Structure-Guided Synthesis and Mechanistic Studies Reveal Sweetspots on Naphthyl Salicyl Hydrazone Scaffold as Non-Nucleosidic Competitive, Reversible Inhibitors of Human Ribonucleotide Reductase

Article abstract of DOI:10.1021/acs.jmedchem.7b00530

Ribonucleotide reductase (RR), an established cancer target, is usually inhibited by antimetabolites, which display multiple cross-reactive effects. Recently, we discovered a naphthyl salicyl acyl hydrazone-based inhibitor (NSAH or E-3a) of human RR (hRR) binding at the catalytic site (C-site) and inhibiting hRR reversibly. We herein report the synthesis and biochemical characterization of 25 distinct analogs. We designed each analog through docking to the C-site of hRR based on our 2.7 ? X-ray crystal structure (PDB ID: 5TUS). Broad tolerance to minor structural variations preserving inhibitory potency is observed. E-3f (82% yield) displayed an in vitro IC₅₀ of 5.3 ± 1.8 μM against hRR, making it the most potent in this series. Kinetic assays reveal that E-3a, E-3c, E-3t, and E-3w bind and inhibit hRR through a reversible and competitive mode. Target selectivity toward the R1 subunit of hRR is established, providing a novel way of inhibition of this crucial enzyme.

Full text of DOI:10.1021/acs.jmedchem.7b00530

Subscriber access provided by University of Florida | Smathers Libraries
Article
Structure-Guided Synthesis and Mechanistic Studies Reveal
Sweetspots on Naphthyl Salicyl Hydrazone Scaffold as Non-Nucleosidic
Competitive, Reversible Inhibitors of Human Ribonucleotide Reductase
Sarah E. Huff, Faiz Ahmad Mohammed, Mu Yang, Prashansa Agrawal, John
Pink, Michael E. Harris, Chris Godfrey Dealwis, and Rajesh Viswanathan
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00530 • Publication Date (Web): 18 Dec 2017
Downloaded from http://pubs.acs.org on December 19, 2017
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 free 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 accessible to all
readers and 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.
Journal of Medicinal Chemistry 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 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structure-Guided Synthesis and Mechanistic Studies Reveal Sweetspots on
Naphthyl Salicyl Hydrazone Scaffold as Non-Nucleosidic Competitive,
Reversible Inhibitors of Human Ribonucleotide Reductase
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
‡
¶
‡
‡
⌠
Sarah E. Huff , Faiz Ahmad Mohammed , Mu Yang , Prashansa Agrawal ; John Pink ; Michael E.
Harris ; Chris G. Dealwis* and Rajesh Viswanathan*, ‡, †
§
,¶,⁑
†
Frank Hovorka Assistant Professor of Chemistry and Scientific Oversight Board Member – Small
Molecule Drug Discovery Core, CWRU, 10900 Euclid Ave, Cleveland, OH 44106.
‡ Department of Chemistry, Case Western Reserve University, College of Arts and Sciences,
Millis Science Center: Rm 216, 2074, Adelbert Road, Cleveland, OH 44106-7078.
¶
Department of Pharmacology, Case Western Reserve University, School of Medicine, 10900 Euclid Ave,
Cleveland, OH 44106.
§ Department of Chemistry, University of Florida, PO Box 117200, Gainseville, FL 32611.
⌠
Case Comprehensive Cancer Center, Case Western Reserve University, School of Medicine, 10900
Euclid Ave, Cleveland, OH 44106.
⁑ Center for Proteomics and the Department of Chemistry, Case Western Reserve University, 10900
Euclid Ave, Cleveland, OH 44106.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 56
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
Ribonucleotide reductase (RR), an established cancer target is usually inhibited by
antimetabolites which display multiple cross-reactive effects. Recently, we discovered a naphthyl
salicyl acyl hydrazone-based inhibitor (NSAH or E-3a) of human RR (hRR) binding at the
catalytic site (C-site) and inhibiting hRR reversibly. We herein report the synthesis and
biochemical characterization of 25 distinct analogs. We designed each analog through docking to
the C-site of hRR based on our 2.7Å X-ray crystal structure (PDB ID: 5TUS). Broad tolerance to
minor structural variations preserving inhibitory potency is observed. E-3f (82% yield) displayed
an in vitro IC50 of 5.3 ± 1.8 µM against hRR, making it the most potent in this series. Kinetic
assays reveal that E-3a, E-3c, E-3t and E-3w bind and inhibit hRR through a reversible and
competitive mode. Target selectivity towards the R1 subunit of hRR is established, providing a
novel way of inhibition of this crucial enzyme.
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
ACS Paragon Plus Environment

