Katsuhiko Mitachi1, Hyun Gi Yun2, Sara M Kurosu1, Shakiba Eslamimehr1, Maddie R Lemieux1, Lada Klaić2, William M Clemons2, Michio Kurosu1. 1. Department of Pharmaceutical Sciences, College of Pharmacy, University of Tennessee Health Science Center, 881 Madison Avenue, Memphis, Tennessee 38163, United States. 2. Division of Chemistry and Chemical Engineering, California Institute of Technology, 1200 E. California Blvd, Pasadena, California 91125, United States.
Abstract
The spectrum of antibacterial activity for the nucleoside antibiotic FR-900493 (1) can be extended by chemical modifications. We have generated a small focused library based on the structure of 1 and identified UT-17415 (9), UT-17455 (10), UT-17460 (11), and UT-17465 (12), which exhibit anti-Clostridium difficile growth inhibitory activity. These analogues also inhibit the outgrowth of C. difficile spores at 2× minimum inhibitory concentration. One of these analogues, 11, relative to 1 exhibits over 180-fold and 15-fold greater activity against the enzymes, phospho-MurNAc-pentapeptide translocase (MraY) and polyprenyl phosphate-GlcNAc-1-phosphate transferase (WecA), respectively. The phosphotransferase inhibitor 11 displays antimicrobial activity against several tested bacteria including Bacillus subtilis, Clostridium spp., and Mycobacterium smegmatis, but no growth inhibitory activity is observed against the other Gram-positive and Gram-negative bacteria. The selectivity index (Vero cell cytotoxicity/C. difficileantimicrobial activity) of 11 is approximately 17, and 11 does not induce hemolysis even at a 100 μM concentration.
The spectrum of antibacterial activity for the nucleoside antibiotic FR-900493 (1) can be extended by chemical modifications. We have generated a small focused library based on the structure of 1 and identified UT-17415 (9), UT-17455 (10), UT-17460 (11), and UT-17465 (12), which exhibit anti-Clostridium difficile growth inhibitory activity. These analogues also inhibit the outgrowth of C. difficile spores at 2× minimum inhibitory concentration. One of these analogues, 11, relative to 1 exhibits over 180-fold and 15-fold greater activity against the enzymes, phospho-MurNAc-pentapeptide translocase (MraY) and polyprenyl phosphate-GlcNAc-1-phosphate transferase (WecA), respectively. The phosphotransferase inhibitor 11 displays antimicrobial activity against several tested bacteria including Bacillus subtilis, Clostridium spp., and Mycobacterium smegmatis, but no growth inhibitory activity is observed against the other Gram-positive and Gram-negative bacteria. The selectivity index (Vero cell cytotoxicity/C. difficileantimicrobial activity) of 11 is approximately 17, and 11 does not induce hemolysis even at a 100 μM concentration.
Clostridium difficile, a Gram-positive
bacterium, is transmitted by the fecal-oral route. C. difficile infection (CDI) can cause illness ranging
from diarrhea, colitis, and toxic inflammatory bowel disease to death. C. difficile has become one of the most common causes
of health care-associated infections in U.S. hospitals.[1] Approximately 250 000 people are hospitalized
in the U.S. every year from CDI.[2] The infective
form of C. difficile is the spore,
and its germination is the first committed step in CDI onset. C. difficile is found in abundance in the environment,
and colonizes in the gut where it produces toxins that cause C. difficile-associated diarrhea (CDAD).[3−5] Frequently, treatment with broad-spectrum antibiotic(s) has the
adverse effect of increasing the incidence of recurrent CDIs because
of the disruption of the normal balance of gut flora.[6] A recently described highly toxic strain, BI/NAP1/027,
exhibits resistance to fluoroquinolones and produces >20 times
more
toxins than historical strains.[7] Antibiotic
treatment of CDI is difficult because of the antibiotic resistance
and bacterial physiology (e.g., spore formation and protective effects
of pseudomembranous colitis).[8,9] Currently, there are
only a limited number of drugs available for the treatment of CDAD.[10] Metronidazole and vancomycin are the primary
therapy options for CDI. Vancomycin is recommended for severe infections
that do not respond to metronidazole. Vancomycin, rifaximin, and fidaxomicin
are used in recurrent or persistent cases.[11] For severe recurrent CDIs, fecal microbiota transplantation, which
involves instillation of stool from a healthy donor into the gastrointestinal
(GI) tract of the patient, has been highlighted to restore the gut
microbiome to a healthy state.[12] At therapeutic
concentrations, currently available drugs for the treatment of CDI
are not effective in inhibiting the germination or outgrowth of C. difficile spores. Only fidaxomicin has been reported
to inhibit spore production in C. difficile.[13] To date, bile acid derivatives and
a few organic molecules have been studied for their efficacy in the
reduction of spore viability or inhibition of spore germination.[14−19]In our continued efforts to identify strong inhibitors of
bacterial
phosphotransferases [phospho-MurNAc-pentapeptide translocase (translocase
I or MraY) and polyprenyl phosphate-GlcNAc-1-phosphate transferase
(WecA)] (Figure ),[20−25] we have generated a small focused library based on the structure
of the known MraY inhibitor FR-900493 (1, Figure )[26] and assayed against the vegetative and spore forms of a C. difficile strain. MraY is specific for UDP-N-acetylmuramyl-pentapeptide (Park’s nucleotide)
forming undecaprenyl diphosphoryl-N-acetylmuramate-pentapeptide
(lipid I). WecA and its homologues (e.g., TagO or TarO) in Gram-negative
bacteria are responsible for the first step of wall teichoic acid
synthesis that anchors the phospho-GlcNAc moiety of UDP-GlcNAc to
undecaprenyl phosphate (C55–P). MraY enzymatic activity
is essential for the growth of both Gram-positive and Gram-negative
bacteria, and thus is considered as a target of interest for the discovery
of novel antibacterial agents. The major source of MraY inhibitors
resides in the nucleoside-based antibiotic group. This group has been
subdivided into four classes: tunicamycins, ribosaminouridines, uridyl
peptides, and capuramycins. Analysis of the pharmacological behaviors
observed with several compounds in these classes shows a broad spectrum
of antibacterial activity, including relevant drug-resistant strains,
and in vivo efficacy without apparent toxicity. Ribosaminouridine
antibiotics, which include muraymycin, liposidomycins, caprazamycin,
and FR-900493 (1), exhibit the most promising biological
profiles.[27] On the other hand, essentiality
of the WecA/TagO transferase subfamily is not completely understood
in the growth of many bacteria.[28,29] Only a few molecules
have been reported to exhibit WecA inhibitory activity.[28,29] Herein, we report a structure–activity relationship (SAR)
obtained from a small focused library of FR-900493 (1) in a range of enzymes (MraY, WecA, and AglH) and bacterial growth
inhibitory assays. Efficacy of these new MraY inhibitors against C. difficile spores along with in vitro toxicity
assessments of the selected anti-C. difficile molecules is discussed.
