Development of des-fatty acyl-polymyxin B decapeptide analogs with pseudomonas aeruginosa-specific antimicrobial activity

Article abstract of DOI:10.1248/cpb.57.332

Twelve N-terminal analogs of des-fatty acyl-polymyxin B (Des-FA-[X ¹]-PMB, X=various amino acids or peptides) were synthesized and examined for their antimicrobial activity against Escherichia coli (E. coli), Salmonella Typhimurium (S. Typhimur

Full text of DOI:10.1248/cpb.57.332

332
Chem. Pharm. Bull. 57(4) 332—336 (2009)
Vol. 57, No. 4
Development of Des-Fatty Acyl-Polymyxin B Decapeptide Analogs with
Pseudomonas aeruginosa-Specific Antimicrobial Activity
Naoko KATSUMA, Yuki SATO, Kazuhiro OHKI, Keiko OKIMURA, Kuniharu OHNISHI, and Naoki SAKURA*
Faculty of Pharmaceutical Sciences, Hokuriku University; Kanazawa 920–1181, Japan.
Received September 30, 2008; accepted January 23, 2009; published online January 26, 2009
Twelve N-terminal analogs of des-fatty acyl-polymyxin B (Des-FA-[X1]-PMB, Xꢀvarious amino acids or
peptides) were synthesized and examined for their antimicrobial activity against Escherichia coli (E. coli), Salmo-
nella Typhimurium (S. Typhimurium) and Pseudomonas aeruginosa (P. aeruginosa). It was found that Des-FA-
[Dap1]-, Des-FA-[Ser1]-, Des-FA-[Dab-Dab-Dab1]- and Des-FA-[Arg-Arg-Arg1]-PMB had potent activity only
against P. aeruginosa, with MIC values of 0.5—1 nmol/ml. Analogs in which X was Lys, Arg, Leu or Ala did not
have increased antimicrobial activity against the three bacterial species tested compared with the lead com-
pounds Des-FA-[Dab1]-PMB and polymyxin B (PMB). Des-FA-[Trp1]-PMB and Des-FA-[Phe1]-PMB had re-
duced activity against P. aeruginosa. The results indicate that compact hydrophilic amino acids (C3) or basic
tripeptides at the N-terminal provide specificity for bactericidal activity towards P. aeruginosa. For LPS-binding
activity, Des-FA-[Dab-Dab-Dab1]-PMB and Des-FA-[Arg-Arg-Arg1]-PMB showed activity comparable to PMB,
while Des-FA-[Ala-Ala-Ala1]-PMB showed very low activity. Reduced acute toxicity of Des-FA-[Dap1]-PMB and
Des-FA-[Trp1]-PMB was demonstrated by a mouse tail intravenous administration test, with LD50 values of 23.5
and 19.0 mmol/kg, respectively, in contrast to 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
Co., Ltd., Tokyo, Japan). Gel column chromatography was carried out using
Toyopearl HW-40-S (Tosoh Corporation, Tokyo, Japan). Fast-atom bom-
bardment mass spectra (FAB-MS) were obtained on a JMS-DX300 mass
spectrometer (JEOL Ltd., Tokyo, Japan). The optical rotations of the pep-
tides were measured with a DIP-370 digital polarimeter (Nippon Bunko Co.,
Ltd., Tokyo, Japan). Peptide deprotection 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, Darm-
stadt, Germany). All reagents, peptide synthesis solvents and 9-fluorenyl-
methoxycarbonyl (Fmoc) amino acids were purchased from Watanabe
Chem. Ind. Ltd. (Hiroshima, Japan).
