Synthesis and Characterization of Tetrahydropyran-Based Bacterial Topoisomerase Inhibitors with Antibacterial Activity against Gram-Negative Bacteria

Article abstract of DOI:10.1021/acs.jmedchem.6b01831

There is an urgent unmet medical need for novel antibiotics that are effective against a broad range of bacterial species, especially multidrug resistant ones. Tetrahydropyran-based inhibitors of bacterial type II topoisomerases (DNA gyrase and topoisomerase IV) display potent activity against Gram-positive pathogens and no target-mediated cross-resistance with fluoroquinolones. We report our research efforts aimed at expanding the antibacterial spectrum of this class of molecules toward difficult-to-treat Gram-negative pathogens. Physicochemical properties (polarity and basicity) were considered to guide the design process. Dibasic tetrahydropyran-based compounds such as 6 and 21 are potent inhibitors of both DNA gyrase and topoisomerase IV, displaying antibacterial activities against Gram-positive and Gram-negative pathogens (Staphylococcus aureus, Enterobacteriaceae, Pseudomonas aeruginosa, and Acinetobacter baumannii). Compounds 6 and 21 are efficacious in clinically relevant murine infection models.

Full text of DOI:10.1021/acs.jmedchem.6b01831

Subscriber access provided by University of Newcastle, Australia
Article
Synthesis and Characterization of Tetrahydropyran-
Based Bacterial Topoisomerase Inhibitors with
Antibacterial Activity Against Gram-Negative Bacteria
Jean-Philippe Surivet, Cornelia Zumbrunn, Thierry Bruyère, Daniel Bur, Christopher Kohl, Hans
H. Locher, Peter Seiler, Eric A. Ertel, Patrick Hess, Michel Enderlin-Paput, Stéphanie Enderlin-
Paput, Jean-Christophe Gauvin, Azely Mirre, Christian Hubschwerlen, Daniel Ritz, and Georg Rueedi
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01831 • Publication Date (Web): 13 Apr 2017
Downloaded from http://pubs.acs.org on April 15, 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
Synthesis and Characterization of TetrahydropyranꢀBased Bacterial Topoisomerase
Inhibitors with Antibacterial Activity Against GramꢀNegative Bacteria
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
Authors: JeanꢀPhilippe Surivet,* Cornelia Zumbrunn, Thierry Bruyère, Daniel Bur, Christopher Kohl,
Hans H. Locher, Peter Seiler, Eric A. Ertel, Patrick Hess, Michel EnderlinꢀPaput, Stéphanie Enderlinꢀ
§
Paput, JeanꢀChristophe Gauvin, Azely Mirre, Christian Hubschwerlen, Daniel Ritz and Georg Rueedi
Actelion Pharmaceuticals Limited, Gewerbestrasse 16, CH-4123 Allschwil
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
There is an urgent unmet medical need for novel antibiotics that are effective against a broad range of
bacterial species, especially multiꢀdrug resistant ones. Tetrahydropyranꢀbased inhibitors of bacterial
type II topoisomerases (DNA gyrase and topoisomerase IV) display potent activity against Gramꢀ
positive pathogens and no targetꢀmediated crossꢀresistance with fluoroquinolones. We report our
research efforts aimed at expanding the antibacterial spectrum of this class of molecules towards
difficultꢀtoꢀtreat Gramꢀnegative pathogens. Physicochemical properties (polarity and basicity) were
considered to guide the design process. Dibasic tetrahydropyranꢀbased compounds such as 6 and 21,
are potent inhibitors of both DNA gyrase and topoisomerase IV, displaying antibacterial activities
against Gramꢀpositive and Gramꢀnegative pathogens (Staphylococcus aureus, Enterobacteriaceae,
Pseudomonas aeruginosa and Acinetobacter baumannii). Compounds 6 and 21 are efficacious in
clinically relevant murine infection models.
0
1
2
3
4
5
6
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
1
Antimicrobial resistance has been recognized as a major threat to human health on a worldwide scale.
Of particular concern are multiꢀdrug resistant (MDR) Gramꢀnegative bacteria (GNB) that cause an
2
increasing number of infections no longer treatable with available chemotherapies. Despite financial
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
3
incentives to promote the discovery and development of new antibiotics, the number of antimicrobials
that were FDAꢀapproved or that are in advanced clinical development (phase 2 or 3) for the treatment
4
5
of Gramꢀnegative infections has remained dramatically low. With rare exceptions, all of these antiꢀ
6
GNB drugs act only on a handful of well characterized targets and belong to old classes discovered
more than fifty years ago (βꢀlactams, tetracyclines, fluoroquinolones and aminoglycosides). Therefore,
crossꢀresistance issues that limit their bacterial coverage and clinical utility, are widespread. To
minimize these challenges and address the growing medical need, truly novel antibacterial drugs that
act via a novel mechanism of action, inhibiting either an unexploited target or addressing a novel site
7
on a validated one, must be discovered.
As many of the targeted cellular pathways components are located in the bacterial cytoplasm,
antibacterial drug discoverers have to manage to pair target potency with sufficient cytoplasmic drug
accumulation. While broad onꢀtarget activity can be assessed at an early phase of the discovery
process considering target site conservation across bacterial species, predicting accumulation of a drug
in bacteria and especially in GNB remains elusive. Indeed, the Gramꢀnegative cell envelope differs
largely from the one of Gramꢀpositive bacteria (GPB), limiting drastically the accessibility of the cell
8
interior to chemicals. Additionally, Gramꢀnegatives have a higher propensity to actively expel
antibiotics than GPB, as a result of a wider array of efflux pumps. These two mechanisms that
9
preclude drug accumulation and decrease susceptibility of GNB, are intimately related to the
10,11
physicochemical properties of antibiotics.
Careful monitoring and tuning of these physicochemical
properties is required to increase accumulation in GNB and enable broad antibacterial coverage.
Among the few biological targets amenable to broad antibacterial activity are DNA gyrase and
topoisomerase IV (Topo IV). These two type II bacterial DNA topoisomerases are heterotetramers of
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
GyrA
2 2 2 2
GyrB (DNA gyrase) and ParC ParE (Topo IV) in Escherichia coli (E. coli), which maintain
12,13
DNA in a proper topological state during DNA replication and transcription.
DNA topoisomerase
function is blocked by fluoroquinolones (FQs) that have been in clinical use since the 1960's for the
treatment of a broad range of infections, including those caused by GNB (e.g urinary tract infections
due to E. coli). Several structurally unrelated classes of DNA topoisomerase inhibitors acting at
0
1
2
3
4
5
6
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
4
binding sites different from the one of FQs have been reported, but only few of them display broad
antibacterial activity. Among those, pyrazole derivative 1 is a potent and dual DNA topoisomerase
inhibitor, with antibacterial activity against both GPB and GNB, while displaying no crossꢀresistance
1
5,16
with FQs.
Figure 1: Chemical structures of some representative nonꢀfluoroquinolone bacterial topoisomerase
inhibitors.
In previous disclosures, we have detailed the discovery and optimization work conducted on novel
(
nonꢀfluoroquinolone) bacterial topoisomerase inhibitors (NBTIs) centered around a tetrahydropyran
1
7
core (henceforth dubbed THPꢀbased NBTIs). This work led to the selection of compound 2 (ACTꢀ
387042) as a preclinical development candidate for the treatment of infections caused by Gramꢀ
18,19
positive pathogens.
Subsequently, related oxabicyclooctaneꢀlinked NBTIs, such as 3 (AM8191)
20
have also been described as broad spectrum antibacterial agents. Here, we report our efforts aimed at
the discovery of THPꢀbased NBTIs with an expanded antibacterial spectrum covering some of the
most problematic GNB in the clinic, i.e carbapenemꢀresistant Enterobacteriaceae (CRE), MDR
Pseudomonas aeruginosa (P. aeruginosa) and MDR Acinetobacter baumannii (A. baumannii), while
21,22
keeping an acceptable pharmacological and safety profile.
4
ACS Paragon Plus Environment

