Novel des-fatty acyl-polymyxin B derivatives with Pseudomonas aeruginosa-specific antimicrobial activity

Article abstract of DOI:10.1248/cpb.59.597

Polymyxin B (PMB) is a cationic cyclic decapeptide antibiotic with a fatty acyl (FA) modification at the αamino group of Dab¹ (Dab: L- α, γ-diaminobutyric acid). In this study, which is part of a series of PMB structure-activity relationship in

Full text of DOI:10.1248/cpb.59.597

May 2011
Regular Article
Chem. Pharm. Bull. 59(5) 597—602 (2011)
597
Novel Des-Fatty Acyl-Polymyxin B Derivatives with Pseudomonas
aeruginosa-Specific Antimicrobial Activity
a
Yuki SATO, ,a Mitsuno SHINDO,b Naoki SAKURA,c Yoshiki UCHIDA,b and Ikuo KATO
*
a Faculty of Pharmaceutical Sciences, Hokuriku University; Kanagawa-machi, Kanazawa, Ishikawa 920–1181, Japan:
b Department of Health and Nutrition, Osaka Shoin Women’s University; Hishiyanishi, Higashi-Osaka, Osaka 577–8550,
Japan: and c Shizuoka Cancer Center Research Institute; Nagaizumi-cho, Sunto-gun, Shizuoka 411–8777, Japan.
Received November 26, 2010; accepted February 23, 2011; published online March 2, 2011
Polymyxin B (PMB) is a cationic cyclic decapeptide antibiotic with a fatty acyl (FA) modification at the a-
amino group of Dab1 (Dab: L-a,g-diaminobutyric acid). In this study, which is part of a series of PMB struc-
ture–activity relationship investigations focused on identifying clinically useful peptide antibiotics, we synthe-
sized ten des-FA PMB derivatives whose N-terminal moieties were changed to basic or hydrophilic amino acids.
The antimicrobial and lipopolysaccharide (LPS) binding activities of these synthetic analogs were tested. The
analogs showed more potent antimicrobial activity against Pseudomonas aeruginosa (P. aeruginosa) compared
with the PMB nonapeptide. In particular, [Ser2-Dap3]-PMB(2—10), Guanyl-[Thr2-Dab3]-PMB(2—10), Guanyl-
[Dab1-Thr2-Dab3]-PMB(1—10), and Na,g-diguanyl-[Dap3]-PMB(3—10) had antimicrobial activity equivalent to
PMB. In LPS binding assays, the displacement curves shifted in a manner proportional to the number of positive
charges available to bind to Escherichia coli (E. coli) and P. aeruginosa. Furthermore, peptides with basic side
chains were comparable to PMB in binding activity assays against E. coli and P. aeruginosa. The acute toxicities
of the peptides were evaluated by intravenously administering the peptides to mice through the tail vein. The tox-
icities of [Ser2-Dap3]-PMB(2—10), [Dap3]-PMB(3—10), and [Ser3]-PMB(3—10) were lower that of PMB (LD50,
4.8 mmol/kg).
Key words polymyxin B analog; antimicrobial activity; lipopolysaccharide binding activity; Pseudomonas aeruginosa speci-
ficity; toxicity; des-fatty acyl-polymyxin B
Polymyxin B (PMB) is a mixture of cationic cyclic de- analogs of Des-FA-[X1]-PMB showed potent and specific an-
capeptide antibiotics isolated from Bacillus polymyxia.1,2) timicrobial activity against P. aeruginosa (MIC: 0.5—
PMB contains six L-a,g-diaminobutyric acid (Dab) residues, 1 nmol/ml), comparable to PMB. In the present study, we
where the g-amino group of the Dab4 residue is acylated by synthesized more compact des-FA-PMB derivatives. We pre-
the C-terminal Thr10 to form a 23-member lactam ring. The dicted that these compact derivatives would have selective
N-terminus of a PMB molecule is acylated with a fatty acid and potent antimicrobial activity against P. aeruginosa and
such as 6-methyloctanoic acid (PMB1), 6-methylheptanoic low acute toxicity by virtue of their smaller amino acid side
acid (PMB2) or octanoic acid (PMB3), which forms a long chain (Ser, Dap) and shortened length of the linear peptide
hydrophobic tail.3,4) PMB exhibits potent antimicrobial activ- portion of des-FA-PMB. Previously, we reported that basic
ity against Gram-negative bacteria. This potency is due to the PMB decapeptides containing a basic tripeptide ([Dab-Dab-
ability of PMB to neutralize endotoxins and change the per- Dab1]- or [Arg-Arg-Arg1]-) in place of the Dab1 in des-FA-
meability of the outer membrane of Gram-negative bacteria. PMB also had potent activity against P. aeruginosa, but their
However, PMB peptides induce serious side effects such as toxicity was significantly higher than that of PMB (data not
nephrotoxicity and neuromuscular blockade, so their thera- shown). Therefore, we synthesized compact and basic des-
peutic applications are limited.3) Therefore, various PMB de- FA-PMB analogs with guanyl-Dab or Na,g-diguanyl-Dap in
rivatives have been synthesized and evaluated for biological place of Dab in order to reduce toxicity.
