Photooxygenation of pregnanes

Helvetica Chimica Acta ± Vol. 87 (2004)

2987

Photooxygenation of Pregnanes

by María A. Ponce and Rosa Erra-Balsells*

CIHIDECAR-CONICET, Departamento de Química Org aÂ nica, Facultad de Ciencias Exactas y Naturales,

Universidad de Buenos Aires, Pabell oÂ n II, 38, Ciudad Universitaria, 1428-Buenos Aires, Argentina

(tel/fax: 54-11-45763346; e-mail: erra@qo.fcen.uba.ar)

and Andrea C. Bruttomesso and Eduardo G. Gros¹)

UMYMFOR-CONICET, Departamento de Química Org aÂ nica, Facultad de Ciencias Exactas y Naturales,

Universidad de Buenos Aires, Pabell oÂ n II, 38, Ciudad Universitaria, 1428-Buenos Aires, Argentina

The course of the singlet-oxygen reaction with pregn-17(20)-enes and pregn-5,17(20)-dienes was studied to

compare the reactivity of the two alkene moieties present in some steroid families. Thus, from commercially

available (3b,5a)-hydroxy-androstan-17-one and (3b)-3-hydroxyandrost-5-en-17-one, the following 3-{[(tert-

butyl)dimethylsilyl]oxy}-substituted, 17(20)-unsaturated pregnanes were prepared (see Fig. 1): (3b,5a)-21-

norpregn-17(20)-ene 1; (3b,5a,17Z)-pregn-17(20)-ene 2, (3b,5a,16a,17E)-pregn-17(20)-en-16-ol 3,

(16b,5a,17E)-pregn-17(20)-en-16-ol 4, (3b,5a,16b,17E)-pregn-17(20)-en-16-olacetate 5, (3b,16a)-21-norpreg-

na-5,17(20)-dien-16-ol 6, (3b,16a,17E)-pregna-5,17(20)-dien-16-ol 7, (3b,17Z)-pregna-5,17(20)-diene 8,

(

3b,17E)-pregna-5,17(20)-dien-21-ol 9 and (3b,17E)-5,17(20)-dien-21-olacetate 10. The oxygenated products

1

3

see Fig. 2) obtained from 1 ± 10 and O

2

, generated by irradiation of Rose Bengalin

2

O -saturated pyridine

1

13

solution, were characterized by H-, C-NMR, and MS (EI, FAB, HR-EI, ESI- and UV-MALDI-TOF) data.

Major products were those formed by the ene reaction involving as intermediates the corresponding

hydroperoxides and the cyclic tautomers of the allylic hydroperoxides, i.e., the corresponding oxiranium oxide-

like intermediate (Scheme 5).

Introduction. ± The phototransformations of steroids owing to the direct absorption

of light or to the interaction with electronically excited species have been calling

continuously the attention of researchers. It is known that sterols, as important lipidic

components of human skin cells, may contribute to phototoxicity by means of

photochemicalreactions in the UVB region through oxygen-dependent reactions in the

UVA region [1]. Sterols with a 5,7,9-triene substructure are the only naturally occurring

1

O sensitizers described in UVA region, but they are also easily oxidized by this

2

species. On the other hand, cholesterol and other sterols with their isolated non-

1

conjugated CC bonds do not absorb UVA or UVB light but quench O . To prevent

2

1

photodamage owing to photooxidation of steroids, the O quenching process needs to

2

be understood.

In an early work (1958), Schenck and co-workers showed that cholesterol reacts

1

with O yielding (3b,5a)-5-hydroperoxycholest-6-en-3-ol, which rearranges, in non-

2

polar solvent, to (3b,7a)-7-hydroperoxycholest-5-en-3-ol [2]. In 1961, Nickon and co-

workers found that hydroperoxidation of steroids with a CC bond is stereospecific,

and the new CÀO bond is in the cis position to the CÀH bond that is broken [3a,b].

¹)

In memory of Dr. Eduardo G. Gros

ꢀ

2004 Verlag Helvetica Chimica Acta AG, Zürich

2

988

Helvetica Chimica Acta ± Vol. 87 (2004)

Since these early works, a dozen or more examples of these phenomena in which an

allyl hydroperoxide rearranges to its allylic isomer, have been identified, half of which

refer to cholestene derivatives. Several years later, Beckwith et al. [4] showed that the

tertiary 5a-yl hydroperoxide originally formed from cholesterol rearranges by a

nondissociative mechanism (sigmatropic 1,3-rearrangement; Schenck suprafacial

rearrangement) yielding the 7a-ylhydroperoxide.

In a previous project, we examined the reaction of stigmasterol, a 5,22-

diunsaturated sterol, and of some 3b-substituted 5,22-diunsaturated stigmastanes with

1

O , i.e., of 3b-F-, 3b-Cl-, 3b-Br-, 3b-I-, 3b-AcO-, 3b-MsO-, and 3a-Br-substituted

2

derivatives; depending on the substituent present at C(3), the initially formed 5a-yl

hydroperoxides rearranged to the 7a- or the 3a-isomeric structures [5]. In an attempt

to understand further some of the structuralfactors that affect the reactivity of steroids

1

with O , we then compared the reactivity of ergosteroland O by using different

2

combinations of photosensitizers and solvents [6].

In the present report, we describe the photooxygenation of CC bonds present in

some steroids of the pregnane family. Thus, a series of new unsaturated (3b)-3-{[(tert-

butyl)dimethylsilyl]oxy}pregnanes were prepared and fully characterized to study their

1

reaction with O . In these new unsaturated pregnanes 1 ± 10, a C(17)C(20) bond with

2

H-atoms in allylic positions and in some cases, also a C(5)C(6) bond is present

(

Fig. 1). The latter CC bond was shown to be quite reactive in the stigmastane family

[

5]. Commercially available (3b,5a)-3-hydroxyandrostan-17-one and (3b)-3-hydroxy-

androst-5-en-17-one were used as starting materials. All new compounds 1 ± 10 were

1

13

fully characterized by H- and C-NMR and MS (EI, FAB, HR-EI, ESI- and UV-

MALDI-TOF) data.

Fig. 1. Pregnanes studied

Results. ± A solution of pregnene 1 and Rose Bengalin pyridine was stirred under

O and irradiated with a tungsten lamp for 24 h. After isolation and purification of the

2

1

13

products, the H- and C-NMR spectra showed that the (5a)-21-norpregn-16-en-20-al

Helvetica Chimica Acta ± Vol. 87 (2004)

2989

1a was obtained in 22% yield besides the starting material (Fig. 2). All spectroscopic

data of 1a supported the structure of an a,b-unsaturated aldehyde (see Exper. Part).

When 2, carrying an ethylidene group at C(17), was oxygenated according to the

generalconditions, (20 S)-pregn-16-en-20-ylhydroperoxide 2a was isolated in 64%

yield, besides trace amounts (<0.5%) of pregn-16-en-20-one 2b and starting material

1

13

(

Fig. 2). Comparison of the H- and C-NMR spectra of 2 and 2a clearly established

the presence of a shifted CC bond as well as a MeCH(OOH) moiety at C(20) of 2a

see Exper. Part). The MS data supported the formula C H O Si and the hydro-

(

27

48

3

peroxide structure of 2a. Compound 2b of formula C H O Si (MS data) showed the

2

7

46

2

1

typical H-NMR signals of the C(16)C(17) bond and of the acetylgroup in viny il c

position at C(17) (see Exper. Part).

Photooxygenation of 3 and 4 under the usual conditions gave the following results:

Pregnenol 3 was recovered quantitatively after 24 h. In contrast, pregnenol 4 gave 17a-

1

and 17b-isomers in 26% yield besides starting material (Fig. 2). Comparison of the H-

13

and C-NMR spectra of 4 and 4a established the presence of a C(16)O and an

OHÀC(20) group in 4a (see Exper. Part).

When the acetylderivative 5 of pregnenol 4 was submitted to the general

oxygenation conditions, a product with similar R value than that of compound 4a and a

f

new product, i.e., 5a and 5b, were isolated in 38 and 40% yield, respectively, besides

starting material. The spectral data of 5a established its identity with 4a (17a/17b

mixture), and those of 5b were compatible with the proposed structure (see Fig. 2 and

Exper. Part).

To compare the reactivity of the H-atoms in allylic position of the C(17)C(20)

and C(5)C(6) bonds located in the pregnane skeleton, five new pregna-5,17(20)-

dienes were subjected to the usualoxygenation conditions. Pregnadienol 6 afforded

one oxidized product, 6a (Fig. 2), which was obtained in 36% yield, besides

starting material. The structure of 6a was established by spectroscopic means and

was also in agreement with the Schenck mechanism described in the literature for the

1

reaction of O with the C(5)C(6) bond of cholesterol [1 ± 4] and stigmastanes [5].

2

Pregnadienol 7 gave on oxygenation the expected (7a,16a,17E)-7-hydroperox-

ypregna-5,17(20)-dien-16-ol 7a in 26% yield besides starting material (Fig. 2). The

NMR and MS data of 7a were in agreement with the proposed structure (see Exper. Part).

It is interesting to note that in 6a and 7a, the C(17)C(20) bond was unchanged as

1

compared to the starting materials. Furthermore, it is well-known that the attack of O

2

at the C(5)C(6) bond in cholesterol-like structures yields only the 5-en-7-yl

hydroperoxide with (7a) configuration by a concerted rearrangement of the primarily

formed 6-en-5-ylhydroperoxide with (5 a) configuration [1 ± 5].

However, when pregnadiene 8, which is devoid of OHÀC(16), was oxygenated, the

obtained products 8a and 8b (36 and 9% yield, resp.) showed an unchanged C(5)C(6)

1

bond region (Fig. 2), i.e., the reaction with O had taken place at C(17)C(20).

2

1

13

Comparison of the H- and C-NMR spectra of compounds 8, 8a, and 8b clearly

1

13

established the positions of the two CC bonds in the products. The H- and C-NMR

1

spectra of 8a and 8b were quite similar, except for major differences in the H-NMR

signals of HÀC(20) and Me(21) (see Exper. Part). Detailed analysis of the MS data of

8a and 8b supported the hydroperoxy structure of 8a and the hydroxy structure of 8b

(see Exper. Part).

2

990

Helvetica Chimica Acta ± Vol. 87 (2004)

Fig. 2. Pregnanes studied and products obtained by photooxygenation

Helvetica Chimica Acta ± Vol. 87 (2004)

2991

Fig. 2 (cont.)

