Synthesis of the hypotensive agent 8-amino-7-[2-hydroxy-3-morpholinopropyl] theophylline (P23) and analogs

Article abstract of DOI:10.1002/jhet.51

Synthesis of a number of 7, 8-disubstituted theophyllines including enantiomers of the hypotensive agent P23 is described.

Full text of DOI:10.1002/jhet.51

March 2009
Synthesis of the Hypotensive Agent 8-Amino-7-[2-hydroxy-3-
morpholinopropyl]theophylline (P23) and Analogs
191
Marek T. Cegla,a* Joanna Potaczek,a Marek Zylewski,a
and Lucjan Strekowskib
aDepartment of Organic Chemistry, Faculty of Pharmacy, Jagiellonian University Medical College,
PL 30-688 Krakow, Poland
bDepartment of Chemistry, Georgia State University, Atlanta, Georgia 30302-4098
*E-mail: mfceglam@cyf-kr.edu.pl
Received July 25, 2008
DOI 10.1002/jhet.51
Published online 23 March 2009 in Wiley InterScience (www.interscience.wiley.com).
Synthesis of a number of 7,8-disubstituted theophyllines including enantiomers of the hypotensive
agent P23 is described.
J. Heterocyclic Chem., 46, 191 (2009).
INTRODUCTION
Compound 2 [2] and the reagent 9 [3] were synthe-
sized by using the published procedures. Treatment of 1
with 2 gave the expected adduct 3 in 43% yield. It was
believed that a selective nucleophilic displacement of
the tosylate function in 3 by the reaction with morpho-
line followed by the addition of ammonia to the
C¼¼NABr moiety of the resultant intermediate product
and then elimination of bromide anion would produce
the desired product P23. Unfortunately, the treatment of
3 with morpholine resulted in the displacement and
addition reactions simultaneously to give a dimorpholino
derivative 4. On the other hand, it was thought that the
addition reaction of less sterically hindered benzylamine
with the C¼¼NABr moiety of 3 would be a selective
process. The benzyl group could be removed later by
hydrogenation. To our surprise a tricyclic product 5 was
formed instead, showing that the third ring system is
formed by the benzylamine-aided ionization of the
hydroxyl function in the side chain followed by addition
of the alkoxide anion to the C¼¼NABr moiety and a
subsequent elimination of bromide. The treatment of 3
with ammonia also resulted in cyclization, however,
without nucleophilic displacement of the tosylsulfonyl
group, to give 6 as the final product. Another surprising
result was obtained upon an attempted nucleophilic dis-
placement of the tosylsulfonyl group in 6 by treatment
with morpholine. Thus, the tosylsulfonyl group was
retained in the bicyclic product 7. The loss of the third
ring can be explained in terms of the addition reaction
N-Methyl derivatives of xanthine including caffeine,
theobromine, and theophylline show diverse biological
activities. An additional substitution of these compounds
results in modulation of their biological responses. In par-
ticular, a 7,8-disubstituted theophylline P23 (structure in
Scheme 1) was synthesized two decades ago and protected
by patents as a potent hypotensive agent that acts on the
cardiovascular system [1]. The original synthesis of P23
was based on a reaction of theophylline with racemic epi-
chlorohydrin to give the racemic intermediate and final
products. No attempts were made to synthesize enantiom-
ers of P23, which is a requirement of the current drug de-
velopment process. More specifically, two enantiomers of
a chiral drug candidate must be biologically assayed.
Since the patent protection of the chemistry leading to the
racemate of P23 has recently expired, the patented infor-
mation is in the public domain. Therefore, it was of inter-
est to review the previously published chemistry in an
attempt to synthesize enantiomers of P23. Unfortunately,
we have found that the use of enantiomerically pure epi-
chlorohydrin instead of a racemic substrate in the reaction
with 8-bromotheophylline (1) results in an extensive race-
mization of the intermediate adduct. Additional methodol-
ogies are illustrated in Scheme 1.
To test the feasibility of the designed synthetic routes,
the first experiments involved the reaction of 1 with ra-
cemic reagents, namely 2,3-epoxypropyl 4-toluenesulfo-
nate (2) and 2,3-epoxypropylmorpholine (9).
C
V
2009 HeteroCorporation

192
M. T. Cegla, J. Potaczek, M. Zylewski, and L. Strekowski
Vol 46
Scheme 1
of morpholine to the C¼¼N function of 6 followed by
elimination of alkoxide anion from the resultant adduct.
