Synthesis of (5-azido-2-nitrobenzoyl)amido, (4-azido-2- nitrophenyl)amino, and (5-azido-2-nitro-3,4,6-trifluorophenyl)amino derivatives of 17α-methylamino-, 17α-ethylamino-, and 17α-propylamino-5α- dihydrotestosterone as reagents of different linker lengths for the photoaffinity labeling of sex hormone binding globulins and androgen receptors

Article abstract of DOI:10.1016/S0039-128X(00)00106-9

The photoactivable aryl azide reagents, N-(5-azido-2- nitrobenzoyl)oxysuccinimide, 4-azido-1-fluoro-2-nitrobenzene, and 4-azido-1- nitro-2,4,5,6-tetrafluorobenzene have been condensed at the extremity of three 17α-aminomethyl, 17α-aminoethyl, and 17α-aminopropyl side-chains introduced on (17S)-spiro-(3,3-dimethoxy)-5α-androstan-17β,2'-oxirane either directly, by ammonolysis, in the first case, or by conversion to nitrile intermediates with cyano or cyanomethyl anions and subsequent reduction to amines with lithium aluminum hydride, in the two other cases. The 3,3-dimethoxy group of these photoreagents was cleaved by acidolysis to a 3-ketone, which was reduced with sodium borohydride to a 3β-alcohol. All of these compounds were characterized by ¹H- and ¹³C-NMR as well as by ¹H, ¹³C heteronuclear 2D NMR, which helped to resolve ambiguous assignments. Significant differences of substituent-induced effects on ¹³C NMR signals were observed according to the 17α-side-chain length, the structure of the terminal aryl azide groups, and the solvent, showing a different behavior of N-5-azido-2-nitrobenzoyl derivatives as compared with 4-azido-2- nitrophenylamino and 5-azido-2-nitro-3,4,6-trifluorophenylamino derivatives. The N-5-azido-2-nitrobenzoyl conjugates of the three 17α-aminomethyl, aminoethyl, and aminopropyl derivatives of 5α-dihydrotestosterone were tested as ligands for purified human sex hormone-binding globulin and for the cytosolic androgen receptor of rat ventral prostate by competition experiments with tritiated 5α-dihydrotestosterone. The increasing lengths of the aminomethyl, aminoethyl, and aminopropyl spacer arms of N-5-azido-2- nitrobenzoyl conjugates were found to correspond to decreasing relative binding affinities for sex hormone-binding globulin (0.76, 0.47, and 0.10, respectively, versus 1.00 for 5α-dihydrotestosterone) while only the longer aminoethyl and aminopropyl conjugates interacted significantly with the androgen receptors (0.05 and 0.10, respectively). (C) 2000 Elsevier Science Inc.

Full text of DOI:10.1016/S0039-128X(00)00106-9

Steroids 65 (2000) 459–481
Synthesis of (5-azido-2-nitrobenzoyl)amido, (4-azido-2-
nitrophenyl)amino, and (5-azido-2-nitro-3,4,6-trifluorophenyl)amino
derivatives of 17␣-methylamino-, 17␣-ethylamino-, and 17␣-
propylamino-5␣-dihydrotestosterone as reagents of different linker
lengths for the photoaffinity labeling of sex hormone binding globulins
and androgen receptors
Elisabeth Mappusa, Christophe Chambona, Bernard Fenetb, Marc Rolland de Ravela,
Catherine Grenota, Claude Y. Cuillerona,*
aINSERM U 329, Pathologie Hormonale et Mole´culaire, Hoˆpital Debrousse, Lyon, France
bCentre de RMN, Universite´ Claude Bernard Lyon 1-ESCPE, Villeurbanne, France
Received 2 February 2000; received in revised form 3 April 2000; accepted 6 April 2000
Abstract
The photoactivable aryl azide reagents, N-(5-azido-2-nitrobenzoyl)oxysuccinimide, 4-azido-1-fluoro-2-nitrobenzene, and 4-azido-1-nitro-
2,4,5,6-tetrafluorobenzene have been condensed at the extremity of three 17␣-aminomethyl, 17␣-aminoethyl, and 17␣-aminopropyl side-chains
introduced on (17S)-spiro-(3,3-dimethoxy)-5␣-androstan-17␤,2Ј-oxirane either directly, by ammonolysis, in the first case, or by conversion to
nitrile intermediates with cyano or cyanomethyl anions and subsequent reduction to amines with lithium aluminum hydride, in the two other cases.
The 3,3-dimethoxy group of these photoreagents was cleaved by acidolysis to a 3-ketone, which was reduced with sodium borohydride to a
1
1
3␤-alcohol. All of these compounds were characterized by H- and 13C-NMR as well as by H, 13C heteronuclear 2D NMR, which helped to
resolve ambiguous assignments. Significant differences of substituent-induced effects on 13C NMR signals were observed according to the
17␣-side-chain length, the structure of the terminal aryl azide groups, and the solvent, showing a different behavior of N-5-azido-2-nitrobenzoyl
derivatives as compared with 4-azido-2-nitrophenylamino and 5-azido-2-nitro-3,4,6-trifluorophenylamino derivatives. The N-5-azido-2-nitroben-
zoyl conjugates of the three 17␣-aminomethyl, aminoethyl, and aminopropyl derivatives of 5␣-dihydrotestosterone were tested as ligands for
purified human sex hormone-binding globulin and for the cytosolic androgen receptor of rat ventral prostate by competition experiments with
tritiated 5␣-dihydrotestosterone. The increasing lengths of the aminomethyl, aminoethyl, and aminopropyl spacer arms of N-5-azido-2-nitroben-
zoyl conjugates were found to correspond to decreasing relative binding affinities for sex hormone-binding globulin (0.76, 0.47, and 0.10,
respectively, versus 1.00 for 5␣-dihydrotestosterone) while only the longer aminoethyl and aminopropyl conjugates interacted significantly with
the androgen receptors (0.05 and 0.10, respectively). © 2000 Elsevier Science Inc. All rights reserved.
Keywords: 17␣-[(5-azido-2-nitrobenzoyl)amidomethyl/ethyl/propyl]-5␣-androstane-3␤,17␤-diol; 17␣-[(4-azido-2-nitrophenyl/5-azido-2-nitro-2,4,6-trifluorophe-
nyl)/aminoalkyl]-5␣-androstane-3␤,17␤-diol; Photoaffinity labeling reagents; Human sex hormone-binding globulin; Rat prostate cytosolic androgen receptor; 1H/
13C NMR
1. Introduction
osterone (DHT) as ligands for affinity chromatography of an-
drogen-binding proteins such as sex hormone-binding
globulins (SHBGs) and androgen receptors. Previous reports
(reviewed in ref. 1) have shown that immobilized 17␣-(2Ј-
carboxyethynyl)-testosterone or 17␣-(6Ј-hexanoic)-DHT both
interact with SHBG, while only 17␣-(2Ј,3Ј-epoxypropyl)-
DHT, but not 17␣-(carboxymethyl)-DHT, can be employed
for the purification of the androgen receptor from rat ventral
Several studies have demonstrated the usefulness of 17␣-
substituted derivatives of testosterone and 5␣-dihydrotest-
* Corresponding author. Tel.: ϩ33 4 78 25 18 08; fax: ϩ33 4 78 25 61
68.
E-mail address: cuilleron@lyon151.inserm.fr (C.Y. Cuilleron).
0039-128X/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved.
PII: S0039-128X(00)00106-9

460
E. Mappus et al. / Steroids 65 (2000) 459–481
prostate cytosol, thus suggesting a differential recognition of
these two proteins according to the structure of the 17␣-sub-
stituent, especially at the level of the first three atoms of the
side-chain. Moreover, 17␣-alkyl substitution has been reported
as a means to increase the oral activity of androgens, possibly
as the result of modifications of binding characteristics with
both receptor and transport proteins [2]. Previous studies from
France). Chromatographic and spectrometric methods were
similar to those previously mentioned in a preceding article
[3], except for IR spectra, which were recorded on a Bruker
Vector 22 spectrometer. Liquid secondary ion mass spec-
trometry (LSIMS) experiments (positive mode) were per-
formed on a ZAB-2-SEQ mass spectrometer (Micromass,
Manchester, UK) equipped with a Csϩ gun operating at 40
keV. The accelerating voltage was 8 kV, and the instrumen-
tal resolution was 1000 or 5000 (10% valley) for low- or
high-resolution measurements, respectively. Calibrations
were performed with (CsI)nCsϩ cluster ions. Steroid sam-
ples were dissolved in 2 ␮l of methanol and added to 2 ␮l
of 1-thioglycerol matrix (except for 3,3-dimethoxy-17␣-
derivatives, which were added to a 3-nitrobenzyl alcohol
matrix) introduced on a stainless-steel probe tip. Spectra
were recorded in the mass range 100-1000. Data for exact
mass measurements were acquired using data system con-
trol in the multichannel analyser mode (MCA), and 10 scans
1
this laboratory have reported the synthesis and the H NMR
and 13C NMR characterization of 17␣-hexanoic derivatives of
DHT and testosterone and of their 17␣-hexynyl synthetic pre-
cursors [1], useful as ligands for affinity chromatography of
both androgen receptors and SHBGs as well as the synthesis
and characterization of the corresponding 17␣-hemiglutar-
amidomethyl derivatives [3], useful for affinity chromatogra-
phy of SHBGs only.
These affinity chromatography ligands may be readily con-
verted to aryl azide photoaffinity labeling reagents, potentially
able to interact with amino acids in the vicinity of the binding
site of the steroid skeleton, provided that the addition of an
aromatic photoactivable chromophore at the extremity of the
17␣-aminoalkyl side-chain does not alter too much the binding
properties or the degree of selectivity for androgen receptors
and SHBGs. However, although photoaffinity labeling re-
agents have been claimed to have a theoretical ability to react
with any interacting amino acid, numerous reports (reviewed
in ref. 4) have often identified similar amino acid targets for a
same structural class of photoreagents such as aryl azides.
Therefore, optimization of the spacer arm structure might also
favor an appropriate positioning of the photoactivable chro-
mophore toward a more reactive amino acid, thus augmenting
the probability of a successful labeling in a region vicinal to the
steroid binding site where a much lesser degree of immobili-
zation can be expected for the side-chain as compared with that
of the steroid skeleton in the steroid binding site.
1
were summed to give the final spectrum. H and 13C NMR
spectra were recorded on Bruker DRX 300 and DRX 500
spectrometers at 300.13 and 500.13 MHz, respectively, for
1
1H, and at 75.47 and 125.75 MHz, for 13C. H chemical
shifts were measured relative to tetramethylsilane. Samples
were prepared by dissolving 10–25 mg of steroid in 0.75 ml
of CDCl3 (99.8%) or C5D5N (99.5%) solvents purchased
1
from CEA (Saclay, France). The H chemical shifts are
estimated to be accurate to Ϯ0.01 ppm and coupling con-
stants to Ϯ0.5 Hz. The 13C chemical shifts are estimated to
be accurate to Ϯ0.05 ppm. DEPT experiments [5] were
systematically performed in order to differentiate quater-
nary, methine, methylene, and methyl carbon atoms. The
1H, 13C heteronuclear single quantum correlation (HSQC)
experiments, the HSQC-TOCSY (total correlation spectros-
1
copy) experiments, and the H, 13C heteronuclear multiple
This work was undertaken with the view of preparing pho-
toactivable derivatives of 17␣-(aminoalkyl)-17␤-hydroxy de-
rivatives in both DHT and 5␣-androstane-3␤,17␤-diol series as
potential reagents for photoaffinity labeling of androgen bind-
ing proteins such as SHBG and androgen receptors. Three
different 17␣-aminomethyl, aminoethyl, and aminopropyl
side-chains of increasing lengths were introduced on the ste-
roid skeleton and each one was linked to three 5-azido-2-
nitrobenzoyl, 4-azido-2-nitrophenyl, and 5-azido-2-nitro-3,4,6-
trifluorophenyl chromophores. Preliminary experiments were
undertaken with radioinert 5-azido-2-nitrobenzoyl derivatives
in order to estimate the effects of the side-chain length on the
binding affinities for SHBG and androgen receptors, using
competitive displacement of tritiated DHT.
bond correlation (HMBC) experiments were performed on
the Bruker DRX 500 spectrometer, using the standard gra-
dient pulse sequences from Bruker [6–8]. The relaxation
delay was D1 ϭ 1.5 s in all cases. The pulse sequences were
optimized using coupling constants of 145 Hz for HSQC
experiments, and 10 Hz for HMBC experiments, except for
3
the JCCNH coupling of NH with the aromatic C-1Љ carbon
atom of compound 11, which could be observed only at 6
Hz, but not at 10 or 4 Hz. The mixing time for HSQC-
TOCSY experiments was 15 ms. All 2D spectra were re-
corded at high resolution (td1 ϭ 1024) for carbon, in order
to improve separation of carbon signals. 13C chemical shift
increments were considered to be significant only if they
were observed in all three 3,3Јdimethoxy-, 3-oxo-, and
3␤-hydroxysteroid series and if the increment had a mag-
nitude equal to or above 0.10 ppm for at least two of these
three series.
2. Experimental
2.1. General methods
2.2. 3,3-(Dimethoxy)-5␣-androstan-17-one (1)
5␣-Androstane-3,17-dione and other chemicals were
A solution of 5␣-androstane-3,17-dione (10 g, 34.7
purchased from Sigma-Aldrich (St Quentin Fallavier,
mmol) in 250 ml of methanol was stirred for 1 h at 40°C in

E. Mappus et al. / Steroids 65 (2000) 459–481
461
the presence of p-toluenesulfonic acid monohydrate (100
mg, 0.53 mmol). The reaction mixture was neutralized with
MeONa (60 mg, 1.1 mmol), and the solvent was evaporated
under reduced pressure. Water (300 ml) was added, and the
aqueous layer was extracted with dichloromethane. The
combined organic layers were washed with water and evap-
orated under reduced pressure. The product was analyzed by
TLC on silica gel (Rf 0.45, petroleum ether-ethyl acetate 4:1
v/v) and recrystallized from methanol to give the pure
dimethoxy derivative (7.2 g, 21.5 mmol, 62%): m.p. 120–
121°C, recrystallized from methanol; (reported [9] 128–
129°C); [␣]D ϭ ϩ82.1° (CHCl3) (reported [9] ϩ100.2° in
CHCl3); ␯max (CCl4): 1740 (17 Ϫ C ϭ 0), 1110–1055
washed with 100 ml of petroleum ether/ethyl acetate 4:1 v/v
mixture, and the pure 17␣-aminomethyl derivative was then
eluted with a chloroform/methanol/NH4OH 100:20:5 v/v/v
mixture. The white solid product (0.75 g, 2.05 mmol, 72%)
was analyzed by TLC on silica gel (Rf 0.55, chloroform/
methanol/NH4OH 100:20:5 v/v/v): m.p. 177–179°C, recrys-
tallized from diethyl ether; [␣]D ϭ - 8° (CHCl3); ␯
max
(CCl4): 1105–1055 (3,3-dimethoxy), 1085 cmϪ1 (OCH2);
1H NMR (CDCl3) ␦ 0.80 (3H, s, 19-CH3), 0.88 (3H, s,
18-CH3), 2.51 and 2.92 (2H, dd: J ϭ 12.5 and 12.5 Hz,
1
CH2N), 3.13 and 3.19 (2 ϫ 3H, 2s, dimethoxy); H NMR
(C5D5N) ␦ 0.77 (3H, s, 19-CH3), 1.11 (3H, s, 18-CH3), 2.77
and 3.15 (free amine [3]) (2H, 2d: J ϭ 12.3 and 12.8 Hz,
CHaNH and CHbNH), 3.20 and 3.23 (2 ϫ 3H, 2s, dime-
thoxy); MS (LSIMSϩ) m/z (relative intensity %) 366 (MHϩ,
100), 334 (MHϩ-32, 16). High-resolution MS calculated for
C22H40O3N1 (MHϩ): 366.3008; found: 366.3002.
1
cmϪ1 (3,3-dimethoxy); H NMR (CDCl3) ␦ 0.82 (3H, s,
19-CH3), 0.86 (3H, s, 18-CH3), 3.14 and 3.19 (2 ϫ 3H, 2s,
dimethoxy); 1H NMR (C5D5N) ␦ 0.73 (3H, s, 19-CH3), 0.79
(3H, s, 18-CH3), 3.19 and 3.22 (2 ϫ 3H, 2s, dimethoxy).
2.3. (17S)-Spiro-3,3-(dimethoxy)-5␣-androstan-17␤,2Ј-
oxirane (2)
2.5. 17␣-Cyanomethyl-3,3-(dimethoxy)-5␣-androstan-17␤-
ol (4)
A solution of 17-oxosteroid 1 (6.9 g, 20.6 mmol) and of
trimethylsulfonium iodide (6.8 g, 33.3 mmol) in 160 ml of
DMF was stirred for 1 h at 22°C in the presence of potas-
sium t-butoxide (5.1 g, 45.4 mmol). The solvent was con-
centrated to 20 ml under reduced pressure (Ͻ0.01 mm Hg,
t Ͻ 37°C) and ice-cooled water (1 liter) was added. The
aqueous suspension was extracted with ethyl acetate. The
combined organic layers were washed with water and evap-
orated under reduced pressure. The product was analyzed by
TLC on silica gel (Rf 0.60, petroleum ether/ethyl acetate 4:1
v/v) and purified by flash-chromatography on silica gel
(petroleum ether/ethyl acetate 4:1 v/v) to give the oxirane
derivative (6.5 g, 18.6 mmol, 90%) as a white crystalline
solid: m.p. 80–82°C, recrystallized from diethyl ether; [␣]D
A solution of oxirane 2 (1.5 g, 4.3 mmol) in a mixture of
38 ml of ethanol 95% and 7.5 ml of water was stirred for
18 h at 25°C in the presence of KCN (3.9 g, 59.9 mmol).
The reaction mixture was evaporated under reduced pres-
sure. The precipitated product was extracted in dichlo-
romethane, washed with water, and dried under reduced
pressure to give the pure 17␣-cyanomethyl derivative as a
white solid (1.6 g, 4.26 mmol, 99%). The product was
analyzed by TLC on silica gel (Rf 0.25, petroleum ether/
ethyl acetate 4:1 v/v; Rf 0.35, chloroform/ethyl acetate 5:1
v/v): m.p. 148–151°C; [␣]D ϭ - 2.8° (CHCl3); ␯max (CCl4):
3605 (OH), 2250 weak (CϵN), 1105–1055 cmϪ1 (dime-
thoxy); 1H NMR (CDCl3) ␦ 0.81 (3H, s, 19-CH3), 0.90 (3H,
s, 18-CH3), 2.55 (2H, dd: J ϭ 37.1 and 16.2 Hz, CH2CN),
1
ϭ ϩ8.0° (CHCl3); ␯
(CCl4): 1105–1055 (3,3-dime-
3.14 and 3.19 (2 ϫ 3H, 2s, dimethoxy); H NMR (C5D5N)
max
thoxy), 1085 cmϪ1 (OCH2); 1H NMR (CDCl3) ␦ 0.87 (3H,
s, 18-CH3), 0.81 (3H, s, 19-CH3), 2.60 and 2.89 (2H, d: J ϭ
5.1 Hz, CH2O), 3.14 and 3.19 (2 ϫ 3H, 2s, dimethoxy); 1H
NMR (C5D5N) ␦ 0.74 (3H, s, 19-CH3), 0.90 (3H, s, 18-
CH3), 2.57 and 2.90 (2H, d: J ϭ 5.3 Hz, CH2O), 3.18 and
3.22 (2 ϫ 3H, 2s, dimethoxy); MS (LSIMSϩ) m/z (relative
intensity %) 371 (MNaϩ, 10), 317 (MHϩ-32, 100). High-
resolution MS calculated for C22H36O3Na (MNaϩ):
371.2562; found: 371.2537.
␦ 0.75 (3H, s, 19-CH3), 1.09 (3H, s, 18-CH3), 2.92 (2H, dd:
J ϭ 16.4 Hz and 31.3 Hz, CH2CN), 3.20 and 3.23 (2 ϫ 3H,
2s, dimethoxy); MS (LSIMSϩ) m/z (relative intensity %)
398 (MNaϩ, 53), 344 (MHϩ-32, 100). High-resolution MS
calculated for C23H37O3N1Na (MNaϩ): 398.2671; found:
398.2661.
2.6. 17␣-Aminoethyl-3,3-(dimethoxy)-5␣-androstan-17␤-ol
(5)
2.4. 17␣-Aminomethyl-3,3-(dimethoxy)-5␣-androstan-17␤-
ol (3)
A solution of cyanomethyl derivative 4 (1.0 g, 2.66
mmol) in 25 ml of anhydrous tetrahydrofuran was added
dropwise to a suspension of LiAlH4 (1.8 g, 47.4 mmol) in
125 ml of anhydrous tetrahydrofuran, at 4°C, under a nitro-
gen atmosphere, and stirred overnight at 22°C. The excess
of LiAlH4 was destroyed by successive dropwise additions
of water (2.0 ml), 15% aqueous NaOH (2.0 ml), and water
(6.0 ml). The reaction mixture was filtered, and the solid
fraction was rinsed with tetrahydrofuran (4 times, 50 ml).
After evaporation of the solvent, the crude residue (1.02 g)
A solution of oxirane 2 (1.0 g, 2.87 mmol) in 125 ml of
ethanol 95% was stirred for 72 h at 40°C in the presence of
50 ml of NH4OH (25%) and of 1.5 ml of acetic acid. The
reaction mixture was evaporated under reduced pressure.
The precipitate was dissolved in a mixture of petroleum
ether/ethyl acetate 4:1 v/v and filtered on a column of silica
gel 70–230 Mesh in the same solvent. The column was

