Novel Human Aminopeptidase N Inhibitors: Discovery and Optimization of Subsite Binding Interactions

Article abstract of DOI:10.1021/acs.jmedchem.9b00757

Aminopeptidase N (APN/CD13) is a zinc-dependent M1 aminopeptidase that contributes to cancer progression by promoting angiogenesis, metastasis, and tumor invasion. We have previously identified hydroxamic acid-containing analogues that are potent inhibitors of the APN homologue from the malarial parasite Plasmodium falciparum M1 aminopeptidase (PfA-M1). Herein, we describe the rationale that underpins the repurposing of PfA-M1 inhibitors as novel APN inhibitors. A series of novel hydroxamic acid analogues were developed using a structure-based design approach and evaluated their inhibition activities against APN. N-(2-(Hydroxyamino)-2-oxo-1-(3′,4′,5′-trifluoro-[1,1′-biphenyl]-4-yl)ethyl)-4-(methylsulfonamido)benzamide (6ad) proved to be an extremely potent inhibitor of APN activity in vitro, selective against other zinc-dependent enzymes such as matrix metalloproteases, and possessed limited cytotoxicity against Ad293 cells and favorable physicochemical and metabolic stability properties. The combined results indicate that compound 6ad may be a useful lead for the development of anticancer agents.

Full text of DOI:10.1021/acs.jmedchem.9b00757

Subscriber access provided by UNIV AUTONOMA DE COAHUILA UADEC
Article
Novel Human Aminopeptidase N Inhibitors: Discovery
and Optimization of Subsite Binding Interactions
Jisook Lee, Natalie Vinh, Nyssa Drinkwater, Wei Yang, Komagal Kannan Sivaraman,
Luke Steven Schembri, Michelle Gazdik, Peter Grin, Georgina Butler, Christopher
M Overall, Susan A. Charman, Sheena McGowan, and Peter J. Scammells
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00757 • Publication Date (Web): 28 Jun 2019
Downloaded from http://pubs.acs.org on June 28, 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 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Novel Human Aminopeptidase N Inhibitors: Discovery and
Optimization of Subsite Binding Interactions
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
†
†
‡
‡
Jisook Lee , Natalie B. Vinh , Nyssa Drinkwater , Wei Yang , Komagal Kannan
‡
†
†
∇⊥
Sivaraman , Luke S. Schembri , Michelle Gazdik , Peter M. Grin , Georgina S.
Butler∇⊥O, Christopher M. Overall ∇⊥O, Susan A. Charman , Sheena McGowan* ,
§
‡
Peter J. Scammells*†
†
Medicinal Chemistry, Monash Institute of Pharmaceutical Sciences, Monash University, Parkville
Campus, VIC 3052, Australia.
‡
Infection and Immunity Program, Biomedicine Discovery Institute and Department of Microbiology,
Monash University, Clayton Campus, VIC 3800, Australia.
§
Centre for Drug Candidate Optimisation, Monash Institute of Pharmaceutical Sciences, Monash
University, Parkville Campus, VIC 3052, Australia
∇
O
Department of Biochemistry & Molecular Biology and Department of Oral Biological & Medical
Sciences, University of British Columbia, Vancouver, BC V6T 1Z3, Canada
Centre for Blood Research, Life Sciences Institute, University of British Columbia, Vancouver, BC
V6T 1Z3, Canada
⊥
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 88
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
ABSTRACT: Aminopeptidase N (APN/CD13) is a zinc-dependent M1 aminopeptidase that contributes
to cancer progression by promoting angiogenesis, metastasis and tumor invasion. We have previously
identified hydroxamic acid containing analogues that are potent inhibitors of the APN homologue from
the malarial parasite, Plasmodium falciparum M1 aminopeptidase (PfA-M1). Herein we describe the
rationale which underpins the repurposing of PfA-M1 inhibitors as novel APN inhibitors. A series of
novel hydroxamic acid analogues were developed using a structure-based design approach and evaluated
for their inhibition activities against APN. N-(2-(Hydroxyamino)-2-oxo-1-(3',4',5'-trifluoro-[1,1'-
biphenyl]-4-yl)ethyl)-4-(methylsulfonamido)benzamide (6ad) proved to be an extremely potent inhibitor
of APN activity in vitro, selective against other zinc dependent enzymes such as matrix metalloproteases,
and possessed limited cytotoxicity against Ad293 cells and favorable physicochemical and metabolic
stability properties. The combined results indicate that compound 6ad may be a useful lead for the
development of anti-cancer agents.
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
ACS Paragon Plus Environment

Page 3 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
 INTRODUCTION
Aminopeptidase N (APN/CD13; EC 3.4.11.2) is an ubiquitous transmembrane ectoenzyme that is widely
present in different types of cells including renal, intestinal, fibroblast, endothelial and tumor cells.1, 2
APN is described as a “moonlighting” enzyme due to its multi-functional roles: an enzyme to cleave
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
3
peptide substrates, a receptor, and a signaling molecule. The enzyme cleaves hydrophobic and basic
4
amino acid residues from the N-terminus of polypeptides with broad substrate specificity. For example,
APN catalyses the metabolism of angiotensin III to generate angiotensin IV to regulate the renin-
5
6, 7
angiotensin system and levels of neuropeptides such as enkephalins. APN also acts as a viral receptor
for mammalian coronavirus, and is a signaling molecule in phagocytosis, angiogenesis, and cell
adhesion.8-13
APN is a member of the zinc-dependent M1 aminopeptidase superfamily of enzymes (protease clan
4
MA) that can be found in all kingdoms of life except viruses. M1 aminopeptidases are characterized by
a thermolysin fold and two consensus sequence motifs; a HEXXHX E zinc binding motif and a GXMEN
1
8
substrate-guiding motif.4, 14-16 Wong et al. reported the X-ray crystal structure of human APN, as well
as structure of APN bound to generic inhibitors bestatin and amastatin, and an endogeneous peptide
4
substrate, Angiotensin IV. Human APN consists of a short intracellular tail, a transmembrane anchor,
a small serine/threonine-rich extracellular stalk and a large ectodomain, comprised of four domains (I –
IV) characteristic of M1 aminopeptidase superfamily.4, 14 The catalytic domain II contains the consensus
motifs 352GXMEN356 and 388HEXXHX E , the latter of which includes the catalytic triad His , His
411
388
392
1
8
and Glu411 that coordinates the essential zinc ion.4
APN has been extensively studied due to its significant role in the regulation of metastasis and
17
angiogenesis. A significant body of evidence supports the rationale that APN is an effective therapeutic
target for malignancies.17-20 Dysregulated activity of APN has been found to develop into a wide
spectrum of human malignancies.21-27 Metastasis is a complex multistep process of cell migration, cell
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 88
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
invasion, and angiogenesis and it is the major cause of cancer related deaths worldwide.28-31 Multiple
studies have shown that APN activity is involved in extracellular matrix (ECM) degradation, an essential
step for metastasis, which was later determined to increase tumour cell migration and invasion by
stimulating MAPK/PI3K signalling cascade.24, 32, 33 Thus, there is an on-going interest to develop potent
APN inhibitors as effective anti-cancer drug candidates.
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 variety of APN inhibitors have been developed as potential anti-cancer candidates.18, 19, 34, 35
Among them, a natural peptidomimetic, bestatin, is the most widely studied competitive APN inhibitor.
Originally isolated from Streptomyces olivoreticuli as an immunomodulating agent, bestatin was found
to have anti-tumor activity36, 37 as well as clinical efficacy against acute myeloid leukemia and lung
cancer in clinical trials.38-41 Another APN inhibitor Tosedostat (CHR2797) is an orally bioavailable
prodrug that is converted to a pharmacologically active drug (CHR79888) inside cells.42 Tosedostat
demonstrated significant anti-leukemic activity in phase II clinical trials in elderly or relapsing patients
with acute myeloid leukaemia.43
Previous work by our group generated a series of hydroxamic acid-containing compounds that were
inhibitors of the Plasmodium falciparum M1 aminopeptidase, PfA-M1.44-46 We described N-(2-
(hydroxyamino)-2-oxo-1-[3',4',5'-trifluoro(1,1'-biphenyl)-4-yl]ethyl)pivalamide (1) (Figure 1) as a
potent inhibitor of PfA-M1, exhibiting an inhibitory constant (Ki(app)) in the nanomolar range.44 Here we
show that 1 is also active towards APN and is more potent than both bestatin and Tosedostat. We have
repurposed compound 1 as a novel APN inhibitor and developed a new series of analogues with improved
potency against APN through structure-based approaches.
ACS Paragon Plus Environment

