Allylic oxidation of steroidal olefins by vanadyl acetylacetonate and tert-butyl hydroperoxide

Article abstract of DOI:10.1016/j.steroids.2015.06.005

Abstract Readily available vanadyl acetylacetonate was found to oxidize the allylic sites of Δ⁵ steroidal alcohols without protection of hydroxyl groups. Cholesterol, dehydroepiandrosterone, cholesterol benzoate, cholesterol acetate, pregnenolone, and 5-pregnen-3,20-diene were oxidized to 7-keto products using vanadyl acetylacetonate in one pot reactions at room temperature in the presence of oxygen and water.

Full text of DOI:10.1016/j.steroids.2015.06.005

Steroids 101 (2015) 103–109
Contents lists available at ScienceDirect
Steroids
journal homepage: www.elsevier.com/locate/steroids
Allylic oxidation of steroidal olefins by vanadyl acetylacetonate
and tert-butyl hydroperoxide
⇑
Wendell S. Grainger, Edward J. Parish
Department of Chemistry and Biochemistry, College of Science and Mathematics, Auburn University, Auburn, AL 36849-5319, United States
a r t i c l e i n f o
a b s t r a c t
Readily available vanadyl acetylacetonate was found to oxidize the allylic sites of D5 steroidal alcohols
without protection of hydroxyl groups. Cholesterol, dehydroepiandrosterone, cholesterol benzoate,
cholesterol acetate, pregnenolone, and 5-pregnen-3,20-diene were oxidized to 7-keto products using
vanadyl acetylacetonate in one pot reactions at room temperature in the presence of oxygen and water.
Ó 2015 Elsevier Inc. All rights reserved.
Article history:
Received 7 April 2015
Received in revised form 14 May 2015
Accepted 9 June 2015
Available online 16 June 2015
Keywords:
Oxidation
7-Ketocholesterol
Vanadyl acetylacetonate
tert-Butyl hydroperoxide
1. Introduction
Steroidal oxidations are often proceeded by esterification of the
alcohol at the C3 carbon to prevent its oxidation [16]. Attempts to
Derivatives of dehydroepiandrosterone (DHEA) oxidation have
shown beneficial health related potential [1], such as boosting
memory [2] and the resting metabolic rate of overweight adults
[3]. Several derivatives of pregnenolone oxidation have also been
reported to inhibit undesired cortisone effects [4]. Even 7-ketoc-
holesterol, the major auto-oxidation product of cholesterol, has
shown anti-cancer potential [5,6]. With the goal of producing ben-
eficial steroidal derivatives, an innovative approach to a common
set of reagents, vanadyl acetylacetonate and tert-butyl hydroper-
oxide, has been successfully used and is reported here.
remove the ester, when possible, result in loss of product.
However, vanadium catalysts do not significantly oxidize hydroxyl
groups past initial complexation [13,17,18]. Without needing a
protecting group, oxidation via vanadium removes two entire steps
from total synthesis, increasing overall yield. Vanadyl acetylaceto-
nate (VO(acac)2), in particular, is also relatively inexpensive, safe,
shelf stable and was found to oxidize several allylic carbons of ster-
oidal olefins in the presence of oxygen and at room temperature
using tert-butyl hydroperoxide (TBHP) as the oxidant in benzene
solvent.
Vanadium catalysts are recognized for their epoxidation of ole-
fins [7–14]. The carbon–hydrogen oxidation that occurs simultane-
ously has mostly been discarded as a side reaction from free radical
formation [7,14]. However, carbon–hydrogen oxidation potential
was realized when several benzylic carbon were oxidized in high
yields using vanadium catalysts [15]. In the present study, we have
modified this approach and extended it to steroidal olefins. (We
explicitly acknowledge that others have achieved carbon-hydrogen
activation before us. Albeit, we are unaware of any studies, such as
the one presented in this report, focused on the allylic oxidation of
steroidal compounds using vanadium catalysts and TBHP.)
2. Experimental
2.1. Optimization
Reaction conditions were optimized using cholesterol as a pri-
mary substrate. Conversion was qualitatively determined by TLC
(silica gel on glass plates) and molybdic acid [19] after eluting with
50:50 ethyl acetate and toluene. Full conversion using minimal
concentrations of TBHP and VO(acac)2 was considered optimized.
2.2. General procedure
⇑
Corresponding author at: Department of Chemistry and Biochemistry, College
Substrate, 2.6 mmol (only 1.3 mmol for (5), (12), (13), and (14)
for lack of solubility), and 0.38 mmol of VO(acac)2 were put into a
round bottom flask. The flask was then charged with 25 mL of
of Science and Mathematics, Auburn University, 179 Chemistry Bldg., Auburn, AL
36849-5319, United States. Tel.: +1 334 844 4043; fax: +1 334 844 6959.
E-mail address: parisej@auburn.edu (E.J. Parish).
http://dx.doi.org/10.1016/j.steroids.2015.06.005
0039-128X/Ó 2015 Elsevier Inc. All rights reserved.

104
W.S. Grainger, E.J. Parish / Steroids 101 (2015) 103–109
benzene. After dissolving as much substrate as possible, 24 mmol
of TBHP were added. The reaction flask was loosely capped to pre-
vent loss of solvent and the reaction was allowed to run for 5 days.
