Metal coordination-directed hydroxylation of steroids with a novel artificial P-450 catalyst

Article abstract of DOI:10.1021/ol052589i

A novel catalyst has been synthesized in which a manganese-porphyrin unit is linked to two 2,2′-bipyridyl groups and two pentafluorophenyl groups in trans fashion on its four meso positions. Relative to a previous catalyst in which the manganese-porphyrin had four 2,2′-bipyridyl groups, the new catalyst, in the presence of Cu²⁺ ions as coordinating linkers, catalyzes the oxidation of a steroid substrate with much better regioselectivity and higher turnover numbers.

Full text of DOI:10.1021/ol052589i

ORGANIC
LETTERS
2006
Vol. 8, No. 2
251-254
Metal Coordination-Directed
Hydroxylation of Steroids with a Novel
Artificial P-450 Catalyst
Zhenglai Fang and Ronald Breslow*
Department of Chemistry, Columbia UniVersity, New York, New York 10027
rb33@columbia.edu
Received October 25, 2005
ABSTRACT
A novel catalyst has been synthesized in which a manganese
−
porphyrin unit is linked to two 2,2
′
-bipyridyl groups and two pentafluorophenyl
groups in trans fashion on its four meso positions. Relative to a previous catalyst in which the manganese
−
porphyrin had four 2,2
′
-bipyridyl
+
groups, the new catalyst, in the presence of Cu2 ions as coordinating linkers, catalyzes the oxidation of a steroid substrate with much better
regioselectivity and higher turnover numbers.
We have described hydroxylation of steroids by mimics of
the enzyme cytochrome P-450 class. In our earliest examples,
we used metalloporphyrins carrying cyclodextrin groups to
bind hydrophobic substrates in water. When the cyclodextrins
were linked to p-phenylene groups attached to the porphyrin
meso positions in catalyst 1, we saw highly selective
hydroxylation at the C-6 R position of a steroid derivative 2
doubly bound into the cyclodextrins (Figure 1).1 However,
there were only about 10 turnovers before the catalyst was
itself oxidatively destroyed. When instead the phenylene
linkers in the catalyst were replaced by tetrafluorophenylene
groups in catalyst 3, we still saw hydroxylation selectively
at C-6 R, but now with 187 turnovers before destruction of
the catalyst (Figure 1).2 That same perfluorinated catalyst
was also able to hydroxylate the steroid in the C-9 alpha
position when a substrate was hydrophobically bound to three
cyclodextrins of the catalyst.3
There could be advantages in changing the mode of
substrates/catalyst binding. Hydrophobic binding is somewhat
ill-defined geometrically, and it is seen only in water solvent.
We had used double metal-ion coordination to bind a sub-
strate to a metalloporphyrin and a metallosalen, and had ob-
served good selective epoxidation of the substrate.4 Thus,
we used manganese porphyrin 4 to bind and oxidize sub-
strates, e.g., 5, in the presence of coordinating metals such
as Cu2+ ions (Figure 2).5 However, the oxidation was not
selective, with observed formation of six almost evenly dis-
tributed products, including 6. Moreover, the turnover num-
ber was only 5, low compared to the 187 turnovers we
observed with catalyst 3.
The poor catalytic performance in this biomimetic study
might result from two factorssthe intrinsic reactivity and
(1) Breslow, R.; Zhang, X.; Huang, Y. J. Am. Chem. Soc. 1997, 119,
4535.
(2) (a) Breslow, R.; Gabriele, B.; Yang, J. Tetrahedron Lett. 1998, 39,
2887. (b) Yang, J.; Breslow, R., Angew. Chem. Int. Ed. 2000, 39, 2692; (c)
Breslow, R.; Fang, Z., Tetrahedron Lett. 2002, 43, 5197; (d) Yang, J.;
Gabriele, B.; Belvedere, S.; Huang, Y.; Breslow, R. J. Org. Chem. 2002,
67, 5057.
(3) Breslow, R.; Yan, J.; Belvedere, S. Tetrahedron Lett. 2002, 43, 363.
(4) Breslow, R.; Brown, A. B.; McCullough. R. D.; White, P. W. J. Am.
Chem. Soc. 1989, 111, 4517.
(5) Belvedere, S.; Breslow, R. Bioorg. Chem. 2001, 29, 321.
10.1021/ol052589i CCC: $33.50
© 2006 American Chemical Society
Published on Web 12/21/2005

