Pentamidine, an FDA-approved antiparasitic drug, was recently identified as an outer membrane disrupting synergist that potentiates erythromycin, rifampicin, and novobiocin against Gram-negative bacteria. The same study also described a preliminary structure-activity relationship using commercially available pentamidine analogues. We here report the design, synthesis, and evaluation of a broader panel of bis-amidines inspired by pentamidine. The present study both validates the previously observed synergistic activity reported for pentamidine, while further assessing the capacity for structurally similar bis-amidines to also potentiate Gram-positive specific antibiotics against Gram-negative pathogens. Among the bis-amidines prepared, a number of them were found to exhibit synergistic activity greater than pentamidine. These synergists were shown to effectively potentiate the activity of Gram-positive specific antibiotics against multiple Gram-negative pathogens such as Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Escherichia coli, including polymyxin- and carbapenem-resistant strains.
Pentamidine, an FDA-approved antiparasitic drug, was recently identified as an outer membrane disrupting synergist that potentiates erythromycin, rifampicin, and novobiocin against Gram-negative bacteria. The same study also described a preliminary structure-activity relationship using commercially available pentamidine analogues. We here report the design, synthesis, and evaluation of a broader panel of bis-amidines inspired by pentamidine. The present study both validates the previously observed synergistic activity reported for pentamidine, while further assessing the capacity for structurally similar bis-amidines to also potentiate Gram-positive specific antibiotics against Gram-negative pathogens. Among the bis-amidines prepared, a number of them were found to exhibit synergistic activity greater than pentamidine. These synergists were shown to effectively potentiate the activity of Gram-positive specific antibiotics against multiple Gram-negative pathogens such as Acinetobacter baumannii, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Escherichia coli, including polymyxin- and carbapenem-resistant strains.
The growing
threat of antimicrobial
resistance (AMR) has led to projections that by 2050 the world may
be confronted with as many as 10 million annual AMR-associated deaths.[1] Society is already dealing with the rising tide
posed by this global health challenge: each year, 700,000 people die
due to infections with drug-resistant pathogens.[2] At present, the most critical threats are presented by
Gram-negative pathogens, including Acinetobacter baumannii (carbapenem-resistant), Pseudomonas aeruginosa (carbapenem-resistant), and the Enterobacteriaceae (carbapenem-resistant and ESBL-producing strains), such as Escherichia coli and Klebsiella pneumoniae, according to the World Health Organization (WHO).[3]In treating infections due to Gram-negative bacteria,
there is
an increased interest in strategies aimed at disrupting the outer
membrane (OM) so as to potentiate a number of clinically used antibiotics
that on their own are only effective against Gram-positive bacteria.[4−6] In an elegant approach recently reported by Brown and coworkers,
a panel of 1440 previously approved drugs were screened to identify
compounds capable of disrupting the OM of Gram-negative bacteria.[7] The assay used in the screen was based on findings
that at low temperatures, OM synthesis is altered in E. coli making it more susceptible to vancomycin.[8,9] This led to the hypothesis that compounds that antagonize vancomycin
in E. coli grown at 15 °C would
likely also impact the OM integrity.[7,10] Among the
hits identified using this innovative screen, the small-molecule bis-amidine
pentamidine (1) (Figure ) exhibited the most effective capacity to antagonize
the activity of vancomycin.[7]
Figure 1
Structures
of pentamidine (1) and analogues 2 and 3 previously found to exhibit synergy with
Gram-positive antibiotics against Gram-negative species.[7]
Structures
of pentamidine (1) and analogues 2 and 3 previously found to exhibit synergy with
Gram-positive antibiotics against Gram-negative species.[7]Pentamidine is used clinically
to treat Pneumocystis
jiroveci pneumonia, trypanosomiasis, and leishmaniasis.[11−13] Apart from its antiprotozoal activity, pentamidine is also known
to have moderate antibacterial activity against Gram-positive species.[14,15] Furthermore, pentamidine has also been shown to have anti-cancer
activity by restoring the tumor-suppressing activity of p53, is capable
to bind A/T-rich regions of double-stranded DNA, and can non-specifically
bind and disrupt tRNA secondary structures.[16−19] Unsurprisingly, this broadly
active compound has a high incidence of side effects such as nephrotoxicity,
hypotension, hypoglycaemia, or local reactions to the injection.[11−13] The Brown group’s discovery that pentamidine potentiates
the anti-Gram-negative activity of rifampicin, erythromycin, and novobiocin
further highlights the multifaceted nature of the compound.[7]It is well established that the disruption
of the Gram-negative
OM, for example, with the well-studied polymyxin B nonapeptide (PMBN),
can potentiate the activity of hydrophobic, Gram-positive specific
antibiotics.[7,20] In keeping with these findings,
it is also known that polymyxin-resistance also reduces the synergistic
potential of PMBN.[7,20] In this regard, it is notable
that the synergistic activity of pentamidine in combination with novobiocin,
when evaluated against wild-type and polymyxin-resistant strains of A. baumannii, was observed both in vitro and in vivo.[7]In addition to pentamidine, Brown and co-workers also examined
the synergistic activity of other commercially available bis-amidines
by performing checkerboard assays, from which the fractional inhibitory
concentration index (FICI) was derived, serving as a measure of synergistic
activity.[7,21] These studies highlighted the necessity
of two amidine groups for effective potentiation of Gram-positive
antibiotics against an E. coli indicator
strain.[7] In addition, the linker used to
connect the benzamidine moieties was also found to play a key role
in the determining the activity of the compounds evaluated.[7] Based on these studies, two analogues were identified
as having enhanced synergistic activities relative to pentamidine
(compounds 2 and 3, Figure ). The conclusions drawn from these studies
suggest that increased linker length and hydrophobicity, along with
decreased linker flexibility, contributes to an increase in synergistic
activity for these bis-amidines.[7]Inspired by these findings, we here describe structure–activity
relationship (SAR) studies designed to provide a broad understanding
of the structural features required for potent and selective synergy
by bis-amidines. While the previous study of Brown and coworkers evaluated
the synergistic potential of commercially available bis-amidines,
we here report the design, synthesis, and evaluation of a number of
novel bis-amidines. In addition to screening for synergistic activity,
the new compounds here studied were also assessed for their capacity
to selectively target the Gram-negative OM membrane rather than act
as non-specific membrane disruptors. Our findings serve to both validate
published accounts, while also revealing new, more potent, and selective
bis-amidine-based synergists.
Results and Discussion
Synthesis and Initial Screening
Linear
Linkers
To further explore the correlation between
linker length and synergistic activity, a set of linear pentamidine
analogues was selected. In addition to the previously reported nonamidine
(2) and propamidine (9), we also synthesized
heptamidine (10), octamidine (11), and undecamidine
(12) analogues (Scheme A). Pentamidine (1) was also synthesized
by the same route to allow for comparison with the commercial material
(Supporting Information, Scheme S1), which
subsequently revealed no difference in the synergistic activity of
the in-house prepared and commercial materials (data not shown).
Scheme 1
Synthesis of Pentamidine Analogues Containing Different Linear Spacers
between the Benzamidine Groups
Reagents and conditions:
(a)
4-cyanophenol, NaH, DMF, 80 °C, 1 h (59%-quant.); (b) (i) LHMDS,
THF, 48 h, rt, (ii) HCl, 0 °C to rt, overnight (49%-quant.);
(c) K2CO3, DMF, 100 °C, 5 h (43%); (d)
Na2S·9H2O, DMSO, 115 °C, 1 h (93%);
(e) (i) LHMDS, THF, rt, 48 h; (ii) HCl, rt, overnight (64%); (f) m-CPBA, DCM, 0 °C, 2 h (32%).
Synthesis of Pentamidine Analogues Containing Different Linear Spacers
between the Benzamidine Groups
Reagents and conditions:
(a)
4-cyanophenol, NaH, DMF, 80 °C, 1 h (59%-quant.); (b) (i) LHMDS,
THF, 48 h, rt, (ii) HCl, 0 °C to rt, overnight (49%-quant.);
(c) K2CO3, DMF, 100 °C, 5 h (43%); (d)
Na2S·9H2O, DMSO, 115 °C, 1 h (93%);
(e) (i) LHMDS, THF, rt, 48 h; (ii) HCl, rt, overnight (64%); (f) m-CPBA, DCM, 0 °C, 2 h (32%).As shown in Scheme A, the dibenzonitrile intermediates were prepared from the commercially
available α,ω-dibromo-alkanes via a Williamson
ether synthesis according to literature protocols.[22] Crystallization from ethanol resulted in the pure intermediates 4–8 in good to excellent yields. The transformation
of the nitrile groups into the corresponding amidine is classically
performed via the Pinner reaction followed by treatment
with ammonia.[23−27] However, recent publications have described the same transformation
by the more convenient use of a lithium bis(trimethylsilyl)amide (LHMDS)
solution followed by an acidic quench.[28−31] In the synthesis of pentamidine
we therefore evaluated the treatment of the corresponding bis-nitrile
precursor with LHMDS [1 M in tetrahydrofuran (THF)] followed by a
quench with saturated ethanolic HCl, 4 M HCl in dioxane, or 1 M HCl
(aq) (see Supporting Information, Scheme
S1 and S2). These trial experiments revealed that quenching with 4
M HCl in dioxane resulted in the highest yield, and these conditions
were therefore also applied in the preparation of the bis-amidines 2, 9–12, which were subsequently isolated
in good yields after high-performance liquid chromatography (HPLC)
purification. In addition to probing linker length, we also explored
the impact of heteroatom substitution in the linker. Notably, thioether
analogue 15 has been previously prepared and tested for
antimicrobial activity.[15,32] Thioether 15 was therefore synthesized, as indicated in Scheme B, also providing ready access to the more
hydrophilic sulfone analogue 16 obtained by m-CPBA treatment of 15.The inherent antibacterial
activities of pentamidine (1) and the bis-amidines 2, 3, 9–12, 15, and 16 were first assessed against
an indicator strain E. coli BW25113.
This revealed a trend wherein compounds containing linkers of eight
or more carbons exhibited moderate antibacterial activity with minimum
inhibitory concentration (MIC) values of 50 μg/mL (see Table ). Neither the thioether
linked species 15 or sulfone linked 16 showed
any inherent activity up to the maximum concentration tested (200
μg/mL). Next, the synergistic activity of the compounds was
assessed in combination with both erythromycin and rifampicin using
the same indicator E. coli strain.
