Reagent-dependent regioselective control in multiple carbohydrate esterifications

Article abstract of DOI:10.1021/jo0620821

(Chemical Equation Presented) Regioselective control in organotin-mediated multiple acylation of carbohydrates is presented. The acylation reagent could be efficiently used to direct the product formation. Reagent-dependent thermodynamic and kinetic contr

Full text of DOI:10.1021/jo0620821

easily prepared and generally lead to intermediate structures with
predictable reactivities. In these reactions, stoichiometric amounts
of organotin reagent are normally used.
Reagent-Dependent Regioselective Control in
Multiple Carbohydrate Esterifications
However, of particular importance in this respect is the
possibility of acquiring multiple protections in single-step
processes, and so far, no efficient, general methods have been
developed. Interestingly, a protocol was recently described
where products with one or two free hydroxyl groups were
produced by use of excess organotin reagent.9 This potentially
general approach is very convenient and efficient for multiple
protection schemes. Combining this organotin method with the
Lattrell-Dax (nitrite-mediated) carbohydrate epimerization method
addressed recently,10 very convenient and highly efficient
methods to modify carbohydrate structures that traditionally
require many steps can be developed.11-16 Such structures are
increasingly important for many biological processes or as drug
candidates.17,18
Hai Dong,† Zhichao Pei,† Styrbjo¨rn Bystro¨m,†,‡ and
Olof Ramstro¨m*,†
KTH - Royal Institute of Technology, Department of Chemistry,
Teknikringen 30, S-10044, Stockholm, Sweden, and BioVitrum
AB, Department of Medicinal Chemistry, S-11276 Stockholm,
Sweden
ramstrom@kth.se
ReceiVed October 7, 2006
In order to advance the organotin-mediated multiple carbo-
hydrate protection method, a study of regioselective single-step
acylations of unprotected pyranosides was initiated. During this
study, it was, however, found that the multiple esterification
processes were highly dependent on the acyl reagent used. These
interesting results prompted us to further explore the mecha-
nisms how the multiple carbohydrate esterification process was
regioselectively controlled and to predict the outcome of
structures that normally require many steps with traditional
synthesis.
The general multiprotection process is outlined in Scheme
1. The unprotected glycoside was first treated with an excess
amount (2-3 equiv) of dibutyltin oxide, producing a stannylene
intermediate that was not isolated. This intermediate was
subsequently treated with the acylation reagent to yield the
protected products in a one-pot process.
Regioselective control in organotin-mediated multiple acy-
lation of carbohydrates is presented. The acylation reagent
could be efficiently used to direct the product formation.
Reagent-dependent thermodynamic and kinetic control and
dynamic assistance mechanisms are suggested, resulting in
the efficient preparation of building blocks that normally
require many steps with traditional synthesis.
When acetyl chloride was used in the acylation process, the
main product using galactoside 1 as reactant was found to be
the 4,6-di-O-acetyl product 2 (35%), together with lower
amounts of triacetylated products (Scheme 2, see also the
Supporting Information). To our surprise, a different pattern was,
however, formed when acetic anhydride was instead used as
the acylating reagent. In this case, the 3,6-di-O-acetyl product
3 (46%) proved to be the major product of the process, whereas
no 4,6-di-O-acetyl product 2 could be isolated. Similar differ-
ences were found when 3 equiv of organotin reagent was used
in the process, yielding the 3,4,6-tri-O-acetyl product 4 (57%)
when acetyl chloride was used and the 2,3,6-tri-O-acetyl product
Regioselectivity is a prominent challenge in carbohydrate
chemistry since carbohydrates contain several hydroxyl groups
of similar reactivity. Selective protecting groups and efficient
protecting group strategies are therefore of crucial importance
to efficiently obtain desired carbohydrate structures. Carbohy-
drate hydroxyl groups differ somewhat in reactivity depending
on whether they are anomeric, primary, or secondary and also
depending on their configurations. These differences in reactivity
can sometimes be utilized so that a desired protection pattern
can be achieved in one step without the use of more complex
reaction sequences.1,2 For obtaining monosubstituted compounds
in one or a few steps, the use of organotin reagents such as
tributyltin oxide or dibutyltin oxide3 provides a useful means
to efficient regioselective acylations,4 alkylations,5 silyations,6
sulfonylations,7 and glycosylations.8 Stannylene acetals are
(9) Zhang, Z. Y.; Wong, C. H. Tetrahedron 2002, 58, 6513-9.
