Biomimetic hydroxylation of saturated carbons with artificial cytochrome P-450 enzymes - Liberating chemistry from the tyranny of functional groups

Article abstract of DOI:10.1016/S0040-4020(01)01098-5

Five mimics of cytochrome P-450 have been prepared and examined as catalysts for the specific hydroxylation of steroids. Reactions occur dictated by the geometries of the complexes, overcoming the intrinsic reactivity of a carbon-carbon double bond and of

Full text of DOI:10.1016/S0040-4020(01)01098-5

TETRAHEDRON
Pergamon
Tetrahedron 58 (2002) 653±659
Biomimetic hydroxylation of saturated carbons with arti®cial
cytochrome P-450 enzymesÐliberating chemistry from the
tyranny of functional groups
Ronald Breslow,p Jerry Yang and Jiaming Yan
Department of Chemistry, Columbia University, New York, NY 10027, USA
Received 11 May 2001; revised 15 August 2001; accepted 16 August 2001
AbstractÐFive mimics of cytochrome P-450 have been prepared and examined as catalysts for the speci®c hydroxylation of steroids.
Reactions occur dictated by the geometries of the complexes, overcoming the intrinsic reactivity of a carbon±carbon double bond and of a
secondary carbinol group. In some cases as many as 3000 catalytic turnovers are observed. q 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
selectivity, which could permit chemists to achieve
reactions now possible only by the use of enzymes, for
example in fermentation processes. We called this ®eld
`biomimetic chemistry',3±5 a name that has now been
extended to cover all respects in which chemists invent
new processes inspired by biological chemistry. Our ®rst
example was catalytic. Anisole bound into a cyclodextrin
in water and was then selectively chlorinated as the result of
geometric control.6 However, most of our early work used
covalently attached reagents or templates.
Synthetic chemistry and biochemistry achieve selective
reactions using very different approaches. Selectivity in
synthesis is dominated by the intrinsic reactivity of the func-
tional groups of the substrate. For example, it is normally
not possible to reduce a ketone if a more reactive aldehyde is
part of the molecule. In addition, it is not possible to oxidize
an unactivated saturated carbon if the molecule contains
secondary carbinol or ole®nic groups.
For example, the photolysis of benzophenonecarboxylic
acid esters of steroidal alcohols led to the selective function-
alization of carbon atoms that were within reach of the
benzophenone carbonyl group.7 The geometric control
was not perfectÐonly carbons within reach were attacked,
but the ¯exibility of the substrate and the ¯exibility resulting
from the single attachment of the benzophenone ester to the
substrate often permitted more than one position to be
functionalized.
Some exceptions exist, as in the Barton functionalization of
a steroid methyl group by photolysis of a nearby nitrite
ester.1 However, this depends on the presence of a particular
functional group in the molecule, so it is a special case of
functional group domination of the chemistry.
By contrast, enzymes can override the intrinsic reactivity
of their substrates, deriving their selectivities from the
geometries of the enzyme/substrate complexes. As a typical
example, in the biosynthesis of cholesterol, cytochrome
P-450 enzymes oxidize three methyl groups of the triterpene
lanosterol even though those methyls sit on saturated
carbons and are not near any special functional groups.2
This occurs even though lanosterol has two intrinsically
more reactive carbon±carbon double bonds and an easily
oxidizable secondary carbinol group. Only the methyl
groups are within reach of the oxidizing unit in the
enzyme±substrate complex.
In a second approach, we linked iodophenyl carboxylic
acids to steroids and used them to direct free radical
chlorinations to particular positions in the steroid.8 Again
the reactions were dominated by the geometries in the
esters, but again, there was some ¯exibility that at times
diminished the selectivity. Also, the chlorinations directed
by the iodophenyl groups, and those directed by other types
of templates,9 had a preference for attack on tertiary C±H
bonds rather than the less reactive methylene or methyl
groups of the substrate. Intrinsic substrate reactivity was
playing some role.
Many years ago, we set out to imitate such geometric
Keywords: cytochrome P-450; enzyme mimics; steroids; remote oxidation;
cyclodextrins.