462
E. Mappus et al. / Steroids 65 (2000) 459–481
was purified by TLC on silica gel (Rf 0.35, chloroform/
methanol/NH4OH 60:10:2 v/v/v) to give 17␣-aminoethyl
derivative (0.61 g, 1.61 mmol, 60.5%) as a white solid: m.p.
138–141°C, recrystallized from diethyl ether; [␣]D ϭ - 14 °
tetrahydrofuran (4 times, 50 ml). After evaporation of the
solvent, the crude residue (550 mg) was purified by TLC on
silica gel (Rf 0.45, chloroform/methanol/NH4OH 100:20:5
v/v/v) to give the 17␣-aminopropyl derivative (0.35 g, 0.89
mmol, 69%) as a white solid: m.p. 66–69°C, recrystallized
(CHCl3); ␯
(CCl4): 3605 (OH, NH2), 1105–1055 cmϪ1
max
1
(dimethoxy); H NMR (CDCl3) ␦ 0.81 (3H, s, 19-CH3),
0.87 (3H, s, 18-CH3), 1.90 (2H, m, CH2CH2NH2), 3.14 and
from diethyl ether; [␣]D ϭ Ϫ2.4° (CHCl3); ␯
(CHCl3):
max
3620 (OH), 1105–1050 cmϪ1 (dimethoxy); 1H NMR (CDCl3)
␦ 0.81 (3H, s, 19-CH3), 0.87 (3H, s, 18-CH3), 2.82 and 2.65
(4H, m, CH2NH2), 3.14 and 3.19 (2 ϫ 3H, 2s, dimethoxy); 1H
NMR (C5D5N) ␦ 0.79 (3H, s, 19-CH3), 1.13 (3H, s, 18-CH3),
2.9 (2H, m, CH2NH2), 3.21 and 3.23 (2 ϫ 3H, 2s, dimethoxy);
MS (LSIMSϩ) m/z (relative intensity %) 394 (MHϩ, 100), 376
(MHϩ-18, 25), 362 (MHϩ-32, 21), 344 (MHϩ-18–32, 63).
High-resolution MS calculated for C24H44O3N1 (MHϩ):
394.3321; found: 394.3330.
1
3.19 (2 ϫ 3H, 2s, dimethoxy); H NMR (C5D5N) ␦ 0.78
(3H, s, 19-CH3), 1.10 (3H, s, 18-CH3), 2.0 (2H, m,
CH2CH2NH2), 3.20 and 3.23 (2 ϫ 3H, 2s, dimethoxy); MS
(LSIMSϩ) m/z (relative intensity %) 380 (MHϩ, 100), 362
(MHϩ -18, 68), 348 (MHϩ-32, 15), 330 (MHϩ-18–32, 7).
High-resolution MS calculated for C23H42O3N1 (MHϩ):
380.3165; found: 380.3163.
2.7. 17␣-Cyanoethyl-3,3-(dimethoxy)-5␣-androstan-17␤-ol
(6)
2.9. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidomethyl]-3,3-
dimethoxy-5␣-androstan-17␤-ol (8)
A solution of anhydrous diisopropylamine (3.04 g, 30.04
mmol) in 60 ml of anhydrous tetrahydrofuran was stirred for
2 min at Ϫ76°C in the presence of n-butyl lithium (18.8 ml
of a 1.6 M solution in hexane, 30.08 mmol) and let to react
at Ϫ76°C with a solution of anhydrous acetonitrile (2 ml) in
6.0 ml of tetrahydrofuran. After a few minutes, an orange
color was produced. A solution of oxirane 2 (2 g, 5.74
mmol) in 10 ml of anhydrous tetrahydrofuran was added
and stirred for 2 h at 22°C. Water (5 ml) was added, and the
reaction mixture was neutralized with HCl. After evapora-
tion under reduced pressure, the product was extracted in
ethyl acetate, washed with water, and dried under reduced
pressure to give the 17␣-cyanoethyl derivative as a white
solid (2.1 g, 5.3 mmol, 92%) analyzed by TLC on silica gel
(Rf 0.30, chloroform/ethyl acetate 10:1 v/v): m.p. 120–
123°C, recrystallized from diethyl ether; [␣]D ϭ Ϫ3.8°
A solution of 17␣-aminomethyl derivative 3 (55 mg, 0.15
mmol), in a mixture of 1 ml of tetrahydrofuran and 0.25 ml of
a NaHCO3 0.1 M solution, was acylated by progressive addi-
tion of a solution of N-(5-azido-2-nitrobenzoyloxy)succinim-
ide (42 mg, 0.14 mmol, ANBNOS reagent, purchased from
Pierce) in 1 ml of tetrahydrofuran. The reaction mixture was
stirred for 24 h at 22°C in the dark and evaporated under a
nitrogen stream. The crude product was analyzed by TLC on
silica gel (Rf 0.4, petroleum ether/ethyl acetate 1:1 v/v; Rf
0.85, chloroform/methanol/NH4OH 100:20:5 v/v/v) and puri-
fied by TLC on silica gel to give the 17␣-(5-azido-2-nitroben-
zoyl)amidomethyl derivative (60 mg, 0.108 mmol, 72%) as a
white solid: m.p. 171–173°C, recrystallized from ethyl acetate;
[␣]D ϭ Ϫ23° (CHCl3); ␯
(CCl4): 3610–3470-3415 (OH
max
and NH), 2120 (N3), 1655 (CONH), 1520 (C-NO2), 1105–
(CHCl3); ␯
(CCl4): 3630(OH), 2250 weak (CϵN),
1050 cmϪ1 (dimethoxy); ␭
(EtOH) ϭ 305 nm (⑀ ϭ
max
max
1105–1055 cmϪ1 (dimethoxy); H NMR (CDCl3) ␦ 0.81
(3H, s, 19-CH3), 0.86 (3H, s, 18-CH3), 2.38 and 2.43 (4H,
m, CH2CH2CN), 3.14 and 3.19 (2 ϫ 3H, 2s, dimethoxy); 1H
NMR (C5D5N) ␦ 0.78 (3H, s, 19-CH3), 1.05 (3H, s, 18-
CH3), 1.89 and 2.12 (2H, m, CH2CH2CN), 2.68 and 2.83
(2H, m, CH2CH2CN), 3.21 and 3.23 (2 ϫ 3H, 2s, dime-
thoxy); MS (LSIMSϩ) m/z (relative intensity %) 412
(MNaϩ, 90), 358 (MHϩ-32, 100). High-resolution MS cal-
culated for C24H39O3N1 Na (MNaϩ): 412.2827; found:
412.2818.
14 200); 1H NMR (CDCl3) ␦ 0.88 (3H, s, 18-CH3), 0.81 (3H,
s, 19-CH3), 3.15 and 3.20 (2 ϫ 3H, 2s, dimethoxy), 3.35 and
3.77 (2H, d: J ϭ 13.8 Hz and dd: J ϭ 13.4 and 7.9 Hz, CHa NH
and CHb NH), 6.33 (1H, m, NHCO), 7.09 (1H, d: J ϭ 2.5 Hz,
H-6Ј), 7.15 (1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 8.12 (1H, d:
J ϭ 8.8 Hz, H-3Ј); 1H NMR (C5D5N) ␦ 0.77 (3H, s, 19-CH3),
1.12 (3H, s, 18-CH3), 3.24 and 3.25 (2 ϫ 3H, s, dimethoxy),
3.78 and 4.32 (2H, d: J ϭ 13.4 Hz and dd: J ϭ 5.5 and 5.5 Hz,
CHaNH and CHbNH), 7.18 (1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј),
7.52 (1H, d: J ϭ 2.5 Hz, H-6Ј), 8.10 (1H, d: J ϭ 8.8 Hz, H-3Ј),
9.38 (1H, m, NHCO); MS (LSIMSϩ) m/z (relative intensity %)
578 (MNaϩ, 15), 556 (MHϩ, 62), 524 (MHϩ-32, 100), 506
(MHϩ-18–32, 71). High-resolution MS calculated for
C29H42O6N5 (MHϩ): 556.3135; found: 556.3162.
1
2.8. 17␣-Aminopropyl-3,3-(dimethoxy)-5␣-androstan-17␤-
ol (7)
The solid cyanoethyl derivative 6 (500 mg, 1.29 mmol) was
added to a suspension of LiAlH4 (1.5 g, 4.08 mmol) in 40 ml
of anhydrous tetrahydrofuran, at 4°C, under a nitrogen atmo-
sphere, and stirred for 12 h at 22°C. The excess of LiAlH4 was
destroyed by successive dropwise additions of water (2 ml),
15% aqueous NaOH (2 ml), and water (6 ml). The reaction
mixture was filtered, and the solid fraction was rinsed with
2.10. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidomethyl]-17␤-
hydroxy-5␣-androstan-3-one (9)
The 3,3-dimethoxy derivative 8 (30 mg, 0.053 mmol)
was stirred at 22°C, for 4 h, in the dark, in 3 ml of a
dioxane/H2O/conc. HCl 90:10:1 v/v/v mixture. The reaction

E. Mappus et al. / Steroids 65 (2000) 459–481
463
mixture was neutralized by addition of solid NaHCO3. The
solvent was evaporated under reduced pressure, and the
product was extracted in ethyl acetate, washed with water,
and dried under reduced pressure to give the pure 17␣-[(N-
5-azido-2-nitrobenzoyl)amidomethyl]-3-oxo derivative (24
mg, 0.047 mmol, 89%), which was recrystallized from di-
ethyl ether, and analyzed by TLC on silica gel (Rf 0.45,
petroleum ether/ethyl acetate 1:1 v/v; Rf 0.5, toluene/ethyl
acetate 1:2 v/v): m.p.170–172°C; [␣]D ϭ ϩ3.7° (EtOH);
High-resolution MS calculated for C27H38O5N5 (MHϩ):
512.2873; found: 512.2897.
2.12. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidoethyl]-3,3-
dimethoxy-5␣-androstan-17␤-ol (11)
The 17␣-aminoethyl steroid 5 (65 mg, 0.17 mmol) was
condensed in the dark with N-(5-azido-2-nitrobenzoyloxy)
succinimide (40 mg, 0.13 mmol), as described above, for
the synthesis of the 17␣-aminomethyl derivative 8, and the
product was purified by TLC on silica gel (Rf 0.50, petro-
leum ether/ethyl acetate 1:2 v/v) to give the 17␣-(5-azido-
2-nitrobenzoyl)amidoethyl derivative (45 mg, 0.086 mmol,
66%): m.p.151–152°C, recrystallized from ethyl acetate;
␯
(CHCl3): 3600–3430 (OH and NH), 2120 (N3), 1705
max
(3-C ϭ O), 1670 (CONH), 1525 cmϪ1 (C-NO2); ␭
max
1
(EtOH) ϭ 306 nm (⑀ ϭ 10 600); H NMR (CDCl3) ␦ 0.92
(3H, s, 18-CH3), 1.03 (3H, s, 19-CH3), 3.36 and 3.76 (2H,
d: J ϭ 11.3 Hz and dd: J ϭ 13.4 and 7.7 Hz, CHaNH and
CHbNH), 6.49 (1H, m, NHCO), 7.09 (1H, d: J ϭ 2.5 Hz,
H-6Ј), 7.15 (1H, dd: J ϭ 8.8 and 2.4 Hz, H-4Ј), 8.12 (1H, d:
J ϭ 8.8 Hz, H-3Ј); 1H NMR (C5D5N) ␦ 0.93 (3H, s,
19-CH3), 1.13 (3H, s, 18-CH3), 3.76 and 4.30 (2H, d: J ϭ
11.2 Hz and dd: J ϭ 13.2 and 7.5 Hz, CHaNH and CHbNH),
7.15 (1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 7.52 (1H, d: J ϭ
2.5 Hz, H-6Ј), 8.10 (1H, d: J ϭ 8.8 Hz, H-3Ј), 9.44 (1H, m,
NHCO); MS (LSIMSϩ) m/z (relative intensity %) 510
(MHϩ, 100), 492 (MHϩ-18, 36). High-resolution MS cal-
culated for C27H36O5N5 (MHϩ): 510.2716; found:
510.2730.
[␣]D ϭ ϩ9.9° (EtOH); ␯
(CCl4): 3510–3440 (OH and
max
NH), 2120 (N3), 1655 (CONH), 1520 (C-NO2), 1105–1045
cmϪ1 (dimethoxy); ␭
(EtOH) ϭ 306 nm (⑀ ϭ 12 300);
max
1H NMR (CDCl3) ␦ 0.81 (3H, s, 19-CH3), 0.83 (3H, s,
18-CH3), 3.14 and 3.19 (2 ϫ 3H, 2s, dimethoxy), 3.45 and
3.88 (2H, d: J ϭ 3.8 Hz and dd: J ϭ 13.8 and 7.2 Hz, CHa
NH and CHbNH), 6.92 (1H, m, NHCO), 7.10 (1H, d: J ϭ
2.3 Hz, H-6Ј), 7.12 (1H, dd: J ϭ 8.4 and 2.3 Hz, H-4Ј,
superimposed to H-6Ј d at 7.10), 8.09 (1H, d: J ϭ 8.4 Hz,
H-3Ј); 1H NMR (C5D5N) ␦ 0.77 (3H, s, 19-CH3), 1.10 (3H,
s, 18-CH3), 3.22 and 3.23 (2 ϫ 3H, 2s, dimethoxy), 4.22
(2H, dd, CH2NH), 7.11 (1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј),
7.58 (1H, d: J ϭ 2.5 Hz, H-6Ј), 8.06 (1H, d: J ϭ 8.8 Hz,
H-3Ј), 9.03 (1H, m, NHCO); MS (LSIMSϩ) m/z (relative
intensity %) 570 (MHϩ, 25), 538 (MHϩ-32, 60), 520
(MHϩ-18–32, 100). High-resolution MS calculated for
C30H44O6N5 (MHϩ): 570.3292; found: 570.3264.
2.11. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidomethyl]-5␣-
androstane-3␤,17␤-diol (10)
A solution of the 3-oxo derivative 9 (35 mg, 0.07 mmol)
in 8 ml of methanol was reduced for 5 min, at 4°C, in the
dark, with a solution of NaBH4 (5 mg, 0.13 mmol) in 250 ␮l
of methanol. The reaction was stopped by addition of water
(175 ␮l) and acetic acid (1.5 ␮l). The solution was evapo-
rated under reduced pressure, and the product was purified
by TLC on silica gel (Rf 0.40, chloroform/ethyl acetate 1:1
v/v) to give the 17␣-[(N-5-azido-2-nitrobenzoyl)amidom-
ethyl]-5␣-androstane-3␤,17␤-diol derivative (15 mg, 0.029
mmol, 42%) as a white solid that was analyzed by TLC on
silica gel (Rf 0.49, petroleum ether/ethyl acetate 1:1 v/v; Rf
0.34, chloroform/ethyl acetate 1:1 v/v): m.p. 114–115°C,
recrystallized from ethyl acetate; [␣]D ϭ Ϫ15.6° (EtOH);
2.13. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidoethyl]-17␤-
hydroxy-5␣-androstan-3-one (12)
The 3,3-dimethoxy derivative 11 (45 mg, 0.086 mmol)
was acidolyzed, as described above for the synthesis of the
3-oxo derivative 9, and the product was purified by TLC on
silica gel (Rf 0.30, petroleum ether/ethyl acetate 1:2 v/v) to
give the 17␣-[(N-5-azido-2-nitrobenzoyl)amidoethyl]-3-
oxo derivative (40 mg, 0.076 mmol, 95%): m.p. 98–99°C,
recrystallized from ethyl acetate; [␣]D ϭ ϩ14.7° (EtOH);
␯
(CHCl3): 3700–3610-3430 (OH and NH), 2120 (N3),
max
1665 (CONH), 1530 cmϪ1 (C-NO2); ␭
(EtOH) ϭ 306
␯
(CHCl3): 3690–3420 (OH and NH), 2120 (N3), 1705
max
max
nm (⑀ ϭ 10 700); 1H NMR (CDCl3) ␦ 0.83 (3H, s, 19-CH3),
0.89 (3H, s, 18-CH3), 3.36 and 3.78 (2H, d: J ϭ 13.0 Hz and
dd: J ϭ 13.3 and 7.8 Hz, CHaNH and CHbNH), 3.61 (1H, m,
H-3␣), 6.38 (1H, m, NHCO), 7.09 (1H, d: J ϭ 2.5 Hz,
H-6Ј), 7.16 (1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 8.13 (1H, d:
J ϭ 8.8 Hz, H-3Ј); 1H NMR (C5D5N) ␦ 0.86 (3H, s,
19-CH3), 1.12 (3H, s, 18-CH3), 3.76 and 4.33 (2H, d: J ϭ
12.0 Hz and dd: J ϭ 13.2 and 7.7 Hz, CHaNH and CHbNH),
3.93 (1H, m, H-3␣), 7.14 (1H, dd: J ϭ 8.7 and 2.4 Hz,
H-4Ј), 7.52 (1H, d: J ϭ 2.5 Hz, H-6Ј), 8.10 (1H, d: J ϭ 8.7
Hz, H-3Ј), 9.43 (1H, m, NHCO); MS (LSIMSϩ) m/z (rela-
tive intensity %) 512 (MHϩ, 100), 494 (MHϩ-18, 33).
(3-C ϭ O), 1670 (CONH), 1525 cmϪ1 (C-NO2); ␭
max
1
(EtOH) ϭ 305 nm (⑀ ϭ 11 200); H NMR (CDCl3) ␦ 0.86
(3H, s, 18-CH3), 1.03 (3H, s, 19-CH3), 3.46 and 3.89 (2H,
m, and dd: J ϭ 14.1 and 7.5 Hz, CHaNH and CHbNH),
6.83 (1H, m, NHCO), 7.10 (1H, d: J ϭ 2 Hz, H-6Ј, 7.12
(1H, dd: J ϭ 11.41 and 2.5 Hz, H-4Ј, superimposed to signal
at 7.10), 8.10 (1H, d: J ϭ 8.5 Hz, H-3Ј); 1H NMR (C5D5N)
␦ 0.92 (3H, s, 19-CH3), 1.11 (3H, s, 18-CH3), 4.22 (2H, dd:
J ϭ 7.5 and 13.3 Hz, CHaNH and CHbNH), 7.12 (1H, dd:
J ϭ 8.8 and 2.5 Hz, H-4Ј), 7.57 (1H, d: J ϭ 2.5 Hz, H-6Ј),
8.07 (1H, d: J ϭ 8.8 Hz, H-3Ј), 9.56 (1H, m, NHCO); MS
(LSIMSϩ) m/z (relative intensity %) 524 (MHϩ, 72), 506

