A new efficient synthesis of the immunosuppressive agent (±)-15-deoxyspergualin

Article abstract of DOI:10.1016/S0040-4020(01)00130-2

(±)-15-Deoxyspergualin is a new promising immunosuppressive agent, recently marketed in Japan for the control of cortico-resistant acute renal graft rejection. We describe here a nine-step synthesis starting from 7-bromoheptanenitrile and N¹,N⁴-bis(benzyloxycarbonyl)spermidine, suitable for the production of multigram quantities of this unstable highly polar compound with an 18% overall yield. Furthermore, it opens the way for the synthesis of α-hydroxyglycine analogues.

Full text of DOI:10.1016/S0040-4020(01)00130-2

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 2757±2760
A new ef®cient synthesis of the immunosuppressive agent
(^)-15-deoxyspergualin
Philippe Durand,p Philippe Peralba and Patrice Renaut
Laboratoires Fournier S.A., Immunology Research, 50 route de Dijon, 21121 Daix, France
Received 27 October 2000; revised 26 January 2001; accepted 31 January 2001
AbstractÐ(^)-15-Deoxyspergualin is a new promising immunosuppressive agent, recently marketed in Japan for the control of cortico-
resistant acute renal graft rejection. We describe here a nine-step synthesis starting from 7-bromoheptanenitrile and N1,N4-bis(benzyl-
oxycarbonyl)spermidine, suitable for the production of multigram quantities of this unstable highly polar compound with an 18% overall
yield. Furthermore, it opens the way for the synthesis of a-hydroxyglycine analogues. q 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
and the puri®cation of this hygroscopic unstable salt deriva-
tive a real challenge.
(^)-15-Deoxyspergualin (DSG) 1 has aroused considerable
interest1 as an immunosuppressive agent in transplantation.
This synthetic derivative of the natural antitumor and immu-
nosuppressive antibiotic spergualin 2 (Fig. 1) is usually
found more effective than the popular immunosuppressants
Cyclosporin A, FK 506 or Rapamycin,2 and routinely elicits
fewer side effects and a different mechanism of action.3 (^)-
DSG has a rather unstable peptidomimetic4 structure
containing an asymmetric carbon. Although in vivo experi-
ments clearly showed that (2)-DSG is the only immuno-
suppressive enantiomer and that both isomers contribute at
least to the acute toxicity of the compound,4,5 most of the
biological, pharmacological and clinical data have been
obtained with the more readily available racemic tris-
hydrochloride DSG. Furthermore, the racemic form
(Spanidinw) received marketing approval in Japan for the
treatment of corticoresistant acute renal graft rejection
episodes. We have been involved for some years in a
structure±activity relationship study in this series in order
to better de®ne its structural requirements and to design
more stable analogues.6 To evaluate the potency of our
compounds in several pharmacological experiments, we
needed multigram quantities of (^)-DSG as reference.
DSG consists of a rather unstable a-hydroxyglycine central
part, connecting two highly polar moieties: guanidinohepta-
noic acid and spermidine (Fig. 1). Owing to the unusual
hemiaminal structure of the a-hydroxyglycine unit, DSG
hydrolyses gradually, in basic or acidic aqueous solution,
into 7-guanidinoheptanamide and hydrated glyoxylspermi-
dine.1b,7 These different properties have made the synthesis
2. Results and discussion
The synthesis of (^)-DSG was ®rst described in a low over-
all yield (3.5%) from 7-aminoheptanoic acid through
condensation of a glyoxylspermidine derivative with
7-guanidinoheptylamide.5a,8 This synthesis involves amino
and guanidino intermediates simply protected as their
hydrochloride salts. The drawback of this strategy is the
puri®cation, which requires the use of ion exchange
chromatography at each step of the synthesis, making the
scale up dif®cult. A more elegant synthesis of (^)-DSG in a
7.2% overall yield from non commercially available
7-aminoheptylamide used condensation of Boc-protected
7-guanidinoheptylamide with methyl-2-hydroxy-2-meth-
oxy acetate for the synthesis of the a-hydroxyglycine
central part.9 This protective group strategy avoids the use
of ion exchange chromatography for the puri®cation of the
intermediates, but the condensation step could not be
improved by heating due to the thermal instability of the
Keywords: a-hydroxyglycine; polyamine; deoxyspergualin; immunosup-
pressive agent.
p
Corresponding author. Fax: 133-3-8044-7538;
e-mail: p.durand@fournier.fr
Figure 1.
