P. Vinš et al. / Steroids 109 (2016) 56–59
57
2.2. Chemical synthesis
3.68–3.63 m, 2 H (H-50, H-17
a
); 3.57–3.43 m, 3 H (2 ꢀ H-700,
1 ꢀ H-40); 3.39–3.36 m, 1 H (H-30); 3.21 t, 1 H, J = 8.8 (H-20); 2.82
br t, 2 H, J = 6.6 (2 ꢀ H-600); 1.99–1.90 m, 2 H (H-12a, H-16a);
1.73–1.66 m, 2 H (H-7a, H-2a); 1.64–1.58 m, 5 H (H-5, H-8,
H-11a, H-15a, H-16b); 1.52 dt, 1 H, J1 = 13.2, J2 = 2.2 (H-4a);
1.48–1.40 m, 2 H (H-1a, H-2b); 1.39–1.35 dd, 1 H, J1 = 13.7,
J2 = 3.3 (H-11b); 1.29–1.21 m, 3 H (H-6a, H-6b, H-15b); 1.15 dt, 1
H, J1 = 12.6, J2 = 3.8 (H-12b); 1.03–0.91 m, 2 H (H-14, H-7b);
0.84 s, 6 H (3 ꢀ H-18, 3 ꢀ H-19); 0.77 dt, 1 H, J1 = 12.1, J2 = 3.8
2.2.1. Methyl 2,3,4-tri-O-acetyl-1-O-(3a-acetyloxy-5a-androstan-
17b-yl)-b-D-glucopyranosiduronate (2)
17b-Hydroxy-5 -androstan-3 -yl acetate [4–6] (1, 100 mg,
1 eq, 0.299 mmol) was dissolved in dry freshly distilled benzene
(8 ml) and treated by methyl 2,3,4-tri-O-acetyl-1 -bromo-1-
a
a
a
deoxy-b-D-glucuronate (6) (300 mg, 2.5 eq, 0.75 mmol) in pres-
ence of silver carbonate (200 mg, 2.4 eq, 0.725 mmol) and ground
molecular sieves (250 mg). Reaction mixture was stirred in dark,
under argon at room temperature for 3 days. Then, the mixture
was diluted by dichloromethane, filtered, and the solvent was
evaporated. The crude product was chromatographed on silica
gel column using gradient of diethyl ether (1–10% v/v) in dichlor-
omethane. Glycoside 2 (36 mg, 18%) was obtained as white amor-
phous solid. 1H NMR (CDCl3, 300 MHz): 5.24–5.15 m, 2 H; 5.01 m, 2
H; 4.58 d, 1 H, J = 7.9; 3.99 d, 1 H, J = 9.7; 3.75 s, 3 H; 3.57 t, 1 H,
J = 8.8; 2.05 s, 3 H; 2.04 s, 3 H; 2.01 m, 6 H; 0.79 s, 3 H; 0.70 s, 3
H. The title compound was previously prepared by another method
[7] and was found to have identical structure.
13
(H-9). C NMR (CD3OD, 150 MHz): 170.49 (C-60); 134.74, 2 C (C-
200, C-400); 117.99 (br.[8], C-500); 103.58 (C-10); 89.22 (C-17); 76.17
(C-30); 74.80 (C-50); 73.50 (C-20); 72.15 (C-40); 65.78 (C-3); 54.62
(C-9); 50.85 (C-14); 42.93 (C-13); 38.93 (C-5); 38.59 (C-700);
37.31 (C-12); 35.81 (C-10); 35.33 (C-4); 35.31 (C-8); 32.12 (C-1);
31.46 (C-7); 28.56 (C-16); 28.28 (C-2); 28.19 (C-600); 28.18 (C-6);
22.97 (C-15); 20.08 (C-11); 10.69 (C-18); 10.30 (C-19). HRMS ESI
+ (m/z): [M+H]+ calcd. for C30H47N3O7: 562.349341. Found:
562.34900.
3. Results and discussion
2.2.2. 3
a-Hydroxy-5a-androstan-17b-yl b-D-glucopyranosiduronic
For the synthesis of complex target amide 4 from commercial
acid (3)
building blocks of reasonable price, pathway starting with 3
a-ace-
Protected glycoside 2 (165 mg, 1 eq, 0.254 mmol) was dissolved
in methanol (2 ml) and the solution was diluted by 10 ml of water.
Solution of lithium hydroxide in methanol (0.5 M, 10 ml) was then
added. Reaction mixture was stirred in dark at room temperature
for 3 days. After dilution by another 10 ml of methanol, pH was
adjusted to neutral by a solution of acetic acid (10% in methanol).
Resulting solution was passed through a column of catex (Amber-
lite, IR-120, H-cycle, 20–50 mesh, 15 ml) using the methanol/water
(2/1) mixture as eluent. After evaporation, the crude product was
purified by column chromatography on reverse phase (C-18 mod-
ified silica gel), using gradient of MeOH in water. Deprotected glu-
curonide 3 (63 mg, 53%) was obtained as amorphous white solid.