Checkerboard assays were performed in which a dilution series of the
synergist was evaluated in combination with the antibiotic of interest,
also serially diluted. The resulting “checkerboard”
or 2-dimensional MIC readout makes it possible to identify the lowest
concentration of both components that results in the most potent synergistic
effect. The highest concentrations tested among the synergists correspond
to their inherent MIC values (or up to 200 μg/mL in the case
where no antibacterial activity was observed). For erythromycin, the
highest concentration tested was 200 μg/mL and for rifampicin
it was 12 μg/mL.
Table 1
Overview of Synergy
with Erythromycin
against E. coli BW25113 and Hemolysis
Data
Synergy defined as FICI ≤
0.5. See Supporting Tables S1 and S2 for full data used in calculating the FICIs
with erythromycin and rifampicin, respectively.
Hemolytic activity of all compounds
after 20 h of incubation at 200 μg/mL. Values below 10% were
defined as non-hemolytic.[33]
Synergy defined as FICI ≤
0.5. See Supporting Tables S1 and S2 for full data used in calculating the FICIs
with erythromycin and rifampicin, respectively.Hemolytic activity of all compounds
after 20 h of incubation at 200 μg/mL. Values below 10% were
defined as non-hemolytic.[33]In general, a trend was observed
wherein bis-amidines with longer
linker lengths showed a great capacity to potentiate the activity
of erythromycin (Table ). Compared with pentamidine (FICI 0.500), nonamidine (2), and heptamidine (10) were found to be the most effective
synergists with FICI values of 0.094 and 0.125, respectively, while
the shorter propamidine (9) exhibited activity on par
with pentamidine (Figure ). The synergistic activities observed when the same panel
of bis-amidines was evaluated with rifampicin corroborates the findings
with erythromycin (Table and Supporting Information, Figure
S2). These findings highlight the importance of linker length and
hydrophobicity for synergistic activity. All analogues containing
linkers greater than five carbon atoms demonstrated more potent synergy
than observed for pentamidine. By comparison, propamidine (9), containing a three carbon spacer and thioether 15 (isosteric to pentamidine) exhibited synergistic activities comparable
to pentamidine. It is also interesting to note that the introduction
of the more polar sulfone-linker as in 16 led a complete
loss of synergistic activity (Table and Supporting Information, Figures S1 and S2, and Tables S1 and S2).
Figure 2
Representative checkerboard
assays for pentamidine (1), propamidine (9), nonamidine (2), and
heptamidine (10) in combination with erythromycin vsE. coli BW25113. In each
case, the bounded box in the checkerboard assays indicates the combination
of compound and antibiotic resulting in the lowest FICI (see Table ). OD600 values were measured using a plate reader and transformed to a gradient:
purple represents growth, white represents no growth. An overview
of all checkerboard assays with erythromycin can be found in Supporting Information, Figure S1.
Representative checkerboard
assays for pentamidine (1), propamidine (9), nonamidine (2), and
heptamidine (10) in combination with erythromycin vsE. coli BW25113. In each
case, the bounded box in the checkerboard assays indicates the combination
of compound and antibiotic resulting in the lowest FICI (see Table ). OD600 values were measured using a plate reader and transformed to a gradient:
purple represents growth, white represents no growth. An overview
of all checkerboard assays with erythromycin can be found in Supporting Information, Figure S1.Examination of the effect of these bis-amidines on red blood
cells
revealed another feature that correlates with linker length. Specifically,
the enhanced antimicrobial activity and synergistic potential in combination
with erythromycin observed for analogues containing longer linkers
is accompanied by an increase in hemolytic activity (Table and Supporting Information, Figures S17 and S18 and Table S17). While propamidine
(9) and pentamidine (1) have little inherent
antibacterial activity (MIC of 200 μg/mL or higher) and are
moderate synergists with erythromycin (FICI of 0.500), they are also
non-hemolytic (erythrocytes treated with compounds at 200 μg/mL
for 20 h at 37 °C, non-hemolytic defined as <10%[33]). By comparison, the slightly longer heptamidine
(10) has an inherent antimicrobial activity (MIC 200
μg/mL) along with enhanced synergistic activity with erythromycin
(FICI ≤ 0.125) but also a slight increase in hemolytic activity
to 9.2%. However, the longer octamidine (11), nonamidine
(2), and undecamidine (12) exhibit very
significant levels of hemolysis (16–87%), suggesting that both
the inherent antimicrobial activity (MIC 50 μg/mL) and potent
synergistic activity in combination with erythromycin (FICI ≤
0.094–0.156) of these analogues are driven by a general membrane
disruption mechanism and not a selective disruption of the Gram-negative
OM. Based on these findings, it appears that the “tipping point”
associated with the desirable synergistic effects versus the unwanted hemolytic activity appears to be for C7-spaced bis-amidine
analogue heptamidine (10). These findings served to inform
the design of the next series of analogues.
Linkers with Reduced Flexibility
Building from our
initial findings with the linear bis-amidines, we next examined the
effect of reducing the rotational flexibility of the linker. In the
Brown group’s earlier study, it was noted that phenyl-substituted
bis-amidine 3 (Figure ) was an extremely effective synergist, an effect that
was attributed in part to its decreased molecular flexibility.[7] To this end, we prepared a series of bis-amidines
(Scheme , compounds 21–24) that incorporate linkers comprising different
planar, aromatic motifs as a means of even further restricting flexibility.
For purposes of comparison, we also prepared compound 3 (Supporting Information, Scheme S3) and
confirmed its synergistic activity (Table , Supporting Information, Figures S1 and S2). Notable, however, was the finding that compound 3 also exhibits significant hemolytic activity (above 10%[33]) (See Table and Supporting Information, Figure S18 and Table S17) suggesting that impressive synergistic
activity associated with the compound is not selective for the Gram-negative
OM and is due instead to general membrane disruption. The synthetic
route used to access bis-amidines 21–24 is shown
in Scheme and was
based largely on the published preparation of these and similar compounds
previously evaluated as anti-parasitic agents.[22,34−39] The meta-oriented linker in compound 22 most closely
mimics the 5-carbon spacer found in pentamidine, while analogues 21 and 23 differ slightly due to the ortho- and
para-orientations of the benzene core. In the case of compound 24, a 2,7-disubstituted naphthalene motif was envisioned to
mimic the 7-carbon spacer found in heptamidine (10).
The synthesis of compounds 21–24 started from
the corresponding commercially available dibromo-xylenes or 2,7-bis(bromomethyl)naphthalene,
which were transformed into the corresponding bis-nitriles 17–20 by treatment with 4-cyanophenol and NaH in dimethylformamide (DMF)
at 80 °C. In this case, recrystallization of the intermediates 17, 19, and 20 from ethanol was
not successful. However, based on an acceptable purity (as assessed
by NMR), the crude bis-nitriles 19 and 20 could be used directly without a need for further purification,
while bis-nitrile 17 was purified using column chromatography.
Transformation into the corresponding bis-amidines was in turn performed
by treatment with LHMDS[34] followed by acidic
quench with 4 M HCl in dioxane to provide compounds 21–24 in acceptable yields after HPLC purification.
Scheme 2
Synthesis of Bis-Amidines
Containing Rigid Aromatic Spacers
Reagents and conditions:
(a)
4-cyanophenol, NaH, DMF, 80 °C, 1 h (79%-quant.); (b) (i) LHMDS,
THF, 48 h; (ii) HCl, 0 °C to rt, overnight (19–83%).
Synthesis of Bis-Amidines
Containing Rigid Aromatic Spacers
Reagents and conditions:
(a)
4-cyanophenol, NaH, DMF, 80 °C, 1 h (79%-quant.); (b) (i) LHMDS,
THF, 48 h; (ii) HCl, 0 °C to rt, overnight (19–83%).Evaluation of the inherent antimicrobial activity
of compounds 21–24 as well as their ability to
synergize with erythromycin
revealed 22 and 24 to be the most effective
of these four of compounds (FICI of ≤0.094 with erythromycin)
(Figure and Table ). o-Xylene analogue 21 also exhibited enhanced synergistic
activity relative to pentamidine (≤0.125 vs 0.500) while p-xylene analogue 23 showed
less activity (FICI ≤ 0.313). Interestingly, while none of
compounds 21–24 showed any inherent antibacterial
activity up to 200 μg/mL, the 2,7-naphthalene linked analogue 24 was found to exhibit significant hemolytic activity (75%)
(see Table ). These
findings are in line with previous studies in which compound 24 was evaluated as an anti-protozoal, where it was also found
to exhibit significant toxicity against a rat L6 muscle cell line.[38] By comparison, compounds 21 and 22 were found to be non-hemolytic and demonstrate potent synergy
when combined with erythromycin with FICI values of ≤0.125
and ≤0.094, respectively (Table ). Similarly, 21 and 22 were
also found to significantly potentiate the activity of rifampicin
against the same E. coli indicator
strain with FICI values of ≤0.094 and ≤0.188, respectively
(Table ). These findings
support the hypothesis that reduced linker flexibility is beneficial
for synergistic activity and also reveal the importance of the orientation
of the benzamidines on the aromatic nucleus. This is most clearly
demonstrated by the potent synergy exhibited by the ortho- and meta-xylene analogues 21 and 22 (FICI ≤ 0.094–0.188) in contrast to the much
less active para-xylene linked 23 (FICI
≤ 0.313–0.375).
Figure 3
Checkerboard assays for compounds 21–24 in
combination with erythromycin vsE.
coli BW25113. In each case, the bounded box in the
checkerboard assays indicates the combination of compound and antibiotic
resulting in the lowest FICI (see Table ). OD600 values were measured
using a plate reader and transformed into a gradient: purple represents
growth, white represents no growth. An overview of all checkerboard
assays with erythromycin can be found in Supporting Information, Figure S1.
Checkerboard assays for compounds 21–24 in
combination with erythromycin vsE.
coli BW25113. In each case, the bounded box in the
checkerboard assays indicates the combination of compound and antibiotic
resulting in the lowest FICI (see Table ). OD600 values were measured
using a plate reader and transformed into a gradient: purple represents
growth, white represents no growth. An overview of all checkerboard
assays with erythromycin can be found in Supporting Information, Figure S1.
