Synthesis of 13C-Labeled Pyrazinone Thrombin Inhibitors and Elucidation of Metabolic Activation Pathways

Article abstract of DOI:10.1002/hlca.200490063

In support of a research program aimed at discovering orally bioavailable thrombin inhibitors, ¹³C-labeled 3-aminopyrazinone acetamide thrombin inhibitors have been prepared to aid in the structural elucidation of key metabolites during preclin

Full text of DOI:10.1002/hlca.200490063

674
Helvetica Chimica Acta Vol. 87 (2004)
Synthesis of 13C-Labeled Pyrazinone Thrombin Inhibitors and Elucidation of
Metabolic Activation Pathways
by Jonathan Z. Hoa)*, MatthewP. Braun a), Raju Subramanianb), Ying-Duo Gaoc), Dennis C. Deana),
and David G. Melilloa)
a) Department of Drug Metabolism, Merck Research Laboratories, Merck & Co., Inc., 128 East Lincoln
Avenue, Rahway, NJ 07065, USA (E-mail: jonathan_ho@merck.com)
b) Department of Drug Metabolism, P.O. Box 4 Sumneytown Pike, West Point, PA19486, USA
c) Department of Molecular Systems, 128 East Lincoln Avenue, Merck Research Laboratories, Merck & Co.,
Inc., Rahway, NJ 07065, USA
In support of a research program aimed at discovering orally bioavailable thrombin inhibitors, 13C- labeled
3-aminopyrazinone acetamide thrombin inhibitors have been prepared to aid in the structural elucidation of key
metabolites during preclinical and clinical studies. Amethod for the 13C-labeling of pyrazinones utilizing 13C-
glycine as a building block has been developed in a good overall yield. Two targets with labels at different
positions have been prepared. NMR Studies not only led to the structure of a major metabolite, but also
revealed a likely metabolic pathway, which will provide structural guidance for the design and synthesis of future
thrombin inhibitors as well as other pyrazinone-containing compounds.
Introduction. Thrombin is a trypsin-like serine protease that regulates platelet
activation and plays an important role in the blood-coagulation cascade process [1].
Thrombin can convert fibrinogen to fibrin that ultimately combines with platelets and
other components to form a blood clot [2]. Thrombosis, also known as excessive blood
clotting, is a major factor in cardiovascular disease, which is blamed for nearly half of all
deaths in the industrialized world annually [3]. Aberrant thrombosis is linked to
disseminated intravascular coagulation (DIC) [4], deep vein thrombosis (DVT) [5],
acute myocardial infarction (MI; heart attack) [6], unstable angina (serious chest pain
preceding MI) [7], and pulmonary embolism [8]. Therefore, thrombin and other
related coagulant agents such as factor Xa and prothrombinase (Ptase) have attracted
enormous attention in pharmaceutical research [9]. Our focus at Merck has been the
search for potent thrombin inhibitors with good oral bioavailability (OBA) and
inhibition selectivity relative to trypsin [10 13]. Most early thrombin inhibitors were
peptides or peptidomimetics with low OBAand poor pharmacokinetic (PK) profiles
[9]. Recently, compound 1 has been identified as an efficacious thrombin inhibitors
with excellent OBAand PK ( Figure) [13], and it has advanced into Phase-I clinical
trials as a once-daily oral anticoagulant drug candidate.
Absorption, distribution, metabolism, and excretion (ADME) studies of 1 resulted
in isolation of several major metabolites, many of which came from the oxidation of the
pyrazinone ring in 1. Compound 2 and 3 were the synthetic targets with 13C-label at the
indicated position located on the pyrazinone ring; it was expected that they would show
the same behavior as 1 during ADME studies. Metabolism and excretion of 1 varies
considerably across species, but compound 4 (also known as −metabolite number 1× or