Figure 1
Bacterial phosphotransferases and a human glycosyltransferase.
MraY (MurX) is an established drug target for Gram-positive and Gram-negative
bacterial infections. WecA is essential in the growth of Mycobacterium spp. and some Gram-positive bacteria
(TagO or TarO). DPAGT1 is a human glycosyltransferase. Strong inhibition
of DPAGT1 may cause cytotoxicity in mammalian cells.
Figure 2
Structures of FR-900493 (1) and representative
muraymycins,
muraymycin A1 (2) and D1 (3). FR-900493
has only the right half of the muraymycins (the highlighted portion
in blue). The unknown stereochemistries of the 5′- and 6′-positions
are determined unequivocally in this study.
Bacterial phosphotransferases and a human glycosyltransferase.
MraY (MurX) is an established drug target for Gram-positive and Gram-negative
bacterial infections. WecA is essential in the growth of Mycobacterium spp. and some Gram-positive bacteria
(TagO or TarO). DPAGT1 is a human glycosyltransferase. Strong inhibition
of DPAGT1 may cause cytotoxicity in mammalian cells.Structures of FR-900493 (1) and representative
muraymycins,
muraymycin A1 (2) and D1 (3). FR-900493
has only the right half of the muraymycins (the highlighted portion
in blue). The unknown stereochemistries of the 5′- and 6′-positions
are determined unequivocally in this study.
Results and Discussion
Chemistry and SAR of FR-900493
Aminoribosyl-uridyl
peptide antibiotics such as FR-900493 (1) and muraymycins
are an important class of natural products for the development of
novel antibacterial agents.[30,31] Chemical syntheses
of 1 and muraymycin analogues are essential to perform
exhaustive SAR studies.[32] Muraymycin A1
(2) is one of the most active members of this family
against both Gram-positive and Gram-negative bacteria. The fatty acid
side chain (R2) of 2 is critical for antimicrobial
activity as muraymycin D1 (3) and the other related molecules
lacking the R2 group are poorly active (Figure ).[25,33] Interestingly, we have demonstrated that muraymycin D1 shows strong
bacteriostatic activity against Mycobacterium tuberculosis by targeting the bacterial phosphotransferases (MurX and WecA).[25] Although FR-900493 (1) possesses
only one half of the structure of the muraymycins, it displays antistaphylococcal
activity [minimum inhibitory concentration (MIC) 3.13 μg/mL]
in vitro and in vivo. The LD50 value of 1 is
over 500 mg/kg, which was determined via intravenous administrations
in mice, indicating that 1 is an ideal scaffold to develop
into new antibacterial agents.[26] It is
interesting to note that 1 was isolated from the culture
broth of Bacillus cereus (no. 2045),
whereas other aminoribosyl-uridyl peptide antibiotics including muraymycins
were isolated from Streptomyces spp.[31,33] Structurally, the C6′-amino group of FR-900493 is methylated,
whereas O-methylation at the C2″-position
is observed in muraymycin A1 and D1. Absolute stereochemistries at
the C5′- and C6′-positions are speculated to be 5′S and 6′S, respectively, based on
the correlations with stereochemistries of muraymycins and structurally
related molecules, caprazamycins. On the basis of the proposed structure,
FR-900493 was first synthesized by Hirano and co-workers who only
reported physical chemistry data (e.g., NMR and optical rotation)
for the synthetic molecule.[34] We have synthesized
the four diastereomers of FR-900493 with respect to the C5′-
and C6′-stereocenters according to the synthetic scheme developed
for muraymycin D1 (3)[25] with
minor modifications and compared their physical and biological data
with those of the natural product. Chemical shifts and coupling constants
of the 5′S,6′S-diastereomer 1 showed good agreement with those of the natural FR-900493
(see the Supporting Information).[26] The four diastereomers of FR-900493 were evaluated
against the bacterial phosphotransferases MraY and WecA (Table ).[29,35]
Table 1
Inhibitory Activity of Bacterial Phosphotransferases
(MraY and WecA) and C. difficile Growth
by the C5′- and 6′-Diastereomers of FR-900493a
compound
WecA inhibition IC50(μM)b
MraY inhibition IC50 (μM)c
C. difficile ATCC 43596 MIC (μg/mL)d
FR-900493 (1)
5.0 ± 5.44
25.0 ± 8.67
>25.0
5′S,6′R-diastereomer (4)
>100
>100
>25.0
5′R,6′S-diastereomer (5)
>100
>100
>25.0
5′R,6′R-diastereomer (6)
>100
>100
>25.0
FR-900493-amide (7)
5.0 ± 6.34
25.0 ± 9.67
>25.0
8
>100
>100
>25.0
tunicamycin
0.15 ± 7.80
3.38 ± 7.32
>25.0
WecA and MraY assays (see the Supporting Information).
E. coli WecA-containing membrane was used.
Hydrogenivirga spp. MraY
was used.
A microdilution
broth method was
used.
