Optimization of pyrrolamide topoisomerase II inhibitors toward identification of an antibacterial clinical candidate (AZD5099)

Article abstract of DOI:10.1021/jm500462x

AZD5099 (compound 63) is an antibacterial agent that entered phase 1 clinical trials targeting infections caused by Gram-positive and fastidious Gram-negative bacteria. It was derived from previously reported pyrrolamide antibacterials and a fragment-based approach targeting the ATP binding site of bacterial type II topoisomerases. The program described herein varied a 3-piperidine substituent and incorporated 4-thiazole substituents that form a seven-membered ring intramolecular hydrogen bond with a 5-position carboxylic acid. Improved antibacterial activity and lower in vivo clearances were achieved. The lower clearances were attributed, in part, to reduced recognition by the multidrug resistant transporter Mrp2. Compound 63 showed notable efficacy in a mouse neutropenic Staphylococcus aureus infection model. Resistance frequency versus the drug was low, and reports of clinical resistance due to alteration of the target are few. Hence, 63 could offer a novel treatment for serious issues of resistance to currently used antibacterials.

Full text of DOI:10.1021/jm500462x

Subscriber access provided by University of Texas Libraries
Article
Optimization of Pyrrolamide Topoisomerase II Inhibitors Towards
Identification of an Antibacterial Clinical Candidate (AZD5099)
Gregory S. Basarab, Pamela J Hill, C. Edwin Garner, Kenneth Hull, Oluyinka Green, Brian
A. Sherer, P. Brian Dangel, John I. Manchester, Shanta Bist, Sheila Hauck, Fei Zhou,
Maria Uria-Nickelsen, Ruth Illingworth, Richard Alm, Michael Rooney, and Ann E. Eakin
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 24 Jun 2014
Downloaded from http://pubs.acs.org on June 25, 2014
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 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Optimization of Pyrrolamide Topoisomerase II Inhibitors Towards Identification of an Antibacterial
Clinical Candidate (AZD5099)
*
a
b
c
Gregory S. Basarab, Pamela J. Hill, C. Edwin Garner, Ken Hull, Oluyinka Green, Brian A. Sherer, P.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
d
e
Brian Dangel, John I. Manchester, Shanta Bist, Sheila Hauck, Fei Zhou, Maria Uria-Nickelsen, Ruth
f
Illingworth, Richard Alm, Mike Rooney, and Ann E. Eakin
AstraZeneca R&D Boston, Infection Innovative Medicines, 35 Gatehouse Drive, Waltham, MA 02451,
USA
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 68
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: AZD5099 (Compound 63) is an antibacterial agent that entered Phase 1 clinical trials
targeting infections caused by Gram-positive and fastidious Gram-negative bacteria. It was derived from
previously reported pyrrolamide antibacterials and a fragment based approach targeting the ATP binding
site of bacterial type II topoisomerases. The program described herein varied a 3-piperidine substituent
and incorporated 4-thiazole substituents that form a seven member ring intramolecular hydrogen bond
with a 5-position carboxylic acid. Improved antibacterial activity and lower in vivo clearances were
achieved. The lower clearances were attributed, in part, to reduced recognition by the multidrug resistant
transporter Mrp2. Compound 63 showed notable efficacy in a mouse neutropenic Staphylococcus aureus
infection model. Resistance frequency versus the drug was low, and reports of clinical resistance due to
alteration of the target are few. Hence, 63 could offer a novel treatment for serious issues of resistance to
currently used antibacterials.
0
1
2
3
4
5
6
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 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
INTRODUCTION
1, 2
The work presented herein details optimization work in pyrrolamide antibacterial agents culminating in
the identification of compound 63, a drug candidate that progressed to Phase 1 human clinical trials.
There is an urgent need for novel antibacterial agents effective against drug-resistant infections for which
treatment options have become limited. Among the most pressing pathogens in the hospital setting are
MRSA (methicillin resistant Staphylococcus aureus) and VRE (vancomycin resistant Enterococci). These
drug-resistant Gram-positive pathogens included in a group referred to as “ESKAPE” pathogens
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
Enterococcus faecium, S. aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas
3
-6
aeruginosa and Enterobacter spp.) that cause the majority of US hospital infections. Pyrrolamide
antibacterials (Compounds 1 and 2, Figure 1) display potent activity against these Gram-positive
pathogens and function by inhibition of the type II bacterial topoisomerases, DNA gyrase and
Topoisomerase IV (Topo IV), two homologous and essential enzymes that alter DNA topology during
1, 2, 7-10
replication.
DNA gyrase primarily introduces negative supercoils into DNA, while Topo IV
primarily decatenates DNA during replication. Both topoisomerases exist as heterotetrameric A B
2
2
complexes, designated GyrA and GyrB for DNA gyrase, and ParC and ParE for Topo IV. The
pyrrolamides are competitive with ATP, the binding site of which resides on GyrB and ParE subunits of
1
DNA gyrase and Topo IV, respectively. Extensive crystallographic studies of truncated N-terminal
constructs of GyrB and ParE have been used to guide inhibitor design for the pyrrolamides and for a
11-16
variety of other scaffolds that similarly bind to the ATP site of the enzymes (Figure 2).
DNA gyrase and Topo IV are clinically validated targets for the treatment of resistant bacterial
1
7
18
infections, with two compound classes, the fluoroquinolones and aminocoumarin antibiotics,
1
9-21
inhibiting the enzymes via distinct binding sites.
The fluoroquinolones bind in the cleavage-ligation
22
site, where they stabilize double-strand breaks in the bacterial DNA, induce rapid killing of the cells,
and exhibit broad-spectrum Gram-positive and Gram-negative antibacterial activity. The aminocoumarins
bind the ATPase subunits of gyrase (GyrB) and, to a lesser extent, Topo IV (ParE), and exhibit Gram-
positive antibacterial activity. Novobiocin (Figure 2) is the only aminocoumarin ever approved for
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 68
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
clinical use, but due to safety concerns and the introduction of cephalosporins, doxycycline and latter
generation penicillins in the 1960s, it was not widely used and eventually withdrawn from the market.
Widespread availability and use of the fluoroquinolones has driven the emergence of resistance to this
23, 24
25-27
0
1
2
3
4
5
6
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
class;
however, clinical isolates of fluoroquinolone-resistant bacteria are typically susceptible to
ATPase inhibitors with no elevation of the MIC. To our knowledge, there are no well-documented
2
3, 28
examples of bacterial resistance to ATPase inhibitors in the clinic.
The high degree of homology
between GyrB and ParE, particularly in the ATP-binding region, supports the design of dual-targeting
inhibitors which would minimize the development of target based bacterial resistance since the two
2
9
enzymes are independently essential and independently encoded. Importantly, that DNA gyrase and
Topo IV are conserved across prokaryotic species offers the prospect that broad spectrum activity can be
realized. Notably, DNA gyrase, only found in bacteria, is the only known prokaryotic or eukaryotic
topoisomerase that traverses an energetically uphill gradient towards creating negative supercoils in
DNA, rather than cleaving and re-ligating DNA single or double strands with net thermodynamically
29
neutral energetics.
A wide variety of compounds besides pyrrolamides and aminocoumarins have been shown to
7
exhibit antibacterial activity by competing with ATP in DNA gyrase and Topo IV. These include
aminocoumarin mimics such as RU79115 and other natural products such as the macrocyclic
3
0-32
cyclothialidine
pyrimidoindoles,
pharmacophore (Figure 2).
and a broad range of non-natural product based heterocyclic compounds including
13, 33
14
34
16
azaindoles, anilinopyridines, pyrazothiazoles and those with an ethylurea
11, 12, 15, 35, 36
Although the compounds are structurally diverse, they all contain a
hydrogen bond (H-bond) donating and H-bond accepting motif to an active-site aspartate and a tightly
bound water, which contributes to inhibitory potencies. This H-bond donating and accepting motif
mimics that seen for adenine of ATP as determined by X-ray crystallography with the 46-kDa N-terminal
37
GyrB ATP binding domain of E. coli. The ATP competitive inhibitors extend outside of the ATP
binding pocket to a highly conserved region that participates in the dimerization of GyrB subunits
following ATP binding. This dimerization involves interactions with the N-terminus from the adjacent
ACS Paragon Plus Environment

