Antibacterial and Solubility Optimization of Thiomuracin A

Article abstract of DOI:10.1021/acs.jmedchem.6b00726

Synthetic studies of the antimicrobial secondary metabolite thiomuracin A (1) provided access to analogues in the Northern region (C2-C10). Selective hydrolysis of the C10 amide of lead compound 2 and subsequent derivatization led to novel carbon- and nitrogen-linked analogues (e.g., 3) which improved antibacterial potency across a panel of Gram-positive organisms. In addition, congeners with improved physicochemical properties were identified which proved efficacious in murine sepsis and hamster C. difficile models of disease. Optimal efficacy in the hamster model of C. difficile was achieved with compounds that possessed both potent antibacterial activity and high aqueous solubility.

Full text of DOI:10.1021/acs.jmedchem.6b00726

Article
pubs.acs.org/jmc
Antibacterial and Solubility Optimization of Thiomuracin A
Matthew J. LaMarche,*,† Jennifer A. Leeds,⊥ Jason Brewer,† Karl Dean,† Jian Ding,† Joanne Dzink-Fox,⊥
Gabe Gamber,† Akash Jain,∥ Ryan Kerrigan,† Philipp Krastel,# Kwangho Lee,† Franco Lombardo,§
David McKenney,⊥ Georg Neckermann,⊥ Colin Osborne,⊥ Deborah Palestrant,‡ Michael A. Patane,†
Elin M. Rann,† Zachary Robinson,† Esther Schmitt,# Travis Stams,‡ Stacey Tiamfook,⊥ Donghui Yu,⊥
and Lewis Whitehead†
‡
§
†Global Discovery Chemistry, Protein Structure Group, Metabolism and Pharmacokinetics, Novartis Institutes for Biomedical
∥
Research, Chemical and Pharmaceutical Profiling, Novartis Pharmaceuticals, Cambridge, Massachusetts 02139, United States
⊥Infectious Disease Area, Novartis Institutes for Biomedical Research, Emeryville, California 94608, United States
#Natural Products Unit, Novartis Institutes for Biomedical Research, Basel, Switzerland
S
* Supporting Information
ABSTRACT: Synthetic studies of the antimicrobial secondary metabolite thiomuracin A (1) provided access to analogues in the
Northern region (C2−C10). Selective hydrolysis of the C10 amide of lead compound 2 and subsequent derivatization led to
novel carbon- and nitrogen-linked analogues (e.g., 3) which improved antibacterial potency across a panel of Gram-positive
organisms. In addition, congeners with improved physicochemical properties were identified which proved efficacious in murine
sepsis and hamster C. difficile models of disease. Optimal efficacy in the hamster model of C. difficile was achieved with
compounds that possessed both potent antibacterial activity and high aqueous solubility.
these SAR findings to other classes of thiopeptides, such as the
baringolins, have recently been reported by others.6 In addition,
a de novo biosynthesis of the thiomuracin core has also been
reported,7 as well as manipulation of thiopeptide biosynthetic
gene clusters toward structural variation.8 Indeed both semi-
synthetic and biosynthetic efforts have contributed to the
evolving SAR of the thiopeptide class.9
Previously we reported initial synthetic studies on thiomuracin
A, which chemically stabilized the C84 epoxide region via
pyrrolidine cyclization, removed the C2−C10 side chain, and im-
proved the isolation and material supply for continued medicinal
chemistry optimization.10 These studies culminated in the identi-
fication of lead compound pyrrolidine 2, which retained potent
antibacterial activity against five target Gram+ organisms with
MIC values <1 μg/mL and also displayed potent in vivo activity
in the mouse sepsis model of infection. Due to the structural
INTRODUCTION
■
Novel classes of antibiotics with new modes of action and no
cross-resistance with clinically used antibiotics remains a pressing
need in infectious disease care because of the rising incidence of
resistance against marketed antibiotics. In 2009 we reported the
isolation, structural elucidation, antimicrobial activity, and
biosynthetic gene cluster of a new class of thiazoyl actinomycetes
metabolites, the thiomuracins (1, Figure 1).1 These architectur-
ally complex, secondary metabolites were fermented from the
rare soil actinomycetes species Nonomurae and possess highly
modified, sulfur-containing macrocyclic peptide structures. The
thiomuracins are structurally related to the GE2270 A (4) class of
antibiotics2 (Figure 1), another actinomycetes fermentation pro-
duct upon which we have based extensive drug discovery efforts.3
Both the thiomuracins and 4 derive their antibiotic activity via
binding to elongation factor Tu, thereby disrupting prokary-
otic protein synthesis. Medicinal chemistry optimization of 4
resulted in the identification of LFF571 (5),4 an investigational
agent for C. difficile infection in humans.5 General applications of
Received: May 17, 2016
Published: June 29, 2016
© 2016 American Chemical Society
6920
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
Figure 1. Thiomuracin A (1), 4, and 5.
simplification, chemical stability, optimized material supply, and
retention of potency, amide 2 was selected for continued medici-
nal chemistry optimization. Considering our recent SAR findings
on aminothiazolyl analogues of 4, the potential for Northern
region analogues of thiomuracin A was of considerable interest.3,4
Thus, treatment of pyrrolidine 2 with HCl, H2O, and THF at
elevated temperatures (RT → 100 °C) revealed that selective
hydrolysis indeed occurred at 80−100 °C to provide 9 directly
and in acceptable yield (80%). Incidentally, macrocyclic hydro-
lysis of the amides occurred at 110 °C and thus could be avoided.
