Compared behavior of 5-deoxy-5-iodo-D-xylo- and L-arabinofuranosides in the reductive elimination reaction: A strong dependence on structural parameters and on the presence of Zn2+. A combined experimental and theoretical investigation

Article abstract of DOI:10.1021/jo0614590

In the framework of a project devoted to the chemical transformation of monosaccharides from hemicelluloses into higher added value materials, the zinc-induced reductive elimination from 5-deoxy-5-iodo derivatives of D-xylose and L-arabinose was carried o

Full text of DOI:10.1021/jo0614590

Compared Behavior of 5-Deoxy-5-iodo-D-xylo- and
L-Arabinofuranosides in the Reductive Elimination Reaction: A
Strong Dependence on Structural Parameters and on the Presence
2
+
of Zn . A Combined Experimental and Theoretical Investigation
,†
‡
‡
‡
Eric Henon,* Ariane Bercier, Richard Plantier-Royon, Dominique Harakat, and
,
‡
Charles Portella*
Groupe de Spectrom e´ trie Mol e´ culaire et Atmosph e´ rique (GSMA), UMR CNRS 6089 and R e´ actions
S e´ lectiVes et Applications (RSA), UMR CNRS 6519, U.F.R Sciences, UniVersit e´ de Reims
Champagne-Ardenne, Moulin de la housse, B.P. 1039, 51687 Reims Cedex 2, France
eric.henon@uniV-reims.fr; charles.portella@uniV-reims.fr
ReceiVed July 13, 2006
In the framework of a project devoted to the chemical transformation of monosaccharides from
hemicelluloses into higher added value materials, the zinc-induced reductive elimination from 5-deoxy-
5
-iodo derivatives of D-xylose and L-arabinose was carried out. This study gave us the opportunity to
observe surprising behaviors. In particular, the reaction strongly depends on structural parameters
protecting group pattern, configuration at C-4) and on the presence of Zn2+ ions. Collaterally with the
experimental work, water solvent PCM HF-DFT (MPW1K/LANL2DZ) computations were performed
(
2
+
to obtain insight into the mechanism for the reductive part of the reaction sequence. Without Zn , the
zinc insertion reaction was found to proceed through a concerted but non-synchronous process involving
-1
a relatively large energy barrier (32 kcal mol ) that directly leads to the presumed organozinc intermediate.
2+
In the presence of Zn , a three-step mechanism was identified in which the cation coordinates the anomeric
and ring oxygen atoms and also the sugar iodine atom, causing an activating effect on the zinc insertion
2
+
process by facilitating the homolytic rupture of the C-I bond. Complexes between zinc and Zn bound
2
+
carbohydrates were characterized with large stabilization energies, suggesting that Zn might enhance
the affinity of the organic compound with the zinc metal surface.
Introduction
SCHEME 1
The reductive elimination from halogeno sugars (Bernet-
Vasella reaction) is a reaction sequence frequently used in
carbohydrate chemistry. First reported by Vasella’s group with
1
6
-deoxy-6-bromo-D-glucopyranose derivatives, the method was
2
extended to 5-deoxy-5-halo-pentofuranosides (Scheme 1).
The metal that is by far the most used to induce this reaction
3
is zinc, even if other reducing systems such as indium and Zn/
*
Corresponding authors. Tel.: (E.H.) 33 (0)326918774; (C.P.) 33(0)326913234.
Groupe de Spectrom e´ trie Mol e´ culaire et Atmosph e´ rique.
R e´ actions S e´ lectives et Applications.
†
‡
(1) Bernet, B.; Vasella, A. HelV. Chim. Acta. 1979, 62, 1990.
1
0.1021/jo0614590 CCC: $37.00 © 2007 American Chemical Society
Published on Web 03/09/2007
J. Org. Chem. 2007, 72, 2271-2278
2271

Henon et al.
4
Ag on graphite gave good results. Zinc (powder or dust) needs
SCHEME 2
to be activated (acidic washing, addition of TMSCl), and a protic
medium such as an alcohol or THF containing between 3 and
1
1
5% of water is generally required to perform these reactions.
In the framework of a project dedicated to the chemical
transformation of monosaccharides from hemicelluloses into
higher added value materials, we considered the possibility to
apply the reductive elimination methodology to 5-deoxy-5-iodo
derivatives of D-xylose and L-arabinose, the main components
of wheat hemicelluloses. This study gave us the opportunity to
observe surprising behaviors. In particular, 5-deoxy-5-iodo-D-
xylofuranosides can react quantitatively or can be perfectly inert,
depending on the protecting group pattern. On the other hand,
the epimeric L-arabino analogue of the unreactive D-xylo
derivative proved to be quantitatively transformed under the
same conditions. These experiments disclosed that the outcome
of the Vasella reaction of 5-iodo-pentofuranosides depends
strongly on structural parameters and prompted us to study more
deeply the structure-reactivity relationship, in connection with
the reaction conditions. A computational approach helped us
to discriminate some hypotheses and to propose a mechanism
for the reductive part of the reaction sequence. We report in
this paper the results of this investigation.
