Intramolecular sensitization within steroids: Excited-state interaction of C-17 α and β carbon-bromine bonds with a C-6 carbonyl group

Article abstract of DOI:10.1139/v03-055

The synthesis, photochemistry, and photophysics of 17α-bromo-3α -(triphenylsilyloxy)-5α-androstan-6-one (1) and 17β-bromo-3β -(triphenylsilyloxy)-5α-androstan-6-one (2) have been studied in aqueous tetrahydrofuran. The 17α-bromo isomer gives evidence (red

Full text of DOI:10.1139/v03-055

6
60
Intramole cular s e ns itization within s te roids :
Excite d-s tate inte raction of C-17 ␣ and ␤ carbon–
bromine bonds with a C-6 carbonyl group
We n-Sha n Li, Lokm a n Torun, a nd Ha rry Morris on
Abstract: The synthesis, photochemistry, and photophysics of 17α-bromo-3α-(triphenylsilyloxy)-5α-androstan-6-one (1)
and 17β-bromo-3β-(triphenylsilyloxy)-5α-androstan-6-one (2) have been studied in aqueous tetrahydrofuran. The 17α-
bromo isomer gives evidence (reduced φ , τ , and φ for the ketone) for interaction between the ketone and C–Br moi-
f
1
isc
eties in the excited singlet state. Some photodehalogenation is also observed upon excitation of the ketone chromo-
phore. This interaction seems to be absent or minimal for the 17β-bromo isomer.
Key words: photodehalogenation, bromosteroid, ketosteroid, intramolecular singlet–singlet energy transfer (ISSET).
Résumé : On a réalisé la synthèse et on a étudié la photochimie et la photophysique de la 17α-bromo-3α-(triphénylsi-
lyloxy)-5α-androstan-6-one (1) et de la 17β-bromo-3β-(triphénylsilyloxy)-5α-androstan-6-one (2) dans le tétrahydrofu-
rane. Avec l’isomère 17α-bromo, on a observé des résultats (valeurs réduites de φ , τ et φ pour la cétone) suggèrent
f
1
isc
l’existence d’une interaction entre les portions cétone et C–Br dans l’état singulet excité. On a aussi observé la pré-
sence de photodéshalogénation lors de l’excitation du chromophore de la cétone. Cette interaction semble être absente
ou minimale dans l’isomère 17β-bromo.
Mots clés : photodéshalogénation, bromostéroïde, cétostéroïde, transfert d’énergie singulet–singulet intramoléculaire
(
TESSI). Li et al. 668
Introduc tion
In general, our program has explored the chemistry of un-
saturated (e.g., ketone and olefin) acceptor groups resulting
from the transfer of energy from an absorbing (“antenna”)
chromophore. However, in a preliminary communication, we
recently presented evidence that a σ-bonded (e.g., C-Br)
functionality attached to the steroid framework could be ac-
tivated by interaction with a ketone antenna (2). Specifically,
photonic excitation (directly or indirectly) of a 6-keto (i.e.,
ring B) functionality can effect the photochemical cleavage
of an α C-17 (ring D) C—Br bond in the androstane 1. We
now present the details of this chemistry and extend the ear-
lier study to include the β C-17 analogs (2). (Note that TPS
and TBDMS in 2 are triphenylsilyl and tert-butyldimethyl-
silyl, respectively).
For many years chemists have evinced interest in the use
of rigid, well-defined hydrocarbon skeletons to study the
photophysical and photochemical consequences of excited-
state intramolecular energy transfer between distal func-
tional groups. The most extensively studied frameworks
have involved norbornylogous (1) and steroidal (2) back-
bones. The use of steroids to study intramolecular singlet–
singlet and triplet–triplet energy transfer dates back to the
1
960s (3). Though most efforts in the field have emphasized
photophysics, we have had a particular interest in the capa-
bility of the steroid skeleton to facilitate the photochemical
activation of a reactive, non-light absorbing, functional
group (2). Both through-space and through-bond interactions
can effect such a result. Given the interchromophore dis-
tances involved, chemistry resulting from triplet–triplet
intramolecular energy transfer involving rings A and D of
the steroid skeleton most convincingly makes the case for
coupling through the hydrocarbon framework (4, 5).
The somewhat surprising outcome of our most recent ob-
servations is that the interaction between the C-6 ketone and
the C-17 C—Br bond appears to be specific to the 17α iso-
mer.
Received 6 January 2003. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 18 June 2003.
Dedicated to Professor Don Arnold for his contributions to
chemistry.
1
2
W.-S. Li, L. Torun, and H. Morrison. Department of
Chemistry, Purdue University, 560 Oval Drive, West
Lafayette, IN 47907-2038, U.S.A.
1Present address: Institute of Chemistry, Academia Sinica,
1
28 Academia Road, Taipei 11529, Taiwan.
2
Corresponding author (e-mail: hmorrison@purdue.edu).
Can. J. Chem. 81: 660–668 (2003)
doi: 10.1139/V03-055
© 2003 NRC Canada

Li et al.
661
Scheme 1.
