Optimization of LpxC Inhibitors for Antibacterial Activity and Cardiovascular Safety

Article abstract of DOI:10.1002/cmdc.201900287

UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine deacetylase (LpxC) is a Zn²⁺ deacetylase that is essential for the survival of most pathogenic Gram-negative bacteria. ACHN-975 (N-((S)-3-amino-1-(hydroxyamino)-3-methyl-1-oxobutan-2-yl)-4-(((1R,2R)-2-(hydroxymethyl)cyclopropyl)buta-1,3-diyn-1-yl)benzamide) was the first LpxC inhibitor to reach human clinical testing and was discovered to have a dose-limiting cardiovascular toxicity of transient hypotension without compensatory tachycardia. Herein we report the effort beyond ACHN-975 to discover LpxC inhibitors optimized for enzyme potency, antibacterial activity, pharmacokinetics, and cardiovascular safety. Based on its overall profile, compound 26 (LPXC-516, (S)-N-(2-(hydroxyamino)-1-(3-methoxy-1,1-dioxidothietan-3-yl)-2-oxoethyl)-4-(6-hydroxyhexa-1,3-diyn-1-yl)benzamide) was chosen for further development. A phosphate prodrug of 26 was developed that provided a solubility of >30 mg mL^?1 for parenteral administration and conversion into the active drug with a t_1/2 of approximately two minutes. Unexpectedly, and despite our optimization efforts, the prodrug of 26 still possesses a therapeutic window insufficient to support further clinical development.

Full text of DOI:10.1002/cmdc.201900287

Accepted Article
Title: Optimization of LpxC Inhibitors for Antibacterial Activity and
Cardiovascular Safety
Authors: Frederick Cohen, James B. Aggen, Logan D. Andrews, Zahra
Assar, Jen Boggs, Taylor Choi, Paola Dozzo, Ashton N.
Easterday, Cat M. Haglund, Darin J. Hildebrandt, Melissa
C. Holt, Kristin Joly, Adrian Jubb, Zeeshan Kamal, Timothy
R. Kane, Andrei W. Konradi, Kevin M. Krause, Martin S.
Linsell, Timothy D. Machajewski, Olga Miroshnikova, Heinz
E. Moser, Vincent Nieto, Thu Phan, Craig Plato, Alisa W.
Serio, Julie Seroogy, Anton Shakhmin, Adam J. Stein, Alex D.
Sun, Serguei Sviridov, Zhan Wang, Kenneth Wlasichuk, Wen
Yang, Xiaoming Zhou, Hai Zhu, and Ryan T. Cirz
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemMedChem 10.1002/cmdc.201900287
Link to VoR: http://dx.doi.org/10.1002/cmdc.201900287
A Journal of
www.chemmedchem.org

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
Optimization of LpxC Inhibitors for Antibacterial Activity and
Cardiovascular Safety
Frederick Cohen* [a], James B. Aggen[a], Logan D. Andrews[a], Zahra Assar[b], Jen Boggs[a], Taylor
Choi[a], Paola Dozzo[a], Ashton N. Easterday[a], Cat M. Haglund[a], Darin J. Hildebrandt[a], Melissa C.
Holt[b], Kristin Joly[c], Adrian Jubb[a], Zeeshan Kamal[d], Timothy R. Kane[a], Andrei W. Konradi[e], Kevin
M. Krause[a], Martin S. Linsell[a], Timothy D. Machajewski[a], Olga Miroshnikova[d], Heinz E. Moser[a],
Vincent Nieto[a], Thu Phan[d], Craig Plato[c], Alisa W. Serio[a], Julie Seroogy[a], Anton Shakhmin[d], Adam
J. Stein[b], Alex D. Sun[d], Serguei Sviridov[d], Zhan Wang[d], Kenneth Wlasichuk[a], Wen Yang[d],
Xiaoming Zhou[f], Hai Zhu[d], Ryan T. Cirz*[a]
[a]
Dr. F. Cohen 0000-0003-2229-3900, Dr. J. Aggen, Dr. L. Andrews 0000-0002-5092-4530 , J. Boggs, Dr. T. Choi, Dr. P. Dozzo, A. Easterday, Dr. C.
Haglund, D. Hildebrandt, Dr. A. Jubb, T. Kane, K. Krause, M. Linsell, Dr. T. Machajewski, Dr. H. Moser, Dr. V. Nieto, Dr. A. Serio., J. Seroogy, K.
Wlasichuk, Dr. R. Cirz 0000-0002-9848-7421
Achaogen Inc.
1 Tower Place, Suite 400, South San Francisco, CA 94080, United States
E-mail: frederickcohen2017@gmail.com
ryancirz@gmail.com
[b]
[c]
[d]
[e]
[f]
Dr. Z. Assar 0000-0001-6008-6203, M. Holt 0000-0002-7883-924X, Dr. A. Stein 0000-0002-9948-0870
Cayman Chemical Co.
1180 E. Ellsworth, Ann Arbor, MI 48108, United States
K. Joly, Dr. C. Plato
Plato BioPharma, Inc.
7581 W 103rd Ave Unit 300, Broomfield, CO 80021, United States
Dr. Z. Kamal, O. Miroshnikova, T. Phan, Dr. A. Shakhmin, Dr. A. Sun, Dr. S. Sviridov, Z. Wang, Dr. W. Yang, Dr. H. Zhu
Nanosyn, Inc.
3100 Central Expressway, Santa Clara, California, 95051, United States
Dr. A. Konradi
Konradi Molecular
30 Victoria Rd, Burlingame, CA 94010, United States
Dr. X. Zhou
Pharmaron, Inc.
6 Tai-He Road, BDA, Beijing, 100176, China
Supporting information for this article is given via a link at the end of the document.((Please delete this text if not appropriate))
Abstract: LpxC is a Zn2+ deacetylase that is essential for the
survival of most pathogenic Gram(–) bacteria. ACHN-975 (N-((S)-3-
amino-1-(hydroxyamino)-3-methyl-1-oxobutan-2-yl)-4-(((1R,2R)-2-
(hydroxymethyl)cyclopropyl)buta-1,3-diyn-1-yl)benzamide) was the
first LpxC inhibitor to reach human clinical testing and was
discovered to have a dose-limiting cardiovascular toxicity of transient
hypotension without compensatory tachycardia. We report here the
effort beyond ACHN-975 to discover LpxC inhibitors optimized for
enzyme potency, antibacterial activity, pharmacokinetics, and
cardiovascular safety. Based on overall profile, 26 (LPXC-516, (S)-
N-(2-(hydroxyamino)-1-(3-methoxy-1,1-dioxidothietan-3-yl)-2-
oxoethyl)-4-(6-hydroxyhexa-1,3-diyn-1-yl)benzamide) was chosen
for further development. A phosphate prodrug of 26 was developed
that provided solubility of >30 mg/mL for parenteral administration
and conversion to the active drug with a T1/2 of approximately 2
minutes. Unexpectedly, and despite our optimization efforts, the
prodrug of 26 still possesses a therapeutic window insufficient to
growing threat. Patients have succumbed to infections that are
resistant to every approved antibiotic.[3] The combination of scientific
and economic challenges behind this divergence has been
reviewed.[4,5] Current estimates place the mortality due to multidrug-
resistant (MDR) bacteria between 23,000 and 160,000 annually in
the US alone.[1,2]
As part of an on-going effort to introduce antibiotic therapies with
completely novel mechanisms of action, we pursued a research and
development program aimed at bringing a novel inhibitor of the
bacterial
enzyme
UDP-3-O-(R-3-hydroxymyristoyl)-N-
acetylglucosamine deacetylase (LpxC) to patients. This enzyme
catalyzes the first committed step in the synthesis of
lipopolysaccharide (LPS), removal of an N-Acetyl group from the
growing LPS core and is essential to most Gram-negative pathogens
(Figure 1).[6] Additionally, there are no mammalian homologs of LpxC,
eliminating the possibility of on-target toxicity, and increasing interest
in the target.
support further clinical development
.
Introduction
Figure 1. LpxC catalyzes the first committed step of LPS biosynthesis
.
Effective antibiotic therapy is a foundation of modern medicine.