SCHEME 3
Results and Discussion
Experimental Approach. Compounds 1x and 1a, obtained
5
in five steps from D-xylose and L-arabinose, respectively, were
treated by acid activated zinc dust, under sonication, in a mixture
of THF/H2O 4:1, according to the conditions used by Madsen
et al.3 The results are summarized in Scheme 2. D-Xylose-
derived iodo sugar 1x proved to be perfectly inert, whereas the
L-arabinose-derived analogue 1a was completely converted into
the expected aldehyde 2, which was trapped in situ by
O-benzylhydroxylamine to give 3 in 78% overall yield.
a,6a
Since alkyl D-xylofuranoside analogues were reported to react
under such conditions, the first step, the insertion of zinc into
7
the C-I bond, seems to depend on the C-4 configuration and/
6
or on the rigidity of the fused bicyclic system. To estimate
these two parameters, both acetonides 1x and 1a were converted
into the corresponding methylfuranosides 4x and 4a by
5
converted into the expected aldehyde 2 isolated as the O-
benzyloxime 3 in 83 and 88% overall yield, respectively.8
treatment in methanol under hydrogen chloride catalysis, and
the methyl glycosides were then treated with zinc dust under
the same conditions (Scheme 3). The two compounds behaved
similarly: 4x (mixture of anomers) or an equimolar mixture of
epimeric iodo-methylfuranosides 4x and 4a were completely
According to these results, the C-4 configuration is not the
only parameter to be considered to explain the difference of
reactivity between 1x and 1a: the overall shape of the molecule
being of importance, as one could expect for a transformation
of which the first step is a heterogeneous process. These
observations led us to consider more deeply the reaction
mechanism in connection with the reaction conditions.
(
2) (a) Kobori, Y.; Myles, D. C.; Whitesides, G. M. J. Org. Chem. 1992,
7, 5899. (b) Uenishi, J.; Ohmiya, H. Tetrahedron 2003, 59, 7011.
3) (a) Hyldtoft, L.; Madsen, R. J. Am. Chem. Soc. 2000, 122, 8444. (b)
Yadav, J. S.; Reddy, B. V. S.; Srinivasa Reddy, K. Tetrahedron 2003, 59,
333.
4) (a) F u¨ rstner, A.; Weidmann, H. J. Org. Chem. 1989, 54, 2307. (b)
5
(
To assess the influence of zinc treatment, the iodo derivatives
4x, 1a, and 1x were treated with zinc dust under various
conditions (Table 1). Compound 4x was first considered as a
model substrate. In all experiments, 10 equiv of zinc dust was
used under ultrasound irradiation. While a total conversion of
5
(
F u¨ rstner, A.; Weidmann, H. J. Org. Chem. 1990, 55, 1363. (c) F u¨ rstner, A.
Tetrahedron Lett. 1990, 31, 3735. (d) F u¨ rstner, A.; Jumbam, D.; Teslic, J.;
Weidmann, H. J. Org. Chem. 1991, 56, 2213.
(5) Bercier, A.; Plantier-Royon, R.; Portella, C. Green Chem. 2006,
4x took place using acid washed zinc dust (entry 1), the starting
submitted.
material was completely recovered after sonication for 6 h using
unactivated metal (entry 2).
(6) (a) Poulsen, C. S.; Madsen, R. J. Org. Chem. 2002, 67, 4441. (b)
Boyer, F.-D.; Hanna, I.; Ricard, L. Org. Lett. 2001, 3, 3095.
(7) In carbohydrate chemistry, the zinc-induced ring opening reaction
of halogeno sugars is generally referred to as a reductive elimination
promoted by zinc; as pointed out by a referee, the insertion of a metal into
a covalent bond can also be described as an oxidative addition of zinc to
the C-I bond in the standard organometallic nomenclature.
(8) No difference of reactivity was observed between the two anomers
of methylfuranoside 4x. Ratio â/R remained constant (1.54) during the
reaction.
2272 J. Org. Chem., Vol. 72, No. 7, 2007

5
-Deoxy-5-iodo-D-xylo- and L-Arabinofuranosides
TABLE 1. Influence of Reaction Conditions on the Ring Opening
of 4x, 1a, and 1x
DFT molecular orbital computations to obtain an insight into
the mechanism of the zinc insertion reaction.
entry
1
substrate
reaction conditions
activated Zn dust
conversiona
Y
Theoretical Considerations. To the best of our knowledge,
there are no computational studies of the zinc-induced ring
opening of iodo sugars. As suggested by Grob et al.,11 the
organozinc intermediate might be formed in the first step. Next,
the cleavage of the molecule can proceed by three basic
mechanisms differing in the order in which the fragments are
released through a one- or a two-step process. For the concerted
4x
THF/H2O 4:1
2
3
4x
4x
untreated Zn dust THF/H2O 4:1
untreated Zn dust
THF/H2O 4:1
N
Y
ZnCl2 (1 equiv)
4
5
6
7
4x
1a
1a
1a
untreated Zn dust THF
ZnCl2 (1 equiv)
untreated Zn dust
THF/H2O 4:1
activated Zn dust
THF/H2O 4:1
untreated Zn dust
THF/H2O 4:1
N
N
Y
Y
1
2
fragmentation, Grob has derived rules according to which an
antiperiplanar arrangement of the bonds that will be broken is
required. Such stereoelectronically controlled fragmentations
1
3
14,15
have been the subject of experimental and computational
studies. It has been shown that the reduction of the carbon-
4d
halogen bond competes in some cases with the ring opening
ZnCl2 (1 equiv)
reaction. This can be taken as an indication of the existence of
the organozinc intermediate. As explained before, the studied
L-arabinose and D-xylose-derived iodo compounds behave quite
differently in their reaction with zinc: they undergo either a
ring opening Grob-type elimination or they are completely
unreactive. It is to be noticed that no reduction products were
formed in the latter case. The possibility that the acetal opens
first (before the elimination) was ruled out. Actually, although
8
9
1x
1x
1x
activated Zn dust
THF/H2O 4:1
untreated Zn dust
THF/H2O 4:1
untreated Zn dust THF/H2O 4:1
ZnCl2 (1 equiv)
N
N
N
1
0
a
N ) no reaction. Y ) conversion into enal 2.
