Discovery and Structure-Based Optimization of Next-Generation Reversible Methionine Aminopeptidase-2 (MetAP-2) Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.9b00041

Co- and post-translational processing are crucial maturation steps to generate functional proteins. MetAP-2 plays an important role in this process, and inhibition of its proteolytic activity has been shown to be important for angiogenesis and tumor growth, suggesting that small-molecule inhibitors of MetAP-2 may be promising options for the treatment of cancer. This work describes the discovery and structure-based hit optimization of a novel MetAP-2 inhibitory scaffold. Of critical importance, a cyclic tartronic diamide coordinates the MetAP-2 metal ion in the active site while additional side chains of the molecule were designed to occupy the lipophilic methionine side chain recognition pocket as well as the shallow cavity at the opening of the active site. The racemic screening hit from HTS campaign 11a was discovered with an enzymatic IC₅₀ of 150 nM. The resynthesized eutomer confirmed this activity and inhibited HUVEC proliferation with an IC₅₀ of 1.9 μM. Its structural analysis revealed a sophisticated interaction pattern of polar and lipophilic contacts that were used to improve cellular potency to an IC₅₀ of 15 nM. In parallel, the molecular properties were optimized on plasma exposure and antitumor efficacy which led to the identification of advanced lead 21.

Full text of DOI:10.1021/acs.jmedchem.9b00041

Subscriber access provided by IDAHO STATE UNIV
Article
Discovery and Structure-Based Optimization of Next Generation
Reversible Methionine Aminopeptidase-2 (MetAP-2) Inhibitors
Timo Heinrich, Jeyaprakashnarayanan Seenisamy, Beatrix Blume, Jörg Bomke,
Michel Calderini, Uwe Eckert, Manja Friese-Hamim, Rainer Kohl, Martin
Lehmann, Birgitta Leuthner, Djordje Musil, Felix Rohdich, and Frank T. Zenke
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00041 • Publication Date (Web): 02 Apr 2019
Downloaded from http://pubs.acs.org on April 3, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery and Structure-Based Optimization of Next Generation Reversible Methionine
Aminopeptidase-2 (MetAP-2) Inhibitors
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
1
1
1
Timo Heinrich,* Jeyaprakashnarayanan Seenisamy, Beatrix Blume, Jörg Bomke, Michel Calderini, Uwe
1
1
1
1
1
1
Eckert, Manja Friese-Hamim, Rainer Kohl, Martin Lehmann, Birgitta Leuthner, Djordje Musil, Felix
1
1
Rohdich, Frank T. Zenke
1
2
Merck Healthcare, Merck KGaA, Frankfurter Str. 250, 64293 Darmstadt, Germany; Syngene
International Ltd., Biocon Park, Plot 2&3, Bommasandra-Jigani Link Road, Bangalore 560 099, India
Abstract
Co- and post-translational processing are crucial maturation steps to generate functional proteins.
MetAP-2 plays an important role in this process and inhibition of its proteolytic activity has been shown
to be important for angiogenesis and tumor growth suggesting that small molecule inhibitors of MetAP-2
may be promising options for the treatment of cancer. This work describes the discovery and structure-
based hit optimization of a novel MetAP-2 inhibitory scaffold. Of critical importance, a cyclic tartronic
diamide coordinates the MetAP-2 metal ion in the active site while additional side chains of the molecule
were designed to occupy the liphophilic methionine side chain recognition pocket as well as the shallow
cavity at the opening of the active site. The racemic screening hit from a HTS campaign 11a was
discovered with an enzymatic IC50 of 150 nM. The re-synthesized eutomer confirmed this activity and
inhibited HUVEC proliferation with an IC50 of 1.9 µM. Its structural analysis revealed a sophisticated
interaction pattern of polar and lipophilic contacts that were used to improve cellular potency to an IC50
of 15 nM. In parallel, the molecular properties were optimized on plasma exposure and anti-tumor
efficacy which led to the identification of advanced lead 21.
Introduction
Methionine aminopeptidase-2 (MetAP-2) is responsible for cleaving N-terminal methionine residues of
proteins which is an important step of protein maturation during protein synthesis. Inhibition of MetAP-
2
effectively impairs angiogenesis and tumor growth making it an attractive target for anticancer
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
,2
therapy. Recently, analogs of the natural product MetAP-2 inhibitor fumagillin have been evaluated in
3
clinical trials in oncology and obesity. These fumagillin analogs are known to covalently bind MetAP-2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
and have to be administered subcutaneously or intravenously and, thus, have limited therapeutic utility.
One fumagillin analog, TNP-470, was extensively evaluated in clinical trials but was finally discontinued,
probably due to unfavorable pharmacokinetics as well as neurotoxicological side effects which were
5
considered to be drug-related. However, it is conceivable that non-covalent, non-fumagillin small
molecule inhibitors of MetAP-2 may have different tolerability in a clinical setting. The chemical diversity
of small molecule MetAP-2 inhibitors described until now is quite broad and X-ray crystallography has
helped to identify critical aspects how MetAP-2 inhibitors bind in the proteolytic cleft and to the
catalytically important metal ions. Figure 1 depicts the ligands for which the binding mode has been
identified by X-ray structure analyses. The center(s) reported to coordinate the metal ion(s) are marked
in red (distance <3 Å).
Figure 1: reversible MetAP-2 inhibitors. X-Ray structure analyses published. Metal cofactor binding
atoms marked in red; PDB codes added: 1 (SB-587094) 2OAZ; 2 (A-357300) 1R58; 3 (LAF-153) 1QZY; 4 -
JFR; 5 (A-849519) 1YW9; 6 - 5LYX.
5
6
This collection shows that one to three heteroatoms can bind to the metal ions in the active site of
MetAP-2, and the arrangement of either oxygen or nitrogen or both in one scaffold is accomplished with
ACS Paragon Plus Environment

Page 3 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
diverse chemical matter. For the triazole 1 (SB-587094) it was shown that both of the adjacent nitrogens
are necessary to achieve single digit nanomolar biochemical inhibition constant (K ) values with cobalt as
i
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
co-factor. Methylation of either nitrogen spoils affinity significantly. The 2-hydroxy-3-aminoamide
8
2+
6
grouping of the bestatin like 2 (A-357300) interacts with the active site Mn metal ions, and the 2-
hydroxy substituent is bridging between the ions. The thioether-containing sidechain largely fills the
adjacent hydrophobic site, while the 3-chlorophenyl aromatic group lies face to face with the His339
9
imidazole. An IC50 value of 110 nM in the enzyme assay was reported for this inhibitor. The t-butyl group
of bengamide analog 3 (LAF-153) extends into the innermost portion of the cleft, whereas the lactam
ring is coordinated in the shallow pocket at the surface of the enzyme. All three hydroxyl groups
1
0, 11, 12
coordinate to the two cobalt ions in the published structure.
In contrast to these very flexible open
chain MetAP-2 inhibitors compound 4 from Takeda is much more rigid as two heterocyclic moieties are
linked via just one rotatable bond. One nitrogen of the pyrazole interacts directly with one metal ion
while the other protonated nitrogen builds up a water mediated interaction with the other manganese
ion. The phenyl ring of the core is well situated to make an edge-to-face interaction with the aromatic
side chain of Tyr444. The 7-fluoro substituent protrudes into the lipophilic pocket occupied by the
thioether of the endogenous substrates, making the ligand very shape complementary with the target.
The aza-benzimidazole substituent at the N-4 position points into a hydrophobic space between Tyr444-
13
and His339-side chain. This compound was reported with a pIC50 of 8.3. The carboxylate 5 (A-849519)
was published with two digit nanomolar IC50 in enzyme inhibition assay. Only one oxygen of the acid
function is in direct contact to one of the manganese ions but the other heteroatom, like in case of the
14
pyrazole 1, is binding to a water molecule positioned over both metal ions. For the substituted triazolo-
triazine 6 as well as for analogous derivatized purines it was demonstrated that both southern nitrogens
15
are in direct contact to manganese ions, even though the assignment changes (180° flip). In addition to
these published MetAP-2 co-structures of reversible ligands, further literature known inhibitors were
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
synthesized and crystallized to extend the structural knowledge of reversible MetAP-2 inhibitors (Figure
2
).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2: reversible MetAP-2 inhibitors. Synthesis and structural analyses performed in house. PDB
codes added: 7 (JNJ-4929821) 6QEJ; 8 (GSK-2229238) 6QEI, 9 (X = Cl; Clioquinol) 6QEH.
16
As 1, 7 (JNJ-4929821) contains a 4H-[1,2,4]-triazole as metal coordinating heterocycle, but 8 (GSK-
17
2
229238) is the first carboxamide described as MetAP-2 inhibitor. The terminal triazole from 7
superimposes quite well with the di-substituted derivative 1, supporting the hypothesis that the adjacent
nitrogens are necessary for MetAP-2 binding. The carboxamide in 8 is substantially differently oriented
than the carboxylate of 5. Only the oxygen of the amide function is binding to one metal ion but the
nitrogen is not involved in any water mediated hydrogen bond network. It is in 3.35 Å distance to the
imidazole of His331. Clioquinol 9 and the nitro analog nitroxoline 10 (X = NO ; APL-1202) are known
2
1
8,19
antibiotics for the treatment of urinary tract infections and are described to inhibit MetAP-2 activity.
Nitroxoline 10 is reported with an IC50 value of 55 nM in the enzymatic MetAP-2 inhibition assay and
interferes in HUVEC (human umbilical vein endothelial cells) proliferation, but only with micromolar
potency (IC50: 1.9 µM). In the context of repurposing activities clinical investigations are on-going to
20,21
develop 10 for the treatment of invasive bladder cancer.
The hydroxy-quinoline coordinates both
metal ions with the oxygen binding to the more buried ion in 2 Å, and the quinoline nitrogen to the
second ion in 2.51 Å distance as found for 9. The chlorine is directed into the lipophilic substrate
recognition pocket and is in vdW contacts to Ala414 (4.32 Å), His231 (3.82 Å) and His382 (4.18). During
inhouse work with purine based MetAP-2 inhibitors it was not possible to optimize biochemical activity
ACS Paragon Plus Environment