Polymyxin B (PMB) is a cyclic decapeptide antibiotic, and
the N-terminal a,g-diaminobutyric acid residue (Dab1) is
Na-acylated by fatty acids such as 6-methyloctanoic acid
(PMB1), 6-methylheptanoic acid (PMB2) and octanoic acid
(PMB3).1,2) PMB has potent antimicrobial activity against
Gram-negative bacteria, but the toxicity of PMB limits its
therapeutic application.1) It is well-known that the enzymatic
removal of the fatty acyl-Dab1 of PMB with ficin yields PMB
nonapeptide (PMBN),3,4) which retains considerable lipo-
polysaccharide (LPS)-binding activity5) and has reduced toxi-
city.6) However, PMBN has low bactericidal activity, with a
minimum inhibitory concentration (MIC) value of ꢀ256
nmol/ml towards Escherichia coli (E. coli).4) Thus, it appears
that the fatty acyl-Dab1 moiety of PMB greatly affects an-
Synthesis of Peptides Des-FA-[X1]-PMB analogs (1—12, Fig. 1)
were synthesized as reported previously.7) In brief, the protected peptide
was constructed on 4-hydroxylmethylphenoxymethyl-resin10) (HMP-resin,
0.74 mmol/g, Novabiochem, Läufelfingen, Switzerland) by a solid phase
methodology11) using an ABI 433A peptide synthesizer (Applied Biosys-
tems, Foster City, CA, U.S.A.). Protected amino acids used were Fmoc-
timicrobial activity and toxicity, and has some effect on LPS Thr(Bzl)-OH, Fmoc-Dab(2-ClZ)-OH, Fmoc-Dab(Boc)-OH, Fmoc-D-Phe-
binding activity.7) Our previous studies on the structure–
OH, Fmoc-Leu-OH, Fmoc-Lys(2-ClZ)-OH Fmoc-Arg(Tos)-OH, Fmoc-
Dap(2-ClZ)-OH (Dap; L-a,b-diaminopropionic acid), Fmoc-Ser(Bzl)-OH,
Fmoc-Phe-OH, Fmoc-Trp(Boc)-OH, Fmoc-Ala-OH and Fmoc-Glu(OcHex)-
antimicrobial activity relationship of PMB showed that des-
fatty acyl-PMB (Des-FA-[Dab1]-PMB) has considerable an-
OH. The Fmoc groups were removed with 25% piperidine in N-methylpy-
timicrobial activity, with MIC values of 8, 16 and 4 nmol/ml
against E. coli, Salmonella Typhimurium (S. Typhimurium)
and Pseudomonas aeruginosa (P. aeruginosa).7—9) Appar-
ently, the introduction of Dab at the N-terminal of PMBN
greatly increases its activity towards E. coli, from an MIC
value of 256 to 8 nmol/ml. We therefore selected the cyclic
decapeptide Des-FA-[Dab1]-PMB as a lead compound for
developing a peptide antibiotic with increased antimicrobial
activity and decreased toxicity. In the present study, synthetic
peptides with bactericidal activity specific towards P. aerugi-
nosa were obtained by introducing compact amino acids with
two or more basic or hydrophilic functions into the X posi-
tion of des-FA-[X1]-PMB (Fig. 1).
rolidone (NMP).
1) Preparation of Fmoc-[X1]-Thr(Bzl)-Dab(2-ClZ)-Dab-Dab(2-ClZ)-
D-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)-OH (1L—12L) Starting
from Fmoc-Thr(Bzl)-O-HMP-resin (1.132 g, 0.6 mmol eq), Fmoc-amino
acids (1.5 mmol) and O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluro-
Experimental
General HPLC was performed on an apparatus equipped with two 510
pumps (Waters Corp., Milford, MA, U.S.A.), a U6K injector (Waters), a
Lambda-Max Model 481 LC Spectrophotometer (Waters), a 680 Automated
Gradient Controller (Waters), and a chromatocorder 21 (System Instruments
Fig. 1. Structures of the Synthetic Peptides, Des-FA-[X1]-PMB
∗ To whom correspondence should be addressed. e-mail: n-sakura@hokuriku-u.ac.jp
© 2009 Pharmaceutical Society of Japan

April 2009
333
nium hexafluorophosphate (HATU) (1.5 mmol) were dissolved in NMP uct peak was collected, combined, evaporated and lyophilized. The product
(10 ml), and N,N-diisopropylethylamine (3.0 mmol, 523 ml) was added. The was chromatographed on a Toyopearl HW-40-S column (1.5ꢁ57 cm) using
mixture was transferred to a reaction vessel containing the resin and allowed 25% CH3CN in 5 mmol/l HCl. Purified product was obtained as the hy-
to react for 2—5 h prior to acetylation with acetic anhydride (3 ml)–pyridine drochloride salt by lyophilization. Yield of 1—12 was 30—45% (calculated
(3 ml) in NMP (10 ml). Coupling of Fmoc-Thr(Bzl)-OH at position 2 fol-
lowed by deprotection with piperidine yielded H-Thr(Bzl)-Dab(2-ClZ)- synthetic peptides are shown in Table 2. The purity of the products was
Dab(Boc)-Dab(2-ClZ)-D-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)-O- ꢀ98% as determined by HPLC analysis, monitoring at 210 nm. A represen-
HMP-resin, which was divided into four portions and stored in NMP at 4 °C tative analytical HPLC chromatogram of Des-FA-[Dap1]-PMB (3) is shown
from partially purified 1C—12C). FAB-MS data and characteristics of the
until coupling with various protected N-terminal amino acids (Fmoc-X).