WecA and MraY assays (see the Supporting Information).E. coli WecA-containing membrane was used.Hydrogenivirga spp. MraY
was used.A microdilution
broth method was
used.Surprisingly, 1 exhibited a weak MraY inhibitory activity
(IC50 25.0 μM) but a moderate WecA inhibitory activity
(IC50 5.0 μM). All unnatural diastereomers (4, 5, and 6) did not display either
MraY or WecA inhibitory activity, even at a 100 μM concentration.
The results of these enzymatic assays unequivocally determined the
absolute stereochemistry of FR-900493 to be 5′S and 6′S configurations. We have observed
that amidation of the C6′-carboxylic group in muraymycin D1
(3) does not decrease the MraY/WecA activity.[25] Similarly, FR-900493-amide (7)
exhibited an MraY/WecA inhibitory activity equal to that of 1. The N-methyl group of 1 is
essential to inhibit the MraY and WecA enzymes; the de-N-methyl analogue 8 completely lost the MraY/WecA inhibitory
activities. FR-900493 and its analogues shown in Table were not effective in killing C. difficile (ATCC 43596) at 25.0 μg/mL or
lower concentrations. As exemplified in the antibacterial activity
of the muraymycin family molecules,[30] the
hydrophobic residues appended on the 3-aminopropylamine portion play
a key role in selectivity and susceptibility against bacteria (Figure ). We have generated
a small forced library based on the core structures of 7 and 8, and the generated molecules were assayed against C. difficile (ATCC 43596) at a single concentration
of 50.0 μg/mL. Four molecules (9, 10, 11, and 12) displayed anti-C. difficile activity (Table ), and sufficient amounts of these molecules
were resynthesized for thorough in vitro profiling.
Table 2
Inhibitory Activity of Bacterial Phosphotransferases
(MraY and WecA) and C. difficile Growth
by 9–12a
compound
WecA inhibition
IC50 (μM)b
MraY inhibition IC50 (μM)c
C.
difficile ATCC
43596 MIC (μg/mL)d
FR-900493 (1)
5.0 ± 5.44
25 ± 8.67
>50.0
UT-17415 (9)
0.85 ± 7.50
0.69 ± 7.32
25.0 ± 7.64
UT-17455 (10)
12.5 ± 6.01
0.25 ± 5.80
12.5 ± 5.93
UT-17460 (11)
0.32 ± 4.09
0.08 ± 4.33
3.25 ± 4.40
UT-17465 (12)
12.3 ± 5.75
7.70 ± 5.47
50.0 ± 5.81
UT-01320
0.035 ± 9.14
>100
>50.0
tunicamycin
0.15 ± 7.80
3.38 ± 7.32
>50.0
vancomycin
>50.0
>50.0
2.5 ± 7.25
WecA and MraY assays
(see the Supporting Information).
E. coli WecA-containing membrane was used.
Hydrogenivirga spp. MraY
was used.
A microdilution
broth method was
used.
WecA and MraY assays
(see the Supporting Information).E. coli WecA-containing membrane was used.Hydrogenivirga spp. MraY
was used.A microdilution
broth method was
used.
Syntheses of FR-900493
Analogues, 9, 10, 11, and 12
In our synthesis
of muraymycin D1 (3), β-ribosylation of the C2-ether-protected
ribosedonor (R1 = Me in 15) and Strecker
reaction with mono-protected 1,3-diaminopropane to form secondary amine were performed stereoselectively.[25] Our revised synthetic routes for FR-900493 analogues (9–12) are illustrated in Schemes and 2. The monomethoxytetrachlorodiphenylmethoxymethyl
(MTPM)-protected uridine 13 was prepared according to
the previously reported procedure.[36] The primary alcohol of 13 was oxidized by a modified
Swern condition to provide the corresponding aldehyde in a quantitative
yield, which was then subjected to Carreira’s asymmetric alkynylation
reaction using (−)-N-methylephedrine,[37] furnishing the (S)-propargyl
alcohol 14 in 80% yield with a selectivity of >98:2.
The stereochemistry of the secondary alcohol of 14 generated via Carreira’s alkynylation was determined
by the advanced Mosher’s method (see the Supporting Information).[38] The
newly devised glycosyl donor 15 was designed to provide
β-riboside whose protecting groups can be deprotected under
mild acidic conditions. N-Iodosuccinimide (NIS)–AgBF4-promoted ribosylation of (S)-propargyl alcohol 14 with 15 furnished the β-riboside 16 exclusively in 95% yield.[23,25,39] The azido group of 16 was reduced with
Zn metal in the presence of aq NH4Cl, and the generated
free-amine was protected with (Boc)2O to furnish 17 in 90–92% overall yield. The alkyne moiety of 17 was subjected to a three-step procedure (partial reduction
with a Lindlar catalyst, osmylation, and oxidative cleavage with Pb(OAc)4), providing the crude aldehyde 18. In a systematic
screening of catalysts for Strecker reaction of 18 with
the 4-aminobutanamide derivatives 19 (Scheme ) or 20 (Scheme ), it was realized
that (BnO)2P(O)CH2P(O)(OBn)OH provided a ∼4:1
mixture of the 6′S- and 6′R-diastereomers (21 and 21) in greater than 80% yield. The desired diastereomer 21 was subjected to hydration reaction with
HgCl2-acetaldoxime to furnish the amide 22 in 95% overall yield. N-methylation of 22 was performed via reductive amination with paraformaldehyde and
NaB(CN)H3 to afford 23 in 95% yield. Global
deprotection of 22 and 23 to form the desired
products 9 and 10, respectively, was performed
in a one-pot two-step reaction using 30% trifluoroacetic acid (TFA)
followed by 80% TFAat 40 °C; the crude products were purified
by C18 reverse-phase high-performance liquid chromatography
(HPLC) (MeOH/H2O = 75:25) to yield 9 and 10 in 85–90% yield. The stereochemistry of the
C6′-stereocenter for 21 generated
via Strecker reaction was unequivocally determined via its conformational
analyses (see the Supporting Information). Similarly, the analogues 11 and 12 were
synthesized with the primary amine 20 via the synthetic scheme developed for 9 and 10 (Scheme ).
