Contribution of each amino acid residue in polymyxin b3 to antimicrobial and lipopolysaccharide binding activity

Article abstract of DOI:10.1248/cpb.57.240

This study on the structure-activity relationship of polymyxin B, a cyclic peptide antibiotic, used sixteen synthetic polymyxin B₃ analogs including alanine scanning analogs to elucidate the contribution of the side chains to antimicrobial acti

Full text of DOI:10.1248/cpb.57.240

240
Chem. Pharm. Bull. 57(3) 240—244 (2009)
Vol. 57, No. 3
Contribution of Each Amino Acid Residue in Polymyxin B3 to
Antimicrobial and Lipopolysaccharide Binding Activity
Kazushi KANAZAWA,a Yuki SATO,a Kazuhiro OHKI,a Keiko OKIMURA,a Yoshiki UCHIDA,b
Mitsuno SHINDO,b and Naoki SAKURA*,a
a Faculty of Pharmaceutical Sciences, Hokuriku University; Kanagawa-machi, Kanazawa 920–1181, Japan: and
b Department of Food Science and Nutrition, Osaka Shoin Women’s University; Hishiyanishi, Higashi-Osaka 577–8550,
Japan. Received August 25, 2008; accepted December 25, 2008; published online January 9, 2009
This study on the structure–activity relationship of polymyxin B, a cyclic peptide antibiotic, used sixteen
synthetic polymyxin B3 analogs including alanine scanning analogs to elucidate the contribution of the side
chains to antimicrobial activity and lipopolysaccharide (LPS) binding. Of these analogs, [Ala5]-polymyxin B3
showed greatly reduced antimicrobial activity against Escherichia coli (E. coli), Salmonella Typhimurium (S. Ty-
phimurium) and Pseudomonas aeruginosa (P. aeruginosa) with MIC values of 4—16 nmol/ml, suggesting that the
Dab (a,g-diaminobutyric acid) residue at position 5 is the most important residue contributing to bactericidal
activity. The antibacterial contribution of Dab when located within the lactam ring (positions 5, 8 and 9) was
greater than when located outside the ring (positions 1 and 3). [D-Ala6]-, [L-Phe6]-, [Ala7]-, and [Gly7]-polymyxin
B3 analogs retained potent antimicrobial activity, indicating that neither the reduction of hydrophobic character
of the D-Phe6-Leu7 region nor the D-configuration at position 6 is indispensable for antimicrobial activity. LPS
binding studies showed that decreased hydrophobicity of the lactam ring had little effect, but the Ng-amino func-
tion of the Dab residues at position 1, 3, 5, 8 and 9 greatly affected LPS binding, with the contribution of Dab5
being the most significant.
Key words polymyxin B; alanine scanning; antimicrobial activity; lipopolysaccharide binding activity; diaminobutyric acid
Polymyxin B is an N-terminally fatty acylated peptide an- Japan). Gel column chromatography was carried out using Toyopearl HW-
tibiotic isolated from Bacillus polymyxa.1,2) Polymyxin B
40-S (Tosoh Corporation, Tokyo, Japan). Fast-atom bombardment mass
spectra (FAB-MS) were obtained on a JMS-DX300 mass spectrometer
(JEOL Ltd., Tokyo, Japan). Amino acid analysis of peptide acid hydrolysates
was conducted on a model 7300 amino acid analyzer (Beckman Instruments
contains six a,g-diaminobutyric acid (Dab) residues. The g-
amino group of Dab4 is acylated by C-terminal Thr10 to form
a 23-member lactam ring3,4) resulting in a peptide that has
Ltd., Fullerton, CA, U.S.A.). The optical rotations of the peptides were
measured with a DIP-370 digital polarimeter (Nippon Bunko Co., Ltd.,
Tokyo, Japan). Deprotection reaction of protected peptides with anhydrous
HF was carried out in a Teflon HF apparatus (Peptide Institute Inc., Osaka,
Japan). HP-TLC was performed on precoated silica gel plates (Kieselgel 60;
Merck, Darmstadt, Germany). All reagents, solvent used for peptide synthe-
sis and Fmoc-amino acids were obtained from Watanabe Chem. Ind. Ltd.,
Hiroshima, Japan.