Page 3 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Introduction
Ribonucleotide reductase (RR) catalyzes the conversion of ribonucleoside substrates into
deoxyribonucleosides (Figure 1A), the rate-limiting step in the formation of all
deoxyribonucleoside 5'-triphosphates (dNTPs) as building blocks for DNA synthesis and
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
replication. Therefore, RR crucially maintains a balanced nucleotide pool in the cell.
Consequently, RR has been an established target for many proliferative diseases, including
2
cancer. Additionally, RR is a relatively complex cellular target for anti-cancer therapeutic
development due to the fact that inhibition of this critical enzyme is likely to cause cell death even
in normal cells. This presents an additional layer of challenge to develop RR inhibitors that can
target cancer cells selectively.
Foundational knowledge on allosteric regulation of this multi-subunit enzyme has been
extensively probed through a series of studies conducted on prokaryotic and some eukaryotic
1
RRs. Mechanistic details of substrate regulation through allostery also remains a focus of ongoing
3
investigations. Recent studies have provided structural insights underlying allosterically driven
dNTP regulation by RR.1d-e, 4, 5c Though previous studies establish the molecular details of
5
multimerization as a basis for development of novel anticancer agents, there has been little focus
on identifying non-nucleosidic, reversible and competitive inhibitors of RR.
With the exceptions of hydroxyurea, triapine, and related radical scavengers that target the R2
subunit, most small-molecule anticancer agents targeting RR belong to the antimetabolite class of
antineoplastic agents (see Table S1 for a list of FDA-approved drugs). For example gemcitabine,
FDA approved for treatment of breast, ovarian, non-small cell lung and pancreatic cancer, is a
6
nucleoside analog that structurally mimics ribonucleoside diphosphate substrates. Gemcitabine
inactivates the RR enzymatic machinery by covalently and irreversibly modifying an active site
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 56
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
6
nucleophilic Cys sulfhydryl moiety, causing significant abrogation of enzyme function. There is
little distinction if any for the action of gemcitabine between RRs in normal cells versus RRs in
6
cancer cells, resulting in toxic side effects in chemotherapy. Clofarabine and others inhibit hRR
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
through a non-covalent mode; however, in general the antimetabolite class of inhibitors
significantly cross react with essential nucleotide metabolic enzymes causing indiscriminate
toxicity. For example, clofarabine triphosphate causes chain termination of DNA polymerase
activity.7
Given this background, we hypothesized that reversible, non-covalent inhibitors of hRR that
are non-nucleosidic may show reduced cross reactivity, and thereby reduced side effects and
toxicity. In 2015, we reported an integrated approach consisting of in silico rapid throughput
screening of small molecules, in vitro biochemical assays against hRR, cellular growth inhibition
studies against select cancer cell lines, and X-ray crystallography to guide us with future rational
8
design. This integrated approach afforded identification of several new classes of non-nucleosidic
modulators of hRR with in vitro hRR inhibitory potencies in the micromolar range. Detailed X-
ray crystallography revealed the binding of a phthalimide-containing class of modulators
maintaining non-covalent stabilizing interactions at the hexamerization interface of hRR.8,9
4
ACS Paragon Plus Environment

Page 5 of 56
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. A. Conversion of ribonucleoside diphosphates into deoxyribonucleoside diphosphates
by RR and its redox coupled activity with thioredoxin; B. Comparison of lead compound
Naphthyl Salicyl Acyl Hydrazone (NSAH or E-3a) and natural substrate GDP bound to the C-site
1
0
of hRR (PDB ID: 5TUS) ; C. Overall approach leading to identification of novel hRR inhibitor
whose analogs (E-3a-z) are synthesized and biochemically characterized in this report.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 56
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
Subsequent to the 2015 report, we discovered a lead compound, Naphthyl Salicyl
Acylhydrazone (NSAH or E-3a) that binds to the catalytic site (C-site) of hRRM1 (Figure 1B).10
E-3a demonstrated significant levels of selective cytotoxicity against multiple cancer cell lines
while displaying little cytotoxicity against normal mobilized peripheral blood progenitor cells.10
This result presented us with a lead compound possessing favorable properties worthy of further
exploration. In the present study we use the previously determined 2.7 Å crystal structure of this
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
0
lead inhibitor (E-3a) in complex with hRRM1 (PDB ID: 5TUS) as a template for structure-based
ligand optimization. We designed suitable analogs of E-3a, docked them to the C-site of hRR using
the Schrödinger software suite, and evaluated their docking poses to predict which analogs would
show increased inhibitory efficacy relative to the lead compound, E-3a. We then synthesized and
biochemically characterized a series of twenty five analogs (E-3a-z) as hRR inhibitors and
established preliminary structure-activity evaluations (Figure 1C). A five step, modular synthetic
scheme was developed that generally afforded analogs in yields ranging between 63% and 85%,
with few exceptions. E-3f was obtained in 82% yield and displayed an in vitro IC50 of 5.3 ± 1.8
µM against hRR, making it the most potent in this series of analogs. Ranking of the analogs by
their enzymatic IC50 value reveals that the inclusion of a polar substituent in the ortho position of
the salicylic moiety is critical for biological activity, while the substitution of an indole ring for a
naphthalene ring showed no loss in activity. Detailed kinetics assays reveal that these inhibitors
bind and inactivate hRR through a reversible and competitive mode, consistent with our recently
1
0
reported finding for the lead inhibitor E-3a. Remarkably, this class of compounds are the first
non-nucleosidic inhibitors of hRR displaying a competitive, reversible mode of inhibition,
consistent with C-site binding.
Results
6
ACS Paragon Plus Environment

Page 7 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In silico modeling based upon the X-ray crystal structure of E-3a in complex with hRRM1
(
PDB ID: 5TUS) was used to design a focused library of hydrazone-based hRR inhibitors.10
Approximately one hundred distinct analogs with unique substitution patterns surrounding the
central hydrazone core were docked to the C-site, where the grid was defined as a 5 Å cube
centered around the lead hydrazone ligand. Evaluation of the docking poses for interactions with
residues previously established to interact with nucleoside substrates, such as Ser 202, Ser 606,
Gly 246, Cys 218, Cys 429 and Ser 217, helped define a subset of twenty five analogs most likely
to bind at the C-site and inhibit the enzyme. We noted the key residues involved in NDP binding
in the C-site of hRR and catalysis, such as: Glu 431, Cys 429, Cys 444, Cys 218 and Asn 427.
Upon comparing the close hydrogen bonding contacts between parent inhibitor (E-3a) at the C-
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
10
site of hRR in our recent X-ray structure (PDB ID: 5TUS) , Ser 217 and Cys 218 made relevant
contacts through hydrogen bonds with E-3a. Cys 429 offered weaker Van der Waals contact with
inhibitor E-3a, however it did not contribute strongly to stabilize binding to the inhibitor. The
docking poses along with predicted affinities of representative candidates to the C-site of hRR, are
presented in Figure 2A-F.
The library of compounds (E-3a-z) were docked against the S-site and A-site of hRRM1 to
determine if any analogs were likely to bind to allosteric sites other than the C-site. The
Schrödinger docking scores for all sites are reported in Table S3. In general, a lower score is
considered favorable, with scores between 0 and -5 considered poor and scores of -10 or lower
considered an indication of strong binding interactions. The docking scores for the C-site ranged
from -6.26 to -7.76, indicating favorable interactions between the ligands and C-site residues. The
scores for the A-site ranged from -2.99 to -5.23, with only two compounds reporting a score below
-
5 (E-3m and E-3j). Scores for the S-site ranged from -1.95 to -5.41 with all but one compound
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 56
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
(
E-3r) reporting poor docking scores. While E-3m, E-3j, and E-3r showed some favorable
interactions with either the A-site or S-site, all analogs reported much higher docking scores for
the C-site, suggesting that these analogs are more likely to inhibit hRR by binding to the C-site.
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 using the now well-defined pan-assay interference compounds11 (PAINS) filter as a
gateway parameter, we monitored the presence or absence of potentially problematic functional
groups in our analog design. Despite being a potential PAINS candidate, the validity of parent E-
3
a as a lead compound was greatly strengthened by the crystal structure data that showed binding
10
of E-3a to the C-site of RR. A subset of analogs, E-3a, E-3c, E-3f, E-3s, E-3t and E-3u, displayed
favorable docking poses as shown in Figure 2 upon docking to the C-site of hRR. They were
primary synthetic targets in our design leading to twenty additional analogs.
8
ACS Paragon Plus Environment