Scheme 1
Syntheses of Anti-C. difficile FR-900493
Analogues 9 and 10
Scheme 2
Syntheses of Anti-C. difficile FR-900493
Analogues 11 and 12
Enzyme and Bacterial Growth Inhibitory Activities of 9, 10, 11, and 12
Despite the extensive efforts to obtain membrane fractions from C. difficile (ATCC 43596), we could not obtain sufficient
amounts of active C. difficile membrane
containing MraY and WecA homolog. In this study, for assays, we used
purified MraY from the thermophilic bacterium Hydrogenivirga (HyMraY) and the active membrane fraction (P-60)
from Escherichia coli. Protein sequence
alignment via BLAST[40] of HyMraY and EcWecA against C. difficile (strain F501) revealed that MraY between Hydrogenivirga spp. and C. difficile showed 57%
similarity/38% identity and WecA/putative WecA homolog between E. coli and C. difficile showed 47% similarity/28% identity. We previously confirmed that
the activities of inhibitor molecules against MraY and WecA are similar
across various species: Mycobacterium smegmatis, M. tuberculosis, E. coli, and Hydrogenivirga spp.[21,29,35] The synthetic
molecules (9, 10, 11, and 12) were evaluated against the phosphotransferases (MraY and
WecA) and C. difficile strain (ATCC
43596) (Table ). The N-methyl analogues, 9 and 12,
exhibited weak anti-C. difficile growth
inhibitory activity, although their MraY inhibitory activities were
over 30- and 3-fold, respectively, more potent than that of FR-900493
(1). In sharp contrast, the de-N-methyl
analogues, 10 and 11, were over 100 and
300-times more potent in the MraY inhibitory activity than 1. The WecA inhibitory activity of 11 was ∼15-fold
more than that of 1. Therefore, the anti-C. difficile activity of 11 is likely
attributed to the inhibition of the MraY enzyme. This hypothesis was
further supported by the fact that a selective WecA inhibitor, UT-01320,[20] did not inhibit growth of C.
difficile at >50.0 μg/mL concentrations.
The
analogue 11 was identified to be a strong MraY/WecA inhibitor,
whose activity was significantly better than that of a known MraY/WecA
inhibitor, tunicamycin.[28,29,35] The anti-C. difficile activity is
correlated with the enzyme inhibitory activity of MraY; 10 and 11 displayed MIC values of 12.5 and 3.25 μg/mL,
respectively, against C. difficile (e.g.,
an MIC of 2.50 μg/mL for vancomycin).
Effect of 9, 10, and 11 on C. difficile Spores
The
effect of 9, 10, and 11 on C. difficile spores was determined by counting colony-forming
units (CFUs) of the spore germination on taurocholate-containing agar
plates after the treatment of the C. difficile spores with these analogues (2× MIC) for 24 h. C. difficile spores show resistance to most known
anti-C. difficile agents.[19] Indeed, in these studies, vancomycin, metronidazole,
and linezolid did not inhibit the spore outgrowth, even at 5×
MIC. On the contrary, the new MraY inhibitors 9, 10, and 11 prevented the outgrowth of the C. difficile spores into colonies at 2× MIC
(Figure ).
Figure 3
Viability of C. difficile (ATCC
43596) spores treated with the MraY inhibitors, 9, 10, and 11, and representative anti-C. difficile drugs. The spores treated with molecules
(2× or 5× MIC) in a BHI medium or saline (24 h) were plated
on a BHI agar plate containing sodium taurocholate and incubated anaerobically
for 48 h. Vancomycin, metronidazole, and linezolid were treated with
5× MIC, and 9, 10, and 11 were treated with 2× MIC. Germinated spores were counted (CFUs)
(p ≤ 0.05).
Viability of C. difficile (ATCC
43596) spores treated with the MraY inhibitors, 9, 10, and 11, and representative anti-C. difficile drugs. The spores treated with molecules
(2× or 5× MIC) in a BHI medium or saline (24 h) were plated
on a BHI agar plate containing sodium taurocholate and incubated anaerobically
for 48 h. Vancomycin, metronidazole, and linezolid were treated with
5× MIC, and 9, 10, and 11 were treated with 2× MIC. Germinated spores were counted (CFUs)
(p ≤ 0.05).
Physicochemical Properties and in Vitro Toxicity of 11
UT-17460 (11) exhibited pharmacological characteristics
superior to those of UT-17455 (10); (1) water solubility
of 11 (22.0 mg/mL) is 200 times greater than that of 10 (0.11 mg/mL), (2) 11 is approximately 8 times
less cytotoxic than 10 against Vero cells (African green
monkey kidney epithelial cells, IC50 65.0 μM for 10), and (3) 11 showed relatively low induction
of hemolysis (IC50 205.6 μM), whereas 10 lysed blood cells at a much lower concentration (IC50 70.0 μM) than 11 (Table ). Thus, we selected 11 for
further in vitro pharmacological evaluation. WecA enzyme inhibitors
have the potential to interfere with a human homologue, dolichyl-phosphate
GlcNAc-1-phosphotransferase 1 (DPAGT1), which catalyzes the first
step of the protein N-glycosylation process in humans
(Figure ). Thus, strong
inhibition of DPAGT1 may cause cytotoxicity in vitro and in vivo.[41] We have established a counterselection assay
using a thermophilic dolichyl-phosphate GlcNAc-1-phosphotransferase
(AglH from Methanocaldococcus jannaschii) to assess the toxicity level of antibacterial phosphotransferase
inhibitors (Table ). Interestingly, 11 exhibited a relatively stronger
AglH inhibitory activity (IC50 3.61 μM) than 10 and tunicamycin; however, the IC50 level of 11 against Vero cells was much higher than those of 10 and tunicamycin. The FR-900493 analogue 11 exhibited low permeability across Caco-2 epithelial monolayers (Papp rate coefficient < 1 × 10–6 cm/s) with moderate efflux (an efflux ratio of 5.5), predicting
that 11 is poorly absorbed from the GI tract.