2

992

Helvetica Chimica Acta ± Vol. 87 (2004)

Pregnadienol 9 yielded, after photooxidation under the usual conditions, isolation,

and purification, the four new photoproducts, 9a (6%), 9b (27%), 9c (10%), and 9d

1

13

(

17%), besides starting material. Comparison of the H- and C-NMR spectra of

compounds 9, 9a, and 9b showed that the C(17)C(20) bond of 9 was shifted to

C(16)C(17) in 9a and 9b, that the CH (21)OH group was retained, and that the

2

C(5)C(6) was unaffected by the oxygenation. In addition, 9a carried an OOH group

at C(20), while 9b showed a conjugated oxo group at C(20). The presence of these

1

functionalgroups were confirmed by MS and UV data ( Exper. Part). The H- and

1

3

C-NMR data of 9c and 9d clearly indicated that only the C(5)C(6) bond was

present, besides an aldehyde functionality at C(21) and an a-OH group at C(17) in the

case of 9c, and a corresponding a,b-unsaturated aldehyde structure in the case of 9d

(

see data in Exper. Part) MS and UV data were in agreement with the proposed

structures.

Finally, when the acetyl derivative 10 of pregnadienol 9 was oxygenated in the usual

manner, two products 10a and 10b were isolated, besides starting material. The R value

f

of 10b corresponded to that of authentic 9d, and its spectraldata estabi ls hed the

1

13

identity of the two compounds (see Exper. Part). The H- and C-NMR spectra of 10a

showed the presence of the unmodified C(5)C(6) bond and of a C(16)C(17) bond

with an oxo function at C(20). Additionaldata supported the a,b-unsaturated keto

structure of 10a (see Exper. Part).

1

13

It is interesting to note that both the H- and C-NMR and the MS data (EI-, HR-,

and ESI-TOF-MS) revealed the presence of the unmodified 3-{[(tert-butyl)dimethyl-

1

silyl]oxy}substituent in all the products obtained by reaction of O₂with the

unsaturated pregnanes 1 ± 10; this stability was also observed in the pregn-5-ene series

6

± 10, even when the ene reaction took place at the C(5)C(6) bond.

1

Discussion. ± Preamble. The reaction of O with unsaturated olefins provides a

2

valuable and convenient route to introduce O-functionality in the reactants [7a]. The

major reaction pathway for alkenes with H-atoms in allylic position is the ene reaction

[

7a,b], which is a powerful method for the synthesis of allyl hydroperoxides. The

hydroperoxides may then be transformed into allyl alcohols, allylic carbonyl

compounds (ketones, aldehydes), or may be used as O-transfer agents in epoxidation

reactions (Scheme 1) [7c].

Scheme 1. Singlet-Oxygen Ene Reaction and Photooxygenated Products

The reaction shows low regioselectivity when simple trisubstituted alkenes are used

[

7d] but shows a noticeable side selectivity called ꢁcis effectꢂ [7e], (the higher

substituted side of the trisubstituted alkene is more reactive) for the same compounds

Scheme 2). Additionalregio- and stereose el ctivity phenomena observed in the ene

(

Helvetica Chimica Acta ± Vol. 87 (2004)

2993

reaction suggest a perepoxide intermediate or a perepoxide-like transition-state

geometry [7a,f ± g], making this reaction usefulin stereose el ctive synthesis. Further-

more, for chiral allyl alcohols and analogous substrates, the ꢁhydrogen-directing effectꢂ

described by Adam et al. [7g,h] helps the reaction to proceed with high regioselectivity

and high syn diastereoselectivity (Scheme 3). Thus, depending on the nature of the

functional groups at the allylic position around a CC bond and on the experimental

conditions (i.e., solvent polarity), photooxygenation of ene compounds can be a

valuable tool in stereoselective synthesis [7g,i].

Scheme 2. The cis Effect in the Singlet Ene Reaction with 2-Methylbut-2-ene. The percentages of H-abstraction

are indicated.

Scheme 3. The Hydroxy-Directing Effect in the Ene Reaction of the Allyl Alcohol 4-Methylpent-3-en-2-ol

Behavior of Pregn-17(20)-enes 1 ± 5 and Pregna-5,17(20)-dienes 6 ± 10. The low

regioselectivity observed when noncyclic trisubstituted alkenes are studied changes, as

1

a rule, when cyclic trisubstituted alkenes are attacked by O . The photosensitized

2

oxidation in pyridine solution of pregnene 1 with the exocyclic C(17)C(20) bond

yielded as the sole product the 16-en-20-al derivative 1a. This is the stable carbonyl

derivative formed from the expected primary C(20) hydroperoxide (not observed)

after elimination of H O.

2

1

As the 17-keto derivative formed by a [2 2]-cycloaddition of O and the CC

2

bond was not detected in any of the unsaturated pregnanes studied, the ene reaction is

the preferentialreaction pathway when the exocyc il c C(17) C(20) bond reacts with

1

O .

2

Pregnene 2, the (Z)-17-ethylidene analogue of 1, gave on photooxygenation the

now secondary allyl hydroperoxide 2a as a stable product together with its dehydrated

product 2b. Dehydration of allyl hydroperoxides in pyridine solution has been

previously reported [7g]. Although the C(17)C(2) bond in pregnene 2 represents a

trisubstituted alkene, with two potentially active allylic H-atoms at C(16) and C(21)

and with the higher substituted side involving Me(21), only 2a and 2b were isolated.

Neither the 20-en-17-ylhydroperoxide nor any product formed from this hydro-

peroxide intermediate was observed in the reaction mixture. Thus, the so-called ꢁcis

effectꢂ [7e] is not operating in the photooxygenation of C(17)C(20) in pregnene 2.

To evaluate the so-called ꢁhydrogen-directing effectꢂ [7g,h] in the photooxygenation

of the C(17)C(20) bond, the behavior of the unsaturated hydroxypregnanols 3, 4, 6, 7,

and 9 and of two of their corresponding O-acetylderivatives, 5 and 10, was compared

2

994

Helvetica Chimica Acta ± Vol. 87 (2004)

(

Fig. 2). The formation of hydroxy ketone 4a can be rationalized by an ene reaction

involving the allylic H-atom located at C(16) of 17(20)-en-16-ol 4, yielding the 20-

hydroperoxy-16-en-16-ol intermediate, which leads to the more stable 20-hydroxy-16-

one 4a (Scheme 4). As mentioned above, the absence of the corresponding 17-

hydroperoxy-20-en derivative and of products formed from this intermediate in the

product mixture obtained from 4 suggests again that the so-called ꢁcis effectꢂ [7e] is not

operating in the photooxygenation of the C(17)C(20) bond. The 16-O-acetyl

derivative 5 of 4, also yielded 4a (ꢀ5a) i.e., deacetylation had occurred under the

photooxygenation conditions. In the second product 5b, the C(17)C(20) bond of 5

was shifted to C(20)C(21) and the AcO still located at C(16); this can be rationalized

1

by the interaction of O with the allylic HÀC(21) of 5 leading to the 20-en-17a-yl

2

hydroperoxide intermediate (not detected) which yields the stable 16-(acetyloxy)-20-

en-17a-ol 5b.

Scheme 4. Singlet-Oxygen Attack at the Pregn-17(20)-ene Double Bond. Formation of 20-Hydroxypregnan-16-

one

The results obtained with 4 and 5 and the fact that no photooxygenated product was

isolated from pregnen-16-ol 3, the 16a-isomer of 4, suggests that a concrete effect is

operating due to the presence of the a-OH group at C(16). As it is known,

stereoselective transformation with asymmetric induction by an allylic stereogenic

1,3

1,2

center is rationalized in terms of allylic A and A strain (Fig. 3) [7j,k]. This concept

presupposes the eclipsed alignment of an allylic CÀH bond and the adjacent CC

bond. These stereoelectronic factors determine, in flexible structures, the preferential

1

conformers which interact with O . On the basis of the allylic-strain model, it may be

2

assumed that in all substrates, a nearly perpendicular alignment of the allylic OH group

is adopted. From the allylic functional group, two major effects may be anticipated:

electronic influence on the reactivity of the adjacent CC, and noncovalent

1

interactions with the incoming electrophilic O molecule. The second effect has been

2

described as activating or de-activating [7j]. In the case of unsaturated pregnanes, due

1

,3

to the reduced flexibility of the cyclopentane ring, we propose that both the allylic A

1,2

1

and A strains are not operating, and the ene reaction of compounds 2 and 4 with O

2

proceeds with high regioselectivity, yielding the 20-yl hydroperoxide. Furthermore, the

inertness of pregnen-16a-ol 3 shows that a cis-effect is not operating in this case, and, as

a consequence, the major effect of the hydroxyallyl moiety is deactivating the a-attack

1

of O at C(17)C(20). As the b-attack does not occur due to the fixed axial

2

1

MeÀC(18), the hydroxyallyl moiety of 3 is not attacked by O . When the substituent

2

at C(16) is located on the b-side as in the pregnen-16b-ol 4, this deactivating effect is not

1

operating, and the O a-attack at the hydroxyallyl system occurs yielding the

2

corresponding 20-ylhydroperoxide and, hence, the 20-hydroxy-16-one 4a.

Helvetica Chimica Acta ± Vol. 87 (2004)

2995

Fig. 3. Lowest-energy conformers due to allylic A¹^,³and A¹^,²strain

The apparent deactivation or protective effect of the 16a-OH substituent was also

observed in the photooxidation of pregnadienols 6 and 7 on comparison with the

behavior of pregnadiene 8. Moreover, the results obtained with 6 ± 8 allowed also a

comparison with the recently described sensitized photooxygenation of several 3-

substituted derivatives of cholesterol [5]; this family of steroids represent trisubtituted

cyclic alkenes whose C(5)C(6) bond can be regarded as exocyclic with respect to ring

A as well as endocyclic with respect to ring B. Thus, the photooxygenation of pregna-

5

7

,17(20)-dien-16a-ols 6 and 7 occurred at the allylic HÀC(7) leading to the expected

a-ylhydroperoxides 6a and 7a, respectively. However, the pregna-5,17(20)-diene 8,

1

which carries no OH group at C(16), reacted with O in a regioselective way at the

2

C(17)C(20) bond, yielding the 16-en-20-yl hydroperoxide 8a and the 16-en-20-ol 8b.

In a similar way, the C(5)C(6) bond was kept intact on photooxygenation of the

1

pregna-5,17(20)-dienes 9 and 10, although the O attack at C(17)C(20) was not

2

regioselective. Not only the 16-en-20-yl hydroperoxide 9a and its dehydrated 16-en-20-

one derivative 9b but also 17-hydroxy-21-al 9c and 17(20)-en-21-al 9d were isolated.

