Raising the ceiling of diastereoselectivity in hydrogen transfer on acyclic radicals

Article abstract of DOI:10.1021/jo802634w

We report here that the monodentate complexation of Me₂AlCl to an ester group significantly enhances the selectivity of hydrogen transfer on acyclic radicals flanked by both an ester functionality and a stereogenic center, leading to C-2,C-3-an

Full text of DOI:10.1021/jo802634w

Raising the Ceiling of Diastereoselectivity in Hydrogen Transfer on
Acyclic Radicals
Irina Denissova,† Luca Maretti,| Brian C. Wilkes,† J.C. Scaiano,| and Yvan Guindon*,‡,§
Institut de Recherches Cliniques de Montre´al (IRCM), 110 aVenue des Pins Ouest, Montre´al, Que´bec,
Canada H2W 1R7, Department of Chemistry, UniVersite´ de Montre´al, C.P.6128, succursale Centre-Ville,
Montre´al, Que´bec, Canada H3C 3J7, Department of Chemistry, McGill UniVersity, 801 rue Sherbrooke
Ouest, Montre´al, Que´bec, Canada H3A 2K6, and Department of Chemistry, UniVersity of Ottawa,
10 Marie Curie, Ottawa, Ontario, Canada K1N 6N5
yVan.guindon@ircm.qc.ca
ReceiVed NoVember 30, 2008
We report here that the monodentate complexation of Me2AlCl to an ester group significantly enhances
the selectivity of hydrogen transfer on acyclic radicals flanked by both an ester functionality and a
stereogenic center, leading to C-2,C-3-anti products with high diastereoselectivity. In certain cases improved
ratios were obtained using bulkier aluminum Lewis acid such as MAD (methylaluminum-di(di-2,6-tert-
butyl-4-methylphenoxide). Electron spin resonance studies on these acyclic radicals indicate that Lewis
acid complexation leads to conformational changes in the radicals. The stereochemistry of the preferred
enolate radicals complexed with Lewis acids and their impact on the acyclic transition states involved
are suggested.
Introduction
diastereoselectivity of processes involving radical intermediates.1
Our group has studied extensively the influence of Lewis acids
on the reactivity of tertiary acyclic radicals flanked both by an
ester and a stereocenter at C-3 in hydrogen transfer and allylation
reactions.2 These radicals are formed by the homolytic cleavage
of either a carbon-halogen or a carbon-selenium bond of the
corresponding R-halo or seleno-esters. The relative stereochem-
istry varies depending on the reaction conditions used. Thus,
the C-2,C-3-syn products are obtained when the hydrogen
transfer reactions are conducted in the presence of bidentate
Lewis acid (the endocyclic effect) as illustrated in Scheme 1
Many natural products of significant biological interest (e.g.,
polyketides) contain an acyclic subunit, consisting of a defined
sequence of stereogenic centers with various chemical func-
tionalities, synthesis of which is often very problematic. Indeed,
inducing stereoselectivity in reactions involving acyclic mol-
ecules poses a significant challenge due to their high flexibility,
allowing for multiple transition states of similar energy and poor
stereoselectivity. The use of a Lewis acid can significantly
improve the selectivity either by restricting the number of
conformations via temporary ring formation or by changing the
steric and stereoelectronic factors around the site being trans-
formed, thus favoring one transition state. Since the early 1990s
Lewis acids have been widely employed to improve the
(1) For the most recent review, see: Guérin, B.; Ogilvie, W. W.; Guindon,
Y. Lewis Acid-Mediated Diastereoselective Radical Reactions. In Radicals in
Organic Synthesis; VCH: Weinheim, 2001; Vol. 1, pp 441-460.
(2) (a) Cardinal-David, B.; Brazeau, J-F.; Katsoulis, I.; Guindon, Y. Curr.
Org. Chem. 2006, 10, 1939–1961. (b) Guérin, B.; Chabot, C.; Mackintosh, N.;
Ogilvie, W. W.; Guindon, Y. Can. J. Chem. 2000, 78, 852–867. (c) Guindon,
Y.; Jung, G.; Guérin, B.; Ogilvie, W. W. Synlett 1998, 3, 213–220. (d) Guindon,
Y.; Guérin, B.; Rancourt, J.; Chabot, C.; Mackintosh, N.; Ogilvie, W. W. Pure
Appl. Chem. 1996, 68, 89–96.
† Institut de Recherches Cliniques de Montre´al.
‡ Universite´ de Montre´al.
§ McGill University.
| University of Ottawa.