Synthesis of Peptides. 1) Solid Phase Synthesis of Protected Pep-
tide-Resins (1R—16R) Polymyxin B3 analogs (1—16, Fig. 1) were synthe-
sized according to the route representatively shown for [Ala3]-polymyxin B3
(3, Fig. 2). The synthetic strategy was essentially as reported previously.16)
In brief, the protected peptide was constructed on 4-hydroxylmethylphe-
noxymethyl-resin (HMP-resin or Wang-resin, 0.74 mmol/g, Novabiochem-
läufelfingen, Switzerland) by a solid phase methodology using an ABI 433A
peptide synthesizer (Applied Biosystems, Foster City, CA, U.S.A.). Pro-
tected amino acids used were Fmoc-Dab(2-ClZ)-OH, Fmoc-Dab(Boc)-
OH, Fmoc-Dab(Ac)-OH, Fmoc-Thr(Bzl)-OH, Fmoc-Phe-OH, Fmoc-D-Phe-
antimicrobial activity against Gram-negative bacteria. The
mechanism of action of Polymyxin B is believed to be due to
the amphiphilic character of the molecule, with the basic
Dab side chains interacting with the negative charges of the
lipid A portion of lipopolysaccharides (LPS) on the cell sur-
face of bacteria, and the hydrophobic D-Phe-Leu in the lac-
tam ring and the fatty acyl group at the N-terminus interact-
ing with the lipophilic part of LPS.5,6) These interactions lead
to disordering of Gram-negative bacterial cell membranes,
resulting in cell death.7) Since the total solid phase synthesis
of polymyxin B1 was first reported by Sharma in 1999,8) vari-
ous polymyxin B analogs have been synthesized and evalu-
ated for biological activity.9—17) However, the structure–activ-
ity relationship of polymyxin B peptides is not understood in
detail, due to the lack of extensive works employing highly OH, Fmoc-Leu-OH, Fmoc-Ala-OH, Fmoc-Gly-OH, Fmoc-Trp(Boc)-OH
and Fmoc-D-Trp(Boc)-OH. Starting from Fmoc-Thr(Bzl)-O-HMP-resin (0.2
mmol, 196 mg), the Fmoc group was removed with 20% piperidine in N-
methylpyrolidone (NMP), and the peptide chain was sequentially elongated
with the appropriate Fmoc-amino acid (1.0 mmol), O-(7-azabenzotriazol-1-
pure peptides. We previously reported a study aimed at clari-
fying the contribution of the N-terminal fatty acyl groups of
various polymyxin B family peptides to biological activity, as
well as the development of N-terminal analogs without fatty
acyl groups.16,18,19) The aim of the present study was to clar-
ify the structural requirements of the side chain of each
amino acid residue by means of alanine scanning, and to fur-
ther examine the role of the hydrophobic portion (D-Phe-Leu)
of the polymyxin B3 lactam ring in antimicrobial activity and
LPS binding, employing sixteen synthetic peptides (Fig. 1).
yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HATU) (1.0 mmol)
and diisopropylethylamine (2.0 mmol) in NMP. Representatively, for the
preparation of [Ala3]-polymyxin B3 (3), after the introduction of octanoic
acid to the a-amino function of Dab(2-ClZ) at position 1, the protected pep-
tide-resin was washed successively with three portions of dimethylform-
amide (DMF), dichloromethane, MeOH and ether, and dried in vacuo to
give octanoyl-Dab(2-ClZ)-Thr(Bzl)-Ala3-Dab(Boc)-Dab(2-ClZ)-D-Phe-Leu-
Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)-O-HMP-resin (3R). Yield, 511 mg. For the
preparation of the other analogs (1, 2 and 4—16), various protected peptide-
resins (1R, 2R and 4R—16R) were constructed in the same manner as de-
scribed for 3R.