The former ene reaction involved HÀC(16) and the latter HÀC(21). Similar results

were obtained by photooxygenation of the acetylderivative 10 of 9.

Concerning the diastereoselectivity of the photooxygenation of pregnanes 1 ± 10,

1

the O attack showed a-facialstereose el ctivity for compounds 2, 4, 5, and 8. As a

2

consequence, the products 2a, 4a, 5a, and 8a were obtained as the corresponding

hydroperoxy and hydroxy compounds with (20S)-configuration.

Finally, the photooxygenation of pregnadienol 9 showed low regioselectivity but

both ene reactions proceeded with very high diastereoselectivity. Steric factors explain

the preferential a-p-facialattack when 9c was formed via the 17-hydroperoxy

intermediate, while a hydroxy-directing effect would account for the diastereoselec-

tivity of the formation of 21-hydroperoxy derivative 9a.

Stereoselectivity and Regioselectivity at C(17)C(20). The a-p-facialstereose el c-

1

tivity of the reaction of O with C(17)C(20) of the unsaturated pregnanes 1 ± 10 to

2

form the oxiranium-oxide-like intermediate (Scheme 5) can be rationalized by the fact

that MeÀC(18) is on the b-side (axial) (Fig. 4). Thus, as expected, the photosensitized

oxidation of pregnen-16a-ol 3 did not afford any product because of the stereo-

electronically deactivating effect of the a-OH substituent at C(16). Similarly,

pregnadien-16a-ols 6 and 7 yielded only the expected 7a-ylhydroperoxides via the

normalene reaction at C(5) C(6) keeping the C(17)C(20) bond intact.

From the oxiranium-oxide-like intermediate, the reaction can follow two different

pathways: i) involvement of the allylic HÀC(16) on the a-face (compounds 1, 2, 4, 5,

and 8 ± 10) and/or ii) involvement of one HÀC(21) (compounds 2, 4, 5, and 8 ± 10). In

2

996

Helvetica Chimica Acta ± Vol. 87 (2004)

Scheme 5. Singlet-Oxygen Attack and the para-Epoxy-like Intermediate

Fig. 4. Singlet-oxygen attack at the pregn-17(20)-ene double bond. R, R'H, OH.

the case of 2, 4, 5, and 8, the products obtained showed that the former was the only

mechanism operating. For 9 and 10, both mechanisms were operating, the former

leading to 9a and 9b and the latter to 9c and 9d in the case of 9.

1

Regioselectivity of the Reaction of 8 and 9. The clear regioselectivity of the O ene

2

reaction at C(17)C(20) of 8 and 9 was a bit surprising. Analysis of the geometry-

optimized structures (semi-empirical(AM1 and PM3); ab initio (HF/3-21G and

1

B3LYP/3-21G) level) showed that O attack at the C(5)C(6) bond would be feasible

2

from the steric point of view. However, the HOMOs of 8 and 9 were calculated

(

2

ZINDO/S//AM1, ZINDO/S//PM3, ZINDO/S/HF/3-21G, and ZINDO/S//B3LYP/3-

1G level), and the shapes of the calculated orbitals were almost the same in all cases,

the largest coefficients being found at C(17)C(20), and no important (or any)

1

contribution of the C(5)C(6) p-system to the HOMO was observed. As the O ene

2

reaction with alkenes can be considered an electrophilic attack forming an exciplex in

the rate-determining step (Scheme 5), the analysis of the HOMO properties of the

pregnadienes 8 and 9 agrees with the regioselectivity observed. In the same way, when

1

O₂attacks the C(17)C(20) bond region with partialstructures

C(16)C(17)C(20)ÀC(21) or b-HOÀC(16)ÀC(17)C(20)ÀC(21), only the 16-

en-20-hydroperoxy and/or its dehydrated and/or reduced products were obtained.

Conclusions. ± Photooxygenation of CC bonds of unsaturated pregnane

structures depends on i) their location in the skeleton, ii) the location and nature of

the allylic functional groups, and iii) steric factors. The C(17)C(20) bond proved to be

more reactive than the C(5)C(6) bond [8]. This behavior was very different when an

Helvetica Chimica Acta ± Vol. 87 (2004)

2997

OH substituent was present at C(16) on the a-side. For the ene reaction involving the

C(17)C(20) bond, high regioselectivity was found when the allylic OH group was

located at C(16) on the b-side, whereas no regioselectivity was observed when the

allylic OH group was located at C(21). The stereoselectivity of these reactions was

dependent not only on steric factors (i.e., axialMe ÀC(18)) but also on the activating or

deactivating effect of the allylic OH groups at C(16) and C(21), respectively.

We wish to acknowledge the Universidad de Buenos Aires (022), CONICET (PIP/904), and ANPCyT

(

PICT 12312) for partialfinancialsupport, and UMYMFOR (UBA-CONICET) for technicalsupport. R. E.-B.

is a research member of CONICET. ESI-TOF-MS Experiments were performed as part of the academic

agreement between R. E.-B. (FCEyN, Universidad de Buenos Aires, Argentina) and Hiroshi Nonami (CA,

Ehime University No, Japan) with the facilities of the high-resolution liquid-chromatography-integrated mass

spectrometer system of the United Graduated Schoolof Agricu tl uralSciences (Ehime University, Japan).

Experimental Part

General. Toluene, hexane, CH

MeCN, EtOH, MeOH, BuOMe and pyridine were obtained from JT Baker. Rose Bengal(sodium sa tl 95%),

2

Cl

2

, CHCl

3

, Et

2

O, and AcOEt of anal. grade were used after purification.

t

(

3b,5a)-3-hydroxyandrostan-17-one and (3b)-3-hydroxyandrost-5-en-17-one were purchased from Aldrich.

Column chromatography (CC): Merck silica gel 60 (0.040 ± 0.063 mm). TLC: Merck silica gel 60 F254 on

aluminium sheets (0.2 mm thickness). M.p.: Fisher ± Jones apparatus; not corrected. UV Spectra: Shimadzu UV-

À1

1

203 spectrophotometer; lmax in nm. IR Spectra: Nicolet Magna-550-FT/IR spectrophotometer; in cm . H- and

1

3

4

C-NMR Spectra: Bruker AC-200 and Bruker 500 spectrometer; chemicalshifts d in ppm rel. to internal SiMe ,

J in Hz; standard pulse sequences. EI-MS: VG TRIO-2 mass spectrometer; in m/z (rel. int. in %). HR- and FAB-

MS: VG ZAB BEQ instrument. ESI-TOF-MS: Mariner Applied-Biosystems, ESI-TOF spectrometer, in

positive-ion mode (no signals in the negative-ion mode); m/z range 10 ± 1000 Da; infusion of 0.5 ± 50 mlof the

steroid soln. in THF into the ion source by a Harvard PHD-2000 syringe infusion pump at a flow-rate of 5 ml/

min, with MeOH/H

skimmer voltage 11.01 V, nozzle temp. 1408; N

mass calibration with 1 mm standard in MeOH/H

2

O 9 :1 as solvent stream; spray-tip potential 2796.68 V, nozzle potential 245.12 V,

2

flow rate 0.40 l/min, analyzer temp. 29.08, pressure 0.55 MPa;

O 9 :1 (standards: caffeine, tetrabutylammonium bromide, b-

2

estradiol 3-sulfate 17-glucuronide dipotassium salt, g-cyclodextrin, and angiotensin I (z 1, 1296.685 Da; z 2,

48.342 Da).

3b,5a)-3-{[(tert-Butyl)dimethylsilyl]oxy}-21-norpregn-17(20)-ene (1). A mixture of methyltriphenylphos-

6

(

t

phonium bromide (4.2 g, 11.8 mmol) and KO Bu (1.27 g, 11.4 mmol) in dry THF (30 ml) was vigorously stirred

for 5 min. To this suspension, (3b,5a)-3-{[(tert-butyl)dimethylsilyl]oxy}androstan-17-one (prepared from

commercial(3 b,5a)-3-hydroxyandrostan-17-one according to [8][9]; 1 g, 2.5 mmol) in THF (5 ml) was added.

The mixture was heated at 558 for 15 h and partitioned between hexane (30 ml) and 50% aq. MeOH (50 ml).

The H

2

O/MeOH layer was extracted with hexane (3 Â 10 ml), the combined org. phase evaporated, and the

residue purified by filtration through a short column (silica gel, hexane): 1 (0.92 g, 93%). M.p. 203 ± 2058

1

t

(

3

(

hexane/EtOH). H-NMR (CDCl

3

, 200 MHz): 0.05 (s, Me

2

Si); 0.77 (s, Me(18)); 0.82 (s, Me(19)); 0.88 (s, Bu);

Si); 12.4 (C(19)); 18.5

.55 (m, HÀC(3)); 4.61 (br. s, 2 HÀC(20)). ¹³C-NMR (CDCl

, 500 MHz): À4.5 (Me

3

2

.

3 3

C(18)); 18.2 (Me C); 25.9 (Me C); 72.2 (C(3)); 100.6 (C(20)); 162.1 (C(17)). EI-MS (70 eV): 402 (0.6, M ),

t

3

C

87 (1, [M À Me] ), 345 (62, [M À Bu] ), 269 (7), 75 (100, Me

2

SiOH ). HR-MS: 345.2606 ([M À Bu] ,

22

H

37OSi , calc. 345.2604). ESI-TOF-MS (pos.): 425.320 ([M Na] , C26

H46NaOSi ; calc. 425.3204), 403.3384

(

[M H] , C26

H47OSi ; calc. 403.3384).

The corresponding (3b,5a)-21-norpregn-17(20)-en-3-olwas described e sl ewhere [10].

(

3b,5a,17Z)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregn-17(20)-ene (2) was prepared from (3b,5a)-3-{[(tert-

1

butyl)dimethylsilyl]oxy}androstan-17-one (see above) according to [8][9]. H-NMR (CDCl

s, Me

3

, 200 MHz): 0.05

t

(

7

(

2

Si); 0.86 (s, Me(18)); 0.81 (s, Me(19)); 0.88 (s, Bu); 3.55 (m, HÀC(3)); 5.11 (tq, J(20,16) 2, J(20,21)

1

3

, H ÀC(20)); 1.64 (dt, J(21,16) 2, J(20,21) 7, Me(21)). C-NMR (CDCl

3

, 200 MHz): À4.5 (Me

2

Si); 12.4

C(19)); 13.1 (C(18)); 16.9 (C(21)); 18.2 (Me

3

C); 25.9 (Me

3

C); 72.2 (C(3)); 113.2 (C(20)); 150.4 (C(17)). ESI-

TOF-MS (pos.): 439.336 ([M Na] , C27

H

48NaOSi ; calc. 439.3360), 417.354 ([M H] , C27

H49OSi ; calc.

