An effective method for allylic oxidation of Δ5-steroids using tert-butyl hydroperoxide

Article abstract of DOI:10.1021/jo902637k

"Chemical equation presented" An allylic oxidation method for Δ⁵-steroids using TBHP as oxidant with a 2-quinoxalinol salen Cu(II) complex as catalyst is reported. A variety of Δ⁵- steroidal substrates are selectively oxidized to the corresponding enones. Excellent yields are achieved (up to 99% under optimized conditions) while significantly reducing reaction times required as compared to other current methods.

Full text of DOI:10.1021/jo902637k

pubs.acs.org/joc
synthetic modifications of steroids,4,11-15 complications in
An Effective Method for Allylic Oxidation of
Δ5-Steroids Using tert-Butyl Hydroperoxide
applying these methods remain because of the harsh reaction
conditions required and difficult workup and/or purification
procedures. In addition, the accumulations of chromic acid
or chromium salt wastes that are the side products of these
reactions are of great environmental concern.16 To complete
such modifications in a more environmentally friendly yet still
efficacious manner, many other metal complexes/salts have
been employed, such as sodium chlorite,17 copper iodide,5
dirhodium caprolactamate,18 ruthenium trichloride,19 bismuth
salt,20 cobalt acetate,21 palladium(II) salts,22 and manganese-
(III) acetate;23 however, numerous limitations remain. All of
the available methods must strike a balance between good
yields and functional group compatibility while continuing to
suffer from long reaction time requirements. Oxidations using
manganese(III) acetate and dirhodium caprolactamate as
catalysts are two more recent representative examples of
the current methods.17,21 The use of manganese(III) acetate
allowed for excellent yields in Δ5-steroidal oxidation under
ambient temperature when tert-butyl hydroperoxide was used
as oxidant. The reaction times were reduced remarkably when
the reaction mixture was heated, but with a loss in yield. This
method also does not exhibit compatibility with sensitive
functionalities near allylic sites (e.g., hydroxyl group).21 The
use of the dirhodium caprolactamates as catalysts in such an
allylic oxidation showed a wide tolerance for a variety of
functional groups, and yet it also had decreased yields.17
Yuancheng Li, Xianghong Wu, Tae Bum Lee, Eleanor K. Isbell,
Edward J. Parish, and Anne E. V. Gorden*
Department of Chemistry and Biochemistry, College of
Science and Mathematics, Auburn Univeristy, Auburn,
Alabama 36849-5319
gordeae@auburn.edu
Received December 27, 2009
An allylic oxidation method for Δ5-steroids using TBHP
as oxidant with a 2-quinoxalinol salen Cu(II) complex as
catalyst is reported. A variety of Δ5-steroidal substrates
are selectively oxidized to the corresponding enones.
Excellent yields are achieved (up to 99% under optimized
conditions) while significantly reducing reaction times
required as compared to other current methods.
Because of the biological and physiological properties
of their 7-keto-Δ5-steroid oxidation products, Δ5-steroids
have attracted attention as highly useful synthetic building
blocks.1-10 Easy access to the 7-keto-Δ5-steroids can be
achieved via allylic oxidation of Δ5-steroids. Although a
variety of chromium(VI) compounds have been used in the
Previously, we have reported a new catalytic system that
consists of 2-quinoxalinol salen copper(II) complex 2 as a
catalyst using tert-butyl hydroperoxide (TBHP) as an oxidant.24
The 2-quinoxalinol salen copper(II) system was demonstrated
(1) Guardiola, F.; Codony, R.; Addis, P. B.; Rafecas, M.; Boatella, J.
Food Chem. Toxicol. 1996, 34, 193–211.
(2) Schroepfer, G. J. J. Physiol. Rev. 2000, 80, 361–544.
(3) Guardiola, F.; Dutta, P. C.; Codony, R.; Savage, G. P. Cholesterol and
Phytosterol Oxidation Products: Analysis, Occurrence, and Biological Effects;
AOCS Press: Champaign, IL, 2002.
(12) Kasal, A. Tetrahedron 2000, 56, 3559.
(13) Bora, U.; Chaudhuri, M. K.; Dey, D.; Kalita, D.; Kharmawphlang,
W.; Mandal, G. C. Tetrahedron 2001, 57, 2445.
