A solvation-assisted model for estimating anomeric reactivity. Predicted versus observed trends in hydrolysis of n-pentenyl glycosides

Article abstract of DOI:10.1021/jo9601223

An attempt has been made to predict qualitative trends in reactivity at the anomeric center, using N-bromosuccinimide-induced hydrolysis of n-pentenyl glycosides (NPGs) as the experimental model. Calculated relative activation energies based on internal e

Full text of DOI:10.1021/jo9601223

5280
J . Org. Chem. 1996, 61, 5280-5289
A Solva tion -Assisted Mod el for Estim a tin g An om er ic Rea ctivity.
P r ed icted ver su s Obser ved Tr en d s in Hyd r olysis of n -P en ten yl
Glycosid es1
C. Webster Andrews,†, Robert Rodebaugh,‡,§ and Bert Fraser-Reid*,‡
Division of Medicinal Chemistry, Glaxo Wellcome, Research Triangle Park, North Carolina 27709, and
Paul M. Gross Chemical Laboratory, Duke University, Durham, North Carolina 27708
Received J anuary 19, 1996X
An attempt has been made to predict qualitative trends in reactivity at the anomeric center, using
N-bromosuccinimide-induced hydrolysis of n-pentenyl glycosides (NPGs) as the experimental model.
Calculated relative activation energies based on internal energy differences between a reactant
and the associated intermediate are not always in agreement with experimental observations.
However, solvation energies obtained by the generalized Born surface area model in MacroModel
developed by Still et al. give modified activation energies that are in excellent agreement with the
experimentally observed trends. It is shown that the solvation model does not disturb the normally
observed reactivity trends that can be rationalized on the basis of internal energies alone. The
value of the methodology has been demonstrated for several substrates by first calculating their
relative activation energies, then testing them experimentally, and finding excellent agreement
with predictions.
In tr od u ction
differences provide the basis of a protocol whereby two
saccharides equipped with the same anomeric activating
group could be coupled efficiently.6 This plan, dubbed
armed/ disarmed glycoside coupling, has since been ap-
plied to a variety of glycosyl donors, showing that the
phenomenon is not restricted7 to the n-pentenyl glyco-
sides8 that had been used in our initial investigations.
Furthermore, in 1991 we demonstrated that widely used
cyclic acetals also affected anomeric reactivity to such an
extent that an armed/disarmed strategy could be also be
devised around them.9 The latter phenomenon which
arises from torsional strain complements the former
which was ascribed to electronic factors.10 Thus armed/
disarmed strategies could be based upon either electronic
or torsional considerations.
Other studies11 designed to see if glycoside hydrolysis
(a) involves boat conformations12 and (b) requires that
the leaving group be presented with an antiperiplanar
lone pair13 prompted us to carry out ab initio studies11
to probe the energetics and preferred geometries for
glycoside cleavage.14 Axial and equatorial conformers of
2-methoxytetrahydropyran served as models for R- and
For laboratory syntheses of complex oligosaccharides
in the current state of the art,2 protecting groups must
usually be stationed on the glycosyl donor and acceptor
so as to enforce coupling at the desired site(s). Although
some rules of engagement may be gleaned from literature
precedents,3 a prelude of trial and error is usually needed
before the best partners can be ascertained. Such a
procedure is wasteful of manpower, time, and resources,
and the situation could be relieved if it were possible to
evaluate, and thereby screen, potential partners compu-
tationally. Ideally such a computational tool would be
useful (and used) only if it blended simplicity with
reliability. The experimental tool should be similarly
simplesideally thin layer chromatography (TLC) or high-
performance liquid chromatography (HPLC). This is the
long-range objective of a program that has been initiated
in our laboratory, and in this paper we describe some
recent results.
Ba ck gr ou n d
Although protecting groups prevent some sites from
competing during a reaction, it is well-known that they
can also profoundly affect the reactivity of the entire
molecule.4,5 Thus in 1988 we showed that such reactivity
(4) See, for example: Paulsen, H. Angew. Chem., Int. Ed. Engl. 1982,
21, 155.
(5) See, for example: Mehta, S.; Andrewes, J . S. J ohnston, B. D.;
Svenson, B.; Pinto, B. M. J . Am. Chem. Soc. 1995, 117, 9783. Garegg,
P. J .; Hutberg, H.; Wallin, S. Carbohydr. Res. 1982, 10897.
(6) Mootoo, D. R.; Konradsson, P.; Udodong, U.; Fraser-Reid, B. J .
Am. Chem. Soc. 1988, 110, 5583.
† Glaxo Wellcome.
‡ Duke University.
§ Winner of the Charles R. Hauser Fellowship, 1995-1996.
Present address: Natural Products and Glycotechnology (NPG)
Research Institute, 4118 Swarthmore Rd, Durham, NC 27707.
X Abstract published in Advance ACS Abstracts, J uly 1, 1996.
(1) We are grateful to the National Science Foundation (CHE
9311356) and Glaxo Research Institute for financial support of this
work.
(2) For some recent reviews on oligosaccharide syntheses, see:
Khan, S. H., O’Neil, R. A., Rahman, A., Eds. Modern Methods in
Carbohydrate Synthesis; Harwood Academic Publishers: Amsterdam
1996. Fraser-Reid, B.; Madsen, R.; Campbell, A. S.; Roberts, C.;
Merritt, J . R. In Chemical Synthesis of Oligosaccharides; Hecht, S. M.,
Ed.; Oxford University Press: Oxford, 1995, in press. Khan, S. H.;
Hindsgaul, O. In Frontiers in Molecular Biology; Fukuda, M., Hinds-
gaul, O., Eds.; IRL Press: Oxford, 1994; p 206. Toshima, K.; Tatsuta,
K. Chem. Rev. 1993, 15031. Bonoub, J .; Boullanger, P.; Lafont, D.
Chem. Rev. 1992, 92, 1167.
(7) See, for example: Roy, R.; Anderson, F. O.; Letellier, M.
Tetrahedron Lett. 1992, 33, 6053. Sliedregt, L. A. J . M.; Zegelaar-
J aarsvald, K.; van der Marel, G. A.; van Boom, J . H. Synlett 1993,
335. Boons, G. J .; Isles, S. Tetrahedron Lett. 1994, 35, 3593.
(8) Fraser-Reid, B.; Udodong, U. E.; Wu, Z.; Ottosson, H.; Merritt,
R.; Rao, C. S.; Roberts, C.; Madsen, R. Synlett 1992, 927.
(9) Fraser-Reid, B.; Wu, Z.; Andrews, C. W.; Skowronski, E.; Bowen,
J . P. J . Am. Chem. Soc. 1991, 113, 1434.
(10) Fraser-Reid, B.; Wu, Z; Udodong, U. E.; Ottosson, H. J . Org.
Chem. 1990, 55, 6068.
(11) Ratcliffe, A. J .; Mootoo, D. R.; Andrews, C. W.; Fraser-Reid, B.
J . Am. Chem. Soc. 1989, 111, 7661.
(12) Deslongchamps, P. Stereoelectronic Effects in Organic Chem-
istry; Pergamon Press: New York, 1983; pp 30-35. Intramolecular
Strategies and Stereochemical Effects: Glycosides and Orthoesters
Hydrolysis Revisited. In The Anomeric Effect and Associated Stereo-
electronic Effects; Thatcher, G. R. J ., Ed.; ACS Symposium Series 539,
American Chemical Society: Washington, D.C., 1993.
(13) Sinnott, M. L. Adv. Phys. Org. Chem. 1988, 24, 113.
(3) For such a compilation that focuses on anomeric selectivity,
see: Barresi, F.; Hindsgaul, O. J . Carbohydr. Chem. 1995, 14, 1043.
S0022-3263(96)00122-3 CCC: $12.00 © 1996 American Chemical Society

Hydrolysis of n-Pentenyl Glycosides
J . Org. Chem., Vol. 61, No. 16, 1996 5281
Sch em e 1
Sch em e 2
â-glycosides, respectively, and it was found that they
were hydrolyzed through half-chair (4H3) and sofa (4E)
conformers and that they proceeded therefrom to oxo-
carbenium ions having the same conformations, respec-
tively.15
Our calculations15 also agreed with the well-known fact
that â-glucosides generally react faster than their R-coun-
terparts, the â/R ratio of ∼2:1 being found to agree with
experimentally obtained ratios for aqueous hydrolysis of
some methyl glycosides.16 However it should be noted
that subsequent experiments by Wilson and Fraser-Reid
have found that the â/R ratios can vary widely, depending
on substitution and protecting group patterns.17
In connection with the above-described study concern-
ing the effects of torsional strain on glycoside hydrolysis,
we had attempted to correlate the experimentally ob-
served times for hydrolysis with calculated activation
energies computed with the PM3 semiempirical Hamil-
tonian.19 The results were encouraging, but, not surpris-
ingly, hardly adequate given the major approximations
that had been made. For example, the activation ener-
gies were estimated to be the difference between internal
energies of the glycoside I and the corresponding oxo-
carbenium ion II as shown in eq i, Scheme 1. This
estimation involves the assumption that geometries of
the transition states are the same as those of the
associated glycosyl oxocarbenium ion intermediates, this
approximation being justified on the grounds that transi-
tion states for glycoside hydrolyses are known to be late.18
Additionally, our use of the PM3 Hamiltonian rendered
this shortcut even more precarious owing to the fact that
the resulting bond rotation barriers are very low,19 as a
result of which the structures obtained are rather flexible.