At the end of 5 days, the benzene solvent was removed by evap-
oration. Residue was then extracted with deionized (DI) water and
ethyl ether into a separatory funnel. (If time is available, the pre-
cipitate will dissolve into the solution overnight.) The ether layer
was retained and the ether was evaporated under reduced pressure
at 55 °C.
Once the ether was removed, products (15) and (16) were dis-
solved in acetone and recrystallized with DI water. The suspension
was chilled for 5 h with an ice bath. Products (15) and (16) were
then filtered out, rinsed with DI water, and dried under reduced
pressure with P2O5. The filtrate from the first recrystallization of
(16) should be recrystallized again to extract more product.
Beyond work-up of (15) and (16), we advise against recrystalliza-
tion for the following products.
After removing the ether, products (17–20) were separated
with column chromatography (silica gel). Product (17) was
extracted with hexane into the column and eluted using a gradient
(0–40% ethyl acetate in hexane). Also, product (18) was extracted
with 20% ethyl acetate in hexane and eluted with a gradient (20–
50% ethyl acetate in hexane). Products (19) and (20) were
extracted into the column with 50:50 ethyl acetate and hexane
and were isocratically eluted with 50:50 ethyl acetate and hexane.
Fractions were identified by TLC, collected, and dried under
reduced pressure with P2O5. (We found that, particularly for elu-
tion of (17), it is better get the products through the column as
quickly as possible, without losing separation.)
2.4.5. 5-Androgen-3,17-diol (5)
5-Androgen-3,17-diol (5) was obtained through reduction of
DHEA using Luche reduction [21]. Cerium(III) Chloride heptahy-
drate, 7.5 g, was dissolved in 100 mL of methanol. DHEA, 4.59 g,
was then added and dissolved. To the mixture, 1 g of sodium boro-
hydride was slowly added. The solution was stirred overnight with
a magnetic stir bar. DI water was added to the mixture to quench
the reaction. The reaction mixture was then poured into a separa-
tory funnel. Ethyl ether was added to the separatory funnel to
make a second layer. The ether layer was then washed with DI
water and collected. TLC revealed only one band from the ether
layer. The ether was evaporated and the residue was used as sub-
strate. White solid. 13C NMR (100 MHz, CDCl3): d 140.9, 121.4, 81.9,
71.7, 51.3, 50.2, 42.3, 37.3, 36.6, 32.0, 31.5, 30.5, 23.5, 20.7, 19.5,
11.0. TLC showed one band.
2.4.6. 7-Dehydrocholesterol (6)
White solid. 13C (100 MHz, CDCl3): d 141.6, 140.0, 119.8, 116.5,
70.6, 56.1, 54.7, 46.4, 43.1, 41.0, 39.7, 39.4, 38.6, 37.2, 36.4, 36.3,
32.2, 28.3, 28.2, 24.1, 23.0, 22.8, 21.3, 19.1, 16.5, 12.0. TLC showed
one band and spectroscopic data matched expected values [22].
2.4.7. 7-Cholesten-3-ol (7)
7-Cholesten-3-ol (7) was obtained through hydrogenation of 7-
dehydrocholesterol. 7-Dehydrocholesterol, 15 g, was dissolved in
235 mL of ethyl acetate and 15 mL of glacial acetic acid (a 95:5 vol-
ume ratio respectively). The mixture was heated by steam bath and
kept warm throughout the hydrogenation process by heat lamps to
keep the 7-dehydrocholesterol soluble. Platinum dioxide, 1.5 g,
was added to the solution. The solution was then pressurized with
H2 gas (55–60 pounds) for 2 days. Upon completion of the 2 days,
the solution was filtered while warm to remove the catalyst. The
filtrate was then chilled causing 7-cholesten-3-ol to precipitate.
After filtering out the precipitate and rinsing it with DI water,
the 7-cholesten-3-ol was dissolved in acetone and recrystallized
with DI water. The precipitate was filtered out and rinsed with
DI water. 7-Cholesten-3-ol was then dried under reduced pressure.
White solid. 1H NMR (250 MHz, CDCl3): d 5.60, 5.42, 5.18, 3.63,
2.09, 1.81, 1.65, 1.56, 1.37, 1.28, 1.16, 0.92, 0.83, 0.65, 0.58. 13C
NMR (400 MHz, CDCl3): d 139.8, 117.6, 71.3, 56.4, 55.2, 49.6,
43.6, 40.4, 39.8, 39.7, 37.3, 36.4, 36.3, 34.4, 31.6, 29.9, 28.2, 28.2,
24.1, 23.2, 23.0, 22.8, 21.8, 19.1, 13.3, 12.1. 1H NMR showed this
starting material to be 74% 7-cholesten-3-ol and 24% 7-dehdroc-
holesterol. Spectroscopic data matched expected values [23].
2.3. Method of purity analysis
13C proton-decoupled NMR was used in conjunction with TLC to
confirm the purity of starting materials. For identification of oxida-
tion products, 1H NMR, 13C NMR, and mass spectroscopy (Q-TOF
and GCT) were utilized.
