Synthesis of derivatives of the novel cyclophilin-binding immunosuppressant sanglifehrin A with reduced numbers of polar functions

Article abstract of DOI:10.1016/S0960-894X(01)00293-1

The syntheses and the biological activities of 53-deoxo sanglifehrin A (2) and 61-deoxy octahydrosanglifehrin A (3) are described. Compound 2 shows intracellular cyclophilin (CyP)-binding and immunosuppressive activity in the mixed-lymphocyte reaction (MLR) similar to that of sanglifehrin A (1). Compound 3 is much less active in the MLR despite unchanged intracellular CyP-binding. This indicates that the 53-keto group is not necessary for immunosuppressive activity, while the 61 hydroxy group is required.

Full text of DOI:10.1016/S0960-894X(01)00293-1

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1609–1612
Synthesis of Derivatives of the Novel Cyclophilin-Binding
Immunosuppressant Sanglifehrin A with Reduced Numbers of
Polar Functions
Rolf Banteli,* Jurgen Wagner and Gerhard Zenke
Novartis Pharma AG, CH-4002 Basel, Switzerland
Received 4 December 2000;accepted 27 April 2001
Abstract—The syntheses and the biological activities of 53-deoxo sanglifehrin A (2) and 61-deoxy octahydrosanglifehrin A (3) are
described. Compound 2 shows intracellular cyclophilin (CyP)-binding and immunosuppressive activity in the mixed-lymphocyte
reaction (MLR) similar to that of sanglifehrin A (1). Compound 3 is much less active in the MLR despite unchanged intracellular
CyP-binding. This indicates that the 53-keto group is not necessary for immunosuppressive activity, while the 61 hydroxy group is
required. # 2001 Published by Elsevier Science Ltd.
Sanglifehrin A (1) was isolated at Novartis from a fer-
mentation broth of the actinomycete strain Strepto-
myces flaveolus and is the most abundant representative
of a series of related immunosuppressive natural pro-
ducts identified by screening for novel cyclophilin-bind-
ing metabolites.1 Compound 1 exhibits a 20-fold higher
affinity for cyclophilin (CyP) than cyclosporine A in a
cell-free binding assay. However, its immunosuppressive
activity is 10-fold lower as determined by the mixed
lymphocyte reaction (MLR).1a Compound 1 does not
inhibit the phosphate activity of calcineurin, the target
of CyP/cyclosporine A complex, and does not affect IL-
2 gene expression indicating a different mode of action
than cyclosporine A.1a
Sanglifehrin A consists of a 22-membered macrocycle,
bearing in position 23 a nine-carbon chain terminated
by a unique spirobicyclic moiety. The macrocycle
*Corresponding author. Tel.: +41-61-324-5455;fax: +41-61-324-6735;e-mail: rolf.baenteli@pharma.novartis.com
0960-894X/01/$ - see front matter # 2001 Published by Elsevier Science Ltd.
PII: S0960-894X(01)00293-1

