Regioselective oxidations of equilenin derivatives catalyzed by a rhodium(III) porphyrin complex - Contrast with the manganese(III) porphyrin

Article abstract of DOI:10.1016/S0040-4039(00)01402-7

Equilenin acetate and dihydroequilenin acetate were oxidized with iodosobenzene and a rhodium(III) porphyrin catalyst. The selectivity of the reactions differs from that with the corresponding Mn(III) catalyst, or from that of free radical chain oxidation. (C) 2000 Published by Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01402-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 8063–8067
Regioselective oxidations of equilenin derivatives catalyzed
by a rhodium(III) porphyrin complex—contrast with the
manganese(III) porphyrin
Jerry Yang and Ronald Breslow*
Department of Chemistry, Columbia University, New York, NY 10027, USA
Received 27 July 2000; revised 14 August 2000; accepted 16 August 2000
Abstract
Equilenin acetate and dihydroequilenin acetate were oxidized with iodosobenzene and a rhodium(III)
porphyrin catalyst. The selectivity of the reactions differs from that with the corresponding Mn(III)
catalyst, or from that of free radical chain oxidation. © 2000 Published by Elsevier Science Ltd.
We have described the oxidation and amination reactions of equilenin acetate 1 catalyzed by
a manganese porphyrin 2, carrying four pentafluorophenyl substituents, in methylene chloride
1
solution. Oxidation, using iodosobenzene as the reagent, afforded the C6 phenol 3 as the major
product (74%) while the axial C11 benzylic alcohol 4 was a minor product (26%). No
hydroxylation occurred at the other benzylic position, C14, and even with 25 catalytic turnovers
the alcohol 4 was not oxidized to a ketone. With 57% conversion of starting material, only these
two products were formed in quantitative overall yield. With the same catalyst 2, tosylamidation
of 1 using PhꢀIꢁNꢀTs afforded only the axial tosylamide 5, with no attack on the naphthalene
ring (with varying amounts of moisture, the MnꢁNTs intermediate partially hydrolyzed to
MnꢁO and gave some of the oxygenation products). We invoked an explanation in which the
MnꢁO intermediate performs oxygen atom transfer to the naphthalene ring to form the phenol,
while the benzylic hydroxylation and tosylamidation reactions involve hydrogen abstraction to
form a benzylic radical that is then captured with stereoelectronic control to form axial
products.
To explore these interesting selectivities further, we have examined related reactions catalyzed
2
,3
by the rhodium(III) analog 6 of catalyst 2. No phenol was formed on oxidation of 1 with one
4
equivalent of iodosobenzene and 0.001 equivalents of catalyst 6, only benzylic ketone 7 (43%)
5
and alcohol 8 (57%) in quantitative overall yield with ca. 40% conversion. Neither of these
*
Corresponding author.
0
040-4039/00/$ - see front matter © 2000 Published by Elsevier Science Ltd.
PII: S0040-4039(00)01402-7

8064
products had been formed with catalyst 2. We also examined the oxidation of dihydroequilenin
6
7
diacetate 9 with iodosobenzene and rhodium catalyst 6. Again an 11-keto product 10 was
formed, but not the 14-hydroxy compound. In contrast to our findings with the manganese
catalyst 2, the rhodium catalyst 6 did not catalyze any amidation of the steroid with PhꢀIꢁNꢀTs,
but instead the catalyst was rapidly destroyed by the reagent.
Rhodium(III) porphyrins are known to catalyze carbonꢀcarbon bond formations between
8
–10
carbenes (generated from diazo compounds) and olefins to form cyclopropanes.
They also
11
catalyze the insertion of carbenes into the CꢀH bonds of saturated compounds. In both cases
the substrates were used as solvent. We examined the reaction of ethyl diazoacetate with a
concentrated solution of 1 in CHCl with catalyst 6, but under anaerobic conditions there was
3
no reaction. We found that ethyl diazoacetate in 2-methylnaphthalene solution did form two
1
2,13
cyclopropane products, 11 and 12.
