Correlation of the antibacterial activities of cationic peptide antibiotics and cationic steroid antibiotics

Article abstract of DOI:10.1021/jm0105070

The antibacterial activities of cationic steroid antibiotics and cationic peptide antibiotics have been compared. Depolarization of bacterial membranes, activation of bacterial stress-related gene promoters, and changes in bacterial morphologies caused by

Full text of DOI:10.1021/jm0105070

J . Med. Chem. 2002, 45, 663-669
663
Cor r ela tion of th e An tiba cter ia l Activities of Ca tion ic P ep tid e An tibiotics a n d
Ca tion ic Ster oid An tibiotics1
Bangwei Ding, Qunying Guan, J oshua P. Walsh, J . Scott Boswell, Tim W. Winter, Erica S. Winter,
Stephanie S. Boyd, Chunhong Li, and Paul B. Savage*
Department of Chemistry and Biochemistry, Brigham Young University, Provo, Utah 84602
Received November 2, 2001
The antibacterial activities of cationic steroid antibiotics and cationic peptide antibiotics have
been compared. Depolarization of bacterial membranes, activation of bacterial stress-related
gene promoters, and changes in bacterial morphologies caused by these antibiotics suggest
that cationic steroid and peptide antibiotics share mechanistic aspects. Modified cationic steroid
antibiotics display improved selectivity for prokaryotic cells over eukaryotic cells presumably
due to increased charge recognition.
In tr od u ction
Cationic peptide antibiotics (CPAs) have received
increasing attention because they are found in organ-
isms ranging from bacteria to mammals and because
of their potent antimicrobial activities.2 Reported CPAs
can be classified according to their structure; the two
major classes include CPAs that adopt R-helical confor-
mations and those that form â-sheets.2,3 In general,
CPAs in both classes adopt facially amphiphilic4 con-
formations, with cationic groups on one face of the
molecule and hydrophobic groups on the other. Dem-
onstrating the importance of cationic/hydrophobic seg-
regation, amphiphilic helices comprised of â-peptides
have been prepared and have shown potent antibacte-
rial properties.5 Models for the mode of action of CPAs
include the “carpet model” and the “barrel-and-stave
F igu r e 1. Structures of 1 and 2 and sequences of magainin
I and cecropin A (one letter amino acid code is used).
model”.6 In the former model, amphiphilic peptides
associate with the negatively charged bacterial mem-
brane, and once the local concentration of the peptide
reaches a sufficient level, patches of the membrane are
removed and the membrane structure is compromised.
In the latter model, cationic peptides act as the staves
of a barrel in forming stable pores in bacterial mem-
branes.
Four general characteristics of CPAs have been used
to describe their activities.7 (i) Many CPAs display
selective toxicity for prokaryotic cells over eukaryotic
cells. This selectivity is likely caused by affinity of the
cationic peptides for the net negative charge found on
bacterial cells in contrast to eukaryotic lipid bilayers,
which are typically made up of zwitterionic phospho-
lipids. (ii) Most CPAs exhibit rapid bacterial killing
times. In general, these antibiotics are believed to be
membrane active, as opposed to other types of antibiot-
ics that target a single biochemical pathway and require
longer periods of time to inhibit cell growth and division.
(iii) Examples of CPAs display a broad spectrum of
antibacterial activity. Because they are targeted to the
charged character of bacterial membranes, they ef-
fectively kill multiple types of bacteria. (iv) CPAs
generally do not induce the formation of resistant
organisms. To become resistant to CPAs, bacteria have
to change their membrane structure, which alters the
permeability barrier provided by the membrane and can
make the bacteria more susceptible to other antibiotics.
An additional characteristic of certain CPAs is the
ability to effectively permeabilize the outer membranes
of Gram-negative bacteria, sensitizing these organisms
to hydrophobic antibiotics.8
Because of their complexity and size, many CPAs are
difficult to synthesize and derivatize,2 which complicates
their biological study and application as antibiotics.
Simpler compounds with similar activities would be
desirable provided that they could be easily prepared
and derivatized. One of the features of some CPAs that
has impeded their development for clinical use is their
hemolytic activity (disruption of red blood cells). Smaller
molecules that display controlled hemolytic activity
could provide a better understanding of the factors that
govern prokaryote vs eukaryote membrane selectivity.9
We have developed multiple antimicrobial compounds
(e.g., 1 and 2, Figure 1), comprised of steroids,10 that
display antibacterial behaviors that are similar to those
of CPAs. These cationic steroid antibiotics (CSAs) have
proven to be broad-spectrum antibiotics active even
* To whom correspondence should be addressed. Tel: 801-378-4020.
Fax: 801-378-5474. E-mail: paul_savage@byu.edu.
10.1021/jm0105070 CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/22/2001

664 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 3
Ding et al.
F igu r e 2. Helix wheel representation of magainin II demon-
strating the segregation of basic and hydrophobic residues.
Perspective drawing of 1.
against multidrug-resistant strains of Gram-negative
and Gram-positive bacteria.11 In addition, examples of
CSAs permeabilize the outer membranes of Gram-
negative bacteria effectively sensitizing them to hydro-
phobic antibiotics.8,12 Because of this activity, we have
postulated that the bactericidal properties of CSAs are
due to membrane disruption, and we have found that
CSAs display a moderate degree of selectivity for
prokaryotic over eukaryotic membranes.12
In an effort to better understand the mechanism of
action of two representative CSAs (1 and 2), we have
directly compared their activities to those of one of the
best-studied groups2 of CPAs, the magainin and ce-
cropin antibiotics (Figure 1). The magainins, isolated
from the skin secretions of the frog Xenopus laevis,13
and cecropins, isolated from the giant silk worm moth
Hyalaphora cecropia,14 adopt R-helical conformations in
the presence of bacterial membranes15 and are believed
to act via the carpet model.6 Studies performed include
measurements of depolarization of bacterial membranes
by CSAs and CPAs, observation of bacterial gene
promoters activated in response to the antibiotics, and
changes in bacterial morphology as a consequence of
antibacterial action. In addition, we have also pre-
pared and characterized new CSAs that display im-
proved selectivity for prokaryotic membranes over their
eukaryotic counterparts.
