1,3-Diastereocontrol in acyclic radical allylations

Article abstract of DOI:10.1016/S0040-4039(02)02596-0

The radical allylation of an acyclic α-hydroxyketone with allyltributyltin under chelation-controlled conditions is reported. Several reaction conditions were explored, including radical initiators, solvents, and temperatures to improve the yield and the diastereomeric ratio. Some Lewis acids, like magnesium bromide etherate and zinc chloride, gave superior diastereomeric ratios (up to 100:1) and good yields.

Full text of DOI:10.1016/S0040-4039(02)02596-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 531–534
1,3-Diastereocontrol in acyclic radical allylations
Eric J. Enholm,* Sophie Lavieri, Tanya Cord o´ va and Ion Ghiviriga
Department of Chemistry, University of Florida, Gainesville, FL 32611, USA
Received 21 October 2002; revised 13 November 2002; accepted 14 November 2002
Abstract—The radical allylation of an acyclic a-hydroxyketone with allyltributyltin under chelation-controlled conditions is
reported. Several reaction conditions were explored, including radical initiators, solvents, and temperatures to improve the yield
and the diastereomeric ratio. Some Lewis acids, like magnesium bromide etherate and zinc chloride, gave superior diastereomeric
ratios (up to 100:1) and good yields. © 2002 Elsevier Science Ltd. All rights reserved.
Adequate methods for diastereoselective reactions in a1
wide variety of acyclic systems have been examined.
Similar concepts can now be applied to radical reac-
tions, where high levels of diastereoselectivity can be
metric synthesis, close proximity is generally the key to
good stereochemical control. The second, and more
poorly understood variation of this process, involves
1,3-diastereocontrol, shown in Scheme 1B as 3“4. This
more difficult situation does not have the close proxim-
ity as before; now the new carbon center is further
away and less able to be directed by the R , R and R
3
2
realized. It is also clear that excellent acyclic diastereo-
control in free radical reactions is possible by the use of
chelation with Lewis acids, following from seminal
1
2
3
studies by Guindon and co-workers.
substituents.
The most common form of diastereocontrol emanates
from a single chiral center exerting its influence on an
adjacent carbon-centered radical. As shown in Scheme
One of our goals was to design radical precursors that
would employ Lewis acids in chelation effects of a-
hydroxyketones. This work seeks to more clearly define
the effects of asymmetric 1,3-diastereocontrol over a
more demanding two-carbon span separated by a
1A, the more studied transformation of 1“2 involves a
halogen-substituted carbon leading to 1,2-diastereocon-
trol next to the chiral center. In radical-mediated asym-
4
ketone, as shown in the transformation in Scheme 2.
For the best chances for success, a large steric bias,
such as a t-butyl group, would block one face of a
metal-chelated or hydrogen-bonded ring, as shown in 6
in Scheme 2. Degrees of freedom and rotation about
the two-carbon span were to be disfavored by placing a
ketone between the two interacting groups. The ketone
Scheme 1.
Keywords: free radical; diastereoselective; allyltributyltin.
*
Corresponding author. Tel.: +1-352-392-9700; fax: +1-353-846-0296;
e-mail: enholm@chem.ufl.edu
Scheme 2.
0
040-4039/03/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02596-0

5
32
E. J. Enholm et al. / Tetrahedron Letters 44 (2003) 531–534
would also play an important role in the chelation and
at the same time rigidify the substrate. Note that the
carbon bearing the bromide will eventually become a
carbon centered radical that, unlike other chelation-
based five-membered asymmetric methodology, is out-
side of the metal-mediated ring in 6 and it is free to
tions, hydrogen bonding in 5 between the alcohol and
the ketone forms a five-membered ring conformation
but still favors 7 over 8. It is interesting to note that
Et B in the absence of Lewis acids resulted in substan-
3
tial decomposition of the starting material 5 affording
only modest yields at best. The low yields in the
absence of Lewis acids might result from the enoxyl
radical undergoing homolytic substitution at boron to
make a boron enolate which may react further with the
ketone.
