Highly potent inhibitors of methionine aminopeptidase-2 based on a 1,2,4-triazole pharmacophore

Article abstract of DOI:10.1021/jm061182w

High-throughput screening for inhibitors of the human metalloprotease, methionine aminopeptidase-2 (MetAP2), identified a potent class of 3-anilino-5-benzylthio-1,2,4-triazole compounds. Efficient array and interative synthesis of triazoles led to rapid SAR development around the aniline, benzylthio, and triazole moeities. Evaluation of these analogs in a human MetAP2 enzyme assay led to the identification of several inhibitors with potencies in the 50-100 picomolar range. The deleterious effects on inhibitor potency by methylation of the anilino-triazole nitrogens, as well as the X-ray crystal structure of triazole 102 bound in the active site of MetAP2, confirm the key interactions between the triazole nitrogens, the active site cobalt atoms, and the His-231 side-chain. The structure has also provided a rationale for interpreting SAR within the triazole series. Key aniline (2-isopropylphenyl) and sulfur substituents (furanylmethyl) identified in the SAR studies led to the identification of potent inhibitors (103 and 104) of endothelial cell proliferation. Triazoles 103 and 104 also exhibited dose-dependent activity in an aortic ring tissue model of angiogenesis highlighting the potential utility of MetAP2 inhibitors as anticancer agents.

Full text of DOI:10.1021/jm061182w

J. Med. Chem. 2007, 50, 3777-3785
3777
Highly Potent Inhibitors of Methionine Aminopeptidase-2 Based on a 1,2,4-Triazole
Pharmacophore†
Joseph P. Marino, Jr.,* Paul W. Fisher,‡ Glenn A. Hofmann, Robert B. Kirkpatrick, Cheryl A. Janson,§ Randall K. Johnson,|
Chun Ma, Michael Mattern,# Thomas D. Meek, M. Dominic Ryan,X Christina Schulz, Ward W. Smith,O David G. Tew,
Thaddeus A. Tomazek, Jr., Daniel F. Veber,0 Wenfang C. Xiong,) Yuuichi Yamamoto, Keizo Yamashita, Guang Yang,[ and
Scott K. Thompson
Departments of Medicinal Chemistry, Enzymology, Oncology, and Structural Biology, GlaxoSmithkline, King of Prussia, PennsylVania 19406
ReceiVed October 11, 2006
High-throughput screening for inhibitors of the human metalloprotease, methionine aminopeptidase-2
(MetAP2), identified a potent class of 3-anilino-5-benzylthio-1,2,4-triazole compounds. Efficient array and
interative synthesis of triazoles led to rapid SAR development around the aniline, benzylthio, and triazole
moeities. Evaluation of these analogs in a human MetAP2 enzyme assay led to the identification of several
inhibitors with potencies in the 50-100 picomolar range. The deleterious effects on inhibitor potency by
methylation of the anilino-triazole nitrogens, as well as the X-ray crystal structure of triazole 102 bound in
the active site of MetAP2, confirm the key interactions between the triazole nitrogens, the active site cobalt
atoms, and the His-231 side-chain. The structure has also provided a rationale for interpreting SAR within
the triazole series. Key aniline (2-isopropylphenyl) and sulfur substituents (furanylmethyl) identified in the
SAR studies led to the identification of potent inhibitors (103 and 104) of endothelial cell proliferation.
Triazoles 103 and 104 also exhibited dose-dependent activity in an aortic ring tissue model of angiogenesis
highlighting the potential utility of MetAP2 inhibitors as anticancer agents.
Introduction
have reversible human MetAP2 inhibitors been reported, many
of which possess structures derived from peptides.5 There are
very few examples of metalloprotease or MetAP2 inhibitors that
rely on key binding interactions between a heterocycle and
active-site metal ions, along with amino acid residues.7 In this
paper, we disclose an anilino-1,2,4-triazole class of heterocyclic
metalloprotease inhibitors, detailing the array and iterative
synthesis of these compounds as well as key structural activity
relationships. X-ray crystallographic studies are reported that
reveal key binding interactions between the heterocycle and the
enzyme. Several compounds were found to be potent inhibitors
of endothelial cell proliferation and also showed promising
activity in an angiogenesis assay.
Angiogenesis is a key process in the progression of a number
of diseases such as diabetic retinopathy, rheumatoid arthritis,
and cancer.1 In the cancer area, it has been demonstrated that
tumors cannot metastasize without the formation of new blood
vessels.1 Biological transformations and pathways that play a
role in angiogenesis are, therefore, particularly attractive targets
as potential methods for inhibiting solid tumor growth/metasta-
sis. Methionine aminopeptidase type 2 (MetAP2a) has been
identified as the intracellular target for the potent antiangio-
genesis agent fumagillin (1; Figure 1), a natural product from
Aspergillus fumigatus.3 Early drug discovery efforts have
focused on analogs of fumagillin, which irreversibly inhibit
MetAP2 through covalent modification of an epoxide. Fumag-
illin analogs such as 2a (TNP-470; Figure 1) and 2b (PPI-2458;
Figure 1), potent selective inhibitors of MetAP2 proteolytic
activity and endothelial cell proliferation, have been in clinical
trials for the treatment of a variety of tumors.4 Only recently
Synthesis
High-throughput screening (GSK compound collection) of
the cobaltous form of methionine aminopeptidase-2 led to the
discovery of triazole inhibitor 6. In an effort to follow up on
this hit, we investigated two efficient synthetic routes for the
preparation of 3-anilino-5-benzylthio-1,2,4-triazoles. The first
strategy was developed by Kurzer and Douraghi-Zadeh8 and is
initiated by the S-benzylation of acetone thiosemicarbazone (3)
followed by treatment of the resulting acetone S-benzylisothi-
osemicarbazone with diphenylcarbodiimide (Scheme 1). Expo-
sure of the N,N′-diarylamidine product (5) to acid led to
cyclization, affording the target molecule 6 in good yield with
concomitant expulsion of acetone and aniline. Although this
route is short, it was clear that a diverse set of anilino-triazoles
could not be synthesized using this approach because the
commercial availability of substituted diphenylcarbodiimides
was limited.
† Coordinates and structure factor files for methionine aminopeptidase-2
complexed to 102 have been deposited in the Protein Data Bank under the
access code 2OAZ.
* To whom correspondence should be addressed. Phone: (610)-270-
5496. Fax: (610)-270-4490. E-mail: joseph.p.marino@gsk.com.
‡ Current address: Centocor, Inc., Radnor PA.
§ Current address: Shamrock Structures, Woodridge, IL.
| Current address: Independent consultant, Santa Fe, NM.
Current address: Hoffman-La Roche, Inc., Nutley, NJ.
# Current address: Progenra, Inc., Malvern, PA.
X Current address: Millennium Pharmaceuticals, Cambridge, MA.
O Current address: Argonne National Laboratory, Argonne, IL.
0 Current address: Independent consultant, Ambler, PA.
) Current address: Finnegan Henderson, Palo Alto, CA.
[ Current address: Tanabe Research Laboratories U.S.A., Inc., San
Diego, CA.
The second approach undertaken, first discovered by David-
son,9 was found to be more readily ammenable to rapid
analoging. Davidson found that nucleophilic attack on isobiuret
9 by hydrazine led to the diplacement of ethanethiol, followed
by cyclization of the amino group of the guanidine onto the
a Abbreviations: MetAP2, methionine aminopeptidase type 2; MetAP1,
methionine aminopeptidase type 1; HUVEC, human umbilical vein endot-
helial cell; MS-1, murine stroma cell type 1; XTT, cell growth inhibition
assay employing 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylami-
no)carbonyl]-2H-tetrazolium hydroxide sodium salt.
10.1021/jm061182w CCC: $37.00 © 2007 American Chemical Society
Published on Web 07/18/2007

3778 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16
Marino et al.
Figure 1. Irreversible inhibitors of methionine aminopeptidase-2.
Scheme 1. Kurzer and Douraghi Methoda
of potassium carbonate to afford the 1,2,4-triazole 6 in good
yield. Because a wide diversity of arylisothiocyanates and benzyl
bromides were commercially available, the chemistry was ideal
for two-dimensional parallel array synthesis.
To explore the SAR around the aniline group of the 1,2,4-
triazole 6, eight substituted phenyl-isothiocyanates were selected
that contained both electron-rich and electron-poor aromatic
substituents. The requisite eight mercapto-triazoles were ob-
tained via the Davidson synthesis, and the resulting thiols were
then treated with twelve alkyl halides in the presence of
potassium carbonate to provide the desired triazole targets.