both the manganese porphyrin and the substrate are easily
oxidized to N-oxides under the hydroxylation conditions.6
Once N-oxides are formed, the nitrogens lose the ability to
coordinate to the metal center. Indeed, we isolated an ini-
tial undesired substrate oxidation product, the bis N-oxide
of 5 from the hydroxylation experiments before basic
hydrolysis;7 this was not further hydroxylated under the same
conditions.
To avoid the undesired substrate oxidation, we attached
different metal ligands to the steroid, two R-phosphonoacetyl
groups in substrate 7 (Figure 3).8,9 When 7 was subjected to
Figure 1. Selective hydroxylation by a catalyst with attached
cyclodextrins.
Figure 3. Hydroxylation of substrate 7 with catalyst 4.
the oxidative stability of catalyst 4 are significantly lower
compared with catalyst 3, and the nitrogen atoms from
hydroxylation with iodosobenzene and the manganese por-
phyrin 4 in a mixture of H2O and t-BuOH (1:1), products
resulting from hydroxylation on the 5R, 6R, 7R, 7â, 12R,
1
and 14R positions were identified by H NMR.10,11 Ten
catalytic turnovers were observed.
To further address the poor regioselectivity as well as the
low turnover in the metal coordination-directed oxidation,
we designed a new catalyst 8 with the introduction of two
C6F5 groups on its meso positions.
(6) Posakony, J. J.; Pratt, R. C.; Rettig, S. J.; James, B. R.; Skov, K. A.
Can. J. Chem. 1999, 77, 182.
(7) 1H NMR (300 MHz, CDCl3): δ ) 8.80 (d, 2H, J ) 6.0 Hz), 8.37
(d, 2H, J ) 6.2 Hz), 7.90 (d, 2H, J ) 7.8 Hz), 7.40 (t, 2H, J ) 6.5 Hz),
4.97 (m, 1H, C3R-H), 4.86 (t, 1H, J ) 7.7 Hz, C17R-H), 0.91 (s, 1H,
C19-Me), 0.84(s, 1H, C18-Me). CI-MS: 535.3 [M + H]+.
(8) Da Costa, C. P.; Song, B.; Gregan, F.; Sigel, H. J. Chem. Soc., Dalton
Trans. 2000, 899.
(9) Compound 7 was synthesized from the commercially available 3â,-
17â-androstanediol and 2-dibenzylphosphonate acetic acid in a three-step
procedure. 1H NMR (400 MHz, D2O): δ ) 4.47 (m, 1H, C3R-H), 4.40
(t, J ) 8.0 Hz, 1H, C17R-H), 2.55-2.41 (dd, 4H), 0.670 (s, 3H, C19-Me),
0.639 (s, 3H, C18-Me).
(10) The chemical shifts of C19 and C18 angular methyl protons in the
identified steroid products (CDCl3, 400 MHz). 3â,17â-Androstanediol
(starting material): δ ) 0.819 (C19-Me), 0.732 (C18-Me). 3â,5R,17â-
Androstanetriol: δ ) 1.003 (C19-Me), 0.738 (C18-Me), calcd 1.00, 0.73.
3â,6R,17â-Androstanetriol: δ ) 0.833 (C19-Me), 0.738 (C18-Me). Identical
to authentic sample; 3â,7R,17â-androstanetriol: δ ) 0.823 (C19-Me),
0.741 (C18-Me); calcd 0.81, 0.74. 3â,7â,17â-Androstanetriol: δ ) 0.850
(C19-Me), 0.763 (C18-Me), calcd 0.85, 0.76; 3â,14R,17â-Androstane-
triol: δ ) 0.844 (C19-Me), 0.864 (C18-Me), calcd 0.82, 0.85.
(11) Bhacca, N. S.; and Williams, D. H. Applications of NMR Spectros-
copy in Organic Chemistry: Illustrations from the Steroid Field; Holden-
Day: San Francisco, 1964; p 21.
Figure 2. Metal coordination-directed oxidation.
252
Org. Lett., Vol. 8, No. 2, 2006