Page 5 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
To gain some insights into the relationship that exists between biochemical potency on DNA
topoisomerases and antibacterial activity against GNB, we determined the inhibitory profile of our
17
preclinical candidate 2 and its analogs 4 against DNA gyrase and Topo IV enzymes isolated from
wildꢀtype E. coli, A. baumannii and P. aeruginosa. Antibacterial activity against antibioticꢀsusceptible
E. coli and P. aeruginosa strains and an FQꢀresistant A. baumannii strain (due to mutations in the
quinoloneꢀresistance determining regions of GyrA and Par C) were determined. Ciprofloxacin (CIP), a
FQ with potent antiꢀGNB activity was used as a reference although its mode of action (poisoning 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
23
DNA topoisomerases) differs from the one of NBTIs (catalytic inhibitors).
Table 1: Antibacterial and topoisomerases activities of CIP, NBTIs 2 and 4
MIC (ꢀg/mL)
Gyrase SCIA (IC50, µM)
TopoIV RIA (IC50, µM)
a
b
c
d
e
f
d
e
f
Ec
Ab
Pa
Ec
Ab
Pa
Ec
Ab
Pa
2
4
>8
0.5
2
>8
4
8
2
0.5
0.125
0.5
0.125
0.03
8
0.125
0.125
32
0.125
0.125
8
0.125
>8
0.125
0.5
0.125
2
CIP
≤0.03
0.25
a
b
Escherichia coli ATCC 25922. Acinetobacter baumannii T6474 containing a Ser83ꢀLeu mutation in GyrA as well as a Ser80ꢀLeu
c
d
e
f
mutation in the ParC. Pseudomonas aeruginosa ATCC 27853. wildꢀtype (wt) Escherichia coli. wt Acinetobacter baumannii. wt
Pseudomonas aeruginosa.
As indicated in Table 1, DNA gyrases of GNB were ~16ꢀfold more sensitive to CIP than Topo IV
enzymes. With IC50s = 0.5 ꢀM against E. coli and P. aeruginosa DNA gyrases, CIP exhibits a potent
antibacterial activity against wildꢀtype E. coli and P. aeruginosa strains. The wildꢀtype A. baumannii
DNA gyrase enzyme was also inhibited by CIP (IC = 2 ꢀM). As expected, the A. baumannii T6474
5
0
strain was resistant to CIP (MIC > 8 ꢀg/mL). NBTI 2 featuring an oxathiinoꢀpyridazine based rightꢀ
hand side (RHS) displayed potent inhibitory activity against all Gramꢀnegatives Topo IV enzymes
(
0.125 µM), whereas weaker inhibitory potencies against DNA gyrase enzymes were measured,
ranging from 0.5 µM (P. aeruginosa) to 8 µM (E. coli). In contrast to CIP, the enzymatic inhibition
profile of 2 did not result in antibacterial activity in vitro, except in case of A. baumannii that was
susceptible to 2 (MIC = 2 ꢀg/mL). On the other hand, pyridoꢀoxazinoneꢀbased compound 4 displayed
potent, dual and balanced inhibition of all DNA topoisomerases isolated from GNBs, that tracked well
with good MICs measured against Gramꢀnegative E. coli and A. baumannii (MICs = 0.5 and
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
0.125 ꢀg/mLrespectively). Despite its better inhibitory potency on isolated P. aeruginosa DNA
topoisomerases than CIP, NBTI 4 exhibited a weaker wholeꢀcell activity against P. aeruginosa
(
MIC = 4 ꢀg/mL). The noticeable difference in susceptibility observed for 4 against P. aeruginosa in
comparison to other GNB tested, was attributed to insufficient cytoplasmic accumulation in this
particular organism. Because of its broad, potent inhibitory and antibacterial activities as well as its
lack of crossꢀresistance with FQs, compound 4 was selected as a starting point.
0
1
2
3
4
5
6
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
DESIGN HYPOTHESIS
Our working hypothesis to optimize antiꢀGNB activity of proprietary THPꢀbased NBTIs was based on
observations made by O'Shea and Moser that anti Gramꢀnegative agents are generally more polar
(lower clogD7.4 values) and more positively charged at pH 7.4 than agents active against Gramꢀpositive
1
0
pathogens only. Therefore, we sought analogs of NBTI 4 with increased polarity hoping that drug
accumulation and, consecutively susceptibility of GNB (and in particular P. aeruginosa) to THPꢀbased
NBTIs will improve. The pyridoꢀoxazinone RHS, resulting in dual and potent DNA topoisomerase
inhibition was prioritized throughout the course of our study. Putting aside oral absorption
2
4
considerations, we first envisioned to increase the polarity of THPꢀbased NBTIs by designing
molecules that would be more protonated at physiological pH. As indicated in Figure 2, with a
a a
calculated pK (cpK ) of 7.6 for the secondary amine, about 50% of NBTI 4 should be protonated at
pH 7.4. Taking into account the constraints set by the specific binding mode of NBTIs (see thereafter,
Figure 3), an option to render 4 more polar was to swap the hydroxyl group present on the twoꢀatom
linker (that connects the THPꢀcore to the leftꢀhand side (LHS), Figure 1) for an amino group, creating
molecules with a dibasic character. This change would likely have little influence on the binding
affinity of the inhibitor but derivatives 11 and 16 with a ꢀC
respectively, are seemingly more polar with a clogD7.4 around ꢀ1.5 (compared to a clogD7.4 =0.24 for 4)
Figure 2). As the cpK for the primary amine group present in compounds 11 and 16 reached 9.3, a
α 2 2 2 β 2
H(NH )ꢀCH ꢀ or a ꢀCH ꢀC H(NH )ꢀ linker
(
a
charged species was expected to be their major component at pH ~7.4.
6
ACS Paragon Plus Environment

Page 7 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
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
Figure 2: Envisioned structural modifications of NBTI 4 resulting in NBTIs with increased polarity
We also envisioned to increase polarity of our series by utilizing naphthyridinones or analogous LHS,
attached to the THPꢀcore from a Nꢀatom through a ꢀCH ꢀCH(OH)ꢀ linker. Even though the polarity of
2
the resulting structures increased in a limited extent (decreased of clogD7.4 from 0.24 for 4 to ꢀ0.62 for
a
5) and basicity remained constant (cpK = 7.6), the construction of such novel molecules via epoxide
opening was an attractive prospect. It indeed provided a rapid entry point to novel entities and an
opportunity to subsequently reach dibasic entities featuring as well a ꢀCH ꢀC H(NH )ꢀ type linker.
2
β
2
Thus, dibasic derivatives such as derivative 6 had a marked increase in polarity (clogD7.4 = ꢀ1.85),
being mostly present as charged species at pH 7.4 (with a cpK
featuring a ꢀC H(NH )ꢀCH
a
= 8.6 for the primary amine). NBTI 11
17
α
2
2
ꢀ linker was synthesized from known compound 7 following Scheme 1.
CHEMISTRY
Bromination of compound 7 at benzylic position was performed in refluxing carbon tetrachloride
using Nꢀbromosuccinimide and azobisisobutyronitrile as a radical initiator, leading to bromide 8 as a
3
mixture of epimers. Treatment with sodium azide (NaN ) in hot dimethylformamide (DMF) and
subsequent removal of the tertꢀbutyloxycarbonyl (Boc) group using trifluoroacetic acid (TFA)
afforded free amine 9. The latter, upon reaction with aldehyde 10 in MeOH, led to an imine that was
further reduced with sodium borohydride. The resulting azide derivative was finally reduced using
3
polymerꢀbound triphenylphosphine (PPh ) in aqueous tetrahydrofuran (THF) to afford NBTI 11,
isolated as an equimolar mixture of C ꢀepimers. Chiral HPLC separation was performed, leading to
α
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
both 11a and 11b (for details see Supporting Information). The absolute stereochemistry of the C
α
(carbon bearing the primary amine) of both epimers was not assigned.
a
Scheme 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
2 β 2
THPꢀbased NBTIs 16a,b and 20ꢀ23 featuring a ꢀCH ꢀC HNH ꢀ linker were obtained as depicted in
18
Schemes 2 and 3, from compounds 12aꢀc. NꢀBoc groups were cleaved using TFA, and resulting free
amines were treated with benzyl chloroformate (CbzꢀCl) in presence of aqueous sodium bicarbonate
(
NaHCO ). Subsequent oxidation of secondary alcohols using ParikhꢀDoering conditions (combining
3
sulfur trioxideꢀpyridine complex, dimethylsulfoxide and N,Nꢀdiisopropylethylamine (DIPEA)) in
dichloromethane (DCM) led to ketones 13aꢀc. A reductive amination reaction, performed with
ammonium acetate and sodium cyanoborohydride gave rise to free amines that were further NꢀBoc
protected upon treatment with diꢀtertꢀbutyl carbonate (Boc
2
O). Removal of the benzyloxycarbonyl
(
Cbz) protecting groups under normal hydrogen atmosphere in presence of 20% palladium hydroxide
2
on carbon (20% Pd(OH) on C) as a catalyst afforded free amines 14a,c. Subsequently, amines 14a,b
were reductively aminated using aldehyde 10 to afford compounds 15a,b as equimolar mixtures of C
epimers. Intermediates 15a,b were treated with TFA to afford NBTIs 16a,b, as equimolar mixtures of
ꢀepimers.
β
ꢀ
C
β
8
ACS Paragon Plus Environment