464
E. Mappus et al. / Steroids 65 (2000) 459–481
(MHϩ-18, 100). High-resolution MS calculated for
C28H38O5N5 (MHϩ): 524.2873; found: 524.2875.
MS (LSIMSϩ) m/z (relative intensity %) 584 (MHϩ, 11),
552 (MHϩ-32, 49), 534 (MHϩ-18–32, 100). High-resolu-
tion MS calculated for C31H46O6N5 (MHϩ): 584.3448;
found: 584.3417.
2.14. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidoethyl]-
hydroxy-5␣-androstane-3␤,17␤-diol (13)
2.16. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidopropyl]-17␤-
hydroxy-5␣-androstan-3-one (15)
The 3-oxo derivative 12 (28 mg, 0.053 mmol) was re-
duced with NaBH4, as described above for the synthesis of
the diol derivative 10, and the product was purified by TLC
on silica gel (Rf 0.40, chloroform/ethyl acetate 1:5 v/v) to
give the 17␣-[(N-5-azido-2-nitrobenzoyl)amidoethyl]-
3␤,17␤-diol derivative (22 mg, 0.042 mmol, 79%): m.p.
The 3,3-dimethoxy derivative 14 (25 mg, 0.043 mmol)
was acidolyzed, as described above for the synthesis of the
3-oxo derivative 9, to give the 17␣-[(N-5-azido-2-nitroben-
zoyl)amidopropyl]-3-oxo derivative, which was purified by
TLC on silica gel (Rf 0.35, petroleum ether/ethyl acetate 1:2
v/v) as a white solid (22 mg, 0.076 mmol, 95%): m.p.
81–83°C, recrystallized from ethyl acetate; [␣]D ϭ Ϫ17.4°
162–164°C, recrystallized from ethyl acetate; [␣]D
ϭ
Ϫ11.9° (EtOH); ␯
(CHCl3): 3690–3620 (OH and NH),
max
2120 (N3), 1670 (CONH), 1525 cmϪ1(C-NO2); ␭
max
1
(EtOH) ϭ 305 nm (⑀ ϭ 11 750); H NMR (CDCl3) ␦ 0.83
(3H, s, 18-CH3), 0.82 (3H, s, 19-CH3), 3.46 and 3.90 (2H,
dd: J ϭ 3.75 and 9.7 Hz, and dd: J ϭ 7.2 and 13.8 Hz, CHa
NH and CHbNH), 3.6 (1H, m, H-3␣), 6.91 (1H, m, NHCO),
7.10 (1H, d: J ϭ 2 Hz, H-6Ј, 7.12 (1H, dd: J ϭ 11.41 and
2.5 Hz, H-4Ј, superimposed to H-6Ј d at 7.10), 8.10 (1H, d:
J ϭ 8.5 Hz, H-3Ј); 1H NMR (C5D5N) ␦ 0.85 (3H, s,
19-CH3), 1.10 (3H, s, 18-CH3), 3.90 (1H, m, H-3␣), 4.22
(2H, dd: J ϭ 7.5 and 13.0 Hz, CHaNH and CHbNH), 7.11
(1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 7.58 (1H, d: J ϭ 2.1 Hz,
H-6Ј, superimposed to signal of pyridine), 8.06 (1H, d: J ϭ
8.8 Hz, H-3Ј), 9.54 (1H, m, NHCO); MS (LSIMSϩ) m/z
(relative intensity %) 548 (MNaϩ, 27), 526 (MHϩ, 53), 508
(MHϩ-18, 100), 500 (MHϩ-28, 32). High-resolution MS
calculated for C28H40O5N5 (MHϩ): 526.3029; found:
526.2998.
(EtOH); ␯
(CHCl3): 3600–3455-3300 (OH and NH),
max
2120 (N3), 1705 (3-C ϭ O), 1670 (CONH), 1530 cmϪ1
1
(C-NO2); ␭
(EtOH) ϭ 305 nm (⑀ ϭ 11 000); H NMR
max
(CDCl3) ␦ 0.87 (3H, s, 18-CH3), 1.02 (3H, s, 19-CH3), 3.44
and 3.54 (2H, d: J ϭ 13Hz, and dd: J ϭ 6.2 and 6.6 Hz, CHa
NH and CHbNH), 6.42 (1H, m, NHCO), 7.10 (1H, d: J ϭ
2.3 Hz, H-6Ј), 7.13 (1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 8.10
(1H, d: J ϭ 8.8 Hz, H-3Ј); 1H NMR (C5D5N) ␦ 0.94 (3H, s,
19-CH3), 1.12 (3H, s, 18-CH3), 3.87 (2H, dd: J ϭ 6.0 and
6.5 Hz CHaNH and CHbNH), 7.11 (1H, dd: J ϭ 8.8 and 2.5
Hz, H-4Ј), 7.51 (1H, d: J ϭ 2.5 Hz, H-6Ј), 8.06 (1H, d: J ϭ
8.8 Hz, H-3Ј), 9.66 (1H, m, NHCO); MS (LSIMSϩ) m/z
(relative intensity %) 538 (MHϩ, 34), 520 (MHϩ-18, 100).
High-resolution MS calculated for C29H40O5N5 (MHϩ):
538.3029; found: 538.3047.
2.15. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidopropyl]-3,3-
dimethoxy-5␣-androstan-17␤-ol (14)
2.17. 17␣-[(N-5-Azido-2-nitrobenzoyl)amidopropyl]-5␣-
androstane-3␤,17␤-diol (16)
The 17␣-aminopropyl steroid 7 (60 mg, 0.153 mmol)
was condensed in the dark with N-(5-azido-2-nitrobenzoy-
loxy)succinimide (40 mg, 0.13 mmol), as described above,
for the synthesis of 17␣-aminomethyl and 17␣-aminoethyl
derivatives 8 and 11, and the product was purified by TLC
on silica gel (Rf 0.35, petroleum ether/ethyl acetate 1:2 v/v)
to give the 17␣-(5-azido-2-nitrobenzoyl)amidopropyl deriv-
ative (50 mg, 0.086 mmol, 56%): m.p. 97–98°C, recrystal-
The 3-oxo derivative 15 (20 mg, 0.037 mmol) was re-
duced with NaBH4, as described above for the synthesis of
the diol derivative 10, and the product was purified by TLC
on silica gel (Rf 0.30, chloroform/ethyl acetate 1:1) to give
the 17␣-[(N-5-azido-2-nitrobenzoyl)amidopropyl]-3␤,17␤-
diol derivative (18 mg, 0.033 mmol, 90%), which was
analyzed by TLC on silica gel (Rf 0.40, chloroform/ethyl
acetate 1:3 v/v): m.p. 166–167°C, recrystallized from di-
lized from diethyl ether; [␣]D ϭ Ϫ7.4° (EtOH); ␯
ethyl ether; [␣]D ϭ Ϫ23.8° (EtOH); ␯
(CHCl3): 3690–
max
max
(CCl4): 3630–3450-3330 (OH and NH), 2120 (N3), 1685
3610-3530–3430-3310 (OH and NH), 2120 (N3), 1670
(CONH), 1530 (C-NO2), 1105–1055 cmϪ1 (dimethoxy);
(CONH), 1530 cmϪ1 (C-NO2); ␭max (EtOH) ϭ 305 nm (⑀ ϭ
1
1
␭
(EtOH) ϭ 305 nm (⑀ ϭ 11 390); H NMR (CDCl3) ␦
11 700); H NMR (CDCl3) ␦ 0.82 (3H, s, 19-CH3), 0.84
max
0.80 (3H, s, 19-CH3), 0.83 (3H, s, 18-CH3), 3.14 and 3.19
(2 ϫ 3H, 2s, dimethoxy), 3.44 and 3.52 (2H, 2 m, CH2NH),
6.42 (1H, m, NHCO), 7.11 (1H, d: J ϭ 2.5 Hz, H-6Ј), 7.14
(1H, dd: J ϭ 8.5 and 2.5 Hz, H-4Ј, superimposed to H-6Ј d
at 7.11), 8.10 (1H, d: J ϭ 8.5 Hz, H-3Ј); 1H NMR (C5D5N)
␦ 0.78 (3H, s, 19-CH3), 1.10 (3H, s, 18-CH3), 3.20 and 3.23
(2 ϫ 3H, 2s, dimethoxy), 3.87 (2H, m, CH2NH), 7.10 (1H,
dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 7.50 (1H, d: J ϭ 2.5 Hz,
H-6Ј), 8.04 (1H, d: J ϭ 8.8 Hz, H-3Ј), 9.62 (1H, m, NHCO);
(3H, s, 18-CH3), 3.45 and 3.55 (2H, 2 m, CH2NH), 3.59
superimposed to CH2NH m at 3.55 (1H, m, H-3␣), 6.43
(1H, m, NHCO), 7.11 (1H, d: J ϭ 2.4 Hz, H-6Ј), 7.13 (1H,
dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 8.10 (1H, d: J ϭ 8.8 Hz,
H-3Ј); 1H NMR (C5D5N) ␦ 0.85 (3H, s, 19-CH3), 1.12 (3H,
s, 18-CH3), 3.87 (3H, m, CH2NH), 3.44 (1H, m, H-3␣), 7.12
(1H, dd: J ϭ 8.8 and 2.5 Hz, H-4Ј), 7.50 (1H, d: J ϭ 2.4 Hz,
H-6Ј), 8.07 (1H, d: J ϭ 8.8 Hz, H-3Ј), 9.72 (1H, m, NHCO);
MS (LSIMSϩ) m/z (relative intensity %) 540 (MHϩ, 55),

E. Mappus et al. / Steroids 65 (2000) 459–481
465
522 (MHϩ-18, 100). High-resolution MS calculated for
C29H42O5N5 (MHϩ): 540.3186; found: 540.3174.
and 3.69 (2H, dd: J ϭ 3.1 and 5.5 Hz and d, J ϭ 12.3 Hz,
CHaNH and CHbNH), 7.32 (2H, m, H-5Ј ϩ H-6Ј), 7.97 (1H,
d: J ϭ 2.6 Hz, H-3Ј), 6.60 (1H, m, NH); MS (LSIMSϩ) m/z
(relative intensity %): 482 (MHϩ, 100). High-resolution MS
calculated for C26H36O4N5 (MHϩ): 482.2767; found:
482.2772.
2.18. 17␣-[(N-4-Azido-2-nitrophenyl)aminomethyl]-3,3-
dimethoxy-5␣-androstan-17␤-ol (17)
A solution of 17␣-aminomethyl steroid 3 (100 mg, 0.27
mmol) in 2 ml of tetrahydrofuran and 1 ml of an ethanol/
ether/triethylamine 20:10:0.2 v/v/v mixture was stirred
overnight in the dark, in the presence of 4-azido-1-fluoro-
2-nitrobenzene (100 mg, 0.55 mmol) dissolved in 3 ml of
tetrahydrofuran. The reaction mixture was evaporated under
reduced pressure, in the dark. The product was purified with
efficient separation from the reagent by TLC on silica gel
(Rf 0.45, chloroform stabilized with 0.5% ethanol) to give
the 17␣-(N-4-azido-2-nitrophenyl)aminomethyl derivative
(130 mg, 0.24 mmol, 91%), which was also analyzed by
TLC on silica gel (Rf 0.70, chloroform/ethyl acetate 3:1
v/v): m.p.138–139°C, recrystallized from ethyl acetate;
2.20. 17␣-[(N-4-Azido-2-nitrophenyl)aminomethyl]-5␣-
androstane-3␤,17␤-diol (19)
The 3-oxo derivative 18 (40 mg, 0.083 mmol) was re-
duced with NaBH4, as described above for the synthesis of
the diol derivative 10, and the product was purified by TLC
on silica gel (Rf 0.35, chloroform/ethyl acetate 3:1 v/v) to
give the 17␣-[(N-4-azido-2-nitrophenyl)aminomethyl]-
3␤,17␤-diol derivative (26 mg, 0.054 mmol, 66%): m.p.
84–89°C, recrystallized from diethyl ether; [␣]D ϭ Ϫ10.7°
(EtOH); ␯
(CHCl3): 3690–3610-3380 (OH and NH),
max
2120 (N3), 1705 (3-C ϭ O), 1520 cmϪ1 (C-NO2); ␭
max
1
[␣]D ϭ Ϫ7.9° (EtOH); ␯
(CHCl3): 3690–3610-3380
(EtOH) ϭ 260 nm (⑀ ϭ 22 500), 459 nm (⑀ ϭ 7400); H
NMR (CDCl3) ␦ 0.84 (3H, s, 19-CH3), 0.94 (3H, s, 18-
CH3), 3.26 and 3.41 (2H, dd: J ϭ 3.3 and 5.9 Hz and d, J ϭ
12.4 Hz, CHaNH and CHbNH), 3.58 (1H, m, H-3␣), 6.90
(1H, d: J ϭ 9.2 Hz, H-6Ј), 7.11 (1H, dd: J ϭ 9.2 and 2.7 Hz,
H-5Ј), 7.87 (1H, d: J ϭ 2.8 Hz, H-3Ј); 8.39 (1H, m, NH); 1H
NMR (C5D5N) ␦ 0.86 (3H, s, 19-CH3), 1.17 (3H, s, 18-
CH3), 3.57 and 3.72 (2H, dd: J ϭ 3.3 and 5.7 Hz and d, J ϭ
12.2 Hz, CHaNH and CHbNH), 3.89 (1H, m, H-3␣), 6.60
(1H, m, NH), 7.33 (2H, m, H-5Јϩ H-6Ј), 7.97 (1H, d: J ϭ
2.8 Hz, H-3Ј); MS (LSIMSϩ) m/z (relative intensity %) 484
(MHϩ, 100). High-resolution MS calculated for
C26H38O4N5 (MHϩ): 484.2924; found: 484.2916.
max
(OH and NH), 2120 (N3), 1520 (C-NO2), 1105–1050 cmϪ1
(dimethoxy); ␭max (EtOH) ϭ 261 nm (⑀ ϭ 21 900), 466 nm
(⑀ ϭ 5700); 1H NMR (CDCl3) ␦ 0.82 (3H, s, 19-CH3), 0.94
(3H, s, 18-CH3), 3.14 and 3.19 (2 ϫ 3H, 2s, dimethoxy),
3.25 and 3.43 (2H, 2 m, CH2NH), 6.90 (1H, d: J ϭ 9.2 Hz,
H-6Ј), 7.12 (1H, dd: J ϭ 9.2 and 2.6 Hz, H-5Ј), 7.88 (1H, d:
1
J ϭ 2.7 Hz, H-3Ј); 8.4 (1H, m, NH); H NMR (C5D5N) ␦
0.78 (3H, s, 19-CH3), 1.17 (3H, s, 18-CH3),), 3.21 and 3.24
(2 ϫ 3H, 2s, dimethoxy), 3.60 and 3.74 (2H, dd: J ϭ 12.2
and 5.6 Hz and d, J ϭ 12.2 Hz, CHaNH and CHbNH), 7.35
(1H, dd: J ϭ 9.2 and 2.6 Hz, H-5Ј), 7.44 (1H, d: J ϭ 9.2 Hz,
H-6Ј), 7.98 (1H, d: J ϭ 2.6 Hz, H-3Ј); MS (LSIMSϩ) m/z
(relative intensity %) 528 (MHϩ, 41), 496 (MHϩ-32, 100).
High-resolution MS calculated for C28H42O5N5 (MHϩ):
528.3186; found: 528.3172.
2.21. 17␣-[(N-4-Azido-2-nitrophenyl)aminoethyl]-3,3-
dimethoxy-5␣-androstan-17␤-ol (20)
2.19. 17␣-[(N-4-Azido-2-nitrophenyl)aminomethyl]-17␤-
hydroxy-5␣-androstan-3-one (18)
The 17␣-aminoethyl steroid 5 (65 mg, 0.171 mmol)
was condensed with 4-azido-1-fluoro-2-nitrobenzene as
described above for the 17␣-aminoethyl derivative 17,
and the product was purified by TLC on silica gel (Rf
0.55, chloroform/ethyl acetate10:1 v/v) to give the 17␣-
(N-4-azido-2-nitrophenyl)aminoethyl derivative (30 mg,
0.055 mmol, 33%): m.p. 115–118°C, recrystallized from
The 3,3-dimethoxy derivative 17 (130 mg, 0.24 mmol)
was acidolyzed, as described above for the synthesis of the
3-oxo derivative 9, and the product was purified by TLC on
silica gel (Rf 0.35, chloroform stabilized with 0.5% ethanol)
to give the 17␣-[(N-4-azido-2-nitrophenyl)aminomethyl]-3-
oxo derivative (60 mg, 0.124 mmol, 50%), which was also
analyzed by TLC on silica gel (Rf 0.65, chloroform/ethyl
acetate 3:1 v/v): m.p. 153–155°C, recrystallized from di-
diethyl ether; [␣]D ϭ Ϫ20.4° (EtOH); ␯
(CCl4):
max
3690–3630-3380 (OH and NH), 2115 (N3), 1520 (C-
NO2), 1105–1055 cmϪ1 (dimethoxy); ␭max (EtOH) ϭ 260
1
nm (⑀ ϭ 20 800), 458 nm (⑀ ϭ 5600); H NMR (CDCl3)
ethyl ether; [␣]D ϭ Ϫ7.5° (EtOH); ␯
(CHCl3): 3690–
␦ 0.82 (3H, s, 19-CH3), 0.88 (3H, s, 18-CH3), 3.14 and
3.19 (2 ϫ 3H, 2s, dimethoxy), 3.50 and 3.55 (2H, 2 m,
CH2NH), 6.93 (1H, d: J ϭ 9.2 Hz, H-6Ј), 7.14 (1H, dd:
J ϭ 9.2 and 2.7 Hz, H-5Ј), 7.87 (1H, d: J ϭ 2.7 Hz, H-3Ј);
8.42 (1H, m, NH); 1H NMR (C5D5N) ␦ 0.80 (3H, s,
19-CH3), 1.11 (3H, s, 18-CH3), 3.21 and 3.24 (2 ϫ 3H,
2s, dimethoxy), 3.69 and 3.80 (2H, 2 m, CH2NH), 7.11
(1H, d: J ϭ 9.2 Hz, H-6Ј), 7.19 (1H, dd: J ϭ 9.2 and 2.6
Hz, H-5Ј), 7.95 (1H, d: J ϭ 2.7 Hz, H-3Ј), 9.07 (1H, m,
max
3610-3380 (OH and NH), 2120 (N3), 1705 (3-C ϭ O), 1520
cmϪ1 (C-NO2); ␭
(EtOH) ϭ 260 nm (⑀ ϭ 19 200), 464
max
1
nm (⑀ ϭ 5000); H NMR (CDCl3) ␦ 0.97 (3H, s, 18-CH3),
1.04 (3H, s, 19-CH3), 3.27 and 3.43 (2H, dd: J ϭ 3.6 and 6.0
Hz and d, J ϭ 12.4 Hz, CHaNH and CHbNH), 6.90 (1H, d:
J ϭ 9.2 Hz, H-6Ј), 7.11 (1H, dd: J ϭ 9.2 and 2.7 Hz, H-5Ј),
7.87 (1H, d: J ϭ 2.7 Hz, H-3Ј); 8.4 (1H, m, NH); 1H NMR
(C5D5N) ␦ 0.95 (3H, s, 19-CH3), 1.18 (3H, s, 18-CH3), 3.59