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)00130-2

2758
P. Durand et al. / Tetrahedron 57 (2001) 2757±2760
Ê
Scheme 1. Reagents and reaction conditions: (a) see Ref. 10a, 62%; (b) (MeO)(HO)CHCO2Me, CH2Cl2, 408C, 24 h, molecular sieves 4 A; (c) SOCl2, CH2Cl2,
408C, 2 h; (d) benzyl alcohol, Et3N, CH2Cl2, rt, 24 h, 67% overall from 4; (e) 1N NaOH, DME, rt, 1.5 h; (f) H2N(CH2)4N(Cbz)(CH2)3NHCbz (9), DCC, HOBT,
CH2Cl2, rt, 48 h, 72% overall from 5; (g) (i) Ph3P, H2O, THF, 658C, overnight; (ii) CbzNvC(SMe)NHCbz (10), THF, rt, 3 h, 79% overall from 7; (h) (i)
Pd(OH)2/C (20%), MeOH/AcOH, H2 1 atm, (ii) 1N HCl pH 3, 08C, and lyophilization, 2 cycles, 77%.
Boc-protected intermediates. Furthermore, in this synthesis,
the hemiaminal hydroxyl had to be protected by a silyl ether
group in order to avoid a retro reaction during the coupling
of Boc-protected spermidine with the glyoxylic ester. The
®nal acidic deprotection step required a TFA treatment,
which led to the tri¯uoro acetate salt of DSG in moderate
yield. Ion exchange chromatography followed by a desalt-
ing chromatography were necessary to obtain the drug as its
tris-hydrochoride salt. Moreover, the ®rst synthesis of the
optically active (2)-DSG involved several steps with many
dif®cult puri®cations and could not be adapted for our pur-
pose.5a However, we have recently described10 an enantio-
selective synthesis of (2)-DSG, with a good overall yield,
using only benzyl type protective group strategy for the
three different functional groups of the molecule and intro-
ducing the guanidino function in the last step before depro-
tection. This approach made the synthesis more ef®cient and
the puri®cation of the intermediates easier, but we still
required ion exchange chromatography for the obtention
of the ®nal tris-hydrochloride salt. However, we felt that
we could use a similar strategy for a multigram synthesis
of the racemic form and that the ®nal deprotective step
could be modi®ed in order to obtain the hydrochloride salt
without using ion exchange chromatography.
sequence.9 Saponi®cation of 5 gave the corresponding
acid 6 which could be either puri®ed or used directly in
the following step. Coupling of N1,N4-bis(benzyl-
oxycarbonyl)spermidine12 with 6 using the usual DCC/
HOBT methodology provided the polyamine derivative 7.
Conversion of 7 to 8 was made in a two-step one-pot
process. First the azido function was reduced without affect-
ing the benzyloxy and the benzyloxycarbonyl protective
groups using Ph3P/H2O.13 The primary amine was then
reacted in situ with N,N0-bis(benzyloxycarbonyl)-S-methyl-
isothiourea14 to give 8 in 79% yield. The product could be
puri®ed by a simple crystallization. The high purity of this
intermediate made it possible to avoid chromatography for
puri®cation. Final deprotection was accomplished using
Perlman's catalyst in 1N AcOH in methanol under 1 atm
of hydrogen. We did not attempt to carry out the deprotec-
tion in the presence of hydrochloric acid because of the
instability of the product in acid and the dif®culty of adjust-
ing the pH of the solution. In such conditions, the a-hydroxy-
glycine moiety could either be hydrolyzed or converted to
its corresponding methyl ether. At this stage, we found that
the so formed tris-acetate derivative could be easily
converted to its corresponding tris-hydrochloride using a
simple salt exchange procedure. It simply consists of adjust-
ing a cold aqueous solution of the tris-acetate derivative to
pH 3 with hydrochloric acid solution, followed by lyophili-
7-Azidoheptylamide 4 was obtained from 7-bromoheptane-
nitrile 3 in 62% yield as described previously (Scheme 1).10a
This amide was then converted in a three-step one-pot
process into the desired a-benzyloxy glycine derivative 5.