Page 9 of 56
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 2. Predicted binding interactions for top-ranked NSAH candidates at the C-site of
1
0
hRR based on the X-ray recently published structure PDB ID: 5TUS. A. E-3a; B. E-3c;
C. E-3f; D. E-3s; E. E-3t; F. E-3u.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 56
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
Synthesis of Naphthyl Salicylacylhydrazone analogs:
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 1. A. Modular synthesis of hRR inhibitor library.; Yields and other structural
characterization per each analog is provided in the experimental section. Thermodynamic
control of the condensation step leading to E-isomer is observed. B. UV profile for E-3a.
λ
max at 320 nm is attributed to acyl hydrazone in trans configuration. The varying
concentrations for E-3a are 2.0, 4.0, 8.0, 16.0 and 32.0 µM respectively as measured in
MeOH as a solvent. Data corrected for solvent and background.
We constructed a library of salicyl-derived acyl hydrazones (Scheme 1A) based upon in silico
screening results. The synthetic method yielded a library of analogs through a five-step sequence
from commercially available acids in high overall yields. Specifically, salicylic acid derivatives
(1) were converted to their corresponding methyl esters and subsequently transformed into
corresponding acylhydrazides (2) under the treatment with hydrazine hydrate in methanol.
Naphthyl ring containing acids (4), were reduced to their primary alcohol using lithium aluminum
hydride in diethyl ether and the resulting alcohols were re-oxidized with pyridinium
chlorochromate (PCC) in dichloromethane at room temperature to yield several aldehydes as
10
ACS Paragon Plus Environment

Page 11 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
represented by 5. The resulting aldehydes were condensed with acyl hydrazides (2) with catalytic
amount of glacial acetic acid under reflux over 3-4 h. This step resulted in the generation of E-3a
1
13
as the major product whose identity was verified using H, C-NMR and high resolution mass
spectrometry (See SI). Detailed step-wise description of the synthesis is provided in the SI. Our
synthetic pathway, by virtue of employing higher temperatures during the final condensation step
involving 2 and 5, uniformly afforded the thermally more stable E-isomer in all cases. Though
Scheme 1 outlines the theoretical possibility of a photoisomerization equilibrium that may exist
between the two configurational isomers of 3A (E and Z isomers), in practice we observe only E-
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
,3
isomers to be formed. We observe a possible influence of severe allylic (A ) strain in the Z-isomer
(which forces 6 atoms including the hydrazone nitrogens and the naphthyl carbons and the
hydroxyl group to be in one plane) thermodynamically favoring the E-isomer to be the sole product
in our synthesis.
Recently, acylhydrazones with unique λmax per geometrical isomers were characterized as new
1
2
class of highly tunable photoswitching compounds. Through UV emission spectroscopy, we
measured a λmax of 320 nm attributable to the acyl hydrazone functionality in the trans
configuration (Scheme 1B). It is the E-isomer that crystallized with hRRM1 in the recent report
1
0
we published and used as a template herein. The E and Z isomers are expected to show distinct
biological efficacy, and therefore throughout this work we have characterized the major E-isomer
as a single component, in each analog case, and depict them with the E-notation.
Synthesis of most of the analogs followed the sequence as shown in Scheme 1, with the
exception of sulfonyl hydrazones (E-3s, E-3o, E-3r and E-3q) that were obtained by a slightly
modified procedure (See SI). A few diimino hydrazones were obtained through a bi-directional
imine condensation with aldehydes that deviated slightly from general procedures described for
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 56
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
reactions in Scheme 1 (See SI). Nevertheless, all analogs were neatly obtained as purified products
in moderate to high yields (> 60-90 %) and the reactions were easy to perform on multi-gram
quantities. Overall, twenty-five analogs were prepared for biological evaluations and their
inhibitory properties are listed in Table 1.
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
12
ACS Paragon Plus Environment