activity.5—8) It is well known that enzymatic removal of
the fatty acyl (FA)-Dab1 from PMB by ficin yields PMB
Experimental
nonapeptide (PMBN),9,10) which retains considerable
General HPLC was performed on an apparatus equipped with two 510
pumps (Waters Corp., Milford, MA, U.S.A.), a U6K injector (Waters), a
lipopolysaccharide (LPS)-binding activity11) but has reduced
Lambda-Max Model 481 LC Spectrophotometer (Waters), a 680 Automated
toxicity.12) However, PMBN has a lower antimicrobial activ-
Gradient Controller (Waters), and a Chromatocorder 21 (System Instruments
Co., Ltd., Tokyo, Japan). Gel filtration column chromatography was carried
out using Toyopearl HW-40-S (Tosoh Corp., Tokyo, Japan). Fast-atom bom-
bardment mass spectrometry (FAB-MS) was conducted on a JMS-DX300
mass spectrometer (JEOL Ltd., Tokyo, Japan). All reagents, peptide synthe-
sis solvents, and 9-fluorenylmethoxycarbonyl (Fmoc) amino acids were pur-
chased from Watanabe Chem., Ind., Ltd. (Hiroshima, Japan).
ity, with an minimum inhibitory concentration (MIC) value
of 256 nmol/ml towards Escherichia coli (E. coli).10) Our pre-
vious studies on the structure–antimicrobial activity relation-
ships of PMB showed that des-FA-PMB (Des-FA-[Dab1]-
PMB) had a considerable antimicrobial activity, with MIC
values of 8, 16, and 4 nmol/ml against E. coli, Salmonella
Typhimurium (S. Typhimurium), and Pseudomonas aerugi-
nosa (P. aeruginosa), respectively.13—17) N-terminal analogs
of Des-FA-[X1]-PMB (Xꢀany one of a number of amino
acids or peptides) were synthesized and examined for their
antimicrobial activity. Des-FA-[Dap1]-PMB (Dap; L-a,b-di-
Synthesis of Peptides PMB analogs ((1)—(10)) were synthesized ac-
cording to the route representatively shown for [Ser2-Dap3]-PMB(2—10)
(Fig. 1) (2). The synthetic strategy was essentially as reported previously.13)
In brief, the protected peptide was constructed on 4-hydroxylmethylphe-
noxymethyl-resin (HMP-resin or Wang-resin, Novabiochemläufelfingen,
Switzerland) via standard Fmoc solid phase chemistry using an ABI 433A
peptide synthesizer (Applied Biosystems, Foster City, CA, U.S.A.). Pro-
aminopropionic acid) and Des-FA-[Ser1]-PMB N-terminal tected amino acids used were Fmoc-Dab(2-ClZ)-OH, Fmoc-Dab(Boc)-OH,
∗ To whom correspondence should be addressed. e-mail: y-satoh@hokuriku-u.ac.jp
© 2011 Pharmaceutical Society of Japan

598
Vol. 59, No. 5
Fig. 1. Structure of Polymyxin B, Polymyxin B Nonapeptide, and Synthetic Peptides ((1)—(10))
Fmoc-Thr(Bzl)-OH, Fmoc-D-Phe-OH, Fmoc-Leu-OH, Fmoc-Ser(Bzl)-OH,
and Fmoc-Dap(2-ClZ)-OH. The Fmoc group was removed with 20% piperi-
dine in N-methylpyrolidone (NMP). As an example, the synthesis of [Ser2-
Dap3]-PMB(2—10) (2) and guanyl-[Dab1-Thr2-Dab3]-PMB(1—10) (5) are
described below.
similar to that described in (2).
Preparation of Guanyl-[Dab1-Thr2-Dab3]-PMB(1—10) (5) Guanyl-
[Dab1-Thr2-Dab3]-PMB(1—10) was synthesized essentially as described
above on Fmoc-Thr(Bzl)-O-HMP-resin (0.14 mmol, 100—200 mesh) using
appropriate Fmoc-derivatives. The desired protected peptide-resin, Fmoc-
Dab(2-ClZ)-Thr(Bzl)-Dab(2-ClZ)-Dab(Boc)-Dab(2-ClZ)-D-Phe-Leu-Dab(2-
ClZ)-Dab(2-ClZ)-Thr(Bzl)-O-HMP-resin (785 mg) thus obtained was
treated with TFA (9.5 ml) in the presence of H2O (0.5 ml) at room tempera-
ture for 1 h to give Fmoc-Dab(2-ClZ)-Thr(Bzl)-Dab(2-ClZ)-Dab4-Dab(2-
ClZ)-d-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)-OH (271 mg), which was
then chromatographed on a Toyopearl HW-40-S column (1.6ꢁ95 cm) using
DMF : H2O (9 : 1). The obtained linear, partially protected peptide (221 mg,
0.095 mmol) was cyclized with DPPA (103 ml, 0.48 mmol) and NMM (97 ml,
0.95 mmol) in the same manner as described above to yield Fmoc-Dab(2-
ClZ)-Thr(Bzl)-Dap(2-ClZ)-Dab4*-Dab(2-ClZ)-D-Phe-Leu-Dab(2-ClZ)-
Dab(2-ClZ)-Thr(Bzl)10* (*–*: amide bond between the a-COOH of Thr10
and the g-NH2 of Dab4) (173 mg). The Fmoc group of the protected cyclic
peptide was removed with 20% piperidine in DMF.
