Efficient synthesis of (S)-(-)- and (R)-(+)-enantiomers of 15-deoxyspergualin (15-DSG)

Article abstract of DOI:10.1016/S0957-4166(00)00350-5

An efficient synthesis of the (S)-(-)- and (R)-(+)-enantiomers of 15-deoxyspergualin (15-DSG) is reported. The synthesis involves preparative HPLC separation and subsequent hydrogenolysis of two diastereoisomers 10a and 10b of fully protected 15-DSG penultimates. Alternatively, diastereomerically pure amide 10b (97% de) was also prepared from acid 8b (97% de), which was obtained via crystallization of a 1:1 diastereomeric mixture of 8a and 8b. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0957-4166(00)00350-5

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 11 (2000) 3665–3669
Efficient synthesis of (S)-(−)- and (R)-(+)-enantiomers of
15-deoxyspergualin (15-DSG)
Xuebao Wang* and John Thottathil
Bristol-Myers Squibb Pharmaceutical Research Institute, Department of Process Research, New Brunswick,
NJ 08903-0191, USA
Received 14 August 2000; accepted 17 August 2000
Abstract
An efficient synthesis of the (S)-(−)- and (R)-(+)-enantiomers of 15-deoxyspergualin (15-DSG) is
reported. The synthesis involves preparative HPLC separation and subsequent hydrogenolysis of two
diastereoisomers 10a and 10b of fully protected 15-DSG penultimates. Alternatively, diastereomerically
pure amide 10b (97% de) was also prepared from acid 8b (97% de), which was obtained via crystallization
of a 1:1 diastereomeric mixture of 8a and 8b. © 2000 Elsevier Science Ltd. All rights reserved.
15-Deoxyspergualin trihydrochloride (15-DSG) is a potent immunosuppressive agent, which
has been used in the racemic form to treat acute renal graft rejection episodes in Japan since
1994.1 The (S)-(−)-enantiomer of 15-DSG was initially derived from spergualin isolated from the
culture filtrates of bacillus laterosporus.1a,2 (S)-(−)-15-Deoxyspergualin was found to have
stronger antitumor activity against mouse leukemia L-1210 than the natural spergualin.2c,d In
both compounds, the (S)-(−)-enantiomers are more potent than the (R)-(+)-enantiomers. For
example, antibacterial and antitumor activity of the racemic mixture of spergualin is about half
of that of the natural (S)-spergualin.2c The (S)-(−)-enantiomer of 15-DSG has strong antitumor
activity against mouse leukemia L1210, while the (R)-(+)-enantiomer is almost inactive.3
* Corresponding author.
0957-4166/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(00)00350-5

3666
X. Wang, J. Thottathil / Tetrahedron: Asymmetry 11 (2000) 3665–3669
Synthesis of racemic 15-DSG was reported by Umeda,2d Bergeron4 and Dischio.5 The key step
in the racemic synthesis is the condensation of an amide with an aldehyde to construct the C-11
acylaminol.
Asymmetric synthesis of chiral 15-DSG is nontrivial due to chemical instability of the C-11
acylaminol. A relatively low ee in the product indicates that partial racemization at C-11 may
have occurred in the original synthesis of (S)-(−)-15-DSG from natural spergualin.2c An
alternative approach based on enzymatic resolution of a stable N-(7-guanidoheptanoyl)-a-
alkoxyglycine intermediate is applicable only to small scale preparation of optically active
15-DSG.3a
Herein, we report an efficient route for the preparation of both the (S)-(−)- and (R)-(+)-enan-
tiomers of 15-DSG.3b This route involves separation of the diastereoisomeric ethers 10a and 10b
derived from (R)-a-methylbenzyl alcohol by preparative HPLC. Alternatively, crystallization of
the precursor acids 8a and 8b afforded diastereomerically pure 8b, which was subsequently
converted to 10b without epimerization.
The overall synthesis is outlined in Scheme 1. The sulfate salt of 2-methyl-2-thiopseudourea 1
was reacted with excess CbzCl to afford 2 which, on reaction with 7-aminoheptanoic acid in the
presence of NaHCO3, gave N,N-(di-Cbz-guanidino)heptanoic acid 3. Reaction of 3 with
ammonia via the N-hydroxysuccinimide ester produced amide 4 in 87% crystallized yield. Amide
4 was heated with 1.4 equivalents of methyl glyxolate at 45°C to afford racemic amino acetal 5
in 85% crystallized yield.
Attempted conversion of 5–7 directly with a-methylbenzyl alcohol under various acidic
,
conditions (conc. H2SO4 or CSA, 4 A MS) was unsuccessful as the Cbz groups did not survive
the acidic conditions. A successful route involved conversion of alcohol 5 to the acetate 6 with
acetic anhydride and pyridine. Substitution of acetate 6 with (R)-a-methylbenzyl alcohol was
performed in the presence of a Lewis acid and Hunig’s base. Various Lewis acids were screened
and EtAlCl2 was found to be the most effective for this transformation. The Hunig’s base was
added to ensure a neutral or slightly basic reaction condition, thus suppressing cleavage of the
acid labile Cbz groups. The acetate formation step and the substitution reaction were carried out
in the same reaction flask with an overall yield of 74%.6
The product from the above substitution reaction was a mixture of the diastereomers 7a and
7b in approximately 1:1 ratio. Attempted separation of methyl ester 7a from 7b by crystalliza-
tion in various organic solvents was unsuccessful. It was difficult to separate them by normal
phase chromatography. Hence, saponification of the mixture to acids 8a and 8b followed by
coupling to di-CBZ-spermidine 93 afforded tetra-CBZ-11-(R)-a-methylbenzyloxyl-DSG
diastereomers 10a and 10b. These fully protected DSG penultimates were separated effectively
by preparative HPLC.7
Alternatively, a crystallization protocol to separate 8b from 8a was developed. The
diastereomeric acids 8a and 8b (1:1 mixture) on crystallization from toluene at rt produced 8b
in 40% yield (50% maximum). The ratio of isomer 8b to 8a in the crystallized material was 96:4
(1H NMR), while the ratio in the mother liquor was 20:80. A second crystallization from toluene
increased the diastereomeric ratio to 98.5:1.5 with 94% recovery. It was observed that isomer 8a
was not as stable as isomer 8b, with epimerization of 8a at C-11 slowly occurring at rt. For
example, the 8a enriched mother liquor became a 1:1 mixture of 8a and 8b after storage at rt for
about a week. The configuration of diastereomer 8b was confirmed by conversion to 10b and
ultimately to the (S)-(−)-enantiomer of 15-DSG.

