIEs. A 5 mm NMR tube
was charged with a solution of 1 (∼0.3 mg), 1a (∼0.3 mg), 1b (∼0.3
mg), and 1f (∼0.6 mg) in DMSO-d (50 μL), 1,4-difluorobenzene (0.2
6
μL), 2,6-lutidine (4.8 μL), and HFIP (600 μL). A protocol identical to
2
2
13
that described above was used to measure 1- H, 2- H, and 1- C KIEs
simultaneously.
Solvent KIE. A solution of 1 (16 mg) in DMSO-d (75 μL) was
6
Solvolysis Product Stability. To show that α-GluHFIP is stable
to the solvolysis conditions, α-GluHFIP (10 mg) was dissolved in
HFIP (2 mL) and the solution sealed in an ampule that was heated in
added to a flask charged with HFIP (12 mL) and 2,6-lutidine (25 μL).
The solution was distributed evenly into 14 prescored ampules that
were subsequently sealed. Samples containing HFIP-d were prepared
in an identical manner. These two sets of ampules were heated in a
steam bath, and every 5 min, samples were removed and cooled in an
ice bath. The observed rate constants were calculated as described in
Activation Parameters.
1
a steam bath for 12 h. Subsequent H NMR spectroscopic analysis in
D O of the nonvolatiles showed no 1,6-anhydroglucose. In addition,
2
1
,6-anhydroglucose was shown to be stable under the reaction
conditions by being heated a sealed ampule containing it (∼1 mg),
DMSO (15 μL), and 2,6-lutidine (4 μL) in HFIP (600 μL) at 100 °C
for 2 h. After removal of the volatiles, the residue was dissolved in
1
THEORETICAL CALCULATIONS
D O, and analysis by H NMR spectroscopy showed that no reaction
2
■
had occurred.
Computational Analysis. Calculations for the HFIP
solvolysis of α-GluF were performed using Gaussian 09 and
the B3LYP method with a 6-31G* basis set. All TS structures
were calculated in the gas phase at 373 K. The α-GluF ground
Activation Parameters. Ampules were charged with 1 (1 mg) in
DMSO (5 μL), HFIP (1 mL), and 2,6-lutidine (1.5 μL). After being
sealed, the ampules were heated in a “reflux chamber” equipped with a
still pot containing an appropriate solvent. The temperature
dependence of the reaction was determined by monitoring the rate
of solvolysis at 56, 65, 82, and 100 °C by suspending the ampules in
the vapors of boiling acetone, methanol, 2-propanol, and water,
respectively. Samples were removed after various time intervals and
cooled in an ice bath. The contents were transferred to a flask, from
24
4
state structure was optimized beginning from several C
1
conformations to ensure that the calculated structure is a
local minimum. In addition, ground state structures with a
hydrogen bond between the C2 hydroxyl group and the
anomeric fluorine atom were not considered to ensure that the
intramolecular H−O···H hydrogen bonding arrangement,
which mainly results from the system being calculated in
vacuo, is the same in the TS as it is in the ground state. With
regard to the TS structures for ionization of α-GluF, an HFIP
which the volatiles were removed under reduced pressure. The residue
1
was dissolved in D O (650 μL) and analyzed by H NMR
2
spectroscopy. The anomeric proton chemical shifts for 1 (5.62
ppm), 4 (5.19 ppm), and 5 (5.38 ppm) were integrated and
normalized. The normalized ratio of 1 was fit to a standard exponential
one-phase decay using a nonlinear regression algorithm.
́
molecule was positioned 4 Å from the fluoride leaving group
2
19
and allowed to optimize. Subsequently, the C−F bond distance
was incrementally increased and constrained to locate TS
structures that had one imaginary frequency. In addition, a TS
structure with an HFIP catalyzing the departure of the fluoride
KIE Measurements. The α- H KIE was determined via F NMR
2
0
spectroscopy on a Bruker AVANCE III 500 MHz spectrometer. A 5
mm NMR tube was charged with 1 (∼1.5 mg), 1a (∼1.5 mg), DMSO-
d (15 μL), 1,4-difluorobenzene (1.5 μL), 2,6-lutidine (4.8 μL), and
6
HFIP (650 μL). The NMR tube was flame-sealed, and an initial NMR
in a backside S 2 reaction that generates 1,6-anhydro-β-D-
N
spectrum was recorded with the sample temperature maintained at 298
glucopyranose was located by constraining the O6−C1
distance, a protocol that was necessary to prevent re-formation
of the starting material. The ground state for 1,6-anhydroglu-
cose was optimized from several starting geometries.
