Miraziridine A: Natures blueprint towards protease class-spanning inhibitors

Article abstract of DOI:10.1016/j.bmcl.2003.12.030

The natural product miraziridine A isolated from the marine sponge Theonella aff. mirabilis unifies within one molecule three structurally privileged elements: (i) (2R,3R)-aziridine-2,3-dicarboxylic acid, (ii) (3S,4S)-4-amino-3-hydroxy-6-methylheptanoic acid (statine), and (iii) (E)-(S)-4-amino-7-guanidino-hept-2-enoic acid (vinylogous arginine). The alignment of them realized in the tetrapetide allows for a simultaneous inhibition of the proteolytic activity of trypsin-like serine proteases, papain-like cysteine proteases, and pepsin-like aspartyl proteases. Therefore, this unique compound represents a blueprint for the design of protease class-spanning inhibitors.

Full text of DOI:10.1016/j.bmcl.2003.12.030

Bioorganic & Medicinal Chemistry Letters 14 (2004) 855–857
Miraziridine A: natures blueprint towards protease
class-spanning inhibitors
Norbert Schaschke*
Max-Planck-Institut fu¨r Biochemie, D-82152 Martinsried, Germany
Received 21 October 2003; revised 28 November 2003; accepted 3 December 2003
Abstract—The natural product miraziridine A isolated from the marine sponge Theonella aff. mirabilis unifies within one molecule
three structurally privileged elements: (i) (2R,3R)-aziridine-2,3-dicarboxylic acid, (ii) (3S,4S)-4-amino-3-hydroxy-6-methylheptanoic
acid (statine), and (iii) (E)-(S)-4-amino-7-guanidino-hept-2-enoic acid (vinylogous arginine). The alignment of them realized in the
tetrapetide allows for a simultaneous inhibition of the proteolytic activity of trypsin-like serine proteases, papain-like cysteine
proteases, and pepsin-like aspartyl proteases. Therefore, this unique compound represents a blueprint for the design of protease
class-spanning inhibitors.
#
2003 Elsevier Ltd. All rights reserved.
Miraziridine A (1) (Fig. 1) is a natural product that has
been isolated from the marine sponge Theonella aff.
Correspondingly, three building blocks, i.e., (i) (2R,3R)-
aziridine-2,3-dicarboxylic acid mono ethyl ester (2),
(ii) Bpoc-Leu-(3S,4S)-Sta-Abu-OH (4), and (iii) H-
1
mirabilis. Its structure was elucidated by a combination
1
of chemical and 2D NMR techniques. Among the
vArg(Boc) -OEt (5) were selected for the total synthesis
2
great variety of protease inhibitors from natural sour-
ces, this tetrapeptide seems to be unique because it uni-
fies within one molecule three inhibitory elements: (i)
of 1. The half ester 2 was obtained in a six-step sequence
(2R,3R)-aziridine-2,3-dicarboxylic acid, (ii) (3S,4S)-4-
amino-3-hydroxy-6-methylheptanoic acid (statine), and
(iii) (E)-(S)-4-amino-7-guanidino-hept-2-enoic acid
(vinylogous arginine). Whereas the former ones con-
stitute known cysteine and aspartyl protease inhibi-
2
,3
tors, the latter one possibly inhibits trypsin-like serine
proteases. To investigate this hypothesis whether the
alignment of the inhibitory elements present in 1 yields
an inhibitor blocking proteases belonging to three dif-
ferent classes in a simultaneous fashion, the compound
1
was synthesized and the inhibitory properties assessed.
The retrosynthetic analysis (Fig. 1) was guided by the
idea to introduce 2 at a late stage of the synthesis before
the final deprotection step since the aziridine moiety is
known to be sensitive to nucleophilic ring opening. To
realize this approach, the regime of protecting groups
within tetrapeptide 3 has to allow for a selective
unmasking of the N-terminus. Thus, the Bpoc/Boc-
4
combination was selected that is also compatible with
the double bond present in the vinylogous arginine.
*
Tel.: +49-89-85783910; fax: +49-89-85782847; e-mail: schaschk@
biochem.mpg.de
Figure 1. Disconnetion of miraziridine A (1) into subunits. The amino
acid code for each compound is given below the formula respectively.
0
960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.12.030