(14) Salvador, J. A. R.; Silvestre, S. M.; Moreira, V. M. Curr. Org. Chem.
2006, 10, 2227–2257.
(4) Parish, E. J.; Kizito, S. A.; Qiu, Z. Lipids 2004, 39, 801–804.
(5) Arsenou, E. S.; Koutsourea, A. I.; Fousteris, M. A.; Nikolaropoulos,
S. S. Steroids 2003, 68, 407–414.
(6) Lee, J. W.; Fuda, H.; Javitt, N. B.; Strott, C. A.; Rodriguez, I. R. Exp.
Eye Res. 2006, 83, 465.
(7) Berthier, A.; Lemarie-Ewing, S.; Prunet, C.; Monier, S.; Athias, A.;
Bessade, G.; Paisde Barros, J.-P.; Laubriet, A.; Gambert, P.; Lizard, G.;
Nacel, D. Cell Death Differ. 2004, 11, 897.
(15) Liu, J.; Zhu, H. Y.; Cheng, X. H. Synth. Commun. 2009, 39, 1076–
1083.
(16) Manahan, S. E. Environmental Chemistry; 7th ed.; Lewis Pub.: Boca
Raton, FL, 2000.
(17) Silvestre, S. M.; Salvador, J. A. R. Tetrahedron 2007, 63, 2439.
(18) Choi, H.; Doyle, M. P. Org. Lett. 2007, 9, 5349–5352.
(19) Miller, R. A.; Li, W.; Humphrey, G. R. Tetrahedron Lett. 1996, 37,
3429–3432.
(8) Deckert, V.; Duverneuil, L.; Poupon, S.; Monier, S.; de Guen, N.;
Lizard, G.; Masson, D.; Lagrost, L. Br. J. Pharmacol. 2002, 137, 655.
(9) Steffen, Y.; Wiswedel, I.; Peter, D.; Schewe, T.; Sies, H. Free Radicals
Biol. Med. 2006, 41, 1139.
(10) Javitt, N. B. Curr. Opin. Lipidol. 2007, 18, 283.
(11) Parish, E. J.; Sun, H.; Kizito, S. A. J. Chem. Res. Synop. 1996, 12,
544–545.
(20) Salvador, J. A. R.; Silvestre, S. M. Tetrahedron Lett. 2005, 46, 2581–
2584.
(21) Salvador, J. A. R.; Clark, J. H. Chem. Commun. 2001, 33-34, 33.
(22) Yu, J.-Q.; Corey, E. J. Org. Lett. 2002, 4, 2727.
(23) Shing, T. K. M.; Yeung, Y.-Y.; Su, P. L. Org. Lett. 2006, 8, 3149–
3151.
(24) Wu, X.; Gorden, A. E. V. Eur. J. Org. Chem. 2009, 503–509.
DOI: 10.1021/jo902637k
r
Published on Web 02/08/2010
J. Org. Chem. 2010, 75, 1807–1810 1807
2010 American Chemical Society

Li et al.
JOCNote
TABLE 1. Oxidation of Steroids Using Optimized Condition
entry
substrate
yield (%)
1
2
3
4
5
6
7
R1 = OAc, R2 = Ac, R3 = H
R1 = OAc, R2 = C8H17, R3 = H
R1 = OBz, R2 = Ac, R3 = H
R1 = OTHP, R2 = C8H17, R3 = H
R1 = OH, R2 = C8H17, R3 = H
R1 = OH, R2 = Ac, R3 = H
R1 = Cl, R2 = C8H17, R3 = H
64
62
55
61
44
53
49
FIGURE 1. Isolated yields with reaction time and isolated yields
with oxidant ratio.
to oxidize benzylic methylenes into carbonyl groups in near-
quantitative yields with the use of catalyst loading of 1 mol %
and 3 equiv of TBHP when different functional groups
remained intact. On the basis of our previous success in this
related study, we sought to modify this system to develop mild
conditions that would retain satisfying yields, efficacious for
allylic oxidations.
slowly decompose in most organic solvents25 and the observa-
tion that the color of solution changes from its initial red to
yellowish, it would appear that the ligand-metal complex 1
was consumed as the reaction proceeded.
To verify this hypothesis, we carried out the oxidation of
cholesteryl chloride in CH3CN and CHCl3 independently.