The carboxylic acid (9, Scheme 2) was then used as a strategic
intermediate for the synthesis of Northern region analogues via
amide couplings (e.g., 9 → 10−12) and via a Curtius rearrange-
ment sequence (e.g., 9 → 16, 3). Accordingly, a variety of
structural motifs were examined, partially guided by our previous
studies of aminothiazolyl-based analogues of 4.3,4 For C-linked
analogues, acyclic acids (e.g., 10), cyclohexylcarboxylic acids
(e.g., 11), and diacid structures were prepared (e.g., 12). Other
polar functional groups (e.g., amines, alcohols) were not con-
sidered, as previous studies demonstrated these functional
groups had little impact on the aqueous solubility of this class.3,4
For N-linked analogues, carboxylate 9 was activated via ethyl
chloroformate, and azide displacement furnished intermediate 13.
Interestingly, carbonate protection of the tyrosine residue
occurred and proved chemically stable to isolation and further
chemistry (vide supra). Curtius rearrangement using methyl
trans-4-hydroxycyclohexane-1-carboxylate afforded urethane 14.
Exhaustive hydrolysis of the ester and carbonate moieties
(LiOH) afforded acid 16. To selectively alkylate the Northern
urethane, the remaining hydroxyl functional groups of 14 were
protected as acetates (e.g., 15). Alkylation of the urethane and
global deprotection then furnished diacid 3, which structurally
resembled 5.
Antibacterial Activity. Thiomuracin A (1), lead amide (2),
and novel analogues 9−12, 16, and 3 were then evaluated in MIC
(minimum inhibitory concentration) assays for G+ bacterial
growth inhibition (Table 1). Five species comprised our anti-
bacterial screen and included Enterococcus faecalis, Enterococcus
faecium, Staphylococcus aureus, Streptococcus pyogenes, and the
anaerobic intestinal pathogen Clostridium difficile. Thiomuracin
A (1) and amide 2 were exquisitely potent in all five organisms
tested with MICs ranging from 0.03 to 1 μg/mL. Carboxylic acid
(9) retained moderate antibacterial activity in all of the target
organisms but was less potent as compared to thiomuracin A (1).
Unfortunately, the aqueous solubility of9 was poor (0.001 mg/mL
RESULTS
■
Synthesis. During the course of identifying lead amide 2,
an acid-mediated hydrolysis of the C10 didehydroalanine side
chain and acetate protecting group strategy emerged (Scheme 1,
1 → 2). Three hydroxyls of chlorohydrin 6 were protected as
acetate esters using standard conditions at room temperature
(Ac2O, DMAP). Subsequent DBU cyclization formed the N81
pyrrolidine, and the esters were then removed to provide 2.
Acetylation at elevated temperature (60 °C), however, resulted in
selective imide formation at C10 (7). The C10 primary amide is
sterically more accessible than the macrocyclic secondary amides
and perhaps more acidic than the C34 primary amide due to
the northern thiazolyl ring (via pseudo-benzylic stabilization).
Attempts at further derivatization of 7 at the C10 imide were
initially unfruitful due to acid and base sensitivity of the imide,
chlorohydrin, hydroxyphenylalanine, and acetate protecting
groups. For example, NaOH treatment of 7 led to 2, but
retroaldol decomposition of the C44 hydroxyphenylalanine was
also observed (not shown). In addition, HCl treatment of 7 led to
unproductive imide hydrolysis, providing the C10 amide.
After significant experimentation, selective allylation of the
imide was possible (NaH, DMF, allyl bromide), although the
yields were variable. This provided the tertiary imide with con-
comitant N81 to C84 pyrrolidine cyclization (8). The tertiary
imide was then hydrolyzed to the C10 carboxylic acid (9).
Although we had achieved an initial synthesis of the carboxylic
acid 9, a key intermediate en route to Northern carbonyl-based
analogues, the yields (<10%) were untenable. We hypothesized
that selective C10, acid-mediated hydrolysis would be possible
because the C10 amide in 2 was prone to selective imide for-
mation and was the most sterically accessible (vide supra)10 and
pseudo-benzylic. We had previously demonstrated that thiomuracin
A was quite stable to strong acid conditions (HCl, MeOH, 50 °C),
aside from the C2−C10 hydrolysis and C84 epoxide opening.10
6921
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
Scheme 1. Synthesis of Intermediate 9
in pH 7.4 buffer). Straight-chained and cycloalkanes terminating
in carboxylic acids proved exquisitely potent and soluble in the
aminothiazole series of 4.3,4 In the current C-linked studies of the
thiomuracin scaffold, these functional groups also retained potent
biological activity against the target organisms (e.g., 10: MICs
0.06−0.5 μg/mL). Likewise, a cyclohexane-based carboxylic acid
functional group was evaluated (11) and proved equipotent as
the natural product thiomuracin A (1) and the acyclic analogue
10. Importantly, the aqueous solubility was also improved
(11: 0.75 mg/mL in pH 7.4 buffer). C-linked dicarboxylic acids
were also evaluated (e.g., 12) but significantly lowered anti-
bacterial activity across organisms (MICs = 0.125−1 μg/mL).