5
rotate. In these studies we hoped that diastereoselec-
tive induction would occur in a completely predictable
fashion so that one product, 7, would be favored from
conformation 6. In addition, we also predicted that
without Lewis acids, hydrogen bonding might favor the
same products expected with Lewis acids but only
experimental scrutiny would bear this out.
When a Lewis acid was employed, diastereomeric ratios
increased substantially. The initiator system of triethyl-
borane and O2 was used in these reactions at low
(
−78°C) temperatures. We predicted that the steric bulk
Preparation of racemic bromide 5 began (Scheme 3) by
conversion of commercially available pinacolone 9 to
the a-hydroxyketone 10 by permanganate oxidation to
the corresponding acid followed by reduction with
sodium borohydride using a modification of a literature
provided by the t-butyl group would enhance our selec-
tivities to some degree (vide infra). Lewis acids ZnCl2
and MgBr ·OEt2 proved to be superior Lewis acids
producing high 100:1 diastereomeric ratios in good
yields.
2
6
,7
procedure. Ketone 11 was next obtained by treatment
with excess n-BuLi, followed by Br in acetic acid to
2
Lewis acid complexation appears to accelerate many
examples of radical reactions. This occurs by weaken-
ing the CꢀBr bond or increasing the electrophilic nature
of the radical. Porter and co-workers have found that a
CꢀBr bond next to a carbonyl, as in 6, is weakened by
chelation with a Lewis acid. Clark and Wayner have
also studied this bond-weakening effect suggesting that
the carbonyl destabilizes the CꢀBr bond dipole in the
parent compound, stabilizing the radical with respect to
the parent. Thus, a chelated carbonyl with a carbon
bearing bromine (CHBr) has a dramatic effect on the
normal CꢀBr dipole and weakens the CꢀBr bond.
give 5. We reacted a-hydroxyketone 5 with allyl-
tributyltin under various free radical conditions of ini-
tiators, temperatures, Lewis acids and solvents. We are
pleased to report excellent results with diastereoselectiv-
ities up to 100:1. Of the two products, 7 was indeed
favored over 8 and in some cases by a wide margin, as
shown in Table 1.
5a
8
A brief examination of Table 1 shows several trends. In
the absence of Lewis acids, lowering the temperature of
the reaction surprisingly only slightly increased
diastereoselectivity. The reaction at −78°C gave a 5:1
ratio versus 3:1 at higher temperatures. In these reac-
There were several lines of evidence that confirmed the
major stereochemistry in 7. First, an analogy to a very
similar series of Diels–Alder reactions performed by
Masamune and co-workers in 1983 produced products
trans to the t-butyl group by attacking the back face of
alkene 12a (Scheme 4). In the related sp to sp radical
reaction of 12c, the rotation in the n-butyl side chain is
also limited by the bulky t-butyl group. Because of this,
7
2
3
1
2c is favored over 12d in the equilibrium for the two
radical species. This leads to the trans product 7 pre-
dominating over the cis product 8.
A second line of evidence used to determine the stereo-
chemistry of 7 was based on energy minimizations and
Scheme 3.
Table 1. Results of radical allylations
Radical initiator
Temp. (°C)
Solvent system
Lewis acid (equiv.)
Ratio of 7:8
Yield (%)
AIBN
80
25
0
0
0
Benzene
CH Cl
None
None
None
ZnCl2 (2)
MgBr ·OEt (2)
Yb(CF SO ) (2)
None
ZnCl2 (2)
MgBr ·OEt (2)
Yb(CF SO ) (2)
4:1
3:1
3:1
6:1
6:1
5:1
5:1
100:1
100:1
9:1
90
40
23
75
91
66
33
76
81
48
Et B, O2
3
2
2
2
Et B, O2
CH Cl
3
2
Et B, O2
CH Cl /THF
3
2
2
Et B, O2
CH Cl /Et O
3
2
2
2
2
2
2
2
Et B, O2
0
CH Cl
3
2
3
3 3
Et B, O2
−78
−78
−78
−78
CH Cl
3
2
Et B, O2
CH Cl /THF
3
2
2
Et B, O2
CH Cl /Et O
3
2
2
2
2
Et B, O2
3
CH Cl
2
2
3
3 3

E. J. Enholm et al. / Tetrahedron Letters 44 (2003) 531–534
533
Guidelines and Synthetic Applications; VCH: Weinheim,
996.