This commercial product was described previously [4,7]. 1H NMR
(CD3OD, 300 MHz): 4.36 d, 1 H, J = 7.91 (H-10); 3.95 br t, 1 H
(H-3); 3.77–3.64 m, 2 H (H-50, H-17); 3.50 t, 1 H, J = 9.1 (H-40);
3.34 t, 1 H, J = 9.1 (H-30); 3.20 t, 1 H, J = 7.91 (H-20); 0.82–0.81
2 ꢀ s, 6 H (3 ꢀ H-18, 3 ꢀ H-19). HRMS ESI + (m/z): [M+Na]+ calcd.
for C25H40O8: 491.262087. Found: 491.26158.
toxy-5 -androstan-17-one was selected. Protected androstanolone
a
was reduced to secondary alcohol (Scheme 1), glycosylated by
acetylated halogenose, and after hydrolysis of protecting ester
groups subjected to the conjugation with histamine. Glycosylation
and amide formation with subsequent purification were critical
steps.
Ketone group of starting material was readily reduced to obtain
corresponding alcohol 1 in very good yield by common procedure
using sodium borohydride in mixture of methanol and ethyl acet-
ate. As reported previously [4–6], only 17b isomer was isolated.
The suitable halogenose 6 was prepared (see Scheme 2) with
the method of Bowering and Timell [9] from b-D-glucuronolactone.
From obtained building blocks, glycoside 2 was prepared by
modified Koenigs–Knorr reaction [10], following to method of
Thevis et al. [11] using silver carbonate as catalyst (even though
in stoichiometric excess).
It is noteworthy that this reaction is very sensitive and requires
absolutely dry solvents, giving increasing ratio of corresponding
orthoester 7 (Fig. 1) hand in hand with rising moisture content.
During our early experiments with this method, the orthoester
was even isolated in good yield as main product, using ‘‘dry”
reagent-grade toluene. After such experience, toluene was substi-
tuted by freshly dried and distilled benzene over grinded activated
molecular sieves. Nevertheless, the orthoester remained trouble-
some side product and thorough purification via column chro-
matography was necessary. This caused additional losses and the
overall yield was low (about 20% in total). Orthoester content can
be detected by characteristic signal of its methyl group in 1H
NMR spectrum (1.71 ppm in CDCl3).
2.2.3. 3a-Hydroxy-5a-androstan-17b-yl N-(4-imidazolyl)ethyl-b-D-
glucopyranosiduronamide (4)
To a solution of glucuronide 3 (9 mg, 1 eq, 19.2 nmol) in DMF
(200 ll), hydroxybenzotriazole (hydrate, 90%, 6 mg, 3.5 eq,
6.7 nmol) was added. DCC (14 mg, 3.5 eq, 6.7 nmol) was being
added in 3 portions. 6 mg were added at the beginning of the reac-
tion, another 6 mg were added after 2 days and at last, 2 mg were
added after 2 more days. The reaction was completed after 6 days
of slow stirring at room temperature. The reaction mixture was fil-
tered from crystals of dicyclohexylurea, which were washed by
another 100
6.7 nmol) in DMF (100
l
l of DMF. Solution of histamine (7.5 mg, 3.5 eq,
The deprotection of acetyl groups was achieved by treatment by
solution of LiOH in water/methanol over two days [12]. Despite the
general swiftness and ease of deacetylation reactions, ordinary
protocols using MeONa or NaOH with reaction times of several
hours did not lead to complete deacetylation of the substrate.
Several methods of amide formation were tried out for the final
step. First, T3PÒ [13] (propylphosphonic anhydride, 50% solution in
EtOAc) was employed in DMF/pyridine. Even with an excess (4–
6 mol. eq.) of the reagent, only starting steroid 3 was detected in
reaction mixture. Then, experiments with DEPC (diethyl phospho-
ryl cyanide) were made [14]. This reagent readily promoted amide
group formation in DMF with DIPEA, but only phosphate of target
structure 4 was retrieved (single product was obtained, showing
MW = 698 in MS spectrum and two excessive ethyl groups in 1H
l
l) and DIPEA (20 l) were added. After stir-
l
ring in dark at room temperature for 3 days, the reaction was com-
pleted. Reaction mixture was evaporated to dryness and residue
was submitted to the column chromatography on reverse phase,
using gradient of MeOH in water as eluent. Fractions containing
target compound were combined and evaporated, yielding 7 mg
of compound 4 which was further purified by column chromatog-
raphy on silica gel (CH2Cl2/MeOH). Target compound 4 (3.7 mg,
34%) was obtained as amorphous white solid. Prior to measure-
ments of NMR spectra, the compound was further purified from
trace impurities by semi-preparative HPLC (gradient of MeOH in
1
water). H NMR (CD3OD, 600 MHz): 7.60 s, 1 H (H-200); 6.89 br s,
1 H (H-500); 4.38 d, 1 H, J = 7.7 (H-10); 3.97 br s, 1 H (H-3b);