Procedures that are routine and life-saving such as coronary stents,
joint and organ transplants, and cancer chemotherapy rely on
effective antibiotic therapy. Due to bacterial evolution and misuse of
antibiotics, we are in danger of losing this foundation.[1,2] In addition,
the pipeline of new antibiotic therapies has not kept pace with the
The first LpxC inhibitor to reach clinical trials from this program was
ACHN-975. In a single-ascending dose study, we discovered a Cmax
driven dose-limiting toxicity (DLT) of transient hypotension without
tachycardia. Herein we describe our work to identify a second
-
1
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
clinical candidate with a wider therapeutic window for cardiovascular
(CV) safety.
hydroxyl group makes an unfavorable interaction with the protein. A
minimum of two carbons between the acetylene and polar
functionality is required to fully escape the hydrophobic tunnel and
restore potent (<5 nM) enzyme inhibition (5). Small, hydrophobic
branching at the exit of the tunnel is generally tolerated. For example,
the cyclopropyl found in ACHN-975 and the substituted azetidines
found in 4 and 10 both result in potent inhibitors of the enzyme (<5
nM).
Turning our attention to the head group, we found a number of polar
substituents to occupy the sugar pocket and make key interactions
with the protein. For instance, ACHN-975 makes a number of water-
mediated hydrogen bonds to protein, while one of the methyl groups
packs against Phe191. Inhibitors with non-basic substituents are
generally less potent enzyme inhibitors (cf 5 vs 9) unless they make
additional specific interactions. The methyl sulfone replacement 15 –
17 can make additional interactions with the protein and has a
similar IC50. Further substitution of the sulfone afforded a vector to
reach additional protein contacts. Replacing the methylsulfone of 17
(IC50 = 4.6 nM) with amine containing groups (20, IC50 =4 nM and 21,
IC50 = 4 nM) were unable to make additional productive protein
contacts. The oxetane containing substitution found in 19 makes
additional hydrophobic contacts, as well as an additional hydrogen
bond with the backbone amide of Phe193 Figure 3 B), and results in
more than a 10-fold improvement in binding affinity (IC50 = 0.3 nM).
Oxetane 13 and thietanedioxides 23 – 26 were designed to interact
with Lys238. As such, modification of the dimethyl of 4 (IC50 = 2 nM)
to the oxetane, 13 (IC50 = 0.76 nM) resulted in a 2–3-fold increase in
LpxC inhibition. Furthermore, introduction of the thiethanedioxide
During the development of ACHN-975 we discovered that the
respiratory pathogen Pseudomonas aeruginosa was more sensitive
to LpxC inhibition in in vivo models than Enterobacteriaceae species
such as Escherichia coli and Klebsiella pneumoniae. Furthermore,
the LpxC substrate binding pocket in Enterobacteriaceae species
has a longer hydrophobic tunnel than P. aeruginosa. This structural
difference results in a requirement for longer, more amphiphilic
inhibitors, which we thought would make solving the toxicity
problems more challenging. Finally, the pipeline of novel anti-
pseudomonal
agents
is
even
sparser
than
for
the
Enterobacteriaceae. For these reasons, we focused our second
optimization program exclusively on P. aeruginosa.
Results and Discussion
SAR of Bis Alkyne Scaffold. LpxC inhibitors are generally composed
of three segments, a hydroxamic acid, or other bidentate metal
ligand to chelate the catalytic Zn2+ atom, a polar moiety that
occupies the sugar-binding pocket (the ‘head group’), and
a
hydrophobic tail to fill the hydrophobic tunnel that is usually occupied
by the C10-C16 aliphatic chain in the natural substrate.[7,8] In this work,
we focused on a benzamide bis-acetylene linker to span the
hydrophobic tunnel and explored the effect of structural modification
of the head group and of the portion of the tail that extended out of
the hydrophobic tunnel. Compounds were initially evaluated for their
ability to inhibit purified LpxC and anti-bacterial potency and the data
are summarized in Table 1. The screening panel contained wild-type
strains of P. aeruginosa (PAO1). To measure the ability of
compounds to cross the outer membrane in P. aeruginosa, we
added polymyxin B nonapeptide (PMBN) to permeabilize the outer
membrane.[9] We used a strain of P. aeruginosa with its major efflux
systems (MexAB-oprM, MexCD-oprJ, MexEF-oprN, and MexXY)
deleted or rendered inactive to measure each compound’s efflux
liability.[10] We also evaluated the activity of compounds against five
recent clinical isolates of P. aeruginosa. The geometric mean of the
minimum inhibitory concentrations (MICs) of these five strains (listed
as “MIC5” in Table 1) was an accurate predictor of the MIC90 against
a larger panel of 40 clinical isolates.[11] To get an accurate estimate
of the total drug levels that would be required for efficacy, we also
measured plasma protein binding (PPB) as a primary assay in our
testing funnel.[12] We then looked to the PPB adjusted MIC as our
primary measure of activity.
We further evaluated these compounds for on-target antibacterial
activity by assaying them against a strain of P. aeruginosa that
overexpresses LpxC. All compounds in Table 1 show ≥ 8-fold
increase in MIC against this strain (data not shown), indicating that
they remain on-target.
SAR for LpxC Inhibition. Compounds that lack polarity at the solvent-
exposed end of the tail, such as 1, are generally potent enzyme
inhibitors (IC50 < 1 nM) with potent antibacterial activity across
species and low (≤ 4-fold) shift in MIC upon addition of PMBN,
indicating high permeability across the outer membrane. Larger
shifts were seen in the efflux-deficient strain (≥ 10-fold), indicating
that these compounds are relatively good substrates for efflux
pumps. As a group, these compounds also demonstrated a number
of toxicities including cytotoxicity to endothelial cells and hemolysis
of red blood cells (data not shown).
resulted in more than a 10-fold increase in LpxC inhibition (22, IC50
=
10 nM; 26, IC50 0.71 nM). In this case, the quaternary methoxy
substitution was optimal. The single atom change of carbon for
oxygen (24, IC50 = 4.3 nM) resulted in a 5-fold loss of binding affinity.
Modeling studies suggested this was the result of a conformation
change effect. Decreasing the size from ethyl to a methyl group (23,
IC50 = 3.5 nM) did not substantially improve inhibition and increasing
the size to a cyclopropyl (25, IC50 = 1.6 nM), while increasing the
hydrophobic interactions, was still 2-fold higher than the methoxy
group.
Antibacterial Activity. Consistent and predictable structure-
antibacterial activity relationships are made more difficult due to the
multitude of effects that play a role in the ability of an inhibitor to get
to the target location. Often, a narrow window of physicochemical
parameters defines the ability of a molecule to cross the outer
membrane via porin, be effluxed by efflux pumps, and cross the
inner membrane to gain access to the target in the cytoplasm.
Therefore, it is not always the case that a compound with a potent
inhibitory activity will result in acceptable MICs. By using the
engineered strains of P. aeruginosa described above, either with or
without the addition of PMBN, we are able to glean important
information about the properties of the molecule that improve the
likelihood of target engagement.
Some trends emerge from this work. Firstly, compounds from this
scaffold were generally highly permeable across the outer
membrane, with MIC shifts of <5 upon PMBN addition. Exceptions to
this trend include 8, 16 and 19, which lack a basic amine. These
observations are consistent with other work in the field, which found
that compounds containing a basic amine had higher cellular
accumulation than non-basic compounds.[13, 14] For example, the ratio
of the parental to pump knock-out P. aeruginosa strain is 8.6 for
amine containing head group 5, but 20 for the hydroxyl containing
head group, 9, and 39 for the sulphone containing head group, 17.
Reinstalling a basic group on the tail of compounds with non-basic
head groups removes some of the efflux liability as demonstrated by
the comparison of 9 (non-basic head and tail, ratio 20) with 12 (non-
basic head, basic tail, ratio 11).
Rat cardiovascular (CV) Safety. To screen compounds for
hemodynamic effects, we developed a high content assay in rats.