Page 5 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of compounds based on purine-and related scaffolds like 6 to a low nanomolar level. More important,
the relatively weak enzymatic potency did not translate into favorable HUVEC proliferation inhibition in a
consistent manner, considered to be required to elicit significant in vivo efficacy (tumor growth
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
5
inhibition). Accordingly, focus was switched to another HTS hit series comprised during a screening
campaign not taken into account during first result evaluations. This hit matter was re-considered
because of missing compelling alternatives. Insufficient purity was the reason for initial de-prioritization,
even though sub-micromolar activity was measured (IC50 of 150 nM for the most potent derivative).
Accordingly, a more detailed analysis of the series was conducted, and it became obvious that these hits
relinquished a consistent pattern responsible for their failure in post-HTS control. In all cases, the
detected molecular mass was 16 units larger than the mass for the registered structure. Oxidation was
responsible for this deviation between expected and observed molecular mass, and NMR analysis
revealed the position of the oxidized center (Figure 3).
Figure 3: The stored α-amidated pyrrolidinone 11 was oxidized to the cyclic tartronic di-amide 11a
under storage in DMSO over approximately 2 years.
Cα of the pyrrolidinone in 11 is substituted with an amide residue, activating this carbon for oxidation.
The screening sample had been stored for 2 years as DMSO stock solution in a sealed tube at -20°C.
Obviously, these conditions were not sufficient to prevent quantitative oxidation. Registration analytics
of the compound confirmed the envisioned original structure of 11, and a newly synthesized batch did
not show any activity in the biochemical assay. The dedicated synthesized compound 11a delivered an
IC50 of 57 nM confirming the initial screening result. Attempts to mimic the oxidation of 11 to 11a during
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
storage by stirring the reduced precursor 11 in DMSO at higher temperature for a couple of days were
not successful (no oxidation could be detected). The compound had been included in many HTS initiatives
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
shortly before the MetAP-2 screen and afterwards, but never showed any activity. This suggested a very high
overall target selectivity. Thus, persistence and serendipity helped to identify the starting point for the
herein outlined optimization.
The structural analysis of the hit 11a challenges virtually every atom of the ligand. Our systematic
optimization was graded by pragmatic considerations and scaffold dissections (Figure 4).
Figure 4: Screening hit rac-11a with indications of different parts of the molecule. The first
optimization cycle focused on the amide residue highlighted by a blue circle; second improvement
round considered the N-lactam residue marked with a red circle. The investigation of lactam size and
additional substitutions (brown circle) as well as the heteroatoms (indicated with arrows) are not
discussed here.
Chemistry
All compounds described herein belong to the same pyrrolidinone class but were prepared via
alternative approaches, depending on the optimization purpose. The preparation of optimized amides
followed the procedure out-lined in Scheme 1.
a
Scheme 1: Synthetic approach to optimize amide residue R . Derivatives from tables 1 and 2.
1
ACS Paragon Plus Environment

Page 7 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Reagents and conditions: a) ACN, aniline, 60°C, 12h, 85%; b) EtOH, SOCl
2
, 0°C – reflux, 4h, N
2
; 92% c)
CeCl
1
3
-(H
2 7
O) , iPrOH, O
2
, 12h, RT; Chiralpak AD-H; d) LiOH, THF/H
2
O, 4h, RT; 84% e) R
1
-NH , EDCI, HOBt,
2
60°C, MW, 20 min 40 – 80%.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
This approach started with the reaction of cyclopropyl Meldrum’s acid 12 with aniline to give N-Phenyl-
pyrrolidinone-3-carboxylic acid 13 which was then transferred into the corresponding ethyl ester 14 with
thionyl chloride in ethanol. C3 oxidation of the pyrrolidinone-3-carboxylate 14 was achieved with cerium
chloride in isopropanol. The functionalized pyrrolidinone was separated into the enantiomers S-15 and
R-15 via chiral chromatography and the desired enantiomer R-15 was saponified with lithium hydroxide
in a THF/water mixture of 3:1 to the corresponding acid 16. The hydroxy-acid 16 was coupled under
standard conditions with amines to the amides 11a-11r.
The second optimization round focused on the lactam nitrogen substituent. In this context an alternative
synthesis was established. In the initial approach aniline was reacted with Meldrum’s acid, introducing
the lactam substituent in the first reaction. However, this procedure was considered to be unpractical for
an optimization of the lactam nitrogen residue which should be added in the last step of the synthesis.
Accordingly, the sequence outlined in Scheme 2 using the identified 3-chloro-5-fluoro-benzylamine as
suitable residue R was established.
1
a
Scheme 2: Synthetic approach to optimize lactam residue R . Derivatives from table 3
2
a
Reagents and Conditions: a) BOC
78°C – RT, 42%; c) Pd/charcoal, methanol, H
DCM, 5h, 0°C – RT, 86%; e) TFA, DCM, 3h, 0°C, 90%; f) MMPP, DMF, 4h, 70°C, 37%; g) R
2
O, DMAP, ACN, 3h, RT, 70%; b) benzyl chloroformate, LiHMDS, THF, 2h,
, 2 bar, 2h RT, 86%; d) 3-F, 5-Cl benzylamine, T3P, TEA,
-halide, K CO
-
2
2
2
3
,
CuI, N,N’-diemethyl-ethane-1,2-diamine, dioxane, 2h, 140°C, MW; h) Chrialpak AD-H.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Commercially available pyrrolidinone 17 was first BOC protected. Different options were worked out to
introduce the carboxylic moiety in position 3. Reproducible results were obtained by deprotonation with
lithium bis(trimethylsilyl)amide (LiHMDS) and quenching with benzyl chloroformate. Hydrogenation of
the benzyl ester, subsequent amidation of the acid with 3-fluoro, 5-chloro benzylamine in the presence
of 2,4,6-Tripropyl-[1,3,5,2,4,6]trioxatriphosphinane 2,4,6-trioxide (T3P) and deprotection of the lactam
nitrogen proved to be necessary to oxidize the chiral carbon with magnesium monoperoxyphthalate
hexahydrate (MMPP) to give the key intermediate 18. Established Buchwald conditions were used to
arylate the lactam nitrogen. The racemic products from this approach were separated into the
enantiomers via chiral chromatography giving the final products 11ra-11rj, 19 and 20.
Results and Discussion
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
For the determination of potency of MetAP2 inhibitors an enzyme-coupled assay to measure the
proteolysis of a tripeptide substrate (Met-Ala-Ser) was established. Recombinant purified His-TEV-
6
MetAP-2 (AA2-478) with manganese as a co-factor was used. Proliferation of HUVEC primary
endothelial cells was applied as mechanistical assay to assess compound potency in the cellular
22
environment.
Crystallization of the human MetAP-2 protein fragment, comprising amino acid residues 108 – 478 had
15
been established during earlier work so this tool could be used to understand the binding mode of 11a
(Figure 5 & Figure 6:).
A
B
ACS Paragon Plus Environment