(10) Dong, H.; Pei, Z. C.; Ramstro¨m, O. J. Org. Chem. 2006, 71, 3306-
9.
(11) Yu, H.; Ensley, H. E. Tetrahedron Lett. 2003, 44, 9363-6.
(12) Graziani, A.; Passacantilli, P.; Piancatelli, G.; Tani, S. Tetrahedron
Lett. 2001, 42, 3857-60.
† KTH - Royal Institute of Technology.
‡ Biovitrum AB.
(1) Kattnig, E.; Albert, M. Org. Lett. 2004, 6, 945-8.
(2) Kurahashi, T.; Mizutani, T.; Yoshida, J. J. Chem. Soc., Perkin Trans.
1 1999, 465-73.
(3) David, S.; Hanessian, S. Tetrahedron 1985, 41, 643-63.
(4) Peri, F.; Cipolla, L.; Nicotra, F. Tetrahedron Lett. 2000, 41, 8587-
90.
(5) Jenkins, D. J.; Potter, B. V. L. Carbohydr. Res. 1994, 265, 145-9.
(6) Glen, A.; Leigh, D. A.; Martin, R. P.; Smart, J. P.; Truscello, A. M.
Carbohydr. Res. 1993, 248, 365-9.
(7) Takeo, K.; Shibata, K. Carbohydr. Res. 1984, 133, 147-51.
(8) Hodosi, G.; Kovac, P. J. Am. Chem. Soc. 1997, 119, 2335-6.
(13) Liakatos, A.; Kiefel, M. J.; von Itzstein, M. Org. Lett. 2003, 5,
4365-8.
(14) Pei, Z. C.; Dong, H.; Ramstro¨m, O. J. Org. Chem. 2005, 70, 6952-
5.
(15) Wu, X. Y.; Bundle, D. R. J. Org. Chem. 2005, 70, 7381-8.
(16) Huang, X. F.; Huang, L. J.; Wang, H. S.; Ye, X. S. Angew. Chem.,
Int. Ed. 2004, 43, 5221-4.
(17) Witczak, Z. J.; Culhane, J. M. Appl. Microbiol. Biot. 2005, 69, 237-
44.
(18) Rich, J. R.; Szpacenko, A.; Palcic, M. M.; Bundle, D. R. Angew.
Chem., Int. Ed. 2004, 43, 613-5.
10.1021/jo0620821 CCC: $37.00 © 2007 American Chemical Society
Published on Web 01/19/2007
J. Org. Chem. 2007, 72, 1499-1502
1499

SCHEME 1. Example of Organotin-Mediated Multiple
Carbohydrate Esterification
SCHEME 3. Reagent-Dependent Multiprotection of Methyl
â-D-Glucopyranoside 9a
a Reagents and conditions: (a) (i) Bu2SnO (2.2 equiv), MeOH, 70 °C, 2
h, (ii) Ac2O, toluene, 0-22 °C, 6 h; (b) (i) Bu2SnO (2.2 equiv), MeOH, 70
°C, 2 h, (ii) AcCl, toluene, 0-22 °C, 6 h.
explaining the effect of benzoyl migration at elevated temper-
ature whereas acetyl groups migrate already at room tempera-
ture.