1
3
. (a) Guindon, Y.; Houde, K.; Prevost, M.; Cardinal-
David, B.; Landry, S. R.; Daoust, B.; Bencheqroun, M.;
Guerin, B. J. Am. Chem. Soc. 2001, 123, 8496–8501; (b)
Guindon, Y.; Prevost, M.; Mochirian, P. Org. Lett. 2002,
4
, 1019–1022; (c) Guindon, Y.; Rancourt, J. J. Org.
Chem. 1998, 63, 6554–6565; (d) Guindon, Y.; Guerin, B.;
Rancourt, J.; Chabot, C.; Mackintosh, N.; Ogilvie, W.
W. Pure Appl. Chem. 1996, 68, 89–96.
4
. Longer range acyclic diastereoselective radical reactions
using an oxazolidinone ring as a control element are
studied, see: (a) Sibi, M. P.; Johnson, M. D.; Punniya-
murthy, T. Can. J. Chem. 2001, 79, 1546–1555; (b) Sibi,
M. P.; Manyem, S. Tetrahedron 2000, 56, 8033–8061; (c)
Sibi, M. P.; Ji, J.; Sausker, J. B.; Jasperse, C. P. J. Am.
Chem. Soc. 1999, 121, 7517–7526; (d) Sibi, M. P.; Porter,
N. A. Acc. Chem. Res. 1999, 32, 163–171; (e) Renaud, P.;
Gerster, M. Angew. Chem., Int. Ed. Engl. 1998, 37,
Scheme 4.
2
562–2579; (f) Porter, N. A.; Feng, H.; Kavrakova, I. K.
9
molecular dynamics (MD) simulations. Since it was
not possible to separate products 7 and 8, the calcula-
tions were used in conjunction to NMR studies to
determine stereochemistry by long distance proton–pro-
ton coupling constants for the methine hydrogens on
Tetrahedron Lett. 1999, 40, 6713–6716; (g) Wu, J. H.;
Radinov, R.; Porter, N. A. J. Am. Chem. Soc. 1995, 117,
1
. Two examples of diastereoselective radical reactions out-
side a chelated ring include: (a) Feng, H.; Kavrakova, I.
K.; Pratt, D. A.; Tellinghuisen, J.; Porter, N. A. J. Org.
Chem. 2002, 67, 6050–6054; (b) Mero, C. L.; Porter, N.
A. J. Am. Chem. Soc. 1999, 121, 5155–5160.
. Jaeger, D. A.; Broadhurst, M. D.; Cram, D. J. J. Am.
Chem. Soc. 1979, 101, 717–731.
. Masamune, S.; Ree, L. A., II; Davis, J. T.; Choy, W. J.
Org. Chem. 1983, 48, 4441–4444.
. Clark, K. B.; Wayner, D. D. M. J. Am. Chem. Soc. 1991,
1029–11030.
5
10
each side of the ketone. This third line of stereochem-
ical evidence indicates a small, yet clearly observable,
long distance H–H coupling in the minor product 8
because these hydrogens are nearly coplanar (4.62 ppm,
6
7
8
9
11
d of d, J=7.2, 1.5 Hz). Major product 7 gave no long
distance H–H coupling (3.93 ppm, d, J=7.2 Hz).
In summary, the 1,3-diastereocontrolled radical allyla-
tion of acyclic a-hydroxyketones using allyltributyltin
at low-temperature conditions (−78°C) is reported. A
number of reaction conditions were studied to improve
the yield and the diastereomeric ratio, that ranged from
113, 9363–9365.