Page 9 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 2
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
As described in Scheme 3, the asymmetric synthesis of 20, the (S)ꢀconfigured epimer of 16a was
1
8
prioritized over its (R)ꢀepimer based on our previous findings. Starting from alcohol 12a, NꢀBoc
protecting group was exchanged to NꢀCbz via a deprotection (TFA) ꢀ protection (CbzꢀCl, NaHCO3)
sequence to afford alcohol 17.
a
Scheme 3
Gratifyingly, when treated with diphenyl phosphoryl azide (DPPA) under typical Mitsunobu reaction
conditions (combining PPh
as a single (S)ꢀconfigured enantiomer. Subsequent Staudinger reduction (PPh
azide 18, Bocꢀprotection (Boc O) and hydrogenolysis (H , 20% Pd(OH) on C) of the remaining NꢀCbz
3
and diisopropyl azodicarboxylate (DIAD)), alcohol 17 afforded azide 18
3
in aqueous THF) of
2
2
2
group led to intermediate 19. The latter was transformed to corresponding NBTI 20 via reductive
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
amination reaction employing aldehyde 10 and Boc deprotection (TFA). Unfortunately, this sequence
did not lead to analogous NBTIs 21 and 23, as attempted Mitsunobu reactions on alcohols 12b,c or
their NꢀCbz protected analogs (not shown) failed, returning only elimination products. Therefore,
β
previous intermediate 15b was subjected to a chiral HPLC separation, and resulting C epimers were
0
1
2
3
4
5
6
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
NꢀBoc deprotected (TFA), giving rise to NBTIs 21 and 22. Considering that the H NMR spectrum
1
recorded for 20 showed an excellent fit with the one of compound 21 (comparatively to the H NMR
25
spectrum obtained for 22), the absolute stereochemistry of the carbon bearing the primary amine in
compound 21 was similarly assigned to (S). Intermediate 14c was subjected to chiral HPLC separation
2
6
to afford (S)ꢀ14c. Latter amine was alkylated via a reductive amination with aldehyde 10 and treated
with TFA, affording NBTI 23.
a
Scheme 4
The asymmetric synthesis of Nꢀlinked THPꢀbased NBTIs relies on chiral epoxides (R)ꢀ27 and (S)ꢀ27.
18
As illustrated in Scheme 4, alkene 24 easily available in large amounts was chosen as a starting
material. Direct epoxidation with m-CPBA gave rise to a mixture of (R) and (S)ꢀepimers of epoxide
27. Both epimers could be separated by tedious chromatography to afford (R)ꢀ27 and (S)ꢀ27 in 21 and
29% yield, respectively. Sharpless asymmetric dihydroxylation of alkene 24 using commercial ADꢀ
mixꢀβ gave rise to diol 25 in a low diastereoselectivity (de ~30%) in favor of the enantiomer (S)ꢀ25.
On the other hand, the use of commercial ADꢀmixꢀα returned diol 25 with an acceptable level of
diastereoselectivity (de ~70%) in favor of the enantiomer (R)ꢀ25. Diol 25 was then treated with
pivaloyl chloride (PivꢀCl) in presence of DMAP to cleanly mask the primary alcohol. Subsequent
reaction with methanesulfonyl chloride (MsꢀCl) in presence of triethylamine (TEA) led to compound
10
ACS Paragon Plus Environment

Page 11 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
6. Upon treatment with sodium methoxide in methanol, 26 formed the intermediate primary sodio
alcoholate that displaced via a clean S
N
2 reaction the adjacent mesylate, closing the epoxide 27, finally
isolated as a 8 to 1 mixture of epimers. The minor epimer (R)ꢀ27 was converted back to diol (R)ꢀ25 via
2
7
its selective hydrolysis using (R,R)ꢀJacobsen's catalyst, thereby allowing the isolation 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
enantiomerically pure epoxide (S)ꢀ27. Diol (R)ꢀ25 was transformed to epoxide (R)ꢀ27 using the oneꢀ
2
8
pot methodology developed by Sharpless (for experimental details see Supporting Information).
a
Scheme 5
2
9
As shown in Scheme 5, epoxide (S)ꢀ27 and naphthyridinone 28a were reacted together in hot DMF
in presence of cesium carbonate to afford alcohol 29a. After Boc deprotection using TFA, the
resulting amine 30 was subjected to reductive amination reactions with aldehydes 10 or 31ꢀ34, giving
rise to NBTIs 5 or 35ꢀ38, respectively.
As shown in Scheme 6, the synthesis of NBTIs 6 and 42ꢀ44 started from alcohols 29aꢀd obtained by
opening of epoxide (S)ꢀ27 by naphthyridinones 28a,b, quinolinone 28c, or naphthyridinone 28d,
respectively, using cesium carbonate in hot DMF. Alcohols 29aꢀd were transformed to their
corresponding mesylates by treatment with MsꢀCl in presence of TEA. The latter intermediates reacted
with NaN in hot DMF gave rise to azides 39aꢀd. The nucleophilic substitution reaction took place
3
with an overall net retention of configuration as the intramolecular formation of an oxonium species
occurs faster than the subsequent backside attack by the azide anion. However, this sequence failed in
delivering azides 39e,f. Therefore, the synthesis of NBTIs 45ꢀ46 had to start from alcohols 29e,f
obtained in this instance by opening of epoxide (R)ꢀ27 by pyridoꢀpyrazinones 28e,f respectively using
potassium carbonate in hot DMF. A subsequent Mitsunobuꢀtype reaction (DIAD, PPh
3
) using DPPA
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
as an azide source was performed on (R)ꢀdiols 29e,f, returning (S)ꢀconfigured azides 39e,f in low
yields. Azides 39aꢀf were Bocꢀdeprotected under acidic conditions, affording amines 40aꢀf, which
after reductive amination with aldehyde 10 gave rise to intermediates 41aꢀf. Staudinger reduction led
3
0
to NBTIs 6 and 42ꢀ46, isolated as single diastereomers.
0
1
2
3
4
5
6
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
Scheme 6
RESULTS AND DISCUSSION
For all novel chemical entities, antibacterial activity, expressed as MIC (minimal inhibitory
concentration, [ꢀg/mL]), was measured against a panel of bacterial strains including drugꢀsusceptible
Staphylococcus aureus (S. aureus), E. coli, and P. aeruginosa strains and a quinoloneꢀresistant A.
baumannii. Inhibition of DNA gyrase and Topo IV enzymes isolated from wildꢀtype E. coli,
A. baumannii and P. aeruginosa was determined.
12
ACS Paragon Plus Environment