2438 J. Org. Chem. 2009, 74, 2438–2446
10.1021/jo802634w CCC: $40.75 2009 American Chemical Society
Published on Web 02/18/2009

Raising the Ceiling of DiastereoselectiVity
SCHEME 1. Proposed Transition States Leading to the Corresponding 2,3-syn and 2,3-anti Products
SCHEME 2. Proposed Transition States for Substrates Not Having a Disubstituted C-4 Center and without a ꢀ-Heteroatom
(entry 1).3 When no Lewis acid is added, C-2,C-3-anti diaster-
eomers are generated (entry 2).4
dipole-dipole repulsion. As suggested before,5 this leads to the
dihedral angle θ1(TS2) being close to 30°. The top face attack,
leading to the syn product, imposes the θ1 of TS3 to be smaller
than θ1 in TS2. The dipole-dipole repulsion is not completely
minimized in TS3, increasing the energy of the transition state.
Moreover, the trajectory of the incoming hydrogen donor would
be more sterically encumbered in TS3 than in TS2, increasing
even more the difference in energy between these transition
states. When R ) H, this difference is decreased as reflected
by lower ratios.
In the absence of the heteroatom at C-3 and thus of
dipole-dipole repulsion, the minimization of 1,3-allylic strain
then becomes the major controlling factor (Scheme 2, entry 2).
It favors the conformation in which the ꢀ-hydrogen is placed
in the plane with the ester bringing the dihedral angle θ1 close
to 0°. Thus, the dihedral angle θ1 in TS4 is smaller than θ1 in
both TS2 and TS3. As a result, significant anti-ratios in these
cases could only be achieved in the presence of a very bulky
The lowest energy transition states TS1 and TS2 were
proposed to rationalize the major products obtained in these
reactions. The anti-diastereoselectivity noted in the latter case
was dependent on the presence of an ester (or amide) at C-1
and maximized by the presence of a heteroatom at C-3. A
substituent at C-4 (R) was also found to be critical, a lower
ratio being observed when R ) H (entry 1, Scheme 2).
In kinetically controlled reactions, such as the hydrogen
transfer reaction in this study, the ratio of observed products is
determined by the difference of energy between the correspond-
ing transition states leading to the anti and syn products (TS2
versus TS3). TS2, the lowest energy transition state, is the result
of the synergy between steric and electronic factors: the
minimization of the allylic-1,3 strain and of the intramolecular
(3) (a) Guindon, Y.; Rancourt, J. J. Org. Chem. 1998, 63, 6554–6565. (b)
Guindon, Y.; Guérin, B.; Chabot, C.; Ogilvie, W. W. J. Am. Chem. Soc. 1996,
118, 12528–12535. (c) Guindon, Y.; Lavallee, J.-F.; Llinas-Brunet, M.; Horner,
G.; Rancourt, J. J. Am. Chem. Soc. 1991, 113, 9701–9702.
(4) (a) Guindon, Y.; Lavallee, J.-F.; Boisvert, L.; Chabor, C.; Delorme, D.;
Yoakim, C.; Hall, D.; Lemieux, R.; Simoneau, B. Tetrahedron Lett. 1991, 32,
27–30. (b) Guindon, Y.; Yoakim, C.; Lemieux, R.; Boisvert, L.; Delorme, D.;
Lavallee, J.-F. Tetrahedron Lett. 1990, 31, 2845–2848.
(5) Giese, B.; Damm, W.; Wetterich, F.; Zeic, H.-G.; Rancourt, J.; Guindon,
Y. Tetrahedron Lett. 1993, 34, 5885–5888.
(6) (a) Giese, B.; Damm, W.; Wetterich, F.; Zeic, H.-G. Tetrahedron Lett.
1992, 33, 1863–1862. (b) Durkin, K.; Liotta, D.; Rancourt, J.; Lavallee, J.-F.;
Boistvert, L.; Guindon, Y. J. Am. Chem. Soc. 1992, 114, 4912–4914.
J. Org. Chem. Vol. 74, No. 6, 2009 2439

Denissova et al.
substituent (e.g., t-Bu group).6 These steric requirements,
unfortunately, severely restrict the synthetic usefulness of these
hydrogen transfer reactions, and the search for solutions has
become an objective.
In the past we reported that linking the substituent at C-3
and C-5 through a ring (the exo-cyclic effect)7 is a strategy that
could be used to circumvent some of these problems (where R
) H) as seen in entry 3, Scheme 2. The present study offers
another solution. Herein we report that the formation of a
monodentate complex with the ester group, using a bulky
aluminum Lewis acid, can significantly improve the hydrogen
transfer diastereoselectivity.8 A mechanistic rationale, supported
by electron spin resonance (ESR) studies, is suggested.
interesting in its own right, suggesting that a combination of a
tertiary centered free radical, an ester, and a sequence of
contiguous tertiary carbon centers maybe sufficient to foster
diastereoselectivity. A mixture of side products 11 and 12
resulting from the oxygen addition was also isolated in 20%
yield. When the reaction was performed using either 1.5 equiv
of Me2AlCl in toluene (entry 11) or employing a larger excess
of Me2AlCl in CH2Cl2 (entry 12), the anti:syn product ratio
increased to >20:1. The formation of the side product was also
suppressed.