Page 13 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. Structures, % yields, in vitro IC50 values, and predicted solubility and
permeability properties for E-3a and its analogs. Cellular IC50 values are averaged
from HCT116 and MDA-MB-231 cell lines. ClogP and membrane permeability
parameters were predicted using Qikprop. Permeability is reported as a diffusion
rate in nanometers/second. Other related properties are provided in Table S2 (SI).
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
HO
O
HO
OH
HO
O
N
O
NH
O
A
N
O
O
N
N
H
S
N
N
N
N
H
N
N
H
H
H
2
N
H
H
H
H
H
H
N
OH
Cl
OH
Analog
IC50
E-3f
.3 ± 1.8 µM
2 %
.758
E-3c
E-3t
E-3s
E-3z
5
7.3 ± 1.5 µM
72 %
6.1 ± 1.7 µM
74 %
6.8 ± 1.5 µM
51 %
10.19 ± 2.1 µM
15 %
against hRR
yield
%
8
from hydrazide
ClogP
octanol/water)
Permeability (nm/s)
Caco-2
2
3.485
333
3.145
3.248
3.788
(
685
329
552
346
358
605
288
150
260
MDCK
p-amino
p-methyl
sulfonyl
7-methyl
indolyl
Alteration from 3A
pyridyl
C6-hydroxyl
m
-chloro
HO
HO
O
HO
O
O
NH
O
NH
O
O
N
Br
N
N
N
N
N
N
N
N
B
H
H
N
H
H
H
H
H
Br
O
OH
O
2
N
OH
NO
2
OH
Cl
Analog
E-3w
E-3v
E-3u
E-3y
E-3n
IC50
1
3.56 ± 1.9 µM
16.63 ± 2.3 µM
85 %
16.65 ± 1.8 µM
63 %
20.7 ± 1.9 µM
85 %
21.14 ± 2.4 µM
64 %
against hRR
%
yield
85 %
from hydrazide
ClogP
octanol/water)
4.362
4.216
3.428
2.742
2.629
(
Permeability (nm/s)
Caco-2
1061
1392
1053
1392
137
118
62
171
73
24
MDCK
p-bromo
o-methoxy
m-bromo
o-methoxy
m-bromo
o-methoxy
5-nitro
indolyl
5-hydroxy
indolyl
Alteration from 3A
HO
HO
HO
HO
O
Br
HO
O
S
O
C
N
N
N
N
N
H
N
N
N
N
H
N
H
H
O2
N
NH
2
O
E-3e
E-3l
21.5 ± 3.6 µM
E-3k
= 39.42 ± 2.7 µM
E-3q
= 43.65 ± 3.1 µM
E-3i
=
IC = 94.11 ± 3.9 µM
IC50 = 23.8 ± 2.3 µM
IC50
IC50
IC50
50
HO
HO
HO
O
HO
O
HO
O
O
H
N
N
S
N
O
O
N
N
N
S
N
N
N
H
N
H
H
N
HO
O
E-3r
E-3g
E-3h
E-3j
E-3o
=
121.8 ± 4.0 µM
IC
IC50
IC50 = 297.7 4.3 µM
IC50 = 686.8 ± 5.2 µM
= 95.2 ± 6.1 µM
IC50 = 86.83 ± 9.4 µM
50
HO
N
O2
N
HO
HO
O
O
O
S
N
N
N
O
N
N
H
H
H
E-3x
= 120.3 µM
E-3m
= 70.1 µM
E-3p
= 130.7 µM
IC50
IC50
IC50
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 56
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
Enzymatic inhibition assays reveal favourable substitutions for increasing potency
The parent inhibitor and its analogs (E-3a-z) were subjected to enzymatic inhibition assays and
ranked according to their IC50s. All of the analogs that were evaluated reported IC50s in the
micromolar range, with subtle variations in structure leading to observable differences in
potencies. Of the top ranked hydrazones based on their IC50 values, five compounds (E-3f, E-3c,
E-3t, E-3s and E-3z as listed in Table 1, group A (above) showed ~ 2-4-fold improvement from
the initial IC50 of 19 µM for our lead compound E-3a. Pyridyl ring-containing E-3f (IC50 ~ 5.3
µM), E-3c consisting of a 6-hydroxy naphthyl ring (IC50 ~ 7.3 µM), p-methyl substituted aryl
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
sulfonamide containing E-3s (IC50 ~ 6.8 µM), and p-amino-m-chloro substituted E-3t (IC50 ~ 6.1
µM) analogs showed the greatest improvement from E-3a with IC50s below 10 µM, as listed in
Table 1, group A (above). Interestingly, E-3z (IC50 ~ 10.2 µM) consisting of an C3-derived indole
ring system in place of the naphthyl ring, retained its potency despite the change in the heterocyclic
ring system. The single point structural change occurring on either of the two flanking aromatic
rings on either side of the hydrazone core impacted IC50s of analogs. The importance of these
substitutions were predicted earlier through the in silico screening, where these five analogs were
identified as top choices. For example, the C-site Ser 606 and Thr 607 display shorter hydrogen
bonding interactions with the acyl group of E-3a and at least one more stabilizing interaction
comes from an additional hydrogen bond due to the presence of polar hydroxyl or amino groups
on the aromatic ring system on these inhibitors as depicted in Figure 2B-F (above). For E-3s, a
sulfonyl hydrazone, both O atoms connected to S engage in hydrogen bonding interactions
favoring the binding of these analogs at the C-site of hRR. For naphthalene substitutions containing
a C6-hydroxy group (E-3c), the main chain of Pro 294 engages in a stronger hydrogen bonding
interaction as depicted in Figure 2B. E-3w, E-3u, E-3v, E-3y and E-3n showed modest
14
ACS Paragon Plus Environment