4

17.3540).

2

998

Helvetica Chimica Acta ± Vol. 87 (2004)

(

3b,5a,16a-17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregn-17(20)-en-16-ol (3) was prepared from 2 accord-

1

t

ing to [8][9]. H-NMR (CDCl

3

, 200 MHz): 0.06 (s, Me

2

Si); 0.87 (s, Me(18)); 0.82 (s, Me(19)); 0.90 (s, Bu); 3.57

(

m, HÀC(3)); 5.58 (dq, J(20,16) 1, J(20,21) 7, H ÀC(20)); 1.74 (dd, J(21,16) 2, J(20,21) 7, Me(21)); 4.41

1

3

br. d, HÀC(16)). C-NMR (CDCl

3

, 200 MHz): À4.6 (Me

2

Si); 12.3 (C(19)); 13.1 (C(18)); 17.5 (C(21)); 18.2

Me

3

C); 25.9 (Me

3

C); 72.1 (C(3)); 119.2 (C(20)); 155.5 (C(17)); 74.31 (C(16)). ESI-TOF-MS (pos.); 455.330

[M Na] , C27

H

48NaO

2

49 2

Si ; calc. 455.3309), 433.348 ([M H] , C27H O Si ; calc. 433.3489).

(

3b,5a,16b,17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregn-17(20)-en-16-ol (4). A mixture of (3b,5a,17E)-3-

{

1

[(tert-butyl)dimethylsilyl]oxy}pregn-17(20)-en-16-one (prepared from 3 according to [8][9]; 526.9 mg,

i

.22 mmol) and NaBH

4

(200 mg, 5.26 mmol) in MeOH/ PrOH 3 :1 (20 ml) was stirred at r.t. for 3 h. The

reaction was quenched by addition of H

2

2 2

O/AcOH 100 :5 (105 ml) and extracted with CH Cl . The org. layer was

washed with H

2

O, dried (Na

2

SO

4

), and evaporated. The solid obtained was recrystallized: 4 (524 mg, 100%).

1

M.p. 215 ± 2188 (dec.; hexane/EtOH). H-NMR (CDCl

3

, 200 MHz): 0.05 (s, Me

2

Si); 0.82 (s, Me(19)); 0.89

t

(

s, Bu); 1.02 (s, Me(18)); 1.72 (dd, J(16,21) 1.7, J(20,21) 7.2, Me(21)); 3.55 (m, HÀC(3)); 4.34 (br. t,

1

3

HÀC(16)); 5.49 (dq, J(20,16) 1.7, J(20,21) 7.2, HÀC(20)). C-NMR (CDCl

3

, 500 MHz): À4.5 (Me

2

Si); 12.3

(C(19)); 13.2 (C(18)); 18.2 (Me C); 25.9 (Me C); 72.1 (C(3)); 75.4 (C(16)); 118.5 (C(20)); 155.1 (C(17)). EI-

3

.

t

MS (70 eV): 432 (1.4, M ), 417 (1.7, [M À Me] ), 375 (22, [M À Bu] ), 281 (7), 75 (100, Me

2

SiOH ). HR-MS:

t

3

75.2711 ([M À Bu] , C23

H

39

O

2

Si ; calc. 375.2709). ESI-TOF-MS (pos.): 455.330 ([M Na] , C27

2

H48NaO Si ;

calc. 455.3309), 403.3384 ([M H] , C27

49 2

H O Si ; calc. 433.3489).

The (3b,5a,16b)-pregn-17(20)-ene-3,16-dioldiacetate ana ol gue has been described as a compound with

unkown double-bond geometry [11].

(

3b,5a,16b,17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregn-17(20)-en-16-ol Acetate (5). A soln. of 4 (300 mg,

.69 mmol) in Ac O (1 ml, 10 mmol) and pyridine (2 ml) was stirred at r.t. for 15 h. The mixture was diluted with

cold H O, and dil. HCl soln. was added to neutrality. The mixture was extracted with CH Cl , the combined org.

phase washed with H

0

2

O, dried (Na

2

SO

4

), and evaporated, and the solid obtained recrystallized: 5 (297.5 mg,

1

9

(

1%). M.p. 195 ± 1978 (hexane/EtOH). H-NMR (CDCl

3

, 200 MHz): 0.05 (s, Me

2

Si); 0.82 (s, Me(19)); 0.89

t

s, Bu); 1.01 (s, Me(18)); 1.71 (br. d, J 7, Me(21)); 2.05 (s, MeCO)); 3.55 (m, HÀC(3)); 5.42 (m, HÀC(20),

1

3

t

HÀC(16)). C-NMR (CDCl

3

, 500 MHz): À4.5 (Me

2

Si); 12.3 (C(19)); 13.5 (C(18)); 17.5 (C(21)); 18.3 ( Bu);

2

(

Me

1.5 (MeCO); 26.0 (Me C); 72.1 (C(3)); 77.0 (C(16)); 120.4 (C(20)); 149.4 (C(17)); 171.3 (MeCO). EI-MS

3

.

t

70 eV): 474 (5, M ), 459 (2, [M À Me] ), 431 (10, [M À Me À CO] ), 399 (32, [M À Bu] ), 281 (7), 75 (100,

t

2

SiOH ). HR-MS: 417.2816 ([M À Bu] , C25

H

41

O

3

Si ; calc. 417.2814). ESI-TOF-MS (pos.): 497.000 ([M

Na] , C29

For a description of this compound in the literature, see 4.

3b)3-{[(tert-Butyl)dimethylsilyl]oxy}androst-5-en-17-one was prepared by a procedure different from that

H

50NaO

3

Si ; calc. 497.0001), 475.000 ([M H] ; C29

51 3

H O Si ; calc. 475.0000).

(

previously described [12]. To a soln. of commercial (3b)-3-hydroxyandrost-5-en-17-one (1 g, 3.5 mmol) in

t

CH

2

Cl

2

(25 ml) was added BuMe

2

SiCl(1.5 g, 9.7 mmo l) , Et

3

N (1.1 ml, 7.9 mmol), and N,N-dimethylpyridin-4-

O (300 ml), and extracted with

Clso nl . and brine, dried (Na SO ), and evaporated, and

amine (914 mg, 7.5 mmol). The mixture was stirred at r.t. for 15 h poured into H

CH Cl . The org. layer was washed with 10% aq. NH

2

4

2

4

the residue purified by CC (silica gel, 5% AcOEt/hexane): pure product (1.4 g, 100%). M.p. 151 ± 1528 (MeOH)

1

(

[12]: 146 ± 1478). FT-IR (KBr): 1748 (CO), 1092 (SiÀOÀC), 837, 774 (Me

2

Si). H-NMR (CDCl

3

, 200 MHz):

t

0

.06 (s, Me

2

Si); 0.88 (s, Me(18)); 1.03 (s, Me(19)); 0.89 (s, Bu); 3.49 (m, HÀC(3)); 5.34 (br. d, J 5, HÀC(6)).

1

3

C-NMR (CDCl

3

), 500 MHz): À4.6 (Me

2 3 3

Si); 13.5 (C(18)); 18.2 (Me C); 19.4 (C(19)); 25.9 (Me C); 72.4

.

(

[

C(3)); 120.3 (C(6)); 141.8 (C(5)); 220.8 (C(17)). EI-MS (70 eV): 402 (0, M ), 387 (3, [M À Me] ), 345 (99,

t

.

M À Bu] ), 269 (10), 75 (100, Me

2 42 2

SiOH ). HR-MS: 402.2956 (M , C25H O Si ; calc. 402.2954). Anal. calc.

for C25

42 2

H O

Si: C 74.6, H 10.5; found: C 74.4, H 10.6.

(

3b,16a)-3-{[(tert-Butyl)dimethylsilyl]oxy}-21-norpregna-5,17(20)-dien-16-ol (6). As described for 1, with

t

methyltriphenylphosphonium bromide (4.2 g, 11.8 mmol), KO Bu (1.27 g, 11.4 mmol), dry THF (30 ml), (3b)-

3

-{[(tert-butyl)dimethylsilyl]oxy}androst-5-en-17-one (1 g, 2.5 mmol), and THF (5 ml): (3b)-3-{[(tert-butyl)di-

methylsilyl]oxy}-21-norpregna-5,17(20)-diene (0.93 g, 93%), which was not isolated and characterized.

To a suspension of selenium dioxide (40 mg, 0.36 mmol) in CH Cl (10 ml) at 08 was added dropwise 70%

2

tert-butyl hydroperoxide soln. 0.07 ml, 0.7 mmol). After stirring for 1 h at r.t., a homogeneous soln. was

obtained. A soln. of (3b)-3-{[(tert-butyl)dimethylsilyl]oxy}-21-norpregna-5,17(20)-diene (300 mg, 0.75 mmol)

in CH Cl (10 ml) was added dropwise, and the mixture was stirred for 30 min at r.t. Toluene was added (40 ml),

2 2

and the resulting mixture was washed, with 10% aq. NaOH soln. (3 Â 10 ml) and brine (3 Â 10 ml), the org.

phase dried (Na

2

SO

4

) and evaporated, and the crude solid thus obtained purified by CC (silica gel, 3% AcOEt/

1

hexane): 6 (218 mg, 70%). M.p. 199 ± 2008 (hexane/EtOH). H-NMR (CDCl

3

, 200 MHz): 0.06 (s, Me

2

Si); 0.80

t

(

s, Me(18)); 0.89 (s, Bu); 1.02 (s, Me(19)); 3.49 (m, HÀC(3)); 4.67 (dd, HÀC(16)); 4.90 (d, J 2, 1 HÀC(20));

Helvetica Chimica Acta ± Vol. 87 (2004)

2999

5

.10 (d, J 2, 1 HÀC(20)); 5.32 (br. d, J 5, HÀC(6)). ¹³C-NMR (CDCl

3

, 500 MHz): À4.6 (Me

2

Si); 18.0

3 3

(C(18)); 18.2 (Me C); 19.4 (C(19)); 25.9 (Me C); 72.5 (C(3)); 72.7 (C(16)); 103.8 (C(20)); 141.6 (C(5)); 120.8

.

t

(

Me

C(6)); 165.9 (C(17)). EI-MS (70 eV): 416 (1, M ), 399 (2, [M À OH] ), 359 (20, [M À Bu] ), 281 (7), 75 (100,

t

2

SiOH ). HR-MS: 359.2399 ([M À Bu] , C22

H

35

O

2

Si ; calc. 359.2397). ESI-TOF-MS (pos.): 439.299 ([M

Na] , C26

H

44NaO

2

Si ; calc. 439.2997), 417.317 ([M H] , C26

45 2

H O Si , calc. 417.3177).