Aryl ester 13 was studied next. Unsurprisingly, for this ester
having an sp2 carbon center at C-4, the diastereoselectivity of
the hydrogen transfer reaction in the absence of the Lewis acid
was low in both DCM and toluene (3:1 and 2.3:1 respectively,
entries 13 and 14). Adding 1.5 equiv of MAD in toluene
improved the diastereoselectivity by 2-fold, a 6:1 ratio being
obtained (entry 15). CollectiVely, these results underline the
utility of aluminum Lewis acids in enhancing the diastereose-
lectiVity of the hydrogen transfer eVen in case of “problematic”
substrates. A mechanistic rationale was then sought.
Results and Discussion
We have observed an increase in anti-diastereoselectivity and
in relative rates of the hydrogen transfer when studying the
radical reduction of various ꢀ-alkoxy-R-halo-esters9 with
Bu3SnH and Lewis acid in dichloromethane (DCM) or toluene
at -78 °C (Table 1).
Mechanistic Considerations. Upon addition of Me2AlCl to
ꢀ-alkoxy-esters, complexes 16 and 17 between the Lewis acid
and the substrate may be formed (Scheme 3).
As shown in Table 1 (entries 1-3), a significant increase of
diastereomeric anti:syn ratio going from 10:1 to 20:1 was
observed in the reduction of ester 1 when 1.0 equiv of Me2AlCl
was added. The yields were also improved. One will note that
the time needed to complete the reaction was significantly
reduced (from 6 to 1 h). Studying the radical reduction of the
ꢀ-alkoxy ester not having a tertiary carbon center at C-4 (a
disubstituted carbon center) such as 4 was also of great interest.
Not surprisingly, its reaction in the absence of Lewis acid gave
almost no diastereoselectivity (entries 4 and 5) in both DCM
and toluene. Remarkably, almost a 3-fold increase of diaste-
reoselectivity was noted in the presence of 1.0 equiv of Me2AlCl
in toluene, the ratio rising from 1.5:1 to 4:1 (entry 6). This ratio
was further improved to 6.5:1 by replacing Me2AlCl with a
bulkier aluminum Lewis acid such as MAD (methylaluminum-
di(di-2,6-tert-butyl-4-methylphenoxide), Figure 1)10,11 in toluene
(entry 7). A ratio of 9:1 was obtained when 4 equiv of MAD
was used (entry 8). No reaction was observed in the absence of
the initiator (Et3B) (entry 9), attesting to the involvement of a
free-radical-based process in the reactions previously described
(vide supra).
These complexes are in equilibrium with the noncomplexed
substrate. Complex 18 could also be formed.12-14 One will note
that (as illustrated in Scheme 1, entry 1) bidentate complexes
(17 or 18) should lead to syn products as the result of hydrogen
transfer reaction. However, in our experiments (using 1 equiv
of Me2AlCl) we have observed an improvement of the ratios
favoring the anti isomer, suggesting empirically that the
bidentate reaction pathways were neither dominant nor important.
We decided to verify qualitatively the nature of the complexes
induced by Me2AlCl and the substrate in solution using 13C
NMR (Figure 2A–C).
The 13C NMR performed on ester 1 with 1 equiv of Me2AlCl
in CD2Cl2 at -40 °C points in the direction of the formation of
a monodentate carbonyl-Me2AlCl complex as the most popu-
lated intermediate (Figure 2B). Indeed, there is a downfield shift
of the signal belonging to the ester quaternary carbon ac-
companied by a significant line-broadening of this peak. Line-
broadening suggests a dynamic equilibrium between the
carbonyl-Me2AlCl complex and the free carbonyl group.15 The
methyl of the ester is also shifted downfield (∆δ ) 4.7 ppm).
Remarkably, the 13C NMR of ester 1 with 4 equiv of Me2AlCl
at -40 °C (Figure 2C) reveals the formation of a distinct new
species, the signals of the ester, methyne (C-3), and benzylic
methylene carbons having shifted downfield with ∆δ ) 11.7,
7.2, and 8.1 ppm, respectively. This new species was identified
as bidentate ionic complex 19 (Scheme 4). To our knowledge,
this is a first observation of such a complex for ꢀ-alkoxy esters.
Additional experimental evidence of the bidentate complex
formation was obtained from the radical reduction of ester 1 in
the presence of 4 equiv of Me2AlCl in DCM at -78 °C (Scheme
We then turned our attention to substrates lacking the alkoxy
group at C-3 and therefore dependent solely on steric factors in
their hydrogen transfer reactions. In the absence of the Lewis
acid the reduction of a C-2,C-3 diastereomeric mixture of
R-iodo-ꢀ-methyl-esters 7 and 8 gave a 12:1 selectivity favoring
the anti-diastereomer (entry 10). This result is extremely
(7) (a) Guindon, Y.; Pre´vost, M.; Mochirian, P.; Gue´rin, B. Org. Lett. 2002,
4, 1019–1022. (b) Guindon, Y.; Liu, Z.; Jung, G. J. Am. Chem. Soc. 1997, 119,
9289–9290. (c) Guindon, Y.; Faucher, A.-M.; Bourque, E.; Caron, V.; Jung, G.;
Landry, S. J. Org. Chem. 1997, 62, 9276–9283.