Helvetica Chimica Acta Vol. 87 (2004)
675
Figure. A Non-Covalent Thrombin Inhibitor 1 and Its Metabolites 4
6
M1) is the predominant metabolite in all species studied (Figure), and its structure was
determined based on both NMR and mass-spectrometric data.
Two distinct metabolic activation pathways (a and b) were proposed (Scheme 1),
both of which included oxidation and subsequent rearrangement of the pyrazinone
heterocycle. As part of our effort to determine which pathway occurred during
metabolic process, 2 and 3 were needed because metabolism of either would generate
the corresponding 13C-labeled metabolite, which would reveal the metabolic mecha-
nism. These studies were important because identification of principal sites associated
with drug metabolic breakdown and understanding how it happened would help us
design future drug candidates containing a pyrazinone ring and to improve its duration
of action, i.e., PK profile.
Scheme 1. Two Plausible Biotransformation Pathways to the Major Metabolite 4

676
Helvetica Chimica Acta Vol. 87 (2004)
Results and Discussion. The retrosynthetic analysis is outlined in Scheme 2. 13C-
Labeled compound 2 could be prepared from 10 (nine steps) according to the method
previously developed by Burgey et al. [13]. The crucial pyrazinone ring would be
constructed from its precursor 11. Unlabeled compound 12 is a known one obtained by
ozonolysis of the corresponding olefin. Since construction of such an olefin with the
label at the desired position was not obvious, we envisioned that 12 would come from
readily available [1-13C]glycine. The key to this synthesis was to develop an efficient
procedure to obtain the aldehyde 12, as was shown in Scheme 3.
Scheme 2. Retrosynthetic Analysis of 13C-Labeled 2
Scheme 3. Synthesis of Amino[1-13C]acetaldehyde Dimethyl Acetal 15.
a) (Benzyloxy)carbonyl chloride (CbzCl), NaOH, H2O, toluene, r.t., 18 h; 77%. b) N,O-Dimethylhydrox-
ylamine ¥ HCl, N-[3-(dimethylamino)propyl]carbodiimide ¥ HCl (EDC), 1-hydroxy-1H-benzotriazole (BtOH),
Et3N, CHCl3; 97%. c) LiAlH4, THF, 08; 80%. d) HC(OMe)3, conc. H2SO4, MeOH; 95%. e) H2, Pd/C, i-PrOH.
Acylation of [1-13C]glycine with benzyl chloroformate in the presence of aqueous
NaOH gave acid 13 in good yield. Reduction of 13 to the corresponding alcohol was
readily accomplished but the subsequent oxidation to aldehyde 12 was found to be very
difficult. Several oxidation methods were tried but they always resulted in low yields.
Both the high cost of the labeled glycine and the long linear synthesis sequence (nine
steps from 12 to 2) prompted us to develop a better way to 12. Coupling 13 with N,O-
dimethylhydroxylamine ¥ HCl led to the Weinreb amide 14 in 97% yield, and LiAlH4
reduction of 14 in THF yielded aldehyde 12. It is known that LiAlH4 can also reduce
the (benzyloxy)carbonyl (Cbz) protecting group but the rate of reduction is slower than
that for the Weinreb amide. We felt that we could carry out the reduction step under
kinetic control. Using an undercharged LiAlH4 (90 mol-%) together with a short
reaction time (0.5 h) at 08 was the best condition found for this reaction to achieve a
consistently good yield (75 84%). Treatment of trimethyl orthoformate followed by