Alkylation reactions of the eight mercaptotriazoles (employing
12 commercially available alkyl-halides) were carried out in a
96-well plate. Automated liquid handling instrumentation
facilitated filtration and liquid transfer. A total of 83 compounds
out of a possible 96 were successfully synthesized (highlighted
in Table 1). Yields ranged from 50% to 90%, with greater than
95% purity after reverse phase HPLC purification.
a Reagents and conditions: (a) BnCl, NaOH, EtOH, 60%; (b) diphenyl-
carbodiimide, acetone, reflux, 40%; (c) 10% HCl (EtOH), 70%.
To determine the importance of the heterocyclic and aniline
nitrogens for binding, a series of N-methylated triazoles were
synthesized, as shown in Scheme 3. Treatment of 6 with sodium
hydride led to deprotonation of the more acidic heterocyclic
protons. Addition of methyl iodide led to the formation of two
products in which alkylation had occurred on nitrogens 2 and
1 of the triazole ring to provide a 1:1 mixture of N-methyl-
triazoles 95 and 96, respectively (Scheme 3). Regiochemistry
was unambiguously established by heteronuclear correlations
and NOE (based on the position of the methyl group in relation
to the aniline proton). Following a protecting group strategy
commonly used for imidazole rings,11 the triazole ring was
readily alkylated with chloromethyl ethyl ether and protected
as the ethoxymethyl ether. The aniline nitrogen was methylated
by treatment with sodium hydride and iodomethane in DMF
and subsequently deprotected with trifluoroacetic acid to afford
the N-Me-aniline 98.12
Scheme 2. Davidson Method via Thioureaa
a Reagents and conditions: (a) phenylisothiocyanate, NaOH, H2O,
CH3CN 60%; (b) EtI, Et3N, DMF, 70%; (c) NH2NH2, EtOH, 80 °C, 60%;
(d) BnBr, K2CO3, DMF, 80%.
Biological Activity
In addition to the discovery that 1,2,4-triazole 6 was a potent
inhibitor (Ki,app ) 0.5 nM) of the Co2+-MetAP2 enzyme, it was
thiourea (Scheme 2). Aromatization of the triazole ring through
the loss of ammonia afforded the mercapto-triazole 11 in 70%
yield. The diaminotriazole product 10 was also formed as a
minor product (20%), however, 11 could be selectively pre-
cipitated by trituration of the crude reaction mixture with
aqueous HCl. Since an efficient synthesis of isobiuret 9 had
not been reported in the literature, a sequence was developed
which utilized common synthetic building blocks. The synthesis
of the key biuret intermediate 7 and the triazole 6 is summarized
in Scheme 2. In the first step, thiourea is reacted with
phenylisothiocyanate10 in aqueous sodium hydroxide to produce
1-phenyldithiobiuret 7. The biuret 7 was treated with ethyl iodide
in the presence of triethylamine to give a 2:1 mixture of a
S-alkylation products 8 (minor) and 9 (major), which were
separable by column chromatography. Heating the isobiuret 9
in ethanol in the presence of hydrazine hydrate afforded the
cyclization product 11. Mercapto-triazole 11 was then selectively
alkylated with one equivalent of benzyl bromide in the presence
also found to be quite selective versus Co2+-MetAP1 (Ki,app
)
3900 nM). Selectivity over MetAP1 is striking because both
enzymes share relatively similar active site topography.13
A
structural rationale for this selectivity will be discussed later.
Mechanistic enzymology studies determined MetAP2 binding
for this 1,2,4-triazole to be competitive.
In analyzing the MetAP2 inhibitory activities from the 12 ×
8 array, there are several SAR trends that were evident. First,
with respect to substitution on the aniline ring of lead 6, addition
of a methyl group to either the ortho- or the para-positions of
the aniline led to a 10-fold increase in potency, as evidenced
by analogs 13 and 23. Varying the electronic nature of the
aniline substituent produced modest effects on potency. Sub-
stitution of the para-position of the aniline with an electron
donating group such as methoxy (46) led to a 10-fold enhance-
ment in activity (Ki,app ) 50 pM) relative to lead 6. Addition of

Highly Potent Inhibitors of Methionine Aminopeptidase-2.
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16 3779
Table 1. SAR for 8 × 12 Triazole Array
a
a
MetAP2 Ki,app
(nM)
MetAP2 Ki,app
cmpd
R1
R2
cmpd
R1
R2
(nM)
6
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
Ph
Ph
Ph
0.5
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
4-OMe-Ph
4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
3,4-OMe-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
4-CO2Me-Ph
2-(Ph)-Ph
2-(Ph)-Ph
2-(Ph)-Ph
2-(Ph)-Ph
2-(Ph)-Ph
2-(Ph)-Ph
2-(Ph)-Ph
2-(Ph)-Ph
2-(Ph)-Ph
3-pyridyl
3-pyridyl
3-pyridyl
3-pyridyl
3-pyridyl
3-pyridyl
3-pyridyl
3-pyridyl
-CHdC(CH3)2
5-methyl-3-isoxazol-3-yl
Ph
2-Me-Ph
2-MeO-Ph
2-F-Ph
3,4-difluoro-Ph
2-pyridyl
4-pyridyl
cyclohexyl
-CHdC(CH3)2
2-methyl-thiazo-4-yl
5-methyl-3-isoxazol-3-yl
Ph
2-Me-Ph
2-MeO-Ph
2-F-Ph
3,4-difluoro-Ph
2-pyridyl
4-pyridyl
cyclohexyl
-CHdC(CH3)2
2-methyl-thiazo-4-yl
5-methyl-3-isoxazol-3-yl
Ph
2-Me-Ph
2-MeO-Ph
2-F-Ph
3,4-difluoro-Ph
4-pyridyl
0.16
1.6
1.7
thiophen-2-yl
Ph
2-Me-Ph
2-MeO-Ph
2-F-Ph
0.75
0.04
2.5
61
0.23
2.4
3.7
3.3
0.15
3.8
3.6
2-Me-Ph
2-Me-Ph
2-Me-Ph
2-Me-Ph
2-Me-Ph
2-Me-Ph
2-Me-Ph
2-Me-Ph
2-Me-Ph
2-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Me-Ph
4-Cl-Ph
18
>1000
1.3
4.0
24
3,4-difluoro-Ph
4-pyridyl
cyclohexyl
-CHdC(CH3)2
2-methyl-thiazo-4-yl
5-methyl-3-isoxazol-3-yl
Ph
2-Me-Ph
2-MeO-Ph
2-F-Ph
3,4-difluoro-Ph
2-pyridyl
9.6
8.3
0.9
21
7.0
0.7
135
10
0.41
1.3
0.9
0.9
2.7
0.5
3.8
0.8
0.07
6.5
>10 000
0.65
3.3
7.8
5.7
4-pyridyl
cyclohexyl
-CHdC(CH3)2
2-methyl-thiazo-4-yl
5-methyl-3-isoxazol-3-yl
thiophen-2-yl
Ph
2-Me-Ph
2-MeO-Ph
2-F-Ph
3,4-difluoro-Ph
2-pyridyl
100
1.2
7.8
3.5
5.2
0.43
740
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
23
>10 000
0.61
>2000
>2000
350
>2000
>2000
650
1.5
4.0
2.5
5.2
0.52
4.8
1.8
0.05
34
0.13
2.1
4-pyridyl
cyclohexyl
cyclohexyl
-CHdC(CH3)2
2-methyl-thiazo-4-yl
5-methyl-3-isoxazol-3-yl
Ph
2-MeO-Ph
2-F-Ph
3,4-difluoro-Ph
2-pyridyl
4-pyridyl
-CHdC(CH3)2
5-methyl-3-isoxazol-3-yl
Ph
2-Me-Ph
2-F-Ph
3,4-difluoro-Ph
2-pyridyl
4-pyridyl
cyclohexyl
2-methyl-thiazo-4-yl
5-methyl-3-isoxazol-3-yl
480
1700
4-Cl-Ph
4-Cl-Ph
0.04
1.7
0.3
0.5
5.7
3.0
6.5
9.1
1.5
4-OMe-Ph
4-OMe-Ph
4-OMe-Ph
4-OMe-Ph
4-OMe-Ph
4-OMe-Ph
4-OMe-Ph
18
6.1
3.1
cyclohexyl
3-pyridyl
a The statistical error for Ki,app values was in the range of (5-10% of the reported value from two independent experiments (separate study dates).
Scheme 3. Synthesis of N-methyl-triazolesa
to lead 6. In contrast to the ortho-methyl aniline analogs (13,
14, 16-22), the ortho-phenyl anilines (77-85) were found to
be much less active by two to three orders of magnitude in
relation to lead 6. Replacement of the aniline phenyl ring in 6
with a 3-pyridyl ring (86) resulted in a 10-fold enhancement in
potency.