Re s ults a nd dis c us s ion
using a route totally analogous to that used for 2a. The other
was 17β-bromo-3β-methoxy-5α-androstan-6-one (2c), wherein
the 3-ol was methylated prior to the reduction of ring D and
hydroboration–oxidation in ring B.
Preparation of steroid substrates
1
3
7␣-Bromo-3␣-(triphenylsilyloxy)-5␣-androstan-6-one (1;
␣-TPSO–6ketone–17␣-Br)
Photochemistry of compounds 1 and 2a–2c
This substrate was prepared as outlined in Scheme 1,
starting with 17β-hydroxy-5α-androstane-3,6-dione that had
itself been prepared from 17β-hydroxyandrost-4-en-3-one
The photolysis of 1 was studied using both 266 and
308 nm laser light. A mixture of THF and water was em-
ployed to facilitate solubility and ammonium hydroxide was
present to neutralize released HBr. A summary of our obser-
vations, reproduced from our communication (2), is pre-
sented in eq. [1].
(
6).
The assignment of structure to 1 was confirmed by X-ray
analysis which clearly established the α orientations of both
the 3-TPSO and 17-Br functionalities (2).
1
7␤-Bromo-3␤-(triphenylsilyloxy)-5␣-androstan-6-one
[
1]
(
2a; 3␤-TPSO–6ketone–17␤-Br)
The title compound was prepared in six steps from com-
mercially available 5-androsten-3β-ol-17-one (dehydroisoan-
drosterone) as outlined in Scheme 2. The selective reduction
of the vinylbromide in ring D in the presence of the
cycloalkene in ring B is noteworthy. The conversion to the
cycloalkyl bromide is virtually quantitative. An expanded
discussion of this reaction as a synthetic tool will be pre-
sented elsewhere. As with compound 1, X-ray analysis
(
Fig. 1) was employed to confirm the β bromide stereo-
chemistry at C-17 as well as the β configuration of the 3-
TPSO group. The latter was a matter of convenience in the
availability of starting material. The stereochemistry in ring
A is of no consequence to the photophysics discussed in this
paper. The axial and pseudo-axial configurations of the
methine protons at C-3 and C-17, respectively, were also ev-
ident in the chemical shifts of their resonances at 3.6–3.8 δ,
characteristically upfield of the resonances for their equato-
rial and pseudo-equatorial counterparts in 1 that appear at
4
.2–4.3 δ (7).
Two analogs of 2a were also prepared. One was 17β-
bromo-3β-(tert-butyldimethylsilyloxy)-5α-androstan-6-one (2b),
©
2003 NRC Canada

6
62
Can. J. Chem. Vol. 81, 2003
Scheme 2.
The assignment of 4 and 5 as the 6α and 6β alcohols, re-
spectively, rests on the (axial) 6H resonance (3.14–3.28 δ) in
bond formation by the radical pair. We cannot distinguish
among the possible diastereomers that could result from
such a mechanism.
4
being located well upfield of the (equatorial) 6H resonance
(
3.54–3.60 δ) in 5 (7). The debromination characteristic of 8
Among the products shown in eq. [1], the most interesting
is the debrominated steroid (8). To confirm that cleavage of
the C-17 C—Br bond is occurring by intramolecular sensiti-
zation, we photolysed the 3α-OH–17α-Br steroid, both alone
and in conjunction with 1, using both 266 and 300 nm irra-
diation. No debromination was observed in any of these
cases. Nor could debromination of this steroid be sensitized
by the 3α-OH–6ketone (9) or by acetone (as solvent) using
300 nm irradiation.
was readily evident in the mass spectrum and in the disap-
pearance of the downfield C-17 methine proton and C-17
carbon resonances characteristic of 1. Compounds 4, 5, and
8
were independently synthesized, the alcohols by sodium
borohydride reduction of 1, and 8 by reduction of 5α-
androstan-3,6-dione with K-Selectride followed by silyl-
ation. Compounds 6 and 7 were inseparable and handled as
a mixture. The mass spectrum confirms the net addition of
THF to 1 and the proton NMR is consistent with the assign-
ment. These products are expected consequences of hydro-
gen abstraction from THF by the 6-keto group followed by
Quantum efficiencies are included in eq. [1]. Note that the
data for 266 vs. 308 nm irradiation are essentially identical
within experimental error. The former deposits energy in the
©
2003 NRC Canada

Li et al.
663
Fig. 1. X-ray structure of 2a.
point, and address the question of the excited-state multiplic-
ity of the ketone as sensitizer, later in the discussion.
The photochemistry of the β 17-Br series was studied in a
manner identical to that outlined above for 1. We began with
the 3β-TPSO derivative (2a) but quickly found that the
photolysis mixture was less readily resolved by the usual an-
alytical techniques than had been the case with 1. Recog-
nizing that the aryl chromophore at C-3 was, in fact,
superfluous for sensitization at C-17, we turned to the two
analogs (2b and 2c), which retained the requisite 6-keto
functionality and proved more amenable to study by GLC.