(8) There are only few examples in the literature where the use of
monodentate chelates between the carbonyl and the Lewis acid influences the
diastereoselectivity of either a radical reduction or an atom transfer on the radicals.
For example, Sato et al. reported that using a bulky aluminum Lewis acid such
as methylaluminum di(2,4,6-trimethylphenoxide) in the radical addition of BuI
to R-methelene-γ-phenyl-butyrolactone in the presence of Bu3SnH reversed the
selectivity of the hydrogen-transfer step, favoring formation of anti product (dr
) 1.5:1), whereas in the absence of the Lewis acid good syn-selectivity (dr )
9:1) was observed. See: Urabe, H.; Kobayashi, K.; Sato, E. J. Chem. Soc. Chem.
Commun. 1995, 104, 3–1044.
(12) Previous studies in the literature have demonstrated formation of
monodentate complexes 16 when 1 equiv of Me2AlCl was added to ꢀ-alkoxy-
aldehydes, ꢀ-alkoxy-ketones (see ref 13), and ꢀ-alkoxy-oxazolodinones (see ref
14), whereas using an excess of the Lewis Acid led to formation of bidentate
ionic complexes 18.
(13) Evans, D. A.; Allison, B. D.; Yang, M. G.; Masse, C. E. J. Am. Chem.
Soc. 2001, 123, 10840–10852.
(9) The preparation of the esters is described in Supporting Information.
(10) Maruoka, K.; Itoh, T.; Sakurai, M.; Nonoshita, K.; Yamamoto, H. J. Am.
Chem. Soc. 1988, 110, 3588–3597.
(11) Renaud, P.; Moufid, N.; Kuo, L. H.; Curran, D. P. J. Org. Chem. 1994,
59, 3547–3552.
(14) Castellino, S; Dwight, W. J. J. Am. Chem. Soc. 1993, 115, 2986–2987.
(15) Castellino, S. J. Org. Chem. 1990, 55, 5197–5200, and ref 9 therein.
(16) We did also suggest in the past on the basis of the experimental data
that the monodentate complexes reacted more rapidly than the bidentate
complexes in the hydrogen-transfer reaction.
2440 J. Org. Chem. Vol. 74, No. 6, 2009

Raising the Ceiling of DiastereoselectiVity
TABLE 1. Studies on the Radical Reduction
b
a Reaction did not go to completion; ∼50% starting material was still present. Conversion based on H NMR of the crude. c No Et3B was added.
1
pathway is operative in the hydrogen transfer when 1 equiv of
Me2AlCl is used.16
As shown in Scheme 5 complex 20 and the nonchelated ester
could both participate in the formation of the carbon-centered
radical via halogen abstraction and the subsequent hydrogen
transfer reaction. Hence, in the case of the hydrogen transfer
FIGURE 1. Structure of MAD.
reaction, responsible for a diastereoselective outcome, four
competing pathways are possible, as illustrated in Scheme 5.
Two transition states A (nonchelated) and C (monochelated)
would lead to the anti product, while the corresponding B
(nonchelated) and D (monochelated) lead to syn products.
4, step 2). The reaction afforded syn diastereomer in 20:1 ratio,
signifying involvement of bidentate transition state TS1 (Scheme
1, entry 1). One may thus conclude that only the monodentate
J. Org. Chem. Vol. 74, No. 6, 2009 2441

Denissova et al.
SCHEME 3. Formation of Mono- and Bidentate Complexes between the Ester and Me2AlCl
SCHEME 4. Formation of Bidentate Ionic Complex 19 Followed by the Hydrogen Transfer Reaction
We believe that transition states C and D are lower in energy
than their corresponding nonchelated versions. Indeed, the
presence of an electron-withdrawing group R to a carbon-
centered radical is known to increase its reactivity toward
trialkyltin hydrides.17 A complexation of an ester by a Lewis
acid would enhance the electron-withdrawing characteristics of
this functionality. In other words, the complexation of the Lewis
acid to the ester is expected to lower the radical SOMO energy,
thus allowing for a better overlap with the HOMO energy of
Bu3SnH, which will decrease the energy of the transition states.
A similar argument would suggest that the halogen abstraction
step (by the electropositive tin radical) also accelerates when
the ester becomes more electrophilic upon coordination with a
Lewis acid.18 Taken together, this should result in the increase
of the relative reaction rates, which indeed was observed in our
studies from reduced reaction times in cases when Me2AlCl was
used.
