Thermodynamic and kinetic considerations in the chemoselective O-acylation by mixed anhydrides. A semiempirical MO approach

Article abstract of DOI:10.1016/j.tetlet.2004.03.045

A simple methodology to achieve high chemoselective O-acylation of primary hydroxy groups, in the presence of secondary ones, was performed by means of the quick generation of a mixed anhydride using mild and inexpensive conditions. Steric and electronic

Full text of DOI:10.1016/j.tetlet.2004.03.045

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 3349–3353
Thermodynamic and kinetic considerations in the chemoselective
O-acylation by mixed anhydrides. A semiempirical MO approachq
,
Antonio J. Mota,* Rafael Robles,* Luis Alvarez de Cienfuegos and Alberto Lamenca
ꢀ
Departamento de Quꢀımica Orgꢀanica, Facultad de Ciencias, Universidad de Granada. Campus de Fuentenueva, 18071 Granada, Spain
Received 23 January 2004; revised 3 March 2004; accepted 8 March 2004
Abstract—A simple methodology to achieve high chemoselective O-acylation of primary hydroxy groups, in the presence of sec-
ondary ones, was performed by means of the quick generation of a mixed anhydride using mild and inexpensive conditions. Steric
and electronic factors are involved in this preference. As far as the geometries are concerned, semiempirical calculations were carried
out in order to determine the nature of the contribution to the thermodynamics and kinetics. From a qualitative point of view, it was
found that the thermodynamics are sugar dependent whereas the kinetics are mixed anhydride dependent. These conclusions fit
quite well with the experimental results allowing the use of semiempirical MO calculations as a tool for modulating the selectivity
ratio for this process since different systems may be modeled in the same way.
Ó 2004Elsevier Ltd. All rights reserved.
The selective protection of functional groups is one of
the most critical aspects of many synthetic strategies.
Satisfactory chemoselectivity is often achieved by taking
advantage of electronic differences of the functional
groups or of their distinct steric environments. Since the
latter factors constitute usually the most important
contribution for the selectivity in organic reactions,
many strategies employing either hindered substrates or
bulky reagents (or both) have been developed.1 How-
ever, it is very difficult to introduce small protecting
groups with acceptable selectivity through direct and
inexpensive methods because reagents involved in the
protecting process (as acetyl chloride in sugar acetyla-
tion) readily react with all the unprotected positions,
having no way to discriminate them on the basis of steric
and electronic factors, leading to mixtures. This is the
reason why per-acylated sugars are common starting
materials in carbohydrate chemistry.2
are readily available, versatile and economical reagents.
For instance, mixed anhydrides synthesized with sodium
[13C] acetate and pivaloyl chloride, have been employed
particularly for the synthesis of labelled compounds.4
Among other different strategies leading to the title
compounds it is worth mentioning the reaction of thal-
lium carboxylates with acyl chlorides,5 of arenediazo-
nium salts and sodium carboxylates under palladium
catalysis,6 the reaction between the corresponding free
acid and ketenes,7 or from the free acid and acyl chloride
in a solid-phase co-polymer.8 Related to mixed anhy-
drides, different interesting studies have been carried
out, for example, their stabilization by the triethylbor-
oxin complex formation,9a photochemical cleavage
reactions,9b as well as their use in the electroreduction of
benzylic and allylic halides to afford ketones.9c Although
numerous reports on the synthetic utility of anhydrides
have been reported, only very limited information is
available concerning their reactivity as a function of
their structure.10
Anhydrides are in many cases reactive enough to be
employed instead of acyl halides, which usually react
using well controlled conditions.3 Moreover, anhydrides
This work describes the use of mixed anhydrides as a
source of different acyl protecting groups (acetyl and
benzoyl), the acylation process being chemoselective on
account of structural parameters.
q Supplementary data associated with this article can be found, in the
online version, at doi:10.1016/j.tetlet.2004.03.045
Thus, mono-saccharides 1,2-O-isopropylidene-a-
(1), 3-O-benzoyl-1,2-O-isopropylidene-a- -gluco- (2),
1-O-benzyl-2,3-O-isopropylidene-a- -sorbo- (3) and 2,3-
O-isopropylidene-a- -sorbo-furanose (4), have been
selected as starting materials in order to study the scope
D-xylo-
* Corresponding authors. Tel.: +34-958-243185; fax: +34-958-248437;
e-mail addresses: mota@quantix.u-strasbg.fr; rrobles@ugr.es
Present address: Laboratoire de Chimie Quantique, Institut Le Bel,
D
L
L
ꢀ
Universite Louis Pasteur, 4rue Blaise Pascal, 67070 Strasbourg,
France.