Calculations were performed using the same conditions and
level of theory. All TS structures were recalculated at the
MP2(full)/6-31G* level of theory using the B3LYP geometric
19
K. Each quantitative proton-decoupled and -coupled F NMR
spectrum was acquired using a gated pulse sequence. Spectra
consisting 128 scans (acquisition time of 0.35 s) were recorded with
a recycle delay of 2.5 s between scans (6.1 min per spectrum). The
baseline was corrected to remove baseline distortions that are typically
associated with 19F spectra using the Whittaker Smoother method
19
found in MestReNova version 6.2. The F signals corresponding to
1
214
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
constraints, although no further matching of experimental KIEs
was performed in this study. KIEs for each putative TS
structure were calculated using ISOEFF98. Of note, scale
The identification of the two products of solvolysis of α-GluF
(1) in HFIP at 100 °C were, after separation and purification,
25
1
shown by H NMR spectroscopy to be 1,1,1,3,3,3-hexafluoro-2-
factors of 0.9614 and 0.9427 were used for the B3LYP/6-31G*
propyl α-D-glucopyranoside 4 (α-GluHFIP = 54.6 ± 2.2%) and
26
and MP2(full)/6-31G* calculations, respectively.
1,6-anhydro-β-D-glucopyranose 5 (45.4%) (Figures S1 and S2
1
13
The starting geometry used to probe computationally
whether the epoxide (1,2-anhydro-α-D-glucopyranose) is a
viable intermediate en route to the 1,6-anhydro product via a
of the Supporting Information show the H and C NMR
spectra of α-GluHFIP, respectively). Of note, an identical
product ratio, within experimental error, was noted in the NMR
spectra for these solvolyses at all temperatures between 56 and
standard S 2 reaction was generated by incorporating the
N
1
following modifications into TS3: (i) the fluorine atom was
removed, and (ii) the 2-OH group was transformed into an
epoxide ring. All calculations were performed at the B3LYP/6-
100 °C. For example, integration of the H NMR spectrum for
the solvolysis performed at 56 °C showed that α-GluHFIP
constituted 55.8% of the binary mixture, while the remaining
44.2% was 1,6-anhydro-β-D-glucopyranose. Moreover, an
identical product mixture was obtained when the solvolysis
reaction was performed at 100 °C in HFIP-D (α-GluHFIP =
53.7 ± 1.2%; 1,6-anhydro-β-D-glucopyranose = 46.3%). The
solvent deuterium KIE for the HFIP solvolysis of α-GluF at 100
°C was measured to be 1.68 ± 0.07 (see Materials and Methods
for full details).
3
1G* level of theory without constraints. The bond distance
between the epoxide oxygen and the hydrogen atom of a
solvating HFIP molecule was incrementally varied by 0.05 Å
from 1.79 to 1.04 Å. In addition, the O6−C1 distance was
varied between 1.63 and 1.25 Å.
The equilibrium isotope effects (EIEs) for the formation of a
solvent-equilibrated oxacarbenium ion could be computed from
the vibrational frequencies of the cationic intermediate, the
structure of which was calculated by optimizing the cation−
HFIP complex in which the structure of TS1 without the
fluoride ion was used as the starting point for minimization.
The formation of 1,6-anhydro-β-D-glucopyranose was shown
not to involve a base-catalyzed S 2 reaction involving 2,6-
N
lutidine, which was added to neutralize the HF produced
during solvolysis, as the rate of solvolysis was invariant as a
function of base concentration (data not shown).
25
The EIEs were then calculated using ISOEFF98.
KIE Measurements. A slight modification to the “Direct
NMR Method” developed by Chan et al. to measure C and
1
3
RESULTS AND DISCUSSION
■
18
O KIEs on the Vibrio cholerae sialidase-catalyzed hydrolysis of
Rate Constants and Product Studies. The rate constants
for solvolysis of α-GluF in HFIP were measured at 56, 65, 82,
and 100 °C and are summarized in Table S1 of the Supporting
Information. The corresponding Eyring plot is shown in Figure
20
natural substrate analogues was used in this study to evaluate
1
9
the KIEs for the solvolysis of α-GluF in HFIP. Specifically,
F
was the NMR active probe nucleus chosen to report on the
changes in isotopologue ratios in the remaining starting
material. That is, a pronounced 19F chemical shift perturbation
(
∼0.7 ppm) occurs upon substitution of the C1 proton with a
deuteron (Figure 2). During the course of the solvolysis
Figure 2. Proton-decoupled 19F NMR spectrum of a mixture of α-D-
2
glucopyranosyl fluoride (1) and α-D-(1- H)glucopyranosyl fluoride
(
1a) in 1,1,1,3,3,3-hexafluoro-2-propanol.
Figure 1. Eyring plot for the solvolysis of α-D-glucopyranosyl fluoride
in 1,1,1,3,3,3-hexafluoro-2-propanol. The dashed line is the best linear
fit to the rate constant data.
reaction, the 19F signals corresponding to the internal standard
1,4-difluorobenzene) and the remaining starting isotopo-
(
logues, for example, 1 and 1a, were integrated. Figure 3
shows a stacked plot of the change in relative intensities of 1
and 1a at various points throughout the course of the reaction.
⧧
1
8
, and the derived activation parameters are as follows: ΔH =
−1
⧧
−1 −1
1.4 ± 1.7 kJ mol , and ΔS = −90.3 ± 4.6 J mol K . The
1
9
calculated activation parameters suggest that the solvolysis
involves a highly ordered transition state, which likely results
from solvation of the leaving group fluoride ion by the bulky
HFIP solvent. Indeed, solvolysis of α-GluF in HFIP displays a
negative entropy of activation greater than that of the
To ensure that the KIE values derived from the F NMR
2
spectroscopic experiments were accurate, the α- H KIE was
measured seven times. Table S2 of the Supporting Information
lists the individual KIE values measured for each experiment
23
and the corresponding standard deviations. In addition, it was
verified that the KIE value was unaffected by nuclear
−1
−1 27
corresponding reaction in water (−37 J mol K ).