However this would not explain per se an increase in the
anti-preferred ratio, since both the syn- and the anti-predictive
transition states would be lowered upon complexation with
Me2AlCl. Such an increase could only be explained if the
difference of energy between C and D (complexed transition
states) is greater than between A and B (noncomplexed transition
states).
What are the factors at the origin of the increase in energy
difference between transition states C and D as compared to
the one between A and B? One hypothesis would be to postulate
a conformational change of the complexed free radicals in the
transition state rendering the addition from the top face more
difficult. As suggested by Fischer and others, a carbon-centered
free radical R to the ester is planar and could therefore exhibit
an enol-like character.19 Hence, such an enol-like radical can
exist in the two geometrical forms Z (21) and E (22) (Scheme
6, entry 1).
Recently, Spichty et al. have reported UB3LYP/6-31G*
calculations of the minimum geometry of [D3]methyl-2-methyl-
4,4,4-trichlorobutanoate-2-yl radical.20 The E-isomer was found
to be 0.33 kcal/mol more stable than the corresponding Z-isomer.
Moreover, values for the dihedral angle of θ1(E-isomer) ) 15.9°
and θ1(Z-isomer) ) 18.0° were obtained. One could attribute the
angle increase to the increased 1,3-allylic strain in the Z-isomer
as compared to the E-isomer (OR > O•). The situation may be
different in our case, the oxygen radical being complexed with
a sterically demanding Lewis acid. In the isomer E (23) (note
change of ordering priority due to complexing L.A.) strong 1,2-
interaction is now present, whereas in the Z-isomer (24) the
allylic 1,3-strain is further increased (Scheme 6, entry 2). What
will be the net directing effect of such complexation on the
equilibrium between E- and Z-isomers?
To address this question, theoretical calculations21 were
performed on model olefins 25-Z and 26-E (Figure 3, entry 1)
and 25′-Z and 26′-E where a tert-butyl group replaced the Lewis
(17) Avila, D. V.; Ingold, K. U.; Lusztyk, J.; Dolbier, W. R.; Pan, H. Q.;
Muir, M. J. Am. Chem. Soc. 1994, 116, 99–104.
(18) (a) Xiao, X.; Pan, H.-Q.; Dolbier, W. R. J. Am. Chem. Soc. 1994, 116,
4521–4522. (b) Blackburn, E. V.; Tanner, D. D. J. Am. Chem. Soc. 1980, 102,
692–697. (c) Coats, D. A.; Tedder, J. M. J. Chem. Soc., Perkin Trans. 2 1973,
1570–1574. (d) See also ref 3a.
(19) Roduner, E.; Strub, W.; Burkhard, P.; Hockmann, J.; Percival, P. W.;
Fischer, H.; Ramos, M.; Webster, B. C. Chem. Phys. 1982, 67, 275–285.
(20) Spichty, M.; Giese, B.; Matsumoto, A.; Fischer, H.; Gescheidt, G.
Macromolecules 2001, 34, 723–726.
FIGURE 2. 13C NMR of ester 1 with 0.0 (A), 1.0 equiv (B), and 4.0
equiv (C) of Me2AlCl recorded at -40 °C.
2442 J. Org. Chem. Vol. 74, No. 6, 2009

Raising the Ceiling of DiastereoselectiVity
SCHEME 5. Proposed Complexed and Noncomplexed Pathways for the Hydrogen Transfer
SCHEME 6. Z and E Enol Radicals with and without Me2AlCl
angle between the hydrogen at C-3 and the plane of the ester in
24. This, in turn, would change the orientation of R3 and raise
the energy of transition state D (Scheme 5), therefore increasing
the energy difference between C and D as opposed to the one
between A and B. (As discussed before, when R2 ) OBn, the
minimization of the dipole-dipole electronic repulsion will
further increase the θ1 angle).
It is obviously difficult to experimentally support a hypothesis
on an increased θ1 angle in the transition state. However, the
radicals being known to react through early transition states,
the conformational changes in the ground state of these radicals
may be reflected in the corresponding transition state. Thus we
decided to study the conformational changes in the ground states
of various radicals occurring upon complexation with Me2AlCl
with low-temperature ESR techniques.
The ESR spectrum of an acyclic radical represents averaged
information obtained from a number of low energy conformations.
Complexation with Me2AlCl is expected to raise the rotation
barriers (vide supra), the conformation with a larger dihedral angle
becoming dominant. One should then detect an increase in the
corresponding coupling constant.22 The low-temperature ESR
studies were performed on radicals 27-30 (Figure 5).
Indeed, in the presence of Me2AlCl an increase in the
ꢀ-hyperfine coupling constant with the proton at C3 was
observed for both 29 and 30 compared to 27 and 28, respec-
tively, thus signifying increase in θ1 upon complexation. The
ESR experiments are described in the following section.