Preparation of [Ser2-Dap3]-PMB(2—10) (2) The peptide chain was
elongated on Fmoc-Thr(Bzl)-O-HMP-resin (0.25 mmol, 100—200 mesh)
and Fmoc-amino acid derivatives (1.0 mmol) were successively introduced
on the peptide chain using O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl-
uronium hexafluorophosphate (HATU) (1.0 mmol) as the coupling reagent.
After construction of the desired sequence, the protected peptide-resin,
Fmoc-Ser(Bzl)-Dap(2-ClZ)-Dab(Boc)-Dab(2-ClZ)-D-Phe-Leu-Dab(2-ClZ)-
Dab(2-ClZ)-Thr(Bzl)-O-HMP-resin (685 mg) was treated with trifluo-
roacetic acid (TFA) (9.5 ml) in the presence of H2O (0.5 ml) for 1 h at room
temperature, cleaving the peptide from the HMP-resin and removing the Ng-
Boc group from Dab4 to yield Fmoc-Ser(Bzl)-Dap(2-ClZ)-Dab4-Dab(2-
ClZ)-D-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)-OH. The excess TFA
was evaporated in vacuo and the residue was lyophilized from dioxane. The
product (433 mg) was dissolved in dimethylformamide (DMF) and purified
by gel filtration on a column (1.6ꢁ95 cm) of Toyopearl HW-40-S using
DMF : H2O (9 : 1) as the eluent. Fractions containing the desired product
were combined, DMF was evaporated, and the peptide was dissolved in
dioxane and lyophilized. The linear, partially protected [Ser2-Dap3]-
PMB(2—10) (381 mg, 0.19 mmol) was dissolved in ice-cold DMF–di-
methylsulfoxide (DMSO) (1 : 1), and then diphenyl phosphorazidate (DPPA)
(161 ml, 0.75 mmol) and 4-methylmorpholine (NMM) (150 ml, 1.50 mmol)
were added. The mixture was reacted for 18 h at 4 °C to form the lactam
ring. The cyclized product Fmoc-Ser(Bzl)-Dap(2-ClZ)-Dab4*-Dab(2-ClZ)-
D-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)10* (*–*: amide bond between
the a-COOH of Thr10 and the g-NH2 of Dab4) was purified on a Toyopearl
HW-40-S column (1.6ꢁ95 cm). Fractions containing the designed product
(362 mg) were combined, the solvent was evaporated, and the peptide was
re-solubilized in dioxane and lyophilized. The Fmoc group was removed
with 20% piperidine in DMF for 5 min at room temperature, the solvent was
evaporated, and the peptide was again re-solubilized in dioxane and
lyophilized. The obtained product was treated with anhydrous HF (5 ml) in
the presence of anisole (500 ml) for 1 h at 0 °C. After evaporation of the HF,
the residue was dissolved in H2O, washed with three portions of ether, then
lyophilized. The crude product (175 mg) was further purified by HPLC em-
ploying a Capcell Pak C18 UG-80 column (2ꢁ15 cm, Shiseido Co., Ltd.,
Tokyo, Japan) using an acetonitrile (CH3CN)–0.1% TFA solvent system.
The main product was collected, the fractions were combined, the solvent
was evaporated, and the product was re-solubilized in water and lyophilized.
The product (121 mg) was chromatographed on a Toyopearl HW-40-S col-
umn (1.5ꢁ57 cm) using 25% CH3CN in 5 mmol/l HCl to afford the final
preparation of compound (1) (108 mg, 0.11 mmol).