Helvetica Chimica Acta Vol. 87 (2004)
677
catalytic hydrogenation gave arise to 2-amino[1-13C]acetaldehyde dimethyl acetal (15)
in quantitative yield. At this point, application of the literature procedures [13] made it
possible for us to obtain the final tracer 2 (Scheme 4). Another 13C-labeled tracer 3 was
prepared in the same manner with [2-13C]glycine as the starting material.
Scheme 4. Synthesis of the 13C-Labeled Target 2
a) Glycine, Et3N, ClCH2CH2Cl, r.t., 18 h, 95%. b) 15, i-PrOH, r.t., 18 h, 98%. c) CF3COOH (TFA), AcOH,
reflux, 20 h. d) (COCl)2, DMF, MeCOO(i-Pr), MeCN. e) EtN(i-Pr)2 (DIEA), NaI, MeCN, reflux. f) NCS,
MeCN. g) NaOH, THF, H2O. h) EDC, BtOH, Et3N.
The biological assays were carried out with several important human serine
proteases including the digestive enzyme trypsin, the procoagulant factor Xa,
thrombin, and the thrombolytic protease plasmin. The labeled compounds 2 and 3
showed the same activity as that for unlabeled 1 [13]. For metabolic studies, 2 was
administered at 20 mg/kg to bile-duct-cannulated male SpragueÀDawley rats. The rat
bile and urine were collected at specific time intervals as required by the in vivo study
protocol. Each metabolite was isolated with a preparative HPLC column, and the
identity of the 13C-labeled M1 metabolite was confirmed by co-elution with authentic
sample of 4 on an analytical column [13b].
The results are highlighted in Scheme 5. Compound 2 has the 13C-label at position
13 whereas 3 with the label at position 12. The sole resultant M1 metabolite from 2
turned out to be 19, while 3 led to the isolation of 20 as the only M1 metabolite. The
label locations for 2, 3, 19, and 20 were confirmed by 13C-NMR spectroscopy, and this
metabolic transformation was accompanied by two significant changes in the 13C-NMR
spectra (recorded in CDCl3). In the first instance, the absorption attributable to the
13C-label of 2 at position 13 (118.6 ppm) was displaced by the signal from 19 at lower
field (171.5 ppm), clearly as a direct consequence of the label at the carboxylic acid C-
atom. In the second case, when the metabolic experiment was performed with 3 with
13C-label at position 12 (119.9 ppm), its metabolite 20 (M1) revealed that the label was
moved to higher field (101.2 ppm), which was assigned as a OH-bearing C-atom. Based
on these results and other evidence for biotransformation involving pyrazinone-ring
oxidation on similar systems [14], we proposed a mechanism (Scheme 6) for the
formation of metabolite M1 (compound 4).

678
Helvetica Chimica Acta Vol. 87 (2004)
Scheme 5. Metabolism of Labeled Thrombin Inhibitors 2 and 3
Scheme 6. Proposed Mechanism for the Formation of Metabolite M1