Page 13 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2: Antibacterial and topoisomerases activities of NBTIs 2, 11, 16a,b and 20ꢀ23
O
O
O
O
H
H
N
N
N
NH2 O
N
N
H
O
N
O
N
N
H
O
N
W
U
NH2
11
NH2
20 W = N
NH2
16a W = N
6b W = N, U = C-F
1
21 W = N, U = C-F 22 W = N, U = C-F
3 U = C-F
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
when unspecified U, W represents CH
2
MIC (ꢀg/mL
)
Gyrase SCIA (IC50, µM)
TopoIV RIA (IC50, µM)
a
b
c
d
e
f
g
e
f
g
Sa
Ec
Ab
Pa
Ec
Ab
Pa
Ec
Ab
Pa
2
≤0.03
2
>8
2
8
>8
8
8
2
0.5
0.5
0.125
0.125
0.5
0.125
0.125
0.5
0.125
0.125
0.5
1
1
1
8
2
16a
6b
0.125
0.03
0.06
0.03
1
0.125
0.25
0.125
0.125
1
0.5
0.25
0.25
0.125
1
2
0.5
0.125
0.125
0.125
0.5
0.5
0.5
1
4
0.125
0.125
0.125
0.5
0.03
0.125
0.03
0.125
0.03
0.03
0.03
0.03
0.03
0.03
0.125
0.03
2
0
1
2
0.5
1
0.03
0.03
2
0.03
0.125
0.125
2
>8
1
0.125
≤0.01
23
0.03
≤0.03
≤0.03
0.03
0.03
0.03
a
b
c
d
Staphylococcus aureus ATCC 29213. Escherichia coli ATCC 25922. Acinetobacter baumannii T6474. Pseudomonas
e
f
g
aeruginosa ATCC 27853. wildꢀtype (wt) Escherichia coli enzyme. wt Acinetobacter baumannii enzyme. wt Pseudomonas
aeruginosa enzyme.
Our investigations on dibasic Cꢀlinked THPꢀbased NBTIs started with the characterization of
α 2 2
compound 11, built on a ꢀC H(NH )ꢀCH ꢀ linker. As shown in Table 2, compound 11, a mixture of
two epimeric amines at the benzylic position, displayed an inhibitory profile similar to NBTI 2, with
potent activities against Topo IV enzymes isolated from Gramꢀnegatives (~0.125 ꢀM), and more
scattered activities against DNA gyrase enzymes with IC50s ranging from 0.5 to 8 ꢀM. However, in
contrast to NBTI 2, compound 11 showed a promising MIC of 1 ꢀg/mL against wildꢀtype E. coli.
Despite potent inhibition of A. baumannii and P. aeruginosa Topo IV enzymes, limited antibacterial
activity against A. baumannii and P. aeruginosa strains was observed with 11 (MIC = 8 ꢀg/mL). The
antibacterial activity of 11 against S. aureus was also weaker than expected (MIC = 2 ꢀg/mL). Both
epimers 11a and 11b showed activities comparable to 11, behaving as dual but unbalanced DNA
gyraseꢀTopo IV inhibitors (data not shown). Naphthyridineꢀbased derivatives 16a (U = CH) and 16b
2 β 2
(U = CF), featuring ꢀCH ꢀC H(NH )ꢀ linkers were also evaluated as mixtures of two epimeric amines.
Both 16a and 16b elicited dual and balanced inhibitory activity against DNA gyrases and Topo IV
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
enzymes. DNA topoisomerase inhibition for all species was tightly distributed in comparison to those
observed for 11, with IC50s for 16a and 16b ranging between 0.03 ꢀM and 0.5 ꢀM. The slightly more
potent DNA topoisomerase inhibition observed for compound 16b in comparison to 16a did not
translate into a better antibacterial activity. Both compounds 16a and 16b inhibited growth of
P. aeruginosa with MICs = 2 and 4 ꢀg/mL respectively. For the first time in the series, we could
observe decent antiꢀpseudomonal activities for biochemically potent compounds. Compounds 20 and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
β
21, (S)ꢀconfigured epimers at C (the carbon bearing the primary amine), present in compound 16a
and 16b, show potent, dual and balanced inhibitory activities against all DNA topoisomerases with
IC50s within a narrow range (0.03 to 0.125 ꢀM). Gratifyingly, exquisite antibacterial activities against
S. aureus, E. coli and A. baumannii (all MICs < 0.25 ꢀg/mL) as well as good antibacterial activity
against P. aeruginosa were measured for compounds 20 and 21, with MICs = 0.5 and 1 ꢀg/mL,
respectively. The (R)ꢀconfigured epimer 22 was found to be a >10ꢀfold less potent antibiotic against
Gramꢀnegative strains than 21. This could in part be due to its 4ꢀfold lower inhibitory potency against
DNA topoisomerases compared to 21. Compound 23 featuring a fluoroquinoline LHS displayed
exquisite inhibitory activity against DNA topoisomerases and good antibacterial activity again
P. aeruginosa. From this limited dataset, we could establish that the (S)ꢀconfigured ꢀCH
linker shows superior coverage of both Gramꢀpositive and Gramꢀnegative bacteria as compared to the
R)ꢀconfigured linker and ꢀC H(NH )ꢀCH ꢀ linkers. We could also confirm the absence of crossꢀ
2 β 2
ꢀC H(NH )ꢀ
(
α
2
2
resistance with FQs in this subꢀseries as demonstrated with good MICs measured against a FQꢀ
resistant A. baumannii strain. Analogues of NBTIs 20 or 21, featuring dioxinoꢀ or oxathiinoꢀpyridine
(
or pyridazine) RHS were also prepared but displayed MICs ≥ 8 ꢀg/mL against P. aeruginosa despite
potent DNA topoisomerases inhibition (see Supporting Information, table Sꢀ9).
To understand if cytoplasmic accumulation drives the better cellular activity of NBTIs featuring the
2 β 2 2
(S)ꢀconfigured ꢀCH ꢀC H(NH )ꢀ linker and a pyridoꢀoxazinone RHS over their ꢀCH ꢀCH(OH)ꢀ
analogs, compounds 4 and 21 sharing a similar topoisomerase inhibitory profile on E. coli enzymes,
were chosen for MIC determination against a panel of genetically defined E. coli mutants. These
E. coli mutants possess either a decreased ability to expel compounds (tolC and acrAB knockout
14
ACS Paragon Plus Environment