2.4. Compounds (1–14): substrates
2.4.1. Estrone (1)
White solid. 13C NMR (150 MHz, CDCl3): d 153.6, 138.1, 132.2,
126.6, 115.4, 112.9, 50.5, 48.2, 44.0, 38.5, 36.0, 31.6, 29.5, 26.6,
21.6, 14.0. TLC showed one band and spectroscopic data matched
expected values [20].
2.4.8. 24,25-dihydrolanosterol (8)
2.4.2. 4-Cholesten-3-one (2)
White solid. 13C NMR (100 MHz, CDCl3): d 200.0, 172.1, 123.9,
56.2, 56.0, 54.0, 42.5, 39.8, 39.6, 38.8, 38.8, 36.3, 35.9, 35.1, 35.7,
34.1, 33.1, 32.2, 28.3, 28.2, 24.3, 24.0, 23.0, 22.7, 21.2, 18.8, 17.5,
12.1. TLC showed one band and spectroscopic data matched
expected values [20].
24,25-dihydrolanosterol (8) was obtained through hydrogena-
tion of a lanosterol/24,25-dihydrolanosterol (40:60 respectively)
mixture that was commercially obtained. The mixture’s identity
and purity were confirmed by 1H NMR, 13C NMR, and TLC.
Lanosterol mixture, 10 g, was dissolved in 300 mL of glacial acetic
acid. To the solution, 1.75 g of a platinum dioxide and platinum
black (46:54 respectively) mixture was added. The lanosterol mix-
ture was warmed by steam bath and kept warm during the hydro-
genation process with heat lamps. To the mixture, 55–60 pounds of
H2 gas were applied for 17 h. The mixture was then filtered while
warm to remove the catalyst. Recrystallization occurred by cooling
the reaction solution and adding DI water. The precipitate was fil-
tered out and rinsed with water. It was then dissolved in acetone
and recrystallized with DI water. The precipitate was filtered out
and allowed to dry using P2O5 under reduced pressure. White solid.
1H (400 MHz, CDCl3): 3.25, 2.07, 1.72, 1.52, 1.36, 1.27, 1.00, 0.89,
0.82, 0.70, 0.58. 13C (62.5 MHz, CDCl3): 134.4, 130.9, 125.3, 79.0,
50.4, 79.8, 44.5, 38.9, 37.0, 35.6, 31.0, 30.9, 29.7, 28.0, 27.8, 26.5,
25.8, 24.3, 21.0, 19.2, 18.3, 17.7, 15.8, 15.5. TLC showed one band.
2.4.3. 1,4-Androstadiene-3,17-dione (3)
White solid. 13C NMR (62.5 MHz, CDCl3): d 186.4, 168.5, 155.5,
127.9, 124.3, 52.5, 50.6, 47.9, 43.6, 35.8, 35.3, 32.7, 32.5, 31.4,
22.3, 22.1, 18.9, 14.0. TLC showed one band and spectroscopic
value matched expected values [20].
2.4.4. Cholestan-3-ol (4)
White solid. 13C NMR (MHz 100, CDCl3): d 71.3, 56.5, 56.3, 54.4,
44.9, 42.6, 40.1, 39.5, 38.2, 37.0, 36.2, 35.8, 35.5, 35.5, 32.1, 31.5,
28.8, 28.3, 28.0, 24.2, 23.9, 22.9, 22.6, 21.3, 18.7, 12.3, 12.1. TLC
showed one band and spectroscopic data matched expected
values [20].

W.S. Grainger, E.J. Parish / Steroids 101 (2015) 103–109
105
2.4.9. Cholesteryl benzoate (9)
2.5. Compounds (15–20): products
2.5.1. 7-Ketocholesteryl benzoate (15)
Cholesteryl benzoate (9) was obtained by slowly adding 100 mL
of benzoyl chloride to 50 g of cholesterol dissolved in 200 mL of
pyridine at 40 °C for 3.5 h. That mixture was poured into chilled
DI water. Cholesteryl benzoate precipitated and was filtered out.
It was rinsed with DI water and dried. The precipitate was dis-
solved in warm chloroform. That solution was put into a freezer
for 5 h, during which time the cholesteryl benzoate precipitated.
The precipitate was filtered, rinsed with methanol, and dried.
White solid. 13C NMR (62.5 MHz, CDCl3): d 166.2, 139.8, 132.9,
131.0, 129.7, 128.4, 123.0, 74.8, 56.9, 56.3, 50.2, 42.5, 39.8, 38.7,
37.3, 36.2, 32.1, 28.3, 24.2, 23.2, 21.5, 19.6, 12.3. TLC showed one
band and spectroscopic data matched expected values [24].
