3-Amino-2-hydroxyamides and related compounds as inhibitors of methionine aminopeptidase-2

Article abstract of DOI:10.1016/j.bmcl.2003.12.031

Substituted 3-amino-2-hydroxyamides and related hydroxyamides and acylhydrazines were identified as inhibitors of human methionine aminopeptidase-2 (MetAP2). Examination of substituents through parallel synthesis and iterative structure-based design allow

Full text of DOI:10.1016/j.bmcl.2003.12.031

Bioorganic & Medicinal Chemistry Letters 14 (2004) 865–868
3-Amino-2-hydroxyamides and related compounds as
inhibitors of methionine aminopeptidase-2
George S. Sheppard,* Jieyi Wang, Megumi Kawai, Nwe Y. BaMaung, Richard A. Craig,
Scott A. Erickson, Linda Lynch, Jyoti Patel, Fan Yang, Xenia B. Searle, Pingping Lou,
Chang Park, Ki H. Kim, Jack Henkin and Richard Lesniewski
Abbott Laboratories, Global Pharmaceutical Research and Development, Cancer Research,
100 Abbott Park Road, Abbott Park, IL 60064, USA
Received 9 October 2003; accepted 3 December 2003
Abstract—Substituted 3-amino-2-hydroxyamides and related hydroxyamides and acylhydrazines were identified as inhibitors of
human methionine aminopeptidase-2 (MetAP2). Examination of substituents through parallel synthesis and iterative structure-
based design allowed the identification of potent inhibitors with good selectivity against MetAP1. Diacylhydrazine 3t (A-357300)
was identified as an analogue displaying inhibition of methionine processing and cellular proliferation in human microvascular
endothelial cells (HMVEC).
# 2003 Elsevier Ltd. All rights reserved.
Methionine aminopeptidases are metalloenzymes
responsible for the removal of the N-terminal initiator
methionine residue of nascent proteins. Humans have
two forms of the enzyme, MetAP1 and MetAP2.1
MetAP2 was demonstrated to be a molecular target of
the antiangiogenic natural product fumagillin,2 and its
semisynthetic analogue TNP-470,3 which irreversibly
inactivate the enzyme by covalent modification of a
histidine residue in the active site.4 TNP-470 displays
inhibitory activity in animal models of angiogenesis and
tumor growth,5 supporting the potential of MetAP2
inhibition as an approach to the treatment of cancer
and other diseases with an angiogenic component.
the 2S,3R diastereomer was demonstrated to be the
most active (data not shown). A series of substituted
(2SR,3R)-3-amino-2-hydroxy-butanoic amides coupled
to l-alanine ethyl ester were prepared from available
d-amino acids using the route shown for 1a in Scheme
1, and evaluated for inhibition of human MetAP2 and
MetAP1 (Table 1) by a previously described method.11
As expected based on the published X-ray structure4
and our own crystallography, the active site was able to
accommodate sidechains larger than methionine, and
required extended side-chains for good activity. Larger
As part of a program to discover reversible inhibitors of
MetAP2, we examined several structural types reported
in the literature as inhibitors of aminopeptidases,
including 3-amino-2-hydroxyamides such as bestatin6
and amastatin.7 Contemporaneously with these investi-
gations, others have examined related approaches to
inhibition of MetAP1,8 as well as MetAP2.9 Analogues
such as 1a containing a methionine side chain at
the 3-position provided reversible MetAP2 inhibitors
(Table 1). As expected based on the bestatin analogy,10
Scheme 1. Reagents and conditions: (a) REDAL, toluene, 0–25 ꢀC; (b)
SO3 pyridine, Et3N, DMSO; (c) NaHSO3, KCN, EtOAc/H2O; (d) 6M
HCl, dioxane/H2O, reflux, 18 h; (e) BOC-ON, Et3N, dioxane/H2O; (f)
EDCI, HOBT, NMM, l-AlaOEt; (g) HCl/dioxane.
* Corresponding author. Fax.: +1-847-935-5165; e-mail: george.
sheppard@abbott.com
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.12.031