1610
R. Banteli et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1609–1612
¨
contains a tripeptide subunit consisting of valine and
the unusual amino acids meta-tyrosine and piperazic
acid.
of 7 was effected by treatment with Bu3SnH in the pre-
sence of AIBN to give 8 in 36% yield. Final desilylation
led to 53-deoxo sanglifehrin A (2).3 Attempted deoxy-
genation of 6 under the same conditions as above was
unsuccessful. Instead, a diastereomeric mixture of the
five-membered ring products 9 was obtained, which had
formed through reaction of the intermediate radical
with the 26–27 double bond in a 5-exo-trig reaction. No
further attempts were made towards 31-deoxy sangli-
fehrin A.
Our aim was to determine which functional groups of
sanglifehrin A (1) are essential for immunosuppressive
activity as well as to improve the biological activity. The
relatively low immunosuppressive activity of 1 in com-
parison to its high affinity for CyP could be the result of
poor cell permeation due to the presence of polar
hydroxy and keto functions. We anticipated that redu-
cing the number of polar groups could lead to increased
cell permeation and would reveal their importance for
the immunosuppressive activity. To this end, we fol-
lowed synthetic approaches2 as well as natural product
derivation. In this paper we report the syntheses of 53-
deoxo sanglifehrin A (2) and 61-deoxy octahydro san-
glifehrin A (3) as well as attempts towards preparing 31-
deoxy sanglifehrin A.
In order to remove the phenolic hydroxy group of 1, a
Pd catalyzed reduction of the corresponding triflate
using formic acid as hydrogen donor was envisioned.
Thus, 1 was first converted to its triflate 10 using N-
phenylbis(trifluoromethanesulfonimide)4 in 77% yield
(Scheme 2), which was treated with catalytic amounts of
Pd(OAc)2 and 1,10-bis(diphenylphosphino)ferrocene
(dppf) and excess formic acid.5 No reaction occurred at
room temperature and at 50 ꢀC 10 decomposed. The
poor reactivity was attributed to the complexation of
the palladium catalyst to the 1,3-diene systems. There-
fore we decided to remove the double bonds. Thus, 1
was hydrogenated using palladium on charcoal to give
octahydro sanglifehrin A (11) in 34% yield as an epi-
meric mixture at position 24. By the same method as for
1, 11 was converted to its triflate 12. Treatment of 12
with a combination of catalytic amounts of Pd(OAc)2
For the synthesis of 53-deoxo sanglifehrin A (2) the phe-
nolic hydroxy group of 1 was silylated by treatment with
N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide to
give 4 in 94% yield (Scheme 1). NaBH4 reduction of the
ketone 4 gave the alcohol 5 as a 1:1 epimeric mixture.
Treatment of 5 with phenylchlorothionoformate led to
the 31-thionocarbonate 6 and the 53-thionocarbonate 7
in 11 and 18% yield, respectively. Barton-deoxygenation
Scheme 1. Reagents and conditions: (a) N-tert-butyldimethylsilyl-N-methyltrifluoroacetamide (2 equiv), acetonitrile, 50 ꢀC, 3 h, 94%;(b) NaBH
4
(1.5 equiv), MeOH, 0 ꢀC, 30 min, quant;(c) ClCSOC H5 (2 equiv), DMAP (cat), ClCH2CH2Cl, 3 days, rt, 6: 11%, 7: 18%;(d) Bu SnH (2 equiv),
6
3
AIBN (0.25 equiv), benzene, reflux, 45 min, 36%;(e) TBAF (1.5 equiv), THF, 0 ꢀC, 30 min, quant;(f) Bu SnH (2 equiv), AIBN (0.25 equiv), ben-
zene, reflux, 45 min.
3

R. Banteli et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1609–1612
¨
1611
and (dppf)PdCl2 dichloromethane 1:1 complex and