Page 5 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
38
monomer and results in clamping of the ATPase subunits around the bacterial DNA. Due to functional
differences, this region is variable compared to mammalian homologues as well as to other ATPases of
39
the GHKL protein family. This is an important feature towards attaining selectivity against a myriad of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ATP utilizing enzymes. The pyrrolamide compounds incorporate a pharmacophore that interacts with two
2
arginine residues critical for dimer formation in this region of the topoisomerases. Since there are
significant deviations between prokaryotic and eukaryotic type II topoisomerases in this region, it was
anticipated that inhibition of the latter would be minimal for analogues utilizing these binding
interactions, thus reducing the risk of target-based mammalian toxicity.
The genesis of the pyrrolamide class of compounds via fragment based NMR screening against a
1
24-kDa N-terminal ATP binding domain of DNA gyrase from E. coli has been published previously. A
more interesting fragment, a 5-methylpyrrole carboxylic ester (Figure 1), was modeled to bind in a deep
pocket of the enzyme where adenine of the ATP resides, and a program to optimize inhibitory potency
against ATPase activity of DNA gyrase was initiated. A series of corresponding carboxamides was
designed to extend from the pyrrole into space non-overlapping with ATP otherwise, improving
inhibitory potency and leading to compound 1. Notably, 1 demonstrated in vivo efficacy in a S.
pneumoniae mouse lung infection model, validating the scaffold as a lead series for further optimization.
A fluorine substituent on the 3-piperidine position affords two chiral centers; de-convolution of the four
possible diastereomers showed highest activity in the (3S,4R)-diastereomer 2 with improved DNA gyrase
inhibitory potency and antibacterial activity. A crystal structure of 2 with the N-terminal 23-kDa ATP
binding domain of Staphylococcus aureus GyrB (PDB code 3TTZ, Figure 1) showed the pyrrole NH
positioned to donate an H-bond to Asp81 and the carboxamide carbonyl positioned to accept an H-bond
from a water molecule that is otherwise associated with the sidechain functionality of Asp81 and Thr173
in an H-bond array. The thiazole ring of 2 stacked against the guanidinium of Arg84 for what is
4
0
considered a favorable cation-π interaction. Arg84 was otherwise locked in a salt bridge with Glu58,
together forming a ceiling over the inhibitor binding pocket. The thiazole carboxylic acid was positioned
for a salt bridge with Arg144. The fluorine atom faced but did not closely contact a hydrophobic region of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 68
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 enzyme and may have induced a conformational bias to the piperidine ring for more favorable
binding. Compound 2 also demonstrated in vivo efficacy via oral administration in a S. pneumoniae
mouse infection model, resulting in similar reductions in measured colony forming units (CFU) at notably
lower doses relative to 1. However, species-to-species variation for the in vivo clearance (Cl) and
bioavailability (F) of 2 was difficult to reconcile (see below), and an exploratory Phase 1 evaluation of the
drug was carried out with oral dosing in healthy human volunteers to assess drug exposure. Ultimately,
adequate drug exposure in the blood deemed necessary for reduction of bacterial burdens during an
infection based on pre-clinical models could not be achieved (unpublished work). The efforts reported
herein address improvement in antibacterial activity and drug exposure relative to 2 towards identification
of a more efficacious drug candidate.
0
1
2
3
4
5
6
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
CHEMISTRY
Compounds were synthesized according to the general procedure shown in Scheme 1. The
4
1
dichloropyrrole carboxylic acid 3 was prepared according to literature methods and was coupled with
aminopiperidines 4 to prepare 4-amino piperidines 5, N-protected as a carbamate. Deprotection of the
piperidine nitrogen to afford compounds 6 was followed by reaction with a substituted bromo- or
chlorothiazoles 7 to afford corresponding esters of the final products. Hydrolysis of the esters led to the
desired target carboxylic acids.
The 3-methoxy group on the piperidine ring imparted the best overall profile for final products
(
Results and Discussion), and the synthesis of the corresponding 4-amino-3-methoxypiperidine linker 4 is
4
2, 43
shown in Scheme 2, considerably shortening previous synthetic routes.
Ethyl 4-oxopiperidine-1-
carboxylate was converted to methoxypiperidine ketal 8 upon reaction with iodosobenzene diacetate and
4
4
potassium hydroxide according to established procedures followed by methylation with sodium hydride
and methyl iodide. The intermediates for the other R3 ethers shown in Table 2 (compounds 28- 31) were
prepared analogously as detailed in the Supporting Information. Deprotection of the ketal followed by
ACS Paragon Plus Environment

Page 7 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reductive amination with benzylamine provided 9 as a racemic mixture in a 9:1 ratio of cis- to trans-
diastereomers that could be separated by chromatography to afford clean cis material. The benzyl group
was removed via hydrogenolysis to yield the corresponding racemic 4-amino-3-methoxypiperidine 4,
which was transformed into compound 22 following the sequence of Scheme 1. Direct reductive
amination with ammonium acetate in place of benzylamine to afford 4 was non-diastereoselective
affording difficult to separate cis- and trans-isomers. Enantiomers of the cis-isomers could be obtained by
chiral chromatography of 9, but an improved separation and recovery was achieved after capping racemic
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
with the Cbz group followed by chiral chromatography to separate 10a and 10b, each pure to the limit
of chiral HPLC analysis. The Cbz of the (+)-isomer 10b with the (3R,4S) absolute configuration was
removed en route to compound 23, which proved considerably more active than the opposite enantiomer
24 derived from 10a. The absolute configurations of 10a and 10b were inferred from previous work that
2
led to compound 2 in that the signs of the rotations for the analogous intermediates correlated;
confirmation was achieved by a single molecule X-ray crystal structure of compound 63 (deposited in the
Cambridge Crystallographic Data Center as CCDC 993908, see Supporting Information). The
stereocontrolled syntheses of the other two diastereomeric 4-amino-3-methoxypiperidine linkers that were
incorporated in 24 and 25 are detailed in the Supporting Information. The syntheses of piperidine linkers
relevant to compounds otherwise in Table 2 are also detailed in the Supporting Material.
In the course of our investigation, a series of acyclic and cyclic ketals at the piperidine 3-position
were also prepared and evaluated (Scheme 3 and Table 3). The preparation of the dimethyl ketal 39 began
by conversion of ethyl 3-oxopiperidine-4-carboxylate in two steps via protection of the piperidine N-H
with methylchloroformate followed by acid catalyzed ketalization with trimethylorthoformate. Hydrolysis
of the ethyl ester followed by Curtius rearrangement in the presence of benzyl alcohol furnished the Cbz-
protected aminopiperidine 12. After deprotection, this was coupled to the dichloropyrrole carboxylic acid
and was followed by removal of the methylcarbamate protecting group to prepare 14. S Ar reaction
N
between 14 and methyl 2-bromothiazole-5-carboxylate and subsequent hydrolysis led to 39. The sequence
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 68
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
to compounds 40 and 41 was carried out in an analogous fashion, and details for the synthesis of the
corresponding intermediates are given in the Supporting Information. However, to synthesize a larger
variety of analogs, ketal exchange from the methyl ester of 39 was more productive. To this end, the ester
of 39 was heated with a given diol and an acid catalyst to produce the methyl esters of compounds 42-46,
which were subsequently hydrolyzed to the acids. The ketalization with 2-methoxy-1,3-propanediol
produced two separable diastereomers, the esters of 42 and 43. Extensive NMR correlations (COSY,
NOESY, HMQC) failed to assign the configuration of the diastereomers, and the assignment was made
only after co-crystallization of the more potent 43 with the 20.4 kDa N-terminal ParE construct of H.
influenzae showing the methoxyl group to be positioned syn to the piperidine carboxamide on the ketal
spirocycle (unpublished results).
0
1
2
3
4
5
6
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 carboxamides adjacent to acid of the thiazole ring were prepared via the procedure of
Scheme 4. In one reaction step, thiourea and 3-chlorotetrahydrofuran-2,4-dione were cyclized in ethanol
with concomitant ethanolysis to form aminothiazole 15. TBS-protection of the alcohol followed by a
Sandmeyer reaction with tert-butylnitrite and CuCl yielded 16, which was deprotected and oxidized to
2
the chlorothiazole carboxylic acid 17 under Jones reaction conditions. Displacement of the chloride of 17
with 6a to afford 18 enabled HATU mediated conversion to a variety of ester-amides that were
hydrolyzed to the acid-amides. The amides proved to be of particular interest due to the capability of
intramolecular H-bonding to the acid. The final hydrolysis step required significant optimization to avoid
formation of regioisomeric acid-amides. Treatment of ester-amides with aqueous LiOH or NaOH resulted
in a mixture of regioisomers often in as much as a 1:1 ratio resulting from the intermediacy of the cyclic
diimide and subsequent hydrolysis. The use of Ba(OH) in aqueous methanol enabled clean hydrolysis
2
without cyclization to the diimide or regioisomer formation. The intermediate 16 was also used to prepare
other derivatives shown in Table 4 (i.e. compounds 49 and 54-56, 59) via straightforward chemistry, and
the synthetic details are outlined in the Supporting Information. Also analogously prepared from
ACS Paragon Plus Environment