ESR Experiments. Nitrogen-equilibrated toluene solutions
of halo-ester 1 or the mixture of 7 and 8 in the presence of
Me2AlCl (when applicable) and (Bu3Sn)2 were irradiated in the
ESR spectrometer cavity at -78 °C (7 and 8 mixture) and -40
acid (Figure 3, entry 2). For the olefins without tert-butyl, the
E-isomer was found to be 0.24 kcal/mol more stable than the
Z-isomer, a value similar to the one reported by Spichty et al.
for enolate-like [D3]methyl-2-methyl-4,4,4-trichlorobutanoate-
2-yl radical (vida supra). With the tert-butyl group added (25′
and 26′), the inverse was observed: the Z-isomer was found to
be 1.46 kcal more stable than the E-isomer, suggesting that 1,2-
allylic strain was more important than 1,3-allylic strain.
By analogy with the olefin model one could then conclude
that the Z-enolate radical (24) (Figure 4) would be the prevalent
specie involved in both the anti- (C) and syn-predicitive (D)
transition states of the hydrogen transfer. Importantly, this
complexed Z-radical 24 would suffer from greater allylic 1,3-
strain, which has to be alleviated, compared to noncomplexed
21 and 22. This would therefore result in increase of the θ1
(21) See Experimental Section for details.
(22) These coupling constants are directly dependent on the <cos2 θ2>, where
θ2 is the dihedral angle formed by the direction of the 2pz axis (p orbital of the
radical) and the direction of the corresponding C-Hꢀ bond; the increase in the
coupling constant reflects smaller angles between the p-orbital and the C-Hꢀ
bond. This means that the dihedral angle between this C-Hꢀ bond and the plane
of the ester increases (θ2 ) 90°-θ1). See: Gerson, F.; Huber, W. Electron Spin
Resonance Spectroscopy of Organic Radicals; VCH: Weinheim, 2003; p 61.
FIGURE 3. Comparison of the 26-E and 25-Z enolates.
J. Org. Chem. Vol. 74, No. 6, 2009 2443

Denissova et al.
FIGURE 4. Increase of dihedral angle θ1 occurring upon complexation of Me2AlCl.
FIGURE 5. Series of radicals studied by ESR technique.
FIGURE 7. ESR spectrum of radical 29 (bottom) and its simulation
(top).
FIGURE 6. ESR spectrum of radical 27 (bottom) and its simulation
(top).
FIGURE 8. Possible low-energy ground-state conformations for
radicals 27 and 28.
°C (ester 1).23 Radicals were generated by continuous lamp
irradiation upon (Bu3Sn)2 photolysis. As a control experiment,
the ESR spectra were recorded for several minutes prior to lamp
irradiation to ensure that radicals were not forming via any other
pathway.
The ESR spectra of radicals 27 and 29 together with the
spectra simulations24 are shown on Figures 6 and 7, respectively.
The ESR spectrum of radical 27 originating from esters 7 and
8 (Figure 6) features eight lines as expected from the coupling
with the three equivalent protons of the methyl group at C-2
and the proton at C-3. Two coupling constants of 22.0 and 7.0
G for CH3 and CH, respectively, can be calculated from the
spectrum. However, the intensity distribution slightly differs
from that of expected binomial 1:1:3:3:3:3:1:1. This might
indicate a coexistence of several low energy radical conforma-
tions with different ꢀ-hyperfine coupling constants. This makes
it more difficult to explain the preference of the anti products
in this series in the absence of the Lewis acid, which is
reinforced by the fact that the spectrum of 27 does not exactly
fit the simulation performed under assumption of a single
conformation. Thus conformations F, G, and H could be
populated in the ground state (Figure 8).
FIGURE 9. Possible resonance structures of radical 29.
a distinctly different spectrum is detected (Figure 7). The
ꢀ-coupling constant for the proton at C3 is now 11.5 G, which
indicates significant changes in the conformation as well as an
increase in dihedral angle θ1. There is also an additional smaller
splitting with the coupling constant of 3.0 G. The coupling with
the three protons of the methyl group remains the same (22.0
G). The simulation of the spectrum for radical 29 using these
values gives the best overlap with the experimental spectrum
when the additional splitting (3.0 G) arises from the coupling
with two equivalent protons. These protons could come from
the CH2 belonging to ethoxy group, indicating a full delocal-
ization of the unpaired electron (Figure 9). It is reasonable to
suggest that donation of the electron to form the delocalized
structure should occur more readily from the most electron-
rich oxygen thus making J (the Z-enol ether radical complex)
a prevalent resonance structure. The g-value of 2.0045 calculated
from the simulations also points to the delocalization of the
unpaired electron on the oxygen. The spectrum of 29 fits well
with simulation for a single conformation and thus could support
the higher diastereoselectivity observed.
When the mixture of 7 and 8 is irradiated with Me2AlCl in
the solution, new radical species (presumably, radical 29) with
(23) See Experimental Section for further details.
(24) The computer simulations were performed using the EasySpin Software.
See: Stoll, S.; Schwieger, A. J. Magn. Reson. 2006, 178, 42–55.