excess formic acid now led to the desired 61-deoxy
octahydro sanglifehrin A (3)6 in 81% yield.
53-Deoxo sanglifehrin A (2) shows similar values in cell-
free and intracellular CyP-binding and hence no
improved permeation (cell-free/intracellular Cyp-bind-
ing ratio 0.8). Thus, the removal of the 53-keto group of
1 does not improve cell permeation. Octahydro sangli-
fehrin A (11) has a 26-fold weaker affinity for CyP than
sanglifehrin A (1). The decreased binding affinity of 11
compared to 1 could be due to the increased flexibility
of the compound.
The biological activities of the compounds with respect
to CyP-binding and immunosuppression are summar-
ized in Table 1. CyP-binding was analyzed in both a
cell-free assay and a cellular assay. Cell-free CyP-bind-
ing was assessed by competition of the compounds with
the interaction of CyP-A and 1.7 Intracellullar CyP-
binding was determined in a competitive format with
61-Deoxy octahydro sanglifehrin A (3) has a relative
IC50 of 300 in the cell-free CyP assay, which indicates
that the removal of the phenolic hydroxy group results
in an additional 11-fold loss in binding activity. How-
ever, the intracellular CyP-binding of 11 and 3 is
roughly unchanged, which indicates a better permeation
of both compounds in comparison to 1 (cell-free/intra-
cellular CyP-binding ratio of 65 and 270, respectively).
3
whole cells and H-cyclosporine A.8 The ratio cell-free/
intracellular CyP-binding is a way to estimate the cell
permeation of sanglifehrin derivatives. For example, if
cell-free binding decreases, that is if the cell-free IC50
increases, and if intracellular binding is unchanged, the
ratio cell-free/intracellular binding will be >1 and is
interpreted as an increased cell-permeation. Likewise, a
ratio <1 indicates a decrease in cell-permeation. The
immunosuppressive activity of the compounds was
assessed by MLR as described.1a To compensate for
variation of absolute IC50 values between independent
experiments of 1 and its derivatives, 1 was used as a
reference in each experiment. Thus, the IC50 values
given are relative values compared to sanglifehrin A (1)
(rIC50=IC50 of compound/IC50 of 1).
The MLR activity of 2 is slightly increased indicating
that the 53-keto group is not essential for immunosup-
pressive activity. Compound 11 is almost equipotent to
1 in the MLR despite a 26-fold loss in cell-free CyP
binding. This indicates that the increased permeation
compensates to some degree for the reduced CyP
binding. Alternatively, the immunosuppressive activity
Scheme 2. Reagents and conditions: (a) Tf2NPh (3.6–5.4 equiv), EtN(i-Pr)2 (20 equiv), ClCH2CH2Cl, rt, 3–6 days, 10: 85%, 12: 77%;(b) Pd(OAc) 2
(0.1 equiv), dppf (0.1 equiv), HCOOH (12.6 equiv), EtN(i-Pr)2 (25.2 equiv), DMF, rt, 5.5 h, no reaction;(c) Pd/C, H , EtOH, 35 min, rt, 34%;(d)
2
Pd(OAc)2 (0.1 equiv), (dppf)PdCl.2CH2Cl2 1:1 (0.1 equiv), HCOOH (12.6 equiv), EtN(i-Pr)2 (25.2 equiv), DMF, rt, 5.5 h, 81%.
Table 1. Biological activities of compounds
CyP binding
Substance
Cell-free7
rIC50
Intracellular8
rIC50
‘Permeation’ ratio
cell free/intracellular
MLR1a
rIC50
1
1
1
1
1
2
11
3
1.5ꢁ0.1
2.0ꢁ0.6
0.4ꢁ0.2
1.1ꢁ0.1
0.8
65
270
0.5ꢁ0.1
1.7ꢁ0.5
22ꢁ4
26ꢁ2
300ꢁ50
Mean relative IC50 values (rIC50)ꢁstandard error of 3–4 experiments are shown. Absolute IC50 values of sanglifehrin A (1) are as follows: cell-free
CyP-binding, 0.001–0.011 mM;intracellular CyP-binding, 4–16 mM;and MLR, 0.07–0.42 mM.