With respect to substitution of the benzyl group of lead 6,
substitution at the ortho-position was not well tolerated. A 10-
fold decrease in activity was observed for 2-methyl-benzyl
analogs (24, 36, and 56), and at least a 100-fold decrease or
more was observed for 2-methoxy-benzyl analogs (15, 25, and
37). Substitution of the benzyl group of 6 with electron-
withdrawing groups such as 2-fluoro or 3,4-difluoro (16, 38,
and 89) did not lead to a significant change in potency (∼5-
fold or less) across aniline templates. Replacement of the benzyl
group with the smaller but equally rigid and lipophilic isoprenyl
group in most cases maintained activity (20, 43, 53, and 74)
compared to the lead 6. However, substitution with the more
conformationally flexible, lipophilic, cyclohexylmethyl group
generally led to a 5-10-fold decrease in activity across each
aniline series (19, 42, 73, and 92) relative to lead 6. Replacement
a Reagents and conditions: (a) NaH, DMF; MeI, rt, 81%; (b) NaH, DMF;
ClCH2OEt, rt, 66%; (c) NaH, DMF; MeI, rt, 52%; (d) trifluoroacetic acid,
rt, 82%.
a second methoxy group to 46, exemplified by 3,4-dimethoxy-
aniline 55, decreased potency slightly relative to the lead 6. para-
Substitution of the aniline ring with electron withdrawing groups
such as chloride (35) and ester (66) afforded activity comparable

3780 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16
Table 2. 5-Alkylthio Analogs
Marino et al.
a
MetAP2 Ki,app
cmpd
R
(nM)
Figure 2. Thiophene 102.
6
99
100
101
Bn
Me
0.5
1000
66
-CH2-CH2-CH(CH3)2
-CH2-CHdCH(CH3)2
1
a The statistical error for Ki,app values was in the range of (5-10% of
the reported value from two independent experiments (separate study dates).
Table 3. N-methylated Analogs
R2
a
MetAP2 Ki,app
Figure 3. Triazole 102 bound in the active site of MetAP2; distances
between inhibitor and active site residues are shown.
cmpd
R1
N1
Me
N2
Me
(nM)
95
96
98
H
H
Me
1300
>10 000
The remaining nitrogen of the triazole ring is in suitable
proximity (2.63 Å) to the histidine nitrogen of His-231 for
hydrogen bonding. This interaction was found to be critical
because mutation of His-231 to Ala in human MetAP2 (data
not shown) completely abolished inhibitory activity in this
chemical series. The S-thienylmethyl side-chain of 102 occupies
a hydrophobic pocket defined structurally by Phe-219, His-331,
and Tyr-444 as well as His-231. The enhanced potency observed
for compounds containing benzylic or allylic substituents on
the sulfur atom appears to be the result of favorable π-π
interactions between the inhibitor and these aromatic/heteroaro-
matic amino acid residues. The close proximity of His-231 and
His-331 to the thiophene ring also confirms why phenyl
substituents such as ortho-methoxy (compounds 25, 37, and 57)
are not well tolerated on the phenyl group of the S-benzyl side-
chain. The S-benzyl binding pocket may indeed be responsible
for the specificity observed for triazole inhibition of MetAP2
versus MetAP1, as this pocket is narrower in MetAP1 and may
only be able to accept a methionine side-chain as opposed to a
larger S-benzyl group.13 The ortho-isopropyl-aniline ring,
positioned to the right of Tyr-441 and flanked on its left by the
protein backbone, is oriented toward the pocket opening and,
thus, is largely exposed to solvent. It is not too surprising that
a variety of alkyl and polar groups may be accommodated at
the ortho-, meta-, and para-positions of the aniline in lead 6,
as evidenced by the excellent potency observed for compounds
13, 23, 35, 46, 55, and 66. However ortho-phenyl-aniline
analogs of lead 6, such as 77 and 84, are much weaker in activity
as compared to ortho-methyl-anilines 13 and 20, suggesting that
the ortho-phenyl substituent may interact unfavorably with the
nearby protein. This hypothesis is supported by the close
proximity of the isopropyl group of 102 to the backbone (Asn
329, Leu 328) of the protein (Figure 3).
15
a The statistical error for Ki,app values were in the range of (5-10% of
the reported value from two independent experiments (separate study dates).
of the phenyl ring in the sulfur side-chain of 6, with heterocyclic
rings such as isoxazole, pyridine, thiazole, and thiophene, led
to decreases in binding affinity of 5-15-fold, depending on the
aniline group with which it was combined. This is best
exemplified by analogs 18, 32, 44, 51, 65, and 93.
To firmly establish whether or not the benzyl group in 6 is
the optimal thio-ether side-chain, a series of simplified S-
alkylated analogs of 6 were synthesized (Table 2). Replacement
of the benzyl group with methyl (99) led to a 1000-fold decrease
in potency. Installation of a simple aliphatic chain such as
3-methyl-butyl (100; Ki,app ) 66 nM) regained a portion of the
activity back, however, the net result was still a 100-fold loss
of activity in relation to lead 6. Introduction of unsaturation
(isoprenyl; 101) restored activity relative to the lead 6. These
results reveal that a benzylic group or allylic olefins are key
functionalities required for tight inhibitor binding.
Substitution on the aniline or triazole nitrogens resulted in a
significant decrease in potency (Table 3). This is not surprising
given the presumed coordination of one or more of these
nitrogens to the active-site cobalt ions. Methylation of the aniline
nitrogen, exemplified by 98, resulted in only a 30-fold decrease
in activity relative to the initial screening lead 6, suggesting
that the aniline N-H is not critical for inhibitor binding.
However, substitution on N2 (95) gave rise to a 1000-fold
decrease in potency, while installation of a methyl group at N1
(96) of the heterocyclic ring abolished activity completely, thus
indicating that these two triazole nitrogen atoms play an
important role in binding.
The X-ray crystal structure of inhibitor 102 (Figure 2) bound
in the active site of human MetAP2 confirmed the critical role
the two contiguous triazole nitrogens play in coordinating to
the Co2+ ions, with close contacts (2.19, 2.16 Å) observed
between both cobalt atoms, N1, and N2 of the triazole ring.
Cellular and Tissue Activity
Angiogenesis, the process by which blood vessels form, is
known to be essential for the growth of solid tumors.2 Liu’s
laboratory has shown that the irreversible MetAP2 inhibitor,

Highly Potent Inhibitors of Methionine Aminopeptidase-2.
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16 3781
Table 4. Functional Activity of Triazole Analogs in Endothelial Cells
mMetAP2a hMetAP2b MS-1 HUVEC
c
c
cmpd
R1
H
R2
Ph
Ki,app
Ki,app
IC50
IC50
6
103
104
2a
2
5
20
0.5
3
24
1
3000
100
140
3000
30
40
Figure 4. Dose-dependent inhibition of angiogenesis by 104 in an
aortic ring assay.
i-Pr 2-furan
i-Pr 3-furan
0.2
a m ) murine. All data in nM. b h ) human. All data in nM. c The
statistical error was in the range of (5-10% of the reported value from at
least three independent growth inhibition experiments.
do not shed light on which metallo-enzyme activity more
accurately predicts activity in cells. The observation that
MetAP2 is catalytically active when complexed to a variety of
endogenous metal cofactors,2e,7c such as Co2+, Fe2+, Mn2+, Ni2+
,
Table 5. M2+-MetAP2 Enzyme vs HUVEC Activitya
and Zn2+, highlights the difficulty in determining the physi-
ologically relevant metalloprotease. Adding an additional ele-
ment of complexity, Hanzlik and co-workers6 have recently
shown that monometalated E. coli Mn2+-MetAP exhibits full
catalytic activity, a discovery that raises the possibility that
human MetAP2 could also be monometalated. In a cellular
environment, predicting which endogenous metal-enzyme spe-
cies is present may not be that straightforward and could be
dependent not only on cell type, but also surrounding tissue
and the concentration of different metal ions found within. The
implication is that there may be multiple metallo-MetAP2
species that are naturally occurring, and therefore, pinpointing
the most relevant one that affects angiogenesis or cell prolifera-
tion may be a challenge. However, the lack of a clear correlation
within the cellular results does also suggest other possibilities,
such as whether or not this particular class of inhibitors is truly
exhibiting their antiproliferative effects through MetAP2 (off-
target activity)? This is difficult to determine unequivocally
given the issues discussed above, but the observation that
analogs such as 103 and 104 exhibit good activity across
enzyme, cellular, and angiogenesis assays (data disclosed below)
is promising and merits continued focus on MetAP2 as the
putative target for this class of anilino-triazoles.
To further investigate the potential of these molecules as
antiangiogenic agents, triazole 104 was evaluated in an aortic
ring tissue model, an angiogenesis model that examines the
ability of a compound to inhibit the growth of blood vessels
sprouting from aortic tissue localized in culture. Confocal
microscopy was used to quantitate the vessel/sprout area;
fluorescent labeling of the sprout area is depicted in Figure 5.