Experimental
General HPLC was performed using two 510 pumps (Waters Corp.,
Milford, MA, U.S.A.), a U6K injector (Waters), an S310 model II UV detec-
tor (Soma Optics Ltd., Tokyo, Japan), a 680 Automated Gradient Controller
(Waters), and a chromatocorder 21 (System Instruments Co., Ltd., Tokyo,
2) Preparation of Linear Partially Protected [Ala3]-Polymyxin B3
(3L) and Various Linear Partially Protected Polymyxin B3 Analogs (1L,
2L and 4L—16L) Protected peptide-resin (3R, 0.2 mmol eq) was treated
∗ To whom correspondence should be addressed. e-mail: n-sakura@hokuriku-u.ac.jp
© 2009 Pharmaceutical Society of Japan

March 2009
241
Fig. 1. Synthetic Polymyxin B3 Analogs Used in This Study
HW-40-S column (1.6ꢀ95 cm) as described above for 3L. Fractions contain-
ing the main product were combined, evaporated, and the partially purified
product 3C (the purity: 86% by HPLC analysis measured at 210 nm) was
lyophilized from dioxane. Yield, 133 mg (94%). FAB-MS; Found (for the
most abundant isotopic variant of 3C; formula C100H125N15O21Cl4); 2014.8
[MꢁH]ꢁ, 2036.8 [MꢁNa]ꢁ. The other cyclic protected peptides (1C, 2C,
4C—16C) were obtained almost quantitatively in the same manner as de-
scribed for 3C and their FAB-MS data (not shown) confirmed the structure of
the main products. The partially purified peptides (1C, 2C, 4C—16C) were
used without further purification for the deprotection reaction with HF.
4) Preparation of [Ala3]-Polymyxin B3 (3) and Various Polymyxin B3
Analogs (1, 2, 4—16) Cyclic protected [Ala3]-polymyxin B3 (3C, 101 mg,
0.05 mmol) was treated with ice-cold anhydrous HF (2 ml)–anisole (0.2 ml)
for 1 h, then excess HF was removed in vacuo. The residue was dissolved in
H2O (15 ml), washed with three portions of ether and lyophilized. The crude
[Ala3]-polymyxin B3 (3) was purified by HPLC on a CAPCELL PAK C18
UG-80 5 m (2ꢀ15 cm) column using acetonitrile–0.1% TFA as eluent. The
main peak was collected and the combined eluents were evaporated and
lyophilized. The product was chromatographed on a Toyopearl HW-40-S
column (1.5ꢀ57 cm) using 25% CH3CN in 5 mmol/l HCl as eluent. Purified
product 3 was obtained by lyophilization as a hydrochloride salt; yield 29 mg
(44.4%). The other polymyxin B3 analogs, 1, 2 and 4—16, were prepared in
the same manner except for the L-Trp or D-Trp-containing peptides (14, 15),
which were obtained by treating 14C and 15C with HF (2 ml) containing
anisole (0.2 ml) and ethanedithiol (0.2 ml). For amino acid analysis of syn-
thetic peptides, acid hydrolysis was carried out with vapor of 6 mol/l HCl
containing 3% phenol at 130 °C for 3 h. The peak of Dab on the analysis was
determined as Lys. Data of amino acid analysis are shown in Table 1. Char-
acterization and analytical data of synthetic peptides 1—16 are shown in
Table 2.
Bacteria and Susceptibility Test Escherichia coli (E. coli) IFO 12734,
Salmonella Typhimurium (S. Typhimurium) IFO 12529 and Pseudomonas
aeruginosa (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).
Antibacterial activity of the synthetic peptides was evaluated by compari-
son with commercially available polymyxin B (Sigma Chemical Co., St.
Louis, MO, U.S.A.). Minimum inhibitory concentrations (MIC) of the syn-
thetic peptides against the bacterial strains were determined using a standard
microplate dilution method (nꢂ6—8). One hundred microliters of each seri-
ally diluted peptide in distilled water (0.25—512 nmol/ml) was added to a
mixture of 10 ml bacterial suspension (approximately 106 CFU/ml) and 90 ml
Mueller-Hinton broth (Becton Dickinson and Company Sparks, Cock-
eysville, MD, U.S.A.) in each well of a flat- bottomed microplate (Corning
Fig. 2. Synthetic Route of [Ala3]-Polymyxin B3 (3)
with trifluoroacetic acid (TFA)–H2O (95 : 5, 5 ml) for 1.5 h at room tempera-
ture, cleaving the peptide from the HMP-resin and the Boc protecting group
from Dab4 and yielding octanoyl-Dab(2-ClZ)-Thr(Bzl)-Ala3-Dab4-Dab(2-
ClZ)-D-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)-OH (3L). The excess TFA
was evaporated in vacuo and the residue was lyophilized from dioxane. The
products were dissolved in dimethylsulfoxide (DMSO) (2 ml) and fraction-
ated on a column (1.6ꢀ95 cm) of Toyopearl HW-40-S using DMF : H2O
(9 : 1) as eluent. Fractions containing the main product (3L) were combined,
evaporated, and the partially purified product (the purity: 82% by HPLC
analysis measured at 210 nm) was lyophilized from dioxane. Yield, 170 mg
(41.8% calculated from Fmoc-Thr(Bzl)-O-HMP-resin). FAB-MS; Found
(for the most abundant isotopic variant of 3L; formula C100H127N15O22Cl4);
2032.8 [MꢁH]ꢁ, 2054.7 [MꢁNa]ꢁ.