466
E. Mappus et al. / Steroids 65 (2000) 459–481
NH); MS (LSIMSϩ) m/z (relative intensity %) 542
(MHϩ, 53), 510 (MHϩ-32, 100). High-resolution MS
calculated for C29H44O5N5 (MHϩ): 542.3342; found:
542.3336.
2.24. 17␣-[(N-4-Azido-2-nitrophenyl)aminopropyl]-3,3-
dimethoxy-5␣-androstan-17␤-ol (23)
The 17␣-aminopropyl steroid 7 (90 mg, 0.228 mmol)
was condensed with 4-azido-1-fluoro-2-nitrobenzene as de-
scribed above for the 17␣-aminomethyl derivative 17, and
the product was purified by TLC on silica gel (Rf 0.40,
chloroform/ethyl acetate 15:1 v/v) to give the 17␣-(N-4-
azido-2-nitrophenyl)aminopropyl derivative (60 mg, 0.108
mmol, 47%) which was also analyzed by TLC on silica gel
(Rf 0.60, chloroform/ethyl acetate 4:1 v/v): m.p.
142–143°C, recrystallized from diethyl ether; [␣]D ϭ Ϫ6.9°
(EtOH); ␯max (CCl4): 3695–3610 (OH and NH), 2120 (N3),
2.22. 17␣-[(N-4-Azido-2-nitrophenyl)aminoethyl]-17␤-
hydroxy-5␣-androstan-3-one (21)
The 3,3-dimethoxy derivative 20 (13 mg, 0.024 mmol)
was acidolyzed, as described above for the synthesis of
the 3-oxo derivative 18, and the product was purified by
TLC on silica gel (Rf 0.50, chloroform/ethyl acetate 10:1
v/v) to give the 17␣-[(N-4-azido-2-nitrophenyl)amino-
ethyl]-3-oxo derivative (10 mg, 0.02 mmol, 84%): m.p.
1520 C-NO2), 1105–1055 cmϪ1 (dimethoxy);
␭
max
1
129–131°C, recrystallized from ethyl acetate; [␣]D
ϭ
(EtOH) ϭ 260 nm (⑀ ϭ 22 400), 463 nm (⑀ ϭ 5600); H
NMR (CDCl3) ␦ 0.81 (3H, s, 19-CH3), 0.87 (3H, s, 18-
CH3), 3.14 and 3.19 (2 ϫ 3H, 2s, dimethoxy), 3.36 (2H, m,
CH2NH), 6.90 (1H, d: J ϭ 9.2 Hz, H-6Ј), 7.13 (1H, dd: J ϭ
9.2 and 2.7 Hz, H-5Ј), 7.87 (1H, d: J ϭ 2.7 Hz, H-3Ј), 8.12
ϩ24.0° (EtOH); ␯
(CHCl3): 3690–3630-3380 (OH
max
and NH), 2120 (N3), 1705 (3-C ϭ O), 1520 cmϪ1 (C-
NO2); ␭max (EtOH) ϭ 260 nm (⑀ ϭ 19 200), 463 nm (⑀ ϭ
1
5200); H NMR (CDCl3) ␦ 0.92 (3H, s, 18-CH3), 1.04
1
(3H, s, 19-CH3), 3.50 and 3.55 (2H, 2 m, CH2NH), 6.94
(1H, d: J ϭ 9.1 Hz, H-6Ј), 7.13 (1H, dd: J ϭ 9.1 and 2.6
Hz, H-5Ј), 7.86 (1H, d: J ϭ 2.6 Hz, H-3Ј), 8.40 (1H, m,
(1H, m, NH); H NMR (C5D5N) ␦ 0.80 (3H, s, 19-CH3),
1.12 (3H, s, 18-CH3), 3.21 and 3.24 (2 ϫ 3H, 2s, dime-
thoxy), 3.47 (2H, m, CH2NH), 7.07 (1H, d: J ϭ 9.3 Hz,
H-6Ј), 7.19 (1H, dd: J ϭ 9.2 and 2.6 Hz, H-5Ј), 7.93(1H, d:
J ϭ 2.7 Hz, H-3Ј), 8.52 (1H, m, NH); MS (LSIMSϩ) m/z
(relative intensity %): 556 (MHϩ, 41), 538 (MHϩ-18, 5),
524 (MHϩ-32, 100). High-resolution MS calculated for
C30H46O5N5 (MHϩ): 556.3499; found: 556.3502.
1
NH); H NMR (C5D5N) ␦ 0.96 (3H, s, 19-CH3), 1.12
(3H, s, 18-CH3), 3.68 and 3.78 (2H, 2 m, CH2NH), 7.11
(1H, d: J ϭ 9.3 Hz, H-6Ј), 7.18 (1H, dd: J ϭ 9.3 and 2.6
Hz, H-5Ј), 7.95 (1H, d: J ϭ 2.6 Hz, H-3Ј), 9.04 (1H, m,
NH); MS (LSIMSϩ) m/z (relative intensity %) 496
(MHϩ,100). High-resolution MS calculated for
C27H38O4N5 (MHϩ): 496.2924; found: 496.2914.
2.25. 17␣-[(N-4-Azido-2-nitrophenyl)aminopropyl]-17␤-
hydroxy-5␣-androstan-3-one (24)
2.23. 17␣-[(N-4-Azido-2-nitrophenyl)aminoethyl]-5␣-
androstane-3␤,17␤-diol (22)
The 3,3-dimethoxy derivative 23 (35 mg, 0.63 mmol)
was acidolyzed, as described above for the synthesis of the
3-oxo derivative 12, and the product was purified by TLC
on silica gel (Rf 0.35, chloroform/ethyl acetate 15:1 v/v) to
give the 17␣-[(N-4-azido-2-nitrophenyl)aminopropyl]-3-
oxo derivative (31 mg, 0.061 mmol, 97%), which was also
analyzed by TLC on silica gel (Rf 0.55, chloroform/ethyl
acetate 4:1 v/v): m.p. 132–134°C, recrystallized from di-
The 3-oxo derivative 21 (22 mg, 0.044 mmol) was re-
duced with NaBH4, as described above for the synthesis of
the diol derivative 10, and the product was purified by TLC
on silica gel (Rf 0.30, chloroform/ethyl acetate 4:1 v/v) to
give the 17␣-[(N-4-azido-2-nitrophenyl)aminoethyl]-3␤,17␤-
diol derivative (9 mg, 0.018 mmol, 42%): m.p. 136–137°C,
recrystallized from diethyl ether; [␣]D ϭ Ϫ28.3° (EtOH); ␯
ethyl ether; [␣]D ϭ ϩ8.7° (EtOH); ␯
(CHCl3): 3740–
max
max
(CHCl3): 3695–3610- (OH and NH), 2120 (N3), 1525 cmϪ1
(C-NO2); ␭max (EtOH) ϭ 260 nm (⑀ ϭ 18 800), 463 nm (⑀ ϭ
5200); 1H NMR (CDCl3) ␦ 0.83 (3H, s, 19-CH3), 0.88 (3H, s,
18-CH3), 3.49 (1H, m, H-3␣), 3.50 and 3.54 (2H, 2 m,
CH2NH), 6.94 (1H, d: J ϭ 9.1 Hz, H-6Ј), 7.13 (1H, dd: J ϭ 9.1
and 2.6 Hz, H-5Ј), 7.86 (1H, d: J ϭ 2.6 Hz, H-3Ј), 8.40 (1H, m,
NH); 1H NMR (C5D5N) ␦ 0.88 (3H, s, 19-CH3), 1.12 (3H, s,
18-CH3), 3.69 and 3.79 (2H, dd, J ϭ 9.4 and 2.6 Hz and d: J ϭ
12.5 Hz, CHaNH and CHbNH), 3.92 (1H, m, H-3␣), 7.11 (1H,
d: J ϭ 9.2 Hz, H-6Ј), 7.18 (1H, dd: J ϭ 5.0 and 7.5 Hz, and d:
H-5Ј), 7.95 (1H, d: J ϭ 2.6 Hz, H-3Ј), 9.07 (1H, m, NH); MS
(LSIMSϩ) m/z (relative intensity %) 498 (MHϩ, 100). High-
resolution MS calculated for C27H40O4N5 (MHϩ): 498.3080;
found: 498.3093.
3640-3370 (OH and NH), 2120 (N3), 1715 (3-C ϭ O), 1520
cmϪ1 (C-NO2); ␭
(EtOH) ϭ 261 nm (⑀ ϭ 21 300), 460
max
1
nm (⑀ ϭ 5300); H NMR (CDCl3) ␦ 0.90 (3H, s, 18-CH3),
1.03 (3H, s, 19-CH3), 3.35 (2H, dd: J ϭ 5.8 and 6.0 Hz and
d, J ϭ 12.2 Hz, CHaNH and CHbNH), 6.89 (1H, d: J ϭ 9.2
Hz, H-6Ј), 7.13 (1H, dd: J ϭ 9.2 and 2.6 Hz, H-5Ј), 7.86
1
(1H, d: J ϭ 2.6 Hz, H-3Ј), 8.11 (1H, m, NH); H NMR
(C5D5N) ␦ 0.95 (3H, s, 19-CH3), 1.13 (3H, s, 18-CH3), 3.46
(2H, dd: J ϭ 5.8 and 6.6 Hz and d, J ϭ 12.1 Hz, CHaNH and
CHbNH), 7.06 (1H, d: J ϭ 9.3 Hz, H-6Ј), 7.18 (1H, dd: J ϭ
9.3 and 2.6 Hz, H-5Ј), 7.92 (1H, d: J ϭ 2.6 Hz, H-3Ј), 8.51
(1H, m, NH); MS (LSIMSϩ) m/z (relative intensity %) 510
(MHϩ, 100). High-resolution MS calculated for
C28H40O4N5 (MHϩ): 510.3080; found: 510.3086.

E. Mappus et al. / Steroids 65 (2000) 459–481
467
2.26. 17␣-[(N-4-Azido-2-nitrophenyl)aminopropyl]-5␣-
androstane-3␤,17␤-diol (25)
2.28. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminomethyl]-17␤-hydroxy-5␣-androstan-3-one (27)
The 3,3-dimethoxy derivative 26 (121 mg, 0.21 mmol)
was acidolyzed, as described above for the synthesis of the
3-oxo derivative 9, and the product was purified by TLC on
silica gel (Rf 0.40, chloroform/ethyl acetate 2:1 v/v) to give
the 17␣-[(N-5-azido-2-nitro-3,4,6-trifluorophenyl)amino-
methyl]-3-oxo derivative (100 mg, 0.187 mmol, 89%): m.p.
111–114°C, recrystallized from diethyl ether; [␣]D ϭ Ϫ3.9°
The 3-oxo derivative 24 (30 mg, 0.059 mmol) was re-
duced with NaBH4, as described above for the synthesis of
the diol derivative 10, and the product was purified by TLC
on silica gel (Rf 0.30, chloroform/ethyl acetate 4:1 v/v) to
give the 17␣-[(N-4-azido-2-nitrophenyl)aminopropyl]-
3␤,17␤-diol derivative (12 mg, 0.024 mmol, 41%): m.p.
118–120°C, recrystallized from diethyl ether; [␣]D
Ϫ15.3° (EtOH); ␯max (CHCl3): 3690–3600-3390- (OH and
NH), 2120 (N3), 1520 cmϪ1 (C-NO2); ␭
(EtOH) ϭ 260
ϭ
(EtOH); ␯
(CHCl3): 3690–3615-3370 (OH and NH),
max
2130 (N3), 1705 (3-C ϭ O), 1505 cmϪ1 (C-NO2); ␭
max
(EtOH) ϭ 235 nm (⑀ ϭ 22 500), 298 nm (⑀ ϭ 9500), 420
nm (⑀ ϭ 4400); H NMR (CDCl3) ␦ 0.92 (3H, s, 18-CH3),
max
nm (⑀ ϭ 20 500), 463 nm (⑀ ϭ 5000); 1H NMR (CDCl3) ␦
0.83 (3H, s, 19-CH3), 0.87 (3H, s, 18-CH3), 3.36 (2H, m,
CH2NH), 3.59 (1H, m, H-3␣), 6.86 (1H, m, NHCO), 6.88
(1H, d: J ϭ 9.3 Hz, H-6Ј), 7.12 (1H, dd: J ϭ 9.2 and 2.8 Hz,
H-5Ј), 7.87 (1H, d: J ϭ 2.6 Hz, H-3Ј); 8.12 (1H, m, NH); 1H
NMR (C5D5N) ␦ 0.88 (3H, s, 19-CH3), 1.12 (3H, s, 18-
CH3), 3.46 (2H, dd: J ϭ 6.7 and 13 Hz, CHaNH and CHb
NH), 3.91 (1H, m, H-3␣), 7.07 (1H, d: J ϭ 9.2 Hz, H-6Ј),
7.19 (1H, dd: J ϭ 9.2 and 2.7 Hz, H-5Ј), 7.93 (1H, d: J ϭ
2.7 Hz, H-3Ј), 8.53 (1H, m, NH); MS (LSIMSϩ) m/z (rel-
ative intensity %) 512 (MHϩ, 100). High-resolution MS
calculated for C28H42O4N5 (MHϩ): 512.3236; found:
512.3255.
1
1.03 (3H, s, 19-CH3), 3.37 and 3.54 (2H, 2 m, CH2NH),
1
6.87 (1H, m, NH); H NMR (C5D5N) ␦ 0.94 (3H, s, 19-
CH3), 1.13 (3H, s, 18-CH3), 3.67 and 3.89 (2H, dd: J ϭ 3.0
and 4.0 Hz and d, J ϭ 12.2 Hz, CHaNH and CHbNH), 7.67
(1H, m, NH); MS (LSIMSϩ) m/z (relative intensity %) 536
(MHϩ, 100), 518 (MHϩ-18, 9). High-resolution MS calcu-
lated for C26H33O4N5F3 (MHϩ): 536.2485; found:
536.2507.
2.29. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminomethyl]-5␣-androstane-3␤,17␤-diol (28)
The 3-oxo derivative 27 (35 mg, 0.065 mmol) was re-
duced with NaBH4, as described above for the synthesis of
the diol derivative 10, and the product was purified by TLC
on silica gel (Rf 0.30, petroleum ether/ethyl acetate 2:1) to
give the 17␣-[(N-5-azido-2-nitro-3,4,6-trifluorophenyl)ami-
nomethyl]-3␤,17␤-diol (19 mg, 0.035 mmol, 54%): m.p.
2.27. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminomethyl]-3,3-dimethoxy-5␣-androstan-17␤-ol (26)
A solution of 17␣-aminomethyl steroid 3 (50 mg,
0.137 mmol) in 3 ml of tetrahydrofuran was stirred for 8
days in the dark, in the presence of a solution of 4-azido-
1-nitro-2,3,5,6-tetrafluorobenzene [10] (50 mg, 0.212
mmol) in 1 ml of tetrahydrofuran and 1 ml of an ethanol/
diethyl ether/triethylamine 20:10:0.2 mixture. The reac-
tion mixture was evaporated under reduced pressure, and
the product was purified by TLC on silica gel (Rf 0.30,
petroleum ether/ethyl acetate 5:1 v/v) to give the 17␣-
(5-azido-2-nitro-3,4,6-trifluorophenyl)aminomethyl de-
rivative (37 mg, 0.064 mmol, 47%): m.p. 126.5–127.5°C,
recrystallized from diethyl ether; [␣]D ϭ Ϫ15.7° (EtOH);
156–157°C, recrystallized from ethyl acetate; [␣]D
Ϫ17.9° (EtOH); ␯ (CHCl3): 3690–3615 (OH and NH),
ϭ
max
2130 (N3), 1505 cmϪ1 (C-NO2); ␭
(⑀ ϭ 20 300), 300 nm (⑀ ϭ 8600), 421 nm (⑀ ϭ 3900); H
NMR (CDCl3) ␦ 0.82 (3H, s, 19-CH3), 0.89 (3H, s, 18-
CH3), 3.39 and 3.55 (2H, 2 m, CH2NH), 3.59 superimposed
(EtOH) ϭ 236 nm
max
1
1
to CH2NH m at 3.55 (1H, m, H-3), 6.85 (1H, m, NH); H
NMR (C5D5N) ␦ 0.85 (3H, s, 19-CH3), 1.13 (3H, s, 18-
CH3), 3.69 and 3.90 (2H, 2 m, CH2NH), 3.90 superimposed
to CH2NH m at 3.90 (1H, m, H-3␣), 7.68 (1H, m, NH); MS
(LSIMSϩ) m/z (relative intensity %) 538 (MHϩ, 100), 520
(MHϩ-18, 16). High-resolution MS calculated for
C26H35O4N5F3 (MHϩ): 538.2641; found: 538.2662.
␯
(CCl4): 3630–3380 (OH and NH), 2125 (N3), 1505
max
(C-NO2), 1105–1055 cmϪ1 (dimethoxy); ␭
(EtOH) ϭ
max
237 nm (⑀ ϭ 19 400), 301 nm (⑀ ϭ 8500), 420 nm (⑀ ϭ
3800); H NMR (CDCl3) ␦ 0.81 (3H, s, 19-CH3), 0.89
1
2.30. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminoethyl]-3,3-dimethoxy-5␣-androstan-17␤-ol (29)
(3H, s, 18-CH3), 3.13 and 3.19 (2 ϫ 3H, 2s, dimethoxy),
1
3.33 and 3.55 (2H, 2 m, CH2NH), 6.90 (1H, m, NH); H
The 17␣-aminoethyl steroid 5 (100 mg, 0.264 mmol)
was condensed with 4-azido-1-nitro-2,3,5,6-tetrafluoroben-
zene, as described above for the synthesis of the 17␣-
aminomethyl derivative 26, and the product was purified by
TLC on silica gel (Rf 0.60 chloroform/ethyl acetate 20:1
v/v) to give the 17␣-(5-azido-2-nitro-3,4,6-trifluorophe-
nyl)aminoethyl derivative (70 mg, 0.118 mmol, 45%): m.p.
NMR (C5D5N) ␦ 0.77 (3H, s, 19-CH3), 1.12 (3H, s,
18-CH3), 3.21 and 3.23 (2 ϫ 3H, 2s, dimethoxy), 3.71
and 3.93 (2H, 2 m, CH2NH), 7.72 (1H, m, NH); MS
(LSIMSϩ) m/z (relative intensity %) 582 (MHϩ, 31), 550
(MHϩ-32, 100). High-resolution MS calculated for
C28H39O5N5 F3 (MHϩ): 582.2903; found: 582.2938.

468
E. Mappus et al. / Steroids 65 (2000) 459–481
66–69°C, recrystallized from diethyl ether; [␣]D ϭ ϩ34.2°
(CCl4); ␯ (CHCl3): 3620–3345 (OH and NH), 2130
MS (LSIMSϩ) m/z (relative intensity %) 552 (MHϩ, 100),
534 (MHϩ-18, 26). High-resolution MS calculated for
C27H37O4N5F3 (MHϩ): 552.2797; found: 552.2811.
max
(N3), 1505 (C-NO2); 1105–1050 cmϪ1 (dimethoxy); ␭
max
(EtOH) ϭ 237 nm (⑀ ϭ 21 300), 298 nm (⑀ ϭ 8500), 419
1
nm (⑀ ϭ 4000); H NMR (CDCl3) ␦ 0.81 (3H, s, 19-CH3),
2.33. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
0.85 (3H, s, 18-CH3), 3.14 and 3.19 (2 ϫ 3H, 2s, dime-
thoxy), 3.63 (2H, 2 m, CH2NH), 6.94 (1H, m, NH); H
yl)aminopropyl]-3,3-dimethoxy-5␣-androstan-17␤-ol (32)
1
NMR (C5D5N) ␦ 0.78 (3H, s, 19-CH3), 1.08 (3H, s, 18-
CH3), 3.21 and 3.23 (2 ϫ 3H, 2s, dimethoxy), 3.79 and 3.90
(2H, 2 m, CH2NH), 7.75 (1H, m, NH); 19F NMR (CDCl3)
␦ 144.31 (1F, s broad, F-6), 146.83 (1F, dd: J ϭ 22.8 and
11.3 Hz, F-2), 163.06 (1F, d: J ϭ 24.0 Hz, F-3); MS
(LSIMSϩ) m/z (relative intensity %) 596 (MHϩ, 36), 578
(MHϩ-18, 6), 564 (MHϩ-32, 100), 546 (MHϩ-32–18, 25).
High-resolution MS calculated for C29H41O5N5F3 (MHϩ):
596.3060; found: 596.3069.
The 17␣-aminopropyl steroid 7 (100 mg, 0.254 mmol) was
condensed with 4-azido-1-nitro-2,3,5,6-tetrafluorobenzene, as
described above for the synthesis of 17␣-aminomethyl and
17␣-aminoethyl derivatives 26 and 29, and the product was
purified by TLC on silica gel (Rf 0.50, chloroform/ethyl ace-
tate 20:1 v/v) to give the 17␣-(5-azido-2-nitro-3,4,6-trifluoro-
phenyl)aminopropyl derivative (80 mg, 0.131 mmol, 52%):
m.p. 127–129°C, recrystallized from diethyl ether; [␣]D
Ϫ8.3°(CHCl3); ␯
ϭ
(CCl4): 3625–3340 (OH and NH), 2125
max
(N3), 1505 (C-NO2), 1105–1055 cmϪ1 (dimethoxy); ␭
max
(EtOH) ϭ 237 nm (⑀ ϭ 18 700), 301 nm (⑀ ϭ 8300), 417 nm
2.31. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminoethyl]-17␤-hydroxy-5␣-androstan-3-one (30)
1
(⑀ ϭ 3500); H NMR (CDCl3) ␦ 0.81 (3H, s, 19-CH3), 0.85
(3H, s, 18-CH3), 3.14 and 3.19 (2ϫ3H, 2s, dimethoxy), 3.44
(2H, m, CH2NH), 6.57 (1H, m, NH); 1H NMR (C5D5N) ␦ 0.79
(3H, s, 19-CH3), 1.09 (3H, s, 18-CH3), 3.20 and 3.23 (2 ϫ 3H,
2s, dimethoxy), 3.58 (2H, m, CH2NH), 7.68 (1H, m, NH); MS
(LSIMSϩ) m/z (relative intensity %) 610 (MHϩ, 25), 592
(MHϩ-18, 4), 578 (MHϩ-32, 100), 560 (MHϩ-18–32, 39).
High-resolution MS calculated for C30H43O5N5F3 (MHϩ):
610.3193; found: 610.3216.
The 3,3-dimethoxy derivative 29 (45 mg, 0.076 mmol) was
acidolyzed, as described above for the synthesis of the 3-oxo
derivative 9, and the product was purified by TLC on silica gel
(Rf 0.35, chloroform/ethyl acetate 20:1 v/v) to give the 17␣-
[(N-5-azido-2-nitro-3,4,6-trifluorophenyl)aminoethyl]-3-oxo
derivative (40 mg, 0.073 mmol, 96%): m.p. 125–127°C, re-
crystallized from diethyl ether; [␣]D ϭ ϩ23.8° (EtOH); ␯
max
(CHCl3): 3680–3610-3345 (OH and NH), 2130 (N3), 1705
(3-C ϭ O), 1505 cmϪ1 (C-NO2); ␭max (EtOH) ϭ 238 nm (⑀ ϭ
19 200), 299 nm (⑀ ϭ 7200), 420 nm (⑀ ϭ 3300); H NMR
2.34. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminopropyl]-17␤-hydroxy-5␣-androstan-3-one (33)
1
(CDCl3) ␦ 0.89 (3H, s, 18-CH3), 1.03 (3H, s, 19-CH3), 3.63
(2H, m, CH2NH), 6.95 (1H, m, NH); 1H NMR (C5D5N) ␦ 0.94
(3H, s, 19-CH3), 1.08 (3H, s, 18-CH3), 3.80 and 3.90 (2H, 2 m,
CH2NH), 7.74 (1H, m, NH); MS (LSIMSϩ) m/z (relative
intensity %) 550 (MHϩ, 100), 532 (MHϩ-18, 28). High-reso-
lution MS calculated for C27H35O4N5F3 (MHϩ): 550.2641;
found: 550.2649.
The 3,3-dimethoxy derivative 32 (54 mg, 0.088 mmol)
was acidolyzed, as described above for the synthesis of the
3-oxo derivative 12, and the product was purified by TLC
on silica gel (Rf 0.40, chloroform/ethyl acetate 20:1 v/v)
to give the 17␣-[(N-5-azido-2-nitro-3,4,6-trifluorophenyl)-
aminopropyl]-3-oxo derivative (48 mg, 0.085 mmol, 97%):
m.p. 53–54°C, recrystallized from diethyl ether; [␣]D
ϩ16.6° (EtOH); ␯
ϭ
2.32. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminoethyl]-5␣-androstane-3␤,17␤-diol (31)
(CHCl3): 3690–3610-3390 (OH and
max
NH), 2130 (N3), 1720 (C ϭ O), 1505 cmϪ1 (C-NO2); ␭
max
(EtOH) ϭ 237 nm (⑀ ϭ 17 000), 302 nm (⑀ ϭ 7300), 413
1
The 3-oxo derivative 30 (22 mg, 0.04 mmol) was reduced
with NaBH4, as described above for the synthesis of the diol
derivative 10, and the product was purified by TLC on silica
gel (Rf 0.35, chloroform/ethyl acetate 4:1 v/v) to give the
17␣-[(N-5-azido-2-nitro-3,4,6-trifluorophenyl)aminoethyl]-
3␤,17␤-diol (13 mg, 0.024 mmol, 59%): m.p. 97–100°C, re-
nm (⑀ ϭ 2850); H NMR (CDCl3) ␦ 0.88 (3H, s, 18-CH3),
1.03 (3H, s, 19-CH3), 3.45 (2H, m, CH2NH), 6.57 (1H, m,
1
NH); H NMR (C5D5N) ␦ 0.95 (3H, s, 19-CH3), 1.10 (3H,
s, 18-CH3), 3.58 (2H, m, CH2NH), 7.28 (1H, m, NH); MS
(LSIMSϩ) m/z (relative intensity %) 564 (MHϩ, 100), 546
(MHϩ-18, 27). High-resolution MS calculated for
C28H37O4N5F3 (MHϩ): 564.2797; found: 564.2784.
crystallized from diethyl ether; [␣]D ϭ ϩ47° (EtOH); ␯
max
(CCl4): 3625–3340 (OH and NH), 2125 (N3), 1505 cmϪ1
(C-NO2); ␭max (EtOH) ϭ 238 nm (⑀ ϭ 18 200), 302 nm (⑀ ϭ
6900), 431 nm (⑀ ϭ 7900); 1H NMR (CDCl3) ␦ 0.82 (3H, s,
19-CH3), 0.85 (3H, s, 18-CH3), 3.64 (2H, m, CH2NH), 3.59
superimposed to CH2NH m at 3.64 (1H, m, H-3␣), 6.95 (1H,
m, NH); 1H NMR (C5D5N) ␦ 0.87 (3H, s, 19-CH3), 1.08 (3H,
s, 18-CH3), 3.79 and 3.91 (2H, 2 m, CH2NH), 3.91 superim-
posed to CH2NH m at 3.91 (1H, m, H-3␣), 7.77 (1H, m, NH);
2.35. 17␣-[(N-5-Azido-2-nitro-3,4,6-trifluorophen-
yl)aminopropyl]-5␣-androstane-3␤,17␤-diol (34)
The 3-oxo derivative 33 (36 mg, 0.064 mmol) was reduced
with NaBH4, as described above for the synthesis of the diol
derivative 10, and the product was purified by TLC on silica
gel (Rf 0.35, chloroform/ethyl acetate 4:1 v/v) to give the