Page 9 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
commercial thiazole and benzothiazole materials were compounds 47, 48, 57 and 58 (Table 4) as
described in the Supporting Information.
Finally, Scheme 5 summarizes the synthesis of thiazole carboxylic acids with heterocycles
adjacent to the carboxylic acid towards engineering an intramolecular H-bond. A series of heterocyclic β-
keto esters 19 were prepared and converted to aminothiazoles 20 by iodination and reaction with thiourea.
Sandmeyer conversion to chloride or bromide and displacement with 6a led to the desired heterocyclic
thiazole esters that were converted to the acids 73-90. The experimental details for the syntheses of
intermediates represented by 21 are given in the Supporting Information.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
Enzyme inhibition (IC or the half maximal inhibitory concentration) was measured against the S. aureus
50
12
GyrB and the E. coli ParE isozymes, and values are listed in Tables 1-5. The rank ordering of
compounds versus the two isozymes was similar, although the absolute IC s were generally 1-2 orders of
5
0
magnitude lower in GyrB, wherein the IC s of more active inhibitors approached the enzyme
50
concentration, limiting resolution below 10 nM in the routine screening assay. Thus, the ParE assay
proved more robust for driving the optimization of binding affinity. A similar shift of IC s in E. coli ParE
5
0
1
2, 33
relative to other isozymes has been noted for other chemical series of ATPase inhibitors,
and is
partially explained by differences in residues in the hydrophobic floor that forms part of the ribose
45
binding pocket. Not all of the pyrrolamides for which IC shifts are observed interact with the
5
0
hydrophobic floor, and thus this explanation is incomplete. In any case, the inhibitory potency difference
between the two isozymes is not necessarily reflected in the cellular response. Both assays measure ATP
hydrolysis via Malachite Green detection of phosphate release in a purified, reconstituted system which
does not account for regulatory mechanisms present in the bacterial cell, or for relative expression levels
26
of GyrB and ParE. The relative importance of these enzymes for any given pathogen is best determined
by genomic analysis of mutants generated to a specific inhibitor or inhibitor class.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 68
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
Microbiological activity of the pyrrolamides was measured against a panel of bacterial pathogens,
-
including two E. coli strains, a wild-type strain (Eco523)and a tolC (outer membrane transport channel)
strain disabled in compound efflux capability (Tables 1-5). The pyrrolamide compounds herein showed
weak activity against other serious Gram-negative pathogens such as Pseudomonas aeruginosa and
Acinetobacter baumannii (data not shown). Considerably lower MIC values were recorded for E. coli
0
1
2
3
4
5
6
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
-
tolC compared to wild-type, demonstrating that achieving effective intracellular concentrations in this,
and likely other, Gram-negative organisms is problematic. Overall, the strength of the pyrrolamide class
lies primarily in its Gram-positive spectrum, thereby positioning the drug class for use against skin and
skin structure infections (SSSI) that are mainly caused by S. aureus and Streptococcus pyogenes.
Compounds that showed activity 1 µg/mL or less versus two strains of S. aureus, a methicillin-sensitive
strain (MSSA) and a methicillin-resistant, quinolone-resistant strain (MRQR) support the prospect of
battling clinical resistance. The added activity of compounds versus S. pneumoniae, Haemophilus
influenzae, Moraxella catarrhalis and other susceptible bacterial (data not shown) such as Mycoplasma
pneumoniae and atypical bacterial pathogens could support treatment of community acquired bacterial
pneumonia (CABP) and respiratory tract infections (RTI).
Influence of piperidine substitution and stereochemistry. The stereochemistry on the piperidine
scaffold had a profound effect on activity as seen by comparing the four possible diastereomers having a
C3-methoxyl substituent (compounds 23-26, Table 1). The highest activity resided in the (3S,4R)-
diastereomer 23, the same configuration reported for 2. As mentioned for 2, the carboxamide substituent
with the R-configuration on the piperidine ring is thought to better align the key interactions of the pyrrole
NH with Asp81 and the thiazole carboxylate with Arg84 in the chiral environment of the enzyme. The
methoxyl substituent in the S-configuration was presumed to afford better contact with a hydrophobic
region below the plane of the piperidine ring from the perspective diagrammed in Figure 1, accounting for
the improved potency relative to 1. The second most active diastereomer 25 (3R,4R), epimeric at the 3-
position, placed the methoxyl substituent in an allowable but unproductive environment resulting in a 2-
ACS Paragon Plus Environment

Page 11 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
fold drop in S. aureus GyrB inhibitory potency relative to 1. Modeling suggested that the piperidine ring
of the two 4S-carboxamide diastereomers 24 and 26 would rotate 180° at both the 1- and 4-positions,
maintaining the favorable orientations and alignments of both the pyrrole NH and the thiazole carboxylate
(Figure 3). This would position the methoxyl substituent towards the inside surface of the enzyme binding
pocket pushing against the well-ordered crystallographic water molecule, particularly for the 3S,4S-
diastereomer. Overall, placing a methoxyl group on the piperidine C3-positon in the 4S-configuration led
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
to lower biochemical potency than fluorine with the 4S-configuration, which, in turn, was less potent
than hydrogen. The ramification of these findings was that subsequent analogues focused on either
racemic compounds or chiral compounds with the carboxamide in the R-configuration and R3 piperidine
substituents oriented cis in the S-configuration. Table 2 shows the data for pyrrolamides with a variety of
other substituents at the piperidine 3-position, illustrating a dilemma for the design of improved
compounds. More polar compounds, as reflected in logD values, that were well accommodated for
binding to GyrB and ParE might be favored in efforts to improve solubility and human plasma protein
binding (PPB expressed as fraction unbound, f ) important for clinically useful drugs. However,
u
antibacterial activity trended lower with the increased polarity, presumably due to decreased membrane
permeability. For example, incorporation of the more hydrophilic methoxyethoxyether of compound 31 at
the 3-position afforded higher MIC values than the more lipophilic ethers of compounds 28-30, even
though inhibitory potencies were comparable. Similarly, sulfide 36 and sulfone 37 showed similar
inhibitory potencies, but the greater polarity inherent in the latter led to much higher MIC values.
Incorporating the chlorine atom of 35 afforded a quite potent compound, but one with exceedingly low fu
contraindicating what is deemed necessary for in vivo efficacy. Compound 38 having an azide substituent
demonstrated highly potent antibacterial activity relative to 23, which was counteracted, however, by
lower f and higher in vivo clearance (79 mL/min/kg) in the rat. Compound 23 retained the most
u
interesting profile taking into account antibacterial activity in combination with the other characteristics,
such as PPB and clearance, important for the demonstration of efficacy in vivo.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 68
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
Since the methoxyl group of 23 improved inhibitory potency and the epimeric methoxyl group of
5 was tolerated with only about a 2-fold decreased potency relative to 1 (see Table 1), incorporating
2
cyclic or acyclic ketals was targeted. Modeling showed that ketals at the piperidine 3-position extend into
a binding region of the enzyme where the adenine ribose of ATP resides. As the ketals were not resolved
into optical antipodes, comparisons to 22, the racemate of 23, are made in Table 3. The MIC values for 39
with the dimethoxyketal was consistently higher than for 22 despite approximately equal enzyme
inhibitory potency versus E. coli ParE. Activity deteriorated by formation of 5-membered ring ketal 40,
especially with regards to antibacterial activity. The 6-membered ring ketal 41 improved activity relative
to the 5-membered ring ketal and 22, albeit overall lower antibacterial activity was seen. The methoxyl-
substituted 6-membered ring ketals 42 and 43 were made as a mixture of separable diastereomers, and the
diastereomer 43 with the methoxyl group oriented towards the piperidine carboxamide proved more
potent versus ParE than 42 with the methoxyl group oriented towards the piperidine thiazole substituent.
However, the measured logD’s of both compounds were lower, and MIC values were higher especially
versus S. aureus, H. influenza and M. catarrhalis. Making spiroketal compounds more lipophilic with the
dimethyl substituents of 44, the cyclopropane of 45 and the vinyl group of 46 improved antibacterial
activity. The cyclopropane ketal 45 improved inhibitory potency 20-fold relative to 22 versus ParE and
displayed lower MIC values versus H. influenzae and M. catarrhalis. Despite the interesting observations
with the ketals, improvements across combined considerations of antibacterial activity, PPB and clearance
were not achieved. Additionally, with the increased complexity of synthesis and diastereomer resolution,
the series’ prospects were diminished relative to compound 22 towards identifying a drug candidate and
the 3-methoxyl substituent remained optimal.
0
1
2
3
4
5
6
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
Intramolecular H-bonding with the carboxylic acid. Placement of substituents on the carbon adjacent
to the thiazole carboxylate was envisioned to reduce the acidity of the carboxylate as an avenue towards
reducing polarity and thereby improving bacterial membrane permeability and consequently antibacterial
activity. It appeared from modeling that this thiazole 4-position carbon would be tolerant to a variety of
ACS Paragon Plus Environment