0040-4039/$ - see front matter Ó 2004Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.03.045

3350
A. J. Mota et al. / Tetrahedron Letters 45 (2004) 3349–3353
HO
HO
O
O
HO
OH
OBz
O
O
O
O
O
O
HO
HO
OH
O
O
OBn
OH
OH
O
O
1
2
3
4
Figure 1. Compounds used as starting materials.
Table 1. Yields afforded in the acylation reaction
of the selective 5-(6-) O-acetylation and 5-(6-) O-ben-
zoylation. Hence, compound 2 is a 1,2-diol, whereas
compounds 1, 3 are 1,3-diols. Moreover, the 1,3-diol 4
has an additional more hindered primary hydroxy group
(Fig. 1).
Compounds
1
2
3
4
5-(6-)-O-Acetyl
(%Yield)
5 (77)
7 (75)
9 (76)
11 (63)
5-(6-)-O-Benzoyl
(%Yield)
6 (68)
8 (66)
10 (70)
––a
Acylations were performed by reaction of the corre-
sponding sugars (1–4) with a solution of either acetic
pivalic (for acetylation) or benzoic pivalic (for benzoyl-
ation) mixed anhydride, which were generated in situ.11
Both acetylation and benzoylation took place in a highly
chemoselective way, since acylation of the secondary
hydroxyl group was negligible in all cases, affording the
corresponding 5-O-acyl or 6-O-acyl derivatives,
depending on the sugar (Fig. 2).
a Complex mixture.
Calculations were carried out as close-shell type (RHF)
using the Polak–Ribiere optimization algorithm, except
for transition states for which eigenvector-following
algorithm was employed. Convergence limits for all the
evaluated systems were fixed at 0.01 kcal Aꢀ1 molꢀ1
ꢁ
(RMS gradient) for the geometry optimization process
and 0.001 kcal molꢀ1 for the iterative SCF calculation.
Geometries for the starting molecules were modeled
before by molecular mechanics (Amber 99 force field).
It is worth mentioning that compound 4 reacts to give 11
as the major product. In this case, small amounts of
1-O-acetyl and 1,6-di-O-acetyl derivatives were
observed. Actually, it was found in all cases that the
concurrent reaction was the protection of the primary
hydroxy functionality by the pivaloyl moiety. As
expected, this undesired reaction was more important
for benzoylation (for which chemoselectivity decreased)
and was directly responsible for the moderate yields
achieved for the major products. This data is summa-
rized in Table 1.
In order to evaluate the thermodynamics of the different
processes, geometry optimization calculations (ground
state) were carried out by choosing the simplest 1,2- and
1,3-diols, namely compounds 1 and 2, since these reac-
tions seems to be similar for the other derivatives. In
addition, both primary and secondary hydroxy group
mono-substituted derivatives have also been considered
for comparison, including the undesired pivaloylation
reaction (Tables 2 and 3).
We can explain these results on the basis of a double
steric induction for the acyl-transfer reaction: first, the
more accessible and reactive primary hydroxyl group
and second, the more accessible acetyl or benzoyl group
in the mixed anhydride (Fig. 3). Obviously, the larger
size of the benzoyl group with respect to the acetyl one
leads to a lack of selectivity in the acylation process.