F igu r e 3. Depolarization of the membrane of M. luteus
measured by an increase in fluorescence of 3,3′-diethylthiodi-
carbodyanine iodide: (, magainin (50 µg/mL or 21 µM); 9, 1
(2 µg/mL or 3.5 µM); 2, 2 (0.5 µg/mL or 0.74 µM).
As a consequence of membrane activity of CPAs,
bacterial membranes are rapidly depolarized. The ac-
tivities of the magainins and nisin, both CPAs, have
been characterized using the fluorescent, membrane-
potential-sensitive probe 3,3′-diethylthiodicarbocyanine
iodide.17 In these experiments, the fluorescent dye is
added to a bacterial culture, and the dye incorporates
into the bacterial membrane. Depolarization of the
membrane results in a large increase in fluorescence of
the dye.18 For our experiments, we used Micrococcus
luteus (Presque Isle Cultures 456), a Gram-positive
organism that is susceptible to the magainins. The
addition of magainin I or CSAs 1 or 2 caused a rapid
depolarization of the bacterial membrane (Figure 3).
These results demonstrate the rapidity with which the
antibiotics act and suggest that the CSAs have mem-
brane activity comparable to magainin I, although the
CSAs are active at lower concentrations.
To verify that CSAs rapidly kill bacteria, the times
required for compound 2 to lower bacterial popu-
lations in culture were determined. To initial bac-
terial populations of 105 colony forming units (CFUs)
per milliliter of Escherichia coli (ATCC 25922) or
Staphylococcus aureus (ATCC 25923), compound 2 was
added at its minimum inhibition concentrations (MICs)
(0.31 and 0.59 µg/mL with E. coli and S. aureus,
respectively). The numbers of CFUs were reduced by
half from E. coli within 15 min and from S. aureus
within 75 min.
CPAs have been effectively characterized by the
bacterial promoters that they activate.19 In response to
external inducing agents, bacteria activate promoters
for a number of stress response genes. These promoters
have been cloned and coupled to a bacterial lumines-
cence reporter operon (luxCDABE) on a plasmid intro-
duced into E. coli.19 Strains of these engineered bacteria
respond to specific inducing agents by producing bacte-
rial luciferase and its substrate, which results in bacte-
rial luminescence. For example, strain DPD2170 re-
sponds to osmotic stress caused by high concentrations
(0.5 M) of sucrose or sodium chloride by activating the
Resu lts a n d Discu ssion
The possibility that CPAs and CSAs share mechanis-
tic aspects is better understood as the morphologies of
both groups of antibiotics are considered. Both types of
antibiotics display cationic facial amphiphilicity: the
CPAs by virtue of their secondary (and in some cases
tertiary) structure (for example, see the helix wheel of
magainin II in Figure 2) and the CSAs due to the
stereochemical orientation of the groups linking amines
to the steroid scaffolding (Figure 2). While the facial
amphiphilicity of the CSAs can be influenced by rotation
about the bonds in the propyleneoxy linkers, due to the
stereochemistry of the oxygen atoms on the steroid
scaffolding, the amine groups are predisposed to be on
one face of the molecule.16 Considering this common
feature of CPAs and CSAs, amphiphilic morphology
appears essential for membrane activity. While in
general the CSAs are smaller than the CPAs, the CPAs
that have been identified display a wide range of
molecular weights suggesting that “structure is more
important than size”.2b

Antibacterial Activities of Antibiotics
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 3 665
to CSA 1 in a manner similar to the response to the
cecropin CPAs.
Proposed steps in the carpet model of CPA activity
include (i) binding of peptide to anionic phospholipids
via ionic interactions and (ii) reorientation of the CPA
juxtaposing hydrophobic residues with the hydrophobic
core of the membrane, followed by (iii) disintegration
of the membrane.20 Disintegration of bacterial mem-
branes can be observed via transmission electron mi-
croscopy (TEM).
TEM was used to observe the morphological changes
that bacteria undergo when treated with CSAs. It was
necessary to use high concentrations of bacteria (∼109
CFU/mL) to prepare samples for TEM. To ensure that
the ratios of CSAs to bacteria were similar in MIC
measurements21 and in preparing the TEM samples,
MIC values were converted from units of micrograms
per milliliter to moles of CSA per bacterium. Using MIC
values with these units, TEM images were captured of
E. coli (ATCC 25922) treated with 1 and 2 at one-half
of their MIC values (Figure 5). As compared to images
of untreated bacteria, the TEM images of treated
bacteria clearly show disintegration of the bacterial
membranes. In fact, the formation of vesicles apparently
lacking intracellular bacterial components can be seen.
Disintegration of the bacterial membranes by the CSAs
is consistent with a carpet model of action.
The desirable activities of CPAs (i.e., broad spectrum
of activity, rapid killing times, and small likelihood of
development of resistant bacteria) are due, in general,
to their membrane activity; however, membrane activity
presents potential problems.22 The most important of
which is selectivity for binding and perturbation of
bacterial (prokaryotic) membranes. A measure of mem-
brane selectivity can be obtained by comparing the MIC
of an antibiotic to its minimum hemolytic concentration
(MHC, minimum concentration required to lyse red
blood cells). Compounds that display similar MIC and
MHC values offer little or no membrane selectivity;
compounds that are selective for prokaryotic membranes
have low MIC values and high MHC values.
