An unexpected synthesis of novel oxygen-bridged 1,5-benzothiazepine derivatives and their reductive five-membered-ring opening

Article abstract of DOI:10.1007/s007060050320

A convenient procedure is reported for the preparation of benzofuro-annelated 2-phenyl-1,5-benzothiazepine derivatives by oxidative cyclocondensation of phenolic β-diketones with o-aminothiophenol in DMSO. The regiochemistry of these compounds is proven by HMBC signals and the existence of a five-bond ¹⁹F,¹³C-2 coupling. Surprisingly, treatment with LiAlH₄ at room temperature led to a double reduction under opening of the five-membered ring. Refluxing such solutions with a higher amount of LiAlH₄ gave rise to a further reduced derivative possessing the trans-configuration. All structures (regio-and stereochemistry) were assigned on the basis of NMR spectroscopic data.

Full text of DOI:10.1007/s007060050320

Monatshefte fuÈr Chemie 131, 393±400 (2000)
An Unexpected Synthesis of Novel
Oxygen-Bridged 1,5-Benzothiazepine
Derivatives and their Reductive
Five-Membered-Ring Opening
Roshan Ahmad1;Ã, Mohammad Zia-ul-Haq1, Shahid Hameed1,
Humaira Akhtar1, and Helmut Duddeck2;Ã
1
Department of Chemistry, Quaid-i-Azam University Islamabad, Islamabad, Pakistan
2
University of Hannover, Institute of Organic Chemistry, D-30167 Hannover, Germany
Summary. A convenient procedure is reported for the preparation of benzofuro-annelated 2-phenyl-
1,5-benzothiazepine derivatives by oxidative cyclocondensation of phenolic ꢀ-diketones with o-
aminothiophenol in DMSO. The regiochemistry of these compounds is proven by HMBC signals and
the existence of a ®ve-bond 19F,13C-2 coupling. Surprisingly, treatment with LiAlH4 at room
temperature led to a double reduction under opening of the ®ve-membered ring. Re¯uxing such
solutions with a higher amount of LiAlH4 gave rise to a further reduced derivative possessing the
trans-con®guration. All structures (regio- and stereochemistry) were assigned on the basis of NMR
spectroscopic data.
Keywords. Benzofuro-annelated 1,5-benzothiazepines; 4,5-Dihydro-1,5-benzothiazepines; 2,3,4,5-
Tetrahydro-1,5-benzothiazepines; Cyclocondensation; Regiochemistry; Stereochemistry; Conforma-
tional analysis.
Introduction
Substituted benzothiazepines are known for their calcium antagonistic activity. A
well-known example is diltiazem [1], an accepted drug for the treatment of
cardiovascular disorders. During the last ®ve years a number of compounds have
been patented mostly in Japan and USA which have shown prominent calcium
antagonistic activity. Mostly, they are derivatives of benzo-1,5-thiazepine-2-one
bearing a substituted phenyl ring at C-2 [2] and alkoxycarbonyl or alkoxysulfonyl
groups at C-3 [3]. The nitrogen atom at position 5 carries in most cases disubstituted
aminoalkyl groups [4]. Oxygen-bridged sterically constrained heterocycles with
complicated conformational behaviour are also gaining interest [5]. The syntheses of
oxygen-bridged 1,5-benzothiazepine derivatives and dihydro analogues have been
reported starting from chalcones [6] and from ꢁ,ꢀ-enones [7].
Ã
Corresponding authors

394
R. Ahmad et al.
Results and Discussion
Cyclocondensation reaction of some phenolic ꢀ-diketones (1) with o-aminothio-
phenol in DMSO has been reported to give 1,5-benzothiazepines with an o-
hydroxyphenyl side chain (2) [8]. Adopting this procedure we attempted to react
the phenolic ꢀ-diketones (1a±1d) [9] with o-aminothiophenol. Surprisingly,
however, the cyclocondensation proceeded under oxidation to give oxygen-bridged
1,5-benzothiazepines 3a±d in reasonable yields (Fig. 1), whereas the respective
compounds 2a±d were not detected. This offers a simple and effective synthetic
pathway to this heterocyclic system.
Fig. 1. Synthesis of compounds 3a±d, 5a, b, and 7a, b