1
215
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
Figure 3. Stacked plot of proton-decoupled 19F NMR spectra containing a mixture of 1 and 1a in 1,1,1,3,3,3-hexafluoro-2-propanol as a function of
the fraction of reaction of the unlabeled isotopologue (1). The two parallel dotted lines are for visualization of the change in relative peak intensity as
the reaction progresses.
Overhauser effects (NOE) by performing the 19F NMR
spectroscopic experiment while accumulating data with an
inverse-gated pulse sequence (Table S2, entry 7).
Although isotopic substitutions at positions farther from the
fluorine atom do not necessarily produce a chemical shift
perturbation sufficient to observe baseline resolution, it is
2
possible to determine all other KIEs relative to the 1- H KIE.
The relative ratio of each isotopologue (R) was determined by
routine integration, and the fraction of reaction (F ) was
1
evaluated throughout the time course of the reaction using 1,4-
difluorobenzene as an innocuous internal standard. These two
parameters, R and F , are then fit to eq 1 to determine the
1
Rel
KIE (k /k ). The KIE of interest can then be calculated
x
1D
using eq 2 (Table S3 of the Supporting Information). Panels a
and b of Figure 4 show the least-squares fits to eq 1 for typical
Figure 5. Proton-decoupled 19F NMR spectrum of a solution of 1
(
∼0.3 mg), 1a (∼0.3 mg), 1b (∼0.3 mg), and 1f (∼0.6 mg) in HFIP.
All peaks in the spectrum are labeled with the appropriate assignment.
spectra were combined using the “fidadd” function in TopSpin
2
.1 when the signal-to-noise ratio deteriorated to an
unsatisfactory level, a protocol that reduced the total number
of experimental data points used in the fit to eq 1. The
determined KIE values for this experiment are also listed in
Table 1, while the fits to the experimental data are shown in
Figures S3−S5 of the Supporting Information.
Table 1. Mean Individual Measurements of KIEs on the
Solvolysis of α-D-Glucopyranosyl Fluoride in HFIP at 100
a
Figure 4. Plots of the change in integrated peak intensity ratio (R/R0)
vs F1 for the competitive KIE measurements: (a) data from an
°C and the Corresponding Standard Deviations
compd
site of substitution
KIE and SD
simultaneous KIE
experimental measurement of k /k1D and (b) data from an
H
2
experimental measurement of k /k .
1a
1- Hα deuterium
1.185 ± 0.006
1.080 ± 0.010
0.987 ± 0.007
0.997 ± 0.006
1.003 ± 0.007
1.008 ± 0.007
1.190 ± 0.006
1.076 ± 0.007
2D
1D
2
1
b
2- Hβ deuterium
2
data sets for the competition reactions between 1 and 1a and
between 1b and 1a, respectively.
1c
1d
1e
1f
5- Hγ deuterium
18
5- O ring oxygen
18
To verify that the calculated KIE values were indeed accurate,
6- O nucleophile oxygen
2
13
we performed a single NMR experiment in which the 1- H,
1- C anomeric
1.000 ± 0.002
2
13
a
2
- H, and 1- C KIEs were measured simultaneously. That is,
Also given are the results from a single simultaneous measurement of
three KIEs.
the NMR solution contained three isotopologues in addition to
unlabeled α-GluF (Figure 5). Of note, as it was necessary to
reduce the initial concentrations of each α-GluF isotopologue
by approximately 6-fold to limit signal overlap, sequential
13C KIE. The anomeric 13C KIE for the solvolysis of α-GluF
in HFIP is 1.008 ± 0.007 (Table 1; the value from a single
1
216
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
multiple-KIE experiment is 1.000 ± 0.002), which is in the
reactions, while inverse values (kSOH/kSOD < 1) indicate the
occurrence of specific catalysis. The observed KIE (kSOH/kSOD
= 1.68 ± 0.07) is consistent with proton transfer occurring as
part of the rate-limiting transition state for solvolysis, and this
typical range associated with a dissociative S 1-like reac-
N
1
8,28,29
tion.
If the 1,6-anhydroglucose that results from these
solvolyses is formed by a concurrent S 2 pathway, a reaction
N
2
7−29
39
that should exhibit KIEs between 1.03 and 1.08,
it would
certainly involves protonation of the departing fluoride ion.