Many CPAs, including the magainins, exhibit very
good selectivity for prokaryotic over eukaryotic mem-
branes.2 Other examples, such as mellitin, are much less
selective. The membrane selectivity of the magainins
has been attributed to recognition of the bulk charges
on bacterial membranes comprised of anionic phospho-
lipids, whereas mellitin associates strongly with zwit-
terionic phospholipids.23 Differences in helix dipoles
have also been used to explain the difference in hemo-
lytic activities of the magainins and mellitin.9
The membrane selectivity of reported CSAs varies;
some have demonstrated good selectivity, while others
have proven to be potent hemolytic agents.12,24 For
example, 1 exhibits low MIC values (>5 µg/mL) against
Gram-positive bacteria, while its MHC is relatively high
(100 µg/mL).12 Compound 1 is also a potent sensitizer
of Gram-negative bacteria to other hydrophobic anti-
biotics. Compound 2 is very active against both Gram-
negative and Gram-positive bacteria (MICs of <4 µg/
mL) and is more strongly hemolytic (MHC of 29 µg/mL).
F igu r e 4. Luminescence of E. coli strains DPD2170 and
DPD2194 induced by 1. (A) E. coli (DPD2170) (osmY -
luxCDABE): ∆, sucrose (0.5 M); 0, NaCl (0.3 M); ), control; (,
1 (0.10 × MIC); 9, 1 (0.05 × MIC). (B) E. coli (DPD2194) (micF
- luxCDABE): ∆, methyl viologen (0.1 µg/mL); ), control; (,
1 (0.10 × MIC).
promoter for the osmY gene. Activation of this promoter,
which is coupled to the luminescence operon in this
strain, causes an increase in luminescence. Similarly,
strain DPD2191 responds to oxidative stress caused by
methyl viologen by activation of the micF promoter,
which causes an increase in luminescence.
At sublethal doses, CPAs magainin I and magainin
II cause activation of the osmY promoter but not the
micF promoter19 where activation is defined by a
doubling of background luminescence. In contrast, ce-
cropins A and B activate both the osmY and the micF
promoters.19 Using the strain DPD2170, we found that
sublethal doses (one-tenth or one-twentieth of the MIC)
of 1 caused strong activation of the osmY (Figure 4)
suggesting that the CSA caused the bacteria to ex-
perience osmotic stress in a manner similar to the
magainins and cecropins. Higher concentrations caused
bacterial death and a loss of luminescence. CSA 2 only
caused a small increase in luminescence over the
background (data not shown). The difference in the
abilities of 1 and 2 to cause activation of the osmY
promoter may be related to their differing abilities to
traverse the outer membrane of Gram-negative bac-
teria.10c CSA 2 effectively traverses the outer membrane
and is bactericidal at relatively low concentrations. The
ability of 1 to activate the micF promoter was also
measured; a low concentration of the CSA caused a large
increase in the luminescence of DPD2191 (Figure 4).
Taken together, results from experiments with strains
DPD2170 and DPD2191 suggest that bacteria respond
Because bacterial membranes are generally nega-
tively charged,23 cationic compounds would be expected
to display affinity for prokaryotic membranes. To en-

666 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 3
Ding et al.
F igu r e 5. TEM images of E. coli (ATCC 25922). (A) Control (magnification 7000). (B) Treated with 1 at 0.5 × MIC (magnification
7000). (C) Treated with 2 at 0.5 × MIC (magnification 15 000).
F igu r e 6. Structures of CSAs 3, 4, squalamine, and 5.
Sch em e 1a
a
Reagents (yields in parentheses): (a) MsCl, Et3N, CH2Cl2; 3,3′-iminodipropionitrile, NaI, Na2CO3, THF (85%). (b) H2, PtO2, EtOH;
H2, PtO2, HCl, EtOH (85%).
hance the ability of CSAs to bind to bacterial mem-
branes, we prepared compounds 3 and 4 (Figure 6) in
which polyamine groups, in a branched and a linear
fashion, are appended at C-24. Polyamines, extending
from a steroid scaffolding, are found in squalamine
(Figure 6), a naturally occurring CSA obtained first from
the Dogfish shark25 and in squalamine mimics (e.g., 5,
Figure 6) developed by Regen and co-workers.24
polyamine, the methanesulfonate of 710c was prepared
and then reacted with 3,3′-iminodipropionitrile to give
8 in good yield. Reduction of the azides with hydrogen
and platinum oxide was possible without affecting the
nitrile groups. Under acidic conditions, the nitriles were
reduced giving 3 in pure form. In preparing 4, the linear
polyamine was added in a stepwise manner. The meth-
anesulfonate of 7 was reacted with 3-aminopropanol,
and the resulting amine was protected as the benzyl-
oxycarbamate giving 9. Mesylation of the resulting
The syntheses of 3 and 4 are outlined in Schemes 1
and 2, respectively. To prepare the CSA with a branched

Antibacterial Activities of Antibiotics
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 3 667
Sch em e 2a
a
Reagents (yields in parentheses): (a) MsCl, Et3N, CH2Cl2; 3-aminopropanol, NaI, Na2CO3, THF (75%). (b) Benzyl chloroformate,
Et3N, CH2Cl2 (80%). (c) MsCl, Et3N, CH2Cl2; N-Z-Propylenediamine, NaI, Na2CO3, DMSO (76%). (d) H2, PtO2, EtOH; H2, Pd(OH)2, AcOH,
EtOH (95%).
Ta ble 1. MIC Values of 1, 3, and 4 with Gram-Negative and
Gram-Positive Bacteria and MHC Values.
rapidly kill bacteria, they are active against a broad
spectrum of bacteria, and because they are membrane
MIC and MHC values (µg/mL)
active, there is little likelihood of formation of resistant
organisms. In addition, CSAs 1 and 2 depolarize bacte-
rial membranes at rates similar to that of magainin I,
and 1 causes activation of the same bacterial gene
promoters as the cecropins. TEM images of bacteria
subjected to CSAs are consistent with membrane activ-
ity similar to that of the facially amphiphilic CPAs
magainins and cecropins. Taken together, these results
suggest that CSAs share at least aspects of the mech-
anisms of action of CPAs that function via the carpet
model. Because in general CSAs are simpler to prepare
than CPAs and are easier to derivatize and purify, they
are well-suited for further investigations elucidating
how small molecules can selectively disrupt bacterial
membranes. Enhanced cell selectivity of 3 and 4, most
likely caused by charge recognition, increases the prob-
ability that CSAs can be used to fight infection and
demonstrates a means by which membrane active
compounds can be made more cell selective.
