Mechanism of the hydrogen chloride/methanol-catalyzed mutarotation reaction of N-(p-chlorophenyl)-β-D-glucopyranosylamine

Article abstract of DOI:10.1002/1099-0690(200111)2001:22<4269::AID-EJOC4269>3.0.CO;2-G

The rate of the hydrogen chloride/methanol-catalyzed mutarotation of N-(p-chlorophenyl)-β-D-glucopyranosylamine has been studied polarimetrically at 16, 20, 25, and 30 °C. Rate constants and activation parameters have been determined for two parallel reac

Full text of DOI:10.1002/1099-0690(200111)2001:22<4269::AID-EJOC4269>3.0.CO;2-G

FULL PAPER
Mechanism of the Hydrogen Chloride/Methanol-Catalyzed Mutarotation
Reaction of N-(p-Chlorophenyl)-β- -glucopyranosylamine
D
Kazimiera Smiataczowa,*[a] Katarzyna Maj,[a] and Piotr Skurski[a]
Keywords: N-Glucosides / Mutarotation / Homogeneous catalysis / Thermodynamics
The rate of the hydrogen chloride/methanol-catalyzed muta-
rotation of N-(p-chlorophenyl)-β- -glucopyranosylamine has
been studied polarimetrically at 16, 20, 25, and 30 °C. Rate
constants and activation parameters have been determined between the experimental and calculated activation para-
for two parallel reactions. Enthalpies of formation, ∆Hof , and
meters are discussed. A relatively stable acyclic immonium
for the various structures involved in the anomerization.
The electrostatic potential around the p-chloroaniline N-
glucoside molecule has also been calculated. The differences
D
D-
entropies, S°, have been determined using the PM3 method ion was found to be an intermediate of the reaction.
Introduction
Elucidation of the mechanism of anomerization reactions
of biologically important N-glycosylamines has been the
subject of studies and speculation by many authors. Some
divergence in their views specifically concerns the nature of
the intermediate, in which a change in the configuration of
the anomeric carbon atom is feasible. The reaction is
strongly catalyzed by acids, thus suggesting a cationic form
of the N--glucoside as the intermediate.
The formation of a pseudoacyclic immonium ion (A in
Scheme 1), as proposed by Isbell and Frush[1,2] and also
noted by Paulsen and Pflughaupt,[3] appears to be the most
Scheme 1. Various structures for the intermediate of the mutarot-
probable reaction course. However, the results of certain ation of N-glycosides proposed in the literature
studies indicate the involvement of the ion-pair B (AϪ de-
notes a benzoate ion) rather than the cation.[4] The use of
a
culations were to determine the protonation site on the N-
14C tracer has shown that the cyclic ion C is actually in-
(p-chlorophenyl)-β--glucopyranosylamine molecule and to
seek confirmation of the interpretation that the acyclic im-
monium ion is an intermediate in the mutarotation of N-
glucosides.
volved in the mutarotation of N--glycosides,[5] as was also
found to be the case in the anomerization of ethyl α- and
β--xylopyranoside.[6] Furthermore, the authors of a num-
ber of reports have suggested that mutarotation of N--gly-
cosides proceeds via an acyclic Schiff base D,[7] as is the
case in the hydrolysis of N-glucosides,[1,8] and is analogous
to the mutarotation of sugars occurring via an aldehydic
form.[9]
Results and Discussion
The hydrogen chloride-catalyzed mutarotation reaction
of N-(p-chlorophenyl)-β--glucopyranosylamine at a ca.