The photolysis of 2b is outlined in eq. [3]. Note that only di-
rect photolysis of the ketone with 311 nm light was em-
ployed since the TBDMSO functionality is transparent to
aryl chromophore while the latter selectively excites the car-
bonyl group. The lack of any significant dependence on the
initial site of excitation is reasonable, since we have earlier
demonstrated that singlet–singlet energy transfer from an α-
arylsilyloxy group at C-3 to a ketone at C-6 is 90% efficient
2
66 nm light.
(
8). (The efficiency drops to 77% when the arylsilyloxy
The products proved to be analogous to those reported in
group is β at C-3 (8)). These efficiencies are readily calcu-
lated from eq. [2] (or an analogous equation utilizing singlet
lifetimes) (8). Inserting the TPSO fluorescence quantum
eq. [1], with the exception that no dehalogenated steroid
could be detected. The alcohols (12 and 13) were formed in
a ca 1:1 ratio and together constituted 73% of the product
mixture by area count. They were each identified by GC–
MS spectrometry as a 6-ol but the spectral data in hand did
not permit a specific assignment of stereochemistry to a spe-
cific peak. Three other peaks, totaling 27% of the product,
were also resolvable by GLC. Each gave molecular ions by
GC–MS corresponding to a THF adduct and are assigned as
such, but again the data in hand did not allow one to differ-
entiate among the possible diastereoisomers and they are
grouped in eq. [3] as 14.
–
4
efficiency for 1 of 9.5 × 10 , and using either that for 10
–
2
–3
(
1.0 × 10 ) or 11 (9.3 × 10 ) as the reference (“r”), like-
wise provides an efficiency of singlet energy transfer from
1
C-3 to C-6 (φ intra3 → 6) of 90%.
Compound 2c was likewise photolyzed with 311 nm light
see eq. [4]). Again, no dehalogenated steroid was detect-
(
able. In this case only a single peak, constituting 41% of the
product area count, could be attributed to a 6 hydroxy prod-
uct (15); the GC–MS spectrum did not allow us to distin-
guish between the two possible diastereomers. Four peaks at
longer retention times, which together constituted 53% of
the product, each gave mass spectra consistent with a THF
adduct. They are grouped in eq. [4] as the mixture of iso-
mers (16). Finally, a sixth peak (17, 6%) was detected with a
retention time intermediate between the alcohols and the
[
2]
φ1
= [φf(r) – φf(1)]/φf(r)
intra3 → 6
The highly efficient transfer of singlet energy to C-6 from
the TPSO group, and the ability of the ketone, when directly
excited, to initiate cleavage of the C-17 C—Br bond, argue
for activation of the C—Br bond as involving the 6-keto
group as the sensitizing chromophore. We return to this
©
2003 NRC Canada

6
64
Can. J. Chem. Vol. 81, 2003
[
3]
[
4]
THF adducts that has not been identified. Its mass spectrum
indicates that it still retains bromine.
the carbonyl singlet excited state and the C-Br functionality.
The reduction in singlet lifetime should also be manifested
in the relative quantum efficiencies for intersystem crossing
In summary, much of the photochemistry of the two C-17
stereoisomers derives from reactions at C-6 that one would
anticipate as resulting from initial hydrogen abstraction from
the THF by the ketone exited state. Only in 1 do we see any
indication of activation at C-17. Mechanistic and photo-
physical studies were conducted to further clarify the nature
and extent of this interaction.
at C-6 (φ ) in these two steroids. Knowing that the reduc-
isc
tion products at C-6 derive from the triplet state, and assum-
ing that the ketone φisc would be ca 1.0 in the model (9), the
relative conversions to C-6 reduction products (at identical
photon flux) should provide the φisc for C-6 in 3. In the
event, we observed a 37% diminution (by GLC) of alcohols
and THF adducts from photolysis of the halogen-containing
steroid (3) relative to that observed with the model (9), thus
leading us to estimate that φisc for 3 is 0.63. The ratio of 1.6
Mechanistic and photophysical studies
Our first concern was to establish the multiplicity of the
ketone excited state(s) responsible for the chemistry depicted
in eq. [1]. This was done using cis-piperylene (1.2 mM) as a
potential triplet quencher for the 300 nm photolysis of 6 mM
for the relative φ s for 9:3 compares well with the ratio of
isc
1.4 for their respective singlet lifetimes. The two sets of
data, taken together, lead to the conclusion that there is an
additional mode of singlet decay present in 3 that causes a
reduction in this compound’s singlet lifetime and φisc.
Perhaps most notable in this series of studies is that the
measured singlet lifetime for the C-6 ketone in the 17β C-Br
steroid (2c) is 3.2 ± 0.2 ns. It is debatable as to whether the
11% reduction relative to 9 is statistically significant, but it
seems clear that the α C-17 C-Br functionality in 3 has a
much great effect on the C-6 ketone singlet state than does a
β C—Br bond in the same location. This conclusion is sup-
ported by a comparison of the photoreactivity of 2c relative
to the non-brominated model (9) as was done for the φisc
study for 3 described above. We found that the relative
photoreactivities of compounds 2c vs. 9 was 1.06, essen-
tially identical within experimental error. It is also interest-
1
7
. The contrasting effects on the products were striking; 50–
0% of the alcohol and THF adduct (4–7) formation was
quenched whereas there was no effect on the formation of
the C-17 cleavage product (8). That the hydrogen abstraction
chemistry should derive from the ketone triplet state is to be
expected. What was not so obvious is that the C-17 cleavage
chemistry would involve the ketone excited singlet state.