From this data, several interesting conclusions can be
drawn. Firstly, in the case of the reactions on the pri-
mary hydroxy group (Table 2), all the processes are
thermodynamically favored and only slight differences
can be observed along the acetylation–benzoylation–
pivaloylation series, both methods AM1 and PM3 being
almost coincident. In addition, for the 1,2-diol 2 these
processes are between 1.7 and 2.7 kcal molꢀ1 less
favoured than they are for compound 1. This fact can
be explained considering that the entering group in the
The computational study was performed by means of
AM1 and PM3 semiempirical MO calculations, both
methods implemented in the Hyperchem 7.5 package.12
RO
RO
O
O
HO
OH
OBz
O
O
O
O
O
O
RO
RO
OBn
OH
O
O
OH
OH
O
O
11 R = Ac
5 R = Ac
6 R = Bz
7 R = Ac
8 R = Bz
9
R = Ac
10 R = Bz
Figure 2. Main acylation products.

A. J. Mota et al. / Tetrahedron Letters 45 (2004) 3349–3353
3351
HO
R
ROCO
O
OH
OH
O
O
O
OH
O
O
O
O
O
O
Figure 3. The schematic acyl-transfer process.
Table 2. Acetylation, benzoylation and pivaloylation on the primary
hydroxy group for compounds 1 and 2
between the reactions on 1 and 2. This clearly means
that PM3 takes into account factors that AM1 simply
ignores, that is to say, AM1 overestimates the stability
of these compounds. On the other hand, PM3 shows
frequently the opposite tendency. Such a behavior is
widely known for these methods.13 Considering that in
this case compound 2 has two neighboring secondary
substituted hydroxy groups (at 3 and 5 positions), the
results afforded by PM3 become quite reasonable on
account of steric factors.
Reaction
Substrate
AM1a
PM3a
Acetylation
1
2
ꢀ9.58
ꢀ7.84
ꢀ8.86
ꢀ6.23
Benzoylation
1
2
ꢀ9.84
ꢀ7.71
ꢀ7.99
ꢀ6.26
Pivaloylation in the
acetylation
1
2
ꢀ9.26
ꢀ7.53
ꢀ8.58
ꢀ5.88
Finally, due to the fact that only very slight energy
differences can be observed along the series when vary-
ing the size of substituents (acetyl–benzoyl–pivaloyl),
and that energy differences come only from different
substituted positions on the sugar, it can be concluded
that the thermodynamics take mainly into account the
environment of the sugar, that is to say, it is sugar-
sterically determined. This is the origin of the high
chemo-selectivity observed.
Pivaloylation in the
benzoylation
1
2
ꢀ9.67
ꢀ7.94
ꢀ9.00
ꢀ6.30
a Energy values in kcal molꢀ1. See the supplementary material section
for further details.
Table 3. Acetylation, benzoylation and pivaloylation on the secondary
hydroxy group for compounds 1 and 2
On the other hand, the steric environment at the mixed
anhydride (the acylating agent) has been shown to
influence the regioselectivity observed since the major
products correspond to a transfer of the smaller com-
ponent of the mixed anhydride to the sugar skeleton.
Since the different size of these substituents is not a crit-
ical factor in the energy of the final products, the envi-
ronment at the mixed anhydride must be the determinant
step in the reaction, controlling kinetically the preferred
acetyl or benzoyl addition towards the pivaloylation
process. Therefore, for these reactions, kinetics is mainly
mixed anhydride-sterically determined. In order to
evaluate this influence, it was necessary to establish the
corresponding transition states.
Reaction
Substrate
AM1a
PM3a
Acetylation
1
2
ꢀ8.06
ꢀ6.10
ꢀ6.52
ꢀ1.87
Benzoylation
1
2
ꢀ8.04
ꢀ6.40
ꢀ6.73
ꢀ2.13
Pivaloylation in the
acetylation
1
2
ꢀ7.95
ꢀ6.01
ꢀ6.31
ꢀ1.65
Pivaloylation in the
benzoylation
1
2
ꢀ8.36
ꢀ6.26
ꢀ6.73
ꢀ2.07
a Energy values in kcal molꢀ1. See the supplementary material section
for further details.
These calculations aimed at transition states with forms
that could be considered as six-membered half chair-like
intermediates, where the hydroxy group participates
doubly in a simultaneous process: the hydrogen atom is
being transferred to the most distant carbonyl on the
mixed anhydride while acylation takes place over the
oxygen (Fig. 4).
molecule exhibits a greater hindrance when the hydroxy
group is placed closer (that is, for 1,2-diols). Note also
that, as expected, the pivaloylation process is slightly
more favored when a benzoylation reaction is carried
out because of the greater hindrance of the benzoyl
group with respect to the acetyl group.