Altering the Position of the Amidine Moiety
The rigidity
of the xylene-based linkers described above not only affects the spacing
but also the positioning of the amidine groups. In the case of pentamidine
(1) and compounds 21–23, the amidine
moieties are positioned para relative to the linker.
We, therefore, next prepared a series of analogues wherein the positioning
of the amidine groups was shifted to either the meta- or ortho-positions (Scheme ). While the meta-amidine
analogues 1b, 21b–23b are known in
the literature,[24,35,38−41]ortho-amidine analogues 1c, 21c–23c have not been previously described. The synthesis
of the meta-amidine analogues was performed following
the same protocol employed for the preparation of the corresponding para-amidines but using 3-cyanophenol in place of 4-cyanophenol
(Scheme ). For the
preparation of the ortho-amidine analogues, the intermediate
bis-nitriles were prepared in an analogous fashion, however, conversion
to the product bis-amidines required a different set of conditions.
Unlike the route used in the preparation of the para- and meta-bis-amidines, treatment of the ortho-bis-nitrile intermediates 29–32 with LHMDS failed to yield the expected amidine product. For this
reason, an alternative, previously reported three-step procedure for
the conversion of nitriles to amidines, was instead employed.[42] In doing so, the nitrile is first converted
to the corresponding N-hydroxyamidine by treatment
with hydroxylamine hydrochloride. The N-hydroxy group
is then acetylated with Ac2O followed by reduction to the
amidine product using zinc powder (Scheme ). After HPLC purification, the ortho-bis-amidines (1c, 21c–23c) were
obtained in yields suitable for subsequent evaluation.
Scheme 3
Synthesis
of bis-amidine analogues 1b, 21b–23b and 1c, 21c–23c
Reagents
and conditions: (a)
3-cyanophenol, NaH, DMF, 80 °C, 1 h (63%-quant.); (b) (i) LHMDS,
THF, 48 h, (ii) HCl, 0 °C to rt, overnight (72%-quant.); (c)
2-cyanophenol, NaH, DMF, 80 °C, 1 h (83–99%); (d) (i)
DIPEA, NH2OH·HCl, EtOH, 85 °C, 6 h; (ii) Ac2O, AcOH, rt, 4 h; (iii) Zn powder, AcOH, 35 °C, 6 h (12–48%).
Synthesis
of bis-amidine analogues 1b, 21b–23b and 1c, 21c–23c
Reagents
and conditions: (a)
3-cyanophenol, NaH, DMF, 80 °C, 1 h (63%-quant.); (b) (i) LHMDS,
THF, 48 h, (ii) HCl, 0 °C to rt, overnight (72%-quant.); (c)
2-cyanophenol, NaH, DMF, 80 °C, 1 h (83–99%); (d) (i)
DIPEA, NH2OH·HCl, EtOH, 85 °C, 6 h; (ii) Ac2O, AcOH, rt, 4 h; (iii) Zn powder, AcOH, 35 °C, 6 h (12–48%).As for pentamidine (1) and the other para-bis-amidines 21–23, no inherent
antimicrobial
activity or hemolysis was observed for the meta-substituted analogues 1b, 21b–23b or the ortho-substitute analogues 1c, 21c–23c (Table ). Assessment of synergy with erythromycin
showed that the meta-bis-amidines maintain a reasonable
degree of synergistic activity (Figure ) while the ortho-bis-amidines show
no such ability (Table ).
Figure 4
Checkerboard assays for compounds 1b, 21b–23b in combination with erythromycin vsE. coli BW25113. In each case, the bounded box in
the checkerboard assays indicates the combination of the compound
and antibiotic resulting in the lowest FICI (see Table ). OD600 values were
measured using a plate reader and transformed into a gradient: purple
represents growth, white represents no growth. An overview of all
checkerboard assays with erythromycin can be found in Supporting Information, Figure S1.
Checkerboard assays for compounds 1b, 21b–23b in combination with erythromycin vsE. coli BW25113. In each case, the bounded box in
the checkerboard assays indicates the combination of the compound
and antibiotic resulting in the lowest FICI (see Table ). OD600 values were
measured using a plate reader and transformed into a gradient: purple
represents growth, white represents no growth. An overview of all
checkerboard assays with erythromycin can be found in Supporting Information, Figure S1.In general, the meta-orientated bis-amidines
are
less effective synergists than the corresponding para-oriented compounds,
a trend also observed in synergy studies with rifampicin (Table ). An exception to
this was observed for compounds 23 and 23b both containing the p-xylene linker. In this case,
the placement of the amidine groups at the meta-position relative
to the linker results in a slight decrease in FICI from 0.313 for
compound 23 to 0.250 for 23b when tested
in combination with erythromycin. An even more pronounced potentiation
effect was seen when these compounds where evaluated with rifampicin.
In this case, compound 23 was found to have an FICI value
of 0.375 while for 23b, the FICI value calculated was
0.156, making it one of the most potent, non-hemolytic, rifampicin
synergists identified (Table ). Collectively, these findings indicate that both the geometry
of the linker and the positioning of the amidines in the benzamidine
moieties are interrelated structural features that play a key role
in dictating optimal synergistic activity.
Increasing Linker Hydrophobicity
As described above,
bis-amidines with more hydrophobic linkers typically show enhanced
synergistic activity but often at the cost of increased hemolysis.
In this light, compounds 21 and 22 were
deemed to be particularly interesting given that they exhibit potent
synergistic activity with both erythromycin and rifampicin while displaying
no appreciable hemolytic activity. To examine the possibility of further
enhancing these compounds, we next prepared analogues wherein an additional
phenyl group, as for compound 3, was added as a substituent
to the aromatic linkers in both 21 and 22 to give analogues 38 and 44 (Scheme ). The synthetic
route used also provided ready access to brominated intermediates 35 and 41. Given the hydrophobic character of
halogen atoms,[43] we opted to also convert
these intermediates to the corresponding bis-amidines 37 and 43. The synthesis of meta-linked analogues 37 and 38 started with the reduction of dimethyl
5-bromoisophthalate to give diol 33.[44] An Appel reaction was then applied to transform the diol
into tribromide 34,[45] followed
by reaction with 4-cyano phenol to yield bis-nitrile 35.[22] A portion of 35 was subsequently
used in a Suzuki coupling employing phenylboronic acid, resulting
in intermediate 36.[46−48] Both 35 and 36 were then converted to the corresponding bis-amidines
by treatment with LHMDS followed by HCl quench and HPLC purification
to give 37 and 38. The preparation of 43 and 44 followed a similar synthetic strategy
but started with the reduction of 4-bromophthalic anhydride using
lithium aluminum hydride and ZnCl2.[49] The resulting diol 39 was cleanly converted
to tribromide 40, which was subsequently transformed
into the brominated bis-nitrile intermediate 41. A portion
of 41 was then transformed into intermediate 42 using the same Suzuki conditions applied in the previous preparation
of 36.[46−48] Notably, while bis-nitrile 42 was readily
transformed into the desired bis-amidine 44 using the
LHMDS protocol, when the same conditions were applied to 41 an unexpected dehalogenation occurred. As an alternative, the same
three-step process, described above for the preparation of 21b–23b, was successfully applied to convert the bis-nitrile to the desired
bis-amidine 43.[42]
Scheme 4
Synthesis
of (A) Meta-Linked or (B) Ortho-Linked Bis-Amidines Containing
Bromo (37, 43) or Phenyl Substitution (38, 44) on the Central Aromatic Core
Reagents and conditions: (a)
(i) DIBALH, DCM, 0 °C, 1 h; (ii) Rochelle salt (quench), rt,
overnight (96%); (b) PPh3, CBr4, DCM, rt, 2
h (55–74%); (c) 4-cyanophenol, NaH, DMF, 80 °C, 1 h (87–99%);
(d) phenylboronic acid, Pd(dppf)Cl2·DCM, THF/Na2CO3, 65 °C, 8–18 h (8–80%);
(e) (i) LHMDS, THF, rt, 48 h; (ii) HCl (quench), 0 °C–rt,
overnight (17–75%); (f) (i) LAH, ZnCl2, THF, rt,
6 h; (ii) Rochelle salt (quench), rt, overnight (95%); (g) (i) DIPEA,
NH2OH·HCl, EtOH, 85 °C, 6 h; (ii) Ac2O, AcOH, rt, 4 h; (iii) Zn powder, AcOH, 35 °C, 6 h (7%).
Synthesis
of (A) Meta-Linked or (B) Ortho-Linked Bis-Amidines Containing
Bromo (37, 43) or Phenyl Substitution (38, 44) on the Central Aromatic Core
Reagents and conditions: (a)
(i) DIBALH, DCM, 0 °C, 1 h; (ii) Rochelle salt (quench), rt,
overnight (96%); (b) PPh3, CBr4, DCM, rt, 2
h (55–74%); (c) 4-cyanophenol, NaH, DMF, 80 °C, 1 h (87–99%);
(d) phenylboronic acid, Pd(dppf)Cl2·DCM, THF/Na2CO3, 65 °C, 8–18 h (8–80%);
(e) (i) LHMDS, THF, rt, 48 h; (ii) HCl (quench), 0 °C–rt,
overnight (17–75%); (f) (i) LAH, ZnCl2, THF, rt,
6 h; (ii) Rochelle salt (quench), rt, overnight (95%); (g) (i) DIPEA,
NH2OH·HCl, EtOH, 85 °C, 6 h; (ii) Ac2O, AcOH, rt, 4 h; (iii) Zn powder, AcOH, 35 °C, 6 h (7%).Compounds 37, 38, 43, and 44 were found to show no significant
inherent antimicrobial
activity when tested against E. coli BW25113 (Table ).
As expected, the introduction of the hydrophobic side chains improved
the synergistic activity with FICI values ranging from 0.047 to 0.094
(Figure and Table ). Unfortunately,
however, and not entirely unexpectedly, the increased hydrophobicity
of these analogues was also found to result in a severe increase in
hemolytic activity (Table ) indicating that the enhanced synergistic activity observed
is likely due to non-specific membrane disruption.
Figure 5
Checkerboard assays for
compounds 37, 38, 43, and 44 in combination with erythromycin vsE. coli BW25113. In each
case, the bounded box in the checkerboard assays indicates the combination
of the compound and antibiotic resulting in the lowest FICI (see Table ). OD600 values were measured using a plate reader and transformed into a
gradient: purple represents growth, white represents no growth. An
overview of all checkerboard assays with erythromycin can be found
in the Supporting Information, Figure S1.