. Modeling experiments were carried out in the Brain
Institute Computer Center, University of Florida,
Gainesville, on a Silicon Graphics Indigo XZ4000 work-
station using Discover as implemented in Insight II from
Molecular Simulations, Inc., 1998. All energy minimiza-
tions and MD simulations were carried out using CVFF
force field (MSI, Inc., 1998) in vacuum. Prior to MD, all
structures were optimized using conjugated gradient
algorithm to final gradients better than 0.005 kcal/mol A.
Equilibration for 10 ps followed energy minimization.
Structures were subject to a series of MD simulations for
10,000 steps, T=300 K, time step 1.00 fs, saving confor-
mations every ten steps. Energetically accessible confor-
mations were generated using MD simulations for the
two possible products of the reaction: compounds 7 and
8. A total of 1011 structures were analyzed for each
molecule to determine lowest energy and average confor-
mations. Selected conformers were analyzed. We deter-
mined that for 8, both minimum energy and average
structures show a conformation in which the two hydro-
gens in the chiral centers are closer to be in a plane than
the equivalent hydrogens in product 7. Starting with
conformations in which the hydroxyl and the carbonyl
group are closer to a hydrogen-bond five-ring conforma-
tion, further MD simulations were performed for each
isomer. Lowest energy structures were selected and dihe-
dral angles were measured. Despite the differences in the
values of the dihedrals, the same tendency is observed;
for isomer 8 the preferred conformation places the two
3:1 to 100:1.
References
. (a) Kim, Y. H. Acc. Chem. Res. 2001, 34, 955–962; (b)
Concellon, J. M.; Rodriguez-Solla, H.; Huerta, M.;
Perez-Andres, J. A. Eur. J. Org. Chem. 2002, 1839–1847;
1
(
c) Ward, D. E.; Sales, M.; Man, C. C.; Shen, J.; Sasmal,
P. K.; Guo, C. J. Org. Chem. 2002, 67, 1618–1629; (d)
Matsukawa, Y.; Dileep, K. J. S.; Yoneyama, Y.; Kata-
giri, T.; Uneyama, K. Tetrahedron: Asymmetry 2000, 11,
521–4528; (e) Soloshonok, V. A.; Cai, C.; Hruby, V. J.;
Van Meervelt, L.; Yamazaki, T. J. Org. Chem. 2000, 65,
688–6696.
. (a) Dakternieks, D.; Schiesser, C. H. Aust. J. Chem. 2001,
4, 89–91; (b) Miyabe, H.; Fujii, K.; Ueda, M.; Naito, T.
Tennen Yuki Kagobutsu Toronkai Koen Yoshishu 2000,
01–605; (c) Villar, F.; Andrey, O.; Renaud, P. Tetra-
hedron Lett. 1999, 40, 3375–3378; (d) Lee, E.; Jeong, J.;
Yu, Y. Tetrahedron Lett. 1997, 38, 7765–7768; (e) Yim,
A.-M.; Vidal, Y.; Viallefont, P.; Martinez, J. Tetrahedron:
Asymmetry 2002, 13, 503–510; (f) Radicals in Organic
Synthesis, Vols. 1–2; Renaud, P.; Sibi, M., Eds.; Wiley-
VCH: Weinheim, 2001; (g) Curran, D. P.; Porter, N. A.;
Giese, B. Stereochemistry of Radical Reactions, Concepts,
4
6
2
5
6

5
34
E. J. Enholm et al. / Tetrahedron Letters 44 (2003) 531–534
1
hydrogens in the chiral centers closer to the same plane
mmol, 59%) of a yellowish oil: H NMR (300 MHz,
compared to the preferred conformation of isomer 7.