Thus the oxocarbenium ions could distort away from the
Syn th esis of Su bstr a tes
It is well-known that R- and â-anomers (can) have
different rates of hydrolysis, and in order to avoid this
and other configurational effects, we initially confined our
attention to derivatives of pent-4-enyl â-D-glucopyrano-
side (1âb). However, for reasons that will become clear
below some R-D anomers (see Table 2) were subsequently
studied. Compound 1âb obtained by de-O-acetylation of
pent-4-enyl 2,3,4,6-tetra-O-acetyl-â-D-glucopyranoside (1âa)
was prepared by the procedure of Rodriguez and Stick.20
Compounds 2, 4, 5, 6b, and 8 (Scheme 2) were prepared
by routine procedures, or by adapting literature methods
developed for the corresponding methyl glucosides.21
Ley’s procedure22 was followed for preparation of the
dispiroketal (dispoke) derivatives 9 and 10 (Scheme 3),
from which 11 and 12 were obtained without event. The
structure of 11 was confirmed by X-ray analysis, and the
ORTEP diagram is shown in Figure 1.
The benzylidene derivative 13 was problematic. We
first attempted its preparation by treating the diol 3 with
bisdihydropyran and camphorsulfonic acid in the usual
way.22 However these conditions caused cleavage of the
benzylidene ring. Next an attempt was made to ben-
zylidenate the dispoke diol 9 with R,R-dimethoxytoluene
and camphorsulfonic acid, but cleavage of the dispoke
moiety was now a major problem and only traces of 13
were obtained.
4
4
normal H3 and E ring conformations that are obtained
at a higher level of theory.15
A second approximation, as is evident from Scheme 1
eq i, was to ignore solvation factors in order to simplify
the calculations. This was not inconsequential, in view
of the obvious expectation that the glycosides I and
oxocarbenium ions II should have very different solvation
energies and that protecting groups should have different
solvation effects on I and II.
We now report that substantial improvements in the
predictive value of the calculated energies are achieved
by use of both internal and solvation energies of species
I and II.
(14) Andrews, C. W.; Bowen, J . P.; Fraser-Reid, B. J . Chem. Soc.,
Chem. Commun. 1989, 1913.
(15) Andrews, C. W.; Fraser-Reid, B.; Bowen, J . P. J . Am. Chem.
Soc. 1991, 113, 8293.
(16) (a) See Table I in ref 15. (b) Isbell, H. S.; Frush, H. L. J . Res.
Natl. Bur. Std. (U.S.) 1949, 24, 125.
(17) Wilson, B. G.; Fraser-Reid, B. J . Org. Chem. 1995, 60, 317.
(18) Bennett, A. J .; Sinnott, M. L. J . Am. Chem. Soc. 1986, 108,
7287.
(20) Rodriquez, F. B.; Stick, R. V. Austr J . Chem. 1990, 43, 673.
(21) Cesare, P. D.; Gross, B. Carbohydr. Res. 1976, 483. Garegg, P.
J .; Iverson, T.; Oscarson, S. Carbohydr. Res. 1976, 50, C-12.
(22) Ley, S. V.; Leslie, R.; Tiffin, P. D.; Woods, M. Tetrahedron Lett.
1992, 33, 4767. Ley, S. V.; Boons, G.-J .; Leslie, R.; Woods, M.;
Hollinshead, D. Synthesis 1993, 689. Ley, S. V.; Downham, R.;
Edwards, P. J .; Innes, J . E.; Woods, M. Contemp. Org. Synth. 1995, 2,
365.
(19) Stewart, J . J . Comp.-Aided Mol. Design 1990, 4, No. 1.

5282 J . Org. Chem., Vol. 61, No. 16, 1996
Andrews et al.
Sch em e 3
for geometry optimization with the 6-31G* basis set or
for post Hartree-Fock treatment necessary to deal with
electron correlation.24,27
In keeping with the work which had been done in our
previous report,15 the structures of the glycosides in Table
1 and their corresponding oxocarbenium ions were built
using MacroModel28 and subjected to PM3 geometry
optimization and energy evaluation. The activation
energies were calculated using eq i of Scheme 1 and then
made relative to the reference compound 2â, the results
being shown in column 2 of Table 1. It is seen that these
energies correlate poorly with the experimentally ob-
served hydrolysis times in column 7, since shorter reac-
tion times should have correlated with lower activation
energies.
Our method of estimating activation energies therefore
needed to be improved. It was noted above that there
are usually problems with PM3 calculations of oxocar-
benium ion structures. By contrast, these structures are
generally easily and reliably obtained by geometry opti-
mizations using the 3-21G ab initio basis set and Mac-
roModel starting structures.28 A set of 3-21G activation
energies was thereby obtained,25,26 and once again, these
were made relative to reference compound 2â, as shown
in column 3 of Table 1. However, the results so obtained
were found to follow the PM3 energies shown in column
2, which meant that the discrepancies with the experi-
mental data had not been alleviated.
We next resorted to the bis(pent-4-enyl) acetal transfer
reagents which have been recently developed in our
laboratory for use under mild conditions.23 In that report
it was found that the acetalization reactions did proceed
under the agency of N-halosuccinimides, but were greatly
accelerated by trace amounts of protic or Lewis acids.
Nevertheless the process has been shown to be driven
by halonium ionssnot by protons.23 Accordingly, treat-
ment of diol 9 with benzaldehyde bis(pent-4-enyl) acetal,
NBS, and camphorsulfonic acid for 10 min at room
temperature afforded a 72% yield of 13.
For preparation of the R-anomers, D-glucose was
subjected to Fischer glycosidation with pent-4-enol, and
the product was acetylated in order to separate the
anomers (Scheme 4). The desired R-anomers were then
processed as described above for the corresponding
â-counterparts.
In a second attempt at improving the correlation, we
decided to include solvation energies. As illustrated in
Figure 2, solvation in water stabilized the oxocarbenium
(24) The relative activation energies in column 5 of Table 1, column
4 of Table 2, and the last column of Table 3 deserve comment. By
calculating relative activation energies (equivalent to the use of
isodesmic reactions involving reactants and intermediates), it is
possible to cancel out many of the errors inherent in using low levels
of theory. 25,26 The principal error is that electron correlation is ignored
at the Hartree-Fock level for both ground states and oxocarbenium
ions. However, the change in relative activation energy from molecule
to molecule is quite large relative to the rate changes and reflects the
fact that some errors are not cancelled out, as well as the fact that
real electronic and steric differences exist. In the end, the relative
ordering, which is what this study requires, is satisfactory. These
relative activation energies (kinetic properties) should not be compared
against known ground state energy differences (thermodynamic prop-
erties).
(25) Hehre, W. J .; Radom, P.; Schleyer, P. v. R.; Pople, J . A. Ab Initio
Molecular Orbital Theory; J ohn Wiley & Sons: New York, 1986 (see p
271, 298, 307).
Exp er im en ta l Hyd r olyses
(26) Wiberg, K. B.; Murcko, M. A. J . Am. Chem. Soc. 1989, 111, 4821.
J effrey, G. A.; Pople, J . A.; Binkley, J . S.; Vishveshwara, S. J . Am.
Chem. Soc. 1978, 100, 373. Andrews, C. W. Ph.D. Thesis, Duke
University, 1989.
The standard conditions for NBS-induced hydrolysis
of the test substrates are described in detail in the
Experimental Section. The times shown in Table 1,
column 6, are for disappearance of the starting materials
as judged by TLC, and these were subsequently con-
firmed by HPLC analysis of the crude reaction mixtures
(see Experimental Section). These times are shown as
relative data in column 7. (The data assembled in Tables
1 and 2 are simplified, more user-friendly versions of
those presented in Table 3).
(27) Due to the cyclic nature of these compounds, conformational
differences between solution and in vacuo are expected to be small.
This allowed us to do in vacuo quantum mechanics with the expectation
that the same conformation would be obtained in aqueous solution.
This will not always be the case. A reviewer points out that a general
scheme for more flexible molecules would be to do geometry optimiza-
tion (or indeed, conformational searching) using one of the continuum
solvation models that are now available in semiempirical or ab initio
codes. A further assumption inherent in this work is that the single
minimized conformation found for each structure (formally at 0 K)
represents the ensemble of structures at the temperature of the rate
determination. This is also reasonable due to the cyclic nature of the
compounds. An additional point relates to the use of 6-31G* charges
but not energies. High-level charges are required to obtain accurate
solvation energies from the GB/SA model. Since the charges are not
expected to change a great deal with geometry optimization, we used
6-31G*//3-21G charges and did not pursue 6-31G* geometry optimiza-
tion (and more importantly, could not afford the cpu time involved).
We did not use the single-point 6-31G* energies in the rate correlations
because these energies are not minimized on the 6-31G* energy
surface.