SCHEME 2. Reagent-Dependent Multiprotection of Methyl
â-D-Galactopyranoside 1a
However, this tentative mechanism does not take into account
the apparent effects of acetic anhydride. In contrast to benzoyl
chloride and acetyl chloride, it was found that migration with
acetic anhydride does not seem to take place at room temper-
ature. The high yield of, for example, the 2,3,6-tri-O-acetyl
product 5 using acetic anhydride is rather an effect of kinetic
control. Thus, the experimental data indicate that migration could
only be observed with acetyl chloride at room temperature
(thermodynamic control), whereas acetic anhydride proved
inefficient in this reaction (kinetic control). More subtle differ-
ences were recorded for glucoside 9 due to similar reactivities
of the stannylene acetals (Scheme 3). However, this interpreta-
tion also accounts for why, in addition to the 2,3,6-tri-O-acetyl
product 10 mainly formed, the 3,4,6-tri-O-acetyl product 11 was
formed when glucoside 9 was treated with acetic anhydride,
whereas the 2,4,6-tri-O-acetyl product 12 was formed when
acetyl chloride was used under the same conditions. No 2,4,6-
tri-O-acetyl product 12 could be formed using acetic anhydride
on account of its thermodynamically controlled formation.
a Reagents and conditions: (a) (i) Bu2SnO (2.2 equiv), MeOH, 70 °C, 2
h, (ii) AcCl, toluene, 0-22 °C, 6 h; (b) (i) Bu2SnO (2.2 equiv), MeOH, 70
°C, 2 h, (ii) Ac2O, toluene, 0-22 °C, 6 h; (c) (i) Bu2SnO (3.3 equiv), MeOH,
70 °C, 2 h, (ii) AcCl, toluene, 0-22 °C, 6 h; (d) (i) Bu2SnO (3.3 equiv),
MeOH, 70 °C, 2 h, (ii) Ac2O, toluene, 0-22 °C, 6 h.
5 (70%) when acetic anhydride was employed, respectively. This
difference in product formation was subsequently further
explored. Recently, an organotin-mediated benzoyl group
migration pattern was suggested for different diols and also
carbohydrates.19,20 This effect could also be used in the present
case, starting from galactoside 1 and employing benzoyl group
migration at higher temperature. The 4,6-di-O-benzoyl product
6 (80%) was thus obtained (see the Supporting Information).
At lower temperature, the 3,6-di-O-benzoyl product 7 (90%)
was instead formed suggesting that no migration can take place
under these conditions. Benzoyl migration to the 3,4,6-tri-O-
benzoyl product 8 (90%) was also formed at higher temperature
when 3 equiv of organotin reagent was used. By comparison
with the results for the acetylating reagents, this seems to suggest
that an organotin-mediated acetate group migration can in this
case take place already at room temperature, but only when
acetyl chloride is used. The relatively high yields of the 4,6-
di-O-acetyl product 2 and the 3,4,6-tri-O-acetyl product 4 are
in this case thermodynamically controlled by acetyl migration
at lower temperature.
A tentative explanation as to why migration takes place with
acetyl chloride, whereas it is inefficient with acetic anhydride,
is outlined in Figure 1. With acetyl chloride, the migration
follows the route presented in Figure 1a, based on previous
studies,20 where the resulting chloride anion coordinates to the
tin center. However, with acetic anhydride, bidentate coordina-
tion of the resulting acetate to the organotin intermediate takes
place instead,21-23 dramatically reducing the Lewis acidic effect
of the organotin-intermediate so that transesterification will not
take place (Figure 1b), and the kinetic product is formed.
According to this explanation, it was hypothesized that this
kinetically controlled reaction should become thermodynami-
cally controlled at higher temperature from temperature-de-
pendent reversibility of the acetoxystannane complex. With
acetic anhydride, the yield of the kinetic product 2,3,6-tri-O-
acetylgalactoside 5 also decreased from 70% at room temper-
ature to 18% at 90 °C due to the acetyl group migration from
the 3-position to the 4-position. In this case, however, low
selectivity between the 2- and 3-positions yielded a mixture of
2,4,6-tri-O-acetylgalactoside 13 and 3,4,6-tri-O-acetylgalactoside
4. Similar results were in this case obtained for benzoyl group
According to the tentative organotin benzoyl group migration
mechanism suggested,19,20 the resulting alkoxystannane inter-
mediate is able to attack the neighboring acyl carbonyl group
with acyl transfer as a consequence. If this mechanism is valid,
it is reasonable to assume that acyl groups in general are able
to migrate under these conditions, however showing different
migration rates. It is well-known that acetyl groups are more
reactive than benzoyl groups in migration reactions, thus
(21) Okawara, Y. M. a. R. J. Organomet. Chem. 1967, 247-56.