Corresponding author. Tel.: 11-212-854-2170; fax: 11-212-854-2755;
e-mail: rb33@columbia.edu
With attached templates acting as catalysts for the chlorina-
tion of steroids and other substrates, it was not reasonable to
diminish the ¯exibilities so as to achieve better geometric
p
0040±4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)01098-5

654
R. Breslow et al. / Tetrahedron 58 (2002) 653±659
control of the reactions. Better rigidity would require
multiple interactions between catalyst and substrate, and a
covalently attached catalyst does not achieve any turnovers
until it is chemically removed and re-attached to a new
substrate. Such a cumbersome process makes complicated
catalysts unattractive in terms of the work needed to achieve
a limited result. We needed true turnover catalysts to
improve our biomimetic approach to selectivity.
ship in the complex, and very recently others have reported
such selectivity with analogs of our catalyst.19
The enzymes in the class called cyctochrome P-450 are able
to attack any hydrocarbon group, activated or not, even
methyls on saturated carbon. These enzymes contain a
heme group, and they form an iron-oxo species that inserts
the oxygen into C±H bonds. The mechanism apparently
involves abstracting the hydrogen with the oxygen atom,
and then transferring the resulting hydroxyl group from
the iron to the newly formed substrate radical.
Other research groups have studied models for cytochrome
P-450 enzymes,10 but not with the well de®ned reversible
substrate binding that we wanted to develop. However, their
®ndings were important to us. They have found that the
intermediate iron-oxo species can be generated by oxygen
transfer from iodosobenzene and many other oxidizing
species, not just from the O2 used biochemically.11 Also,
Groves has found in model systems that a manganese(III)
porphyrin is more effective than is the Fe(II) porphyrin of
heme.12 Both of these ®ndings aided our own work.
We focused our efforts on achieving the selective hydroxyl-
ation of saturated unactivated carbons, at ®rst with steroid
substrates. The rigid steroid skeleton makes geometric
control easier, and the selective hydroxylation of steroids
is of practical importance. Such selective hydroxylations for
the synthesis of steroid medicinals are currently achieved by
fermentation, taking advantage of enzymatic selectivity. In
addition, when steroids are hydroxylated, there are charac-
teristic shifts in the proton NMR, including the signals of the
angular methyl groups, which are quite diagnostic. These
spectra can establish whether more than one product is
formed, and they are the critical evidence for the nature of
the products, including stereochemistry (Scheme 1).
Recently, we have developed some arti®cial cyctochrome
P-450 enzymes that do achieve the desired result. They bind
their substrates with non-covalent interactions and achieve
true turnover catalysis. Furthermore, they have more than
one catalyst±substrate interaction, so ¯exibility is mini-
mized. As a result, they are indeed able to direct reactions
to unactivated saturated carbons in the presence of other-
wise more reactive functionality.
We prepared an ester derivative 3 of androstan-3,17-diol 2
that carried two tert-butylphenyl hydrophobic groups for
binding into the CDs, as well as sulfonate groups for
water solubility.14 We saw that the Mn(III) complex of
porphyrin 1 catalyzed the hydroxylation of a steroid posi-
tion, and with remarkable selectivity. The only product was
the 6a-hydroxy derivative of the substrate, which was
hydrolyzed to 3b-6a-17b-androstantriol 4. No other posi-
tion was hydroxylated, the stereochemistry at C-6 was
exclusively equatorial a, and no ketone was formed even
though the hydroxylation proceeded with four turnovers. In
a control reaction, a substrate like 3 but lacking the tert-
butylphenyl groups was not oxidized at all under our reac-
tion conditions, so the catalyst/substrate binding produced
both reactivity and selectivity.
Some of our work has been reported in preliminary com-
munications13±16 and in a recent review.17 In this full paper,
we describe the details and some striking new advances.
In our ®rst approach we used reversible metal ion coordina-
tion to hold a substrate in a well de®ned position relative to
the oxidizing species.18 An iron porphyrin was synthesized
carrying four 8-hydroxyquinoline groups, which bound to
added Cu(II). Then a substrate was added with nicotinate
ester groups at both ends, and we saw that the two pyridine
rings of the substrate bound through the Cu(II) bridges to the
porphyrin. This promoted the catalyzed epoxidation of the
substrate, under conditions in which an uncoordinated
substrate was not attacked. Thus, we were indeed achieving
the imitation of cytochrome P-450 in catalyzing the oxida-
tion of a bound substrate, but there was no special positional
selectivity involved.