Ca lcu la ted Activa tion En er gies
It should be noted that the rigors of computing absolute
activation energies are too involved to be undertaken in
support of our synthetic work. The intent here is to obtain
relative reactivities as an aid to synthesis. For the same
reason, “reaction times” are determined by TLC and
HPLC. The molecules under consideration are too large
(28) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J . Comput. Chem.
1990, 11, 440.
(23) Madsen, R.; Fraser-Reid, B. J . Org. Chem. 1995, 60, 772.

Hydrolysis of n-Pentenyl Glycosides
J . Org. Chem., Vol. 61, No. 16, 1996 5283
F igu r e 1. ORTEP diagram of the X-ray structure for pent-4-enyl 4,6-di-O-benzyl-2,3-O-(octahydro-2,2′-bi-2H-pyran-2,2′-diyl)-
â-D-glucopyranoside (11). Diagram (40% probability ellipsoids) showing the crystallographic atom numbering scheme and solid-
state conformation; C(8) and C(9) are disordered over two positions. Small circles represent hydrogen atoms.
Sch em e 4
energy and is a function of the conformation and solvent
accessible surface area (SA) of the solute. Term two is
an electrostatic polarization energy computed using
generalized Born (GB) theory. Solvent is treated as a
continuum, and explicit solvation is not required.
A
reasonable conformation and high-quality charges are
necessary to compute an accurate GB/SA solvation
energy.29 High-quality charges require the use of a good
basis set and electrostatic potential fitting, so we have
used 6-31G* ESP-fit charges computed with the Pop )
CHELPG option in the Gaussian 92 program.30 The
solute structures used are 3-21G-optimized, however,
since the size of the dispoke protecting group prohibits
6-31G* geometry optimization.
The charges so obtained were used with the 3-21G-
optimized structures to compute the solvation energy
difference between reactant and reaction intermediate
(see Scheme 1, eq ii) and then made relative to 2â. The
relative solvation energy differences (rel ∆Gs) shown in
column 4 of Table 1, were then added to the relative
activation energies based on internal energies [rel Ea(g),
i.e., column 3 of Table 1] to give the solvation-corrected
activation energies [rel Ea(s), eq iii of Scheme 1 and
column 5 of Table 1].
ion much more than the glycoside owing to the charged
nature of the former. In the context of the thermody-
namic cycle in Figure 3, we use the approximation that
the internal energy change associated with activation,
Ea (internal), is equal to the gas phase activation energy,
Ea(g). Furthermore, the activation energy in solution
Ea(s) is equal to Ea(internal) modified by solvation. It is
clear that protecting groups influence the activation
energy both from an internal energy standpoint, since
some stabilize the oxocarbenium ion better than others,
and from a solvation standpoint, since some are solvated
better than others. The energy cycle in eq iv of Figure 3
has a net change of zero. The terms can be rearranged
to yield an expression for the relative activation energy
in solution composed of a term for the relative gas phase
activation energy and another for the relative solvation
energy shown in eq vi of Figure 3. The data resulting
from eq vi are shown as rel Ea(s) in Tables 1 and 2.
Our solvation energies are calculated using the GB/
SA solvation model of Still et al.29 This is a two-term
model. Term one consists of a cavity plus van der Waals
Arrhenius analyses,31 done by plotting log of the
reaction times in column 7 versus the activation energies
before and after solvation correction, are shown in
(29) GB/SA solvation model: Still, W. C.; Tempczyk, A.; Hawley,
R. C.; Hendrickson, T. J . Am. Chem. Soc. 1990, 112, 6127.
(30) Frisch, M. J .; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.;
Wong, M. W.; Foresman, J . B.; J ohnson, B. G.; Schegel, H. B.; Robb,
M. A.; Replogel, E. S.; Gomperts, R.; Andres, J . L.; Raghavachari, K.;
Binkley, J . S.; Gonzalez, C.; Martin, R. L.; Fox, D. J .; Defrees, D. J .;
Baker, J .; Stewart, J . J . P.; Pople, J . A. Gaussian Inc., Pittsburgh, PA,
1992.
(31) From the Arrhenius equation, plotting ln 1/k versus Ea gives a
line with slope 1/RT ) 1.7 (at 300 K), and since reaction time is
proportional to 1/k, ln(reaction time) is plotted against ∆Ea in Figure
4. Note, this is not a traditional Arrhenius analysis involving variation
of temperature.

5284 J . Org. Chem., Vol. 61, No. 16, 1996
Andrews et al.
Ta ble 1. NBS-In d u ced Hyd r olysis of n -P en ten yl
Glycosid es: Com p u ted a n d Obser ved Rea ctivity Tr en d s
Rela tive to 2âa
Ta ble 2. P r ed icted Rela tive Rea ctivities of n -P en ten yl
Glycosid es: Im p or ta n ce of Solva tion Con tr ibu tion sa
d
a
b
For meaning of column headings, see Scheme 1. Same as
in Table 1. c Experimental times given are (0.2 h as verified by
HPLC. Note: all relative energies in kcal.
5â. Thus, calculated and experimentally observed ratios
in columns 5 and 7 show the same trends.
(iv) The 3-21G activation energies for the dispoke
derivatives 11 and 12 (Table 1) were computed to be
approximately the same, -4.1 and -4.0, respectively,
which implied that both should be hydrolyzed faster than
the reference material 2â. On the other hand, the
corresponding experimentally observed values in column
7 showed (a) that 11 was hydrolyzed faster than 12 and
(b) that both were slower than 2â. However, as shown
in column 5, by allowing for solvation energies, both
observations, namely, the slower reaction of 11 and 12
vis-a-vis 2â and the greater reactivity of 11 vis-a-vis 12,
are qualitatively predicted.
Items i-iv above relate to the results in Table 1, where
the calculated energies were determined after the ex-
perimental data had been obtained. It was therefore im-
portant to see whether the calculations would have a pre-
dictive value. Accordingly, the molecules in Table 2 were
designed to test two predictions: (a) first, that R-anomers
react more slowly than â-anomers; (b) second, that by
breakdown of the data in Table 1, the disarming effect
of a dispoke protecting group can be dissected into elec-
tronic and solvation factors. Thus for 11 and 12, column
3 shows arming electronic effects (-4.1 and -4.0 kcal,
respectively) while column 4 shows disarming solvation
effects (i.e., positive contributions, 5.3 and 8.2 kcal, re-
spectively). The same trends are observed for 5â and 13,
in that the dispoke residue in the latter lowers the elec-
tronic energy difference from 11.5 to 8.1 kcal but raises
the solvation energy difference from -2.4 to 3.6 kcal.
(v) It is significant to note that the activation energy
values [rel Ea(g) and rel Ea(s)] in Tables 1 and 2 predict
correctly that 2â, 4â, and 5â are hydrolyzed faster than
2r, 4r, and 5r, respectivelysin keeping with well-known
experimental evidence11 and our previous calculations.15
Thus, the predicted activation energies for the R-anomers
are greater than those for the â-anomers.
a
b
For meaning of column headings, see Scheme 1.
Use of
relative activation energy values means that structures with Ea′
lower than that of 2â will have negative rel Ea(s) values. c Ex-
perimental times given are (0.2 h as verified by HPLC. Note:
all relative energies in kcal. In column 7, reaction times are
shown relative to that for 2â.
d
Figures 4a and 4b, respectively. It is clear that the latter
is a vast improvement over the former, and that a
distinctive trend has now emerged, the desired propor-
tionality between predicted activation energies and ob-
served reaction times being evident. The major differ-
ence between the solvated and unsolvated data in Figure
4 is that the activation energies for species 11, 12, and
13 (the dispoke derivatives) are not predicted well when
solvation is ignored. Indeed, Figure 5 shows a linear
regression line with an excellent R2 statistic (R2 ) 0.98)
indicating that the predicted activation energies explain
well the experimental reaction times.32
Support of this conclusion is exemplified in the obser-
vations below.
Obser va tion s
(i) The torsional “disarm” effect of a 4,6-O-benzylidene
group is apparent by comparing the Ea values for 2â and
4â in Table 1, which are 0 and 6.1 and 0 and 3.8 before
and after solvation adjustment. The cyclic protecting
group in 4â raises the activation barrier by opposing the
flattening that is required in the oxocarbenium ion.
(ii) The same trends, as in (i), are seen with 11 and
13. Thus the rel Ea(s) sum values are 1.2 and 11.7, in
agreement with the trend in the hydrolysis time ratio
1.8:13.1. The relative reaction time should decrease as
the relative activation energy decreases.