X. Wang, J. Thottathil / Tetrahedron: Asymmetry 11 (2000) 3665–3669
3667
Scheme 1. Reagents and conditions: a CbzCl, 1N NaOH, CH2Cl2, H2O, rt, 2 days, 97%; b 7-aminoheptanoic acid,
NaHCO3, THF, H2O, 65°C, 6 h, 81%; c (i) NHS, DCC, CH3CN, THF, rt, 16 h and (ii) NH3, DMF, rt, 2 h, 87%;
d methyl glyxolate, THF, 40°C, 20 h, 89%; e Ac2O, pyridine, CH2Cl2; f R-a-methyl benzyl alcohol, EtAlCl2, Hunig’s
base, CH2Cl2, 0°C to rt, 10 h, 74% two steps; g 0.1N LiOH, THF, H2O, rt, 2 h, 93%; h PPA, TEA, CH2Cl2, 0°C,
1 h, 94%; i (i) H2, Pd/C; Pd(OH)2/C, 2N AcOH/MeOH, rt, 20 h, (ii) CM-Sephadex C-25 and (iii) Sephadex LH-20
Cleavage of all protecting groups of the penultimates 10a and 10b was accomplished by mild
hydrogenolysis. Since 15-DSG is unstable in its free amine form, the hydrogenolysis was
performed under acidic conditions. Interestingly, complete racemization occurred when
hydrogenolysis was performed in 1N hydrochloric acid and methanol with the best result being
obtained in acetic acid and methanol. To prevent epimerization during hydrogenolysis, the
concentration of acetic acid in methanol was maintained at 2N or less. Under the acetic acid
conditions, the Pearlman catalyst was used and the Cbz groups were easily cleaved in about 3
h at rt, while cleavage of the a-methyl benzyl protecting group was slower.
The crude tri-acetate salt from the hydrogenolysis reaction was converted to the tri-hydro-
chloride salt on a CM-Sephadex C25 column using aqueous sodium chloride as the mobile
phase. The resulting tri-hydrochloride salt was contaminated with sodium chloride and was
desalted on a CM-Sephadex LH20 column. After desalting, a pure final drug substance was
obtained. The optical rotation of purified (S)-(−)-15-DSG tri-hydrochloride {[h]D=−13.7 (c 1.0,