Page 15 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
mutants) or an increased permeability to lipophilic substances due to a truncated lipopolysaccharide
(LPS) layer (rfaC).
Table 3: Antibacterial activities of compounds 4 and 21 against a panel of genetically modified strains
b
c
d
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
a
E.coli ꢁrfaC
E.coli ꢁtolC
E.coli ꢁacrAB
logD7.4
E.coli wt
Permeability ↑
Efflux ↓;
Efflux ↓
e
e
f
e
f
e
f
MIC
MIC
ꢁMIC
MIC
ꢁMIC
MIC
ꢁMIC
4
2.0
0.9
0.5
0.031
↓, 16 x
0.004
↓, 100 x
0.004
↓, 100 x
21
0.063
0.031
↓, 2 x
0.008
↓, 8 x
0.008
↓, 8 x
a
b
c
E. coli MG1655, wildꢀtype strain. Derivative of E. coli MG1655 with truncated LPS. Derivative of E. coli MG1655 lacking the efflux
d
e
system TolC of the outer membrane. Derivative of E. coli MG1655 lacking the AcrAB efflux system of the cytoplasmic membrane. MIC
in ꢀg/mL. Fold change compared to MIC measured on wildꢀtype strain.
f
Compound 4 displayed good antibacterial activity against wildꢀtype E. coli, but its amineꢀcontaining
analog compound 21, was still 8ꢀfold more potent with a MIC of 0.063 ꢀg/mL (Table 3). Firstly, the
marked potency difference observed against wildꢀtype strains between compounds 4 and 21 , was
absent with an E. coli permeable mutant as MICs measured for both compounds were identical
(0.031 ꢀg/mL). Therefore ꢁMIC for compound 4 is significantly larger (16ꢀfold difference) than the
one observed for compound 21 (2ꢀfold change) and, in contrast to compound 4, compound 21 retains
more of its activity when an intact outer membrane (OM) is present. Secondly, there was only a little
difference in potency against E. coli strains with impaired efflux systems (MICs of 0.008 and 0.04
ꢀg/mL, respectively). ꢁMIC for compound 4 (100ꢀfold difference) was again larger with these pairs of
strains than the one observed for compound 21 (8ꢀfold change), suggesting that compound 21 is less
affected by efflux via the AcrABꢀTolC system than the corresponding hydroxyl analog 4. As indicated
in Table 3, primary amine 21 is more polar (logD7.4 = 0.9) than its hydroxyl counterpart 4
(
logD7.4 = 2.0). Furthermore, the measured pK
a
of the primary amine present in compound 21 was 1ꢀ
= 7.4).
log higher (pK = 8.4) than the one of the secondary amine present in both compounds (pK
a
a
Consequently, compound 21 is expected to be mostly protonated (> 90%) at pH 7.4, whereas
compound 4 is only 50% charged at this pH. The better entry into the cytoplasm of the GN E. coli of
dibasic compounds such as 21 may be the result of either specific uptake through porins or selfꢀ
3
1,32
promoted uptake through the outer membrane of Gramꢀnegative bacteria.
Overall, the enhanced
antibacterial activity of compound 21 on wildꢀtype E. coli over its analog 4 can probably be ascribed
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
to higher cytoplasmic accumulation, resulting from the combination of a better permeability through
the cell envelope and a lower propensity to be effluxed.
0
1
2
3
4
5
6
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: Compound 21 (blue) is docked into the DNA – gyrase dimer. (DNA cyan; chain 1 red; chain
2
magenta). The surface of the active site is mimicked by an orange net. Hydrogen bond interactions
are shown in black dotted lines.
We tried to rationalize the broad enzymatic inhibitory activity of compound 21 by docking it into the
3
3
Xꢀray structure of 3 complexed to a dimer of S. aureus DNA gyrase A and a 20ꢀbp DNA duplex
2
0
34
(PDB accession code: 4PLB, see Supporting Information) using the software Moloc. Molecule 21
runs along the internal C2ꢀsymmetry axis of the GyrAꢀDNA complex with the LHS being intercalated
between a GpC pair of base pairs while the RHS is embedded in a cavity located at the interface of the
two GyrA monomers. A linker with appropriate length, geometry and flexibility carrying functional
groups is connecting these two units. The protonated primary amine present in the linker of 21
improves solubility a lot (>100ꢀg/mL at pH varying from 1 to 9) and increases polarity but has no
obvious Hꢀbond partner in the protein DNA complex unlike a hydroxyl group in 3 located in an
identical position in the template structure. However, the secondary amine in the linker is in an ideal
position to form a strong Hꢀbond with Asp83 of one GyrA monomer and a second water mediated Hꢀ
bond with the same amino acid of the second protein monomer. The pyridoꢀoxazinone RHS nicely
complements the shape of a hydrophobic pocket establishing strong van der Waals interactions but
clearly lacking any direct Hꢀbonds with the protein. All structural features involved in binding of 21
16
ACS Paragon Plus Environment

Page 17 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
are present in both DNA gyrase and topoIV and are well conserved across key bacterial species
thereby providing the basis for its broad antibacterial activity. This structural insight also provides a
3
3
rational for the lack of crossꢀresistance of NBTIs with FQs that bind at a different site.
Table 4: Antibacterial and topoisomerases activities of NBTIs 5, 35ꢀ38.
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
MIC (ꢀg/mL
)
Gyrase SCIA (IC50, µM)
TopoIV RIA (IC50, µM)
a
b
c
d
e
f
g
e
f
g
Sa
Ec
Ab
Pa
Ec
Ab
Pa
Ec
Ab
Pa
5
≤0.03
≤0.03
≤0.03
0.06
0.5
≤0.03
0.5
8
0.125
8
0.125
0.125
0.125
8
0.125
0.125
0.5
0.03
0.125
0.03
2
0.125
0.125
0.03
8
0.03
0.125
≤0.01
0.5
3
5
4
4
>8
>8
>8
>8
36
0.5
0.5
32
3
3
7
8
>8
>8
>8
8
0.031
8
8
2
0.5
0.125
0.5
0.125
a
b
c
d
Staphylococcus aureus ATCC 29213. Escherichia coli ATCC 25922. Acinetobacter baumannii T6474. Pseudomonas
e
f
g
aeruginosa ATCC 27853. wildꢀtype (wt) Escherichia coli enzyme. wt Acinetobacter baumannii enzyme. wt
Pseudomonas aeruginosa enzyme.
As our working hypothesis seemed to hold some promise, we turned our attention to NBTIs featuring
a polar LHS. Thanks to a modular synthesis we had a rapid access to such novel Nꢀlinked THPꢀbased
2
molecules built on a ꢀCH ꢀCH(OH)ꢀ linker. A set of compounds, all featuring the 1,8ꢀnaphthyridinone
LHS, was synthesized to study within this limited subꢀseries, the relationship that exists between the
RHS and the antibacterial coverage. As illustrated in Table 4, these compounds, i.e compounds 5 and
3
5ꢀ38, displayed good antibacterial activities against S. aureus wildꢀtype strains. However, as in the Cꢀ
linked THPꢀbased series, only compound 5, featuring a pyridoꢀoxazinone RHS exhibited activity
against the Gramꢀnegative E. coli and A. baumannii, as a result of dual and balanced inhibitory
activity against DNA topoisomerases (IC s ranged between 0.03 and 0.125 ꢀM). However its
50
antibacterial activity against P. aeruginosa remained weak (MIC = 8 ꢀg/mL). Compounds 35 and 36,
featuring a dioxinoꢀpyridine and an oxathiinoꢀpyridine RHS, were also dual DNA gyraseꢀ Topo IV
inhibitors, although some variations in IC50s were observed. Both inhibited A. baumannii growth
potently (MIC = 0.5 ꢀg/mL) but were mostly inactive against E. coli and P. aeruginosa. Compound
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
3
7, featuring a dioxinoꢀpyridazine showed weak to no inhibitory activity against DNA topoisomerases
isolated from Gramꢀnegative strains, failing in turn to inhibit GNB growth. Compound 38, a close
analog of our preclinical candidate 2, shared with it similar enzymatic and antibacterial properties,
being only active against S. aureus in this panel. The measured logD7.4 values for 3 and 5 were
identical (logD7.4 = 2.0). Introduction of a supposed polar LHS did not translate into NBTIs with
increased polarity (lower logD ). Compounds 4 and 5, with identical inhibitory potency on DNA
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
.4
topoisomerases and similar physicochemical properties, exhibited ultimately a similar antibacterial
potency and spectrum. The lack of activity against Gramꢀnegative strains for DNA topoisomerases
inhibitors 5, 35, 36 and 38 could be attributed to a poor cytoplasmic accumulation because, similarly
to compound 4, MICs recorded on E. coli mutants with decreasing efflux capability or increased
permeability were dramatically improved, when compared to MICs obtained on wildꢀtype E. coli
strains (see Supporting Information, table Sꢀ11).
Table 5: Antibacterial and topoisomerases activities of NBTIs 6, 42ꢀ46.
MIC (ꢀg/mL
)
Gyrase SCIA (IC50, µM)
TopoIV RIA (IC50, µM)
a
b
c
d
e
f
g
e
f
g
Sa
Ec
Ab
Pa
Ec
Ab
Pa
Ec
Ab
Pa
6
0.25
1
0.25
1
4
2
2
2
2
1
0.125
0.5
0.125
0.03
0.125
0.03
0.03
0.03
0.03
0.03
0.125
0.03
0.125
0.125
0.125
0.125
0.125
0.03
0.03
0.5
4
2
3
4
5
6
1
4
4
4
2
1
4
4
4
4
0.25
0.063
0.125
0.5
0.25
0.25
0.5
0.125
0.5
0.125
0.03
0.125
0.03
0.03
0.03
0.03
0.5
0.125
0.125
0.125
0.5
0.5
0.125
0.03
a
b
c
d
Staphylococcus aureus ATCC 29213. Escherichia coli ATCC 25922. Acinetobacter baumannii T6474. Pseudomonas
.e
f
g
aeruginosa ATCC 27853. wildꢀtype (wt) Escherichia coli enzyme. wt Acinetobacter baumannii enzyme.
wt
Pseudomonas aeruginosa enzyme.
To strengthen our earlier findings on the beneficial effect of increasing polarity by modification of the
twoꢀatom linker, a series of dibasic Nꢀlinked THPꢀbased molecules was synthetized and evaluated. As
18
ACS Paragon Plus Environment