866
G. S. Sheppard et al. / Bioorg. Med. Chem. Lett. 14 (2004) 865–868
sidechains also showed a trend toward greater selectiv-
ity relative to MetAP1. Although we lack an X-ray
structure of human MetAP1, this observation is con-
sistent with a prediction of a smaller side chain binding
area based on modeling of MetAP1, which lacks a 60
amino acid insertion forming part of the active site in
MetAP2.
The initial series of analogues defined the approximate
size and shape requirements of the enzyme active site.
Investigation next turned to exploration of the amide
group by parallel synthesis. Based on initial results,
analogues of cyclohexylmethyl analogue 6a were tar-
geted for synthesis due to the potency and selectivity
observed, as well as the ready availability of quantities
of the intermediate BOC protected 3-amino-2-hydroxy
acid with good diastereomeric purity at C2 due to
selective crystallization of an intermediate. Parallel
synthesis was carried out in 48 well reactors by activ-
ation of the acid as the HOAT ester using HATU (O-(7-
azabenzotriazol-1-yl)-tetramethyluronium hexafluoro-
phosphate) and DIEA in DMA for 10 min at ambient
temperature, followed by amine addition and shaking at
ambient temperature for 60–75 h. Following workup
and solid phase extraction, the product amides were
deprotected with 4 N HCl in dioxane, and when neces-
sary purified by RPHPLC. Over 500 compounds were
Scheme 2. Reagents and conditions: (a) Na, i-PrBr, NH3; (b) BOC2O,
i-PrOH, H2O; (c) HN(OCH3)CH3 HCl, EDCI, HOBT, NMM,
CH2Cl2; (d) LiAlH4, E tO; (e) TMSCN, ClCH2CH2Cl, 90 ꢀC; (f) HCl,
2
dioxane, H2O, 100 ꢀC; (g) BOC2O, NMM, dioxane, H2O; (h) 3-chloro-
benzhydrazide, EDCI, HOBT, NMM, CH2Cl2; (i) Diastereomer
separation by HPLC (Silica, 20%acetone/hexanes); (j) HCl, dioxane.
N-terminal methionine and the relatively small active
site opening observed in X-ray crystal structures of the
enzyme, it was expected that large amide substituents
would be poorly accommodated. While this seemed to
hold for alkyl substituents (6e–i), aryl substituents were
found to provide potent inhibitors (6j–m). Interestingly,
a variety of linkers between the carbonyl and aryl moi-
eties were accepted, including heteroatom linkers such
as hydrazines and alkoxyamides (6n–p). Similar trends
were seen for over 200 analogues of 2a in a parallel
study (results not shown).
1
characterized by H NMR and MS and judged to have
HPLC purity >90% prior to biological testing.
A subset of the testing results is shown in Table 2. A
quick survey indicated both similarities and differences
between our SAR and that reported for inhibition of
aminopeptidase B by bestatin.10 As in the bestatin case,
the stereochemistry at C2 of alanine was not crucial
(compounds 6b and 6c). However, in this case, presence
of a carboxylate (6d) was detrimental to activity, in
contrast to the aminopeptidase B case where the car-
boxyl is reported to be important for inhibition. Based
on the reported substrate preference of MetAP2 for
small amino acids such as alanine adjacent to the
To further explore these findings, a matrix synthesis was
carried out combining side chains and carbonyl sub-
stituents (see Table 3 for monomer list). In this case a
set of BOC-protected 3-amino-2-hydroxy acids12 were
activated as their HOBT esters using DCC resin in a
mixture of DMA and dichloromethane for 5 min in
48 well blocks, followed by addition of the coupling
Table 2. Inhibition of human methionine aminopeptidases
Table 1. Inhibition of human methionine aminopeptidases
Compd
NHR2
MetAP2
IC50, mMa
MetAP1
IC50, mMa
Compd
R1
MetAP2
IC50, mMa
MetAP1
IC50, mMa
6a
6b
6c
6d
6e
6f
6g
6h
6i
6j
6k
6l
6m
6n
6o
6p
l-AlaOEt
d-AlaOEt
0.60
1.1
0.97
14
35
84
>100
>100
>100
NT
NT
NT
NT
NT
90
HNC(CH3)2CO2Et
l-AlaOH
1a
2a
3a
4a
5a
6a
7a
8a
CH2CH2SCH3
CH2CH2SCH2CH3
CH2CH2SCH(CH3)2
CH2CH2CH2CH3
C6H11
CH2C6H11
CH2CH2C6H5
CH2CH2C6H11
CH2SC(CH3)3
CH2SCH(CH3)2
CH2SCH2CH2CH3
CH2SCH2CH(CH3)2
CH2SCH2C6H5
CH2SCH2C6H11
5.7
1.3
1.1
8.0
12.5
0.6
0.46
0.59
1.6
1.8
1.4
1.2
51
63
69
NT
100
NT
84
101
137
14
13
57
l-LeuOMe
l-ValOMe
4.3
l-IleOMe
14
>100
16
HNCH(CH3)CH(CH3)2
HN(CH2)9CH3
HNPh-3-OCH3
HNCH2(2-Naphthyl)
HNCH(CH3)(2-Naphthyl)
HN(CH2)2Ph-2,4-diCl
HNNHC6H4-4-CH3
HNOC6H5
0.46
0.16
0.31
0.32
0.17
0.20
0.29
12
5.3
63
14
9a
10a
11a
12a
13a
14a
66
NT
>100
20
>100
HNNHCOC6H3-2,5-diCl
74
a Values are means of two-three experiments (NT=not tested).
a Values are means of two-three experiments (NT=not tested).