Synthesis of Novel Benzothiazepine Derivatives
395
The structures of compounds 3a±d were con®rmed by spectroscopic
techniques. The molecular ion peak in the mass spectrum of 3a was observed at
m/z ˆ 327 instead of m/z ˆ 329 for 2a; the molecular formula based on the high
resolution mass spectrum was found to be C21H13NOS (and not C21H15NOS). The
mass spectra of products 3b±d showed analogous differences. Elemental analyses
Table 1. 1H and 13C chemical shifts (ppm) of compounds 3a±d; in CDCl3, relative to internal TMS;
n
n
values in parentheses are coupling constants J(19F,1H) or J(19F,13C), respectively, in Hz
3a (X ˆ H)
3b (X ˆ Cl)
3c (X ˆ F)
3d (X ˆ OCH3)
7.15 dd
H-6
7.16 dd
7.01 td
6.90 td
6.74 dd
7.09 dd
7.43 td
7.25 td
8.25 dd
7.71 (AA0)
7.16 dd
7.03 td
6.92 td
6.76 dd
7.08 dd
7.43 td
7.23 td
8.23 dd
7.68 (AA0)
7.16 dd
7.02 td
6.91 td
6.75 dd
7.09 dd
7.43 td
7.25 td
8.24 dd
7.72 (AA0)
(8.9)
H-7
7.01 td
H-8
6.90 td
H-9
6.75 dd
7.09 dd
7.43 td
H-30
H-40
H-50
H-60
H-200/600
7.24 td
8.24 dd
7.681 (AA0)
H-300/500
7.49 (XX0)
7.44 (XX0)
7.16 (XX0)
6.99 (XX0)
(8.5)
H-400
C-2
7.47 (Y)
106.1
±
±
±
106.4
106.1
(0.9)
105.2
C-3
146.6
148.7
141.4
128.9
127.1
127.0
125.0
121.1
120.4
154.5
117.0
132.0
124.9
125.0
132.5
145.3
148.4
141.3
129.1
127.3
127.1
125.1
120.8
120.3
154.4
117.0
130.9
125.1
125.1
132.1
145.3
148.9
141.3
129.0
127.2
127.1
125.1
120.8
120.3
154.4
117.0
132.1
125.0
125.1
128.6
(3.5)
146.6
148.9
141.5
128.8
127.1
126.9
125.0
121.2
120.4
154.5
117.0
132.0
124.8
125.0
124.8
C-4
C-5a
C-6
C-7
C-8
C-9
C-9a
C-10
C-20
C-30
C-40
C-500
C-6
C-100
C-200/600
C-300/500
C-400
128.0
128.5
130.1
128.8
129.4
135.9
130.2
(8.6)
129.6
113.9
160.8
115.7
(21.9)
163.3
(251.3)
±
OCH3
OCH3
±
±
±
±
3.86
55.4
±