(

3b,16b,17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregna-5,17(20)-dien-16-ol (7). As described for 6 (second

part), with selenium dioxide (40 mg, 0.36 mmol), CH Cl (10 ml), 70% tert-butyl hydroperoxide soln. (0.07 ml,

2

0

.7 mmol), 8 (350.6 mg, 0.85 mmol), and CH

2

Cl

2

(10 ml): 7 (297.3 mg, 82%). M.p. 149 ± 1508 (MeOH/H

2

O). FT-

Si); 0.88

1

IR (KBr): 3278 (OH), 1098 (SiÀOÀC), 837; 774 (Me

2

Si). H-NMR (CDCl

3

, 200 MHz): 0.06 (s, Me

2

t

(

5

s, Me(18)); 0.89 (s, Bu); 1.02 (s, Me(19)); 1.74 (dd, J(21,16) 1, J(20,21) 7, Me(21)); 3.49 (m, HÀC(3)); 4.43

1

3

dd, HÀC(16)); 5.32 (br. d, J 5, HÀC(6)); 5.59 (dq, J(20,16) 1, J(20,21) 7, H ÀC(20)). C-NMR (CDCl

3

,

00 MHz): À4.6 (Me Si); 13.2 (C(18)); 17.3 (C(21)); 18.2 (Me C); 19.4 (C(19)); 25.9 (Me C); 72.5 (C(3)); 74.4

2

3

(

C(16)); 119.6 (C(20)); 120.9 (C(6)); 141.5 (C(5)); 155.4 (C(17)). EI-MS (70 eV): 430 (0, M ), 373 (38, [M À

t

Bu] ), 355 (7, [M À 57 À 18] ), 75 (100, Me

2

SiOH ). HR-MS: 373.2555 ([M À Bu] , C23

H

37

O

2

Si ; calc.

3

73.2553). ESI-TOF-MS (pos.): 453.315 ([M Na] , C27

H

46NaO

2

Si ; calc. 453.3153), 431.333 ([M H] ,

C

27

H

47

O

2

Si ; calc. 431.3333).

The corresponding, 3b,16a,17E)-pregna-5,17(20)-diene-3,16-diolhas been described [13].

(

3b,17Z)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregna-5,17(20)-diene (8). As described for 1, with ethyl-

t

triphenylphosphonium bromide (4 g, 10.7 mmol), KO Bu (1.15 g, 10.3 mmol), dry THF (30 ml), (3b)-3-{[(tert-

butyl)dimethylsilyl]oxy}androst-5-en-17-one (1 g, 2.5 mmol), and THF (5 ml): 8 (0.98 g, 95%). M.p. 129 ± 1308

1

(

2

CH

2

Cl

2

/MeOH). FT-IR: (KBr): 3030 (CC), 1086, 890 (SiÀOÀC), 850, 773 (Me

2

Si). H-NMR (CDCl

3

,

t

00 MHz): 0.06 (s, Me

2

Si); 0.88 (s, Me(18)); 0.89 (s, Bu); 1.01 (s, Me(19)); 1.66 (dd, J(21,16) 2, J(20,21) 7,

Me(21)); 3.49 (m, HÀC(3)); 5.13 (dq, J(20,16) 2, J(20,21) 7, H ÀC(20)); 5.32 (br. d, J 5, HÀC(6)).

1

3

C-NMR (CDCl

3

, 500 MHz): À4.6 (Me

2 3

Si); 13.1 (C(18)); 16.6 (C(21)); 18.2 (Me C); 19.4 (C(19)); 26.0

.

(

C

Me C); 72.6 (C(3)); 113.5 (C(20)); 121.0 (C(6)); 141.6 (C(5)); 150.2 (C(17)). EI-MS (70 eV): 414 (0, M ), 399

3

t

0.5, [M À Me] ), 357 (30, [M À Bu] ), 281 (6), 75 (100, Me

2

SiOH ). HR-MS: 357.2606 ([M À Bu] ,

23

H

37OSi ; calc. 357.2604). ESI-TOF-MS (pos.): 437.325 ([M Na] , C27

H46NaOSi ; calc. 437.3258), 415.338

(

[M H] , C27

The corresponding (3b,17Z)-pregna-5,17(20)-diene-3-ol has been fully described [14].

3b,17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregna-5,17(20)-dien-21-ol (9). A mixture of diethoxyphosphin-

ylacetate (3 ml, 15.1 mmol) and NaOEt (344 mg of Na/5 ml of abs. EtOH, 15.0 mmol of NaOEt in abs. EtOH

20 ml) was vigorously stirred for 5 min. To this soln. was added (3b)-3-{[(tert-butyl)dimethylsilyl]oxy}androst-

-en-17-one (600 mg, 1.5 mmol) in EtOH (5 ml). The mixture was heated under reflux for 15 h and then poured

into H O (100 ml) and extracted with CH Cl SO ) and

H47OSi ; calc. 415.3384).

(

5

2

(3 Â 25 ml). The combined org. phase was dried (Na

2

4

evaporated and the residue purified by CC (silica gel, 1% AcOEt/hexane to give the corresponding 5,17(20)-

1

dien-21-oic acid ethylester (650 mg, 92%). M.p. 187 ± 189 8 (CH

(

2

Cl

2

/MeOH). H-NMR (CDCl

3

, 200 MHz): 0.06

); 3.48 (m, HÀC(3)); 4.15

3

t

s, Me

q, J 7, MeCH

Si); 12.4 (C(18)); 19.4 (C(19)); 18.2 (Me

2

Si); 0.84 (s, Me(18)); 0.89 (s, Bu); 1.02 (s, Me(19)); 1.28 (t, J 7, MeCH

2

1

2

); 5.32 (br. d, J 5, HÀC(6)); 5.54 (t, J(20,16) 2, HÀC(20)). C-NMR (CDCl

3

, 500 MHz):

À4.6 (Me

2

3

C); 26.8 (Me

3

C); 72.6 (C(3)); 108.4 (C(20)); 121.0 (C(6));

.

1

41.6 (C(5)); 167.5 (C(17)); 176.5 (C(21)). EI-MS (70 eV): 472 (0.2, M ), 457 (2.4, [M À Me] ), 415 (100, [M À

t

Bu] ), 75 (93, Me

A mixture of LiAlH

10 ml) was vigorously stirred for 30 min. To this suspension was added H

2

SiOH ). HR-MS: 415.2214 ([M À Bu] , C29

H

48 3

O Si ; calc. 415.2212.

4

(95 mg, 2.5 mmol) and 5,17(20)-dien-21-oic ethyl ester (232 mg, 0.5 mmol) in dry THF

O (50 ml). The aq. phase was

(

2

extracted with AcOEt (3 Â 10 ml), the combined org. phase evaporated, and the residue purified by CC (silica

1

gel, 5% AcOEt/hexane: 9 (200 mg, 95%). M.p. 210 ± 2128 (MeOH). H-NMR (CDCl

3

, 200 MHz): 0.06

t

(

s, Me

2

Si); 0.78 (s, Me(18)); 0.89 (s, Bu); 1.02 (s, Me(19)); 3.49 (m, HÀC(3)); 4.13 (m, 2 HÀC(21)); 5.25

1

3

m, HÀC(20)); 5.32 (br. d, J 5, HÀC(6)). C-NMR (CDCl

3

, 500 MHz): À4.6 (Me

2

Si); 18.5 (C(18)); 19.4

C(19)); 18.2 (Me C); 25.9 (Me C); 60.4 (C(21)); 72.6 (C(3)); 115.6 (C(20)); 120.9 (C(6)); 141.7 (C(5)); 155.8

3

.

t

C(17)). EI-MS (70 eV): 430 (0.3, M ), 415 (2.8, [M À Me] ), 355 (9), 373 (100, [M À Bu] ), 281 (10), 75 (65.5,

t

Me

Na] , C27

2

SiOH ). HR-MS: 373.2555 ([M À Bu] , C23

H

37

O

2

Si ; calc. 373.2553). ESI-TOF-MS (pos.): 453.315 ([M

H

46NaO

2

Si ; calc. 453.3153), 431.333 ([M H] , C27

47 2

H O Si ; calc. 431.3333).

The (3b,17E)-pregna-5,17(20)-diene-3,21-diol3-acetate has been described [15].

3b,17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}-5,17(20)-dien-21-ol Acetate (10). As described for 5, with 9

(

300 mg, 0.70 mmol), Ac

2

O (1 ml, 10 mmol), and pyridine (2 ml): 10 (300 mg, 90%). M.p. 173 ± 1758 (hexane/

1

t

EtOH). H-NMR (CDCl

3

, 200 MHz): 0.06 (s, Me

2

Si); 0.79 (s, Me(18)); 0.89 (s, Bu); 1.02 (s, Me(19)); 2.05

(

s, MeCO); 3.49 (m, HÀC(3)); 4.54 (m, 2 HÀC(21)); 5.18 (m, HÀC(20)); 5.32 (br. d, J 5, HÀC(6)).

3

000

Helvetica Chimica Acta ± Vol. 87 (2004)

Si); 18.2 (Me C); 18.4 (C(18)); 19.4 (C(19)); 21.1 (MeCO); 25.9

¹³C-NMR (CDCl

, 500 MHz): À4.6 (Me

3

2

3

(

Me C); 62.4 (C(21)); 72.5 (C(3)); 110.7 (C(20)); 120.9 (C(6)); 141.7 (C(5)); 158.1 (C(17)); 171.1 (MeCO). EI-

3

.

t

MS (70 eV): 472 (1, M ), 457 (2.8, [M À Me] ), 429 (7, [M À Me À CO] ), 415 (31, [M À Bu] ), 413 (23, [M À

t

Me À COO] ), 399 (22, [M À MeCOOCH

2

] ), 75 (100, Me

2

SiOH ). HR-MS: 415.2658 ([M À Bu] ,

C

4

25

H

39

O

3

Si; calc. 415.2658). ESI-TOF-MS (pos.): 495.325 ([M Na] , C29

3

H48NaO Si ; calc. 495.3258),

73.343 ([M H] , ([C29

49 3

H O Si ; calc. 473.3438).