The successful preparation of P23 that is also amend-
able to the synthesis of individual enantiomers is based
on the reaction of 8-bromotheophylline (1) with N-(2,3-
epoxypropyl)morpholine (9). Reagent 9 is not a com-
mercial product and was obtained by the reaction of
morpholine with epichlorohydrin (8) which is available
in the form of a racemate or individual enantiomers. It
has been suggested previously that the reaction of mor-
pholine with an optically active substrate (S)-(þ)-8 is
accompanied by extensive racemization of the epoxya-
mino product (S)-(ꢀ)-9 [3-5]. We have shown that the
racemization can be minimized by using mild condi-
tions. As can be seen from the experimental part, the
absolute values of the optical rotations for the enantiom-
ers of 9 obtained by us are virtually identical. The
greater optical purity of our sample of (ꢀ)-9 is indicated
by the absolute rotation, [a]D ¼ ꢀ25.4ꢁ (c ¼ 10% in
toluene), as opposed to the value published previously,
[a]D ¼ ꢀ20.2ꢁ, and measured under similar conditions
[3]. The optical rotation of the (þ)-9 enantiomer has not
been reported previously.
We have published recently that the treatment of 8-
bromotheophylline (1) with the reagent 9 yields the oxa-
zoline derivative 10 and, depending on conditions, vary-
ing amounts of the rearrangement product 11. Relatively
short reaction time and temperature favour the formation
of 10 which can be obtained in an 85% yield. Con-
versely, heating the mixture at an elevated temperature
for a prolonged period of time results in an efficient
rearrangement of the initially formed compound 10 to
the oxazine derivative 11 [6].
The racemic compound 10 was treated with ammonia
in aqueous ethanol to give racemic product P23.
After optimization of the conditions a similar treat-
ment of the enantiomers of 10 furnished optically active
samples of P23. Their enantiomeric excess (defined as
Journal of Heterocyclic Chemistry
DOI 10.1002/jhet

March 2009
Synthesis of the Hypotensive Agent 8-Amino-7-[2-hydroxy-
3-morpholinopropyl]theophylline (P23) and Analogs
193
NMR: d 2.46 (m, 4H), 2.63 (m, 2H), 3.22 (m, 2H), 3.46 (m,
2H), 3.37 (s, 3H), 3.54 (s, 3H), 3.70 (m, 4H), 3.83 (m, 4H),
4.04 (m, 2H), 4.18 (m, 1H), 4.30 (m, 1H); ei-ms: m/z 408
(Mþ, 10), 308 (50), 265 (10), 126 (50), 100 (100). High reso-
lution ms. Calcd. for C18H27N6O5 (Mꢀ ꢀ 1): m/z 407.2043.
Found: m/z 407.2055.
the percent excess of the enantiomer over the racemate)
1
was calculated by analysis of the H NMR spectra of
diastereomeric esters of the samples of P23 with (S)-
(þ)-a-methoxyphenylacetic acid. The procedure has
been published previously [7], and the relative intensity
of the singlets for the methine protons MeOCH(Ph)COO
(d 4.70-4.80) is the most informative. The obtained val-
ues of the enantiomeric excess for the samples contain-
ing (þ)-P23 and (ꢀ)-P23 as the major enantiomers
were 54% and 58%, respectively. These values corre-
spond to the enantiomerically enriched samples contain-
ing 77% and 79% of the major enantiomer, respectively.
The partial racemization observed for the synthesis of
P23 may be related to the partial racemization of 9, as
already mentioned [3]. Nevertheless, on the basis of the
stereospecific chemistry involved, it can be safely
assumed that the absolute configuration of the major
enantiomer in the samples, (R)-(ꢀ)-P23 and (S)-(þ)-
P23, are as shown. More specifically, the chemical
transformations 8 ! 9 ! 10 ! P23 do not involve
stereocenters in these compounds and, as a result, the
absolute configurations, R or S, are retained in the inter-
mediate and final products.
7-[(Benzylamino)methyl]-1,3-dimethyl-6,7-dihydro-oxaz-
olo[2,3-f]purine-1H,3H-2,4-dione [(ꢂ)-5]. A solution of 2
(0.05 g, 0.10 mmol) and benzylamine (0.02 g, 0.20 mmol) in
toluene (5 mL) was heated under reflux. The reaction was
completed after 1.5 hours, as judged by TLC analysis on silica
gel eluting with chloroform/methanol (9:1). The resultant pre-
cipitate of benzylammonium 4-toluenosulfonate was filtered
off and the solution was concentrated on a rotary evaporator.
