Ozonolysis of cholesterol and other Δ5-steroids in the presence of alcohols: A revised mechanism and hydroperoxide structure of the solvent-participated product, confirmed by X-ray analysis

Article abstract of DOI:10.1039/a702633a

The structure of the product formed in the course of the reaction of cholesterol acetate and other Δ⁵-steroids with ozone in alcohol-containing solvents has been revised. The solvent-participated products are hydroperoxides, 5α-hydroperoxy-7α-alkoxy-5α-B-homo-6-oxasteroids 3 and not the previously claimed cyclic hemiperacetals, 5α-hydroxy-7α-alkoxy-B-dihomo-6,7-dioxacholestane derivatives 1. The final evidence has been obtained from X-ray crystal structure analysis of the selected hydroperoxides 3a, 3c and 3j. It is proposed that hydrogen bonding involving the hydroperoxy and the alkoxy group is responsible for the stability of these hydroperoxides and also exerts a directive effect in the nucleophilic attack on the Criegee intermediate 12 by the alcohol.

Full text of DOI:10.1039/a702633a

View Article Online / Journal Homepage / Table of Contents for this issue
Ozonolysis of cholesterol and other Ä5-steroids in the presence of
alcohols: a revised mechanism and hydroperoxide structure of the
solvent-participated product, confirmed by X-ray analysis
Zdzisław Paryzek† and Urszula Rychlewska
Faculty of Chemistry, A. Mickiewicz University, 60-780 Poznan´, Poland
The structure of the product formed in the course of the reaction of cholesterol acetate and other
Ä5-steroids with ozone in alcohol-containing solvents has been revised. The solvent-participated products
are hydroperoxides, 5á-hydroperoxy-7á-alkoxy-5á-B-homo-6-oxasteroids 3 and not the previously claimed
cyclic hemiperacetals, 5á-hydroxy-7a-alkoxy-B-dihomo-6,7-dioxacholestane derivatives 1. The final
evidence has been obtained from X-ray crystal structure analysis of the selected hydroperoxides 3a, 3c and
3j. It is proposed that hydrogen bonding involving the hydroperoxy and the alkoxy group is responsible for
the stability of these hydroperoxides and also exerts a directive effect in the nucleophilic attack on the
Criegee intermediate 12 by the alcohol.
C8H17
Y
Oxysterols are compounds of biochemical and biomedical
interest,1 and are commonly encountered in human blood,
other tissues and foods.2 Their variety of biological activities
and important effects on cell membranes have also been recog-
nized.3 The formation of oxysterols results from facile oxid-
O
O
O
ation of cholesterol and related sterols by active oxygen species.4
7a
OR
R1O
X
7
O
OH
Ozone might be a potential oxidizing reagent. In 1949, Criegee
postulated the existence of carbonyl oxides as key intermediates
in the ozonolysis of alkenes.5 Since that time investigations
have been carried out, both theoretical and experimental,
toward understanding the electronic nature and the structure of
carbonyl oxides, as well as their chemical reactivity.6 The known
reactions of carbonyl oxides include isomerization, nucleophilic
trapping, cycloadditions and oxygen transfer.7
The reaction of cholesterol with ozone results in formation
of the expected ozonide, which could be reduced to the 5,6-
secosterol.8 The 5,6-epoxide was also found among the prod-
ucts of cholesterol ozonolysis in ethyl acetate and other polar
solvents.9 However, the reaction takes a different pathway in
hydroxylic solvents, like water or alcohols. In 1957, the reaction
of cholesterol with ozone carried out in non-polar solvents in
the presence of alcohols (MeOH, EtOH or C5H11OH) was
reported by Lettre and Jahn.10 The major product of this reac-
tion was assigned the structure 1 (the hydroxy-peroxide) on the
basis of its chemical properties. The structure represented by
the general formula 1, with the hydroxy-peroxide moiety, was
then adopted for the principal products of ozonolysis of other
steroidal ∆5-olefins reacting in similar conditions.11.
OR
OH
3
1
R1 = H or Ac
R1 = CH3 or C2H5 or C5H11
a
b
c
d
e
f
X = β-OAc, H
R = Me
Y = C8H17
R = Et X = β-OAc, H Y = C8H17
R = Pri X = β-OAc, H
Y = C8H17
Y
R = But X = β-OAc, H Y = C8H17
R = Me
R = Me
Y = C8H17
Y = C8H17
Y = C8H17
Y = C8H17
Y = C2H5
Y = H
X = β-Cl, H
X = β-OEt, H
R = Me X = O
g
h
i
X = H2
R = Me
R = Me
R = Me
X
X = β-OAc, H
X = β-OAc, H
2
j
a
b
c
d
e
f
X = β-OAc, H
X = β-Cl, H
X = β-OEt, H
X = O
Y = C8H17
Y = C8H17
Y = C8H17
Y = C8H17
Y = C8H17
Y = C2H5
OOH
(CH2)n
O
X = H2
X = β-OAc, H
4
g
X = β-OAc, H Y = H
C8H17
C8H17
Gumulka and Smith published an extensive study on the
ozonation of cholesterol in participating9 and non-particip-
ating12 solvents. Their elaborate analysis of the extensive spec-
tral data13 was interpreted in favour of the structure 1 for the
major product formed in the reaction of cholesterol with ozone
carried out in hydroxylic solvents (H2O or alcohols). In the 1H
NMR spectra of compounds with the proposed general struc-
ture 1 the presence of a low-field signal, that was originally not
observed,9 at ca. δ 10 constitutes a weakness in their structural
assignment. The chemical evidence‡ supports either peroxy-
hemiacetal structure 1 or structure 3 with hydroperoxy group at
C-5.