10Ϫ6  concentration of the catalyst in methanol is a com-
plex process, as illustrated by the following equilibria:[10]
Studies on the hydrogen chloride/methanol-catalyzed
mutarotation of N-(p-chlorophenyl)-β--glucopyranosyla-
mine have shown that the process is complex, yielding a
relatively stable protonated N-glucoside as an interme-
diate.[10]
The purpose of this study was to determine the activation
parameters of two parallel reactions: the fast protonation
reaction of the β-anomer (with rate constant kr) and the
slow product formation reaction (α anomer; rate constant
ks). The results have been compared with those of theoret-
ical calculations carried out by means of the semi-empirical
PM3 procedure.[11] The aims of these experiments and cal-
(1)
The rate constants for the two parallel reactions, deter-
mined at different temperatures, are shown in Table 1. The
rate constant of the faster reaction (kr ϭ k1) refers to con-
version of the hydrated β anomer (Sβ) into its protonated
form SHϩ. The rate constant ks refers to the slow reaction,
i.e. that of the formation of the α anomer (Sα) from the
[a]
´
Faculty of Chemistry, University of Gdansk,
´
80Ϫ952 Gdansk, Poland
Eur. J. Org. Chem. 2001, 4269Ϫ4274
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
1434Ϫ193X/01/1122Ϫ4269 $ 17.50ϩ.50/0
4269

K. Smiataczowa, K. Maj, P. Skurski
FULL PAPER
SHϩ ion. Generally, it is a consecutive reaction involving
reversible steps, where ks ϭ (k1k2)/(kϪ1 ϩ k2). The depend-
ence of log (kr/T) and log (ks/T) on 1/T is shown in Fig-
ure 1, whereas the activation parameters calculated from ex-
Table 2. Experimental activation parameters for the fast and slow
reactions proceeding during the anomerization of N-(p-chloro-
phenyl)-β--glucopyranosylamine
perimental data for both kinetically controlled reactions, i.e. Reaction steps ∆H#/
∆S#/
∆G#/
(kcal·molϪ1
)
(cal·molϪ1·KϪ1
)
(kcal·molϪ1
)
the slow and the fast one, are presented in Table 2.
ks ϭ k1
kr ϭ k1·k2/
(kϪ1 ϩ k2)
16.195Ϯ0.115 6.58Ϯ1.25
19.639Ϯ0.348 14.28Ϯ3.97
14.234Ϯ0.389
15.380Ϯ0.1233
Table 1. Rate constants for the fast, kr, and slow, ks, reactions as
determined during the mutarotation of N-(p-chlorophenyl)-β--
glucopyranosylamine at different temperatures in methanol/HCl at
a HCl concentration of 3.72·10Ϫ6

The experimental thermodynamic activation parameters
were then compared with the results of theoretical calcula-
tions obtained by the semi-empirical PM3 method for an
identical process proceeding in the gas phase. The calcu-
lated heats of formation, ∆Hof,298, and entropies, ∆S2o98, for
particular structures involved in the anomerization of N-
(p-chlorophenyl)-β--glucopyranosylamine, as catalyzed by
solvated Hϩ ions, are summarized in Table 3. Table 4 lists
activation parameters for particular elementary steps of the
reaction, calculated on the basis of the data in Table 3.
Both the experimental characteristics and the results of
theoretical calculations are supportive of the mechanism
described by Equation (1). Nonetheless, it should be ex-
panded to Equation (2).
Temperature/
(°C)
K/
(T)
kr/
ks/
(10Ϫ2 minϪ1
)
(10Ϫ3 minϪ1
)
16
3.458·10Ϫ3
2.247Ϯ0.183
2.384Ϯ0.892
2.096Ϯ0.973
Ϫ
2.25Ϯ0.008
2.10Ϯ0.002
2.78Ϯ0.006
3.16Ϯ0.002
2.57Ϯ0.001
4.296Ϯ0.011
5.143Ϯ0.013
5.047Ϯ0.022
4.942Ϯ0.028
4.857Ϯ0.007
6.219Ϯ0.023
8.253Ϯ0.043
7.226Ϯ0.020
7.233Ϯ0.014
2.242Ϯ0.176
3.274Ϯ0.093
3.261Ϯ0.309
2.356Ϯ0.108
4.433Ϯ0.054
3.331Ϯ0.042
6.179Ϯ0.035
4.882Ϯ0.078
4.000Ϯ0.138
5.020Ϯ0.031
9.372Ϯ0.106
9.109Ϯ0.107
6.780Ϯ0.430
Ϫ
20
3.411·10Ϫ3
25
30
3.354·10Ϫ3
3.299·10Ϫ3
Accordingly, the mutarotation reaction is initiated by an
12.294Ϯ0.110 electrophilic attack of the CH3OH2ϩ ion on the in-ring oxy-
11.880Ϯ0.040
17.745Ϯ0.035
11.969Ϯ0.027
13.177Ϯ0.151
gen atom of the β-anomer molecule. The outcome of the
exothermic and spontaneous process occurring in the gas
phase is that activated complex III is formed with a bifurc-
ated (O···H····N) H-bond, as shown in Scheme 2.