(iii) The comparison between 5â and 13 focuses on the
effect of the dispoke protecting group. Thus, the internal
activation energy in column 3 predicts that 13 should be
hydrolyzed faster than 5â, whereas the experimental
observation in column 7 indicates the opposite. It is
gratifying to see that the solvation adjusted values in
column 5 predict that 13 should indeed react slower than
(vi) Comparison of 11 and 12 in Table 1 with 6 and 8,
respectively, in Table 2 bears out prediction (b) above that
the dispoke protecting group is disarming. Thus, removal
of dispoke from 11 gives 6, which is more disarmed (i.e.,
internal energy barrier is raised from -4.1 to -0.7 kcal)

Hydrolysis of n-Pentenyl Glycosides
J . Org. Chem., Vol. 61, No. 16, 1996 5285
F igu r e 2. Impact of solvation on the activation energy of glycoside hydrolysis for substrate 2â. Solvation in water stabilizes the
oxocarbenium ion much more than the glycoside due to the fact that the oxocarbenium ion is charged (via protonation). In the
context of the thermodynamic cycle below, we use the approximation that the internal energy change associated with activation,
Ea(internal), is equal to the gas phase activation energy, E(g). Furthermore, the activation energy in solution E(s) is equal to
E(internal) modified by solvation. Protecting groups influence the activation energy both from an internal energy standpoint,
since some stabilize the oxocarbenium ion better than others, and from a solvation standpoint, since some are solvated better
than others.
F igu r e 3. Thermodynamic cycle applied to activation energies. The energy cycle has a net change of zero. The terms can be
rearranged to correspond to the columns in Table 2. Species I is the glycoside and species II is the oxocarbenium ion, based on
the approximation that the energy of the oxocarbenium ion is similar to the energy of the transition state. In the general case,
species I is the reactant and species II is the transition state. The activation energies can be absolute (v) or relative (vi). Column
numbers refer to Table 1.
but has a lower solvation energy difference (5.3 f 0.1
kcal). Similarly, in going from 12 to 8, the structure
becomes more disarmed electronically (-4.0 f -2.5 kcal)
but is better solvated (8.2 f 0.3 kcal). In both cases, the
dispoke group is disarming because of its poor solvation.
(vii) It is possible to further pinpoint the origin of some
of these effects by breakdown of the solvation energy data
in Table 3. Thus, these solvation energy values indicate
that in the case of 11, 12, and 13 ionic intermediate II is
relatively less stabilized vis-a-vis the corresponding start-
ing glycoside I, than is the case with the non-dispoke
substrates. The result in item vii confirms the finding
in item vi, that it is difficult to solvate a glycosyl cation
containing a dispoke protecting group.
differ only in the fact that the O2-O3 bridge is more
difficult to solvate in 13 than in 5â. The energy values
in column 3 indicate that 13 should be hydrolyzed faster
than 5âswhich is opposite to the experimentally obtained
results in column 7. Inspection of the values in column
4 reveal that the effect of solvation is to increase the
activation energy of 13 but to decrease that of 5â.
Inclusion of these adjustments (see column 5) now
correctly predicts the observed faster hydrolysis of 5â.
From the result in item v, it is gratifying to see that
the solvation model does not disturb the normal trends
in reactivity (â > R) that are based on internal energies.
Evidence in support of the foregoing conclusion is found
in the data for 5â and 13, Table 1. These compounds
(32) The treatment shown in footnote 31 predicts a slope of 1.7.
However, the slope in Figure 5 is 0.21, owing to the fact that the
regression analysis compensates for the exaggerated 3-21G activation
energies.24 Hehre et al. have shown the need for electron correlation
in reactivity modeling25 which has not been addressed in this study.
Finally, the disarming effect of a 4,6-O-benzylidene

5286 J . Org. Chem., Vol. 61, No. 16, 1996
Andrews et al.
F igu r e 4. Plot of ln(rel experimental time) vs computed Ea
(a) without solvation and (b) with solvation. (a) ) Table 1 (ln-
[column 7] vs 3) or Table 2 (ln[column 6] vs 2). (b) ) Table 1
(ln[column 7] vs 5) or Table 2 (ln[column 6] vs 4).
F igu r e 5. Arrhenius analysis of glycoside hydrolysis data.
to be done. Further testing of the solvation model is
therefore underway and will be reported in due course.
ring was discussed in items i and ii above. By contrast,
could bridging cyclic protecting group have an arming
effect if poor solvation could be avoided? Our 6-31G*
calculations15 had shown that â-glycosides hydrolyze
through a 4E sofa transition state, III. If so, a trans-
fused 6-membered ring should impede reactivity at C2/
C3 much more than at C3/C4sa fact which is obvious
from simple chemical models. Indeed the data for
regioisomers 6 and 8 (Table 2) indicate that in both cases
the rings reduce the internal activation energies (rel Ea
) -0.7 and -2.5, respectively) relative to 2â. However,
the C3/ C4 ring is so effective that it may be considered
to have an arming effect. Unfortunately, the experimen-
tal values in column 6 for 6 and 8 do not reflect this
dramatic difference, suggesting that there is more work
Exp er im en ta l Section
Gen er a l. Please see ref 17 for general procedures.
Gen er a l P r oced u r e for Hyd r olysis Stu d ies. Rate stud-
ies on the n-pentenyl glycosides were carried out by accurately
weighing 25-50 mg of the glycosides into separate flasks
wrapped in aluminum foil. An accurately weighed amount of
NBS was added to a standard solution of 1% H2O/MeCN to
make a solution that contained 3 mmol of NBS in 40 mL of
solution. Portions of this solution (40 mL per mmol glycoside)
were pipetted into the reaction flasks and the mixtures stirred
at room temperature. The reaction times were measured by
TLC (4:1 light petroleum ether: EtOAc) by looking for the
disappearance of the starting materials. Relative times were
then calculated on the basis of absolute reaction times.
In order to confirm that all of the starting material had
reacted fully, an aliquot of each reaction was quenched by
Ta ble 3. In ter n a l a n d Solva tion En er giesa Ca lcu la ted by P M3 a n d 3-21G P r ogr a m s
internal energies
solvation energies
GB/SA (kcal)
PM3 (kcal)
3-21G (kcal)
internal plus solvation
energies (kcal)
Ea(g)
internal
rel Ea(g)
internal
Ea(g)
internal
rel Ea(g)
internal
solvation
energy
structure
energy
energy
∆Gs
rel ∆Gs
rel Ea(a)
2â
-237.6
-2.4
235.2
0
-873.604 54
-759.504 27
-833.640 04
-719.530 12
-832.480 95
-718.362 32
-1329.611 18
-1215.517 42
-1329.613 35
-1215.519 42
-1289.648 39
-1175.535 23
-872.448 09
-758.347 37
-872.448 09
-758.351 79
-873.605 75
-759.504 27
-833.644 33
-719.530 12
-832.487 48
-718.362 32
114.100 27
114.109 92
114.116 83
114.093 76
114.093 93
114.113 16
114.099 18
114.096 31
114.101 48
114.114 21
114.125 16
0
-8.3
-46.8
-9.7
-50.5
-12.6
-53.5
-9.7
-42.9
-13.4
-43.7
-10
-44.9
-11
-49.4
-10.5
-48.7
-8.6
-46.8
-10.2
-50.5
-12.5
-53.5
-38.5
0
0
4â
4â
11
12
13
6
6.1
-40.8
-40.9
-33.2
-30.3
-34.9
-38.4
-38.2
-38.2
-40.3
-41
-2.3
-2.4
5.3
3.8
9.1
-224
23.1
247.1
228.4
232.8
241.9
11.9
-6.8
-2.4
6.7
11.5
-4.1
-4
-322.6
-94.2
-323.6
-90.6
-316.6
-74.7
1.2
8.2
4.2
8.1
3.6
11.7
-0.6
-2.2
1.1
-0.7
-2.5
0.8
0.1
8
0.3
4r
4r
5r
0.3
8.7
-1.8
-2.5
6.9
15.6
13.1
a
The energy values are shown in pairs: the upper one for the glycoside and the lower one for the corresponding oxocarbenium ion.

Hydrolysis of n-Pentenyl Glycosides
J . Org. Chem., Vol. 61, No. 16, 1996 5287
treatment with 10% aqueous Na2S2O3. Extraction of this
aqueous layer with CH2Cl2 was then performed, followed by
extraction of the organic phase with brine. The CH2Cl2 extract
was dried (Na2SO4), filtered, and evaporated in vacuo. The
residue was then analyzed for the presence of starting materi-
als via injection onto a Rainin Dynamax HPLC column.17
P en t-4-en yl 2,3,4,6-Tetr a -O-a cetyl-r-D-glu cop yr a n osid e
(1ra ). Camphorsulfonic acid (300 mg) was added to a mixture
of D-glucose (25.0 g, 139 mmol) and 4-penten-1-ol (80 mL). The
mixture was heated at 100 °C for 3 days under argon. The
reaction was quenched with Et3N, and the bulk of the pentenyl
alcohol was distilled under vacuum using a dry ice condenser.