CDCl and some drops of D O): l 3.85 (s, 1H), 2.62–2.43
3
2
1
0. These findings suggest that a long distance H–H coupling
(m, 2H), 1.65–1.55 (m, 2H), 1.39–1.27 (m, 2H), 0.98(s,
13
(H1 and H6 in the chiral centers) would be observed for
9H), 0.91 (t, J=7.2, 3H); C NMR (75 MHz, CDCl ): l
3
product 8 rather than product 7 because the hydrogens
are closer to coplanarity. Signals in the NMR spectra
indicate that this long distance coupling is taking place,
i.e. signal at approx. 4.6 Hz corresponding to H1, in the
minor product 8, shows a coupling (doublets) constant of
213.9, 84.0, 41.9, 35.7, 26.4, 26.1, 22.5, 14.0. Anal. calcd
for C H O : C, 69.72; H, 11.70. Found: C: 69.72; H,
11.82%.
10
20
2
5-Bromo-3-hydroxy-2,2-dimethyl-4-octanone (5): A 50 mL,
three-necked round-bottomed flask was charged with 335
mg (1.95 mmol) of 11 which was diluted in 20 mL of
CH Cl . This mixture was stirred under reflux with a
1
.5 Hz. This signal is equivalent to the signal at ca. 3.9
Hz in the major product 7 with no coupling observed.
This observation suggests that a long distance H–H cou-
pling occurs in minor product 8. Since coplanarity is a
requirement for this coupling, the minor product is
assigned as compound 8, while the major product is
assigned as compound 7.
2
2
magnetic stirrer. Acetic acid (5 mL) was added. Bromine
(312 mg, 1.95 mmol) diluted in CH Cl was added drop-
2
2
wise to the heated solution over a period of 30 min. The
reaction was stirred for a further 3 h under reflux. To the
cooled solution was added H O and the organic phase
2
1
1. 2-Hydroxy-3,3-dimethylbutanoic acid (10): To a stirred
was separated. The aqueous phase was extracted three
times with CH Cl . The combined organic phase was
solution of 93.0 g (0.588 mol) of KMnO and 30.0 g (0.75
4
2
2
mol) of NaOH in 720 mL of H O at 0°C was added a
washed three times with H O, saturated aqueous
2
2
slurry of 30.0 g (37.5 mL, 0.300 mol) of 3,3-dimethyl-2-
NaHCO , brine, and dried over anhydrous Na SO . The
3
2
4
butanone (pinacolone) in 400 mL of H O over a 40 min
solvent was removed under reduced pressure to give 477
mg of a yellow oil which was pre-adsorbed onto silica
and purified by flash chromatography (20% ethyl acetate/
hexane) to yield 445 mg (1.77 mmol, 91%) of a light
yellowish oil. Note that 5 must be stored in a frozen
2
period. After the addition, the mixture was stirred for 1 h
at 0°C and for 6 h at room temperature. The reaction
mixture was filtered, and the filtrate was acidified to pH
2
with concentrated HCl and extracted four times with
ether. The organic phase was dried over anhydrous
benzene matrix to avoid decomposition. Data for major
1
Na SO4 and the solvent was removed under reduced
diastereomer: H NMR (300 MHz, CDCl
3
and some
2
pressure to give an oil which was fractionally distilled to
give 28.35 g (0.218 mol, 73%) of a liquid, which upon
drops of D
2
O): l 4.54 (t, J=7.2, 1H), 4.28 (s, 1H),
2.03–1.95 (m, 2H), 1.42–1.37 (m, 2H), 0.99 (s, 9H), 0.96
(t, J=7.5, 3H); C NMR (75 MHz, CDCl ): l 206.9,
3
81.8, 51.3, 37.0, 34.6, 26.9, 21.3, 14.2; IR: 3477.1, 2959.1,
2873.3, 1702.5 cm ; MS (FAB positive): m/z 251.06.
13
standing became a white solid (mp 32°C) 3,3-dimethyl-2-
1
oxobutanoic acid: H NMR (300 MHz, CDCl ): l 10.47
3
13
−1
(
s, 1H), 1.31 (s, 9H); C NMR (300 MHz, CDCl ): l
3
2
01.37, 164.86, 42.49, 25.55. Next, a modification of an
Anal. calcd for C10
47.87; H, 7.97.