Ϫ
8.420Ϯ0.074
13.401Ϯ0.018
Figure 1. Eyring plot for the fast [log(kr/T) ϭ f(1/T)] and slow [log(ks/T) ϭ f(1/T)] reactions proceeding during the anomerization of N-
(p-chlorophenyl)-β--glucopyranosylamine
4270
Eur. J. Org. Chem. 2001, 4269Ϫ4274

Mutarotation Reaction of N-(p-Chlorophenyl)-β-D-glucopyranosylamine
FULL PAPER
Table 3. Heats of formation and entropies calculated using the
PM3 method
The calculated activation enthalpy for this step is 26.7
kcal·molϪ1, which is 10 kcal·molϪ1 higher than that for the
ring-opening of the α-anomer (16.7 kcal·molϪ1) and higher
than the experimental value (16.2 kcal·molϪ1). The en-
hanced activation enthalpy for the β-anomer might be due
to the necessity of breaking the bifurcated H bond in com-
plex III. In solution, this effect can be compensated for by
the solvation enthalpy of the β-anomer. As a result of the
protonation, an acyclic intermediate IV is formed; this is
likely to be stable, as indicated by the deep minimum in Fig-
ure 2.
Structure
∆Hof /(kcal·molϪ1
)
S°/(cal·molϪ1·KϪ1
)
I
II
Ϫ196.491
156.564
142.251
58.502
157.416
47.703
149.443
57.065
160.457
156.310
142.473
III
TSβ
IV
V
TSα
VI
VII
Ϫ88.682
Ϫ61.946
Ϫ41.712
Ϫ51.878
Ϫ53.991
Ϫ70.656
Ϫ196.802
In the next step, the ring closes to form the α anomer
and there is simultaneous proton transfer to the methanol
Alternatively, the nitrogen atom of the N-glucoside might molecule, which is linked by a single H-bond to the N-gluc-
be the target of electrophilic attack. This, however, would oside molecule VI. The energy barrier for this step is higher
produce a stable ammonium ion, which could not undergo (40 kcal·molϪ1). The experimental value of the activation
further transformation[3] unless protonation of the nitrogen energy for the slow reaction is as low as 20 kcal molϪ1, but
atom led to the release of an amine molecule.[5] In this case, this is insufficient for ring closure to the α-configuration.
anomerization would proceed via a cyclic cation (C in Activation energies were determined for the rate constant
Scheme 1).[5] To solve this problem, the distribution of the of a two-step reversible reaction, where ks ϭ (k1·k2)/
electrostatic potential around the N-glucoside molecule was (kϪ1 ϩ k2). The kinetic barriers specified in Table 4 and
analyzed. This showed the lowest potential to be located on Figure 2 lead to the assumption that, to a good approxi-
the in-ring oxygen atom of the β-anomer. The potential on mation, ks ϭ (k1·k2)/(kϪ1). Consequently, the calculated ac-
the nitrogen atom is almost six times as high (values of tivation enthalpy for the slow reaction is ∆H#s ϭ ∆H1#
Ϫ34.34 and Ϫ5.52, respectively). This means that the oxy- ∆HϪ# 1 ϩ ∆H2#. In addition, if one assumes that the activa-
gen rather than the nitrogen atom is the most probable site tion energy for sugar ring-opening, ∆H1#, is equal to ∆HϪ# 2
Ϫ
,
of protonation on the N-glucoside.