Page 13 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substituents relative to enzyme inhibitory potency. The measured pK of 23 was 3.6 as determined by
a
potentiometric titration, and engineering a higher pK would increase the propensity to populate the
a
neutral carboxylic acid rather than the anionic carboxylate for better permeation through bacterial
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
phospholipid membranes. The inductive effect of electron donating substituents should increase pK , and
a
indeed compound 47 with a methyl substituent had a slightly higher measured pK of 3.8, which, in
a
combination with the higher lipophilicity, were presumed to contribute to the improved antibacterial
activity (Table 4). However, the added methyl substituent increased PPB and led to a propensity for
decarboxylation (data not shown). It was presumed that transient protonation of the weakly basic
aminothiazole provides an electron sink for decarboxylation, which would be promoted by such electron
donating substituents. Increasing the donating character of substituents led to greater instability; for
example, the 4-methoxyl-5-carboxylate thiazole analogue of 1 could be not isolated due to facile
decarboxylation. Placement of electron withdrawing groups such as CF H on the thiazole led to more
2
acidic compounds (pK = 3.3 for 49) and, correspondingly, lower antibacterial activity.
a
It was therefore hypothesized that an electron-withdrawing group that offered an opportunity for
an intramolecular H-bond would maintain stability relative to decarboxylation and increase the pK . It
a
was further hypothesized that rapid clearance of 2 was a key contributing factor to the low drug
concentrations in the Phase 1 evaluation due to recognition by one or more organic anion transporters
(
OATs) or multidrug resistance proteins (MRPs) that shuttle the drug out of circulation either from the
46, 47
liver or the kidneys.
Hence, thiazole 4-position substituents adjacent to the carboxylate were
envisioned to block recognition by a putative anion transporter disfavoring the interaction via steric
constraints, electronic alteration of the carboxylate acidity, and/or intramolecular H-bonding to
appropriately positioned heteroatoms. This led to the design and synthesis of compounds 50 and 51
wherein the geometry of thiazole carboxylate allowed for a seven-member ring H-bonding array with the
carbonyl substituent as shown for the carboxamide in Figure 4. Both 50 and 51 showed improved
antibacterial activity relative to 23, in particular for S. aureus. This was offset by a 3.5-fold lower f in
u
human plasma for the two compounds relative to 23. Note that although human plasma f measurements
u
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 68
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
were used routinely for assessing compounds, select compounds were assessed for f across mouse, rat
u
and dog plasma (see Table 7), and rank ordering of compounds were similar, supporting the influence on
in vivo clearance. Tracking the ratio MIC/f became a useful parameter for assessing net improvement of
u
0
1
2
3
4
5
6
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 properties acknowledging that increasing f often leads to greater susceptibility to clearance.
u
This strategy is consistent with recent literature that cautions against optimizing protein binding without
4
8-51
considering other parameters that would be affected by fu.
Protein binding variations do not
necessarily alter the exposure of pharmacologically relevant free drug concentration in that a lower fu
would also lead to lower exposure due to in vivo clearance mechanisms. Modifications that lower the
MIC/f ratio are desirable as long as they do not increase clearance. In Tables 4, 5 and 6, the calculated
u
MIC/f is included for one pathogen (MRQR S. aureus) as a guide for analogue progression. For example,
u
although compound 48 showed potent enzyme inhibition (especially versus ParE) and potent antibacterial
activity, the high logD correlated with a decreased f affording a higher MIC/f . This, combined with high
u
u
rat clearance led to diminished interest in the compound. Relative to 23, compounds 50 and 51 showed
lower MIC/f and lower clearance in mouse, rat and dog (rat clearance shown in Table 4). The lower
u
clearance for these compounds could be due partially to the lower f , but as previously mentioned, it was
u
also hypothesized that analogues with intramolecular H-bonds would diminish recognition of the thiazole
carboxylate by hepatic or renal transporters. The pK s of 50 and 51 were 3.29 and 4.7, respectively, the
a
latter being higher than the 3.6 pK of 23 lacking a thiazole 4-position substituent. The amide of 51 would
a
better accept a hydrogen bond than the acetyl group of 50, accounting for the higher pK . The pK of 50 is
a
a
presumed to be lower as the electron withdrawing nature of the acetyl group has a larger influence than its
capacity for donation of the electron pair in the intramolecular H-bond. The NMRs of 50 and 51 in
DMSO-d supported the existence of an intramolecular H-bond as the acidic hydrogen peak shifted by 0.5
6
and 0.4 ppm, respectively, downfield relative to 23. Other amide analogues besides 51 showed a
considerable shift downfield in the NMR relative to 23 including a shift of 3.4-4.0 ppm for compounds
62-67 and 69-71. The large shift for the acid proton was seen universally across the thiazole
carboxamides. Moreover, at higher pH where the carboxylic acid would be expected to be ionized as
ACS Paragon Plus Environment

Page 15 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
shown in Figure 4, there is evidence for an intramolecular H-bond between the amide NH and the
carboxylate anion. In the NMR of the sodium salts of secondary carboxamides, there was a downfield
shift of the NH resonance by a notable 4.4 ppm (compare NMR data for compounds 51, 62, 63 and 65
and their corresponding sodium salts in the Experimental Procedures). This H-bond is also associated
with improved antibacterial activity as it presumably improved bacterial membrane permeability at
physiological pH, correlating with a higher logD for the molecules. The tertiary amides of 52 and 53 are
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
capable of receiving an H-bond from the carboxylic acid at lower pH, thereby raising the pK . At higher
a
pH, the carboxylate anion H-bonding capability is removed accounting for the lower logD and higher fu
that led to less potent antibacterial activity and higher MIC/f values relative to secondary amides like 51.
u
Similarly, compounds 54-56 suffered from sufficiently high MIC values and MIC/f to diminish interest
u
in continued profiling of biological properties.
Density Functional Theory (DFT) calculations showed that amides adjacent to carboxylic acids
can form favorable intramolecular H-bonds for 5-membered ring aromatic compounds such as thiazole
but not for 6-membered aromatic rings due to the geometrical considerations. This larger C-C-C
exocyclic bond angles (135.8˚) for the carbonyl substituents on a thiazole ring allow for co-planarity with
the resultant 7-membered ring, thereby achieving improved orbital overlap for the intramolecular H-bond.
By contrast, the smaller 128.3˚ angles for such substituents on a 6-membered aromatic ring did not allow
for co-planarity without significant strain. Figure 5 shows the comparison of optimized geometries for the
substituents on thiazole and pyridine for both the carboxylic acid species and for the anionic carboxylate
species. The side view shows graphically the distortion from co-planarity for the benzene example.
Whereas the pK is increased in the case of the thiazole ring, the pK for benzoic acid decreased from 4.2
a
a
5
2
to 3.8 on incorporation of a vicinal carboxamide.
The alternative thiazole substitution pattern with the carboxylate at the thiazole 4-position and
substituents at the 5-positon (compounds 57-60, Table 4) was also investigated. A model maintaining a
salt bridge between the carboxylate of the compounds and Arg144 pointed the thiazole nitrogen atom
towards the enzyme surface. A less favorable interaction is expected between the thiazole nitrogen atom
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 68
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
and the backbone carbonyl of the GyrB Gly83 relative to a thiazole sulfur atom due to the polarizability
of the sulfur allowing for a weak non-bonding interaction as supported by analyses of the Protein Data
5
3, 54
Bank.
Hence, 57-60 were less active without improving PPB or clearance. Therefore, further
0
1
2
3
4
5
6
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
analogue work focused on identifying substituents on thiazole-5-carboxylic acids. To this end the series
of primary carboxamides akin to 51 were evaluated targeting a relatively narrow logD range of 1-2. In
line with previously mentioned observations, more polar compounds such as 70 and 72 trended to higher
MIC/f due to higher MIC values. Compounds 67-69 exhibited moderately higher rat in vivo clearance
u
relative to 51 and 62-66. Compounds 62-66 maintained a relatively narrow logD range, higher
solubilities, favorably low MIC/f ratios and low clearances, leading to their selection for further profiling
u
as drug candidates.
Table 6 shows a series of 5-thiazole carboxylic acids substituted at the 4-position with
heterocycles having nitrogen atoms positioned to accept an intramolecular H-bond (except 77). The range
of more basic heterocycles expands upon the quite similar H-bonding capabilities of the Table 5
carboxamides. With this, the heterocycles were envisioned to raise the pK of the acids leading, as
a
suggested previously, to improved bacterial membrane permeability and antibacterial activity. That the
position of the nitrogen atom is important was illustrated by pyrimidine 77, which lacks the capability for
intramolecular H-bonding and showed higher MIC values and a notably higher MIC/f ratio than the
u
isomeric pyrimidines 75 and 76. Overall, incorporation of pyrazines, pyrimidines and triazoles as in 73-76
offered a favorable combination of antibacterial activity, PPB, solubility and in vivo clearance.
Presumably a balance is needed between the basicity of the heterocycle and the resultant physicochemical
properties. The pK s were indeed elevated for the heterocycles with values of 4.5 for triazole 74, 5.2 and
a
5.7 for pyrimidines 75 and 76, and 5.4 for pyrazine 73. The more highly basic heterocycles such as
imidazole and pyridine (compounds 81, 82, 84 and 88) presumably form stronger H-bonds with the
carboxylate, raising the pK even higher. Imidazole compounds 81 and 84 had measured pKa’s of 8.0 and
a
7
.1, respectively. However, increasing the pK led to poorly soluble compounds in physiologically
a
relevant pH 7.4 buffer, which is detrimental for advancement of a drug candidate.
ACS Paragon Plus Environment