The other linear partially protected peptides (1L, 2L and 4L—16L) were
obtained in the same manner as described for 3L and their FAB-MS data (not
shown) confirmed the structure of the main products. The partially purified
peptides (1L, 2L and 4L—16L) were used without further purification for the
cyclization reaction.
3) Preparation of Cyclic Protected [Ala3]-Polymyxin B3 (3C) and
Various Cyclic Protected Polymyxin B3 Analogs (1C, 2C and 4C—16C)
Linear partially protected [Ala3]-polymyxin B3 (3L, 142 mg, 0.07 mmol) was
dissolved in ice-cold DMSO (1 ml)-DMF (2 ml), then diphenyl phosphorazi-
date (DPPA)20) (66 ml, 0.35 mmol) and 4-methylmorpholine (72 ml, 0.70
mmol) were added. The mixture was allowed to react overnight at 4 °C to
form the lactam ring. The reaction mixture containing the cyclization prod-
uct Na-octanoyl-Dab(2-ClZ)-Thr(Bzl)-Ala3-cyclic[Dab4*-Dab(2-ClZ)-
D-Phe-
Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)10*] (*–*: amide bond between a-
COOH of Thr10 and g-NH2 of Dab4) (3C) was fractionated on a Toyopearl

242
Vol. 57, No. 3
Inc., Corning, NY, U.S.A.). The plates were then incubated overnight at
37 °C for MIC evaluation. The MIC value was expressed as the lowest final
concentration (nmol/ml) at which no growth was observed (Table 3).
LPS Binding Assay of Synthetic Peptides to LPS Following a method
described previously,16) a solution of [Dab(Ng-dansyl-Gly)1]-polymyxin B3
Results and Discussion
Synthesis The synthesis of polymyxin B3 analogs (1—
16) shown in Fig. 1 was carried out as reported previously.16)
The peptide chain was constructed by solid-phase synthesis
in H2O (1 mmol/ml) (4 ml, 4 nmol) was added to a quartz cuvette containing using Fmoc-amino acids bearing benzyl-type protecting
N-(2-hydroxyethyl)piperazine-Nꢃ-(2-ethanesulfonic acid) buffer (HEPES; 5
mmol/l, pH 7.2) (1 ml), followed by a solution of E. coli (serotype 055:B5)
lipopolysaccharide (LPS, Sigma Chemical Co., St. Louis, MO, U.S.A.) in
H2O (3 mg/ml) (10 ml, 30 mg). The solution was kept at 30 °C for 60 min
groups on the side chain functional groups, except for the
Dab at position 4, whose g-amino function was protected by
Boc. After the cleavage of the solid support and the Boc pro-
tecting group on Dab4 with TFA, the linear partially pro-
tected peptides (1L—16L) were cyclized between the a-
COOH of Thr(Bzl)10 and the g-NH2 of Dab4 using DPPA20)
at 4 °C in minimal amounts of DMSO–DMF (1 : 1). The cy-
clization reaction proceeded quantitatively in a high concen-
tration (ca. 0.01 mmol/ml) solution of 1L—16L to yield 1C—
16C, which were treated with HF to yield polymyxin B
until the fluorescence intensity of [Dab(Ng-dansyl-Gly)1]-polymyxin B3
plateaued. A solution of each polymyxin B3 analog (1 mmol/ml) (4 ml each)
was added cumulatively to the quart cuvette at 5-min intervals to obtain
eight data points (4—32 nmol). The change in fluorescence intensity was
measured after each addition with a fluorescence spectrophotometer (F-850,
Hitachi Instrument Co., Tokyo, Japan) using an excitation wavelength of
330 nm and an emission wavelength of 490 nm. The initial intensity of fluo-
rescence was taken as 100%. The percent fluorescence intensity was plotted
as a function of peptide concentration. The binding experiments were re-
peated at least three times for each peptide to test the reproducibility of the
results. The concentrations required for 50% quenching of [Dab(Ng-dansyl-
Gly)1]-polymyxin B3 bound to LPS (IC50) were derived from the quenching
curves of the synthetic peptides (Figs. 4, 5, Table 3).