Checkerboard assays for
compounds 37, 38, 43, and 44 in combination with erythromycin vsE. coli BW25113. In each
case, the bounded box in the checkerboard assays indicates the combination
of the compound and antibiotic resulting in the lowest FICI (see Table ). OD600 values were measured using a plate reader and transformed into a
gradient: purple represents growth, white represents no growth. An
overview of all checkerboard assays with erythromycin can be found
in the Supporting Information, Figure S1.
Exploring the Synergistic Range
Erythromycin, rifampicin,
novobiocin, and vancomycin are typically used to treat Gram-positive
infections.[50−55] However, when combined with OM disrupting agents, these antibiotics
can also display efficacy against Gram-negative bacteria.[6,20] The Brown group’s recent study with pentamidine showed that
erythromycin, rifampicin, and novobiocin were most effectively potentiated
by this bis-amidine.[7] With this in mind,
we next investigated the broader synergy of the most promising compounds
identified in our present study, namely, compounds 21, 22, and 23b. As noted above, these three
compounds were all found to be more active than pentamidine in potentiating
the activity of erythromycin and rifampicin against an indictor E. coli stain while showing no hemolytic activity.
To this end, compounds 21, 22, and 23b were evaluated against an expanded panel of organisms,
including several E. coli strains (including
carbapenem- and polymyxin-resistant strains) and ATCC strains of A. baumannii, K. pneumoniae, and P. aeruginosa. In addition,
the well-studied OM disruptor PMBN and pentamidine itself were taken
along as benchmarks in the expanded assessment of compounds 21, 22, and 23b.
Synergy with Novobiocin
and Vancomycin
Building from
the synergy studies with erythromycin and rifampicin described above,
compounds 21, 22, and 23b were
next tested for the ability to potentiate novobiocin and vancomycin,
along with pentamidine (1) and PMBN (Figure and Supporting Information Figures S3 and S4). In agreement with previous
studies, novobiocin and vancomycin showed no antimicrobial activity
against the indicator E. coli BW25113
strain at the highest concentration tested of 200 μg/mL.[7,56] Checkerboard assays with compounds 21, 22, and 23b in combination with novobiocin revealed the
compounds to be superior synergists compared to pentamidine (Table , Figure ), a finding in line with the
results obtained when the same bis-amidines were evaluated with erythromycin
and rifampicin. In general, PMBN was found to be a more potent synergist
than the bis-amidines with the exception of compound 22 in combination with erythromycin which resulted in very effective
growth prevention of the E. coli indicator
strain. In line with expectation, when tested in combination with
vancomycin, none of the bis-amidines showed any synergistic activity,
while PMBN maintained a potent effect (Table ). These findings are in line with previously
reported observations in which pentamidine was found not to synergize
with vancomycin.[7]
Figure 6
Checkerboard assays of
compounds pentamidine (1), 21, 22, and 23b in combination with
(A) rifampicin and (B) novobiocin against E. coli BW25113. In each case, the bounded box in the checkerboard assays
indicates the combination of the compound and antibiotic resulting
in the lowest FICI (see Table ). OD600 values were measured using a plate reader
and transformed into a gradient: purple represents growth, white represents
no growth. The poor aqueous solubility of novobiocin results in the
background signal observed in the OD600 read-out at when
tested at concentrations ≥100 μg/mL. An overview of all
checkerboard assays with rifampicin, novobiocin, and vancomycin can
be found in Supporting Information, Figures
S2–S4.
Table 2
FICI Values of Pentamidine
(1), 21, 22, 23b, and
PMBN against E. coli BW25113 in Combination
with Gram-Positive-Specific Antibiotics Rifampicin, Novobiocin, and
Vancomycina
erythromycin
rifampicin
novobiocin
vancomycin
pentamidine (1)
0.500
0.375
≤0.281
>0.5b
21
≤0.125
≤0.094
≤0.125
>0.5b
22
≤0.094
≤0.188
≤0.078
>0.5b
23b
≤0.250
≤0.156
≤0.188
>0.5b
PMBN
≤0.125
≤0.039
≤0.047
≤0.156
MIC and minimal synergistic concentrations
(MSCs) data can be found in Supporting Information, Tables S1–S4.
Synergy defined as an FICI ≤
0.5.[21]
Checkerboard assays of
compounds pentamidine (1), 21, 22, and 23b in combination with
(A) rifampicin and (B) novobiocin against E. coli BW25113. In each case, the bounded box in the checkerboard assays
indicates the combination of the compound and antibiotic resulting
in the lowest FICI (see Table ). OD600 values were measured using a plate reader
and transformed into a gradient: purple represents growth, white represents
no growth. The poor aqueous solubility of novobiocin results in the
background signal observed in the OD600 read-out at when
tested at concentrations ≥100 μg/mL. An overview of all
checkerboard assays with rifampicin, novobiocin, and vancomycin can
be found in Supporting Information, Figures
S2–S4.MIC and minimal synergistic concentrations
(MSCs) data can be found in Supporting Information, Tables S1–S4.Synergy defined as an FICI ≤
0.5.[21]
Synergy against Other E. coliStrains
The next phase of our investigation
involved assessing the synergistic activity of the most promising
compounds identified against an expanded panel of E.
coli strains. For these screens, we opted to focus
on rifampicin as the companion antibiotic given that it is bactericidal
while erythromycin is considered to be bacteriostatic.[11,57] In our initial screens, a more clear-cut distinction of growth versus no growth was indeed observed for rifampicin, possibly
due to its bactericidal nature (see Figures and 6A). Furthermore,
given that the MIC of rifampicin is significantly lower against the
Gram-negative strains used versus the MICs of erythromycin
or novobiocin, potential solubility issues at the highest antibiotic
concentrations tested were not a problem.In selecting an expanded
panel of E. coli strains, we sought
to examine a variety of features ranging from the OM composition to
resistance profile. In the case of E. coli, the structure of the lipopolysaccharide (LPS) layer is known to
affect their susceptibility to antibiotics[58] and we therefore reasoned that it could also play a role in the
synergistic activity of compounds targeting the OM. This was seen
as particularly relevant for the pentamidine analogues investigated
here, given that previous studies have suggested that pentamidine
interacts with lipid A.[7] With this in mind, E. coli ATCC25922 (smooth LPS) and E. coli W3110 (rough LPS) were selected, along with
the indicator lab strain E. coli BW25113
also known to possess a rough LPS layer.[59−61] Additionally,
a clinical isolate E. coli 552060.1
was included, which, like most clinical isolates, has a smooth LPS
layer.[58,62] The inherent antimicrobial activity of rifampicin,
pentamidine (1), compounds 21, 22, 23b, and PMBN was first established against these E. coli strains (Supporting Information, Figures S5–S7 and Tables S5–S7). In keeping with
our initial checkerboard assays with rifampicin and the E. coli BW25113 strain (Table ), compound 21 in nearly all
cases showed the lowest FICI values among the bis-amidines evaluated
against the expanded E. coli panel
(Figure A and Table ). In general, the
bis-amidines tested all showed effective synergy with little difference
observed for the rough or smooth LPS strains.
Figure 7
Checkerboard assays of
compounds pentamidine (1), 21, 22, and 23b in combination with
rifampicin vs (A) E. coli ATCC25922, (B) E. coli EQASmcr-1,
and (C) E. coli RC0089. In each case,
the bounded box in the checkerboard assays indicates the combination
of the compound and antibiotic resulting in the lowest FICI (see Table ). OD600 values were measured using a plate reader and transformed into a
gradient: purple represents growth, white represents no growth. An
overview of all checkerboard assays with rifampicin with the E. coli strains can be found in Supporting Information, Figures S5–S13.
Table 3
FICI Values of Pentamidine (1), 21, 22, 23b, and
PMBN in Combination with Rifampicin against Different E. coli Strains Including Polymyxin- and Carbapenem-Resistant
Strainsa
strain
pentamidine (1)
21
22
23b
PMBN
wild-type
BW25113
0.375
≤0.094
≤0.188
≤0.156
≤0.039
ATCC25922
0.313
≤0.125
0.094
0.156
≤0.047
W3110
≤0.188
≤0.188
0.313
≤0.188
≤0.031
552060.1
0.375
≤0.094
0.250
≤0.188
≤0.047
polymyxin-resistant
BW25113 mcr-1
≤0.250
≤0.094
≤0.156
≤0.188
≤0.156
mcr-1
≤0.188
≤0.188
≤0.188
≤0.188
≤0.094
EQASmcr-1
≤0.250
≤0.125
0.188
≤0.188
≤0.125
EQASmcr-2
0.375
≤0.125
0.313
≤0.125
≤0.156
EQASmcr-3
≤0.188
≤0.125
≤0.188
≤0.188
≤0.094
carbapenem-resistant
RC0089
≤0.375
≤0.250
≤0.156
≤0.375
≤0.188
MIC and MSCs data
can be found in Supporting Information,
Table S2, S5–S13.
Checkerboard assays of
compounds pentamidine (1), 21, 22, and 23b in combination with
rifampicin vs (A) E. coli ATCC25922, (B) E. coli EQASmcr-1,
and (C) E. coli RC0089. In each case,
the bounded box in the checkerboard assays indicates the combination
of the compound and antibiotic resulting in the lowest FICI (see Table ). OD600 values were measured using a plate reader and transformed into a
gradient: purple represents growth, white represents no growth. An
overview of all checkerboard assays with rifampicin with the E. coli strains can be found in Supporting Information, Figures S5–S13.MIC and MSCs data
can be found in Supporting Information,
Table S2, S5–S13.The expanded screening was continued with E. coli bearing mcr-1, mcr-2, and mcr-3 genotypes known to confer polymyxin resistance. For
this purpose, a lab strain E. coli BW25113
mcr-1, transformed with the pGDP2 plasmid, was also included to directly
assess the effect of the phosphoethanolamine transferase responsible
for lipid A modification.[63−65] The bis-amidines displayed synergy
with rifampicin against all mcr-positive strains
evaluated (Figure B, Table , Supporting Information, Figures S8–S12,
and Tables S8–S12). Again, in nearly all cases, compound 21 gave the lowest FICI values among the bis-amidines evaluated,
with synergy comparable to that of PMBN, which was found to be generally
less effective against mcr-positive strains than
non-mcr strains (Table ).In addition, carbapenem-resistant E. coli RC0089, a clinical isolate producing New
Delhi β-lactamase
1 (NDM-1), was also evaluated to assess whether this resistance mechanism
affected the synergistic activity of the bis-amidines here studied.