Page 5 of 88
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
1.
Structure
of
N-(2-(hydroxyamino)-2-oxo-1-[3',4',5'-trifluoro(1,1'-biphenyl)-4-
yl]ethyl)pivalamide (1).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 88
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
 RESULTS AND DISCUSSION
Compound 1 can inhibit recombinant human APN. PfA-M1 is a homologue of APN found in
Plasmodium falciparum which is one of the parasites that causes malaria. Being part of the same M1
aminopeptidase family, PfA-M1 shares a number of structural similarities with APN, particularly within
the catalytic domain II. Overall, PfA-M1 and APN share 19% sequence identity
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
(35% similarity), however, in the highly conserved catalytic domain
II, this sequence identity increases to 24 % (43% similarity). This
conserved catalytic domain II adopts a thermolysin-like fold and in PfA-M1, contains a
4
96
519
490
16
H EYFHX KE zinc-binding motif as well as a G AMEN substrate-guiding motif. The catalytic
1
7
zinc in the active site is coordinated by a catalytic triad His , His , and Glu in the unbound state.16
Therefore, we were interested to see how the potent PfA-M1 inhibitor 1 would interact with human APN.
Our standard fluorescence-based aminopeptidase inhibition assay was used to measure the inhibitory
activity of bestatin, Tosedostat and 1 against human APN. This assay used recombinant human APN
and a commercially available fluorophore 7-amino-4-methylcoumarin as the competitive substrate to
determine an inhibitory constant (Ki(app)). We compared the inhibitory activity of bestatin, Tosedostat
and 1 against human APN as well as PfA-M1 (Table 1). Bestatin showed a moderate loss in potency
toward APN compared to PfA-M1, whereas Tosedostat exhibited a 6-fold improved potency toward APN
when compared to PfA-M1. Interestingly, compound 1 was significantly more potent than Tosedostat
and bestatin, displaying a 10 to 20-fold increase in APN inhibition activity.
4
96
500
519
Table 1: Ki(app) comparison of bestatin, Tosedostat, and compound 1 against human APN and PfA-M1.
Compound
Ki(app)(APN) ± SEM (nM)
Ki(app)(PfA-M1) ± SEM (nM)
Bestatin
Tosedostat
1
2370 ± 350
1180 ± 8
118 ± 3
1530 ± 58
6150 ± 275
331 ± 12
ACS Paragon Plus Environment

Page 7 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In order to understand the mechanism by which 1 was able to inhibit both APN and PfA-M1, we
investigated the interactions the compound made with the active site of the enzymes. To do this, we used
4
4
the X-ray crystal structure of PfA-M1 bound to compound 1 (PDB ID 4ZX4) as a scaffold to dock 1
4
into the active site of APN (PDB 4FYQ). The catalytic domains of the two proteins share 24 % sequence
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
identity and have an RMSD of only 1.244 Å (over 257 C atoms in domain II). The co-crystal structure
of 1 bound to PfA-M1 revealed extensive water-mediated interactions of the 3,4,5-trifluorophenyl ring
with backbone residues located at the S1 substrate binding pocket, as well as key hydrophobic
interactions with the biaryl system of 1 (Figure 2A).44 Our docking analysis of 1 bound to APN showed
a similar pose to that observed when bound to PfA-M1 (Figure 2B). Each of the poses obtained from
docking 1 were similar and showed the 3,4,5-trifluorophenyl ring in the same position, located deep
within the S1 pocket of APN and minor rotations of the position of the N-pivaloyl group were observed.
In general, there were significantly less interactions observed between 1 and the active site residues of
APN than that of PfA-M1 (Figure 2B).
To evaluate any potential dynamics of 1 bound within the active site of APN as well as any effect
from water-mediated interactions, we performed molecular dynamics (MD) simulations of 1 docked into
APN. Molecular modelling for metallo-proteins presents significant challenges and traditional force-
fields are often not appropriate for simulation.47 In the case of APN, the presence of zinc in the active
site means that this problem cannot be computationally ignored. Recently, our team produced the
necessary parameters to use the zinc Amber force-field (ZAFF) to simulate the active site of PfA-M1.48,
4
9
We used this system to simulate our docking of APN bound to 1. MD simulations (n=3) were
performed for the duration of 50 ns, which should be sufficient to observe the movements of small
molecules. The results were surprising and showed that in two of the three MD runs, 1 experienced
significant movement within the active site (Figure S2). Investigation into the motion of 1 indicated that
the 3,4,5-trifluorophenyl ring engaged in interactions with a single water molecule and largely
maintained face-face stacking interactions with relatively rigid Phe472 at the S1 pocket of APN (Figure
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 88
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
2C). In addition, the pivaloyl group was facing the aromatic sidechain of Tyr477 residue. We
hypothesised therefore that compounds that are capable of improved interactions in the S1ʹ subsite, as
well as extend further to engage residues beyond the S1’ pocket of APN may contribute to improved
4
77
381
inhibitory activity toward APN. To this end, we turned our attention to residues Tyr , Arg , and
Arg442 that are located within the S1′ and beyond the S1′ pocket of APN, and that may allow the formation
of hydrophobic interactions as well as polar interactions with inhibitor compound(s).
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. (A) X-ray crystal structure of 1 bound to PfA-M1 (PDB ID 4ZX4). (B) A predicted binding
pose of 1 bound to APN obtained from docking (PDB ID 4FYQ). (C) MD simulation of 1 docked to
APN showed that the trifluorophenyl group formed stable interactions in the S1 pocket. The ligand is
shown in magenta and residues of PfA-M1 and APN are coloured in blue and dark green, respectively.
Interactions between the ligand and proteins are depicted as black dashed lines. The zinc is represented
by the grey sphere.
ACS Paragon Plus Environment