Irradiation Conditions, Isolation, and Characterization of the Products: General Procedure (G.P.). A soln.

of the steroid (190 mg) and Rose Bengal(1 mg) in pyridine (5 m l) , contained in a Pyrex tube kept in a water-

cooled bath, was saturated with oxygen. The soln., vigorously stirred by the gas bubbling, was irradiated with

light from a 220-V-400-Im-CE-4-300 W lamp placed at a distance of 15 cm. The oxygen bubbling was stopped

after 24 h. The irradiated soln. was analyzed by TLC (precoated silica gel 60 F254 plates, hexane/AcOEt and

AcOEt/EtOH mixtures): The irradiated soln. was then evaporated, the solid residue dried and dissolved in a

minimum volume of CH

2

Cl

2

, and this soln. adsorbed on silica gel, which was added to the top of a prep. column.

1

After flash CC (silica gel, 2% MeOH/CH

2

Cl

2

), the isolated products were characterized by m.p., H- and

1

3

C-NMR, EI-MS, HR-MS, and ESI-TOF-MS.

Irradiation of 1: (3b,5a)-3-{[(tert-Butyl)dimethylsilyl]oxy}-21-norpregn-16-en-20-al (1a). The G.P. yielded

1

a (22%). M.p. 162 ± 1648 (hexane/EtOH). UV (EtOH): 248. H-NMR (CDCl

3

, 200 MHz): 0.05 (s, Me

2

Si); 0.77

t

(

s, Me(18)); 0.82 (s, Me(19)); 0.88 (s, Bu); 3.56 (m, HÀC(3)); 5.73 (br. s, HÀC(16)); 9.70 (s, HÀC(20)).

1

3

C-NMR (CDCl

3

, 500 MHz): À4.5 (Me

2

Si); 12.3 (C(19)); 18.5 (C(18)); 18.2 (Me

3

C); 25.9 (Me

3

C); 72.1 (C(3));

.

1

52.9 (C(16)); 157.2 (C(17)); 190.0 (C(20)). EI-MS (70 eV): 416 (1.2, M ), 387 (3, [M À HCO] ), 361 (50,

t

[

(

M À Bu] ), 75 (100, Me

2

SiOH ). HR-MS: 359.2399 ([M À Bu] , (C22

35 2

H O Si ; calc. 359.2397). ESI-TOF-MS

pos.): 439.299 ([M Na] , C26

H

44NaO

2

Si ; calc. 439.2997), 417.317 ([M H] , C26

45 2

H O Si ; calc. 417.3177).

Irradiation of 2: (3b,5a,20S)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregn-16-en-20-yl Hydroperoxide (2a) and

(

3b,5a)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregn-16-en-20-one (2b). The G.P. yielded 2a (64%) and 2b (<0.5%).

1

Data of 2a: M.p. 182 ± 1838 (dec.; EtOH). H-NMR (CDCl

3

, 200 MHz): 0.05 (s, Me

2

Si); 0.85 (s, Me(18));

t

0

.88 (s, Me(19)); 0.88 (s, Bu); 1.35 (d, J 6.5, Me(21)); 3.54 (m, HÀC(3)); 4.59 (q, J 6.5, HÀC(20)); 5.73

1

3

(

4

br. s, HÀC(16)). C-NMR (CDCl

3

, 500 MHz): À4.5 (Me

2

3

Si); 12.4 (C(19)); 16.8 (C(18)); 18.2 (Me C); 25.9

.

Me

3

C); 18.9 (C(21)); 72.2 (C(3)); 79.1 (C(20)); 126.4 (C(16)); 154.9 (C(17)). EI-MS (70 eV): 448 (0, M ), 433

t

0.6, [M À Me] ), 431 (5, [M À OH] ), 391 (42, [M À Bu] ), 75 (100, Me

2

SiOH ). FAB-MS (propylene glycol):

t

49 ([M H] ), 432. HR-MS: 391.2659 ([M À Bu] , C23

39 3

H O Si ; calc. 391.2658). ESI-TOF-MS (pos.): 471.325

(

[M Na] , C27

H

48NaO

3

Si ; calc. 471.3258), 449.343 ([M H] , C27

49 3

H O Si ; calc. 449.3438).

1

t

Data of 2b: H-NMR (CDCl

3

, 200 MHz): 0.05 (s, Me Si); 0.78 (s, Me(18)); 0.82 (s, Me(19)); 0.88 (s, Bu);

2

.

2

4

(

.22 (s, Me(21)); 2.30 (m, 2 HÀC(15)); 3.54 (m, HÀC(3)); 5.93 (br. s, HÀC(16)). EI-MS (70 eV): 430 (2, M ),

t

15 (6, [M À Me] ), 387 (0.5, [M À MeCO] ), 373 (23, [M À Bu] ), 75 (100, Me

2

SiOH ). HR-MS: 373.2555

t

[M À Bu] , C23

H

37

O

2

Si ; calc. 373.2553). ESI-TOF-MS (pos.): 453.315 ([M Na] , C27

2

H46NaO Si ; calc.

4

53.3153), 431.333 ([M H] ; C27

47 2

H O Si ; calc. 431.3333).

Data of 2b: The compound has been previously described [16].

Irradiation of 3. The G.P. yielded only nonconverted 3.

Irradiation of 4: (3b,5a,17a,20S)- and (3b,5a,17b,20S)-3-{[(tert-Butyl)dimethylsilyl]oxy}-20-hydroxypreg-

1

nan-16-one (4a). The G.P. yielded 4a (26%). M.p. 203 ± 2058 (MeOH). H-NMR (CDCl

3

, 500 MHz): 0.05

t

(

2

s, Me

2

Si); 0.84 (s, Me(19)); 0.89 (s, Bu); 1.04 (s, Me(18)); 1.39 (d, J 2.3, Me(21)); 2.31 (d, J 2.7), 2.34 (d, J

1

3

.6) (HÀC(17)); 2.50 (m, 2 HÀC(15)); 3.45 (q, J 2.3, HÀC(20)); 3.56 (m, HÀC(3)). C-NMR (CDCl

3

,

5

(

00 MHz): À4.5 (Me

2

Si); 12.3 (C(19)); 14.3 (C(18)); 15.1 (C(21)); 18.2 (Me

3

C); 25.9 (Me

3

C); 69.6 (C(20)); 71.9

.

C(3)); 215.4 (C(16)). EI-MS (70 eV): 448 (4, M ), 433 (1.5, [M À Me] ), 431 (3, [M À OH] ), 420 (2, [M À

t

CO] ), 391 (53, [M À Bu] ), 75 (100, Me

2

SiOH ). HR-MS: 391.2659 ([M À Bu] , C23

H

39

O

3

Si ; calc.

3

C

91.2658). ESI-TOF-MS (pos.): 471.326 ([M Na] , C27

H

48NaO

3

Si ; calc. 471.3258), 449.344 ([M H] ,

27 49 3

H O Si ; calc. 449.3438).

The (3b,5a,17b,20S)-3-(acetyloxy)-20-hydroxypregnan-16-one analogue has been described [17].

Irradiation of 5: 5a (ꢀ4a) and (3b,5a,16b,17a)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregn-20-ene-16,17-diol

1

6-Acetate (5b). The G.P. gave 5a (38%) and 5b (40%).

Data of 5a (ꢀ4a): identicalto those of 4a (see above).

Data of 5b: M.p. 193 ± 1958 (dec.; hexane/EtOH). ¹H-NMR (CDCl

3

2

, 200 MHz): 0.05 (s, Me Si); 0.88

t

(

5

s, Me(19)); 0.89 (s, Bu); 0.91 (s, Me(18)); 2.07 (s, MeCO)); 3.53 (m, HÀC(3)); 5.13 (d, J 8, 1 HÀC(21)); 5.53

1

3

d, J 19, 1 HÀC(21)); 6.04 (t, J 7.8, HÀC(16)); 5.52 (dd, J 8 and 19, HÀC(20)). C-NMR (CDCl

00 MHz): À4.5 (Me Si); 12.3 (C(19)); 14.2 (C(18)); 18.3 (Me C); 21.2 (MeCO); 26.0 (Me C); 72.1 (C(3)); 76.7

C(17)); 95.2 (C(16)); 115.8 (C(21)); 133.2 (C(20)); 169.8 (MeCO). EI-MS (70 eV): 490 (7, M ), 475 (2,

3

,

2

3

.

(

Helvetica Chimica Acta ± Vol. 87 (2004)

3001

t

[

M À Me] ), 447 (2, [M À MeCO] ), 432 (37, [M À Bu] ), 75 (100, Me

2

SiOH ). HR-MS: 433.2765 ([M À

t

Bu] , C25

4

H

41

O

4

Si ; calc. 433.2763). ESI-TOF-MS (pos.): 513.346 ([M Na] , C29

4

H50NaO Si ; calc. 513.3465),

91.354 ([M H] , C29

51 4

H O Si ; calc. 491.3543).

Irradiation of 6: (3b,7a,16a)-3-{[(tert-Butyl)dimethylsilyl]oxy}-7-hydroperoxy-21-norpregna-5,17(20)-

1

dien-16-ol (6a): The G.P. yielded 6a (36%). M.p. 170 ± 1728 (dec.; EtOH). H-NMR (CDCl

3

, 200 MHz): 0.06

t

(

s, Me

2

Si); 0.79 (s, Me(18)); 0.89 (s, Bu); 1.00 (s, Me(19)); 3.57 (m, HÀC(3)); 4.20 (br. t, J 4.8, HÀC(7)); 4.68

1

3

dd, HÀC(16)); 4.89 (d, J 1.5, 1 HÀC(20)); 5.08 (d, J 1.5, 1 HÀC(20)); 5.67 (d, J 4.8, HÀC(6)). C-NMR

CDCl Si); 19.4 (C(19)); 18.2 (C(18)); 18.2 (Me C); 25.9 (Me C); 72.5 (C(3)); 72.7

, 500 MHz): À4.6 (Me

3

2

3

.

C(16)); 77.9 (C(7)); 103.8 (C(20)); 120.3 (C(6)); 147.6 (C(5)); 165.9 (C(17)). EI-MS (70 eV): 448 (0, M ), 431

t

(

2, [M À OH] ), 430 (10, [M À H

2

O] ), 391 (25, [M À Bu] ), 75 (100, Me

2

SiOH ). HR-MS: 391.2297 ([M À

t

Bu] , C22

H

35

O

4

Si ; calc. 391.2295). ESI-TOF-MS (pos.): 471.289 ([M Na] , C26

4

H44NaO Si ; calc. 471.2895),

4

49.308 ([M H] , C26

45 4

H O Si ; calc. 449.3075).