then ∆H#s ϭ 16.7 Ϫ 31.7 ϩ 39.6 ϭ 24.6 kcal·molϪ1. The
Sugar ring-opening takes place simultaneously with calculated value of ∆H#s is thus slightly higher than the ex-
proton transfer from the CH3OH2ϩ ion to the oxygen atom. perimental one (19.6 kcal·molϪ1). However, it is common
Table 4. Activation parameters calculated using the PM3 method
Eur. J. Org. Chem. 2001, 4269Ϫ4274
4271

K. Smiataczowa, K. Maj, P. Skurski
FULL PAPER
Scheme 2. Reaction scheme for the transformation of N-(p-chlorophenyl)-β--glucopyranosylamine into N-(p-chlorophenyl)-α--glucopy-
ranosylamine suggested on the basis of the results of this work
knowledge that kinetic barriers for chemical reactions de- phase are lower than the experimental ones. Nevertheless,
termined by semi-empirical procedures are invariably over- the directions of the changes are consistent in both cases.
estimated.
The changes in the activation entropy have been found
Our calculations were performed on the elementary steps to be crucial for a thermodynamic analysis of anomeriz-
of the gas-phase reaction and did not take into account the ation.[4] It can thus be hypothesized that the entropy term
influence of solvation on the complex reaction proceeding should also be considered at particular steps of the reaction
in solution. To estimate the activation barriers more pre- catalyzed by strong acids. For this reason, the net effect of
cisely, one would have to use either a density functional changes in enthalpy and entropy has been presented in Fig-
(DFT) procedure or techniques enabling the thermodyn- ure 3 by showing the Gibbs’ free-energy change: ∆G° ϭ
amic parameters to be determined with greater accuracy. ∆H° Ϫ T∆S°.
Nevertheless, the principal aim of our calculations was to
The plot in Figure 3 is wholly different from the analog-
find supporting evidence in favor of experimental findings ous hypothetical relationship for the mutarotation of
and for this purpose the semi-empirical procedure adopted sugars.[1,2] The difference is that it reveals a low-energy
seems quite satisfactory.
stable intermediate IV. Its identification was possible owing
The experimental activation entropy for the fast reaction to the slightly stronger basicity of the nitrogen derivative of
is 6.6 cal·molϪ1·KϪ1 (Table 2). Such a small ∆S# value is glucose as compared to that of glucose itself. The pKb
characteristic of ring-opening producing an acyclic spe- values of the p-chloroaniline N-glucoside, α--glucose, and
cies.[12] In our case, this refers to the conversion of III into β--glucose are 12.33,[13] 12.83, and 12.53, respectively.[14]
IV and V. The experimental activation entropy for the slow Figure 3 shows that the α-anomer is less stable and that it
reaction is roughly ∆S1#
Ϫ ϩ ϭ 14.3 is protonated more readily than the β-anomer. Our calcula-
∆SϪ# 1 ∆S2#
cal·molϪ1·KϪ1. This means that ∆S2# Ͼ ∆SϪ# 1. The activa- tions have also demonstrated that one of the reasons for the
tion entropies calculated for the mutarotation in the gas stability of the β-anomer is the possibility of bifurcated H-
4272
Eur. J. Org. Chem. 2001, 4269Ϫ4274

Mutarotation Reaction of N-(p-Chlorophenyl)-β-D-glucopyranosylamine
FULL PAPER
On the basis of the experimental results and theoretical
calculations, it can be stated that the HCl-catalyzed mutar-
otation of N-(p-chlorophenyl)-β--glucopyranosylamine in
methanol proceeds via a low-energy acyclic immonium ion
formed by protonation of the in-ring oxygen atom.
Experimental Section
Reagents: Methanol was dried first with Na2SO4, then with iodine-
activated magnesium metal, and finally distilled from tartaric acid
to remove basic impurities.[16]
p-Chloroaniline glucoside was synthesized by
a methanolic
method[17] and crystallized from 96% ethanol; m.p. 145Ϫ147 °C
(ref.[16] 146 °C); initial specific rotation [α]52456 ϭ Ϫ124.3 (c ϭ 0.579,
methanol) (ref.[16] [α]
ϭ Ϫ137). Ϫ C12H16NO5Cl·H2O (307.7):
25
546
calcd. C 46.98, H 5.91, N 4.57; found C 46.51, H 5.89, N 4.62.
Methanolic HCl solution was prepared by saturating freshly puri-
fied methanol with hydrogen chloride; its concentration was later
determined potentiometrically.