Page 19 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
shown in Table 5, compounds 6 and 42ꢀ46, featuring the pyridoꢀoxazinone RHS and various LHS
2 β 2
linked to the (S)ꢀconfigurated ꢀCH ꢀC H(NH )ꢀ motif through a nitrogen atom, were all dual and
balanced DNA gyrase and TopoIV inhibitors with activities below 0.5 ꢀM. Gratifyingly, compound 6
displayed good antibacterial activity against S. aureus and E. coli with MICs = 0.25 mg/L and
difficultꢀto treat Gramꢀnegative pathogens A. baumannii and P. aeruginosa with MICs = 1 ꢀg/mL.
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
NBTIs 42ꢀ46 were not as potent antibiotics as 6 against P. aeruginosa and A. baumannii, but
consistently displayed MICs ≤ 4 ꢀg/mL against these strains. Despite potent and dual inhibitory
activities against E. coli DNA topoisomerases, NBTI 42 exhibited a slightly weaker antibacterial
activity against wildꢀtype E. coli (MICs = 1 ꢀg/mL) than its congeners NBTIs 43ꢀ46 (MICs
3
5
≤
0.5 ꢀg/mL). A weaker antibacterial activity against S. aureus was also observed for NBTI 42.
Compound 6 exhibits a measured logD7.4 of 0.3, significantly lower than that of NBTI 4
(logD7.4 = 2.0). A large proportion (>95%) of 6 is positively charged at pH 7.4 as measured pK s of its
a
3
6
primary and secondary amines were 8.7 and 7.0 respectively. The direct comparison of antibacterial
activity of compounds 5 and 6 confirmed unambiguously that the amine function present in the linker
3
5
ameliorates the activity against Gramꢀnegative pathogens, by favoring cytoplasmic accumulation. As
demonstrated with compound 21, MICs recorded for compounds 6 and 42ꢀ46 on E. coli mutants with
impaired efflux systems or increased permeability were less improved than those measured with
compound 5 on these strains (see ꢁMIC in Supporting Information, table Sꢀ10). It turned out that in
case of THPꢀbased NBTIs featuring a pyridoꢀoxazinone RHS, our working hypothesis of increasing
concomitantly polarity and basicity by addition of an amino group at a position not detrimental for onꢀ
target activity, indeed led to significant gain in antibacterial activity against difficultꢀtoꢀtreat GNB,
thanks to a sufficient cytoplasmic accumulation.
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
Preliminary microbiological profile of two representative molecules
Compounds 6 and 21 were selected as representative molecules of the Nꢀ and Cꢀlinked THPꢀbased
series respectively, and were evaluated in vitro against a panel of clinically relevant Gramꢀpositive and
Gramꢀnegative organisms, including multidrug resistant ones. This microbiological evaluation was
performed to compare the broadness of their antibacterial spectrum and their in vitro potency (MIC
data and bacteria resistance phenotypes are detailed in Supporting Information).
0
1
2
3
4
5
6
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 shown in Table 6, compound 21, featuring a fluoronaphthyridineꢀLHS, exhibited a broad and
potent cellular activity against a wide range of organisms. MICs against representative Gramꢀpositive
strains of S. aureus, enterococci (E. faecium and E. faecalis), streptococci (S. agalatiae,
S. pneumoniae, and S. pyogenes) were ≤ 2 ꢀg/mL. Remarkably, with the only exception of one
Serratia marcescens strain, antibacterial activity against Gramꢀnegative pathogens such as
Enterobacteriaceae, P. aeruginosa and A. baumannii, with clinically relevant resistance phenotypes
(
FQꢀresistance, carbapenemꢀresistance), were also good, all strains being susceptible to 21 with
MICs ≤ 4 ꢀg/mL.
Table 6: MIC ranges obtained for NBTIs 6 and 21 against selected strains (MIC in ꢀg/mL).
Strains
Nb of strains
Ciprofloxacin
6
21
S. aureus
5
0.25 ꢀ >16
0.25 ꢀ 2
0.125 ꢀ 0.5
Enterococcus faecium
Enterococcus faecalis
Streptococcus agalactiae
Streptococcus pneumoniae
2
1
1
2
> 16
1, 16
0.063, 0.5
0.25
1
1
1
4
2
1, >16
0.25, 8
0.125, 2
Streptococcus pyogenes
E. coli
1
4
0.25
≤0.03
≤0.03
≤0.03 ꢀ >16
0.5 ꢀ 2
0.125 ꢀ 0.5
Citrobacter freundii
1
3
7
≤0.03
1
0.25
0.25 ꢀ 4
1 ꢀ 4
Enterobacter cloacae
Klebsiella pneumoniae
≤0.03 ꢀ 0.5
0.06 ꢀ >16
2 ꢀ 16
8 ꢀ 16
P. aeruginosa
A. baumannii
4
0.125 ꢀ >16
1 ꢀ 2
1 ꢀ 2
4
2
> 16
1 ꢀ > 16
8, >16
0.125 ꢀ 4
1, 8
Serratia marcescens
0.06, 1
In comparison, naphthyridinoneꢀbased compound 6 which also exhibited broad antibacterial spectrum,
was generally less potent against Gramꢀpositive and Gramꢀnegative bacteria (roughly 2ꢀ to 4ꢀfold
lower). While compound 6 had weak or no activity against some MDR E. faecium and A. baumannii
20
ACS Paragon Plus Environment