Nevertheless, this conclusion is supported by singlet life-
time data obtained for a pair of 3α-OH 6-keto steroids in
which the C-6 ketone is the only emitting chromophore, i.e.,
compounds 3 (with Br at C-17) and 9 (no Br at C-17).
Values of 2.5 ± 0.2 and 3.6 ± 0.1 ns were measured for these
two compounds, respectively. We attribute the ca 31% re-
duction in ketone singlet lifetime in 3 to interaction between
©
2003 NRC Canada

Li et al.
665
ing to note that the C-6 ketone triplet seems to be ineffective
in sensitizing cleavage of the C-17 C—Br bond, despite the
fact that its estimated (using cyclohexanone) energy of ca
Expe rim e nta l
General
–
1
7
8 kcal mol (9) should be well in excess of the estimated
A considerable amount of the experimental details is
available in the doctoral dissertation of L. Torun (16) or as
supplemental material for the preliminary communication
(
using isopropyl bromide) bond energy for the C—Br bond
–
1
of ca 68 kcal mol (10). Additional studies in a solvent less
capable of intercepting the carbonyl triplet state are neces-
sary to confirm that triplet energy remains totally localized
at the ketone. The challenge, of course, will be to do this
while simultaneously efficiently trapping the C-17 radical
generated by C-Br homolysis.
(
2). Thus, only the most salient features are presented below
2
or as supplemental material.
13
C NMR spectra were obtained in CDCl with a Varian
3
spectrometer operating at 50 MHz with chemical shifts rel-
ative to the residual chloroform peak at 77 ppm unless oth-
1
erwise noted. H NMR spectra were obtained in CDCl3
Further consideration of the interaction between the C-6
carbonyl group and the C-17 C—Br bond in 1
with a Varian spectrometer operating at 200 MHz with
chemical shifts relative to the residual chloroform peak at
7.24 ppm. In the case of two synthetic intermediates, the
proton count in the steroid aliphatic region read ca 5–10%
high relative to the downfield protons; this is indicated in
the listing of the proton spectra. Mass spectrometry utilized
a Finnigan 4000 mass spectrometer operating at a source
temperature of 250°C. Electron impact (EI) mass spectra
and chemical ionization (CI) mass spectra were recorded at
an ionization energy of 70 eV, with the latter utilizing iso-
butane at a pressure of 0.30 torr (1 torr = 133.322 Pa). Ele-
mental analysis was done by HR-MS using either FAB or
EI ionization on samples analytically pure by GLC. Ultra-
violet absorption spectra were recorded using 1-cm quartz
cells on a dual beam Cary model 100 UV–vis spectro-
photometer interfaced to a computer using Cary software.
Steady-state fluorescence spectra were obtained with 1 cm
square fluorescence cells on a SLM Aminco SPF-500
spectrophotometer using a 250 W xenon arc lamp operating
in the A/B mode. Fluorescence quantum efficiencies were
obtained using toluene as the reference. The fluorescence
lifetime measurements were obtained using a PTI model
100 spectrophotometer at room temperature. All the fluo-
rescence samples were purged with argon for 20 min prior
to analysis. For the X-ray structural analysis, crystals of
17α-bromo-3α-(triphenylsilyloxy)-5α-androstan-6-one (1) were
prepared by recrystallization of column-purified material
from hexane–CH Cl . A colorless plate with dimensions of
Given that the singlet lifetime of the C-6 ketone has been
shortened from 3.6 to 2.5 ns by the presence of an α C-17
C—Br bond (see above), if one assumes that this shortening
is entirely due to new decay involving an interaction be-
tween the two moieties, one calculates a rate constant for
8
–1
this decay mode of 1.2 × 10 s . An analogous calculation
using the 37% reduction in φisc yields a rate of singlet decay
8
–1
because of interaction of 1.5 × 10 s . We have already esti-
mated that ca 37% of the ketone singlets are diverted from
intersystem crossing by this interaction. Only 0.0066 (the
quantum by efficiency for formation of 8 (eq. [1])) is ac-
counted for by product formation resulting from C-Br cleav-
age. Clearly, a very small fraction (ca 2%) of the diverted
singlets surface as product, either because cleavage of the in-
teracting C—Br bond is minimal or because there is efficient
cleavage but also efficient radical-pair recombination.