Product 5 was isolated by chromatography on silica gel eluting
with chloroform/methanol (9:1), yield 29%, mp 252–253ꢁC;
1H NMR: d 3.30 (s, 3H), 3.35 (s, 2H), 3.49 (s, 3H), 3.80 (s,
1H), 4.40 (m, 2H), 4.47 (m, 1H), 4.80 (m, 2H), 7.30 (m, 5H);
ei-ms: m/z 341 (Mþ, 100), 250 (25), 91 (40). High resolution
ms. Calcd. for C17H20N5O3 (Mþ þ 1): m/z 342.1566. Found:
m/z 342.1570.
7-[(4-Tolylsulfonyloxy)methyl]-1,3-dimethyl-6,7-dihydro-
oxazolo[2,3-f]purine-1H,3H-2,4-dione [(ꢂ)-6]. Gaseous am-
monia was slowly bubbled at 0ꢁ through a solution of 3 (0.27
g, 0.6 mmol) in anhydrous ethanol. After 3 hours the resultant
precipitate of 6 was collected by filtration and dried at 25ꢁ;
1
yield 80%, mp 215–220ꢁC; H NMR: d 2.50 (s, 3H), 3.42 (s,
3H), 3.55 (s, 3H), 4.40 (m, 4H), 5.68 (m, 1H), 7.42 (d, J ¼
8.5 Hz, 2H), 7.80 (d, J ¼ 8.5 Hz, 2H); ei-ms: m/z 406 (Mþ,
100), 234 (25), 91 (30). High resolution ms. Calcd. for
C17H17N4O632S (Mꢀ ꢀ1): m/z 405.0869. Found: m/z
405.0883.
EXPERIMENTAL
Melting points (Pyrex capillary) are uncorrected. Electron-
impact mass spectra (ei-ms) were recorded at 70 eV. High re-
solution time-of-flight mass spectra were recorded using elec-
tron-spray ionization (high resolution ms) with 0.5% ammonia
in methanol for the negative ion mode and 0.1% formic acid
7-[2-Hydroxy-3-(4-tolylsulfonyloxy)propyl]-8-morpholino-
theophylline (ꢂ)-7]. A mixture of (ꢂ)-6 (0.03 g, 0.08 mmol),
morpholine (0.015 g, 0.16 mmol), and ethanol (2 mL) was
stirred at 25ꢁC for 24 h. Concentration of the mixture on a ro-
tary evaporator followed by chromatography of the residue on
silica gel eluting with dichloromethane/methanol (49:1) gave
1
in methanol for the positive ion mode. H NMR spectra were
obtained at 300 MHz in deuteriochloroform solution with the
solvent used as an internal standard. All commercial reagents
were purchased from Aldrich or Fluka and used without
purification.
1
product 7 in a 20% yield, mp 163–164ꢁC; H NMR: d 2.50 (s,
3H), 3.31 (m, 4H), 3.40 (s, 3H), 3.57 (s, 3H), 3.88 (m, 4H),
3.94 (m, 1H), 4.25 (m, 4H), 5.52 (br s, 1H), 7.42 (d, J ¼ 8.2
Hz, 2H), 7.82 (d, J ¼ 8.2 Hz, 2H). High resolution ms. Calcd.
for C21H28N5O732S (Mþ þ 1): m/z 494.1709. Found: m/z
494.1685.
8-Bromo-7-[2-hydroxy-3-(4-tolylsulfonyloxy)propyl]theo-
phylline [(ꢂ)-3]. A solution of (ꢂ)-2,3-epoxypropyl 4-tolue-
nesulfonate (2, 0.27 g, 1.2 mmol), 8-bromotheophylline (1,
0.21 g, 0.8 mmol), and a catalytic amount of pyridine (0.02 g,
0.2 mmol) in n-propanol (2.5 mL) was heated under reflux for
1 hour. After cooling the mixture was concentrated on a rotary
evaporator, and the residue was subjected to silica gel chroma-
tography eluting with dichloromethane/methanol (49:1). Prod-
uct (ꢂ)-3 was obtained in a 43% yield, mp 78–79ꢁC; 1H
NMR: d 2.46 (s, 3H), 3.38 (s, 3H), 3.56 (s, 3H), 3.80 (br s,
1H), 4.12–4.18 (m, 3H), 4.48 (d, J ¼ 1.4 Hz, 2H), 7.38 (d, J
¼ 8.5 Hz, 2H), 7.85 (d, J ¼ 8.5 Hz, 2H). High resolution ms.