6922
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
Scheme 2. Synthesis of C-Linked and N-Linked Northern Analogues
In the N-linked series of analogues, two congeners (16, 3) were
evaluated and were more potent than the C-linked diacid (12),
and equipotent to the C-linked monoacids (10, 11) across the
panel of organisms. Overall, Clostridium difficile was very sen-
sitive to all the analogues. N-linked dicarboxylic acid 3 was the
only analogue which maintained potent antibacterial activity
(MICs 0.03−0.25 μg/mL) while significantly improving aqueous
solubility (9.7 mg/mL in pH 7.4 buffer).
Crystallography. Several compounds from these lead opti-
mization studies were selected for cocrystallographic studies with
6923
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
a
Table 1. Antibacterial and Solubility Data for 1 and Analogues
a
BLD: below limit of detection.
Figure 2. Cocrystal structure of E. coli EF-Tu and 11.
EF-Tu derived from E. coli, including cyclohexanecarboxylic
acid 11.11 The cocrystal structure of 11 (Figure 2, 2.0 Å, PDB
code 5JBQ) confirmed the location of binding within domain 2
of EF-Tu and also displayed the same macrocyclic conformation
as previously reported for 4 and amide 2 in their respective
bound conformations.10,12 Additionally, similar interactions were
observed between 11 and EF-Tu as compared to 4, including
H-bonding with EF-Tu at several locations (Asn273, Phe261).
Intramolecular H-bonding across the macrocycle from the
primary amide N35 (donor) to the tyrosine carbonyl O87
(acceptor) was observed (3.0 Å distance). Likewise, amide
carbonyl O91 (acceptor) interacts with tyrosine OH68 (donor)
in an intramolecular H-bond (3.4 Å distance). Presumably, both
through-space interactions allow for conformational stabilization
of the macrocyle. In the Northern region, the amide linker
orients the cyclohexane carboxylic acid side chain toward solvent
and Arg223. We speculate that both the cyclohexane functional
motif and the appropriate linker distance orient the carboxylate
favorably for optimal interaction with Arg223.
Pharmacokinetics and Efficacy Studies. The cyclohexane
carboxylic acid 11 was investigated in rat pharmacokinetic and
efficacy experiments. Accordingly, acid 11 was dosed intravenously
to male Sprague−Dawley rats (1 mg/kg).13 Acceptable exposure
(AUC = 13187 nM·h), low clearance (0.98 mL/min/kg), low
volume of distribution (Vss = 0.032 L/kg), and a moderate half-
life (T1/2 = 1.8 h) were observed. Furthermore, in the mouse
sepsis model of infection,14 acid 11 displayed dose-dependent
efficacy (S. aureus ED50 = 3.42 mg/kg, 95% confidence interval =
2.48−4.90 mg/kg; E. faecalis ED50 = 0.71 mg/kg, 95% confidence
interval = 0.7−2.0 mg/kg). This compares favorably with the
6924
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
previously described efficacy of amide 2 (S. aureus ED50
=
While there was 0% survival with C-linked monoacid 11,
N-linked diacarboxylic acid 3 provided improved results.
At 5 mg/kg, 25% survival was observed, while at 10 mg/kg,
compound 3 appeared superior (63% survival) to vancomycin
(38% survival) and approximately equipotent to 5 (75% survival,
5 mg/kg), albeit at 2-fold higher dose.
8.0 mg/kg, 95% confidence interval = 4.8−53.1 mg/kg; E. faecalis
ED50 = 1.5 mg/kg, 95% confidence interval = 0.31−5.0 mg/kg),
presumably due to the improvement in MIC against these
organisms.8 These experiments were controlled with both
daptomycin (S. aureus ED50 = 0.21 mg/kg, 95% confidence
interval = 0.07−0.36 mg/kg; E. faecalis ED50 = 2.60 mg/kg,
confidence interval not calculatable) and vancomycin (S. aureus
ED50 = 1.1 mg/kg, 95% confidence interval = 0.77−1.47 mg/kg;
E. faecalis ED50 = 7.74 mg/kg, 95% confidence interval =
6.7−20 mg/kg). Thus, acid 11 was commensurate with vanco-
mycin in the S. aureus model of infection and less effective than
daptomycin. But in the E. faecalis sepsis model, acid 11 proved
superior to both vancomycin and daptomycin.
DISCUSSION AND CONCLUSIONS
■
A selective, acid-mediated hydrolysis of the C10 amide of the
thiomuracin scaffold has been achieved which enabled the
synthesis and biological evaluation of Northern region analogues.
Cyclic and acyclic C- and N-linked carboxylic acid analogues
were evaluated for antibacterial activity and aqueous solubility.
Congeners with improved physicochemical properties were
identified and evaluated in murine sepsis and hamster C. difficile
models of disease. C-linked cyclohexanecarboxylic acid 11 was
evaluated in pharmacokinetic experiments and proved efficacious
in the mouse sepsis model of infection. Observed EC50s were
either superior or comparable to the standards of care (e.g.,
vancomycin, daptomycin) and dependent on the model
organism. Multiple analogues were evaluated in the Golden
Syrian hamster model of C. difficile, and optimal efficacy was
achieved with compounds that possessed both potent anti-
bacterial activity and high aqueous solubility. These studies
culminated in the identification of N-linked dicarboxylic acid 3,
which proved significantly more soluble than thiomuracin A and
more potent than vancomycin in the hamster model of C. difficile
infection.