Irradiation of 7: (3b,7a,16a,17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}-7-hydroperoxypregna-5,17(20)-dien-

1

6-ol (7a). The G.P. yielded 7a (26%). M.p. 168 ± 1718 (dec.; MeOH). H-NMR (CDCl

3

, 200 MHz): 0.06

t

(

4

s, Me

2

Si); 0.87 (s, Me(18)); 0.90 (s, Me(19)); 0.89 (s, Bu); 1.75 (d, J 5.9, Me(21)); 3.50 (m, HÀC(3)); 4.15

1

3

br. t, HÀC(7)); 4.45 (br. d, HÀC(16)); 5.58 (br. d, J 8, HÀC(6)); 5.76 (q, J 5.9, HÀC(20)). C-NMR

CDCl

3

, 500 MHz): À4.6 (Me

2

Si); 13.2 (C(18)); 17.3 (C(21)); 18.2 (Me C); 19.4 (C(19)); 25.9 (Me C); 72.5

3

C(3)); 74.4 (C(16)); 78.0 (C(7)); 119.6 (C(20)); 120.9 (C(6)); 141.5 (C(5)); 155.4 (C(17)). EI-MS (70 eV): 462

.

t

0, M ), 445 (8, [M À OH] ), 405 (28, [M À Bu] ), 387 (7, [M À 57 À 18] ), 75 (100, Me

2

SiOH ). HR-MS:

t

05.2453 ([M À Bu] , C23

H

37

O

4

Si ; calc. 405.2451). ESI-TOF-MS (pos.): 485.305 ([M Na] , C27

4

H46NaO Si ;

calc. 485.3051), 463.323 ([M H] , C27

47 4

H O Si ; calc. 463.3231).

Irradiation of 8: (3b,20S)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregna-5,16-dien-20-yl Hydroperoxide (8a)

and (3b,20S)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregna-5,16-dien-20-ol (8b). The G.P. yielded 8a (36%) and 8b

(

9%).

Data of 8a: M.p. 192 ± 1948 (dec.; EtOH). H-NMR (CDCl

.03 (s, Me(18)); 1.04 (s, Me(19)); 1.35 (d, J 6.6, Me(21)); 3.50 (m, HÀC(3)); 4.50 (q, J 6.6, HÀC(20)); 5.31

1

t

3

, 200 MHz): 0.06 (s, Me

2

Si); 0.89 (s, Bu);

1

3

(br. d, HÀC(6)); 5.75 (br. d, J 1.5, HÀC(16)). C-NMR (CDCl

(Me C); 19.4 (C(19)); 19.8 (C(21)); 26.0 (Me C); 72.6 (C(3)); 83.5 (C(20)); 121.0 (C(6)); 128.9 (C(16)); 141.6

(C(5)); 148.2 (C(17)). EI-MS (70 eV): 446 (0, M ), 429 (5, [M À OH] ), 414 (0.5, [M À Me À HO] ), 389 (30,

[

3

, 500 MHz): À4.6 (Me

2

Si); 13.1 (C(18)); 18.2

3

.

t

M À Bu] ), 75 (100, MeO

2

SiOH ). HR-MS: 389.2504 ([M À Bu] , C23

37 3

H O Si ; calc. 389.2502). ESI-TOF-

MS (pos.): 469.310 ([M Na] , C27

47.3282).

Data of 8b: H-NMR (CDCl

H

46NaO

3

Si ; calc. 469.3102), 447.328 ([M H] , C27

47 3

H O Si ; calc.

4

1

t

3

, 200 MHz): 0.06 (s, Me

2

Si); 0.89 (s, Bu); 0.90 (s, Me(18)); 1.05 (s, Me(19));

1

.47 (d, J 8.4, Me(21)); 3.50 (m, HÀC(3)); 4.60 (q, J 8.4, HÀC(20)); 5.34 (br. d, J 5.0, HÀC(6)); 5.75

13

(

2

br. d, J 1.5, HÀC(16)). C-NMR (CDCl

3

, 500 MHz): À4.6 (Me

2 3

Si); 13.1 (C(18)); 18.2 (Me C); 19.4 (C(19));

3.0 (C(21)); 26.0 (Me C); 72.6 (C(3)); 73.5 (C(20)); 120.0 (C(6)); 127.5 (C(16)); 143.6 (C(5)); 150.2 (C(17)).

3

.

t

EI-MS (70 eV): 430 (1.1, M ), 413 (8, [M À OH] ), 398 (0.9, [M À Me À HO] ), 373 (25, [M À Bu] ), 75 (100,

t

Me

Na] , C27

2

SiOH ). HR-MS: 373.2555 ([M À Bu] , C23

H

37

O

2

Si ; calc. 373.2553). ESI-TOF-MS (pos.): 453.315 ([M

H

46NaO

2

Si ; calc. 453.3153), 431.333 ([M H] , C27

47 2

H O Si ; calc. 431.3333)).

The (3b,20R/S)-pregna-5,16-diene-3,20-diolhas been described [18].

Irradiation of 9: (3b,20R)-3-{[(tert-Butyl)dimethylsilyl]oxy}-20-hydroperoxypregna-5,16-dien-21-ol (9a),

3b)-3-{[(tert-Butyl)dimethylsilyl]oxy}-21-hydroxypregna-5,16-dien-20-one (9b), (3b,17a)-3-{[(tert-Butyl)di-

(

methylsilyl]oxy}-17-hydroxypregn-5-en-21-al (9c), and (3b,17E)-3-{[(tert-Butyl)dimethylsilyl]oxy}pregna-

5

,17(20)-dien-21-al (9d). The G.P. yielded 9a (6%), 9b (27%), 9c (10%), and 9d (17%).

1

Data of 9a: M.p. 179 ± 1808 (dec.; EtOH). H-NMR (CDCl

3

, 200 MHz): 0.06 (s, Me

2

Si); 0.86 (s, Me(18));

t

0

.89 (s, Bu); 1.04 (s, Me(19)); 3.48 (m, HÀC(3)); 3.68 (m, 2 HÀC(21)); 5.33 (br. d, J 5.1, HÀC(6)); 5.72

1

3

(

[

t, HÀC(20)); 6.74 (br. d, J 1.8, HÀC(16)). C-NMR (CDCl

C(19)); 18.2 (Me C); 25.9 (Me C); 66.1 (C(21)); 69.84 (C(20)); 72.6 (C(3)); 120.8 (C(6)); 126.4 (C(16)); 141.9

C(5)); 154.4 (C(17)). EI-MS (70 eV): 462 (0, M ), 445 (1, [M À OH] ), 431 (0.9, [M À MeO] ), 415 (0.9,

3

, 500 MHz): À4.6 (Me

2

Si); 16.6 (C(18)); 19.4

3

.

t

M À MeO À O] ), 405 (33, [M À Bu] ), 387 (10, [M À MeO À O À CO] ), 75 (100, Me

2

SiOH ). HR-MS:

t

4

05.2453 ([M À Bu] , C23

H

37

O

4

Si ; calc. 405.2451). ESI-TOF-MS (pos.): 485.305 ([M Na] , C27

4

H46NaO Si ;

calc. 485.3051), 463.323 ([M H] , C27

47 4

H O Si ; calc. 463.3231).

1

Data of 9b: M.p. 158 ± 1618 (dec.; EtOH). UV (EtOH): 246. H-NMR (CDCl

3

, 200 MHz): 0.06 (s, Me

2

Si);

t

0

.96 (s, Me(18)); 0.89 (s, Bu); 1.04 (s, Me(19)); 3.29 (t, J 4.7, disappeared after adding D

2

O, OH); 3.50

(

m, HÀC(3)); 4.39 (dd, J 4.7, 17.8, ! d(J 17) after adding D

2

O, 1 HÀC(21)); 4.51 (dd, J 4.7, 17.8,

!

d, J 17 after adding D

2

O, 1 HÀC(21)); 5.33 (br. d, J 5.1, HÀC(6)); 6.74 (br. d, J 1.8, HÀC(16)).

3

002

Helvetica Chimica Acta ± Vol. 87 (2004)

¹³C-NMR (CDCl

, 500 MHz): À4.6 (Me

3

2 3 3

Si); 15.9 (C(18)); 19.4 (C(19)); 18.2 (Me C); 25.9 (Me C); 65.1

(

C(21)); 72.6 (C(3)); 120.4 (C(6)); 142.2 (C(5)); 144.6 (C(16)); 151.6 (C(17)); 196.6 (C(20)). EI-MS (70 eV):

.

t

4

7

4

44 (0, M ), 427 (1, [M À OH] ), 413 (0.9, [M À MeO] ), 387 (26, [M À Bu] ), 385 (20, [M À MeO À CO] ),

t

5 (100, Me

2

SiOH ). HR-MS: for 387.2348 ([M À Bu] , C23

35 3

H O Si ; calc. 387.2346). ESI-TOF-MS (pos.):

67.294 ([M Na] , C27

H

44NaO

3

Si ; calc. 467.2946), 445.312 ([M H]

27 45 3

C H O Si ; calc. 445.3126).

The (3b)-3,21-bis(acetyloxy)pregna-5,16-dien-20-one analogue has been described [19].

1

Data of 9c: M.p. 189 ± 1928 (dec.; EtOH). H-NMR (CDCl

3

, 200 MHz): 0.06 (s, Me

2

Si); 0.83 (s, Me(18));

t

0

9

(

3

.89 (s, Bu); 1.00 (s, Me(19)); 3.22 (d, J 5.9, 2 HÀC(20)); 3.50 (m, HÀC(3)); 5.32 (br. d, J 5.4, HÀC(6));

1

3

.23 (br. d, J 5.9, HÀC(21)). C-NMR (CDCl

3

, 500 MHz): À4.6 (Me

2

Si); 15.9 (C(18)); 19.5 (C(19)); 18.2

Me C); 25.9 (Me C); 72.5 (C(3)); 75.8 (C(17)); 120.4 (C(6)); 141.7 (C(5)); 199.7 (C(21)). EI-MS (70 eV): 446

3

.

t

0.8, M ), 426 (3, [M À H

2

O] ), 417 (9, [M À HCO] ), 398 (32, [M À Bu] ), 75 (100, Me

2

SiOH ). HR-MS:

t

98.2504 ([M À Bu] , C23

H

37

O

3

Si ; calc. 398.2502. ESI-TOF-MS (pos.): 469.310 ([M Na] , C27

H

46NaO

3

Si ;

calc. 469.3102), 447.320 ([M H] , C27

47 3

H O Si ; calc. 447.3204).