Then, in the case of the reactions on the secondary
hydroxy group (Table 3), the same tendency can be seen
for the energies, which only vary slightly along the ser-
ies, showing again that pivaloylation is slightly more
favored when benzoylation is carried out. However, now
the energy differences exhibited between AM1 and PM3
for the thermodynamically less stable substituted 1,2-
diol are too large (4.0–4.5 kcal molꢀ1), although is the
same difference (4.5–5 kcal molꢀ1) that PM3 itself shows
A complete chemoselectivity between primary and sec-
ondary hydroxy groups is expected since the complex is
too voluminous. However, it is possible to have transi-
tion states where either the acetyl (or benzoyl) or the
pivaloyl moieties are the reaction targets, the latter
resulting in the observed side reaction. Both possibilities
have been studied for acetic pivalic and benzoic pivalic
mixed anhydrides in the selective 5-O-acylation of
compound 1 (Fig. 5).

3352
A. J. Mota et al. / Tetrahedron Letters 45 (2004) 3349–3353
Table 4. AM1 and PM3 energy differences between A–B states for the
acylation process (kcal molꢀ1
)
AM1 A–B energy
PM3 A–B energy
Acetylation
Benzoylation
ꢀ2.97
ꢀ1.04
ꢀ1.31
þ0.31
The energy difference between the A and B states is a
qualitative measure of the kinetic preference for the
smaller substituents. The results afforded fit quite well
with the selectivity encountered experimentally (Table
4).
These values show that the AM1 method always con-
siders the A states as the lowest energy complexes, the
benzoylation being less favorable as expected. In con-
trast, PM3 calculations place pivaloylation (B state)
only a little more favorably than benzoylation.
In summary, we have presented a new method for
selective protection of primary hydroxy groups in sugar
derivatives using an extremely simple protocol. This
selectivity is based on the fact that introducing a steri-
cally hindered component (namely pivalic acid) in the
mixed anhydride leads to a bulky structure that reacts
with the less hindered group in the substrate transferring
their smaller component. This concept of double
induction can be considered separately in both ther-
modynamic and kinetic factors. Semiempirical MO
calculations have been carried out in order to determine
the nature of each contribution, finding that the ther-
modynamics are not affected by the different size of the
substituents but the specific environment in the sugar
moiety (sugar control). Conversely, the kinetics deter-
mine the type of substituent that will be introduced
(mixed anhydride control). These results indicate the
fact that a wide number of different mixed anhydrides
can be used in the same manner, being able to modulate
both chemo- and regioselectivity.
Figure 4. AM1 transition-state geometrics for the acetylation (left) and
benzoylation (right) of 1 with the corresponding mixed anhydrides.
Supplementary material
Energies and Cartesian coordinates of all the calculated
structures are available.
Acknowledgements
ꢀ
We thank to the Ministerio de Ciencia y Tecnologıa of
Spain for the grant to L. Alvarez de Cienfuegos, and to
ꢀ
ꢀ
the Junta de Andalucıa for the research grant to A.J.M.
References and notes
Figure 5. Above, the two B intermediates (acetyl and benzoyl) leading
to pivaloylation (AM1). Below, a schematic representation of the
transition states leading to the major products (A) and the undesired
pivaloylation reaction (B).
1. (a) Hanessian, S. Total Synthesis of Natural Products: The
‘Chiron’ Approach; Pergamon: Oxford, 1986; (b) Collins,

A. J. Mota et al. / Tetrahedron Letters 45 (2004) 3349–3353
3353
P. M.; Ferrier, R. J. Monosaccharides. Their Chemistry and
their Roles in Natural Products; John Wiley and Sons:
Chichester, England, 1995.
8. Fife, W. K.; Zhang, Z. Tetrahedron Lett. 1986, 27, 4933–
4936.