White solid. 1H NMR (400 MHz, CDCl3): d 8.05 (d, 2H, J = 7.3 Hz),
7.57 (t, 1H, J = 7.4 Hz), 7.46 (t, 3H, J = 7.6 Hz), 5.75 (m, 1H), 4.98 (m,
1H), 2.75–0.97 (comp, 28H), 0.94 (d, 3H, J = 6.5 Hz), 0.87 (dd, 6H,
J = 6.7 Hz, J = 1.8 Hz), 0.70 (s, 3H). 13C NMR (100 MHz, CDCl3): d
202.1, 165.9, 164.0, 133.1, 130.3, 128.4, 127.0, 72.9, 54.8, 50.0,
49.8, 45.5, 43.1, 39.5, 38.4, 37.9, 36.2, 35.7, 28.6, 28.0, 26.3, 23.9,
22.9, 22.6, 18.9, 17.3, 12.0. HRMS (Q-TOF, ESI+) m/z: [M+1] calcu-
lated for C34H49O3 505.3682; found 383.3228. MS (GCT) m/z:
504.4295, 502.0367, 382.3056, 269.2161, 218.9884, 174.1001,
161.1125, 105.0313, 77.0457. (1.28 g, 98%) Spectroscopic data
matched expected values [24].
2.4.10. Cholesteryl acetate (10)
Cholesteryl acetate (10) was obtained by slowly adding 12 g of
acetic anhydride to 10 g of cholesterol dissolved in 200 mL of pyr-
idine at 0 °C. The reaction was allowed to warm to room tempera-
ture after 3 h and was left overnight with magnetic stirring. To
quench the reaction, DI water was added, also causing the choles-
teryl acetate to precipitate. The precipitate was filtered out, rinsed
with DI water, and dried. After drying, the solids were dissolved in
acetone and recrystallized by adding DI water. The precipitate was
filtered out, rinsed with DI water, and dried. White solid. 13C NMR
(62.5 MHz, CDCl3): d 170.5, 139.7, 122.7, 74.0, 56.7, 56.1, 50.0, 42.3,
39.5, 36.6, 35.8, 31.9, 31.9, 28.0, 24.3, 22.8, 22.6, 19.3, 18.7, 11.9.
TLC showed one band and spectroscopic data matched expected
values [20].
2.5.2. 7-Ketocholesteryl acetate (16)
White solid. 1H NMR (400 MHz, CDCl3): d 5.66 (s, 1H), 4.74–4.64
(m, 1H), 2.57–0.94 (comp, 32H), 0.90 (d, 3H, J = 6.4 Hz), 0.86–0.82
(m, 6H), 0.66 (s, 3H). 13C NMR (100 MHz, CDCl3): d 201.9, 170.2,
163.9, 126.7, 72.2, 54.8, 49.9, 49.8, 45.4, 43.1, 39.5, 38.7, 38.3,
37.7, 36.2, 36.0, 35.7, 28.5, 28.0, 27.3, 26.3, 23.8, 22.8, 22.6, 21.3,
21.2, 18.6, 17.2, 12.0. MS (Q-TOF, ESI+) m/z: [M+1] calculated for
C29H47O3 443.3525; found 885.6874, 443.3505, 383.3316, MS
(GCT) m/z: 382.1558, 173.9955. (0.96 g, 83%) Spectroscopic data
matched expected values [25].
2.5.3. 7-Ketocholesterol (17)
White solid. 1H NMR (400 MHz, CDCl3): d 5.70 (d, 1H, J = 1.9 Hz),
3.74–3.62 (m, 1H), 2.56–0.96 (comp, 30H), 0.93 (d, 3H, J = 6.6 Hz),
0.87 (dd, 6H, J = 6.6 Hz, J = 2.0 Hz), 0.69 (s, 3H). 13C NMR (100 MHz,
CDCl3): d 202.8, 166.0, 126.0, 70.4, 54.9, 50.1, 50.0, 45.5, 43.2, 41.9,
39.6, 38.8, 38.4, 36.5, 36.3, 35.8, 31.1, 28.7, 28.1, 26.4, 24.0, 23.0,
22.7, 21.3, 19.0, 17.4, 12.1. MS (Q-TOF, ESI+) m/z: [M+1] calculated
for C27H43O3 401.3420; found 401.3335. MS (GCT) m/z: 400.3997,
346.3411, 161.1247, 81.0792. Rf = 0.60 (ethyl acetate/-
toluene = 50/50). (0.47 g, 45%) Spectroscopic data matched
expected values [25].
2.4.11. Cholesterol (11)
White solid. 13C NMR (62.5 MHz, CDCl3): d 140.8, 121.7, 71.7,
56.8, 56.2, 50.2, 42.3, 42.2, 39.8, 39.5, 37.3, 36.5, 36.2, 35.8, 31.9,
31.6, 28.3, 28.0, 24.3, 23.9, 22.8, 22.6, 21.1, 19.4, 18.7, 11.9. TLC
showed one band and spectroscopic data matched expected values
[20]. Rf = 0.85 (ethyl Acetate/Toluene = 50/50).
2.4.12. Pregnenolone (12)
White solid. 13C NMR (100 MHz, CDCl3): d 209.8, 140.9, 121.2,
71.5, 63.7, 56.9, 49.9, 44.0, 42.2, 38.8, 37.3, 36.5, 31.8, 31.7, 31.4,
31.5, 24.5, 22.8, 21.1, 19.4, 13.2. TLC showed one band and spectro-
scopic data matched expected values [20].