Notably, the MIC of rifampicin was significantly elevated against
this strain (MIC of >192 μg/mL, see Supporting Information, Figure S13 and Table S13). While the bis-amidines
were again found to synergize with rifampicin, the FICI values calculated
were elevated, with the exception of compound 22 (Figure C and Table ). Interestingly, this strain
also resulted in an increased FICI for PMBN.
Synergy against A. baumannii, K. pneumoniae, and P. aeruginosa
In addition
to studying the synergistic activity of the
selected bis-amidines against the E. coli strains described above, we also investigated their capacity to
potentiate the activity of rifampicin against the selected strains
of A. baumannii, K.
pneumoniae, and P. aeruginosa (Figure , Table ). As for the E. coli strains, the inherent antimicrobial activities
of rifampicin, pentamidine (1), compounds 21, 22, 23b, and PMBN were first established
against each strain (Supporting Information, Tables S14–S16). Full checkerboard assays with the A. baumannii and K. pneumoniae strains tested showed the bis-amidines and PMBN to be effective
synergists. In general, compounds 21, 22, and 23b were found to be more potent than pentamidine
(1), while PMBN was found to be an even more effective
synergist. Among the bis-amidines tested, compound 22 displayed the most effective potentiation of rifampicin. Interestingly,
when tested against P. aeruginosa,
the FICIs determined for pentamidine and compounds 21, 22, and 23b were significantly elevated
while PMBN maintained potent synergistic activity.
Figure 8
Checkerboard assays of
pentamidine (1), 21, 22, and 23b in combination with rifampicin
and vs (A) A. baumannii ATCC17978 and (B) K. pneumoniae ATCC13883.
In each case, the bounded box in the checkerboard assays indicates
the combination of compound and antibiotic resulting in the lowest
FICI (see Table ).
OD600 values were measured using a plate reader and transformed
into a gradient: purple represents growth, white represents no growth.
An overview of all checkerboard assays with rifampicin with the E. coli strains can be found in Supporting Information, Figures S14–S16.
Table 4
FICI Values of Pentamidine (1), 21, 22, 23b, and
PMBN in Combination with Rifampicin against Different Gram-Negative
Pathogensa
strain
pentamidine (1)
21
22
23b
PMBN
A. baumannii ATCC17978
≤0.125
≤0.094
≤0.094
≤0.094
≤0.023
K. pneumoniae ATCC13883
≤0.125
≤0.094
≤0.078
≤0.125
≤0.070
P. aeruginosa ATCC27853
≤0.500
≤0.313
≤0.250
≤0.375
0.031
MIC and
MSCs data can be found in Supporting Information, Tables S14–S16.
Checkerboard assays of
pentamidine (1), 21, 22, and 23b in combination with rifampicin
and vs (A) A. baumannii ATCC17978 and (B) K. pneumoniae ATCC13883.
In each case, the bounded box in the checkerboard assays indicates
the combination of compound and antibiotic resulting in the lowest
FICI (see Table ).
OD600 values were measured using a plate reader and transformed
into a gradient: purple represents growth, white represents no growth.
An overview of all checkerboard assays with rifampicin with the E. coli strains can be found in Supporting Information, Figures S14–S16.MIC and
MSCs data can be found in Supporting Information, Tables S14–S16.
Mechanistic Studies
To characterize the mechanism of
action of the bis-amidines here studied, we next investigated the
capacity of the most active compounds to disrupt the Gram-negative
OM. This line of investigation was based in part on the previously
noted interaction of pentamidine with lipid A and also on the knowledge
that the potentiation of antibiotics like erythromycin, rifampicin,
and novobiocin generally relies on OM disruption.[7,20,66] To this end, we employed an established
assay relying on the fluorescent properties of N-phenyl-napthalen-1-amine
(NPN) allowing for the real-time monitoring and quantification of
OM disruption.[67] In the presence of intact
bacterial cells, NPN exhibits relatively low levels of fluorescence.
However, in the event that the OM is disrupted, NPN can gain entry
to the phospholipid layer resulting in a detectable increase in fluorescence
that can, in turn, be measured.[67] For this
assay, we selected compounds 21 and 22 based
on their consistently potent activity in the various synergy assays
described above. The bacterial strain used was E. coli BW25113 and pentamidine (1) and PMBN were taken along
as benchmarks. As illustrated in Figure , a clear, dose-dependent increase in the
fluorescent signal is observed for both 21 and 22, indicating effective OM disruption. In general, both compounds
appear to outperform pentamidine in their ability to disrupt the OM
with compound 22 also exhibiting a stronger effect than
PMBN (see Supporting Information, Figure
S19 for NPN fluorescence at higher concentrations of bis-amidines
and PMBN).
Figure 9
OM permeabilization assay of pentamidine (1), compounds 21, 22, and PMBN with E. coli BW25113 using N-phenyl-1-naphthylamine (NPN) as
a fluorescent probe. The read-out was performed after 60 min of incubation
using a plate reader with λex 355 nm and λem 420 nm. The NPN uptake values shown are relative to the
uptake signal obtained upon treating the cells with 100 μg/mL
colistin as previously reported.[68] All
values corrected for the background signal of the negative control. Error bars represent the standard
deviation based on n = 3 technical replicates.
OM permeabilization assay of pentamidine (1), compounds 21, 22, and PMBN with E. coli BW25113 using N-phenyl-1-naphthylamine (NPN) as
a fluorescent probe. The read-out was performed after 60 min of incubation
using a plate reader with λex 355 nm and λem 420 nm. The NPN uptake values shown are relative to the
uptake signal obtained upon treating the cells with 100 μg/mL
colistin as previously reported.[68] All
values corrected for the background signal of the negative control. Error bars represent the standard
deviation based on n = 3 technical replicates.
Conclusions
We here describe SAR
studies aimed at delivering new insights into
the capacity for small-molecule bis-amidines to potentiate the activity
of Gram-positive specific antibiotics against Gram-negative bacteria.
Inspired by the finding that anti-parasitic drug pentamidine disrupts
the Gram-negative OM to synergize with antibiotics like erythromycin,
rifampicin, and novobiocin, we prepared a number of structurally similar
bis-amidines and characterized their synergistic potential with the
same antibiotics. Our studies confirm that the length, rigidity, and
hydrophobicity of the linker unit present in these bis-amidines play
an important role in determining their ability to potentiate Gram-positive
specific antibiotics.[7] Also of note, however,
is the finding that the potent synergy exhibited by bis-amidines containing
long, hydrophobic linkers is likely driven by nonspecific membrane
disruption as indicated by the strong hemolytic activity associated
with these analogues. Further assessment of the linker motif also
revealed that, in general, a single aromatic ring provides a desirable
balance of enhanced synergistic activity relative to pentamidine,
without introducing hemolytic activity. Further examination of the
relative positioning of the benzamidine groups on the aromatic linker
and as well as the ortho-, meta-, and para-geometry of the amidine
moieties themselves identified compounds 21, 22, and 23b as most promising. These compounds were found
to consistently outperform pentamidine in their ability to potentiate
the activity of erythromycin, rifampicin, and novobiocin against a
number of E. coli strains including
polymyxin-resistant and carbapenem-resistant variants. Additional
screening showed that among the bis-amidines here studied, compounds 21, 22, and 23b maintain their superior
synergistic activity against other Gram-negative pathogens including A. baumannii, K. pneumoniae, and P. aeruginosa. Mechanistic studies
also confirm that these bis-amidines effectively induce Gram-negative
OM disruption. Taken together, the findings here reported provide
a broader understanding of the potential for bis-amidines to be used
as synergists in expanding the activity of Gram-positive specific
antibiotics against Gram-negative bacteria.
Methods
General Procedures
All reagents employed were of American
Chemical Society (ACS) grade or finer and were used without further
purification unless otherwise stated. For compound characterization, 1H NMR spectra were recorded at 400 MHz with chemical shifts
reported in parts per million (ppm) downfield relative to CHCl3 (7.26) or dimethyl sulfoxide (DMSO) (δ 2.50). 1H NMR data are reported in the following order: multiplicity
(s, singlet; d, doublet; t, triplet; q, quartet and m, multiplet),
coupling constant (J) in hertz (Hz), and the number
of protons. Where appropriate, the multiplicity is preceded by br,
indicating that the signal was broad. 13C NMR spectra were
recorded at 101 MHz with chemical shifts reported relative to CDCl3 (δ 77.16) or DMSO (δ 39.52). HRMS analysis was
performed on a Shimadzu Nexera X2 UHPLC system with a Waters Acquity
HSS C18 column (2.1 × 100 mm, 1.8 μm) at 30 °C and
equipped with a diode array detector. The following solvent system,
at a flow rate of 0.5 mL/min, was used: solvent A, 0.1% formic acid
in water and solvent B, 0.1% formic acid in acetonitrile. Gradient
elution was as follows: 95:5 (A/B) for 1 min, 95:5 to 15:85 (A/B)
over 6 min, 15:85 to 0:100 (A/B) over 1 min, 0:100 (A/B) for 3 min,
and then reversion back to 95:5 (A/B) for 3 min. This system was connected
to a Shimadzu 9030 QTOF mass spectrometer (ESI ionization) calibrated
internally with an Agilent’s API-TOF reference mass solution
kit (5.0 mM purine, 100.0 mM ammonium trifluoroacetate and 2.5 mM
hexakis(1H,1H,3H-tetrafluoropropoxy)phosphazine) diluted to achieve a mass count
of 10,000. Compounds 13, 14, 33, and 34 were synthesized as previously described and
had NMR spectra and mass spectra consistent with the assigned structures.[32,69] Compounds 1, 2, 4–6, 8–11, 15, 18, 19, 21–23, 1b, 21b–23b, 39, and 40 were synthesized using optimized
protocols as described below and gave NMR spectra and mass spectra
consistent for the same compounds previously described in the literature.[22,29,32,34,38,39,70−72] Purity of the final compounds 1–3, 9–12, 15, 16, 21–24, 1b, 21b–23b, 1c, 21c–23c, 37, 38, 43, and 44 was confirmed to
be ≥95% by analytical RP-HPLC using a Shimadzu Prominence-i
LC-2030 system with a Dr. Maisch ReproSil Gold 120 C18 column (4.6
× 250 mm, 5 μm) at 30 °C and equipped with a UV detector
monitoring at 214 nm. The following solvent system, at a flow rate
of 1 mL/min, was used: solvent A, 0.1% trifluoroacetic acid (TFA)
in water/acetonitrile, 95/5 and solvent B, 0.1% TFA in water/acetonitrile,
5/95. Gradient elution was as follows: 95:5 (A/B) for 2 min, 95:5
to 0:100 (A/B) over 30 min, 0:100 (A/B) for 1 min, then reversion
back to 95:5 (A/B) over 1 min, 95:5 (A/B) for 3 min. The compounds
were purified via preparative HPLC using a BESTA-Technik
system with a Dr. Maisch Reprosil Gold 120 C18 column (25 × 250
mm, 10 μm) and equipped with a ECOM Flash UV detector monitoring
at 214 nm. The following solvent system, at a flow rate of 12 mL/min,
was used: solvent A, 0.1% TFA in water/acetonitrile 95/5 and solvent
B, 0.1% TFA in water/acetonitrile 5/95. Unless stated otherwise in
the protocol, the gradient elution was as follows: 100:0 (A/B) to
0:100 (A/B) over 25 min, 0:100 (A/B) for 3 min, then reversion back
to 100:0 (A/B) over 1 min, 100:0 (A/B) for 1 min.