Molecular models explained these selectivities.20 The
steroid is bound to CDs on opposite sides of the porphyrin,
holding the steroid above the Mn(III). It can tip edge down
to present the a-hydrogen at C-6 to an oxygen atom on the
porphyrin Mn, but hydrogens of C-5 or C-7 are not within
reach, nor is the b-hydrogen at C-6. The fact that the
ordinarily reactive C-6 carbinol is not oxidized to the ketone
even with turnover catalysis is striking. We propose that
such oxidations require attack by the MnvO species on
the CH bond of the carbinol, which is out of reach in our
substrate/catalyst complex.
We then synthesized a porphyrin 1 carrying four b-cyclo-
dextrin (b-CD) units, and saw that its Mn(III) complex
catalyzed the epoxidation of ole®nic substrates that bound
into the CDs in water solution.13 Again, no particular
positional selectivity was involved. It was obvious that
with a polyene there should be selectivity among the double
bonds because of the catalyst/substrate geometrical relation-
We saw only a few turnovers with this catalyst, but it was
clear how to improve the situation.14 Previous work showed

R. Breslow et al. / Tetrahedron 58 (2002) 653±659
655
Scheme 1.
that penta¯uorophenyl groups on the meso carbons stabilize
metalloporphyrins to oxidation,21 so we devised a particu-
larly easy synthesis of such a ¯uorinated catalyst, compound
5.16 This showed the same selectivity we had seen
previously with catalyst 5-Mn(III), but now the hydroxyla-
tion of the substrate at C-6 occurred with 187 turnovers
before the catalyst was destroyed. Again there were no
other products than the C-6 a-alcohol, which was not
oxidized to the ketone.
2. Results and discussion
In cytochrome P-450 enzymes, there is an additional ligand
on the iron atom, the sulfur of a cysteine. In our work, we
have added an additional external ligand for two purposes: it
shields one face of the metalloporphyrin so that both the
oxygen atom and the substrate are bound on the same
porphyrin face, and to some extent it activates the oxo-
porphyrin derivative for oxygen atom transfer to the
substrate. We originally used adamantanecarboxylate ion

656
R. Breslow et al. / Tetrahedron 58 (2002) 653±659
The hydroxylation of substrate 3 to the 6-hydroxy derivative
involved the use of only two of the CDs in the Mn(III)
catalysts derived from 1, 5, 6, or 7, so binding the substrate
at positions 3 and 17 permitted it to rotate and present the
edge of the steroid to the oxidizing catalytic species. We
realized that we could attach an additional binding ester
group to the androstane nucleus, at the C-6 position that
had just been hydroxylated, and that this triester 8 would
be able to insert three tert-butylphenyl groups into three CD
units of the catalyst. Models showed that this would present
the face of the steroid to the porphyrin ring, and in particular
that the hydrogen at C-9 should now be the one attacked and
converted to hydroxyl.20
This was the result.20 Reaction of substrate 8 with iodoso-
benzene in the presence of catalyst 5-Mn(III) afforded the
9-hydroxy derivative 9, hydrolyzed to the androstane tetraol
10. This was the only signi®cant product, but a trace amount
of an isomer was formed that we have now identi®ed as the
15b hydroxysteroid 11. This isomer was also formed to
some extent with catalysts 6 and 7. Apparently, there is
not only the triple binding that produces the 9-hydroxy-
steroid 9, but also some double binding by the ligands
attached at C-6 and C-17 of the steroid. Models show that
this is possible, and that in the complex it can present posi-
tion 15 to the MnvO oxidizing group. Thus further work is
needed to encourage only the highly selective triple binding
interaction between 8 and 5-Mn(III).
as the extra ligand,13 but more recently have added pyridine
to the solution. However, we wanted to create a catalyst in
which this ligand was already present, so we synthesized
compound 6. Models showed that the pyridine ring in 6
could coordinate to the Mn(III).