396
R. Ahmad et al.
also favour the structures 3. Eight tertiary carbons instead of seven were observed
1
in the 13C NMR spectra. The assignment of the H and 13C NMR signals is based
on extensive 2D spectroscopy (COSY, HMQC, and HMBC). The regiochemistry of
the reaction, i.e. the fact that the products are of type 3 and not of type 4, is
unambiguously proven by two spectroscopic evidences: (a) there is a HMBC peak
correlating C-4 with its high ꢂ-value (ca. 148.5 to 149 ppm; typical for C=N) to H-
60; (b) in 3c a small but signi®cant 19F,13C coupling of 0.9 Hz can be observed for
the signal of C-2 at ꢂ ˆ 106.1 ppm (typical for a ꢀ-carbon within an enol fragment).
It should be noted that the compounds 3 differ from related ones reported earlier
[5±7] which contain reduced atoms within the thiazepin ring.
Compounds 3a and 3b ± as examples ± were reduced by of LiAlH4 (3 M) to
yield 5a and 5b (Fig. 1). Although we expected the reduction of both double bonds
(6a and 6b), only the C=N double bond was reduced and, to our surprise, the
oxygen bridge was opened as well. In the 13C NMR spectrum of 5a there is only
1
one peak in the aliphatic region at ꢂ ˆ 61.2 ppm (CH). The H NMR spectrum
shows two sharp signals at ꢂ ˆ 5.95 and 4.97 ppm besides the signals in the
aromatic and ole®nic region. The proton at ꢂ ˆ 4.97 ppm is attached to the carbon
at ꢂ ˆ 61.2 ppm as shown by the HMQC spectrum. This proves the presence of a
-(H-4)-(C-4)-N- fragment. The hydrogen at ꢂ ˆ 5.95 ppm is attached to the carbon
at ꢂ ˆ 126.0 ppm which is a typical sp2-CH region. The assignment of this CH
fragment to H-3/C-3 is proven by the HMBC spectrum correlating C-3 and H-4. H-
4 showed a correlation (three-bond coupling) to the aromatic carbon carrying the
1
OH group (C-20). Finally, the H NMR signals at ꢂ ˆ 8.30 and 4.45 ppm are
assigned to OH and NH protons. Analogous spectroscopic results were obatined for
the chloro derivative 5b.
The suprisingly small vicinal coupling constant 3J(H-3,H-4) ˆ 1.3 Hz indicates
that the respective torsion angle (H-3/C-3/C-4/H-4) is close to 90ꢀ. This allows to
de®ne the preferred conformation of 5a and 5b as a half-chair with a quasi-
equatorial phenol substituent at C-4 (Fig. 2).
In an attempt to achieve a fully reduced 2,3,4,5-tetrahydro-1,5-benzothiazepine
derivative, a higher amount of LiAlH4 was used for the reduction of 3b as an
example, and the reaction mixture was re¯uxed for 30 minutes after stirring
overnight at room temperature. Thereby, a new reduction product 7b (Fig. 1) was
isolated. This compound contains an aliphatic fragment -NH-C4H-C3H2-C2H-S-
Fig. 2. Preferred conformation of 5a, b

Synthesis of Novel Benzothiazepine Derivatives
397
Fig. 3. Conformations of 7b
Table 2. Torsion angles (ꢀ) of conformers A (chair) and B (twist-boat) of 7b, calculated by the
3
semiempirical AM1 method; experimental coupling constants JH,H (Hz) are given for comparison
Fragment
A (chair)
B (twist-boat)
3JH,H/Hz
C9a-S-C2-C3
S-C2-C3-C4
76.3
78.0
69.3
73.1
165.4
76.5
58.1
57.9
0.1
25.9
83.6
40.5
54.2
156.9
86.2
35.2
80.5
0.4
C2-C3-C4-N
C5a-N-C4-C3
H2-C2-C3-H3
H2-C2-C3-H3a
H4-C4-C3-H3
H4-C4-C3-H3a
12.0
3.1
4.2
4.0
0
0
C4-C1 -C2 -O
H-N-C4-H4
41.7
66.8
173.4
94.8
±
00
00
S-C2-C1 -C2
which could be identi®ed from 1D and 2D NMR spectra. The corresponding 1H,1H
2
3
3
coupling constants are: J(H-3,H-3a) ˆ 15.3 Hz, J(H-2,H-3) ˆ 12.0 Hz, J(H-2,H-
3
3
3a) ˆ 3.1 Hz, J(H-3,H-4) ˆ 4.2 Hz and J(H-3a,H-4 ˆ 4.0 Hz The most diagnostic
one of them is 3J(H-2,H-3) ˆ 12.0 Hz clearly indicates the anti-periplanar orientation
of the hydrogens involved. There are only two conceivable conformations of the
trans-con®gurated 7b ± the cis-con®guration can be excluded ± which ®t to this
combination of couplings, structures A (chair) and B (twist-boat), as depicted in Fig. 3.
Semiempirical calculations (AM1) of 7b slightly favour the twist-boat
conformation B, but the difference of ca 1 kJ/mol is possibly within the error limits
of the calculation. If, however, the geometry of the calculated structures (Table 2)
and the experimental vicinal H,1H coupling constants are evaluated using the
1
Karplus equation, the chair conformation is more plausible. Especially, the coupling
constant of 4.0 Hz for the H4-C4-C3-H3a fragment ®ts much better to the chair
(torsion angle ˆ 58ꢀ) than to the half-boat conformation (torsion angle ˆ 81ꢀ).
Experimental
Melting points were determined on a Gallenkamp digital melting point apparatus and are
uncorrected. IR spectra were recorded on a Hitachi 270-50 IR spectrophotometer. 1H and 13C NMR