Page 9 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Substitution of the N-pivaloyl group to optimize binding at the S1′ subsite of APN. The
molecular modelling revealed the potential to achieve an enhanced binding interactions at, and beyond,
the S1′ pocket of APN by replacing the N-pivaloyl group of 1 with aromatic groups which target the
Tyr477 residue. We designed and produced a series of hydroxamic acid analogues that contained
elongated alkyl and aryl linkers to reach deeper into the pocket (Figure 3). Various hydrogen bonding
groups were also incorporated to capture additional interactions with Arg381 or Arg442 residues and
improve inhibition activities of the designed compounds. In addition, we were interested to investigate
the effect of various heteroaromatic groups, such as indole and indoline. Analogues with benzyl linkers
that contain an additional methylene group were also designed to increase flexibility and allow the phenyl
group to penetrate more deeply beyond the S1′ pocket.
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 3. Structures of targeted hydroxamic acid analogues.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 88
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
Chemistry. The synthesis of the key intermediate, the phenyl glycine derivative 4 (Scheme 1), was
46
adapted from Mistry et al. Installation of the 3,4,5-trifluorophenyl ring was successfully achieved using
a Suzuki coupling reaction between 4´-bromoacetophenone and (3,4,5-trifluorophenyl)boronic acid
under reported conditions producing acetophenone 2 in excellent yield (99%).44 Oxidation of 2 with
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
50
selenium dioxide in anhydrous pyridine afforded the corresponding α-keto acid 3 in quantitative yield.
The product produced from the reductive amination of 3 underwent an acid-catalysed esterification and
subsequent debenzylation to successfully afford key precursor 4 in 47% yield over three steps.
The key intermediate (4) was then used to incorporate a range of functionalities in place of the N-
pivaloyl group present in 1, which was predominately achieved with traditional peptide coupling reagents
such as HCTU or EDCI. The first synthetic attempt to obtain benzamide
analogues 5r and 5s involved in the synthesis of their respective acid
chlorides in situ, then subsequent acylation with intermediate 4.
However, benzonitriles 5n and 5o were produced instead by dehydration
of the carboxamide group. Ring-opening reactions of cyclic acid anhydrides were also
performed to synthesise alkyl carboxylate analogues. Cyclic anhydrides such as succinic anhydride and
Meldrum’s acid have been commonly used in literature to form amide bonds through ring-opening
reactions.51-59 The butyric acid analogue (5e) was synthesised from intermediate 4 via nucleophilic attack
on the carbonyl π system present in succinic anhydride. Subsequent PyBOP amide coupling of 5e
converted the acid moiety to the corresponding carboxamide 5f. However, reaction with Meldrum’s acid
under identical reaction conditions resulted in decarboxylation, generating the acetamide analogue (5a)
instead of the expected propionic analogue.
ACS Paragon Plus Environment

Page 11 of 88
1
Journal of Medicinal Chemistry
Scheme 1. Synthesis of hydroxamic acid inhibitorsa
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
F
F
F
F
F
F
F
F
F
F
F
O
F
F
Br
(
a)
(b)
(c)
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
HO
HO
O
O
O
NHBn
NHBn
O
O
O
2
3
F
F
F
F
F
F
F
F
F
F
F
(
d) or (e)
or (f)
(
h) or (i)
(g)
O
O
O
H
N
H
N
O
O
HO
N
H
R1
HO
N
R1
N
R1
NH2
H
H
O
O
O
O
6ag, 6ar,6av
6
5
4
6
ax, 6az
a
b
c
d
e
f
-CH3
r
s
t
-C H C(O)NH (m)
aj -indol-5-yl
6
4
2
t
-CH C(O)O Bu
-C H C(O)NH (p)
ak -indazol-5-yl
2
6
4
2
-CH C(O)OH (6c from 5b)
-C H C(O)NHMe (p)
al -benzotriazol-5-yl
2
6
4
-CH C(O)NHOH (6d from 5b)
u
v
w
x
y
z
-C H C(O)NMe (p)
am -benzimidazol-5-yl
an -(2-oxoindolin-5-yl)
2
6
4
2
-CH CH C(O)OH
-C H C(O)NHEt (p)
6 4
2
2
i
-CH CH C(O)NH
-C H C(O)NHPr (p)
6 4
2
2
2
-C H C(O)NHCH CH OH (p)
ao -CH C H
2 6 5
6
4
2
2
g
h
i
-C H
-C H SO NH (p)
ap -CH C H F (m)
2 6 4
6
5
6
4
2
2
-C H F (m)
-C H NHBoc (m)
aq -CH C H F (p)
2 6 4
6
4
6
4
-C H F (p)
aa -C H NHBoc (p)
ar -CH C H OMe (p)
2 6 4
6
4
6
4
j
-C H OMe (m)
ab -C H NH (m) (6ab from 5z)
as -CH C H OH (p) (6as from 6ar)
2 6 4
6
4
6
4
2
k
l
-C H OMe (p)
ac -C H NH (p) (6ac from 5aa)
at -CH C H C(O)NH (p)
2 6 4 2
6
4
6
4
2
-C H OH (m) (6l from 6j)
ad -C H NHSO Me (p)
au -CH C H (3,4-OMe)
2 6 3
6
4
6
4
2
m -C H OH (p) (6m from 6k)
ae -C H NHSO NH (p)
av -CH C H (3,4-OH) (6av from 6au)
2 6 3
6
4
6
4
2
2
n
o
p
q
-C H CN (m)
af -C H (3,4-OMe)
aw -CH C H (3-F,4-OMe)
2 6 3
6
4
6
3
-C H CN (p)
ag -C H (3,4-OH) (6ag from 6af)
ax -CH C H (3-F,4-OH) (6ax from 6aw)
2 6 3
6
4
6
3
-C H C(NHOH)NH (m) (6p from 5n) ah -C H (3-F, 4-OH)
ay -CH C H (4-F,3-OMe)
2 6 3
6
4
2
6
3
-C H C(NHOH)NH (p) (6q from 5o)
ai -C H (4-F, 3-OH)
az -CH C H (4-F,3-OH) (6az from 6ay)
2 6 3
6
4
2
6
3
a
Reagents and conditions: (a) 3,4,5-trifluorophenylboronic acid, Pd(PPh ) Cl , 1 M Na CO , THF,
reflux, 2 h; (b) SeO , anhyd. pyridine, reflux, 24 h; (c) (i) benzylamine, Na(OAc) BH, DCE, rt, 24 h; (ii)
conc. H SO , MeOH, reflux, 24 h; (iii) H , 10% Pd/C, cat. HCl MeOH, rt, 24 h; (d) (i) carbamoylbenzoic
acid, (COCl) , DMF, DCM, rt, 1 h; (ii) 4, DIPEA, DCM, rt, 30 min; (e) carboxylic acid, HCTU, DIPEA,
DMF, DCM, rt, 24 h; (f) carboxylic acid, EDCI, DMAP, DCM, rt, 24 h; (g) NH OH.HCl, 5 M KOH in
anhyd. MeOH, rt; (h) 20 % TFA in DCM, rt; (i)1 M BBr in DCM, -78 ºC to rt, 2-24 h.
3 2
2
2
3
2
3
2
4
2
2
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 88
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 methyl ester in 5 was converted to the corresponding hydroxamic acids (6) using hydroxylamine
hydrochloride and methanolic potassium hydroxide. In some cases, minor conversion to the carboxylic
acid through base-mediated hydrolysis was detected. Occasionally, complete conversion to the desired
product was not successful due to unexpected side reactions. For instance, LC-MS analysis of the tert-
butyl propanoate analogue (5b) indicated only partial conversion. Interestingly, when more of the
reagents were added to the reaction mixture, condensation of both methyl and tert-butyl esters occurred,
generating the desired mono-hydroxamic acid 6b and dihydroxamic acid analogue 6d. For the
benzonitrile compounds 5n and 5o, nucleophilic attack of hydroxylamine on the electrophilic nitrile
carbon resulted in the formation of amidoxime compounds 6p and 6q, respectively.60, 61 The APN
inhibitory activity of these unintended analogues was still evaluated due to their potential ability to form
hydrogen bonds or ionic interactions at the S1′ pocket. Compounds with acidic functionality (5c, and
phenolic analogues) showed poor solubility in methanol when deprotonated by potassium hydroxide,
which consequently resulted in a longer reaction times and poor reaction yields.
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
Further deprotection reactions were required in order to produce carboxylic acid 6c, anilines 6ab
and 6ac and phenols 6ag, 6as, 6av, 6ax and 6az. The tert-butyl propanoate analogue (5b) and Boc-
protected aniline analogues (5z, 5aa) were hydrolysed under mild acidic conditions to give propionic
acid 6c and anilines 6ab and 6ac, respectively. O-Demethylation was carried out in the presence of
boron tribromide to form the corresponding phenols / catechols 6ag, 6as, 6av, 6ax and 6az.
In addition to the analogues with an amide linker, a benzylamine analogue (9) was also synthesised
(Scheme 2). Further physicochemical studies on the parent compound (1) indicate it has an aqueous
4
4
solubility ranging between 12 – 25 μM and a LogD of 3.0. This study demonstrated that there was room
to improve its physicochemical properties for optimal pharmacokinetic profiles. Incorporating bulkier
groups inevitably increases the hydrophobicity of molecules, consequently decreasing aqueous
solubility. However, the introduction of a secondary amine would allow the molecule to be formulated
as a salt to overcome solubility issues. As described previously, reductive amination of intermediate 4
ACS Paragon Plus Environment