Guanylation of the cyclic partially protected peptide was performed in the
following manner.18) To a solution of the cyclic, partially protected peptide
(73 mg, 0.035 mmol) in DMF (1 ml), NMM (7.9 ml, 0.077 mmol), HgCl2
(10.5 mg, 0.039 mmol) and 1,3-bis(benzyloxycarbonyl)-2-methyl-2-thio-
pseudourea (13.8 mg, 0.039 mmol) were added at 0 °C. The reaction mixture
was stirred for 18 h at 4 °C, and then centrifuged. The supernatant was puri-
fied by gel filtration on a column (1.6ꢁ95 cm) of Toyopearl HW-40-S using
DMF as the eluent. Fractions containing the designed peptide were collected
and lyophilized. The resulting product (55 mg) was treated with anhydrous
HF (3 ml) containing anisole (0.3 ml) for 1 h at 0 °C. The crude peptide
(29 mg) thus obtained was purified by HPLC on a Capcell Pak C18 UG-80
column (2ꢁ15 cm) using a CH3CN–0.1% TFA solvent system. The product
(20 mg) was chromatographed on
(1.5ꢁ57 cm) using 25% CH3CN in 5 mM HCl to yield the desired compound
(5) (16 mg, 0.014 mmol).
The other guanyl PMB derivatives ((6)—(10)) were synthesized in a simi-
lar manner as described above. The purity of each synthetic peptide was
confirmed by analytical HPLC and FAB-MS analysis (Table 1).
Bacteria, and Antimicrobial Assay E. coli IFO 12734, S. Ty-
phimurium IFO 12529 and P. aeruginosa IFO 3080 were purchased from the
Institute for Fermentation, Osaka (IFO), Japan. These bacterial strains were
grown overnight at 37 °C on nutrient agar medium and harvested in sterile
saline. The densities of the bacterial suspensions were determined at 600 nm
using a standard curve relating absorbance to the number of colony-forming
units (CFU). The antibacterial activity of each synthetic peptide was evalu-
ated by comparing its activity with that of commercially available PMB
(Sigma Chemical Co., St. Louis, MO, U.S.A.). The MIC of each synthetic
peptide against each of the bacterial strains was determined using a standard
microplate dilution method (nꢀ6—8). One hundred microliters of each pep-
a Toyopearl HW-40-S column
The other PMB derivatives ((1), (3) and (4)) were synthesized in a manner

May 2011
599
Table 1. FAB-MS Data of the Synthetic Peptides ((1)—(10))
[MꢄH]ꢄ, m/z
Formula
Calculated
Found
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
Des-FA-[Dap1-Ser2-Dap3]-PMB(1—10)
[Ser2-Dap3]-PMB(2—10)
C44H76N16O12
C41H70N14O11
C38H65N13O9
C38H64N12O10
C48H84N18O12
C44H76N16O11
C44H77N17O10
C40H69N15O9
C36H61N13O8
C40H69N17O9
1021
935
848
1021
935
848
[Dap3]-PMB(3—10)
[Ser3]-PMB(3—10)
849
849
Guanyl-[Dab1-Thr2-Dab3]-PMB(1—10)
Guanyl-[Thr2-Dab3]-PMB(2—10)
Guanyl-[Dab2-Dab3]-PMB(2—10)
Guanyl-[Dab3] -PMB(3—10)
Guanyl-PMB(4—10)
1105
1005
1004
904
804
932
1105
1005
1004
904
804
932
N
a,g-diguanyl-[Dap3]-PMB(3—10)
Table 2. Antimicrobial Activity of the Synthetic Peptides ((1)—(10))
MIC (nmol/ml)
E. coli
S. Typhimurium
P. aeruginosa
Polymyxin B (PMB)
1
ꢃ256
16
0.5
ꢃ256
32
1
128
2
1
4
2
1
1
2
Polymyxin B nonapeptide (PMBN)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
Des-FA-[Dap1-Ser2-Dap3]-PMB(1—10)
[Ser2-Dap3]-PMB(2—10)
64
64
[Dap3]-PMB(3—10)
256
ꢃ256
4
256
ꢃ256
2
[Ser3]-PMB(3—10)
Guanyl-[Dab1-Thr2-Dab3]-PMB(1—10)
Guanyl-[Thr2-Dab3]-PMB(2—10)
Guanyl-[Dab2-Dab3]-PMB(2—10)
Guanyl-[Dab3] -PMB(3—10)
Guanyl-PMB(4—10)
2
8
2
8
16
ꢃ256
16
16
ꢃ256
32
2
4
0.5
N
a,g-diguanyl-[Dap3]-PMB(3—10)
tide serially diluted with distilled water to 0.25—256 nmol/ml was added to
a mixture of 10 ml of bacterial suspension (approximately 106 CFU/ml) and
90 ml of Mueller–Hinton broth (Becton Dickinson and Company Sparks,
Cockeysville, MD, U.S.A.) in each well of a flat-bottom microplate (Corning
Inc., Corning, NY, U.S.A.). The plates were incubated overnight at 37 °C to
allow evaluation of the MIC. The MIC value was expressed as the lowest
final concentration (nmol/ml) at which no growth was observed (Table 2).