1
3
be expected that the C KIE would be significantly larger than
that measured. As such, these results are inconsistent with the
notion that two independent pathways are involved. Moreover,
the anomeric 13C KIE for the solvolysis of α-GluF is of the
same magnitude as those reported for the acid-catalyzed
General Mechanistic Scheme. The important observa-
tions that the product ratio is independent of the solvent
(HFIP or HFIP-d) yet the rate of solvolysis is dependent
require either that the rate-determining and product-determin-
ing steps in a sequential S 1-like mechanism are not the same
N
18
hydrolyses of methyl α-D-glucopyranoside (1.007 ± 0.001)
or that if the products are formed in two parallel reactions they
have the same SKIE. All data are consistent with a unified
mechanism that involves formation of a solvent-separated
oxacarbenium ion intermediate that partitions to either α-
30
and methyl α-D-xylopyranoside (1.006 ± 0.001). It is
therefore likely that these reactions have similar TS structures
and that the solvolysis of α-GluF in HFIP has a late TS with
significant C−F bond cleavage.
⧧
GluHFIP by collapse of the ion pair [D *A (Scheme 1)] or
N
Nss
18
18
18
O KIEs. The experimental 5- O and 6- O KIEs are 0.997
0.006 and 1.003 ± 0.007, respectively. Of note, the small
inverse endocyclic O KIE is of a magnitude similar to that
reported for the hydrolysis of methyl α-D-glucopyranoside,
and these values are significantly different from those of the
conformationally more flexible methyl xylopyranosides or
1,6-anhydro-β-D-glucopyranose by dissociation to give a
±
solvent-equilibrated ion pair in which the oxacarbenium ion
1
8
has a significant lifetime before intramolecular capture by the
18
⧧
C6 hydroxyl group (DN + A ). The critical observation that
N
partitioning to the product is independent of both temperature
and solvent (HFIP or HFIP-d) requires that the processes that
determine the partition ratio have similar activation enthalpies
and solvent deuterium isotope effects. In the current example,
we suggest that this involves solvent reorganization at the stage
of the solvent-separated ion pair to give either the retained
30
2
0
those from sialidase-catalyzed hydrolysis reactions, where
more extensive charge delocalization onto the ring oxygen atom
1
8
is the likely cause of the reported O KIEs of 0.978−0.986. In
other words, these small inverse KIE measurements for the
solvolysis of α-GluF indicate that the TS geometry possesses a
product by intramolecular capture (k ) or 1,6-anhydroglucose
c
18
poor n → π orbital overlap.
via a dissociation (k ) into a solvent-equilibrated oxacarbe-
p
−d
With respect to the measured value for the C6 oxygen atom,
nium ion that ultimately undergoes intramolecular capture
(Scheme 2).
1
8
few examples of O KIEs have been reported in the literature
3
1,32
where the isotopic substitution is in the nucleophile;
it is
clear that on the basis of the measured value of 1.003 ± 0.007
Scheme 2. Possible Mechanisms for the Solvolysis Reactions
of α-D-Glucopyranosyl Fluoride in Hexafluoro-2-propanol
18
for the 6- O isotopologue and those calculated for the various
possible transition states (vide infra) it is impossible to
differentiate between TSs solely on the basis of the magnitude
of this effect.
Secondary Deuterium KIEs. The measured α-secondary
deuterium KIE (α-SDKIE) for the solvolysis of α-GluF in HFIP
is 1.185 ± 0.006 (Table 1; the value from a single multiple-KIE
experiment is 1.190 ± 0.006), whereas the corresponding α-
2
7
SDKIE for its spontaneous hydrolysis is 1.142 ± 0.007,
a
2
7,33
reaction that occurs via an exploded S 2 (A D ) TS.
N
N
N
These KIEs are primarily associated with weakening of an out-
of-plane bending vibration as the steric crowding at the reaction
becomes weaker on approach to a carbenium ion-like transition
state, although it important to remember that maximal α-
3
4−36
SDKIE values depend on the identity of the leaving group.
Indeed, carbenium ion lifetimes in HFIP have been reported
4
8
The β-secondary deuterium KIEs (β-SDKIE), which
originate from hyperconjugative weakening of the C−H/(D)
bond with the developing empty p orbital at the anomeric
center, for HFIP solvolysis and hydrolysis of α-GluF are 1.080
0.010 (Table 1; the value from a single multiple-KIE
experiment is 1.076 ± 0.007) and 1.067 ± 0.008, respectively.
The magnitude of β-SDKIEs has been used to provide insights
to be extended by factors of approximately 10 and 10 in
40
comparison to those in TFE and water, respectively. For
instance, the rate constants for solvent capture (k ) of the
s
9
diphenylmethyl carbenium ion, at 20 °C, are 1.3 × 10 , 3.2 ×
6
1
−1
41
42
43
±
10 , and ∼1 × 10
s
in water, TFE, and HFIP,
27
respectively. Given that the estimated lifetime of the
44,45
glucopyranosylium ion in aqueous solutions is ∼1−3 ps,
1
8,27
into TS conformations,
as hyperconjugation is an angle-
the expected lifetime of this cation in HFIP is certainly in the
range that would allow for the formation of solvent-equilibrated
ions.
2
dependent phenomenon that varies as a function of cos (θ),
where θ is the dihedral angle between the C−H/(D) bond and
the developing p orbital on the anomeric center.