Page 9 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 5: X-ray structure analysis of 11a in MetAP-2; A: 2D plot to show relevant polar interactions,
Imidazole
indicated with dashed lines, oxygen-manganese contacts in purple, heteroatom-His
in green,
distances are given in Ångstrom; B: Conformation of 11a in the active site; interacting imidazoles of
His331 and His339 and the manganese ions are pointed out (PDB: 6QEG).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The schematic 2D plot of Figure 5A depicts the polar interactions of 11a with the protein. The oxygen at
Cα of the tartonic di-amide coordinates the two manganese ions in the active site of MetAP-2. The
distance to the first, less buried ion is 2.36 Å and to the second ion, which is deeper located in the
pocket, 2.08 Å. The lactam oxygen is also in close contact to the first metal ion with 2.68 Å, and the
amide oxygen binds to the second ion with a distance of 2.32 Å. The three oxygens of the α-hydroxy-
23
malonic di-amide form a unique metal binding element that is unprecedented for MetAP-2 inhibitors.
The binding of compound 11a to MetAP-2 is characterized by additional polar interactions. In addition to
its metal coordination, the lactam oxygen interacts with the protons from His339-imidazole (2.70 Å) and
His331-imidazole (3.31 Å). The amide NH is in H-bond distance to His231 side chain (3 Å). The high
electron density of this imidazole is documented by its nucleophilic attack on the epoxide of covalently
24
binding fumagillin type natural products and derivatives. Figure 5B shows the active site of MetAP-2
with 11a. The lactam ring is the planar template of this MetAP-2 inhibitor class from which the residues
are positioned for relevant protein interaction. The N-phenyl ring of compound 11a covers the entrance
to the active site and is rotated by 62° out of the lactam plane. The hydroxy function is pointing
downwards and is coordinated to both manganese ions in the active site. The amide residue is pointing
up-wards and is allowing the thiophene-ethyl residue to orient itself in the substrate recognition area of
MetAP-2.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A
B
Figure 6: X-ray structure analysis of 11a in MetAP-2: A: 2D plot to show relevant lipophilic interactions,
indicated with dashed lines, distances are given in Ångstrom; B: identical view on the conformation of
1a in the active site as in figure 3 B; the lipophilic trait of the protein surface is color coded. Brown
1
indicates a lipophilic surface, green a polar one (PDB: 6QEG).
Figure 6 A illustrates the van der Waals (vdW) distances of the ligand 11a to the amino acid side chains in
the active site. The thiophene is located in the nonpolar area that accommodates the methionine side
chain upon hydrolysis of natural substrates. Not only the side chains of Ala414 and Ile338 are in vdW
contacts with the thiophen but also the imidazole of His231. In addition, this imidazole has also lipophilic
interactions with the ethyl bridge between the amide nitrogen and the thiophene. These observations of
polar and lipophilic interactions of the imidazole heterocycle of His231 demonstrate that this is the most
crucial amino acid for inhibitor recognition. The lactam N-phenyl ring of 11a faces an amphiphilic
surrounding. The properties of the bottom part of the active site (as oriented in Figure 6B) are driven by
the backbone -oxygen and -nitrogen of Asn329. These heteroatoms are responsible for constituting a
polar region whereas the protein surface in the upper-right region is liphophilic due to the domination of
the methyl and iso-butyl side chains of Ala230, Leu447 and Leu328 (not explicitly indicated in Figure 6B).
The protein surface around the unsubstituted carbons C4 and C5 of the lactam ring of 11a is amphiphilic,
too. Carbon C5 is in vdW contacts to the side chains of Ala230 (3.49 Å) and Phe399 (3.9 Å). Carbon C4 is
pointing towards a polar cavity that occupies two crystal water molecules (not shown in Figure 6) and
this carbon is virtually without any measurable interaction.
ACS Paragon Plus Environment

Page 11 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The structural analysis indicates that the S-enantiomer (S-11a) is bound to the protein. Accordingly, rac-
1
1a was separated into the enantiomers via chiral chromatography on a SFC device. Both enantiomers
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
were assessed in the enzyme inhibition assay and the eutomer was found with an IC50 value of 0.053 µM.
Our structural analyses (NMR, X-ray) indicated that the stereochemistry of the active isomer is confirmed
to have S configuration at the chiral carbon (Table 1). R-11a inhibits MetAP-2 with an IC50 of 13 µM,
which is >200 times less active than the S enantiomer. During the optimization process we continued to
separate enantiomers via chiral chromatography from racemic syntheses but for production of larger
25
compound quantities we established an asymmetric synthesis route for most efficient drug production.
Amide Chain length and substitution
The easily adjustable and deeply buried amide residue was considered at first during optimization, as a
strong impact on potency was anticipated. The residue attached to the lactam nitrogen was deemed to
play an important role for binding affinity, too. Though not as submerged as the amide substituent, the
X-ray analysis revealed contacts for the phenyl ring with the protein surface as insinuated above (Figure
5
and Figure 6B). This awareness suggested a dedicated optimization of the lactam N-substituent during
second approach.
The expansion of the thiophene with a methyl group beside the sulfur (11b) should fill in the pocket
superior to the unsubstituted thiophene of 11a. Apparently, the contrary is the case as the methylated
analog 11b of the original hit 11a is three times less active (Table 1).
a
Table 1: initial SAR of amide residue
HUVEC
MSC
Enzyme IC50 [µM] proliferation IC50
[
µM]
R
b
rac-11a
0.057 +/- 0.08
NT
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
S-11a
R-11a
0.050 +/- 0.006
17 +/- 2.1
1.9 +/- 0.93
c
ND
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1b
0.22 +/- 0.07
0.38 +/- 0.05
2.9 +/- 0.21
18 +/- 4.2
11c
11d
0.41 +/- 0.09
0.74 +/- 0.33
18 +/- 1.3
5.2 +/- 2.7
1
1e
1
1f
0.48 +/- 0.18
0.17 +/- 0.06
15 +/- 1.4
9.7 +/- 5.2
1
1g
1
1h
0.14 +/- 0.03
2.5 +/- 0.57
a
b
c
values are mean of three repetitions; NT: not tested; ND: no IC50 detectable
Comparison of 11a with a published structure of methionine in MetAP-2 (PDB: 1wkm) revealed
differences in the positions of the sulfur atoms of the thiophen in 11a and of the thioether from
methionine side chain of about 3 Å (Supplement Information, Figure S1). It was anticipated that
shortening the ethyl linker between the amide nitrogen and the thiophene should allow for better
substrate mimicking and pocket occupancy. The truncated linker in 11c was less potent then the
methylated hit-derivative 11b. Systematic methylation of the thiophene-carbons either confirmed (11d,
1
1f) or even decreased activity (11e). Reduction of thiophene ring size by introduction of a furan resulted
in a three-fold less potent compound (11g) compared to the screening hit. Further five membered rings
with up to three heteroatoms and an array of substituents were investigated but for none the activity
ACS Paragon Plus Environment

Page 13 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
could be optimized. Hence, these derivatives are not discussed in detail here. The saturated amide
residue of 11h was by factor of three less active as 11a and does not suggest that reduction of aryl load is
beneficial for increased potency. This observation was confirmed by additional analogs which are not
out-lined explicitly. Neither five membered aromatic nor saturated linear nor cyclic moieties allowed a
potency increase and consequently larger residues were investigated (table 2). The unsubstituted benzyl
derivative 11i was also not a very promising residue as it showed an IC50 of 440 nM what was factor eight
less potent than the screening hit 11a. Systematic permutation of one chlorine atom on the benzyl ring
proved that ortho (11j) and meta (11k) substitution delivered more potent compounds than para
substitution, with an IC50 of 5.3 µM for the latter and sub-micromolar IC50 for the other two. Accordingly,
further investigations focused on the ortho and meta positions. Also, with fluorine as substituent the
meta (11n) substitution showed slightly higher potency than the ortho (11m) position. The activity in the
HUVEC proliferation assay was also better for the ortho substituted isomers compared to their meta
analogs (11j vs 11k and 11m vs 11n).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Table 2: second SAR of amide residue
HUVEC
Enzyme IC50
No
proliferation IC50
µM]
[
µM]
[
R
11i
H
0.44 +/- 0.04
0.27 +/- 0.04
0.24 +/- 0.01
5.3 +/- 1.2
8.9 +/- 3
18 +/- 5.9
1.5 +/- 0.42
1
1j
1k
1l
1m
11n
2-Cl
3-Cl
4-Cl
2-F
3-F
1
b
1
ND
1
0.28 +/- 0.04
0.14 +/- 0.001
5.2 +/- 0.92
1.6 +/- 0.071
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
11o
3-Me
3,5-Cl2
3,5-F2
0.55 +/- 0.04
0.06 +/- 0.02
0.08 +/- 0.01
4.0 +/- 2.8
0.35 +/- 0.14
0.25 +/- 0.081
0.1 +/- 0.085
1
1
1p
1q
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1r
3-Cl, 5-F
0.06 +/- 0.006
a
b
values are mean of three repetitions; ND: no IC50 detectable
It was expected that the fluorine substitution of the benzylic moiety should be favorable in terms of
compound properties compared to aliphatic moieties. Nevertheless, the meta-methyl benzylamide 11o
was prepared as favorable lipophilic contacts had been anticipated. Surprisingly, 11o had reduced
biochemical and cellular activity compared to the meta-F derivative 11n. This observation motivated the
further analyses of halogen substitutions. The introduction of a second meta halogen improved
biochemical activity. In the enzymatic testing the bis-chloro (11p) and bis-fluoro (11q) benzylamides
proved to be as active as the fluoro, chloro mixed analog 11r. This similarity was confirmed in the anti-
proliferative activity in the HUVEC assay. All three bis-halogen substituted benzyl amides (11p, 11q, 11r)
impaired HUVEC proliferation with a low three digit nanomolar IC50
.
The structural analysis of “stopover” 11r from the first optimization cycle clearly demonstrates how both
halogen substituents are involved in vdW interactions in the lipophilic pocket (Figure 7). The fluorine
atom (brown) is pointing upwards and towards the protein surface. It is involved in lipophilic interactions
with the methylene group of His382 (3.41 Å) and the methyl-side chain of Ala414 (3.22 Å). The chlorine
atom (light green) is pointing down-wards to be in contact with the imidazole side chain of His339 (3.51
Å). In addition, this chlorine is sandwiched between the side chain of Ile338 (3.68 Å) and Tyr444 (3.50 Å).
ACS Paragon Plus Environment