Briefly, 4 male Sprague-Dawley (SD) rats were anesthetized with
isoflurane. The femoral artery was cannulated and interfaced with a
pre-calibrated pressure transducer for continuous collection of heart
rate and blood pressure. The right femoral vein of each rat was
cannulated for infusion of test article, and the left jugular vein was
cannulated for collection of toxicokinetic samples. After equilibration,
These observations led us to reduce the amphiphilic nature of the
compounds by adding polar moieties on the solvent-exposed end of
the hydrophobic tail. Proper placement of these groups is key to
maintaining potent enzyme inhibition. For instance, ACHN-975
containing a hydroxymethyl substituted cyclopropane is equipotent
to 1. Placing a hydroxy group close the hydrophobic tunnel as in
compounds 2 and 3 results in a 10-fold increase in IC50, and a
concomitant loss in antibacterial activity. Analysis of the X-ray co-
structure of ACHN-975 bound to LpxC (Figure 2) suggested that the
carbon atom attached to the second acetylene lies at the exit point of
the hydrophobic tunnel. Therefore, the introduction of the polar
all
animals
2
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
Table 1. Structure, Potency, and Activity of LpxC Inhibitors.
O
Tail
Head
HN
MIC[f]
(µg/mL)
Head
Tail
1
0.62
0.68
0.048
0.061
0.017
0.0041
0.0055
0.0017
0.0038
0.16
0.38
7/ND
2.3
3.0
ACHN-
975
0.035
13/13[a]
2
3
NH2
19
12
0.71
1.1
0.63
1
0.11
0.22
0.079
0.16
3.5
8.8
72/77[b,c]
72/77[b,c]
4.9
12
NH
OH
O
4
5
2.2
2.2
0.41
0.12
0.79
0.12
0.062
0.014
0.079
0.013
2.0
41/51[a]
54/49[b]
4.9
1.5
0.81
6
1.8
19
0.25
0.063
0.013
0.25
0.0039
0.25
1.1
16
15/14[a]
7.3
–
7
8
2
4
ND
3.7
13
0.22
0.24
0.043
0.13
0.012
0.012
0.0032
0.01
1.1
1.8
7/9[a]
16
9
(LPXC-
289)
10
(LPXC-
313)
31/28[b]
5.8
4.8
0.38
0.2
0.022
0.008
2.4
34/27[b]
7.1
11
12
9.1
2.7
0.66
0.3
1
0.19
0.25
4
74/59[b]
49/49[a]
5.4
4.3
0.25
0.027
0.039
2.1
13
14
0.76
53
0.5
6
0.35
0.031
0.5
0.031
0.35
3.7
31
51/54[a]
37/40[a]
7.3
84
8
3
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
O
Tail
Head
HN
MIC[f]
(µg/mL)
Head
Tail
15
0.63
0.5
0.25
0.026
0.016
4
39/42[a]
10
16
17
1.4
4.6
0.5
0.018
0.11
0.031
0.0039
0.0025
3.4
2.4
40/26[a]
31/29
8.5
7.7
0.38
0.0098
18
19
5.3
0.3
4
2
2
0.13
0.063
24
9
ND
–
0.25
0.061
0.0049
21/23
43
O2
S
20
4
4
8
4
0.25
0.044
0.012
0.13
0.025
0.0086
32
4.4
1.9
ND
–
NH
NH
O
OH
21
22
0.63
0.28
0.35
0.098
28/25[a]
34/31[a]
16
5.6
10
23
24
25
3.5
4.3
1.6
0.5
0.5
0.25
0.25
0.25
0.28
0.022
0. 022
0.022
0.027
0.0078
0.011
2.8
2.5
2.6
2.3
31/23[a]
24/20[a]
21/14[a]
9
O
O
S
10
12
7.4
NH
OH
O
0.5
0.0078
0.0081
26
(LPXC-
516)
0.71
0.51
31/25[a]
nd/15
27
>1000
na
4
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
O
Tail
Head
HN
MIC[f]
(µg/mL)
Head
Tail
28
70[d]
nd
nd/14
[a] Determined by ultrafiltration, r = rat, h = human; [b] determined by equilibrium dialysis; cdetermined for a 1:1 mixture of 2 and 3. [d] 98% ee. [e]
Geometric mean of five MICs, run in at least duplicate. [f] Geometric mean of MIC, run in at least duplicate.
Figure 2. Structure of ACHN-975 co-crystallized with P. aeruginosa LpxC at a resolution of 2.2 Å (PDB 6MOO). (A) The hydroxamate moiety of
ACHN-975 (magenta) at the head group of the molecule is observed coordinating the active site Zn2+ ion (grey sphere) and the tail side of the
molecule extends through a hydrophobic tunnel. The hydroxymethyl cyclopropyl tail group protrudes entirely out of the surface of the tunnel exit. (B) A
view of ACHN-975 within the active site. Three crystallographic water molecules shown as red spheres are in position to accept hydrogen bonds from
the amine-containing head group. Distances between the oxygen atoms of the waters and the amine nitrogen atom of ACHN-975 are labelled and
highlighted by dashed yellow lines.
Figure 3. Structures of key inhibitors co-crystallized with P. aeruginosa LpxC. (A) Inhibitor 17 (green sticks) co-crystallized at 1.84 Å (PDB 6MO4). The
amine side chain of Lys238 and backbone amide of Phe193 appeared to be to in close enough proximity to serve as hydrogen-bond donors for
expanded versions of the 17 head group. (B) Inhibitor 19 (pink sticks) was co-crystallized at 1.85 Å (PDB 6MO5). The addition of the methyl oxetane
moiety resulted in the oxygen atom of the oxetane positioned within hydrogen-bonding distance to the Phe193 backbone amide. (C) Inhibitor 26 (cyan
sticks) was co-crystallized at 1.85 Å (PDB 6MOD). The presence of the thietane dioxide moiety results in a small conformational change of the Lys238
side chain. One of the oxygen atoms of the thietane dioxide moiety appears within hydrogen-bonding distance to the nitrogen atom of the re-
positioned Lys238.
received intravenous vehicle (HPCD: 20% hydroxypropyl
cyclodextrin, 250 mM Tris, in water, pH 9; or A5D: 2.5% dextrose in
20 mM acetate buffer, pH 5) for 30 minutes, followed by infusion of
test article in increasing doses for 30 minutes at each dose level, or
5
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
until pre-determined stopping criteria were met (see Methods
section). Toxicokinetic samples were taken at the end of the low, mid,
and high doses, and at 2, 6, 10, and 30 minutes following cessation
of infusion. These data were used to determine a maximum free
drug concentration that resulted in < 20% change in CV parameters.
This concentration was defined as the maximum tolerated
concentration (MTC). Plotting these data as in Figure 4, showed that
this assay reproduced the effect observed in humans in the Phase 1
trial of ACHN-975 at the same free drug concentration (manuscript
in preparation). As an estimate of the therapeutic window for CV
safety, we divided the MTC by the antibacterial potency, represented
by the MIC5, and these data are summarized in Table 2, along with
the measured or calculated pKa of the test compounds and the
doses administered.
We systematically investigated the structure-toxicity relationships for
this series of compounds using the rat CV assay. Understanding the
essentiality of the hydroxamic acid to antibacterial potency and the
perceived liabilities of this functionality,[15] our first modification was
to remove it by reduction to the amide 27. While this compound had
no antibacterial activity, its MTC was nearly identical to ACHN-975
(3 vs 5 µg/mL). We thus believe that it is unlikely that the
hydroxamate alone is responsible for the observed toxicity. We next
hypothesized that the amine functionality on the head group in
proximity to the aromatic core was an ion-channel pharmacophore.
This inspired us to move the amine to the distal end of the molecule
in analogs 12 and 15. These compounds had improved CV
tolerability, but lost antibacterial potency, and thus still had
unacceptably narrow safety windows (<30x). Neither sterically
hindering the amine nor modulating the pKa resulted in an
acceptable window; any compound with
a basic pKa ≥ 6.5
demonstrated a similar effect on CV parameters. It was not until we
tested compounds devoid of basic amines that we found compounds
the met our criteria for CV safety (>30x). As a control for vehicle
effects, ACHN-975 was also dosed with the HPCD formulation and
found to have a similar MTC.
In Vitro Pharmacology. To attempt to identify the target(s)
responsible for the observed CV effects, and develop a higher-
throughput assay, ACHN-975 was profiled against a CEREP panel
(Eurofins) of 123 human receptors, ion channels and enzymes at a
single concentration of 10 µg/mL. Of these assays, only the site 2
sodium channel assay showed a significant signal of 50% inhibition.
We followed up this result by testing a series of compounds for
which we had rat in vivo data in the sodium channel assay at 100
µg/mL and the results are listed in Table 3. We observed no
correlation between activity in this assay and the MTC or CV safety
window. The caveat with this analysis is that we are comparing rat in
vivo data to human ion channel data.