Several analogues were also evaluated in the C. difficile Golden
Syrian hamster model of infection (Figures 3 and 4).15 After oral
Figure 3. Evaluation of compounds 2, 9, and 10 in the C. difficile
hamster model. 5 and vancomycin data are the compiled average of
seven independent experiments;4b data for compounds 2, 9, and 10 are
from discrete experiments.
The thiomuracins and associated analogues contained herein
represent a novel class of antibiotics with a new mode of action
and no cross-resistance with clinically used antibiotics,1a pressing
need in infectious disease care because of rising incidence of
resistance against marketed antibiotics. These interesting and
architecturally complex structures have proven amenable to
semisynthetic derivatization, gene cluster manipulation, and de
novo biosynthesis. Optimization of physicochemical properties
and antibacterial activity and translation of SAR from one series
to another within the thiopeptide class have been successfully
demonstrated. Novel analogues with improved properties and
antibacterial activity have enabled efficacy studies in multiple
animal models of infection, and a leading thiopeptide 5 has
demonstrated clinical utility. Furthermore, this evolving knowl-
edge and now considerable body of literature around the
thiopeptide class perhaps serves as a roadmap for other novel
classes of antibiotics yet to be discovered and characterized.
Figure 4. Evaluation of compounds 11 and 3 in the C. difficile hamster
model. 5 and vancomycin data are the compiled average of seven
independent experiments;4b data for compounds 3 and 11 are from
discrete experiments.
EXPERIMENTAL SECTION
■
gavage of C. difficile inoculum, hamsters were treated with 2, 9, or
10 (25 mg/kg), or controls which included saline, vancomycin
(20 mg/kg), or 5 (5 mg/kg) for 5 days. The animals were then
observed for an additional 16 days. While no animals survived
after saline treatment, 38% survival was observed with
vancomycin (20 mg/kg) and 71% survived with 5 (5 mg/kg).
Considering the first three analogues, animal survival was the
highest with acid 9 (75%) versus alkyl acid 10 (13%) and amide 2
(0%). This result was surprising considering the antibacterial
activity of 2 (MIC, C. difficile = 0.008 μg/mL). We speculate that
low solubility of 2 precluded growth inhibition in the gut.
Particularly noteworthy was the high dose of 9 (25 mg/kg) which
provided similar survival as 5 at a lower dose (5 mg/kg).
Compound Synthesis and Characterization. Synthetic proce-
dures and compound characterization data are found in the Supporting
Information. Compound purity was assessed by HPLC to confirm >95%
purity.
MIC Assays. MIC assays were conducted according to broth
microdilution methods described by the Clinical and Laboratory
Standards Institute with the exception that broth microdilution methods
were utilized for all species tested: CLSI Methods for Dilution
Antimicrobial Susceptibility Tests for Bacteria That Grow Aerobically;
Approved StandardEighth Edition. CLSI document M07-A8; Clinical
and Laboratory Standards Institute: Wayne, PA, 2009. Methods for
antimicrobial susceptibility testing of anaerobic bacteria: Methods for
Antimicrobial Susceptibility Testing of Anaerobic Bacteria; Approved
StandardEighth Edition (M11-A8); Clinical and Laboratory Stand-
ards Institute: Wayne, PA, 2012. Bacterial strains included E. faecalis
(ATCC 29212), E. faecium (Prof. Chopra, Univ. Leeds UK, strain
7130724), S. aureus (MRSA from Prof. Willinger, isolated from a
Two additional analogues were evaluated in the C. difficile
hamster model of infection (Figure 4): C-linked cyclo-
hexanecarboxylic acid 11 and N-linked dicarboxylic acid 3.
6925
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
pharyngeal smear, AKH Vienna A4.018), S. pyogenes (ATCC BAA-595),
and C. difficile (ATCC 43255).
At 24 h after clindamycin pretreatment, hamsters were infected by oral
gavage with approximately 106 CFU of the C. difficile culture. All culture
work was completed within 30 min, and all cultures were back in
anaerobic conditions by this time. Briefly, strains were resuspended from
overnight plates (in reinforced clostridial medium + 1% oxyrase
(Oxyrase, Inc. Mansfield, OH) and diluted to 1 × 107 cfu/mL (OD600
nm = 1.1−1.3, dilute 1:100). Hamsters were immediately infected with
0.75 mL of inoculum (∼7.5 × 106 CFU/hamster final). An aliquot of
inoculum was diluted 1:1000 and spiral-plated (100 μL, slow deposi-
tion) on TSAB or RCMA (×2) and incubated at 37 °C anaerobically for
determination of infection titer. Treatment: hamsters were administered
the test compound (eight animals per dose level) starting 24 h after
infection. Single antibiotic doses were administered orally and
formulated as described below. Antibiotics were administered 1 time
per day and continued for up 5 days. The control group was admin-
istered the solution vehicle alone. Animals were observed two times a
day for the duration of the experiment. General observations included
signs for mortality and morbidity, for the presence of diarrhea (“wet
tail”), overall appearance (activity, general response to handling, touch,
ruffled fur), and recorded. Animals judged to be in a moribund state
were euthanized. Criteria used to assign a moribund state are extended
periods (5 days) of weight loss, progression to an emaciated state,
prolonged lethargy (more than 3 days), signs of paralysis, skin erosions
or trauma, hunched posture, and a distended abdomen. Observations
continued, with any deaths or euthanasia recorded for a period up to
21 days postinfection (for relapse). Any animal that died during the
observation period was necropsied and the contents of their cecums
removed, diluted with an equal volume of PBS, and frozen at −80 °C
until processing (efficacy study). All surviving animals were euthanized
by CO2 inhalation and sampled in a similar manner as above.