After introduction of Fmoc-X to this protected nonapeptide resin
(0.15 mmol eq), the resin was treated with trifluoroacetic acid (TFA)–H2O
in Fig. 2.
Bacteria, and Bactericidal Test E. coli IFO 12734, S. Typhimurium
IFO 12529 and P. aeruginosa IFO 3080 were purchased from the Institute
(95 : 5, 3 ml) for 1.5 h at room temperature to simultaneously cleave the pep- for Fermentation, Osaka (IFO), Japan. These bacterial strains were grown
tide from the HMP-resin support and the Boc-protecting group(s) from Dab4
or Trp1. TFA was removed in vacuo and the residue was lyophilized from
dioxane. The crude product was partially purified by column chromatogra-
phy on a Toyopearl HW-40-S column (1.6ꢁ95 cm, Tosoh Co.) using di-
methylformamide (DMF) : H2O (9 : 1). Fractions containing the main prod-
uct were combined, evaporated and lyophilized from dioxane to give Fmoc-
[X1]-Thr(Bzl)-Dab(2-ClZ)-Dab-Dab(2-Cl)-D-Phe-Leu-Dab(2-ClZ)-Dab(2-
ClZ)-Thr(Bzl)-OH [XꢂLys(2-ClZ) (1L), Arg(Tos) (2L), Dap(2-ClZ) (3L),
Ser(Bzl) (4L), Phe (5L), Trp (6L), Leu (7L), Ala (8L), Glu(OcHex) (9L),
Dab(2-ClZ)-Dab(2-ClZ)-Dab(2-ClZ) (10L), Arg(Tos)-Arg(Tos)-Arg(Tos)
(11L) and Ala-Ala-Ala (12L)]. The structures of 1L—12L were confirmed by
FAB-MS (Table 1). The purity of the products 1L—12L was 80—90% as de-
termined by HPLC analysis, monitoring at 210 nm. The partially purified
peptides (1L—12L) were used without further purification for the cyclization
reaction.
overnight at 37 °C on nutrient agar medium and harvested in sterile saline.
Table 1. FAB-MS Analysis of Linear Partially Protected Peptides (1L—
12L) and Cyclic Protected Peptides (1C—12C)
Linear protected [X1]-peptide
Found
X
Formula
[MꢃH]ꢃ
[MꢃNa]ꢃ
1L
2L
3L
4L
5L
6L
7L
8L
9L
C118H135N16O25Cl5
C117H136N18O25S1Cl4
C115H129N16O25Cl5
C114H129N15O24Cl4
C113H127N15O23Cl4
C115H128N16O23Cl4
C110H129N15O23Cl4
C107H123N15O23Cl4
C115H135N15O25Cl4
C140H157N20O31Cl7
C143H172N26O31S3Cl4
C113H132N17O25Cl4
2356.8
2369.8
2314.8
2234.8
2214.8
2244.8
2171.8
2129.8
2269.9
2864.9
2991.1
2271.9
2378.8
2391.8
2336.8
2256.8
2227.8
2266.8
2193.8
2151.8
2291.9
2886.9
3013.1
2293.9
2) Preparation of Fmoc-[X1]-Thr(Bzl)-Dab(2-ClZ)-cyclic-[Dab*-
Dab(2-ClZ)-D-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)*-] (*–*;
Amide Bond between the g-NH2 of Dab4 and a-COOH of Thr10)] (1C—
12C) The linear partially protected peptide (1L—12L, 0.02—0.07 mmol)
was dissolved in ice-cold DMSO (1 ml)–DMF (2 ml), and diphenyl phos-
phorazidate12) (DPPA; 0.1—0.35 mmol, 5-fold eq) and 4-methylmorpholine
(0.2—0.70 mmol, 10-fold eq) were added. The mixture was allowed to react
overnight at 4 °C to form the lactam ring, then the cyclized products (1C—
12C) were purified on a Toyopearl HW-40-S column (1.6ꢁ95 cm) as de-
scribed above for 1L—12L. Fractions containing the main product were com-
bined, evaporated, and the product was lyophilized from dioxane. Yields of
all cyclized products (1C—12C) were almost quantitative (ꢀ90%). FAB-MS
data are shown in Table 1; the purity of the products was shown to be 85—
90% by HPLC analysis. The partially purified peptides (1C—12C) were used
without further purification for the deprotection reaction.