2+
this reaction could be promoted by Lewis acidic Zn , 1x
9
According to Luche, the activation accomplished by ultra-
sound does not require, in most cases, a pre-activation of the
metal before addition of the organic co-reagents. Hence, in the
studied Vasella fragmentation, the role of the acidic treatment
is not a simple activation of the zinc surface with elimination
of zinc oxide. The HCl treatment induces the formation of zinc
chloride at the metal surface. Such a Lewis acid could interact
with the sugar derivative and participate in the overall activation
process. Indeed, the addition of zinc chloride (1 equiv) to
untreated zinc dust allowed a complete conversion of 4x into
the expected aldehyde 2 within 1.5 h under sonication (entry
2+
remains intact in the presence of Zn (Table 1, entry 10). Thus,
it appears from our experimental results that the rate-limiting
step for the studied molecules is the insertion reaction of zinc
into the C-I bond. That is the reason why the subsequent
fragmentation step was not modeled and why the possible
organized aggregation of the RZnI species was not examined
as well. We know of only one theoretical study in which the
insertion reaction of zinc into the C-I bond has been examined.
1
6
In this work, the B3LYP/PVTZ//B3LYP/6-311G** barrier
height computed within the gas-phase model for the insertion
-1
reaction of Zn into the C-I bond of CH2I2 was 29.5 kcal mol .
Moreover, our experimental results have shown that the use of
zinc chloride is required to obtain the fragmentation products.
Some Zn(II) promoted reactions are known in the literature.1
For instance, complexation with Lewis acids is known to play
a role in the radical conformation control in free radical
3
). This is exemplified by mass spectrometric monitoring of
the reaction (Figure 1), which clearly exhibits the activating
role of ZnCl2. Actually, while the starting material remained
inert under sonication using untreated zinc, a strong peak
appeared (m/z ) 435) in the spectrum after addition of ZnCl2
that corresponds to a dimeric form of the hydroxy aldehyde 2
7-20
1
8
2+
reactions. The presence of Zn in the Vasella elimination of
bromo derivatives obtained from D-glucose allowed the authors19
to explain a subsequent rearrangement according to the chelation
of the Lewis acid with the OH and carbonyl oxygens in the
enal intermediate. Thus, our main objective in this theoretical
study has been to contribute to the understanding of the role of
(
the minor peak m/z ) 261 is related to the C-I bond reduction
compound).
Moreover, it can be noticed that the presence of water in the
reaction medium is necessary since no reaction was observed
in anhydrous THF (entry 4). These experiments show for the
first time, in the Vasella fragmentation, the crucial role of zinc-
2
+
the Lewis acid Zn in the mechanism of the organozinc
formation and to try to rationalize the observed reactivities.
(II) previously formed at the metal surface during acidic
washing. In the same way, we have checked that the reductive
elimination reaction of the L-arabinose-derived acetonide 1a was
also controlled by acidic washing of zinc metal or by zinc
chloride addition (entries 5-7). In contrast, the D-xylose-derived
acetonide 1x failed to react with whatever the reaction conditions
(10) After sonication for 1 h and 30 min, 4x totally disappeared, and
aldehyde 2 was observed as a dimer using mass spectrometry (m/z ) 435
+
) [2 × 2 + Na] ).
(
(
(
11) Grob, A.; Schiess, P. W. Angew. Chem., Int. Ed. Engl. 1967, 6, 1.
12) Grob, A. Angew. Chem., Int. Ed. Engl. 1969, 8, 535.
13) Bernet, B.; Vasella, A. HelV. Chim. Acta 1984, 67, 1328.
(
entries 8-10). Could the nature of the interaction between iodo
2+
sugar and Zn ions explain the lack of reactivity of compound
x (as opposed to the reacting species 1a and 4x)? Following
(14) Andr e´ s, J.; Queralt, J. J.; Safont, V. S. J. Phys. Org. Chem. 1996,
9
, 371.
1
(15) Queralt, J. J.; Andr e´ s, J.; Mois e´ s Canle, L.; Hermogenes Cobas, J.;
these experimental studies, a fundamental interest appeared for
the understanding of the factors that influence the selectivity
Santaballa, J. A.; Sambrano, J. R. Chem. Phys. 2002, 280, 1.
(16) Fang, W.-H.; Phillips, D. L.; Wang, D.-q.; Li, Y.-L. J. Org. Chem.