1
3
4
Gram-Negative Rods
36
E. coli (ATCC 25922)
Salmonella typhimurium
(ATCC 14028)
6.6
25
7.3
12
43
Gram-Positive Cocci
S. aureus (ATCC 25923)
Streptococcus pyogenes
(ATCC 19615)
minimum hemolytic
concentrations
2.0
4.2
4.6
3.0
2.0
1.6
100
>200
>200
alcohol and reaction with the monobenzyloxycarbamate
of propylenediamine completed incorporation of the
linear polyamine. Deprotection to give 4 was accom-
plished by first reducing the azides with hydrogen in
the presence of platinum oxide followed by removal of
the carbamate groups by hydrogenation with palladium
hydroxide.
The antibacterial activities of 3 and 4 were deter-
mined using both Gram-negative and Gram-positive
bacteria (Table 1). As compared to 1, both 3 and 4 were
more active against Gram-negative organisms and
displayed comparable activity against Gram-positive
bacteria. As observed with 1, CSAs 3 and 4 gave lower
MIC values against Gram-positive bacteria. This dif-
ference in the activities of 3 and 4 against the two
classes of organisms is presumably due to the perme-
ability barrier provided by the outer membrane.10c
As anticipated, the hemolytic properties of 3 and 4
were better than other CSAs reported. That is, hemoly-
sis was not observed up to concentrations of 200 µg/mL
(Table 1), and due to limited solubility, it was not
possible to use higher concentrations. Consideration of
the relatively low MIC values of 3 and 4 and high MHC
values suggests that the cationic character of these
compounds causes them to be selective for prokaryotic
membranes over eukaryotic membranes. This level of
membrane selectivity increases the likelihood that CSAs
may find therapeutic uses.
Exp er im en ta l Section
Ma ter ia ls a n d Sp ectr oscop ic Meth od s. 1H nuclear mag-
netic resonance (NMR) and 13 C NMR were recorded on Varian
Unity 500 MHz or Varian Unity 300 MHz instruments. Mass
spectrometric data were obtained on a J EOL SX 102 A
spectrometer. Optical density measurements were made using
an HP 8453 spectrophotometer. Fluorescence measurements
were made using a Perkin-Elmer LS50B fluorimeter. Lumi-
nescence experiments were made using an MGM Optocomp 1
luminometer. Tetrahydrofuran (THF) and CH2Cl2 were dried
over Na or CaH. Chemicals were obtained from Fluka, Aldrich,
and ARCOS and were used as received unless otherwise noted.
Mea su r em en t of Mem b r a n e Dep ola r iza t ion . An M.
luteus culture was grown up in Mueller-Hinton broth, and
cells were harvested by centrifugation, washed twice in a
buffer containing 250 mM sucrose and 5 mM MgSO4, and
suspended in the same buffer at an optical density (600 nm)
of 0.085. The dye 3,3′-diethylthiodicarbocyanine iodide was
added giving a 1 µM concentration. The dye was allowed to
incorporate for 15 min before measurements. An excitation
wavelength of 600 nm and an emission wavelength of 660 nm
were used to monitor depolarization. Samples were slowly
stirred during the measurement, and measurements were
taken every 30 s.
Con clu sion s
Esta blish in g Killin g Tim es for E. coli a n d S. a u r eu s.
The microorganisms were grown to a concentration of 105 CFU/
mL by incubating colonies for 120 min (E. coli) or 20 min (S.
CSAs display the general activities attributed to
CPAs: they display selectivity for bacterial cells, they

668 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 3
Ding et al.
aureus) in Mueller-Hinton broth at 37 °C. The bacterial
suspensions were challenged with the CSA, and at specific
time intervals, samples were plated out on Mueller-Hinton
agar plates. The plates were then incubated for 24 h at 37 °C,
and CFUs were counted.
out, and the filtrate was concentrated under reduced pressure.
Compound 3 was obtained (58.1 mg, 95%) as a yellow solid.
1
H NMR (dimethylsulfoxide (DMSO)-d6): δ 4.01 (broad, 10H),
3.72-2.81 (m, 20H), 2.18-165 (m, 20H), 1.51-1.12 (m, 19H),
13
0.96 (d, J ) 6.4 Hz, 3H), 0.89 (s, 3H), 0.69 (s, 3H). C NMR
(CD3OD): δ 82.53, 81.89, 77.96, 67.89, 67.78, 67.23, 58.12,
56.89, 54.21, 47.56, 44.41, 43.85, 40.92, 36.89, 36.77, 36.59,
35.96, 31.99, 30.09, 29.60, 29.28, 24.94, 24.72, 24.10, 23.71,
23.12, 22.68, 22.19, 20.06, 17.55, 11.80. HRFAB-MS (thio-
glycerol + H+ matrix): m/e ([M+ + H+]) 679.6229, calcd
679.6214.
Obser va tion of Activa tion of osm Y a n d m icF P r om ot-
er s. E. coli strains DPD 2170, DPD 2191, and DPD 2194 were
provided by Tina K. Van Dyk (Dupont). Cultures (12 mL) were
grown to optical densities (600 nm) of 0.05 in LB broth
containing ampicillin (50 µg/mL). Inducers were added at the
concentrations described, and the samples were incubated at
30 °C. At 30 min intervals, 1 mL samples were removed and
their luminescence was measured. Samples were discarded
after the measurements.