10L
11L
12L
Cyclic protected [X1]-peptide
Found
X
Formula
[MꢃH]ꢃ
[MꢃNa]ꢃ
1C
2C
3C
4C
5C
6C
7C
8C
9C
C118H133N16O24Cl5
C117H134N18O24SCl4
C115H127N16O24Cl5
C114H127N15O23Cl4
C113H125N15O22Cl4
C115H126N16O22Cl4
C110H127N15O22Cl4
C107H121N15O22Cl4
C115H133N15O24Cl4
C140H155N20O30Cl7
C143H170N26O30S3Cl4
C113H131N17O24Cl4
2338.8
2351.8
2296.8
2216.8
2187.8
2226.8
2153.8
2111.8
2251.9
2846.9
2972.1
2253.8
2360.8
2373.8
2318.8
2238.8
2209.8
2248.8
2175.8
2133.8
2273.9
2868.9
2995.1
2275.8
3) Preparation of [X1]-Polymyxin B (1—12) Each cyclic protected
[X1]-peptide (1C—12C, 40—70 mg) was dissolved in DMF (2 ml), and
piperidine (0.5 ml) was added. The mixture was stirred for 5 min at room
temperature and evaporated. The piperidine treatment was repeated three
times and the product was lyophilized from dioxane. Lyophilized product
was treated with anhydrous HF (2 ml)–anisole (0.2 ml) for 1 h on ice, then
excess HF was removed in vacuo. The residue was dissolved in H2O (15 ml),
washed with three portions of ether (15 ml) and lyophilized. The crude prod-
uct was purified by HPLC employing a CAPCELL PAK C18 UG-80 5 mm
(2ꢁ15 cm, Shiseido Co., Ltd.) using acetonitrile–0.1% TFA. The main prod-
10C
11C
12C
Table 2. Characterization of Synthetic Peptides 1—12
FAB-MS
RP-HPLCa)
HP-TLCc)
[a]D20
Peptide
(cꢂ0.5, 12% AcOH)
Formula
[MꢃH]ꢃ
[MꢃNa]ꢃ
tRb) (min)
Rf 1
Rf 2
1
2
3
4
5
6
7
8
9
ꢄ63.6°
ꢄ64.0°
ꢄ70.3°
ꢄ65.2°
ꢄ59.3°
ꢄ52.9°
ꢄ65.3°
ꢄ74.7°
ꢄ67.3°
ꢄ78.1°
ꢄ62.7°
ꢄ92.8°
C49H86N16O12
C49H86N18O12
C46H80N18O12
C49H79N15O12
C52H83N15O12
C54H84N16O12
C49H85N15O12
C49H79N15O12
C48H81N15O14
C55H98N20O14
C61H110N26O14
C52H89N17O14
1092.7
1120.7
1050.6
1050
1111.6
1150.7
1077.6
1035.7
1093.6
1264.8
1432.9
1177.7
1114.7
1142.7
1072.6
1072
1133.6
1172.7
1099.6
1057.7
1115.6
1286.8
1454.9
1199.7
17.0
17.0
16.7
16.3
21.7
25.5
19.2
16.3
16.3
16.6
18.0
16.7
0.14
0.09
0.07
0.25
0.34
0.42
0.35
0.28
0.29
0.08
0.06
0.29
0.09
0.05
0.06
0.12
0.14
0.15
0.09
0.04
0.06
0.05
0.04
0.08
10
11
12
a) Conditions: column, CAPCELL PAK C18 UG120 (6.0ꢁ150 mm); elution, a linear gradient from 12 to 20% MeCN in 0.1% TFA over 30 min; flow rate, 1.5 ml/min; detec-
tion, 210 nm. b) Retention time. c) Rf 1; n-butanol : pyridine : AcOH : H2O (30 : 20 : 6 : 24). Rf 2; n-butanol : AcOH : AcOEt : H2O (1 : 1 : 1 : 1).