Page 15 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7: X-Ray structure analysis of the optimized amide residue in 11r. Depicted are only the
distances of the halogens (F, top, brown; Cl, bottom, light green) on the benzylamide to the amino acid
side chains of MetAP-2 (6QEF).
The oxygen atom of the alcohol function is coordinating the two manganese ions with distances of 2.08 Å
and 2.19 Å, respectively, virtually identical to the position of this oxygen of the initial screening hit
(
distances of oxygen to metal ions are not explicitly shown in Figure 7).
Lactam residue
Benzylamide 11r was considered to be a good starting point for optimizing the lactam nitrogen
substituent (Table 3). The unsubstituted building block 18 was still active and inhibited the isolated
enzyme with an IC50 in the sub-micromolar range. Replacement of the proton with a methyl group (19)
had virtually no effect on biochemical activity. The sub-micromolar activity of 18 and 19 in the enzyme
assay clearly indicates the importance of tartronic-diamide motive for strong binding to MetAP-2.
Subsequent efforts focused on exploring the phenyl ring SAR through systematic evaluation of electron
donating and electron withdrawing substituents. (Figure 5B).
a
Table 3: SAR of the N-lactam residue
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
HUVEC
Enzyme IC50 [µM] proliferation IC50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
No
[
µM]
R
b
R-18
H
0.59 +/- 0.02
0.7 +/- 0.007
0.54 +/- 0.17
0.09 +/- 0.001
0.09 +/- 0.02
0.86 +/- 0.31
0.09 +/- 0.02
0.13 +/- 0.04
0.08 +/- 0.04
0.08 +/- 0.01
0.06 +/- 0.03
0.06 +/- 0.01
0.05 +/- 0.01
0.074 +/- 0.03
ND
ND
b
1
9
Me
1
1ra
1rb
2-Me-phenyl
3-Me-phenyl
4-Me-phenyl
2-CN-phenyl
17 +/- 2.8
0.3 +/- 0.12
1
11rc
0.24 +/- 0.078
29 +/- 1.3
1
1rd
1re
1
3-CN-phenyl
1.7 +/- 0.32
1
1rf
4-CN-phenyl
6.1 +/- 0.35
1
1rg
3-Acetamido-phenyl
4-Acetamido-phenyl
3-carbamoyl-phenyl
4-carbamoyl-phenyl
Oxindol-5-yl
0.28 +/- 0.095
0.032 +/- 0.013
0.31 +/- 0.035
1.6 +/- 0.74
11rh
11ri
1
1rj
2
0
1
0.17 +/- 0.07
0.015 +/- 0.01
2
2-Oxo-tetrahydro-quinolin-6-yl
a
b
values are mean of three repetitions; ND: no IC50 detectable
Even though all ortho-substituted derivatives (2-Me 11ra, 2-CN 11rd) were still sub-micromolar active in
the enzyme assay, their cellular activity was hardly measurable. The meta and para methyl isomers 11rb
and 11rc were equally potent having nanomolar IC50 values in the biochemical and cellular assays. The
polar cyano moiety was also not discriminating between the two positions, as the meta derivative 11re
ACS Paragon Plus Environment

Page 17 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
was as potent as the para isomer 11rf. Cellular potency of these derivatives (11ra – 11rf) was in all cases
less favorable compared to the unsubstituted compound 11r. The structural analysis of 11a had
disclosed that the carboxamide side chain of Asn329 is responsible for a polar protein surface in the
bottom area depicted in Figure 5B. This observation served as rationale to provide a functionality
capable of H-bonding interaction to this amide group. The donor/acceptor orientation of the amide
residue of Asn329 should be flexible to allow for optimal interaction. Based on this prerequisite and the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
26
anticipated distance an amide was considered to be a good H-bonding partner. Both, para and meta
position phenyl carboxamides (11ri, 11rj) and acetamides (11rg, 11rh) were examined. While the enzyme
data for these derivatives was not significantly differentiated, the para-acetamide 11rh led to the best
cellular response with an IC50 in the two digit nanomolar range. Reduction of conformational flexibility of
the amide should minimize entropic penalties and should reduce rotational freedom contributing to
improved binding. Accordingly, the donor/acceptor- substitution pattern was captured in the oxindole 20
and the more expanded analog 21. The biochemical data looked quite the same for the open chain
(
11rh) and both cyclic derivatives (20, 21) so that the working hypothesis to improve enzyme inhibition
had to be challenged. Surprisingly, a remarkable difference (factor of 10) was observed in the effect of 20
vs 21 on HUVEC proliferation for these compounds. Efficient cellular growth impairment was achieved
with the 2-oxo-tetrahydro-quinolinyl substitution of 21. The reason for the better proliferation inhibition
can only be hypothesized but different binding kinetics might be responsible. The enzyme used for in
vitro enzyme inhibition is a recombinant His-tagged form and may behave differently than the native,
full-length MetAP-2 protein. Since a quite decent concentration of MetAP-2 enzyme had to be applied in
the biochemical assay the measured IC50 values might already have reached the resolution for high
potency compounds. In addition, MetAP-2 associates with ribosomal subunits which may influence the
conformation of the protein and consequently, the potency of certain MetAP-2 inhibitors in cells. These
factors may account for some of the discrepancies observed between enzyme proliferation inhibition.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The optimized ligand was crystallized together with MetAP-2 and the structural examination of 21 is
depicted in Figure 8.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8: X-ray structure analysis of 21 with focus on the additional polar interaction of the tetrahydro-
quinone-NH with the ASN329 side chain amide of MetAP-2 (6QED).
In Figure 8 only the additional polar interaction between the lactam N-H with the oxygen of the Asn329
side chain amide is shown. The H-bond (3.06 Å) is probably contributing to the interaction of 21 with
MetAP-2.
For the covalent binding of irreversible inhibitors to the imidazole of His231 the nature of the metal ion
is of limited relevance. The ion might only function as Lewis acid, coordinating to the spiro-epoxide
2
oxygen to facilitate His231-imidazole attack. All reversible ligands coordinate the metal ion co-factor.
Accordingly, identification of physiological metal cofactors for MetAP-2 is critical because literature
known inhibitors like A-310840 (3-{[(naphthalen-2-yl)methyl]sulfanyl}-4H-1,2,4-triazole) differentiate this
enzyme with different metal cofactors with potencies varying up to 1000-fold. In vitro investigations of
eight different metal ions on purified recombinant human MetAP-2 demonstrated that manganese, as
well as cobalt, maximally stimulate MetAP-2 activity indicating that either manganese or cobalt might be
6
the MetAP-2 cofactor in vivo. As such, compound 21 was evaluated in the enzyme assay with both
ACS Paragon Plus Environment