Interestingly, the reaction in CH3CN had an isolated yield of
57% while the one in CHCl3 only had 28% yield, despite the
fact that complex 1 has much better solubility in CHCl3 than
in CH3CN. According to the consensus radical mechanism
of current allylic oxidation methods using TBHP,18,23,26 the
reason behind the low yield is likely stabilizers present in the
CHCl3 that can function as radical scavengers. The reaction
was then carried out again with CDCl3 as solvent, achieving
48% yield. Although a negative factor was found, the
mystery of complex 1 appearing to be consumed over the
course of the reaction remained.
It was thus decided to change the reaction procedure
conditions to limit the exposure time of the catalyst, so in
additional experiments complex 1 and TBHP were added in
portions to the reaction mixture every 2 h. With this further
optimization, new reaction conditions were developed in
which CH3CN heated to 70 °C was used as solvent, and
0.5 mol % of complex 1 and 5 equiv of TBHP were added
simultaneously with multiple portions used.
For each substrate, different numbers of portions were
tested to find optimal conditions. Cholesteryl acetate needs 3
portions (each addition contains complex 1 (1.8 mg, 0.0025
mmol) and t-BuOOH (0.5 mL, 2.5 mmol)) to achieve 99%
yield (Table 2, entry 1); benzoyl protected pregnenolone
required 5 portions to get 77% yield (Table 2, entry 2).
Simple cholesterol only needs 2 portions (Table 2, entry 3)
while pregnenolone and cholesteryl chloride need 4 portions
(Table 2, entries 4 and 5). Generally excellent yields were
obtained albeit unprotected substrates have generally lower
yields than the protected ones. The addition of subsequent
portions of catalyst and oxidant did not increase yields. The
benzoyl protected cholesterol is oxidized in significantly
lower yield than other protected steroids although 5 portions
were added due to the insolubility of substrate in CH3CN.
We also performed a theoretical study and computational
modeling to investigate the origins of regioselectivity.
We began by using 2-quinoxalinol salen copper(II) com-
plex 1 as catalyst and pregnenolone acetate 3 as substrate
(eq 1). Two factors were varied to determine the optimal
reaction conditions: reaction time and oxidant ratio. The
oxidation was performed on a millimole scale using 1 mol %
of catalyst complex 1. CH3CN/CHCl3 (50/50) was chosen as
solvent to dissolve steroids. The solution was heated to 70 °C,
and then the TBHP was added. When the reaction was
completed, the solution was concentrated under reduced
pressure, and the residue was purified by flash column
chromatography using hexane/ethyl acetate as eluent. Initi-
ally, we used 10 equiv of TBHP to determine the optimal
reaction time (Figure 1). The isolated yields slightly
increased when the reaction time was extended from 3 to
12 h, and did not change when the time was elongated from
12 to 16 h. There is little difference in isolated yields between
6 and 16 h, so we set the time to 6 h to optimize the oxidant
ratio (Figure 1).
A generally increasing trend of isolated yields was
observed as the oxidant ratio was increased from 3 equiv to
20 equiv. The highest yield was achieved when 10 equiv of
TBHP was used, and the additional amount of oxidant did
not result in a higher yield. It is notable that the remaining
starting material, pregnenolone acetate, was recovered after
this time. When carried out at ambient temperature, the
reaction rate was found to be much slower (51% yield for
24 h and 65% for 48 h). In the absence of complex 1, only
39% yield was obtained after heating 6 h at 70 °C.
A variety of Δ5-steroidal substrates were oxidized using
the optimized condition with 1 mol % of catalyst loading and
10 equiv of oxidant TBHP heated to 70 °C for 12 h (Table 1).
The even more challenging oxidation of 3-hydroxy-Δ5-steroids
can also be successfully realized with a small decrease in yields
(entry 5 and 6); however, in all cases, unreacted steroidal
starting materials were recovered. Given the fact that structu-
rally similar salen copper(II) complexes have been shown to
(25) Pratt, R. C.; Stack, T. D. P. J. Am. Chem. Soc. 2003, 125, 8716–8717.
(26) McLaughlin, E. C.; Choi, H.; Wang, K.; Chiou, G.; Doyle, M. P.
J. Org. Chem. 2009, 74, 730–738.
1808 J. Org. Chem. Vol. 75, No. 5, 2010

Li et al.