As to the specific nature of the ketone–C-Br interaction,
we suggest that this may occur through TBI coupling of the
ketone n,π* singlet state with the C-Br n,σ* state, by anal-
ogy with the proposed explanation for the sensitized photo-
activation of a β C—Cl bond in chloronorbornenes (11) and
chlorobenzobicyclics (12). The strong stereoelectronic de-
pendence of the interaction is not surprising, there being am-
ple precedent in systems having fewer σ bonds between the
two moieties (13). However, the apparent preference for in-
teraction with the α, pseudo-axial C—Br bond at C-17 was
unexpected, since the “all trans” rule is generally accepted
as leading to enhanced coupling of σ-bonded equatorial vs.
axial substituents with the steroid framework (14). Neverthe-
less, though we have likewise observed a greater rate of
intramolecular triplet–triplet energy transfer (intraTTET)
from an equatorial (DPSO) donor at C-3 to an olefin at C-
2
2
0.41 × 0.35 × 0.22 mm was placed on a glass fiber in a ran-
dom orientation. X-ray data were collected with MO Kα
radiation (λ = 0.71073 Å) on a Nonius Kappa CCD com-
puter-controlled diffractometer equipped with a graphite
crystal, incident beam monochromator. Cell constants and
an orientation matrix for data collection were obtained
from least-squares refinement, using the setting angle of
17 224 reflections in the range 4 < θ < 27°. Data were col-
lected at a temperature of 173 ± 1°C to a maximum of 2θ of
55.8°. Melting points were determined with a Fisher-Johns
melting point apparatus and are uncorrected. Laser irradia-
tions were conducted at 266 nm using a Continuum NY-61
Nd:YAG laser equipped with a frequency quadrupler
(10 Hz, ca 3.0 mJ/pulse). Samples (3 mL, 10.0 mM) were
degassed with argon for at least 25 min prior to irradiation
1
7, relative to its axial C-3 counterpart (4), the inverse was
observed when the donor was σ-bonded to C-17. In that
event, intraTTET from C-17 to C-3 was found to be almost
1
0-fold faster from the α-DPSO group (15). Thus, the evalu-
ation of additional steroidal ketobromides, both by experi-
ment and by theory, is needed to determine whether the
stereoelectronics associated with the C-6–C-17 interaction
studied here is prototypical or peculiar to the specific place-
ment of the C—Br bond at C-17 in ring D.
2
Supplementary data may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research Council
Canada, Ottawa, ON K1A 0S2, Canada (http://www.nrc.ca/cisti/irm/unpub_e.shtml for information on ordering electronically). These data
can be obtained, free of charge, via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge CB2 1EZ, U.K.; fax +44 1223 336033; or deposit@ccdc.cam.ac.uk).
©
2003 NRC Canada

6
66
Can. J. Chem. Vol. 81, 2003
and kept closed with septa. They were photolyzed in square
vycor photolysis cells with continuous stirring using a
36.54, 36.19, 32.82, 32.44, 31.50, 24.53, 20.63, 19.38,
13.96. EI (m/z): 352/354 (M) .
+
2
mm stirring bar. The power of the laser was measured
with an OPHIR power meter, model AN/2. Data were cor-
rected for scatter and reflectance of light from the cell
walls. For 311 nm irradiations, a reactor was equipped with
17␤-Bromo-3␤-(triphenylsilyloxy)-5-androsten (VIII)
17β-Bromo-5-androsten-3β-ol (VII) (52 mg, 0.14 mmol)
was added to a round-bottomed flask containing DMF
(2.0 mL). To this solution was added triphenylsilyl chloride
(0.65 mg, 0.22 mmol), imidazole (20 mg, 0.28 mmol), and a
catalytic amount of 4(dimethylamino)pyridine, and the solu-
tion was stirred under argon for 12 h. Water (10 mL) was
added and the solution was extracted with ether. The organic
extracts were washed with water. After drying and concen-
trating, the crude product was chromatographed on a
Chromatotron plate prepared with silica gel using hexane to
8
× 311 nm Phillips TL01 (20 W) lamps and a merry-go-
round turntable apparatus that positioned 8 cylindrical
quartz tubes (10 mm i.d.) approximately 2 cm from the
lamps. The 300 nm irradiation employed a Rayonet Reactor
(
New England Ultraviolet Co.) equipped with lamps having
maximal output at 300 nm. All solutions were deoxygen-
ated with a stream of argon prior to photolysis. Uranyl
oxalate actinometry was used for the quantum efficiency
determinations. 4-Androsten-3,17-dione, 3β-hydroxy-5α-
androstan-17-one, 17β-hydroxy-5α-androtan-3-one, 3β-
hydroxy-5-androsten-17-one, 5-androsten-3β-ol-17-one
1
obtain 0.27 g (80%) as an oily product. H NMR (CDCl3,
200 MHz) δ: 7.65–7.58 (m, 6H), 7.48–7.32 (m 9H), 5.25–
5.10 (bs, 1H), 3.85–3.65 (m, 2H), 2.60–0.80 (m, ca 25H).
1
3
(
dehydroisoandrosterone), and testosterone were purchased
C NMR (CDCl3, 50 MHz) δ: 141.24, 135.40, 134.94,
134.89, 130.10, 129.97, 127.90, 127.76, 73.31, 62.03, 51.34,
9.96, 43.82, 42.30, 37.16, 36.56, 36.19, 32.80, 32.45,
31.77, 31.53, 24.53, 20.59, 19.40, 13.97. EI (m/z): 610/612
from Sigma.