Assay of LPS Binding of the Synthetic Peptides As described previ-
ously,13) a solution of [Dab(Ng-dansyl-Gly)1]-PMB3 in H2O (1 mmol/ml)
Table 3. Acute Toxicity of the Synthetic Peptides ((2)—(6) and (10))
LD50 value 95% confidence
(mmol/kg) interval (mmol/kg)
Polymyxin B (Sigma)
4.8
31.5
40.9
ꢃ50
ꢃ50
5.4
4.3—5.3
25.5—39.0
36.0—49.7
—
Polymyxin B nonapeptide (PMBN)
(2) [Ser2-Dap3]-PMB(2—10)
(3) [Dap3]-PMB(3—10)
(4) [Ser3]-PMB(3—10)
—
(4 ml, 4 nmol) was added to
a quartz cuvette containing N-(2-
(5) Guanyl-[Dab1-Thr2-Dab3]-PMB(1—10)
(6) Guanyl-[Thr2-Dab3]-PMB(2—10)
(10) Na,g-diguanyl-[Dap3]-PMB(3—10)
2.8—12.1
4.3—11.5
6.2—16.7
hydroxyethyl)piperazine-Nꢂ-(2-ethanesulfonic acid) buffer (HEPES;
5 mmol/l, pH 7.2) (1 ml), followed by a solution of E. coli LPS (serotype
055:B5) (Sigma) in H2O (3 mg/ml) (10 ml, 30 mg). The solution was kept at
30 °C for 60 min, by which time the fluorescence intensity of [Dab(Ng-dan-
syl-Gly)1]-PMB3 reached a plateau. A solution of each PMB3 analog
(1 mmol/ml) (4 ml aliquots) was added cumulatively to the quartz cuvette at
5-min intervals to obtain eight data points (4—32 nmol). The change in fluo-
rescence intensity was measured after each addition at an excitation wave-
length of 330 nm and an emission wavelength of 490 nm using an F-850 flu-
orescence spectrophotometer (Hitachi Instrument Co., Tokyo, Japan). The
initial intensity of fluorescence was taken as 100%. The percent fluorescence
intensity was plotted as a function of the concentration of peptide. The bind-
ing experiments were repeated at least three times for each peptide to estab-
lish the reproducibility of the results. To study the inhibitory effect on P.
aeruginosa LPS (Sigma), a larger volume of [Dab(Ng-dansyl-Gly)1]-PMB3
in H2O (1 mmol/ml) (20 ml, 20 nmol) was added to the quartz cuvette.
7.0
10.4
Analogs (1) and (7)—(9) were not evaluated.
ber of mice dying following each dose.
All animal procedures were approved by the Animal Care and Use Com-
mittee of Hokuriku University.
Results and Discussion
Synthesis of Peptides Synthesis of des-FA-PMB
analogs ((1)—(10)) was carried out as reported previously
(Fig. 1).13,17) The product was extensively purified by HPLC
and gel-filtration prior to examination of its biological activ-
ity. The purity of the synthetic peptides was ꢃ98%.
Antimicrobial Activity The antimicrobial activity of the
synthetic peptides against E. coli, S. Typhimurium and P.
aeruginosa is summarized in Table 2.
Acute Toxicity The LD50 of each synthetic peptide (PMB, PMBN,
(2)—(6) and (10)) was determined in male ddY mice (4 weeks old; Japan
SLC, Hamamatsu, Japan). The following peptide solutions in saline were
prepared: PMB sulfate (Sigma); 0.25, 0.50, 0.75, and 1.0 mmol/ml, and
PMBN; 2.0, 2.5, 3.0, 3.5, and 4.0 mmol/ml. The concentration of each syn-
thetic peptide solution was determined from the results of a pretest. The so-
lutions were injected through the lateral tail vein at a rate of 0.1 ml/30 s to a
volume of 0.1 ml/10 g body weight. Five to ten mice were used per dose. The
The MIC values of Des-FA-[Dap1-Ser2-Dap3]-PMB(1—
LD50 was determined by the Litchfield–Wilcoxon method19) from the num- 10) (1), [Ser2-Dap3]-PMB(2—10) (2), [Dap3]-PMB(3—10)

600
Vol. 59, No. 5
(3), and [Ser3]-PMB(3—10) (4) were respectively 16, 64, membranes were designed. Only Guanyl-PMB(4—10) (10),
256, and ꢃ256 nmol/ml against E. coli and 32, 64, 256, and which is the lowest molecular weight analog studied, had an-
ꢃ256 nmol/ml against S. Typhimurium. Compared with timicrobial activity against P. aeruginosa.
these hydrophilic small des-FA-PMB analogs containing Ser
The difference of their antimicrobial specificity against the
and/or Dap ((1)—(4)), the peptides with shorter chain three bacterial species tested can be largely explained by
lengths were less active against E. coli and S. Typhimurium, structural differences in the bacterial cell membranes, espe-
but retained potency against P. aeruginosa (MIC: 1— cially the lipid A portions, which anchor the LPS molecule to
4 nmol/ml). Des-FA-[Dap1-Ser2-Dap3]-PMB(1—10) (1) has the lipid layer of the membranes.23) Lipid A of E. coli and S.