The ®rst step consisted of a condensation with methyl-2-
hydroxy-2-methoxy acetate, trapping the formed methanol
with molecular sieves.11 The resulting a-hydroxyglycine
was converted in situ into a labile a-chloroglycine inter-
mediate by treatment with thionyl chloride at 408C. The
absence of Boc-protected guanidine group allowed us to
use these rather acidic but ef®cient conditions for the
condensation step and allowed warming to further improve
the yield. The chloro compound was then treated with
benzyl alcohol in presence of triethylamine to afford the
a-benzyloxy derivative in 67% overall yield for the three
steps. It favorably compares with the 29% yield reported
using the Boc protective group strategy for a similar
1
zation. The salt exchange could be followed by H NMR
spectroscopy checking the disappearance of the methylic
protons of acetic acid. Two cycles of this simple procedure
enabled a complete salt exchange and gave the tris-
hydrochloride salt of (^)-DSG with high purity in 77%
yield.
3. Conclusion
The present synthesis affords an easy access to racemic DSG
which is the prototype of a new family of immunosuppres-
sor agents5b,6,15 from 7-bromoheptanenitrile in a nine-step
process and an 18% overall yield. Furthermore, it opens new
possibilities for the synthesis of analogues modi®ed on the
polyamine and/or on the guanidine moieties.

P. Durand et al. / Tetrahedron 57 (2001) 2757±2760
2759
4. Experimental
pressure. EtOAc (100 mL) was added and the insoluble
dicyclohexyl urea was eliminated by ®ltration and washed
with EtOAc (50 mL). The ®ltrate was washed with saturated
aqueous NaHCO3, brine, and the organic layer was dried
(MgSO4). Concentration under reduced pressure of the
organic layer and ¯ash chromatography on silica gel
(EtOAc±i-Pr2O, 8:2) gave 7 (17.5 g, 72%) as an oil; Rf:
4.1. General
All chemicals were purchased from commercial sources and
used without further treatment except when speci®ed. THF
was freshly distilled from sodium benzophenone ketyl.
Melting points were uncorrected. Proton magnetic reso-
nance spectra were determined either at 250 or 300 MHz
with TMS as internal standard. Carbon magnetic resonance
spectra were recorded at 62.9 or 75 MHz. The chemical
shifts are expressed in d values relative to TMS. IR spectra
were obtained on KBr, 3M Disposable IR card (type 61) or
in solution in the speci®ed solvent. Rf values were measured
after thin layer chromatography performed with precoated
silica plates (Kieselgel 60 F254, Merck). Elemental analyses
were performed with an elemental analyzer, Perkin±Elmer
2400 CHN.
0.38 (EtOAc±i-Pr2O, 8:2); IR (CHCl3) 1680, 2100 cm21
;