2.5.4. 5-Pregnen-3-ol-7,20-dione (18)
White solid. 1H NMR (400 MHz, Deuterated DMSO): d 5.70 (s,
1H), 3.72–3.62 (m, 1H), 2.59–0.50 (comp, 28H). 13C NMR
(100 MHz, Deuterated DMSO): d 210.3, 201.8, 166.0, 126.2, 70.7,
62.5, 50.2, 45.6, 42.1, 38.5, 36.6, 32.0, 26.8, 23.8, 21.3, 17.5, 13.7.
MS (Q-TOF, ESI+) m/z: [M+1] calculated for C29H31O3 331.2273;
found 331.2200. MS (GCT) m/z: 329.9635, 245.1098, 161.1338,
91.0515, 79.0365. (0.20 g, 23%) Spectroscopic data matched
expected values [25].
2.4.13. Dehydroepiandrosterone (13)
White solid. 13C NMR (100 MHz, CDCl3): d 141.2, 120.7, 71.3,
51.7, 50.2, 47.5, 42.1, 37.2, 35.8, 31.5, 21.9, 20.3, 19.4, 13.5. TLC
showed one band and spectroscopic data matched expected values
[20].
2.4.14. 5-Pregnen-3,20-diol (14)
2.5.5. 3-Hydroxyandrost-5-ene-7,17-dione (19)
5-Pregnen-3,20-diol (14) was obtained through reduction of
pregnenolone using Luche reduction [21]. Cerium(III) Chloride
heptahydrate, 7.5 g, was dissolved in 100 mL of methanol.
Pregnenolone, 5 g, was then added and dissolved. To the solution,
1 g of sodium borohydride was added. The solution was stirred
overnight with a magnetic stir bar. DI water was then added to
quench the reaction. The reaction mixture was poured into a sep-
aratory funnel and ethyl ether was added forming a second layer.
The ether layer was washed with DI water and collected. TLC
revealed only one band from the ether layer, which was subse-
quently evaporated leaving 5-pregnen-3,20-diol. White solid. 1H
NMR (400 MHz, CDCl3): d 5.28 (s, 1H), 4.64 (s, 1H), 4.12 (m, 1H),
3.50 (m, 1H), 3.27 (m, 1H), 2.52 (s, 1H), 2.14 (m, 3H), 2.00–0.50
(comp, 25H). 13C NMR (100 MHz, CDCl3): d 140.8, 121.6, 71.7,
70.6, 70.4, 58.4, 58.4, 56.6, 56.2, 50.1, 50.1, 42.2, 41.6, 39.9, 38.8,
37.3, 36.5, 31.9, 31.7, 31.6, 31.5, 25.8, 25.6, 24.2, 23.5, 20.9, 20.8,
19.4, 12.4, 12.4.
White solid. 1H NMR (400 MHz, CDCl3): d 5.75 (s, 1H), 3.75–3.64
(m, 1H), 2.87–2.77 (m, 1H), 2.60–0.80 (comp, 23H). 13C NMR
(100 MHz, CDCl3): d 201.1, 166.1, 126.0, 70.3, 50.1, 47.9, 45.7,
44.3, 41.9, 38.4, 36.3, 35.7, 31.1, 30.7, 24.2, 20.6, 17.5, 13.8. MS
(Q-TOF, ESI+) m/z: [M+1] calculated for C19H27O3 303.1960; found
303.1958. MS (GCT) m/z: 301.8061, 161.534, 91.0481, 79.0190
(0.14 g, 19%).
2.5.6. 5-Pregnen-3, 20-diol-7-one (20)
White solid. 1H NMR (400 MHz, Deuterated DMSO): d 5.58 (s,
1H), 4.90 (d, 1H, J = 4.4 Hz), 4.15–4.10 (m, 1H), 3.54–3.25 (m,
2H), 2.49 (m, 5H), 2.42–0.95 (comp, 18H), 0.67 (s, 3H), 0.60 (s,
1H) 13C NMR(100 MHz, Deuterated DMSO): d 201.7, 167.4, 125.3,
69.5, 68.9, 56.8, 50.1, 45.2, 43.1, 38.5, 36.4, 31.5, 26.9, 24.4, 21.1,
17.4, 12.5. MS (Q-TOF, ESI+) m/z: [M+1] calculated for C21H33O3
333.2430; found 333.2350. MS (GCT) m/z: 332.1760, 161.0795,
91.0423 (0.11 g, 19%).

106
W.S. Grainger, E.J. Parish / Steroids 101 (2015) 103–109
2.6. Use of other vanadium complexes to produce 7-ketocholesterol
from cholesterol
D5 steroid precursor. Cholesterol, 1 g (2.6 mmol), dissolved in
25 mL of benzene, was almost completely oxidized using 3.09 g
(24 mmol) of 70% TBHP (0.85 M TBHP overall) and 0.1 g
(0.38 mmol) VO(acac)2. When the overall concentration of
TBHP was below 0.85 M, conversion diminished while higher
concentrations seemed to have no gained effect. Lower concen-
trations of VO(acac)2 also led to lower conversion percentages.
Only slight oxidation of cholesterol occurred when VO(acac)2
was not present. Nor did any significant oxidation occur when
TBHP was substituted with hydrogen peroxide or tert-butyl
peroxide.