Solubility Experiments. One to two milligrams of compound was
weighed into 1 mL glass tubes, and a fixed volume of each vehicle was
added to yield approximately 20 mg/mL of compound. After initial
mixing using brief vortexing and sonication (5−10 min), samples were
equilibrated by shaking for 24 h at room temperature. After
equilibration, the vials were visually examined. If a clear solution was
obtained, solubility was reported as >X mg/mL (where X is the starting
concentration in that sample) and the pH of the sample was recorded.
Suspensions or solutions with visible particles were filtered through
0.22 μm PVDF membrane filters, their pH was recorded, and the
dissolved drug concentration was analyzed using an RP-HPLC assay.
Dosing Solution Formulation. Compound 11. Concentration:
5 mg/mL; ingredients (amount for 1 mL (%)): 0.1 N NaOH [100 μL
(10% v/v)], PEG-300 [100 μL (10% v/v)], purified water [770 μL (77%
v/v)], 0.1 N HCl [30 μL (3% v/v)]. Procedure: the compound was
weighed in a vial, PEG-300 was added, and the mixture was stirred/
sonicated for 15 min at 50 °C until clear. NaOH (0.1 N) was added and
stirred for 15 min. Water was added and mixed thoroughly, and the pH
was adjusted to 7.0−7.5 with 0.1 N HCl. The final formulation was a
clear solution with pH 7.3. The solution was physically stable for at least
24 h at room temperature.
Rat Pharmacokinetic Studies. PK studies (N = 2, iv) were
performed with male Sprague−Dawley rats, weighing 220−270 g and
approximately 6−10 weeks old. They were obtained from Harlan
Research Laboratories (South Easton, MA), each bearing dual
implanted jugular vein cannula. The rats were fasted overnight before
use and for 8 h after dosing. Blood samples were taken into K2-EDTA-
coated tubes and then centrifuged to yield plasma samples for analysis by
LC-MSMS. Bioanalysis of rat plasma from in-life experiments were per-
formed by LC-MSMS using a system with the following configuration:
Agilent liquid chromatograph (Santa Clara, CA), LEAP Technologies
CTC-PAL autosampler (Carrboro, NC), and Applied Biosystems API
4000 mass spectrometer (Framingham, MA). LC was performed in gra-
dient mode with reversed phase C18 columns (2.1 mm × 30−50 mm ×
3.5−5 μm particle size). Mobile phase A was 0.1% formic acid in water,
and mobile phase B was 0.1% formic acid in acetonitrile. Gradients were
run from 5% B to 95% B for ∼3.5 min. Plasma samples were protein-
precipitated with acetonitrile containing glyburide as the internal
standard (Sigma-Aldrich, St. Louis, MO).
Mouse Systemic Infection Model. The studies were approved by
the Institutional Animal Care and Use Committee of the Novartis
Institutes for BioMedical Research Inc., Cambridge MA. Animals were
maintained under controlled conditions with free access to food
and water. Female CD1 mice (21−25 g, Charles River Laboratories,
Wilmington, MA) were used for infections with S. aureus (ATCC
49951) and E. faecalis (NB04025, a clinical isolate from the Novartis
bacterial collection that is resistant to erythromycin, tetracycline, and
gentamicin, courtesy of Dr. B. Willinger, Vienna General Hospital,
Austria). MICs of amide 2 were 1 μg/mL and 0.125 μg/mL for animal
strains S. aureus (ATCC 49951) and E. faecalis (NB04025), respectively.
Lethal infections were induced by intraperitoneal injection of a freshly
prepared bacterial suspension of 1 × 108 CFU/mouse in either 50%
sterile rat fecal extract (E. faecalis) or 0.9% NaCl (S. aureus). The injected
bacterial dose corresponded to 10−100 times the minimal lethal dose as
determined from previous lethal dose titration studies. Against E. faecalis
infections the mice were treated once immediately following the
bacterial inoculation, while against S. aureus infections mice were treated
1 and 5 h after inoculation. Compound 2 was formulated in 5% DMA,
30% PEG400, 10% cremophor EL, 5% 0.1 N NaOH, and 50% pH 7.4
buffer. Daptomycin was formulated in saline. Compounds were
administered iv via tail vein bolus injection at several dose levels to
groups of six mice each. Following bacterial inoculation, the mice were
observed for 5 days. In addition, body temperature was monitored by
electronic microtransponders (Bio Medic Data Systems, Inc. Seaford,
Delaware) that were implanted in mice subcutaneously prior to
infection. Mice that became moribund as indicated by a combined score
based on clinical observations and drop in body temperature were
preemptively euthanized. The 50% effective dose (ED50), the dose
providing protection to 50% of mice, and the 95% confidence limits
(95% C.I.) were calculated from the survival data at day 5 by probit
analysis using the program Systat (SPSS Inc.). All animals in the vehicle-
treated control groups developed lethal infections.