Page 13 of 88
1
Journal of Medicinal Chemistry
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
with 4-fluoro-3-methoxybenzylaldehyde yielded 7, which was converted to the corresponding
hydroxamic acid (8). The methoxy group in 8 was readily deprotected using boron-tribromide, then
reacted with hydrochloric acid to obtain the 4-fluoro-3-hydroxybenzylamine analogue (9) as a
hydrochloride salt.
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
Scheme 2. Synthesis of benzylamine analogue 9a
F
F
F
F
F
F
F
F
F
F
F
F
(a)
(b)
(c)
Cl
H
N
H
N
O
O
OCH3
F
OCH3
F
OH
F
NH2
N
HO
N
HO
N
H
H
H
2
O
O
O
O
4
7
8
9
a
Reagents and conditions: (a) benzaldehyde, Na(OAc) BH, anhyd. DCE, rt, 24 h; (b) NH OH.HCl, 5
3
2
M KOH in anhyd. MeOH, rt; (c) (i) 1 M BBr in DCM, -78 ºC to rt, 2-24 h; (ii) 1 M HCl, MeOH, rt, 24
3
h;
A substituted aromatic group is important for potency toward human APN. The inhibitory
activity (Ki(app)) of the synthesised hydroxamic acids were measured against recombinant human APN
(Table 2). Initial experimental triplicates were performed for all compounds to determine compounds of
higher priority with Ki(app) value of ≤ 100 nM, which were assessed further in biological triplicate. In
general, the aliphatic carboxamides (6a-f) were less potent than the lead compound (1), but more potent
than bestatin and Tosedostat (Table 2). This suggests that the appended carboxylic acid and carboxamide
moieties (in the case of 6a, 6e and 6f) were unable to make the intended polar interactions with the S1′
subsite of APN, potentially due to the short linker length. The unsubstituted benzamide 6g showed over
4
-fold loss in inhibitory activity compared to 1. However, introduction of hydrogen bond donating
groups to the phenyl ring led to a recovery in potency. For example, hydroxyl, amidoxime and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 88
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
carboxamide mono-substituted benzamides 6m-s were all more potent than compound 1. On the other
hand, hydrogen bond acceptors such as fluorine (6h and 6i) and methoxy (6j and 6k) were not well
tolerated. We also examined eight functionalized carboxamides (6t-x), sulfonamides (6y and 6ad) and
sulfamide (para-(sulfamoylamino)benzoic acid) (6ae) analogues which have the capacity to increase
ligand binding interactions in the S1′ pocket via hydrogen bonding interactions. The benzamide
derivatives 6t-x exhibited a decrease in activity ranging from 2 to 8-fold compared to carboxamide 6s.
A trend of decreasing activity was observed from the benzamide derivatives as the size of hydrophobic
group increases from methyl, dimethyl, ethyl and isopropyl, indicating that a loss in polar contacts may
result in reduced binding to APN. Additionally, the inhibitory activity of the methyl (6t) and dimethyl
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
(6u) analogues were essentially identical, suggesting the hydrogen bond donating capability of
benzamide does not play a significant role in potency. 4-Methylsulfonamide 6ad (Ki(app) = 4.5 nM) and
sulfamide 6ae (Ki(app) = 8.2 nM) were the most potent inhibitors of the series and showed a greater than
1
0-fold improvement in potency compared to 1, potentially due to its multiple hydrogen bond forming
capacity for strong binding interactions with APN.
Among the di-substituted benzamides (6af-ai), the fluorohydroxyl analogues (6ah and 6ai) showed
improved activity relative to 1. When compared with the corresponding mono-substituted benzamide
derivatives, di-substituted analogues showed stronger inhibition activity.
For example, 3-
fluorobenzamide 6h and 4-hydroxybenzamide 6m displayed activities of 919 nM and 366 nM
respectively, whereas the potency of the 3-fluoro-4-hydroxyl analogue (6ah) significantly increased to
2
9 nM, suggesting both fluoro and hydroxyl are making important interactions with APN. A similar
result was observed for the 4-fluoro-3-hydroxy analogue (6ai), which exhibited significantly greater
potency (Ki(app) = 40 nM) than the corresponding mono-substituted analogues; 4-fluorobenzamide 6i
(
Ki(app) = 704 nM) and 3-hydroxybenzamide 6l (K (app) = 102 nM). Similarly, the 3,4-dimethoxy analogue
i
(6af) exhibited an increased potency compared to 3- and 4-methoxy analogues, 6j and 6k, respectively.
ACS Paragon Plus Environment