Page 19 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
manganese and cobalt to determine if the hydroxy-malonic motif discriminates between the two metals.
Derivative 21 was found to be similarly active under cobalt and manganese assay conditions: IC50 = 160
nM vs 74 nM. Like other compounds from this new MetAP-2 binding class the development candidate is
not active in the MetAP-1 enzyme assay.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
7
28
The pKa of 21 was determined to be around 11 and the logD at pH 7.4 is 2, explaining the good
solubility profile in bio-relevant media (FaSSIF: 158 µg/ml; FeSSIF: 334 µg/ml). A favorable molecular
weight of 431 Da and the tPSA of 99 contribute to good passive permeability and acceptable efflux in the
Caco-2 assay (a>b: 16 1E-6 cm/s; ER: 5). The metabolic stability in liver microsomes from mouse and
human was <10 µl/min/mg protein qualifying the compound for further in vivo PK profiling. After iv
dosing in mice, the compound showed favorable PK characteristics, i.e. clearance (CL), volume of
1/2
distribution (Vss) and half-life (t ) were determined to be 0.18 L/h/kg, 0.78 L/kg and 3 h, respectively.
The oral bioavailability was 35% at 0.5 mg/kg and increased to ~100% at 50 mg/kg suggesting the
saturation of a clearance mechanism (Figure 9). The favorable potency and PK data of 21 prompted us to
investigate its tumor growth inhibition potential in vivo.
Figure 9: Dose-normalized plasma concentrations of 21 after 0.2 mg/kg iv and 50 mg/kg po dosing
For investigating in vivo efficacy of 21 the human glioblastoma U87-MG model was used. The tumor cells
were subcutaneously injected in nude mice. Once tumors were established with a mean tumor volume
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
of 300 mm³, the animals were orally treated daily with different doses of 21 (5, 10 and 20 mg/kg QD).
The results are shown in Figure 10.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 10: in vivo tumor growth inhibition of 21 in human U87-MG glioblastoma model. Tumor
volumes for the 12-day treatment period (volumes for the individual animals at the end of the study
are shown in the Supporting Information Figure S2).
Dose dependent anti-tumor activity could be shown with compound 21. Best efficacy was achieved at
doses of 10 and 20 mg/kg daily (T/C value of 45%), while the lowest dose of 5 mg/kg resulted in a T/C
value of 78%. Applying the highest daily dose but using an alternative scheduling (10 mg/kg BID) did not
enhance the anti-tumor activity. All regimens were well tolerated, and no body weight loss was observed
during the treatment period of 12 days (data not shown).
Conclusion
29
The anti-angiogenic properties of the natural product fumagillin were discovered already in 1990.
However, it took seven additional years until the molecular target of this compound class could be
30
identified. Since then many covalent and reversible ligands were investigated to understand the target
biology and develop new drugs for indications with high medical need like cancer or obesity. For the
latter condition at least one NCE (ZGN-1061) is in active clinical development. Compelling data render
ACS Paragon Plus Environment

Page 21 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MetAP-2 to be also a promising drug target for oncology. All clinical investigations in oncology so far
were only considering derivatives of fumagillin were discontinued due to unknown reasons. Using a
structure-based approach the successful optimization of the screening hit 11a resulted in the advanced
lead 21 which was well tolerated and showed anti-tumor efficacy in a mouse glioblastoma model.
Furthermore, the derivative 21 set the basis for further optimizations to establish the clinical compound
M8891.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Experimental Section
Chemistry General Information. All reactions were carried out under nitrogen atmosphere or in sealed
vials unless noted otherwise. Dry solvents and reagents were of commercial quality and were used as
purchased. Reactions were magnetically stirred and monitored by thin-layer chromatography (TLC) using
Merck silica gel 60 F254 by fluorescence quenching under UV light or by high-pressure liquid
chromatography with subsequent mass detection (LCMS) detection, except if indicated otherwise.
LCMS-Analysis: Method A: A-0.1 % TFA in H
50 x 4.6 mm, 3.5µm), +ve mode. Method B: A- 0.1% HCOOH, B- Methanol: flow: 1.0 ml/min, column:
Eclipse plus C18 (50 x 4.6 mm); Method C: A- 0.1% HCOOH, B- Acetonitrile: flow: 1.0ml/min, column:
Atlantis dC18 (50 x 4.6 mm.) + ve mode. HPLC-analysis: A-0.1 % TFA in H O, B-0.1 % TFA in ACN: flow: 2.0
2
O, B-0.1 % TFA in ACN: flow: 2.0 mL/min, column: XBridge C8
(
2
mL/min, column: XBridge C8 (50 x 4.6 mm, 3.5 µm). Chiral HPLC analysis: Method A: Mobile phase: 0.1%
DEA in hexane/IPA = 60:40, column: Chiralpak AD-H (250 x 4.6) mm, 5 µm, flow: 1.0mL\min; Method B:
Mobile phase: n-heptane/IPA = 60:40, column: Chiralpak AD-H (250 x 4.6) mm, 5 μm, flow: 1.0 mL/min;
Method C: Mobile phase: 0.1% TFA in hexane/IPA = 80:20, column: Chiralcel OJ-H (250 x 4.6) mm, 5 μm,
flow: 1.0mL/min; Method D: Mobile phase: heptane/IPA = 50:50, column: Chiralpak AD-H (250 x 4.6)
mm, 5 μm, flow rate: 0.8mL\min; Method E: Mobile phase: 0.1% DEA in hexane/IPA = 70:30, column:
Chiralpak AD-H (250 x 4.6) mm, 5 µm, flow: 1.0mL\min; Method F: Mobile phase: 0.1% DEA in
hexane/IPA = 70:30, column: Chiralpak OD-H (250 x 4.6) mm, 5 µm, flow: 1.0mL\min; Method G: Mobile
phase: hexane/IPA = 70:30, column: Chiralpak AD-H (250 x 4.6) mm, 5 µm, flow rate: 1.0mL\min.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Retention times (Rt) are given in minutes. In addition, TLC plates were stained using phosphomolybdic
acid or potassium permanganate stain. Chromatographic purification of products (flash chromatography)
was performed on Isco Combiflash systems using Redisep columns and ethyl acetate/heptanes gradients.
Concentration under reduced pressure was performed by rotary evaporation at 40 °C at the appropriate
pressure unless otherwise stated. The purity of the compounds reported in the manuscript was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
established through HPLC-MS methodology. HPLC analyses were run according to the LCMS method. H
NMR (in DMSO-d ) and mass spectra are in agreement with the structures and were recorded on a
6
Bruker NMR 400 MHz spectrometer if not stated differently (Tetramethylsilane as an internal standard)
and Vacuum Generators VG 70-70 or 70-250 at 70 eV, respectively. Elemental analyses (obtained with a
Perkin-Elmer 240 BCHN analyzer) for the final products were within 0.4% of calculated values if not
stated otherwise. All compounds reported in the manuscript have a purity ≥95% unless noted otherwise.
(S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid (2-thiophen-2-yl-ethyl)-amide (S-11a)
Synthesis of 2-oxo-1-phenyl-N-[2-(thiophen-2-yl)ethyl]pyrrolidine-3-carboxamide
1
-Phenyl-2-oxo-3-pyrrolidine carboxylic acid (200 mg, 0.95 mmol), 2-(thiophen-2-yl)-ethan-1-amine
hydrochloride (155 mg, 0.95 mmol), EDCI (549 mg, 2.84 mmol), HOBt (192 mg, 1.42 mmol) and TEA (132
µL, 0.95 mmol) were dissolved in DMF (5 mL) and heated in the microwave to 160°C for 5 min. After the
addition of water, the precipitate was filtered and dried, giving the product in 80 % (239 mg) yield as
colorless solid.
Synthesis of 3-hydroxy-2-oxo-1-phenyl-N-[2-(thiophen-2-yl)ethyl]pyrrolidine-3-carboxamide
2
-Oxo-1-phenyl-N-[2-(thiophen-2-yl)ethyl]pyrrolidine-3-carboxamide (100 mg, 0.32 mml), tert-butyl
hydro peroxide (70 % in water, 62 mg, 0.48 mmol) and sodium ethylate (20 % in ethanol, 182 µL, 0.48
mmol) were mixed with tert-butanol (10 ml) and heated to 90°C for 4 h. After concentration in vacuum
the residue was distributed between ethyl acetate and water, the organic phase dried over Na
2
SO ,
4
filtered and after removal of the solvent purified by chromatography, giving the product in 54 % (57 mg)
yield. 3-Hydroxy-2-oxo-1-phenyl-N-[2-(thiophen-2-yl)ethyl]pyrrolidine-3-carboxamide (40 mg) was
ACS Paragon Plus Environment