JOCNote
TABLE 2. Oxidation of Steroids Using Further Optimized Condition
becomes more significant (6.21 kcal/mol) when calculations
with the more complicated steroid molecule (pregnenolone
acetate 3) were performed with the same level of theory. A
series of calculations were carried out toward the model
compound 4 bearing different functional groups. In each
case, compounds with the radical at the 7-position were more
stable than those at the 4-position. It is interesting that the
trend of energy difference bearing different functional
groups is in agreement with the trend of numbers of portions
needed for the corresponding substrates (see Table 2, entries
1, 3, and 5).
In summary, we have demonstrated a novel system for
allylic oxidation of Δ5-steroids. This method is tolerant
with a variety of functional groups, has low sensitivity to
air or water, and excellent yields (up to 99%) can be
realized.
entry
susbtrate
yield (%)
1
2
3
4
5
R1 = OAc, R2 = C8H17, R3 = H
R1 = OBz, R2 = Ac, R3 = H
R1 = OH, R2 = C8H17, R3 = H
R1 = OH, R2 = Ac, R3 = H
R1 = Cl, R2 = C8H17, R3 = H
97a
77b
69c
88d
99d
a3 portions of TBHP and complex 1 were used. b5 portions of TBHP
and complex 1 were used and 1 mL of CHCl3 was added to help dissolve
the substrate. c2 portions of TBHP and complex 1 were used. d4 portions
of TBHP and complex 1 were used.
Experimental Section
Synthesis of 2-Quinoxalinol Salen Copper(II) Complex 1. The
2-quinoxalinol salen ligand was synthesized following a pub-
lished procedure.27 In a 100 mL round-bottomed flask charged
with a stirrng bar, 0.20 mmol of 2-quinoxalinol salen ligand
(133 mg) and 0.24 mmol of Cu(OAc)2 H2O (48 mg) were
3
dissolve with 20 mL of DCM and 20 mL of methanol. After
1.2 mmol of triethylamine (0.20 mL) was added, the reaction
was stirred for 2 h at refluxing temperature. After solvent was
removed, the resulting dark red solid was washed with water and
cold ether 3 times each. A product of 127 mg of solid was
obtained (87%).
FIGURE 2. Resonance structures of radical species.
2-Quinoxaninol Salen Copper(II) Complex 1. IR (KBr): 3429
(br), 2955, 2909, 2868, 1661, 1556, 1524, 1495, 1462, 1423, 1385,
1202 cm-1. UV-vis (CHCl3) 250 (ε = 35 600 M-1 cm-1), 285
(ε = 27 140 M-1 cm-1), 330 (ε = 31 340 M-1 cm-1), 460 nm
(ε = 39 280 M-1cm-1). Formula (M þ H): C42H55N4O3Cu.
HRMS: found 726.3575 (M þ H), calcd 726.3570 (M þ H).
Representative Procedure for Allylic Oxidation Using Opti-
mized Condition (Table 1). To a 50 mL round-bottomed flask
charged with a stirring bar, were sequentially added pregneno-
lone acetate (3) (358 mg, 1 mmol), complex 1 (7.3 mg, 0.01
mmol), CH3CN (10 mL), CHCl3 (10 mL), and t-BuOOH (2 mL,
10 mmol). After the reaction was stirred at 70 °C for 12 h, solvent
was removed under reduced pressure. The residue was purified
by flash column chromatography with hexane/ethyl acetate
(5:1) as eluent to yield product as a white solid (239 mg).
3β-Acetoxypregn-5-ene-7,20-dione (Entry 1, Table 1). IR
TABLE 3. Calculations of Free Energies of Model Radicals
aValue reported as ΔG at 298K.
1
(solid): 1730, 1707, 1672 cm-1. H NMR (400 MHz, CDCl3):
For Δ5-steroids, the two allylic sites do not have much
difference in reactivity at first glance; however, only 7-keto
products were obtained exclusively. From the resonance
structures, the difference becomes clear (Figure 2). Species
bearing a radical at the 7-position will have two possible
resonance structures, one of which is a tertiary radical that
can significantly lower the energy, while the species bearing a
radical at the 4-position also has two possible resonance
structures, but neither contributes to the lower energy state.
Of course, the preference of reactivity is also affected by the
conformation of the sterol.