2002, 67, 154.
of the reaction and the role of Zn2 in the Vasella reaction.
+
(
17) Nakamura, E.; Hirai, A.; Nakamura, M. J. Am. Chem. Soc. 1998,
Therefore, in parallel with the experimental work, we performed
1
20, 5844.
(
18) Renaud, P.; Gerster, M. Angew. Chem., Int. Ed. 1998, 37, 2562.
(9) Luche, J.-L.; Sarandeses, L. A. In Organozinc Reagents; Knochel,
(19) Holt, D. J.; Barker, W. D.; Ghosh, S.; Jenkins, P. R. Org. Biomol.
P., Jones, P., Harwood, L. M.; Moody, C. J., Eds.; Oxford University Press
Inc.: New York, 1999; pp 307-323.
Chem. 2004, 2, 1093.
(20) Yamamoto, H.; Futatsugi, K. Angew. Chem., Int. Ed. 2004, 44, 1924.
J. Org. Chem, Vol. 72, No. 7, 2007 2273

Henon et al.
FIGURE 1. Mass spectrometry monitoring of the reductive elimination of 4x: (a) with untreated zinc; THF/H
addition of ZnCl ; sonication for 1 h and 30 min.
2
2
O, sonication for 6 h and (b)
1
0
Reaction pathways were explored on the potential energy surface
for both 1,2-O-isopropylidene L-arabino- and D-xylofuranose
compounds. Solvent effects were assessed by the water solvent
polarized continuum model (PCM). Full PCM geometry opti-
mization calculations were performed for estimating the solvent
effects on structures and the changes in energy for the zinc
insertion reaction. As the reactions described herein were carried
out in a polar solvent (water was added), the Zn2 bound sugar
species were assumed to be dissociated from the counterions
because their charges are partially screened out by the solvent.
+
2
+
The study of the solvent-separated complex between the Zn
bound sugar and the counterions would require simulations with
explicit water molecules, which is more computationally
expensive. One of the limitations of the present calculations is
the simulation of an isolated Zn(0) atom. Metal zinc bulk could
be further investigated by means of cluster models or the
periodic slab21 approach. Consequently, the structures and
energy profile determined here with a continuum solvent model
represent the first step toward a description of the metal zinc
FIGURE 2. Reactants and transition state HF-DFT(MPW1K/
LANL2DZ) optimized structures involved in the zinc insertion reaction
via a concerted mechanism (distances in angstroms; TS56 stands for
the transition state connecting 5 and 6). Geometries were fully optimized
within the water PCM model.
2
+
insertion reaction into the C-I bond in presence of Zn . For
reasons of computational cost, the calculations reported in the
present work were performed on a model system obtained by
replacing the benzyl group with the methyl group. The x and a
notations have been reserved for the results concerning the
structures derived from D-xylose and L-arabinose, respectively.
TSij stands for the transition state connecting species i and j.
Although a partial charge appears on the zinc metal and ion in
the studied systems, these atoms are still designated as Zn and
In both cases (TS56x and TS56a), the incoming zinc atom
approaches the carbon site from the opposite direction to the
rest of the sugar. No transition state for the insertion reaction
was found with the zinc atom placed near the oxygen atom of
the ring. While the insertion process is concerted, it is non-
synchronous with two important events taking place in transition
state TS56. The first event is the C-I bond cleavage and the
Zn-I bond formation. Actually, the breaking C-I bond (2.85
Å) in the transition state is significantly longer (by about 31%)
than that in the unperturbed iodo sugar (2.17 Å), and the forming
Zn-I bond length (2.83-2.87 Å) is only slightly different (by
12%) from the final bond length (2.56-2.59 Å) in the
organozinc intermediate. The second event is the C-Zn bond
formation (length of the forming C-Zn bond is 3.2-3.3 Å in
TS56 and 2.02 Å in the organozinc intermediate). This predicted
non-synchronous behavior of the insertion process is consistent
with the computed electron spin density distribution in the
transition state, mainly localized on the carbon and zinc atoms
2+
Zn in the next part of the discussion according to their initial
charge.
2+
Zinc Insertion Reaction without Zn . The insertion
reaction without Zn2+ was found to occur through a concerted
mechanism that directly leads to the organozinc intermediate.
This result agrees favorably with the one-step insertion mech-
anism predicted by Fang et al.16 for the insertion of Zn into the
C-I bond of CH2I2. The solvent-optimized geometries of both
reactants and transition states (TS) are shown in Figure 2, where
relevant geometrical parameters have been indicated.
(but not on the iodine atom). A summary of the potential energy
profiles is shown in Figure 3. The PCM zero-point corrected
relative energies are given with respect to the considered
(
21) Bocquet, M-L.; Michaelides, A.; Loffreda, D.; Sautet, P.; Alavi,
A.; King, D. A. J. Am. Chem. Soc. 2003, 125, 5620.
2274 J. Org. Chem., Vol. 72, No. 7, 2007

5
-Deoxy-5-iodo-D-xylo- and L-Arabinofuranosides
FIGURE 3. Water PCM HF-DFT(MPW1K/LANL2DZ) potential
2+
energy profiles for the zinc insertion reaction (with and without Zn ),
including zero-point vibrational energy. x and a designate species
derived from D-xylose and L-arabinose, respectively. Energies are in
-
1
kcal mol
.