O
AcO
O
O
AcO
OH
OCH3
O
OOH
OH
5
6
In general, hydroperoxides are characterized by a low-field
signal in their spectra at δH 7–10.14,15 In oxabicyclic peroxides of
partial structure 4, a low-field signal at δ 8.3–9.3 was assigned
to a hydrogen of a hydroperoxy group.16 Recently, we have
reported17 the steroidal hydroperoxides 5 (δH 9.80) and 6 (δH
9.95). These results together with the literature data cast some
† E-Mail: zparyzek@chem.amu.edu.pl
‡ The negative test for the presence of the hydroperoxide group was
most probably a result of its tertiary character.14
J. Chem. Soc., Perkin Trans. 2, 1997
2313

View Article Online
doubt on the correctness of the structure 1. In order to obtain
unequivocal evidence for the structure of the solvent-par-
ticipated product, we have reinvestigated the ozonolysis of
cholesterol acetate18 and other ∆5-steroids.
Results and discussion
In the reaction of cholesterol acetate 2a with ozone carried
out in chloroform–alcohol (MeOH, EtOH PriOH or ButOH)
1:1 mixtures at Ϫ70 ЊC we obtained a series of crystalline
compounds 3a–3d. Their physical and spectral properties were
similar to those previously described.9,10,13 The evidence pre-
sented below shows that these and other peroxidic products are
in fact hydroperoxides and have the structure depicted in the
general formula 3.
Fig. 1 Perspective view of the hydroperoxide 3j
2
after the addition of H2O), and a triplet of one proton (δ 4.58,
J 7.6 Hz) arising from the CH(OR)2 group. The IR spectrum of
9a showed absorption for both a free and an associated hydroxy
group at 3555 and 3420 cmϪ1, respectively. Similarly, the
ethoxy-hydroperoxide 3b was reduced to the bisacetal 9b.
In the 1H NMR spectrum of compounds 3a–3d an important
signal at δ ca. 10 should be assigned to the proton of the OOH
group. Also, the strong band in the range 3250–3315 cmϪ1 in the
IR spectra of these compounds (taken as chloroform solutions)
is consistent with the presence of a hydroperoxy group.19 In
order to confirm this assumption, the chemical evidence for the
presence of a hydroperoxy group in 3 was required.
1
In the H NMR spectra, the similar positions for the 7-H
signal (δ 4.60 in 3a, δ 4.58 in 9a; δ 4.72 in 3b, δ 4.69 in 9b) as well
as a similar splitting pattern for this signal (for example: both
triplets, J 7.8 and 7.6 Hz for 3a and 9a, respectively) are indic-
ative of the same size and conformation of ring B in com-
pounds 3 and 9.
C8H17
C8H17
Moreover, the formation of 5α-hydroperoxy compounds
in the course of ozonolysis of ∆5-steroids was found to be
independent of their substitution at position 3. This was proved
when cholest-5-ene derivatives 2b–2e were treated with ozone in
chloroform–methanol solution. As a result, the respective
hydroperoxides 3e–3g were isolated. Similarly, cholest-5-ene 2e
gave the crystalline hydroperoxide 3h in 64% yield. All the
O
O
AcO
O
CHO
O
1
7a 3β-OAc
b 3β-OH
c ∆3
8
compounds gave IR and H NMR spectra characteristic of
5-hydroperoxy-7α-alkoxy-6-oxa B-homosteroids (see Experi-
mental).
The final evidence for the structure of ozonation products 3
was obtained from X-ray analysis of the selected compounds:
3a, 3c and 3j (Table 1). Since crystals of 3a were poorly diffract-
ing and crystals of 3c showed some signs of disorder (see
Experimental), it was hoped that steroidal hydroperoxides with
the shorter side chain would give crystals suitable for satis-
factory measurements. Therefore, ozonolyses of 3β-acetoxy-
pregn-5-ene 2f and of 3β-acetoxyandrost-5-ene 2g were carried
out to give hydroperoxides 3i, and 3j, respectively. Of these two
compounds, hydroperoxide 3j gave crystals suitable for X-ray
studies. Thus, two crystal structure analyses were fully com-
pleted, i.e. for compounds 3c and 3j.
Both compounds 3c and 3j crystallize in the orthorhombic
system, the former with one, and the latter with two molecules
in the asymmetric unit. The skeleton geometry of the three
independent molecules does not differ significantly. Their close
conformational similarity can be seen from Table 2 which lists
endocyclic torsion angles. The perspective view of one of the
two crystallographically independent molecules of compound
3j is shown in Fig. 1. The stereochemistry of the four-ring
skeleton is trans–transoid–trans–transoid–trans. The six-
membered A and C rings have chair conformations slightly
flattened at the junction with the seven-membered B ring. In
both rings mirror symmetry dominates; in the A rings, about
the planes through C(3) and C(10), and in the C rings about the
planes through C(9) and C(13). Independent of the presence
or lack of substituent at C(17), the five-membered D rings have
a conformation intermediate between a 13β-envelope and
13β,14α half-chair. The side chain in 3c has an extended con-
formation and shows some signs of disorder at its end. In fact
C(24) has been refined in two alternative positions, labelled as
C(24) and C(24B), with the ratio of site-occupation factors
0.70:0.30, respectively.