We then used this catalyst to perform the hydroxylation of
substrate 3 without added pyridine, and as before it
selectively hydroxylated the 6a position of the substrate.
However, catalyst 6-Mn(III) now did this oxidation with
2000 turnovers, compared to the 187 turnovers for catalyst
5-Mn(III). Interestingly, porphyrin 6 was prepared from an
intermediate 7 with an o-nitrophenyl group, and the Mn(III)
catalyst prepared from this performed the 6a hydroxylation
of substrate 3 with 3000 turnovers. These are very effective
catalysts.
We have examined the hydroxylation of lithocholic acid 12
and of D11-lithocholenic acid 13 catalyzed by compound
5-Mn(III). We had earlier found that lithocholic acid binds
strongly to b-CD, and proposed that the strong binding
re¯ects an extra contribution from ring A of the steroid
because of the AB cis junction.24 We ®nd that both litho-
cholic acid and lithocholenic acid are indeed oxidized by
our catalyst 5-Mn(III) without the necessity of any added
ester groups, but the reaction is not completely selective.
With both substrates, we see hydroxylation on C-15 of ring
D. There is also some oxidation of the 3-hydroxyl to a
We wanted to learn how to perform the selective hydroxyl-
ations at C-9 of the steroid. Since C-9 alcohols can be
dehydrated to introduce the 9(11) double bond,22 this
permits the synthesis of 9-¯uoro-11-hydroxy steroids that
are the basis for effective corticosteroid anti-in¯ammatory
pharmaceuticals.23 From model building we saw how to
direct our hydroxylation reaction to this desirable position.

R. Breslow et al. / Tetrahedron 58 (2002) 653±659
657
of double bonds and secondary carbinol groups, which are
not attacked for geometric reasons. Thus in these reactions
the normal dominance of functional group reactivity has
been overcome, as it is in typical enzymatic reactions.
Arti®cial enzymes of the type we have prepared have the
potential to permit processes that have hitherto been
possible only by fermentation procedures.
4. Experimental
4.1. Catalysts
Porphyrin 1 was prepared as previously described,13 as was
porphyrin 5. They were converted to their Mn(III)
complexes in the standard fashion, as described below for
the Mn complexes of 6 and 7.
ketone, but this is not the dominant reaction. There is some
oxidation at C-1 and C-6, but there is no oxidation into or
next to the double bond in lithocholenic acid.
4.1.1. Porphyrin 7. Porphyrin 7 was prepared as follows.
The porphyrin carrying three penta¯uorophenyl groups and
one o-nitrophenyl group on the meso positions was prepared
as described in the literature by reaction of a mixture of
penta¯uorobenzaldehyde and o-nitrobenzaldehyde with
pyrrole.25 Then, treatment with an excess of 6-mercapto-6-
deoxycycloheptaamylose attached the three cyclodextrin
rings to yield 7.
In detail, 6-deoxy-6-mercapto-b-cyclodextrin (792 mg,
0.688 mmol) and the above porphyrin (160 mg,
0.172 mmol) were dissolved in 10 ml anhydrous DMF.
The solution was stirred under argon for 10 min, then
K2CO3 (280 mg) was added. The mixture was stirred at
room temperature in the dark overnight and dissolved in
200 ml water. The aqueous solution was neutralized and
loaded on a reverse-phase column. Washing the column
with a 40±100% aqueous CH3OH gradient system provided
the pure 7 in 82% yield.
1H NMR (DMSO-d6, 300 MHz) d: 9.30 (6H), 8.87 (2H),
8.64 (1H), 8.51 (1H), 8.21 (2H), 5.77±5.69 (42H, 28-OH),
5.03±4.31 (39H, H1 and 18-OH), 3.66±3.22 (126H), 23.06
(2H). MALDI MS: 4350.09 [M1Na]1 UV±vis (H2O): 415
(Soret), 508, 578 nm.