1H NMR (CD3OD) d 1.3±1.73 (m, 14H), 2.25 (t, Jˆ
7.2 Hz, 2H), 3±3.3 (m, 10H), 4.55±4.67 (m, 2H), 5.04 (s,
2H), 5.07 (s, 2H), 5.52 (s, 1H), 7.2±7.4 (m, 15H); 13C NMR
(CD3CN) d 25.49, 26.56, 28.81, 36.07, 38.35, 44.76, 51.50,
62.21, 66.92, 70.22, 78.06, 128.01, 128.16, 128.18, 128.28,
128.41, 128.77, 137.83, 137.94, 138.34, 156.80, 168.04,
174.33; Anal. calcd for C39H51N7O7: C, 64.18; H, 7.04; N,
13.44. Found: C, 64.55; H, 7.09; N, 13.35.
4.1.3. Benzyl 21-[(benzyloxy)carbonyl]amino-11-(benzyl-
oxy)-4-[(benzyloxy)carbonyl]-10,13,23-trioxo-25-phenyl-
24-oxa-4,9,12,20,22-pentaazapentacos-21-en-1-yl carba-
mate (8). A mixture of triphenyl phosphine (6.75 g,
25.7 mmol), 7 (17 g, 21.4 mmol) and water (0.46 mL,
25.7 mmol) in THF (200 mL) was heated at re¯ux over-
night. After cooling at room temperature, N,N0-bis(benzyl-
oxycarbonyl)-S-methylisothiourea 10 (9.21 g, 25.7 mmol)
was added and the mixture was stirred for 6 h at room
temperature. Concentration of the mixture gave a crude
product, which was puri®ed by chromatography on silica
gel (EtOAc). 8 (17.5 g, 79%) was obtained as a white
solid by crystallization from Et2O and recrystallization
from MEK±i-Pr2O, 7:10; mp 58±648C; Rf: 0.43 (EtOAc±
i-Pr2O, 8:2); IR (KBr) 1640, 1690, 1710, 1730, 2830, 3020,
4.1.1. Methyl [(7-azidoheptanoyl)amino](benzyloxy)-
acetate (5). A solution of 4 (9.74 g, 52.2 mmol) and methyl-
2-hydroxy-2-methoxy acetate (6.8 mL, 68.7 mmol) in
250 mL of CH2Cl2 was heated under re¯ux for 24 h in a
¯ask connected to a Soxhlet apparatus ®lled with 40 g of
Ê
4 A molecular sieves. After cooling, the Soxhlet apparatus
was replaced by a re¯ux condenser. Thionyl chloride
(5.4 mL, 74.4 mmol) was added and the resulting mixture
was re¯uxed for 2 h. The reaction mixture was concentrated
under reduced pressure. The crude chloroglycine derivative
was dissolved in 100 mL of CH2Cl2, then benzyl alcohol
(7.1 mL, 68.7 mmol) and triethylamine (9.55 mL, 68.7
mmol) in 50 mL of CH2Cl2 were added dropwise for 24 h
at room temperature. The reaction mixture was then washed
with 1N HCl (100 mL) and brine (100 mL). The organic
layer was dried (MgSO4), ®ltered and concentrated under
reduced pressure, and the compound was puri®ed by ¯ash
chromatography on silica gel (hexane±i-PrOH, 9:1) to give
13.4 g (67%) of 5 as a colorless oil. IR (neat) 1660, 1750,
1
3060, 3300 cm21; H NMR (d6-DMSO) d 1.23±1.64 (m,
14H), 2.16±2.18 (m, 2H), 2.96±2.98 (m, 2H), 3±3.4 (m,
8H), 4.44±4.54 (m, 2H), 4.99±5.20 (4s, 8H), 5.46 (d, Jˆ
9 Hz, 1H), 7.2±7.4 (m, 26H), 8.08 (brt, 1H), 8.38 (brt, 1H),
8.59 (d, Jˆ9 Hz, 1H), 11.57 (s, 1H); 13C NMR (CD3CN) d
25.62, 26.73, 28.99, 36.21, 39.99, 41.22, 66.31, 67.02,
68.45, 70.33, 78.17, 128.09, 128.26, 128.35, 128.38,
128.41, 128.50, 128.61, 128.87, 129.00, 129.11, 129.19,
135.90, 137.87, 137.93, 138.03, 138.43, 154.08, 156.58,
156.89, 164.36, 168.14, 174.43; Anal. calcd for
C56H67N7O11: C, 66.32; H, 6.66; N, 9.67. Found: C, 66.21;
H, 6.65; N, 9.69.
1
2100, 2930, 2970 cm21; H NMR (CDCl3) d 1.34±1.4 (m,
4H), 1.55±1.7 (m, 4H), 2.3 (t, Jˆ7.5 Hz, 2H), 3.26 (t, Jˆ
6.6 Hz, 2H), 3.8 (s, 3H), 4.67±4.76 (m, 2H), 5.74 (d, Jˆ
7.8 Hz, 1H), 6.5 (d, Jˆ7.8 Hz, 1H), 7.26±7.37 (m, 5H); 13C
NMR (CD3CN) d 25.39, 26.55, 28.78, 35.93, 51.51, 52.63,
70.26, 77.23, 128.26, 128.47, 128.80, 138.17, 168.88,
173.88 (2); Anal. calcd for C17H24N4O4: C, 58.61; H,
6.94; N, 16.08. Found: C, 58.75; H, 6.97; N, 16.22.