E. Mappus et al. / Steroids 65 (2000) 459–481
469
17␣-[(N-5-azido-2-nitro-3,4,6-trifluorophenyl)aminopropyl]-
3␤,17␤-diol (20 mg, 0.035 mmol, 55%): m.p. 110–111°C,
2.37. Competitive binding assays with androgen receptors
of rat ventral prostate
recrystallized from diethyl ether; [␣]D ϭ ϩ6.0° (EtOH); ␯
max
Aliquots of cytosol of prostates from rats castrated 24 h
before being sacrificed (200 ␮l in Tris⅐HCl 10 mM, pH 7.4, 1.5
mM EDTA, 10% glycerol, and 20 mM sodium molybdate,
containing 0.1 mM leupeptin and 0.1 mM bacitracin) in trip-
licate tubes were incubated in the dark with [1,2,4,5,6,7-
3H6]DHT (100 ␮l, 30 000 dpm, 0.3 nM) and seven concen-
trations (ranging from 1.25 to 125 nM) of radioinert steroid
competitors (DHT and 17␣-aminomethyl-, 17␣-aminoethyl-,
and 17␣-aminopropyl-ANB derivatives of DHT) in 300 ␮l of
the same buffer for 4 h at 4°C. Nonspecific binding was
estimated by incubation in the same conditions in the presence
of 625 nM of each of the radioinert competitors. Free and
bound steroid fractions were separated by dextran-coated char-
coal. For each competitor concentration, the radioactivity of
bound tracer (B) was expressed as percent of radioactivity
bound in the absence of competitor (B0), after substraction of
nonspecifically bound radioactivity.
(CHCl3): 3690–3610-3380 (OH and NH), 2130 (N3), 1510
cmϪ1 (C-NO2); ␭max (EtOH) ϭ 237 nm (⑀ ϭ 17 900), 302 nm
(⑀ ϭ 7400), 417 nm (⑀ ϭ 3400); 1H NMR (CDCl3) ␦ 0.82 (3H,
s, 19-CH3), 0.85 (3H, s, 18-CH3), 3.45 (2H, m, CH2NH), 3.59
(1H, m, H-3␣), 6.59 (1H, m, NH); 1H NMR (C5D5N) ␦ 0.87
(3H, s, 19-CH3), 1.10 (3H, s, 18-CH3), 3.58 (2H, m, CH2NH),
3.91 (1H, m, H-3␣), 7.30 (1H, m, NH); MS (LSIMSϩ) m/z
(relative intensity %) 566 (MHϩ, 100), 548 (MHϩ-18, 19).
High-resolution MS calculated for C28H39O4N5F3 (MHϩ):
566.2954; found: 566.2960.
2.36. Competitive binding assays with SHBG
Aliquots (100 ␮l) of purified human SHBG (5.5 nM in
10 mM phosphate buffer pH 7.4, 0.15 M NaCl, 0.1% gelatin
and 0.1% NaN3) in triplicate tubes were incubated in the
dark with [1,2,4,5,6,7-3H6]DHT (100 ␮l, 100 000 dpm, 1
nM, s.a. 114 Ci/mmol, purchased from Amersham) and
seven concentrations (ranging from 8.6 to 172 nM) of ra-
dioinert steroid competitors (DHT and 17␣-aminomethyl-,
17␣-aminoethyl-, and 17␣-aminopropyl-ANB derivatives
of DHT) in 300 ␮l of the same buffer for 1 h at 22°C. Free
and bound steroid fractions were separated by dextran-
coated charcoal (prepared from 5 g of Norit A and 0.5 g of
Dextran T-70 in 1 liter of 10 mM phosphate buffer, pH 7.4,
0.15 M NaCl). For each competitor concentration, the ra-
dioactivity of bound tracer (B) was expressed as percent of
radioactivity bound in the absence of competitor (B0).
3. Results and discussion
In this study, the synthetic route chosen for the access to
17␣-aminoalkyl derivatives of 5␣-dihydrotestosterone
(Scheme 1) was established from 5␣-androstane-3,17-di-
one, which was converted to a 3,3-(dimethoxy)-5␣-andro-
stan-17-one precursor 1. The choice of this derivative rather
than that of the corresponding 3,3-ethylenedioxy analog,
employed in a previous synthesis of 17␣-hexanoic deriva-
tives [1], was made with the view of preparing radioactive
Scheme 1.

470
E. Mappus et al. / Steroids 65 (2000) 459–481
Scheme 2.
analogs from commercially available radiolabeled 3-oxos-
teroid precursors without the difficulty of performing a
microscale dioxolanation reaction that requires a device for
removing water formed in the reaction. The formation of a
dimethoxyacetal can be easily performed with high selec-
tivity for the 3-oxo group of 5␣-androstane-3,17-dione by
refluxing in methanol in the presence of p-toluenesulfonic
acid catalyst [11]. Reduction of the 17-ketone with sodium
borohydride led to the 3,3-(dimethoxy)-5␣-androstan-
17␤-ol derivative prepared also directly from DHT [9] (data
not shown).
Condensation of the 17-ketone with the dimethylsulfo-
nium methylide reagent [12] generated from trimethylsul-
fonium iodide and potassium tert-butoxide in dimethylfor-
mamide [13] led to the formation of the 17␤-spirooxirane 2
as the major product [14], in 90% yield, while only traces of
the 17␣-isomer [15] could be detected.
The 17␤-spirooxirane 2 was transformed by nucleophilic
opening into either the 17␣-aminomethyl-17␤-hydroxy in-
termediate 3, obtained in 72% yield after treatment with
concentrated ammonium hydroxide in the presence of a low
amount of acetic acid [16], or the 17␣-cyanomethyl-17␤-
hydroxy intermediate 4, obtained in nearly quantitative
yield after reaction with potassium cyanide in ethanol [16].
The 17␣-cyanomethyl-17␤-hydroxy compound 4 was con-
verted into the 17␣-aminoethyl-17␤-hydroxy intermediate 5
in 60% yield, by reduction with lithium-aluminum hydride
in tetrahydrofuran, at room temperature. No noticeable for-
mation of unsubstituted 17␤-hydroxy side-product could be
detected in this reaction despite the thermal unstability of
17␣-cyanomethyl-17␤-hydroxy derivatives, which may re-
generate the 17-ketone in basic conditions [10,16]. The
17␤-spirooxirane 2 was also transformed into the 17␣-
cyanoethyl-17␤-hydroxy intermediate 6 in 92% yield, by
direct opening with the cyanoethyl anion generated from
acetonitrile at Ϫ76°C in the presence of butyllithium and
diisopropylamine. The 17␣-cyanoethyl-17␤-hydroxy com-
pound 6 was then reduced with lithium-aluminum hydride
in tetrahydrofuran, at room temperature, to the 17␣-amin-
opropyl-17␤-hydroxy intermediate 7 in 69% yield. This
synthetic scheme provides a more straightforward access to
17␣-aminomethyl, 17␣-aminoethyl, and 17␣-aminopropyl
spacer arms than other methods previously reported for
estradiol derivatives, using transformations of 17␣-cyano
and 17␣-ethynyl precursors [17]. The reduction of 17␣-
cyano intermediates had also been employed in this labo-
ratory as an alternative method for an earlier synthesis of
17␣-aminomethyl derivatives of both DHT and testosterone
series [3].
The three 17␣-aminomethyl-, 17␣-aminoethyl-, and
17␣-aminopropyl-17␤-hydroxy intermediates 3, 5, and 7
(Scheme 1) were either acylated with the commercially
available N-(5-azido-2-nitrobenzoyloxy)succinimide re-
agent to give 17␣-(5-azido-2-nitrobenzoyl)amidoalkyl
(ANB) derivatives 8, 11, and 14, in 72, 66, and 56% yields,
respectively, or condensed with 4-azido-1-fluoro-2-nitro-
benzene to give 17␣-(4-azido-2-nitrophenyl)aminoalkyl
(ANP) derivatives 17, 20, and 23, in 91, 33, and 47% yields,
respectively, while a similar condensation with 4-azido-1-
nitro-2,3,5,6-tetrafluorobenzene [10] led to the correspond-
ing 17␣-(5-azido-2-nitro-3,4,6-trifluorophenyl)aminoalkyl
(ANF) derivatives 26, 29, and 32, in 89, 45, and 52% yields,
respectively. In all cases, higher yields were obtained for
17␣-aminomethyl derivatives than for the aminoethyl and
aminopropyl analogs. The 3,3-dimethoxy protecting groups
of each of these compounds was then removed by mild
acidolysis (1% HCl in 10% aqueous dioxane), monitored by
thin-layer chromatography, to regenerate the corresponding
3-oxosteroids, 9, 12, and 15, for ANB derivatives, 18, 21,
and 24, for ANP derivatives, and 27, 30, and 33, for ANF

E. Mappus et al. / Steroids 65 (2000) 459–481
471
Table 1
13C NMR data (CDCl3) for 5␣-androstane-3␤,17␤-diol (A-diol) and 17␤-hydroxy (3-MeO-DHT), 17-oxo (1), 17␤-spirooxirane (2), 17␣-cyanomethyl (4)
and 17␣-cyanoethyl (6), 17␣-aminomethyl (3), 17␣-aminoethyl (5), and 17␣-aminopropyl (7) derivatives of 3,3-dimethoxy-5␣-androstane
A-diol
3-MeO-DHT
1
2
4
6
3
5
7
Carbon
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
37.04
31.52
71.31
38.19
44.92
28.58
31.63
35.56
54.49
35.58
20.84
36.75
43.00
51.02
23.39
30.55
81.98
11.14
12.36
35.06
28.33
100.38
35.55
42.50
28.33
31.54
35.53
54.20
35.83
20.78
36.76
43.01
51.02
23.40
30.53
81.95
11.15
11.66
47.50
47.46
34.97
28.27
100.25
35.49
42.43
28.14
30.81
35.04
54.14
35.89
20.44
31.55
47.81
51.42
21.78
35.85
221.38
13.82
11.62
47.51
47.45
35.04
28.30
100.33
35.50
42.47
28.30
31.42
35.62
54.11
35.83
20.55
29.05
40.16
52.82
23.53
33.98
70.54
14.38
11.62
47.50
47.45
53.63
35.01
28.30
100.28
35.45
42.42
28.19
31.61
36.28
53.87
35.79
20.70
31.90
46.36
51.03
23.40
37.19
81.86
14.09
11.63
47.52
47.46
35.04
28.32a
100.33
35.50
42.44
28.27a
31.62a
36.36
53.92
35.81
20.79
31.66a
46.57
50.55
23.55
34.78
82.68
14.26
11.64
47.52
47.47
35.07
28.35
100.37
35.50
42.49
28.35
31.76
36.12
54.10
35.81
20.82
32.61
45.66
51.00
23.54
34.80b
81.65b
14.52
11.63
47.49
47.44
35.09
28.35
100.41
35.52
42.54
28.41
31.82
36.24
54.22
35.83
20.86
31.82
46.47
50.27
23.88
35.01
84.16b
14.00
11.65
47.50
47.45
35.08
28.34
100.40
35.51
42.51
28.39
31.64
36.34
54.14
35.81
20.89
31.83
46.53
50.53
23.86
34.50
82.22
14.62
11.65
47.49
47.44
17
18
19
CH3O
CH3O
CCH2O
C§N
CH2CN
CH2CH2CN
CH2NH
CH2CH2NH
CH2CH2CH2
118.78
28.14
121.02
12.08
32.69
47.44b
38.18b
36.17b
42.26b
34.88b
26.83b
a Assignments may be exchanged.
b Chemical shifts corresponding to the free amine [3].
derivatives, all obtained in high yields (84–97%) except for
the 17␣-aminomethyl ANP derivative 18 recovered only in
50% yield owing to losses during chromatographic purifi-
cations. These 3-oxosteroids were then reduced in mild
conditions using approximately a two-molar excess of a 2
mM solution of sodium borohydride in methanol (prepared
extemporaneously) to generate the corresponding 3␤-hy-
droxy steroids, 10, 13, and 16, for ANB derivatives, 19, 22,
and 25, for ANP derivatives, and 28, 31, and 34, for ANF
derivatives, most of them isolated in moderate yields (42–
66%) except for 17␣-aminoethyl and 17␣-aminopropyl
ANB derivatives 13 and 16, obtained in higher yields (79
and 90%, respectively). Using these mild conditions and
careful monitoring by thin-layer chromatography, this re-
duction step did not alter the structures of the three aryl
azide chromophores, as confirmed by spectrometric charac-
terizations and by comparison with reference samples ob-
tained without reduction step, via direct coupling of the aryl
azide chromophores to 17␣-aminoalkyl derivatives of 5␣-
androstane-3␤,17␤-diol (data not shown).
unsuccessful. Despite several modifications aimed at
optimizing experimental conditions, the formation of the
17␤-spirooxirane from the 3,3-(dimethoxy)-5␣-[16-2H2]-
androstan-17-one precursor obtained from the nondeuter-
ated analog 1 by exchange in the presence of deuterated
sodium methylate, as previously reported for 3,3-(ethyl-
enedioxy)-5␣-androstan-17-one [3], failed to maintain a
significant amount of deuterium at the 16-position (data not
shown).
The stereoselectivity of the conversion of a 17-oxo ste-
roid to a 17␤-2Ј-spirooxirane has been verified by compar-
ing the 1H and 13C NMR spectra of 17␣-aminomethyl
derivatives, obtained in this study by direct ammonolysis,
with those of 17␣-aminomethyl derivatives previously syn-
thesized by reduction of 17␣-cyanohydrins of established
configurations [3]. Moreover, nucleophilic opening of a
17␤-2Ј-spirooxirane with either cyanide or cyanoethyl an-
ions and subsequent reduction to ethyl and propylamine
derivatives cannot be expected to generate 17␤-aminoalkyl-
17␣-hydroxy intermediates.
1
Attempts were made to synthesize the 17␣-aminoalkyl
analogs deuterated at the 16 position, useful as reference
compounds for 13C NMR characterizations, but remained
The assignments of the H NMR signals for 18- and
19-methyl protons of 17␣-substituted derivatives of the
three 3,3-dimethoxy-, 3-oxo-, and 3␤-hydroxy-5␣-andro-