Table 3. Antimicrobial Activity and LPS-Binding Activity of Synthetic
Peptides 1—16
MIC (nmol/ml)
IC50
(nmol/ml)
Table 1. Amino Acid Analysis of Synthetic Peptides
Escherichia
coli
Salmonella
Typhimurium
Pseudomonas
aeruginosa
Peptide
Thr
Dab
Leu
Phe
Ala
Gly
Polymyxin B
(Sigma)
1
0.5
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1.99 (2) 4.83 (5) 1.00 (1) 0.99 (1) 1.19 (1)
0.99 (1) 5.73 (6) 1.06 (1) 1.04 (1) 1.19 (1)
1.98 (2) 4.65 (5) 1.05 (1) 1.05 (1) 1.28 (1)
2.04 (2) 4.77 (5) 1.01 (1) 1.01 (1) 1.18 (1)
—
—
—
—
—
—
—
—
—
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2
8
1
16
1
1
2
4
2
2
2
2
1
2
256
2
2
4
1
16
1
1
4
4
1
2
2
16
2
2
2
2
4
1
1
2
2
1
2
2
1
2
4
20
12
19
—
6
2.04 (2) 5.75 (6) 1.01 (1)
2.02 (2) 5.73 (6)
1.98 (2) 5.93 (6) 1.03 (1) 1.05 (1)
—
1.20 (1)
—
1.04 (1) 1.21 (1)
—
5
2.00 (2) 4.68 (5) 0.99 (1) 1.13 (1) 1.20 (1)
1.01 (1) 5.77 (6) 0.99 (1) 0.98 (1) 1.23 (1)
26
19
5
6
5
11
11
8
20
10
2.05 (2) 6.02 (6) 0.94 (1)
1.71 (2) 6.20 (6) 1.08 (1)
1.77 (2) 6.03 (6) 1.28 (1)
—
—
—
—
—
—
—
—
0.91 (1)
1.01 (1)
—
1.97 (2)
—
1.69 (2) 6.18 (6)
1.77 (2) 6.27 (6)
1.96 (2) 6.07 (6)
—
—
—
1.11 (1)
—
—
—
1.96 (2)
—
32
1.93 (2) 5.88 (6) 1.08 (1) 1.08 (1)
—
ꢅ512
2
ꢅ512
2
Hydrolysis was carried out with vapor of 6 mol/l HCl containing 3% phenol at
130 °C for 3 h. Number in parentheses are theoretical values.
Table 2. Characterization of Synthetic Peptides 1—16
FAB-MS
HPLCa)
HP-TLCb)
[a]D20
(cꢂ0.5, 12% AcOH)
Formula
[MꢁH]ꢁ
[MꢁNa]ꢁ
tR/min
Rf 1
Rf 2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
ꢄ75.2°
ꢄ64.4°
ꢄ78.8°
ꢄ65.2°
ꢄ65.6°
ꢄ62.0°
ꢄ63.4°
ꢄ66.0°
ꢄ70.0°
ꢄ57.7°
ꢄ50.7°
ꢄ45.1°
ꢄ43.6°
ꢄ52.0°
ꢄ40.7°
ꢄ54.0°
C54H93N15O13
C54H94N16O12
C54H93N15O13
C54H93N15O13
C49H92N16O13
C52H90N16O13
C57H98N16O14
C54H93N15O13
C54H94N16O12
C57H97N17O13
C60H95N17O13
C48H90N16O13
C51H88N16O13
C46H86N16O13
C44H82N16O13
C55H96N16O13
1160
1159
1160
1160
1113
1147
1231
1160
1159
1228
1262
1099
1133
1071
1043
1189
1182
1181
1182
1182
1135
1169
1253
1182
1181
1250
1284
1121
1155
1093
1065
1211
23.1
19.1
21.7
22.4
14.9
16.0
21.8
23.5
20.8
20.6
21.7
15.2
15.1
12.5
11.5
17.2
0.47
0.45
0.49
0.49
0.21
0.21
0.51
0.49
0.42
0.47
0.47
0.16
0.17
0.07
0.07
0.06
0.45
0.38
0.47
0.47
0.30
0.30
0.46
0.47
0.43
0.43
0.43
0.32
0.37
0.24
0.19
0.44
a) HPLC conditions: column; YMC-Pack pro C18 (4.6ꢀ250 mm), elution; linear gradient elution from 16 to 32% MeCN in 0.1% TFA (20 min), flow rate; 1 ml/min, detec-
tion; 210 nm. b) TLC solvent systems: Rf 1; BuOH : Pyridine : AcOH : H2O (30 : 20 : 6 : 24). Rf 2; BuOH : AcOH : AcOEt : H2O (1 : 1 : 1 : 1).