The residue was purified by flash chromatography (5 f >10%
methanol/CH2Cl2) to yield the pentenyl glucoside as a dark
orange oil (18.58 g, 75 mmol, 54%). This material was
dissolved in pyridine (80 mL) under argon, acetic anhydride
(42 mL, 450 mmol) was added, and the solution was stirred
overnight. The reaction was quenched by addition of metha-
nol, and the solvent was evaporated under high vacuum. The
residue was azeotroped with toluene (3 × 200 mL) and then
flash chromatographed (20 f 30% EtOAc/petroleum ether) to
give 1ra (16.83 g, 40.4 mmol, 54%) and 1âa (4.02 g, 9.65 mmol,
13%) and a mixture (ca. 2:1 â: R) of anomers (8.65 g, 20.8 mmol,
and concentrated. The crude residue was purified by flash
chromatography (10 f 15% EtOAc/petroleum ether) to give
2r as a colorless oil (2.106 g, 3.46 mmol, 85%): [R]21 +33.1°
D
(c 1.34, CHCl3); 1H NMR δ 1.77 (m, 2H), 2.18 (m, 2H), 3.44
(dd, 1H), 3.58-3.85 (m, 6H), 4.02 (m, 1H), 4.53 (m, 2H), 4.67
(m, 2H), 4.70-4.78 (m, 4H), 4.98-5.07 (m, 3H), 5.85 (m, 1H),
7.13-7.48 (m, 20H); 13C NMR δ 138.93, 138.35, 138.26, 138.11,
137.98, 128.49, 128.24, 128.20, 127.99, 127.75, 127.65, 114.94,
97.02, 82.14, 80.15, 77.78, 75.77, 75.14, 73.51, 73.25, 70.15,
68.50, 67.53, 30.33, 28.59.
Anal. Calcd for C39H44O6: C, 76.95; H, 7.28. Found: C,
76.81; H, 7.25.
4-P en ten yl 2,3,4,6-Tetr a -O-ben zyl-â-D-glu cop yr a n osid e
(2â). To a solution of 4-pentenyl 2,3,4,6-tetra-O-acetyl-â-D-
glucopyranoside (1âa ) (6.58 g, 15.8 mmol) in anhydrous
methanol (50 mL) was added a catalytic amount of K2CO3 (ca.
200 mg), and the mixture was stirred for 4.5 h. The reaction
mixture was filtered through a Celite pad and concentrated
to give crude pentenyl glucoside 1âa as a yellow foam. To a
stirred solution of this material in DMF at 0 °C under argon
was added NaH (60% dispersion, 3.16 g, 79 mmol, 5 equiv).
Benzyl bromide (11.3 mL, 6 equiv) was added dropwise at 0
°C, and the mixture was then stirred at room temperature
overnight. The reaction was quenched with methanol, diluted
with diethyl ether (200 mL), and washed consecutively with
cold H2O (1 × 150 mL), saturated aqueous NaHCO3 (2 × 100
mL), and brine (2 × 100 mL). The organic layer was dried
over MgSO4, filtered, concentrated in vacuo, and the residue
was flash chromatographed (95:5 f 90:10 petroleum ether/
EtOAc) to yield 2â as a white solid (10.0630 g, 10.74 mmol,
68% yield from 1âa ). Rf 0.69 (4:1 petroleum ether/EtOAc); [R]D
+5.48° (c 1.06, CHCl3); mp 70-71 °C (ethanol); 1H NMR
(CDCl3) δ 1.72 (m, 2H) 2.11 (m, 2H), 3.38 (m, 1H), 3.45-3.69
(m, 5H), 3.91 (m, 1H), 4.32 (d, J ) 7.76 Hz, 1H), 4.50-4.61
(m, 4H), 4.78 (m, 3H), 4.98 (m, 4H), 5.82 (m, 1H), 7.35 (m,
20H); 13C NMR (CDCl3) δ 138.67, 138.48, 138.21, 138.10,
128.42, 128.29, 128.03, 127.92, 127.81, 127.64, 114.97, 103.67,
89.8, 84.76, 82.30, 77.94, 75.77, 75.08, 74.91, 73.53, 69.42,
30.32, 29.04.
28%). For 1ra : mp 62-63 °C (petroleum ether/ethyl ether);
1
[R]21 +124° (c 1.12, CHCl3); H NMR δ 1.65 (m, 2H), 2.00-
D
2.12 (4 s, 12H), 2.14 (m, 2H), 3.41 (ddd, 1H), 3.68 (ddd, 1H),
4.00 (m, 1H), 4.08 (dd, 1H), 4.23 (dd, 1H), 4.83 (dd, J ) 3.74
Hz, 1H), 4.94-5.15 (m, 4H), 5.46 (t, 1H), 5.79 (m, 1H); 13C NMR
δ 170.57, 170.09, 169.55, 137.65, 115.30, 95.63, 70.84, 70.15,
68.54, 67.78, 67.10, 61.86, 30.03, 28.34, 20.98, 20.65.
Anal. Calcd for C19H28O10: C, 54.80; H, 6.78. Found: C,
54.85; H, 6.83.
4-P en ten yl 2,3,4,6-Tetr a -O-a cetyl-â-D-glu cop yr a n osid e
(1âa ). Following the procedure of Rodriquez and Stick,20 tetra-
O-acetyl-R-D-glucopyranosyl bromide (24.87 g, 60.48 mmol) and
4-penten-1-ol (18.3 mL, 2.5 equiv) were dissolved in dry,
distilled CH2Cl2 (130 mL) under argon. Powdered, freshly
activated 4 Å molecular sieves (30 g) were added, and the
mixture was stirred for 30 min. Silver carbonate (20.34 g,
73.76 mmol, 1.2 equiv) was quickly added, and the reaction
was stirred under darkness for 72 h. The reaction mixture
was then diluted with CH2Cl2 (150 mL), filtered through a wet
Celite pad, and washed consecutively with saturated aqueous
NaHCO3 (2 × 125 mL) and brine (1 × 150 mL). The organic
layer was dried over anhydrous Na2SO4, concentrated, and
flash chromatographed (85:15 f 60:40 petroleum ether/EtOAc)
to give 1âa (20.05 g, 76%) as a white solid: Rf 0.48 (3:2
Anal. Calcd for C39H44O6: C, 76.95, H, 7.28. Found: C,
76.74, H, 7.33.
P en t -4-en yl 4,6-O-Ben zylid en e-r-D-glu cop yr a n osid e
(3r). To a solution of pentenyl R-glucoside 1rb (1.2875 g, 5.2
mmol) in DMF (12 mL) were added PPTS (50 mg) and
benzaldehyde dimethyl acetal (0.898 mL, 5.98 mmol). The
mixture was heated at 80 °C for 3.5 h under a stream of argon
to remove methanol. The reaction was quenched with Et3N
(10 drops), concentrated under high vacuum, and flash chro-
matographed (25 f 30% EtOAc/CH2Cl2) to give 3r as a white
crystalline solid (1.6417 g, 4.87 mmol, 94%): mp 90 °C (ethyl
petroleum ether/EtOAc); [R]D -19.5° (c 1.06, CHCl3) [lit.20
)
-19.4°], mp 47-48 °C (petroleum ether:diethyl ether) [lit.20
mp 45-46 °C]; 1H NMR (CDCl3) δ 1.69-1.61 (m, 2H), 1.99 (s,
3H), 2.02 (s, 3H), 2.05 (s, 3H), 2.09 (s, 3H), 3.46 (m, 1H), 3.67
(m, 1H), 3.85 (m, 1H), 4.13 (dd, J ) 2.44 Hz, 1H), 4.24 (dd, J
) 4.68 Hz, J ) 12.24 Hz, 1H), 4.48 (d, J ) 7.92 Hz, 1H), 4.94
(m, 2H), 5.01 (m, 1H), 5.05 (t, 1H), 5.18 (t, J ) 9.55 Hz, 1H),
5.78 (m, 1H); 13C NMR (CDCl3) δ 170.70, 170.32, 169.41,
169.29, 137.79, 115.10, 100.81, 72.85, 71.66, 71.35, 69.33,
68.39, 61.96, 29.81, 28.54, 20.77, 20.69, 20.64, 20.58.
acetate/petroleum ether); [R]21 +96.5° (c 1.13, CHCl3); 1H
D
NMR δ 1.73 (m, 2H), 2.14 (m, 2H), 2.27-2.88 (broad s, 2H),
3.48 (m, 2H), 3.57 (dd, 1H), 3.65-3.88 (m, 4H), 3.89 (t, 1H),
4.25 (dd, 1H), 4.85 (d, J ) 3.93 Hz, 1H), 4.97-5.09 (m, 2H),
5.51 (s, 1H), 5.79 (m, 1H), 7.31-7.50 (m, 5H); 13C NMR δ
137.82, 137.04, 129.26, 128.34, 126.33, 115.27, 101.90, 98.76,
80.92, 72.91, 71.83, 68.93, 67.96, 62.62, 30.35, 28.58.
Anal. Calcd for C18H24O6: C, 64.27; H, 7.19. Found: C,
64.10; H, 7.21.
Anal. Calcd for C19H28O10: C, 54.80, H, 6.78; Found: C,
54.80, H, 6.80.
P en t-4-en yl 4,6-O-Ben zyliden e-â-D-glu copyr an oside (3â).