334
Vol. 57, No. 4
The densities of 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 in com-
peptide. The binding experiments were repeated at least three times for each
peptide to obtain reproducible results.
Acute Toxicity of Des-FA-[Dap1]-PMB (3) Des-FA-[Trp1]-PMB (6)
parison with that of commercially available PMB (Sigma Chemical Co., St The lethal dose (LD50) of synthetic peptides (PMB, PMBN, Des-FA-[Dap1]-
Louis, MO., U.S.A.). MIC of the synthetic peptides against the bacterial
strains were determined using a standard microplate dilution method
PMB (3) and Des-FA-[Trp1]-PMB (6)) were determined in male ddY mice
(4 weeks old; Japan SLC, Hamamatsu, Japan). The following peptide solu-
(nꢂ6—8). One hundred microliters of each peptide serially diluted with dis- tions in saline (mmol/ml) were prepared: PMB sulfate (Sigma); 0.25, 0.50,
tilled water to 0.25—256 nmol/ml was added to a mixture of 10 ml of bacter- 0.75, and 1.0. PMBN; 2.0, 2.5, 3.0, 3.5, and 4.0. Des-FA-[Dap1]-PMB (3);
ial suspension (approximately 106 CFU/ml) and 90 ml of Mueller–Hinton
1.0, 1.5, 2.0, 2.5, and 3.0. Des-FA-[Trp1]-PMB (6); 1.0, 1.5, 2.0, 2.5, and
broth (Becton Dickinson and Company Sparks, Cockeysville, MD, U.S.A.) 3.0. The rate and volume of injection through the lateral tail vein were
in each well of a flat-bottomed microplate (Corning Inc., Corning, NY, 0.1 ml/30 s and 0.1 ml/10 g body weight, respectively. Five to ten mice per
U.S.A.). The plates were incubated overnight at 37 °C for MIC evaluation. dose were used. The LD50 was determined by the Litchfield–Wilcoxon
The MIC value was expressed as the lowest final concentration (nmol/ml) at method13) from the number of mice dying following each dose.
which no growth was observed (Table 3).
All animal procedures were approved by the Animal Care and Use Com-
mittee of Hokuriku University.
LPS Binding Assay of Synthetic Peptides As described previously,7)
a
solution of [Dab(Ng-dansyl-Gly)1]-PMB3 in H2O (1 mmol/ml) (4 ml, 4 nmol)
was added to a quartz cuvette containing N-(2-hydroxyethyl)piperazine-Nꢅ-
(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.) 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 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 fluo-
rescence intensity was measured after each addition at an excitation wave-
length of 330 nm and at an emission wavelength of 490 nm using an F-850
fluorescence spectrophotometer (Hitachi Instrument Co., Tokyo, Japan). The
initial intensity of fluorescence was taken as 100%. The percent fluorescence
Results and Discussion
Synthesis of N-terminal fatty acyl-free and substitution
analogs of N-terminal Dab1 of PMB, Des-FA-[X1]-PMB, was
readily achieved using the synthetic route reported previ-
ously.7) Fmoc-[X1]-Thr(Bzl)-Dab(2-ClZ)-Dab(Boc)4-Dab(2-
ClZ)-D-Phe-Leu-Dab(2-ClZ)-Dab(2-ClZ)-Thr(Bzl)10-O-
HMP-resin was synthesized by solid phase techniques using
Fmoc-amino acids and HATU as the coupling reagent. Treat-
ment of the protected peptide resin with TFA simultaneously
intensity was plotted as a function of the concentration of peptide. The con- cleaved the peptide from the solid support and the Boc group
centration required for 50% quenching of [Dab(Ng-dansyl-Gly)1]-PMB3
bound to LPS (IC50) was derived from the quenching curve of each synthetic
from Dab4. Lactam formation on the resulting linear partially
protected peptide was achieved with DPPA reagent,12) yield-
ing cyclic protected Des-FA-[X1]-PMB quantitatively. Final
deprotection was performed by piperidine treatment, fol-
lowed by HF. 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 ꢀ98%, as shown
representatively by the HPLC chromatogram of Des-FA-
[Dap1]-PMB (3) (Fig. 2).