2444 J. Org. Chem. Vol. 74, No. 6, 2009

Raising the Ceiling of DiastereoselectiVity
Conclusion
We have demonstrated that using either Me2AlCl or a bulkier
aluminum Lewis acid such as MAD can significantly enhance
the diastereoselectivity of the hydrogen transfer reaction on
acyclic radicals, thus expanding the scope of the previously
developed methodology, which will certainly find further
applications in synthesis. The low-temperature ESR studies
confirmed that the complexation of the Lewis acid induces
conformational changes in the radicals. This may be at the origin
of the greater difference between relative energies of the
corresponding anti- and syn-predictive transition states and
therefore of the greater diastereoselectivity in the hydrogen
transfer reaction.
FIGURE 10. ESR spectrum of radical 28 (bottom) and its simulation
(top).
Experimental Section
Theoretical Conformational Analysis. All calculations were
performed using the molecular modeling software SYBYL version
7.0 (Tripos Associates, St. Louis, MO). The standard Tripos force
field25 was used for energy calculations, and a dielectric constant
of 78 was used. Molecules were constructed using the standard
fragment library and were minimized using the conjugate gradient
approach.26 A grid search was performed on each compound in
which the dihedral angles of all rotatable bonds were systematically
varied by 30° increments.27 For each search, the resulting conform-
ers were grouped into families based on similarity of their dihedral
angles ((30°). The lowest energy member of each conformational
family was then minimized extensively, and the resulting conform-
ers were combined and ranked in order of increasing energy.
ESR Measurements. All measurements were carried out on a
JEOL (FA 100) EPR spectrometer equipped with ES-DVT4 variable
temperature controller. The 5-mm tube was placed into the cavity
of the spectrometer and irradiated in situ by means of a Xenon
illuminator (Luzchem). An H2O infrared filter (10 cm in length)
was placed between the lamp and the sample in order to avoid
heating of the sample.
FIGURE 11. ESR spectrum of radical 30 (bottom) and its simulation
(top).
Materials and Sample Preparation for the ESR Experiments.
Toluene (99.8%, anhydrous), Me2AlCl (1 M solution in hexanes),
and Bu6Sn2 were purchased from Aldrich and used as supplied.
A. Preparation of a Sample without Me2AlCl. A correspond-
ing ester (1.00 mmol) was preweighed in a dry vial. To the vial
was added toluene (500 µL), and the solution was transferred into
the oven-dried tube. Then Bu6Sn2 (50 µL, 1.00 mmol) was added
via syringe. The resulting solution was capped and degassed with
N2 for at least 30 min.
B. Preparation of a Sample with Me2AlCl. A corresponding
ester (1.00 mmol) was preweighed in a flame-dried round-bottom
flask and kept under N2 or argon. To the ester was added toluene
(n µL). The flask was cooled with dry ice, followed by consequent
addition of Bu6Sn2 (50 µL, 1.00 mmol) and Me2AlCl (n mmol, 1
M in hexanes). [Quantities used: for ester 1, toluene (400 µL),
Me2AlCl (100 µL, 1.00 mmol); for the mixture of esters 7 and 8,
toluene (350 µL), Me2AlCl (150 µL, 1.50 mmol).] The resulting
yellow solution was then quickly transferred via canula to the
precooled (dry ice) flame-dried tube under N2. (Alternatively, the
tube can be dried in the oven at 140 °C for at least 2 h and cooled
under N2 flow.) The solution was degassed with N2 for 20-30 min;
the tube was kept in the beaker with dry ice at all times during the
sample preparation.
The ESR spectrum of radical 28 (Figure 10) with ꢀ-coupling
constants of 6.5 and 22.3 G exhibits the expected pattern with
an appropriate binomial intensity distribution. This is consistent
with the dual (minimization of allylic 1,3-strain and dipole-dipole
repulsion) steric and electronic effects that have to be minimized
to favor conformation G (Figure 8). When ester 1 is irradiated
in the presence of Me2AlCl to form radical 30, a different
spectrum is obtained (Figure 11). There is a line broadening
and an increase in the second coupling constant to 8.5 G, also
implying that changes in the conformation of the radical occur
upon complexation of Me2AlCl to the ester with an increase of
the θ1 angle. The broadening could be attributed to the poorly
resolved third hyperfine coupling constant (resulting from the
delocalization on the methyl ester) as well as to an existing
dynamic equilibrium between complexed and noncomplexed
radicals.
In conclusion, from comparison of the ESR spectra of radicals
27 and 28 (Figures 6 and 10, respectively), we suggest that when
R2) Me (radical 27), a larger number of low energy conforma-
tions is accessible than when R2) OBn (radical 28) at the same
temperature. When Me2AlCl is added, the ESR spectra of 29
fits well the simulation for a single conformation, which could
be at the origin of the increased diastereoselectivity of hydrogen
transfer noted. In the case of 30 (R2 ) OBn), a conformational
preference is already established in the ground state, which angle
θ1 is further increased in the presence of Me2AlCl.