Page 15 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
improvements from the lead compound as ranked in Table 1, group B (above). Analogs E-3y, E-
3
n and E-3z (from group A) contain an indole ring system replacing the naphthalene core of lead
compounds. This change resulted in preserving, or even improving (E-3z), the efficacy afforded
by the naphthalene ring system. These analogs are a favorable choice for further refinement
because it is easier to derivatize indole ring-containing compounds than the naphthalene ring core
of these inhibitors. Besides the six most potent inhibitors shown in Figure 2, docking poses
revealing contacts between the remaining nine inhibitors and the hRR C-site are presented in
Figures S1-S9 (see SI). Together, these data provide a roadmap for designing second generation
non-nucleosidic RR 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
Seven of the 12 analogs reporting IC50s below 24 µM have a polar group positioned ortho to
the hydrazone chain on the benzene ring, while another three feature a polar group in the meta
position, as shown in Table 1, group C (above). The majority of para-substituted analogs,
however, showed a marked decrease in activity. This suggests that the ortho-polar substitution is
critical to enhancing hRR inhibition. While E-3s showed improvement in potency towards hRR,
all other sulfonyl hydrazones became less potent than the lead compound, suggesting that this
functional change is sensitive to the pendant hydrophobic groups present on either side of the
hydrazone core. Moving the hydroxyl group of the naphthalene ring from carbon 2 to carbon 6
improved the potency toward RR by approximately 3-fold, supporting the prediction made by
structure guided design that this substitution would hydrogen bond more favorably within the
catalytic C-site of hRR. Additionally, substituting the naphthalene ring with an indole derivative
showed no loss in potency, suggesting that the two ring systems are equivalent toward hRR
inhibition. Of the five halogenated analogs tested, four showed an increased potency toward RR.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 56
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
By these observations, a broad pharmacophore for these analogs can be developed
incorporating ortho substitution of polar groups on both the benzene and naphthalene ring systems.
Substitutions which interact with Ser 606, Thr 607, and Pro 294 are considered favorable for
improving inhibitory potency. Additionally, substitution of the naphthalene ring for an indole ring
can be used to access additional substitutions not readily available from a naphthalene ring system
while fragment growth can be directed from the para position of the benzene ring system. Together,
these observations provide a guide for future lead optimization of this class of hRR inhibitors.
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
NSAHs do not inhibit hRR through sequestering catalytically essential Fe
There are multiple reports of small molecule agents that inhibit RR through the scavenging of
1
3
free radicals that are required for the reduction of nucleoside diphosphates. Nitric oxide and
1
4
15
hydroxyurea (that quench the tyrosyl free radicals), and iron chelating small molecules such as
desferrioxamine (that irreversibly chelate to Fe (II) that is essential for housing the free radical)
are a few examples. Because the basic scaffold of these inhibitors possessed a possible chelating
functional group that involves the 2-OH substituent on the naphthyl ring along with the
hydrazonyl-imino N, we wondered if a metal chelating mechanism may contribute to its inhibitory
mode of action against hRR. Therefore, to test this hypothesis, the propensity of E-3a to bind to
3
+
2+
Fe and Fe was studied by UV spectroscopy. In 10 mM ammonium acetate buffer (pH 7.0)
under anaerobic conditions, a 100 µM solution of E-3a was treated with FeCl ∙6H O or
FeCl ∙6H O at varying concentrations of metal ranging from 100 µM to 5 mM. The stoichiometry
2
2
3
2
X+
of E-3a: M tested were 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10, 1:20, and 1:50. A scanning
UV emission spectrum was recorded between 200-800 nm at each concentration. Noticeable
changes occurred when E-3a was subjected to a titration with FeCl
2
∙6H O at concentrations of
2
metal ranging from 400 µM to 1 mM. At concentrations above 1 mM, the UV detector was
saturated and the UV profile of the ligand-metal complex could not be readily obtained. Once the
16
ACS Paragon Plus Environment

Page 17 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2+
E-3a:Fe ion stoichiometry reached, or exceeded a 1:4 ratio, the UV emission spectrum revealed
2
+
mild changes reflecting that the ligand was in a bound state with Fe . Similarly, a titration
3
+
experiment was carried against Fe salt at concentrations of metal ranging from 100 µM to 5 mM
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
at stoichiometry of 1:1, 1:2, 1:3, 1:4, 1:5, 1:6, 1:7, 1:8, 1:9, 1:10, 1:20, and 1:50. There were no
3
+
noticeable changes in UV emission spectra at any of these stoichiometry for Fe ions, indicating
3
+
that E-3a does not chelate Fe ions. Taken together, these results support the conclusion that while
3
+
E-3a does not inhibit hRR through metal chelation or sequestration of Fe , E-3a may participate
2+
in a chelation mechanism in the presence of excess Fe . Results supporting these observations are
presented in Figures S10-S11 (See SI). We reviewed the potential of these inhibitors to chelate
2
+
Mg ions as well, because these ions are present in the in vitro enzyme assay buffer for hRR
inhibition. Per spectra shown in Figure S12 (See SI) we conclusively show that these compounds
2
+
do not bind and sequester Mg ions.
NSAH analogs inhibit hRR competitively and reversibly
1
0
In a recently published study , using multiple lines of investigations involving steady-state
inhibition kinetics, a jump-dilution assay and an in vitro hRRM1 fluorescence-quenching assay,
we have gathered evidence that the lead inhibitor, E-3a inhibited hRR in a reversible, competitive
10
manner with an IC50 of 32 ± 10.2 µM and a dissociation constant (K
D
) of 37 ± 4.7 µM. We
initially suspected that due to the presence of an unsubstituted electrophilic azomethine carbon
(C=N) on the parent pharmacophore, this inhibitor could irreversibly inactivate the enzyme, by
forming covalent bonds with nucleophilic residues at the C-site. Though the 2.7 Ǻ X-ray structure
did not support this model, further detailed kinetic analyses were undertaken for further
confirmation. In order to verify the potential of E-3a to reversibly inhibit hRRM1 at its C-site, a
10
preincubation-dilution experiment was performed as outlined in the parallel study. Results from
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 56
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
that analysis point to reversible inhibition of the enzyme whose activity could be resuscitated upon
diluting away the inhibitor. Further analyses using steady state-kinetics revealed a model
consistent with competitive inhibition, and correspondingly the K
i
value of E-3a was evaluated to
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
be 5.0 µM.
To verify that new analogs of the parent inhibitor (E-3a) follow the same mode of inhibition
of hRRM1in this study, we subjected the most potent analogs, E-3a (as positive control), E-3c, E-
3
t, and E-3w analysis by steady-state inhibition kinetics. Double-reciprocal plots were prepared
for wild type hRRM1 at varying substrate concentrations and hRRM1 in the presence of these each
of the four inhibitors, independently at concentrations of 100 µM, 50 µM, 25 µM, 10 µM, and 5
µM respectively. As shown in Figure 3A (below), the double-reciprocal plot of E-3a inhibition
against hRRM1 indicates that all data sets converge upon a common y-intercept near the origin,
indicative of the fact that K
max is independent of [I], consistent with a competitive mechanism of inhibition. As shown in
Figure 3B, C and D (below), a similar trend was observed for E-3c, E-3t, and E-3w respectively,
indicating that like E-3a, the K of these analogs is dependent upon [I] while Vmax is independent
m
of the inhibitor is dependent upon inhibitor concentration [I] while
V
m
of [I]. Overall, these observations suggest that E-3c, E-3t, and E-3w are competitive inhibitors in
the same manner as the parent inhibitor E-3a.
18
ACS Paragon Plus Environment