6
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
Figure 4. Rat hemodynamic assay data. (A) Effects of ACHN-975 in rats. (B) Effects of compounds 5 and 9. Compounds 5 and 9 differ by a single heavy atom
(NH2 → OH).
Table 2. Summary of Cardiovascular Safety in Rat.
MTC
MIC5
(µg/mL)
0.38
NA
Compound
ACHN-975[a]
27[a]
pKa
7.1/8.3[c]
6.9[d]
Doses (mg/kg/h)
12.5/37.5/75
50/150/450
37.5/75/150
100/200/400
75/300/450
50/150/450
50/150/450
300
(µg/mL)[f]
CV window
5
3
13
-
-
26
<6
4.2
16
6
28[a]
nd
3.5
26
<12
3
18
12
29
50
NA
6.2
2.0
0.81
1.1
2.1
4
4.4
2/3[a, e]
4[a]
7.2/8.2[d]
7.4/8.2[d]
7.2/8.3[c]
nd
5[a]
6[a]
12[a]
7.4/8.0[c]
7.4/8.6
nd
15[a]
50/150/400
50/150/450
7
11
21[a]
7
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
9[b]
10[b]
9.1[c]
7.9[c]
7.6[c]
nd
50/150/450
50/150/450
50/150/450
75/150/300
>82
>81
>74
>73
1.1
2.4
2.4
2.3
>75
>34
>31
>32
17[b]
26[b, g]
[a] Formulated in A5D (A5D: 2.5% dextrose in 20 mM acetate buffer, pH 5. [b] Formulated in HPCD: 20% hydroxypropyl cyclodextrin, 250 mM Tris, in water, pH 9.
[c] Measured pKa by aqueous titration. [d] Calculated with ACD labs model with training set of measured data.1:1 [e] mixture of 2 and 3. [f] MTC is the maximum
tolerated concentration, defined as the highest measured free drug concentration that results in <20 % change in blood pressure or heart rate. [g] Tested as a
racemic mixture
.
We further assayed a series of compounds in a FastPatch® ion
current assay in CHO cells (ChanTest). Compounds with wide or
narrow CV safety windows in the rat in vivo assay showed no
inhibition of the L-type or T-type calcium channels, Kir6.2/SURAA
potassium channel, or Nav1.5 sodium channel at a concentration of
100 µg/mL (data not shown).
appeared to have a wider TW, it displayed higher clearance than 10
and 26 in dog and monkey PK studies. We chose 26 for further
investigation as the primary alcohol on the tail provided a more
convenient handle for prodrug development relative to the tertiary
alcohol on the head of 10. Additional microbiological characterization
of ACHN-975, 9, 10 and 26 that also contributed to candidate
selection will be described in an accompanying manuscript.[16]
Prodrug Development. Our Target Product Profile called for an
injectable formulation for treatment of seriously ill patients with an
anticipated dose >1 g per day. We thus required high solubility (>20
mg/mL) in a vehicle suitable for injection. ACHN-975 met this goal
via the basic amine and has a solubility of >50 mg/mL at pH 5.5. Our
second lead, 26 has a solubility of 1 mg/mL at pH 7.4 but does not
increase at lower pH. The basic HPCD formulation used for PK and
CV studies would not be a cceptable for further development, due to
instability of the hydroxamate at elevated pH. We thus undertook the
development of a phosphate prodrug to increase solubility.[16] The
synthesis of this prodrug is illustrated in Scheme 2.
Table 3. Sodium Channel Binding Data
% Inhibition
MTCa
(µg/mL)
5
CV
window [a]
13
26
>75
Compound
ACHN-975
(100 µg/mL)
97
31
11
41
35
2/3
9
10
21
26
>82
>81
50
>34
11
[a] from rat hemodynamic assay
Prodrug 55 was soluble at >30 mg/mL at pH 6.4 in 0.9% saline, a
formulation that allowed dosing of rats up to 500 mg/kg. The
pharmacokinetics of 55 were measured in rats and compared to
direct dosing of active drug. The results are shown in Figure 5 and
Table 5. The phosphate converted rapidly to active drug with a half-
life of 2.4 minutes. Bioavailability of active drug from the prodrug was
100%. Based on these results, 55 was chosen for further
development.
Pharmacokinetics. Compounds were screened for pharmacokinetics
(PK) in rats. In general, the compounds displayed moderate
clearance values 20 – 40% of liver blood flow (Table 4). Two notable
exceptions are ACHN-975 and 14. The low clearance of 14 was the
inspiration for synthesis of other 4-member ring head groups that
lacked basic amines, and eventually led to candidate 26. To effect
comparison between compounds of differing potency and exposure,
we normalized these values by looking at the ratio of exposure (free
AUC) from a constant dose to antibacterial activity as measured by
the MIC5. Of the compounds that met our threshold for CV safety,
10 and 26 had the highest exposure:potency ratio. Although 9
Table 4. Pharmacokinetic Parameters
AUC 0-∞
(µg*h/mL)
12
Dose
(mg/kg)
10[a]
Cl
Vss
(L/kg)
PPB
freeAUC:MIC5
Compound
species
SD rat
(L/hr/kg)
(% free)[c]
ratio
4.1
ACHN-975
0.83
2.4
1.9
1.3
3.5
0.63
2.1
3.6
2.3
2.2
1.1
0.86
1.3
0.82
1.70
0.73
0.48
0.95
0.27
0.46
0.91
1.0
13
51
32
32
49
40
42
31
23
20
25
37
31
5[a]
2.1
5.1
7.7
0.53
0.82
0.86
2.6
4
10[b]
10[b]
10[a]
10[a]
10[a]
5[b]
9
10
12
14
15
22
23
16
4.8
1.4
4.4
4.6
8.1
11.7
8.2
0.21
0.50
0.24
0.35
0.37
0.85
NA
10[b]
10[b]
10[b]
10[b]
10[b]
24
26
0.68
0.34
0.38
0.57
SD rat
dog
monkey[d]
NA
[a] formulated in A5D: 2.5% dextrose in 20 mM acetate buffer, pH 5. [b] formulated in 20% HPCD 20% hydroxypropyl cyclodextrin, 250 mM Tris, in water, pH 9.
[c] determined by ultracentrifugation.[d] Cynomolgus monkey
.
Prodrug Rat CV Safety. As a final evaluation prior to initiating IND-
enabling studies on 55, we re-tested the prodrug in the rat CV safety
formulation was used. Surprisingly, dosing of this prodrug resulted in
the same CV effects observed upon dosing the basic compounds,
model. Given its increased solubility,
a simple 0.9% saline
8
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
resulting in a CV safety window of <30-fold. The underlying cause
for the different outcomes is not known.
executed an Ireland-Claisen reaction to install a vinyl group that
could be further functionalized by reduction or cyclopropanation to
provide analogs 24 and 25, respectively.
Synthesis of the methoxy substituted thietane required
a
modification of our standard route, illustrated in Scheme 2. We
discovered that hydroxamate formation from a methyl ester in the
final step resulted in racemization of the α-stereocenter. We thus
developed a route in which a protected hydroxylamine was coupled
to acid 50, then carried through hydrogenation of the azide and
coupling to the fully elaborated tail 53 to afford protected prodrug 54.
The protecting groups were removed with sequentially with TMSBr
and HCl to afford prodrug 55, which could be conveniently purified
by precipitation.
Conclusions
ACHN-975 is the first LpxC inhibitor to be tested in humans. A
Phase-I clinical trial revealed transient Cmax-associated hypotension
without concomitant tachycardia as a DLT that would limit its utility
as a drug. We further optimized LpxC inhibitors for enzyme potency,
antibacterial activity, pharmacokinetics, and therapeutic window in a
rat CV safety assay. We discovered that compounds from this class
with a basic pKa > 6.5 all displayed an unacceptably low TW of <30-
fold. Extensive investigation into the mechanism of this unusual CV
effect failed to identify a single molecular target responsible for the
observed effect in vivo. It remains possible that weak activity against
several targets causes the effect. However, given the diversity of
structures that show the same effect (including additional data not
shown), we believe it unlikely that they all act on the same target(s).
Rather we believe it to be a nonspecific effect of these basic amines.