(22) Mokal, V. B.; Jain, V. K. J. Organomet. Chem. 1992, 441, 215-
26.
(19) Bredenkamp, M. W.; Spies, H. S. C. Tetrahedron Lett. 2000, 41,
543-6.
(20) Roelens, S. J. Org. Chem. 1996, 61, 5257-63.
(23) Hasha, D. L. J. Organomet. Chem. 2001, 620, 296-302.
1500 J. Org. Chem., Vol. 72, No. 4, 2007

Information). This was especially pronounced in chloroform,
where the 2,6-di-O-benzoyl glucoside 15 was obtained in 51%
yield.
As reported, poor selectivity was obtained for acylation of
the hydroxyl groups in the 2- and 3-positions of â-D-glucosides
when the hydroxyl groups in the 4- and 6-positions were
protected with benzylidene groups.7,24 However, in the examples
with the different acylation reagents mentioned above, good
selectivity was instead obtained between the 2- and 3-positions
of the free glucoside 9. Based on these findings, a tentative
dynamic assistance mechanism is suggested (Figure 2). Thus,
for the acylation reagent benzoyl chloride, by virtue of the large
organotin group in the 4-position obstructing the approaching
benzoyl group, the accessibility of the 3-position is disfavored.
As a result, high selectivity was obtained for the 2-position
(Figure 2a). For the acylation reagent acetic anhydride on the
other hand, the generated acetate group can bridge the two tin
centers, inducing stannylene migration from the 2,3-position to
the 3,4-position. As a result of this tentative dynamic assistance
mechanism, the 3-position was acylated with high selectivity
(Figure 2b).
In the course of the studies, it was also found that the solvent
played an important role for the outcome of the multiprotection
reactions (cf. Scheme 4). To further explore how the solvent
polarity influences the selectivity, a range of solvents of different
polarity were evaluated. To avoid acyl group migration, acetic
anhydride was chosen as acylation reagent. The results are
displayed in Table 1, clearly indicating the solvent importance
in the reactions. For galactoside 1, although the 3,6-di-O-acetyl
product 3 was obtained in moderate yield in toluene (46%),
this product was obtained in 90% yield in DMF and 75% yield
in acetonitrile (Table 1).
FIGURE 1. Difference in organotin-mediated acetyl migration: (a)
acetyl chloride used as reagent, generating a monodentate chlorostan-
nane complex; (b) acetic anhydride used as reagent producing a less
reactive bidentate acetoxystannane.
SCHEME 4. Reagent-Dependent Multiprotection of Methyl
â-D-Glucopyranoside 9a
For the glucoside 9, a similar pattern was seen. Whereas the
3,6-di-O-acetyl product 14 was obtained in very low yield in
toluene (22%), it could be obtained in 70% yield in DMF. The
experimental data clearly indicate that the solvent can indeed
control the selectivity of the organotin-mediated esterification
reactions. Good selectivity was always obtained when the
esterification reactions were done in the more polar solvent.
The reason for this effect is likely due to decreased reactivity
of the esterification reagent from solvent-induced destabilization
of the stannylene intermediates.25
Finally, the results obtained for the â-D-galactosides and the
â-D-glucosides were applied to two other carbohydrate struc-
tures, methyl R-D-mannoside 17 and methyl R-D-glucoside 18.