4.1.4. 7-{[Amino(imino)methyl]amino}-N-[2-({4-[(3-amino-
propyl)amino]butyl}amino)-1-hydroxy-2-oxoethyl]hep-
tanamide tris-hydrochloride (1). A ¯ask containing a solu-
tion of 8 (3 g, 2.96 mmol) in 1N AcOH in methanol (10 mL)
was purged with nitrogen. To this solution were added
300 mg (50 wt%) of palladium hydroxide (Pearlman's cata-
lyst, 20% on carbon/50% H2O). The mixture was stirred for
8 h under 1 atm of hydrogen at rt. The mixture was purged
with N2 and ®ltered. The ®ltrate was purged with N2,
300 mg (50 wt%) of Pearlman's catalyst were added again
and the mixture was treated as above overnight. The mixture
was purged with N2, ®ltered and water was added. The
solution was concentrated under reduced pressure and
extracted three times with CH2Cl2. The aqueous phase
was lyophilized. The tris-acetate form of the compound
was transformed into its tris-hydrochloride form by the
following procedure. The powder was dissolved in H2O
(40 mL) and the pH of the solution was adjusted to 3 by a
4.1.2. Benzyl 3-{(4-{[[(7-azidoheptanoyl)amino](benzyl-
oxy)acetyl]amino}butyl)[(benzyloxy)carbonyl]amino}
propyl carbamate (7). A mixture of 1N NaOH (37 mL,
37 mmol) and 5 (10.7 g, 30.6 mmol) in 200 mL of DME
was stirred for 1.5 h at room temperature, then diluted
with brine (100 mL) and Et2O (200 mL). The aqueous
phase was acidi®ed to pH 2 with 1N HCl, extracted with
Et2O (3£100 mL) and the organic phases were dried
(Na2SO4) and concentrated to yield 10.2 g (100%) of the
crude acid. This product was dissolved in CH2Cl2
(200 mL), and hydroxybenzotriazole (4.13 g, 30.6 mmol)
followed by DCC (7.06 g, 34.2 mmol) were added. After
stirring at room temperature for 30 min, N1,N4-bis(benzyl-
oxycarbonyl)spermidine12
9 (14.18 g, 34.2 mmol) in
CH2Cl2 (100mL) was added. The mixture was stirred at
room temperature for 20 h and concentrated under reduced

2760
P. Durand et al. / Tetrahedron 57 (2001) 2757±2760
4. Sawa, H.; Kondo, S.; Ikeda, D.; Takeuchi, T.; Umezawa, H.
J. Antibiot. 1982, 35, 1665.
slow addition of 1N HCl at 08C and the solution was lyo-
philised. This procedure was repeated to give the tris-
hydrochloride 1 as a hygroscopic white powder (1.14 g,
5. (a) Umeda, I.; Moriguchi, M.; Ikai, K.; Kuroda, H.; Nakaruma,
T.; Fujii, A.; Takeuchi, T.; Umezawa, H. J. Antibiot. 1987, 40,
1316. (b) Maeda, K.; Umeda, Y.; Saino, T. Ann. N.Y. Acad.
Sci. 1993, 685, 123 (and references cited therein).
6. (a) Lebreton, L.; Annat, J.; Derrepas, P.; Dutartre, P.; Renaut,
P. J. Med. Chem. 1999, 42, 277. (b) Lebreton, L.; Jost, E.;
Carboni, B.; Annat, J.; Vaultier, M.; Dutartre, P.; Renaut, P.