G. S. Sheppard et al. / Bioorg. Med. Chem. Lett. 14 (2004) 865–868
867
partners and agitation for 18 h. Following a quench
with trisamine resin followed by isocyanate resin the
mixtures were shaken for 4 h, filtered and washed with
dichloromethane, then concentrated and purified by
RPHPLC. Deprotection with 1:1 TFA/dichloromethane
provided the test compounds, which were characterized
1
by H NMR and MS and judged to have HPLC purity
>90% prior to biological testing (see Table 3.)
The trends identified earlier among the R1 substituents
generally were maintained, with groups approximating
or slightly exceeding the volume of the methionine
sidechain providing the best activity. Extended analo-
gues such as 3 and 7 displayed a trend toward greater
selectivity between MetAP2 and MetAP1. The aromatic
amides were generally superior in activity to the alanine
containing counterparts, but the effects of the link and
aromatic group on potency and selectivity showed
dependence on R1 in a manner that was not immedi-
ately obvious. Crystallographic investigation of a series
of enzyme–inhibitor complexes revealed that while the 3-
amino-2-hydroxy amide portion of the inhibitors bound
in a consistent manner across many structures, the por-
tion of the inhibitors projecting out of the active site
opening displayed considerable variation (see Fig. 1).13
In particular, it was noted that the imidazole sidechain of
His339 can adopt a number of conformations depending
on the structure of the inhibitor. Use of a matrix synth-
esis was instrumental in optimizing this series in view of
the variable binding modes that were observed.
Figure 1. Overlaid MetAP2 complexes of 2l (green), 3t (red) and 6n
(yellow) from X-ray crystal structures of enzyme–inhibitor complexes.
Residues forming the front of the active site were omitted for clarity.
The active site Mn2+ atoms are shown as purple spheres. His339 is
shown in lavender, in the conformation observed for the complex with
3t.
Table 3. Enzyme inhibition assay results for matrix compoundsa
R2=a
l
m
n
o
q
r
s
t
u
R1=2
3
5
6
1.3/69
1.1/NT
13/NT
0.6/84
0.46/>100
1.4/57
1.2/66
0.03/0.37
0.19/12
0.24/76
0.31/5.3
0.02/12
0.04/1.6
0.05/2.2
2.6/45
—
0.05/1.3
0.05/11
4.9/NT
0.18/11
0.80/16
0.43/5.0
0.06/6.8
2.4/NT
2.1/NT
0.05/1.3
0.05/4.7
0.83/95
0.20/20
0.31/13
0.05/1.7
0.03/1.5
3.6/NT
0.7/>100
0.02/12
0.13/81
1.6/NT
0.5/78
0.19/>100
1.0/NT
0.12/19
5.3/NT
4.1/NT
—
0.05/8.9
—
0.27/8.7
0.11/56
0.9/46
1.5/>100
2.7/NT
0.29/11
0.13/8.1
42/NT
0.21/15
0.25/18
51/NT
0.56/63
1.6/NT
—
0.07/NT
—
—
0.48/69
3.2/NT
0.32/63
1.1/NT
0.62/24
0.27/53
3.5/>100
1.3/NT
0.25/32
0.9/79
0.12/68
0.42/75
0.85/12
0.14/11
—
1.0/89
0.14/58
0.50/93
1.7/NT
0.12/34
8.6/NT
6.1/NT
7
11
12
14
15
74/>100
—
—
1.5/NT
1.2/NT
a Values reported are IC50 for MetAP2/MetAP1, and are means of two–three experiments (NT=not tested, —=compound not prepared).