Crystallization: Compound 11·EF-Tu Complex. To form the
EF-Tu·compound 11 complex, 10 mg/mL (227 μM) E. coli EF-Tu
protein in a buffer containing 50 mM Tris pH 8 and 50 mM NaCl was
incubated with 1 mM compound 11 for 1 h at 4 °C. The sample was
centrifuged at 20 000g to remove any resulting precipitate. Crystal-
lization was carried out using 300 nL of the protein sample plus 300 nL
of crystallization solution containing 0.1 mM Tris pH 8.2, 18%
PEG3350, and 0.2 M MgSO4, using a sitting drop format and equili-
brated against a reservoir of the crystallization solution. The crystal was
flash frozen with liquid nitrogen after being stabilized in a cryobuffer
containing 0.1 mM Tris pH 8.2, 18% PEG3350, 0.2 M MgSO4, and 20%
ethylene glycol.
X-ray Data Collection and Structure Determination. Initial data
from a single crystal of the EF-Tu·compound 11 complex were collected
on a Rigaku Saturn 92 detector using Cu−Kα radiation (λ = 1.5418 Å)
from a Rigaku FR-E rotating anode generator. The data were measured
from a single crystal maintained at 100 K, and the reflections were
indexed, integrated, and scaled using XDS.11 Crystals were diffracted
to 2.0 Å resolution in the space group P2(1)2(1)2. The structure of the
Golden Syrian Hamster Model of C. difficile Infection.
Hamsters were kept under controlled conditions with 12 h dark−12 h
light cycles, 68−72 °F constant temperature, 50% relative humidity, and
10−15 exchanges of a fresh HEPA filter air per hour. Animals were kept
in 7.5 in. wide by 6 in. deep cages (Alternative Design Manufacturing
Silom Springs, AR) with sterilized Bed O-Cob bedding (corn cob) and
free access to water and standard rodent chow (Harlan). Animal details:
Hamster-Golden Syrian (Mesocricetus auratus), Michigan-206 Golden
Syrian wild-type from Harlan, male, 90−110 g, 5−6 weeks old.
Statement on Animal Welfare: Studies described were approved by the
Institutional Animal Care and Use Committee (IACUC) of the Novartis
Institutes for BioMedical Research Inc., Cambridge, under protocol
number OS 20,061.
Experimental Conditions. Infection: C. difficile ATCC 43255 was
received from American Type Culture Collection and stored at −80 °C in
Brucella broth supplemented with 20% glycerol. On day −1, all animals
received a single subcutaneous injection of clindamycin (10 mg/kg).
6926
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
EF-Tu·GDP·compound 11 complex was determined by molecular
replacement as implemented in PHASER,11 using E. coli EF-Tu protein
as a search model (1D8T). The resulting molecular replacement
solution contained one EF-Tu·GDP·compound 11 protein complex in
the asymmetric unit. Structure determination was achieved through
iterative rounds of positional refinement using BUSTER,11 with model
building using COOT.11 Individual B-factors were refined using an
overall anisotropic B-factor refinement along with bulk solvent
correction. The solvent, GDP, and compound 11 were built into the
density in later rounds of the refinement. Data collection and refinement
statistics are shown in Table 2.
REFERENCES
■
(1) Morris, R. P.; Leeds, J. A.; Naegeli, H. U.; Oberer, L.; Memmert, K.;
Weber, E.; LaMarche, J. M.; Parker, C. N.; Burrer, N.; Esterow, S.; Hein,
A. E.; Schmitt, E. K.; Krastel, P. Ribosomally Synthesized Thiopeptide
Antibiotics Targeting Elongation Factor. J. Am. Chem. Soc. 2009, 131,
5946−5955.
(2) (a) Selva, E.; Beretta, G.; Montanini, N.; Saddler, G. S.; Gastaldo,
L.; Ferrari, P.; Lorenzetti, R.; Landini, P.; Ripamonti, F.; Goldstein, B. P.;
Berti, M.; Montanaro, I.; Denaro, M. Antibiotic GE2270 A: A Novel
Inhibitor of Bacterial Protein Synthesis I. Isolation and Characterization.
J. Antibiot. 1991, 44, 693−701. (b) The structure was later corrected and
verified: Tavecchia, P.; Gentili, P.; Kurz, M.; Sottani, C.; Bonfichi, R.;
Selva, E.; Lociuro, S.; Restelli, E.; Ciabatti, R. Degradation studies of
antibiotic MDL 62,879 (GE2270 A) and Revision of the Structure.
Tetrahedron 1995, 51, 4867−4890 (for acidic decomposition see
above)..