Presumably the problem with 1 was that it was diluted in
solvent, not neat. However, when oxygen was present the same reaction conditions with 1
1
4,15
afforded the two isomeric 14-hydroxy steroids, 8 and 13,
in a 1:1 ratio. Under the same
16
conditions substrate 9 afforded only the 14-a-hydroxy steroid 14. In both cases no oxidation
occurred when the diazo compound was omitted.
Some sense can be made of these varied reactions. Apparently manganese catalyst 2 forms a
phenol 3 by oxygen addition from the MnꢁO intermediate to the naphthalene ring, the
17
dominant reaction. In contrast, the presumed RhꢁO intermediate formed from rhodium
catalyst 6 is selective for benzylic oxidation, not oxygen transfer. Also, catalyst 2 does not

8065
oxidize the benzylic alcohol 4 to a ketone, presumably because it cannot remove the equatorial
hydrogen at C11 in 4, but rhodium catalyst 6 oxidizes that alcohol more readily than it oxidizes
the substrate, so only the ketones 7 and 10 are seen as C11 products.
Two different mechanisms for benzylic oxidation seem to be operating. The hydroxylations at
C14 with ethyl diazoacetate in air must be free radical chain oxidations, and indeed we find that
heating substrate 9 in benzene under reflux with azo-bis-isobutyronitrile in air without catalysts
afforded the same 14-a-hydroxy product 14, with no identified product from oxidation at C11.
A presumed intermediate hydroperoxide must be converted to this carbinol, probably on
workup or during chromatographic isolation. However, in our catalytic oxidations of 1 with
catalyst 2 and of substrate 9 with both catalysts we see benzylic reaction at C11, not C14.
This contrast may reflect a steric difference between hydrogen atom removal by the small
oxygen radical intermediate in the oxygen reactions and the bulky MꢁO intermediate in the
catalyzed reactions. However, another possibility is that some of the metalꢀporphyrin catalyzed
benzylic oxidations may involve electron transfer from the naphthalene system, followed by
proton loss from the benzylic position to yield a benzylic radical. Such a mechanism can make
sense of the finding that hydroxylation by catalyst 6 occurs preferentially at C14 in the ketone
1
, while it occurs exclusively at C11 with substrate 9. By this mechanism, product is determined
by the preferential loss in the cation radical of 1 of the more acidic C14 proton, which is nearer
the carbonyl dipole, but without the ketone group the cation radical of 9 deprotonates at C11.
Regardless of the detailed reasons for these results, the contrast in behavior of the two
different catalysts 2 and 6 is striking. Also, it is clear that the two catalysts have different useful
selectivities. With equilenin acetate 1 the manganese catalyst produces amidation only at C11, in

8066
product 5. With dihydroequilenin acetate 9, the rhodium catalyst directs selective oxidation to
form the 11-keto derivative 10, along with some oxidative decomposition products. Thus for
selective functionalizations in the equilenin series, both catalysts have a role.
Acknowledgements
This work has been supported by the NIH and the NSF. J.Y. acknowledges support from a
Bristol-Myers Squibb Graduate Fellowship and an EPA NCERQA STAR Graduate
Fellowship.
References
1
2
. Yang, J.; Weinberg, R.; Breslow, R. Chem. Commun. 2000, 531–532.
. Rhodium porphyrin 6 was synthesized using a similar procedure to that reported by Wayland et al. A 0.24 mM
3
solution of 5,10,15,20-tetrakis(pentafluorophenyl)porphyrin(TFPP-H ) in 100 ml of 1,2-dichloroethane was
2
charged with 1.83 mmol NaOAc under argon. Rh (CO) Cl (0.32 mmol) dissolved in 20 ml of 1,2-dichloroethane
2
4
2
was added dropwise to the porphyrin solution under argon. The solution was heated at reflux under argon for
8 h. After cooling to room temperature, 90 mg of I was added and the solution was stirred for 5 h. The solution
4
2
was then concentrated and filtered to remove the inorganic salts. The product was isolated by column
1
chromatography using chloroform as eluent to give the bright red RhTFPPI in 94% overall yield. H NMR (400
19
MHz, CDCl ): l 8.99 ppm (singlet). F NMR (300 MHz, CDCl ): l (ppm) −135 (2F, m, ortho to porphyrin),
3
3
−
150 (1F, t, para to porphyrin), −160 (2F, m, meta to porphyrin). UV–vis (CH Cl ): u 414 (Soret), 543, 577 nm.