2+
reactants (with or without Zn ). The insertion process was
-
1
found to be exothermic by about 48-57 kcal mol . The
-
1
computed barriers are 32.3 and 32.0 kcal mol (33.3 and 32.6
-
1
kcal mol in the gas-phase model) above the reactants for
FIGURE 4. Zn2+ bound reactants and transition state HF-DFT-
MPW1K/LANL2DZ) optimized structures involved in the zinc inser-
tion reaction through an indirect mechanism (distances in angstroms;
TSij stands for the transition state connecting i and j). Geometries were
fully optimized within the water PCM model.
TS56x and TS56a, respectively, in compliance with the barrier
(
-
1
16
of 29.5 kcal mol predicted by Fang et al. at the gas-phase
B3LYP/6-311G** level of theory for the insertion reaction of
Zn into the C-I bond of CH2I2. These very close activation
energies are, however, clearly inconsistent with the difference
in reactivity observed between the L-arabinose and the D-xylose-
derived iodo compounds in their reaction with zinc.
Geometries (Figure S1) and energy profiles (Figure S2)
obtained in the gas phase are reported in the Supporting
Information. From these results, it appears that the nonspecific
solvent effect is small for the concerted mechanism.
product, could be found by our computations. Instead, two
intermediates (referred to as 8 and 9) were identified and
confirmed to be connected with the reactants, TSs, and products
according to the reaction scheme
2
+
2+
Zn ‚‚‚RI (7) + Zn f Zn ‚‚‚RI‚‚‚Zn
2
+
•
•
2+
2+
(8) f TS89 f Zn ‚‚‚R ‚‚‚IZn
Zinc Insertion Reaction with Zn . As the Lewis acid Zn
2
+
has been experimentally found to play a role in the zinc-induced
Vasella reaction mechanism, we examined how the presence
(
9) f TS9 (10) f Zn ‚‚‚RZnI (10)
2
+
of one Zn ion could affect the structure and energy of the
stationary points presented previously for the insertion step. The
complexation of Zn2 in the sugar was assumed to take place
in the vicinity of the C-I bond. More precisely, the mode of
bonding of Zn2 with the studied iodo sugars was chosen
according to the previously mentioned factors that seem to
influence the reactivity. Hence, we envisaged the chelation of
Selected optimized geometries involved in this mechanism
are presented in Figures 4 and 5.
+
2+
In this mechanism, after the coordinate Zn intermediate 8
is formed (step 1), the insertion reaction starts via the transition
state TS89, which describes (step 2) both the homolytic rupture
of the C-I bond and the Zn-I bond formation. This concerted
process leads to the intermediate 9, the equilibrium structure
of which is described in Figure 5. Interestingly, the two unpaired
electrons of 9 are mainly located on the carbon atom (1.1) and
the zinc atom (0.9), thus facilitating the formation of a C-Zn
bond to produce the organozinc intermediate. From an entropy
point of view, it should be noted that the cation-halogen bond
+
2+
Zn by the anomeric and ring oxygen atoms as depicted in
structures 7x and 7a (Figure 4).
These two similar complexes involve three chelating atoms,
including the iodine atom and resulting in a large stabilization
-
1
energy of -98.2 and -112.5 kcal mol (-212.4 and -226.9
-
1
•
kcal mol in the gas phase), respectively. As a consequence,
interaction in 9 keeps the newly formed IZn fragment close to
the C-I bond in the reactant is lengthened by approximately
the region of the sugar after the dissociation, in particular, near
the carbon radical center. This also facilitates the next associa-
tion step. Step 3 is the association of the two fragments through
the transition structure TS9 (10).
.03-0.05 Å when Zn2 coordinates with the organic com-
+
0
pound, thus favoring a possible cleavage. Unlike the previous
concerted two-stage reaction,22 in the presence of Zn , the zinc
insertion was found to proceed through a three-step process.
Actually, no indication for the occurrence of a direct reaction,
characterized by a single barrier between the reactants and the
2+
For the discussion, we will first focus on the influence of
Zn2 on the insertion process (TS89 and TS9 (10)). The role
of the precursor intermediate 8 is addressed at the end of the
discussion.
+
In comparison with TS56, the relative energies of TS89 and
TS9 (10) lie near and below the reactant levels, respectively
(
22) Dewar, J. S.; Olivella, S.; Stewart, J. J. P. J. Am. Chem. Soc. 1986,
1
08, 5771.
J. Org. Chem, Vol. 72, No. 7, 2007 2275

Henon et al.
noting that species 8x and 8a slightly differ in structure. Both
2+
involve the Zn‚‚‚Zn coordinate bound associated with a rather
large stabilization energy, but the former is tri-coordinated while
the latter is tetra-coordinated. This might support the idea that
the reaction selectivity experimentally observed in the Vasella
reaction of the D-xylose-derived acetonide 1x and the L-
arabinose-derived acetonide 1a could originate from the behavior
2
+
of the corresponding Zn bound carbohydrates at the metal
surface.