856
N. Schaschke / Bioorg. Med. Chem. Lett. 14 (2004) 855–857
Scheme 1. Synthesis of tripeptide 4. Reaction conditions: (i) HOSu,
ꢁ
.
DCC, acetonitrile, 0 C!rt; (ii) H-(3S,4S)-Sta-OH HCl, DIEA,
CHCl
3
, rt (77% over two steps); (iii) HOSu, DCC, acetonitrile,
O (1:1, v/v), rt (79%
ꢁ
0 C!rt; (iv) H-Abu-OH, 1N NaOH, dioxane/H
over two steps); Bpoc=2-(biphenyl-4-yl)prop-2-yloxycarbonyl;
2
Scheme 3. Assembly of the building blocks. Reaction conditions: (i)
ꢁ
HOSu=N-hydroxy-succinimide; Sta=statine, (3S,4S)-4-amino-3-
hydroxy-6-methylheptanoic acid; DIEA=N-ethyldiisopropylamine;
.
5
HCl, DIEA, EDC, HOAt, CHCl
ꢁ
3
, 0 C!rt (86%); (ii) TFA/CHCl
3
0
(0.5:99.5, v/v), 0 C!rt (71%); (iii) 2, DPPA, NEt , DMF (86%); (iv)
DCC=N,N -dicyclohexyl-carbodiimide;
acid.
Abu=(S)-2-aminobutyric
3
ꢁ
TFA/CHCl
3
(5:95, v/v), 0 C!rt (46%); (v) Porcine liver esterase,
.
0
Tris HCl buffer (50 mM, pH 8.0) (65%); EDC=N-ethyl-N -(3-dime-
thylaminopropyl)carbodiimide hydrochloride; HOAt=1-hydroxy-7-
aza-1H-benzotriazole; TFA=trifluoroacetic acid; DPPA=diphenyl
phosphorazidate; Tris=tris(hydroxymethyl)aminomethane.
Table 1. Inhibitory properties of Miraziridine A
Protease class
Protease
Affinity
ꢀ
5
Serine protease
Cysteine protease
Trypsin
Cathepsin L
Cathepsin B
Pepsin
6Â10
M
ꢀ1 ꢀ1
6
1.0Â10 M
s
4
ꢀ1 ꢀ1
1.5Â10 M
s
M
ꢀ
8
Aspartyl protease
1.4Â10
Scheme 2. Synthesis of vinylogous arginine 5. Reaction conditions: (i)
3.
2
2
0
CuCO Cu(OH)
azole, DIEA, formamide, rt (78%); (iii) EDTA, NaHCO
O/acteone (1:1, v/v), rt (84%); (iv) TBTU, HOBt,
3.
, H
O, reflux (88%); (ii) N,N -bis-Boc-1-guanylpyr-
3
, Fmoc-OSu,
procedure was performed. First attempts to saponify
the ester functions by aqueous LiOH failed. The ‘N-
terminal’ ester function could be removed smoothly
H
2
ꢁ
HN(CH
3
)OCH HCl, DIEA, CHCl
3
, 0 C!rt (80%); (v) LiAlH
4
,
ꢁ
Et
ꢀ
2
O, 0 C!rt (67%); (vi) triethyl phosphonoacetate, NaH, THF,
ꢁ
ꢁ
50 C!rt; (vii) HNEt
2
/DMF (1:4, v/v) (42% over two steps);
with one equivalent LiOH at 0 C, whereas saponifica-
Fmoc=9-fluorenylmethoxycarbonyl; EDTA=ethylenediaminetetra-
acetic acid; Fmoc-OSu=N-(9-fluorenylmethoxycarbonyloxy)-succin-
imide; TBTU=2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium
tetrafluoroborate; HOBt=1-hydroxy-1H-benzotriazole; THF=tetra-
hydrofurane; DMF=N,N-dimethylformamide.
tion of the ‘C-terminal’ one was possible only by pro-
longed exposure to an excess of base which was
accompanied by the nucleophilic ring opening of the
aziridine moiety. To circumvent this problem, a
enzyme-assisted ester hydrolysis using pig liver esterase
was applied to give smoothly 1. Starting from 4, mir-
aziridine A (1) was obtained in five steps in an overall
according to the literature starting from (2S,3S)-tartaric
,5
2
acid diethyl ester. The tripeptide 4 was synthesized
6
1
from Bpoc-Leu-OH
succinimide ester activation for each coupling step
Scheme 1). The orthogonally protected amino acid
using common hydroxy-
yield of 16%. The H NMR data of synthetic 1 are in
good agreement with those reported for the natural
1
(
product. The optical rotation however, differs slightly
ꢁ
2
0
10
aldehyde 11, selected as starting material for the synthesis
of the vinylogous arginine 5, was obtained via the
corresponding Weinreb amide 10 in five steps from
between synthetic ([a] ꢀ89.9 (c 0.087, MeOH)) and
D
2
0
ꢁ
1
natural ([a]D ꢀ74 (c 0.085, MeOH)) 1.
7
,8
l-ornithine (Scheme 2). According to a procedure of
Bastiaans, the crude aldehyde 11 was immediately
The capability of 1 to inhibit proteases belonging to
different classes was assassed using trypsin, cathepsin B,
cathepsin L, as well as papain (Table 1). The inhibitory
potency of 1 against the serine protease trypsin is in the
same order of magnitude as that of benzamidine (60
verus 18 mM, respectively) suggesting that the inhibition
is mainly due to an P1/S1 interaction. Moreover, 1 is a
potent and irreversible inhibitor of the papain-like
cysteine proteases cathepsin B and L. Comparing the
second order rate constants with those of HO-(2R,3R)-
Azy-Leu-HN-CH -CH -CH(CH ) (cathepsin B: k2/
9
converted by a Horner–Emmons–Wadsworth olefina-
tion using triethyl phosphonoacetate/NaH at
ꢁ
. The latter one is readily deprotected using diethyl-
ꢀ
50 C!rt to give a mixture of 5 and Fmoc-protected
5
amine/DMF (1:4, v/v). Scheme 3 shows the assembly of
the building blocks. Using EDC/HOAt, the tripeptide 4
was coupled with 5 followed by the cleavage of the
Bpoc-group with TFA/CHCl (0.5:95.5, v/v). Then, the
3
N-terminus was acylated with 2 by the DPPA-method
2
2
2
3 2
4
ꢀ1 ꢀ1
5
according to Martichonok et al. yielding protected
miraziridine A (14). Finally, a two-step deprotection
K =1.7Â10
M
that structurally resembles the N-terminal
s
; cathepsin L: k /K =1.2Â10
i
2
i
ꢀ
1 ꢀ1 2
M
s
)