P r ep a r a tion of TEM Sa m p les. Cultures of E. coli (ATCC
25922) were grown to 109 CFU/mL in Mueller-Hinton broth,
and the bacteria were pelleted by centrifugation. The super-
natant was removed, and the CSA was added in Mueller-
Hinton broth. The samples were incubated for 90 min at 37
°C, and the bacteria were centrifuged. The supernatant was
removed, and 2% glutaraldehyde was added for fixing. After
the mixture was centrifuged and the supernatant was re-
moved, a solution of 1.5% low-melting agarose was added at
37 °C. A pellet of the cells in agarose was formed via
centrifugation and was used for TEM sectioning.
P r ep a r a t ion of 3r,7r,12r-Tr is(3-a zid op r op oxy)-5â-
ch ola n -24-(N,N-bis(2-cya n oeth yl))a m in e (8). MsCl (27.4
mg, 0.12 mmol) in CH2Cl2 (5 mL) was added dropwise to a
solution of 3R,7R,12R-tris(3-azidopropoxy)-5â-cholan-24-ol (7)
(77.0 mg, 0.12 mmol) and Et3N (48.5 mg, 0.48 mmol) in CH2Cl2
(20 mL) at 0 °C under N2 during 10 min. The mixture was
stirred for another 1 h and then washed with brine (3 × 10
mL), and the organic phase was dried over Na2SO4. The solvent
was removed under reduced pressure, and the resulting
material was used in the next step. A mixture of the mesylate,
3,3′-iminodipropionitrile (2 mL, 16.6 mmol), NaI (5 mg), and
Na2CO3 (127 mg, 1.20 mmol) in THF (2 mL) was refluxed
under N2 overnight. Water (10 mL) was added, and the product
was extracted by ethyl acetate (3 × 10 mL). The combined
extracts were dried over Na2SO4. The product 8 (76.4 mg, 85%)
was obtained as an oil after column chromatography (hexane/
EtOAc 2/1). 1H NMR (CDCl3): δ 3.68-3.59 (m, 2H), 3.55-3.37
(m, 8H), 3.28-3.09 (m, 4H), 2.86 (t, J ) 6.20 Hz, 4H), 2.57-
2.46 (m, 6H), 2.19-2.08 (m, 4H), 1.96-1.60 (m, 15H), 1.58-
1.48 (m, 3H), 1.46-1.21 (m, 9H), 1.09-1.01 (m, 2H), 0.96 (d,
J ) 6.4 Hz, 3H), 0.89 (s, 3H), 0.69 (s, 3H). 13C NMR (CDCl3):
δ 118.55, 80.43, 79.60,75.84, 64.80, 64.27, 64.11, 60.32, 54.02,
49.68, 48.80, 48.78, 48.56, 46.36, 46.23, 42.51, 41.72, 39.62,
35.58, 35.24, 35.13, 34.89, 33.36, 31.52, 29.57, 29.50, 28.89,
27.68, 27.47, 23.79, 23.34, 22.83, 22.65, 17.90, 16.94, 12.38.
High-resolution fast atom bombardment mass spectroscopy
(HRFAB-MS) (thioglycerol + Na matrix): m/e ([M+ + Na])
771.5140 (100%), calcd 771.5122.
P r ep a r a t ion of 3r,7r,12r-Tr is(3-a zid op r op oxy)-5â-
ch ola n -24-(N-(ben zyloxyca r bon yl)-N-(3-h yd r oxyp r op yl))-
a m in e (9). MsCl (105 mg, 0.92 mmol) in CH2 Cl2 (10 mL) was
added dropwise to a solution of 7 (294 mg, 0.46 mmol) and
Et3N (186 mg, 1.84 mmol) in CH2Cl2 (30 mL) at 0 °C under N2
during 10 min. The mixture was stirred for another 1 h, then
it was washed with brine (3 × 10 mL), and the organic phase
was dried over Na2SO4. The solvent was removed under
reduced pressure, and the resulting material was used for the
next step reaction. The mesylate, 2-aminopropanol (3 mL), NaI
(10 mg), and Na2CO3 (488 mg, 4.6 mmol) in THF (4 mL) were
refluxed under N2 overnight. Water (20 mL) was added, and
the product was extracted by ethyl acetate (3 × 20 mL) and
dried over Na2SO4. After column chromatography (CH2Cl2/
MeOH/NH4OH 10/1.5/0.2), the desired alcohol was obtained
as a brown oil (243 mg, 75%). 1H NMR (CDCl3): δ 3.78 (t, J )
5.37 Hz, 2H), 3.68-3.37 (m, 14H), 3.23-3.05 (m, 6H), 2.07 (t,
J ) 5.37 Hz, 2H), 2.60-2.51 (m, 2H), 2.17-2.11 (m, 3H), 1.93-
1.51 (m, 17H), 1.48-1.15 (m, 6H), 1.07-0.98 (m, 2H), 0.88 (d,
J ) 6.84 Hz, 3H), 0.86 (s, 3H), 0.63 (s, 3H). 13 C NMR (CDCl3):
δ 80.41, 79.59, 75.87, 64.82, 64.25, 64.21, 64.10, 50.25, 49.88,
48.82, 48.76, 48.56, 46.29, 46.20, 42.48, 41.79, 40.16, 39.62,
35.57, 35.24, 35.10, 34.89, 33.42, 30.43, 29.57, 29.52, 28.89,
27.67, 27.62, 27.45, 26.18, 23.35, 22.84, 22.65, 17.86, 12.38.