Table 3
Counterselection Assays Based on Prenyl
phosphate-GlcNAc-1-phosphotransferases (WecA vs AglH) and in Vitro
Cytotoxicities
WecA
inhibition IC50 (μM)a
compound
E. coli WecA
M. smegmatis WecA
AglH inhibition
IC50 (μM)aM. jannaschii AglH
Vero
cells IC50 (μM)a
hemolysis IC50 (μM) sheep blooda
UT-17455 (10)
12.5 ± 6.01
12.5 ± 5.42
12.5 ± 5.18
7.08 ± 5.67
70 ± 5.71
UT-17460 (11)
0.32±4.09
0.25±4.84
3.61±4.43
65±4.65
205.8±4.33
tunicamycin
0.15 ± 7.80
0.15 ± 7.22
13.27 ± 8.10
0.12 ± 7.78
15 ± 7.83
All assay procedures are described
in the Supporting Information.
All assay procedures are described
in the Supporting Information.
Spectrum of Antibacterial Activity of 11
As summarized in Figure , UT-17460 (11) exhibited a
very narrow spectrum
of antibacterial activity; 11 also killed the other Clostridium spp., Clostridium perfringens (MIC 3.1 μg/mL), Bacillus subtilis (MIC 1.6 μg/mL), and M. smegmatis (MIC 0.2 μg/mL) but did not inhibit the growth of Lactobacillus spp., Staphylococcus
aureus, Staphylococcuspneumoniae, and all Gram-negative bacteria tested at 25.0 μg/mL or higher
concentrations.
Figure 4
Antibacterial activity of UT-17460 (11, line
in red).
Antibacterial activity of UT-17460 (11, line
in red).
Conclusions
FR-900493
(1) has long been known as an MraY inhibitor
for which we have firmly established the stereochemistry. In our thorough
enzyme inhibitory assays of the phosphotransferases, we concluded
that the antibacterial activity of 1 is attributed to
a combination of WecA and MraY inhibitory activities, but 1 is a weak MraY inhibitor (IC50 25.0 μM). Because
WecA is not essential in the growth of many bacteria except for a
few pathogens (e.g., Mycobacterium spp.),[28,42] improving MraY inhibitory activity of 1 should be a
beneficial direction to identify effective antibacterial agents. Stereochemistry
of the C5′- and C6′-positions of 1 is required
to be the natural configuration (5′S and 6′S) to exhibit bacterial growth and MraY/WecA enzyme inhibitory
activities. The C6′-CO2H group of 1 can be masked as its primary amide without decreasing MraY/WecA
activity; however, the N-methyl group is essential
for 1 to display phosphotransferase inhibitory activities.
Muraymycins lack the N-methyl group but exhibit a
strong MraY enzyme inhibitory activity. Importantly, only the muraymycin
analogues having a hydrophobic side chain [e.g., muraymycin A1 (2)] exhibit a strong antimicrobial activity. We have generated
a small focused library based on the structures of 7 and 8 and screened against the vegetative form of C. difficile strain (ATCC 43596) under anaerobic
conditions. We identified four anti-C. difficileFR-900493 analogues, 9, 10, 11, and 12. These analogues were resynthesized via the
synthetic procedures reported previously with modifications of selective
ribosylation (14 → 16) and Strecker
reactions (18 → 21 and 24) (Schemes and 2). The resynthesized analogues were characterized
by C. difficile growth and the bacterial
phosphotransferase inhibitory assays. The analogue 11 exhibited strong MraY/WecA inhibitory activity (IC50 0.08
and 0.30 μM, respectively) and killed the vegetative state of C. difficile with an MIC value of 3.25 μg/mL.
In vitro cytotoxicity level of 11 was much lower than
those of tunicamycin and 10; although 11 inhibited AglH at low concentrations. Toxicity of tunicamycin and 10 in mammalian cells may be largely due to the membrane disruption
as demonstrated by the hemolysis assay. On the other hand, the hemolytic
activity of 11 is attenuated by the pharmacologically
benign hydrophobic group.[43] Interestingly,
the MraY inhibitors 9, 10, and 11 inhibited the outgrowth of the C. difficile spores (endospores) at 2× MIC concentrations (Figure ). The inhibitory mechanism
of outgrowth of the spores by the MraY inhibitors is far from completely
understood. Recently, a strong lipid II-binding antibacterial peptide,
nisin, was reported to inhibit the viability of the C. difficile spores at high concentrations.[44] The effect of fidaxomicin was also observed
in the inhibition of the C. difficile spore outgrowth at high concentrations (4.0–14 μg/mL,
MIC90 0.5 μg/mL against vegetative C. difficile).[45]C. difficile spores are metabolically dormant and
are known to exhibit impermeability to a wide range of organic molecules.[46] Thus, organic molecules with a high molecular
mass (e.g., Mw = 881.3 for 11) are not likely to permeate the spore walls and directly inhibit
the viability of the spores. Certain peptides or nucleoside derivatives
are recognized by the Bacillus anthracis germination machinery, increasing their susceptibility to geminated
spores or outgrowing spores.[47,48] On the other hand,
germination machineries of C. difficile spores remain less well-understood because of the significantly
lower levels of homologues of the spore coat proteins than those of
other well-studied bacteria (e.g., B. subtilis and B. anthracis).[48] Recent studies demonstrated that C. difficile uses subtilisin-like serine proteases (Csp) that regulate spore
germination.[49] In C. difficile, CspC is likely to transmit the bile acid signal to CspB, which
activates the spore cortex lytic enzyme (i.e., SleC) by cleavage of
pro-SlecC. SlecC alters the structure of the spore cortex and induces
Ca-dipicolinate release, which triggers the outgrowth of C. difficile spores.[50] Our MraY inhibitors may be involved in the induction from the spore
forms to the germinating state. The germinated or outgrowing spores
require MraY, which will be inhibited by the new inhibitors 9–12. We are currently evaluating the effectiveness
of structurally distinct MraY inhibitors against the C. difficile spores. Correlation of nucleoside-based
MraY inhibitor and ability of the C. difficile spore germination are being investigated. In preliminary bacterial
growth inhibitory assays against a series of bacteria, 11 displayed a very narrow-spectrum antibiotic activity (Figure ), which is advantageous in
the development of selective anti-C. difficile agents that do not disrupt the gut microbiota in humans. In addition,
in vitro pharmacokinetic data obtained via the Caco-2 permeability
assay may indicate that 11 is poorly absorbed from the
GI tract, which may reduce the toxicity of 11 when given
orally. Pharmacokinetics and oral bioavailability, susceptibility
to clinically isolated C. difficile and in vivo efficacy of 11 (using preclinical animal
models), and selectivity of antibacterial activity of 11 at higher concentrations (1000 μg/mL) are the object of future
studies.