These simple substrates are expected to use only one of the
CD groups of 5-Mn(III) for binding, so the geometry of
the substrate/catalyst complex is ambiguous. However, the
geometry of binding does prevent attack on the double bond
and minimizes attack on the alcohol function. Thus, these
results indicate that even with these simple singly bound
substrates the intrinsic reactivity of double bonds, and to
some extent of alcohol groups, has been overcome by
geometric control.
4.1.2. Porphyrin 6. Porphyrin 6 was prepared from the
porphyrin carrying three penta¯uorophenyl groups and
one o-aminophenyl group on the meso positions, prepared
by reduction of the above o-nitrophenyl derivative as
described in the literature.25 In detail, (5,10,15-tri(penta-
¯uorophenyl)-20-((o-(N-3-pyridinepropionyl)amino)phenyl)
porphyrin) 3-pyridinepropionic acid (68 mg, 0.444 mmol)
was dissolved in 12 ml CH2Cl2/toluene (1:1). To this solu-
tion was added oxalyl chloride (3 ml). The solution was
stirred at 508C for 24 h, and the solvent and excess oxalyl
chloride were removed under reduced pressure. The residue
was dissolved in 10 ml anhydrous THF and added dropwise
to a solution of the above porphyrin (200 mg, 0.222 mmol)
and pyridine (1 ml) in 200 ml anhydrous THF. The mixture
was stirred under argon at room temperature overnight. The
THF was distilled off and the residue was dissolved in
200 ml chloroform. The organic layer was washed with
water and dried over anhydrous Na2SO4. The pure product
was obtained in 83% yield by column chromatography
3. Conclusion
We have prepared mimics of cytochrome P-450 that achieve
the selective hydroxylations of steroids directed by the
geometries of the catalyst/substrate complexes. The best
catalysts achieve these selective reactions with as many as
3000 turnovers. The selectivities are such that saturated
carbons of the substrates are hydroxylated in the presence

658
R. Breslow et al. / Tetrahedron 58 (2002) 653±659
(SiO2) with CHCl3/CH3OH (10:1) as eluent. 1H NMR
(CDCl3, 300 MHz) d: 8.93±8.81 (m, 7H), 8.62 (m, 2H),
8.06 (d, 1H, Jˆ7.5 Hz), 7.85 (m, 2H), 7.65 (m, 2H), 6.69
(m, 2H), 6.46 (m, 1H), 2.25 (t, 2H, Jˆ7.4 Hz), 1.56 (t, 2H,
Jˆ7.4 Hz), 22.89 (s, 2H). FAB-MS: 1033.3 [M1H]1. UV±
vis (CH2Cl2): 414 (Soret), 508, 579 nm.
catalyst, the conversion to the single product 4 was
quantitative.
Substrate 3 (0.06 mmol) was dissolved in 60 ml water and
the pH was adjusted to 5.5. To this solution, 0.00020 equiv.
catalyst 6 or 7 in water was added by syringe. The solution
was stirred at room temperature in the dark for 10 min. A
solution of iodosylbenzene (66 mg, 0.30 mmol) in 3.0 ml
methanol was added in several aliquots over 4 h. The
mixture was stirred for 6 h at room temperature. After the
excess oxidant was quenched, with sodium thiosulfate 25%
aqueous KOH (20 ml) was added and the mixture was stir-
red overnight. The solution was extracted with EtOAc
(4£50 ml). The organic layer was washed with saturated
brine and dried over Na2SO4. Evaporation of the solvent
and column chromatography (SiO2) gave the product 4
and recovered starting material.
Then, porphyrin 6 was obtained from this porphyrin in 72%
yield by reacting it (190 mg, 0.184 mmol) with 6-deoxy-6-
mercapto-b-cyclodextrin (847 mg, 0.736 mmol) using the
same procedure as described above for 7.
