Design, synthesis and structure-activity relationship of a series of arginine aldehyde factor Xa inhibitors. Part 1: Structures based on the (D)- Arg-Gly-Arg tripeptide sequence

Article abstract of DOI:10.1016/S0960-894X(99)00582-X

A series of arginine aldehyde inhibitors was designed as transition state (TS) analogues based on the known factor Xa specific substrate Cbz-D- Arg-Gly-Arg-pNA. BnSO₂-(D)Arg-Gly-Arg-H (20) was found to be the most potent and selective inhibitor of factor Xa and prothrombinase activity in this series.

Full text of DOI:10.1016/S0960-894X(99)00582-X

Bioorganic & Medicinal Chemistry Letters 10 (2000) 13±16
Design, Synthesis and Structure±Activity Relationship of a Series
of Arginine Aldehyde Factor Xa Inhibitors. Part 1: Structures
Based on the (D)-Arg-Gly-Arg Tripeptide Sequence
Charles K. Marlowe,* Uma Sinha, Alice C. Gunn and Robert M. Scarborough
COR Therapeutics, Inc., 256 East Grand Avenue, South San Francisco, CA 94080, USA
Received 8 July 1999; accepted 1 October 1999
AbstractÐA series of arginine aldehyde inhibitors was designed as transition state (TS) analogues based on the known factor Xa
speci®c substrate Cbz-d-Arg-Gly-Arg-pNA. BnSO2-(d)Arg-Gly-Arg-H (20) was found to be the most potent and selective inhibitor
of factor Xa and prothrombinase activity in this series. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
P3, as a well as an Ile residue in P4, re¯ecting the nat-
ural cleavage sequence found in prothrombin. Interest-
ingly, S-2765, containing an unnatural d-Arg-residue in
P3, has a higher relative rate of cleavage by factor Xa
than S-2222 yet is the shorter tripeptide sequence. Both
substrates show some 30- to 40-fold greater reactivity
with factor Xa compared to thrombin, which is known
to prefer a Pro-residue in P2. We therefore reasoned
that the d-Arg-Gly-Arg sequence would be a good
starting point for the preparation of TS analogues
that might display both potent and speci®c inhibitory
activity for factor Xa.
There is a continuing need for the development of novel
anticoagulant agents for the treatment of cardiovascular
diseases to replace existing therapy.1±4 Direct and spe-
ci®c inhibition of the trypsin-like serine proteases in the
coagulation cascade, especially thrombin, have been the
focus of many e€orts over the last decade to design
novel anticoagulants. Thrombin occupies the ®nal step
in the conversion of ®brinogen to ®brin while also being
the most potent platelet agonist and, therefore, plays a
central role in both thrombosis as well as normal
hemostasis. An attractive alternative to thrombin inhi-
bition may be inhibition of the synthesis of thrombin
itself. Prothrombin conversion is catalysed most e-
ciently by factor Xa in conjunction with factor Va in the
prothrombinase complex which is primarily localized to
activated platelet surfaces.4
The use of arginine aldehydes as TS analogues inhibi-
tors for serine proteases with trypsin-like speci®city
dates back to the work of Bajusz et al.6 with their series
of thrombin and trypsin inhibitors. More recently,
Semple et al.,7 have utilized argininals for potent and
speci®c thrombin inhibitors. For preparation of argin-
inals by solution phase, we utilized the reduction of an
arginine lactam (Scheme 1a) as reported by Shuman et
al.8,9 The parent lactam 1 is readily prepared from Boc-
Arg(Cbz)OH,10 which is easily synthesized using Schot-
ten±Baumann conditions employing Cbz-Cl and Boc-
Arg. Coupling of the N-deprotected lactam with a sui-
tably protected dipeptide gives the tripeptide lactams
which can be reduced with LAH followed by careful
removal of protecting groups by hydrogenation to
a€ord the tripeptide argininals.11
We have utilized a transition state (TS) inhibitor
approach to design speci®c, potent inhibitors of factor
Xa recognizing that structural information found in
good substrates of a target enzyme often a€ord potent
TS inhibitors.5 The relative rates of hydrolysis for two
of the best known, commercially available, para-nitro-
anilide (pNA) factor Xa substrates are shown in Table 1.