Helvetica Chimica Acta Vol. 87 (2004)
679
This mechanism excluded the pathway a and was consistent with pathway b
(Scheme 1), in which the pyrazinone-ring opening and contraction occurred at the
CÀN bond between the Cl-bearing C-atom and its neighboring N-atom in 1. In other
words, the metabolic pathway involved breaking the CÀN bond between position 13
and 14, which indicates that this bond is weaker than that between position 11 and 12
(Scheme 6).
AM1, a semi-empirical quantum-mechanical calculation [15], was performed to
understand the energetic difference between the two proposed intermediate structures
8 and 9. Atoms from position 9 to 15 of 1 in Scheme 6 are called the core, and they were
kept free during energy minimization, except that distance constraints of 2.5 ä were
applied between the atoms N(11) and C(13) in structure 8, and N(14) and C(12) in
structure 9. The energy minimizations were carried out until the geometries converged.
The results are very interesting. Not only the free energy of the non-core fragments in 8
and 9 were exactly the same, but they also were completely fixed during the energy-
minimization process. Therefore, the energy difference between the minimized
structures comes only from the different cores of 8 and 9. The final energies for 8
and 9 are À220.36 and À227.46 kcal/mol, respectively. In other words, the calculation
results indicated that path-b intermediate 9 is by ca. 7 kcal/mol more stable than path-a
intermediate 8. Since the experimental observations are clearly supportive for path b
over a, the mechanism of this pyrazinone-ring-opening reaction must have involved
late (product-like) transition states for the slow steps of the hydrolyses.
Conclusions.
We reported an efficient synthesis of the 13C-labeled amino-
acetaldehyde dimethyl acetal 15, which was utilized to construct a labeled thrombin
inhibitor for an ADME study to elucidate an important metabolic mechanism. A
mechanism for the formation of metabolite M1 (also known as 4) from 1 was proposed
based on the experimental results, which were further strengthened by AM1 molecular
calculations. This study can provide some guidance on structural design for drugs
candidates with a pyrazinone motif.
Experimental Part
General. All reactions were conducted under N2. Column chromatography (CC) was performed with 230
400-mesh silica gel. Unless otherwise noted, all NMR data were obtained at 400 MHz for 1H and 100 MHz for
13C, in CDCl3 with TMS as an internal standard.
N-[(Benzyloxy)carbonyl][1-13C]glycine (13) Benzyl chloroformate (15 ml, 100 mmol) was added to an ice-
cold soln. of [1-13C]glycine (5.0 g, 67 mmol) in NaOH (5.0m; 50 ml, 250 mmol) and toluene (50 ml). The mixture
was stirred vigorously at r.t. for 18 h. The toluene layer was separated, and the aq. soln. was cooled and acidified
with conc. HCl to pH 2. The precipitate was filtered off, washed with one volume of cold H2O, and dried. Pure 13
was obtained (13.2 g, 95%). 1H-NMR ((D6)DMSO): 3.68 (t, J ˆ 5.6, 2 H); 5.04 (s, 2 H); 7.36 (m, 5 H); 7.57 (br. s,
NH); 12.70 (br. s, OH). 13C-NMR ((D6)DMSO): 41.67; 65.50; 127.76; 127.85; 128.40; 137.06; 156.55; 171.61. MS:
‡
‡
211 ([M ‡ H] ), 234 ([M ‡ Na] ).
2-{[(Benzyloxy)carbonyl]amino}-N-methoxy-N-methyl[1-13C]acetamide (14). Amixture of 13 (5 g,
23.9 mmol), EDC (4.9 g, 25.6 mmol), HOBt (3.9 g, 25.5 mmol), and Et3N (13 ml, 9.4 g, 93.4 mmol) in CHCl3
(250 ml) was stirred at r.t., while N,O-dimethylhydroxylamine hydrochloride (2.5 g, 25.6 mmol) was added. The
resulting mixture was stirred at r.t. for 18 h under N2. The mixture was washed with H2O (250 ml), aq. NaHCO3
(250 ml), 1m HCl (250 ml), and brine (50 ml), and dried (Na2SO4). Evaporation of the solvent afforded a yellow
residue, which was purified by FC (AcOEt/hexanes 1 :1) to give 14 (5.87 g, 97%). White solid. 1H-NMR: 3.20 (s,