C8H17
C8H17
O
+
AcO
AcO
O
OH
O
OR
O–
9 a R = Me
b R = Et
13
The reduction of 3a with lithium aluminium hydride in
diethyl ether at Ϫ60 ЊC resulted in formation of the known
B-seco steroids 7a–7c,8,9,12,20 while the reaction of 3a with
potassium iodide in acetic acid gave 7a quantitatively. Treat-
ment of 3a with trifluoroacetic acid in a benzene–hexane
mixture at 0 ЊC resulted in formation of a mixture, from which
the ozonide12 8 was isolated by chromatography (15% yield).
The ozonide 8 was quite stable when pure and its spectral prop-
erties were in accordance with those published.12 The formation
of 8 suggested the 5α-configuration of the hydroperoxy group
in 3a. It was expected that mild reduction of the hydroperoxides
3 would enable the isolation of the respective hydroxy deriv-
atives 9 without cleavage of the ring B. Indeed, brief treatment
of the methoxy-hydroperoxide 3a with dimethyl sulfide without
solvent gave, after evaporation of the excess of Me2S, a product
of higher polarity [RF = 0.25 as compared with RF = 0.51 of 3a
(in benzene–ethyl acetate, 5:1)], which decomposed to the seco
compound (7a) on contact with silica gel. Also, addition of
hydroxylic solvents (H2O, ROH, AcOH) transformed it immedi-
ately to 7a. However, when the crude reduction product was
dissolved in ethanol-free [2H3]chloroform, it gave a H NMR
1
spectrum perfectly consistent with structure 9a. It showed the
following characteristic signals: a singlet for the methoxy (δ
3.39), a singlet for the hydroxy (δ 3.00, the signal disappears
Of special interest in the two structures (three molecules) are
the seven-membered B rings. Their conformation can be
described as a twist chair with a twofold symmetry axis passing
2314
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
Table 1 Crystal data and details of X-ray measurements
3a
3c
3j
Empirical formula
Formula mass
Temperature/K
Wavelength/Å
Crystal system
Space group
C30H52O6
508.1
293
Cu-Kα (1.541 78)
Monoclinic
P21
C32H56O6
536.2
293
Cu-Kα (1.541 78)
Orthorhombic
P212121
C22H36O6
396.5
293
Cu-Kα (1.541 78)
Orthorhombic
P212121
Unit cell dimensions/Å
a = 18.976(4)
b = 6.398(1)
c = 25.519(5)
β = 102.47(2)Њ
3025(2)
a = 10.131(2)
b = 10.494(3)
c = 30.036(8)
—
3193(2)
4
a = 6.490(1)
b = 21.126(4)
c = 31.319(6)
Volume/Å3
4294(2)
8
1.227
Z
4
Dc/g cmϪ3
1.117
0.57
1.115
0.59
Absorption coefficient/mmϪ1
Crystal size/mm
2θmax for data collection
Index ranges
0.68
0.62 × 0.05 × 0.01
120Њ
Ϫ21 р h р 21
0 р k р 7
0 р l р 28
4940
0.32 × 0.24 × 0.12
120Њ
Ϫ11 р h р 11
0 р k р 11
0 р l р 33
4607
0.62 × 0.22 × 0.17
127Њ
Ϫ7 р h р 7
0 р k р 24
0 р l р 36
7013
Reflections collected
Observed reflections [F > 4σ (F)]
Extinction parameter
Weighting scheme
866
1657
3959
not applied
w = 1/[σ2 (Fo)2 ϩ (0.1151P)2]
P = (maxFo2 ϩ 2Fc2)/3
0.805
2.6(2) × 10Ϫ3
w = 1/[σ2 (Fo)2 ϩ (0.0858P)2]
P = (maxFo2 ϩ 2Fc2)/3
0.988
Goodness-of-fit
Final R indices
R1 = 0.067
wR2 = 0.160
Ϫ0.16
R1 = 0.048
wR2 = 0.125
Ϫ0.16
Largest diff. peak
and hole/e ÅϪ3
0.20
0.20
Table 2 Endocyclic torsion angles (Њ)
Compound 3
Compound 2
Molecule I
Molecule II
Ring A
C(10)᎐C(1)᎐C(2)᎐C(3)
C(1)᎐C(2)᎐C(3)᎐C(4)
C(2)᎐C(3)᎐C(4)᎐C(5)
C(3)᎐C(4)᎐C(5)᎐C(10)
C(4)᎐C(5)᎐C(10)᎐C(1)
C(5)᎐C(10)᎐C(1)᎐C(2)
Ϫ54.7(9)
54.3(8)
Ϫ55.4(8)
56.1(8)
Ϫ50.2(8)
50.4(8)
Ϫ54.3(5)
54.5(4)
Ϫ56.3(5)
56.2(4)
Ϫ51.1(4)
50.3(4)
Ϫ55.9(5)
54.9(5)
Ϫ55.0(4)
54.9(4)
Ϫ50.8(4)
51.6(4)
Ring B
C(10)᎐C(5)᎐O(6)᎐C(7)
C(5)᎐O(6)᎐C(7)᎐C(7A)
O(6)᎐C(7)᎐C(7A)᎐C(8)
C(7)᎐C(7A)᎐C(8)᎐C(9)
C(7A)᎐C(8)᎐C(9)᎐C(10)
C(8)᎐C(9)᎐C(10)᎐C(5)
C(9)᎐C(10)᎐C(5)᎐O(6)
Ϫ89.9(7)
37.8(8)
40.6(9)
Ϫ79.7(7)
67.8(7)
Ϫ58.6(7)
76.8(7)
Ϫ89.1(4)
40.0(4)
40.4(4)
Ϫ82.1(4)