Of note, no evidence was obtained to indicate that any
inverted solvolysis product (β-GluHFIP) was formed during
these reactions. On the basis of the analysis described above
that the key partitioning steps between the two reaction
products occur at the stage of the solvent-separated ion pair,
the two possibilities that exist for the formation of the 1,6-
anhydro-β-D-glucopyranose are (i) direct intramolecular
Finally, the measured γ-secondary deuterium KIE (γ-SDKIE)
is 0.987 ± 0.007, an effect that traditionally has been considered
to originate from inductive effects; however, it is now clear that
37,38
inductive isotope effects are exceedingly small.
Solvent Kinetic Isotope Effect. Normal solvent deuterium
KIEs (kSOH/kSOD > 1) are associated with general catalyzed
1
217
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
capture of the solvent-equilibrated glucopyranosylium ion by
the 6-hydroxyl group and (ii) neighboring group capture to give
the 1,2-epoxide that in a subsequent step forms the anhydro
product. At present, we are experimentally unable to distinguish
between these two possibilities; however, we tried computa-
tionally to investigate whether an S 2-type reaction was feasible
N
between the 1,2-anhydro (epoxide) and the 1,6-anhydro
product. Of note, all calculations, including starting geometries
expected to favor formation of the 1,6-anhydro product, gave
instead an optimized structure of the 1,2-epoxide that was
hydrogen bonded to an HFIP molecule. At present, we suggest,
on the basis of these calculations, that the 1,6-anhydro-β-D-
glucopyranose results from an intramolecular capture of the
Figure 7. Transition state structure TS2 for the frontside A D
N
N
reaction of α-D-glucopyranosyl fluoride in 1,1,1,3,3,3-hexafluoro-2-
propanol. The transition state structure was determined in vacuo by
hybrid density functional theory implemented in Gaussian 09 using the
B3LYP functional and the 6-31G* basis sets. The distances among the
leaving group fluoride, the anomeric carbon, and the catalytic solvent
molecule are shown.
glucopyranosylium ion rather than by an intramolecular A D
N
N
reaction of 1,2-anhydro-α-D-glucopyranose.
Computational Analysis. Two TS structures for the
solvolysis reaction, TS1 and TS2, were located computationally
with the suite of experimental KIEs being used as restraints.
First, TS1 involves an HFIP-catalyzed cleavage of the carbon−
fluorine glycosidic bond, and this transition state precedes the
formation of an oxacarbenium ion intermediate (Figure 6).
Information (Figure S6 and Tables S12−S15). Listed in Tables
S16 and S17 of the Supporting Information are selected bond
distances and angles of the calculated TSs.
In addition, the structure of a glucopyranosylium ion−HFIP
complex was computed (Table S18 and Figure S7 of the
Supporting Information), and the EIEs calculated for the
formation of this oxacarbenium ion−molecule complex are
listed in Table 2.
Transition State Analysis. Using the experimental KIEs as
⧧
constraints, the two possibilities for the S i (D *A and
N
N
Nss
A D ) mechanism were modeled by varying the C−F bond
N
N
distance (r ) in 0.05 Å increments between 2.30 and 2.60 Å
C−F
and between 2.20 and 2.40 Å for TS1 and TS2, respectively.
Both structures adopt a flattened chair conformation with the
main difference between the two TS models being the position
and orientation of HFIP. Of these two models, the predicted
⧧
KIEs for TS1 (D *ANss) more closely match the experimental
N
values (Table 2). Critical features of TS1 include (i) extensive
⧧
C−F bond fission (r
C−F
= 2.45 Å) and (ii) the moderate
Figure 6. Transition state structure TS1 for the D *A reaction of
N
Nss
transfer of protons from the solvating HFIP to the fluoride ion.
When rC−F is increased beyond 2.60 Å, the computed structure
is minimized to an oxacarbenium ion−hexafluoroisopropoxide
ion pair and HF. With regard to TS2, an archetype of a coupled
α-D-glucopyranosyl fluoride in 1,1,1,3,3,3-hexafluoro-2-propanol. The
transition state structure was determined in vacuo by hybrid density
functional theory implemented in Gaussian 09 using the B3LYP
functional and the 6-31G* basis sets. The distances among the leaving
group fluoride, the anomeric carbon, and the catalytic solvent molecule
are shown.
“frontside” S i mechanism, the computed TS involves a
N
noticeably different geometry at the anomeric center, with
the sum of the three bond angles at C1 being 351.8° and 353.4°
for the B3LYP and MP2 calculations, respectively, whereas, for
TS1, in which the anomeric center is almost planar, the
corresponding sums are 358.5° and 359.0° for the B3LYP and
MP2 calculations, respectively. For the TS2 model when the
C−F bond distance is increased past 2.25 Å, an extrusion of HF
occurs with simultaneous formation of α-GluHFIP.
Second, TS2 entails a coupled fluoride ion departure and a
nucleophilic attack of the catalyzing HFIP (Figure 7). The
computed KIE values for both transition states are summarized
in Table 2 along with the experimentally determined values.
A single transition state (TS3) was located for formation of
the second reaction product 1,6-anhydroglucopyranose, via a
Of note, the experimental anomeric 13C KIE (k /k
=
concerted frontside S 2 reaction (Figure 8 and Table 2). For
N
1
2
13
this model, it was necessary to fix the O6−C1 bond distance to
prevent starting material, α-GluF, from being generated. TS3
incorporates an HFIP solvent molecule into the model and thus
allows proton transfer to occur during fluoride ion departure.