Page 21 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
strains (MICs ≥ 16 ꢀg/mL), it surprisingly showed similar potency than compound 21 against
P. aeruginosa strains (same MIC range 1 ꢀ 2 ꢀg/mL).
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
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
In vivo efficacy screening experiments with NBTI 16b, 21 and 6
The antibacterial activity of compounds 6 and 21 was explored in neutropenic murine thigh infection
models in vivo using Gramꢀpositive and Gramꢀnegative pathogens. Because compound 21 was only
available in limited amounts, in vivo efficacy experiments were mostly performed with compound 16b
as a surrogate (16b is a 1ꢀ1 mixture of compound 20 and 21, see table 2). As summarized in Table 7,
pharmacokinetic properties of compounds 6, 16b and 21 were determined in rodents. In mouse,
compounds 6 and 21 showed comparable profiles with high clearance and large volume of distribution
resulting in a long terminal halfꢀlife (~6 H). Using the methicillinꢀsensitive S. aureus (MSSA) strain as
an infecting pathogen, and administering subcutaneously (sc) compound 16b (MIC against the MSSA
infecting strain = 0.06 ꢀg/mL) at a dose of 40 mg/kg, a net 2ꢀlog reduction of colony forming units
0
1
2
3
4
5
6
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
(
(
CFU) in thighs was observed compared to CFU at treatment start when measured 24h postꢀinfection
22h post treatment). To confirm that the strong antiꢀstaphylococcal activity of compound 16b actually
resulted from its more active epimer compound 21, the latter was assayed using the same experimental
setꢀup (dose and duration), and it reduced bacterial load by more than 2.5ꢀlog unit in thighs. From a
pharmacokineticꢀpharmacodynamic (PKꢀPD) standpoint, the AUC/MIC ratio associated with
compound 21 at the dose of 40mg/kg/day is 23.5 (derived from iv PK data generated at 10mg/kg).
Subsequently, compound 16b was tested under the same conditions against the methicillinꢀresistant
S. aureus (MRSA) (MIC against the MRSA in vivo strain = 0.03 ꢀg/mL). The net bacterial load was
reduced by 1ꢀlog unit in thighs after sc administration of a 40 mg/kg dose. To explore the broadness of
the antibacterial coverage, in vivo efficacy of compound 16b was also determined against a Gramꢀ
negative K. pneumoniae strain. After sc administration of a dose of 30 mg/kg, compound 16b (MIC
against K. pneumoniae in vivo strain = 0.125 ꢀg/mL) nearly exhibited a net static efficacy. At a dose
of 90mg/kg, compound 16b elicited a 1ꢀlog reduction in colony forming unit (CFU) in thighs
compared to CFUs at treatment start. These efficacy data were determined 7.5h post infection (5.5h
postꢀtreatment) to ensure that during the dosing interval (5.5h), the free drug exposure of 16b was
sufficient.
22
ACS Paragon Plus Environment

Page 23 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Because of its lower in vitro potency against K. pneumoniae, in vivo efficacy of compound 6 was
assessed against a susceptible Gramꢀnegative E. coli strain. When administered subcutaneously at
doses of 120 and 360 mg/kg, compound 6 (MIC against E. coli in vivo strain = 0.125 ꢀg/mL) showed
a net static and 2ꢀlog reduction in colony forming units in thighs respectively, when measured 7.5 h
post infection. Technical procedures, graphical representations and statistical details of in vivo
experiments are reported in 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
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
Table 7: Mouse, Rat and Dog Pharmacokinetic Parameters of selected THPꢀBased NBTIs.
0
1
2
3
4
5
6
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
Compounds
4
5
6
16b
21
a
Mouse
b
AUC
n
(
ꢀ
g.h/(mL.dose))
186
0.369
68.6
86
260
0.154
40.0
64
236
0.15
35.4
71
fu
b
n
fAUC (ꢀg.h/(mL.dose))
Cl/F (mL/min/kg)
Vdss (L/kg)
t1/2 (h)
n.d
6.0
12.3
5.6
5.2
1.2
c
Rat
b
AUC
n
(
ꢀ
g.h/(mL.dose))
457
0.086
39
378
0.167
63
145
0.236
34
194
0.15
29
f
u
b
fAUC
Cl (mL/min/kg)
Cl (mL/min/kg)
Vdss (L/kg)
n
(ꢀg.h/(mL.dose))
37
44
110
144
21
86
u
430
2.5
263
2.5
1
573
25
t1/2 (h)
1.3
9.1
8.8
d
Dog
b
AUC
n
(
ꢀ
g.h/(mL.dose))
376
45
34
490
61
Cl (mL/min/kg)
Vdss (L/kg)
2.5
1.3
t1/2 (h)
2.9
a
b
Mouse dosed iv at 10 mg/kg (compounds 16b and 21) or sc at 120 mg/kg (compound 6). Doseꢀnormalized AUC obtained after
c
d
sc or iv dosing. Rat dosed iv at 10mg/kg (compounds 6 and 21) or 1mg/kg (compounds 4 and 5). Dog dosed iv at 1mg/kg
(compounds 6 and 21).
The pharmacokinetic profiles of 6 and 21 in rat were also comparable with both dibasic compounds
displaying high clearances, long halfꢀlives and large volumes of distribution. Compounds 4 and 5,
2 β
their respective monobasic analogs featuring a ꢀCH ꢀC H(OH)ꢀ linker, shared also comparable
pharmacokinetic parameters in rat, with moderate clearances (Cl 37ꢀ44 mL/min/kg), moderate volume
of distribution and short terminal halfꢀlife. Free AUC, the parameter that, together with MIC, is
24
ACS Paragon Plus Environment

Page 25 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
9
inferred to drive the therapeutic dose, was similar for 4, 6 and 21, and only slightly advantageous for
compound 5. The high clearances of the dibasic compounds 6 and 21 observed in rodents were
replicated in the dog, raising concerns of potentially inferior pharmacokinetic profiles, necessitating
nonꢀviable dosing regimens in the clinic.
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
Cardiovascular safety assessment of representative NBTIs 6 and 21
We already reported that, contrary to compound 2, compound 4 featuring a pyridoꢀoxazinone RHS,
+
18
exhibited an unsatisfactory cardiovascular safety profile, partly due to hERG K channel inhibition.
+
As shown in Table 8, a block of the hERG K channel was also observed for compounds 5, 6 and 21,
with 71%, 37% and 86% inhibition at 30ꢀM compound concentration, respectively. The
cardiovascular safety profile of these compounds was assessed further in vitro using cardiomyocytes
derived from human induced pluripotent stem cells (iPSCꢀCM) and in vivo with anesthetized guinea
pigs.
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
+
Table 8: Effects of selected THPꢀbased NBTIs on hERG K channels, on the spontaneous rhythm of
human inducedꢀpluripotentꢀstemꢀcellꢀderivedꢀcardiomyocytes (iPSCꢀCM), and on MAP, HR and ECG
parameters in anesthetized guinea pigs.
Compounds
2
4
5
6
21
0
1
2
3
4
5
6
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
hERG block (%)
20
92
71
37
86
hERG IC50/IC20 (ꢀM)
128/32
2.0/0.5
15/3.7
52/13
5/1.2
b,e
0.27ꢀ0.81
0.31ꢀ0.93)
iPSCꢀCM (ꢀM)
13 (22)
0.21 (0.42)
0.86 (1.4)
0.28 (0.50)
(
b,c
QT (ꢀM)
>9.0 (>39)
>9.0 (>39)
>9.0 (>39)
>9.0 (>39)
>9.0 (>39)
0.23
0.36 (2.1)
1.5 (8.9)
4.8 (29)
1.5 (8.9)
4.8 (29)
0.168
1.5 (3.5)
6 (14)
2.7 (5.0)
11 (19)
28 (53)
2.7 (5.0)
2.7 (5.0)
0.54
0.53 (1.6)
0.53(1.6)
1.8 (5.6)
1.8 (5.6)
1.8 (5.6)
0.33
b,c
HR (ꢀM)
b,c
MAP (ꢀM)
21 (49)
21 (49)
>21 (>49)
0.43
b,c
QRS (ꢀM)
b,c
PR (ꢀM)
d
f
u
a
+
b
Reduction of K current measured at 30 ꢀM compound concentration. Lowest active free (total) concentration (see methods in Supporting
Information for the definition of active) Compared to vehicleꢀtreated animals. Unbound fraction measured in guinea pig plasma by rapid
equilibrium dialysis. Free/total corrected based on the presence of 20% serum in the test solution.
c
d
e
In line with their hERGꢀblocking activities, compound 2 affected the spontaneous rhythm of iPSCꢀCM
only at high concentrations (~13 ꢀM free compound) while compound 4 was already inhibiting at
~
0.21 ꢀM free compound (Table 8). While a similar observation could be made for compounds 5 and
+
21, compound 6 appears different in that it affected iPSCꢀCM much more potently than hERG K
channels. During the in vivo experiment, doses of 3, 10 and 30 mg of compound per kg of bodyweight
were administered intravenously over three consecutive 20ꢀmin periods. Mean arterial pressure
(MAP), heart rate (HR) and several ECG parameters (including PR, QRS, RR and QT intervals) were
recorded continuously, and blood samples were taken at the end of each period for determination of
drug levels. For each cardiovascular parameter, Table 8 reports the lowest drug plasma concentration
(
free and total) at which the parameter is different between compoundꢀtreated and vehicleꢀtreated
animals. At the highest dose tested, compound 2 had no effect on any parameter while compounds 4
and 21, based on naphthyridineꢀLHS, affected QTꢀinterval already at subꢀmicromolar free drug
concentrations and compounds 5 and 6, based on naphthyridinoneꢀLHS, at an intermediate
concentration, in line with their hERGꢀblocking activities. HR is also changed potently by compounds
26
ACS Paragon Plus Environment