Table 2. | Crystallographic Data and Refinement Statistics
parameters
space group
EF-Tu·compound 11 complex
P21212
unit cell (Å)
a = 80.35, b = 125.35, c = 45.40
(3) (a) LaMarche, M. J.; Leeds, J. A.; Dzink-Fox, J.; Gunderson, K.;
Krastel, P.; Memmert, K.; Patane, M. A.; Rann, E. M.; Schmitt, E.;
Tiamfook, S.; Wang, B. 4-Aminothiazolyl Analogues of GE2270 A:
Antibacterial Lead Finding. J. Med. Chem. 2011, 54, 2517−2521.
(b) LaMarche, M. J.; Leeds, J. A.; Dzink-Fox, J.; Mullin, S.; Patane, M. A.;
Rann, E. M.; Tiamfook, S. 4-Aminothiazolyl analogs of GE2270 A:
Design, Synthesis and Evaluation of Imidazole Analogs. Bioorg. Med.
Chem. Lett. 2011, 21, 3210−3215. (c) LaMarche, M. J.; Leeds, J. A.;
Amaral, K.; Brewer, J. T.; Bushell, S. M.; Dewhurst, J. M.; Dzink-Fox, J.;
Gangl, E.; Goldovitz, J.; Jain, A.; Mullin, S.; Neckermann, G.; Osborne,
C.; Palestrant, D.; Patane, M. A.; Rann, E. M.; Sachdeva, M.; Shao, J.;
Tiamfook, S.; Whitehead, L.; Yu, D. Antibacterial Optimization of 4-
Aminothiazolyl Analogues of the Natural Product GE2270 A:
Identification of the Cycloalkylcarboxylic Acids. J. Med. Chem. 2011,
54, 8099−8109.
(4) (a) LaMarche, M. J.; Leeds, J. A.; Amaral, A.; Brewer, J. T.; Bushell,
S. M.; Deng, G.; Dewhurst, J. M.; Ding, J.; Dzink-Fox, J.; Gamber, G.;
Jain, A.; Lee, K.; Lee, L.; Lister, T.; McKenney, D.; Mullin, S.; Osborne,
C.; Palestrant, D.; Patane, M. A.; Rann, E. M.; Sachdeva, M.; Shao, J.;
Tiamfook, S.; Trzasko, A.; Whitehead, L.; Yifru, A.; Yu, D.; Yan, W.; Zhu,
Q. Discovery of LFF571: An Investigational Agent for Clostridium
difficile Infection. J. Med. Chem. 2012, 55, 2376−2387. (b) Trzasko, A.;
Leeds, J. A.; Praestgaard, J.; LaMarche, M. J.; McKenney, D. Efficacy of
LFF571 in a Hamster Model of Clostridium difficile Infection.
Antimicrob. Agents Chemother. 2012, 56, 4459−4462. (c) Leeds, J. A.;
Sachdeva, M.; Mullin, S.; Dzink-Fox, J.; LaMarche, M. J. Mechanism of
Action and Mechanism of Reduced Susceptibility to the Novel Anti-
Clostridium difficile Compound LFF571. Antimicrob. Agents Chemother.
2012, 56, 4463−4465.
(5) (a) Mullane, K.; Lee, C.; Bressler, A.; Buitrago, M.; Weiss, K.;
Dabovic, K.; Praestgaard, J.; Leeds, J. A.; Blais, J.; Pertel, P. Multicenter,
Ramdomized Clinical Trial To Compare the Safety and Efficacy of
LFF571 and Vancomycin for Clostridium difficile infections. Antimicrob.
Agents Chemother. 2015, 59, 1435−1440. (b) Bhansali, S. G.; Mullane,
K.; Ting, L. S.; Leeds, J. A.; Dabovic, K.; Praestgaard, J.; Pertel, P.
Pharmacokinetics of LFF571 and Vancomycin in Patients with
Moderate C. difficile Infections. Antimicrob. Agents Chemother. 2015,
59, 1441−1445. (c) Ting, L. S.; Praestgaard, J.; Grunenberg, N.; Yang, J.
C.; Leeds, J. A.; Pertel, P. A First in Human, Double Blind, Placebo
Controlled, Single and Multiple Ascending Dose Study to Assess the
Safety and Tolerability of LFF571 in Healthy Volunteers. Antimicrob.
Agents Chemother. 2012, 56, 5946−5951.
resolution range (Å)
total observations
unique reflections
42.69−2.006 (2.012−2.006)
144646
31079
a
completeness (%)
98.5 (66.2)
multiplicity
4.7 (2.6)
a
⟨I/σ(I)⟩
15.7 (4.2)
ab
Rmerge
Rcryst/Rfree
,
0.065 (0.250)
c
0.205/0.253 (0.224/0.235)
protein atoms
3027
124
heterogen atoms
solvent molecules
average B-factor (Å2)
rmsd bond lengths (Å)
rmsd bond angle (deg)
224
27.4
0.01
1.16
a
b
Numbers in parentheses are for the highest resolution shell. Rmerge
=
∑|Ih − ⟨Ih⟩|/∑Ih over all h, where Ih is the intensity of reflection h.
c
Rcryst and Rfree = ∑∥Fo| − |Fc∥/∑|Fo|, where Fo and Fc are observed
and calculated amplitudes, respectively. Rfree was calculated using 5% of
data excluded from the refinement.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the ACS
Publications website at DOI: 10.1021/acs.jmedchem.6b00726.