March 2009
243
23-member lactam ring. On the other hand, three Dab
residues in the lactam ring were found to play an important
bactericidal role since the antimicrobial activities of [Ala5]-,
[Dab(Ac)8]-, and [Ala9]-polymyxin B3 (4, 7, 8) were less than
those of 1 and 3. Substitution of Ala for Dab at position 5
markedly reduced the activity, to 1/16 (E. coli), 1/32 (S. Ty-
phimurium) and 1/4 (P. aeruginosa) that of native polymyxin
B. It should be emphasized that the side chain of Dab5 was
the most important of the five Dab amino functions in
polymyxin B3 for biological activity.
[D-Ala6]- and [Ala7]-polymyxin B3 (5, 6) retained the same
or half the antimicrobial activity of polymyxin B toward
three bacteria tested, showing that the reduction in hydropho-
bicity by replacing either D-Phe6 or Leu7 with D-Ala or Ala
had little effect on reducing antimicrobial activity. Further
studies revealed that even [Gly6]- and [Gly7]-polymyxin B3
(12, 13) retained the same or half the activity of polymyxin B
against E. coli and P. aeruginosa. Thus the hydrophobic side
chains in the lactam ring (D-Phe6-Leu7) do not appear to be
essential for the antimicrobial activity of polymyxin B3
against these two bacteria. However, S. Typhimurium showed
low susceptibility to the Gly6-analog (12), which had only
Fig. 3. Analytical HPLC of [Ala3]-Polymyxin B3 (3)
analogs (1—16). The purity of the synthetic peptides was 1/32 the antimicrobial activity of polymyxin B. [D-Ala6,
ꢅ97% as shown representatively in Fig. 3 for analog 3. The Ala7]-polymyxin B3 (14) showed greatly reduced activity, es-
results demonstrated the usefulness of the synthetic protocols pecially toward S. Typhimurium (1/64). The low susceptibil-
described herein. However, during the present Ala-scanning ity of S. Typhimurium to polymyxin B3 analogs with reduced
study, it was noted that no cyclization product was obtained hydrophobicity at positions 6 and 7 suggests that the cell
from octanoyl-Dab(2-ClZ)1-Thr(Bzl)-Dab(2-ClZ)-Dab4-Dab(2- membrane of this bacterium may have different characteris-
ClZ)-D-Phe-Leu-Ala8-Dab(2-ClZ)-Thr(Bzl)10-OH (7ꢃL), which tics from that of E. coli and P. aeruginosa. [Gly6, Gly7]-
was prepared as an intermediate for the synthesis of [Ala8]- polymyxin B3 (15) was inactive against three bacterial
polymyxin B3 (7ꢃ). No considerable decease of the starting species tested. The complete loss of antimicrobial activity of
material (7Lꢃ) was proved by HPLC analysis of the reaction 15 might be explained by the complete loss of hydrophobic-
mixture in the presence of a large excess of DPPA for 48 h. ity simultaneously at positions 6 and 7, or by the drastic
When HATU was used as a cyclization reagent, the decrease change of the whole peptide conformation, caused possibly
of 7ꢃ peak was observed without the appearance of desired by substitution of Gly-Gly for D-Phe-Leu. Conversely, an in-
L
product peak (data not shown), suggesting that polymer for- crease in hydrophobicity in terms of the Nozaki–Tanford hy-
mation occurred under the reaction condition. Then an Ng- drophobicity scale21) at position 6 or 7 did not increase bacte-
acetylated peptide at Dab8 residue (7) was designed to evalu- ricidal activity, since [D-Trp6]-polymyxin B3 (10) and [Trp7]-
ate the biological role of the side chain. When Dab(Ac) was polymyxin B3 (11) showed moderately reduced (1/2—1/4)
introduced at position 8, the intra-molecular reaction of 7L activity, with MIC values of 2 nmol/ml, against the three
proceeded smoothly to give a lactam 7C. It can therefore be bacterial species tested. A similar result was reported regard-
assumed that Ala8-analog (7Lꢃ) has an exceptionally special ing polymyxin B nonapeptide analogs: [D-Trp6]-polymyxin
conformation under the reaction condition, by which the lac- B (2—10) had moderately reduced activity and [Phe7]-
tam formation is completely disturbed.