Both substrates re¯ect the natural preference of factor
Xa for having an Arg-residue in P1 for the substrate
primary speci®city site and a Gly-residue in P2. The
substrate S-2222 contains the negatively charged Glu in
Alternatively, multiple analogues of tripeptide argininals
can also be prepared on solid phase utilizing a mod-
i®cation of the method of Webb et al.12 (Scheme 1b).
*Corresponding author.
0960-894X/00/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(99)00582-X

14
C. K. Marlowe et al. / Bioorg. Med. Chem. Lett. 10 (2000) 13±16
Table 1. Second order rate: kcat/kma
Results and Discussion
Substrate
Substrate
numberb
Factor Xa
(bovine)
Thrombin
Initially, because of synthetic accessibility, we prepared
the N-terminal Boc-derivatized tripeptide analogues,
rather than the Cbz-protected analogues, and identi®ed
Boc-d-Arg-Gly-Arg-H (3) as a very potent inhibitor of
factor Xa with IC50=50 nM.13 This analogue also dis-
played a desirable selectivity versus thrombin (Table 2).
Curiously, this inhibitor gave poor activity toward the
prothrombinase complex (abbreviated IIase).14±16 Eight
analogues were prepared in this initial series of P3
modi®cations (Table 2), which illustrates a stringent
requirement at P3 for both residue con®guration and
the highly basic guanidino group. For example, repla-
cement of P3 with l-Arg (4) results in a 40-fold loss of
inhibitory activity while replacement with d-Lys (5)
reduced activity by a factor of 200. The shorter and less
basic sidechain residue, d-Orn (6), also loses signi®cant
inhibitory activity. Substitution of the basic guanidine
functionality of the Arg sidechain with the urea-
containing residue d-citruline (d-Cit, 7) or with the
amidine containing residue d-Orn((-CˆNH)±CH3) 8,
diminishes activity compared to the Boc-d-Arg-Gly-Arg-H
compound. Either the dual hydrogen bond donating
ability of guanidine is critical or a cation-ꢀ interaction,
Bz-Ile-Glu-Gly-Arg-pNA
Cbz-d-Arg-Gly-Arg-pNA
S-2222
S-2765
0.35
2.27
0.018
0.106
aRates are expressed in terms of L/mmols.
bNumbering is that used by the manufacturer: Chromogenix. S-2222 is
supplied as a 50% mixture of the Glu(g-OMe) and the acid although
the methyl ester is some 30% less active.
Following their procedure, Boc-Arg(NO2)-semicarbazone
linker 2a (R2, R3=NO2, H) was synthesized and loaded
onto resin, followed by Boc-deprotection and amino
acid coupling using standard coupling techniques.
Cleavage from the resin was achieved by an exchange
reaction with formaldehyde. However, ®nal removal of
the nitro group from the Arg(NO2) residue via reductive
methods is often complicated by reduction of the alde-
hyde moiety. Arginine bis-Cbz-protection provides a
more reductively labile protecting group giving con-
sistently better yields of pure tripeptide aldehydes. Thus,
Boc-Arg(Cbz)2-linker 2b (R2, R3=Cbz) was prepared in
a similar fashion and utilized as shown in Scheme 1b.
Scheme 1a. Solution-phase synthesis via arginine lactam.
Scheme 1b. Modi®ed solid-phase synthesis.