J. Med. Chem. 1999, 42, 4749.
1
77%); Rf: 0.6 (R.P.18, CH3CN±H2O±TFA, 3:6.5:0.5); H
NMR (D2O) d 1.3±1.45 (m, 4H), 1.5±1.8 (m, 8H), 2±2.2
(m, 2H), 2.3 (t, Jˆ4.5 Hz, 2H), 3.05±3.25 (m, 8H), 3.27±3.3
(m, 2H), 5.45 (s, 1H); 13C NMR (D2O) d 23.64, 24.46,
25.57, 26.22 (2C), 28.42, 28.47, 36.20, 37.28, 39.22,
41.81, 45.16, 48.08, 72.63, 157.39, 171.97, 178.08. HRMS
found: 388.3036; calcd for (M1H)1ˆC18H38N7O3:
388.30361. The chemical purity was determined to be
.98% by HPLC analysis (Inertsil OSD 2, Gl Sciences
Inc.; Solvent A: water with 0.05% TFA; Solvent B:
CH3CN with 0.05% TFA; 2% Solvent B, 5 min; 2±80%
Solvent B in 15 min, 1 mL/min, 308C).
7. Brockman, R. W.; Wheeler, G. P. Proc. Am. Assoc. Cancer
Res. 1985, 26, 255.
8. Umeda, Y.; Moriguchi, M.; Nakamura, T.; Fujii, A.;
Takeuchi, T.; Umezawa, H. Ger Offen DE, 3 506 330, 1985;
Chem. Abstr. 1996, 104, 168275h.
9. Bergeron, R. J.; McManis, J. S. J. Org. Chem. 1987, 52, 1700.
10. (a) Durand, P.; Richard, P.; Renaut, P. J. Org. Chem. 1999, 63,
9723. (b) Lebreton, L.; Renaut, P.; Durand, P. PCT Int. Appl.
WO 9624579, August 15, 1996. (c) A similar approach had
been described since: Wang, J. K.; Thottahil, J. K. Eur. Pat.
Appl. EP 765, 866, April 2, 1997.
Acknowledgements
We are grateful to A. Dhennequin, G. Martin Gousset and
Dr V. Lucas for the analytical studies.
11. Daumas, M.; Vo Quaang, L.; Le Gof®c, F. Synth. Commun.
1990, 20, 3395.
References
12. Nishizawa, R.; Takei, Y.; Yoshida, M.; Tomiyoshi, T.; Saino,
T.; Nishikawa, K.; Nemoto, K.; Takahashi, K.; Fujii, A.;
Nakamura, T.; Takita, T.; Takeuchi, Y. T. J. Antibiot. 1988,
41, 1629.
1. (a) Groth, C. G. Ann. N.Y. Acad. Sci. 1993, 685, 193.
(b) Thomas, F. T.; Tepper, M. A.; Thomas, J. M.; Haisch,
C. E. Ann. N.Y. Acad. Sci. 1993, 685, 175. (c) Gores, P. F.;
Kaufman, D. B.; Sutherland, D. E. R. Ann. N.Y. Acad. Sci.
1993, 696, 377.
Â
13. Knouzi, N.; Vaultier, M.; Carrie, R. Bull. Soc. Chim. Fr. 1985,
5, 815.
14. Tian, Z.; Edward, P.; Roeske, R. W. Int. J. Pept. Protein Res.
1992, 40, 119.
2. (a) Gores, P. F.; Field, J. F.; Sutherland, D. E. R.; Chou, T. C.
Transplant. Proc. 1994, 26, 745. (b) Ohlman, S.; Zilg, H.;
Schindel, F.; Lindholm, A. Transplant. Int. 1994, 7, 5.
3. (a) Nadeau, K.; Nadler, S. G.; Saulnier, M.; Tepper, M. A.;
Walsh, C. T. Biochemistry 1994, 33, 2561. (b) Mazzuco, C. E.;
Nadler, S. G. Ann. N.Y. Acad. Sci. 1993, 685, 202 (and refer-
ences cited therein).
15. (a) Renaut, P.; Lebreton, L.; Dutartre, P.; Derrepas, P.;
Samreth, S. Eur. Pat. Appl. EP 600762, June 8, 1994.
(b) Lebreton, L.; Renaut, P.; Dumas C. Eur. Pat. Appl. EP
743300, November 20, 1996.