This conclusion is further supported by the CV data for 28, the
enantiomer of ACHN-975, which displays and identical MTC. We
chose 26 for further investigation based on its favorable overall
profile, and developed a solubility-enhancing phosphate prodrug that
provided rapid conversion to parent. Sadly, the final lead prodrug
compound, when tested in the rat CV safety assay (formulated in
A5D) demonstrated unacceptable CV toxicity, in stark contrast to the
results we obtained dosing the parent compound in an HPCD
formulation. The cause for this discrepancy is still not understood.
Figure 5. Figure 5. Pharmacokinetic profiles of 26, 55, and 26 from 55 in SD
rat.
Table 5. Pharmacokinetic Parameters of 26, 55, and 26 from 55 in SD Rat.
AUC
AUC 0-∞
(µg/mL) (µg*h/mL)
Compound
(dose
mg/kg)
26
(12)
55
Cmax
T1/2
(h)
Cl
(L/h/kg)
Vss
(L/kg)
46
7.8
0.7
1.5
0.32
29
25
3.4
5.6
0.04
0.7
4.3
-
0.12
-
(10)
26 from 55
Chemistry. The synthesis of the methyl substituted thietane head
began by addition of LiMe2Cu to olefin 30 with BF3·Et2O as a
promoter (Scheme 1). We relied on an Evans azide transfer scheme
to set the amino stereocenter. The sulfide was oxidized to sulfone 35,
which was converted to amine 38 through a standard 3-step
sequence. Coupling to tail 39 and treatment with hydroxylamine
provided analog 23.
Conjugate addition of groups other than methyl to olefin 30 failed to
provide any product. Thus, to expand the SAR at this position, we
9
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
Scheme 1. Synthesis of carbon-substituted thietane-Headgroup LpxC Inhibitors (a) LiMe2Cu, BF3·Et2O, Et2O, –78 °C to 0 °C, 88%; (b) 1M NaOH, THF, MeOH,
97%; (c) PivCl, TEA, THF, –15 °C then Li(S)-(+)-4-phenyl-2-oxazolidinone, –78 °C to 0 °C, 67%; (d) NaHMDS, THF, –78 °C then TrisylN3, 80%; (e) AcOOH, THF,
H2O, 0 oC; (f) LiOOH, THF, H2O, –5 °C; (g) TMS-diazomethane, DCM, MeOH; quant, 3 steps; (h) H2, Pd·C, Pd(OH)2·C, EtOAc, MeOH; quant; (i) 39, HATU,
DIPEA, DMF; (j) 50% aq. NH2OH, iPrOH, 0 °C; (k) DIBALH, PhMe, –78 °C, 95%; (l) triethyl orthoacetate, propionic acid, PhMe, µW 175–180 °C, 1.2 h, 78%; (m)
CH2N2, Pd(OAc)2, Et2O, 88%.
10
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
Scheme 2. Synthesis of Methoxy-thietane Head Group and Prodrug 55. (a) NaOMe, MeOH, then water, 65%; (b) PivCl, TEA, THF, –15 °C then (S)-(+)-4-phenyl-
.
2-oxazolidinone, LiCl, –10 °C to RT, 62%; (c) KHMDS, THF, –65 °C then TrisylN3, 62%; (d) RuCl3 3H2O, NaIO4, MeCN, 86%; (e) LiOOH, THF, H2O, –10 °C, 85%;
(f) H2NOTHP, DIC, THF; (g) PPh3, THF, water, 68%, two steps; (h) dibenzyl N,N-diisopropylphosphoramidite, tetrazole, MeCN; (i) H2O2, 38%, two steps; (j) HATU,
DIPEA, DMF, 58%; (k) N,O-bis(trimethylsilyl)acetamide, TMSBr, DCM; (l) HCl, MeOH, 50%, two steps; (m) 52, HATU, DIPEA, DMF; (n) HCl, MeCN, 25%, two
steps.
then stirred at 50 °C for 2 hrs, at which time TLC and GC indicated
complete consumption of starting material. The reaction is then
cooled to 10 °C, followed by addition of water (250 L, 10 volumes) in
portions, maintaining the temperature ≤ 10 °C. Upon completion of
addition, the mixture was maintained at 20 °C for 2 h, The MeOH is
Experimental Section
removed under reduced pressure and the aq phase washed with
Chemistry. All final compound for biological assay were determined
EtOAc (1 X 200 L, 2 X 100 L). The aq phase is then acidified to pH
to be >95% by HPLC-MS. All commercially available solvents and
1–2 with 12N HCl, and extracted with EtOAc (1 X 100 L, 2 X 50 L).
reagents were used as received.
The combined organic phases were washed with brine (2 X 50 L),
dried (Na2SO4), filtered and concentrated to afford 27 Kg (65%) of
Compound 1 is described in WO2008154642[18]; Compounds 2, 3, 5,
9, 39 are described in WO2013170030[19]; Compounds 6, 8 are
described in WO2012154204[20]; Compounds 4, 10, 11, 12 are
described in WO2014165075[21]; Compounds 7, 14, are described in
WO2017223349[22]
the title compound as a yellow solid: 1H NMR 400 MHz CDCl3 δ 7.59-
7.57 (m, 1H), 6.95-9.93 (m, 1H), 4.23-4.18 (m, 3H), 4.04 (s, 2H),
3.97 (s, 3H), 1.29-1.26 (m, 3H).
(S)-3-(2-(3-methoxythietan-3-yl)acetyl)-4-phenyloxazolidin-2-one
(47). A reactor with overhead stirrer was charged with THF (30 L,
20.0 volumes), 2-(3-methoxythietan-3-yl)acetic acid (1.64 Kg, 10.1
mol, 1.1 eq), and cooled to –10 °C. Pivaloyl chloride (1.22 Kg, 10.1
mol, 1.1 eq) and triethyl amine (2.42 Kg, 23.9 mol, 2.6 eq) were
charged sequentially at –10 °C and the mixture held at –10 °C for 3h.
(S)-4-phenyloxazolidin-2-one (1.5 Kg, 9.19 mol, 1.00 eq.) and LiCl
(456 g, 10.8 mol, 1.17 eq.) were added sequentially at –10 °C. The
mixture allowed stirred at 10–20 °C for 16 h. The mixture was cooled
Additional Schemes and Procedures for preparation of 13, 15, 17–19,
23–26 and ACHN-975 are provided in the SI.
2-(3-methoxythietan-3-yl)acetic acid (46). A reactor with an overhead
stirrer was charged sequentially with methanol (125 L, 5 volumes),
methyl 2-(thietan-3-ylidene)acetate (25.0 Kg, 173.4 mol, 1 eq.) and
30% NaOMe (53.1 Kg, 1.7 eq, 2.12 weight eq.). The mixture was
11
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
to –10 °C, then water (10 L) was added. The mixture was extracted
with EtOAc (2 X 10 L), the combined organic phases were washed
with brine (1 X 10 L) and concentrated under vacuum to a solid. The
product from 3 batches was combine to afford 7.1 Kg (62%) of the
title compound as an off-white solid. 1H NMR 400 MHz CDCl3 δ 7.29-
7.38 (m, 5H), 5.44-5.47 (m, 1H), 4.69 (t, J = 8.6 Hz, 1H), 4.26-4.29
(m, 1H), 3.73 (s, 2H), 3.63 (t, J = 9.6 Hz, 1H), 3.26 (s, 3H), 3.08-3.11
(m, 2H).
dioxidothietan-3-yl)acetic acid (510 g, 2.17 mol) was dissolved in
THF (500 ml) and cooled (Tint= –8.5 °C). O-(tetrahydro-2H-pyran-2-
yl)hydroxylamine (254.1 g, 2.17 mol, 1.0 eq.) was added and the
mixture was stirred for 10 min. N,N′-Diisopropylcarbodiimide (273.9
g, 2.17 mol, 1.0 eq.) was then added dropwise, and the mixture was
stirred for 0.5 h, at which time it was filtered and triphenylphosphine
(1023 g, 3.91 mol, 1.8 eq) was added to the filtrate and stirred for 1 h.