The triazole 104 demonstrated dose-dependent inhibition of new
blood vessel growth, with near complete inhibition of angio-
genesis at 1.0 µM (Figure 4). Triazole 103 also exhibited similar
dose-dependent activity in the same model (data not shown).
These results establish the utility of anilino-1,2,4-triazole
MetAP2 inhibitors as potential therapeutic agents for inhibiting
the process of angiogenesis.
Co2+-hMetAP2b
Mn2+-hMetAP2b
HUVEC
IC50
c
cmpd
Ki,app
Ki,app
6
12
102
103
104
0.5
0.8
2
3
24
180
630
1500
1500
1800
1000
2000
150
30
40
a M ) metal. b hMetAP2 ) human MetAP2. All data in nM. c The
statistical error was in the range of (5-10% of the reported value from at
least three independent growth inhibition experiments.
2a, inhibits the growth or proliferation of endothelial cells, the
primary cell type involved in vascularization.2 The lead 1,2,4-
triazole compound, 6, was also found to be an inhibitor of
endothelial cell proliferation when studied in an XTT14 assay,
however, it unexpectedly exhibited a three to four order of
magnitude shift between human MetAP2 enzyme and endot-
helial cell (HUVEC) potency. A similar shift in potency was
observed between murine MetAP2 enzyme and murine stroma
cell (MS-1) activity. Interestingly, furan analogs of 6 (103 and
104) were found to be potent inhibitors of human endothelial
cell proliferation. Unlike lead 6, the cellular activity for triazoles
103 and 104 compared favorably (one order of magnitude shift)
with both the murine and the human Co2+-MetAP2 enzyme
inhibitory activity (Table 4). A possible explanation for the close
similarity between potencies obtained for enzyme and cell
growth inhibition in the furan series may be a favorable change
in physiochemical properties, thus enabling these analogs more
facile access (cell permeability) to the intracellular site of action.
Recent reports suggest the physiologically relevant human
MetAP2 metal to be manganese, which may explain the
discrepancy between enzyme and cellular potency observed for
our lead 6.2e,7c,7e In light of these reports, we also evaluated a
set of compounds against recombinant Mn2+-MetAP2 (Table
5). HUVEC cellular potency for analogs 6 and 12 was
discovered to be in line with their Mn2+ enzyme inhibitory
potencies (only 5-10-fold attenuated), however, greater than a
100-fold difference exists versus Co2+ enzyme inhibitory
potencies. By itself, this data would seem to indicate that Mn2+
is the relevant metal. However, the cellular potency for
thiophene 102 was found to be similar to enzyme potency for
both metalloenzymes (roughly one order of magnitude differ-
ence, although closer to the Mn2+-MetAP2 inhibitory potency),
bringing into question that the Mn2+-MetAP2 is the relevant
cellular form and suggesting that neither enzyme activity is
correlative with cellular activity. Interestingly, HUVEC activity
for furan analogs 103 and 104 was found to be more correlative
(e10-fold difference) with the Co2+-MetAP2 enzyme activity,
and conversely, the corresponding Mn2+-MetAP2 enzyme
activity was attenuated by 50-fold. It is clear that these results
Conclusion
Highly potent anilino-1,2,4-triazole MetAP2 inhibitors have
been discovered through high-throughput screening. An efficient
parallel array synthesis of triazoles led to rapid SAR develop-
ment. A variety of sulfur substitutions were combined with
different anilino-triazole templates to produce analogs with
subnanomolar human MetAP2 inhibitory activity. The deleteri-
ous effects on inhibitor potency by methylation of the anilino-
triazole nitrogens, as well as the X-ray crystal structure of
triazole 102 bound in the active site of MetAP2, confirm the
key interactions between the triazole nitrogens, the active site

3782 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16
Marino et al.
centroided data using a Micromass Q-Tof 2 hybrid quadrupole time-
of-flight mass spectrometer equipped with a Z-spray interface over
a mass range of 80-1200 Da, with a scan time of 0.95 s and an
interscan delay of 0.07 s (ionization was achieved with a spray
voltage of 3 kV, a cone voltage of 30 V, with cone and desolvation
gas flows of 5-10 and 500 L/hour, respectively).
General Procedure. Preparation of 3-Anilino-5-benzylthio-
1,2,4-triazole (6). (A) 1-Phenyl-2,4-dithiobiuret (7). To a stirring
solution of NaOH (0.52 g, 13.1 mmol) in 60 mL of 10% H2O in
CH3CN was added thiourea (1.0 g, 13.1 mmol). The resulting
suspension was heated to 40 °C for 20 min and then cooled to
room temperature. To this mixture was added phenylisothiocyanate
(1.5 mL, 13.1 mmol), and the clear yellow solution was stirred
overnight. After stirring for 12 h, 1 N HCl was added until a white
precipitate formed. The precipitate was filtered, washed with H2O,
and dried under vacuum to produce the title compound as a light
1
yellow powder (0.78 g, 30%). H NMR (400 MHz, DMSO-d6) δ
7.25 (t, 2H, J ) 7.3 Hz), 7.40 (t, 2H, J ) 7.9 Hz), 7.56 (d, 1H, J
) 7.9 Hz), 9.13-9.29 (br s, 1H), 10.26-10.79 (br s, 2H).
(B) 2-Ethyl-1-phenyl-2-isodithiobiuret (9). To a stirring solution
of compound 7 from step A (150 mg, 0.70 mmol) in 4 mL of DMF
was added triethylamine (57 µL, 0.70 mmol). The resulting yellow/
green solution was stirred for 10 min at room temperature. To this
solution was added ethyl iodide (100 µL, 0.70 mmol), and the
reaction mixture was stirred for 2 h at room temperature. The yellow
solution was poured into 20 mL of H2O and extracted four times
with EtOAc. The organic extracts were dried over Na2SO4, filtered,
and concentrated, and the crude residue was subjected to column
chromatography (silica gel; ethyl acetate/hexane) to afford the title
Figure 5. Inhibition of angiogenesis by 104 in the rat aortic ring model.
Fluorescent (calcein) labeling of aortic ring sprouts under confocal
microscopy is shown (0.1-1.0 µM concentrations).
cobalt ions, and His-231. The X-ray structure has also provided
a rationale for interpreting SAR within the triazole series. Well-
defined side-chain pockets that flank the cobalt ions in the
MetAP2 active site prevent extensive ortho-substitution of the
aromatic rings on the S-benzyl and aniline groups contained
within lead 6. Based on the crystal structure, enhanced potency
for analogs containing benzylic or allylic substituents on the
sulfur atom appears to be the result of favorable π-π interac-
tions between the inhibitor and the aromatic/heteroaromatic
amino acid residues, which define a hydrophobic pocket in the
MetAP2 active site. Triazole analogs 103 and 104, which
possess furanylmethyl side-chains, were discovered to be potent
inhibitors of human endothelial cell proliferation. Analogs 103
and 104 were also found to dose-dependently inhibit new blood
vessel growth in an aortic tissue model of angiogenesis.
Although a tight correlation between enzyme (single metallo-
MetAP2 species) and cellular activity could not be established,
the successful identification of new human MetAP2 inhibitors
that inhibit cell proliferation and angiogenesis does underscore
the value in considering multiple metalloenzyme species when
MetAP2 screening. This work demonstrates the utility of
triazoles as potent inhibitors of methionine aminopeptidase-2
and, in addition, validates the strategy of employing heterocycles
as key pharmacophores for drug design around metalloenzymes.
1
compound as a white solid (108 mg, 64%). H NMR (400 MHz,
DMSO-d6) δ 1.22 (t, 3H, J ) 7.2 Hz), 2.96 (q, 2H, J ) 7.2 Hz),
6.85 (d, 1H, J ) 7.6 Hz), 7.16 (t, 1H, J ) 7.2 Hz), 7.29-7.41 (m,
3H), 8.27 (br s, 1H), 9.89 (br s, 1H), and 10.99 (br s, 1H).
(C) 3-Anilino-5-mercapto-1,2,4-triazole (11). To a stirring
solution of 2-ethyl-1-phenyl-2-isodithiobiuret 9 (250 mg; 1.03
mmol) in 2.5 mL of EtOH and 0.25 mL of AcOH was added 150
µL of anhydrous hydrazine. The reaction mixture was heated at
80 °C for 1 h and then cooled to room temperature. To the reaction
mixture was added 5 mL of H2O, and excess concd HCl was added
dropwise until a precipitate formed. The precipitate was collected
and subsequently dried (under vacuum) to afford the title compound
as a white solid (90 mg, 45%). MS (ESI) 190.90 (M - H)+.