Page 23 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
separated into the enantiomers via chiral HPLC (ChiralCel OH; n-heptane/methanol = 7:3, giving S-11a in
1
5
1 % (20 mg) yield. Rt 7.19 min. H NMR (500 MHz, DMSO-d6) δ 8.08 (t, J = 5.9, 1H), 7.71 (s, 1H), 7.69 (d,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
J = 8.8, 1H), 7.48 – 7.37 (m, 2H), 7.33 (dt, J = 10.3, 5.1, 1H), 7.23 – 7.13 (m, 1H), 6.96 (dd, J = 5.0, 3.5, 1H),
6
2
.90 (t, J = 7.1, 1H), 6.61 (d, J = 17.5, 1H), 3.91 – 3.80 (m, 2H), 3.48 – 3.38 (m, 1H), 3.33 – 3.22 (m, 3H),
+
.98 (t, J = 7.3, 2H), 2.58 – 2.52 (m, 1H), 2.10 (m, 1H); HRMS calced for C17
H
19
N
O
2 3
S (M+H ) 331.1111
found 331.1118. R-11a was also obtained in 51 % (20 mg) yield. Rt 15.21 min. HRMS calced for
+
C
17
H
19
N
2
O
3
S (M+H ) 331.1111 found 331.1119.
Compounds 11b-11r were prepared as described below:
To a solution of (R)-3-hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 16 (800 mg, 3.62mmol) and
the respective amine (4.34 mmol) in dry DMF was added EDCI (2.07 g, 10.86 mmol) and HOBt (731 mg,
5
.43 mmol). The reaction mass was heated for 20 min at 160°C under microwave. After completion of
the reaction, DMF was removed under vacuum. The residue was diluted with water and extracted with
ethyl acetate. The organic layer was dried over Na SO and concentrated under reduced pressure. The
crude was purified by flash column chromatography to afford the title compounds.
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid [2-(5-methyl-thiophen-2-yl)-ethyl]-amide (11b)
2
4
(
The title compound was obtained in 15 % (12 mg) yield as light brown solid.
1
H NMR (400 MHz, DMSO) δ 8.07 (t, J = 5.9, 1H), 7.73 – 7.65 (m, 2H), 7.40 (dd, J = 10.8, 5.3, 2H), 7.17 (t, J
= 7.4, 1H), 6.64 (d, J = 3.3, 1H), 6.63 (s, 1H), 6.59 (dd, J = 3.3, 1.1, 1H), 3.91 – 3.75 (m, 2H), 3.30 – 3.33 (m,
1
H), 3.15 (m, 1H), 2.85 (t, J = 7.4, 2H), 2.53 (m, 1H), 2.34 (s, 3H), 2.07 (dt, J = 12.9, 8.0, 1H). LCMS:
(
(
(
Method A) 345.3 (M+H), Rt 3.85 min, 99.16 (max), 98.99 (254nm); HPLC: Rt 3.92 min, 98.77 (max), 98.67
+
254nm); HRMS calced for C18
H
21
N
2
O
3
S (M+H ) 345.1267 found 345.1274.
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid (thiophen-2-ylmethyl)-amide (11c)
The title compound was obtained in 47 % (44 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO) δ 8.58 (t, J = 6.2, 1H), 7.70 (d, J = 1.1, 1H), 7.68 (d, J = 1.0, 1H), 7.45 – 7.33 (m,
3
H), 7.22 – 7.14 (m, 1H), 6.99 – 6.87 (m, 2H), 6.66 (s, 1H), 4.44 (d, J = 6.4, 2H), 3.91 – 3.78 (m, 2H), 2.59 –
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2.52 (m, 1H), 2.09 (dt, J = 12.9, 8.5, 1H). LCMS: (Method A) 317 (M+H), Rt 3.34 min, 96.06 (max), 98.75
(
254nm); HPLC: Rt 3.28 min, 98.41 (max), 99.10 (254nm); Chiral HPLC: (Method B) Rt 6.57 min, area %
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
+
9
9.05 (210nm); HRMS calced. for C16
H
16
N
2
O
3
S (M+H ) 317.0954 found 317.0966.
(
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid (3-methyl-thiophen-2-ylmethyl)-amide (11d)
The title compound was obtained in 46 % (45 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO) δ 8.45 (t, J = 6.1, 1H), 7.74 – 7.66 (m, 2H), 7.46 – 7.37 (m, 2H), 7.25 (d, J = 5.1,
1H), 7.22 – 7.14 (m, 1H), 6.79 (d, J = 5.1, 1H), 6.64 (s, 1H), 4.39 (dd, J = 15.2, 6.4, 1H), 4.33 (dd, J = 15.4,
6
.4, 1H), 3.84 (ddd, J = 14.2, 9.2, 3.7, 2H), 2.60 – 2.52 (m, 1H), 2.16 (s, 3H), 2.09 (dt, J = 12.9, 8.5, 1H).
LCMS: (Method A) 331 (M+H), Rt 3.52 min, 96.54 (max), 95.65 (254nm); HPLC: (Method A) Rt 3.65 min,
+
9
6.02 (max), 94.41 (254nm); HRMS calced for C17
H
19
N
2
O
3
S (M+H ) 331.1111 found 331.1117.
(
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid (4-methyl-thiophen-2-ylmethyl)-amide (11e)
The title compound was obtained in 52 % (43 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO) δ 8.52 (t, J = 6.2, 1H), 7.70 (dd, J = 5.4, 3.4, 2H), 7.44 – 7.37 (m, 2H), 7.22 –
7
–
9
3
.14 (m, 1H), 6.94 – 6.90 (m, 1H), 6.76 (s, 1H), 6.66 (s, 1H), 4.45 – 4.30 (m, 2H), 3.93 – 3.78 (m, 2H), 2.56
2.51 (m, 1H), 2.13 (s, 3H), 2.12 – 2.05 (m, 1H). LCMS: (Method A) 331 (M+H), Rt 3.74 min, 92.09 (max),
+
2.33 (254nm); HPLC: Rt 3.66 min, 96.29(max), 96.27(254nm); HRMS calced for C17
H
19
N
2
O
3
S (M+H )
31.1111 found 331.1121.
(S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid (5-methyl-thiophen-2-ylmethyl)-amide (11f)
The title compound was obtained in 44 % (37 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO) δ 8.48 (t, J = 6.2, 1H), 7.69 (dd, J = 8.7, 1.1, 2H), 7.44 – 7.36 (m, 2H), 7.23 –
7
2
.09 (m, 1H), 6.71 (d, J = 3.3, 1H), 6.65 (s, 1H), 6.63 – 6.54 (m, 1H), 4.44 – 4.28 (m, 2H), 3.92 – 3.76 (m,
H), 2.57 – 2.51 (m, 1H), 2.36 (d, J = 0.8, 3H), 2.15 – 2.04 (m, 1H). LCMS: (Method C) 331 (M+H), Rt 2.52
+
min, 91.54 (max); HPLC: Rt 3.65 min, 92.26 (max), 90.62 (254nm); HRMS calced for C17
H
19
N
2
O
3
S (M+H )
331.1111 found 331.1118.
(
S)- 3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid (2-furan-2-yl-ethyl)-amide (11g)
ACS Paragon Plus Environment