Comparison with the gas-phase results (reported in the
Supporting Information) indicates that the nonspecific interaction
of the solvent exerts only a small effect on both the energetic
and the geometrical parameters. The main characteristics of the
gas-phase profile remain valid in solution except for TS89x,
which in the gas phase (Figure S3) does not involve any
decoordination during the Zn-I bond making/C-I bond break-
ing process. Indeed, the location of the zinc atom in TS89x
changes significantly without solvent effect (the ZnIZn bond
angles in the solvent and gas phase are 94.2 and 155.6°,
respectively). As a consequence, the gas-phase minimum energy
pathway (Figure S2) connects TS89x to the relatively loosely
bound tri-coordinated complex 8′x in the reverse direction. In
this intermediate complex, the zinc atom is not strongly bound
FIGURE 5. HF-DFT(MPW1K/LANL2DZ) optimized structures of
2+
intermediates involved in the zinc + Zn bound sugar reaction system
distances in angstroms). Geometries were fully optimized within the
water PCM model.
(
2
+
(Figure 3). Interestingly, step 2 is concerted but non-synchro-
to Zn but interacts with the iodine atom with a stabilization
-
1
nous. Actually, according to the small degree of C-I bond
lengthening in the transition state TS89 (12-15%), there is no
doubt that the C-I bond breaking takes place only in the second
half of step 2 between TS89 and 9. Thus, in contrast with TS56,
the reaction energy begins to fall in step 2 before the bond
cleavage is fully developed. Part of the reason is that the C-I
bond cleavage allows for the beneficial strengthening of the
energy of only -10.6 kcal mol . The 8′x f TS89x barrier
-1
height is then predicted to be much lower (9.6 kcal mol ) than
-
1
2+
the one (36-43 kcal mol ) required to break the Zn‚‚‚Zn
coordination bond. These gas-phase results emphasize the Zn2
+
assistance for the insertion process when no decoordination is
2
+
required. A Zn‚‚‚Zn complex 8x (Figure S4, analogue to 8x
in Figure 5) was also found in the gas-phase model with a
2+
-1
Zn ‚‚‚I coordinate bond (2.77-2.69 Å in 7x and 7a and 2.55
stabilization energy of -44.6 kcal mol , but it is not directly
2+
Å in 9x and 9a). So, it appears that the presence of Zn causes
the C-I bond to be weakened and facilitates its dissociation.
On the other side, it is the Zn-I bond making that takes place
in the first half of step 2 (the forming Zn-I bond length in
TS89 is only slightly different by 1-2% from its final value in
connected to the C-I dissociation pathway in this case.
It results from our calculations that the coordinate bond
2+
between Zn and iodine atom in the parent carbohydrate can
play a substantial role in lowering the activation energy for the
insertion process. Such a bond was also found at the gas-phase
2
+
9
). However, for this bond to be formed, the Zn‚‚‚Zn
2+
level of theory in the Zn bound methylfuranosides (R/â)
coordinate bond in 8 must be cleaved, resulting in a large energy
derived from 7x and 7a (noted 11xR, 11xâ, 11aR, and 11aâ in
-
1
barrier for step 2 (41-43 kcal mol ). Actually, due to the
proximity of the two zinc atoms in TS89, the minimum energy
pathway (Figure 3) has been found to connect TS89 to the stable
Zn‚‚‚Zn2 coordinated complex 8 in the reverse direction
the Supporting Information), indicating that these compounds
2+
may also be influenced by the presence of Zn in their reaction
with metal zinc as suggested by our experimental findings.
Obviously, further theoretical work, especially on the organic
molecule at the zinc surface, is required before a deeper
understanding can be gained in the reactivity difference observed
between the two sugars. Actually, the investigated processes
occur most certainly on the surface of the bulk metal, involving
the zinc-organic species. It should be mentioned that in
+
(
further information is given in the Experimental Section). Thus,
a chemically activated intermediate 8 can be formed from 7 in
a barrier-less association between Zn2 bound sugar and zinc
atom. However, owing to the large stabilization potential energy
+
-
1
(
39-48 kcal mol ) of 8, it is likely that this intermediate species
be stabilized by energy transfer to the solvent precluding the
insertion reaction of the zinc metal atom coordinated to the Zn
center. Hence, it is not unreasonable to assume that the
4d
employing the zinc/silver-graphite reagent, F u¨ rstner et al.
2+
achieved the reduction of the C-I bond from the iodo sugar 1x
(which failed to react in our case).
2+
coordination interaction (between Zn and Zn) first governs
the chemisorption of the Zn2 bound substrate on the metal
surface. Nevertheless, this does not rule out a subsequent
insertion mechanism involving another reacting zinc metal atom.
It then would be interesting to study the effect of the presence
of more than a single zinc metal atom on the previously
presented indirect mechanism. More precisely, one may wonder
if a Zn decoordinationless mechanism could exist for the zinc
insertion process starting from the chemisorbed precursor that
would involve an activation energy lower than the large 8-TS89
barrier. Our gas-phase results suggest the possible existence of
such a mechanism (discussed hereafter). At this stage, it is worth
+
Conclusion
Experimental evidence shows that the zinc-induced reductive
elimination of 5-deoxy-5-iodo-pentofuranosides is strongly
dependent on both structural parameters and reaction conditions.