Page 17 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Towards selecting compounds for evaluation for in vivo efficacy and to enable allometric scaling
to predict human PK properties, cross-species (mouse, rat and dog) PK properties were measured (Table
7). This, in combination with in vitro data (hepatocyte clearance and PPB) was used towards predicting
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
human PK and bioavailability. The trend already described, associating improved in vivo clearance with
introduction of H-bonding substituents adjacent to the carboxylate, held across the species. For example,
the clearances of 63 were 4- to 40-fold lower than those of 23 in mouse, rat and dog, which represented a
significant effect even accounting for the lower f values for 63 across the species. To address whether the
u
lower clearances with the added substituents could be attributed at least in part to hepatic clearance, the
5
5
influence of Mrp2, the major hepatobiliary transporter for non-bile organic anions, was assessed. Six
-
compounds showed significantly lower clearance in TR rats, which lack functional Mrp2 relative to the
56, 57
wild-type Wistar rats.
By contrast, the differential in vivo clearance for moxifloxacin between Wistar
-
and TR rats was minimal despite the observation that biliary excretion and efflux into the perfusate were
-
58
diminished in the TR rats. Although there is considerable variability in the triplicate experiments
carried out, the clearances of compounds 1 and 23 without the additional thiazole substituent were less
-
differentiated from the substituted thiazole compounds in the TR rats relative to clearances in the wild-
type. Clearly, other transporters, metabolism and conjugation of the drugs contributed to overall
clearance. Towards assessing more directly the influence of pyrrolamides on Mrp2 efflux, isolated
membrane preparations of the isozymes from human and rat transformed into an ovarian cell line of the
55, 59
insect Spodoptera frugiperda (Sf9) were characterized for the ATPase activity.
Figure 6 shows the
dose response curve for 23 and 63 after normalizing for background ATPase activity, the two compounds
being examples of activators of both human MRP2 and rat Mrp2 ATPase activity and inhibitors of the
activity, respectively. MRP2 and Mrp2 are ABC-transporters, and a compound that activates ATP
hydrolysis correlates with its being a substrate. Reduction of the stimulatory effect of the reference
activators, sulfasalazine for human MRP2 and probenecid for rat Mrp2, indicate slowly transported
compounds or those that inhibit basal ATPase activity. As a negative control, an Sf9 cell line expressing
β-galactosidase showed no activation of ATPase activity for any of the compounds tested. The higher rate
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
of in vivo clearance for 23 is supported to the degree that it might activate rat Mrp2, while the lower
clearance for 63 might therefore correlate with its inhibition of Mrp2 activity. The data for six compounds
surveyed for inhibition and activation of human MRP2 and rat Mrp2 are shown in Table 9 with a rough
correlation of activators being more highly cleared in vivo and inhibitors less highly cleared. The caveat
that expression in an insect cell line may not directly relate to the situation in mammals is well
understood, but the trends do correlate. Further ramifications of the data would involve drug-drug
interactions as an inhibitor like 63 might slow the clearance of other drugs that are transported by MRP2
and Mrp2.
0
1
2
3
4
5
6
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
S. aureus neutropenic thigh infection model. Compounds 63, 65, 74 and 75 were selected for
evaluation of efficacy against S. aureus in a neutropenic mouse thigh infection model due to a
combination of superior antibacterial activity, physical properties, f and in vivo PK. Compound 63 was
u
efficacious against S. aureus (MSSA) in the model causing a dose-dependent decrease in viable counts in
the thigh (Figure 7). A maximum response of about 2-log reduction, relative to the bacterial burden at the
start of treatment and a 4-log reduction relative to the control was obtained at 30 mg/kg, whereas a dose
of 10-15 mg/kg led to stasis. Other compounds in the series were evaluated in the same model and the
data comparing efficacy at 30 mg/kg are summarized in Table 10. Compounds 63 and 74 were equally
efficacious at 30 mg/kg, resulting in a 1.5 log reduction in CFU in the thigh relative to the initial pre-
treatment inoculum. Typically the untreated control shows a 2.0 log (100-fold) growth in CFU between
infection and the start of treatment. These compounds had an MIC of 0.03 µg/mL against MSSA and
similar f values in mouse. Higher compound exposure, including a higher free AUC at 30 mg/kg for
u
compound 74, did not result in improved efficacy, indicating that the maximal response in the model may
have been achieved. Compound 65 achieved stasis at the 30 mg/kg dose and showed a further drop in
CFU at higher doses (data not shown). Compound 75 did not perform well in the infection model and
gave variable results, despite adjustments to the dosing regimen. The lower efficacy for 75 was generally
6
0
consistent with its slightly higher MIC and higher clearance, lowering the AUC and free AUC.
ACS Paragon Plus Environment

Page 19 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Biological profiling of 63. Compound 63 was chosen for progression to human clinical trials and
61
designated AZD5099. Pharmacokinetic properties and bioavailability were favorable across species,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
positioning the compound for both parenteral and oral administration. As reported for compound 1, the S.
aureus antibacterial activity of 63 correlated with the inhibition of DNA synthesis as determined by the
1
2, 34, 62
attenuation of radiolabled precursor incorporation into DNA, RNA, protein and cell wall.
Thus, at a
3
3.6
sub-lethal concentration of 63 in S. aureus cultures, incorporation of H-thymidine was reduced by 10 -
3
2.8
14
fold, considerably higher than H-uridine incorporation (10 -fold), C-leucine incorporation (no
14
14
14
decrease), C-valine incorporation (no decrease), C-acetic acid incorporation (no decrease) and C-N-
3
acetylglucosamine incorporation (no decrease). It follows that RNA biosynthesis (decreased H-uridine
incorporation) would be perturbed by the leading inhibition of DNA biosynthesis. Compound 63 did not
exhibit cross-resistance in a variety of drug-resistant bacteria isolated from the clinic and selected to
include those resistant to the fluoroquinolone levofloxacin, the β-lactam antibiotic methicillin and the
glycopeptide antibiotic vancomycin in MIC determinations shown in Table 11. As yet, there has been no
9
0
indication of pre-existing resistance to 63 due to genetic alteration of bacterial topoisomerases from any
clinical isolates. The frequencies of spontaneous resistance to 63 in multiple isolates of S. pneumoniae
-10
and S. aureus were all less than the detection limit (<9.6 x 10 ) at 4 and 8 times the concentration that
prevented confluent bacterial growth, with no resistant variants emerging. The frequencies of spontaneous
resistance were also very low in five Enterococcus spp. isolates tested, with the rates being less than or
-9
equal to 3.7 x 10 , and resistant variants could only be isolated in three strains at the 4 times the
concentration that prevented confluent bacterial growth. In the two strains of S. pyogenes tested, the
-7
-10
frequencies of spontaneous resistance to 63 ranged from 1.4 x 10 and less than 3.4 x 10 at 4 and 8
times the concentration that prevented confluent bacterial growth, respectively. Resistant variants
developed at twice the MIC from four S. aureus isolates from the AstraZeneca Research Collection were
examined by extraction of the genomic DNA followed by amplification and sequencing of the type II
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 68
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
topoisomerase subunit genes GyrB and ParE. Increases in MICs of 8- to 16-fold were observed, whereas
there was no shift in the MIC of linezolid, consistent with the target-based mode of resistance (Table 12).
Sequence analysis identified that the decrease in susceptibility coincided with a mutation in the gyrB
encoded polypeptide, suggesting that GyrB is the primary target of 63 in S. aureus. Two different amino
acid mutations were observed: T173A in the ARC516 and ARC2381 variants and R144I in the ARC1692
and ARC2381 variants. Both mutations are expected to disrupt key interactions with the inhibitor as
Thr173 was H-bonded to the water molecule otherwise H-bonded the pyrrole carboxamide and Arg144
formed a salt-bridge with the thiazole carboxylate. Both the T173A and R144I variations were seen
0
1
2
3
4
5
6
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
previously for pyrrolamides and have been linked to novobiocin and coumermycin A1 resistance in S.
6
3
aureus, where the MIC values against these compounds increased 8- to 128-fold. The T173A and R144I
mutants had also been identified in variants resistant to the urea benzimidazole and urea pyrimidine
scaffolds that target the ATPase subunits of both GyrB and ParE in S. aureus, suggestive of a more
12,14
balanced dual targeting.
The nature of the resistant mutants to 63 has not been fully characterized
since whole genome sequencing was not carried out, and what was produced in the laboratory does not
necessarily correspond to what might develop in the clinic. Overall, the low frequency of spontaneous
resistance in these species suggests a reduced probability of rapid development of resistance against 63 in
clinical practice.
While 63 was a potent inhibitor of bacterial type II topoisomerases, a greater than 10,000-fold
selectivity was measured versus the human topoisomerase IIα where the IC in a gel-based supercoiling
5
0
assay was greater than 50 µM. The compound showed no signs of mutagenicity at the highest
6
4
concentrations tested in an Ames mutagenicity assay, an in vitro micronucleus assay using mouse
65
66
lymphoma cells, and an in vitro mouse lymphoma TK assay. Compound 63 showed no hERG
inhibition or inhibition of other ion channels at the highest concentration tested (100 µM), representing a
greater than 200-fold margin to predicted free Cmax. Compound 63 showed no inhibition at the highest
concentration of 50 µM across a series of five of the most prevalent human cytochrome P450 enzymes
ACS Paragon Plus Environment