Water (780 mL, 43.4 mol, 20.0 eq.) was added, and the mixture was
heated to 70 °C and stirred overnight. The cooled solution was
evaporated and water (5 L) was added to the residue and stirred at
60 °C for 1 h. The cooled solution was filtered, and sodium chloride
(300 g) was added. Extraction with 16% iPrOH in CHCl3 (5 x 4 L)
and evaporation of the combined solvent followed by drying under
(S)-3-((S)-2-azido-2-(3-methoxythietan-3-yl)acetyl)-4-
phenyloxazolidin-2-one (48). A reactor with an overhead stirrer was
charged with THF (20 L, 13.3 volumes) and (S)-3-(2-(3-
methoxythietan-3-yl)acetyl)-4-phenyloxazolidin-2-one (1.5 Kg, 4.88
mol, 1.00 eq.). The solution was cooled to –65 °C then KHMDS
(5.37 L of a 1.0M solution in THF, 5.37 mol, 1.1 eq) was slowly
added, then the solution held at –65 °C for 1 h. A solution of trisyl
azide (1.81 Kg, 5.86 mol, 1.2 eq.) in THF (4 L) was added at –65 °C
and the mixture held at that temperature for 1 h. Acetic acid (1.76 Kg,
29.3 mol, 6.00 eq.) and Me4NOAc (2.6 Kg, 19.5 mol, 4 eq.) were
added sequentially. The mixture was allowed to warm to 20 °C and
held for 16 h. The mixture was diluted with water (10 L) and EtOAc
(10 L), and the phases separated. The aq phase was extracted with
EtOAc (10 L), and the combined organic phases were washed with
sat. aq. NaHCO3. The organic phase was concentrated under
vacuum < 45 °C to a solid. The product from two batches was
combined and slurried with EtOH (15 L) for 2 h. The solid was
collected by filtration, dried under vacuum to afford 4.5 Kg (62%) of
high
vacuum,
afforded
(2S)-2-amino-2-(3-methoxy-1,1-
dioxidothietan-3-yl)-N-((tetrahydro-2H-pyran-2-yl)oxy)acetamide as
an off-white foam (458 g, 68 %).
4-(6-((bis(benzyloxy)phosphoryl)oxy)hexa-1,3-diyn-1-yl)benzoic acid
(53). To a solution of compound 39 (1000 g, 4.67 mol, 1.00 equiv) in
acetonitrile (12.5 L) was added tetrazole (655 g, 9.35 mol, 2 equiv)
and the resultant mixture was stirred for 30 min. To this was added
dibenzyl N,N-diisopropylphosphoramidite (1937 g , 5.61 mol, 1.2
equiv) drop wise at 15–20 °C and the reaction mixture was stirred at
RT for additional 16 h. The reaction mixture was cooled to 0 °C in an
ice/water bath and 30% hydrogen peroxide (5.29 L, 46.72 mol, 10
equiv) was added drop wise (Note: highly exothermic reaction during
initial addition) and stirred for 1 h. The mixture was quenched with
5% HCl solution (pH ~2) and diluted with ethyl acetate (20 L) and
1
the title compound as an off-white solid: H NMR 400 MHz CDCl3 δ
7.37-7.42 (m, 3H), 7.31-7.33 (m, 2H), 5.55 (s, 1H), 5.45-5.48 (m, 1H), brine (4 L). This was stirred for 30 min and then filtered through a
4.74 (t, J = 8.8 Hz, 1H), 4.33-4.36 (m, 1H), 3.41-3.51 (m, 8H).
pad of Celite®. The organic layer was separated from the filtrate and
washed with brine (2 X 2000 ml) and dried (Na2SO4). The solvent
was concentrated under vacuum and the crude compound obtained
was diluted with acetonitrile (3000 mL), stirred for 2 h at 0 °C and
filtered. The solid obtained was further slurred in methanol (2000
mL) for 1 h, filtered and dried under vacuum to afford compound 53
(856 g, 38%) as off white solid.
(S)-3-((S)-2-azido-2-(3-methoxy-1,1-dioxidothietan-3-yl)acetyl)-4-
phenyloxazolidin-2-one (49). A reactor with an overhead stirrer was
charged sequentially with MeCN (3.5 L, 7.0 volumes) and s(S)-3-
((S)-2-azido-2-(3-methoxythietan-3-yl)acetyl)-4-phenyloxazolidin-2-
one (0.50 Kg, 1.4 mol, 1.0 eq.) water (3.5 L, 7.0 volumes), and
.
RuCl3 3H2O (0.75 g, 2.87 mol, 0.002 eq.). NaIO4 (629 g, 2.94 mol,
2.05 eq.) was added in portions, maintaining the internal
temperature ≤ 30 °C. The mixture was stirred for 3 h. Aq. NaHCO3
(0.5 L) was added and the mixture stirred for 15 min. The mixture
was extracted with DCM (2 X 2 L) and filtered. The filtrate was
extracted with DCM (2 L). The combined organic phases were
washed with Na2S2O3, brine, dried and evaporated under vacuum at
40°C to afford a solid. The solid was slurried with EtOH (2 L) for 1h,
the solid was collected by filtration, dried under vacuum to afford 470
g (86%) of the title compounds as a colorless solid: 1H NMR 400
MHz CDCl3 δ 7.41-7.45 (m, 3H), 7.30-7.39 (m, 2H), 5.91 (s, 1H),
5.47-5.50 (m, 1H), 4.80 (t, J = 9.0 Hz, 1H), 4.43-4.51 (m, 3H), 4.22-
4.40 (m, 2H), 3.46 (s, 3H).
dibenzyl (6-(4-(((1S)-1-(3-methoxy-1,1-dioxidothietan-3-yl)-2-oxo-2-
(((tetrahydro-2H-pyran-2-yl)oxy)amino)ethyl)carbamoyl)phenyl)hexa-
3,5-diyn-1-yl) phosphate (54). (2S)-2-amino-2-(3-methoxy-1,1-
dioxidothietan-3-yl)-N-((tetrahydro-2H-pyran-2-yl)oxy)acetamide (4)
(458.5 g, 1.49 mol, 1.30 eq.) was dissolved DMF (4.5 L) at –5 °C
and 4-(6-((bis(benzyloxy)phosphoryl)oxy)hexa-1,3-diyn-1-yl)benzoic
acid (5) (540 g, 1.14 mol, 1.00 eq.) was added and stirred for 10
min. DIEA (5.95 L, 3.43 mol, 3.00 eq.) was then added in one
portion followed by HATU (477 g, 1.26 mol, 1.11 eq.) in one portion
(Tint = 11 °C). The resulting mixture was stirred for 0.5 h after which
time the reaction mixture was diluted with H2O (8 L) and extracted
with EtOAc (8 L). The org layer was concentrated in vacuo to
provide a dark green oil (~1.2 L) which was purified by flash
chromatography on silica gel, eluting with 0 – 90 % EtOAc in hexane
to provide 81 (509 g, 58 %) as a greenish solid.
(S)-2-azido-2-(3-methoxy-1,1-dioxidothietan-3-yl)acetic acid (49). A
solution of (S)-3-((S)-2-azido-2-(3-methoxy-1,1-dioxidothietan-3-
yl)acetyl)-4-phenyloxazolidin-2-one (500 g, 1.31 mol), THF (11.25 L)
and H2O (3.75 L) was cooled to –10 °C and 30% H2O2 (806 mL, 7.89
mol) was added. The mixture was stirred for 60 min, at which time
LiOH (69.0 g, 2.88 mol) in H2O (750 mL) was added, and the
solution was stirred at –10 °C for 0.5 h. The reaction mixture was
quenched by the addition of 10% Na2S2O3 in H2O (9.0 L) and
warmed to room temperature. The mixture was extracted twice with
DCM (15 L, 11 L) and the aqueous solution was acidified to pH ~1
with 6N HCL and extracted with 15% iPrOH in DCM (6 x 8 L). The
organic layers were concentrated to minimum volume and filtered to
provide white solid which was removed by filtration. The filtrate was
concentrated to provide (S)-2-azido-2-(3-methoxy-1,1-dioxidothietan-
3-yl)acetic acid (2) (261 g, 85%) as an off-white solid.