680
Helvetica Chimica Acta Vol. 87 (2004)
3 H); 3.71 (s, 3 H); 4.15 (d, J ˆ 4.0); 5.13 (s, 2 H); 5.61 (br. s, 1 H); 7.35 (m, 5 H). 13C-NMR (CDCl3): 32.27;
‡
40.03; 61.39; 66.79; 127.96; 128.01; 128.42; 136.36; 156.28; 169.67. MS: 254.1 ([M ‡ H] ).
2-{[(Benzyloxy)carbonyl]amino}[1-13C]acetaldehyde (12) LiA lH (1.0m in THF; 18 ml, 18.0 mmol) was
4
added to a mixture of 14 (5.04 g, 20.0 mmol) in THF (150 ml) at 08 under N2. The mixture was stirred at 08 for
1 h. Asoln. of KHSO (3.44 g, 25.3 mmol) in H2O (80 ml) was added, and THF was removed under reduced
4
pressure. The residue was extracted with CHCl3 (4 Â 75 ml). The combined org. layers were washed with 1m aq.
HCl soln., NaHCO3 (100 ml), and brine (100 ml), and dried (MgSO4). Evaporation of the solvent in a rotary
evaporator gave 12 (3.14 g, 80%). 1H-NMR (CDCl3): 4.04 (d, J ˆ 5.2, 2 H); 5.11 (s, 2 H); 5.65 (br. s, 1 H); 7.30
‡
(m, 5 H). 13C-NMR (CDCl3): 57; 62.14; 128.12; 128.24; 128.48; 136.33; 156.28; 196.46. MS: 195.1 ([M ‡ H] ).
2-Amino[1-13C]acetaldehyde Dimethyl Acetal (15). Amixture of 12 (1.3 g, 5 mmol), trimethyl orthofor-
mate (20 ml), MeOH (10 ml), and conc. H2SO4 (one drop) was stirred at r.t. for 2 d. K2CO3 (0.5 g) was added,
and, after stirring for 1 h, the solid was filtered off, and the filtrate was used without further purification. An
anal. sample was purified by FC (Et2O/pentane 1:1). 1H-NMR (CDCl3): 3.28 (dd, J ˆ 139.2, 5.4, 2 H); 3.35 (s,
3 H); 3.36 (s, 3 H); 4.36 (t, J ˆ 5.4, 1 H), 4.61 (br. s, 1 H); 5.08 (s, 2 H); 7.35 (m, 5 H). 13C-NMR (CDCl3): 44.09;
‡
680.7; 104.36; 129.44; 129.52; 129.98; 138.47; 157.81. MS: 255.1 ([M ‡ H] ).
Ethyl 2-({2-[(2,2-Dimethoxy[2-13C]ethyl)amino]-1,2-dioxoethyl}amino)acetate (16). Ethyl 2-[(2-ethoxy-
1,2-dioxoethyl)amino]acetate (1.1 g, 5.5 mmol) and Et3N (0.77 ml, 5.6 mmol) were added to an i-PrOH soln. of
15 (4.8 mmol). The resulting mixture was stirred at r.t. for 18 h, concentrated, and purified by FC (hexane/
AcOEt 85 :15) to yield 16 (2.1 g, 98%). Oil. The spectral data are in agreement with those reported in the
‡
literature [13a]. MS: 423.1 ([M ‡ H] ).
6-Chloro-3-{[2,2-difluoro-2-(pyridin-2-yl)ethyl]amino}-1-({N-[(3-fluoropyridin-2-yl)methyl]carbamoyl}-
methyl) [6-13C]pyrazin-2(1H)-one (2). [13C]-labeled-16 (2 g) was converted to 2 (123 mg, 6% overall yield for six
steps) using literature procedure [13a].
6-Chloro-3-{[2,2-difluoro-2-(pyridin-2-yl)ethyl]amino}-1-({N-[(3-fluoropyridin-2-yl)methyl]carbamoyl}-
methyl) [5-13C]pyrazin-2(1H)-one (3). 2-[13C]-glycine (5 g) was converted to 3 (114 mg) using the same method
as for preparation of 2.
J. Z. H. dedicates this paper to the memory of his postdoctoral advisor, Professor Henry Rapoport, Ph. D.,
Department of Chemistry, University of California, Berkeley, CA94720, USA. We thank Dr. Allen Jones and
Mr. Herb Jenkins for analytical support, and Dr. David Hands for providing synthetic intermediates.
REFERENCES
[1] F. V. Dosprders, −Thrombosis in Cardiovascular Disorders×; Ed. M. Verstraete, W. B. Saunders Co.,
Philadelphia, 1992.
[2] R. W. Coleman, V. J. Marder, E. W. Salzman, J. Hirsch, in − Hemostasis and Thrombosis, Basic Principles
and Clinical Practice×, 3rd edn., Eds. R. W. Colman, J. Hirsh, V. J. Marder, E. W. Salzman, J. B. Lippincott,
Philadelphia, 1994.
[3] J. P. Vacca, Curr. Opin. Chem. Biol. 2000, 4, 394.
[4] H. Terada, Jpn. J. Clin. Med. 1986, 44, 1170.
[5] B. O. Mustafa, S. W. Rathbun, T. L. Whitsett, G. E. Raskob, Arch. Int. Med. 2002, 162, 401.
[6] T. Ibbotson, K. L. Goa, Drugs 2002, 62, 1407.
[7] L. Oltrona, P. A. Merlini, A. Spinola, D. Ardissino, D. Granata, M. Lombardo, G. Zaini, P. Cioffi, Coronary
Artery Disease 1996, 7, 377.
[8] R. Hein-Rasmussen, C. D. Tuxen, N. Wiinberg, Thrombosis Res. 2000, 100, 287.
[9] a) J. M. Fevig, R. R. Wexler, in −Ann Rep. Med. Chem×, Ed. A. M. Doherty, Academic Press, 1999, 34,
p. 81 121; b) L. A. Robinson, E. M. K. Saiah, in −Ann Rep. Med. Chem×, Ed. A. M. Doherty, Academic
Press, 2002, 37, p. 85 143.
[10] T. J. Tucker, W. C. Lumma, S. D. Lewis, S. J. Gardell, B. J. Lucas Jr., E. P. Baskin, R. Woltmann, J. J. Lynch
Jr., E. A. Lyle, S. D. Appleby, I. W. Chen, K. B. Cancheck, J. P. Vacca, J. Med. Chem. 1997, 40, 1565.
[11] T. J. Tucker, S. F. Brandy, W. C. Lumma, S. D. Lewis, S. J. Gardell, A. M. Naylor-Olsen, Y. Yan, J. T. Sisko,
K. J. Stauffer, B. J. Lucas, J. J. Lynch Jr., J. J. Cook, M. T. Stranier, M. A. Holahan, E. A. Lyle, E. P. Baskin,
I. W. Chen, K. B. Dancheck, J. A. Krueger, C. M. Cooper, J. P. Vacca, J. Med. Chem. 1998, 41, 3210.
[12] E. J. Sanderson, T. A. Lyle, K. J. Cutrona, D. L. Dyer, B. D. Dorsey, C. M. McDonough, A. M. Naylor-
Olsen, E. W. Chen, Z. Chen, J. J. Cook, C. M. Cooper, S. J. Gardell, T. R. Hare, J. A. Krueger, S. D. Lewis,