Page 21 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
Cyp1A2, Cyp2C19, Cyp2C9, Cyp2D6 and Cyp3A4) mitigating one mode of deleterious drug-drug
interactions. A more thorough analysis of the preclinical in vitro and in vivo safety characteristics of 63
will be published separately.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSIONS
Compound 63 is a novel antibacterial that entered Phase 1 clinical trials in humans via intravenous
administration to target Gram-positive infections. The pyrrolamide scaffold was initiated in a fragment-
based approach that identified the pyrrole pharmacophore key to binding to an active site aspartate
residue. It is noteworthy that since initial disclosures of the pyrrolamides, the structure of the natural
product antibiotic kibdelomycin was shown to have a dichloropyrrole carboxamide moiety akin to that in
the pyrrolamides. Kibdelomycin also showed Gram-positive antibacterial activity ascribed to its inhibition
6
7
of DNA gyrase and binding to its ATP binding site. It is tempting to speculate that the dichloropyrrole
part of the much larger kibdelomycin molecule (molecular weight = 940) binds in a similar fashion to
Asp81 as seen with the pyrrolamides. The starting pyrrole discovered via fragment-based screening has
been elaborated considerably before arriving at 63. Exploration of the 3-position of the piperidine scaffold
identified the methoxyl substituent that was found to be optimal in terms of intrinsic enzyme inhibitory
activity, translation to antibacterial activity and maintenance of optimal physical properties. It was
important to position a carboxylic acid for a salt bridge interaction with Arg84 in a region that is not
associated with ATP binding to minimize off-target toxicity that might be associated with other ATP-
utilizing enzymes. There is sufficient differentiation from the ATP binding region of human
topoisomerases that selectivity for pyrrolamides was achieved early during the optimization program
culminating in a clear selectivity for bacterial topoisomerases for 63. Problems were encountered,
however, with higher clearances for the initially reported pyrrolamides where a lone carboxylic acid
substituent was placed on a thiazole ring. Reasoning that anion transporters might recognize the
carboxylic acid, the compounds were designed to have an H-bond accepting moiety on the carbon
adjacent to the carboxylic acid. Modeling (and NMR data) showed a clear propensity to form such an
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 68
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
intramolecular H-bond with the acid and H-bond accepting group attached vicinal on the thiazole ring.
Importantly, in vivo clearances across species were improved in line with diminished ATPase activity of
insect cell line expressed human MRP2 and rat Mrp2 and decreased in vivo clearance recorded with Mrp2
-
0
1
2
3
4
5
6
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
knockout TR rats, thereby supporting the hypothesis that transporter recognition had been partially
mitigated. Although only a mouse neutropenic thigh S. aureus infection model is disclosed herein, 63
demonstrated efficacy in a variety of animal models of infection, positioning the compound as a Gram-
positive agent for treatment of hospital lung and skin infections, especially those involving drug resistant
organisms such as MRSA, VRE and those that are quinolone resistant. A more extensive characterization
of the pre-clinical PK-PD characteristics of the compound, the predicted PK and dose estimates and the
results of the human clinical trials will be presented in a separate publication.
EXPERIMENTAL SECTION
General Considerations All of the solvents and reagents used were obtained commercially and
1
used as such unless noted otherwise. H NMR spectra were recorded in CDCl or DMSO-d solutions at
3
6
300 K using a Brucker Ultrashield 300 MHz instrument or a Brucker Ultrashield 400 MHz instrument.
13
C NMR spectra were recorded in DMSO-d solutions at 300 K and 126 MHz using a Brucker DRX-500
6
500 MHz instrument with a QNP cryoprobe or at 101 MHz using a Brucker Ultrashield 400 MHz
1
9
instrument or at 75.5 MHz using a Brucker Ultrashield 300 MHz instrument. F NMR spectra were
recorded at 282 MHz in CDCl or DMSO-d solutions at 300 K using a Brucker Ultrashield 300 MHz
3
6
1
13
instrument. Chemical shifts are reported as parts per million relative to TMS (0.00) for H and C NMR
1
9
and CFCl for F NMR. High-resolution mass spectra (HRMS) were obtained using a hybrid quadrupole
3
+
time-of-flight mass spectrometer (microTOFq II, Bruker Daltonics) in ESI mode. Silica gel
chromatographies were performed on an ISCO Combiflash Companion Instruments using ISCO
RediSep Flash Cartridges (particle size: 35-70 microns) or Silacycle SiliaSep Flash Cartridges
(
particle size: 40-63 microns). Preparative reverse phase HPLC was carried out using YMC Pack ODS-
AQ (100 × 20 mm ID, S-5 ꢀ particle size, 12 nm pore size) on Agilent instruments. All compounds tested
ACS Paragon Plus Environment

Page 23 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
possessed a purity of ≥ 95%. When not indicated, compound intermediates and reagents were purchased
from chemical supply houses. All final compounds (Compounds 1-2, 23-90) were determined to be
greater than 95% pure via analysis by one of two methods. First, reversed phase LC-MS was used with a
Varian Polaris C18A, 2mm x 50mm, 3mm particle size column with DAD and ELSD and aUPLC HSS
T3, 2.1 x 30 mm, 1.8 um column (peak retention time, RT, in minutes). An Agilent HP1100 (Wimington,
DE, USA) LC system was used with gradient of 5 to 95% acetonitile in water with 0.1% formic acid over
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
.5 minutes at 1 mL/min, then hold at 95% acetonitrile at 1.5 mL/min to 6 minutes. Injection volume was
µL and column temperature 30 °C. Detection was based on electrospray ionization (ESI) in positive and
2
negative polarity using Waters ZQ mass spectrometer (Milford, MA, USA), diode-array UV detector
from 210 to 400 nm, and evaporative light scattering detector (Sedex 75, Sedere, Alfortville Cedex,
France). Second, reversed phase UPLC-MS (rentention times, RT, in minutes) was used with a Waters
Acquity UPLC instrument with DAD and ELSD and a UPLC HSS T3, 2.1 x 30 mm, 1.8 um column and
a gradient of 2 to 98% acetonitrile in water with 0.1% formic acid over 2.0 minutes at 1 mL/min. Injection
volume was 1 µL and the column temperature was 30 °C. Detection was based on electrospray ionization
(ESI) in positive and negative polarity using Waters ZQ mass spectrometer (Milford, MA, USA), diode-
array UV detector from 210 to 400 nm, and evaporative light scattering detector (Sedex 75, Sedere,
Alfortville Cedex, France).
Ethyl 3-hydroxy-4,4-dimethoxypiperidine-1-carboxylate (Scheme 2) A solution of ethyl 4-
oxopiperidine-1-carboxylate (26.4 mL, 29.96 g, 175 mmol) in dry MeOH (75 mL) was added via syringe
to a stirred solution of KOH (42 g, 752 mmol) in dry MeOH (100 mL) at 0 °C and under an atmosphere
of N . After stirring form 30 min, iodobenzene diacetate (84.6 g, 262 mmol) was added portionwise over
2
90 min maintain a temperature near 0 °C. The reaction mixture was stirred overnight with warming to rt.
After removal of solvent, the residue was diluted with water, which was extracted 3 times with EtOAc.
The organic layers were combined, dried (MgSO ), and concentrated. Chromatography on silica gel (10-
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 68
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
5% EtOAc gradient in hexanes) afforded 26.74 g (66%) of the title compound as a pale yellow oil. MS
+
1
(
ESI) m/z (M+H) : 233 for C H NO . H NMR (300 MHz, CDCl ) δ 1.22 (t, 3H), 1.69-1.86 (m, 2H),
10 19 5 3
2
.20 (m, 2H), 2.86 (t, 1H), 3.22 (s, 3H), 3.23 (s, 3H), 3.74 (m, 1H), 3.95 (m, 2H), 4.11 (q, 2H).
0
1
2
3
4
5
6
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
Ethyl 3,4,4-trimethoxypiperidine-1-carboxylate (8, Scheme 2) A solution of the preceding compound
(52 g, 223 mmol) in 125 mL THF was added over 30 min to a stirred solution of NaH (6.96 g, 290 mmol)
in 100 mL dry THF at 0 °C and under an atmosphere of N . The resulting solution was stirred at 0 °C
2
under for 15 min before adding iodomethane (18.0 mL, 290 mmol) via syringe over 5 minutes. The
reaction was stirred overnight, gradually warming to rt. The reaction mixture was quenched with a small
volume of water and then concentrated under vacuum. The residue was diluted with water, which was
extracted with 3 times with EtOAc. The organic layers were combined, dried (MgSO ), and concentrated.
4
The crude material (54.9 g) was used without further purification in the subsequent reaction. MS (ESI)
+
1
m/z (M+H) : 247 for C H NO . H NMR (300 MHz, CDCl ) δ 1.24 (t, 3H), 1.72-1.83 (m, 2H), 2.82 (m,
11
21
5
3
2H), 2.98 (t, 1H), 3.20 (s, 3H), 3.21 (s, 3H), 3.41 (s, 3H), 4.10 (q, 2H), 4.20-4.37 (m, 2H).
Ethyl 3-methoxy-4-oxopiperidine-1-carboxylate (Scheme 2) A stirred solution of the preceding
compound (54.9 g, 223 mmol) and 190 mL aq H SO in THF (300 mL) was heated at 60 °C for 2 h. After
2
4
cooling to rt THF solvent was removed. Solid NaHCO was added to the aq solution until it became basic.
3
The solution was saturated with NaCl and extracted with CH Cl 3 times. The organic layers were
2
2
combined, dried (MgSO ) and concentrated to afford 44.6 g of the title compound used without further
4
+
1
purification in the subsequent step. MS (ESI) m/z (M+H) : 201 for C H NO . H NMR (300 MHz,
9
15
4
CDCl ) δ 4.17 (q, 2H), 4.05 (m, 2H), 3.69 (m, 1H), 3.45 (s, 3H), 3.30-3.4 (m, 2H), 2.4-2.60 (m, 2H), 1.28
3
(t, 3H).
Ethyl (3S,4R)-rel-4-(benzylamino)-3-methoxypiperidine-1-carboxylate (9, Scheme 2) Benzylamine
(
0.130 mL, 1.19 mmol) was added dropwise to a solution of the preceding compound (200 mg, 0.99
mmol) in 1 mL THF maintaining a temperature between 0-10 ˚C. After 30 min stirring at 0 ˚C, sodium
triacetoxyborohydride (232 mg, 1.09 mmol) was added portionwise over 5 min, and the mixture was
ACS Paragon Plus Environment