Experimental Section
Chemistry: General Information
All
chemicals were purchased
from commercial sources and used without further purification unless
otherwise noted. Tetrahydrofuran (THF), CH2Cl2, and N,N-dimethylformamide were
purified via an Innovative Technology’s Pure-Solve System.
All reactions were performed under an argon atmosphere. All stirring
procedures were performed with an internal magnetic stirrer. Reactions
were monitored by thin-layer chromatography (TLC) using 0.25 mm coated
commercial silica gel plates (EMD, silica gel 60F254).
TLC spots were visualized by UV light at 254 nm or developed with
ceric ammonium molybdate or anisaldehyde or copper sulfate or ninhydrin
solutions by heating on a hot plate. Reactions were also monitored
by using Shimadzu LCMS-2020 with the following solvents: A: 0.1% formic
acid in water and B: acetonitrile. Flash chromatography was performed
with SiliCycle silica gel (Purasil 60 Å, 230–400 mesh).
Proton magnetic resonance (1HNMR) spectral data were recorded
on 400 and 500 MHz instruments. Carbon magnetic resonance (13CNMR) spectral data were recorded on 100 and 125 MHz instruments.
For all NMR spectra, chemical shifts (δH, δC) were quoted
in parts per million (ppm) and J values were quoted
in Hz. 1H and 13CNMR spectra were calibrated
with a residual undeuterated solvent (CDCl3: δH =
7.26 ppm, δC = 77.16 ppm; CD3CN: δH = 1.94
ppm, δC = 1.32 ppm; CD3OD: δH = 3.31 ppm, δC
= 49.00 ppm; DMSO-d6: δH = 2.50
ppm, δC = 39.52 ppm; D2O: δH = 4.79 ppm) as
an internal reference. The following abbreviations were used to designate
the multiplicities: s = singlet, d = doublet, dd = double doublets,
t = triplet, q = quartet, quin = quintet, hept = heptet, m = multiplet,
and br = broad. Infrared (IR) spectra were recorded on a PerkinElmer
FT1600 spectrometer. HPLC analyses were performed with a Shimadzu
LC-20AD HPLC system. All compounds were purified by reverse HPLC to
be ≥95% purity. High-resolution mass spectrometry (HRMS) data
were obtained from a Waters SYNAPT G2-Si (ion mobility mass spectrometer
with nano-electrospray ionization).
A suspension of 16 (286 mg, 0.19 mmol), NH4Cl (305 mg, 5.70 mmol),
and Zn (373 mg, 5.70 mmol) in EtOH/H2O (9:1, 9.5 mL) was
stirred at 80 °C for 12 h and cooled to rt. The precipitates
were filtered, and the combined organic solution was concentrated
in vacuo. The crude mixture was used for the next reaction without
purification. To a stirred solution of the crude amide in THF (9.5
mL) were added saturated aq NaHCO3 (9.5 mL) and Boc2O (124 mg, 0.57 mmol). The reaction mixture was stirred for
6 h at rt, and the aqueous layer was extracted with EtOAc. The combined
organic extracts were dried over Na2SO4 and
concentrated in vacuo. The crude mixture was purified by silica gel
column chromatography (hexane/EtOAc 85:15 to 80:20 to 67:33) to afford 17 (258 mg, 0.16 mmol, 86% for 2 steps): TLC (hexane/EtOAc
70:30) Rf = 0.30; [α]D21 +0.012 (c = 0.90, CHCl3); IR (thin film) νmax: 3387 (br), 3090, 2941, 2866, 1742, 1720, 1676, 1600, 1556,
1505, 1456, 1366, 1278, 1219, 1161, 1100, 1070, 1049, 1013, 998, 882,
806, 772, 745, 681 cm–1; 1HNMR (400
MHz, CDCl3) δ: 7.54 (dd, J = 19.9,
8.5 Hz, 1H), 7.33–7.27 (m, 4H), 7.24–7.16 (m, 4H), 6.85
(d, J = 7.3 Hz, 2H), 6.51 (d, J =
4.8 Hz, 1H), 5.72–5.64 (m, 2H), 5.60–5.48 (m, 2H), 5.26
(d, J = 6.0 Hz, 1H), 5.17 (d, J =
8.6 Hz, 2H), 5.13–5.08 (m, 1H), 4.82–4.76 (m, 1H), 4.65
(t, J = 7.0 Hz, 1H), 4.51 (dd, J = 13.8, 6.0 Hz, 1H), 4.31–4.26 (m, 1H), 4.23–4.17
(m, 1H), 3.78 (d, J = 2.7 Hz, 3H), 3.74 (d, J = 6.9 Hz, 4H), 3.48–3.40 (m, 1H), 3.36–3.26
(m, 1H), 2.83 (t, J = 7.4 Hz, 2H), 2.56 (t, J = 7.5 Hz, 2H), 2.27 (t, J = 2.6 Hz, 2H),
2.23 (t, J = 3.0 Hz, 2H), 1.62–1.55 (m, 7H),
1.42 (s, 9H), 1.37 (d, J = 2.6 Hz, 3H), 1.11–0.99
(m, 54H); 13CNMR (101 MHz, CDCl3) δ:
159.46, 150.79, 136.91, 131.26, 129.33, 128.48 (2C), 128.37 (2C),
126.48, 126.11, 125.44, 115.30, 115.26, 79.95, 59.98, 55.68, 46.42,
46.16, 45.95, 44.83, 44.78, 42.49, 34.51, 34.49, 32.60, 32.55, 31.91,
29.69, 28.71, 28.40, 28.34, 27.37, 27.33, 27.29, 27.26, 27.09, 27.05,
25.36, 22.68, 22.63, 20.87, 18.06 (6C), 17.90 (6C), 14.12, 11.92 (3C),
11.78 (3C); HRMS (ESI+) m/z: calcd for C79H116Cl4N3O17Si2 [M + H], 1574.6597; found, 1574.6609.