This protocol was based on the synthesis
of structurally similar amidine containing compounds previously described
in the literature.[28−31] 4,4′-(pentane-1,5-diylbis(oxy))dibenzonitrile (94 mg, 0.3
mmol) was dissolved in dry THF (2 mL) under an argon atmosphere and
LHMDS (1.2 mL, 1 M THF solution, 4.0 equiv) was added. The reaction
was stirred at room temperature for 48 h or longer until complete
conversion to the bis-amidine [monitored by liquid chromatography–mass
spectrometry (LCMS)]. The solution was cooled to 0 °C and quenched
with HCl (4.5 mL, 4 M dioxane solution, 60 equiv). The mixture was
stirred at room temperature overnight, then diluted with diethyl ether,
and filtered. The precipitate was purified by preparative HPLC with
the gradient 0–100% in 30 min to give pentamidine (1) (120 mg, quant). 1H NMR (400 MHz, DMSO-d6): δ 9.14 (s, 4H), 9.06 (s, 4H), 7.81 (d, J = 8.9 Hz, 4H), 7.15 (d, J = 8.9 Hz, 4H),
4.12 (t, J = 6.4 Hz, 4H), 1.88–1.75 (m, 4H),
1.65–1.52 (m, 2H). 13C NMR (101 MHz, DMSO): δ
164.70, 163.06, 130.19, 119.50, 114.79, 68.05, 28.21, 22.09. HRMS
(ESI): calcd for C19H24N4O2 [M + H]+, 341.1977; found, 341.1977.
4,4′-((3-Phenylpentane-1,5-diyl)bis(oxy))dibenzonitrile
(109 mg, 0.28 mmol) was dissolved in the LHMDS solution (1.1 mL, 1
M THF solution, 4.0 equiv) under an argon atmosphere. The reaction
mixture was stirred at room temperature for 48 h or longer until complete
conversion to the bis-amidine (monitored by LCMS). The solution was
cooled to 0 °C and quenched with HCl (4.5 mL, 4 M dioxane solution,
60 equiv). The mixture was stirred at room temperature overnight,
then diluted with diethyl ether, and filtered. The precipitate was
purified by preparative HPLC with the gradient 20–100% in 30
min to give compound 3 (27.4 mg, 23%). 1H
NMR (400 MHz, DMSO-d6): δ 9.11 (d, J = 12.6 Hz, 8H), 7.77 (d, J = 8.9 Hz,
4H), 7.34–7.16 (m, 5H), 7.05 (d, J = 9.0 Hz,
4H), 4.00–3.90 (m, 2H), 3.83 (dd, J = 15.0,
8.9 Hz, 2H), 3.14–3.04 (m, 1H), 2.29–2.16 (m, 2H), 2.13–2.00
(m, 2H). 13C NMR (101 MHz, DMSO): δ 164.81, 162.92,
143.38, 130.21, 128.62, 127.69, 126.58, 119.64, 66.21, 38.31, 35.10.
HRMS (ESI): calcd for C25H28N4O2 [M + H]+, 417.2291; found, 417.2287.
4,4′-(Propane-1,3-diylbis(oxy))dibenzonitrile
(4)
These conditions were based on literature
protocols.[22] 4-cyanophenol (0.29 g, 2.4
mmol, 2.4 equiv)
was suspended in dry DMF (3 mL) under an argon atmosphere. The suspension
was cooled to 0 °C using an ice bath and NaH (96 mg, 60% dispersion
in mineral oil, 2.4 equiv) was slowly added. The reaction mixture
was stirred until a clear solution appeared, the ice bath was removed
and 1,3-dibromopropane (202 mg, 1 mmol) was added. The reaction mixture
was heated to 80 °C for 1 h and then cooled to room temperature.
Water (10 mL) was added to the mixture to obtain precipitation. The
precipitate was filtered, washed with water, and recrystallized from
EtOH to give compound 4 as white crystals (164 mg, 59%). 1H NMR (400 MHz, CDCl3): δ 7.59 (d, J = 8.9 Hz, 4H), 6.95 (d, J = 8.9 Hz, 4H),
4.21 (t, J = 6.0 Hz, 4H), 2.37–2.27 (m, 2H). 13C NMR (101 MHz, CDCl3): δ 162.09, 134.19,
119.26, 115.29, 104.39, 64.56, 28.96.
The
protocol is as described in literature.[69] 1,2-dibromoethane (4.3 mL, 50 mmol, 5 equiv), 4-cyanophenol (1.2
g, 10 mmol), and K2CO3 (4.2 g, 30 mmol, 3 equiv)
were suspended in dry DMF (20 mL) under an argon atmosphere. The mixture
was stirred at 100 °C for 5 h, cooled to room temperature, and
EtOAc and water were added. The organic layer was separated, washed
with brine, dried over Na2SO4, and concentrated in vacuo. The crude product was purified by column chromatography
(petroleum ether/EtOAc = 9:1) to afford compound 13 (0.97
g, 43%). 1H NMR (400 MHz, CDCl3): δ 7.60
(d, J = 9.0 Hz, 2H), 6.96 (d, J =
8.9 Hz, 2H), 4.33 (t, J = 6.1 Hz, 2H), 3.66 (t, J = 6.1 Hz, 2H). 13C NMR (101 MHz, CDCl3): δ 161.44, 134.24, 119.11, 115.44, 104.88, 68.08, 28.47.
The protocol is as described in literature.[32] Compound 13 (0.96 g, 4.3 mmol,
2 equiv) and Na2S·9H2O (0.51 g, 2.1 mmol)
were dissolved in DMSO (5 mL), and the mixture was heated to 115 °C
under an argon atmosphere. After 1 h, the mixture was poured into
ice water (25 mL) and left for 24 h in a fridge. The precipitate was
filtered, washed with cold water, and recrystallized from EtOH to
obtain compound 14 (0.65 g, 93%). 1H NMR (400
MHz, CDCl3): δ 7.58 (d, J = 8.9
Hz, 4H), 6.94 (d, J = 8.9 Hz, 4H), 4.23 (t, J = 6.4 Hz, 4H), 3.05 (t, J = 6.4 Hz, 4H). 13C NMR (101 MHz, CDCl3): δ 161.73, 134.20,
119.15, 115.31, 104.61, 68.35, 31.71.
Compound 15 (100 mg, 0.22
mmol) was dissolved in dry dichloromethane (DCM) (10 mL) under an
argon atmosphere. The solution was cooled to 0 °C using an ice
bath and m-CPBA (54 mg, 77% aqueous solution, 1.1
equiv) was added. The mixture was stirred at 0 °C for 2 h and
then concentrated in vacuo. After HPLC purification
with a 0–100% gradient in 30 min to obtain compound 16 (27 mg, 32%). 1H NMR (400 MHz, DMSO-d6): δ 9.18 (s, 4H), 8.99 (s, 4H), 7.83 (d, J = 8.9 Hz, 4H), 7.21 (d, J = 9.0 Hz, 4H),
4.52 (t, J = 5.5 Hz, 4H), 3.79 (t, J = 5.5 Hz, 4H). 13C NMR (101 MHz, DMSO): δ 164.65,
162.00, 130.29, 120.39, 114.94, 62.19, 53.38. HRMS (ESI): calcd for
C18H22N4O4S [M + H]+, 391.1441; found, 391.1434.
Following the procedure as described above
for compound 4, using 1,2-bis(bromomethyl)benzene (1.0
g, 3.8 mmol) afforded the title compound as a crude product. No precipitation
occurred upon the addition of water. Therefore, the mixture was concentrated
under reduced pressure and the crude product was purified by column
chromatography (petroleum ether/EtOAc = 19:1) to obtain compound 17 (1.2 g, 96%). 1H NMR (400 MHz, CDCl3): δ 7.59 (d, J = 8.0 Hz, 4H), 7.46 (dd, 4H),
6.99 (d, J = 8.0 Hz, 4H), 5.21 (s, 4H). 13C NMR (101 MHz, CDCl3): δ 161.76, 134.26, 134.12,
129.56, 129.27, 119.12, 115.57, 104.79, 68.46.
Following the procedure as described above
for compound 4, using 1,4-bis(bromomethyl)benzene (0.92
g, 3.5 mmol) afforded compound 19 as a crude product.
The crude product was not recrystallized due to insolubility issues
and was used in the next step without further purification based on
a purity assessment (NMR) (1.2 g, 97%). 1H NMR (400 MHz,
DMSO-d6): δ 7.78 (d, J = 8.8 Hz, 4H), 7.48 (s, 4H), 7.18 (d, J = 8.9 Hz,
4H), 5.22 (s, 4H). 13C NMR (101 MHz, DMSO): δ 161.74,
136.11, 134.24, 128.08, 119.13, 115.92, 103.05, 69.36.