Scheme 1
Compared to the previous manganese porphyrin catalyst
As hoped, 8 catalyzed the hydroxylation of substrate 7
very regioselectively (90% of the product) at the 6R position
with a significantly improved turnover number of 32 (Figure
4). The significant increase in the turnover number is not
surprising, considering the nonoxidizable nature of the
substrate’s coordination groups. The formation of an N-oxide
of 8, along with other oxidative degradations, are probably
the major factors that limit the turnover number to 32 in the
hydroxylation of 7.
4, the new artificial P-450 catalyst 8 brought the fol-
lowing advantages: (1) The incorporation of two strongly
electron-withdrawing pentafluorophenyl groups (C6F5) in-
creased the enzyme’s reactivity and oxidative stability, and
(2) the “trans” geometrical relationship between the two
bipyridyl ligands eliminates the possibility of undesired cis-
binding between the artificial P-450 enzyme and the sub-
strate, which may be the cause of some of the side reactions
with catalyst 4.
The synthesis of catalyst 8 was first attempted using either
a statistical approach under standard Alder-Longo condi-
tions12 or an approach developed by Lindsey toward the
synthesis of trans-A2B2-porphyrins prepared by reaction of
an aldehyde and a dipyrromethane.13 However, these two
attempts failed. Finally, 8 was successfully synthesized
following a Negishi-coupling approach (Scheme 1). The
dipyrromethane14 and the commercially available pentafluo-
robenzaldehyde were condensed under acid conditions fol-
lowed by DDQ oxidation to form a porphyrin which was
immediately protected as the zinc complex 9.15 Then 9 was
regioselectively brominated with NBS16 to give compound
10,17 which was coupled with a bipyridyl zinc bromide18
under Negishi conditions to give zinc porphyrin 11.19 This
was demetalated with trifluoroacetic acid to give porphyrin
12,20 which was then metalated with MnCl2 in air to give
catalyst 8.21
(12) Alder, A. D.; Longo, F. R.; Finarelli, J. D.; Goldmacher, J.; Assour,
J.; Korsakoff, L. J. Org. Chem. 1967, 32, 476.
(13) Rao, P. D.; Dhanalekshmi, S.; Littler, B. J.; Lindsey, J. S. J. Org.
Chem. 2000, 65, 7323.
Figure 4. Hydroxylation of substrate 7 with catalyst 8.
(14) Wang, Q. M.; Bruce, D. W. Synlett 1995, 1267.
(15) 1H NMR (400 MHz, acetone-d6): δ ) 10.51 (s, 2H, meso proton),
9.61 (d, 2H, J ) 4.5 Hz), 9.27 (d, 2H, J ) 4.5 Hz). CI-MS: 705.7 [M]+.
(16) DiMagno, S. G.; Lin, V.; Therien, M. J. J. Org. Chem. 1993, 58,
5983.
Both the selectivity and turnover number of metal coor-
dination directed steroid hydroxylation are not yet quite as
(17) 1H NMR (400 MHz, acetone-d6): δ ) 9.77 (d, 4H, J ) 4.7 Hz),
9.16 (d, 4H, J ) 4.7 Hz). CI-MS: 863.0 [M]+.
(18) Simkovsky, N. M.; Erman, M.; Roberts, S. M.; Parry, D. M.; Baxter,
A. D. J. Chem. Soc., Perkin Trans. 1 2002, 1847.
(19) 1H NMR (300 MHz, CDCl3): δ ) 9.48 (s, 2H), 9.06 (d, 4H, J )
4.7 Hz), 8.96 (d, 4H, J ) 4.7 Hz), 8.87 (m, 4H), 8.70 (m, 4H), 8.00 (m,
2H), 7.51 (m, 2H). FAB-MS: 1013 [M]+.
Org. Lett., Vol. 8, No. 2, 2006
253

good as the hydrophobic interaction directed steroid hydroxy-
lation by 3. However, both the regioselectivity and catalytic
turnover in the hydroxylation of steroids are probably good
enough for synthetic applications. This system may also be
more practical since it will not require a water solvent and
substrate water solubility for effective catalyst/substrate
binding.
Acknowledgment. This work has been supported by the
NSF and NIH.
Supporting Information Available: Experimental pro-
cedures for the syntheses and characterization of the products.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(20) 1H NMR (300 MHz, CDCl3): δ ) 9.50 (s, 2H), 8.99 (d, 4H, J )
4.8 Hz), 8.86 (m, 8H), 8.68 (m, 4H), 7.96 (m, 2H), 7.45 (m, 2H), -2.80 (s,
2H, pyrrole NH). MS (FAB): 951 [M]+. UV-vis (CH2Cl2, nm): 376
(Soret), 437, 470, 504, 544.
OL052589I
(21) UV-vis (CH2Cl2, nm): 331, 435 (Soret), 535, 573.
254
Org. Lett., Vol. 8, No. 2, 2006