Page 25 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
stirred at 5 °C for 4 h and overnight at rt. After quenching with water (1.2 mL) and stirred for 10 min at 5
°C, 5 mL of 2N aq. NaOH was added. The mixture was diluted with EtOAc and washed with water. The
combined aq layers were extracted with additional EtOAc, which was washed with brine. The combined
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
organic layers were dried (MgSO ), and solvent was removed to afford the title compound (256 mg, 88
4
+
1
%
) as an oil. MS (ESI) m/z (M+H) : 293 for C H N O . H NMR (300 MHz, CDCl ) δ 1.16 (t, 3H),
16 24 2 3 3
1
.50 (m, 2H), 1.89 (s, 2H), 2.66 (m, 1H), 2.89 (dd, 2H), 3.26 (s, 3H), 3.67-3.83 (m, 3H), 3.99 (m, 3H),
.18-7.37 (m, 5H).
7
4
3
Ethyl (3R,4S)-rel-4-amino-3-methoxypiperidine-1-carboxylic acid hydrochloride salt (4, Scheme 2)
Ammonium formate (31.50 g, 500 mmol) was added as a solid to a stirred solution of the preceding
compound (36.45 g, 125 mmol) and 10% palladium on activated carbon (50% wet; approximately 4 g) in
methanol (250 mL) at rt and under an atmosphere of N . The mixture was heated to 70 °C overnight. The
2

reaction mixture was filtered through Celite , and solvent was removed from the filtrate. The residue was
diluted with water, which was extracted with a solution of ~3% methanol in chloroform (4 times). The
combined organic layers were dried (MgSO ) and concentrated to afford 24.18 g (96%) of the title
4
+
compound as an off-white solid. MS (ESI) m/z (M+H) : 202 for C H N O .
9
18
2
3
Ethyl (3R,4S)-rel-4-{[(benzyloxy)carbonyl]amino}-3-methoxypiperidine-1-carboxylate (10, Scheme
) Benzyl chloroformate (3.3 mL) was added dropwise to a cold solution of 4 (5 g, 24.9 mmol) in sat. aq.
2
NaHCO . The mixture was stirred at rt for 14 h. The white precipitate was filtered, washed with water,
3
+
dried in vacuo to give the title compound as a white solid (6.66 g, 80%). MS (ESI) m/z (M+Na) : 359 for
1
C H N O ; H NMR (300 MHz, CDCl ) δ 1.21 (t, 3H), 1.52-1.67 (m, 2H), 3.08 (m, 2H), 3.28 (s, 3H),
1
7
24
2
5
3
3
.41 (s, 2H), 3.74-3.92 (m, 2H), 3.96 (m, 2H), 4.14 (m, 1H), 5.10 (s, 2H), 7.24 (d, 1H), 7.44 (m, 5H).
Ethyl (3R,4S)-4-{[(benzyloxy)carbonyl]amino}-3-methoxypiperidine-1-carboxylate (10a, Scheme 2)
and ethyl (3S,4R)-4-{[(benzyloxy)carbonyl]amino}-3-methoxypiperidine-1-carboxylate (10b, Scheme
2) The preceding compound 10 (6.2 g, 18.5 mmol) was separated into its enantiomers by chiral
chromatography over a Chiralcel OJ column (eluant: hexanes/MeOH/EtOH;70/15/15;0.1% diethylamine;
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 68
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
complete separation was achieved, see Supplementary Materials). The fractions corresponding to the first
chromatographic peak 10a (3R,4S (-)isomer) were collected and evaporated yielding the title compound
as a white solid (2.62 g, 42% recovery). The fractions corresponding to the second chromatographic peak
10b (3S,4R (+)isomer) were collected and evaporated yielding the title compound as a white solid (2.71 g,
0
1
2
3
4
5
6
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
+
44% recovery). MS (ESI) m/z (M+Na) : 359 for C H N O ; NMR: 1.21 (t, 3H), 1.5-1.7 (m, 2H), 3.08
1
7
24
2
5
(
m, 2H), 3.28 (s, 3H), 3.41 (s, 2H), 3.74-3.92 (m, 2H), 3.96 (m, 2H), 4.14 (m, 1H), 5.10 (s, 2H), 7.24 (d,
H), 7.44 (m, 5H).
1
Ethyl (3S,4R)-4-amino-3-methoxypiperidine-1-carboxylate hydrochloride salt (4b, Scheme 2) A
mixture of ethyl (3S,4R)-4-(benzylamino)-3-methoxypiperidine-1-carboxylate 10b (3.98 g, 11.8 mmol),
50 mL 1N HCl and 600 mg 10% Pd on carbon in 100 mL MeOH was degassed and placed under an H2
atmosphere for 3 h. The mixture was filtered through a bed of diatomaceous earth, and the filtrated was
+
concentrated to afford 2.8 g (96%) of the title compound as an off-white solid. MS (ESI) m/z (M+H) :
2
02 for C H N O .
9 18 2 3
Ethyl (3R,4S)-4-amino-3-methoxypiperidine-1-carboxylate hydrochloride salt (4a, Scheme 2)
Prepared as described for 4b from (3R,4S)-4-(benzylamino)-3-methoxypiperidine-1-carboxylate 10a (3.0
g, 8.85 mmol), 50 mL 1N HCl and 600 mg 10% Pd on carbon in 100 mL MeOH to afford 2.1 g (96%) of
+
product. MS (ESI) m/z (M+H) : 202 for C H N O .
9
18
2
3
Ethyl (3S,4R)-rel-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-methoxy-piperidine-1-
carboxylate (Scheme 1: 5, R1=OMe, R3 = Et, racemic) 3,4-Dichloro-5-methyl-1H-pyrrole-2-carboxylic
acid (304 mg, 1.57) was dissolved in anhydrous DMF. HATU (596 mg, 1.57 mmol), HOAt (213 mg, 1.57
mmol) and DIEA (274 ꢀl, 1.76 mmol) were added and stirred at ambient for 15 min. Ethyl (3R,4S)-rel-4-
amino-3-methoxypiperidine-1-carboxylic acid hydrochloride salt (317 mg,1.33 mmol) was added and the
mixture was stirred at ambient temperature for 18 h. The mixture was diluted with EtOAc and washed
with water, 1 N HCl, aq NaHCO , water and brine before drying over Na SO . The solution was
3
2
4
+
concentrated to give the title compound as a brown solid (503 mg, 100%). MS (ESI) m/z (M+H) : 378,
ACS Paragon Plus Environment