Pent-4-enyl â-D-glucopyranoside (1âb) (7.2000 g, 28.92 mmol)
was treated as described above for 3r. The crude residue was
flash chromatographed (3:2 EtOAc/petroleum ether) to give 3â
P en t -4-en yl 2,3,4,6-Tet r a -O-b en zyl-r-D-glu cop yr a n o-
sid e (2r). To a solution of pent-4-enyl 2,3,4,6-tetra-O-acetyl-
R-D-glucopyranoside (1ra ) (16.83 g, 40.4 mmol) in anhydrous
methanol (90 mL) was added K2CO3 (2.02 g), and the mixture
was stirred overnight at room temperature. The solution was
neutralized with Amberlite IR-120 (H+), filtered through Celite
and concentrated. The residue, 1rb, was flash chromato-
graphed (7 f 13% MeOH/CH2Cl2) to give the tetrol as a clear
colorless oil (9.66 g, 97%). A portion of this material (1.01 g,
4.07 mmol) was then dissolved in DMF (20 mL) under argon
cooled to 0 °C, and the solution was treated with sodium
hydride (586 mg, 24.4 mmol) and benzyl bromide (2.9 mL, 24
mmol) and allowed to warm to room temperature while stirring
overnight. The reaction was quenched with methanol, diluted
with ether (75 mL), and washed with water (1 × 40 mL). The
organic phase was washed with saturated aqueous NaHCO3
(1 × 50 mL) and brine (1 × 45 mL), dried (Na2SO4), filtered,
(4.766 g, 49%) as a white solid: [R]21 -43.8° (c 1.16, CHCl3);
D
mp 144-145 °C (ethanol); 1H NMR (CDCl3) δ 1.71 (m, 2H),
2.11 (m, 2H), 3.02 (s, 1H), 3.20 (s, 1H), 3.38-3.59 (m, 4H), 3.74
(m, 2H), 3.88 (ddd, 1H), 4.25-4.36 (m, J ) 7.62 Hz, J ) 4.82
Hz, 2H), 4.92-5.11 (m, 2H), 5.49 (s, 1H), 5.81 (m, 1H), 7.31-
7.50 (m, 5H); 13C NMR δ 137.98, 137.00, 129.29, 128.36, 126.32,
115.09, 103.19, 101.90, 80.57, 74.53, 73.10, 69.88, 68.68, 66.35,
30.12, 28.71.
Anal. Calcd for C18H24O6: C, 64.27,; H, 7.19. Found: C,
64.10; H, 7.23.
P en t-4-en yl 2,3-Di-O-ben zyl-4,6-O-ben zylid en e-r-D-glu -
cop yr a n osid e (4r). To a solution of pent-4-enyl 4,6-O-
benzylidene-R-D-glucopyranoside (3r) (1.16 g, 3.43 mmol) in

5288 J . Org. Chem., Vol. 61, No. 16, 1996
Andrews et al.
DMF (14 mL) under argon at 0 °C was added sodium hydride
(394.5 mg of a 60% dispersion in mineral oil, 10.3 mmol), and
after 15 additional minutes benzyl bromide (1.22 mL, 10.3
mmol) was added. The solution was allowed to warm to room
temperature and gradually stirred overnight. The reaction
was then quenched with glacial acetic acid and taken up in
diethyl ether (125 mL). The ether layer was successively
washed with H2O (2 × 75 mL), saturated aqueous NaHCO3 (2
× 75 mL), and brine (1 × 75 mL). The organic phase was dried
(Na2SO4), filtered, and concentrated. The residue was flash
chromatographed (7 f 15% EtOAc/petroleum ether) to give
P en t-4-en yl 6-O-Ben zyl-2,3-O-eth ylen e-â-D-glu cop yr a -
n osid e (6a ). Pent-4-enyl 4,6-O-benzylidene-2,3-O-ethylene-
â-D-glucopyranoside (5â) (126.7 mg, 0.346 mmol) and NaCN-
BH3 (195.7 mg, 3.11 mmol; 9 equiv) were stirred with
anhydrous THF (7 mL) over freshly activated, powdered 3Å
molecular sieves under argon. A saturated Et2O/HCl solution
(60 mL) was added until the solution stopped bubbling.
Saturated aqueous NaHCO3 (10 mL) was added, and the
mixture was taken up in CH2Cl2 (125 mL) and subsequently
washed with brine (2 × 50 mL). The water layer was
re-extracted with CH2Cl2 (2 × 25 mL) and CHCl3 (1 × 30 mL).
The organic layers were combined, dried over Na2SO4, filtered,
and evaporated to dryness. The residue was flash chromato-
graphed (50% petroleum ether/EtOAc f 40% petroleum ether/
4r as a white crystalline solid (1.6058 g, 3.09 mmol, 90%): mp
1
82-83 °C (ethanol); [R]21 +0.74° (c 5.52, CHCl3); H NMR δ
D
1.76 (m, 2H), 2.17 (m, 2H), 3.46 (ddd, 1H), 3.51-3.76 (m, 4H),
3.88 (ddd, H-5), 4.05 (t, 1H), 4.27 (dd, 1H), 4.72 (m, 2H, J )
3.77 Hz, 1H), 4.79-5.10 (m, 5H), 5.58 (s, 1H), 5.81 (m, 1H),
7.21-7.53 (m, 15H); 13C NMR δ 138.81, 138.27, 137.92, 137.36,
128.88, 128.39, 128.26, 128.19, 127.93, 127.80, 127.51, 125.97,
115.00, 101. 18, 98.02, 82.21, 79.38, 78.59, 75.28, 73.50, 69.04,
67.66, 62.38, 30.21, 28.49.
EtOAc) to give 6a (97.3 mg, 77%) as a colorless oil: [R]21
D
-55.5° (c 1.11, CHCl3); 1H NMR δ 1.79 (m, 2H), 2.13 (m, 2H),
2.92 (s, 1H), 3.32 (t, 1H), 3.50-3.72 (m, 4H), 3.79-3.92 (m,
6H), 4.41 (d, J ) 7.57 Hz, 1H), 4.62 (d, 2H), 5.01 (m, 2H), 5.81
(m, 1H), 7.32-7.50 (m, 5H); 13C NMR δ 138.03, 137.75, 128.50,
127.86, 127.75, 114.91, 100.49, 80.79, 78.42,76.86, 74.74, 73.69,
70.19, 69.58, 69.28, 66.96, 66.77, 30.04, 28.73.
Anal. Calcd for C32H36O6: C, 74.39; H, 7.02. Found: C,
74.21; H, 7.03.
Anal. Calcd for C20H28O6: C, 65.92; H, 7.74. Found: C,
65.95; H, 7.55.
P en t-4-en yl 2,3-Di-O-ben zyl-4,6-O-ben zylid en e-â-D-glu -
cop yr a n osid e (4â). Pent-4-enyl 4,6-O-benzylidene-â-D-glu-
copyranoside (3â) (2.20 g, 6.54 mmol) was benzylated as
described above for 4r. The residue was flash chromato-
graphed (95:5 petroleum ether/EtOAc) to yield 4â as a clear
oil which crystallized overnight under vacuum (2.7368 g, 5.30
mmol, 81%): [R]21D -36.4° (c 1.05, CHCl3); mp 76 °C (ethanol);
1H NMR δ 1.75 (m, 2H), 2.16 (m, 2H), 3.37-3.48 (m, 2H), 3.58
(ddd, 1H), 3.64-3.81 (m, 3H), 3.95 (ddd, 1H), 4.34 (dd, J )
4.91 Hz, J ) 10.45 Hz, 1H), 4.49 (d, J ) 7.76 Hz, 1H), 4.78
(m, 2H), 4.88-5.04 (m, 4H), 5.56 (s, 1H), 5.80 (m, 1H), 7.21-
7.53 (m, 15H); 13C NMR δ 138.54, 138.36, 137.94, 137.36,
128.98, 128.37, 128.33, 128.15, 128.10, 127.77, 127.68, 126.05,
115.11, 104.15, 101.13, 82.17, 81.54, 80.93, 75.44, 75.19, 69.86,
68.85, 66.04, 30.22, 28.99.
P en t -4-en yl 4,6-Di-O-b en zyl-2,3-O-et h ylen e-â-D-glu -
cop yr a n osid e (6b). To a solution of pent-4-enyl 6-O-benzyl-
2,3-O-ethylene-â-D-glucopyranoside (6a ) (90.0 mg, 0.245 mmol)
in DMF (4 mL) at 0 °C was added NaH (19.5 mg of a 60%
dispersion in mineral oil, 0.49 mmol). After 10 min, benzyl
bromide (0.010 mL, 0.49 mmol, 2 equiv) was added to the
mixture, and the solution was allowed to slowly come to room
temperature overnight. The reaction was quenched with
methanol, taken up in EtOAc, washed with saturated aqueous
NaHCO3 (1 × 10 mL) and brine (2 × 15 mL), dried (Na2SO4),
filtered, and evaporated to dryness. The residue was flash
chromatographed to give 6b (75.0 mg, 67%) as a clear, light
1
yellow oil: [R]21 -7.73° (c 2.20, CHCl3); H NMR δ 1.79 (m,
D
2H), 2.12 (m, 2H), 3.32 (m, 1H), 3.50-3.72 (m, 4 H), 3.79-
3.92 (m, 7H), 4.41 (d, J ) 7.76 Hz, 1H), 4.51-4.62 (m, 3H),
4.87 (d, 1H), 5.01 (m, 2H), 5.81 (m, 1H), 7.22-7.50 (m, 10H);
13C NMR δ 138.21, 138.11, 128.37, 127.96, 127.77, 127.73,
127.63, 114.88, 100.37, 82.19, 77.27, 75.42, 75.05, 74.57, 73.47,
72.13, 69.23, 68.98, 66.94, 66.80, 30.08, 28.77.