polymyxin B (2—10) had the same activity in an outer mem-
Antimicrobial Activity The antimicrobial activity of the brane permeabilizing assay.14)
synthetic peptides (1—16) was estimated by the standard
It is known that in the polymyxin B family of peptides,
micro plate dilution method. The minimum inhibitory con- which includes colistin, D-amino acids are commonly located
centration (MIC) values of peptides toward three Gram-nega- at position 6. However, both [L-Phe6]-polymyxin B3 (16) and
tive bacteria, E. coli, S. Typhimurium, and P. aeruginosa, are [Gly6]-polymyxin B3 (12) showed considerable potency against
summarized in Table 3. Polymyxin B (Sigma Co.) was used E. coli and P. aeruginosa with MIC values of 1—2 nmol/ml.
as a standard for antimicrobial potency. [Ala1]-polymyxin Therefore, the D-configuration at position 6 is not a prerequi-
B3 (1) had slightly reduced antimicrobial potency against site for the antimicrobial activity of polymyxin B3.
the three Gram-negative bacteria, with a MIC value of
LPS Binding Activity The LPS binding activity of syn-
2 nmol/ml, which corresponds to 1/2—1/4 the activity of na- thetic polymyxin B3 analogs was evaluated as reported previ-
tive polymyxin B. [Ala3]-polymyxin B3 (3) had the same or ously.16) All Ala-substituted analogs for Dab residues at posi-
half the antimicrobial activity as polymyxin B, suggesting tions 1, 3, 5, 8 and 9 showed greatly reduced binding activity
that the side chain amino function of Dab3 is not indispensa- to LPS, indicating the important contribution of the Dab
ble for activity. The antimicrobial activity of [Ala2]-polymyxin basic side chain to binding. On the contrary, Ala-substituted
B3 (2) was 1/2—1/8 that of polymyxin B, demonstrating that analogs for the residues at positions 2, 6, 7 and 10 had the
the Thr2 residue is more important than the basic amino acids same binding activity as polymyxin B3, showing a less strin-
(Dab1 and Dab3) in the N-terminal region located outside the gent requirement for neutral side chains at these positions

244
Vol. 57, No. 3
the substantial loss in hydrophobicity at positions 6 and 7
resulted in a considerable decrease in binding, as seen in the
displacement curve of [Gly6,Gly7]-polymyxin B3 (15) (Fig.
5). Inversely, the increase in hydrophobicity by substitution
of D-Trp or Trp for D-Phe6 or Leu7 resulted in the same LPS-
binding activity as polymyxin B. Replacement of D-Phe6 with
the L-form lowered binding. This result is in contrast to the
high antimicrobial activity of [L-Phe6]-polymyxin B3 (16) to-
wards microorganisms.
In conclusion it was found that the Dab residues located in
the lactam ring of polymyxin B3 are indispensable for antimi-
crobial activity, with Dab5 being the most important residue.
The Dab residues at positions 1 and 3 are not essential for
bactericidal activity. Neither the hydrophobic character of the
D-Phe6-Leu7 region nor the D-configuration at position 6 is
indispensable for antimicrobial activity. For LPS binding ac-
tivity, all the Dab residues, especially Dab5, play an impor-
tant role.