472
E. Mappus et al. / Steroids 65 (2000) 459–481
Table 2
13C NMR data (C5D5N) for 5␣-androstane-3␤,17␤-diol (A-diol) and 17␤-hydroxy (3-MeO-DHT), 17-oxo (1), 17␤-spirooxirane (2), 17␣-cyanomethyl
(4) and 17␣-cyanoethyl (6), 17␣-aminomethyl (3), 17␣-aminoethyl (5), and 17␣-aminopropyl (7) derivatives of 3,3-dimethoxy-5␣-androstane
A-diol
3-MeO-
DHT
1
2
4
6
3
5
7
Carbon
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
37.61a
32.49
70.61
39.34
45.33
29.14
32.04
35.88
54.90
35.93
21.30
37.58a
43.60
51.42
23.85
31.00
81.41
11.94
12.57
35.41
28.88
100.49
35.94
42.73
28.68
32.02
35.84
54.56
36.06
21.25
37.57
43.61
51.43
23.85
31.02
81.40
11.93
11.83
47.32
47.27
35.26
28.81
100.40
35.90
42.59
28.41
31.08a
35.11
54.30
36.06
20.73
32.18a
47.83
51.48
21.94
35.87
219.71
13.85
11.68
47.33
47.27
35.35
28.85
100.44
35.90
42.67
28.57
31.72
35.75
54.27
36.03
20.85
29.26
40.40
52.80
23.73
34.19
70.37
14.69
11.73
47.33
47.27
53.24
35.32
28.87
100.47
35.88
42.64
28.59
32.05
36.54
53.99
36.00
21.15
32.35
47.01
50.97
23.78
36.89
81.47
14.90
11.76
47.35
47.28
35.40
28.91
100.52
35.93
42.69
28.66
32.17
36.58
54.12
36.03
21.26
32.17
47.14
50.96
24.00
34.27
81.93
15.19
11.81
47.35
47.29
35.36
28.90
100.53
35.92
42.65
28.70
32.17
36.45
54.22
36.04
21.21
32.55
46.16
51.26
24.05
34.28b
82.40b
15.25
11.80
47.34
47.28
35.39
28.90
100.54
35.94
42.69
28.72
32.23
36.58
54.29
36.05
21.27
32.28
47.08
50.68
24.23
35.31
83.57b
14.87
11.82
47.34
47.28
35.42
28.92
100.55
35.94
42.70
28.73
32.26
36.66
54.32
36.05
21.35
32.26
47.17
51.01
24.27
34.75
82.30
15.47
11.85
47.34
47.28
17
18
19
CH3O
CH3O
CCH2O
C§N
CH2CN
CH2CH2CN
CH2NH
CH2CH2NH
CH2CH2CH2
120.45
28.87
122.12
12.42
33.97
48.24b
38.40b
37.20b
43.45b
35.31b
28.20b
a Assignments may be exchanged.
b Chemical shifts corresponding to the free amine [3].
stan-17␤-ol series (see Section 2), were made on the as-
sumption that modifications of 3- or 17-substituents should
influence predominantly the chemical shifts of 19- or 18-
methyl protons, respectively, and were confirmed by corre-
lation with those previously established for 17␣-amino-
methyl derivatives of 3-oxo-5␣-androstan-17␤-ol, 3,3-
(ethylenedioxy)androstan-17␤-ol, and 3-oxoandrost-4-en-
17␤-ol [3].
CDCl3, separated only by 0.01–0.04 ppm. However, a more
significant difference was found for the corresponding ANP
derivatives, thus leading to an unambiguous assignment of
the more shielded signals for the 19-methyl groups of these
compounds, which could then be extended to ANB and
ANF analogs. On the other hand, 18- and 19-methyl signals
in C5D5N were easily differentiated, owing to the strong
deshielding effect on the 18-methyl signals at ϳ 1.1 ppm,
resulting from the presence of a vicinal 17␣-hydroxy group.
A more detailed inspection of these results reveals that
17␣-aminomethyl side-chains exert stronger deshielding ef-
fects on 18-methyl signals, in both CDCl3 and C5D5N
solvents, than the two other aminoethyl and aminopropyl
derivatives, while the magnitude of these effects is slightly
larger for ANP derivatives as compared with the two other
ANB and ANF chromophores. The more pronounced
deshielding effect on the 18-methyl signal of the shorter
17␣-aminomethyl link despite insignificant or very slight
downfield shifts increments observed when changing a 17␣-
methyl substituent for larger 17␣-ethyl or 17␣-[5Ј-(me-
thoxycarbonyl)pentyl] groups [1] points out to the hypoth-
esis of a predominant anisotropic effect of the NH group,
1
The H NMR chemical shifts, in CDCl3, of 19-methyl
protons of 3-oxo derivatives (1.02/1.04 ppm) were signifi-
cantly deshielded as compared with those of 3,3-dimethoxy
and 3␤-hydroxy derivatives (0.80/0.82 and 0.82/0.84 ppm,
respectively), while a small deshielding effect of ϩ0.03/
ϩ0.04 ppm was also exerted on the 18-methyl signals as
compared to the chemical shifts observed for 3␤-hydroxy-
steroid derivatives, in agreement with reported substituent-
induced effects for these two groups [18]. In C5D5N, similar
but weaker effects were also often found. The ANB conju-
gates of both 17␣-aminoethyl and 17␣-aminopropyl side-
chains of 3,3-dimethoxy- and 3␤-hydroxy-5␣-androstan-
17␤-ol derivatives as well as the corresponding ANF
analogs, showed very close 18- and 19-methyl signals in

E. Mappus et al. / Steroids 65 (2000) 459–481
473
Table 3
13C NMR data (CDCl3) for 17␣-(N-5-azido-2-nitrobenzoyl)amidomethyl, ethyl, and propyl derivatives of 3,3-dimethoxy- (8, 11, 14), 3-oxo- (9, 12, 15),
and 3␤-hydroxy-5␣-androstan-17␤-ol (10, 13, 16)
8
11
14
9
12
15
10
13
16
Carbon
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
35.03
28.34
100.31
35.50
42.40
28.34
31.57
36.35
53.79
35.81
20.76
31.73
45.88
50.65
23.59
34.61
83.44
14.08
11.63
47.50
47.50
46.12
35.03
28.31a
100.34
35.51
42.44
28.30a
31.66
36.42
53.92
35.81
20.78
31.47
46.70
50.29
23.60
34.61
84.81
13.98
11.64
47.51
47.47
36.84
33.96
35.03
28.33
100.37
35.52
42.46
28.33
31.70
36.41
53.96
35.81
20.82
31.52
46.52
50.44
23.64
34.73
83.64
14.34
11.65
47.51
47.46
40.95
33.32
23.43
165.70
135.62
142.31
126.82
119.10
146.22
119.96
38.45
38.10
211.96
44.61
46.53
28.76
31.28
36.22
53.52
35.69
21.03
31.64
45.87
50.48
23.60
34.48
83.30
14.12
11.46
38.51
38.13
211.96
44.64
46.65
28.79
31.39
36.33
53.64
35.73
21.05
31.39
46.65
50.13
23.61
34.63
84.73
14.00
11.49
38.53
38.16
212.12
44.67
46.67
28.83
31.42
36.29
53.64
35.72
21.09
31.42
46.51
50.22
23.64
34.63
83.56
14.37
11.48
36.99
31.43
71.17
38.09
44.80
28.53
31.63
36.34
54.09
35.54
20.83
31.72
45.85
50.63
23.58
34.60
83.39
14.09
12.31
36.98
31.46
71.24
38.11
44.85
28.55
31.73
36.42
54.18
35.55
20.83
31.46
46.68
50.26
23.59
34.65
84.90
13.96
12.33
36.99
31.50a
71.26
38.13
44.86
28.60
31.78
36.41
54.21
35.55
20.87
31.49a
46.50
50.40
23.64
34.74
83.62
14.34
12.33
17
18
19
CH3O
CH3O
CH2NH
CH2CH2
CH2CH2CH2
NHCO
1Ј
2Ј
3Ј
4Ј
46.18
36.82
34.00
40.90
33.30
23.51
165.74
135.59
142.25
126.83
119.09
146.25
119.99
46.12
36.84
33.88
40.92
33.29
23.47
165.71
135.63
142.28
126.83
119.09
146.23
119.97
166.14
135.53
142.20
126.91
119.03
146.28
119.99
165.40
135.78
142.30
126.80
119.07
146.22
119.87
166.22
135.48
142.16
126.89
119.03
146.28
119.98
165.40
135.78
142.30
126.83
119.08
146.22
119.92
166.14
135.53
142.19
126.92
119.00
146.28
119.98
165.40
135.79
142.27
126.82
119.05
146.23
119.88
5Ј
6Ј
a Assignments may be exchanged.
although other factors, such as hydrogen bonding, hydro-
phobicity, and conformational effects, cannot be ruled out
[19].
ments (vide infra). Therefore, the more deshielded of these
two signals was assigned to C7, in CDCl3, and to C2, in
C5D5N. The rather close resonances of the two pairs of
carbon atoms, C1/C4 of 3,3-dimethoxy-5␣-androstan-17␤-
ol, in both CDCl3 and C5D5N solvents, and C2/C6, in
C5D5N, could be differentiated only by correlation with
those established by 1H, 13C 2D NMR for 17␣-aminoalkyl-
3,3-dimethoxy-5␣-androstan-17␤-ol derivatives (vide in-
fra).
The assignments of 13C NMR resonances for carbon
atoms of the two unsubstituted 17␤-hydroxy reference com-
pounds, 5␣-androstane-3␤,17␤-diol (A-diol) and 3,3-(dime-
thoxy)-5␣-androstan-17␤-diol (3-MeO-DHT), and of the
3,3-(dimethoxy)-5␣-androstan-17-one precursor 1, in both
CDCl3 and C5D5N solvents (Tables 1 and 2), were based on
data reported in previous studies [1,3,20] for 5␣-androstane-
3␤,17␤-diol, DHT, 3,3-(ethylenedioxy)-5␣-androstan-17␤-
ol, 3,3-(ethylenedioxy)-5␣-androstan-17-one, and 3␤-hy-
droxy-5␣-androstan-17-one. The rather close resonances of
C2 and C7 carbon atoms of 5␣-androstane-3␤,17␤-diol,
especially in CDCl3, could be assigned on the assumption
that changes in the structure of 17␣-side-chains should
affect more strongly the signals of C7 carbon atoms than
those of C2 carbon atoms, as confirmed by similar effects
observed for 17␣-aminoalkyl-3,3-dimethoxy-5␣-androstane
derivatives, unambiguously characterized in this study by
1H, 13C heteronuclear 2D NMR HSQC-TOCSY experi-
The assignments for the resonances of the C1 to C10, and
C19 carbon atoms of rings A and B of the steroid skeletons
of 17␣-substituted 3,3-dimethoxy-, 3-oxo-, and 3␤-hy-
droxy-5␣-androstan-17␤-ol derivatives in both CDCl3 and
C5D5N solvents (Tables 1–8), were based on data men-
tioned above for the corresponding unsubstituted 17␤-hy-
droxy derivatives, while assignments for close signals of
methylene carbon atoms (i.e. C2/C6 or C7/C12) were cor-
1
roborated by H, 13C 2D NMR. The assignments for the
resonances of the carbon atoms of rings C and D were made
by correlation with those previously established for both
unsubstituted and 17␣-substituted derivatives of the 3,3-

474
E. Mappus et al. / Steroids 65 (2000) 459–481
Table 4
13C NMR data (C5D5N) for 17␣-(N-5-azido-2-nitrobenzoyl)amidomethyl, ethyl, and propyl derivatives of 3,3-dimethoxy- (8, 11, 14), 3-oxo- (9, 12, 15),
and 3␤-hydroxy-5␣-androstan-17␤-ol (10, 13, 16)
8
11
14
9
12
15
10
13
16
Carbon
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
35.37
28.95
100.56
35.90
42.50
28.65
31.95
36.57
53.96
36.00
21.24
32.16
46.71
51.45
24.16
34.01
83.04
15.05
11.79
47.35
47.32
47.67
35.33
28.91
100.54
35.91
42.58
28.69
32.11
36.64
54.05
36.02
21.29
32.40
47.34
51.04
24.22
34.85
82.54
15.18
11.80
47.34
47.30
37.44
37.02
35.41
28.91
100.55
35.95
42.71
28.73
32.25
36.68
54.25
36.05
21.34
32.25
47.28
50.97
24.20
34.59
82.66
15.42
11.84
47.35
47.28
41.49
35.07
24.75
166.02
136.89
143.65
126.81
120.06
145.99
120.15
38.70
38.37
210.40
44.90
46.76
29.08
31.57
36.43
53.57
35.93
21.40
32.03
46.66
51.20
24.18
33.97
83.03
15.02
11.34
38.73
38.38
210.41
44.92
46.81
29.10
31.71
36.48
53.79
35.92
21.47
32.33
47.30
50.87
24.22
34.83
82.51
15.16
11.36
38.81
38.41
210.45
44.94
46.92
29.15
31.82
36.53
53.89
35.96
21.51
32.15
47.18
50.72
24.21
34.52
82.63
15.40
11.39
37.57
32.58
70.65
39.39
45.22
29.16
32.06
36.66
54.30
35.94
21.29
32.17
46.70
51.46
24.21
34.01
83.06
15.07
12.56
37.57
32.58
70.66
39.41
45.26
29.18
32.19
36.71
54.48
35.93
21.36
32.44
47.33
51.10
24.26
34.86
82.57
15.20
12.57
37.65
32.59
70.66
39.42
45.38
29.24
32.29
36.77
54.61
35.98
21.40
32.29
47.22
50.98
24.24
34.57
82.67
15.43
12.61
17
18
19
CH3O
CH3O
CH2NH
CH2CH2
CH2CH2CH2
NHCO
1Ј
2Ј
3Ј
4Ј
47.61
37.42
36.98
41.46
35.00
24.76
166.05
136.89
143.65
126.83
120.05
145.96
120.18
47.65
37.45
37.00
41.48
35.03
24.76
166.04
136.93
143.61
126.81
120.06
145.95
120.16
166.93
137.01
143.58
126.87
120.16
145.95
120.16
166.09
136.91
143.65
126.86
120.04
146.03
120.19
166.95
136.99
143.56
126.88
120.17
145.96
120.17
166.06
136.87
143.65
126.87
120.02
146.03
120.21
166.94
137.03
143.56
126.88
120.14
145.96
120.17
166.06
136.90
143.67
126.87
120.03
146.03
120.20
5Ј
6Ј
(ethylenedioxy)-, 3-oxo-, and 3␤-hydroxy-5␣-androstan-
17␤-ol series [1,3]. This correlation method led to the iden-
tification of carbon atoms of the steroid skeletons of the
17␤-spirooxirane 2, of 17␣-cyanoalkyl derivatives 4 and 6,
and of aminoalkyl precursors 3, 5, and 7 (Tables 1 and 2),
while the remaining signals were assigned to carbon atoms
derivatives [3]. This correlation could then be extended to
confirm assignments for carbon atoms of rings C and D of
compounds 8 and 10 in the two other 3,3-dimethoxy and
3␤-hydroxy series, as well as for those of the aminomethyl
analogs in the ANP and ANF series 17, 18, 19 (Tables 5 and
6), and 26, 27, 28 (Tables 7 and 8), respectively. However,
the extension of these correlation methods for identifying
carbon atom signals of the steroid skeletons of the ANB,
ANP, and ANF derivatives in the two homologous 17␣-
aminoethyl and 17␣-aminopropyl series (Tables 3–8) led to
several ambiguities that could be resolved only by use of
1H, 13C heteronuclear 2D NMR correlation techniques.
Two-dimensional NMR experiments using in parallel
HSQC and HSQC-TOCSY techniques were performed in
both CDCl3 and C5D5N solvents for the ANB and ANP
derivatives of 17␣-aminoethyl- and 17␣-aminopropyl-3,3-
dimethoxy-5␣-androstan-17␤-ol 11, 20, and 14, 23, respec-
tively. For HSQC spectra, assignments for singlet signals of
the carbon atoms of C18, C19, and 3,3-dimethoxy methyl
groups were easily deduced from the corresponding 1H
NMR spectra, while those for the singlet signals of methine
C5, C8, C9, and C14 carbon atoms were made by a com-
1
of the 17␣-side-chain and differentiated by H, 13C hetero-
nuclear 2D NMR HSQC and HMBC techniques for the
17␣-cyanoethyl derivative 6 and the 17␣-aminoethyl pre-
cursor 5, and by comparing with ANB, ANP, and ANF
analogs for the 17␣-aminoethyl and 17␣-aminopropyl pre-
cursors 5 and 7 (vide infra). Assignments for the carbon
atoms of 17␣-n-propylamine side-chain of compound 7
were also confirmed by correlation with those previously
established by 2D NMR for 17␣-n-propyl and 17␣-n-propyl
alcohol derivatives of estradiol [21], using known substitu-
ent-induced effects of free amine groups on n-alkanes [22].
The assignments for the carbon atoms of the ANB de-
rivative of 17␣-aminomethyl-3-oxo-5␣-androstan-17␤-ol 9
(Tables 3 and 4) were made more precisely by correlation
with those previously reported for the corresponding acet-
amido or hemiglutaramido (either as acid or methyl ester)

E. Mappus et al. / Steroids 65 (2000) 459–481
475
Table 5
13C NMR data (CDCl3) for 17␣-(N-4-azido-2-nitrophenyl)aminomethyl, ethyl, and propyl derivatives of 3,3-dimethoxy- (17, 20, 23), 3-oxo- (18, 21,
24), and 3␤-hydroxy-5␣-androstan-17␤-ol (19, 22, 25)
17
20
23
18
21
24
19
22
25
Carbon
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
35.03
28.24
100.32
35.47
42.47
28.24
31.71
36.31
53.99
35.82
20.80
32.34
46.09
51.32
23.53
35.82
83.18
14.13
11.64
47.53
47.46
49.35
35.03
28.29a
100.32
35.49
42.45
28.28a
31.69
36.40
53.97
35.80
20.79
31.49
46.69
50.39
23.61
34.78
83.48
14.02
11.65
47.52
47.47
39.59
34.99
35.05
28.33
100.36
35.50
42.47
28.33
31.73
36.40
54.03
35.82
20.85
31.68
46.54
50.60
23.67
35.05
83.42
14.37
11.65
47.52
47.46
44.14
33.81
23.54
143.41
131.39
115.60
127.59
128.23
115.97
38.54
38.11
211.90
44.61
46.72
28.72
31.41
36.20
53.67
35.77
21.05
32.26
46.07
51.11
23.54
35.77
83.10
14.15
11.48
38.54
38.12
211.90
44.64
46.69
28.77
31.43
36.32
53.71
35.74
21.06
31.43
46.69
50.23
23.63
34.77
83.43
14.03
11.50
38.55
38.14
212.01
44.65
46.71
28.80
31.44
36.28
53.72
35.73
21.10
31.59
46.52
50.39
23.66
34.99
83.36
14.37
11.48
37.01
31.46
71.24
38.08
44.91
28.50
31.79
36.33
54.25
35.58
20.86
32.33
46.07
51.30
23.54
35.77
83.17
14.14
12.33
37.00
31.50
71.23
38.10
44.87
28.54
31.78
36.41
54.25
35.56
20.85
31.46
46.67
50.38
23.61
34.78
83.43
14.02
12.33
37.01
31.48
71.26
38.13
44.88
28.57
31.80
36.39
54.27
35.56
20.89
31.67
46.51
50.56
23.66
35.03
83.40
14.36
12.34
17
18
19
CH3O
CH3O
CH2NH
CH2CH2
CH2CH2CH2
1Ј
2Ј
3Ј
4Ј
5Ј
6Ј
49.34
39.56
35.02
44.08
33.79
23.48
143.37
131.41
115.57
127.61
128.23
115.94
49.35
39.57
35.03
44.11
33.79
23.48
143.40
131.34
115.59
127.58
128.25
115.94
143.66
131.68
115.61
127.52
128.12
116.06
143.41
131.37
115.52
127.33
128.19
115.96
143.61
131.67
115.61
127.58
128.12
116.03
143.45
131.37
115.50
127.34
128.19
115.98
143.68
131.67
115.62
127.51
128.10
116.05
143.41
131.36
115.53
127.35
128.20
115.95
a Assignments may be exchanged.
bination of results of DEPT experiments and of correlations
with the corresponding 13C NMR signals previously re-
ported for similar 3,3-(ethylenedioxy)-5␣-androstan-17␤-ol
reference structures [1]. The resonances for the remaining
methylene groups of the steroid skeleton carbon atoms C1,
C2, C4, C6, C7, C11, C12, C15, and C16 which are shared
by all 17 substituted derivatives, and thus easily differenti-
ated from those of the 17␣-aminoethyl and aminopropyl
The assignments for the carbon atoms of the 17␣-amino-
methyl, 17␣-aminoethyl, and 17␣-aminopropyl side-chains of
ANB, ANP, and ANF derivatives were straigthtforward for the
CH2N carbon atom, which is either the only methylene carbon
atom of aminomethyl side-chains, or was also readily identi-
fied as the more deshielded signal, in the two other cases, as
1
confirmed by H, 13C 2D NMR correlation with the corre-
sponding deshielded NH proton signal of ANB and ANP
derivatives. Identification of the CH2-1Ј methylene groups of
aminoethyl side-chains and of the CH2-2Ј methylene groups of
aminopropyl side-chains was established by HSQC-TOCSY
from their coupling with the vicinal CH2N methylene carbon
atom, while the residual CH2-1Ј methylene groups of amino-
propyl side-chains were identified from their strong coupling
with the vicinal CH2-2Ј methylene group and their weak cou-
pling with the CH2N methylene carbon atom. The signals for
these side-chain carbon atoms appeared as doublets, or, in
some instances, as broad singlets, as found for the 2-nitro-4-
azido-phenylaminopropyl derivative 23, in both CDCl3 and
C5D5N solvents, and for the two other CH2-1Ј and CH2-2Ј
groups of the 2-nitro-5-azidobenzoylamidopropyl derivative
14, in CDCl3 only.
1
side-chains, were assigned using the H, 13C heteronuclear
HSQC-TOCSY correlation technique via analysis of the
hydrogen-relayed transfer of magnetization through coupled
spins along vicinal carbon atom sequences. The signals for
C4, C6, C7, C11, and C15 carbon atoms were readily
identified from their coupling with unequivocally character-
ized vicinal methine groups, while those for C12 and C16
carbon atoms were established from their coupling with
previously assigned CH2-11 and CH2-15 methylene carbon
atoms. Assignments for the signals of C1 and C2 carbon
atoms were mainly supported by correlation with the cor-
responding 13C chemical shifts of 3,3-(ethylenedioxy)-5␣-
androstan-17␤-ol reference structure [1] and confirmed by a
weak coupling between C2 and C4 signals in C5D5N.