2 2
3
4
. Wayland, B. B.; Ba, S.; Sherry, A. E. J. Am. Chem. Soc. 1991, 113, 5305–5311.
1
. Compound 7: H NMR (400 MHz, CDCl ): l 9.36 (1H, d, C6-H), 8.00 (1H, d, C1-H), 7.55 (1H, d, C4-H), 7.46
3
(1H, d, C7-H), 7.40 (1H, dd, C2-H), 3.44 (1H, dd, C14-H), 2.97 (1H, d, C12-H), 2.74 (1H, d, C12-H), 2.67–2.48
(3H, m, C15-H and C16-H’s), 2.38 (3H, s, acetate Me), 2.18 (1H, m, C15-H), 0.85 (3H, s, C18-Me). ESI-MS: m/z
3
23 (M+1). Product 7 indicated a loss of both C11 protons present in starting material 1. The two protons at C12
were shifted from 2.23 and 1.92 ppm in 1 to 2.97 and 2.74 ppm in 7. COSY spectra showed that the C12 protons
in 7 had no couplings other than geminal coupling. HSQC spectra indicated that both of the assigned protons
at C12 were attached to the same carbon. COSY also indicated that the C14, C15, and C16 H’s present in 1 were
all still present in 7. Finally, the large shift of the C6-H at l 7.68 in 1 to 9.36 in 7 could be attributed to a
deshielding effect from a conjugated ketone at the C11 position.
1
5
. Compound 8: H NMR (500 MHz, CDCl ): l 7.98 (1H, d, C1-H), 7.80 (1H, d, C7-H), 7.74 (1H, d, C6-H), 7.55
3
(1H, d, C4-H), 7.29 (1H, dd, C2-H), 3.15 (2H, m, C11-H’s), 2.59 (1H, m, C16-H), 2.50 (1H, m, C16-H), 2.36 (3H,
s, acetate Me), 2.34 (2H, m, C15-H’s), 2.05 (1H, m, C12-H), 1.81 (1H, m, C12-H), 1.18 (3H, s, C18-Me).
1
FAB-MS: m/z 325 (M+1), 307 (M−H O+1). H NMR, COSY, and HSQC spectra of product 8 indicated that all
2
carbons and protons present in 1 were retained in 8 except for the C14 proton. The downfield shifts of the C7
proton from 7.32 ppm in 1 to 7.80 in 8 as well as the angular methyl group from 0.80 ppm in 1 to 1.18 ppm in
8
2
are consistent with oxidation at the C14a position. Furthermore, the downfield shift of the C15a proton from
.04 ppm in 1 to 2.34 ppm in 8 was consistent with a C14a hydroxylation.
6
7
. Compounds 1 and 9 were synthesized from the commercially available equilenin and dihydroequilenin
Steraloids, Inc.) by acylation with acetic anhydride in pyridine under standard conditions. All products were
(
1
isolated by column chromatography and characterized by H NMR, COSY, HSQC and MS.
1
. Compound 10: H NMR (500 MHz, CDCl ): l 9.44 (1H, d, C6-H), 7.98 (1H, d, C1-H), 7.57 (1H, d, C4-H), 7.37
3
(
1H, dd, C2-H), 7.34 (1H, d, C7-H), 5.03 (1H, t, C17-H), 3.24 (1H, dd, C14-H), 2.94 (1H, d, C12-H), 2.70 (1H,
d, C12-H), 2.51 (1H, m, C16-H), 2.37 (1H, m, C15-H), 2.36 (3H, s, C3-acetate), 2.10 (3H, s, C17-acetate), 2.01
1H, m, C15-H), 1.80 (1H, m, C16-H), 0.79 (3H, s, C18-Me). ESI-MS: m/z 367 (M+1). Similar characterization
(
and arguments were made to identify compound 10 as for compound 7.