The result is clearly a function of the nature of the protecting
group and of the configuration at C-4 as well. At the same time,
experiments clearly show that the reaction is very sensitive to
the presence of the Zn(II) Lewis acid. Our theoretical investiga-
2
+
tion gives new hints to explain the role of Zn , assuming that
the fragmentable substrate is the organozinc compound. From
2276 J. Org. Chem., Vol. 72, No. 7, 2007

5
-Deoxy-5-iodo-D-xylo- and L-Arabinofuranosides
2
+
used in predicting chemical reactivity.25 The reactivity of some
organometallic complexes has suitably been obtained using the HF-
DFT MPW1K/LANL2DZ methodology,30 even though the use of
a basis set larger than LANL2DZ is recommended. As the molecule
investigated is relatively large, it was not possible to increase the
size of the basis set in our case. Reactants, complexes, and transition
state structures were fully optimized using the analytical gradients
our calculations, we find that, in the presence of Zn , the
insertion reaction takes place in three elementary steps with
transition states near or below the reactant energy level. The
cation is likely to coordinate the anomeric and ring oxygen
atoms and also the halogen of the sugar, causing an activating
effect on the zinc insertion process by facilitating the homolytic
rupture of the C-I bond. The dissociation is followed by an
association step via a loose transition state. The whole process
starts off with the barrier-less formation of a complex between
at the HF-DFT MPW1K/LANL2DZ level of theory within both
31-33
the gas-phase model and the PCM.
In the latter case, the
structures (minima and first-order saddle points) were fully
optimized in the field of water using the solvent accessible surface
(SAS). The standard dielectric constant of water implemented in
the Gaussian program was employed (ꢀ ) 78.39). Vibrational
frequencies were determined within the harmonic approximation,
at the same level of theory (vacuum and PCM). Small imaginary
frequencies were observed for some stationary points at the PCM
level of theory. This may originate in the approximate algorithm
used to obtain the analytical PCM second derivatives. Thus, for
these stationary points, the second derivatives were computed
numerically using analytically calculated PCM first derivatives (this
took 110 CPU hours each). Despite these precautions, one small
imaginary frequency remained for species 9a (32i). This low
frequency is likely to be sensitive to the step size for the numerical
differentiation in this flat region of the PES. Because of the
computational cost of such numerical calculations, the gas-phase
ZPE was used in this case. The energies presented in this work
include the zero-point energy correction. Gas-phase entropies and
Gibbs free energies were estimated at 298 K. All transition states
were characterized by one imaginary frequency (first-order saddle
points on the potential energy surface (PES)) associated with the
desired reaction pathway. The unrestricted approach was used for
transition states and biradical species. The resulting spin-contamina-
tions in the unrestricted electronic structure calculations of TS56x,
TS56a, TS89x, and TS89a were 0.72, 0.68, 0.29, and 0.39,
respectively. The spin-contaminations of the HF-DFT wave
functions for TS9 (10)x, TS9 (10)a, 9x, and 9a were in the range
of 0.98-1.00. The quadratically convergent algorithm (much slower
than DIIS) was used for the studied systems because SCF iterations
sometimes had convergence difficulties. Some HF-DFT wave
functions were first obtained with an internal instability for transition
states and biradical species. That is the reason why the stability of
the solution was tested for each stationary point before any geometry
2
+
2+
zinc and Zn bound sugar involving a Zn‚‚‚Zn coordinate
bond in the iodo sugar. No significant difference was found
that could explain the opposite reactivity displayed by the two
iodo sugars. It turns out that the Zn‚‚‚Zn2+ interaction results
2
+
in a large stabilization energy, suggesting that Zn might
enhance the affinity of the organic compound with the zinc metal
surface whatever the considered sugar in our case. Despite our
theoretical work, we have to conclude that the mechanism for
the zinc insertion reaction into the C-I bond of the iodo sugars
derived from D-xylose and L-arabinose is not satisfactorily
known yet. For the final clarification of these hypotheses, the
carbohydrate environment has to be taken into account. More
precisely, it would be interesting to address the role of explicit
water molecules and metal surface in the previously presented
2+
Zn assisted mechanism.
Experimental Section
Acid Activation of Zinc. Zinc dust was stirred in a 3 N aqueous
HCl solution (0.5 mL/mmol of zinc) for 5 min, then filtered, and
washed with water (0.5 mL/mmol), ethanol (0.5 mL/mmol), and
ether (0.5 mL/mmol). Finally, the material was dried under high
vacuum with a heat gun until the powder became pale gray.
Reaction with Activated Zinc. To a solution of 5-iodo
compounds in a 4:1 THF/H
activated zinc dust (10 equiv). The reaction was sonicated at 40
C until the starting material disappeared. The mixture was filtered
through cotton to eliminate zinc, and solvents were evaporated.
Reaction with Untreated Zinc and ZnCl . To a solution of
-iodo compounds in a 4:1 THF/H O mixture (5 mL/mmol) was
added zinc dust (10 equiv). The reaction was sonicated at 40 °C
for 6 h, then anhydrous ZnCl (1 equiv) was added. The reaction
2
O mixture (5 mL/mmol) was added
°
2
5
2
2
(24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K.
N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A.
D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A.
G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
was sonicated at 40 °C until the starting material disappeared.
Computational Details. Calculations were performed using the
23,24
Gaussian 98 and Gaussian 03 packages.