476
E. Mappus et al. / Steroids 65 (2000) 459–481
Table 6
13C NMR data (C5D5N) for 17␣-(N-4-azido-2-nitrophenyl)aminomethyl, ethyl, and propyl derivatives of 3,3-dimethoxy- (17, 20, 23), 3-oxo- (18, 21,
24), and 3␤-hydroxy-5␣-androstan-17␤-ol (19, 22, 25)
17
20
23
18
21
24
19
22
25
Carbon
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
35.36
28.90
100.53
35.89
42.62
28.61
32.10
36.49
54.06
36.01
21.28
32.86
46.71
51.39
24.03
35.52
82.41
15.07
11.78
47.36
47.29
50.51
35.40
28.91
100.54
35.93
42.69
28.69
32.20
36.64
54.18
36.06
21.28
32.06
47.36
50.87
24.12
34.22
82.94
15.06
11.83
47.36
47.29
40.52
35.81
35.43
28.92
100.55
35.93
42.72
28.72
32.27
36.67
54.31
36.07
21.35
32.34
47.24
51.13
24.21
34.87
82.67
15.40
11.84
47.36
47.30
44.50
34.73
24.35
143.89
131.61
116.13
127.32
128.48
116.62
38.78
38.36
210.38
44.89
46.90
29.04
31.70
36.37
53.74
35.95
21.47
32.79
46.68
51.14
24.06
35.55
82.40
15.04
11.35
38.78
38.38
210.35
44.92
46.89
29.11
31.79
36.51
53.86
35.97
21.47
31.97
47.33
50.64
24.13
34.20
82.93
15.02
11.39
38.81
38.39
210.43
44.92
46.89
29.11
31.85
36.52
54.00
35.96
21.52
32.26
47.19
50.93
24.21
34.84
82.64
15.35
11.39
37.60
32.52
70.64
39.32
45.34
29.12
32.17
36.60
54.44
35.95
21.35
32.89
46.71
51.40
24.08
35.53
82.43
15.07
12.54
37.63
32.59
70.64
39.40
45.37
29.19
32.26
36.73
54.56
35.98
21.35
32.08
47.36
50.87
24.16
34.22
82.96
15.07
12.60
37.67
32.59
70.64
39.41
45.38
29.21
32.33
36.75
54.71
35.98
21.41
32.38
47.23
51.17
24.25
34.88
82.67
15.41
12.60
17
18
19
CH3O
CH3O
CH2NH
CH2CH2
CH2CH2CH2
1Ј
2Ј
3Ј
4Ј
5Ј
6Ј
50.45
40.50
35.76
44.43
34.68
24.30
143.86
131.61
116.12
127.32
128.47
116.60
50.48
40.52
35.78
44.46
34.72
24.32
143.88
131.60
116.12
127.30
128.47
116.61
144.54
131.66
116.18
127.18
128.70
117.10
143.97
131.63
116.15
127.01
128.40
116.50
144.49
131.62
116.15
127.17
128.66
117.05
143.94
131.64
116.16
127.02
128.39
116.48
144.51
131.60
116.15
127.14
128.66
117.05
143.96
131.62
116.15
127.03
128.40
116.50
3
The assignments for the 13C chemical shifts of aromatic
carbon atoms of 5-azido-2-nitro benzoylamido and of
4-azido-2-nitrophenylamino groups (Tables 3–6) were
made by using the 1H, 13C heteronuclear HMBC correlation
technique performed, in C5D5N, on 17␣-aminoethyl ANB
and ANP derivatives of 3,3-dimethoxy-5␣-androstan-
17␤-ol 11 and 20, respectively. On the other hand, no 13C
signals could be detected for the 5-azido-2-nitro-3,4,6-trif-
luorophenylamino chromophore, as expected for polyflu-
oroaromatic compounds. For the 17␣-(5-azido-2-nitroben-
zoyl)amidoethyl analog 11, the doublet signals of C3Љ, C4Љ,
and C6Љ methine carbon atoms were identified by correla-
tion with their corresponding 1H NMR signals. Assignment
for the resonance of the C2Љ quaternary carbon atom was
weak JCCNH coupling of the NH proton was detected with
only one of these two carbon atoms, postulated, therefore, as
the C1Љ carbon atom. An HMBC experiment was also per-
formed in CDCl3, which confirmed that the order of the very
close C4Љ and C6Љ signals assigned for compound 11, in
C5D5N, was conserved in CDCl3. For the 17␣-(4-azido-2-
nitrophenyl)aminoethyl derivative 20, the signals of the
C3Љ, C5Љ, and C6Љ methine carbon atoms were readily iden-
1
tified by correlation with their H NMR signals. The stron-
gest couplings were found to occur for JCCCH interactions
3
between H3Љ and C1Љ or C5Љ carbon atoms, H5Љ and C1Љ or
C3Љ carbon atoms, and H6Љ and C2Љ or C4Љ carbon atoms,
while a strong 3JCCNH coupling was also observed between
NH and both C2Љ or C6Љ carbon atoms. These observations
led to an unequivocal identification of the resonances for
C1Љ, C2Љ, and C3Љ carbon atoms, since for H3Љ and H5,Љ one
of the two coupled carbon atoms, C5Љ and C3,Љ respectively,
was already identified, whereas for H6,Љ the C2Љ carbon
atom could be differentiated from C4Љ by its coupling with
NH, which was otherwise coupled with the already identi-
fied C6Љ carbon atom. The carbon signals of the polyfluoro-
aromatic ring of ANF derivatives could not be detected but
the corresponding CH2N signals appeared as narrow dou-
3
made easily from the strongest JCCCH couplings of H4Љ
with C2Љ and C6Љ carbon atoms and of H6Љ with C2Љ and
C4Љ carbon atoms, as expected for aromatic rings, since the
resonances for C4Љ and C6Љ carbon atoms were previously
established through correlation with signals of attached pro-
tons. A strong coupling of H3Љ with the two remaining C1Љ
and C5Љ quaternary carbon atoms was observed but did not
afford useful information for differentiating these two sig-
nals. However, in optimized conditions (see Section 2.1), a

E. Mappus et al. / Steroids 65 (2000) 459–481
477
Table 7
13C NMR data (CDCl3) for 17␣-(N-5-azido-2-nitro-3,4,6-trifluorophenyl)-aminomethyl, ethyl, and propyl derivatives of 3,3-dimethoxy- (26, 29, 32), 3-
oxo- (27, 30, 33), and 3␤-hydroxy-5␣-androstan-17␤-ol (28, 31, 34)
26
29
32
27
30
33
28
31
34
Carbon
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
␦-CDCl3
1
2
3
4
5
6
7
8
34.98
28.22a
100.30
35.46
42.40
28.29a
31.62
36.26
53.79
35.79
20.75
32.04
45.88
51.06
23.48
35.46
83.32
83.35*
14.07
11.62
47.53
47.46
52.15
52.23*
35.04
28.32
100.34
35.50
42.47
28.32
31.69
36.41
53.99
35.81
20.78
31.43
46.69
50.34
23.56
34.68
83.88
35.06
28.34
100.37
35.51
42.48
28.34
31.73
36.40
54.03
35.82
20.84
31.62
46.52
50.56
23.64
34.97
83.32
38.48
38.10
211.98
44.60
46.64
28.73
31.33
36.18
53.49
35.74
21.02
31.98
45.89
50.89
23.51
35.41
83.28
83.31*
14.09
11.46
38.53
38.13
211.95
44.64
46.68
28.76
31.40
36.29
53.68
35.86
21.03
31.34
46.66
50.13
23.56
34.61
83.80
38.56
38.15
211.97
44.67
46.72
28.82
31.44
36.30
53.74
35.75
21.10
31.54
46.50
50.38
23.64
34.96
83.27
36.96
31.45
71.24
38.08
44.84
28.50
31.70
36.29
54.07
35.56
20.82
32.05
45.88
51.07
23.50
35.46
83.34
37.00
31.46
71.24
38.10
44.87
28.53
31.76
36.40
54.24
35.55
20.82
31.41
46.66
50.31
23.55
34.65
83.85
37.01
31.48
71.25
38.13
44.88
28.58
31.79
36.38
54.27
35.55
20.88
31.60
46.48
50.53
23.63
34.92
83.29
9
10
11
12
13
14
15
16
17
18
19
CH3O
CH3O
CH2NH
13.91
11.65
47.52
47.47
42.84
42.99*
35.86
14.33
11.65
47.51
47.46
47.08
47.22*
33.40
25.11
13.92
11.49
14.34
11.50
14.07
12.30
13.91
12.33
14.33
12.33
52.11
52.25*
42.84
42.93*
35.86
47.04
47.19*
33.39
25.07
52.11
52.25*
42.95
42.86*
35.86
47.08
47.17*
33.38
25.07
CH2CH2
CH2CH2CH2
a Assignments may be exchanged.
* Signal corresponding to the second peak of a narrow doublet.
blets, as found also for C17 carbon atoms of the shorter
aminomethyl side-chains. The hypothesis of a contamina-
tion with an o-azido-nitrophenyl isomer was ruled out by
19F NMR of the 4-azido-1-nitro-2,4,5,6-tetrafluorobenzene
reagent (two symmetrical multiplets at 149.7 and 146.3
ppm), as previously reported [10,23], and of the aminoethyl
ANF derivative 29 (see Section 2).
Substituent-induced effects for the shorter 17␣-amino-
methyl ANB, ANP, and ANF side-chains in all three 3,3-
dimethoxy-, 3-oxo-, and 3␤-hydroxy-5␣-androstane series
were calculated by comparing with the 13C NMR spectra of the
three corresponding unsubstituted 17␤-hydroxy analogs, 2,
DHT [1], and 1, respectively. The effects of increasing the
linker length from aminomethyl (n ϭ 1) to aminoethyl (n ϭ 2)
and aminopropyl (n ϭ 3) derivatives with a same terminal
group were estimated using ⌬␦2–1, ⌬␦3–2, and ⌬␦3–1 incre-
ments calculated for each carbon atom of the steroid skeleton.
Modification of the length of the 17␣-side-chain had
either insignificant or very weak effects on the 13C NMR
resonances for carbon atoms C1 to C11, and C19, remote
from the 17-substituent. These weak effects were observed
mainly for C5, C7, C8, and C9 carbon atoms (data not
shown) and differed slightly according to the structures of
ANB, ANP, and ANF chromophores. In most of these
cases, an increase of the side-chain length was found to
produce weak ⌬␦2–1 deshielding effects and, most often,
weaker ⌬␦3–2 deshielding effects, which resulted in larger
⌬␦3–1 deshielding effects, culminating at ϩ0.35/ϩ0.46 ppm
(for C9 of ANF derivatives, in C5D5N).
The effects of lengthening the aminomethyl spacer to an
aminoethyl substituent were, in most cases, much greater on
carbon atoms C12 to C18 (Table 9), as expected from their
vicinity from the modified 17␣-side-chain, than on more
remote carbon atoms C1 to C10, and C19, although only
very slight effects (୏ 0.18 ppm) were observed for C15
(mostly downfield shifts). In CDCl3, C12, C14, and C18
carbon atoms of the three ANB, ANP, and ANF derivatives
and the C16 carbon atom of ANP and ANF derivatives
underwent shielding effects. The strongest shielding effects
were found for C12, C14, and C16 of ANP derivatives, all
slightly larger than those found for ANF derivatives, while
the weakest shielding effects were observed for C18 in all
three series and for C12 and C14 of ANB derivatives. On
the other hand, in the same CDCl3 solvent, C13 and C17
carbon atoms in all series and the C16 carbon atom of ANB
derivatives underwent deshielding effects. The deshielding
effects on C13 were rather similar for the three ANB, ANP,
and ANF derivatives, while those on C17 were much stron-

478
E. Mappus et al. / Steroids 65 (2000) 459–481
Table 8
13C NMR data (C5D5N) for 17␣-(N-5-azido-2-nitro-3,4,6-trifluorophenyl)-aminomethyl, ethyl, and propyl derivatives of 3,3-dimethoxy- (26, 29, 32),
3-oxo- (27, 30, 33), and 3␤-hydroxy-5␣-androstan-17␤-ol (28, 31, 34)
26
29
32
27
30
33
28
31
34
Carbon
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5N
␦-C5D5
N
1
2
3
4
5
6
7
8
35.25
28.87
100.52
35.88
42.50
28.60
32.03
36.47
53.87
36.01
21.25
32.72
46.58
51.48
24.05
35.08
82.67
82.70*
15.08
11.75
47.36
47.29
53.38
53.51*
35.37
28.90
100.53
35.93
42.68
28.68
32.18
36.64
54.16
36.05
21.25
32.01
47.36
50.77
24.07
34.14
83.34
35.44
28.92
100.53
35.94
42.72
28.71
32.29
36.67
54.33
36.06
21.35
32.29
47.19
51.10
24.17
34.87
82.59
38.69
38.34
210.32
44.87
46.77
29.02
31.64
36.33
53.62
35.92
21.45
32.67
46.56
51.30
24.05
35.11
82.65
38.76
38.38
210.34
44.91
46.86
29.09
31.77
36.49
53.86
35.94
21.43
31.93
47.32
50.56
24.06
34.10
83.34
38.80
38.40
210.38
44.93
46.90
29.12
31.84
36.52
53.99
35.96
21.52
32.21
47.16
50.87
24.18
34.83
82.57
37.52
32.54
70.64
39.35
45.23
29.10
32.12
36.57
54.35
35.93
21.34
32.80
46.60
51.57
24.08
35.13
82.72
37.61
32.57
70.63
39.39
45.34
29.17
32.24
36.71
54.56
35.96
21.32
32.03
47.35
50.79
24.09
34.10
83.38
37.66
32.58
70.64
39.40
45.38
29.21
32.33
36.74
54.70
35.97
21.41
32.33
47.20
51.11
24.22
34.86
82.60
9
10
11
12
13
14
15
16
17
18
19
CH3O
CH3O
CH2NH
14.95
11.81
47.36
47.29
43.53
43.66*
36.64
15.34
11.83
47.34
47.28
47.34
47.46*
34.51
25.92
15.05
11.32
14.91
11.37
15.33
11.39
15.08
12.52
14.95
12.58
15.38
12.60
53.33
53.47*
43.49
43.36*
36.54
47.29
47.42*
34.44
25.87
53.38
53.51*
43.53
43.62*
36.56
47.30
47.40*
34.48
25.90
CH2CH2
CH2CH2CH2
* Signal corresponding to the second peak of a narrow doublet.
ger for ANB derivatives than for ANP and ANF derivatives.
The C16 carbon atom of ANB derivatives exhibited a very
slight deshielding effect, while the C15 carbon atom re-
mained almost unaffected in all series. When changing the
CDCl3 solvent for C5D5N, the ⌬␦2–1 increments of ANB
derivatives were significantly modified for carbon atoms
C12, C16, C17 and C18, as shown by a change of sign but
not of magnitude for C12 and C18, a much stronger
deshielding effect for C16, and a change of sign leading to
a much weaker shielding effect for C17. Otherwise, in
C5D5N solvent, ⌬␦2–1 increments of all C12 to C18 carbon
atoms of ANP and ANF derivatives were found to be rather
close to those observed in CDCl3 solvent.
Changing the aminoethyl spacer for an aminopropyl sub-
stituent led also to characteristic upfield or downfield shifts of
carbon atoms C12 to C18 (Table 9). The corresponding ⌬␦3–2
increments for all three ANB, ANP, and ANF derivatives, in
both CDCl3 and C5D5N solvents, were found to exhibit in most
cases a sign opposite to that of the corresponding ⌬␦2–1 incre-
ments and very often a much lower magnitude, thus leading to
⌬␦3–1 increments having the same sign as ⌬␦2–1 increments.
The largest ⌬␦3–2 increments were found for C17 carbon atom
of ANB derivatives, in CDCl3, while rather strong effects were
also observed on C17 (upfield shifts) and C18 (downfield
shifts) carbon atoms of ANF derivatives in both CDCl3 and
C5D5N solvents and on C16 carbon atoms (downfield shift) of
ANP and ANF derivatives, in C5D5N only. Exceptions to the
general rule of change of sign were observed in the case of
ANB derivatives, for ⌬␦3–2 increments of C15 and C16 in
CDCl3, and of C14 and C18 in C5D5N, and, in the case of ANP
and ANF derivatives, for ⌬␦3–2 increments of C15, in both
CDCl3 and C5D5N solvents, which, in all cases, contributed to
an increase of the magnitude of the corresponding ⌬␦3–1 in-
crement of unchanged sign as compared with that of the ⌬␦2–1
increment. Exceptions to the general rule of lower magnitude
leading to a change of sign for the corresponding ⌬␦3–1 incre-
ments, as compared with the ⌬␦2–1 increments, were observed
for ⌬␦3–2 increments of C17 of ANF derivatives and of C18 of
ANP and ANF derivatives, in both CDCl3 and C5D5N sol-
vents, and of C18 of ANB derivatives, in CDCl3 only. Chang-
ing the CDCl3 solvent for C5D5N was found to modify the sign
of ⌬␦3–2 increments only for C12, C14, C15, C16, and C17
carbon atoms of ANB derivatives and also resulted in signif-
icant modifications of magnitudes for C12, C16, and C17
carbon atoms of these derivatives. Significant solvent effects
were also found for C12, C14, C16, and C17 carbon atoms of
ANP and ANF derivatives, leading to larger magnitudes of the
corresponding ⌬␦3–2 increments in C5D5N solvent than in
CDCl3.
Lengthening the 17␣-side-chain from methyl to ethyl and