Listed in Tables S4−S11 of the Supporting Information are the
Cartesian coordinates of the computed ground state and TS
structures and the associated energies and imaginary
frequencies. Of note, a transition state model that did not
contain an HFIP acting as a general acid catalyst for the
intramolecular backside A D reaction was also located, and
1.008) is not modeled particularly well by either of the two
models (TS1 or TS2). Indeed, no TS that possessed a single
imaginary frequency that gave a better calculated agreement
with this KIE was located. However, a calculated structure [TS5
(Figure S8 of the Supporting Information)] that possesses two
imaginary frequencies, which following minimization became
TS1, has a calculated anomeric 13C KIE that is identical to the
experimental value. The two imaginary frequencies associated
−1
with TS5 involve (i) C2−OH bond rotation (−287.9 cm )
−1
and (ii) C−F bond cleavage (−180.1 cm ). The two transition
N
N
⧧
details of this TS (TS4) are given in the Supporting
state models for the D *A reaction (TS1 and TS5) are
N
Nss
1
218
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
Table 2. Experimental KIEs for α-D-Glucopyranosyl Fluoride HFIP Solvolysis and the Corresponding Computational Values
a
b
a
b
a
TS3MP2 ,
bc
a
,
ad
position exptl KIE TS1 B3LYP
TS1MP2
TS2 B3LYP
TS2MP2
TS3 B3LYP
TS5 B3LYP
oxacarbenium−HFIP complex
2
1
2
5
5
6
1
- H
1.185
1.080
0.987
0.997
1.003
1.008
1.1806
1.1111
0.9859
0.9971
1.0000
1.0192
1.1709
1.0982
0.9932
0.9933
1.0001
1.0188
1.1007
1.0913
0.9750
0.9941
0.9996
1.1131
1.0872
0.9740
0.9912
0.9996
1.0189
1.0568
1.0278
0.9980
0.9949
1.0058
1.0373
1.0513
1.0241
1.0037
0.9961
1.0065
1.0383
1.1553
1.1029
0.9802
0.9952
0.9993
1.1342
1.0747
0.9979
0.9928
0.9980
1.0054
2
- H
2
- H
1
8
- O
1
8
- O
13
- C
1.0200
1.0081
a
b
c
Calculated at the B3LYP/6-31G* level of theory. Calculated at the MP2(full)/6-31G* level of theory. Single-point calculation using the geometry
d
from the B3LYP/6-31G* calculation. Calculated equilibrium isotope effects (EIEs).
reaction displayed a moderate sensitivity to the pK of the
a
conjugate acid of the acceptor nucleophile (Brønsted, β
=
nuc
10
0
.54). Taking all these observations together, we suggest that
the reported results for the biological glucose transfer reaction,
when taken in the context of our experimental and computed
KIE values, are consistent with the biological process occurring
in a stepwise fashion with the kinetically significant step being
capture of the glucopyranosylium ion intermediate that is
formed in the presence of the bound nucleophilic acceptor, i.e.,
⧧
46
a D *A mechanism. That is, the observed KIEs for isotopic
N
N
substitution on the UDP-glucose ring are composites of the
equilibrium isotope effect for formation of the glucopyranosy-
lium ion (calculated EIEs for the formation of a glucopyr-
anosylium ion are listed in Table 2) and the kinetic trapping of
this ion by the sugar acceptor, which must occur via an early
TS. A comment on our analysis of why the model reaction has
rate-limiting dissociation while nucleophilic attack is a kineti-
cally significant step in the biological reaction is justified.
Specifically, leaving group departure in the case of α-D-
glucopyranosyl fluoride has been shown by Banait and Jencks
to involve enforced general catalysis because the glucopyr-
anosylium cation does not have a significant lifetime in the
39
presence of a nonsolvated fluoride ion, which in our example
requires the acidic solvent to promote fluoride departure.
However, once at the stage of the solvent-separated ion pair,
the back reaction to re-form α-D-glucopyranosyl fluoride
Figure 8. Transition state structure TS3 for a hypothetical backside
A D reaction of α-D-glucopyranosyl fluoride in 1,1,1,3,3,3-hexa-
N
N
fluoro-2-propanol to give 1,6-anhydro-β-D-glucopyranose as the
product. The transition state structure was determined in vacuo by
hybrid density functional theory implemented in Gaussian 09 using the
B3LYP functional and the 6-31G* basis sets. The distances among the
leaving group fluoride, the anomeric carbon, the catalytic solvent
molecule, and the C6 hydroxyl group are shown.
−
becomes negligible because once the small F anion is solvated
it is only weakly nucleophilic; i.e., formation of the solvent-
separated ion pair is rate-limiting. With regard to retaining
glycosyltransferases, we suggest that leaving group departure
(
UDP in the case of trehalose-6-phosphate synthase) does not
remarkably similar (Table S15 of the Supporting Information)
in spite of the differences in the calculated anomeric 13C KIEs.