(D) 3-Anilino-5-benzylthio-1,2,4-triazole (6). To a stirring
solution of compound 11 from step C (23 mg, 0.12 mmol) in 1.2
mL of DMF was added K2CO3 (17 mg, 0.12 mmol), followed by
benzyl bromide (20 mg, 0.12 mmol). The mixture was stirred
overnight, filtered, and purified by preparative HPLC to afford the
title compound as a white solid (30 mg, 70%). 1H NMR (400 MHz,
DMSO-d6) δ 9.30 (br s, 1H), 7.47 (d, 2H, J ) 8.1 Hz), 7.39 (d,
2H, J ) 7.3 Hz), 7.31 (t, 2H, J ) 7.3 Hz), 7.23 (q, 3H, J ) 7.3
Hz), 6.82 (t, 1H, J ) 7.3 Hz), and 4.3 (s, 2H). Anal. (C15H14N4S)
C, H, N.
3-Anilino-5-benzylthio-2-methyl-1,2,4-triazole (95) and 3-Anili-
no-5-benzylthio-1-methyl-1,2,4-triazole (96). To a stirring solution
of 3-anilino-5-benzylthio-1,2,4-triazole 6 (0.20 g, 0.71 mmol) in
DMF (4 mL) was added NaH (37 mg, 0.92 mmol). To this mixture
was added iodomethane (49 µL, 0.78 mmol), and the solution was
stirred overnight. The reaction mixture was poured into H2O (25
mL) and extracted three times with EtOAc. The EtOAc extracts
were dried over Na2SO4, filtered, and concentrated. The crude
mixture was subjected to column chromatography (silica gel,
EtOAc/hexane) to provide the title compounds. Compound 95 was
isolated as a clear oil (93 mg, 44%). 1H NMR (400 MHz, DMSO-
d6) δ 9.20 (br s, 1H), 7.63 (d, 2H, J ) 7.6 Hz), 7.42-6.93 (m,
8H), 5.44 (s, 2H), 4.30 (s, 2H), 3.51 (q, 2H, J ) 7.1 Hz), 1.07 (t,
3H, J ) 7.0). LCMS (ESI) 341 (M + H)+. Anal. (C16H16N4S) C,
Experimental Section
Chemistry. All reagents and solvents were purchased from
Aldrich Chemical Co. (or other commercial sources) and used as
received. Kieselgel 60 silica gel was used for chromatography.
Preparative HPLC purification was performed with a YMC Com-
biPrep ODS-A column using a gradient of CH3CN/H2O as eluant
and UV detection at 214 nm. Analytical (Anal.) HPLC was
performed using two diverse HPLC systems: Method 1) 15-45%
CH3CN/H2O gradient over 6 min (8 min total run) using a 3.0 ×
50 mm XTERRA RP C18 column, and Method 2) 40-70% MeOH/
0.1 NH4OAc in H2O gradient over 6 min (8 min total run) using a
3.0 × 50 mm GEMINI C18 column. LCMS mass spectra were
taken on a PE Sciex API-150 instrument using electrospray
ionization (ESI). Elemental analysis was performed by Quantitative
1
H, N. Compound 96 was isolated as a light oil (77 mg, 37%). H
NMR (400 MHz, DMSO-d6) δ 9.33 (br s, 1H), 7.51 (d, 2H, J )
8.3 Hz), 7.42-7.22 (m, 8H), 5.23 (s, 2H), 4.47 (s, 2H), 3.43 (q,
2H, J ) 7.2 Hz), 1.04 (t, 3H, J ) 7.0 Hz). Anal. (C16H16N4S‚
0.2H2O) C, H, N.
1
Technologies, Inc. H NMR and 13C NMR were recorded at 400
and 100 MHz, respectively, on a Bruker NMR spectrometer.
Positive ion mass spectra (HRMS) were acquired as accurate mass

Highly Potent Inhibitors of Methionine Aminopeptidase-2.
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16 3783
3-(N-Methyl-anilino)-5-benzylthio-1,2,4-triazole (98). (A)
3-Anilino-5-benzylthio-1-(methyl-ethyl-ether)-1,2,4-triazole and
3-Anilino-5-benzylthio-2-(methyl-ethyl-ether)-1,2,4-triazole (97).
To a stirring solution of 3-anilino-5-benzylthio-1,2,4-triazole 6 (0.68
g, 2.41 mmol) in DMF (8 mL) was added NaH (0.12 g, 3.1 mmol).
To this mixture was added chloromethyl ethyl ether (0.25 g, 2.65
mmol), and the mixture was stirred overnight. The reaction mixture
was poured into H2O (50 mL) and extracted three times with EtOAc.
The EtOAc extracts were dried over Na2SO4, filtered, and concen-
trated. The crude mixture was subjected to column chromatography
(silica gel, EtOAc/hexane) to provide the title compounds as a
mixture (97) of regioisomers as a light yellow oil (0.58 g, 71%).
Isomer 1: 1H NMR (400 MHz, DMSO-d6) δ 9.33 (br s, 1H), 7.51
(d, 2H, J ) 8.3 Hz), 7.42-7.22 (m, 8H), 5.23 (s, 2H), 4.47 (s,
2H), 3.43 (q, 2H, J ) 7.2 Hz), 1.04 (t, 3H, J ) 7.0 Hz). LCMS
(ESI) 341 (M + H)+. Isomer 2: 1H NMR (400 MHz, DMSO-d6)
δ 9.20 (br s, 1H), 7.63 (d, 2H, J ) 7.6 Hz), 7.42-6.93 (m, 8H),
5.44 (s, 2H), 4.30 (s, 2H), 3.51 (q, 2H, J ) 7.1 Hz), 1.07 (t, 3H, J
) 7.0). LCMS (ESI) 341 (M + H)+.
(B) 3-(N-Methyl-anilino)-5-benzylthio-1,2,4-triazole (98). To
a stirring solution consisting of a mixture of 3-anilino-5-benzylthio-
1-(methyl-ethyl-ether)-1,2,4-triazole and 3-anilino-5-benzylthio-2-
(methyl-ethyl-ether)-1,2,4-triazole (97; 50 mg, 0.15 mmol) in THF
(1 mL) was added NaH (11.8 mg, 0.30 mmol), and to this solution
was added iodomethane (0.036 mL, 0.57 mmol). The reaction
mixture was stirred overnight. THF was removed (via rotary
evaporation under vacuum) and trifluoroacetic acid (0.5 mL) was
added to the crude product. The mixture was stirred overnight.
Trifluoroacetic acid was removed under vacuum, and the mixture
was purified by preparative HPLC to afford the title compound as
a clear oil (28 mg, 53%). 1H NMR (400 MHz, DMSO-d6) δ 3.3-
7.25 (m, 10H), 4.27 (s, 2H), 3.40 (s, 3H). LCMS (ESI) 297 (M +
H)+. HRMS calcd for C16H16N4S [M]+, 296.1095; found, 296.1094.
Anal. HPLC: method 1, 99.2%; method 2, 99.2%.
3-Anilino-5-methylthio-1,2,4-triazole (99). Compound 99 was
prepared according to the General Procedure except methyl iodide
was used in step D instead of benzyl bromide. LCMS (ESI) (M +
H)+. Anal. (C9H10N4S) C, H, N.
3-Anilino-5-(3-methyl-butylthio)-methylthio-1,2,4-triazole (100).
Compound 100 was prepared according to the General Procedure
except 1-bromo-3-methylbutane was used in step D instead of
benzyl bromide. LCMS (ESI) (M + H)+. Anal. (C13H18N4S) C, H,
N.
3-(Anilino)-5-(3-methyl-2-butenylthio)-1,2,4-triazole (101). Com-
pound 101 was prepared according to the General Procedure except
1-bromo-3-methyl-2-butene was used in step D instead of benzyl
bromide. LCMS (ESI) (M + H)+. Anal. [(C13H16N4S)‚0.1H2O] C,
H, N.
3-[2-(Isopropyl)-anilino]-5-(thiophen-2-ylmethylthio)-1,2,4-
triazole (102). Compound 102 was prepared according to the
General Procedure except 2-isopropylphenyl-isothiocyanate was
used in step A instead of phenylisothiocyanate, and 2-(chloro-
methyl)-thiophene was used in step D instead of benzyl bromide.