Synthesis of 21
To a
stirred solution of 17 (258 mg, 0.16 mmol) and quinoline
(38.7 μL, 0.33 mmol) in EtOAc (50 mL) and MeOH (50 mL) was added
Lindlar catalyst (300 mg). H2 gas was introduced, and the
reaction mixture was stirred under a H2 atmosphere (600
psi) at rt. The reaction mixture was stirred for 11 h under a H2 atmosphere (600 psi) at rt. The reaction mixture was filtered
through Celite, and the filtrate was washed with 1 NHCl. The combined
organic solution was dried over Na2SO4 and concentrated
in vacuo. The crude mixture was used for the next reaction without
purification. To a stirred solution of the crude mixture in t-BuOH/acetone (1:1, 2.1 mL) were added NMO (192 mg, 1.64
mmol) and OsO4 (4% in water, 1.04 mL, 0.16 mmol) at rt.
After 2 h, the reaction solution was diluted with EtOAc and quenched
with saturated aq NaHCO3/saturated aqNa2SO3 (2:1). The heterogeneous mixture was stirred for 30 min and
extracted with EtOAc. The combined organic extracts were dried over
Na2SO4 and concentrated in vacuo. The crude
mixture was passed through a silica gel pad (hexane/EtOAc 33:67) to
afford the diols as a diastereomeric mixture. This mixture was used
for the next reaction without further purification. To a stirred solution
of the diols (22.1 mg, 0.014 mmol) and NaHCO3 (11.5 mg,
0.14 mmol) in CH2Cl2 (0.7 mL) was added Pb(OAc)4 (12.1 mg, 0.027 mmol) at 0 °C. The reaction mixture was stirred for 2 h at 0 °C,
quenched with saturated aq NaHCO3, and extracted with EtOAc.
The combined organic extracts were dried over Na2SO4 and concentrated in vacuo. The crude aldehyde 18 was used for the next reaction without purification. To a stirred
solution of (BnO)2P(O)–CH2–P(O)(OBn)OH
(30.6 mg, 0.069 mmol) in CH2Cl2 (0.4 mL) was
added a CH2Cl2 (0.3 mL) solution of 18 and 19. To the reaction mixture was added TMSCN (17.1
μL, 0.14 mmol) and stirred for 9 h at rt. After completion,
the reaction mixture was quenched with saturated aq NaHCO3 and extracted with EtOAc. The combined organic extracts were dried
over Na2SO4 and concentrated in vacuo. The crude
product was purified by silica gel column chromatography (hexane/EtOAc
80:20 to 60:40) to afford the Strecker products. To a stirred solution
of the desired Strecker product (8.8 mg, 5.0 μmol) in EtOH/H2O (9:1, 0.5 mL) were added HgCl2 (2.7 mg, 0.010
mmol) and acetaldoxime (3.0 μL, 0.050 mmol) at rt. After being
stirred for 6 h at rt, the reaction mixture was concentrated under
reduced pressure. The residue was quenched with saturated aq NaHCO3 and extracted with CHCl3. The combined organic
extracts were dried over Na2SO4 and concentrated
in vacuo. The crude product was purified by silica gel column chromatography
(hexane/EtOAc 80:20 to 60:40) to afford 21 (16.7 mg, 9.49 μmol, 69% for two steps) and 21 (4.1 mg, 2.34 μmol, 17% for two steps): 21: TLC (hexane/EtOAc 60:40) Rf = 0.40; [α]D21 +0.075 (c = 0.73, CHCl3); IR (thin film) νmax: 3317 (br), 2930,
2865, 1719, 1675, 1600, 1462, 1102, 1071, 882, 772, 683 cm–1; 1HNMR (400 MHz, CDCl3) δ: 7.68 (s,
1H), 7.49 (dd, J = 11.4, 8.8 Hz, 1H), 7.39 (d, J = 7.9 Hz, 2H), 7.32 (s, 1H), 7.19 (d, J = 8.5 Hz, 2H), 7.11 (d, J = 8.0 Hz, 2H), 6.86 (d, J = 9.3 Hz, 2H), 6.50 (d, J = 15.4 Hz,
1H), 5.73 (dd, J = 23.0, 8.0 Hz, 1H), 5.59 (d, J = 5.9 Hz, 1H), 5.54 (d, J = 9.4 Hz, 2H),
5.42 (t, J = 10.1 Hz, 1H), 5.25 (s, 1H), 5.08–5.00
(m, 2H), 4.96–4.82 (m, 2H), 4.50–4.45 (m, 1H), 4.25–4.19
(m, 1H), 4.15–4.06 (m, 1H), 3.94–3.83 (m, 1H), 3.80–3.63
(m, 10H), 3.49–3.41 (m, 1H), 3.39–3.31 (m, 1H), 3.03
(dt, J = 12.0, 6.1 Hz, 1H), 2.71–2.61 (m,
1H), 2.54 (t, J = 7.3 Hz, 2H), 2.51–2.45 (m,
1H), 2.29–2.17 (m, 4H), 1.67–1.51 (m, 10H), 1.41 (s,
9H), 1.28 (dd, J = 15.7, 8.1 Hz, 10H), 1.05 (s, 42H),
1.01 (s, 6H), 0.95 (s, 6H), 0.87 (t, J = 6.4 Hz,
3H); 13CNMR (101 MHz, CDCl3) δ: 170.9,
159.5, 136.9, 136.8, 131.3, 131.2, 129.42, 129.36, 128.84 (2C), 128.82
(2C), 128.80 (2C), 126.4, 126.2, 125.1, 120.09, 120.05, 115.4, 115.31,
115.30, 114.84, 114.81, 84.90, 84.87, 80.84, 80.78, 80.2, 79.8, 79.4,
78.2, 76.1, 74.3, 60.0, 59.9, 55.8, 55.7, 52.0, 46.2, 46.0, 44.83,
44.77, 35.4, 32.56, 32.55, 31.8, 31.5, 29.7, 29.19, 29.16, 28.4, 28.3,
27.3, 27.2, 22.7, 18.1 (12C), 14.1, 11.9 (6C); HRMS (ESI+) m/z: calcd for C88H135Cl4N6O18Si2 [M + H], 1759.8126; found, 1759.8135.