68.5(4)
Ϫ54.4(4)
71.2(3)
Ϫ89.2(4)
41.2(4)
39.1(4)
Ϫ82.4(4)
69.7(4)
Ϫ54.4(4)
69.7(4)
Ring C
Ring D
C(9)᎐C(8)᎐C(14)᎐C(13)
C(8)᎐C(14)᎐C(13)᎐C(12)
C(14)᎐C(13)᎐C(12)᎐C(11)
C(13)᎐C(12)᎐C(11)᎐C(9)
C(12)᎐C(11)᎐C(9)᎐C(8)
C(11)᎐C(9)᎐C(8)᎐C(14)
53.2(8)
Ϫ59.7(8)
58.1(8)
Ϫ56.5(9)
48.3(8)
55.0(4)
Ϫ62.0(4)
57.4(4)
Ϫ53.1(5)
46.4(4)
52.6(4)
Ϫ61.9(4)
59.3(5)
Ϫ52.9(5)
42.6(5)
Ϫ43.7(7)
Ϫ44.3(4)
Ϫ41.0(4)
C(13)᎐C(14)᎐C(15)᎐C(16)
C(14)᎐C(15)᎐C(16)᎐C(17)
C(15)᎐C(16)᎐C(17)᎐C(13)
C(16)᎐C(17)᎐C(13)᎐C(14)
C(17)᎐C(13)᎐C(14)᎐C(15)
Ϫ32.2(7)
6.4(7)
21.0(7)
Ϫ40.1(7)
45.1(7)
Ϫ30.3(4)
5.6(5)
20.9(4)
Ϫ39.4(4)
43.1(4)
Ϫ32.6(4)
7.6(4)
20.1(5)
Ϫ39.7(4)
44.7(4)
through C(7) and the midpoint of the C(10)᎐C(9) bond. Such a
conformation reduces to its lowest value the strain energy
associated with the trans fusion of the 6-membered rings at
C(5)᎐C(10) and C(8)᎐C(9), and the presence of the substituent
at C(7).21 It is additionally stabilised by an intramolecular
hydrogen bond between the hydroperoxy hydrogen and the oxy-
gen atom of the alkoxy substituent at C(7). This internal hydro-
gen bond closes another seven-membered ring whose conform-
ation approximates to a twist-boat form, with an approximate
twofold axis passing through C(5) and the midpoint of the
hydrogen bridge H(5) ؒ ؒ ؒ O(7). Intramolecular hydrogen-bond
parameters are listed in Table 3, and the conformation of the
two bridged seven-membered rings is illustrated in Fig. 2. The
internal hydrogen bond must have a stabilizing effect on
molecular conformation, since it is maintained in the solid state,
with no intermolecular hydrogen bonds being present. Presum-
ably, it might also be responsible for the relative stability of this
group of hydroperoxides.
J. Chem. Soc., Perkin Trans. 2, 1997
2315

View Article Online
Table 3 Geometry of the intramolecular O5A–H5 ؒ ؒ ؒ O7 hydrogen
C8H17
bondsa
d(D ؒ ؒ ؒ A)/
Å
d(H ؒ ؒ ؒ A)/
Å
Compound
ЄD᎐H ؒ ؒ ؒ A/Њ
2
3
3
2.780(6)
Molecule I 2.717(3)
Molecule II 2.679(4)
1.98
1.94
1.83
138
158
150
AcO
2a
O3
a D᎐H bond lengths have been standardized to a value 0.97 Å.
+
AcO
AcO
O
O
O–
O
O
O
11
10
C8H17
O
O
AcO
AcO
+
Fig. 2 Perspective view of the ring B of the hydroperoxide 3j with the
hydrogen-bonded 5α-hydroperoxy group
O
O
OMe
O–
OH
3a
MeOH
Seven-membered carbocycles occur in many classes of natural
products, and most commonly they are fused to at least one
additional ring. Seven-membered ε-lactone rings are present
in brassinosteroids, the recently characterised plant growth-
promoting steroids22 and in kaurane type diterpenoids.23 In the
few crystal structures reported, the atomic coordinates of which
are available through the Cambridge Structural Database,24 the
seven-membered ε-lactone ring invariably adopts the chair con-
formation, with an approximate mirror plane passing either
through C(10)23a or C(9)22c,23d carbon atoms.
12
Scheme 1
and oligomeric structures proposed for products obtained dur-
ing ozonation of cholesterol in non-participating solvents12
should be revised accordingly, since all these compounds show,
in their 1H NMR spectra, a lowfield proton signal characteristic
of a hydroperoxy group.
The isolation of both types of products, a hydroperoxide and
a hydroxyperoxide, from ozonation of cyclobutene19b and
indene25 derivatives has been reported. The double bond in
these substrates is incorporated into four- or five-membered
rings and the resulting hydroperoxides are tetrahydrofuran and
isochroman derivatives, respectively. The present work also
shows that seven-membered, solvent-participated products are
effectively formed in the course of ozonolysis of rigid cyclo-
hexene derivatives.