868
G. S. Sheppard et al. / Bioorg. Med. Chem. Lett. 14 (2004) 865–868
Among the compounds in Table 3, 3t (A-357300) was of
particular interest, and a larger scale synthesis was
achieved by the route shown in Scheme 2.12b The X-ray
crystal structure of the enzyme–inhibitor complex of 3t
(Fig. 1) indicates the mode of binding. The 2-hydroxy-3-
aminoamide grouping interacts with the active site
Mn2+ metal ions,11 with the 2-hydroxy substituent
bridging between them. The thioether-containing side-
chain largely fills the adjacent hydrophobic site, while
the 3-chlorophenyl aromatic group lies face to face with
the histidine-339 imidazole. Both hydrophobic groups
interact near the site of an insertion1 absent in MetAP1.
6. Umezawa, H.; Aoyagi, T.; Suda, H.; Hamada, M.;
Takeuchi, T. J. Antibiot. 1976, 29, 97.
7. Aoyagi, T.; Tobe, H.; Kojima, F.; Hamada, M.; Takeu-
chi, T.; Umezawa, H. J. Antibiot. 1978, 31, 636.
8. Keding, S. J.; Rich, D. H. In Peptides for the New
Millennium; Fields, G. B., Tam, J. P., Barany, G., Eds.;
Kluwer Academic Publishers: Dordrecht/Boston/London,
2000; p 431.
9. Marino, J. P., Jr.; Ryan, M. D.; Thompson, S. K.; Veber,
D. F. PCT International Publication Number WO 01/
36404.
10. Nishizawa, R.; Saino, T.; Takita, T.; Suda, H.; Aoyagi,
T.; Umezawa, H. J. Med. Chem. 1977, 20, 510.
11. Wang, J.; Sheppard, G. S.; Lou, P.; Kawai, M.; Park, C.;
Egan, D. A.; Schneider, A.; Bouska, J.; Lesniewski, R.;
Henkin, J. Biochemistry 2003, 42, 5035. A coupled-
enzyme chromogenic assay was used to measure methio-
nine aminopeptidase activity by monitoring production of
free methionine with l-amino acid oxidase and horse-
radish peroxidase. Compounds in 10 mM DMSO stock
solutions were serially diluted with 50 mM Hepes buffer
pH 7.4 containing 150 mM NaCl and the solutions added
to an enzyme–substrate mixture containing MetAP2 or
MetAP1, MAS peptide, coupling enzymes and substrate.
A450 was recorded every 20 seconds for 20 min at room
temperature.
12. (a) Syntheses described in: Craig, R. A.; Henkin, J.;
Kawai, M.; Lynch, L. L.; Patel, J.; Sheppard, G. S.;
Wang, J. U.S. Patent 6,242,494, June 5, 2001. (b) Craig,
R. A.; Kawai, M.; Lynch, L. M.; Patel, J. R.; Sheppard,
G. S.; Wang, J.; Yang, F.; BaMaung, N. Y. PCT Inter-
national Publication Number WO 01/79157; U.S. Patent
Application US2002002152.
In addition to good potency and selectivity against the
isolated enzymes, compound A-357300 proved to be a
potent inhibitor of HMVEC (human microvascular
endothelial cell) proliferation, displaying an IC50 of 100
nM,14 and of cellular MetAP2 activity in HMVEC.15
Further biological characterization of this compound
and its examination in in vivo models of angiogenesis
and cancer has been recently described.15
In this study, the combined use of a structure-based
approach with high throughput parallel synthesis and
directed matrix combinatorial synthesis allowed optimiza-
tion of the initial lead structure. Improvements to the
potency, selectivity and other properties provided a rever-
sible inhibitor of MetAP2 suitable for evaluation of this
approach to inhibition of angiogenesis and tumor growth.
13. Refined crystallographic coordinates for the structures of
MetAP2 complexed with ligands 2l, 3t and 6n have been
deposited with the Protein Data Bank (www.rcsb.org)
with entry codes 1R58, 1R5G and 1R5H.
14. HMVEC proliferation was measured using MTS reagents
after a period of 3 days in complete growth medium
EGM2 containing a combination of growth factors
(VEGF, bFGF, EGF, IGF) and FBS: Wang, J.; Lou, P.;
Henkin, J. Biochemistry 2000, 77, 465.
References and notes
1. Arfin, S. M.; Kendall, R. L.; Hall, L.; Weaver, L. H.;
Stewart, A. E.; Matthews, B. W.; Bradshaw, R. A. Proc.
Natl. Acad. Sci. U.S.A. 1995, 92, 7714.
2. Sin, N.; Meng, L.; Wang, M. Q.; Wen, J. J.; Bornmann,
W. G.; Crews, C. M. Proc. Natl. Acad. Sci. U.S.A. 1997,
94, 6099.
3. Griffith, E. C.; Su, Z.; Turk, B. E.; Chen, S.; Chang, Y. H.;
Wu, Z.; Biemann, K.; Liu, J. O. Chem. Biol. 1997, 4, 461.
4. Liu, S.; Widom, J.; Kemp, C. W.; Crews, C. M.; Clardy,
J. Science 1998, 282, 1324.
5. Ingber, D. In Cancer Therapeutics; Teicher, B. A., Ed.;
Humana Press: Totowa, NJ, 1997; p 283.
15. Wang, J.; Sheppard, G. S.; Lou, P.; Kawai, M.;
BaMaung, N. Y.; Erickson, S. A.; Tucker-Garcia, L.;
Park, C.; Bouska, J.; Wang, Y.-C.; Frost, D.; Tapang, P.;
Albert, D.; Morgan, S.; Morowitz, M.; Shusterman, S.;
Maris, J. M.; Lesniewski, R.; Henkin, J. Cancer Research
2003, 63, 7861.