Data of 9d: M.p. 162 ± 1648 (hexane/EtOH). UV (EtOH): 248. ¹H-NMR (CDCl

, 200 MHz): 0.06

3

t

(

s, Me

2

Si); 0.94 (s, Me(18)); 0.89 (s, Bu); 1.06 (s, Me(19)); 3.48 (m, HÀC(3)); 5.32 (br. d, J 5.4, HÀC(6));

1

3

6

.78 (m, HÀC(20)); 9.71 (s, HÀC(21)). C-NMR (CDCl

3

, 500 MHz): À4.6 (Me

2

Si); 15.8 (C(18)); 19.4 (C(19));

1

8.2 (Me3C); 25.9 (Me C); 72.6 (C(3)); 120.4 (C(6)); 142.2 (C(5)); 152.9 (C(20)); 157.1 (C(17)); 189.9 (C(21)).

3

.

t

EI-MS (70 eV): 428 (1, M ), 399 (2, [M À HCO] ), 374 (5, [M À HCOCCH] ), 371 (29, [M À Bu] ), 75 (100,

t

Me

Na] , C27

2

SiOH ). HR-MS: 371.3101 ([M À Bu] , C23

H

35

O

2

Si ; calc. 371.3099). ESI-TOF-MS (pos.): 451.299 ([M

H

44NaO

2

Si ; calc. 451.2997), 429.317 ([M H] , C27

45 2

H O Si ; calc. 429.3177).

The analogue of 9d, (3b,17E)-3-(acetyloxy)pregna-5,17(20)-dien-21-al, has been described [20].

Irradiation of 10: (3b)-3-{[(tert-Butyl)dimethylsilyl]oxy}-21-hydroxypregna-5,16-dien-20-one Acetate (10a)

and 10b (ꢀ9d). The G.P. yielded 10a (8%) and 10b (7%).

Data of 10a: M.p. 203 ± 2058 (hexane/EtOH). UV (EtOH): 246. ¹H-NMR (CDCl

, 200 MHz): 0.06

3

t

(

s, Me

2

Si); 0.89 (s, Bu); 0.94 (s, Me(18)); 1.03 (s, Me(19)); 2.17 (s, MeCO); 3.48 (m, HÀC(3)); 4.88 (d, J 16,

1

3

1

HÀC(21)); 5.01 (d, J 16, 1 HÀC(21)); 5.32 (d, J 5.2, HÀC(6)); 6.74 (br. d, J 1.4, HÀC(16)). C-NMR

, 500 MHz): À4.6 (Me Si); 15.8 (C(18)); 18.0 (C(19)); 18.2 (Me C); 19.4 (MeCO); 25.9 (Me C); 65.7

C(21)); 72.6 (C(3)); 120.4 (C(6)); 142.2 (C(5)); 143.9 (C(16)); 152.1 (C(17)); 170.4 (MeCO); 190.5 (C(20)).

(

CDCl

3

2

3

.

t

EI-MS (70 eV): 486 (2, M ), 471 (1, [M À Me] ), 443 (5, [M À MeCO] ), 429 (33, [M À Bu] ), 427 (5, [M À

t

MeCOO] ), 414 (15, [M À MeCOOCH] ), 75 (100, Me

2

SiOH ). HR-MS: 429.2427 ([M À Bu] , C25

H

37

O

4

Si ;

calc. 429.2425). ESI-TOF-MS (pos.): 509.305 ([M Na] , C29

H

46NaO

4

Si ; calc. 509.3051), 487.323 ([M H] ,

C

29

H

47

O

4

Si ; calc. 487.3231).

The (3b)-3,21-bis(acetyloxy)pregna-5,16-dien-20-one analogue has been described [19].

Data of 10b (ꢀ9d): identicalwith those of 9d (see above).

Computational Details. Geometries were fully optimized without imposing any symmetry constraints by ab

initio and semiempiricalmethods. For ab initio DFT calculations, we used the hybrid gradient-corrected

exchange functional combined with the gradient-corrected correlation functional, commonly known as B3LYP,

which has been shown to be quite reliable for geometries. For geometry optimization, the standardized 3-21G

basis set was used. We denote our B3LYP calculations by B3LYP/3-21G. For single-point and frequency

calculations, the standarized 3-21G basis set was used at the B3LYP and HF (Hartree ± Fock) theory level. The 6-

3

11 G(d,p) basis set was used at the B3LYP level for single-point calculations. All the ab initio calculations

were carried out with the Gaussian 98W program.

The ground-state geometry and heat of formation of steroids were also calculated by using the

semiempiricalparametrized AM1 and PM3 methods as imp el mented in the version of the HyperChem Suite

(

professional) 7.05 program. PM3 has been proven effective in studies on molecules containing heteroatoms,

compared with other methods such as MINDO/3 or MNDO.

The molecular orbitals were calculated by the ZINDO/S (Cl, 6 :6) method transferring the optimized

geometries obtained by ab initio calculations and by semiempirical methods. Similarly, HOMO and LUMO

descriptions were obtained from geometries optimized at the AM1, PM3, HF/3-21G, and B3LYP/3-21G levels.

We also calculated DH at the HF/3-21G and B3LYP/21G levels, including the effect of the solvent. These

f

calculations were performed by using, as self-consistent reaction field method (SCRF), the isodensity Tomasiꢂs

polarized continuum model (IPCM) (IPCM/HF/3-21G//B3LYP/21G and IPCM/B3LYP/3-21G//B3LYP/3-

2

1G). For solvent correction, pyridine with relative permittivity (e 12.3) was used.

Helvetica Chimica Acta ± Vol. 87 (2004)

3003

REFERENCES

[1] W. Albro, P. Bilski, J. T. Corbett, J. L. Schroeder, C. F. Chignell, Photochem. Photobiol. 1997, 66, 316.

[2] G. O. Schenck, O. A. Neumuller, W. Eisfeld, Justus Liebigs Ann. Chem. 1958, 618, 202.

[3] a) A. Nickon, J. F. Bagli, J. Am. Chem. Soc. 1961, 83, 1498; b) A. Nickon, W. L. Mendelson, J. Am. Chem.

Soc. 1963, 85. 1849.

[

4] A. L. J. Beckwith, A. G. Davies, I. G. E. Davison, A. Maccoll, M. H. Mruzek, J. Chem. Soc., Perkin Trans. 2

1

989, 815.

5] M. A. Ponce, J. A. Ramirez, L. R. Galagovsky, E. G. Gros, R. Erra-Balsells, J. Chem. Soc., Perkin Trans. 2

000, 2351.

6] M. A. Ponce, J. A. Ramirez, L. R. Galagovsky, E. G. Gros, R. Erra-Balsells, Photochem. Photobiol. Sci.

002, 1, 745.

2

7] a) C. S. Foote, E. L. Clennan, in ꢁActive Oxygen in Chemistryꢂ, Eds. C. S. Foote, J. S. Valentine, A.

Greenberg, and J. F. Liebman, Blackie Academic, New York,1995, p. 105; b) G. O. Schenck, H. Eggert, W.

Denk, Liebigs Ann. Chem. 1953, 584, 177; c) A. G. Griesbeck, T. T. EL-Idreesy, W. Adam, O. Krebs, ꢁEne-

Reaction with Singlet Oxygenꢂ, in ꢁCRC Handbook of Photochemistry and Photobiologyꢂ, Eds. W. M.

Horspool and P.-S. Sing, CRC, Boca Raton, Florida, 2004, Chap. 8, p. 8; d) M. Stratakis, M. Orfanopoulos,

Tetrahedron 2000, 56, 1595; e) M. Orfanopoulos, M. B. Grdina, L. M. Stephenson, J. Am. Chem. Soc. 1979,

101, 275; f) A. G. Leach, K. N. Houk, Chem. Commun. 2002, 1243; g) W. Adam, M. Prein, Angew. Chem.,

Int. Ed. 1996, 35, 477; h) W. Adam, B. Nestler, J. Am. Chem. Soc. 1993, 115, 5041; i) W. Adam, H.-G.

Brünker, A. S. Kumar, E.-M. Peters, K. Peters, U. Schneider, H. G. von Schnering, J. Am. Chem. Soc. 1996,

118, 1899; j) A. G. Griesbeck, W. Adam, A. Bastoschek, T. El-Idreesy, Photochem. Photobiol. Sci. 2003, 2,

877; k) R. W. Hoffman, Chem. Rev. 1989, 89, 1841.

[

8] A. Bruttomesso, D. Doller, E. G. Gros, Synth. Commun. 1998, 28, 4043.

9] A. Bruttomesso, D. Doller, E. G. Gros, Bioorg. Med. Chem. 1999, 7, 943; A. Bruttomesso, E. G. Gros,

Molecules 2000, 5, 564.

[

10] K. Al-Fouti, J. R. Hanson, J. Chem. Res. Sympo. 2003, 4, 232.

11] E. Mincione, C. Iavarone, Gazz. Chim. Ital. 1971, 101, 956.

12] H. Hosoda, D. K. Fukushima, J. Fishman, J. Org. Chem. 1973, 38, 4209.

13] S. V. Kessar, A. L. Rampal, Tetrahedron 1968, 24, 887.

14] L. D. Timberlake, H. Morrison, J. Am. Chem. Soc. 1999, 121, 3618.

15] G. Ortar, E. Morera, A. Romeo, J. Org. Chem. 1978, 43, 2927.

16] H. Duddeck, D. Rosenbaum, M. Hani, A. Elgamal, M. B. E. Fayez, Magn. Reson. Chem. 1986, 24, 999; R.

Mata, I. Rivero-Cruz, B. Rivero Cruz, R. Bye, B. N. Timmeermann, J. Nat. Prod. 2002, 65, 1030.

17] S. Noguchi, M. Imanishi, K. Morita, Chem. Pharm. Bull. 1964, 12, 1184.

18] A. G. Bajaj, S. Dev, Tetrahedron 1982, 19, 2949.

[

19] D. H. R. Barton, W. B. Motherwell, S. Z. A. Zard, J. Chem. Soc., Chem. Commun. 1982, 551.

20] G. Ferrara, A. Ius, C. Parini, G. Sportoletti, G. Vecchio, Tetrahedron 1972, 28, 2461.

Received July 12, 2004

Article Doi

Photooxygenation of pregnanes

DOI: 10.1002/hlca.200490270

Source and publish data:

Authors:

Article abstract of DOI:10.1002/hlca.200490270

Full text of DOI:10.1002/hlca.200490270

Products guided by the article

R&D Labs maybe for 42151-23-5

Relevant to this article

Hot Product