1612
R. Banteli et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1609–1612
¨
might be independent of CyP binding and the region of
1, which interacts with a yet unknown target, is not
affected by hydrogenation. The overall immunosup-
pressive activity of 3 is decreased in comparison to 11
indicating that the 61 hydroxy group is crucial for
immunosuppressive activity. This information is con-
sistent with our view that the tripeptide unit is the
structural motive determining the CyP binding and
immunosuppressive activity.2d
Zenke, G.;Wagner, J. Angew. Chem., Int. Ed. 1999, 38, 2443.
(e) For a total synthesis of the sanglifehrin A, see: Nicolaou,
K. C.;Ohshima, T.;Murphy, F.;Barluenga, S.;Xu, J.;Wins-
singer, N. Chem. Commun. 1999, 9, 809. (f) Nicolaou, K. C.;
Xu, J.;Murphy, F.;Barluenga, S.;Baudoin, O.;Wei, H.-X.;
Gray, D. L. F.;Ohshima, T. Angew. Chem., Int. Ed. 1999, 38,
2447. (g) Nicolaou, K. C.;Murphy, F.;Barluenga, S.;
Ohshima, T.;Wei, H.;Xu, J.;Gray, D. L. F.;Baudoin, O.
J.
Am. Chem. Soc. 2000, 122, 3830. (h) Paquette, L. A.;
Konetzki, I.;Duan, M. Tetrahedron Lett. 1999, 40, 7441.
3. Spectral data for compound 2. The 1H NMR interpretation
1
In summary, our results show that very selective chemi-
cal derivations can be performed on this complex nat-
ural product. The 53-deoxo compound 2 shows similar
activities compared to 1, which indicates that the 53-
keto group is not necessary for immunosuppressive
activity and its removal does not alter cell permeation.
On the contrary, the 61-deoxy compound 3 shows
decreased MLR activity compared to 11 despite
increased cell permeation, which indicates that the 61
hydroxy group is crucial for immunosuppressive activ-
ity. Further work was undertaken to increase the cell
permeation while maintaining or even increasing the
immunosuppressive activity of sanglifehrin A (1).
Results will be reported elsewhere.
was corroborated by an HSQC spectrum. H NMR (DMSO-
d6, 500 MHz) d 9.25 (s, 1H), 8.14 (d, J=6.5 Hz, 1H), 7.89 (s,
1H), 7.44 (d, J=9.0 Hz, 1H), 7.06 (t, J=7.7 Hz, 1H), 6.59 (dd,
J=1.0 Hz, J=7.0 Hz, 1H), 6.57 (d, J=7.2 Hz, 1H), 6.51 (s,
1H), 6.21 (dd, J=11.4 Hz, J=15.0 Hz, 1H), 6.13 (dd, J=11.4
Hz, J=15.0 Hz, 1H), 6.01 (dd, J=10.8 Hz, J=13.8 Hz, 1H),
5.97 (t, J=11.4 Hz, 1H), 5.68 (td, J=7.2 Hz, J=14.4 Hz, 1H),
5.58 (m, 3H), 5.40 (m, 2H), 5.23 (d, J=9.4 Hz, 1H), 4.79 (d,
J=3.5 Hz, 1H), 4.49 (d, J=11.4 Hz, 1H), 4.18 (m, 1H), 4.05
(m, 2H), 3.90 (m, 1H), 3.82 (m, 1H), 3.69 (t, J=9.6 Hz, 1H),
3.56 (m, 1H), 3.51 (m, 1H), 2.78–2.23 (m, 7H), 2.13–1.08
(m, 29H), 1.71 (s, 3H), 0.93–0.75 (m, 24H, 8ꢂaliphatic
CH3 including 53-CH2–CH3), 0.62 (d, J=6.7 Hz, 3H);
C60H93N5O12, MScalcd for MꢃH requires 1074.68, found
1074.6.
4. Hendrickson, J. B.;Bergeron, R. Tetrahedron Lett. 1973,
4607.
5. Cabri, W.;De Bernardinis, S.;Francalanci, F.;Penco, S. J.
Chem. Soc., Perkin Trans. 1 1990, 428.
Acknowledgements
6. Spectral data for compound 3. The 1H NMR interpretation
1
The authors thank L. Walliser, D. Wagner, U. Stritt-
matter-Keller and S. Fuchs for their excellent technical
assistance.
was corroborated by an HSQC spectrum. H NMR (DMSO-
d6, 400 MHz) d 8.48 (s, br, 1H), 7.90 (s, 1H), 7.38–7.12 (m,
4H), 5.80 (m, 1H), 5.59 (d, J=4.8 Hz, 1H), 5.15–4.65 (m, 3H),
4.18 (s, 1H), 4.12 (m, 1H), 3.98 (d, J=6.0 Hz, 1H), 3.70 (m,
2H), 3.58 (s, br, 1H), 3.50 (m, 1H), 3.22 (m, 1H), 2.90–2.60 (m,
ca. 6H), 2.50–2.30 (m, 2H), 2.18 (m, 1H), 2.07 (s, 3H), 2.10–
1.00 (m, ca. 44H), 0.95–0.70 (m, 21H), 0.62 (d, J=6.7 Hz, 3H);
although 3 is a mixture of diastereomers this is not apparent
from the NMR spectrum;C 60H99N5O12, MScalcd for MꢃH
requires 1080.73, found 1080.7.
7. The cell-free CyP-binding was determined in an enzyme-
linked immunosorbent assay analogous to the assay described
in Schneider, H.;Charara, N.;Schmitz, R.;Wehrli, S.;Mikol,
V.;Zurini, M. G.;Quesniaux, V. F.;Movva, N. R. Biochem-
istry 1994, 33, 8218 with the exception that cyclosporine con-
jugated to bovine serum albumin was replaced by sanglifehrin
coupled to bovine serum albumin.
References and Notes
1. (a) Sanglier, J.-J.;Quesniaux, V.;Fehr, T.;Hofmann, H.;
Mahnke, M.;Memmert, K.;Schuler, W.;Zenke, G.;
Gschwind, L.;Maurer, C.;Schilling, W. J. Antibiot. 1999, 52,
466. (b) Fehr, T.;Kallen, J.;Oberer, L.;Sanglier, J.-J.;Schil-
ling, W. J. Antibiot. 1999, 52, 474. (c) Fehr, T.;Oberer, L.;
Quesniaux Ryffel, V.;Sanglier, J.-J.;Schuler, W.;Sedrani, R.
PCT International Patent Application WO 97/02285-A1,
CAN 126:170491.
2. (a) Banteli, R.;Brun, I.;Hall, P.;Metternich, R.
Tetra-
8. Intracellular CyP-binding was assessed in a competitive
format with whole cells and 3H-cyclosporine A as described in
Baumann, G.;Andersen, E.;Quesniaux, V.;Eberle, M. K.
Transplant. Proc. 1992, 24, 43.
hedron Lett. 1999, 40, 2109. (b) Metternich, R.;Denni, D.;
Thai, B.;Sedrani, R. J. Org. Chem. 1999, 64, 9632. (c) Hall, P.;
Brun, J.;Denni, D.;Metternich, R. Synlett 2000, 3, 315. (d)
Martin Cabrejas, L. M.;Rohrbach, S.;Wagner, D.;Kallen, J.;