Measurement of the Rate Constants: The rate constants for the mut-
arotation of N-glucosyl-p-chloroaniline were determined polari-
metrically by measuring the angle of rotation of the plane of polar-
ized light of wavelength 546 nm with time at an accuracy of
Ϯ0.005°. The measurements were made at 16, 20, 25, and 30 °C.
The solutions were placed in water-jacketed polarimetric tubes
thermostated to within Ϯ0.1 °C. The concentration of the N-gluc-
oside was 2·10Ϫ2  throughout. The concentration of the catalyst
(HCl) in methanol was 3.72·10Ϫ6 . The reaction started immedi-
ately after the addition of hydrogen chloride.
Figure 2. Energy barriers to the reactions
Under these conditions, two parallel first-order reactions occurred.
The rate constants for the two reactions were calculated using the
least-squares method. The rate constant for the slow reaction was
calculated from the following Equation (3), where t is the time in
minutes, αϱ is the final optical rotation of the solution, αt is the
optical rotation after time t, and αo is the initial rotation of the so-
lution.
ln(αϱ Ϫ αt) ϭ Ϫ kst ϩ ln(αϱ Ϫ αo)
(3)
The rate of the fast reaction, kr, was calculated as described else-
where[10] using the parameters of Equation (3).
Determination of the Activation Parameters: The activation para-
meters of both steps of the anomerization of p-chloroaniline N-
glucoside were determined from the temperature dependence of the
rate constants as expressed by the Eyring equation:
Figure 3. Profile of free energy changes during the mutarotation of
N-(p-chlorophenyl)-β--glucopyranosylamine
k ϭ (kBT)/h·e∆S#/R·eϪ ∆H#/RT where kB is the Boltzmann constant,
h is Planck’s constant, ∆S# is the entropy of activation, and ∆H#
is the enthalpy of activation.
The error in ∆H# was estimated in terms of the standard deviation
of the plot of the function log (k/T) ϭ f(1/T).
bond formation with the solvent molecule (III). It is prob-
ably for this very reason that the reaction between α--glu-
cose and p-chloroaniline affords a pure monohydrate of the
β-N--glucoside, which was used as an initial reactant in
this study. For the same reason, the β-form constitutes as
much as 82% of the mixture at equilibrium.[15]
The error in ∆S# was estimated by considering the errors in k (δk)
and ∆H# (δ∆H#) according to the following equation:[12]
δ(∆S#) ϭ [(R·{δk}/k)2 ϩ ({δ∆H#}/T)2]1/2
Eur. J. Org. Chem. 2001, 4269Ϫ4274
4273

K. Smiataczowa, K. Maj, P. Skurski
FULL PAPER
[2]
[3]
[4]
Calculations: The semi-empirical PM3 method,[11] as implemented
in a MOPAC-93 molecular orbital package,[18] was used to elucid-
ate the mechanism of the hydrogen chloride-catalyzed mutarotation
reaction of N-(p-chlorophenyl)-β--glucopyranosylamine. Owing
to the complex nature of the systems under study (they consist of
19 heavy atoms), ab initio or DFT calculations would have been
difficult to complete in a reasonable time. In view of this, a PM3
method was chosen, as it allows a reasonable prediction of the
heats of formation, geometries, dipole moments, and other physico-
chemical parameters characterizing a given system.[19] After prelim-
inary optimization of geometries by the molecular mechanics in-
corporated into the SPARTAN 4.0 program package,[20] the geo-
metries of all the systems were fully optimized following the EF
procedure.[21] Final energy gradients were invariably below 0.1
kcal·molϪ1, whereas eigenvalues of the Hessian matrix were all pos-
itive, thus confirming the stationary points on the potential energy
hypersurface corresponding to the energy minima. As far as the
procedure for seeking out the saddle points of the transformations
is concerned, the ‘‘saddle procedure’’ was first adopted to generate
rough structures.[22,23] These were subsequently refined using the
EF procedure.[21] Finally, the frequencies of the harmonic vibra-
tions characterizing the resulting molecules were calculated in order
to demonstrate that for all saddle points there is one and only one
imaginary harmonic frequency.