398
R. Ahmad et al.
spectra were recorded on a Bruker AM-250 (3a±d) and a Bruker DRX-400 spectrometer (5a, b, 7b)
using CDCl3 as solvent. Bruker standard software was employed for the multipulse experiments
(DEPT, COSY, HMQC, HMBC). Electron-impact mass spectra were recorded on Varian Mat CH-5.
UV spectra were recorded on a Shimadzu UV-265 spectrophotometer. Column chromatography was
performed on silica gel (60±120 mesh; Merck).
General procedure for the cyclocondensation of ꢀ-diketones with o-aminothiophenol
1-(20-Hydroxyphenyl)-3-(phenyl)- or the respective -( p-substituted phenyl)-1,3-propanediones (1a±
d) were re¯uxed with o-aminothiophenol using DMSO as solvent. After 5 h the mixture was poured
onto crushed ice and extracted with ethylacetate. The organic layer was dried (anhydrous Na2SO4)
and evaporated in vacuo. The solid obtained was chromatographed on a silica gel column using
benzene as eluent. The compounds were recrystallized from absolute ethanol.
Benzofuro-[10,20-c]-2-phenyl-1,5-benzothiazepine (3a; C21H13NOS)
Yield: 69.7%. m.p.: ˆ 166±167ꢀC; IR (KBr): ꢃ~ˆ 3982, 3940, 3652, 3544, 3334, 3250, 3196, 3034,
2848, 2680, 2260, 2128, 2026, 1962, 1938, 1896, 1839, 1602, 1530, 1488, 1458, 1446, 1347, 1290,
1260, 1221, 1134, 1056, 1020, 915, 741, 681 cm 1; UV (MeOH): ꢄmax(") ˆ 511 (56925), 323
‡
(20780), 286 (21080), 253 (31300) nm; for NMR data, see Table 1; EI-MS: m/z (%) ˆ 327 (M ,
100), 326 (8), 299 (9), 297 (8), 295 (13), 267 (4), 222 (4), 164 (8), 148 (9).
Benzofuro-[10,20-c]-7-(400-chlorophenyl)-1,5-benzothiazepine (3b; C21H12ClNOS)
Yield: 73.4%. m. p.: 182±183.5ꢀC; IR (KBr): ꢃꢀˆ 3970, 3844, 3778, 3670, 3448, 3412, 3190, 3040,
2914, 2800, 2668, 2284, 2146, 1983, 1896, 1602, 1533, 1482, 1461, 1395, 1344, 1290, 1260, 1224,
1134, 1083, 1062, 1014, 738, 717 cm 1; UV (MeOH): ꢄmax(") ˆ 511 (50615), 328 (17380), 287
‡
(17760), 252 (24550) nm; for NMR data, see Table 1; EI-MS: m/z (%) ˆ 361 (M , 100), 360 (7), 333
(8), 297 (10), 329 (8), 250 (10), 222 (6), 164 (5), 148 (23).
Benzofuro-[10,20-c]-2-(400-¯uorophenyl)-1,5-benzothiazepine (3c; C21H12FNOS)
Yield: 64.1%; m.p.: 188.1±189ꢀC; IR (KBr): ꢃ ˆ 3796, 3658, 3262, 3046, 2908, 2842, 2254, 2152,
2086, 1893, 1602, 1536, 1497, 1461, 1404, 1344, 1266, 1218, 1155, 1137, 1077, 1056, 1014,
741 cm 1; UV (MeOH): ꢄmax(") ˆ 508 (50535), 324 (17000), 288 (16920), 251 (25195) nm; for
‡
NMR data, see Table 1; EI-MS: m/z (%) ˆ 345 (M , 100), 344 (9), 317 (10), 313 (10), 285 (4), 250
(6), 222 (4), 148 (4).
Benzofuro-[10,20-c]-2-(400-methoxyphenyl)-1,5-benzothiazepine (3d; C22H15NO2S)
Yield: 72.5%; m.p.: 150.8±152.5ꢀC; IR (KBr): ꢃ~ˆ 3916, 3604, 3472, 3040, 2914, 2830, 2056, 1929,
1899, 1605, 1533, 1500, 1448, 1347, 1302, 1239, 1173, 1137, 1104, 1059, 1023, 930, 813,
744 cm 1; UV (MeOH): ꢄmax(") ˆ 517 (68530), 353 (22310), 297 (19200), 252 (32235) nm; for
‡
NMR data, see Table 1; EI-MS: m/z (%) ˆ 357 (M , 100), 342 (10), 314 (10), 282 (10), 178 (21),
164 (3).
Syntheses of 5a and 5b
To a stirred suspension of 3 mmol LiAlH4 in 30 m3 dry THF, a solution of 1 mmol 3a or 3b in THF
was added dropwise. The reaction mixture was stirred thoroughly and monitored by TLC (petroleum

