ORIGINAL ARTICLES
Scheme 9 (a) KOH, EtOH
O
O
NH
a
S
N
O
+
NH
O
Ts
S
a
1
14
N
O
HO
25
O
O
28
N
2
5
16
21
1
3
15
12
17
6
7
13
18
14
15
22
26
24
N
19
27
20
S
23
4
N
O
10
9
8
11
O
IV
1
that direct detection experiments like H-13C HETCOR or
1H-13C COLOC. Nevertheless resolution in F1 domain
was so adequate to distinguish and assign even the
13C NMR signals of quaternary carbons C-10 and C-25
differs each other only 0.2 ppm. Signals of methyl groups
C-1 and C-34 were also unambiguously assigned on the
basis of combination of HSQC and HMBC optimized for
50% of compound I, ca 10% of compound II and ca 6% of compound
III. The rest comprised pioglitazone and several small and unknown impu-
rities. This material was chromatographed through column with reverse
phase silica gel in the mixture acetonitrile/formate buffer (55 : 45). Fraction
purity was monitored by TLC. Pioglitazone had Rf ¼ 0.74, comp. I had
Rf ¼ 0.51, comp. II had Rf ¼ 0.41 and comp. III had Rf ¼ 0.36. Fractions
containing compound I were combined, concentrated and residue extracted
with dichloromethane. The product received by dichloromethane evaporat-
ing was recrystallized from methanol. White compound I was yielded.
Melting point ¼ 122.5–124 ꢀC. MS m/e: 581. Anal. Calcd. for
3
three-bond coupling constant J(C, H) ¼ 8 Hz.
C
34H35N3O4S: C, 70.20; H, 6.06; N, 7.22; O, 11.00; S, 5.51 found: C,
The identification of pioglitazone impurities, when piogli-
tazone was prepared by Meerwein arylation, gives the op-
portunity for decreasing their content in the active pharma-
ceutical ingredient and consequently to fulfil the drug
authorities requirements for drug production.
69.85; H, 5.85; N, 6.95.
Other compounds II and III were received from other chromatographic
fractions by the similar manner. Compounds I, II and III have on TLC
different retention factors. The difference in retention factors of compounds
II and III is not suitable for quite a full separation on column chromato-
graphy described in this communication. We have got only the compound
II a little contaminated by compound III for NMR determination and the
small amount of compound III contaminated by compound II. For com-
pound I all 1D and 2D techniques for unambiguous assignment of all
NMR signals were used. For compound II only 1D 1H and 13C and H,H-
COSY techniques were used, so some signals are assigned tentatively and
assignment could be interchanged. For compound III only 1D and 2D
1H NMR spectra were measured (Table 1). Nevertheless the structures of
compounds II and III were proven.
3. Experimental
3.1. Chemicals and apparatus
Nuclear magnetic resonance was measured in CDCl3 solutions on BRUKER
AVANCE 500 spectrometer, (Bruker Analytische Messtechnik GmbH) at
500.13 MHz (1H) and 125.77 MHz (13C) respectively. 1D the experiments
were carried out using a 5 mm TXO probe. For the 2D experiments the
inverse TBI probe was used. All experiments were performed in the sol-
vent CDCl3. For the assignment the 1H-13C g-HSQC and 1H-13C g-HMBC
from Bruker library were used. The 1H-13C g-HSQC dataset consisting of
2K and 256 points in F2 and F1 dimensions, respectively was chosen. The
number of scans was 8 for each t1 increment. The time domain data was
zero-filled to 1024 and 2K data points in F1 and F2 dimensions, respec-
tively, and processed with sinusoidal squared sine-bell window function in
both the dimensions. The gradient-selected 1H-13C HMBC was performed
using 4K and 512 points in F2 and F1 dimensions, respectively. The num-
ber of scans was 32 for each t1 increment. The acquired data was zero-
filled to 1024 and 2K data points in F1 and F2 dimensions, respectively,
and processed with sinusoidal squared sine-bell window function in both
the dimensions.
3.3. 1,3-Bis-{4-[2-(5-ethyl-pyridin-2-yl)-ethoxy]-phenyl}-isothiourea (21)
4-[2-(5-Ethylpyridin-2-yl)ethoxy]aniline (6) (40 g, 0.165 mol) and CS2
(25.2 g, 0.332 mol) were dissolved in absolute ethanol (65 ml). The reac-
tion mixture was maintained at 50 ꢀC for 2 h. After finishing the reaction
all residual solvents were removed. The residue was dissolved in dichloro-
methane. Dichloromethane solution was washed with 1 M solution of HCl
acid and dried over sodium sulfate. A slightly yellowish crude solid com-
pound was yielded. The crude product was recrystallized from ethanol to
give 17.8 g (42%) of pure compound (21). M.p. 84.5–86 ꢀC; 1H NMR
(CDCl3, 25 ꢀC) d 9.44 (s, 2), 8.38 (s, 2), 7.60 (dd, J ¼ 7.7, 1.6 Hz, 2),
7.31 (m, 6), 6.91 (d, J ¼ 7.5 Hz, 4), 4.33 (t, J ¼ 6.3 Hz, 4), 3.15
(t, J ¼ 6.3 Hz, 4), 2.52–2.65 (m, 4), 1.19 (t, J ¼ 7.5 Hz, 6); 13C NMR
(CDCl3, 25 ꢀC) d 180.3, 155.9, 155.7, 148.7, 136.9, 135.9, 132.5, 126.2,
123.3, 114.4, 67.1, 37.0, 25.2, 15.6.