LCMS (ESI) (M + H)+. Anal. (C16H18N4S2) C, H, N.
of the artificial substrate L-methionine-7-amido-4-methylcoumarin
(Met-AMC)2e. Assays were performed in 96-well microtiter plates
using recombinant N-terminal truncated human methionine ami-
nopeptidase 2, with Co2+ as a cofactor. In a typical 96-well plate
assay, the increase in the fluorescence emission (λex ) 360 nm,
λem ) 460 nm) of a 50 µL assay solution in each well was
monitored continuously using a Cytofluor series 4000 multiwell
plate reader (PerSeptive Biosystems) and used to calculate the initial
velocity of hMetAP2. Each 50 µL assay solution contained 50 mM
Hepes‚Na+ (pH 7.5), 100 mM NaCl, 10 nM purified MetAP2
enzyme, and 0.4 mM Met-AMC (in 1% DMSO aqueous solution,
Km ) 0.3 mM with Co2+ as a cofactor). Assays were initiated with
the addition of substrate, and the initial rates were corrected for
the background rate determined in the absence of MetAP2. Ki,app
data was fitted to the Dixon competitive inhibition equation (eq 1)
using Grafit (Erithacus software), in which V0 represents the initial
velocity of the reaction, Km represents the Michaelis-Menton
constant of MetAP2, Vmax represents the maximum turnover rate
of MetAP2, I represents the concentration of inhibitors, and S
represents the substrate of MetAP2.
1/V0 ) Km × I/(Vmax × S × Ki,app) + (1/Vmax) × (1 + Km/S)
(1)
For use in the manganese activation assay, the cobalt metal was
removed from human MetAP2 by treatment with EDTA.15 MetAP2
activity was monitored by measuring the initial velocity of turnover
of the artificial substrate L-methionine-7-amido-4-trifluoromethyl-
coumarin (Met-AFC). Assays were performed in 96-well microtiter
plates using recombinant N-terminal truncated methionine ami-
nopeptidase 2 with 25 µM Mn2+ added as cofactor. In a typical
96-well plate assay, the increase in the fluorescence emission (λex
) 405 nm, λem ) 530 nm) of a 50 µL assay solution in each well
was monitored and used to calculate the initial velocity of MetAP2.
Each 50 µL assay solution contained 50 mM Hepes‚Na+ (pH 7.5),
100 mM NaCl, 23 nM purified MetAP2 enzyme, and 2 mM Met-
AFC (in 1% DMSO aqueous solution, Km ) 5 mM with Mn2+ as
a cofactor). Statistical error for Ki,app values were in the range of
(5-10% of the reported value from two independent experiments
(separate study dates).
Kinetic Analysis. Data was fitted to the appropriate rate
equations using Grafit computer software. Initial velocity data
conforming to Michaelis-Menton kinetics was fitted to eq 2.
Inhibition patterns conforming to apparent competitiVe and non-
competitiVe inhibition were fitted to eq 3 and eq 4, respectively.
V ) VA/(Ka + A)
(2)
(3)
(4)
V ) VA/[Ka(1 + I/Kis) + A]
V ) VA/[Ka(1 + I/Kis) + A(1 + I/Kii)]
In eqs 3 and 4, V is the initial velocity, V is the maximum
velocity, Ka is the apparent Michaelis constant, I is the inhibitor
concentration, and A is the concentration of variable substrates. The
nomenclature used in the rate equations for inhibition constants is
that of Cleland,17 in which Kis and Kii represent the apparent slope
and intercept inhibition constants, respectively.
HUVEC and MS-1 Cell Growth Inhibition Assay. The ability
of MetAP2 inhibitors to inhibit cell growth was assessed by the
standard XTT microtiter assay. XTT, a dye sensitive to the pH
change of mitochondria in eukaryotic cells, is used to quantify the
viability of cells in the presence of chemical compounds. Cells
seeded at a given number undergo approximately two divisions on
average in the 72 h of incubation. In the absence of any compound,
this population of cells is in exponential growth at the end of the
incubation period; the mitochondrial activity of these cells is
reflected in the spectrophotometric readout (A450). Viability of a
similar cell population in the presence of a given concentration of
compound is assessed by comparing the A450 reading from the
test well with that of the control well. Flat-bottomed 96-well plates
are seeded with appropriate numbers of cells (4-6 × 103 cells/
3-[2-(Isopropyl)-anilino]-5-(furan-2-ylmethylthio)-1,2,4-tria-
zole (103). Compound 103 was prepared according to the General
Procedure except 2-isopropylphenyl-isothiocyanate was used in step
A instead of phenylisothiocyanate, and 2-(chloromethyl)-furan was
used in step D instead of benzyl bromide. LCMS (ESI) (M + H)+.
HRMS calcd for C16H18N4OS [M]+, 315.1280; found, 315.1279.
Anal. HPLC: method 1, 99.4%; method 2, 99.3%.
3-[2-(Isopropyl)-anilino]-5-(furan-3-ylmethylthio)-1,2,4-tria-
zole (104). Compound 104 was prepared according to the General
Procedure except 2-isopropylphenyl-isothiocyanate was used in step
A instead of phenylisothiocyanate, and 3-(chloromethyl)-furan was
used in step D instead of benzyl bromide. LCMS (ESI) (M + H)+.
Anal. [(C16H18N4OS)‚0.1H2O] C, H, N.
Enzyme Inhibition Assay. Human methionine aminopeptidase
1 and 2, as well as murine methionine aminopeptidase 2 were
isolated as described previously2e,7f in the presence of Co2+. MetAP2
activity was monitored by measuring the initial velocity of turnover

3784 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16
Marino et al.
well in a volume of 200 µL) from trypsinized exponentially growing
cultures. The wells are coated with matrigel prior to establishing
the cultures. To “blank” wells is added growth medium only. Cells
are incubated overnight to permit attachment. On the next day,
medium from wells that contain cells is replaced with 180 µL of
fresh medium. Appropriate dilutions of test compounds are added
to the wells, with the final DMSO concentration in all wells being
0.2%. Cells plus compound are incubated for an additional 72 h at
37 °C under the normal growth conditions of the cell line used.
Cells are then assayed for viability using standard XTT/PMS
prepared immediately before use: 8 mg XTT (Sigma X-4251) per
plate is dissolved in 100 µL DMSO. H2O (3.9 mL) is added to
dissolve XTT and 20 µL of PMS stock solution (30 mg/mL) is
added from frozen aliquoted stock solution (10 mg of PMS
(phenazine methosulfate, Sigma P-9625) in 3.3 mL of PBS without
cations; these stocks are frozen at -20 °C until use). XTT/PMS
solution (50 µL) is added to each well and the plates are incubated
for 90 min (time required may vary according to cell line, etc.) at
37 °C until A450 is > 1.0. Absorbance at 450 nM is determined
using a 96-well UV plate reader. Percent viability of cells in each
well is calculated from these data (having been corrected for
background absorbance). IC50 is that concentration of compound
that reduces cell viability to 50% control (untreated) viability.
Statistical error for IC50 values was in the range of (5-10% of
the reported value from at least three independent growth inhibition
experiments.
Aortic Ring Assay. Animals: 7- to 10-week-old Male Sprague
Dawley rats at 7 to 10 weeks of age were purchased from Nippon
SLC. Reagents: Matrigel, Becton Dickinson Labware; Calcein-
AM, Dojindo Laboratory; endothelial cell basal medium (EBM),
Clonetics; antibiotic antimycotic solution (100×) stabilized, Sigma.
Equipment: Culture slide, Becton Dickinson Labware; surgical
microscope, Nikon; laser scanning microscope, Carl Zeiss.
Rings of rat thoracic aorta were embedded within Matrigel
overlaid with serum-free medium in the presence/absence of
inhibitors and were subsequently cultured for 6 days. Cells within
capillary-like sprouts were vitally labeled with Calsein-AM/
visualized under confocal microscope, thereby enabling quantifica-
tion of levels of angiogenesis by image analysis.
Preparation of Rat Aortic Ring Cultures: Rings of the rat
thoracic aortas were prepared as described previously.16 Thoracic
aortas from male Sprague-Dawley (SD) rats at 7-10 weeks of
age were dipped in phosphate-buffered saline (PBS) in a culture
dish. The periaortic fibroadipose tissue was removed under a
dissecting microscope, while care was given not to damage the
blood vessel wall. After rinsing with PBS to remove intramural
blood clot, rings were made by slicing the aorta with a fine blade
by approximately 1 mm long, and the rings were subsequently
rinsed three times with serum-free EBM medium. Eight-well
chamber slides were kept on ice and coated with 110 µL/well of
Matrigel at 4 °C. Matrigel was allowed to gel by incubation at 37 °C
for 30 min. Each aortic ring was placed on top of the gel and
embedded with an extra 40 µL of ice-cold Matrigel. After incubation
at 37 °C for 30 min, each well was overlaid with 500 µL of serum-
free medium (EBM) supplemented penicillin (100 µg/mL), strep-
tomycin (100 µg/mL), and amphotericin B (0.25 µg/mL). The rings
were cultured for 6 days in an incubator (5% CO2) at 37 °C with
medium change every 3 days.