lower antimicrobial activity (MIC: 2 nmol/ml) than PMB Typhimurium is largely composed of five C14 fatty acids and
against P. aeruginosa. [Ser2-Dap3]-PMB(2—10) (2) retained one C12 fatty acid, whereas that of P. aeruginosa is com-
the same antimicrobial activity against P. aeruginosa (MIC: posed of four C12 and two C10 fatty acids. The membrane
1 nmol/ml) as PMB (MIC: 1 nmol/ml), but showed less activ- lipid components of P. aeruginosa are less hydrophobic than
ity against E. coli and S. Typhimurium (MIC: 64 nmol/ml). those of the two other bacteria, so the cell membrane struc-
Despite having the same number of amino acids, the antimi- ture of P. aeruginosa is less compact. This allows for hy-
crobial activity of [Ser2-Dap3]-PMB(2—10) (2) was different drophobic interactions between the D-Phe and Leu side
from that of PMBN against the three bacteria. [Ser2-Dap3]- chains of PMB analogs and the outer membrane, resulting in
PMB(2—10) had 64-fold higher activity than PMBN against membrane disorder that leads to cell death. Absence of the
P. aeruginosa. We reported earlier that PMB octapeptide N-terminal FA group in PMB is apparently not essential for
(PMB(3—10)) has low antimicrobial activity against E. coli, antimicrobial activity towards P. aeruginosa.
S. Typhimurium, and P. aeruginosa (MIC values: 128, 256,
Tsubery et al.7) showed that the structures of PMB and
and 64 nmol, respectively).13) [Dap3]-PMB(3—10) (3) and PMBN bound to LPS were modeled based on coordinates
[Ser3]-PMB(3—10) (4) had high antimicrobial activity (MIC provided for PMB by Pristovsˇek and Kidricˇ.24) Elimination of
values of 4 and 2 nmol/ml, respectively) compared with PMB the positively charged Dab1 residue, which interacts with the
octapeptide against P. aeruginosa, whereas the MIC values of negatively charged phosphate group in the PMB-LPS com-
these analogs were very low against E. coli and S. Ty- plex, is not likely to significantly weaken the LPS-peptide in-
phimurium. These results show that small hydrophilic amino teraction. This is because of the conformational change of
acids (Dap, Ser) at the N-terminal are essential for antimicro- the side chain at the Dab3 residue, allowing for the efficient
bial specificity towards P. aeruginosa.17) Shortening the side interaction of the Dab3 residue with the negatively charged
chain by a single methylene results in increased activity.
phosphate group. We considered that our synthetic analogs
In PMB derivatives with an introduced guanyl group, induce a conformational change of the LPS-peptide interac-
Guanyl-[Dab1-Thr2-Dab3]-PMB(1—10) (5) retained antimi- tion.
crobial activity against E. coli, S. Typhimurium, and P.
Our previous study shows that PMB octapeptide
aeruginosa (MIC: 4, 2, and 1 nmol/ml, respectively), and its (PMB(3—10)) has low antimicrobial activity against P.
activity was higher than that of des-FA-PMB decapeptide aeruginosa, with an MIC value of 64 nmol/ml.13) [Dap3]-
(MIC: 8, 16, and 4 nmol/ml against the same three bacteria, PMB(3—10) (3) and [Ser3]-PMB(3—10) (4) showed high
previously reported).13) Guanyl-[Thr2-Dab3]-PMB(2—10) (6) antimicrobial activities, with MIC values of 4 and 2 nmol/ml,
and Guanyl-[Dab2-Dab3]-PMB(2—10) (7) have much higher respectively, compared with PMB octapeptide against P.
activity than PMBN (MIC values of ꢃ256, ꢃ256, and aeruginosa. These results show that small hydrophilic amino
128 nmol/ml, respectively) against the three bacteria. acids (Dap, Ser) at the N-terminal are essential for antimicro-
Guanyl-[Dab3] -PMB(3—10) (8) and Na,g-diguanyl-[Dap3]- bial specificity towards P. aeruginosa.17) We consider that
PMB(3—10) (10) also have higher activity (MIC: 16, 16, shortening a single methylene in the side chain of Dab and/or
and 2 nmol/ml and 16, 32, and 0.5 nmol/ml, respectively) Thr allows for the efficient interaction of the linear portion in
than Des-FA-PMB octapeptide (MIC: 128, 256, and PMB analogs with the negatively charged phosphate group.