Page 15 of 88
Journal of Medicinal Chemistry
1
2
3
4
The activities of compounds where benzamide was replaced with various heterocyclic amides (6aj-
an) was also investigated. The indole analogue (6aj) was equipotent with the parent compound (1). A
significant increase in the potency was observed from indazole 6ak (Ki(app) = 19.2 nM) and benzotriazole
6al (Ki(app) = 23.4 nM).
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
2
In general, replacement of benzamide to acetamide to introduce sp characteristics for increased
flexibility resulted into a significant loss in potency (6ao-ay). The 3-fluoro-4-hydroxyl analog 6ax was
the only compound in this series that proved to be more potent than 1 with inhibition activity of 43.1 nM.
Inconsistent relationships were observed between the benzamide and matching acetamide analogues. For
instance, the potency observed for the benzyl analogue (6ao) reduced drastically compared to the phenyl
analogue (6g). This trend was also observed for the 4-methoxy (6k and 6ar), 4-hydroxy (6m and 6as),
4-fluoro-3-hydroxy (6ai and 6az), and 4-benzamide (6s and 6at) pairs. In contrast, an increase in potency
was observed for the fluoro analogues (6h, 6i, 6ap, 6aq), and 3,4-dimethoxy (6af and 6au), while 3-
fluoro-4-hydroxyl compounds 6ah and 6ax were the most potent compounds in each series. The 4-
fluoro-3-hydroxybenzylamine analogue (9) was the weakest inhibitor of this series; full inhibition was
not achieved at a concentration of 1 mM, suggesting that carbonyl oxygen of the amide-linker is crucial
for potent activity, potentially by providing appropriate rigidity and also hydrogen bonding interaction
with nearby residues.
SAR investigations around the S1′ pocket of APN through modifications of the N-pivaloyl group of
compound 1 indicated that substituted aryls were essential for inhibitory activity, where hydrogen bond
donors were more favored than hydrogen bond acceptors. A substantial decrease in activity observed
from acetamide and secondary amine analogues revealed that rigidity and the carbonyl oxygen of the
benzamide were vital. In addition, studies on functionalized carboxamides and sulfonamides showed
that there was a decreasing trend in activity as the size of the hydrophobic group increases while
incorporating polar groups led to a significant rise in activity. This suggested that hydrogen bonding
interactions played a major role to enhance activity against APN.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 88
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
Table 2. Summary of inhibitory activity of hydroxamic acid analogues against APN.
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
Compound
Bestatin
Tosedostat
1
R
-
-
Ki(app) ± SEM (nM)
2370 ± 350
1180 ± 77
118 ± 3
-C(CH3)3
-CH3
-CH COOH
6
a
560 ± 50
188 ± 9
6
c
2
6
6
d
e
-CH C(O)NHOH
348 ± 33
172 ± 9
2
-CH CH COOH
2
2
6
f
-CH CH C(O)NH
2
497 ± 35
522 ± 37
919 ± 160
704 ± 68
462 ± 6
2
2
6
g
-C H
6 5
6
h
-C H F (m)
6 4
6
6
i
j
-C H F (p)
6 4
-C H OMe (m)
6 4
6k
-C H OMe (p)
-C H OH (m)
6 4
-C H OH (p)
6 4
745 ± 53
366 ± 31
102 ± 5
37.3 ± 2.9
49.1 ± 3.7
71.2 ± 6.5
82.1 ± 9.8
185 ± 22
182 ± 5
6
4
6
l
6
m
*
6p
-C H C(NHOH)NH (m)
6
4
2
*
6
q
-C H C(NHOH)NH (p)
6 4 2
6
r
-C H C(O)NH (m)
6 4 2
*
6s
-C H C(O)NH (p)
6 4 2
6
t
-C H C(O)NHMe (p)
6 4
6
u
-C H C(O)NMe (p)
6 4 2
6v
-C H C(O)NEt (p)
-C H C(O)NH Pr (p)
6 4
277 ± 22
631 ± 70
163 ± 26
240 ± 7
131 ± 10
205 ± 17
4.50 ± 0.80
8.20 ± 0.90
175 ± 16
430 ± 42
29.1 ± 3.6
6
4
i
6
w
6
x
-C H C(O)NHCH CH OH (p)
6 4 2 2
6y
ab
-C H SO NH (p)
6 4 2 2
6
-C H NH (m)
6
4
2
6
ac
-C H NH (p)
6 4 2
*
6ad
-C H NHSO Me (p)
6 4 2
*
6
ae
af
6ag
ah
-C H CNHSO NH (p)
6 4 2 2
6
-C H (3,4-OMe)
6 3
C H (3,4-OH)
-C H (3-F,4-OH)
6 3
6
3
*
6
ACS Paragon Plus Environment

Page 17 of 88
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
*
6ai
aj
ak
6al
am
-C H (4-F,3-OH)
40.0 ± 2.2
111 ± 3
19.2 ± 2.5
23.4 ± 2.3
170 ± 17
156 ± 9
4420 ± 720
442 ± 36
158 ± 5
978 ± 120
235 ± 24
604 ± 66
119 ± 12
137 ± 6
6
3
6
6
-indol-5-yl
-indazol-5yl*
-benzotriazol-5-yl*
-benzimidazol-5-yl*
6
6
an
-(2-oxoindolin-5-yl)
-CH C H
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
6ao
2
6
5
6
6
ap
aq
-CH C H F (m)
2 6 4
-CH C H F (p)
2
6
4
6ar
as
at
6au
-CH C H OMe (p)
-CH C H OH (p)
2 6 4
2
6
4
6
6
-CH C H C(O)NH (p)
2 6 4 2
-CH C H (3,4-OMe)
2
6
3
6
6
av
ax
-CH C H (3,4-OH)
2 6 3
-CH C H (3-F,4-OH)
-CH C H (4-F,3-OH)
2 6 3
43.1 ± 5.0
138 ± 5
>1 mM
2
6
3
6
az
a
9
-CH C H (4-F,3-OH)
2
6
3
*
(app)
Biological triplicates were performed for inhibitors with K
screenings. The compound is secondary amine derivative.
less than 100 nM from initial triage
i
a
The core biaryl system engages in hydrophobic interactions with flexible loop at the S1 pocket
of APN. To understand why 6ad and 6ae displayed such an improvement in potency, we undertook
molecular docking followed by MD simulations to generate a model of APN bound to both inhibitors.
A rigid-docking into the APN crystal structure (PDB ID 4FYQ) of compounds 6ad and 6ae was
performed based on the docked pose of compound 1 into APN using Surflex Docking software available
from Sybyl 2.1. The common fragments of the biphenyl core structure and the hydroxamic group were
set as constraints. Similar to the molecular docking of compound 1, the most preferred structures of 6ad
and 6ae bound to APN were selected based on the total docking score (Table S6). MD simulations
showed that the ligands occupy different conformations in a time-
dependent manner (Supp Movie 1 and 2). Compound 6ad participated in
water-mediated hydrogen bonding interactions with the fluorine atoms
and the nearby residues at the S1 subsite of APN, both of which were
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 88
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
also observed in the co-crystal structure of compound 1 bound to PfA-
M1.44 One compelling result from the MD simulation was a flexible loop
located at domain IV acting as an effective ‘cap’ to close the active
site of APN (Figure 4). The flexible loop and the Phe896 residue are
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
believed to have important roles in conformational changes of APN. The
flexible loop consists of 8 amino acid residues (891YGGGSFSF898) and has
been shown to undergo a dramatic change in conformation upon complex
formation with a peptide substrate.4
The binding mode of bestatin,
which binds APN differently to other M1 aminopeptidases (non-canonical
binding pose), is also thought to be related to the flexible loop, in
particular Phe896
.
As opposed to other M1-bestatin complexes in which
the loop is shorter and different in sequence and therefore does not
impede bestatin from binding with canonical geometry, the Phe896 of APN
was positioned away from binding pocket to accommodate the bulky phenyl
group of bestatin at the S1 pocket.4 In our simulations, we saw that
the biphenyl ring system of 6ad maintained face to face stacking
interactions with Phe896 to keep the inhibitor locked in the active site
of APN, however, it was clear that this interaction varied within the
biphenyl system between different poses. Generally, the sidechain of
Phe896 interacted with the top trifluorophenyl (Figure 4B and C), but
interactions with the central aromatic ring of compound 6ad were also
observed (Figure 4A).
Interestingly, stable hydrophobic stacking interactions between the flexible loop and compound 6ae
were not observed (Figure 4). Without the flexible loop stabilising the biaryl core moiety and restricting
the movement of the molecule the flat biphenyl system of 6ae gained freedom to move within S1 pocket.
ACS Paragon Plus Environment