The antimicrobial activity of synthetic peptides against E.
coli, S. Typhimurium and P. aeruginosa is summarized in
Table 3. When the N-terminal Dab of the key compound,
Des-FA-[Dab1]-PMB, was replaced by Dap or Ser, the an-
timicrobial activity against P. aeruginaosa increased four
times. The activity of Des-FA-[Dap1]-PMB (3) and Des-FA-
[Ser1]-PMB (4) was as potent as that of PMB. However,
these analogs did not have increased activity against E. coli
and S. Typhimurium, i.e., 3 showed the same moderate bacte-
ricidal activity as Des-FA-[Dab1]-PMB, and 4 showed low
Fig. 2. Analytical HPLC of Des-FA-[Dap1]-Polymyxin B (3)
Table 3. Antimicrobial Activity and LPS-Binding Activity of Des-FA-[X1]-PMB (1—12)
MIC (nmol/ml)
LPS-binding
Peptide
PMB
X
IC50 (nmol/ml)
Escherichia coli
Salmonella Typhimurium
Pseudomonas aeruginosa
FA-Dab
Dab
1
8
0.5
16
1
4
3
19
1
2
3
Lys
Arg
Dap
16
8
8
32
16
16
32
16
1
22
15
14
4
5
6
Ser
Phe
Trp
64
8
8
64
8
8
1
16
8
ꢀ32
ꢀ32
20
7
8
9
10
11
12
Leu
Ala
Glu
16
32
256
8
16
64
32
64
ꢀ256
32
64
128
64
32
256
0.5
1
ꢀ32
ꢀ32
ꢀ32
3
3
ꢀ32
Dab-Dab-Dab
Arg-Arg-Arg
Ala-Ala-Ala
128

April 2009
335
Fig. 3. Displacement of [Dab(dansyl-Gly)1]-Polymyxin B3 Bound to LPS by Des-FA-[X1]-PMB (1—12)
Table 4. Acute Toxicity
Peptide
activity with the MIC value of 64 nmol/ml. Therefore, both
PMB analogs are P. aeruginaosa-specific antibiotics. On the
other hand, the antimicrobial activity of Des-FA-[Arg1]-PMB
(2) against E. coli and S. Typhimurium did not change, but its
activity against P. aeruginaosa decreased to 1/4 that of Des-
FA-[Dab1]-PMB. Des-FA-[Lys1]-PMB (1) showed poor activ-
ity against the three bacterial species tested. These results
suggest that small (C3) hydrophilic amino acids (Dap and
Ser) are required at the N-terminal for bactericidal specificity
to P. aeruginosa. The hydrophilic character of the side chain
LD50 value
(mmol/kg)
95% Confidence
interval (mmol/kg)
PMB (Sigma)
PMBN
4.8
31.5
23.5
19.0
4.3—5.3
25.5—39
16.7—33.1
17.4—20.7
3
6
is essential, since another C3 amino acid (Ala)-analog (8) bicity than those of the other two bacteria, rendering that the
showed greatly decreased activity. Substitution of basic cell membrane structure is not so tight that the hydrophobic
amino acids with bigger side chains [e.g., C5 (Arg) or C6 interaction of the side chain of D-Phe and Leu of PMB
(Lys)] for C4 (Dab1) resulted in poor specific activity towards analogs with the outer membrane causes the disorder of
P. aeruginaosa.
membrane to lead the cell death. The lack of N-terminal fatty
Among analogs with neutral and aromatic amino acids at acyl group in 10 and 11 seems not to be essential as bacteri-
the N-terminal, Des-FA-[Phe1]-PMB (5) and Des-FA-[Trp1]- cidal toward P. aeruginosa under the presence of Dab-Dab-
PMB (6) retained antimicrobial activity against E. coli and S. Dab or Arg-Arg-Arg, which may contribute to ionic binding
Typhimurium, with MIC values of 8 nmol/ml; however, both to the inner core of LPS. On the other hand, Des-FA-[Ala-
analogs had reduced activity against P. aeruginosa. The hy- Ala-Ala1]-PMB (12) had markedly reduced activity against
drophobic and aromatic functions of Phe and Trp appear to all three bacterial species tested. Although the N-terminal
take the place of the fatty acyl moiety of PMB and appar- tripeptide portions of 10—12 correspond to the fatty acyl-
ently contribute to the structural requirements for bacterici- Dab1 component in the PMB family in terms of size, the neu-
dal activity towards E. coli and S. Typhimurium. Des-FA- tral tripeptide portion of 12 cannot contribute to retaining the
[Glu1]-PMB (9) had no activity against the three bacterial bactericidal nature of the structure. Overall, the present re-
species tested, demonstrating that the introduction of an sults are consistent with the report by Tsubery et al.14) that
acidic group in the N-terminal completely abolishes the bac- Des-FA-[Ala-Ala-Ala1]-PMB is inactive at ꢀ500 mg/ml
tericidal function of the PMB molecule.
against E. coli and P. aeruginosa.