C. K. Marlowe et al. / Bioorg. Med. Chem. Lett. 10 (2000) 13±16
Table 2. Tripeptide argininals with P3 modi®cations Table 4. Tripeptide argininals with P4 modi®cations
15
IC50s (mM)
IIa IIase IIa/Xa
IC50s (mM)
IIa IIase IIa/Xa
Analogue
Boc-P3-Gly-Arg-H
Xa
Analogue
P4-dArg-Gly-Arg-H
Xa
3
4
5
6
7
8
9
10
-dArg-
-(l)Arg-
0.050
2.0
10
12
3.0
1.0
0.10
3.0
180
79 29
4.0
3600
40
>10
42
33
88
4000
30
3
Boc-
H-
EtOCO-
BnSO2-
PhEtSO2-
NapEtSO2-
PhEtCO-
0.050
0.190
180
4
55 0.828
3600
290
18
19
20
21
22
23
24
25
26
27
28
29
-dLys-
>100 68
>500 65
>100 12
88 10
400
>100 12
0.030 >100 0.135 >3300
-dOrn-
-dCit-
0.015
0.028
0.027
0.200
0.220
50 0.012
64 0.038
39 0.020
75
170 0.18
3300
2300
1500
375
770
180
-dOrn((CˆNH)-CH3)-
-dHar-
0.17
2
-dHar(Me)4-
BnNHCO-
2-NapOAc-
HO2C(CH2)2CO-
HO2C(CH2)CO-
HO2CCO-
0.550 >100 0.69
0.224 >100
0.100 >100 0.50
0.056 95 0.63
0.058
2
450
1000
1900
100
as originally proposed by Johnson,17 may explain our
results. Along those lines, d-homoarginine (d-Har, 9)
retains activity although the tetra-methylated version,18
d-Har(Me)4 (10), loses signi®cant binding.
BnSO2-dArg-Sar-Arg-H 0.020
2
toward factor Xa, but a 4-fold increase in activity
toward the prothrombinase complex. Likewise, the
A fairly stringent SAR is also noted for P2 amino acid
residue replacements (Table 3). Substitution with Ala
(11) results in a 3-fold loss of inhibitory activity to
factor Xa as well as a signi®cant loss of thrombin selec-
tivity. Whereas, a d-Ala substitution 12 results in a
somewhat greater loss of activity, insertion of two
methyl groups at the alpha carbon as found in the Aib-
containing analogue 13 a€ords a 10-fold reduction in
activity. The b-Ala residue containing analogue 14
displayed signi®cantly reduced potency compared to the
other analogues. Not unexpectedly, based on substrate
speci®city, substitution of P2 with proline 15 increases
thrombin inhibitory activity (decreased IIa/Xa selectiv-
ity), but did not decrease factor Xa inhibitory activity.
Again, con®guration is extremely important at P2 since
the d-Pro analogue 16 is >1000-fold less active than
the l-Pro analogue. The a-napthylalanine substituted
analogue 17 at P2, which has been shown to be very
advantageous for inhibitors described by Corvas19 was
still active (IC50=0.14 mM) toward factor Xa in our
series but provided no additional advantage. Unfortu-
nately, none of the P2 substitutions resulted in higher
activity toward factor Xa over that provided by our
initial analogue Boc-d-Arg-Gly-Arg-H.
ethylcarbamoyl analogue 19 displayed
a 30-fold
increase in binding to prothrombinase, while thrombin
selectivity was retained. Even more striking is the activ-
ity of the benzylsulphonamide analogue (BnSO2, 20)
which displayed an IC50 value of 15 nM toward factor
Xa and equivalent activity to prothrombinase. Addi-
tional analogues such as the PhEtSO2-containing ana-
logue 21, and the aNapEtSO2-containing analogue
22, all maintain potent activity to both factor Xa and
prothrombinase. Of signi®cance, replacement of the
sulphonamide linkage of inhibitor 21 with an amide
linkage, as found in the PhEtCO-containing analogue
23, displayed almost a 10-fold decrease in factor Xa
inhibition and a 50-fold drop in prothrombinase inhibi-
tory activity. Both the urea linked analogue 24 and the
2-naphthoxyacetyl-containing analogue 25 lose sig-
ni®cant activity. The last series of P4 changes were syn-
thesized to investigate the known natural preference of a
negatively charged residue (i.e. Asp/Glu) albeit in P3.
As the chain length decreases within the succinic to
oxalic-containing analogues, inhibitory activity to fac-
tor Xa becomes equivalent to analogue 3. Finally, com-
bining the best P4 residue, BnSO2, with a Sar-residue at
P2 (analogue 29) again reinforced our earlier observa-
tions that a P2 Gly-residue is critically required to
maintain high factor Xa versus thrombin selectivity.