Cholesterol oxidation occurred in benzene, 20% methanol/80%
benzene, and chloroform. However, use of chloroform as a solvent
was abandoned because of its potential reaction with TBHP to form
phosgene [27]. The addition of methanol lowered the conversion of
the cholesterol, making pure benzene the preferred solvent.
Several solvents in which the reaction did not significantly occur
include tert-butyl alcohol, pyridine, acetonitrile, 75% acetoni-
trile/25% chloroform, and methanol. The addition of heat (50 °C)
caused the solution to become murky and too viscous. Addition
of acetic acid was also detrimental to overall conversion. Adding
1 mL of toluene also decreased the overall conversion of
cholesterol.
For each reaction, 1 g (2.6 mmol) of cholesterol, was dissolved
in 25 mL of benzene with 0.38 mmol of one of the following vana-
dium catalysts: bis(cyclopentadienyl)vanadium (IV) dichloride
(0.1 g), vanadium (IV) oxide sulfate hydrate (0.06 g), vanadium
(III) acetylacetonate (0.13 g), vanadium (IV) carbide (0.02 g),
tris(triphenylsiloxy)vanadium (V) oxide (0.34 g), and vanadium
(V) oxytriisopropoxide (0.09 g). To these, 3.09 g (24 mmol) of 70%
TBHP was added. (0.85 M TBHP overall) Results were analyzed
through TLC and Q-TOF of neat product. This data is presented in
the Supplementary material section with further details.
3. Results/discussion
3.1. Optimization
The D5 steroids are particularly useful substrates due to the
selective oxidation of the allylic C7 carbon over the more steri-
cally hindered C4 carbon [26]. There is also an energetic benefit
of C7 oxidation over C4 oxidation [25]. Therefore, optimization
of this system centered around cholesterol, an easily obtainable
Table 1
D5 Steroidal alcohols and their 7-keto products
R2
R3
R2
VO(acac)2
TBHP
R3
Benzene,
R1
5 days, r.t.,
Ambient
R1
O
O
O
Major Product
Atmosphere
.
R1
R2
R3
H
Substrate
(9)
Product
(15)
Yield %a
98
C8H17
C8H17
C8H17
H
(10)
(16)
83
H
H
H
H
(11)
(12)
(17)
(18)
45
24
H
H
R2 + R3
Ketone
H
(13)
(14)
(19)
(20)
19
26
a
Isolated.

W.S. Grainger, E.J. Parish / Steroids 101 (2015) 103–109
107
The optimal amounts of reagents at room temperature, were
Step 1: Substitution of acac Ligand, Ref. [7].
proposed to be 2.6 mmol substrate, 25 mL benzene, 0.38 mmol
VO(acac)2, and 24 mmol (0.85 M) TBHP for 5 days. For 5-andro-
gen-3,17-diol (5), pregnenolone (12), DHEA (13), and 5-pregnen-
3,20-diol (14), only 1.3 mmol were used due to solubility issues.
Five days was determined by monitoring the progress of the reac-
tion through TLC. Of note, this reaction was upscaled 5 and 10
times with no difference in yields using multiples of these amounts
of reagents.
Step 2: Radical Initiation and Vanadium Oxidation, Ref. [7].
3.2. Oxidation
Fourteen steroidal substrates were tested under optimal con-
ditions. Five steroids showed no significant oxidation: estrone
(1), 4-cholestan-3-one (2), 1,4-androstadiene-3,17-dione (3),
cholestan-3-ol (4), and 5-androgen-3,17-diol (5). The only D5
to not show high conversion was 5-androgen-3,17-diol, which
was probably due to lack of solubility. Most important of this
set, cholestan-3-ol did not show significant oxidation, if any
at all, confirming that this reaction does not oxidize hydroxyl
groups, at least not of cholesterol. Steroidal D8(9) and D7 olefins,
7-dehydrocholesterol (6), 7-cholesten-3-ol (7) and 24,25-dihy-
drolanosterol (8), had high conversion percentages, but gave
complex mixtures of oxidation products. This was expected
given the multiple allylic sites of (7) and (8) and 7-dehydroc-
holesterol is well-known to form numerous oxidation products
[28].
Step 3: Propagation and termination.
Several D5 steroids showed susceptibility to oxidation at the
allylic C7 carbon. These steroids were cholesteryl benzoate (9),
cholesteryl acetate (10), cholesterol (11), pregnenolone (12),
DHEA (13), and 5-pregnen-3,20-diol (14), shown in Table 1. They
yielded as major products, 7-ketocholesterol benzoate (15), 7-ke-
tocholesterol acetate (16), 7-ketocholesterol (17), 5-pregnen-3-
ol-7,20-dione (18), 3-hydroxyandrost-5-ene-7,17-dione (19), and
5-pregen-3,20-diol-7-one (20) respectively.
Table 1 shows a correlation between the isolated yields and the
lipophilic natures of these steroids. Two large drops in yield can be
seen going from the esters to the hydroxyl group, (16–17), and
from side chain to no side chain, (17–18). The large drop in yields
between the esters and the other compounds is partly due to a
simpler work-up. Additionally, the differences in isolated yields
between (18) through (19) can be attributed to work-up
inefficiency. It is not certain why these yield variances appear
otherwise.