Mass spectra were recorded on API 3000 LC MS/MS apparatus equipped
with E-spray ionization, (PE SCIEX).
Preparative column chromatography was made on silica gel 60 (Fluka) and
reverse phase of silica Lichroprep RP-18e (Merck).
Formate buffer was prepared by dissolving of 6.3 g of ammonium formate
in 1 L of distilled water. The pH was adjusted to 6 by the addition of
diluted (10%) formic acid.
Thin layer chromatography was made on FPKG60 F254 plates (Merck) in
the mixture CHCl3-methanol-NH4OH (25%) (30 : 15 : 0.5).
HPLC was made on Merck Hitachi LaChrom apparatus, endecaped (5 mm)
Purospher1STAR RP-18 column, detection Diode Array Detector L-
7450A.
3.4. 5-{4-[2-(5-Ethyl-pyridin-2-yl)ethoxy]benzyl}-3-{4-[2-(5-ethyl-pyridin-
2-yl)ethoxy] phenyl}-1,3-thiazolidine-2,4-dione (20)
Disubstituted thiourea 21 (11.5 g, 0.02 mol) and bromoester 9 (7.9 g,
0.02 mol) were dissolved in ethanol (125 ml). Reaction mixture was heated
for 7 h under reflux and then evaporated to dryness. The rest was extracted
three times with 40 ml of ethylacetate. Ethylacetate solution was evapo-
rated to dryness and the rest was heated for 5 h in 100 ml of 1 M HCl.
Then reaction mixture was cooled to the room temperature and extracted
three times with 50 ml of dichloromethane. Dichloromethane extract was
washed with a saturated aqueous solution of NaHCO3 (10%), dried over
sodium sulfate and concentrated. 6 g of brown solid was obtained after
evaporation of the solvent. A white product (20) was yielded by recrystalli-
zation from cyclohexane: M.p. 82.5–83.5 ꢀC; 1H NMR (CDCl3, 25 ꢀC) d
8.39 (s, 2), 7.45 (dd, J ¼ 7.7, 1.6 Hz, 2), 7.14–7.26 (m, 8), 6.86 (d,
J ¼ 7.5 Hz, 2), 4.54 (dd, J ¼ 8.3, 3.9 Hz, 1), 4.31–4.38 (m, 4), 3.23 and
3.42 (overlapped and dd, J ¼ 13.9, 3.9 Hz, 1), 3.23–3.28 (m, 4), 2.58–
2.67 (m, 4), 1.24 (t, J ¼ 7.5 Hz, 6); 13C NMR (CDCl3, 25 ꢀC) d 173.4,
170.5, 159.3, 158.4, 155.6, 155.4, 149.0, 149.0, 137.7, 137.1, 137.0,
135.7, 132.1, 130.6, 130.6, 128.3, 127.2, 125.2, 115.3, 114.8, 67.5, 67.3,
51.3, 37.8, 37.6, 37.4, 25.7, 15.3.
3.2. 5-{4-[2-(5-Ethyl-6-{4-[2-(5-ethylpyridin-2-yl)ethoxy]fenyl}pyrid-2-yl)-
ethoxy]benzyl}-1,3-thiazolidine-2,4-dione (I);
5-{4-[2-(5-ethyl-4-{4-[2-(5-ethylpyridin-2-yl)ethoxy]fenyl}pyrid-2-yl)-
ethoxy]benzyl}-1,3-thiazolidine-2,4-dione (II);
5-{6,40-bis-[2-(5-ethyl-pyridin-2-yl)ethoxy]biphenyl-3-ylmethyl}-
1,3-thiazolidine-2,4-dione (III)
Pioglitazone prepared by the procedure of Meerwein arylation was crystal-
lized from DMF/H2O (5/1). Mother liquor was concentrated and the resi-
due was extracted by toluene. The rest of pioglitazone was removed from
the mixture by concentration of toluene extract and filtration of solid. Eva-
poration of toluene filtrate finally yielded the solid material containing ca
Pharmazie 62 (2007) 8
583