The LPS-binding activity of the synthetic peptides was
Des-FA-[Dab-Dab-Dab1]- and Des-FA-[Arg-Arg-Arg1]-
PMB (10, 11) were highly bactericidal only against P. aerugi- evaluated by measuring the displacement of [Dab(dansyl-
nosa, with MIC values of 0.5—1 nmol/ml. Substitution of Gly)1]-PMB3 bound to E. coli LPS, using the methodology
the four positive charges in H-Dab-Dab-Dab- or H-Arg-Arg- reported previously.7) The displacement curve of Des-FA-
Arg- for H-Dab- of Des-FA-[Dab1]-PMB or H-Arg- of Des- [Dap1]-PMB (3) shifted slightly to the left compared to Des-
FA-[Arg1]-PMB (2) resulted in a 8—16 fold increase in bac- FA-[Dab1]-PMB, and the LPS-binding activity of 3 was
tericidal activity against P. aeruginosa. However, these slightly higher than that of other basic amino acid analogs
analogs (10, 11) had slightly decreased bactericidal activity such as Des-FA-[Lys1]-PMB (1) and Des-FA-[Arg1]-PMB (2)
against E. coli and significantly decreased activity against S. (Fig. 3A). Des-FA-[Trp1]-PMB (6) showed LPS-binding ac-
Typhimurium. These experimental results of the distinct an- tivity similar to that of Des-FA-[Dab1]-PMB. The other neu-
timicrobial potency against three bacteria tested could be ex- tral amino acid analogs (4—8) had poor LPS-binding activity
plained mainly by the structural difference of the bacterial (IC50 ꢀ32 nmol/ml). Des-FA-[Glu1]-PMB (9) competed neg-
cell membranes especially in lipid A portions, which harbor ligibly with dansylated PMB3 bound to LPS. These results
the whole LPS molecule to the lipid layer of membrane. indicate that the LPS binding activity of hydrophilic analogs
Lipid A of E. coli and S. Typhimurium has common compo- (1—4) paralleled their antimicrobial activity against E. coli.
nents of five C14 and one C12 fatty acids, while, that of P.
Des-FA-[Dab-Dab-Dab1]- and Des-FA-[Arg-Arg-Arg1]-
aeruginosa has four C12 and two C10 fatty acids. The mem- PMB (10, 11) had markedly potent binding activity, similar
brane lipid components of P. aeruginosa have less hydropho- to that of polymyxin B. Interestingly, the potent LPS binding

336
Vol. 57, No. 4
activity of 10 and 11 was not reflected in their bactericidal 31.5 mmol/kg).
potency against E. coli, which was moderate, with MIC val-
The present study of N-terminal analogs of polymyxin B
ues of 8 and 16 nmol/ml, respectively (Table 3, Fig. 3B). It is lacking a fatty acyl moiety lead to the synthesis of Des-FA-
evident from these experimental results that the N-terminal [Dap1]-PMB (3), Des-FA-[Ser1]-PMB (4), Des-FA-[Dab-
basic tripeptide portions of 10 and 11 contribute to binding Dab-Dab1]-PMB (10) and Des-FA-[Arg-Arg-Arg1]-PMB
to the LPS molecule, however, the mode of binding appears (11). These compounds showed potent antimicrobial activi-
to be different from that of PMB, since the cationic charac- ty specifically against P. aeruginosa (MIC, 0.5—1 nmol/ml)
teristics of H-Dab-Dab-Dab1- of 10 and H-Arg-Arg-Arg1- of comparable to PMB. The latter two peptides had the same
11 are quite different from that of fatty acyl-Dab1- of PMB. LPS binding activity as PMB. Des-FA-[Dap1]-PMB (3)
Therefore, the mechanism by which 10 and 11 bind to ex- showed 4.90-fold less acute toxicity than PMB. These results
clude the fluorescent probe bound to LPS is hypothesized to provide clues for developing a potent bactericidal analog spe-
be as follows. First, the N-terminal tripeptide cationic por- cific for P. aeruginosa with low acute toxicity.