Anal. Calcd for C32H36O6: C, 74.39; H, 7.02. Found: C,
74.35; H, 7.02.
P en t-4-en yl 4,6-O-Ben zylid en e-2,3-O-eth ylen e-r-D-glu -
cop yr a n osid e (5r). To pent-4-enyl 4,6-O-benzylidene-R-D-
glucopyranoside (3r) (469 mg, 1.39 mmol) were added tet-
rabutylammonium bromide (89.6 mg, 0.278 mmol) and
dichloroethane (6 mL) according to the published method.21
An aqueous solution of NaOH (35%, 7 mL) was added, and
the mixture was stirred rapidly at 50 °C for 24 h. At this time,
fresh NaOH (35% aq, 4 mL) was added, and the mixture was
stirred for an additional 24 h. The mixture was cooled and
partitioned between H2O:Et2O (1:1; 125 mL), and the H2O
layer was re-extracted with Et2O (2 × 50 mL) and EtOAc (1 ×
50 mL). The organic layers were combined, extracted with
brine (1 × 100 mL), dried (Mg2SO4), and concentrated to a
yellow oil. The oil was flash chromatographed (3:1 petroleum
Anal. Calcd for C27H34O6: C, 71.34; H, 7.45. Found: C,
71.40; H, 7.45.
P en t-4-en yl 2,6-Di-O-ben zyl-â-D-glu cop yr a n osid e (7).
According to the method of Garegg et al.,21 pent-4-enyl 4,6-O-
benzylidene-â-D-glucopyranoside (3â) (1.00 g, 2.97 mmol) was
combined with n-butyltetraammonium hydrogen sulfate (0.200
g, 0.59 mmol, 0.2 equiv) and benzyl bromide (0.57 mL, 4.74
mmol, 1.6 equiv) in dichloromethane (55 mL). A 5% aqueous
solution of NaOH (5 mL) was added, and the mixture was
refluxed for 48 h. The solution was diluted with CH2Cl2 (25
mL), washed with brine (2 × 20 mL), dried (Na2SO4), filtered,
and evaporated. The residue was flash chromatographed (85:
15 light petroleum ether/EtOAc) to give a light yellow oil which
consisted of an inseparable 3:1 mixture of the 2-O-benzyl and
3-O-benzyl regioisomers (1.0176, 80%). The mixture was
directly treated with NaCNBH3 (1.3438 g, 21.38 mmol, 9 equiv)
in anhydrous THF (35 mL) under argon with 4 Å molecular
sieves. A saturated HCl/Et2O solution (250 mL) was added
until the evolution of gas ceased. The reaction was quenched
by slowly adding saturated aqueous NaHCO3 (20 mL), and the
mixture was diluted with CH2Cl2 (400 mL) and filtered
through Celite. The layers were separated, and the aqueous
layer was re-extracted with CHCl3 (3 × 30 mL). The organic
layers were combined, dried (Na2SO4), filtered, and evaporated.
The residue was flash chromatographed (75:25 petroleum
ether/EtOAc f 40:60 petroleum ether/EtOAc) to give 7 (605.1
mg, 59%) as a clear oil: [R]21D -5.98° (c 1.00, CHCl3); 1H NMR
δ 1.79 (m, 2H), 2.18 (m, 2H), 2.99 (s, 1H), 3.19 (s, 1H), 3.22 (t,
1H), 3.50 (m, 4H), 3.70 (m, 2H), 3.93 (m, 1H) 4.40 (d, J ) 7.76
Hz, 1H), 4.65 (m, 3H), 4.87 (d, 1H), 5.01 (m, 3H), 5.81 (m, 1H),
7.22-7.50 (m, 10H); 13C NMR δ 138.31, 138.00, 137.92, 128.55,
128.47, 128.16, 127.94, 127.79, 127.74, 115.03, 103.31, 80.79,
76.06, 74.43, 74.15, 73.64, 71.40, 70.16, 69.37, 30.26, 28.98.
ether/EtOAc) to yield 5r (463.1 mg, 1.27 mmol, 92%) as a clear,
1
slightly yellow oil: [R]21 +63.2° (c 1.37, CHCl3); H NMR δ
D
1.76 (m, 2H), 2.16 (m, 2H), 3.50 (m, 3H), 3.60-4.01 (m, 10H),
4.32 (dd, 1H), 4.88 d, J ) 3.42 Hz, 1H), 4.98 (m, 2H), 5.52 (s,
1H), 5.81 (m, 1H), 7.32-7.50 (m, 5H); 13C NMR δ 137.89,
136.97, 129.21, 128.31, 126.46, 115.17, 102.10, 97.49, 79.33,-
77.37, 73.90, 69.03, 67.78, 67.53, 66.45,63.22, 30.22, 28.53.
Anal. Calcd for C20H26O6: C, 66.28; H, 7.23. Found: C,
66.19; H, 7.25.
P en t-4-en yl 4,6-O-Ben zylid en e-2,3-O-eth ylen e-â-D-glu -
cop yr a n osid e (5â). Pent-4-enyl 4,6-O-benzylidene-â-D-glu-
copyranoside (3â) (418.9 mg, 1.24 mmol) was subjected to the
phase transfer reactions described for 5r. The crude product
was flash chromatographed (3:1 petroleum ether/EtOAc) to
yield 5â (275 mg, 0.756 mmol, 61%) as a white solid: [R]21
D
-62.8° (c 1.01, CHCl3); mp 77-78 °C (ethanol); 1H NMR δ 1.76
(m, 2H), 2.16 (m, 2H), 3.32 (t, 1H), 3.50-3.72 (m, 4H), 3.79-
3.92 (m, 6H), 4.32 (dd, 1H), 4.53 (d, J ) 7.82 Hz, 1H), 5.01 (m,
2H), 5.52 (s, 1H), 5.81 (m, 1H), 7.32-7.50 (m, 5H); 13C NMR δ
138.12, 136.87, 129.24, 128.33, 126.43, 115.03, 101.99, 100.95,
78.42,78.28, 69.61, 68.72, 67.16, 67.00, 66.70, 29.97, 28.70.
Anal. Calcd for C20H26O6: C, 66.28; H, 7.23. Found: C,
66.00; H, 7.28.

Hydrolysis of n-Pentenyl Glycosides
J . Org. Chem., Vol. 61, No. 16, 1996 5289
Anal. Calcd for C25H32O6: C, 70.07; H, 7.53. Found: C,
70.00; H, 7.56.
methanol, taken up in diethyl ether (60 mL), and washed
consecutively with water (2 × 20 mL), saturated aqueous
NaHCO3 (1 × 25 mL), and brine (1 × 25 mL). The organic
layer was dried (Na2SO4), filtered, and concentrated. The
residue was flash chromatographed to give 11 (475.3 mg, 78%)
P en t -4-en yl 2,6-Di-O-b en zyl-3,4-O-et h ylen e-â-D-glu -
cop yr a n osid e (8). Pent-4-enyl 2,6-di-O-benzyl-â-D-glucopy-
ranoside (7) (366.7 g, 0.858 mmol) was dissolved in dichloro-
ethane (6 mL). Tetrabutylammonium hydrogen sulfate (58.2
mg, 0.171 mmol, 0.2 equiv) and a 35% aqueous NaOH solution
(7 mL) were added, and the biphasic solution was stirred
rapidly at 55 °C for 24 h. At this time, additional dichloro-
ethane (3 mL) and NaOH solution (4 mL) were added, and
the solution was stirred for another 24 h. The mixture was
diluted with ethyl acetate (150 mL), washed with saturated
aqueous NaHCO3 (2 × 50 mL) and brine (1 × 75 mL), dried
(Na2SO4), filtered, and evaporated. The residue was flash
chromatographed (85:15 petroleum ether/EtOAc f 75:25
petroleum ether/EtOAc) to give 8 (246.7 mg, 63%) as a colorless
as a white solid: mp 108-109 °C (ethanol); [R]21 -34.1° (c
D
1.31, CHCl3); 1H NMR δ 1.32-1.87 (m, 14H), 2.19 (m, 2H),
3.53 (m, 2H), 3.61 -3.82 (m, 8H), 3.95 (m, 2H) 4.49-4.64 (m,
4H), 4.95 (m, 3H), 5.83 (m, 1H), 7.18-7.39 (m, 10H); 13C NMR
δ 138.33, 138.27, 128.412, 128.34, 128.05, 127.77, 127.54,
114.76, 100.79, 96.88, 96.78, 75.53, 74.83, 73.46, 73.27, 69.27,
69.18, 68.81, 60.74, 30.15, 28.95, 28.55, 28.45, 24.94, 24.85,
18.42, 18.02.
Anal. Calcd for C35H46O8: C, 70.68; H, 7.80. Found: C,
70.45; H, 7.82.