The formation of the hydroperoxides 3 is rationalized in
terms of the Criegee mechanism.5 Furthermore, the unequiv-
ocal evidence for the structure 3 accounts for the revision of the
mechanism of ozonolysis of cholesterol and other ∆5-steroids in
alcohol-containing solvents (Scheme 1). The primary ozonide
10 which, for steric reasons, has most probably the 5α stereo-
chemistry, is the precursor of the more stable, tertiary carbonyl
oxide intermediate 11. The intramolecular partial capture of 11
by the 6-carbonyl oxygen gives the dipolar intermediate 12. The
reaction with methanol then follows, resulting in 3a. Thus, the
previously claimed secondary carbonyl oxide 139 is not an
intermediate in the ozonolysis of cholesterol. The formation of
the 7α-alkoxy derivatives 3a–3d, irrespective of the bulkiness
of the nucleophilic alcohol, suggests that directed25 α-axial
incorporation of the alkoxy group takes place. Otherwise, the
isomer of 3 with the equatorial 7β-alkoxy group would be
expected to arise, possibly as a major product. It was also
shown, that hydroperoxides 3 are not formed from 1,2,4-
trioxolane 8, as this compound is stable in methanol solution at
room temperature.
Experimental
Mp values were determined on a Kofler hot-stage apparatus
and are uncorrected. IR spectra were determined with a Perkin-
Elmer 580 grating spectrophotometer for solutions in chloro-
form. H and 13C NMR spectra were recorded with a JEOL
1
FX90Q or Varian Gemini 300 VT spectrophotometers oper-
ating in the Fourier transform mode using solutions in deut-
eriochloroform. Coupling constants J are given in Hz, and the
chemical shifts (δ) are expressed in ppm relative to tetramethyl-
silane. Electron impact mass spectra were recorded with a JEOL
LMS D-100 spectrometer. Column chromatography was per-
formed by using silica gel 60 (Merck 70–230 mesh, no. 7734).
The progress of reactions was monitored by TLC using a
precoated aluminium-backed silica plates (Merck, no. 5554).
X-Ray analysis and structure refinement
Single crystals of 3a were obtained from the isopropyl alcohol–
diisopropyl ether solution, and 3c and 3j were grown from
methanol. In general, it was difficult to obtain crystals of a
quality suitable for X-ray measurement. Crystals of 3a were
poorly diffracting and although we were able to solve the struc-
ture, it was apparent that the number of observed reflections
was too low to successfully refine it. Therefore, further refine-
ment of compound 3a was not undertaken. X-Ray data for
compound 3c were of higher quality, but it soon became appar-
ent that some of the atoms of the side chain at C(17) showed
large displacement parameters, suggesting positional disorder.
Consequently, C(24) was refined in two alternative positions
labelled as C(24) and C(24B) with a partial occupancy ratio
70:30, respectively. The crystal structure of hydroperoxide 3j,
which lacks the side chain, was obtained with the highest pre-
cision. Crystal data and details of the X-ray measurements are
summarized in Table 1. For all three compounds the reflection
In view of the results presented, all structures 1 assigned to
the principal products obtained in the ozonolyses of cholesterol
in water or alcohols reported previously9,10,11,13 should be re-
vised. Hence, isomeric structures 3 with hydroperoxy group are
correct for these compounds. It also appears, that the dimeric
2316
J. Chem. Soc., Perkin Trans. 2, 1997

View Article Online
intensities were measured on a four-circle KM-4 (KUMA
Diffraction)26 diffractometer equipped with a graphite mono-
chromator. The cell constants and the orientation matrix were
obtained from a least-squares fit of at least 25 centred reflec-
tions. The intensities were measured using the ω–2θ scan tech-
nique with variable scan rate, and a scan range ω 0.75–1.45Њ.
Background measurements were estimated from a 68-step pro-
file. The intensities were corrected for Lorentz and polarisation
effects; absorption corrections were not applied. The structures
were solved by direct methods with SHELXS-8627 and refined
with SHELXL-93.28 Refinement was completed only for
structures 3c and 3j. Heavy atoms (C, O, N) were refined
anisotropically. The positions of the H-atoms attached to the
C-atoms were calculated, with the exception of the OH
H-atoms which were located from difference Fourier maps. All
H atoms were refined using a riding model with an isotropic
temperature factor 1.2 times Ueq of the atom to which they are
bonded. A Siemens Stereochemical Workstation was used to
prepare the drawings.29 Details of the present X-ray analyses
and crystal data are collected in Table 1.§
3â-Ethoxy-5á-hydroperoxy-7á-methoxy-5á-B-homo-6-oxa-
cholestane 3f. (66% yield) mp 137–139 ЊC (from Et2O–MeOH);
νmax/cmϪ1 3280, 1100, 1050, 1035 and 1010; δH 10.00 (1 H, s,
5α-OOH), 4.61 (1 H, t, J 7.8, 7β-H), 3.49 (3 H, s, 7-OCH3), 3.38
(1 H, m, 3α-H), 2.70 (1 H, m, 4α-H), 1.20 (3 H, t, OCH2CH3),
1.05 (3 H, s, 10-Me) and 0.65 (3 H, s, 13-Me); m/z 494, 476,
445, 315 and 263 (Found: C, 73.0; H, 10.9. C30H54O5 requires
C, 72.83; H, 11.00).
5á-Hydroperoxy-7á-methoxy-5á-B-homo-6-oxacholestane-3-
one 3g. (57% yield) mp 124–127 ЊC (from EtOH); νmax
/
cmϪ1 3260, 1720, 1048 and 1030; δH 10.13 (1 H, s, 5α-OOH),
4.55 (1 H, m, 7β-H), 3.48 (3 H, s, OMe), 3.29 (1 H, dd, J
15.6, J 2.1, 4α-H), 2.61 (1 H, d, J 15.6, 4β-H), 1.24 (3 H, s,
10-Me) and 0.69 (3 H, s, 13-Me); m/z 464, 446, 363, 263 and
135 (Found: C, 72.3; H, 10.2. C28H48O5 requires C, 72.36; H,
10.42).