■
S
Analytical methods, experimental synthetic procedures,
and compound characterization data (PDF)
SMILES and MIC data (CSV)
AUTHOR INFORMATION
Corresponding Author
*Telephone: (617) 871-7729. Fax: 617-871-4081. E-mail:
matthew.lamarche@novartis.com.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The authors thank all of the contributors to the EFT project
team.
■
(6) Just-Baringo, X.; Bruno, P.; Pitart, C.; Vila, J.; Albericio, F.; Alvarez,
M. Dissecting the Structure of Thiopeptides: Assessment of Thiazoline
and Tail Moieties of Baringolin and Antibacterial Optimization. J. Med.
Chem. 2014, 57, 4185−4195.
(7) Hudson, G. A.; Zhang, Z.; Tietz, J. I.; Mitchell, D. A.; van der Donk,
W. A. In vitro Biosynthesis of the Core of Scaffold of the Thiopeptide
Thiomuracin. J. Am. Chem. Soc. 2015, 137, 16012−16015.
(8) Young, T. S.; Dorrestein, P. C.; Walsh, C. T. Codon
Randomization for Rapid Exploration of Chemical Space in Thiopeptide
Antibiotic Variants. Chem. Biol. 2012, 19, 1600−1610.
ABBREVIATIONS USED
■
MIC, minimum inhibitory concentration; EC50, 50% effec-
tive dose; G+, Gram positive; MRSA, methicillin-resistant
Staphylococcus aureus; VRE, vancomycin-resistant enterococci;
S. aureus, Staphylococcus aureus; E. faecalis, Enterococcus faecalis; E.
faecium, Enterococcus faecium; S. pyogenes, Streptococcus pyogenes;
EF-Tu, elongation factor-Tu; BLD, below the limit of detection
6927
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928

Journal of Medicinal Chemistry
Article
́
(9) (a) Just-Baringo, X.; Albericio, F.; Alvarez, M. Thiopeptide
Engineering: A Multidisciplinary Effort Towards Future Drugs. Angew.
Chem., Int. Ed. 2014, 53, 6602−6616. For thiopeptide reviews, see
(b) Hughes, R. A.; Moody, C. J. From Amino Acids to Heteroaromatics-
Thiopeptide Antibiotics, Nature’s Heterocyclic Peptides. Angew. Chem.,
Int. Ed. 2007, 46, 7930−7954. (c) Bagley, M. C.; Dale, J. W.; Merritt, E.
A.; Xiong, X. Thiopeptide Antibiotics. Chem. Rev. 2005, 105, 685−714.
(d) Just-Baringo, X.; Albericio, F.; Alvarez, M. Thiopeptide Antibiotics:
Retrospective and Recent Advances. Mar. Drugs 2014, 12, 317.
(10) LaMarche, M. J.; Leeds, J. A.; Dzink-Fox, J.; Gangl, E.; Krastel, P.;
Neckermann, G.; Palestrant, D.; Patane, M. A.; Rann, E. M.; Tiamfook,
S.; Yu, D. Antibiotic Optimization and Chemical Structure Stabilization
of Thiomuracin A. J. Med. Chem. 2012, 55, 6934−6941.
(11) (a) Kabsch, W. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010,
D66, 125−132. (b) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P.
D.; Winn, M. D.; Storoni, L. C.; Read, R. J. Phaser Crystallographic
Software. J. Appl. Crystallogr. 2007, 40, 658−674. (c) Bricogne, G.;
Blanc, E.; Brandl, M.; Flensburg, C.; Keller, P.; Paciorek, W.; Roversi, P.;
Smart, O. S.; Vonrhein, C.; Womack, T. O. BUSTER, version 2.8.0.
2009, Global Phasing Ltd., Cambridge, United Kingdom. (d) Emsley,
P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features and Development of
Coot. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2010, D66, 486−501.
(12) Heffron, S. E.; Jurnak, F. Structure of an EF-Tu Complex with a
Thiazolyl Peptide Antibiotic Determined at 2.35 Å Resolution: Atomic
Basis for GE2270 A Inhibition of EF-Tu. Biochemistry 2000, 39, 37−45.
(13) Due to improved aqueous solubility, formulations with low
percentages of organic excipients were possible: Formulation = 10%
PEG-300, 0.1 N NaOH, water for injection, 0.1 N HCl (pH = 7.3).
(14) (a) Frimodt-Muller, N.; Knudesen, J. D.; Espersen, F. Handbook
̈
of Animal Models of Infection; Zak, O., Sand, M. A., Eds; Academic Press:
San Diego, 1999; pp 127−136. (b) O’Reilly, T.; Cleeland, R.; Squires, E.
L. Antibiotics in Laboratory Medicine, 4th ed.; Lorian, V., Ed; Williams
and Wilkins: Baltimore, 1996; pp 604−765.
(15) Sambol, S. P.; Merrigan, M. M.; Tang, J. K.; Johnson, S.; Gerding,
D. N. Colonization for the Prevention of Clostridium Difficile Disease in
Hamsters. J. Infect. Dis. 2002, 186, 1781−1789.
6928
DOI: 10.1021/acs.jmedchem.6b00726
J. Med. Chem. 2016, 59, 6920−6928