This hypothesis was supported by ab initio computa-
tional studies at the B3LYP/6-31G(2d,p) level using the
Gaussian03 package (Table 3). Compound 4 was selected
as a model to examine the energy difference between the
species bearing radicals at two different sites. The free energy
at 298 K of the 7-position radical is less than that of the
4-position radical by 3.43 kcal/mol. This relative stability
δ 5.72 (d, J = 1.6 Hz, 1H), 4.75-4.69 (m, 1H), 2.59-1.26 (comp,
18H), 2.12 (s, 3H), 2.05 (s, 3H), 1.22 (s, 3H), 0.64 (s, 3H). 13C
NMR (400 MHz, CDCl3): δ 209.7, 201.2, 170.3, 164.3, 126.4,
72.0, 62.2, 49.9, 49.6, 45.2, 44.4, 38.3, 37.7, 37.6, 35.9, 31.6, 27.2,
26.4, 23.5, 21.2, 21.0, 17.2, 13.2. 1H, 13C NMR, and IR data were
in accordance with literature values.18
3β-Acetoxycholest-5-ene-7-one (Entry 2, Table 1). IR (solid):
1728, 1671 cm-1. 1H NMR (400 MHz, CDCl3): δ 5.70 (d, J =
1.6 Hz, 1H), 4.76-4.67 (m, 1H), 2.58-1.00 (comp, 26H), 2.05 (s,
3H), 1.21 (s, 3H), 0.92 (d, J = 6.5 Hz, 3H), 0.87 (d, J = 6.6 Hz,
3H), 0.86 (d, J = 6.6 Hz, 3H), 0.68 (s, 3H). 13C NMR (400 MHz,
CDCl3): δ 201.9, 170.2, 164.0, 126.6, 72.2, 54.7, 49.9, 49.8, 45.4,
43.1, 39.4, 38.6, 38.3, 37.6, 36.2, 36.0, 35.7, 28.5, 28.0, 27.3, 26.3,
23.8, 22.8, 22.6, 21.2, 21.1, 18.8, 17.2, 11.9. 1H, 13C NMR, and
IR data were in accordance with literature values.18
3β-Benzoyloxypregn-5-ene-7,20-dione (Entry 3, Table 1). IR
(solid): 1707, 1672 cm-1. 1H NMR (400 MHz, CDCl3): δ 8.04 (d,
(27) Wu, X.; Gorden, A. E. V. J. Comb. Chem. 2007, 9, 601–608.
J. Org. Chem. Vol. 75, No. 5, 2010 1809

Li et al.
JOCNote
J = 7.2 Hz, 2H), 7.57 (t, 1H), 7.45 (t, 2H), 5.77 (d, J = 1.6 Hz,
1H), 5.02-4.94 (m, 1H), 2.78-0.68 (comp, 24H), 2.14 (s, 3H),
1.27 (s, 3H), 0.68 (s, 3H). 13C NMR (400 MHz, CDCl3): δ 209.9,
201.4, 166.0, 164.4, 133.2, 130.4, 130.3, 129.7, 128.5, 126.8, 72.8,
62.4, 50.1, 49.8, 45.4, 44.6, 38.6, 38.0, 37.8, 36.2, 31.8, 27.6, 26.6,
23.8, 21.3, 17.5, 13.4. 1H, 13C NMR, and IR data were in
accordance with literature values.18
3β-[(Tetrahydropyran-2R,S-yl)oxy]-7-oxo-cholest-5-ene (Entry 4,
Table 1). IR (solid): 1676, 1630 cm-1. 1HNMR(400MHz, CDCl3):
δ 5.68 (d, J = 1.0 Hz, 1H), 4.72 (s, 1H), 3.55-3.64 (m, 1H),
3.45-3.49 (m, 2H), 0.93 (s, 3H), 0.90 (d, J = 6.6 Hz, 3H), 0.85 (d,
J = 6.6 Hz, 6H), 0.68 (s, 3H). 13C NMR (400 MHz, CDCl3): δ
202.2, 165.6, 126.0, 97.4, 75.0, 62.9, 54.8, 50.0, 45.4, 43.1, 40.1, 39.5,
38.8, 38.5, 36.5, 36.2, 31.2, 29.4, 28.5, 28.0, 27.7, 26.3, 25.4, 23.8,
22.6, 21.2, 19.9, 19.0, 17.4, 12.0. 1H, 13C NMR, and IR data were in
accordance with literature values.28
3β-Hydroxycholest-5-ene-7-one (Entry 5, Table 1). IR (solid):