Page 19 of 56
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 3. A. Double-reciprocal plot for E-3a. All data sets converge upon a common y-
intercept, supporting a competitive mechanism of inhibition. B. Double-reciprocal plot for
E-3c also follows a competitive model. C. Double-reciprocal plot for E-3t, depicting
competitive inhibition. D. Double-reciprocal plot for E-3w supports a competitive model of
inhibition.
For the kinetic analysis outlined above, the sigmoidal dose response curves for E-3a (as
positive control), E-3c, E-3t, and E-3w are provided in Figure S13 (see SI). It is noteworthy that
our recent discovery of the crystal structure data of E-3a in complex with hRRM1 shows the
inhibitor bound at the C-site of the enzyme, corroborating our kinetic model of inhibition.10
Because ADP-reductase activity is one of the lowest among all NDPs (<50 U/mg) and thereby
3
usually has a low dynamic range, we further tested these inhibitors with a H-CDP assay method.
-1
-1
Specific activity of wild-type R1 (ranging between 450-481 nmol min mg ) as well as IC50 values
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 56
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
calculated for inhibitors E-3a, E-3c, E-3t, and E-3w correlate with those values obtained through
the 14C-ADP substrate method, negating the need to repeat the remaining 14C-ADP IC50
measurements. Data supporting this correlation is provided in the supporting information (Table
S4).
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
As the acylhydrazone core of this library of inhibitors is structurally similar to the
thiosemicarbazone functionality present in the known R2 inhibitor and metal chelator triapine, it
was necessary to determine if the overall mode of inhibition for these inhibitors in this series
involve targeting of hRRM2. In order to probe this aspect, enzymatic inhibition assays were used
to determine the enzymatic IC50 values for triapine (as a control) and compare its efficacy to that
of E-3a against hRRM2. The specific activity of wild type hRRM2 in the presence of triapine (at
varying concentrations of 100, 50, 5, 0.2, 0.50, and 0.10 µM) and E-3a (at varying concentrations
of 5,000, 400, 100, 50, 5, 0.2, 0.50, and 0.10 µM) were normalized, then plotted against the
logarithm [I]. The data were fitted to a sigmoidal dose-response curve to determine the IC50 values
of triapine and E-3a, as 0.185 µM and 123 µM respectively (Figure 4, below). As the IC50 for E-
3
a is approximately 1,000-fold that observed for triapine, and approximately 6 times greater than
the IC50 value for E-3a against hRRM1, it is unlikely that hRRM2 is a significant cellular target
of E-3a.
20
ACS Paragon Plus Environment

Page 21 of 56
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 4. Sigmoidal dose-response curves for triapine and E-3a against hRRM2. The IC50 values
of triapine and E-3a are 0.185 and 123 µM, respectively.
siRNA knockdown approach further points to RR targeting mode of action in vivo
Previous work from our group demonstrated changes in dNTP pools and cell cycle
1
0
progression after treatment with (NSAH) E-3a. Although generally measurements of
perturbations to cellular dNTP pool, in tandem with cell cycle arrest, has been used to identify
1
6
cellular RR activity (Bianchi 1986; Lagergren 1987) additional confirmation can be derived
1
7
4d
using siRNA by Reid et al. or using inducible shRNA by Wisitpitthaya et al. to repress RR.
The siRNA approach measuring the level of sensitization that a drug renders to RR under
knocked-down conditions has been previously utilized to verify the cellular inhibition of RRM1
1
7
by gemcitabine and RRM2 by hydroxyurea and to delineate the role of RR oligomerization in
4
d
the activity of clinically relevant inhibitors . To identify optimal levels of RRM1 repression for
NSAH sensitivity testing, we transfected a range of validated siRNAs (specific for RRM1, 3
RefSeqs, ThermoFisher Silencer® Select, s12358; and control scrambled siRNAs), and assessed
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 56
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
cell growth over a 48-hour period (SI Figure S14). As expected, significant inhibition of RRM1
resulted in growth inhibition, with an IC50 of approximately 3 pmoles/well, while the control
scrambled siRNAs showed less than 50% growth inhibition even at 20 pmole/well, the highest
amount tested. This guided the selection of the range of siRNA amounts (1-5 pmoles
siRNA/well) used in the inhibitor treatment studies. For drug sensitivity study, cells were
transfected with RRM1 or control scrambled siRNAs, followed 48 hours later by administration
of E-3a. Relative growth inhibition was assessed after incubation for an additional 5 days. As
shown in Figure 5, the relative NSAH-inhibitor induced growth inhibition in cells with siRNA-
mediated knockdown of RRM1 was greater than that in cells with endogenous levels of RRM1.
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 5A and B. Sensitization of MDA-MB-231 cell line toward NSAH inhibitor
under siRNA transfection and RRM1 knockdown conditions. A: Using RRM1 siRNA
and B: Using Control-scrambled siRNA. Line displayed in A at the 4.0 pmol level
represents the maximal sensitization occurring in the presence of E-3a, operating within
an optimized siRNA concentration range.
The relative differences in DNA content in respective treatments were measured and are shown
in SI Figure S15. The relative growth inhibition in cells treated with 4 pmol siRNA (IC50 = 0.347
22
ACS Paragon Plus Environment