Helvetica Chimica Acta Vol. 87 (2004)
681
J. H. Lin, B. J. Lucas Jr., E. A. Lyle, J. J. Lynch Jr., M. T. Stranieri, K. Vastag, Y. Yan, J. A. Shafer, J. P.
Vacca, J. Med. Chem. 1998, 40, 4466.
[13] a) C. S. Burgey, K. A. Robinson, T. A. Lyle, P. E. Sanderson, S. D. Lewis, B. J. Lucas, J. A. Krueger, R.
Singh, C. Miller-Stein, R. B. White, B. Wong, E. A. Lyle, P. D. Williams, C. A. Coburn, B. D. Dorsey, J. C.
Barrow, M. T. Stranier, M. A. Holahan, G. R. Sitko, J. J. Cook, D. R. McMasters, C. M. McDonough, W. M.
Sander, A. A. Wallance, C. C. Franklin, D. Bohn, Y. M. Leonard, T. J. Detwiler Jr., J. J. Lynch Jr., Y. Yan, Z.
Chen, L. Kuo, S. J. Gardell, J. A. Shafer, J. P. Vacca, J. Med. Chem. 2003, 46, 461; b) C. C. Lin, J. Z. Ho, S. M.
Pitzenberger, M. V. Silva-Elipe, C. R. Gibson, M. P. Braun, X. Yu, J. L. Yergey, R. Singh, Drug Metab.
Dispos. 2003, 31, 1437.
[14] R. Singh, M. V. Silva, P. G. Pearson, B. H. Arison, B. K. Wong, R. White, X. Yu, C. S. Burgey, J. H. Lin,
T. A. Baillie, Chem. Res. Toxicol. 2003, 16, 198.
[15] M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. P. Stewart, J. Am. Chem. Soc. 1985, 107, 3902.
Received September 22, 2003