Following the procedure as described above
for compound 4, using 2,7-bis(bromomethyl)naphthalene
(0.20 g, 0.64 mmol) afforded compound 20 as a crude product.
The crude product was not recrystallized due to insolubility issues
and was used in the next step without further purification based on
a purity assessment (NMR) (0.25 g, quant). 1H NMR (400
MHz, CDCl3): δ 7.90 (d, J = 8.5
Hz, 2H), 7.87 (s, 2H), 7.60 (d, J = 8.9 Hz, 4H),
7.54 (dd, J = 8.5, 1.7 Hz, 2H), 7.06 (d, J = 8.9 Hz, 4H), 5.29 (s, 4H). 13C NMR (101 MHz,
CDCl3): δ 134.23, 134.04, 133.08, 128.74, 126.61,
125.65, 119.26, 115.77, 104.56, 70.42.
Following the procedure as described above
for compound 4, using 1,2-bis(bromomethyl)benzene (1.0
g, 3.8 mmol) and 3-cyanophenol (1.1 g, 9.1 mmol, 2.4 equiv) afforded
the title compound as a crude product. The crude product did not precipitate
but had very high viscosity. During filtration, a minimal amount of
acetone was used to prevent clogging. The precipitate was collected
and the filtrate was concentrated under reduced pressure to evaporate
the acetone. The precipitate in aqueous solution was filtered again
with a minimal amount of acetone. This process was repeated three
times to obtain compound 26 (1.1 g, 85%). 1H NMR (400 MHz, CDCl3): δ 7.53–7.47 (m, 2H),
7.45–7.35 (m, 4H), 7.28–7.27 (m, 1H), 7.26–7.25
(m, 1H), 7.20–7.16 (m, 4H), 5.18 (s, 4H). 13C NMR
(101 MHz, CDCl3): δ 158.65, 134.27, 130.67, 129.54,
129.22, 125.21, 120.18, 118.71, 117.74, 113.49, 68.55.
Following the procedure as described above
for compound 4, using 1,3-bis(bromomethyl)benzene (0.92
g, 3.5 mmol) and 3-cyanophenol (1.0 g, 8.4 mmol, 2.4 equiv) afforded
compound 27 as a crude product. The crude product was
not recrystallized due to insolubility issues and was used in the
next step without further purification based on a purity assessment
(NMR) (1.2 g, quant). 1H NMR (400 MHz, CDCl3): δ 7.51–7.34 (m, 6H), 7.28–7.24 (m, 2H), 7.23–7.17
(m, 4H), 5.11 (s, 4H). 13C NMR (101 MHz, CDCl3): δ 158.73, 136.65, 130.56, 129.34, 127.48, 126.41, 125.02,
120.22, 118.77, 117.91, 113.38, 70.14.
These conditions were based on literature
protocols.[42] To a suspension of compound 29 (190
mg, 0.62 mmol) and DIPEA (0.56 mL, 3.2 mmol, 5 equiv) in EtOH (10
mL) was added NH2OH·HCl (208 mg, 3 mmol, 4.8 equiv).
The reaction mixture was stirred at 85 °C overnight. The mixture
was concentrated in vacuo and the residue was dissolved
in AcOH (4.2 mL) and Ac2O (0.29 mL, 3 mmol, 4.8 equiv)
was added. The reaction mixture was stirred for 4 h and then concentrated in vacuo. The residue was co-evaporated with toluene three
times and then suspended in AcOH (7.5 mL) under an argon atmosphere.
Zinc powder (60 mg, 0.92 mmol, 1.5 equiv) was added and the mixture
was stirred at 35 °C overnight. Upon completion, the reaction
mixture was filtered through Celite, the Celite was rinsed with acetone
and all collected fractions were concentrated in vacuo. The crude product purified by preparative HPLC (gradient 20–100%,
30 min) to afford final compound 1c (102 mg, 48%). 1H NMR (400 MHz, DMSO-d6): δ
9.32 (s, 4H), 9.12 (s, 4H), 7.60 (t, J = 7.9 Hz,
2H), 7.51 (d, J = 7.5 Hz, 2H), 7.25 (d, J = 8.4 Hz, 2H), 7.11 (t, J = 7.5 Hz, 2H), 4.09 (t, J = 6.4 Hz, 4H), 1.81 (p, J = 6.9 Hz, 4H),
1.56 (p, J = 7.6 Hz, 2H). 13C NMR (101
MHz, DMSO): δ 164.64, 156.10, 133.82, 129.53, 120.35, 118.55,
113.07, 68.28, 28.01, 21.76. HRMS (ESI): calcd for C19H24N4O2 [M + H]+, 341.1977;
found, 341.1972.
The protocol is as described in literature.[44] Dimethyl 5-bromoisophthalate (2.3 g, 8.3 mmol)
was dissolved in
dry DCM (25 mL) under an argon atmosphere. The solution was then cooled
to 0 °C using an ice bath and DIBALH (40 mL, 1 M hexane solution,
4.8 equiv) was added dropwise. The mixture was stirred from 0 °C
to room temperature for 1 h. The reaction was quenched with Rochelle
salt (60 mL, sat. aq) and the biphasic mixture was stirred at room
temperature overnight. The layers were separated and the aqueous layer
was two times extracted with diethyl ether. The organic layers were
combined, washed with water and brine, dried over Na2SO4, and concentrated in vacuo. The crude product
was purified using column chromatography (DCM/EtOAc = 1:1) and afforded
compound 33 (1.8 g, 96%). 1H NMR (400 MHz,
MeOD): δ 7.42 (s, 2H), 7.28 (s, 1H), 4.58 (s, 4H), 3.35 (s,
2H). 13C NMR (101 MHz, MeOD): δ 145.51, 129.42, 124.82,
123.31, 64.29.
1-Bromo-3,5-bis(bromomethyl)benzene (34)
The protocol is as described in literature.[45] To a solution of compound 33 (1.0
g, 4.6 mmol) in dry
DCM (50 mL) was added PPh3 (2.5 g, 9.7 mmol, 2.1 equiv)
and CBr4 (3.2 g, 9.7 mmol, 2.1 equiv), and the mixture
was stirred at room temperature for 2 h under an argon atmosphere.
The reaction was quenched with water (30 mL) and the product was extracted
from the aqueous layer with DCM three times. The combined organic
layers were washed with water and brine, dried over Na2SO4, and concentrated in vacuo. The crude
product was purified by column chromatography (petroleum ether 100%)
to give compound 34 (0.87 g, 55%). 1H NMR
(400 MHz, CDCl3): δ 7.51–7.45 (m, 2H), 7.34
(s, 1H), 4.53 (d, J = 4.0 Hz, 1H), 4.41 (d, J = 4.2 Hz, 3H). 13C NMR (101 MHz, CDCl3): δ 140.42, 140.11, 132.11, 132.09, 131.64, 128.40, 127.90,
122.83, 44.89, 31.64, 31.59.
Following the procedure as described above
for compound 4, using compound 34 (0.82
g, 2.4 mmol) afforded compound 35 as a crude product.
The crude product was not recrystallized due to insolubility issues
and was used in the next step without further purification based on
a purity assessment (NMR) (1.0 g, quant). 1H NMR (400 MHz,
CDCl3): δ 7.66–7.57 (m, 4H), 7.57–7.53
(m, 2H), 7.43–7.33 (m, 1H), 7.04–6.96 (m, 4H), 5.15–5.05
(m, 4H). 13C NMR (101 MHz, CDCl3): δ 161.57,
138.62, 134.25, 130.33, 124.72, 123.30, 119.09, 115.62, 104.87, 69.18.
Conditions were based on protocols described
in literature.[47,48] Dibenzonitrile intermediate 35 (0.30 g, 0.72 mmol) was dissolved in a 3:1 mixture of THF
and 2 M Na2CO3 (aq) of 8 mL, respectively. Phenylboronic
acid (0.13 g, 1.1 mmol, 1.5 equiv) and Pd(dppf)Cl2·DCM
(58 mg, 0.07 mmol, 0.1 equiv) were added. The reaction mixture was
heated to 65 °C for 18 h and then partitioned between DCM and
NaHCO3 (sat. aq). The aqueous layer was three times extracted
with DCM, the organic layers were combined and dried over Na2SO4. The solvent was evaporated under reduced pressure
and the crude product was purified using column chromatography (petroleum
ether/EtOAc = 4:1) to obtain compound 36 (0.28 g, 94%). 1H NMR (400 MHz, CDCl3): δ 7.66–7.57
(m, 8H), 7.50–7.36 (m, 4H), 7.05 (d, J = 8.8
Hz, 4H), 5.19 (s, 4H). 13C NMR (101 MHz, CDCl3): δ 161.90, 142.59, 140.22, 137.05, 134.21, 129.06, 128.03,
127.31, 126.40, 125.32, 119.19, 115.67, 104.57, 70.12.
Following the procedure as described above
for compound 3, using compound 36 (0.28
g, 0.67 mmol), LHMDS (5.4 mL, 1 M THF solution, 8 equiv), and HCl
(10 mL, 4 M dioxane solution, 60 equiv). HPLC purification using a
30–100% gradient for 30 min afforded compound 38 (0.23 g, 74%). 1H NMR (400 MHz, DMSO-d6): δ 9.15 (d, J = 15.3 Hz, 8H),
7.83 (d, J = 9.0 Hz, 4H), 7.76 (s, 2H), 7.69 (d, J = 7.2 Hz, 2H), 7.58 (s, 1H), 7.50 (t, J = 7.6 Hz, 2H), 7.40 (t, J = 7.3 Hz, 1H), 7.27 (d, J = 9.0 Hz, 4H), 5.33 (s, 4H). 13C NMR (101 MHz,
DMSO): δ 164.83, 162.61, 140.77, 139.54, 137.54, 130.26, 129.14,
127.94, 126.84, 126.19, 126.03, 120.04, 115.25, 69.51. HRMS (ESI):
calcd for C28H26N4O2 [M
+ H]+, 451.2135; found, 451.2130.