tions of 10 and 11 bind to phosphate anions of lipid A and/or
References
carboxylate anions of the adjacent 3-deoxy-D-manno-octu-
1) Storm D. R., Rosenthal K. S., Swanson P. E., Annu. Rev. Biochem., 46,
losonic acid (Kdo) portion in LPS molecule. Next, the ionic
binding of 10 and 11 causes structural changes in LPS, re-
sulting in the loss of hydrophobic binding between the fatty
acyl group(s) of the lipid A portion of LPS and the octanoyl-
Dab1 portion of the probe, octanoyl-[Dab(Ng-dansyl-Gly)1]-
PMB3. The structural changes in the hydrophobic portions of
E. coli LPS caused by this strong ionic binding do not partic-
ipate in the destruction of the outer membrane of the living
E. coli cell, since the bactericidal activity of 10 and 11
against E. coli is not improved. As discussed here, the pres-
ent results demonstrate that the binding activity of 10 and 11
to E. coli LPS does not parallel their bactericidal activity
against E. coli. On the other hand, Des-FA-[Ala-Ala-Ala1]-
PMB (12) shows very low binding activity, possibly because
of the deficiency of both ionic and hydrophobic interaction
with LPS.
723—763 (1977).
2) Suzuki T., Hayashi K., Fujikawa K., Tsukamoto K., J. Biochem.
(Tokyo), 56, 335—343 (1964).
3) Chihara S., Tobita T., Yahata M., Ito A., Koyama Y., Agric. Biol.
Chem., 37, 2455—2463 (1973).
4) Vaara M., Vaara, T., Antimicrob. Agents Chemother., 24, 107—113
(1983).
5) Ofek I., Cohen S., Rahmai R., Kabha K., Tamarkin D., Herzig Y., Ru-
binstein E., Antimicrob. Agents Chemother., 38, 374—377 (1994).
6) 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).
7) Sakura N., Itoh T., Uchida Y., Ohki K., Okimura K., Chiba K., Sato Y.,
Sawanishi H., Bull. Chem. Soc. Jpn., 77, 1915—1924 (2004).
8) Okimura K., Ohki K., Sato Y., Ohnishi K., Uchida Y., Sakura N., Bull.
Chem. Soc. Jpn., 80, 543—552 (2007).
9) Okimura K., Ohki K., Sato Y., Ohnishi K., Sakura N., Chem. Pharm.
Bull., 55, 1724—1730 (2007).
10) Wang S., J. Am. Chem. Soc., 21, 1328—1333 (1973).
11) Merrifield R. B., Adv. Enzymol. Relat. Areas Mol. Biol., 32, 221—296
(1969).
12) Shioiri T., Yamada S., Chem. Pharm. Bull., 22, 855—858 (1974).
13) Litchfield J. T., Wilcoxon F., J. Pharmacol. Exp. Ther., 96, 99—113
(1949).
14) Tsubery H., Ofek I., Cohen S., Fridkin M., Peptides, 22, 1675—1681
(2001).
15) Lake P., DeLeo J., Cerasoli F., Logdberg L., Weetall M., Handley D.,
Antimicrob. Agents Chemother., 48, 2987—2992 (2004).
16) Viswanath D. V., Jenkins H. J., J. Pharm. Sci., 67, 1275—1280 (1978).
17) Singh Y. N., Marshall I. G., Harvey A. L., Br. J. Anaesth., 54, 1295—
1306 (1982).
The acute toxicity of Des-FA-[Dap1]-PMB (3) and Des-
FA-[Trp1]-PMB (6) in comparison with PMB15) and PMBN
was examined by tail intravenous bolus injection in mice,
which resulted in lethality within 24 h. The administration of
PMB and its analogs induced respiratory arrest and death
within 5 min after injection. This effect may be caused by
neuromuscular block.16,17) The LD50 values of 3 and 6 were
23.5 and 19.0 mmol/kg, respectively. The results demonstrate
that the acute toxicity of 3 in mice was improved from that of
native PMB (LD50, 4.8 mmol/kg) (Table 4), i.e., 4.90-fold less
toxic than PMB, though still more toxic than PMBN (LD50,