1670 cm-1. 1H NMR (400 MHz, CDCl3): δ 5.69 (d, J = 1.6 Hz,
1H), 3.71-3.64 (m, 1H), 2.54-1.00 (comp, 26H), 1.21 (s, 3H),
0.93 (d, J = 6.6 Hz, 3H), 0.88 (d, J = 6.6 Hz, 3H), 0.87 (d, J =
6.6 Hz, 3H), 0.68 (s, 3H). 13C NMR (400 MHz, CDCl3): δ 202.6,
165.6, 126.0, 70.5, 54.8, 50.0, 49.9, 45.4, 43.1, 41.8, 39.5, 38.7,
38.3, 36.4, 36.2, 35.7, 31.1, 28.6, 28.0, 26.3, 23.8, 22.8, 22.6, 21.2,
18.9, 17.3, 12.0. 1H, 13C NMR, and IR data were in accordance
with literature values.18
3β-Hydroxypregn-5-ene-7,20-dione (Entry 6, Table1 1). IR
(solid): 1681, 1666 cm-1. 1H NMR (400 MHz, CDCl3): δ 5.68
(d, J = 1.6 Hz, 1H), 3.67-3.63 (m, 1H), 2.52-1.18 (comp, 19H),
2.10 (s, 3H), 1.18 (s, 3H), 0.63 (s, 3H). 13C NMR (400 MHz,
CDCl3): δ 210.0, 201.8, 166.0, 125.8, 70.3, 62.3, 50.0, 49.8, 45.1,
44.4, 41.8, 38.4, 37.7, 36.4, 31.6, 31.0, 26.5, 23.6, 21.1, 17.3, 13.3.
1H, 13C NMR, and IR data were in accordance with literature
values.18
3β-Chlorocholest-5-ene-7one (Entry 7, Table 1). IR (solid):
1669 cm-1 1H NMR (400 MHz, CDCl3): δ 5.68 (s, 1H),
.
3.86-3.82 (m,1H), 2.71 (d, J = 8.3 Hz, 2H), 2.43-2.36 (m,
1H), 2.26-2.15 (comp, 2H), 2.04-1.88 (comp, 4H), 1.57-1.05
(comp, 20H), 1.22 (s, 3H), 0.92 (d, J = 6.5 Hz, 3H), 0.87 (d, J =
6.6 Hz, 3H), 0.86 (d, J = 6.6 Hz, 3H), 0.68 (s, 3H). 13C NMR
(400 MHz, CDCl3): δ 201.9, 163.9, 126.2, 57.8, 54.8, 49.9, 49.8,
45.4, 43.1, 42.6, 39.5, 38.6, 38.1, 36.2, 35.7, 32.8, 28.5, 28.0, 26.3,
23.8, 22.8, 22.6, 21.1, 18.9, 17.2, 12.0. 1H, 13C NMR, and IR data
were in accordance with literature values.18
Representative Procedure for Allylic Oxidation Using Further
Optimized Condition (Table 2). To a 50 mL round-bottomed flask
charged with a stirring bar were sequentially added cholesteryl
chloride (203 mg, 0.5 mmol), complex 1 (1.8 mg, 0.0025 mmol),
CH3CN (10 mL), and t-BuOOH (0.5 mL, 2.5 mmol). Additional
complex 1 (1.8 mg, 0.0025 mmol) and t-BuOOH (0.5 mL,
2.5 mmol) were added three times every 2 h. Solvent was then
removedunderreducedpressure. The residuewaspurifiedby flash
column chromatography with hexane/ethyl acetate (15:1) as
eluent to yield product as a white solid (201 mg).
Acknowledgment. We thank Dr. Peter Livant and
Dr. Xiaoxun Li for helpful suggestions with NMR spectro-
scopy, and gratefully acknowledge funding for this project
from Auburn University.
Supporting Information Available: Procedures, data from
theoretical calculation, and 13C NMR spectra of the of 7-keto-
Δ5-steroid compounds produced (see Tables S1and S2). This
material is available free of charge via the Internet at http://
pubs.acs.org.
(28) Choucair, B.; Dherbomez, M.; Roussakis, C.; Kihel, L. E. Tetra-
hedron 2004, 60, 11477.
1810 J. Org. Chem. Vol. 75, No. 5, 2010