Synthesis of Novel Benzothiazepine Derivatives
399
ether:ethylacetate ˆ 4:1). After stirring overnight at room temperature the reaction mixture was
cooled on ice. Water was added dropwise to destroy LiAlH4, followed by addition of 10% H2SO4
until the solution was clear. THF was evaporated in vacuo, and the remaining aqueous solution was
extracted with ethyl acetate. The organic layer was dried over anhydrous Na2SO4 and evaporated to
dryness. The residue was puri®ed by column chromatography using petroleum ether and ethylaceta
te (7:1).
2-Phenyl-4-(20-hydroxyphenyl)-4,5-dihydro-1,5-benzothiazepine (5a; C21H17NOS)
Yield: 39%; m.p.: 120ꢀC; 1H NMR (CDCl3, ꢂ, 400 MHz): 4.45 (bs, 1H, NH), 4.97 (H-4, d,
J ˆ 1.3 Hz), 5.92 (H-3, d, J ˆ 1.3 Hz), 6.68 (H-30, d), 6.85 (H-8, dd), 6.85 (H-40, dd), 6.92 (H-6, d),
6.97 (H-50, dd), 7.09 (H-60, d), 7.12 (H-9, d), 7.16 (H-400, dd), 7.24 (H-7, dd), 7.28 (H-300/500, dd),
7.34 (H-200/600, d), 8.30 (bs, 1H, OH) ppm; 13C NMR (CDCl3, ꢂ, 100.6 MHz): 61.2 (C-4), 117.6 (C-
6), 118.6 (C-30), 120.3 (C-8), 120.4 (C-10), 122.6 (C-40), 123.1 (C-9a), 123.8 (C-3), 126.0 (C-50),
126.3 (C-60), 127.5 (C-400), 128.4 (C-300/500), 129.1 (C-200/600), 130.4 (C-9), 130.6 (C-7), 131.4 (C-2),
‡
135.6 (C-100), 141.1 (C-5a), 155.8 (C-20) ppm; EI-MS: m/z (%) ˆ 331 (M , 70), 298 (5), 254 (16),
240 (100), 228 (22), 207 (38), 196 (13), 178 (16), 165 (6), 149 (7), 136 (15), 122 (9), 109 (11), 91
(9), 77 (18).
2-(400-Chlorophenyl)-4-(20-hydroxyphenyl)-4,5-dihydro-benzothiazepine (5b; C21H16ClNOS)
Yield: 45%; m.p.: 156ꢀC; 1H NMR (CDCl3, ꢂ, 400 MHz): 4.5 (broad, 1H, NH), 4.97 (H-4, d,
J ˆ 1.3 Hz), 5.92 (H-3, d, J ˆ 1.3 Hz), 6.68 (H-30, d), 6.91 (H-40, dd), 6.91 (H-8, dd), 6.97 (H-6, d),
7.02 (H-50, dd), 7.15 (H-60, d), 7.19 (H-9, d), 7.25 (H-200/600, d), 7.29 (H-7, dd), 7.30 (H-300/500, dd),
ca. 7.8 (broad, 1H, OH) ppm; 13C NMR (CDCl3, ꢂ, 100.6 MHz): 61.5 (C-4), 117.6 (C-6), 118.3 (C-
30), 120.2 (C-8), 120.2 (C-10), 122.6 (C-40), 123.4 (C-3 and C-9a), 126.1 (C-50), 126.3 (C-60), 134.4
(C-400), 129.1 (C-300/500), 130.2 (C-9), 130.4 (C-200/600), 130.5 (C-7), 132.9 (C-2), 134.4 (C-100), 141.6
‡
(C-5a), 155.9 (C-20) ppm; EI-MS: m/z (%) ˆ 365/367 (M ‡32/10), 331 (30), 272/274 (11/3), 254