. Callot, H.; Schaeffer, E. Nouv. J. Chem. 1980, 4, 311–314.
8

8067
9. Callot, H.; Piechocki, C. Tetrahedron Lett. 1980, 21, 3489–3492.
1
1
1
0. Brown, K.; Kodadek, T. J. Am. Chem. Soc. 1992, 114, 8336–8338.
1. Callot, H.; Metz, F. Tetrahedron Lett. 1982, 23, 4321–4324.
1
2. Compounds 11 and 12 were produced as a 1:1 mixture. H NMR (500 MHz, CDCl ): l 7.33–6.95 (3H, m, C1-H,
3
C2-H, and C4-H), 6.36 (1H, m, vinyl H adjacent to the aromatic ring), 6.25 (1H, m, H adjacent to the
cyclopropyl ring), 4.20 (2H, q, ethyl H’s), 3.05 (1H, m, cyclopropyl H adjacent to the aromatic ring), 2.62 (1H,
m, cyclopropyl H), 2.35 (3H, s, Me on one isomer), 2.32 (3H, s, Me on other isomer), 1.26 (3H, t, ethyl H’s), 0.86
(
1H, m, H a to the ester). ESI-MS: 229 (M+1). These products are similar to those reported for cyclopropanation
13
of naphthalene with CuSO and ethyl diazoacetate.
4
1
1
3. Cummins, J. M.; Porter, T. A.; Jones, M. J. J. Am. Chem. Soc. 1998, 120, 6473–6476.
1
4. Compound 13: H NMR (500 MHz, CDCl ): l 8.02 (1H, d, C1-H), 7.69 (1H, d, C7-H or C6-H), 7.72 (1H, d,
3
C7-H or C6-H), 7.56 (1H, d, C4-H), 7.32 (1H, dd, C2-H), 3.23 (1H, dt, C11-H), 2. 88 (1H, m, C11-H), 2.55–2.44
(
2H, m, C16-H’s), 2.36 (3H, s, acetate Me), 2.34 (1H, m, C15-H), 2.22 (1H, m, C12-H), 2.08 (1H, m, C15-H), 1.96
1
(1H, m, C12-H), 1.35 (3H, s, C18-Me). FAB-MS: m/z 325 (M+1), 307 (M−H O+1). The H NMR, COSY, and
2
HSQC spectra were very similar for compound 13 and compound 8. One major difference was the greater
downfield shift in the angular methyl to l 1.35 ppm in 13. Hydroxylations at 14b are known to produce larger
15
downfield shifts than 14a hydroxylations.
15. Bhacca, N. S.; Williams, D. H. Applications of NMR Spectroscopy in Organic Chemistry: Illustrations from the
Steroid Field; Holden-Day: San Francisco, 1964; pp. 22–23.
1
1
6. Compound 14: H NMR (500 MHz, CDCl ): l 7.99 (1H, d, C1-H), 7.78 (1H, d, C7-H or C6-H), 7.70 (1H, d,
3
C7-H or C6-H), 7.52 (1H, d, C4-H), 7.26 (1H, dd, C2-H), 5.12 (1H, dd, C17-H), 3.19 (2H, m, C11-H’s), 2.36–2.30
(
2H, m, C16-H and C15-H), 2.35 (3H, s, C3-acetate), 2.14 (3H, s, C17-acetate), 2.12 (1H, m, C15-H), 1.97 (1H,
m, C16-H), 1.89 (1H, m, C12-H), 1.81 (1H, m, C12-H), 1.07 (3H, s, C18-Me). FAB-MS: m/z 368 (M+), 351
M−H O+1). High resolution FAB-MS (368.21) indicated an elemental composition of C H O . Similar
(
2
22 24
5
characterization and arguments were made to identify compound 14 as for compound 8.
7. Meunier, B. Chem. Rev. 1992, 92, 1411–1456.
1
.