Page 25 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The title compound was obtained in 51 % (50 mg) yield as white solid
1
H NMR (400 MHz, DMSO) δ 8.06 (t, J = 5.9, 1H), 7.70 (t, J = 1.5, 1H), 7.68 (t, J = 1.6, 1H), 7.52 (dd, J = 1.8,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
0
2
.7, 1H), 7.44 – 7.37 (m, 2H), 7.22 – 7.14 (m, 1H), 6.62 (s, 1H), 6.35 (dd, J = 3.1, 1.9, 1H), 6.16 (dd, J = 3.1,
.7, 1H), 3.92 – 3.76 (m, 2H), 3.40 (td, J = 13.7, 7.4, 1H), 2.79 (t, J = 7.3, 2H), 2.53 (dd, J = 6.7, 3.6, 1H),
+
.15 – 2.00 (m, 1H). HRMS calced for C17
H
19
N
2
O
4
(M+H ) 315.1339 found 315.1348.
(S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid (2-cyclopentyl-ethyl)-amide (11h)
The title compound was obtained in 53 % (38 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO) δ 8.72 (t, J = 6.4, 1H), 7.69 (dd, J = 8.7, 1.0, 2H), 7.45 – 7.33 (m, 2H), 7.27 (dt, J
=
4
8.8, 2.1, 1H), 7.18 (dd, J = 13.0, 4.5, 2H), 7.10 (d, J = 9.0, 1H), 6.76 (s, 1H), 4.38 (dd, J = 15.7, 6.6, 1H),
.24 (dd, J = 15.9, 6.0, 1H), 3.89 – 3.82 (m, 2H), 2.62 – 2.54 (m, 1H), 2.12 (dt, J = 12.9, 7.6, 1H); LCMS:
(
Method A) 317.0 (M+H), Rt 4.03 min, 96.61 (max), 98.74 (254nm); HPLC: Rt 4.1 min, 97.63(max), 98.56
+
(254nm); Chiral HPLC: (Method G) Rt 5.52 min Area % 100 (210nm); HRMS calced for C18
17.186 found 317.1868.
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid benzylamide (11i)
H
25
N
2
O
3
(M+H )
3
(
The title compound was obtained in 15 % (12 mg) yield as light brown solid.
1
H NMR (400 MHz, DMSO) δ 8.51 (t, J = 6.0, 1H), 7.69 (dd, J = 8.6, 0.9, 2H), 7.45 – 7.34 (m, 2H), 7.32 –
7
.24 (m, 4H), 7.20 (dt, J = 14.8, 4.6, 2H), 6.68 (s, 1H), 4.33 (dd, J = 15.1, 6.5, 1H), 4.27 (dd, J = 15.1, 6.2,
1H), 3.92 – 3.74 (m, 2H), 2.63 – 2.52 (m, 2H), 2.11 (dt, J = 12.9, 7.9, 1H); LCMS: (Method A) 311.3 (M+H),
Rt 3.29 min, 99.74 (max), 99.55 (254nm); HPLC: (Method A) Rt 3.93 min, 98.92 (max), 99.61 (254nm);
+
Chiral HPLC: (Method E) Rt 8.81 minArea % 100 (210nm); HRMS calced for C18
H
19
N
2
O
3
(M+H ) 311.139
found 311.1399.
(
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 2-chloro-benzylamide (11j)
The title compound was obtained in 37 % (26 mg) yield as brown gum.
1
H NMR (400 MHz, DMSO) δ 8.59 (t, J = 6.3, 1H), 7.70 (dt, J = 9.0, 1.7, 2H), 7.44 – 7.24 (m, 6H), 7.21 – 7.14
m, 1H), 6.79 (s, 1H), 4.41 (dd, J = 16.2, 6.5, 1H), 4.31 (dd, J = 16.1, 6.1, 1H), 3.86 (dd, J = 8.5, 5.6, 2H),
(
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
+
2.65 – 2.57 (m, 1H), 2.14 (dt, J = 12.9, 7.7, 1H); HRMS calced for C18
45.1004.
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 3-chloro-benzylamide (11k)
H
18ClN
2
O (M+H ) 345.1 found
3
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
The title compound was obtained in 34 % (27 mg) yield as brown solid.
1
H NMR (400 MHz, DMSO) δ 8.65 (t, J = 6.2, 1H), 7.69 (dd, J = 8.7, 1.1, 2H), 7.43 – 7.37 (m, 2H), 7.33 (t, J =
7.6, 2H), 7.27 (dd, J = 6.6, 2.1, 1H), 7.22 (d, J = 7.5, 1H), 7.18 (dd, J = 10.5, 4.2, 1H), 6.73 (s, 1H), 4.35 (dd, J
= 15.2, 6.6, 1H), 4.25 (dd, J = 15.4, 6.1, 1H), 3.90 – 3.80 (m, 2H), 2.61 – 2.54 (m, 1H), 2.11 (dt, J = 13.0, 7.8,
1
H); LCMS: (Method A) 345 (M+H), Rt 3.86 min, 93.21 (max), 98.65 (254nm); HPLC: Rt 3.93 min, 90.93
(
max), 98.60 (254nm); Chiral HPLC: (Method D) Rt 8.26 min, area % 93.34 (210nm); HRMS calced for
+
C
18
H
18ClN
2
O
3
(M+H ) 345.1 found 345.1004.
(
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 4-chloro-benzylamide (11l)
The title compound was obtained in 37 % (27 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO) δ 8.60 (t, J = 6.3, 1H), 7.69 (dt, J = 8.9, 1.7, 2H), 7.43 – 7.33 (m, 4H), 7.31 – 7.25
(
m, 2H), 7.21 – 7.13 (m, 1H), 6.70 (s, 1H), 4.31 (dd, J = 15.2, 6.5, 1H), 4.24 (dd, J = 15.2, 6.2, 1H), 3.91 –
+
3
3
.76 (m, 2H), 2.56 (ddd, J = 12.8, 7.1, 4.4, 1H), 2.17 – 2.04 (m, 1H); HRMS calced for C18
H
18ClN
2
O (M+H )
3
45.1 found 345.1004.
(
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 2-fluoro-benzylamide (11m)
The title compound was obtained in 25 % (17 mg) yield as brown gum.
1
H NMR (400 MHz, DMSO) δ 8.53 (t, J = 6.3, 1H), 7.73 – 7.66 (m, 2H), 7.43 – 7.37 (m, 2H), 7.34 (t, J = 7.8,
1
H), 7.31 – 7.25 (m, 1H), 7.20 – 7.11 (m, 3H), 6.73 (s, 1H), 4.38 (dd, J = 15.6, 6.4, 1H), 4.31 (dd, J = 15.5,
5
.9, 1H), 3.91 – 3.80 (m, 2H), 2.58 (ddd, J = 12.9, 7.0, 4.5, 1H), 2.12 (dt, J = 13.0, 7.7, 1H); HRMS calced for
+
C
18
H
18FN
2
O
3
(M+H ) 329.1296 found 329.1303.
(S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 3-fluoro-benzylamide (11n)
The title compound was obtained in 44 % (34 mg) yield as brown solid.
ACS Paragon Plus Environment