Biological Implications. The results of this study clearly
show that in a non-nucleophilic environment the glucopyr-
anosylium ion has a sufficiently long lifetime to permit
conformational interconversions of the pyranosyl ring. The
require significant general acid catalysis from the weakly acidic
bound sugar acceptor hydroxyl group because the pyrophos-
phate component is tightly coordinated by positively charged
residues, which by themselves likely provide sufficient
catalysis for leaving group departure. We note that Ardevol
and Rovira, on the basis of a QM/MM calculation for
trehalose-6-phosphate synthase, recently concluded that the
mechanism involves the formation of a short-lived cationic
intermediate and that the highest point on the reaction free
energy profile involves nucleophilic attack on the anomeric
center by the acceptor hydroxyl group, a process that is
catalyzed by the UDP leaving group, and that Goedl and
Nidetzky transformed the active site of sucrose phosphorylase
into a glycosyltransferase mimic by mutation of the enzymatic
47
̀
1
4
10
reported anomeric C KIE (k /k = 1.023), when
1
2
14
2
8
13
converted to the expected value for the C isotopologue
k /k = 1.012), and the α-secondary deuterium KIE for the
(
12
13
retaining glucosyltransferase (trehalose-6-phosphate syn-
1
0
thase) are remarkably similar to the values reported here
for the solvolysis of α-GluF. In contrast, the transferase reaction
exhibits a β-secondary deuterium KIE, albeit with a non-natural
48
1
0
nucleophile, that is significantly larger than both the
49
active site.
experimental value for solvolysis of α-GluF and the computed
⧧
values for TS1 (and TS5; D *ANss) and TS2 (frontside
N
CONCLUSIONS
A D ). Unfortunately, we are unable to measure a leaving
■
N
N
group isotope effect for solvolysis of α-D-glucopyranosyl
The solvolysis reaction of α-D-glucopyranosyl fluoride in
hexafluoro-2-propanol occurs via rate-limiting C−F bond
cleavage that is coupled to proton transfer for a solvent
1
8
10
fluoride for a comparison with the reported V/K value. In
this recent paper, Lee et al. noted that the glucosyl transfer
1
219
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220

Journal of the American Chemical Society
Article
molecule. The formed solvent-separated ion pair forms the
(23) Taylor, J. R. An introduction to error analysis: The study of
uncertainties in physical measurements; University Science Books: Mill
Valley, CA, 1982.
product via two separate reaction channels: (i) collapse of the
⧧
ion pair gives retained solvolysis product (D *ANss), and (ii)
N
(24) Frisch, M. J. et al. Gaussian 09, revision A.02; Gaussian, Inc.:
dissociation of the ion pair gives a solvent-equilibrated
glucopyranosylium ion that subsequently reacts to give the
Wallingford, CT, 2009.
(25) Anisimov, V.; Paneth, P. J. Math. Chem. 1999, 26, 75−86.
(26) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502−16513.
(27) Zhang, Y.; Bommuswamy, J.; Sinnott, M. L. J. Am. Chem. Soc.
⧧
inverted product 1,6-anhydro-β-D-glucopyranose (DN + AN).
ASSOCIATED CONTENT
Supporting Information
1
(
994, 116, 7557−7563.
■
S
28) Melander, L. C. S.; Saunders, W. H. J. Reaction rates of isotopic
*
1
molecules; Wiley: New York, 1980.
(29) Lee, J. K.; Bain, A. D.; Berti, P. J. J. Am. Chem. Soc. 2004, 126,
Complete citation for ref 24, Tables S1−S16, H (Figure S1)
and 13C (Figure S2) NMR spectra of 1,1,1,3,3,3-hexafluoro-2-
3
(
1
(
769−3776.
30) Indurugalla, D.; Bennet, A. J. J. Am. Chem. Soc. 2001, 123,
0889−10898.
31) Cassano, A. G.; Anderson, V. E.; Harris, M. E. Biochemistry
propyl α-D-glucopyranoside, and Figures S3−S6. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
E-mail: bennet@sfu.ca
2
(
(
004, 43, 10547−10559.
■
32) Marlier, J. F. J. Am. Chem. Soc. 1993, 115, 5953−5956.
33) Banait, N. S.; Jencks, W. P. J. Am. Chem. Soc. 1991, 113, 7951−
7
958.
(
34) Shiner, V. J. Jr.; Rapp, M. W.; Halevi, E. A.; Wolfsberg, M. J. Am.
ACKNOWLEDGMENTS
We thank the Mizutani Foundation for Glycoscience for a grant
to support this research.
■
Chem. Soc. 1968, 90, 7171−7172.
(35) Harris, J. M.; Hall, R. E.; Schleyer, P. v. R. J. Am. Chem. Soc.
1971, 93, 2551−2553.
(
36) Shiner, V. J. Jr.; Dowd, W. J. Am. Chem. Soc. 1971, 93, 1029−
1
(
1
(
030.
37) Perrin, C. L.; Ohta, B. K.; Kuperman, J. J. Am. Chem. Soc. 2003,
25, 15008−15009.
38) Perrin, C. L.; Ohta, B. K.; Kuperman, J.; Liberman, J.; Erdelyi,
M. J. Am. Chem. Soc. 2005, 127, 9641−9647.
39) Banait, N. S.; Jencks, W. P. J. Am. Chem. Soc. 1991, 113, 7958−
963.