In agreement with previous reports,24 it was reasoned that the
1,2-cis-glucoside would possess a reactivity order in the
organotin-mediated multiprotection reactions of 6 > 2 > 3 >
4, due to increased electron density of the 2-hydroxyl group,
and possibly also to diminished dynamic assistance effects. The
1,2-trans-mannoside would on the other hand follow the order
6 > 3 > 2 > 4, mainly due to the axial configuration of the
2-position. These predictions proved valid in these reactions,
and both the di-O-acetyl and the tri-O-acetyl products for both
structures could be produced in good yields. In conclusion, it
has been shown that organotin-mediated multiple carbohydrate
esterifications can be controlled by the acylating reagent and
the solvent polarity. When acetyl chloride is used, the reactions
a Reagents and conditions: (a) (i) Bu2SnO (2.2 equiv), MeOH, 70 °C, 2
h, (ii) Ac2O, toluene or DMF, 0-22 °C, 6 h; (b) (i) Bu2SnO (2.2 equiv),
MeOH, 70 °C, 2 h, (ii) BzCl, toluene or CHCl3, 0-22 °C, 6 h.
migration at high temperature when benzoyl chloride was used
(see the Supporting Information).
Upon analysis of the resulting product compositions when
using acetic anhydride as acylation reagent, a reactivity order
of the hydroxyl groups can be estimated. For â-D-galactoside,
the reactivity order can be arranged as 3 > 6 > 2 > 4.
Furthermore, the stability order can instead be arranged as 6 >
4 > 3 > 2 from the migration results using acetyl chloride as
acylation reagent. Similarly, the reactivity order for the different
hydroxyl groups of the â-D-glucoside in organotin-mediated
esterification can be arranged as 6 > 3 ) 2 > 4 and the stability
order as 6 > 2 > 3 > 4.
However, the results obtained also reveal an additional
reagent-controlled migration mechanism (Scheme 4). For ex-
ample, the 3,6-di-O-acetyl-glucoside 14 was obtained with acetic
anhydride, whereas the 2,6-di-O-benzoyl glucoside 15 was
instead obtained with benzoyl chloride at room temperature.
Since no organotin-mediated migration occurred with neither
acetic anhydride nor benzoyl chloride at room temperature, it
is clear that these results are not brought about by the migration
mechanism mentioned above. The studies also indicated that
only the 3,6-di-O-acetyl glucoside 14 was obtained when the
stannylene intermediate was treated with acetic anhydride and
no other di-O-acetyl products. When the intermediate was
allowed to react with benzoyl chloride on the other hand, the
2,6-di-O-benzoyl glucoside 15 was always obtained in higher
yield than the 3,6-di-O-benzoyl glucoside 16 (see the Supporting
(24) Williams, J. M.; Richardson, A. C. Tetrahedron 1967, 23, 1369-
78.
(25) David, S.; Thieffry, A.; Forchioni, A. Tetrahedron Lett. 1981, 22,
2647-50.
J. Org. Chem, Vol. 72, No. 4, 2007 1501

FIGURE 2. Dynamic assistance controlled acylation: (a) acylation with benzoyl chloride preferentially occurring in the 2-position; (b) acetate-
assisted migration of the stannylene resulting in acetylation in the 3-position.
TABLE 1. One-Pot Organotin-Mediated Multiple Carbohydrate
Esterification Using Acetic Anhydride in Polar Solvents
in 30 mL of methanol and refluxed for 2 h. After evaporation of
the solvent, the residue was dried under vacuum and then dissolved
in 10 mL of solvent (toluene, DMF, or chloroform). After the
solution was cooled to 0-5 °C, a solution of acylation reagent (2.2
mmol or 3.3 mmol) in anhydrous solvent (1 mL) was added
dropwise and then allowed to react at room temperature for 2-12
h. The resulting mixture was directly purified by flash column
chromatography (2:1 to 1:1 hexane-ethyl acetate), yielding the
acylated products.