3668
X. Wang, J. Thottathil / Tetrahedron: Asymmetry 11 (2000) 3665–3669
H2O)} compared well with the literature value ([h]D=−14.1, (c 1.0, H2O)}.8 The (+)-enantiomer
of 15-DSG was prepared from 10b separately following the same procedure.
In summary, a new and practical route for the synthesis of both enantiomers of chiral
15-DSG was established. The synthesis involves ten steps with an overall yield of ꢀ25%. Key
steps are: (1) protection of the C-11 hydroxy group as a-methyl benzyl ether, (2) subsequent
separation of the resulting penultimates as a 1:1 diastereomeric mixture by preparative chro-
matography or by crystallization and (3) removal of the protecting group by mild hydrogenoly-
sis without racemization.
References
1. (a) Maeda, K.; Umeda, Y.; Saino, T. Ann. N.Y. Acad. Sci. 1993, 685, 123. (b) Tepper, M. Ann. N.Y. Acad.
Sci. 1993, 696, 123. (c) Amemiya, H.; Suzuki, S.; Ota, K.; Takahashi, K.; Sonoda, T.; Ishibashi, M.; Omoto,
R.; Koyama, F.; Orita, K.; Takagi, H. Int. J. Clin. Pharm. Res. 1991, 11, 175.
2. (a) Takeuchi, T.; Iinuma, H.; Kunimoto, S.; Masuda, T.; Ishizuka, M.; Hamada, M.; Naganawa, H.; Kondo,
S.; Umezawa, H. J. Antibiotics 1981, 34, 1619. (b) Umezawa, H.; Kondo, S.; Iinuma, H.; Kunimoto, S.;
Iwasaw, H.; Ikeda, D.; Takeuchi, T J. Antibiotics 1981, 34, 1622. (c) Iwasawa, H.; Kondo, S.; Ikeda, D.;
Takeuchi, T.; Umezawa, H. J. Antibiotics 1982, 35, 1665. (d) Umeda, Y.; Moriguchi, M.; Kuroda, H.;
Nakamura, T.; Iimuna, H.; Takeuchi, T.; Umezawa, H. J. Antibiotics 1985, 38, 886.
3. (a) Umeda, Y.; Moriguchi, M.; Ikai, K.; Kuroda, H.; Nakamura, T.; Fujii, A.; Takeuchi, T.; Umezawa, H. J.
Antibiotics 1987, 40, 1316. (b) Recently, a similar approach was described by: Durand, P.; Richard, P.; Renaut,
P. J. Org. Chem. 1998, 63, 9723, which prompted us to disclose our results. Our work was completed in 1993
and patented: Wang, X.; Thottathil, J. K. Eur. Pat. Appl. EP 765,866, Apr. 2, 1997.
4. Bergeron, R.; McManis, J. J. Org. Chem. 1987, 52, 1700.
5. (a) Dischino, D. J. Labelled Compd. Radiopharm. 1993, 36, 1097. (b) Dischino, D. Cook, D.; Saulnier, M.;
Tepper, M. J. Labelled Compd. Radiopharm. 1995, 36, 1097.
6. Preparation of 7-[N,N%-bis(benzyloxycarbonyl)guanidino]heptanoyl-(RS)-2-[(R)-sec-phenethoxy]glycine methyl
ester 7: Compound 5 (100 g, 171 mmol) was dissolved in CH2Cl2 in a 5 L three-necked round bottom flask,
equipped with a mechanical stirrer, an argon line and a 250 mL addition funnel. The solution was cooled to
0°C (internal) with an ice bath. Anhydrous pyridine (225 mL) was added dropwise through the addition
funnel. The internal temperature remained below 3°C during the course of the addition. Neat acetic anhydride
(130 mL, 1.493 mmol) was added dropwise. The resulting reaction mixture was warmed to rt and stirred for 1
h. The reaction was complete as indicated by TLC. The yellowish reaction mixture was cooled to 5°C and
quenched with dropwise addition of anhydrous MeOH (55 mL, 1.453 mmol). After 15 min of stirring at 5°C,
the reaction mixture was evaporated to an oily residue. Anhydrous toluene (500 mL) was added to the residue
and was removed on a rotovap. The addition and removal of anhydrous toluene were repeated three more
1
times until H NMR did not detect any pyridine in the residue. The resulting yellowish oil was re-dissolved in
CH2Cl2 (2 L). The solution was cooled to −78°C and EtAlCl2 (460.8 mL, 1.0 M in hexane, 460.8 mmol) was
added dropwise through an addition funnel under argon. After 10 min of stirring, a mixture of neat R-a-
methylbenzyl alcohol (33.2 mL, 276.4 mmol) and diisopropylethylamine (48.2 mL, 276.4 mmol) was added
dropwise. The reaction mixture was warmed to rt and stirred for 16 h. The reaction was complete as indicated
by TLC. It was cooled to −78°C and quenched with slow addition of Rochelle salt solution. The mixture was
warmed up to 0°C and stirred for 3 h. The cold mixture was filtered through a pad of Celite. The filtrate was
washed with Rochell salt solution (3×500 mL), dried over Na2SO4 and concentrated to give 150 g of crude oil,
which was passed through a pad of silica gel (300 g, 23–40 mm), eluted with 33–50% EtOAc in hexanes. The
solvents were evaporated to give 104 g of spectroscopically pure methyl ester 7 as a thick oil in 74% yield.
7. Chromatographic conditions for the separation of 10a and 10b: HPLC solid support: Porasil, Prepak (50 mm,
,
125 A), column size: 80 mm (d)×400 mm (l), mobile phase: 80–100% EtOAc in hexanes, sample load: 15 g in
50 mL of EtOAc/hexanes (4/1) and 10 mL of CH2Cl2 per injection, flow rate: 350 mL/min, detection: 254 nm

X. Wang, J. Thottathil / Tetrahedron: Asymmetry 11 (2000) 3665–3669
3669
UV, fractions: 500 ml each, collected: 8–9 fractions for the first diastereoisomer 10a and 14–15 for the second
diastereoisomer 10b. The contaminated fractions 10–11 for 10a and 12–13 for 10b (analyzed by TLC) were
combined, evaporated and re-injected.
8. The ee’s of DSG·3AcOH and DSG·3HCl were determined by HPLC analysis of the diastereomeric GITC
(2,3,4,6-tetra-O-acetyl-b-
D
-glycopyranosyl isothiocyanate) derivative according to the published procedure.3a
.