4
1
7-Bromo-3␤-hydroxy-androsta-5,16-diene (VI)
β-Hydroxy-5-androsten-17-hydrazone (V) (2.0 g,
.57 mmol), prepared from commercially available 3β-
+
+
3
(M ), 259 (Ph Si) base peak. HR-MS (EI) calcd. for
3
6
C H BrOSi: 610.2267; found: 610.2250.
3
7
43
hydroxy-5-androsten-17-one (IV), was dissolved in pyridine
30 mL) and the solution cooled in an ice bath. NBS (2.0 g,
6.85 mmol) dissolved in 15 mL of pyridine was added
dropwise to the solution over 5 min. After stirring the yel-
low-colored solution at ice bath temperature for 5 min, the
solution was poured into 50 mL of ice water and the mixture
extracted with ether. The combined ether layers were
(
1
17␤-Bromo-3␤-triphenylsilyoxy-5␣-androstan-6␣-ol (IX)
The general procedure used for hydroboration was to add
the silylated bromoandrostene (ca 0.5 to 0.6 mmol) to a
flask containing 10 mL of dry THF, cool to ice bath temper-
ature, and then add a borane solution (1 M) in THF (0.8 mL)
under argon. After stirring in an ice bath for 1 h, and then at
room temperature for 12 h, the excess borane was destroyed
with water. A 2 N solution of NaOH (1.5 mL) and 30%
H O (1.5 mL) was added and the resulting mixture stirred
washed with 10% HCl (3 × 20 mL), NaHCO solution (3 ×
3
2
0 mL), and water (2 × 20 mL) and dried over Na SO . The
2 4
crude product was chromatographed on silica gel with 2%
2
2
EtOAc in hexane to give a white solid product in 48–65%
at 50°C for 1 h. After cooling to room temperature, the
crude product was extracted with ether (3 × 15 mL). The
combined organic layers were washed with water and dried
over Na SO , and concentrated under reduced pressure.
1
yield; mp: 163 to 164°C. H NMR (CDCl , 200 MHz) δ:
3
5
0
1
3
.90–5.80 (dd, 1H), 5.4–5.3 (m, 1H), 3.6–3.4 (m, 1H), 2.4–
.85 (m, ca 24 H). 13C NMR (CDCl , 50 MHz) δ: 141.2,
3
2
4
35.6, 129.0, 121.1, 71.6, 55.5, 50.4, 48.5, 42.2, 37.1, 36.7,
The residue was chromatographed on a column using
CH Cl :hexanes (1:2) to obtain the products as white solids
4.5, 31.8, 31.5, 31.3, 30.7, 20.6, 19.3, 15.0. CI (m/z): (M +
2
2
+
+
H) : 351/353 (base peak). EI: 350/352 (M) ; 91 (base peak).
in >90% yield. For 17β-bromo-3β-triphenylsilyoxy-5α-
1
androstan-6α-ol: H NMR (CDCl , 200 MHz) δ: 7.72–7.6
3
1
7␤-Bromo-5-androsten-3␤-ol (VII)
Hydrazine monohydrate (3.0 mL) was added to a flask
(m, 6H), 7.50–7.30 (m, 9H), 3.85–3.65 (t, 2H), 3.45–3.3 (m,
1
3
1H), 2.40–0.50 (m, 27 H). C NMR (CDCl , 50 MHz) δ:
3
containing 17-bromo-3β-hydroxy-androsta-5,16-diene (VI)
0.7 g, 1.99 mmol) dissolved in 30 mL of methanol. To this
solution was added K Fe(CN) (1.2 g, 3.64 mmol) and a cat-
135.38, 134.85, 129.85, 127.74, 72.88, 69.14, 61.83, 53.60,
51.81, 50.79, 44.04, 40.90, 37.22, 36.18, 35.06, 32.39,
32.10, 31.34, 24.30, 20.58, 14.09, 13.38. CI (m/z): 353/355
(
3
6
+
alytic amount of Cu(OAc) (~10 mg). The resulting solution
(M – Ph SiO) . HR-MS (FAB) calcd. for C H BrO Si:
2
3
37 45
2
was stirred at room temperature with continuous monitoring
by GC. Hydrazine monohydrate (3.0 mL each time) was
added as needed to eliminate all starting material in the solu-
tion. A total of 9 to 10 mL was added over 3 days. The
milky white solution was filtered and MeOH was removed
in vacuo. The resulting aqueous residue was partitioned be-
tween dichloromethane and water and extracted with di-
chloromethane. The solvent was dried over Na SO ,
628.2372; found: 628.2352.