5á-Hydroperoxy-7á-methoxy-5á-B-homo-6-oxacholestane 3h.
(65% yield) mp 127–129 ЊC (from MeOH); νmax/cmϪ1 3220,
1140, 1030 and 1010; δH 9.59 (1 H, s, 5α-OOH), 4.59 (1 H,
m, 7β-H), 3.48 (3 H, s, OMe), 2.30 (1 H, m, 4α-H), 1.04 (3 H, s,
10-Me) and 0.66 (3 H, s, 13-Me); m/z 450, 432, 401 and
263 (Found: C, 74.4; H, 11.3. C28H50O4 requires C, 74.62; H,
11.18).
The general procedure for ozonolysis of steroidal olefins in an
alcohol–chloroform solution
3â-Acetoxy-5á-hydroperoxy-7á-methoxy-5á-B-homo-6-oxa-
The steroid olefin (2a–2g) was dissolved in a mixture of
chloroform–alcohol (ROH) (in most cases 1:1) to get an
approximately 1% solution, that was cooled to Ϫ78 ЊC. Ozone
was passed through this solution until the substrate dis-
appeared (TLC test). The excess ozone was removed with a
stream of oxygen. Solvents were evaporated under reduced pres-
sure and the residue was recrystallized. The following hydro-
peroxides were obtained [the yield of the chromatographically
pure (TLC) product is given].
3â-Acetoxy-5á-hydroperoxy-7á-methoxy-5á-B-homo-6-oxa-
cholestane 3a. (56% yield) mp 150–153 ЊC (from MeOH) (lit.,9
mp 143–145 ЊC); the IR, 1H NMR and 13C NMR spectra were
in accordance with the data given in ref. 9.
3â-Acetoxy-5á-hydroperoxy-7á-ethoxy-5á-B-homo-6-oxa-
cholestane 3b. (70% yield) mp 132–136 ЊC (from Et2O–MeOH)
(lit.,9 mp 138-140 ЊC); the IR, 1H NMR and 13C NMR spectra
were in accordance with the data given in ref. 9.
3â-Acetoxy-5á-hydroperoxy-7á-isopropoxy-5á-B-homo-6-
oxacholestane 3c. (55% yield) mp 133–135 ЊC (from ethanol);
νmax/cmϪ1 3250, 1725; δH 10.32 (1 H, s, 5α-OOH), 4.83 (3 H, m,
3α-H), 4.79 (1 H, m, 7β-H), 4.05 (1 H, septet, O-CHMe2), 2.62
(1 H, m, 4α-H), 2.03 (3H, s, OAc), 1.06 (10-Me), 0.65 (13-Me);
δC 69.8 (C-3), 98.6 (C-7), 111.4 (C-5), 170.1 (CH3COO); m/z
536, 518, 459, 399, 291 (Found: C, 71.5; H, 10.6. C32H56O6
requires C, 71.60; H. 10.52%).
pregnane 3i. (56% yield) mp 127–130 ЊC (from MeOH); νmax
/
cmϪ1 3250 and 1720; δH 10.04 (1 H, s, 5α-OOH), 4.90 (3 H, m,
3α-H), 4.62 (1 H, t, J 7.9, 7β-H), 3.48 (3 H, s, OMe), 2.68 (1 H,
m, 4α-H), 2.03 (3 H, s, Ac), 1.08 (3 H, s, 10-Me) and 0.55 (3 H, s,
13-Me) (Found: C, 67.8; H, 9.2. C24H40O6 requires C, 67.89; H,
9.50).
3â-Acetoxy-5á-hydroperoxy-7á-methoxy-5á-B-homo-6-oxa-
androstane 3j. (65% yield) mp 138–142 ЊC (from MeOH-
CHCl3); νmax/cmϪ1 3250, 1720, 1250, 1140, 1040 and 1020; δH
10.03 (1 H, s, 5α-OOH), 4.90 (1 H, m, 3α-H), 4.62 (1 H, t, J 7.5,
7β-H), 3.8 (3 H, s, OMe), 2.67 (1 H, m, 4α-H), 2.03 (3 H, s,
OAc), 1.08 (3 H, s, 10-Me) and 0.70 (3 H, s, 13-Me); δC 69.7
(C-3), 102.3 (C-7), 111.5 (C-5) and 170.0 (CH3COO); m/z 396,
378, 347, 287, 243 and 151 (Found: C, 66.4; H, 9.4. C22H36O6
requires C, 66.64; H, 9.15).
Reduction of 3a with lithium aluminium hydride
To a stirred solution of 3a (187 mg) in anhydrous diethyl ether
(100 ml) under argon, lithium aluminium hydride (18 mg) was
added at Ϫ60 ЊC. After 2 h, methanol (30 ml) was added. The
usual workup gave the residue (169 mg), which showed three
1
spots on TLC. The H NMR spectrum indicated the presence
of compounds 7a,20 7b8,12 and 7c20b in the ratio 5:2:3, as estim-
ated from the integration of characteristic low-field signals at
δ 9.61, 6.77, 5.82, 5.38 and 4.51.