1H NMR (DMSO-d6, 300 MHz) d: 9.26±8.87 (m, 9H), 8.15
(d, 1H, Jˆ6.73 Hz), 8.08 (d, 1H, Jˆ8.19 Hz), 7.99 (d, 1H,
Jˆ4.6 Hz), 7.90±7.85 (m, 2H), 7.65 (t, 1H), 6.85 (d, 1H,
Jˆ7.50 Hz), 6.72 (m, 1H), 5.86±5.68 (m, 42H, 28-OH),
5.03±4.32 (m, 39H, H1, 18-OH), 3.66±3.32 (m, 126H),
2.11 (t, 2H, Jˆ8.06 Hz), 1.70 (t, 2H, Jˆ8.06 Hz), 23.03
(s, 2H). MALDI MS: 4452.68 [M1Na]1 UV±vis (H2O):
413 (Soret), 509, 580 nm.
With catalyst 7, the oxidation conversion was 65.4%, turn-
overs were 2900. With catalyst 6, the oxidation conversion
is 50.1%, turnovers are 2200. In the earlier work with
catalyst 5-Mn(III), the turnovers were also determined by
benzoylating the mixture of product and starting polyols,
and assaying them by HPLC and by gas chromatography.
4.2. Substrates
Substrate 3 was prepared as described and fully character-
ized previously.14,15 The procedure and characterization are
parallel to that used for substrate 8, described below.
4.4. Product identi®cation
Product 4 was identi®ed originally from its NMR spectrum,
but was then compared with an authentic sample obtained
from Steraloids Inc., as reported earlier.14,15 Product 10 had
1H NMR (CDCl3, 500 MHz): d: 3.73 (t, 1H, C17-H), 3.58
(m, 1H, C3-H), 3.43 (dt, 1H, C6-H), 0.959 (s, 3H, C19-Me),
0.753 (s, 3H, C18-Me). CI-MS: m/zˆ342 (M111NH3),
323 (M21, neg.). The chemical shifts of the angular methyl
groups were consistent with reported values for 9-hydroxyl-
ation.26 NOESY indicated that the C5 and C7a protons
moved down®eld signi®cantly relative to starting triol 4,
as expected for their 1,3-diaxial relation to the 9-OH.
Product 11 had an additional low-®eld C±H relative to 4,
at d: 4.16. The COSY spectrum showed that is was coupled
to the C-16 protons, which were also coupled to the C-17
proton. The chemical shifts for the C-18-Me at 0.98
(predicted 1.00) and the C-19-Me at 0.88 (predicted 0.89)
were as expected for a 15-bOH group.26
Substrate 8 was prepared by reaction of 136 mg of triol 4,
prepared by our catalytic hydroxylation of 3, with 650 mg of
3(4-tert-butylphenyl)glutaric anhydride in 100 ml methyl-
ene chloride with 460 ml of DBU for 12 h at 238C. The
solution was washed with aqueous ammonium chloride
and dried with sodium sulfate, then directly treated with
608 mg of N-hydroxysuccinimide and 1.0 g of EDC. The
triactivated triester product was isolated by column chroma-
tography in 55% yield, and had 1H NMR (500 MHz,
CDCl3): d: 7.29 (m, 6H, aromatic), 7.15 (m, 6H, aromatic),
4.51±4.39 (m, 3H, C3-H1C6-H1C17-H), 3.65 (m, 3H,
benzylic C-H's), 2.80 (s, 12H, succinimide-H's), 1.28 (s,
27H, tert-butyl), 0.76 (m, 3H, C19-Me from different
diastereomers), 0.56 (m, 3H, C18-Me from different dia-
stereomers), 3.01±2.67 (12H, linker-H's), 2.0±0.51 (23H,
steroid envelope). FAB-MS: m/zˆ1336 (M11).
The triester was then treated with 6 equiv. of taurine and
6 equiv. of triethylamine in 20 ml anhydrous DMF over-
night, the solvent was removed and the product 8 was
Acknowledgements
1
isolated as a precipitate from 10 ml of 2N HCl. H NMR
(500 MHz, d6-DMSO): d: 7.68 (m, 3H, amide-H's), 7.22
(m, 6H, aromatic), 7.07 (m, 6H, aromatic), 4.25 (m, 3H,
C3-H1C6-H1C17-H), 3.20 (m, 6H, taurine-H's adjacent
to the nitrogen), 0.62 (s, 3H, C19-Me), 0.46 (s, 3H, C18-
Me), 3.4±2.28 (21H, linker-H's), 1.85±0.30 (23H, steroid
envelope).