Calcein Staining and Image Analysis: An aliquot of 1 µL of
5 mg/mL calcein in DMSO was mixed with 1 mL of serum-free
EBM medium. After brief removal of culture medium from a well
of culture slide, 200 µL of calcein/EBM solution was gently added
to a well of culture slide. After incubation at 37 °C for 1 h,
fluorescent-labeled sprouts were visualized under confocal micros-
copy. A projected image of aortic-ring sprouts was made from
images of ten optical slices (70 um/each) for each ring. Levels of
angiogenesis were represented by fluorescent areas (pixels) beyond
the outer rim of each aortic ring.
thresholds range was changed from 40 to 255 to segment the image
into objects and background. The sum of measurements for area
of the objects was calculated and saved to an Excel file for statistic
analysis.
Immunohistochemistry: The aortic ring with Matrigel was fixed
in 4% paraformaldehyde/PBS overnight at 4 °C. After incubation
in 30% sucrose/PBS overnight, the ring was embedded in O.C.T.
compound. The sample was frozen on dry ice and cryo-sectioned.
For immunohistochemistry, sections were washed three times with
0.3% Triton-X100/PBS and blocked with 3% normal serum /PBS.
Sections were incubated with primary antibody at 4 °C overnight.
After washing, sections were incubated with Alexa 546-conjugated
secondary antibody at 4 °C and washed with 0.3% Triton-X100/
PBS. The fluorescence image was detected under confocal micro-
scope.
Crystallography. The entire native sequence of the human
MetAP2 protein was employed and was purified as described by
Liu et al.13a Complexes were prepared by introducing solid inhibitor
into the crystal mother liquor after crystal formation and then
allowed to incubate at 4 °C for 24 to 48 h.
Crystals of human methionineaminopeptidase-2 complexed with
inhibitor grew to a size of approximately 0.1-0.2 mm3 in about 6
days at 4 °C. The concentration of inhibited human methionineami-
nopeptidase-2 used in the crystallization was approximately 10-
15 mg/mL. The method of vapor diffusion in sitting drops was
used to grow crystals from the solution of human methionineami-
nopeptidase-2. Crystals grew at 4 °C from drops containing protein
in a solution of 10% glycerol in 10 mM Hepes buffer at pH 7.4
containing 0.15 M NaCl. This solution was mixed in equal volumes
with a reservoir solution of 15-20% tert-butyl alcohol containing
0.4-0.1 M aqueous sodium citrate at pH 5.5. Crystals of the
complex are orthorhombic, space group C2221, with cell constants
of a ) 89.7, b ) 99.7, and c ) 100.6 Å. The crystals contain one
molecule in the asymmetric unit and approximately 54% solvent,
with a Vm value of 2.75 A3/Dalton. X-ray diffraction data were
measured from a single crystal for each inhibitor complex using
synchrotron radiation provided by beamline 17-ID at the Advance
Photon Source, Argonne National Laboratory. The protein structure
for the complex was determined by molecular replacement using
CNX (Molecular Simulations, Inc.). The starting model consisted
of all protein atoms of the published structure of human methion-
ineaminopeptidase-2 published by Liu et al.13a This model was
refined by rigid-body refinement, and the resulting phases were
used to calculate Fourier maps with coefficients |Fo - Fc| and |2Fo
- Fc| into which the atomic model of human methionineaminopep-
tidase-2 was built using the molecular graphics system XtalView
(Molecular Simulations, Inc.). Conventional positional refinement
was carried out during protein model building using CNX. The
density corresponding to the inhibitor was fit to difference Fourier
electron density maps with XtalView, and the resulting inhibitor
complex structure refined using CNX to a final Rc value of 0.22
for the inhibitor complex.
Acknowledgment. We thank Dr. Yie-Hwa Chang and St.
Louis University as the source of the intellectual property for
the type 2 human methionine aminopeptidase and Prof. Jon
Clardy and Dr. Cristina Nonato for guidance in crystallizing
human MetAP2.
Supporting Information Available: Crystal structure deter-
mination/refinement statistics for 102, data for key SAR compounds
12-94, and tabulated elemental analyses/HRMS/HPLC data for
all SAR compounds are given. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) (a) Gimbrone, M. A. J.; Leapman, S. B.; Cotran, R. S.; Folkman, J.
Tumor dormancy in vivo by neovascularization. J. Exp. Med. 1972,
136, 261-276. (b) Folkman, J. Tumor Angiogenesis. AdV. Cancer
Res. 1985, 43, 175-203. (c) Folkman, J. Angiogenesis in cancer,
Calcein-stained image was loaded in the Image-Pro Plus program,
an imaging software. The area outside the outer rim of each aortic
ring was defined using the Area Of Interest (AOI) tool. The

Highly Potent Inhibitors of Methionine Aminopeptidase-2.
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 16 3785
vascular, rheumatoid and other disease. Nat. Med. 1995, 1, 27-31.
(d) Hanahan, D.; Folkman, J. Patterns and emerging mechanisms of
the angiogenic switch during tumorigenesis. Cell 1996, 86, 353-
364. (e) Risau, W. Mechanisms of angiogenesis. Nature 1997, 386,
671-674.
3231-3243. (b) Auge, F.; Hornebeck, W.; Decarme, M.; Laronze,
J.-Y. Improved gelatinase A selectivity by novel zinc binding groups
containing galardin derivatives. Bioorg. Med. Chem. Lett. 2003, 13,
1783-1786. (c) Wang, J.; Sheppard, G. S.; Lou, P.; Kawai, M.; Park,
C.; Egan, D.; Schneider, A.; Bouska, J.; Lesniewski, R.; Henkin, J.
Physiologically relevant metal cofactor for methionine aminopepti-
dase-2 is manganese. Biochemistry 2003, 42 (17), 5035-5042. (d)
Oefner, C.; Douangamath, A.; D’Arcy, A.; Hafeli, S.; Mareque, D.;
Sweeney, A. M.; Padilla, J.; Pierau, S.; Schulz, H.; Thormann, M.;
Wadman, S.; Dale, G. E. The 1.15 ANG. Crystal structure of the
Staphylococcus aureus methionyl-aminopeptidase and complexes
with triazole-based inhibitors. J. Mol. Biol. 2003, 332, 13-21. (e)
Garrabrant, T.; Tuman, R. W.; Ludovici, D.; Tominovich, R.;
Simoneax, R. L.; Galemmo, R. A.; Johnson, D. L. Small molecule
inhibitors of methionine aminopeptidase type 2 (MetAP-2). Angio-
genesis 2004, 7 (2), 91-96. (f) Kallander, L. S.; Lu, Q.; Chen, W.;
Tomaszek, T.; Yang, G.; Tew, D.; Meek, T. D.; Hofmann, G. A.;
Schulz-Pritchard, C. K.; Smith, W. W.; Janson, C. A.; Ryan, M. D.;
Zhang, G.-F.; Johanson, K. O.; Kirkpatrick, R. B.; Ho, T. F.; Fisher,
P. W.; Mattern, M. R.; Johnson, R. K.; Hansbury, M. J.; Winker, J.
D.; Ward, K. W.; Veber, D. F.; Thompson, S. K. 4-Aryl-1,2,3-
triazole: A novel template for a reversible methionine aminopeptidase
2 inhibitor, optimized to inhibit angiogenesis in ViVo. J. Med. Chem.
2005, 48 (18), 5644-5647.
(2) (a) Wingfield, P.; Graber, P.; Turcatti, G.; Movva, N. R.; Pelletier,
M.; Craig, S.; Rose, K.; Miller, C. G. Purification and characterization
of a methionine-specific aminopeptidase from Salmonella typhimu-
rium. Eur. J. Biochem. 1989, 180, 23-32. (b) Bazan, J. F.; Weaver,
L. H.; Roderick, S. L.; Huber, R.; Matthews, B. W. Sequence and
structure comparison suggest that methionine aminopeptidase, pro-
lidase, aminopeptidase, and creatinase share a common fold. Proc.
Natl Acad. Sci. U.S.A. 1994, 91, 2473-2477. (c) Kobayashi, M.;
Shimizu, S. Cobalt proteins. Eur. J. Biochem. 1999, 261, 1-9. (d)
Lowther, W. T.; Matthews, B. W. Structure and function of the
methionine aminopeptidase. Biochim. Biophys. Acta 2000, 1477,
157-167. (e) Yang, G.; Kirkpatrick, R. B.; Ho, T.; Zhang, G.; Liang,
P.; Johanson, K. O.; Casper, D. J.; Doyle, M. L.; Marino, J. P., Jr.;
Thompson, S. K.; Wenfang, C.; Tew, D. G.; Meek, T. D. Steady-
state kinetic characterization of substrates and metal-ion specificities
of the full-length and N-terminally truncated recombinant human
methionine aminopeptidases (type 2). Biochemistry 2001, 40, 10645-
10654.