P en t-4-en yl 2,6-Di-O-ben zyl-3,4-O-(octa h yd r o-2,2′-bi-
2H-p yr a n -2,2′-d iyl)-â-D-glu cop yr a n osid e (12). Pent-4-enyl
3,4-O-(octahydro-2,2′-bi-2H-pyran-2,2′-diyl)-â-D-glucopyrano-
side (10) (235.0 mg, 0.567 mmol) was dissolved in dry DMF (6
mL) under argon at 0 °C, and NaH (68.0 mg of a 60%
dispersion in mineral oil, 1.70 mmol) was added. After 15 min,
benzyl bromide (0.202 mL) was added dropwise, and the
solution was allowed to slowly come to room temperature
overnight. The reaction was quenched with methanol and
concentrated under high vacuum. The crude yellow solid was
flash chromatographed (97.5:2.5 petroleum ether/EtOAc f 60:
40 petroleum ether/EtOAc) to give 12 (288 mg, 85%) as a white
solid: mp 104-105 °C (ethanol); [R]21D +23.1° (c 1.29, CHCl3);
1H NMR δ 1.47-2.00 (m, 14H), 2.19 (m, 2H), 3.45-3.70 (m,
6H), 3.71-3.91 (m, 4H), 3.96 (m, 2H), 4.42 (d, J ) 7.45 Hz,
1H), 4.62 (m, 2H), 4.82 (m, 2H), 4.99 (m, 2H), 5.85 (m, 1H),
7.21-7.51 (m, 10H); 13C NMR δ 138.18, 138.09, 128.39, 128.17,
127.70, 127.62, 127.45, 114.82, 103.72, 96.89, 96.73, 78.93,
74.66, 73.85, 73.55, 71.84, 69.44, 68.61, 65.40, 60.79, 60.62,
30.17, 28.90, 28.44, 24.79, 18.15, 17.94, 17.85.
oil: [R]21 -1.8° (c 1.25, CHCl3); 1H NMR δ 1.75 (m, 2H), 2.19
D
(m, 2H), 3.22-3.45 (m, 3H), 3.52-3.65 (m, 3H), 3.71-3.89 (m,
5H), 3.96 (m, 1H), 4.45 (d, J ) 7.05 Hz, 1H), 4.61 (s, 2H), 4.78
(dd, 2H), 5.01 (m, 2H), 5.81 (m, 1H), 7.22-7.50 (m, 10H); 13C
NMR δ 138.54, 138.09, 128.35, 128.28, 127.88, 127.63, 127.57,
114.94, 103.59, 80.57, 78.84, 74.70, 74.41, 73.76, 73.53, 69.47,
68.82, 67.07, 66.79, 30.23, 28.97.
Anal. Calcd for C27H34O6: C, 71.34; H, 7.54. Found: C,
71.23; H, 7.55.
P en t-4-en yl 2,3-O-(Octah ydr o-2,2′-bi-2H-pyr an -2,2′-diyl)-
â-D-glu cop yr a n osid e (9) a n d P en ten yl 3,4-O-(octa h yd r o-
2,2′-bi-2H-p yr a n -2,2′-d iyl)-â-D-glu cop yr a n osid e (10). Ac-
cording to the method of Ley et al.,22 a catalytic amount of
camphorsulfonic acid (CSA; ca. 75 mg) was added to a stirred
solution of pent-4-enyl â-D-glucopyranoside (1âb) (3.8010 g,
15.30 mmol) and bidihydropyran (4.8121 g, 28.96 mmol, 2.1
equiv) in dry, distilled CHCl3. The mixture was refluxed for
7 h, and then the reaction was quenched by addition of
ethyleneglycol (1.5 mL). The mixture was refluxed for an
additional 0.5 h, and then it was diluted with CH2Cl2 (250 mL)
and extracted with saturated aqueous NaHCO3 (1 × 100 mL).
The aqueous layer was re-extracted with CH2Cl2 (2 × 75 mL)
and CHCl3 (1 × 75 mL). The organic layers were combined,
dried (Na2SO4), filtered, and concentrated. The residue was
flash chromatographed (70:30 petroleum ether/EtOAc f 30:
70 petroleum ether/EtOAc) to give 9 (1.9830 g, 4.78 mmol,
34%) and 10 (869.3 mg, 2.10 mmol, 15%) as light yellow foams.
Anal. Calcd for C21H34O8: C, 70.68; H, 7.80. Found: C,
70.58; H, 7.76.
P en t-4-en yl 4,6-O-Ben zylid en e-2,3-O-(octa h yd r o-2,2′-
bi-2H-p yr a n -2,2′-d iyl)-â-D-glu cop yr a n osid e (13). Accord-
ing to the method of Madsen and Fraser-Reid,23 a solution of
pent-4-enyl 2,3-O-(octahydro-2,2′-bi-2H-pyran-2,2′-diyl)-â-D-
glucopyranoside (9) (515.7 mg, 1.243 mmol, previously co-
evaporated with toluene) and dipent-4-enyl benzaldehyde
acetal (0.414 mL, 1.49 mmol) in acetonitrile (5 mL) was stirred
with NBS (441 mg, 2.48 mmol, 2 equiv) and camphorsulfonic
acid (28 mg) for 10 min at room temperature under argon. The
reaction was quenched with Et3N (20 mL), diluted with CH2-
Cl2 (20 mL), and washed with 10% aqueous sodium thiosulfate
(1 × 10 mL) and saturated aqueous NaHCO3 (1 × 10 mL).
The organic layer was dried (Na2SO4), filtered, and evaporated.
The residue was flash chromatographed (96:4 petroleum ether/
EtOAc f 80:20 petroleum ether/EtOAc) to give 13 (449.8 mg,
For 9: [R]21 -83.3° (c 1.04, CHCl3); Rf 0.2 (1:1 petroleum
D
ether/EtOAc); 1H NMR δ 1.32-1.87 (m, 16H), 2.08 (m, 2H),
3.41-3.82 (m, 11H), 3.92 (dd, 1H), 4.32 (d, J ) 7.47 Hz, 1H),
4.97 (m, 2H), 5.83 (m, 1H); 13C NMR δ 138.13, 114.87, 101.05,
96.80, 96.75, 76.07, 71.80, 69.46, 68.61, 67.79, 62.24, 60.72,
30.02, 28.88, 28.42, 28.36, 24.85, 24.78, 18.02.
Anal. Calcd for C21H34O8: C, 60.85; H, 8.27. Found: C,
60.72; H, 8.26.
For 10: [R]21 +37.3° (c 1.15, CHCl3); Rf 0.5 (1:1 light
D
petroleum ether/EtOAc); 1H NMR δ 1.32-2.03 (m, 16H), 2.19
(m, 2H), 3.41-3.82 (m, 11H), 3.86 (dd, 1H) 4.54 (d, J ) 7.90
Hz, 1H), 5.01 (m, 2H), 5.81 (m, 1H); 13C NMR δ 138.02, 115.00,
103.56, 96.91, 96.80, 74.11, 71.35, 70.96, 69.74, 65.05, 61.50,
60.86, 60.74, 30.09, 28.68, 28.43, 28.38, 24.79, 24.74,18.02,
17.90.
72%) as a white foam: [R]21 -68.7° (c 1.00, CHCl3); 1H NMR
D
δ 1.39-1.99 (m, 14H), 2.15 (m, 2H), 3.45-4.05 (m, 11H), 4.27
(dd, 1H), (d, J ) 8.05 Hz, 1H), 4.95 (m, 2H), 5.51 (s, 1H), 5.80
(m, 1H), 7.21-7.51 (m, 5H); 13C NMR δ 138.08, 137.45, 128.85,
128.40, 128.18, 126.32, 126.09, 114.9, 101.52, 100.99, 97.08,
96.74, 77.94, 69.68, 69.46, 68.74, 67.42, 61.13, 60.90, 60.64,
30.02, 28.91, 28.45, 25.17, 24.89, 18.05.
Anal. Calcd for C21H34O8: C, 60.85; H, 8.27. Found: C,
60.85; H, 8.34.
Anal. Calcd for C28H38O8: C, 66.91; H, 7.62. Found: C,
66.65; H, 7.67.
P en t-4-en yl 4,6-Di-O-ben zyl-2,3-O-(octa h yd r o-2,2′-bi-
2H-p yr a n -2,2′-d iyl)-â-D-glu cop yr a n osid e (11). Pent-4-enyl
2,3-O-(octahydro-2,2′-bi-2H pyran-2,2′-diyl)-â-D-glucopyrano-
side (9) (428.3 mg, 1.03 mmol) was dissolved in dry DMF at 0
°C under argon, and NaH (120 mg of a 60% dispersion in
mineral oil, 2.99 mmol) was added. After 15 min, benzyl
bromide (0.350 mL, 2.88 mmol) was added dropwise to the
mixture, and the reaction was allowed to come slowly to room
temperature overnight. The reaction was quenched with
Ack n ow led gm en t. We are grateful to the National
Science Foundation (CHE 9311356) and Glaxo Research
Institute for financial support of this work. We thank
Professor A. T. McPhail for the X-ray structure shown
in Figure 1.
J O9601223