H19BrO : C, 47.82; H, 7.62. Found: C:
2
5
established procedure was used to prepare the hydroxy-
acid 10. To a cold (0°C) stirred solution of 10.9 g (83.8
mmol) of 3,3-dimethyl-2-oxobutanoic acid was added
3-Hydroxy-2,2-dimethyl-5-propyl-7-octen-4-one (7). Bro-
mide 5 (1.0 equiv.) was allowed to react with allyl-
tributyltin (2.0 equiv.) in the presence of AIBN (0.1
equiv.) at 80°C using benzene as solvent (to 0.5 M). A
second low-temperature procedure reacted 5 (1.0 equiv.)
in the presence of triethylborane (0.3 equiv.) with or
slowly 3.17 g (83.8 mmol) of NaBH . After a further 2 h
4
of stirring at 0°C, the reaction mixture was concentrated
under reduced pressure and HCl 6N was slowly added to
the residue with cooling until pH 2. The resulting mixture
was extracted four times with ether. The organic phase
without different Lewis acids (MgBr
Yb(CF SO ) at room temperature, at 0°C and at
−78°C, using CH Cl (0.5 equiv.) as solvent. Oxygen was
·OEt , ZnCl or
2
2 2
was washed with H O until the pH of a washing became
3
)
3 3
2
4
, dried over anhydrous Na SO4 and the solvent was
2
2
2
removed under reduced pressure. Recrystallization of the
added by syringe at 20-min intervals. A minimum
amount of dry THF or ether was added in some cases to
solubilize Lewis acids. After completing the reaction
(TLC), the reaction mixture was quenched with concen-
residue from ether–hexane solvent gave 7.39 g (55.9
1
mmol, 67%) of white crystals: H NMR (300 MHz,
13
CDCl ): l 7.56 (s, 2H), 3.93 (s, 1H), 1.01 (s, 9H);
C
3
NMR (75 MHz, CDCl ): l 178.7, 78.3, 35.2, 25.7.
trated NaHCO
rated and extracted three times with CH
was removed under reduced pressure and the residue was
purified by column chromatography using silica with KF
on the top and hexane as eluent to give a yellowish oil
3
solution. The organic phase was sepa-
Cl . The solvent
3
3
-Hydroxy-2,2-dimethyl-4-octanone (11): Under a nitro-
gen atmosphere a hexane solution of 2 M n-BuLi (12 mL,
4 mmol) was added dropwise to a cold (−78°C) stirred
2
2
2
ether solution of the hydroxyacid 10 (0.5 g, 3.78 mmol).
The reaction was allowed to warm to room temperature
and stirred for 5 h. The reaction mixture was slowly
transferred, via cannula, into a rapidly stirred aqueous
HCl solution approximately 1.5N. After extraction with
ether (3×30 mL) the combined organic phase was dried
over anhydrous Na SO , filtered, and the filtrate was
(91% yield). Spectroscopic data for the major compound:
1
H NMR (300 MHz, CDCl ): l 5.77–5.63 (m, 1H),
3
5.07–5.01 (m, 2H), 3.93 (d, J=7.8, 1H), 3.17 (d, J=7.5,
1H), 2.87–2.78 (m, 1H), 2.32–2.16 (m, 2H), 1.69–1.59 (m,
1H), 1.5–1.38 (m, 1H), 1.34–1.24 (m, 1H), 1.00 (s, 9H),
0.91 (t, J=7.2, 3H); C NMR (75 MHz, CDCl
217.2, 134.9, 118.2, 83.6, 50.9, 37.6, 36.4, 31.8, 26.8, 21.1,
13
): l
3
2
4
concentrated on a rotary evaporator. The residue was
pre-adsorbed onto silica and chromatographed on flash
silica gel (15% ethyl acetate/hexane) to give 381 mg (2.21
.
+
14.8; MS m/z 213.188 [M +1]. Anal. calcd for C13H O :
C, 73.53; H, 11.39. Found: C: 73.86; H, 11.14%.
24 2