N. Schaschke / Bioorg. Med. Chem. Lett. 14 (2004) 855–857
857
part of 1, it becomes evident that the inhibitory activity
Acknowledgements
can be attributed mainly to this part of the molecule.
Additionally, this is supported from the structural point
of view. An inspection of the X-ray structure of E-64 in
complex with cathepsin L shows that the N-terminal
part of 1 can easily adopt an conformation that is simi-
lar to the binding mode of E-64, where the C-terminal
part of 1 is fully solvent exposed. Finally, also the
aspartyl protease activity of the pepsin is efficiently
The author would like to thank Prof. W. Machleidt
and Prof. C. P. Sommerhoff for the inhibition
kinetics. The study was supported by the SFB 469 of the
1
1
¨
Ludwig-Maximilians-Universitat of Munich.
References and notes
inhibited by 1 (K =14 nM). Interestingly, the central
i
part of the well known aspartyl protease inhibitor
1. Nakao, Y.; Fujita, M.; Warabi, K.; Matsunaga, S.;
Fusetani, N. J. Am. Chem. Soc. 2000, 122, 10462.
pepstatin
OH; K =0.046 nM), that is, -Val-(3S,4S)-Sta-Ala- is
(Iva-Val-Val-(3S,4S)-Sta-Ala-(3S,4S)-Sta-
3
2. Martichonok, V.; Plouffe, C.; Storer, A. C.; Me
Jones, J. B. J. Med. Chem. 1995, 38, 3078.
3. Rich, D. H.; Sun, E. T. O.; Ulm, E. J. Med. Chem. 1980,
3, 27.
´
nard, R.;
i
structurally mimicked by 1. An inspection of the X-ray
1
2
2
structure of pepsin in complex with pepstatin revealed
that particularly this structural motif is essential for the
inhibitory properties of pepstatin. Furthermore, a
4
. (a) Sieber, P.; Iselin, B. Helv. Chim. Acta 1968, 51, 614. (b)
Wang, S.-S.; Merrifield, R. B. Int. J. Peptide Protein Res.
1
. Mori, K.; Iwasawa, H. Tetrahedron 1980, 36, 87.
. Kemp, D. S.; Fotouhi, N.; Boyd, J. G.; Carey, R. I.;
Ashton, C.; Hoare, J. Int. J. Peptide Protein Res. 1988, 31,
969, 1, 235.
comparison of the K -value of a truncated form of pep-
i
5
6
13
statin (Iva-Val-(3S,4S)-Sta-Ala-Iaa; K =3.0 nM) with
i
that of 1 suggests that the inhibitory activity originates
form the tripeptide part-Leu-(3S,4S)-Sta-Abu-.
3
59.
7
. Wu, Y.; Matsueda, G. R.; Bernatowicz, M. Synth.
Commun. 1993, 23, 3055.
8. Fehrentz, J.-A.; Castro, B. Synthesis 1983, 676.
In conclusion, aligning three privileged inhibitory
elements within one molecule, nature provides us with 1
a blueprint how to design efficiently protease inhibitors
of the small molecule type capable to block simulta-
neously serine, cysteine, and aspartyl protease activity.
In the light of pathopysiological conditions like tumor
9
. Bastiaans, H. M. M.; van der Baan, J. L.; Ottenheijm,
H. C. J. J. Org. Chem. 1997, 62, 3880.
1
1
1
0. The concentration is corrected according to quantitative
amino acid analysis.
1. Fujishima, A.; Imai, Y.; Nomura, T.; Fujisawa, Y.;
Yamamoto, Y.; Sugawara, T. FEBS Lett. 1997, 407, 47.
2. Fujinaga, M.; Chernaia, M. M.; Tarasova, N. I.; Mosi-
mann, S. C.; James, M. N. G. Protein Sci. 1995, 4, 960.
metastasis and invasion that is mediated in
a
cooperative manner by an ensemble of proteases
belonging to different classes, protease class-spanning
inhibitors appear as an interesting new approach to
possibly interfere in such processes.
13. Rich, D. H.; Sun, E. T. O. Biochem. Pharmacol. 1980, 29,
2205.