(3) (a) Griffith, E. C.; Su, Z.; Turk, B. E.; Chen, S.; Chang, Y. H.; Wu,
Z.; Biemann, K.; Liu, J. O. Methionine aminopeptidase (type 2) is
the common target for angiogenesis inhibitors AGM-1470 and
ovalicin. Chem. Biol. 1997, 4 (6), 461-471. (b) Sin, N.; Meng, L.;
Wang, M. Q.; Wen, J. J.; Bornmann, W. G.; Crews, C. M. The anti-
angiogenic agent fumagillin covalently binds and inhibits the
methionine aminopeptidase, MetAP-2. Proc. Natl. Acad. Sci. U.S.A.
1997, 1094 (12), 6099-6103. (c) Griffith, E. C.; Shuang, S.;
Niwayama, S.; Ramsay, C. A.; Chang, Y.; Liu, J. O. Molecular
recognition of angiogenesis inhibitors fumagillin and ovalicin by
methionine aminopeptidase 2. Proc. Natl. Acad. Sci. U.S.A. 1998,
95, 15183-15188. (d) Turk, B. E.; Griffith, E. C.; Wolf, S.; Biemann,
K.; Chang, Y.; Liu, J. O. Selective inhibition of amino-terminal
methionine processing by TNP-470 and ovalicin in endothelial cells.
Chem. Biol. 1999, 6 (11), 823-833. (e) Bernier, S. G.; Taghizadeh,
N.; Thompson, C. D.; Westlin, W. F.; Hannig, G. Methionine
aminopeptidase type I and type II are essential to control cell
proliferation. J. Cell. Biochem. 2005, 95 (6), 1191-1203.
(4) (a) Chang, S. Y.; McGary, E. C.; Chang, S. Methionine aminopep-
tidase gene Escherichia is essential for cell growth. J. Bacteriol. 1989,
171, 4071-4072. (b) Folkman, J. Clinical applications of research
on angiogenesis. New Engl. J. Med. 1995, 333, 1757-1763. (c) Li,
X.; Chang, Y. H. Amino-terminal protein processing in Saccharo-
myces cerevisiae is an essential function that requires two distinct
methionine aminopeptidases. Proc. Natl. Acad. Sci. U.S.A. 1995, 92,
12357-12361. (d) Hannig, G.; Lazarus, D. D.; Bernier, S. G.; Karp,
R. M.; Lorusso, J.; Qiu, D.; Labenski, M. T.; Wakefield, J. D.;
Thompson, C. D.; Westlin, W. F. Inhibition of melanoma tumor
growth by a pharmacological inhibitor of MetAP-2, PPI-2458. Intl.
J. of Oncology 2006, 28 (4), 955-963.
(5) (a) Keding, S. J; Dales, N. A.; Lim, S.; Beaulieu, D.; Rich, D. H.
Synthesis of (3R)-amino-(2S)-hydroxy amino acids for inhibition of
methionine aminopeptidase-1. Synth. Commun. 1998, 28 (23), 4463-
4470. (b) Sheppard, G. S.; Wang, J.; Kawai, M.; BaMaung, N.; Craig,
R. A.; Erickson, S. A.; Lynch, L.; Patel, J.; Yang, F.; Searle, X. B.;
Lou, P.; Park, C.; Kim, K. H.; Henkin, J.; Lesniewski, R. 3-Amino-
2-hydroxyamides and related compounds as inhibitors of methionine
aminopeptidase-2. Bioorg. Med. Chem. Lett. 2004, 14 (4), 865-868.
(c) Kawai, M.; BaMaung, S. D.; Fidanze, S. A.; Erickson, S. A.;
Tedrow, J. S.; Sanders, W. J.; Vasudevan, A.; Park, C.; Hutchins,
C.; Comess, K. M.; Kalvin, D.; Wang, J.; Zhang, Q.; Lou, P.; Tucker-
Garcia, L.; Bouska, J.; Bell, R. L.; Lesniewski, R.; Henkin, J.;
Sheppard, G. S. Development of sulfonamide compounds as potent
methionine aminopeptidase type II inhibitors with antiproliferative
properties. Biorg. Med. Chem. Lett. 2006, 16 (13), 3574-3577.
(6) Ye, Q-Z.; Xie, S-X.; Ma, Z-Q.; Huang, M.; Hanzlik, R. P. Structural
basis for catalysis by monometalated methionine aminopeptidase.
Proc. Natl. Acad. Sci. U.S.A. 2006, 103 (25), 9470-9475.
(8) Kurzer, F.; Douraghi-Zadeh, K. Heterocyclic compounds from urea
derivatives. VI. Synthesis and cyclization of 1-amino-3-(N,N′-
diarylamidino)guanidines and some analogs. J. Chem. Soc. C 1965,
932-937.
(9) Davidson, J. S. Action of hydrazine hydrate on isodithiobiurets. J.
Chem. Soc. C 1967, 2471-2472.
(10) Joshua, C. P.; Presannan, E.; Thomas, S. K. A one-step synthesis of
1,5-disubstituted 2,4-dithiobiurets and their oxidation to 3,5-di-
(substituted imino)-1,2,4-dithiazolidines. Indian J. Chem., Sect. B
1982, 21 (7), 649-651.
(11) (a) Whitten, J. P.; Matthews, D. P.; McCarthy, J. R. [2-(Trimethyl-
silyl)ethoxy]methyl (SEM) as a novel and effective imidazole and
fused aromatic imidazole protecting group. J. Org. Chem. 1986, 51
(10), 1891-1894. (b) Lipshutz, B. H.; Huff, B.; Hagen, W.
Metalations of imidazoles. (Poly)functionalization and conversions
to imidazolones. Tetrahedron Lett. 1988, 29 (28), 3411-3414.
(12) Bell, A. S.; Roberts, D. A.; Ruddock, K. S. A synthesis of 2- and
4(5)-(2-pyridinyl)imidazoles by palladium-catalyzed cross-coupling
reactions. Tetrahedron Lett. 1988, 29 (39), 5013-5016.
(13) (a) Liu, S.; Widom, J.; Kemp, C. W.; Crews, C. M.; Clardy, J.
Structure of human methionine aminopeptidase-2 complexed with
fumagillin. Science 1998, 282, 1324-1327. (b) Lowther, W. T.;
Orville, A. M.; Madden, D. T.; Lim, S.; Rich, D. H.; Matthews, B.
W. Escherichia coli methionine aminopeptidase: Implications of
crystallographic analyses of the native, mutant, and inhibited enzymes
for the mechanism of catalysis. Biochemistry 1999, 38, 7678-7688.
(14) Scudiero, D. A.; Shoemaker, R. H., Paull, K. D.; Monks, A.; Tierney,
S.; Nofziger, T. H.; Currens, M. J.; Seniff, D.; Boyd, M. R. Evaluation
of a soluble tetrazolium/formazan assay for cell growth and drug
sensitivity in culture using human and other tumor cell lines. Cancer
Res. 1988, 48 (17), 4827-4833.
(15) Walker, K. W.; Bradshaw, R. A. Yeast methionine aminopeptidase
1 can utilize either Zn2+ or Co2+ as a cofactor: A case of mistaken
identity? Protein Sci. 1998, 7, 2684-2687.
(16) (a) Gho, Y. S.; Kleinman, H. K.; Sosne, G. Angiogenic activity of
human soluble intercellular adhesion molecule-1. Cancer Res. 1999,
59 (20), 5128-5132. (b) Ponce, M. L.; Nomizu, M.; Delgado, M.
C.; Kuratomi, Y.; Hoffman, M. P.; Powell, S.; Yamada, Y.; Kleinman,
H. K.; Malinda, K. M. Identification of endothelial cell binding sites
on the laminin γ1 chain. Circ. Res. 1999, 84 (6), 688-694.
(17) (a) Cleland, W. W. The kinetics of enzyme-catalyzed reactions with
two or more substrates or products. I. Nomenclature and rate
equations. Biochim. Biophys. Acta 1963, 67, 104-137. (b) Cleland,
W. W. The kinetics of enzyme-catalyzed reactions with two or more
substrates or products. II. Inhibitionsnomenclature and theory.
Biochim. Biophys. Acta 1963, 67, 173-187. (c) Cleland, W. W. The
kinetics of enzyme-catalyzed reactions with two or more substrates
or products. III. Prediction of initial velocity and inhibition patterns
by inspection. Biochim. Biophys. Acta 1963, 67, 188-196.
(7) (a) Schroder, J.; Henke, A.; Wenzel, H.; Brandstetter, H.; Stammler,
H. G.; Stammler, A.; Pfeiffer, W. D.; Tschesche, H. Structure-based
design and synthesis of potent matrix metalloproteinase inhibitors
derived from 6H-1,3,4-thiadiazine scaffold. J. Med. Chem. 2001, 44,
JM061182W