Upon adding TBHP to the reaction mixture, a reddish color was
observed, indicating the presence of a vanadium (IV) peroxide
complex [9,12]. Precipitate formed on the glass of the reaction flask
indicating substitution of the acetylacetonate ligand by the perox-
ide in step 1 [29]. In step 2, homolytic cleavage occurs at the
O
OOtBu
O
3.3. Mechanism
Three probable mechanisms for this oxidative reaction are
through free radical formation, singlet oxygen formation [29],
and formation of a reactive vanadium complex [15]. Because sin-
glet oxygen and radicals have short life times and the reaction
could last at least 5 days, a longer lasting vanadyl peroxide com-
plex seemed logical. However, vanadyl peroxide complexes are
thought to lead to epoxides rather than carbon–hydrogen oxida-
tions [7,12,13]. There was the possibility of radicals [7] and singlet
oxygen [29] continuously being generated. Although a singlet oxy-
gen mechanism seemed possible given auto-oxidation of choles-
terol, free radical oxidation was more prominent in the literature
surveyed [7,14]. Therefore, Scheme 1 was proposed as the primary
mechanism based on radical formation. (All three mechanisms
may contribute.)
H
O
O
O
O
H
H
tBuOH
+
O
O
O
Scheme 1. Proposed reaction mechanism.

108
W.S. Grainger, E.J. Parish / Steroids 101 (2015) 103–109
neat
Vacac2 14 (0.574) Cm (12:19-3:7x5.000)
1: TOF MS ES+
6.85e3
401.3398
100
Ambient Condiꢀons;
Vanadyl Acetylacetonate
402.3468
367.3384
357.3119
415.3235
431.3269
935.6646
m/z
141.0963 159.0364
150 200
489.3949
801.6806
800
275.1623
817.6741871.6753
549.3407
0
100
250 300
350
400
450
500
550 600
650
700
750
850
900
950
1000
neat
171_2 15 (0.610) Cm (11:38-3:8x5.000)
1: TOF MS ES+
8.99e3
122.0860
100
401.3394
N2 Atmosphere;
Vanadyl Acetylacetonate
144.0721
158.0880
402.3446
367.3374
417.3349
431.3164
357.3153
923.7072
m/z
245.1211
250
801.6650
800
489.3931
612.4804628.4843
600 650 700
889.7134
900
0
100
150
200
300
350
400
450
500 550
750
850
950
1000
Fig. 1. Mass spectra of neat product from the reaction conducted with optimal conditions, with cholesterol as the substrate, under ambient and nitrogen environments. These
spectra were taken with positive ion mode ESI Q-TOF MS. M/z 401 coincides with 7-ketocholesterol. TLC was used in conjunction to show that the starting material was
completely converted and not able to contribute to the m/z 401 peak.
oxygen–oxygen bond of the vanadium (IV) peroxide complex
ligand forming a tert-butoxide radical, which then abstracts a
hydrogen from the allylic carbon in step 3. A vanadium (V) species
is produced from the homolytic cleavage of the peroxide, which is
consistent with the observation of yellow colored solvent at the
end of the reaction [9]. In turn, the remaining acetylacetonate
ligand is also substituted by TBHP. A peroxyl radical is emitted
from the vanadium (V) complex. Respective of the small molar
amount of vanadyl compound used, it is necessary that the vana-
dium (V) species is regenerated from excess TBHP. We believe
the VVO2(OOtBu) to be regenerated from VIVO2 and TBHP and that
VIVO2 is the actual catalyst. In step 3, the enone product is formed

W.S. Grainger, E.J. Parish / Steroids 101 (2015) 103–109
109
through either subsequent hydrogen abstraction at the allylic site
in conjunction with termination via peroxyl radical and rearrange-
ment or deprotonation leading to rearrangement as shown in
Scheme 1.
established that the Q-TOF analysis had precision. Furthermore,
the data was analyzed two separate ways with similar results also
indicating that to a relative degree, there was accuracy. (The Q-TOF
data is presented in the Supplementary material.)
Although not shown, molecular oxygen is likely used as a source
of oxygen. This is evident because much lower ion counts of pro-
duct appear in mass chromatograms of neat product when the
reaction is conducted under nitrogen atmosphere as opposed to
ambient conditions. Additionally, there is an increase in side pro-
duct formation shown in the mass spectrum of the neat product
from the reaction conducted under a nitrogen atmosphere com-
pared to the spectrum from the ambient atmosphere. (Fig. 1)
This suggests that the presence of oxygen is also beneficial for
the reaction in that it terminates radical side polymerization and
product formation.
4. Conclusion
Vanadyl acetylacetonate can be used as a pre-catalyst to oxidize
steroidal alcohols under ambient conditions. Given the tolerance of
this reaction, it should be found useful in the oxidation of other
complex natural products.
Acknowledgments
We would like to thank Auburn University for funding this
research.
3.4. Other vanadium catalysts
Appendix A. Supplementary data
We also wanted to find out whether other vanadium com-
pounds would oxidize cholesterol to produce 7-ketocholesterol.