We thank some of the previous collaborators on this overall
project, listed in the references, and acknowledge support of
this work by the NIH, the NSF, and the EPA. Jerry Yang
acknowledges support by an NCERQA STAR graduate
fellowship and Bristol-Myers Squibb graduate fellowship.
4.3. Hydroxylations
References
The procedures are described for the oxidation of substrate 3
by catalysts 6-Mn(III) and 7-Mn(III). The procedures are
similar for catalyst 5-Mn(III). The procedures described are
those that gave incomplete conversion, so the turnover by
the catalyst could be determined. With higher amounts of
1. Barton, D. H. R.; Beaton, J. M. J. Am. Chem. Soc. 1961, 83,
750.
2. Meunier, B. Chem. Rev. 1992, 92, 1411.
3. Breslow, R. Chem. Soc. Rev. 1972, 1, 553.
4. Breslow, R. Acc. Chem. Res. 1980, 13, 170.

R. Breslow et al. / Tetrahedron 58 (2002) 653±659
659
5. Breslow, R. Comprehensive Organic Synthesis; Trost, B. M.,
Ed.; Pergamon: Oxford, 1991; Vol. 7, pp 39±52.
6. (a) Breslow, R.; Campbell, P. J. Am. Chem. Soc. 1969, 91,
3085. (b) Breslow, R.; Campbell, P. Bioorg. Chem. 1971, 1,
140±156.
15. Breslow, R.; Huang, Y.; Zhang, X.; Yang, J. Proc. Natl Acad.
Sci. USA 1997, 94, 11156.
16. Breslow, R.; Gabriele, B.; Yang, J. Tetrahedron Lett. 1998,
39, 2887.
17. Breslow, R. Chem. Record 2001, 1, 3±11.
18. Breslow, R.; Brown, A. B.; McCullough, R. D.; White, P. W.
J. Am. Chem. Soc. 1989, 111, 4517.
7. Breslow, R.; Baldwin, S.; Flechtner, T.; Kalicky, P.; Liu, S.;
Washburn, W. J. Am. Chem. Soc. 1973, 95, 3251.
8. Breslow, R.; Corcoran, R. J.; Snider, B. B. J. Am. Chem. Soc.
1974, 96, 6791.
19. French, R. R.; Holzer, P.; Leuenberger, M. G.; Woggon,
W.-D. Angew. Chem. Int. Ed. 2000, 39, 1267.
20. Yang, J.; Breslow, R. Angew. Chem. Int. Ed. 2000, 39, 2692±
2694.
9. Breslow, R.; Brandl, M.; Hunger, J.; Adams, A. D. J. Am.
Chem. Soc. 1987, 109, 3799.
21. Ellis, P. E.; Lyons, J. E. Coord. Chem. Rev. 1990, 105, 181.
22. Fried, J.; Sabo, E. F. J. Am. Chem. Soc. 1957, 79, 1130.
23. Bergstrom, C.; Nicholson, R.; Dodson, R. J. Org. Chem. 1963,
28, 2633 and references therein.
10. Dolphin, D.; Traylor, T. G.; Xie, L. Acc. Chem. Res. 1997, 30,
251.
11. Robert, A.; Meunier, B. New J. Chem. 1988, 12, 885 and
references therein.
24. Yang, Z.; Breslow, R. Tetrahedron Lett. 1997, 38, 6171.
25. Poliar, C.; Artaud, I.; Mansuy, D. Inorg. Chem. 1996, 35, 210.
26. Bhacca, N. S.; Williams, D. H. Applications of NMR
Spectroscopy in Organic Chemistry: Illustrations from the
Steroid Field; Holden-Day: San Francisco, CA, 1964; p 21.
12. Groves, J. T.; Kruper, W. J.; Haushalter, R. C. J. Am. Chem.
Soc. 1980, 102, 6375.
13. Breslow, R.; Zhang, X.; Xu, R.; Maletic, M.; Merger, R.
J. Am. Chem. Soc. 1996, 118, 11678.
14. Breslow, R.; Zhang, X.; Huang, Y. J. Am. Chem. Soc. 1997,
119, 4535.