(S)-6-(4-((2-(hydroxyamino)-1-(3-methoxy-1,1-dioxidothietan-3-yl)-2-
oxoethyl)carbamoyl)phenyl)hexa-3,5-diyn-1-yl dihydrogen phosphate
(55). Dibenzyl (6-(4-((1-(3-methoxy-1,1-dioxidothietan-3-yl)-2-oxo-2-
(((tetrahydro-2H-pyran-2-yl)oxy)amino)ethyl)carbamoyl)phenyl)hexa-
3,5-diyn-1-yl) phosphate 54 (383.3 g, 0.5 mol, 1.00 eq.) in DCM (4 L)
was cooled to 0 °C, and (N,O-bis(trimethylsilyl)acetamide) (510.3 g,
2.5 mol, 5.00 eq.) was added dropwise followed by TMSBr (307 g,
2.0 mol, 4.00 eq). The mixture was stirred at 0 °C for 40 min,
quenched by the addition of Et3N (703 mL, 5.06 mol) and then
stirred for 5 min. The mixture was evaporated, and the residue was
taken up in MTBE (4.0 L), stirred for 5 min and filtered. The cake
was rinsed with MTBE (2 L). The combined MTBE layers were
diluted with MeOH (1 L) followed by DCM (1 L). The mixture was
concentrated under reduced pressure, and the residue was
partitioned between water (4 L) and EtOAc (4 L). The aqueous layer
(2S)-2-amino-2-(3-methoxy-1,1-dioxidothietan-3-yl)-N-((tetrahydro-
2H-pyran-2-yl)oxy)acetamide (50). (S)-2-azido-2-(3-methoxy-1,1-
12
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
was then acidified with 6M HCL (60 ml) and extracted with 2-methyl
THF (6 L). The organic layer was washed with H2O (2 X 40 mL), 6N
HCL (2 X 40 mL), dried over MgSO4 and evaporated to provide 6-(4-
(((1S)-1-(3-methoxy-1,1-dioxidothietan-3-yl)-2-oxo-2-(((tetrahydro-
2H-pyran-2-yl)oxy)amino)ethyl)carbamoyl)phenyl)hexa-3,5-diyn-1-yl
dihydrogen phosphate as a tan solid (191.2 g, which was used
directly for the next step without further purification.
LpxC Enzyme IC50 Assay. The LpxC enzyme inhibition assay was
adapted from the method detailed by Hale et al. and uses
recombinant P. aeruginosa LpxC purified from E. coli.[24] Test
compounds previously dissolved in DMSO are diluted to a 10X stock
concentration in water such that the final DMSO concentration in the
assay is less than 1%. 3 µL of 10X test compound is then aliquoted
in 3-fold dilutions across the row of a 96-well plate. 24 µL of a
buffered LpxC enzyme-containing solution (fresh enzyme is obtained
from a –80 °C previously aliquoted stock) is then added to the test
compound dilutions. The enzyme and test compound are allowed to
incubate for 10 minutes at room temperature. To initiate the reaction,
The product from above was dissolved in MeOH (1.9 L) and 4M
HCl (180 mL) was added. The mixture was stirred for 15 min and
evaporated to give crude material which was dissolved in H2O (12 L).
The mixture was washed with EtOAc (12 L) / toluene (2.5 L). The
aqueous layer was then extracted with two portions of methyl THF
(20 L + 12 L). The combined organic layers were dried (MgSO4),
stirring for 1 h. The mixture was filtered then evaporated to give a
tan solid (123 g, 96.5 % pure). The solid was dissolved in H2O (11 L)
and washed with 16% iPrOH in CH3CL (4 x 5 L). The aqueous layer
was then extracted with EtOAc (10 x 6 L) and the combined organic
layers were evaporated to afford (S)-6-(4-((2-(hydroxyamino)-1-(3-
methoxy-1,1-dioxidothietan-3-yl)-2-oxoethyl)carbamoyl)phenyl)hexa-
3,5-diyn-1-yl dihydrogen phosphate 55 as an off-white solid (92 g,
50 % yield). MS: m/z for C19H21N2O10PS 500.4, found [M+H]+ 501.3.
3
µL
of
400
µM
UDP-3-O-(R-3-hydroxydecanoyl)-N-
acetylglucosamine (CAS# 953426-26-1; obtained from Beijing
Honghui Meditech Co., Ltd) is added. The final reaction conditions of
the assay are 10 mM sodium phosphate, pH 7.5, 0.005% Triton X-
100,
1 mM TCEP, 40 µM UDP-3-O-(R-3-hydroxydecanoyl)-N-
acetylglucosamine, and 10 µM Zn(acetate)2. The final expected
concentration of catalytically active LpxC was estimated to be
approximately 0.1 nM. The reaction was allowed to progress to
approximately 20% completion in the absence of test compound
over 8 hours at which point 3 µL of 1 M HCl is added to the reaction
to terminate enzyme activity.
The progress of the reaction at each concentration of test compound
was measured using LC/MS. For LC/MS analysis, 20 µL of
quenched reaction solution was injected onto Agilent 1290 UPLC
system using a Kinetex C18 100A, 2.1x50 mm, 2.6 µm column
(Phenomenex). Reaction product and substrate were eluted into the
mass spectrometer (Agilent 6120) with a gradient starting at 16%
Acetonitrile, 84% 20 mM Ammonium Acetate, pH 6.7 to 20%
Acetonitrile, 80% 20 mM Ammonium Acetate, pH 6.7 over 1.75
minutes. Product was quantified by monitoring abundance of m/z
734.3 and the substrate was quantified by monitoring abundance of
m/z 776.3 both in negative ion mode. While the substrate peak
remains stable over time, the product of the LpxC reaction is known
to slowly rearrange over time resulting in the formation of two peaks
in the mass spectrum. Both product peaks, in addition to the
substrate peak, are integrated and used to compute the percent
reaction progress at each test compound concentration. The percent
reaction progress at zero test compound is used to normalize the
data at each test compound concentration. A 3-parameter fit is used
to fit the relationship between the normalized reaction progress and
the test compound concentration, yielding a fitted IC50 value. The
lower limit of detection for the IC50 value in this assay is less than 0.3
nM.
Measurement of Antibacterial Activity. MICs were determined by
broth microdilution according to guidelines established by the
Clinical and Laboratory Standards Institute.(22) Compounds were
serially diluted in two-fold increments in 100% DMSO, then diluted
10-fold in water. Inocula were prepared from cells streaked onto
Mueller Hinton agar (MHA) and grown at 35 °C overnight. 10 uL of
antibiotic solution was mixed with 90 uL of inoculum in cation-
adjusted Mueller Hinton broth in 96-well assay plates, with a final
inoculum of approximately 5 x 104 cells/well. When included, PMBN
was diluted in broth prior to addition of cells, to a final assay
concentration of 8 µg/mL. After incubation of assay plates at 35 °C
for 18-20 hours, the lowest concentration of antibiotic that prevented
visible growth was recorded as the MIC.
P. aeruginosa LpxC protein purification. A IPTG-inducible expression
plasmid for WT P. aeruginosa LpxC with an N-terminal 6-His tag was
transformed into RosettaTM(DE3)pLysS E. coli cells (Novagen) for
protein expression. Cells containing the expression plasmid were
grown in 50 µg/mL kanamycin and 17 µg/mL chloramphenicol at
30 °C. The cultures were induced with 0.5 mM IPTG and 10 µM
Zn(acetate)2 at an OD600 of 0.5 absorbance units. After growth at
30 °C for an additional 2.5 h, the cells were harvested by
centrifugation at 10,500 rpm in a JA10.5 rotor for 15 min. Afterwards
the supernatant was removed, the cells resuspended in buffer A [25
mM Tris-HCl, 100 mM NaCl, 100 µM Zn(acetate)2, 5 mM imidazole,
and pH 8.0], and the resuspension stored at -80 °C.