Given that the ligands were merely being substituted as per
Scheme 1, different vanadium complexes were predicted to also
work well regardless of their ligands. Bis(cyclopentadienyl)-
vanadium (IV) dichloride, vanadium (IV) oxide sulfate hydrate,
vanadium (III) acetylacetonate, vanadium (IV) carbide, vanadium
tris(triphenylsiloxy)vanadium (V) oxide, and vanadium (V)
oxytriisopropoxide all worked as substitutes for VO(acac)2 under
the same reaction conditions.
Indeed, all of the vanadium catalysts were usable for oxidation,
albeit to various degrees. Comparison of neat product on direct
injection Q-TOF chromatograms (using mass selection software),
in conjunction with TLC, showed that the vanadium (V) species
gave the highest yield of 7-keto product while vanadium (III)
acetylacetonate gave the lowest. Among the vanadium (IV) species,
VO(acac)2 gave the highest yield. It could be said that all metal-
TBHP reactions should yield the same results given that the metal
catalysts does not, to a large degree, interact with the substrate
directly. However, that is not considering that different catalyst
emit different radicals and ligands which can form side products
rather than the desired product.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.steroids.2015.06.
005.
References and notes
[1] Kihel LE. Steroids 2012;77:10–26.
[2] Shi J, Schulze S, Lardy HA. Steroids 2000;65:124–9.
[3] Zenk JL, Fredstedt JL, Kuskowski MA. J. Nutr. Biochem. 2007:629–34.
[4] Marshall CW, Ray RE, Laos I, Riegel B. J. Am. Chem. Soc. 1957;79(23):6308–13.
[5] De Medina P, Paillasse MR, Segala G, Khallouki F, Brillouet S, Dalenc F, et al.
Chem. Phys. Lipids 2011;164:432–7.
[6] Carvalho JFS, Silva M, Manuel C, Moreira JN, Simoes S, Melo LS. J. Med. Chem.
2010;53:7632–8.
[7] Vandichel M, Leus K, Van Der Voort P, Waroquier M, Van Speybroeck V. J. Catal.
2012;294:1–18.
[8] Da Silva JAL, da Silva JJRF, Pombeiro AJL. Coord. Chem. Rev. 2011;255:2232–48.
[9] Linden GL, Farona MF. Inorg. Chem. 1977;16:3170–3.
[10] Bhunia S, Konar S. J. Porous Mater. 2011;18:399–407.
[11] Zhao K, Wang Y, Billington DC. Tetrahedron: Asymmetry 2001;12:1211–7.
[12] Su C, Reed JW, Gould ES. Inorg. Chem. 1973;12:337–42.
[13] Chong AO, Sharpless KB. J. Org. Chem. 1977;42(9):1587–90.
[14] Kimura M, Muto T. Chem. Pharm. Bull. 1981;29:35–42.
[15] Xia J, Cormier KW, Chen Chuo. Chem. Sci. 2012;3:2240–5.
[16] Parish EJ, Kizito SA, Qiu Z. Lipids 2004;39:801–4.
[17] Makita N, Hoshino Y, Yamamoto H. Angew. Chem. Int. Ed. 2003;42:941–3.
[18] Zhang W, Yamamoto H. J. Am. Chem. Soc. 2007;129:286–7.
[19] Knapp Jr FF, Schroepfer Jr GJ. Steroids 1975;26:339–57.
[20] Frenkel M, Marsh KN. Spectral data for steroids. In: TRC series. College Station,
Texas: Thermodynamics Research Center; 1994.
3.5. Replicability
The Q-TOF data also gave us an idea of the replicability of this
reaction. Oxysterol yields consist of two phases, oxidation and iso-
lation. Isolation is by far the more difficult and most precarious of
the two. Thus, Q-TOF analysis of neat product allowed us to gain a
more variable-free analysis. To test the value of this method, its
precision and accuracy, we ran three trials with VO(acac)2 several
weeks apart with Q-TOF maintenance performed in between and
through comparison of Q-TOF ion counts, found the 7-ketocholes-
terol yield of the reaction to be very consistent, showing both our
reaction and Q-TOF method to be reproducible. Being reproducible
[21] Luche J, Rodriquez-Hahn L, Crabbe P. J.C.S. Chem. Commun. 1978:601–2.
[22] Dugas D, Brunel JM. J. Mol. Catal. A Chem. 2006;253:119–22.
[23] Wilson WK, Sumpter J, Rhea M, Warren JJ, Rogers PS, Ruan B, et al. J. Lipid Res.
1996;37:1529–55.
[24] Parish EJ, Wei T, Livant P. Lipids 1987;22(10):760–3.
[25] Li Y, Wu X, Lee TB, Isbell EK, Parish EJ, Gordan AEV. J. Org. Chem.
2010;75:1807–10.
[26] Dauben WG, Lorber M, Fullerton DS. J. Org. Chem. 1969;34:3587–92.
[27] Tatarova LA, Trofimova KS, Gorban AV, Khaliullin AK. Russ. J. Org. Chem.
2004;40:1403–6.
[28] Xu Libin, Korade Zeljka, Porter Ned A. J. Am. Chem. Soc. 2010;132:2222–32.
[29] Stepovik LP, Gulenova MV. Russ. J. Gen. Chem. 2009;79:1663–70.