64 nmol/ml, respectively, against the same three bacteria).9) The differences may be related to the distance between the
Guanyl-PMB (4—10) (9) has less activity (MIC: ꢃ256 phosphate group in LPS and the linear portion of the PMB
nmol/ml), similar to that of Des-FA-PMB(4—10) (MIC: analog. The guanyl group is a strong basic group. The strong
ꢃ256 nmol/ml) against E. coli and S. Typhimurium, but re- interaction between the negatively charged phosphate group
tains its potency against P. aeruginosa (MIC: 4 nmol/ml). in LPS and the positively charged portion of the guanyl PMB
Their guanylated analogs exhibited reduced antimicrobial ac- derivative influences the antimicrobial activity.
tivity against E. coli and S. Typhimurium as compared with
LPS Binding Activity The LPS-binding activities of the
PMB, but their activities were higher than those of the Ser synthetic peptides were evaluated by measuring the displace-
and Dap analogs. The guanylated peptides showed potent an- ment of [Dab(Ng-dansyl-Gly)1]-PMB3 bound to E. coli and
timicrobial activity against P. aeruginosa (MIC: 0.5— P. aeruginosa LPS, using the methodology reported previ-
4 nmol/ml). Na,g-diguanyl-[Dap3]-PMB(3—10) (10) was 2- ously.13) The mechanism by which these synthetic peptides
fold more potent than PMB against P. aeruginosa. The MIC exclude the fluorescent probe bound to LPS could be ex-
value of Guanyl-PMB(4—10) (9) against P. aeruginosa was plained as follows.25) First, the N-terminal cationic portion of
4 nmol/ml. In a previous study we and Urakawa et al. found the peptide binds to the phosphate anions of lipid A and/or
that PMB heptapeptide and its derivatives have no significant the carboxylate anions of the adjacent 3-deoxy-D-manno-oc-
antimicrobial activity against Gram-negative bacteria, but the tulosonic acid (Kdo) portion of the LPS molecule. This ionic
binding activity to LPS is retained.13,22) Therefore, PMB hep- binding causes structural changes in LPS, resulting in the
tapeptide analogs with better permeability into the outer loss of hydrophobic binding between the FA group(s) of the

May 2011
601
Fig. 2. Displacement of [Dab(Ng-dansyl-Gly)1]-PMB3 Bound to E. coli LPS by Synthetic Peptides ((1)—(10))
Fig. 3. Displacement of [Dab(Ng-dansyl-Gly)1]-PMB3 Bound to P. aeruginosa LPS by Synthetic Peptides ((1)—(10))
lipid A portion of LPS and the octanoyl-Dab1 portion of the crobial activity is presumed that the difference of interaction
probe, [Dab(Ng-dansyl-Gly)1]-PMB3.
of the peptides with cell-free LPS or cell-bound LPS.
N-terminal PMB analogs with Ser and/or Dap ((1)—(4))
Acute Toxicity The acute toxicities of (2)—(6) and (10)
had low E. coli LPS binding activity compared with PMB compared with PMB and PMBN were examined by tail intra-
(Fig. 2). The binding ability was directly proportional to the venous bolus injection in mice. These injections were lethal
number of positive charges on the peptide that could interact within 24 h. The administration of PMB and its analogs in-
with the E. coli LPS. Furthermore, these results indicate that duced respiratory arrest and death within 5 min of injection.
the LPS binding activity of compact hydrophilic analogs par- This effect may be caused by neuromuscular blockade.20,21)
alleled their antimicrobial activity against E. coli. Guanyl- The LD50 values of [Ser2-Dap3]-PMB(2—10) (2), [Dap3]-
PMB analogs ((6)—(10)) had potent LPS binding activity ex- PMB(3—10) (3), and [Ser3]-PMB(3—10) (4) were 40.9
cept for Guanyl-PMB(4—10) (9), and the binding activities mmol/kg, ꢃ50 mmol/kg, and ꢃ50 mmol/kg, respectively. The
of Guanyl-[Dab2-Dab3]-PMB(2—10) (7), Guanyl-[Dab1- results demonstrate that the acute toxicity of these peptides
Thr2-Dab3]-PMB(1—10) (5), and
PMB(3—10) were especially potent, equivalent to PMB 4.8 mmol/kg) (Table 2), and furthermore, they were less toxic
binding activity against E. coli. than PMBN (LD50: 31.5 mmol/kg). The toxicities of the
N
a,g-diguanyl-[Dap3]- in mice was lower compared to native PMB (LD50:
In the case of P. aeruginosa, N-terminal PMB analogs with guanylated peptides were higher than that of PMBN. Na,g
-
Ser and/or Dap ((1)—(4)) had slightly lower LPS binding ac- diguanyl-[Dap3]-PMB(3—10) (10) showed low toxicity com-
tivity compared with PMB, whereas the Guanyl-PMB pared with PMB.
analogs had high LPS binding activity, except for (9) (Fig. 3).