Page 19 of 88
1
Journal of Medicinal Chemistry
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
Although prominent stacking interactions that were observed with 6ad were missing from 6ae; the
phenylalanine rich region provided a favorable environment for the hydrophobic biaryl group to maintain
its overall position within the pocket.
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 4. Three conformations of compound 6ad (A, B and C) observed during MD simulation. The
carbon atoms of the flexible loop is coloured in light pink, and other residues in the binding site are
coloured in wheat. Interactions between compound 6ad (carbon atoms in white) and APN are shown in
black dashed lines. (D) Overlay of two predicted binding poses of compound 6ae. Residues around
compound 6ae (carbon atoms in magenta) are coloured in wheat (PDB ID 4FYQ). Overlaid Arg442
residues are shown in green. Interactions between the compound and APN are depicted in black dashed
lines.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 88
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
Sulfonamide moiety provides multiple hydrogen bonding interactions at the S1′ pocket of
APN. Based on our MD simulations, the large arylsulfonamide derivatives of both 6ad and 6ae were
able to reach the S1ʹ pocket of APN. The bottom aromatic group of 6ad was flexible enough to interact
with the sidechain of Phe898 or Tyr477 (Figure 4A, B, and C). Due to the high flexibility of the sulfonamide
moiety, these interactions were occasionally lost and replaced with a cation–π interaction between the
Arg381 sidechain and the electron-rich aromatic ring of arylsulfonamide. The results also revealed that
the sulfonyl oxygen atoms can engage in dual hydrogen bonds with Asn900 and Arg442 (Table S3), where
the arginine residue stayed in a relatively rigid manner throughout simulations.
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
Compound 6ae behaved similarly to 6ad, exhibiting multiple binding positions. The aromatic ring
located in the S1ʹ pocket of APN displayed extensive interactions with the side chain of Arg442 and
9
00
Asn . As described above, the flexible loop was not accessible to the molecule, but the bottom aromatic
477
was located closer to Tyr , forming hydrophobic edge-face stacking interactions (Figure 4D). The
backbone amide of Ala474 could also participate in a hydrogen bond interaction with a sulfonyl oxygen
atom of 6ae (Table S4). In the case where the sulfamide was pointing deeper into the S1′ subsite of APN,
the amino group is located in close proximity to Asp439.
In order to rationalise our SAR analysis, we also investigated the interactions between one of the
weaker inhibitors, 6ag (Ki(app) = 430 nM) which contains 3,4-hydroxyl group (Figure 5). In contrast to
compounds 6ad and 6ae which showed extensive hydrophobic interactions with the flexible loop, the
biphenyl system of compound 6ag made no interactions with the key phenylalanine residues of the
flexible loop. Throughout most of the simulation, the trifluorophenyl moiety participated in a water-
mediated hydrogen bonding network with Ser469 and hydrophobic interactions between Phe . However,
the flexible loop was located too far away from the biphenyl system to capture essential non-polar
interactions, which may explain the significant decrease in the potency. Hydrogen bonding interactions
between the catechol group and the sidechain of Glu418 residue were also observed. However these
interactions appear not able to compensate the missing stacking interactions between the S1′ aromatic
472
ACS Paragon Plus Environment

Page 21 of 88
1
Journal of Medicinal Chemistry
group and Tyr477 or the Phe898 residues, which may contribute to the reduced inhibition activity of 6ag
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
against APN.
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 5. Comparison of the binding interactions occurred with tight binding inhibitor 6ad (left) and the
weaker inhibitor 6ag (right) (4FYQ). The carbon atoms of the flexible loop is coloured in light pink, and
other residues in the binding site are coloured in wheat. Compound 6ad and 6ag are coloured in white
and green, respectively. Interactions between the ligands and APN are shown in black dashed lines.
Role of Glu389 residue as the zinc-hydroxamic acid complex stabiliser. Another intriguing
phenomenon was observed from the formation of the zinc-hydroxamic acid complex. In the M1
aminopeptidases, inhibitors often bind with a pentahedral coordination to the zinc, which mimic the
transition state of the activated enzyme.4, 62 In the simulations we performed, the hydroxyl group of
hydroxamic acid readily lost contact with the zinc, with bond distance ranging from 2.2–3.4 Å while the
carbonyl oxygen possessed a very stable coordination with a bond distance of 1.9–2.0 Å, resulting in
tetrahedral coordination. Given the challenges associated with metallo-protein simulations, this
observation is possibly biased by the ZAFF parameters used. To accurately simulate the change in ligand
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 88
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
interaction with the zinc ion, a quantum mechanics / molecular mechanics (QM/MM) simulation would
be needed. However, we did observe was that the loss of pentahedral coordination of 6ad and 6ae to the
zinc ion resulted in the formation of a hydrogen bond with Glu389 (Figure 6). This interaction was
extremely stable and maintained a bond distance of 2.4 – 2.9 Å throughout our simulations (Table S3 &
S4). Further computational and experimental investigation is required to determine whether this
interaction plays a role in the catalytic mechanism by stabilizing the zinc-ligand complex.
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 6. Hydroxamic acid – zinc complex in the active site of APN. APN residues are coloured in wheat
and the ligand is shown in magenta. Interactions between the ligand and the catalytic triad and Glu389 are
illustrated in black dashed lines (PDB ID 4FYQ).
Compound 6ad and 6ae show selectivity for APN over matrix metalloproteinases (MMPs). As
the hydroxamic acid moiety is a strong zinc chelator there is a possibility that our potent APN inhibitors
may interact with other zinc-dependent enzymes. To assess the selectivity and potential off-targets
effects of 6ad and 6ae, we performed quenched fluorescence assays with MMP2, 7, 8, 9, and 13 (Supp
ACS Paragon Plus Environment