E. Mappus et al. / Steroids 65 (2000) 459–481
479
Table 9
13C NMR chemical shift increments for 17␣-aminomethyl ANB, ANP, and ANF substituents versus unsubstituted 17␤-OH steroid analogs (⌬␦ 17␣-
CH2NHR-17␤-OH) and for lengthening 17␣-aminoalkyl side chains from aminomethyl to aminoethyl and aminopropyl substitutents and (⌬␦ 2-1, ⌬␦
3-2, and ⌬␦ 3-1)
C
␦-CDCl3
␦-C5D5N
atom
⌬␦
⌬␦ 2-1
⌬␦ 3-2
⌬␦ 3-1
⌬␦
⌬␦ 2-1
⌬␦ 3-2
⌬␦ 3-1
17␣-CH2NHR
-17␤-OH
17␣-CH2NHR
-17␤-OH
12
ANB
ANP
ANF
13
Ϫ5.01/Ϫ5.03
Ϫ4.39/Ϫ4.42
Ϫ4.67/Ϫ4.72
Ϫ0.25/Ϫ0.26
Ϫ0.83/Ϫ0.87
Ϫ0.61/Ϫ0.64
ϩ0.03/ϩ0.05
ϩ0.16/ϩ0.21
ϩ0.19/ϩ0.20
Ϫ0.21/Ϫ0.23
Ϫ0.66/Ϫ0.67
Ϫ0.42/Ϫ0.45
Ϫ5.39/Ϫ5.41
Ϫ4.63/Ϫ4.71
Ϫ4.75/Ϫ4.85
ϩ0.24/ϩ0.30
Ϫ0.80/Ϫ0.82
Ϫ0.71/Ϫ0.77
Ϫ0.15/Ϫ0.18
ϩ0.28/ϩ0.30
ϩ0.28/ϩ0.30
ϩ0.09/ϩ0.12
Ϫ0.51/Ϫ0.53
Ϫ0.43/Ϫ0.47
ANB
ANP
ANF
14
ϩ2.85/ϩ2.89
ϩ3.07/ϩ3.09
ϩ2.87/ϩ2.91
ϩ0.78/ϩ0.83
ϩ0.60/ϩ0.62
ϩ0.77/ϩ0.81
Ϫ0.14/Ϫ0.18
Ϫ0.15/Ϫ0.17
Ϫ0.16/Ϫ0.18
ϩ0.64/ϩ0.65
ϩ0.44/ϩ0.45
ϩ0.60/ϩ0.64
ϩ3.10/ϩ3.12
ϩ3.10/ϩ3.14
ϩ2.97/ϩ3.02
ϩ0.63/ϩ0.64
ϩ0.65.
ϩ0.75/ϩ0.78
Ϫ0.06/Ϫ0.12
Ϫ0.12/Ϫ0.14
Ϫ0.15/Ϫ0.17
ϩ0.52/ϩ0.57
ϩ0.51/ϩ0.53
ϩ0.60/ϩ0.61
ANB
ANP
ANF
15
Ϫ0.34/Ϫ0.39
ϩ0.28/ϩ0.30
ϩ0.04/ϩ0.07
Ϫ0.35/Ϫ0.37
Ϫ0.88/Ϫ0.93
Ϫ0.72/Ϫ0.76
ϩ0.09/ϩ0.15
ϩ0.16/ϩ0.21
ϩ0.22/ϩ0.25
Ϫ0.21/Ϫ0.26
Ϫ0.72/Ϫ0.74
Ϫ0.50/Ϫ0.54
ϩ0.02/ϩ0.05
Ϫ0.01/ϩ0.04
ϩ0.05/ϩ0.15
Ϫ0.33/Ϫ0.41
Ϫ0.50/Ϫ0.53
Ϫ0.71/Ϫ0.78
Ϫ0.07/Ϫ0.15
ϩ0.26/ϩ0.30
ϩ0.31/ϩ0.33
Ϫ0.48
Ϫ0.21/Ϫ0.26
Ϫ0.38/Ϫ0.46
ANB
ANP
ANF
16
ϩ0.19/ϩ0.22
ϩ0.13/ϩ0.16
ϩ0.08/ϩ0.13
ϩ0.01
ϩ0.07/ϩ0.09
ϩ0.05/ϩ0.08
ϩ0.03/ϩ0.05
ϩ0.03/ϩ0.06
ϩ0.08
ϩ0.04/ϩ0.06
ϩ0.12/ϩ0.14
ϩ0.13/ϩ0.16
ϩ0.31/ϩ0.38
ϩ0.18/ϩ0.26
ϩ0.20/ϩ0.25
ϩ0.04/ϩ0.06
ϩ0.07/ϩ0.09
ϩ0.01/ϩ0.02
Ϫ0.01/Ϫ0.02
ϩ0.08/ϩ0.09
ϩ0.10/ϩ0.13
ϩ0.03/ϩ0.04
ϩ0.15/ϩ0.18
ϩ0.12/ϩ0.14
ANB
ANP
ANF
17
ϩ4.05/ϩ4.08
ϩ5.22/ϩ5.36
ϩ4.91/ϩ5.00
0.00/ϩ0.15
Ϫ0.99/Ϫ1.04
Ϫ0.78/Ϫ0.81
0.00/ϩ0.12
ϩ0.22/ϩ0.27
ϩ0.27/ϩ0.35
ϩ0.12/ϩ0.15
Ϫ0.74/Ϫ0.78
Ϫ0.45/Ϫ0.54
ϩ2.99/ϩ3.01
ϩ4.50/ϩ4.58
ϩ4.06/ϩ4.14
ϩ0.84/ϩ0.86
Ϫ1.30/Ϫ1.35
Ϫ0.94/Ϫ1.03
Ϫ0.26/Ϫ0.31
ϩ0.64/ϩ0.66
ϩ0.73/ϩ0.76
ϩ0.55/ϩ0.58
Ϫ0.65/Ϫ0.71
Ϫ0.21/Ϫ0.28
ANB
ANP
ANF
18
ϩ1.41/ϩ1.59
ϩ1.19/ϩ1.39
ϩ1.36/ϩ1.60
ϩ1.37/ϩ1.51
ϩ0.26/ϩ0.33
ϩ0.49/ϩ0.53
Ϫ1.17/Ϫ1.28
Ϫ0.03/Ϫ0.07
Ϫ0.53/Ϫ0.56
ϩ0.20/ϩ0.26
ϩ0.23/ϩ0.26
Ϫ0.03/Ϫ0.05
ϩ1.64/ϩ1.75
ϩ1.01/ϩ1.12
ϩ1.27/ϩ1.37
Ϫ0.49/Ϫ0.52
ϩ0.53
ϩ0.64/ϩ0.69
ϩ0.10/ϩ0.12
Ϫ0.27/Ϫ0.29
Ϫ0.75/Ϫ0.78
Ϫ0.38/Ϫ0.40
ϩ0.24/ϩ0.26
Ϫ0.08/Ϫ0.12
ANB
ANP
ANF
ϩ2.93/ϩ2.96
ϩ2.98/ϩ3.00
ϩ2.92/ϩ2.93
Ϫ0.10/Ϫ0.13
Ϫ0.11/Ϫ0.12
Ϫ0.16/Ϫ0.17
ϩ0.36/ϩ0.38
ϩ0.34/ϩ0.35
ϩ0.42
ϩ0.25/ϩ0.26
ϩ0.22/ϩ0.24
ϩ0.25/ϩ0.26
ϩ3.12/ϩ3.14
ϩ3.13/ϩ3.16
ϩ3.14/ϩ3.17
ϩ0.13/ϩ0.14
0.00/Ϫ0.02
Ϫ0.13/Ϫ0.14
ϩ0.23/ϩ0.24
ϩ0.33/ϩ0.34
ϩ0.39/ϩ0.43
ϩ0.36/ϩ0.38
ϩ0.31/ϩ0.34
ϩ0.26/ϩ0.30
propyl linkers was found to exert parallel effects on 13C NMR
resonances of the CH2N carbon atoms in all three ANB, ANP,
and ANF series, thus suggesting the absence of any significant
influence of the size of the linker on the two possible cis-trans
conformations of alkylamido NH-CO bonds of ANB deriva-
tives as compared with the corresponding alkylamino bonds of
ANP and ANF analogs. This observation points out to the
hypothesis that a same conformation of the amide bond of
ANB derivatives is present for the three different linkers, as
confirmed by the absence of splitted 13C NMR resonances of
carbon atoms, which surround the amide group even after
lowering temperature down to Ϫ60°C in CDCl3 solvent. This
presumed unique conformation corresponds probably to the
trans rotamer (with the NH proton anti to the carbonyl group),
reported as the predominant one for N-monosubstituted amides
[24–26]. On the other hand, both cis and trans isomers are
often observed for N-disubstituted amides as reported for 16-
alkylamido steroidal side-chains [27].
This difference probably results mainly from the carbonyl
group of the amide function of ANB derivatives. The influence
of the conjugated phenyl ring of ANB derivatives could be
estimated for the shorter 17␣-aminomethyl derivatives only,
by comparing the effects of the ANB terminal group with those
reported for aliphatic acetylamido and hemiglutaramido
methyl ester terminal groups [3]. While almost no significant
differences were found between 17␣-acetylamidomethyl and
17␣-hemiglutaramidomethyl derivatives of 3-oxo-androstan-
17␤-ol, the presence of a 17␣-(5-azido-2-nitrobenzoyl)am-
idomethyl substituent led to small but significant differences,
as compared with the acetylamido derivative, for the reso-
nances of C12 to C17 carbon atoms. Deshielding effects rang-
ing from ϩ0.11 to ϩ0.19 ppm were observed either in CDCl3,
for C16, or in C5D5N, for C13 and C14, and in both solvents,
for C12 and C15, while a weak shielding effect was also found
for C17, in C5D5N only. On the other hand, the direct coupling
of the terminal amino group of the side-chain to the nitroaryl
azide rings of ANP and ANF derivatives led to similar behav-
iors, even when changing the side-chain length, thus indicating
that the presence of fluorine atoms on the ANF derivative has
a rather limited influence. Increments were also calculated for
The general trend observed in this study is that the effects of
the 17␣-aminoalkyl-ANB conjugates on 13C NMR resonances
of steroidal carbon atoms are most often very different from
the rather similar effects observed for ANP and ANF groups.

480
E. Mappus et al. / Steroids 65 (2000) 459–481
changing the 17␣-cyanomethyl substituent of compound 4 for
the 17␣-cyanoethyl group of compound 6, which revealed
extremely large ⌬␦2–1 shielding effects on C16 (Ϫ2.41 and
Ϫ2.62 ppm in CDCl3 and in C5D5N, respectively) much
higher than the deshielding effects on C17 (ϩ0.82 and ϩ0.46
ppm in CDCl3 and in C5D5N, respectively). Increments for the
increasing lengths of 17␣-aminomethyl, 17␣-aminoethyl, and
17␣-aminopropyl substituents of unsubstituted amino precur-
sors 3, 5, and 7, showed the highest values for C17 (⌬␦2–1
shielding effects of Ϫ2.51 and Ϫ1.17 ppm in CDCl3 and in
C5D5N, respectively; ⌬␦3–2 deshielding effects of ϩ1.94 and
ϩ1.27 ppm in CDCl3 and in C5D5N, respectively), while large
effects, often different in sign or magnitude from those found
for ANB, ANP, or ANF derivatives, were also observed on
C16, as well as on C12, C13, C14, C15, and C18 carbon atoms.
These results, established for intramolecular interactions, sug-
gest that improvements of the sensitivity of NMR equipment,
which could be expected in the future for developing biologic
studies, may lead to more subtle investigations of intermolec-
ular protein-ligand interactions in a binding site.
Experiments have been made in order to estimate the
effects of increasing the side-chain length of 17␣-amidoal-
kyl-ANB derivatives of DHT on the binding properties with
SHBG. Competition experiments with the three radioinert
17␣-aminomethyl-, 17␣-aminoethyl-, and 17␣-aminopro-
pyl-ANB conjugates of DHT 9, 12, and 15, and unlabeled
DHT were performed on purified human SHBG [28], using
tritiated DHT as tracer (Fig. 1, upper panel). Relative bind-
ing affinities (RBAs) for SHBG were calculated as the ratio
of the concentration of radioinert unsubstituted DHT that
gave 50% displacement of the binding of the radioactive
tracer to the concentration of 17␣-aminoalkyl competitors
giving the same inhibition. Increasing lengths of amidom-
ethyl-, amidoethyl-, and amidopropyl-ANB derivatives
were found to correspond to decreasing RBA values for
SHBG (0.76, 0.47, and 0.10, respectively versus 1.00 ref-
erence value for DHT). The association constants of these
three amidoalkyl-ANB competitors (1.6 ϫ 109 MϪ1, 1.0 ϫ
109 MϪ1, and 0.2 ϫ 109 MϪ1, respectively) were estimated
indirectly from RBA values using the equation Ka (competitor)
ϭ Ka (DHT) /[(1/RBA)(1 ϩ R) Ϫ R], where Ka (DHT) is the
association constant of DHT for SHBG (2.2 ϫ 109 MϪ1 at
22°C) [28] and R the bound to free ratio of tritiated DHT
tracer at 50% displacement (R ϭ 0.036) [29]. Similar pre-
liminary binding assays were also performed on cytosolic
androgen receptors of rat ventral prostate employing a re-
ported protocol [30], modified by the use of a tritiated DHT
tracer more similar to the structures of 17␣-aminoalkyl
competitors than the usual tritiated 17␣-methyltrienolone
(R1881) tracer and a shorter incubation time, which may
limit uncontrolled degradation effects of proteases. Compe-
tition experiments with the three 17␣-amidoalkyl-ANB de-
rivatives of DHT (Fig. 1, lower panel) showed that only
amidoethyl and amidopropyl derivatives interacted signifi-
cantly with the androgen receptors (RBA values of 0.05 and
0.10, respectively). These RBA values are only estimates
Fig. 1. Displacement of tritiated DHT from purified human SHBG (upper
panel) and from cytosolic androgen receptors of rat ventral prostate (lower
panel) by radioinert 17␣-aminomethyl- (n ϭ 1), 17␣-aminoethyl- (n ϭ 2),
and 17␣-aminopropyl- (n ϭ 3) ANB conjugates of DHT 9, 12, and 15, and
unlabeled DHT. Incubations were carried out as indicated in Section 2, at
22°C for SHBG and at 4°C for androgen receptors.
established from rough measurements providing merely
qualitative data, not suitable for calculations of the corre-
sponding assosiation constants [31]. Nevertheless, these re-
sults are consistent with the increasing relative efficiencies
reported for similar methyl, ethyl, and propyl linkers on the
capacity of 17␣-haloacetamidoalkyl estrogenic derivatives
to inhibit the specific binding of estradiol with the estrogen
receptor [17]. Similarly, 17␣-carboxymethylamide deriva-
tives of DHT were unable to bind to androgen receptors [32]
whereas 17␣-iodovinyl derivatives showed much higher
binding affinities [33]. Successful affinity labeling experi-
ments of purified human SHBG with 17␣-aminoalkyl-ANB
derivatives of [3␣-3H]5␣-androstane-3␤,17␤-diol have al-
ready shown that two different peptides are specifically
labeled according to the length of the 17␣-amidoalkyl
linker. Identification of these peptides as well as parallel
studies with 17␣-acetamidoalkyl ANP and ANF analogs, in
order to reveal possible similar effects of the linker size, are

E. Mappus et al. / Steroids 65 (2000) 459–481
481
[14] Cook CE, Corley RC, Wall ME. Steroids. LXXIX. Synthesis and
reactions of oxiranes obtained from 3- and 17-ketosteroids. J Org
Chem 1968;33:2789–93.
in progress in our laboratory, and the results of these studies
will be published in a future article.
[15] Kirk DN, Wilson MA. A novel route to D-homoandrostane deriva-
tives, including new methods for the preparation and reduction of
hydroxy-azides. Chem Commun 1970:64–5.
Acknowledgments
[16] Ponsold von K, Hu¨bner M, Schnabel R., Strecke J. Synthese und
uterotrope wirkung neuartiger gebremster o¨strogene vom typ 17␣-
substituierter derivate des o¨stradiol-3-methyla¨thers. Arzneim-Forsch
(Drug Res) 1974;24:896–900.
[17] El Garrouj D, Aliau S, Aumelas A, Borgna J-L. Steroidal affinity
labels of the estrogen receptor. 2. 17␣-[(haloacetamido)alkyl]estradi-
ols. J Med Chem 1995;38:2339–48.
[18] Bhacca N, Williams DH. Applications of NMR spectroscopy in
organic chemistry. San Francisco: Holden Day, 1966. p. 19–24.
[19] Salman M, Reddy BR, Ray S, Stotter PL, Chamness GC. 17␣-allyl
estradiol analogues as candidates for development of high-affinity
fluorescein-estradiol conjugates. J Steroid Biochem 1986;24:539–
48.
[20] Blunt JW, Stothers JB. 13C NMR spectra of steroids. A survey and
commentary. J Magn Reson 1977;9:439–64.
[21] Dionne P, Poirier P. 13C nuclear magnetic resonance study of 17␣-
substituted estradiols. Steroids 1995;60:830–6.
[22] Atta-ur-Rahman. Nuclear magnetic resonance. New York: Springer
Verlag, 1986. p. 140–82.
[23] Keana JFW, Cai SX. New reagents for photoaffinity labeling: syn-
thesis and photolysis of functionalized perfluorophenyl azides. J Org
Chem 1990;55:3640–7.
[24] Stewart WE, Siddall TH III. Nuclear magnetic resonance studies of
amides. Chem Rev 1970;70:517–51.
[25] Idoux JP, Scandrett JM, Sikorski JA. Conformational influence of
nonacyl groups on acyl group properties in N-monosubstituted
amides and in other carboxylic acid derivatives: a 7-position prox-
imity effect. J Am Chem Soc 1976;99:4577–83.
[26] Jiao D, Barfield M, Hruby VJ. An initio IGLO study of the syn/anti
dependence of the 13C NMR chemical shifts in simple amides. Magn
Reson Chem 1993;31:75–9.
[27] Pelletier JD, Labrie F, Poirier D. N-butyl, N-methyl, 11-[3Ј,17Ј␤-
(dihydroxy)- 1Ј,3Ј,5Ј(10Ј)-estratrien 16Ј␣-yl]-9(R/S)-)bromound-
ecanamide: synthesis and 17Ј␤-HSD inhibiting, estrogenic and an-
tiestrogenic activities. Steroids 1994;59:536–47.
[28] Grenot C, de Montard A, Blache`re T, Rolland de Ravel M, Mappus
E, Cuilleron CY. Characterization of Met-139 as the photolabeled
amino acid residue in the steroid binding site of sex hormone binding
globulin using ⌬6 derivatives of either testosterone or estradiol as
unsubstituted photoaffinity labeling reagents. Biochemistry 1992;31:
7609–21.
[29] Pugeat MM, Dunn JF, Nisula BC. Transport of steroid hormones:
interaction of 70 drugs with testosterone-binding globulin and corti-
costeroid-binding globulin in human plasma. J Clin Endocrinol
Metab 1981;53:69–75.
[30] Radwan F, Carmel M, Elhilali M, Bouthillier F, Lehoux JG. Studies
on the characterization of rat prostate androgen receptors. Mol Cell
Biochem 1989;90:81–9.
[31] Korenman SG. Relation between estrogen inhibitory activity and
binding to cytosol of rabbit and human uterus. Endocrinology 1970;
87:1119–23.
We thank L. Boggio, S. Samiri, J.-C. Rousseau, and B. Marin,
for their helpful participation at different steps of this work, O.
Miani (Center de RMN, Universite´ Claude Bernard Lyon 1-
ESCPE, Villeurbanne, France) for NMR measurements, and D.
Forest, and Dr. M. Becchi (Centre de Spectrome´trie de Masse,
CNRS, Solaize-Vernaison, France) for mass spectra. Thanks are
also due to M. Ainouze for technical assistance and to Dr J. Carew
for help in editing the manuscript.
References
[1] Mappus E, Renaud M, Rolland de Ravel M, Grenot C, Cuilleron CY.
Synthesis and characterization by 1H and 13C nuclear magnetic res-
onance spectroscopy of 17␣-hexanoic derivatives of DHT and tes-
tosterone. Steroids 1992;57:122–34.
[2] Avery MA, Tanabe M, Crowe DF, Detre G, Peters RH, Chong WKM.
Synthesis and testing of 17a␤-hydroxy-7␣-methyl-D-homoestra-
4,16-dien-3-one: a highly potent orally active androgen. Steroids
1990;55:59–64.
[3] Mappus E, Chambon C, Rolland de Ravel M, Grenot C, Cuilleron
CY. Synthesis and characterization by 1H and 13C nuclear magnetic
resonance spectroscopy of 17␣-cyano, 17␣-aminomethyl, and 17␣-
alkylamidomethyl derivatives of 5␣-dihydrotestosterone and testos-
terone. Steroids 1997;62:603–20.
[4] Kotzyba–Hibert F, Kapfer I, Goeldner M. Recent trends in photoaf-
finity labeling. Angew Chem Int Ed Engl 1995;34:1296–1312.
[5] Dodrell DM, Pegg DT, Bendall MR. Distorsionless enhancement of
NMR signals by polarization transfer. J Magn Reson 1982;48:323–7.
[6] Davis AL, Keeler J, Laue ED, Moskau D. Experiments for recording
pure-absorption heteronuclear correlation spectra using pulse field
gradients. J Magn Reson 1992;98:207–16.
[7] Lerner L, Bax A. Sensitivity-enhanced two-dimensional hetero-
nuclear relayed coherence transfer NMR spectroscopy. J Magn Reson
1986;69:375–80.
[8] Bax A, Summers MF. 1H and 13C assignments from sensitivity-
enhanced detection of heteronuclear multiple bond connectivity by
2D multiple quantum NMR. J Am Chem Soc 1986;108:2093–4.
[9] Gasparrini F, Giovannoli M, Misiti D, Palmieri G. Synthesis of
dimethyl acetals, diethyl acetals, and cyclic acetals catalyzed by
aminopropylated silica gel hydrochloride (APSG-HCl). Tetrahedron
1984;40:1490–1500.
[10] Bolton R, Sandall JPB. Nucleophilic displacement in polyhalogeno-
aromatic compounds. Part 9. Kinetics of azidodefluorination and
methoxydefluorination of some polyfluorobenzonitriles. J Chem Soc
Perkin Trans 1978;2:1288–92.
[11] Rao PN, Khan AH, Moore PH Jr. Synthesis of new steroid haptens for
radioimmunoassay. Part III. 15␤-carboxyethylmercaptosteroid-bo-
vine serum albumin (BSA) conjugates. Specific antisera for radioim-
munoassay of 5␣-dihydrotestosterone, 5␣-androstane-3␤,17␤-diol,
and 5␣-androstane-3␣,17␤-diol. Steroids 1976;29:171–84.
[12] Corey EJ, Chaykovsky M. Dimethylsulfonium methylide, a reagent
for selective oxirane synthesis from aldehydes and ketones. J Am
Chem Soc 1962;84:3782–3.
[32] Stobaugh ME, Blickenstaff RT. Synthesis and androgen receptor
binding of dihydrotestosterone hemisuccinate homologs. Steroids
1990;55:259–62.
[33] Hoyte RM, MacLusky NJ, Hochberg RB. The synthesis and testing of
E-17␣-(2-iodovinyl)-5␣-dihydrotestosterone and of Z-17␣-(2-iodovi-
nyl)-5␣-dihydrotestosterone as ␥-emitting ligands for the androgen
receptor. J Steroid Biochem 1990;36:125–32.
[13] Ponsold K, Hu¨bner M, Schade W, Oettel M, Freund R. Progestagene
vom typ 17␣-CH2X-substituierter 19-nortestosteronderivate.
Pharmazie 1978;33:792–8.