17␤-Bromo-3␤-(triphenylsilyloxy)-5␣-androstan-6-one (2a)
17β-Bromo-3β-triphenylsilyoxy-5α-androstan-6α-ol (IX)
(0.24 g, 0.38 mmol) was dissolved in 5 mL of dichloro-
methane and added to a stirred suspension of pyridinium
chlorochromate (0.123 g, 0.58 mmol) and sodium acetate
(0.031 g, 0.44 mmol). The solution was stirred at room tem-
perature for 3 h, filtered, concentrated, and chromatographed
on a silica gel column using hexane to afford 0.16 g (85%
2
4
concentrated, and purified by column chromatography on
silica gel using 15% EtOAc in hexane to afford a white solid
1
in 85–92% yield; mp: 145–147°C. H NMR (CDCl3,
yield) of an oily product that solidified over several days;
1
2
(
00 MHz) δ: 5.4–5.2 (d, 1H), 3.85–3.65 (t, 1H), 3.65–3.40
mp 186–190°C. H NMR (CDCl , 200 MHz) δ: 7.70–7.55
3
m, 1H), 2.40–0.75 (m, 26H). 13C NMR (CDCl , 50 MHz) δ:
(m, 6 H), 7.50–7.30 (m, 9H), 3.80–3.60 (m, 2H), 2.40–2.17
(m, 2H), 2.15–1.50 (m, 12H), 1.50–0.96 (m, 6H), 0.89 (s,
3
1
40.88, 121.10, 71.59, 61.94, 51.34, 50.01, 43.82, 37.21,
©
2003 NRC Canada

Li et al.
667
3
1
4
2
3
H), 0.84 (s, 3H). 13C NMR (CDCL , 50 MHz) δ: 210.14,
Photolysis of 17β-bromo-3β-methoxy-5α-androstan-6-one
(2c) was conducted in a similar manner. Five photoproducts
were detectable by GLC in addition to unreacted starting
material (6.3 min) that made up 69% of the total GLC area
count. The first photoproduct had a retention time at 6.2 min
and constituted 10% of the total area count. The EI mass
spectrum showed a molecular ion peak at m/z = 384/386 cor-
responding to reduction of the ketone. Fragment ions were
3
35.40, 129.92, 127.79, 72.21 61.17, 56.75, 53.77, 51.45,
8.25, 44.44, 40.73, 38.58, 36.88, 35.96, 32.28, 30.97,
+
8.91, 24.25, 20.99, 14.11, 13.15. CI (m/z): 627/629 (M ),
51/353 (base peak). HR-MS (FAB) calcd. for
C H BrO Si: 627.2294; found: 627.22267.
3
7
43
2
1
6
7␤-Bromo-3␤-(tert-butyldimethylsilyloxy)-5␣-androstan-
-one (2b)
The methods used for this preparation were analogous to
+
observable at 366/368 (M – H2O) and 353/355 (base
+
peak; M – OCH ) . The other four photoproducts had rela-
3
tively long GLC retention times (11.9, 12.2, 12.4, and
those described above for 2a. It was obtained as a white
crystalline material, mp 172 to 173°C. H NMR (CDCl ,
1
1
2.7 min), were formed in a ratio of 4.5:4.9:1.7:1.6, respec-
3
tively, and had a combined area count constituting 13% of
the total area. All four products had identical EI mass spec-
tra and very similar CI mass spectra. The former contained
the THF fragment ion (m/z = 71) as the base peak, and in-
2
0
00 MHz) δ: 3.83–3.67 (t, 1H), 3.59–3.40 (m, 1H), 2.40–
.95 (m, 20H), 0.85 (s, 9H), 0.80 (s, 3H), 0.74 (s, 3H), 0.01
1
3
(
s, 6H). C NMR (CDCl , 50 MHz) δ: 210.31, 71.26, 61.21,
3
5
3
1
6.95, 53.91, 51.48, 46.27, 44.46, 40.78, 38.58, 36.83,
6.01, 32.30, 31.24, 30.18, 25.84, 24.27, 21.04, 18.18,
+
cluded fragment ion peaks at m/z = 383/385 (M – THF) and
+
m/z = 351/353 (M – THF – CH OH) . The base peak in the
4.13, 13.18, –4.63, –4.71. HR-MS (CI) calcd. for
3
+
+
CI spectra appeared at m/z = 437/439 [(M + H) – H O) . A
C H BrO Si (M + H) : 483.2294; found: 483.2309.
2
2
5
43
2
sixth photoproduct (9.0 min) constituted 1.3% of the total
area count. Its CI mass spectrum exhibited a fragment ion at
m/z = 368/370 as the base peak.
1
7␤-Bromo-3␤-methoxy-5␣-androstan-6-one (2c)
The methods used for this preparation were analogous to
1
those described above for 2a. H NMR (CDCl , 200 MHz) δ:
3
3
1
3
5
3
1
3
3
.85–3.65 (t, 1H), 3.32 (s, 3H), 3.20–2.90 (m, 1H), 2.40–
.50 (m, 11 H), 1.50–0.95 (m, 7 H), 0.81 (s, 3H), 0.74 (s,
Ac knowle dge m e nts
1
3
We thank Professor Mary Boyd of Loyola University for
the use of her PTI instrument. We are grateful to Phil
Fanwick and Karl Wood for their assistance in obtaining the
X-ray and mass spectral data, respectively.
H). C NMR (CDCl , 50 MHz) δ: 210.05, 78.92, 61.14,
3
6.62, 55.60, 53.80, 51.39, 46.18, 44.41, 40.98, 38.50,
6.57, 35.92, 32.25, 27.44, 25.80, 24.22, 21.00, 14.10,
+
+
3.06. CI (m/z): 383/385 (M + H) . EI: 382/384 (M ),
53/355 (base peak). HR-MS calcd. for C H BrO :
2
0
29
2
82.1507; found: 382.1507.