Page 23 of 88
1
Journal of Medicinal Chemistry
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
Figure S1). The activity of the parent compound (1), Tosedostat and the broad spectrum MMP inhibitor,
Marimastat, were also evaluated as comparison and controls (Supp Table 1). MMP2 and 9 were
effectively inhibited by Marimastat (IC50 0.43 and 3.1 nM, respectively)
and weakly by Tosedostat (IC : 0.19 and 1.5 μM, respectively), whereas
50
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
6ad and 6ae weakly inhibited MMP2 (1.3 μM and 2.1 μM respectively) and
only inhibited MMP9 at relatively high concentrations (IC50 >100 μM).
Similar observations were made for collagenases MMP8 and MMP13 as well
as matrilysin MMP7, which were all inhibited with low nM IC50 by
Marimastat, high nM IC50 by Tosedostat, and μM – mM IC50 by the novel
inhibitors, with 1 demonstrating the lowest extent of inhibition (Supp
Figure S1). Collectively, these findings show that the novel APN
inhibitors demonstrate low off-target inhibitory effects on MMPs.
Cellular toxicity, physicochemical and pharmacokinetic properties of 6ad and 6ae. Evaluation
of cellular toxicity, physicochemical and stability properties provide important early stage data to
determine whether or not a compound has the necessary features to be pursued further as a drug
candidate. In our cellular toxicity study, we used Ad293 cell line, which was used to measure cell toxicity
in other reported literature, is derived from the Human Embryonic Kidney 293 (HEK293) cell line but
transfected with a special gene for improved cell adherence to make handling cells easier during cell
cultures and assays.63 The compounds show limited cytotoxicity against Ad293 cells with CC values
5
0
of 41 ± 2µM for 6ad and 149 ± 13 µM for 6ae. In vitro physicochemical properties
and metabolic/plasma stabilities of two potent APN inhibitors 6ad and 6ae were measured (Table 4).
The kinetic solubility of each compound was determined by nephelometry. Both compounds showed
moderate solubility under pH conditions representative of the stomach (pH 2) and upper fasted state
small intestine (pH 6.5) suggesting that solubility could be a factor that limits oral absorption. The
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 88
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
partition coefficients at pH 7.4 (LogDpH 7.4) were estimated using a chromatographic method and found
to be 2.8 and 2.6, respectively. The metabolic stabilities were assessed by incubating compounds in
mouse and human hepatic microsomes at 37 ºC and a protein concentration of 0.4 mg/mL. The
compounds showed minimal degradation and very long half-lives of longer than 4 h in both mouse and
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
human microsomes, resulting in low in vitro intrinsic clearance values (CL < 7 µL/min/mg protein). In
int
vitro plasma stability studies were performed by incubating compounds in human and mouse plasma at
37 °C for up to 4 h. Both compounds displayed minimal loss indicating that they are not readily
susceptible to the action of hydrolytic enzymes present in the plasma.
Table 4. Physicochemical and metabolic properties of 6ad and 6ae.
Compound
Kinetic Solubility
Liver Microsome Stability
Plasma Stability
SolpH 2.0
SolpH 6.5
Mouse t1/2
(min)
Human t1/2
(min)
Mouse t1/2
Human t1/2
(min)
(min)
>1020
>1020
6
ad
ae
12.5-25
6.3-12.5
12.5-25
12.5-25
>250
>250
>1020
6
>250
>250
>1020
ACS Paragon Plus Environment

Page 25 of 88
1
Journal of Medicinal Chemistry
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
 CONCLUSIONS
Rapid metastasis of cancer cells through complex mechanisms of ECM degradation, angiogenesis, cell
invasion and cell adhesion is a major burden in effective cancer therapy. Therefore, continuous
development of novel anti-cancer agents targeting metastasis is urgently required. Being involved in
mechanisms of angiogenesis and metastasis, APN has been broadly studied as a therapeutic target for
cancer. Here we have reported the design, synthesis and biological evaluations of novel APN inhibitors
that were derived from a potent inhibitor of parasitic homologue PfA-M1. Through comprehensive
structure-based design, we were able to generate a small library of novel hydroxamic acid analogues
targeting APN and discover a potent compound 6ad that showed 527-fold improved inhibition activity
than a known APN inhibitor bestatin. Molecular docking and MD simulations highlighted the
significance of the flexible loop in domain VI in providing hydrophobic interactions at the S1 pocket.
The results also revealed the combination of stacking and dual-hydrogen bonding interactions was crucial
to optimize the binding at the S1′ subunit of APN. In addition, cross-activity studies showed that 6ad and
6ae possessed low off-target activity on MMPs. The evaluation of the cellular activities of 6ad and 6ae
against Ad293 cell line also indicated that they displayed limited cytotoxicity. Moreover, 6ad and 6ae
had favorable metabolic and plasma stability in both human and mouse models. However, the solubility
of these compounds is sub-optimal and should be addressed by further medicinal chemistry approaches.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 88
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
 EXPERIMENTAL SECTION
Chemistry. Chemicals and solvents were purchased from standard suppliers and used without
1
19
13
further purification unless otherwise indicated. H NMR, F NMR and C NMR spectra were recorded
on a Bruker Avance Nanobay III 400 MHz Ultrashield Plus spectrometer at 400.13, 376.46 and 100.61
MHz, respectively. NMR experiments were obtained at the temperature of 298 K. Data acquisition and
processing was managed using Topspin software package version 3. Chemical shifts (δ) are recorded in
parts per million with reference to the chemical shift of the deuterated solvent. Coupling constants (J)
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
and carbon-fluorine coupling constants (J ) are recorded in hertz and multiplicities are described as
CF
singlet (s), doublet (d), triplet (t), multiplet (m), doublet of doublets (dd), doublet of triplets (dt), doublet
13
of doublets of doublets (ddd), and broad (br). Overlapped non-equivalent C peaks were identified by
HSQC and HMBC NMR.
Analytical thin-layer chromatography (TLC) was performed on Merck
silica gel 60F254 aluminium-backed plates and were visualised by
fluorescence quenching under UV lamp at 254 nm or by Fe(III)Cl staining
3
for hydroxamic acid compounds. Flash chromatography was performed with
silica gel 60 (particle size 0.040–0.063 µm).
Analytical HPLC was performed using an Agilent 1260 Infinity Analytical HPLC with a Zorbax
Eclipse Plus C18 Rapid Resolution 4.6 × 100 mm, 3.5 µm column. Buffer A: 0.1 % TFA in H O and
2
buffer B: 0.1 % TFA in MeCN were used. Samples were run at a gradient of 5% buffer B/ buffer A (0–9
min) to 100% buffer B (9–10min) at a flow rate of 1 mL/min. Unless otherwise indicated, all compounds
were > 95% by HPLC (254 nm and 214 nm) prior to biological evaluation.
Preparative HPLC was performed using an Agilent 1260 Infinity instrument coupled with a binary
preparative pump and an Agilent 1260 FC-PS fraction collector using Agilent OpenLAB CDS software
(revision C.01.04) and an Altima C8 22 × 250 mm, 5 μM column and a 1260 Infinity diode array detector
ACS Paragon Plus Environment