HRFAB-MS (thioglycerol + H+ matrix): m/e ([M+ + H+])
701.5186, calcd 701.5190. Benzyl chloroformate (65.0 mg, 0.38
mmol) in CH2Cl2 (5 mL) was added dropwise to a solution of
the alcohol from the previous step (242 mg, 0.35 mmol) and
Et3N (105 mg, 1.04 mmol) in CH2Cl2 (30 mL) at 0 °C under N2
during 10 min. The mixture was stirred for another 2 h and
then washed with brine (3 × 10 mL), and the organic phase
was dried over Na2SO4. The solvent was removed under
reduced pressure, and the crude product was purified by
column chromatography (hexane/EtOAc 2/1) to give 9 (230 mg,
1
80%) as an oil. H NMR (CDCl3): δ 7.39-7.21 (m, 5H), 5.62
(s, 1H), 5.11 (s, 2H), 3.79-3.22 (m, 14H), 3.19-3.02 (m, 6H),
2.17-2.13 (m, 3H), 1.93-1.61 (m, 14H), 1.48-1.15 (m, 16H),
0.88 (d, J ) 6.84 Hz, 3H), 0.86 (s, 3H), 0.63 (s, 3H). 13C NMR
(CDCl3): δ 157.51, 136.68, 128.45, 127.93, 127.70, 80.41, 79.59,
75.87, 67.37, 64.82, 64.25, 64.21, 58.30, 48.76, 48.56, 47.51,
46.46, 46.20, 45.75, 43.25, 42.48, 41.79, 39.62, 35.57, 35.24,
35.10, 34.89, 32.96, 30.58, 29.65, 29.57, 29.48, 28.89, 27.67,
27.62, 27.45, 25.57, 23.35, 22.84, 22.65, 17.86, 12.38. HRFAB-
MS (thioglycerol + H+ matrix): m/e ([M+]) 835.5467, calcd
835.5480.
P r ep a r a t ion of 3r,7r,12r-Tr is(3-a m in op r op oxy)-5â-
ch ola n -24-(N,N-bis(3-a m in op r op yl))a m in e (3). Compound
8 (76.4 mg, 0.10 mmol) in dry ethanol (10 mL) was hydro-
genated (800 psi) in the presence of PtO2 (5 mg, 0.02 mmol)
at room temperature for 48 h. The catalyst was filtered out,
and the filtrate was concentrated under reduced pressure. The
product was separated by column chromatography (CH2Cl2/
MeOH/NH4OH 3/3/1) to give the corresponding triamine (60.4
P r ep a r a t ion of 3r,7r,12r-Tr is(3-a m in op r op oxy)-5â-
c h o la n -24-(N -(N -(3-a m in o p r o p y l))-3-a m in o p r o p y l)-
a m in e (4). MsCl (31.9 mg, 0.28 mmol) in CH2Cl2 (5 mL) was
added dropwise to a solution of 9 (114 mg, 0.14 mmol) and
Et3N (56.6 mg, 0.56 mmol) in CH2Cl2 (20 mL) at 0 °C under
N2 during 10 min. The mixture was stirred for another 1 h,
then it was washed with brine (3 × 10 mL), and the organic
phase was dried over Na2SO4. The solvent was removed under
reduced pressure, and the resulting material was used in the
next step of the reaction. The resulting mesylate (285 mg, 1.40
mmol), N-benzyloxycarbamoylpropylenediamine (148 mg, 1.40
mmol), NaI (10 mg), and Na2CO3 (100 mg) in DMSO (2 mL)
were heated to 90 °C under N2 overnight. Water (10 mL) was
added, and the mixture was extracted by ethyl acetate (3 ×
20 mL). The extracts were dried over Na2SO4. After column
chromatography (CH2Cl2/MeOH/NH4OH 10/0.5/0.1), the de-
sired amine was obtained as a brown oil (109 mg, 76%). 1H
NMR (CDCl3): δ 7.39-7.22 (m, 10H), 5.85 (s, 1H), 5.11 (s, 2H),
1
mg, 89%). H NMR (CDCl3): δ 3.68-3.41 (m, 8H), 3.38-3.05
(m, 6H), 2.88-2.71 (m, 12H), 2.53-2.41 (m, 8H), 2.18-2.04
(m, 3H), 1.81-1.13 (m, 24H), 0.96 (d, J ) 6.4 Hz, 3H), 0.89 (s,
3H), 0.69 (s, 3H). 13C NMR (80% CDCl3,and 20% CD3OD): δ
118.63, 80.30, 79.20, 75.80, 66.36, 66.25, 66.13, 53.77, 49.35,
46.69, 45.88, 42.43, 41.43, 39.64, 39.64, 39.27, 39.09, 35.36,
35.02, 34.82, 34.57, 33.12, 31.96, 31.86, 31.57, 28.36, 27.55,
27.32, 27.21, 24.01, 23.07, 22.46, 22.23, 17.77, 16.60, 12.11.
HRFAB-MS (thioglycerol + H+ matrix): m/e ([M+ + H+])
671.5598, calcd 671.5588. The triamine (60.0 mg, 0.09 mmol)
in dry ethanol (10 mL) and HCl (aqueous, 37%, 1 mL) was
hydrogenated (800 psi) in the presence of PtO2 (5 mg, 0.02
mmol) at room temperature for 48 h. The catalyst was filtered

Antibacterial Activities of Antibiotics
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 3 669
(10) (a) Savage, P. B.; Li, C. Cholic acid derivatives: novel anti-
microbials. Exp. Opin. Invest. Drugs 2000, 9, 263-272. (b) Guan,
Q.; Schmidt, E. J .; Boswell, S. R.; Li, C.; Allman, G. W.; Savage,
P. B. Preparation and characterization of cholic acid-derived
antimicrobial agents with controlled stabilities. Org. Lett. 2000,
2, 2837-2840. (c) Li, C.; Budge, L. P.; Driscoll, C. D.; Willardson,
B. M.; Allman, G. W.; Savage, P. B. Incremental conversion of
outer-membrane permeabilizers into potent antibiotics for gram-
negative bacteria. J . Am. Chem. Soc. 1999, 121, 931-940. (d)
Rehman, A.; Li, C.; Budge, L. P.; Street, S. E.; Savage, P. B.