REFERENCES
■
(
1) Taylor, M. E.; Drickamer, K. Introduction to Glycobiology; Oxford
University Press: Oxford, U.K., 2003.
(
2) Sinnott, M. Comprehensive biological catalysis: A mechanistic
reference; Academic Press: San Diego, 1998.
3) Sinnott, M. Carbohydrate chemistry and biochemistry: Structure and
mechanism; RSC Publishing: Cambridge, U.K., 2007.
4) Bennet, A. J.; Kitos, T. E. J. Chem. Soc., Perkin Trans. 2 2002,
207−1222.
5) Winstein, S.; Appel, B.; Baker, R.; Diaz, A. Spec. Publ.R. Soc.
Chem. 1965, 19, 109−130.
6) Lowry, T. H.; Richardson, K. S. Mechanism and theory in organic
chemistry, 3rd ed.; Harper & Row: New York, 1987.
7) Cowdrey, W. A.; Hughes, E. D.; Ingold, C. K.; Masterman, S.;
(
7
(
(
40) O’Ferrall, R. M. Adv. Phys. Org. Chem. 2010, 44, 19−122.
(
1
(
(41) Chateauneuf, J. E. J. Chem. Soc., Chem. Commun. 1991, 1437−
1
438.
(
42) McClelland, R. A.; Kanagasabapathy, V. M.; Steenken, S. J. Am.
Chem. Soc. 1988, 110, 6913−6914.
43) McClelland, R. A.; Cozens, F. L.; Li, J. H.; Steenken, S. J. Chem.
Soc., Perkin Trans. 2 1996, 1531−1543.
44) Amyes, T. L.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 7888−
900.
45) Huang, X.; Surry, C.; Hiebert, T.; Bennet, A. J. J. Am. Chem. Soc.
995, 117, 10614−10621.
46) Zhang, Y.; Luo, M. K.; Schramm, V. L. J. Am. Chem. Soc. 2009,
31, 4685−4694.
47) Errey, J. C.; Lee, S. S.; Gibson, R. P.; Fleites, C. M.; Barry, C. S.;
(
(
(
(
7
(
1
(
1
(
Scott, A. D. J. Chem. Soc. 1937, 1252−1271.
(
(
2
8) Guthrie, R.; Jencks, W. P. Acc. Chem. Res. 1989, 22, 343−349.
9) Sinnott, M. L.; Jencks, W. P. J. Am. Chem. Soc. 1980, 102, 2026−
032.
(
10) Lee, S. S.; Hong, S. Y.; Errey, J. C.; Izumi, A.; Davies, G. J.;
Davis, B. G. Nat. Chem. Biol. 2011, 7, 631−638.
11) Monegal, A.; Planas, A. J. Am. Chem. Soc. 2006, 128, 16030−
6031.
12) Soya, N.; Fang, Y.; Palcic, M. M.; Klassen, J. S. Glycobiology
011, 21, 547−552.
13) Fritz, T. A.; Hurley, J. H.; Trinh, L. B.; Shiloach, J.; Tabak, L. A.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 15307−15312.
14) Lairson, L. L.; Henrissat, B.; Davies, G. J.; Withers, S. G. Annu.
Rev. Biochem. 2008, 77, 521−555.
15) Schadt, F. L.; Schleyer, P. V.; Bentley, T. W. Tetrahedron Lett.
974, 2335−2338.
16) Perrin, D. D.; Armarego, W. L. F. Purification of laboratory
chemicals, 3rd ed.; Pergamon Press: Oxford, U.K., 1988.
17) Hardick, D. J.; Hutchinson, D. W. Bioorg. Med. Chem. Lett. 1994,
, 409−414.
18) Bennet, A. J.; Sinnott, M. L. J. Am. Chem. Soc. 1986, 108, 7287−
294.
19) Hayashi, M.; Hashimoto, S.; Noyori, R. Chem. Lett. 1984,
747−1750.
20) Chan, J.; Lewis, A. R.; Gilbert, M.; Karwaski, M. F.; Bennet, A. J.
Nat. Chem. Biol. 2010, 6, 405−407.
21) Gramera, R. E.; Park, A.; Whistler, R. L. J. Org. Chem. 1963, 28,
230−3231.
22) Montgomery, E. M.; Richtmyer, N. K.; Hudson, C. S. J. Org.
Chem. 1945, 10, 194−198.
(
1
(
2
Jung, P. M. J.; O’Sullivan, A. C.; Davis, B. G.; Davies, G. J. Angew.
Chem., Int. Ed. 2010, 49, 1234−1237.
(
̀
48) Ardevol, A.; Rovira, C. Angew. Chem., Int. Ed. 2011, 49, 10897−
1
0901.
(
(
49) Goedl, C.; Nidetzky, B. ChemBioChem 2009, 10, 2333−2337.
(
(
1
(
(
4
(
7
(
1
(
(
3
(
1
220
dx.doi.org/10.1021/ja209339j | J. Am. Chem.Soc. 2012, 134, 1212−1220