Page 27 of 88
1
Journal of Medicinal Chemistry
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
VL. The following buffers were used: buffer A: 0.1% TFA in H O and buffer B: 0.1% TFA in MeCN.
2
The sample was run at a gradient of 30% to 100% buffer B over 10 min at a flow rate of 20 mL/min.
LC-MS was performed using system A or B. System A: Agilent 6100 Series Single Quadrupole
instrument coupled to an Agilent 1200 series HPLC instrument fitted with a Luna 120 C8(2) 5 µ 50 ×
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
4
.6 mm column. Samples were run at a flow rate of 0.5 mL/min for 12 min: 5% buffer B/ buffer A (0–
min), 100% buffer B (4–7 min) and 5% buffer B/ buffer A (7–12 min). Mass spectra were obtained
in positive and negative ion modes with a scan range of 100-1000 m/z. UV detection was carried out at
54 nm. System B: Agilent 6120 series Single Quadrupole instrument coupled to an Agilent 1260 series
2
HPLC instrument fitted with a Poroshell 120 EC-C18 50 × 3.0 mm, 2.7 µm column. The following
buffers were used: buffer A, 0.1% formic acid in H O; buffer B, 0.1% formic acid in MeCN. Samples
2
were run at a flow rate of 0.5 mL/min for 5 min: 5% buffer B/ buffer A (0–1 min), 100% buffer B (1–
2
.5 min) and held at this composition until 3.8 min, 5% buffer B/ buffer A (3.8–4 min) and held until 5
min at this composition. Mass spectra were obtained in positive and negative ion modes with a scan
range of 100-1000 m/z. UV detection was carried out at 214 and 254 nm.
HRMS was carried out using an Agilent 6224 TOF LC-MS mass spectrometer coupled to an Agilent
1290 Infinity. All data were acquired and referenced via dual-spray electrospray ionization (ESI) source.
Acquisition was performed using Agilent Mass Hunter Data Acquisition software version B.05.00 Build
5
5
.0.5042.2 and analysis was conducted using Mass Hunter Qualitative Analysis version B.05.00 Build
.0.519.13.
Instant JChem was used for data management; Instant JChem 16.9.12.0, ChemAxon
(http://www.chemaxon.com).
General Procedure A: Amide coupling using HCTU and DIPEA. The carboxylic acid (1.1 eq.)
and HCTU (1.2 eq.) were dissolved in anhydrous DMF (2 mL/mmol) and stirred for 30 min in a N2
flushed microwave vial. DIPEA (2.1eq.) was added dropwise followed by compound 4 (1.0 eq.) in
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 88
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
anhydrous DCM (2 mL/mmol). The reaction mixture was stirred at rt for 1 d. If the reaction did not
reach completion after 1 d, then a further 1.2 eq. HCTU and 2.1 eq. DIPEA was added. After completion,
the reaction mixture was diluted with sat. NaHCO (10 mL) and extracted with DCM (3 × 15 mL). The
3
combined organic layers were washed with water (2 × 10 mL) and brine (15 mL). The organic layer 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
dried over anhydrous Na SO , filtered and concentrated in vacuo. The crude was purified by column
2
4
chromatography using either DCM:MeOH or PE:EtOAc as the eluent.
General Procedure B: Amide coupling using EDCI and DMAP. Compound 4 (1.0 eq.),
carboxylic acid (1.2 eq.), EDCI (1.2 eq.) and DMAP (1.3 eq.) were dissolved in DCM (8 mL/mmol) or
DMF (8 mL/mmol) and stirred at rt overnight. If the reaction did not reach completion, then a further
1
.2 eq. EDCI and 1.3 eq. DMAP was added. The reaction mixture was diluted with sat. NaHCO (10
3
mL) and extracted with DCM (3 × 10 mL). The combined organic layers were washed with a 1 M HCl
solution (15 mL), dried over anhydrous Na SO , filtered and concentrated in vacuo. The crude residue
2
4
was purified by column chromatography using PE:EtOAc as the eluent.
General Procedure C: Direct aminolysis of methyl ester to the hydroxamic acid. To a solution
of the methyl ester (1.0 eq.) in anhydrous MeOH (3 mL/mmol) was added NH OH•HCl (4-10 eq.),
2
followed by KOH (5M in MeOH, 5-10 eq.). The reaction mixture was stirred at rt and was monitored
by LCMS and TLC using an Fe(III)Cl stain. Once the reaction was complete, the suspension was dry-
3
loaded onto silica and purified by column chromatography.
General Procedure D: O-Demethylation using BBr . To a solution of the methyl ether substrate
3
(
1.0 eq.) in DCM (2 mL/mmol) was added BBr (1 M in DCM, 5.0 eq. for mono O-demethylation, 10.0
3
eq. for double O-demethylation) at -78 ºC. The reaction mixture was stirred at rt for 2 h to 1 d. The
reaction was quenched by the addition of a 1 M HCl and stirred vigorously for 10 min. The resulting
precipitate was filtered and purified by preparative HPLC.
1-(3',4',5'-Trifluoro-[1,1'-biphenyl]-4-yl)ethan-1-one (2). To a nitrogen flushed 500 mL round bottom
flask was added 4’-bromoacetophenone (5.00 g, 25.1 mmol), 3,4,5-trifluorophenylboronic acid (5.74 g,
ACS Paragon Plus Environment

Page 29 of 88
1
Journal of Medicinal Chemistry
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
32.7 mmol), anhydrous THF (180 mL) and a 1 M Na CO solution (60 mL). PdCl (PPh ) (529 mg,
2 3 2 3 2
0
.754 mmol) was added and the mixture was heated at reflux for 2 h. The reaction mixture was
concentrated under reduced pressure and extracted with Et O (3 × 50 mL). The combined organic layers
2
were dried over anhydrous Na SO , filtered and concentrated in vacuo. The crude product was purified
2
4
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
by column chromatography (PE:EtOAc 100:0 to 50:50) to afford compound 2 as a yellow-brown solid
1
(
6.29 g, 100%); H NMR (d -DMSO) δ 8.03 (d, J = 8.6 Hz, 2H), 7.89 (d, J = 8.5 Hz, 2H), 7.80 (dd, J =
6
1
9
9.6, 6.8 Hz, 2H), 2.62 (s, 3H); F NMR (d -DMSO) δ -134.6 (d, J = 21.7 Hz), -162.1 (dd, J = 21.7/21.7
6
13
Hz); C NMR (d -DMSO) δ 197.5, 150.6 (ddd, J = 247.0/9.8/4.2 Hz), 141.1–141.0 (m), 138.79 (dt,
6
CF
J = 250.5/15.6 Hz), 136.5, 135.6 (dt, J = 12.8/6.4 Hz), 128.8, 127.1, 112.2–111.3 (m), 26.8; m/z MS
CF
CF
+
C H F O [MH] calcd 251.1, found 251.0.
14 10
3
2-Oxo-2-(3',4',5'-trifluoro-[1,1'-biphenyl]-4-yl)acetic acid (3). Compound 2 (6.29 g, 25.1 mmol)
and SeO (4.18 g, 37.7 mmol) were dissolved in anhydrous pyridine (200 mL). The reaction mixture
2
was sonicated and then heated at 110 ºC overnight under nitrogen. Once the reaction was complete, the
mixture was filtered through CeliteTM and the filtrate was concentrated in vacuo. A 1 M HCl solution (20
mL) was added and the compound was extracted with EtOAc (3 × 50 mL). The combined organic layers
were dried over anhydrous Na SO , filtered and concentrated under reduced pressure to afford compound
2
4
1
3 as a brown solid (7.04 g, 100%); H NMR (d -DMSO) δ 8.05–7.93 (m, 4H), 7.84 (dd, J = 9.5/6.7 Hz,
6
19
13
2
H); F NMR (d -DMSO) δ -134.4 (d, J = 21.7 Hz), -161.3 (dd, J = 21.7/21.7 Hz); C NMR (d -DMSO)
6 6
δ 188.0, 165.8, 150.7 (ddd, JCF = 247.2/9.8/4.1 Hz), 142.9–142.8 (m), 139.1 (dt, JCF = 248.4/14.2 Hz),
-
1
35.8–134.5 (m), 131.6, 130.2, 127.7, 116.5–105.3 (m); m/z MS C H F O [M-H] calcd 279.0, found
14
6
3
3
279.0.
Methyl 2-amino-2-(3',4',5'-trifluoro-[1,1'-biphenyl]-4-yl)acetate (4). Compound 3 (7.04 g, 25.1
mmol) and benzylamine (4.12 mL, 37.7 mmol) were dissolved in anhydrous DCE (200 mL) and stirred
for 30 min. Na(OAc) BH (7.99 g, 56.6 mmol) was added and the mixture was stirred at rt overnight.
3
Once the reaction was complete, water (30 mL) was added and the mixture was stirred vigorously for 5
ACS Paragon Plus Environment