(25) Clark, M.; Cramer, R.D. III; Opdenbosch, N. V. J. Comput. Chem. 1989,
10, 982–1012.
(26) Powell, M. J. D. Math. Program. 1977, 12, 241–254.
(27) (a) Wilkes, B. C.; Nguyen, T.M.-D.; Weltrowska, G.; Carpenter, K. A.;
Lemieux, C.; Chung, N. N.; Schiller, P. W. J. Peptide Res. 1998, 51, 386–394.
(b) Wilkes, B. C.; Schiller, P. W. Biopolymers 1995, 37, 391–400. (c) Wilkes,
B. C.; Nguyen, T.M.-D.; Weltrowska, G.; Carpenter, K. A.; Lemieux, C.; Chung,
N. N.; Schiller, P. W. J. Peptide Res. 1998, 51, 386–394.
J. Org. Chem. Vol. 74, No. 6, 2009 2445

Denissova et al.
General Procedures for the Radical Reduction. All diaster-
eomeric ratios were determined from 500 MHz H NMR of the
crude reaction mixture.
or toluene (0.5 mL) was added Me3Al (75 µL, 0.15 mmol, 2 M in
hexane) at rt. Gas evolution was observed. The resulting solution
(pale yellow or colorless) was stirred for 1 h at rt and was then
cooled to -78 °C. A solution of the ester (0.10 mmol) in CH2Cl2
or toluene (0.5 mL) was added to it via canula, followed by Bu3SnH
(54 µL, 0.20 mmol) and Et3B (25 µL, 0.025 mmol, 1 M in CH2Cl2).
Addition of Et3B (25 µL) was repeated every 20 min, until the
reaction was judged completed by TLC. To the reaction mixture
was then added m-dinitrobenzene, and the mixture was stirred for
10 min at -78 °C, followed by addition of a saturated aqueous
solution of NaHCO3 and dilution with ether. The mixture was then
warmed to rt, and a saturated aqueous solution of Na+/K+ tartrate
(5 mL) was added to facilitate the separation. The phases were
separated, the aqueous phase was extracted with ether (×2), and
the combined organic fractions were dried over MgSO4 and
concentrated under reduced pressure.
1
A. Radical Reduction in the Absence of Lewis Acids. To a
solution of the ester (0.10 mmol) in CH2Cl2 (1 mL) at -78 °C was
added Bu3SnH (54 µL, 0.20 mmol) followed by addition of Et3B
(25 µL, 0.025 mmol, 1 M in CH2Cl2) and dry air (2 mL). Addition
of Et3B/air (25 µL/2 mL) was repeated every 20 min, until reaction
was judged completed by TLC. To the reaction mixture was then
added m-dinitrobenzene, and the mixture was stirred for 15 min at
-78 °C, followed by addition of brine and dilution with ether. The
mixture was then warmed to rt, and the phases were separated.
The aqueous phase was extracted with ether (×2), and the combined
organic fractions were dried over MgSO4 and concentrated under
reduced pressure.
B. Radical Reduction in the Presence of Me2AlCl. To a
solution of the ester (0.10 mmol) in CH2Cl2 or toluene (1 mL) at
-78 °C was added Me2AlCl (n equiv, 1 M in hexanes), stirred for
3-5 min. Then Bu3SnH (54 µL, 0.20 mmol) was added followed
by addition of Et3B (25 µL, 0.025 mmol, 1 M in CH2Cl2) and dry
air (2 mL). Addition of Et3B/air (25 µL/2 mL) was repeated every
20 min, until reaction was judged completed by TLC. To the
reaction mixture was then added m-dinitrobenzene, and the mixture
was stirred for 15 min at -78 °C, followed by addition of a
saturated aqueous solution of NaHCO3 and dilution with ether. The
mixture was then warmed to rt, and the phases were separated.
The aqueous phase was extracted with ether (×2), and the combined
organic fractions were dried over MgSO4 and concentrated under
reduced pressure.
Acknowledgment. Y.G. thanks NSERC and IRCM for their
generous funding. I.D. thanks Group Jean Coutu for a postdoc-
toral fellowship. J.C.S. thanks NSERC and CFI for generous
funding of research and equipment. We further thank Dr. Keith
Ingold for valuable comments on the ESR experiments.
Supporting Information Available: Experimental proce-
dures for the preparation of 1, 4, 7, 8, and 13; characterization
data for 1-12; proof of structure for 1, 3, 5, 6, and 9; X-ray
image for intermediate S-5. This material is available free of
charge via the Internet at http://pubs.acs.org.
C. Radical Reduction in the Presence of MAD. To a solution
of 2,6-di-tert-butyl-4-methylphenol (66 mg, 0.30 mmol) in CH2Cl2
JO802634W
2446 J. Org. Chem. Vol. 74, No. 6, 2009