Page 23 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
µM) was greater than that observed with lesser amounts of RRM1 siRNA (2 pmol IC50 = 1.43
µM) or any quantity of Control siRNA (4 pmol IC50 = 1.94 µM), or no siRNA (IC50 = 1.49 µM),
indicating that suppression of RRM1 caused sensitization to NSAH, further supporting its RRM1
specific mode of action.
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
Cytotoxicity studies of NSAH inhibitor analogs
1
0
Similar to a recently published study from our group , the entire suite of inhibitors (E-3a-z)
were concurrently evaluated for their cellular effects against both human colon cancer (HCT116)
and human breast cancer (MDA-MB-231) cell lines. Continuous 72-hour exposure of E-3a and
gemcitabine (as a control) demonstrated growth inhibition in both human cancer cell lines. IC50
s
for gemcitabine ranged between 30 and100 nM. E-3a displays cytotoxicity against HCT116 cells
with an IC50 of 225 nM while displaying cytotoxicity against MDA-MB-231 cells with an IC50 of
1
0
3
00 nM. Analogs E-3o and E-3v display cytotoxicity against HCT116 cells with IC50’s of 10 and
1 µM respectively. The remaining analogs did not exhibit significant cytotoxicity. Their net
cytotoxicity effect is summarized in Table S2 (see SI). As illustrated in Table S2, not all inhibitors
in this series exhibited significant levels of cytotoxicity. This observation correlated partly with
the fact that not all inhibitors in this series were sufficiently soluble in cell culture media, with
compound precipitation resulting in insignificant cytotoxicity for many analogs. Therefore,
detailed cellular toxicity results could be obtained only for E-3a, E-3o and E-3v. In many cases, a
cancer cell cytoxicity with cellular IC50 of >10 µM may primarily be an effect due to high
insolubility of the inhibitor, underscoring the importance of solubility (in biologically relevant
environments) for future improvements in potency.
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 56
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
Discussion and Conclusion
Although acylhydrazones display a wide range of biological activities, including inhibition of
18
19
20
metazoan proteases , anti-Leishmania properties , herbicidal effects , and inhibition of
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
1
proliferating cellular nuclear antigen (PCNA) , we were the first to discover that naphthyl Salicyl
acylhydrazones (especially the lead inhibitor E-3a) inhibit hRR through binding at the C-site of
10
RRM1. In this study, we establish a hydrazone pharmacophore inactivating hRR in a reversible,
competitive mode through binding at the C-site of the enzyme. As such, we rationalized that
structure-based design could be used in tandem with structure-activity relationship studies to
further develop our understanding of how acylhydrazones could be used to target hRR for cancer
treatment.
Most of the clinically used RR inhibitors are nucleoside analogs such as gemcitabine, which is
currently used to treat pancreatic cancer. Gemcitabine has been shown to inhibit RR in a synergistic
manner, with irreversible hRR inhibition and depleted nucleotide pools allowing for gemcitabine
triphosphate to become incorporated into growing DNA strands by DNA polymerase, causing
delayed chain termination and triggering apoptosis. This mechanism does not distinguish between
normal and cancerous cells, and therefore gemcitabine is associated with serious side effects due
to toxicity towards normal cells resulting from DNA chain termination, irreversible inhibition of
hRR, and inhibition of other enzymes such as topoisomerase-1, thymidylate synthase,
deoxycytidine deaminase, CTP-synthase, which recognize phosphate moieties.6,7 This
combination of indiscriminate cross-reactivity, irreversible inhibition, and DNA chain termination
produces undesirable cytotoxicity in normal cells and thereby results in a relatively low therapeutic
index when use in the treatment of patients. We therefore hypothesize that identification and
24
ACS Paragon Plus Environment

Page 25 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
characterization of specific, reversible, non-nucleoside RR inhibitors could prove advantageous
by reducing the cytotoxicity towards normal cells caused by nucleoside analogs.
While previous RR inhibitors display a diverse range of mechanisms of inhibition, there have
been few efforts to identify reversible, non-nucleosidic small molecules that inhibit the hRRM1
subunit. In Ahmad et.al. 2015, we reported a rapid throughput screening method integrating in
silico docking, cell-based assays, and biochemical experiments which identified ten novel classes
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
8
of non-nucleosidic RR inhibitors. This study marked the first compounds identified to inhibit RR
by binding to a protein-protein interface (M-site) and inducing formation of catalytically inactive
hexamers. Recently, we reported the identification of the first non-nucleoside identified to bind to
10
the C-site of hRR (PDB ID: 5TUS). This lead inhibitor E-3a, an acylhydrazone, was determined
to inhibit RR in a reversible, competitive manner and exhibit potent cytotoxicity towards multiple
cancer cell lines.
1
0
Using the crystal structure of E-3a in complex with hRRM1 (PDB ID: 5TUS) as a template,
a focused library of analogs was designed using in silico modeling. In silico docking with the
Schrödinger software suite was used to evaluate 100 analogs and the top 25 were selected based
upon the predicted docking poses. A modular, E-isomer selective synthetic pathway was
implemented which afforded hydrazones on a multi-gram scale in five steps with isolated yields
ranging between 65 and 85% typically, with a few exceptions. In vitro RR inhibition assays
identified five analogs with 2-4 fold improvement in IC50 from the lead compound, E-3a. All five
analogs were predicted to be more potent toward RR during the in silico modeling, as all of these
analogs showed shorter hydrogen bonding interactions with residues Ser 606, Thr 607, and Ser
2
17, which are known to hydrogen bond with natural nucleoside substrates. It was also determined
that the ortho substitution of a polar group on either ring system was critical for RR inhibition,
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 56
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
while substitution of the naphthalene ring system for an indole ring resulted in no loss in activity,
providing potential access to a broader range of chemical derivatives. Upon evaluating in vitro
inhibitory potencies against hRR, broadly a subset of 15 inhibitors were identified within a narrow
range of IC50s.
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
Due to the presence of a hydrazone functionality on the parent inhibitor structure, we were
concerned that this group had the propensity to be attacked by an active site nucleophile (e.g. Cys
4
29) present in hRR, thereby irreversibly inactivating the enzyme. The data presented in Figure 3
provided us with a mechanistic model of inhibition for these inhibitors. The participation of the
hydrazone azomethine nitrogen and the C=N bond through a covalent mode can be ruled out, and
reversible inhibitory modes can be invoked. Metal chelation assays determined that E-3a does not
3
+
2+
2+
chelate Fe ions, however Fe complexation was observed in the presence of excess Fe (4-fold
excess and higher). While the X-ray crystal data obtained in a previous study strongly suggests
that the biological activity of E-3a-like compounds can be attributed to inhibition of RR, a separate
metal chelation mechanism cannot be completely ruled out, especially when relatively high Fe2+
ion concentrations may be present in cells (typically > 1 mM). Importantly, analogs E-3c, E-3t,
and E-3w were all determined to inhibit hRR through similar competitive, reversible mechanisms
as the lead compound (E-3a), suggesting that novel analogs possessing this pharmacophore follow
the same mechanism of inhibition as the lead compound.
The relatively large difference between the mid-micromolar enzymatic IC50 (in vitro) and
nanomolar cellular IC50 (against cancer cells in vivo) suggested that E-3a may interact with
multiple targets. This raised a concern regarding whether hRR is indeed a cellular target. In a
recently published study, we evaluated the perturbation of cellular dNTP levels specifically in the
presence of E-3a, and compared them to levels in cells treated with hydroxyurea, at their respective
26
ACS Paragon Plus Environment