3â-Acetoxy-5á-hydroperoxy-7á-tert-butoxy-5á-B-homo-6-
oxacholestane 3d. (60% yield) mp 185–189 ЊC (from benzene–
hexane); νmax/cmϪ1 3280, 1720; δH 10.73 (1 H, s, 5α-OOH), 5.00
(1 H, m, 7β-H), 4.91 (1 H, m, 3α-H), 2.63 (1 H, m, 4α-H), 2.02
(3 H, s, Ac), 1.26 (9 H, s, But), 1.05 (3 H, s, 10-Me), 0.65 (3 H,
s, 13-Me); δC 69.9 (C-3), 94.6 (C-7) and 111.3 (C-5) (Found:
C, 72.1; H, 10.4. C33H58O6 requires C, 71.96; H, 10.61%).
3â-Chloro-5á-hydroperoxy-7á-methoxy-5á-B-homo-6-
oxacholestane 3e. (58% yield) mp 111–114 ЊC (from Et2O–
MeOH); νmax/cmϪ1 3270, 1190 and 1030; δH 10.05 (1 H, s,
5α-OOH), 4.61 (1 H, t, J 7.9, 7β-H), 4.03 (1 H, m, 3α-H),
3.48 (3 H, s, 7-OCH3), 2.85 (1 H, m, 4α-H), 1.07 (3 H, s, 10-Me)
and 0.65 (3 H, s, 13-Me); m/z 467, 452, 434, 398 and 249
(Found: C, 69.5; H, 10.0. C28H49ClO4 requires C, 69.32; H,
10.18).
Reaction of 3a with potassium iodide
To a solution of 3a (85 mg) in acetic acid (5 ml) a solution of
potassium iodide (30 mg) in AcOH (1 ml) was added and the
deep colour of iodine developed quickly. After 0.5 h diethyl
ether–benzene mixture (1:1, 10 ml) was added and the solution
was washed with aq. NaHSO3 (5%), aq. Na2CO3 (5%) and
brine. The organic layer was dried (MgSO4) and evaporated to
give chromatographically pure 7a (68 mg). The H NMR and
IR (in CCl4) spectra were in accordance with the data given in
ref. 20.
1
Reaction of 3a with trifluoroacetic acid
To a solution of the hydroperoxide 3a (830 mg) in hexane–
benzene 1:12 mixture (6 ml), trifluoroacetic acid (1 ml) was
added and the reaction mixture was kept at 0 ЊC for 12 h. The
solution was washed with cold aq. NaOH (5%), water, dried
(MgSO4) and evaporated. The residue was chromatographed on
silica gel (40 g) with chloroform as eluent to give 116 mg of the
§ Atomic coordinates, anisotropic displacement parameters and tables
of bond distances and angles for compounds 3c and 3j have been
deposited at the Cambridge Crystallographic Data Centre. For details,
see ‘Instructions for Authors’, J. Chem. Soc., Perkin Trans. 2, 1996,
Issue 1. Any request to the CCDC for this material should quote the
full literature citation and the reference number, 188/93.
1
semicrystalline ozonide 8. Its H NMR, IR and mass spectra
were well in accordance with the literature.12
J. Chem. Soc., Perkin Trans. 2, 1997
2317

View Article Online
L. Brown, W. J. S. Lyall, C. J. Suckling and K. E. Suckling, J. Chem.
Soc., Perkin Trans. 1, 1987, 595.
Reduction of hydroperoxide 3a with dimethyl sulfide to the cyclic
hemiacetal 9a
12 K. Jaworski and L. L. Smith, J. Org. Chem., 1988, 53, 545.
13 K. Jaworski and L. L. Smith, Magn. Reson. Chem., 1988, 26, 104.
14 H. Kropf and H. von Wallis, Synthesis, 1981, 237.
15 D. Swern, A. H. Clements and T. M. Luong, Anal. Chem., 1969, 41,
412.
16 T. S. Lillie and R. C. Ronald, J. Org. Chem., 1985, 50, 5084.
17 Z. Paryzek, J. Martynow and W. Swoboda, J. Chem. Soc., Perkin
Trans. 1, 1990, 1220.
Into an NMR tube compound 3a (20 mg) and Me2S (1 ml) were
added. After 0.5 h the solvent was removed under reduced pres-
sure, dry CDCl3 or C6D6 (0.5 ml) was added and the 1H NMR
spectrum recorded. It showed the following signals of 9a:
δH(CDCl3) 4.90 (1 H, m, 3α-H), 4.58 (1 H, t, J 7.7, 7β-H), 3.39
(3 H, s, OCH3), 3.00 (1 H, s, 5-OH, disappears on addition of
D2O), 0.99 (3 H, s, 10-CH3), 0.66 (3 H, s, 13-CH3); δH(C6H6)
5.21 (1 H, m, 3α-H), 4.55 (1 H, dd, J 10, J 8.1, 7β-H), 3.62 (1 H,
s, OH), 3.26 (3 H, s, OMe), 1.76 (3 H, s, AcO), 1.02 (s, 10-Me),
0.67 (s, 13-Me); νmax/cmϪ1 (in CCl4) 3555, 3420, 1732 and 1237;
m/z 460 (Mϩ Ϫ MeOH), 354, 318, 247 and 110.
18 Z. Paryzek, J. Martynow and W. Swoboda, J. Chem. Soc., Perkin
Trans. 1, 1990, 1222.
19 (a) R. E. Keay and G. A. Hamilton, J. Org. Chem., 1973, 39, 3604;
(b) K. Griesbaum and G. Kiesel, Chem. Ber., 1989, 122, 145.