Fig. 4. Displacement of [Dab(dansyl-Gly)1]-Polymyxin B3 Bound to LPS
by Polymyxin B3 Analogs 1—9
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) Srimal S., Suloria N., Balasubramanian S., Surolia A., Biochem. J.,
315, 679—686 (1996).
6) Pristovsˇek P., Kidricˇ J., J. Med. Chem., 42, 4604—4613 (1999).
7) Hermsen E. D., Sullivan. C. J., Rotschafer J. C., Infect. Dis. Clin.
North. Am., 17, 545—562 (2003).
8) Sharma S. K., Wu A. D., Chandramouli N., Fotsch C., Kardash G.,
Bair K. W., J. Peptide Res., 53, 501—506 (1999).
9) Sakura N., Itho T., Ohki K., Nagai S., Uchida Y., “Peptide Science
1999,” ed. by Fujii N., The Japanese Peptide Society, 2000, pp. 413—
416.
10) Tsubery H., Ofek I., Cohen S., Fridkin M., J. Med. Chem., 43, 3085—
3092 (2000).
11) Tsubery H., Ofek I., Cohen S., Fridkin M., Biochemistry, 39, 11837—
11844 (2000).
12) Kline T., Holub D., Therrien J., Leung T., Ryckman D., J. Peptide Res.,
57, 175—187 (2001).
13) Tsubery H., Ofek I., Cohen S., Fridkin M., Peptides, 22, 1675—1681
(2001).
14) Tsubery H., Ofek I., Cohen S., Eisenstein M., Fridkin M., Mol. Phar-
macol., 62, 1036—1042 (2002).
15) de Visser P. C., Kriek N. M. A. J., van Hooft P. A. V., Van Schepdael
A., Filippov D. V., van der Marel G. A., Overkleeft H. S., van Boom J.
H., Noort D., J. Peptide Res., 61, 298—306 (2003).
16) Sakura N., Itoh T., Uchida Y., Ohki K., Okimura K., Chiba K., Sato Y.,
Sawanishi H., Bull. Chem. Soc. Jpn., 77, 1915—1924 (2004).
17) Tsubery H., Yaakov H., Cohen S., Giterman T., Matityahou A., Fridkin
M., Ofek I., Antimicrob. Agents Chemother., 49, 3122—3128 (2005).
18) Okimura K., Ohki K., Sato Y., Ohnishi K., Uchida Y., Sakura N., Bull.
Chem. Soc. Jpn., 80, 543—552 (2007).
19) Okimura K., Ohki K., Sato Y., Ohnishi K., Sakura N., Chem. Pharm.
Bull., 55, 1724—1730 (2007).
20) Ozawa K., Shioiri T., Yamada S., Chem. Pharm. Bull., 25, 122—125
(1977).
Fig. 5. Displacement of [Dab(dansyl-Gly)1]-Polymyxin B3 Bound to LPS
by Polymyxin B3 Analogs 10—16
(Figs. 4, 5). The positive charges of the five Dab side chains
in this peptide antibiotic are believed to interact with the neg-
ative charges of the phosphate groups of LPS.5,6) The present
results demonstrate that the g-amino function of Dab5 makes
the greatest contribution to binding to LPS by electrostatic
interactions. The primary role of the amino function at posi-
tion 5 was so great that the IC50 value of [Ala5]-polymyxin
B3 (4) could not be assessed under the experimental condi-
tions of this study. The low LPS-binding activity of 4 re-
sulted in very low antimicrobial activity, as described above.
The contribution of the g-amino function of Dab8 was the
second most important contribution to binding, with an IC50
value of 26 nmol/ml. The IC50 values of the other Ala-substi-
tution analogs for Dab residues at position 1, 3 and 9 were
19—20 nmol/ml.
The role of the Thr residue at position 2 appears to be
more important than the Thr at position 10 for LPS binding,
since the displacement curve of [Ala2]-polymyxin B3 (2)
shifted to the right of that of [Ala10]-polymyxin B3 (9). Sub-
stitution of Ala either for Phe or Leu at positions 6 and 7 did
not affect binding to LPS, indicating that the slight decrease
in hydrophobicity in the lactam ring was tolerated. However,
21) Nozaki Y., Tanford C., J. Biol. Chem., 246, 2211—2217 (1971).