Page 27 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
IC50s. The addition of E-3a was shown to induce a pattern of dNTP depletion with greatest
depletion of dATP and dGTP, with little or no effect on dCTP or dTTP levels. This was followed
by an arrest of cells in early S-phase. These observations are similar to patterns observed with
gemcitabine and hydroxyurea and comprise a signature for a cellular hRR inhibitor.22 Using
mobilized peripheral blood progenitor cells (a common target for dose limiting toxicity during
gemcitabine therapy), E-3a was shown to have a superior therapeutic index compared to
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
0
gemcitabine. These observations point to promising selective cytotoxicity attributable to hRR
inhibition by E-3a. Analogs E-3o and E-3v, showing similar cytotoxicity profile as E-3a, may also
be expected to target hRR in vivo. However, we are cognizant of the possibility that inhibitors like
E-3a and its analogs may possibly engage multiple cellular targets that are yet to be identified.
It is generally agreed that RR is a challenging therapeutic target for cancer, as inhibitors are
unlikely to distinguish between malignant and normal cells. We recently demonstrated that in
blood progenitor cells, inhibitor E-3a showed a superior therapeutic index in comparison with
gemcitabine, suggesting a possible discrimination between transformed and untransformed cells.10
We wish to further exploit this new class of inhibitors for the possible treatment of cancer,
especially against gemcitabine-resistant forms. In the long term, identification of non-nucleoside,
reversible RR inhibitors may provide an avenue towards developing safer alternatives to
irreversible nucleoside analogs as chemotherapeutic agents. Furthermore, considering the fact that
RR broadly impacts multiple pathologies, these reversible agents may impact drug discovery for
treatment of a wider range of proliferative diseases.
Experimental Section
In Silico docking with Schrödinger suite for predicting C-site binding affinity
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 56
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
In silico docking of each hydrazone analog was performed using the Glide docking module of
the Schrödinger 9.3 modeling software suite as previously described in Ahmed and Huff et al.8
The docking site was defined as a box of 5 Å centered on the original hydrazone ligand in the hR1
complex (PDB ID: 5TUS). Ligands were then docked to the catalytic site using Glide XP. The
generated docking poses were evaluated for interactions with residues that commonly bind to
natural substrates, such as Ser 202, Ser 606, Gly 246, and Arg 293. Ligands that showed favorable
interactions with these residues were selected for synthesis. Generally, most inhibitors screened
here showed very minimal perturbation or binding interaction with C-site residues involved in
catalysis, such as Glu 431, Cys 429, Cys 444, Cys 218 and Asn 427. Therefore, these sites were
omitted from depiction in the docked poses with inhibitors. Specific docking poses for inhibitors
that showed lower than 5-10 µM potency (IC50) are depicted in Figures S1-S9 (see SI).
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
In Silico docking with Schrödinger suite for predicting site selectivity
Proteins and ligands were prepared as described below. Docking of the S-site and A-site was
performed using the previously determined ATP-TTP-hRRM1 crystal structure (PDB ID: 3HNE).
3
The S-site docking grid was defined as a 5 Å cube (125 Å ) centered on the TTP ligand. The A-
3
site docking grid was defined as a 5 Å cube (125 Å ) centered on the ATP ligand. Docking scores
for each inhibitor docked to each of the three sites are reported in Table S3.
Human ribonucleotide reductase protein expression and purification
The hRRM1 protein was expressed in E. coli BL21-codon plus (DE3)-RIL cells and purified
5
c
using peptide affinity chromatography as previously described in Fairman et. al. The hRRM2
protein was also expressed in E. coli BL21-codon plus (DE3) cells and purified to homogeneity
5c
using Ni-NTA affinity chromatography as described in Fairman et. al. Iron was loaded into the
28
ACS Paragon Plus Environment

Page 29 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
hRRM2 subunit as described in Ahmed and Huff et. al. Concentration of the homogeneous protein
was quantified using UV spectroscopy.
Ribonucleotide reductase in vitro inhibition 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
[14
was determined by C-ADP reduction assays using a reaction mixture of 0.3 μM hRRM1 and 2.1
μM hRRM2 in an activity assay buffer containing 50 mM HEPES (pH 7.6), 15 mM MgCl , 1 mM
C]-ADP reduction assay: The in vitro specific activity of human ribonucleotide reductase
1
4
2
1
4
EDTA, 100 mM KCl, 5 mM DTT, 3 mM ATP, 100 μM dGTP and 1 mM C-ADP (~3000
cpm/nmol). The reaction mixture was pre-incubated for 3 min at 37 ˚C, and 30 μL aliquots were
sampled at fixed time intervals after initiation of the reaction. Reactions were quenched by
immersion in a boiling water bath. The aliquots were then cooled and treated with alkaline
14
phosphatase for 2 hours. The product C-dADP that formed during the reaction was separated
1
4
14
from substrate C-ADP using boronate affinity chromatography. Amount of [ C]-dADP formed
from the reaction was quantified by liquid scintillation counting using a Beckman LS6500 liquid
scintillation counter. The assays were repeated in triplicate and the averaged specific activities
were used for IC50 calculations.
[3
H]-CDP reduction assay: Assays to determine the IC50 values for compounds 3a, 3c, 3t,
3
and 3w in the presence of H-CDP substrate were performed with a reaction mixture consisting of
0.3 μM hRRM1 and 2.1 μM hRRM2 in an activity assay buffer of 50 mM HEPES (pH 7.6), 15
3
mM MgCl
2
, 1 mM EDTA, 100 mM KCl, 5 mM DTT, 3 mM ATP, 100 μM dGTP and 1 mM H-
CDP (~2000 cpm/nmol). After processing of the reaction similar to the procedure described above,
3
the amount of [ H]-dCDP formed from the reaction was quantified by liquid scintillation counting
using a Beckman LS6500 liquid scintillation counter. The assays were repeated in triplicate and
the raw specific activities are reported below in Table S2 (see SI).
29
ACS Paragon Plus Environment