Calcd for C17H2079BrN4O632S (Mþ þ 1): m/z 487.0287.
Found: m/z 487.0290.
(ꢂ)-N-(2,3-Epoxypropyl)morpholine [(ꢂ)-9]. A mixture of
morpholine (8.5 g, 0.1 mol) and water (0.5 mL) was treated
dropwise with (ꢂ)epichlorohydrin (8, 9.25 g, 0.1 mol) in such
a way that the temperature did not exceed 35 ꢁC (exothermic
reaction). After the addition was completed (1.5 h), the mix-
ture was treated with an aqueous solution of sodium hydroxide
(38%, 12.5 g) and then stirred for an additional 1 hour. The
precipitate of sodium chloride was filtered off, and the filter
and the flask was washed with ether (3 ꢃ 30 mL). The ether
solution was dried (anhydrous K2CO3), then concentrated on a
rotary evaporator, and the oily residue was distilled under
reduced pressure (45–48ꢁC/20 mmHg). Racemic product 9 was
7-(2-Hydroxy-3-morpholinopropyl)-8-morpholinotheophyl-
line [(ꢂ)-4]. A solution of (ꢂ)-3 (0.05 g, 0.10 mmol) and mor-
pholine (0.02 g, 0.21 mmol) in toluene (5 mL) was heated
under reflux for 1.5 hours. The mixture was concentrated on a
rotary evaporator, and the residue was subjected to silica gel
chromatography eluting with dichloromethane/methanol (49:1).
1
obtained in a 40% yield and its H NMR spectrum was virtu-
ally identical with that reported [3] for the compound obtained
by using a different synthetic route.
(2R)-(þ)-N-(2,3-Epoxypropyl)morpholine [(2R)-(þ)-9]. The
1
Compound 4 was obtained in a 69% yield, mp 175–176ꢁC; H
reaction with (2R)-(ꢀ)-epichlorohydrin with morpholine and
Journal of Heterocyclic Chemistry
DOI 10.1002/jhet

194
M. T. Cegla, J. Potaczek, M. Zylewski, and L. Strekowski
Vol 46
workup were conducted as described above; [a]D ¼ þ24.1ꢁ (c
¼ 10% in toluene).
1H), 5.74 (br s, 2H); ei-ms: m/z 338 (20), 238 (45), 126 (50),
100 (100). The treatment of (7S)-(ꢀ)-10 with ammonia gave
(S)-(þ)-P23, [a]D ¼ þ5.3ꢁ (c ¼ 2.2% in dichloromethane/
methanol, 49:1). The treatment of (7R)-(þ)-10 with ammonia
gave (R)-(ꢀ)-P23, [a]D ¼ ꢀ8.0ꢁ (c ¼ 2.2% in dichlorome-
thane/methanol, 49:1). The mp, NMR and ms data reported
above are identical for all three products. Anal. Calcd. for
C14H22N6O4 [(þ)-P23]: C, 49.72; H, 6.55; N, 24.84. Found:
C, 49.49; H, 6.90; N, 24.51.
Derivatization of P23 for analysis of the enantiomeric
excess by 1H NMR. Esterification of the secondary alcohol
function of P23 with (S)-(þ)-a-methoxyphenylacetic acid was
conducted by using a published procedure [7]. The ester was
purified by chromatography on silica gel eluting with chloro-
form/methanol (49:1). The 1H NMR spectra of the diastereo-
meric derivatives were taken in deuteriochloroform, and the ra-
tio of the signals of the methine protons MeOCH(Ph)COO (d
4.70–4.80) were analyzed [7].
(2S)-(-)-N-(2,3-Epoxypropyl)morpholine [(2S)-(ꢀ)-9]. The
reaction with (2S)-(þ)-epichlorohydrin with morpholine and
workup were conducted as described above; [a]D ¼ ꢀ25.4ꢁ (c
¼ 10% in toluene); reported [a]D ¼ ꢀ20.2ꢁ in toluene [3].