Page 27 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
and 21 but more weakly by compounds 5 and 6, and all four compounds affect MAP only at higher
drug concentrations. Compounds 4, 5, and 21 also increased PRꢀ and QRSꢀintervals more weakly than
QTꢀinterval whereas compound 6 was equipotent on the three parameters suggesting that it affects
additional cardiovascular targets as potently as hERG; this could be related to its remarkably potent
activity on iPSCꢀCM.
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
The nature and magnitude of these cardiovascular effects were of particular concern and did not
support the advancement of one of these compounds toward preclinical development. Among all THPꢀ
based NBTIs displaying the desired broad antibacterial spectrum, we were not yet able to identify
compounds displaying a cleaner cardiovascular safety profile. The introduction of an extra primary
amine, essential feature in this chemical series to enable anti Gramꢀnegative activity, led so far to
dibasic compounds with no improved cardiovascular safety profile in comparison to earlier
compounds carrying the pyridoꢀoxazinone RHS such as 4. Further investigation on the RHS have to be
carried out in order to improve the cardiovascular safety profile while keeping broad antibacterial
coverage.
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
CONCLUSIONS
We have reported herein efforts to broaden the antibacterial spectrum of THPꢀbased NBTIs towards
problematic Gramꢀnegative pathogens, especially Enterobacteriaceae, P. aeruginosa and
A. baumannii strains. Notably, we could demonstrate that dibasic compounds, designed by introducing
an extra primary amino function within the flexible 2ꢀatom linker on the NBTI scaffold, led to
improved Gramꢀnegative whole cell activity. For broad and potent DNA topoisomerases inhibitors, the
data we presented here suggests that the better antibacterial activity against Gramꢀnegative organisms,
and notably P. aeruginosa could be primarily attributed to a higher cytoplasmic accumulation
compared to earlier compounds. The mechanism facilitating a better permeation across membranes of
these chemicals remains unknown, although it is tempting to speculate that their two cationic groups
0
1
2
3
4
5
6
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
+
2+
(protonated amines) could displace divalent metals (Mg or Ca ) needed to stabilize the LPS layer,
3
2
thereby promoting selfꢀuptake of these molecules. Two subsets of compounds differing in the nature
of the LHS, but featuring the optimized (S)ꢀconfigured ꢀCH ꢀCH(NH )ꢀ linker could be synthesized.
2
2
For each subset, a representative compound with good potency against P. aeruginosa was studied in
more depth. Compound 21, belonging to the Cꢀlinked THPꢀbased NBTI subset, displayed a broad and
potent antibacterial spectrum, including MDR pathogens. Compound 6, on the other hand, a
representative compound of the Nꢀlinked THPꢀbased NBTI subset, had a narrower spectrum and a
lower potency against certain Gramꢀpositive and negative strains. Compounds 6 and 21 not only
showed in vitro activity but were also efficacious in murine models against relevant Gramꢀnegative
organisms. Pharmacokinetic properties of both compounds 6 and 21 were characterized by high
clearances in rodent species that were not detrimental to in vivo efficacy. Within our proprietary
chemical space delineated around the pyridoꢀoxazinone RHS (chosen to ensure broad antibacterial
coverage), compounds devoid of cardiovascular safety liabilities could not yet be identified and further
3
7
optimization around the THPꢀbased NBTI series is necessary. A better understanding of parameters
driving cytoplasmic drug accumulation within GNB as well as tools for its reliable evaluation are
3
8, 39
required to accelerate the discovery and the optimization of urgently needed novel antibacterials.
28
ACS Paragon Plus Environment

Page 29 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
General chemical methods
Starting materials, reagents and solvents were obtained from commercial sources and used as received.
All reactions were carried out with continuous stirring under atmosphere of dry nitrogen. The
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
13
resonance frequency for H ( C) were recorded on a Bruker Ascend 500 is 500MHz (125MHz).
Chemical shifts (δ) are reported in parts per million (ppm) relative to deuterated solvent as the internal
3 6
standard (δH: CDCl 7.26ppm, DMSOꢀd 2.50ppm), coupling constants (J) are in hertz (Hz). Peak
multiplicities are expressed as follows: singlet (s), doublet (d), doublet of doublets (dd), triplet (t),
quartet (q), multiplet (m), broad singlet (br s). Melting points were determined using a Büchi Bꢀ540
apparatus and are uncorrected. Purification of intermediates and final products was carried out on
2
normal phase using an ISCO CombiFlash system and prepacked SiO cartridges eluted with optimized
gradients of either heptaneꢀEtOAc mixture or DCMꢀMeOH (doped with 1% v/v of aqueous NH OH
4
for basic compounds). Progress of the reactions was monitored by thinꢀlayer chromatography (TLC)
analysis (Merck, 0.2 mm silica gel 60 F254 on glass plates) or by LCꢀMS (Methods and equipment are
described in Supporting Information). All target compounds had purity > 95%, established on a
Waters Atlantis T3, 5 ꢁm, 4.6 mm × 30 mm, eluting with a gradient of 5−95% of MeCN in water
containing 0.04% of TFA; or on a Waters XBridge C18, OBD, 5 ꢁm, 4.6 mm × 50 mm (Waters,
Switzerland), eluting with a gradient of 5−95% of MeCN in water containing 13 mM of NH OH UV
4
at 230 nm and 254nm. Purity and identity were further confirmed by NMR spectroscopy.
tertꢀButyl ((3R,6S)ꢀ6ꢀ((RS)ꢀ2ꢀbromoꢀ2ꢀ(6ꢀmethoxyꢀ1,5ꢀnaphthyridinꢀ4ꢀyl)ethyl)tetrahydroꢀ2Hꢀ
1
7
pyranꢀ3ꢀyl)carbamate (8). To a suspension of 7 (0.6 g, 1.55 mmol) in CCl (10 mL) were added
4
NBS (0.42 g, 2.32 mmol) and AIBN (0.03 g, 0.16 mmol). The mixture was heated at 80°C under a
sunlamp beam for 6 h. The reaction mixture was cooled to rt and was diluted with DCM (20 mL),
2 2 3 2 4
washed with 10% aq Na S O (2 x 15 mL), dried over Na SO , filtered and concentrated to dryness.
The residue was purified by chromatography (DCMꢀMeOH gradient) to afford 8 (0.55 g, 76% yield)
1
as a light yellow foam; H NMR (DMSOꢀd
6
) mixture of isomers δ: 8.83 (d, J = 4.6 Hz, 0.5 H), 8.82 (d,
29
ACS Paragon Plus Environment