Page 27 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
3
80 for C H Cl N O ; H NMR (300 MHz, DMSO-d ) δ 1.29 (t, 3H), 1.79 (m, 2H), 2.27 (s, 3H), 3.12
15 21 2 3 4 6
(
m, 2H), 3.30 (s, 3H), 3.37 (m, 1H), 3.83-4.16 (m, 5H), 7.25 (d, 1H), 12.23 (s, 1H).
Ethyl (3S,4R)-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-methoxy-piperidine-1-
carboxylate (Scheme 1: 5, R1=OMe, chiral) NMM (43.7 mL, 354 mmol) was slowly added to a mixture
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4
2
of 3,4-dichloro-5-methyl-1H-pyrrole-2-carboxylic acid (19.6 g, 101 mmol), ethyl (3S,4R)-4-amino-3-
methoxypiperidine-1-carboxylate 4b (1.2 g, 5.9 mmol) and HOBt (13.6 g, 101 mmol) in 400 mL
anhydrous CH Cl . After 1 h stirring, 39.9 g (182 mmol) EDC was added, and stirring was continued
2
2
overnight. The mixture was washed 3 times with water, 1 N HCl, sat. aq. NaHCO and brine. Drying
3
(
MgSO ) and removal of solvent gave a yellow solid that was triturated with EtOAc to afford 31.8 g
4
+
1
(83%) of product as a whte solid. MS (ESI) m/z (M+H) : 378, 380 for C H Cl N O ; H NMR (300
1
5
21
2
3
4
MHz, DMSO-d ) δ 1.29 (t, 3H), 1.79 (m, 2H), 2.27 (s, 3H), 3.12 (m, 2H), 3.30 (s, 3H), 3.37 (m, 1H),
6
3
.83-4.16 (m, 5H), 7.25 (d, 1H), 12.23 (s, 1H).
Ethyl (3R,4S)-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-methoxy-piperidine-1-
carboxylate (Scheme 1: 5, R1 = OMe, chiral) Prepared as described for the preceding compound from 94
4
2
mg (0.49 mmol) of 3,4-dichloro-5-methyl-1H-pyrrole-2-carboxylic acid, ethyl (3R,4S)-4-amino-3-
methoxypiperidine-1-carboxylate 4a (100 mg, 0.49 mmol), 108 mg (0.49 mmol) HATU, 67 mg (0.49)
+
HOAt, 0.86 mL (4.9 mmol) of DIEA to afford 76 mg (41%) of product. MS (ESI) m/z (M+H) : 378, 380
1
for C H Cl N O ; H NMR (300 MHz, DMSO-d ) δ 1.32 (t, 3H), 1.74 (m, 2H), 2.31 (s, 3H), 3.14 (m,
15
21
2
3
4
6
2
3
1
H), 3.18-3.52 (s, 3H), 3.75-4.44 (m, 5H), 7.30 (d, 1H), 12.39 (s, 1H).
,4-Dichloro-N-(3S,4R)-rel-[3-methoxypiperidin-4-yl]-5-methyl-1H-pyrrole-2-carboxamide (Scheme
: racemic 6a, R1 = OMe) ethyl (3S,4R)-rel-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-
methoxy-piperidine-1-carboxylate (3.85 g, 10 mmol) was suspended in anhydrous CH CN.
3
Iodotrimethylsilane (2.2 mL, 15.5 mmol) was added slowly, and the reaction was heated at reflux for 4 h.
The crude reaction mixture was diluted with water and acidified with 1N HCl to pH 3. The solution was
extracted with EtOAc. The aq layer was basified with 50% NaOH to pH 10. The aq layer as saturated
with sodium chloride and extracted with THF, dried with MgSO and concentrated to afford the title
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 68
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
+
1
compound as a tan solid (2.1 g, 69%). MS (ESI) m/z (M+H) : 306 for C H Cl N O ; H NMR (300
1
2
17
2
3
2
MHz, DMSO-d ) δ 1.61 (d, J=3.77 Hz, 2H), 1.76 (dt, J=6.50, 3.16 Hz, 1H), 2.16-2.20 (m, 3H), 2.56 -
6
2
.69 (m, 2H), 2.90 (d, J=13.19 Hz, 1H), 3.18 (dd, J=13.75, 3.01 Hz, 1H), 3.30 - 3.35 (m, 3H), 3.56 - 3.64
0
1
2
3
4
5
6
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
(m, 1H), 4.04 - 4.15 (m, J=8.10, 7.72, 7.72, 3.01 Hz, 1H), 7.14 (d, J=8.29 Hz, 1H).
3,4-Dichloro-N-(3S,4R)-[3-methoxypiperidin-4-yl]-5-methyl-1H-pyrrole-2-carboxamide
hydrochloride salt (Scheme 1: chiral 6a, R1 = OMe) N was bubbled through a solution of 94.3 g (249
2
mmol) ethyl (3S,4R)-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-methoxy-piperidine-1-
carboxylate in 1.8 L CH CN for 10 min. Iodotrimethylsilane (60 mL, 411 mmol) was added dropwise,
3
and the mixture was heated at reflux for 90 min. After cooling to rt, aq. sat. NaHSO was added, CH CN
3
3
was removed under reduced pressure. The residue was dissolved in hot MeOH and the mixture was
acidified with conc. HCl was added until the mixture was acidic (below pH = 7). Insoluble material was
filtered, and the filtrate was concentrated to afford a yellow solid that was triturated with EtOAc to afford
+
1
the title compound (61 g, 71%) as a white solid. MS (ESI) m/z (M+H) : 306; H NMR (300 MHz,
DMSO-d ) δ 1.81 (m, 2H), 2.19 (s, 3H), 3.0-3.2 (m, 3H), 3.40 (s, 3H), 3.6 (m, 2H), 4.24 (m, 1H), 7.22 (d,
6
1H), 12.2 (s, 1H).
Methyl 2-[(3S,4R)-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-methoxy-1-
piperidyl]thiazole-5-carboxylate (methyl ester of 23) A solution of 2.75 g (8.98 mol) chiral 6a, 1.99 g
(
8.98 mmol) of methyl 2-bromothiazole-5-carboxylate and 3.9 mL (22.4 mmol) of DIEA in NMP (20
mL) was heated at 150 °C for 20 min. The mixture was partitioned between water and EtOAc. The
organic layer was separated and washed with water and brine before being dried (MgSO ) and
4
concentrated to afford a solid that was purified by chromatography on silica gel (0-50% gradient of
+
EtOAc in CH Cl ) to afford 2.8 g (70%) of the title compound. MS (ESI) m/z (M+H) :447 for
2
2
1
C H Cl N O S; H NMR (300 MHz, DMSO-d ) δ 1.87 (m, 2H), 2.05 (m, 2H), 2.27 (s, 3H), 3.01 (m,
1
7
20
2
4
4
6
2
H), 3.25 (s, 3H), 3.65 (m, 1H), 3.83 (s, 3H), 4.12 (m, 1H), 4.46 (m, 2H), 7.40 (d, 1H), 8.00 (s, 1H), 12.38
(
s, 1H).
ACS Paragon Plus Environment

Page 29 of 68
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Methyl 5-(2-aminothiazol-4-yl)-2-[(3S,4R)-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-
carbonyl)amino]-3-methoxy-1-piperidyl]thiazole-4-carboxylate (methyl ester of 73) Prepared as
described for the methyl ester of 23 from 100 mg (0.29 mmol) of chiral 6a, 75 mg (0.27 mmol) of methyl
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2'-amino-2-chloro-(4,4'-bithiazole)-5-carboxylate (described in Supporting Information) and 61 mL
+
NaHCO to afford 122 mg (77%) of the title compound.MS (ESI) m/z (M+H) : 545, 547 for
3
1
C H Cl N O S ; H NMR (300 MHz, DMSO-d ) δ 12.16 (s, 1H), 7.18 (s, 1H), 7.15 (s, 2H), 7.07 (s,
2
0
22
2
6
4
2
6
1
H), 4.22-4.35 (m, 2H), 3.95-4.05 (m, 1H), 3.68 (s, 3H), 3.55 (m, 1H), 3.40 (m, 1H), 3.38 (s, 3H), 3.34
m, 3H), 2.18 (s, 3H), 1.7-1.8 (m, 2H).
-[(3S,4R)-4-[(3,4-Dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-methoxy-1-piperidyl]-5-
(
2
ethoxycarbonyl-thiazole-4-carboxylic acid (18, Scheme 4) A mixture of 34.5 g (101 mmol) of chiral 6a,
23.5 g (101 mmol) of 5-ethyl 2-chloro-4,5-thiazolecarboxylate (17, see Supporting Information) and 53
mL (303 mmol) DIEA in 300 mL NMP was heated at 100 ˚C for 85 min. After cooling to rt, the reaction
mixture was poured slowly into 1 L of aq. 1N HCl. The mixture was stirred for 30 min, and solids were
collected by filtration, washed with water and dried in vacuo to afford 44.6 g (87%) of the title
+
1
compound. MS (ESI) m/z (M+H) : 505 for C H Cl N O S; H NMR (300 MHz, DMSO-d ) δ 1.24 (t,
1
9
22
2
4
6
6
3H) 1.74 (s, 2H) 2.18 (s, 3H) 3.36 (s, 4H) 3.56 (s, 1H) 3.93 (s, 1H) 4.19 (s, 3H) 4.29 (s, 2H) 7.13 (s, 1H)
12.11 (s, 1H) 13.51 (s, 1H).
Ethyl 2-[(3S,4R)-4-[(3,4-dichloro-5-methyl-1H-pyrrole-2-carbonyl)amino]-3-methoxy-1-piperidyl]-
4
-(methylcarbamoyl)thiazole-5-carboxylate (ethyl ester of 51) A solution of 18 1.8 g, 3.56 mmol),
methylamine (1.8 mL, 2M solution in THF), HATU (1.35 g, 3.56 mmol) and triethylamine (0.55 mL,
.96 mmol) was stirred at rt for 30 min. The crude reaction mixture was slowly poured into water and the
resulting white precipitate was filtered, washed with water and dried under vacuum to yield pure product
3
+
1
(1.2 g, 65%). MS (ESI) m/z (M+H) : 518 for C H Cl N O S; H NMR (300 MHz, DMSO-d ) δ 1.21 (s,
2
0
25
2
5
5
6
3
H) 1.74 (s, 2H) 2.18 (s, 3H) 2.69 (s, 3H) 3.17 (s, 2H) 3.36 (s, 3H) 3.55 (s, 1H) 3.98 (s, 1H) 4.14 (s, 2H)
.27 (s, 2H) 7.13 (s, 1H) 8.33 (s, 1H) 12.17 (s, 1H).
4
ACS Paragon Plus Environment