20 (a) K. Tanabe and Y. Morisawa, Chem. Pharm. Bull., 1963, 11,
536; P. Morand and M. Kaufman, J. Org. Chem., 1969, 34, 2175;
(b) P. Yates and S. Stiver, Can. J. Chem., 1988, 66, 1209.
21 J. B. Hendrickson, Tetrahedron, 1963, 19, 1387.
22 (a) M. J. Thompson, N. Mandava, J. L. Flippen-Anderson,
J. F. Worley, S. R. Dutky, W. E. Robbins and W. Lusby, J. Org.
Chem., 1979, 44, 5002; (b) M. D. Grove, G. F. Spencer,
W. K. Rohwedder, N. Mandava, J. F. Worley, J. D. Warthen, Jr.,
G. L. Steffens, J. L. Flippen-Anderson and J. C. Cook, Jr., Nature
(London), 1979, 281, 216; (c) L. Kutschabsky, G. Adam and H.-M.
Vorbrodt, Z. Chem., 1990, 30, 136.
23 (a) M. Shaiq Ali, M. K. Baynham, J. R. Hanson and P. B.
Hitchcock, J. Chem. Soc., Perkin Trans. 1, 1991, 2679; (b) M. K.
Baynham, J. M. Dickinson, J. R. Hanson and P. B. Hitchcock,
J. Chem. Soc., Perkin Trans. 1, 1987, 1987; (c) F. Nagashima,
H. Tanaka, S. Takaoka and Y. Asakawa, Chem. Pharm. Bull., 1994,
42, 2656; (d) A. R. H. Kehrli, D. A. H. Taylor and M. Niven,
J. Chem. Soc., Perkin Trans. 1, 1990, 2057.
24 3D Search and Research using the Cambridge Structural Database,
F. H. Allen; O. Kennard, Chemical Design Automation News, 1993,
8, 131.
25 K. J. McCullough, M. Nojima, M. Miura, T. Fujisaka and
S. Kusabayashi, J. Chem. Soc., Chem. Commun., 1984, 35; K. J.
McCullough, T. Fujisaka, M. Nojima and S. Kusabayashi,
Tetrahedron Lett., 1988, 29, 3375; N. Nakamura, T. Fujisaka,
M. Nojima, S. Kusabayashi and K. J. McCullough, J. Am. Chem.
Soc., 1989, 111, 1799.
26 DATARED, KUMA Diffraction, Wrocław, Poland, 1989.
27 G. M. Sheldrick, SHELXS-86, University of Göttingen,1986.
28 G. M. Sheldrick, SHELXL-93, University of Göttingen, 1993.
29 Stereochemical Workstation Operational Manual, Release 3.4
Siemens Analytical X-Ray Instruments, Inc., Medison, Wisconsin,
USA, 1989.
Reduction of hydroperoxide 3b with dimethyl sulfide to the cyclic
hemiacetal 9b
Compound 9b was obtained according to the procedure
described in the preceding experiment: νmax/cmϪ1 (in CCl4)
3555, 3400, 1735 and 1242; δH 4.96 (1 H, m, 3α-H), 4.69 (1 H, t,
J 8.1, 7α-H), 3.86 and 3.46 (2 H, two q, J 7.2, 7-OCH2CH3),
2.98 (1 H, s, 5-OH, disappears on addition of D2O), 1.20 (3 H,
t, J 7.2, 7-OCH2CH3), 0.99 (3 H, s, 10-Me), 0.66 (3 H, s, 13-Me);
m/z 460 (Mϩ Ϫ EtOH), 354, 318 and 110.
References
1 E. P. C. Kilsdonk, D. W. Morel, W. J. Johnson and G. H. Rothblat,
J. Lipid Res., 1995, 36, 505 and references cited therein.
2 M. F. Grey, T. D. V. Lawrie and C. J. W. Brooks, Lipids, 1971, 6, 836;
N. B. Javitt, E. Kok, S. Burstein, B. Cohen and J. Kutscher, J. Biol.
Chem., 1981, 256, 12 644; N. Kumar and O. P. Singhal, J. Sci. Food
Agric., 1991, 55, 497; L. L. Smith and B. H. Johnson, Free Rad. Biol.
Med., 1989, 7, 285.
3 X. Pannecoucke, J.-P. Starck, A. Milon, G. Ourisson and B. Luu,
Bull. Soc. Chim. Fr., 1994, 131, 693 and references cited therein.
4 L. L. Smith, J. Liquid Chromatogr., 1993, 16, 1731 and references cited
therein.
5 R. Criegee and G. Wenner, Chem. Ber., 1949, 564, 9; R. Criegee,
Angew. Chem., Int. Ed. Engl., 1975, 11, 745.
6 P. S. Bailey in Ozonation in Organic Chemistry, Academic Press, New
York, 1978, vol. 1; 1982, vol. 2.
7 W. H. Bunnelle, Chem. Rev., 1991, 91, 335 and references cited
therein.
8 J. W. Cornforth, G. D. Hunter and G. Popjak, Biochem. J., 1953, 54,
590.
9 J. Gumulka and L. L. Smith, J. Am. Chem. Soc., 1983, 105, 1972.
10 H. Lettre and A. Jahn, Angew Chem., 1957, 69, 266.
11 K. Tanabe and Y. Morisawa, Bull. Chem. Soc. Jpn., 1963, 11, 536;
Paper 7/02633A
Received 17th April 1997
Accepted 12th June 1997
2318
J. Chem. Soc., Perkin Trans. 2, 1997