The presence of two
zinc atoms and one iodine atom along with the studied protected
sugars made the use of high level theoretical methods costly. Hence,
the calculations reported in the present work were performed on a
model system obtained by replacing the benzyl group with the
methyl group. Several levels of theory were tested previously to
25
26-29
choose the HF-DFT MPW1K /LANL2DZ
method. In par-
ticular, the standard HF/LANL2DZ or HF-DFT(B3LYP)/LANL2DZ
methods did not allow for locating the transition state such as TS89
or TS9 (10). The MPW1K hybrid density functional was developed
relatively recently by Truhlar and co-workers with the aim of being
03, revision C.02; Gaussian, Inc.: Wallingford, CT, 2004.
(
25) Lynch, B. J.; Fast, P. L.; Harris, M.; Truhlar, D. G. J. Phys. Chem.
(23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
A 2000, 104, 4811.
M. A.; Cheeseman, V. G.; Zakrzewski, J. R.; Montgomery, R. E.; Stratmann,
J. C., Jr.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Salvador, P.;
Dannenberg, J. J.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Ciolowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B.
B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin,
R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,
A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M.
W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople,
J. A. Gaussian 98, revision A.11.1; Gaussian, Inc.: Pittsburgh, PA, 2001.
(26) Dunning, T. H.; Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., III, Ed.; Plenum: New York, 1976; pp 1-28.
(27) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270.
(28) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284.
(29) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(30) Iron, M. A.; Lo, H. C.; Martin, J. M. L.; Keinan, E. J. Am. Chem.
Soc. 2002, 124, 7041.
(31) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117.
(32) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem. Phys. 2002,
117, 43.
(33) Cammi, R.; Mennucci, B.; Tomasi, J. J. Phys. Chem. A 2000, 104,
5631.
J. Org. Chem, Vol. 72, No. 7, 2007 2277

Henon et al.
optimization or frequency calculation and why the wavefunction
was reoptimized to the lower energy solution when an instability
was found. Special care was taken to determine minimum energy
pathways (MEPs), performing intrinsic reaction coordinate analyses
8a in which the zinc atom was strongly bound to Zn2+. We have
made a detailed search on the PES but have not been able to find
a loosely bound complex 8′a (located rigorously as a minimum)
similar to 8′x involving the Zn‚‚‚I interaction. In particular, two
relaxed potential energy scans were carried out starting from the
geometrical parameters of 8′a. First, the energy was observed to
increase continuously as the Zn-I distance was scanned from 3.44
up to 5.0 Å while holding fixed the Zn-I-Zn2 coordinate to its
value in 8′a (103.0°). In the second scan, the representative point
of the system on the PES moved down to the potential energy well
of 8a as the Zn-I-Zn2 coordinate was scanned from 122 to 50°.
(IRC) to confirm that a specific transition state connected to the
designated local minima. Only gas-phase IRC calculations were
carried out because the reaction path following was a time-
consuming procedure. Some problems arose, however, during this
procedure. For transition states TS9 (10) (association pathway),
the vibrational analysis gave a very low imaginary frequency (about
+
+
47i), whose animation is consistent with the formation of the C-Zn
bond but through a very loose TS. This is the reason why the
standard IRC algorithm failed at connecting TS9 (10) to the
corresponding minima, due to the flatness of the PES in this region.
Rotational energy profiles for internal rotations about the I-Zn2
bond in species 9x and 9a were performed to estimate a possible
rotational barrier before the reacting system 9 connected to TS9
Acknowledgment. C.R.I.H.A.N. computing center and the
computational center of the Universit e´ de Reims Champagne-
Ardenne (ROMEO) are acknowledged for the CPU time
donated. The authors thank Prof. F. Bohr for helpful discussions.
This work was supported by the Contrat de Plan Etat-R e´ gion
+
(
10). The energy was found to increase by no more than 1 kcal
(
Program GLYCOVAL-Europol’agro). We are indebted to
CNRS and R e´ gion Champagne-Ardenne for a PhD fellowship
A.B.).
-
1
2+
mol as the Zn-I-Zn -O dihedral angle was scanned from its
value in 9 to its value in TS9 (10). In other respects, the IRC
(
calculation from TS89a in the reverse direction terminated after
/2
1
85 steps (of 0.06 amu1 ) before reaching the complex 8a but had
Supporting Information Available: General methods; experi-
progressed sufficiently in this direction to establish the C-I
1
13
mental procedures and characterization data ( H and C NMR
spectra, electrospray ionization mass spectra) for compound 3; and
computational data: potential energy, zero point vibrational energy
dissociation pathway. The last IRC point (denoted 8′a in the
-
1
following discussion) energy lies only 3.0 kcal mol below the
reactant level with small energy gradients (but nonzero), and the
corresponding geometry 8′a is very similar to that of complex 8′x
(in hartrees), Cartesian coordinates of all stationary points both in
the gas phase and in the water PCM models. This material is
available free of charge via the Internet at http://pubs.acs.org.
(
C-I ) 2.257 Å and Zn-I ) 3.455 Å in 8′a, as compared to 2.238
and 3.427 Å in 8′x). A straightforward geometry optimization of
this 8′a system led to the formation of the notably stable complex
JO0614590
2278 J. Org. Chem., Vol. 72, No. 7, 2007