(4-Bromo-1,2-phenylene)dimethanol
(39)
Conditions were based on a protocol reported
in literature.[49] LAH (15 mL, 1 M THF solution,
2 equiv) and ZnCl2 (0.61 g, 4.5 mmol, 0.6 equiv) were suspended
in dry THF (30
mL) and cooled to 0 °C, then 4-bromophthalic anhydride (1.7 g,
7.5 mmol) was slowly added. The mixture was stirred at room temperature
for 6 h under an argon atmosphere. The mixture was cooled to 0 °C
and quenched with Rochelle salt (30 mL, sat. aq) and the biphasic
mixture was stirred at room temperature overnight. The layers were
separated and the aqueous layer was extracted with diethyl ether two
times and the combined organic layers were washed with water and brine,
dried over Na2SO4, and concentrated in vacuo. The crude product was purified by column chromatography
(DCM/EtOAc = 1:1) to give compound 39 (1.5 g, 95%). 1H NMR (400 MHz, CDCl3): δ 7.48 (d, J = 2.1 Hz, 1H), 7.42 (dd, J = 8.0, 2.1
Hz, 1H), 7.18 (d, J = 8.0 Hz, 1H), 4.62 (d, J = 2.6 Hz, 4H), 3.20 (s, 2H). 13C NMR (101 MHz,
CDCl3): δ 141.49, 138.18, 129.88, 128.77, 127.92,
122.30, 64.53, 64.40, 64.31, 63.49, 63.47, 31.08, 23.80.
Following the procedure as described above
for compound 3, using compound 42 (0.28
g, 0.66 mmol), LHMDS (5.4 mL, 1 M THF solution, 8 equiv), and HCl
(10 mL, 4 M dioxane solution, 60 equiv) afforded the crude product.
The crude product was purified using HPLC with a 30–100% gradient
for 30 min to obtain pure compound 44 (35 mg, 12%). 1H NMR (400 MHz, DMSO-d6): δ
9.14 (d, J = 5.2 Hz, 4H), 9.00 (d, J = 4.8 Hz, 4H), 7.88 (d, J = 2.0 Hz, 1H), 7.81 (dd, J = 8.9, 3.2 Hz, 4H), 7.72–7.61 (m, 4H), 7.48 (t, J = 7.6 Hz, 2H), 7.39 (t, J = 7.3 Hz, 1H),
7.28 (t, J = 8.6 Hz, 4H), 5.44 (d, J = 10.4 Hz, 4H). 13C NMR (101 MHz, DMSO): δ 164.72,
162.49, 162.45, 140.20, 139.34, 135.20, 133.89, 130.25, 129.56, 129.11,
127.37, 126.75, 120.10, 120.08, 115.26, 115.23, 67.64, 67.31. HRMS
(ESI): calcd for C28H26N4O2 [M + H]+, 451.2135; found, 451.2129.
Antimicrobial
Assays
All compounds were screened for
antimicrobial activity against E. coli BW25113. A select group of the pentamidine analogues was further
tested against E. coli ATCC25922, E. coli W3110, E. coli 552060.1, and E. coli BW25113 transformed
with pGDP2-mcr-1 (the plasmid was a gift from Gerard Wright (Addgene
plasmid # 118404; http://www.n2t.net/addgene:118404; RRID: Addgene_118404)[63]), E. coli mcr-1, E. coli EQASmcr-1 (EQAS 2016 412016126), E. coli EQASmcr-2 (EQAS 2016 KP37), E. coli EQASmcr-3 (EQAS 2017 2013-SQ352), E. coli RC0089, K. pneumoniae ATCC13883, P. aeruginosa ATCC27853, and A. baumannii ATCC17978. The antimicrobial assay was performed according to CLSI
guidelines. Bacteria were plated out directly from their glycerol
stocks on blood agar plates, incubated overnight at 37 °C, and
then kept in the fridge. The blood agar plates were only used for
2 weeks and then replaced.
MIC Assays
A single colony from
a blood agar plate
was inoculated in Lysogeny Broth (LB) at 37 °C until a 0.5 optical
density at 600 nm (OD600) was reached (compared to the
sterility control of LB). The bacterial suspension was diluted in
fresh LB to 2.0 × 106 CFU/mL. The serial dilutions
were prepared in polypropylene microtiter plates: a stock of the test
compounds was prepared with a 2× final concentration in LB. 100
μL of the stock was added to the wells of the top row of which
50 μL was used for the serial dilution. The bottom row of each
plate was used as the positive (50 μL of LB) and negative controls
(100 μL of LB) (6 wells each). 50 μL of the 2.0 ×
106 CFU/mL bacterial stock was added to each well except
for the negative controls, adding up to a total volume of 100 μL
per well. The plates were sealed with a breathable seal and incubated
for 20 h at 37 °C and 600 rpm. The MIC was visually determined
after centrifuging the plates for 2 min at 3000 rpm.
Checkerboard
Assays
Dilution series of both the test
compound and antibiotic to be evaluated was prepared in LB media.
To evaluate synergy, 25 μL of the test compound solutions were
added to wells containing 25 μL of the antibiotic solution.
This was replicated in three columns for each combination so as to
obtain triplicates. To the resulting 50 μL volume of the antibiotic
+ test compound was next added 50 μL of the bacterial stock
(see MIC Assays), and the plates were sealed.
After incubation for 20 h at 37 °C while shaking at 600 rpm,
the breathable seals were removed, and the plates shaken using a bench
top shaker to ensure an even suspension of the bacterial cells as
established by visual inspection. The plates were then transferred
to a Tecan Spark plate reader and following another brief shaking
(20 s), the density of the bacterial suspensions was measured at 600
nm (OD600). The resulting OD600 values were transformed into a 2D
gradient to visualize the growth/no-growth results. The FICI was calculated
using eq , with an FICI
≤ 0.5 indicating synergy.[21]Equation calculation of FICI. MSCant = MIC of antibiotic
in combination with synergist; MICant = MIC of antibiotic
alone; MSCsyn = MIC of synergist in combination with antibiotic;
and MICsyn = MIC of synergist alone. In the cases, where
the MIC of the antibiotic or synergist was found to exceed the highest
concentration tested, the next highest concentration in the dilution
series was used in determining the FICI, and the result reported as
≤ the calculated value.
Hemolysis Assays
The hemolytic activity of each analogue
was assessed in triplicate. Red blood cells from defibrinated sheep
blood obtained from Thermo Fisher were centrifuged (400 g for 15 min
at 4 °C) and washed five times with phosphate-buffered saline
containing 0.002% Tween 20 (buffer). Then, the red blood cells were
normalized to obtain a positive control read-out between 2.5 and 3.0
at 415 nm to stay within the linear range with the maximum sensitivity.
A serial dilution of the compounds (200–6.25 μg/mL, 75
μL) was prepared in a 96-well plate. The outer border of the
plate was filled with 75 μL buffer. Each plate contained a positive
control (0.1% Triton-X final concentration, 75 μL) and a negative
control (buffer, 75 μL) in triplicate. The normalized blood
cells (75 μL) were added, and the plates were incubated at 37
°C for 1 or 20 h while shaking at 500 rpm. A flat-bottom plate
of polystyrene with 100 μL of buffer in each well was prepared.
After incubation, the plates were centrifuged (800g for 5 min at room temperature) and 25 μL of the supernatant
was transferred to their respective wells in the flat-bottom plate.
The values obtained from a read out at 415 nm were corrected for the
background (negative control) and transformed into a percentage relative
to the positive control.
Membrane Permeability Assay Using N-Phenylnaphthalen-1-amine
The assay was performed
based on protocols adapted from those described
in literature.[67,68] Bacteria were inoculated overnight
at 37 °C in LB, diluted the next day 50x in LB, and grown to
OD600 of 0.5. The bacterial suspension was then centrifuged
for 10 min at 1000g at 25 °C. The pellet of
bacteria was suspended in 5 mM HEPES buffer containing 20 mM glucose
to a final concentration of OD600 of 1.0. The compounds
were serial diluted (25 μL) in triplicate in a black 1/2 area
clear-bottom 96-well plate. 100 μg/mL final concentration of
colistin in triplicate served as the positive control. Three wells
were filled with 25 μL buffer to serve as the negative control.
Additional controls of the compounds were made in triplicate using
25 μL of the highest concentration to detect interactions of
the compounds with NPN in the absence of bacteria. A stock of 0.5
mM of NPN in acetone was prepared and diluted 12.5× in the buffer.
25 μL of the NPN solution was added to each well. 50 μL
of the 1.0 OD600 bacterial stock was then added to each
well except for the controls of the compounds with NPN. To these wells,
50 μL of buffer was added. After 60 min, the plate was measured
using a Tecan plate reader with λex 355 ± 20
nm and λem 420 ± 20 nm. The fluorescence values
obtained were then transformed into a NPN uptake percentage using
following eqEquation : NPN uptake. The observed
fluorescence (Fobs) is corrected for background
using the negative control
(F0). This value is divided by the positive
control corrected for background (F100 – F0) and multiplied by 100%
to obtain the percentage NPN uptake.[73]
Sample Availability
Samples of all compounds reported are
available from the authors
upon request.
Authors: Dorota Maciejewska; Jerzy Zabinski; Pawel Kaźmierczak; Mateusz Rezler; Barbara Krassowska-Świebocka; Margaret S Collins; Melanie T Cushion Journal: Eur J Med Chem Date: 2011-12-13 Impact factor: 6.514
Authors: Dennis J Doorduijn; Dani A C Heesterbeek; Maartje Ruyken; Carla J C de Haas; Daphne A C Stapels; Piet C Aerts; Suzan H M Rooijakkers; Bart W Bardoel Journal: PLoS Pathog Date: 2021-11-09 Impact factor: 7.464
Authors: Craig R MacNair; Jonathan M Stokes; Lindsey A Carfrae; Aline A Fiebig-Comyn; Brian K Coombes; Michael R Mulvey; Eric D Brown Journal: Nat Commun Date: 2018-01-31 Impact factor: 14.919
Authors: Charlotte M J Wesseling; Thomas M Wood; Kristine Bertheussen; Samantha Lok; Nathaniel I Martin Journal: Molecules Date: 2021-03-30 Impact factor: 4.411
Authors: Jarrod W Johnson; Michael J Ellis; Zoë A Piquette; Craig MacNair; Lindsey Carfrae; Timsy Bhando; Nikki E Ritchie; Paul Saliba; Eric D Brown; Jakob Magolan Journal: ACS Med Chem Lett Date: 2022-01-21 Impact factor: 4.345
Figure 1
Scheme 1
Figure 2
Scheme 2
Scheme 3
Scheme 4
Figure 9