M. smegmatis (ATCC 607), Klebsiella pneumoniae (ATCC 8047), Pseudomonas aeruginosa (ATCC 27853), Acinetobacter baumannii (ATCC 19606), S. aureus (BAA-1683), C. difficile (ATCC 43596), Enterococcus
faecium (ATCC 349), Fusobacterium periodontium ATCC 33693), Bacteroides fragilis (ATCC 25285), Streptococcus pneumoniae (ATCC 6301), Bacillus subtilis (ATCC
6051), C. perfringens (ATCC 13124), Lactobacillus casei (ATCC 393), Lactobacillus
acidophilus (ATCC 4356), and E. coli (ATCC 10798) were obtained from American Type Culture Collection
(ATCC). A single colony of M. smegmatis was obtained on Difco Middlebrook 7H10 nutrient agar enriched with
albumin, dextrose, and catalase. Single colonies of P. aeruginosa, K. pneumoniae, A. baumannii, S.
aureus, E. faecium, and E. coli were grown on tryptic
soy agar for 24 h at 37 °C in a static incubator and cultured
in tryptic soy broth until log phase to be an optical density (OD)
of 0.2–0.5. The OD was monitored at 600 nm using a 96-well
microplate reader. A single colony of C. difficile was obtained on a brain–heart infusion (BHI) agar plate and
incubated at 37 °C under anaerobic conditions for 48 h. Seed
cultures and larger cultures were obtained using a BHI broth. The
flasks were incubated anaerobically for 48 h at 37 °C and cultured
to mid-log phase (OD600 0.4). The other bacteria were cultured
in the recommended conditions by ATCC.[48] The inhibitors were dissolved in polyethylene glycol 300–H2O (1/1, a final concentration of 1 mg per 100 μL).
This concentration was used as the stock solution for all studies.
Bacterial cultures were treated with serial dilutions of inhibitors
and incubated at 37 °C for 48 h. The MIC was determined by a
96-well plate reader (BioTek Synergy XT, Winooski, VT, USA) at 570
and 600 nm. If necessary, viable bacteria in each well (96-well plate)
were measured via CFUs on a BHI agar plate. The absorbance measurements
were also performed using a BioTek Synergy XT (Winooski, VT, USA)
96-well plate reader at 570 and 600 nm.
Cytotoxicity Assays
Cytotoxicity assays were performed
using Vero monkey kidney (ATCC CCL-81) and HepG2humanhepatoblastoma
cell (ATCC HB-8065) lines. Vero or HepG2 cells were cultured in 75
cm2 flasks and transferred to 96-well cell culture plates
using ATCC-formulated Eagle’s minimum essential medium containing
10% fetal bovine serum and penicillin–streptomycin. Serially
diluted aliquots of each test compound at concentrations ranging from
0.78 to 200 μg/mL were added to the cells. Control compounds
with known toxicity, such as tunicamycin, colistin, or tobramycin,
were included on each plate. The plates were incubated, and cytotoxic
effects were determined via the MTT assay.
Spore Preparation
C. difficile (ATCC 43596) was inoculated
on a BHI agar plate and incubated at
37 °C under anaerobic conditions for 14 days. The spores were
collected from the agar using sterile distilled water and purified
according to the procedures described in the literature.[17] The vegetative forms of C. difficile were killed upon heating at 50 °C for 30 min. The prepared
spores were suspended in sterile distilled waterat 4 °C.
Spore
Viability Testing
A solution of test compound
was added to a suspension containing C. difficile spores (2 × 105 mL–1), and the
mixture was incubated at 37 °C for 24 h. The spore suspension
treated with the test compound was centrifuged (4700g), and the pellet was washed with sterile distilled water, plated
on a BHI agar containing 0.1% sodium taurocholate (a germination agent),
incubated at 37 °C for 48 h under anaerobic conations. The resulting
colonies were counted.
MraY (MurX), WecA, and AglH Assays
Preparations of
the membrane fractions from E. coli, M. smegmatis, and Hydrogenivirga spp. were performed according to the
procedures previously described.[27,33] Procedures
for the purification of MraY and AglH and inhibitory assays using
these phosphotransferases are described in the Supporting Information.
Authors: Kristen L Stoltz; Raymond Erickson; Christopher Staley; Alexa R Weingarden; Erin Romens; Clifford J Steer; Alexander Khoruts; Michael J Sadowsky; Peter I Dosa Journal: J Med Chem Date: 2017-04-12 Impact factor: 7.446
Authors: Yvette H van Beurden; Max Nieuwdorp; Pablo J E J van de Berg; Chris J J Mulder; Abraham Goorhuis Journal: Therap Adv Gastroenterol Date: 2017-02-08 Impact factor: 4.409
Authors: Katsuhiko Mitachi; Rita G Kansal; Kirk E Hevener; Cody D Gillman; Syed M Hussain; Hyun Gi Yun; Gustavo A Miranda-Carboni; Evan S Glazer; William M Clemons; Michio Kurosu Journal: J Med Chem Date: 2020-09-18 Impact factor: 7.446
Authors: Katsuhiko Mitachi; David Mingle; Wendy Effah; Antonio Sánchez-Ruiz; Kirk E Hevener; Ramesh Narayanan; William M Clemons; Francisco Sarabia; Michio Kurosu Journal: Angew Chem Int Ed Engl Date: 2022-06-10 Impact factor: 16.823