Tetrahedron Lett. 1999, 40, 1865. (e) Li, C.; Peters, A. S.;
Meredith, E. L.; Allman, G. H.; Savage, P. B. Design and
synthesis of potent sensitizers of Gram-negative bacteria based
on a cholic acid scaffolding. J . Am. Chem. Soc. 1998, 120, 2961-
2962.
(11) Schmidt, E. J .; Boswell, S. R.; Walsh, J . P.; Schellenberg, M.
M.; Winter, T. W.; Li, C.; Allman, G. W.; Savage, P. B. J .
Antimicrob. Chemother. 2001, 47, 671.
(12) Li, C.; Lewis, M. R.; Gilbert, A. B.; Noel, M. D.; Scoville, D. H.;
Allman, G. W.; Savage, P. B. Antimicrobial activities of amine-
and guanidine-functionalized cholic acid derivatives. Antimicrob.
Agents Chemother. 1999, 43, 1347-1349.
5.08 (s, 2H), 3.75 (s, 1H), 3.67-3.03 (m, 22H), 2.74-2.53 (m,
4H), 2.11-2.06 (m, 4H), 1.95-1.15 (m, 29H), 0.89 (d, J ) 6.84
Hz, 3H), 0.85 (s, 3H), 0.66 (s, 3H). 13C NMR (CDCl3): δ 157.28,
156.69, 136.66, 128.61, 128.37, 128.14, 127.90, 127.74, 127.49,
80.41, 79.58, 75.82, 67.37, 66.49, 64.78, 64.27, 64.11, 48.74,
48.50, 47.53, 46.48, 46.22, 45.67, 43.98, 42.49, 41.76, 39.91,
39.60, 38.25, 35.50, 35.22, 35.10, 34.88, 32.92, 29.64, 29.56,
29.48, 28.86, 27.67, 27.62, 27.45, 25.25, 23.32, 22.84, 22.65,
17.80, 12.39. HRFAB-MS (thioglycerol + H+ matrix): m/e ([M+
+ H+]) 1025.6683, calcd 1025.6664. The amine from the
previous step (40.0 mg, 0.04 mmol) in dry ethanol (10 mL) was
hydrogenated (800 psi) in the presence of PtO2 (5 mg, 0.02
mmol) at room temperature for 48 h. The catalyst was filtered
out, and the filtrate was concentrated under reduced pressure.
The residue was further hydrogenated (800 psi) in the presence
of Pd(OH)2 (40 mg, 20% on carbon) for 48 h. The catalyst was
filtrated out, and the filtrate was concentrated under reduced
pressure to give compound 4 (25.5 mg, 95%) as a yellow solid.
1H NMR (CD3OD): δ 3.92-261 (m, 26H), 2.13-1.11 (m, 33H),
13
0.88 (d, J ) 6.84 Hz, 3H), 0.86 (s, 3H), 0.63 (s, 3H). C NMR
(13) Zasloff, M. Magainins, a class of antimicrobial peptides from
Xenopusi skin: isolation, characterization of two active forms,
and partial cDNA sequence of a precursor. Proc. Natl. Acad. Sci.
U.S.A. 1987, 84, 5449-5453.
(CD3OD): δ 80.95, 79.66, 76.49, 65.55, 65.25, 64.95, 48.21,
48.06, 47.91, 46.20, 46.09, 45.45, 44.97, 44.82, 43.06, 41.77,
39.56, 38.28, 36.85, 35.12, 34.93, 34.83, 34.61, 32.44, 29.62,
29.46, 28.32, 28.26, 28.14, 27.43, 27.25, 26.25, 25.59, 24.20,
23.10, 22.39, 21.99, 17.33, 11.52. HRFAB-MS (thioglycerol +
H+ matrix): m/e ([M+ + H+]) 679.6224, calcd 679.6214.
(14) Steiner, J .; Hultmark, D.; Engstro¨m Å.; Bennich, H.; Bowmam,
H. G. Sequence and specificity of two antibacterial proteins
involved in insect immunity. Nature 1981, 292, 246-248.
(15) (a) Gesell, J .; Zasloff, M.; Opella, S. Two-dimensional 1H NMR
experiments show that the 23-residue magainin antibiotic
peptide is an R-helix in dodecylphosphocholine micelles, sodium
dodecyl sulfate micelles, and trifluoroethanol/water solution. J .
Biomol. NMR 1997, 9, 127-135. (b) Holak, T. A.; Engstro¨m, Å.;
Kraulis, P. J .; Lindeberg, G.; Bennich, H.; J ones, T. A.; Gronen-
born, A.; Clore, G. M. The solution conformation of the anti-
bacterial peptide cecropin A: a nuclear magnetic resonance and
dynamic simulated annealing study. Biochemistry 1988, 27,
7620-7629.
(16) For conformational study of a related amphiphile, see Taotafa,
U.; McMullin, D. M.; Lee, S. C.; Savage, P. B. Anionic facial
amphiphiles from cholic acid. Org. Lett. 2000, 2, 4117-4120.
(17) Breukink, E.; Wiedemann, I.; van Kraaij, C.; Kuipers, O. P.; Sahl,
H.-G.; de Kruijff, B. Use of the cell wall precursor lipid II by a
pore-forming peptide antibiotics. Science 1999, 286, 2361-2364.
(18) Sims, P. J .; Waggoner, A. S.; Wang, C.-H.; Hoffman, J . F. Studies
on the mechanism by which cyanine dyes measure membrane
potential in red blood cells and phosphatidylcholine vesicles.
Biochemistry 1974, 13, 3315-3330.