In the present study, we synthesized hydrophilic and com-
However, unlike with the N-terminal analogs, the binding ac- pact des-FA-PMB analogs (Fig. 1). We developed N-terminal
tivities of the Guanyl-PMB analogs did not correlate with analogs of PMB lacking a FA moiety and demonstrated that
their antimicrobial activities, showing that the FA portion of these analogs have antimicrobial activity specific towards P.
the molecule plays only a secondary role in LPS binding ac- aeruginosa. In [Ser2-Dap3]-PMB(2—10) (2), introduction of
tivity towards P. aeruginosa.
a small amino acid (Ser, Dap) resulted in potent antimicro-
The difference between LPS binding activity and antimi- bial activity specific towards P. aeruginosa (MIC: 1 nmol/ml,

602
Vol. 59, No. 5
9) Chihara S., Tobita T., Yahata M., Ito A., Koyama Y., Agric. Biol.
Chem., 37, 2455—2463 (1973).
10) Vaara M., Vaara T., Antimicrob. Agents Chemother., 24, 107—113
(1983).
11) Ofek I., Cohen S., Rahmai R., Kabha K., Tamarkin D., Herzig Y., Ru-
binstein E., Antimicrob. Agents Chemother., 38, 374—377 (1994).
12) Danner R. L., Joiner K. A., Rubin M., Patterson W. H., Johnson N.,
Ayers K. M., Parrillo J. E., Antimicrob. Agents Chemother., 33, 1428—
1434 (1989).
13) Sakura N., Itoh T., Uchida Y., Ohki K., Okimura K., Chiba K., Sato Y.,
Sawanishi H., Bull. Chem. Soc. Jpn., 77, 1915—1924 (2004).
14) Okimura K., Ohki K., Sato Y., Ohnishi K., Uchida Y., Sakura N., Bull.
Chem. Soc. Jpn., 80, 543—552 (2007).
equal to that of PMB) and very low acute toxicity (8.5-fold
less toxic than PMB). The guanylated analogs maintained an-
timicrobial potency, especially against P. aeruginosa, but
they were as toxic as PMB. Of particular interest is [Ser3]-
PMB(3—10) (4), a PMB octapeptide derivative. Our results
show that it is a non-toxic, compact analog with highly selec-
tive antimicrobial activity towards P. aeruginosa. The im-
proved properties of this analog provide a useful base for the
development of PMB analogs with potent antimicrobial ac-
tivity and low toxicity. These results provide clues for devel-
oping a potent antimicrobial analog with low acute toxicity.
15) Okimura K., Ohki K., Sato Y., Ohnishi K., Sakura N., Chem. Pharm.
Bull., 55, 1724—1730 (2007).
Acknowledgement We are grateful to Dr. Naoki Asano for helpful dis-
16) Kanazawa K., Sato Y., Ohki K., Okimura K., Uchida Y., Shindo M.,
Sakura N., Chem. Pharm. Bull., 57, 240—244 (2009).
17) Katsuma N., Sato Y., Ohki K., Okimura K., Ohnishi K., Sakura N.,
Chem. Pharm. Bull., 57, 332—336 (2009).
18) Chandrakumar N. S., Synthetic Commun., 26, 2613—2616 (1996).
19) Litchfield J. T., Wilcoxon F., J. Pharmacol. Exp. Ther., 96, 99—113
(1949).
20) Viswanath D. V., Jenkins H. J., J. Pharm. Sci., 67, 1275—1280 (1978).
21) Singh Y. N., Marshall I. G., Harvey A. L., Br. J. Anaesth., 54, 1295—
1306 (1982).
22) Urakawa H., Yamada K., Komagoe K., Ando S., Oku H., Katsu T.,
Matsuo I., Bioorg. Med. Chem. Lett., 20, 1771—1775 (2010).
23) Raetz C. R., Whitfield H., Annu. Rev. Biochem., 71, 635—700 (2002).
24) Pristovsˇek P., Kidricˇ J., J. Med. Chem., 42, 4604—4613 (1999).
25) Soon R. L., Velkov T., Chiu F., Thompson P. E., Kancharla R., Roberts
K., Larson I., Nation R. L., Li J., Anal. Biochem., 409, 273—283
(2011).
cussions.
References
1) Ainsworth G. C., Brown A. M., Brownlee G., Nature (London), 160,
263—264 (1947).
2) Benedict R. G., Langlykke A. F., J. Bacteriol., 54, 24—25 (1947).
3) Storm D. R., Rosenthal K. S., Swanson P. E., Ann. Rev. Biochem., 46,
723—763 (1977).
4) Suzuki T., Hayashi K., Fujikawa K., Tsukamoto K., J. Biochem.
(Tokyo), 56, 335—343 (1964).
5) Tsubery H., Ofek I., Cohen S., Fridkin M., J. Med. Chem., 43, 3085—
3092 (2000).
6) Tsubery H., Ofek I., Cohen S., Fridkin M., Biochemistry, 39, 11837—
11844 (2000).
7) Tsubery H., Ofek I., Cohen S., Eisenstein M., Fridkin M., Mol. Phar-
macol., 62, 1036—1042 (2002).
8) Tsubery H., Yaakov H., Cohen S., Giterman T., Matityahou A., Fridkin
M., Ofek I., Antimicrob. Agents Chemother., 49, 3122—3128 (2005).