Page 27 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
H NMR (400 MHz, DMSO) δ 8.64 (t, J = 6.4, 1H), 7.69 (dd, J = 8.7, 1.1, 2H), 7.43 – 7.37 (m, 2H), 7.33 (td, J
=
1
1
7.9, 6.1, 1H), 7.21 – 7.14 (m, 1H), 7.10 (d, J = 7.4, 2H), 7.08 – 6.99 (m, 2H), 6.73 (s, 1H), 4.37 (dd, J =
5.5, 6.6, 1H), 4.26 (dd, J = 15.4, 6.1, 1H), 3.92 – 3.75 (m, 2H), 2.63 – 2.54 (m, 1H), 2.12 (dt, J = 12.9, 7.7,
H). LCMS: (Method A) 329 (M+H), Rt 3.50 min, 94.01(max), 96.23(254nm); HPLC: Rt 3.66 min, 91.55
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
max), 94.61 (254nm); Chiral HPLC: (Method D) Rt 7.14 min, area % 93.67 (210nm); HRMS calced for
+
C
18
H
18FN
2
O
3
(M+H ) 329.1296 found 329.1305.
(S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 3-methyl-benzylamide (11o)
The title compound was obtained in 10 % (7 mg) yield as light brown solid.
1
H NMR (400 MHz, DMSO) δ 8.46 (t, J = 6.2, 1H), 7.70 (d, J = 7.8, 2H), 7.46 – 7.35 (m, 2H), 7.17 (t, J = 7.3,
2
6
H), 7.07 (d, J = 4.4, 2H), 7.02 (d, J = 7.8, 1H), 6.68 (s, 1H), 4.29 (dd, J = 15.0, 6.5, 1H), 4.23 (dd, J = 15.1,
.2, 1H), 3.96 – 3.74 (m, 2H), 2.64 – 2.52 (m, 1H), 2.26 (s, 3H), 2.11 (dt, J = 12.8, 7.8, 1H). LCMS: (Method
A) 325.3 (M+H), Rt 3.74 min, 99.67 (max), 99.56 (254nm); HPLC: Rt 3.78 min, 98.77 (max), 99.48
+
(
254nm); HRMS calced for C19
H
21
N
2
O
3
(M+H ) 325.1547 found 325.1557.
(
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 3,5-dichloro-benzylamide (11p)
The title compound was obtained in 40 % (35 mg) yield as brown solid.
1
H NMR (400 MHz, DMSO) δ 8.74 (t, J = 6.4, 1H), 7.70 (dd, J = 8.7, 1.0, 2H), 7.45 (t, J = 1.9, 1H), 7.44 –
7
.36 (m, 2H), 7.32 (d, J = 1.9, 2H), 7.18 (t, J = 7.4, 1H), 6.77 (s, 1H), 4.37 (dd, J = 15.6, 6.7, 1H), 4.23 (dd, J =
15.7, 6.0, 1H), 3.90 – 3.78 (m, 2H), 2.63 – 2.53 (m, 1H), 2.12 (dt, J = 13.0, 7.7, 1H); LCMS: (Method A) 379
(
M+H), Rt 4.43 min, 92.92 (max), 96.66 (254nm); HPLC: Rt 4.43 min, 96.87 (max), 98.91 (254nm); HRMS
+
calced for C18
H
17Cl
2
N
2
O
3
(M+H ) 379.0611 found 379.0621.
(
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 3,5-difluoro-benzylamide (11q)
The title compound was obtained in 56 % (54 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO) δ 8.72 (t, J = 6.4, 1H), 7.73 – 7.67 (m, 2H), 7.43 – 7.37 (m, 2H), 7.18 (dd, J =
10.5, 4.2, 1H), 7.06 (tt, J = 9.4, 2.4, 1H), 7.00 – 6.95 (m, 2H), 6.77 (s, 1H), 4.39 (dd, J = 15.9, 6.8, 1H), 4.24
(
dd, J = 15.8, 6.0, 1H), 3.86 (dd, J = 8.0, 5.7, 2H), 2.63 – 2.55 (m, 1H), 2.18 – 2.07 (m, 1H). LCMS: (Method
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
A) 347.2 (M+H), Rt 3.86 min, 90.39(max), 94.47(254nm); HPLC: Rt 3.82 min, 92.48 (max), 92.88 (254nm);
+
Chiral HPLC: (Method B) Rt 5.14 min Area % 99.84(210nm); HRMS calced for C18
47.1202 found 347.1206.
S)-3-Hydroxy-2-oxo-1-phenyl-pyrrolidine-3-carboxylic acid 3-chloro-5-fluoro-benzylamide (11r)
H
17
F
2
N
2
O
3
(M+H )
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
(
The title compound was obtained in 55 % (43 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO-d ) δ 8.72 (t, J = 6.36 Hz, 1H), 7.69 (d, J = 7.84 Hz, 2H), 7.40 (t, J = 8.36 Hz, 2H),
6
7.28-7.25 (m, 1H), 7.20-7.16 (m, 2H), 7.11-7.09 (m, 1H), 6.76 (s, 1H), 4.41-4.35 (m, 1H), 4.27-4.22 (m, 1H),
3
9
.86 (t, J = 7.44 Hz, 2H), 2.62-2.55 (m, 1H), 2.16-2.09 (m, 1H). LCMS: (Method A) 363 (M+H), Rt 4.02 min,
7.87(max); 99.35(254nm). HPLC: Rt 4.12 min, 97.88 (max); 99.35 (254nm). Chiral HPLC: (Method C) Rt
+
(
min): 12.10 Area % 100. HRMS calced for C18
H
17ClFN
2
O
3
(M+H ) 363.0906 found 363.0907.
Compounds 11ra – 11rj were prepared according to the following general protocol:
(R)-3-Hydroxy-2-oxo-pyrrolidine-3-carboxylic acid 3-chloro-5-fluoro-benzylamide (R-18; 70 mg; 0.24
mmol) and the respective aryl halogenide (0.29 mmol) were dissolved in dry 1,4-Dioxan (4 ml) and
purged with nitrogen for 10min. Then K
2
CO (101 mg; 0.73 mmol), CuI (186 mg; 0.98 mmol) and N,N´-
3
Dimethylethylendiamin (0.13 ml; 1.22 mmol) were added to the reaction mixture and heated to 100°C
for 2 days. Then the reaction was filtered and washed with DCM and MeOH. The filtrate was extracted
with DCM/water. The organic layer was dried over MgSO4, filtered and evaporated in vacuo. The
product was purified on prep. HPLC.
(
S)-3-Hydroxy-2-oxo-1-o-tolyl-pyrrolidine-3-carboxylic acid 3-chloro-5-fluoro-benzylamide (11ra)
The title compound was obtained in 23 % (43 mg) yield as white solid.
1
6
H NMR (400 MHz, DMSO-d ) δ 8.69 (t, J = 6.4, 1H), 7.32 – 7.17 (m, 6H), 7.11 (d, J = 9.7, 1H), 6.70 (s, 1H),
4
.41 (dd, J = 15.9, 6.7, 1H), 4.24 (dd, J = 15.8, 6.0, 1H), 3.68 (dd, J = 8.3, 5.3, 2H), 2.65 – 2.57 (m, 1H), 2.19
(dt, J = 13.0, 7.6, 1H), 2.11 (s, 3H); LCMS: (Method A) 377.0 (M+H), Rt 4.16 min, 96.51(max), 96.73
(220nm). HPLC: Rt 4.97 min, 99.09 (max), 98.88 (220nm); Chiral HPLC (after prep): (Method B) Rt 5.15
+
min; area % 99.16; HRMS calced for C19
H
19ClFN
2
O
3
(M+H ) 377.1063 found 377.1068.
ACS Paragon Plus Environment

Page 29 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(S)-3-Hydroxy-2-oxo-1-m-tolyl-pyrrolidine-3-carboxylic acid 3-chloro-5-fluoro-benzylamide (11rb)
The title compound was obtained in 9 % (24 mg) yield as colorless gum.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
H NMR (400 MHz, DMSO-d ) δ 8.72 (t, J = 6.3, 1H), 7.50 (d, J = 6.5, 2H), 7.31 – 7.24 (m, 2H), 7.20 (s, 1H),
6
7
1
.10 (d, J = 9.7, 1H), 7.00 (d, J = 7.3, 1H), 6.74 (s, 1H), 4.38 (dd, J = 15.8, 6.7, 1H), 4.23 (dd, J = 15.7, 6.0,
H), 3.83 (dd, J = 8.0, 5.6, 2H), 2.63 – 2.53 (m, 1H), 2.11 (dt, J = 13.0, 7.6, 1H); LCMS: (Method A) 377.0
(M+H), Rt 4.16 min, 96.51 (max), 96.73 (220nm). HPLC: Rt 4.44 min, 96.29 (max), 97.81 (254nm); Chiral
+
HPLC (after prep): (Method B) Rt 5.18 min, area % 99.85; HRMS calded for C19
77.1063 found 377.1069.
S)-3-Hydroxy-2-oxo-1-p-tolyl-pyrrolidine-3-carboxylic acid 3-chloro-5-fluoro-benzylamide (11rc)
H
19ClFN
2
O
3
(M+H )
3
(
The title compound was obtained in 15 % (32 mg) yield as white solid.
1
H NMR (400 MHz, DMSO-d ) δ 8.72 (t, J = 6.4, 1H), 7.60 – 7.55 (m, 2H), 7.27 (dt, J = 8.8, 2.2, 1H), 7.20
6
(dd, J = 4.9, 3.3, 3H), 7.10 (d, J = 9.7, 1H), 6.73 (s, 1H), 4.37 (dd, J = 15.8, 6.7, 1H), 4.23 (dd, J = 15.8, 6.0,
H), 3.82 (dd, J = 7.5, 6.1, 2H), 2.63 – 2.54 (m, 1H), 2.28 (s, 3H), 2.10 (dt, J = 13.0, 7.6, 1H); LCMS:
1
(
(
(
(
Method A) 377 (M+H), Rt 4.40 min, 97.39(max), 98.43(254nm); HPLC: (Method A) Rt 4.42 min, 98.42
+
max), 98.40 (254nm); HRMS calced for C19
H
19ClFN
2
O
3
(M+H ) 377.1063 found 377.1064.
S)-1-(2-Cyano-phenyl)-3-hydroxy-2-oxo-pyrrolidine-3-carboxylic acid 3-chloro-5-fluoro-benzylamide
11rd)
The title compound was obtained in 9 % (18 mg) yield as off white solid.
1
H NMR (400 MHz, DMSO-d ) δ 8.74 (t, J = 6.5, 1H), 7.92 (dd, J = 8.0, 1.5, 1H), 7.84 – 7.74 (m, 1H), 7.51
6
(
(
ddd, J = 6.3, 3.7, 1.1, 2H), 7.27 (dd, J = 8.8, 2.2, 1H), 7.21 (s, 1H), 7.11 (d, J = 8.8, 1H), 6.87 (s, 1H), 4.39
dd, J = 15.7, 7.0, 1H), 4.26 (dd, J = 15.8, 6.2, 1H), 3.97 – 3.76 (m, 3H), 2.69 – 2.56 (m, 2H), 2.28 – 2.14 (m,
1
H). LCMS: (Method A) 388.0 (M+H), Rt 3.83 min, 96.26 (max), 97.18 (254nm); HPLC: (Method A) Rt 3.87
min, 97.15 (max), 97.21 (254nm); Chiral HPLC (after prep): (Method B) Rt (min): 5.85 Area % 97.80;
+
HRMS calced for C19
H
16ClFN
3
O
3
(M+H ) 388.0859 found 388.0861.
ACS Paragon Plus Environment