(19) (a) Oh, J .-T.; Cajal, Y.; Skowronska, E. M.; Belkin, S.; Chen, J .;
Van Dyk, T. K.; Sasser, M.; J ain, M. K. Cationic peptide
antimicrobials induce selective transcription of micF and osmY
in Escherichia coli. Biochim. Biophys. Acta 2000, 1463, 43-54.
(b) Oh, J .-T.; Cajal, Y.; Dhurjati, P. S.; Van Dyk, T. K.; J ain, M.
K. Cecropins induce the hyperosmotic stress response in Es-
cherichia coli. Biochim. Biophys. Acta 1998, 1415, 235-245.
(20) Oren, Z.; Hong, J .; Shai, Y. A comparative study on the structure
and function of a cytolytic R-helical peptide and its antimicrobial
â-sheet diastereomer. Eur. J . Biochem. 1999, 259, 360-369.
(21) Bacterial concentrations of 105 CFU are used in determining
MIC values using broth macrodilution techniques (National
Committee for Clinical Laboratory Standards; Methods for
Dilution Antimicrobial Susceptibility Tests for Bacteria that
Grow Aerobically, 4th ed.; Approved Standard M7-A4. NCCLS:
Villanova, PA, 1997).
Mea su r em en t of MIC a n d MHC Va lu es. These values
were measured as reported previously.11
Ack n ow led gm en t. Financial support from the Na-
tional Institutes of Health (GM 54619) and the National
Science Foundation (CAREER) is gratefully acknowl-
edged. The authors thank Drs. Mahendra K. J ain and
Samuel H. Gellman for insightful discussions and Dr.
Tina K. Van Dyk for providing the engineered bacterial
strains.
Refer en ces
(1) In memory of Glenn Walker Allman (1938-2000).
(2) For recent reviews, see (a) Van’t Hof, W.; Veerman, E. C. I.;
Helmerhorst, E. J .; Amerongen, A. V. N. Antimicrobial pep-
tides: properties and applicability. Biol. Chem. 2001, 382, 597-
619. (b) Hancock, R. E. W.; Chapple, D. S. Peptide antibiotics.
Antimicrob. Agents Chemother. 1999, 43, 1317-1323. (c) Lehrer,
R. I.; Ganz, T. Antimicrobial peptides in mammalian and insect
host defence. Curr. Opin. Immunol. 1999, 11, 23-27.
(3) (a) Hwang, P. M.; Vogel, H. J . Structure-function relationships
of antimicrobial peptides. Biochem. Cell Biol. 1998, 76, 235-
246. (b) Epand, R. M.; Vogel, H. J . Diversity of antimicrobial
peptides and their mechanisms of action. Biochim. Biophys. Acta
1999, 1462, 11-28.
(4) (a) McQuade, D. T.; Barrett, D. G.; Desper, J . M.; Hayashi, R.;
Gellman, S. H. Effects of amphiphilic topology on self-association
in solution, at the air-water interface, and in the solid state. J .
Am. Chem. Soc. 1995, 117, 4862-4896. (b) Cheng, Y.; Ho, D.
M.; Gottlieb, C. R.; Kahne, D. Facial amphiphiles. J . Am. Chem.
Soc. 1992, 114, 7319-7320.
(5) (a) Liu, D.; Degrado, W. F. De novo design, synthesis, and
characterization of antimicrobial â-peptides. J . Am. Chem. Soc.
2001, 123, 7553-7559. (b) Porter, E. A.; Wang, X.; Lee, H.-S.;
Weisblum, B.; Gellman, S. H. Nonheamolytic b-amino acid
oligomers. Nature 2000, 404, 565. (c) Hamuro, Y.; Schneider, J .
P.; DeGrado, W. F. De novo design of antibacterial â-peptides.
J . Am. Chem. Soc. 1999, 121, 12200-12201.
(6) Shai, Y. Mechanism of the binding, insertion and destabilization
of phospholipid bilayer membranes by R-helical antimicrobial
and cell nonselective membrane-lytic peptides. Biochim. Biophys.
Acta 1999, 1462, 55-70.
(22) Hancock, R. E. W. Therapeutic potential of cationic peptides.
Exp. Opin. Invest. Drugs 1998, 7, 167-174.
(23) (a) Mozsolits, H.; Wirth, H.-J .; Werkmeister, J .; Aguilar, M.-I.
Analysis of antimicrobial peptide interactions with hybrid bi-
layer membrane systems using surface plasmon resonance.
Biochim. Biophys. Acta 2001, 1512, 64-76. (b) Matsuzaki, K.;
Sugishita, K.; Fujii, N.; Miyajima, K. Molecular basis for
membrane selectivity of an antimicrobial peptide, magainin 2.
Biochemistry 1995, 34, 3423-3429.
(24) Kikuchi, K.; Bernard, E. M.; Sadownik, A.; Regen, S. L.;
Armstrong, D. Antimicrobial activities of squalamine mimics.
Antimicrob. Agents Chemother. 1997, 41, 1433-1438.
(25) Moore, K. S.; Wehrli, S.; Roder, H.; Rogers, M.; Forrest, J . N.,
J r.; McCrimmon, D.; Zasloff, M. Squalamine: an aminosterol
antibiotic from the shark. Proc. Natl. Acad. Sci. U.S.A. 1993,
90, 1354-1358.
(7) Matsuzaki, K. Why and how are peptide-lipid interactions
utilized for self-defense? Magainins and tachyplesins as arche-
types. Biochim. Biophys. Acta 1999, 1462, 1-10.
(8) Savage, P. B. Multidrug-resistant bacteria: overcoming anti-
biotic permeability barriers of Gram-negative bacteria. Ann.
Med. 2001, 33, 167-171.
(9) Sakai, N.; Gerard, D.; Matile, S. Electrostatics of cell membrane
recognition: structure and activity of neutral and cationic rigid
push-pull rods in isoelectric, anionic, and polarized lipid bilayer
membranes J . Am. Chem. Soc. 2001, 123, 2517-2524.
J M0105070