€
ꢀ
€
9. (a) Koster, R.; Sporzynski, A.; Schußler, W.; Blser, D.;
Boese, R. Chem. Ber. 1994, 127, 1191–1199; (b) Penn,
J. H.; Owens, W. H. J. Am. Chem. Soc. 1993, 115, 82–
86; (c) D’Incan, E.; Sibille, S.; Perichon, J.; Moingeon,
M. O.; Chaussard, J. Tetrahedron Lett. 1986, 27, 4175–
4176.
2. For some recent applications: (a) Uriel, C.; Santoyo-
ꢀ
Gonzalez, F. Synlett 1999, 593–595; (b) Gandolfi-Dona-
dio, L.; Gallo-Rodrıguez, C.; de Lederkremer, R. M. J.
ꢀ
Org. Chem. 2002, 67, 4430–4435; (c) Li, M.; Han, X.; Yu,
B. J. Org. Chem. 2003, 68, 6842–6845; (d) Maier, M. A.;
Yannopoulos, C. G.; Mohamed, N.; Roland, A.; Fritz, H.;
Mohan, V.; Just, G.; Manoharan, M. Bioconjugate Chem.
2003, 14, 18–29.
10. (a) Lee, I.; Cha, O.-J.; Lee, B.-S. J. Phys. Org. Chem. 1990,
3, 279–284; (b) Penn, J. H.; Owens, W. H.; Petersen, J. L.;
Finklea, H. O.; Snider, D. A. J. Org. Chem. 1993, 58,
2128–2133.
3. For some recent applications: (a) Izquierdo, I.; Plaza,
ꢀ
ꢀ
M.-T.; Robles, R.; Rodrıguez, C.; Ramırez, A.; Mota,
A. J. Eur. J. Org. Chem. 1999, 1269–1274; (b) Izquierdo,
I.; Plaza, M.-T.; Robles, R.; Mota, A. J. Eur. J. Org.
Chem. 2000, 2071–2078; (c) Carr, J. A.; Bisht, K. S.
Tetrahedron 2003, 59, 7713–7724; (d) Jeannot, F.; Goss-
elin, G.; Math, C. Org. Biomol. Chem. 2003, 1, 2096–2102.
4. (a) Kelly, N. M.; Reid, R. G.; Willis, C. L.; Winton, P. L.
Tetrahedron Lett. 1995, 36, 8315–8318; (b) Le Sann, C.;
Simpson, T. J.; Smith, D. I.; Watts, P.; Willis, C. L.
Tetrahedron Lett. 1999, 40, 4093–4096; (c) Harding, J. R.;
Hughes, R. A.; Kelly, N. M.; Sutherland, A.; Willis, C. L.
J. Chem. Soc., Perkin. Trans. 1 2000, 3406–3416.
5. Taylor, E. C.; McLay, G. W.; McKillop, A. J. Am. Chem.
Soc. 1968, 90, 2422–2423.
11. Pivaloyl chloride (1 mmol) was added to dry CH2Cl2
(20 mL). To this solution was added sodium acetate or
sodium benzoate (10 mmol). The suspension was placed
into an ultrasound bath and after 5 min, a solution of the
corresponding sugar (1 mmol) in CH2Cl2 (10 mL) was
added. After the disappearance of the starting material the
mixture was washed with water (15 mL), the organic
extract was dried (MgSO4) and the solvent was evaporated
under reduced pressure. The residue was purified by
column chromatography to afford the corresponding
acylated products.
12. Hyperchem Release 7.5 Hypercube, Inc., 1115 NW 4th
Street, Gainesville, Florida 32601, USA.
13. (a) Foresman, J. B.; Frisch, A. E. Exploring Chemistry
with Electronic Structure Methods; 2nd ed.; Gaussian:
Pittsburgh, PA, (USA), 1996; (b) Jensen, F. Introduction to
Computational Chemistry; John Wiley and Sons: Chiches-
ter, England, 1999; Chapter 3.
6. Kikukawa, K.; Kono, K.; Nagira, K.; Wada, F.; Matsuda,
T. J. Org. Chem. 1981, 46, 4413–4416.
7. Williams, J. W.; Dickert, Y. J.; Krynitsky, J. A. J. Am.
Chem. Soc. 1941, 63, 2510–2511.