Finally, substitutions in the P4 position were explored
and shown to have a marked e€ect on factor Xa as well
as prothrombinase activity (Table 4). Of interest,
removal of the Boc group a€orded analogue 18, which
showed a 3- to 4-fold decreased inhibitory activity
Conclusion
Starting with the substrate sequence, Cbz-d-Arg-Gly-
Arg, we have designed a series of TS analogues of factor
Xa containing the argininal functionality. Initially Boc-
d-Arg-Gly-Arg-H (3) was found to be a very potent
factor Xa inhibitor as well as an extremely selective
inhibitor with respect to thrombin.20 Replacement of
both the P3 and P2 residues uncovers a very narrow
SAR suggesting an extremely speci®c interaction with
the enzyme. The best P4 replacements are the sulfona-
mide containing analogues (BnSO2- group a€ords the
most potent compound of this series, 20) which are also
e€ective inhibitors of the prothrombinase complex.
Although the P4 site can tolerate larger sulfonamides,
Table 3. Tripeptide argininals with P2 modi®cations
IC50s (mM)
Analogue Boc-d-Arg-P2-Arg-H
Xa
IIa
IIase IIa/Xa
3
-Gly-
-Ala-
-dAla-
-Aib-
-bAla-
-Pro-
0.050 >100
0.167
0.230 >10
0.550 >500
4
>2000
30
>40
>900
45
11
12
13
14
15
16
17
5.0
0.081
0.800
2.0
14
0.021
440
0.34
11
0.046
60
0.14
>500
0.20
>500
42
4
8
300
-d-Pro-
-l-aNapAla-

16
C. K. Marlowe et al. / Bioorg. Med. Chem. Lett. 10 (2000) 13±16
linkages that utilize amide or urea functionality are
substantially less active. The acidity of the N±H of this
linkage may be important in its interaction with this
enzyme.21±23 Concerning the interaction of the d-argi-
nine residue, recent X-ray crystallographic data24 for
the binding of the small molecule DX-9065a to factor
Xa, as well as for trypsin, suggests that the P3 d-argi-
nine sidechain may also bind into the `aryl-binding' (S4)
pocket. This would leave the sulphonamide moiety
available to interact with the backbone residues which
comprise the antiparallel b-sheet.
Sheman, S. L.; Lumpkin, R. H.; Tamura, S. Y.; Webb, T. R.
J. Am. Chem. Soc. 1992, 114, 3156.
13. The factor Xa and thrombin assays were performed at
23ꢀC in 0.02 M Tris±HCl bu€er, pH 7.5, containing 0.15 M
NaCl. The rates of hydrolysis of the factor Xa substrate S-
2765 (Chromogenix) and the thrombin substrate Chromozym
TH (Boehringer Mannheim) were monitored at 405 nm using
the SOFTmax 96-well plate reader (Molecular Devices) at a
concentration of 0.1 mM. The IC50 is determined from a 4-
parameter curve of the data using SOFTmax software. The
prothrombinase inhibition assay was performed in a plasma
free system with modi®cations to the method described by
Sinha, U.; Hancock, T. E.; Nzerem, J. J.; Lin, P.-H; Tomlin-
son, J. E.; Wolf, D. L. Thrombosis Research 1994, 75, 427.
14. No explanation is given for this decreased prothrombinase
activity although the corresponding substrate, Boc-d-Arg-Gly-
Arg-pNA is reportedly a good substrate for free factor Xa,15 it
is not know what the substrate activity is with the pro-
thrombinase complex.
Acknowledgements
We would like to thank Alan Laibelman for early pre-
parations of the resin linkers used in part of this study.
15. Gastavsson, S. I.; Arielly, S. US Patent No. 4,797,472,
110:232106. Chem. Abstr. 1989, 110, 250510.
16. Claeson, G. Blood Coag. Fibrin. 1994, 5, 411.
References and Notes
17. Lin, Z.; Johnson, M. E. FEBS Lett. 1995, 370, 1.
18. The tetramethylated guanidine analogue d-Har(Me)4 is
prepared by reacting the free amino group of the lysine side-
chain of Boc-d-Lys-Gly-Arg-H with excess HBTU and DIEA.