(16), 240 (100), 227 (29), 207 (60), 178 (38), 149 (21), 136 (48), 122 (34), 109 (27), 97 (38), 71 (38),
207 (38), 196 (13), 165 (6), 91 (9), 77 (18), 69 (45), 57 (47).
trans-2-(400-Chlorophenyl)-4-(20-hydroxyphenyl)-2,3,4,5-tetrahydro-1,5-benzothiazepine
(7b; C21H18ClNOS)
The synthetic procedure was analogous to those for 5a and 5b except for a higher amount of LiAlH4
employed (4 mmol) and the fact that after overnight stirring at room temperature the reaction mixture
was re¯uxed for 30 min before work-up.
Yield: 65%; m.p.: 170ꢀC; IR (KBr) ꢃ~ˆ 3350 (br, NH/OH), 3044, 2944, 1584, 1488, 1456, 1248,
1
1092, 1012, 824, 752 cm 1; H NMR (CDCl3, ꢂ, 400 MHz): 2.22 (H-3a, ddd), 2.48 (H-3, ddd), 4.1
(NH), 4.39 (H-2, dd), 4.81 (H-4, dd), 6.88 (H-50, td), 7.01 (H-40, td), 7.01 (H-30, dd), 7.04 (H-60, dd),
7.15 (H-8, td), 7.15 (H-200/600, m), 7.24 (H-300/500, m), 7.28 (H-6, dm), 7.31 (H-7, td), 7.57 (H-9, dd),
10.1 (OH) ppm; 13C NMR (CDCl3, ꢂ, 100.6 MHz): 118.1 (C-30), 120.3 (C-50), 122.5 (C-6), 125.3 (C-
8) 125.5 (C-10), 126.4 (C-9a), 128.7 (C-200/600), 128.8 (C-60), 129.2 (C-300/500), 129.7 and 129.8 (C-7
and C-40), 133.2 (C-400), 134.2 (C-9), 144.0 (C-100), 146.4 (C-5a), 159.6 (C-20) ppm; EI-MS: m/z
‡
(%) ˆ 368/370 (M , 31/13), 334/336 (32/11), 243 (41), 229 (40), 228 (100), 149 (7), 136 (73), 125
(12), 109 (14), 77 (8).
Acknowledgements
The authors are thankful to Pakistan Science Foundation for ®nancial assistance. This work was
supported by Deutscher Akademischer Austauschdienst and Fonds der Chemischen Industrie.

400
R. Ahmad et al.: Synthesis of Novel Benzothiazepine Derivatives
References
[1] Kurokawa M, Sato F, Honda Y, Uno H (1991) J Med Chem 34: 593
[2] Bousquet A, Charpiol A (1991) Eur Pat Appl Ep 381, 570; Chem Abstr (1991) 114: 102070n
[3] Takatani M (1987) Jpn Kobai Tokkyo Koho JP 61, 130, 277; Chem Abstr (1987) 106: 50265t
[4] Mohecsi E, O'Brien PJ, US 4, 912, 102; Chem Abstr (1990) 113: 172061p
[5] Svetlik J, Turecek F (1988) J Chem Soc Perkin Trans 2053
[6] Svetlik J, Hanus V, Bella J (1989) Liebigs Ann Chem 91
Â
Â
Â
È È
[7] Toth G, Szollosy A, Levai A, Duddeck H (1982) Org Magn Reson 20: 133
[8] Umesh CP, Gaur BS, Kumar V (1986) J Indian Chem Soc 63: 680
[9] Ahmad R, Malik MA, Zia-ul-Haq M (1990) J Chem Soc Pak 12: 352
Received November 22, 1999. Accepted December 21, 1999