Re fe re nc e s
1
. T.A. Smith, N. Lokan, N. Cabral, S.R. Davies, M.N. Paddon-
Row, and K.P. Ghiggino. J. Photochem. Photobiol. A, 149, 55
Photolyses
The details for the preparative photolysis of compound 1
have been presented (2). For the study of 17β-bromo-3β-(tert-
butyldimethylsilyloxy)-5α-androstan-6-one (2b), 5.0 mM solu-
tions in THF:H O (9:1) containing ca 5 mM of NH OH
were degassed with argon for 20 min and irradiated in quartz
tubes with 311 nm light for 5 h. The reactions were analyzed
by GLC using a 30 m DB-5 column at 280°C under isocratic
conditions. The starting material, with a retention time of
(
2002), and refs. therein.
2
. W.-S. Li and H. Morrison. Org. Lett. 2, 15 (2000), and
refs. therein.
2
4
3. (a) D.E. Breen and R.A. Keller. J. Am. Chem. Soc. 90, 1935
(1968); (b) R.A. Keller. J. Am. Chem. Soc. 90, 1940 (1968);
(c) R.A. Keller and L.J. Dolby. J. Am. Chem. Soc. 91, 1293
(1969).
4. L.D. Timberlake and H. Morrison. J. Am. Chem. Soc. 121,
3618 (1999).
. (a) C.-H. Tung, L.-P. Zhang, Y. Li, H. Cao, and Y. Tanimoto.
J. Phys. Chem. 100, 4480 (1996); (b) C.-H. Tung, L.-P. Zhang,
Y. Li, H. Cao, and Y. Tanimoto. J. Am. Chem. Soc. 119, 5348
1
1.9 min, was the major component of the reaction mixtures
5
and constituted 65% of the total area counts. Five photo-
products were also observed in the chromatograms. Two ma-
jor photoproducts appeared with retention times of 11.1 and
(
1997).
. J.H. Fried, A.N. Nutile, and G.E. Arth. J. Am. Chem. Soc. 82,
704 (1960).
1
1.2 min in a 12.1:13.6 relative ratio. There were three other
6
7
peaks with retention times of 20.5, 21.5, and 21.9, with rela-
tive ratios of 6.0:1.51:1.5. The five products were character-
ized by GC–MS analyses. The two major photoproducts had
5
. J.E. Bridgeman, P.C. Cherry, A.S. Clegg, J.M. Evans, E.R.H.
Jones, A. Kasal, V. Kumar, G.D. Meakins, Y. Morisawa, E.E.
Richards, and P.D. Woodgate. J. Chem. Soc. C, 250 (1970).
. J.K. Agyin, L.D. Timberlake, and H. Morrison. J. Am. Chem.
Soc. 119, 7945 (1997).
+
very similar EI and CI mass spectra, with a (M + H) molec-
ular ion evident in the CI mass spectrum at m/z 485/487.
Loss of water was evident at 467/469. There were three ad-
ditional photoproducts with retention times of 20.5, 21.5,
and 21.9 min. One (20.5 min) exhibited a small molecular
ion peak in the CI spectrum at m/z = 555/557 corresponding
to the addition of THF to 2b. The spectrum also contained a
fragment ion peak at m/z = 537/539 as the base peak, corre-
sponding to the loss of water. The other two products had
very similar EI and CI mass spectra but lacked the small mo-
lecular ion peak.
8
9
. N.J. Turro. Modern molecular photochemistry. Benjamin/
Cummings, Menlo Park, Calif. 1978. p. 290.
1
1
0. S.W. Benson. Thermochemical kinetics. Wiley, New York.
968.
1
1. D.M. Pearl, P.D. Burrow, J.J. Nash, H. Morrison, and K.D.
Jordan. J. Am. Chem. Soc. 115, 9876 (1993); J.J. Nash, D.V.
Carlson, A.M. Kaspar, D.E. Love, K.D. Jordan, and H. Morri-
son. J. Am. Chem. Soc. 115, 8969 (1993); B.D. Maxwell, J.J.
©
2003 NRC Canada

6
68
Can. J. Chem. Vol. 81, 2003
Nash, H.A. Morrison, M.L. Falcetta, and K.D. Jordan. J. Am.
Chem. Soc. 111, 7914 (1989).
14. N. Liang, J.R. Miller, and G.L. Closs. J. Am. Chem. Soc. 112,
5353 (1990), and preceding papers in this series.
15. L. Timberlake. Doctoral Dissertation, Purdue University, W.
Lafayette, Indiana, December 1998.
16. L. Torun. Doctoral Dissertation, Purdue University, W.
Lafayette, Indiana, December 2001.
1
1
2. A.J. Post, J.J. Nash, D.E. Love, K.D. Jordan, and H. Morrison.
J. Am. Chem. Soc. 117, 4930 (1995), and other papers in this
series.
3. H. Morrison, T.-V. Singh, L. De Cardenas, D. Severance, K.
Jordan, and W. Shaefer. J. Am. Chem. Soc. 108, 3862 (1986).
©
2003 NRC Canada