1,3-Dimethyl-7-morpholinomethyl-6,7-dihydrooxazolo-[2,
3-f]-purine-1H,3H-2,4-dione [(ꢂ)-10, (7S)-(ꢀ)-10 and (7R)-
(þ)-10]. A mixture of 8-bromotheophylline (1, 0.39 g, 1.5
mmol), racemic compound 9 (0.31 g, 2 mmol), and pyridine
(0.08 g, 1 mmol) in ethanol (5 mL) was stirred at room tem-
perature for up to 10 days until TLC analysis (silica gel, chlo-
roform/triethylamine, 9:1) showed the absence of 1 and/or the
presence of a small amount of a byproduct 11. Then the mix-
ture containing precipitate of 10 was cooled to 0ꢁC for an
additional 24 h and filtered. Product (ꢂ)-10 was obtained in an
analytically pure form by chromatography on silica gel eluting
with chloroform/methanol (9:1): yield 85%, mp 230–231ꢁC. A
similar treatment of 1 with (R)-(þ)-9 gave (R)-(þ)-10, mp
176–178ꢁC, [a]D ¼ þ55.3ꢁ (c ¼ 3% in dichloromethane/meta-
nol, 49:1). A similar treatment of 1 with (S)-(ꢀ)-9 gave (S)-
(ꢀ)-10, mp 180–182ꢁC, [a]D ¼ ꢀ49.7ꢁ (c ¼ 3% in dichloro-
methane/methanol, 49:1). The 1H NMR spectra of these three
products were virtually identical with that reported previously
for (ꢂ)-10 [6]; ei-ms (identical in all cases): m/z 321 (Mþ,
30), 234 (40), 126 (55), 113 (20), 100 (100). Anal. Calcd. for
C14H19N5O4 [(ꢂ)-10]: C, 52.33; H, 5.95; N, 21.79. Found: C,
52.06; H, 5.93; N, 21.53.
REFERENCES
[1] (a) Pawlowski, M.; Gorczyca, M.; Bobkiewicz-Kozlowska,
T.; Chodera, A.; Mrozikiewicz, A. Polish Pat PL 165,986 B1, 1991;
(b) Pawlowski, M.; Gorczyca, M.; Bobkiewicz-Kozlowska, T.; Cho-
dera, A.; Mrozikiewicz, A. Polish Pat PL167,407 B1, 1991; (c) Paw-
lowski, M.; Gorczyca, M.; Bobkiewicz-Kozlowska, T.; Chodera, A.;
Mrozikiewicz, A. Chem Abstr 1995, 124, 55689.
[2] Ammazzalorso, A.; Amoroso, R.; Bettoni, G.; Fantacuzzi, M.;
De Philippis, B.; Giampietro, L.; Maccallini, C.; Paludi, D.; Tricca, M. L.
Farmaco 2004, 59, 685.
8-Amino-7-(2-hydroxy-3-morpholinopropyl)theophylline
[(ꢂ)-P23, (S)-(þ)-P23, and (R)-(ꢀ)-P23]. A mixture of (ꢂ)-
10 (0.32 g, 1.0 mmol), aqueous ammonia (25%, 10 mL), and
ethanol (15 mL) was heated in a pressure vessel to 110ꢁC for
6 hours. After cooling the mixture was concentrated on a ro-
tary evaporator and subjected to silica gel chromatography
eluting with chloroform/methanol (49:1), yield 88%, mp 221–
222ꢁC (reported [1] mp 226–228ꢁC for the racemic product);
1H NMR: d 2.44 (m, 2H), 2.66 (m, 4H), 3.37 (s, 3H), 3.50 (s,
3H), 3.66 (m, 4H), 4.18 (m, 1H), 4.40 (m, 2H), 4.50 (br s,
[3] Achmatowicz, O.; Malinowska, I.; Wisniewski, A.; Gladecki,
Z.; Fruzinski, A. Polym J Chem 1996, 70, 891.
[4] Swirska, A. Acta Pol Pharm 1962, 19, 317.
[5] Hanson, R. M. Chem Rev 1991, 91, 437.
[6] Cegla, M. T.; Potaczek, J.; Zylewski, M.; Strekowski, L.
Tetrahedron Lett 2005, 46, 3561.
[7] Seco, J. M.; Quinoa, E.; Riguera, R. Tetrahedron: Asymme-
try 2001, 12, 2915.
Journal of Heterocyclic Chemistry
DOI 10.1002/jhet