Methyl 4,6-di-O-benzoyl-â-D-galactopyranoside (6): 1H NMR
(CDCl3, 400 MHz) δ 7.38-8.14 (m, 10 H), 5.66 (d, 1 H, J3,4 3.2
Hz), 4.55 (dd, 1 H, J6a,6b 11.3 Hz, J6a,5 6.7 Hz), 4.34 (dd, 1 H, J6a,6b
11.3 Hz, J6b, 5 6.7 Hz), 4.27 (d, 1 H, J1,2 7.8 Hz), 4.04 (t, 1H, J6,5
6.7 Hz), 3.89 (dd, 1 H, J2,3 9.62 Hz, J3,4 3.2 Hz), 3.77 (dd, 1 H, J1,2
7.8 Hz, J2,3 9.62 Hz), 3.58 (s, 3 H); 13C NMR (CDCl3, 125 MHz)
δ 166.6, 166.3, 133.7, 133.4, 130.2, 129.9, 129.7, 129.4, 104.1,
72.5, 72.1, 71.5, 70.2, 62.6, 57.5; [R]20D ) -57 (c ) 1.8, CHCl3).
Anal. Calcd for C21H22O8·1/2H2O: C, 61.31; H, 5.64. Found: C,
61.26; H, 5.30.
reagent
(equiv)
yielda
(%)
reactant
solvent
DMF
product
1
1
1
9
9
17
17
18
18
Ac2O (2.2) 3,6-di-OAc-â-D-Gal 3
90
75
85
70
85
80
90
80
90
CH3CN Ac2O (2.2) 3,6-di-OAc-â-D-Gal 3
CH3CN Ac2O (3.3) 2,3,6-tri-OAc-â-D-Gal 5
DMF
Ac2O (2.2) 3,6-tri-OAc-â-D-Glc 14
Methyl 3,4,6-tri-O-benzoyl-â-D-galactopyranoside (8): 1H
NMR (CDCl3, 400 MHz) δ 7.2-8.1 (m, 15 H), 5.91 (d, 1 H, J3,4
3.45 Hz), 5.39 (dd, 1 H, J3,2 10.1 Hz, J4,3 3.45 Hz), 4.65 (dd, 1 H,
J6a,6b 11.3 Hz, J6a,5 6.5 Hz), 4.47 (d, 1 H, J1,2 7.6 Hz), 4.37 (dd, 1
H, J6a,6b 11.3 Hz, J6b,5 6.5 Hz), 4.24 (d, 1 H, J5,6 6.5 Hz), 4.10 (dd,
1 H, J2,3 10.1 Hz, J2,1 7.6 Hz), 3.64 (s, 3 H); 13C NMR (CDCl3 125
MHz) δ 166.2 166.1 165.6, 133.7, 133.4, 133.3, 130.1 130.0, 129.9,
129.6, 129.4, 129.3, 128.7, 128.6, 128.4, 104.5, 73.5, 71.4, 70.1,
CH3CN Ac2O (3.3) 2,3,6-tri-OAc-â-D-Glc 10
CH3CN Ac2O (2.2) 3,6-di-OAc-R-D-Man 19
CH3CN Ac2O (3.3) 2,3,6-tri-OAc-R-D-Man 20
CH3CN Ac2O (2.2) 2,6-di-OAc-R-D-Glc 21
CH3CN Ac2O (3.3) 2,3,6-tri-OAc-R-D-Glc 22
a Isolated yield.
are under thermodynamic control, whereas when acetic anhy-
dride is employed, kinetic control takes place. A dynamic
assistance control has also been suggested for stannylene
migration induced by acetate. Very good selectivity can
furthermore be obtained in more polar solvents. These results
have been used in the efficient preparation of a range of
prototype carbohydrate structures.
68.3, 62.2, 57.7; [R]20 ) -3 (c ) 2.6, CHCl3). Anal. Calcd for
D
C28H26O9: C, 66.40; H, 5.17. Found: C, 66.17; H, 5.08.
Acknowledgment. This study was supported by the Swedish
Research Council, the Swedish Foundation for International
Cooperation in Research and Higher Education, and the Carl
Trygger Foundation.
Supporting Information Available: General methods and
product distributions for acylation reactions in toluene. This material
is available free of charge via the Internet at http://pubs.acs.org.
Experimental Section
General Acylation Procedure. Methyl â-D-glycoside (194 mg,
1 mmol) and dibutyltin oxide (550 mg, 2.2 mmol) were dissolved
JO0620821
1502 J. Org. Chem., Vol. 72, No. 4, 2007