H. S. Isbell, W. Pigman, Adv. Carbohydr. Chem. Biochem. 1969,
24, 13.
H. Paulsen, K. W. Pflughaupt, Carbohydr. Chem. Biochem.
1980, I B, 881.
´
T. Jasinski, K. Smiataczowa, J. Sokołowski, Roczn. Chem.
´
1968, 42, 107; T. Jasinski, K. Smiataczowa, J. Sokołowski, Pol-
ish J. Chem. 1979, 53, 1235.
[5]
[6]
´
Z. Pawlak, E. Gorska, Roczn. Chem. 1969, 43, 1237.
R. I. Ferrier, L. R. Hatton, W. G. Overend, Carbohydr. Res.
1968, 8, 1237.
[7]
R. Kuhn, L. Bifkofer, Ber. Dtsch. Chem. Ges. 1938, 71, 1535;
G. P. Ellis, J. Honeyman, Adv. Carbohydr. Chem. 1955, 10, 104.
B. Capon, B. E. Connett, J. Chem. Soc. 1965, 3, 4497.
J. M. Los, L. B. Simpson, K. Wiesner, J. Chem. Soc. 1956, 78,
1564; P. W. Wertz, J. C. Garver, L. Anderson, J. Am. Chem.
Soc. 1981, 103, 3916.
[8]
[9]
[10]
[11]
K. Smiataczowa, Polish J. Chem. 1999, 73, 1513.
J. J. P. Stewart, J. Comput. Chem. 1989, 10, 209; J. J. P. Stewart,
J. Comput. Chem. 1989, 10, 221.
K. Schwetlick, Kinetische Methoden zur Untersuchen von Reac-
tionsmechanismen, PWN, Warszawa, 1975, p. 92, 109 (in Pol-
ish).
[12]
[13]
[14]
[15]
K. Smiataczowa, K. Maj, T. Widernik, M. Nesterowicz, Polish
J. Chem. 1998, 72, 587.
J. M. Los, L. B. Simpson, Recl. Trav. Chim. Pays-Bas 1956,
75, 267.
´
K. Smiataczowa, T. Jasinski, J. Sokołowski, Roczn. Chem.
1971, 45, 329.
The thermodynamic parameters (entropy, thermal energy contribu-
tions) were calculated by applying the methods of statistical ther-
modynamics.[24] The heats of formation (∆Hof ) and entropies (S°)
were then used to calculate the enthalpies (∆Hor ), entropies, and
free energies (∆Gor ) of reactions at 298.15 K under a standard pres-
sure of 1 atmosphere.[25]
[16]
[17]
[18]
K. Smiataczowa, J. Chem. Soc., Perkin Trans. 2 1989, 1593.
B. Helferich, W. Portz, Chem. Ber. 1955, 10, 141.
J. J. P. Stewart, MOPAC-93, Copyright Fujitsu, Tokyo, Japan,
1993.
[19]
[20]
J. J. P. Stewart, J. Comput. -Aided Mol. Design 1990, 4, 1.
Available from Wavefunction, Inc., 18401 Von Karman, Suite
370, Irvine, CA 92715.
J. Baker, J. Comput. Chem. 1986, 7, 385; 1988, 9, 465.
M. J. S. Dewar, E. F. Healy, J. J. P. Stewart, J. Chem. Soc.,
Faraday Trans. 2 1984, 80, 227.
M. J. S. Dewar, Int. J. Quantum Chem. Quantum Chem. Symp.
1988, 22, 557.
M. J. S. Dewar, G. P. Ford, J. Am. Chem. Soc. 1977, 99, 7822.
P. W. Atkins, Physical Chemistry, 5th ed., Oxford University
Press, Oxford, 1994.
All calculations were carried out using an IBM RS/6000 3CT
workstation.
[21]
[22]
[23]
Acknowledgments
This research was supported by the Polish State Committee for
[24]
[25]
Scientific Research under grant BW/8000-5-0306Ϫ9.
Received November 2, 2000
[1]
H. S. Isbell, H. L. Frush, J. Org. Chem. 1958, 23, 1309.
[O00554]
4274
Eur. J. Org. Chem. 2001, 4269Ϫ4274