Anesthetized Rat Methodology: All animals studies were performed
with the approval of the Institutional Animal Care and Use
Committee at Plato Biopharma, Inc. and in compliance with the
Guide for the Care and Use of Laboratory Animals. Four male
Sprague Dawley rats (approximately 7 weeks of age) per study were
anesthetized using inhaled isoflurane (ISO) on spontaneous nose-
cone rebreathing with 2.0% ISO in 100% O2 during surgery, and
1.5% ISO for maintenance. Body temperature was maintained with
rectal probe-controlled heating pads. Cannulae were surgically
implanted in the femoral artery, vein, and jugular vein for blood
pressure measurement, IV infusion of test article and blood sampling,
respectively. Test articles were delivered by intravenous infusion at
a rate of 10 mL/kg/h in either HPCD/Tris (20% HPCD, 250 mM Tris)
or A5D2.5 (2.5% Dextrose in 20 mM Acetate buffer). In dose
escalation (DE) studies, vehicle and 3 dose levels were infused for
30 minutes per dose level. In maximum tolerated dose (MTD)
studies, vehicle was infused for 30 minutes followed by one dose
level of test article for 60 minutes. Blood samples were drawn to
assess plasma levels with the following time points: DE studies:
@30, 60 90, and 120 minutes of infusion, and 2, 6, 12 and 30
minutes post-infusion cessation; MTD studies: @30, 45, 60, 75, 90
minutes of infusion, and 2, 6, 12, and 30 minutes. Steady-state and
time-dependent hemodynamic parameters (mean arterial pressure
The resuspended cells were lysed using an microfluidzer
(MicrofluidicsTM LM10; 16,000 psi, 4 passages). The lysate was
centrifuged at 18,000 rpm in a JA25.5 rotor for 45 min and the
resulting supernatant was filtered through a 0.45 uM syringe-
operated filter. The filtered supernatant was applied via gravity flow
to a 3 mL HisPurTM (Thermo Scientific). The column was washed
with approximately 80 mL of buffer A. The bound enzyme was eluted
with 20 mL of buffer B 25 mM Tris-HCl, 100 mM NaCl, 100 µM
Zn(acetate)2, 150 mM imidazole, and pH 8.0). The elution volume
was dialyzed overnight at 4 C into 4L of storage buffer (25 mM
imidazole, 100 mM KCl, 2 mM DTT, 20% Glycerol, and pH 7.0). The
enzyme solution was then filtered through a 0.2 µm syringe-operated
filter, aliquoted, and stored at –80 °C to be later used in the LpxC
IC50 assay and for x-ray crystallography studies. The presence of the
desired enzyme and final purity was estimated by SDS-PAGE to be
>95%. The concentration of total enzyme was determined by A280
using a calculated extinction coefficient of 15,390 M-1cm-1 or by
Bradford assay.
13
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
(MAP), systolic and diastolic blood pressure (SBP, DBP), and heart
rate (HR) were collected and expressed as minute averages over
the study duration. Parameters were expressed as actual values
(both individual and mean) and as percent change from vehicle
where the final 10 minutes of the vehicle infusion data were
averaged for the baseline value. Test article plasma levels were
expressed as total concentrations and free concentrations using the
protein binding data for the test article per time point. Stopping
criteria for test article infusion of MAP decrease of 30 mmHg or 30%
decrease in HR for 2 continuous minutes were established. All
animals were euthanized under anesthesia at the end of the study
after inspection of the infusion catheter tip site via laparotomy.
[20] R. Kasar, M. S. Linsell, J. B. Aggen, Q. Lu, D. Wang, T. Church, H. E.
Moser, P. A. Patten (Achaogen), WO2012154204, 2012.
[21] M. S. Linsell, J. B. Aggen, P. Dozzo, D. J. Hildebrandt, F. Cohen, R. A.
Kasar, T. R. Kane, M. J. Gliedt, G. A. McEnroe (Achaogen),
WO2014165075, 2014.
[22] F. Cohen, A. W. Konradi, T. A. Choi, T. Machajewski, T. R. Kane, D. J.
Hildebrandt (Achaogen), WO2017223349, 2017.
[23] CLSI. Methods for Dilution Antimicrobial Susceptibility Tests for
Bacteria That Grow Aerobically; Approved Standard—Tenth Edition.
CLSI document M07-A10. Wayne, PA: Clinical Laboratory Standards
Institute; 2015.
[24] M. R. Hale, P. Hill, S. Lahiri, M. D. Miller, P. Ross, R. Alm, N. Gao, A.
Kutschke, M. Johnstone, B. Prince, J. Thresher, W. Yang, Bioorg. Med.
Chem. Lett. 2013, 23, 2362-2367.
Acknowledgements
This work was supported by The Defense Threat Reduction Agency,
Department of Defense, under Contract No. HDTRA107C0079, the
Military Medical Research and Development Program, Defense
Health Program, Department of Defense, under Grant No.
W81XWH122000040 and by the National Institute of Allergy and
Infectious Diseases, National Institutes of Health, Department of
Health
and
Human
Services,
under
Contract
No.
HHSN272201500009C. We thank Phil Patten, J. Kevin Judice, and
Marc Navre for leadership and guidance.
Keywords: Amines, Antibiotics, Inhibitors, LpxC, Prodrugs,
Pseudomonas, Zwitterions
References:
[1]
[2]
J. Burnham, M. Olsen, M. Kollef, Infect. Control Hosp. Epidemiol. 2019,
40, 112-113.
Centers for Disease Control and Prevention. (2013). Antibiotic
resistance threats in the United States, 2013. Retrieved Feb. 27, 2019,
from CDC website: www.cdc.gov/drugresistance/threat-report-2013.
L. Chen, R. Todd, J. Kiehlbauch, M. Walters, A. Kallen, MMWR Morb.
Mortal Wkly. Rep. 2017, 66, 33.
[3]
[4]
[5]
R. Tommasi, R. Iyer, A. A. Miller, ACS Infect. Dis. 2018, 4, 686-695..
D. J., Payne, M. N., Gwynn, D. J. Holmes, D. L. Pompliano, Nat. Rev.
Drug Disc. 2007, 6, 29-40.
[6]
C. R. Raetz, C. M. Reynolds, M. T. Trent, R. E. Bishop, Ann. Rev.
Biochem. 2007, 76, 295-329.
[7]
[8]
[9]
A. L. Erwin, Cold Spring Harb. Perspect. Med. 2016, 6, a025304.
F. Liu, S. Ma, Mini-Rev. Med. Chem. 2018, 18, 310-323.
M. Vaara, Microbiol. Rev. 1992, 56, 395-411.
[10] O. Lomovskaya, A. Lee, K. Hoshino, H. Ishida, A. Mistry, M. S. Warren,
E. Boyer, S. Chamberland, V. J. Lee, Antimicrob. Agents Chemother.
1999, 43, 1340-1346.
[11] Davies, B. I., Antimicrob. Agents Chemother. 1990, 25, 471-472.
[12] L. Xingrong, M. Wright, C. E. Hop, J. Med. Chem., 2014, 57(20), 8238-
8248.
[13] M. F. Richter,B. S. Drown, A. P. Riley, A. Garcia, T. Shirai, R. L. Svec,
P. J. Hergenrother, Nature, 2017, 545, 299-304.
[14] M. F. Richter, P. J. Hergenrother, Annals of the New York Academy of
Sciences, 2019, 1435, 18-38.
[15] G. F. Smith, Progress Med. Chem. 2011, 50, 1-47.
[16] K. Krause, C. M. Haglund, C. Hebner, A. W. Serio, G. Lee, V. Nieto, F.
Cohen, T. R. Kane, T. Machajewski, D. J. Hildebrandt, A. Burek, L. D.
Andrews, E. Armstrong, L. Swem, A. Jubb, R. T. Cirz. Manuscript
Accepted.
[17] V. Stella, R. Borchardt, M. Hageman, R. Oliyai, H. Maag, J. Tilley,
Prodrugs: Challenges and Rewards Part 1 and 2. Springer, New York,
NY, 2007.
[18] H. Moser, Q. Lu, P. A. Patten, D. Wang, R. Kasar, S. Kaldor, B. D.
Patterson (Achaogen), WO2008154642, 2008.
[19] B. D. Patterson, Q. Lu, J. B. Aggen, P. Dozzo, R. A. Kasar, M. S.
Linsell, T. R. Kane, M. J. Gliedt, D. J. Hildebrandt, G. A. McEnroe, F.
Cohen (Achaogen), WO2013170030, 2013.
14
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900287
ChemMedChem
FULL PAPER
Entry for the Table of Contents
It’s not the hydroxamate!: Antibacterial LpxC inhibitors have optimized for enzymatic potency, anti-Pseudomonal activity,
pharmacokinetics and cardiovascular safety. We demonstrate that the much-maligned hydroxamic acid is not responsible for the CV
toxicity observed with LpxC inhibitors. Development of a phosphate prodrug enabled high solubility and rapid conversion to active
drug upon i.v. administration.
15
This article is protected by copyright. All rights reserved.