The product was con®rmed by ES±MS.
1. Scarborough, R. M. Annu. Rep. Med. Chem. 1995, 30, 71.
2. Edmunds, J. J.; Rapundalo, S. T. Annu. Rep. Med. Chem.
1996, 31, 51.
3. Scarborough, R. M. J. Enzyme Inhibition 1998, 14, 15.
4. Zhu, B.-Y.; Scarborough, R. M. Curr. Opin. Cardiovasc.
Pulm. Renal Invest. Drugs 1999, 1, 63.
19. Brunck, T. K.; Webb, T. R.; Ripka, W. WO9413693.
Chem. Abstr. 1994, 122, 133851.
5. Bartlett, P. A.; Marlowe, C. K. Biochemistry 1983, 22, 1085.
6. Bajusz, S.; Barabas, E.; Tolnay, P.; Szell, E.; Bagdy, D. J.
Peptide Protein Res. 1978, 12, 217.
7. Semple, J. E.; Minami, N. K.; Tamura, S. Y.; Brunck, T.
K.; Nutt, R. F.; Ripka, W. C. Bioorg. Med. Chem. Lett. 1997,
7, 2421.
8. Shuman, R. T.; Rothenberger, R. B.; Campbell, C. S.;
Smith, G. F.; Gi€ord-Moore, D. S.; Paschal, J. W.; Gesell-
chen, P. D. J. Med. Chem. 1995, 38, 4446.
9. Shuman, R. T. EP 0 542 525 A2. Chem. Abstr. 1993, 119,
250510.
20. These compounds are also quite selective with respect to a
number of other relevant proteases. The following IC50s (mM)
are noted for argininals 3, 20: bovine trypsin (0.048, 0.024), t-
PA (32.2, 4.4), plasmin (8.1, 1.1), aPC (56.4, 6.9), kallikrein
(4.1, 0.38).
21. Semple, J. E.; Rowley, D. C.; Brunck, T. C.; Ha-Uong, T.;
Minami, N. K.; Owens, T. D.; Tamura, S. Y.; Goldman, E.
A.; Siev, D. V.; Ardecky, R. J.; Carpenter, S. H.; Yu, G.;
Richard, B. M.; Nolan, T. G.; Hakanson, K.; Tulinsky, A.;
Nutt, R.; Ripka, W. C. J. Med. Chem. 1996, 39, 4531.
22. Krishnan, R.; Zhang, E.; Hakansson, K.; Arni, R. K.;
Tulinsky, A.; Wilby-Lim, M. S. L.; Levy, O. E.; Semple, J. E.;
Brunck, T. K. Biochemistry 1998, 37, 12094.
10. Boc Arg(Cbz)OH is also commercially available from
Bachem, A-4055.
11. The tripeptide aldehydes were identi®ed by their char-
acteristic HPLC trace showing at least three peaks presumably
re¯ecting the equilibrium between the free aldehyde, the
hydrate, and the two diasteriomeric cyclic aminal forms. Pur-
i®cation is performed by RP-HPLC using a water/acetonitrile/
0.1% TFA gradient. All ®nal compounds were characterized
by electrospray mass spectrometry as well as by their ability to
react with DNPH.
23. Sanderson, P. L.; Cutrona, K. J.; Dorsey, B. D.; Dyer, D.
L.; McDonough, C. M.; Naylor-Olsen, A. M.; Chen, I.-W.;
Chen, Z.; Cook, J. J.; Gardell, S. J.; Krueger, J. A.; Lewis, S.
D.; Lin, J. L.; Lucus, B. J., Jr.; Lyle, E. A.; Lynch, J. J., Jr.;
Stranieri, M. T.; Vastag, K.; Shafer, J. A.; Vacca, J. P. Bioorg.
Med. Chem. Lett. 1997, 8, 817.
24. Brandstetter, H.; Kuhne, A.; Bode, W.; Huber, R.; vd Saal,
W.; Wirthensohn, K.; Engh, R. A. J. Biol. Chem. 1996, 271,
29988.
12. Murphy, A. M.; Dagnino, R.; Vallar, P. L.; Trippe, A. J.;