Structure - Function studies of Polymyxin B nonapeptide: Implications to sensitization of Gram-negative bacteria

Article abstract of DOI:10.1021/jm0000057

Polymyxin B nonapeptide (PMBN), a cationic cyclic peptide derived by enzymatic processing from the naturally occurring peptide polymyxin B, is able to increase the permeability of the outer membrane of Gram-negative bacteria toward hydrophobic antibiotics probably by binding to the bacterial lipopolysaccharide (LPS). We have synthesized 11 cyclic analogues of PMBN and evaluated their activities compared to that of PMBN. The synthetic peptides were much less potent than PMBN in their capacity to sensitize Escherichia coli and Klebsiella pneumoniae toward novobiocin and to displace dansyl-PMBN from Escherichia coli LPS. Moreover, unlike PMBN, none of the analogues were able to inhibit the growth of Pseudomonas aeruginosa. The structural - functional features of PMBN were characterized and identified with regard to the ring size, the distance between positive charges and peptide backbone, the chirality of the DPhe-Leu domain, and the nature of the charged groups. Apparently, the structure of PMBN is highly specific for efficient perturbation of the outer membrane of Gram-negative bacteria as well as for LPS binding. The present study further increases our understanding of the complex PMBN-LPS and may, potentially, enable the design of compounds having enhanced permeabilization potency of the Gram-negative outer membrane.

Full text of DOI:10.1021/jm0000057

J . Med. Chem. 2000, 43, 3085-3092
3085
Str u ctu r e-F u n ction Stu d ies of P olym yxin B Non a p ep tid e: Im p lica tion s to
#
Sen sitiza tion of Gr a m -Nega tive Ba cter ia
Haim Tsubery,†,‡ Itzhak Ofek, Sofia Cohen, and Mati Fridkin*
‡
‡
,†
Department of Organic Chemistry, The Weizmann Institute of Science, Rehovot 76100, Israel, and Department of Human
Microbiology, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv 69978, Israel
Received J anuary 5, 2000
Polymyxin B nonapeptide (PMBN), a cationic cyclic peptide derived by enzymatic processing
from the naturally occurring peptide polymyxin B, is able to increase the permeability of the
outer membrane of Gram-negative bacteria toward hydrophobic antibiotics probably by binding
to the bacterial lipopolysaccharide (LPS). We have synthesized 11 cyclic analogues of PMBN
and evaluated their activities compared to that of PMBN. The synthetic peptides were much
less potent than PMBN in their capacity to sensitize Escherichia coli and Klebsiella pneumoniae
toward novobiocin and to displace dansyl-PMBN from Escherichia coli LPS. Moreover, unlike
PMBN, none of the analogues were able to inhibit the growth of Pseudomonas aeruginosa.
The structural-functional features of PMBN were characterized and identified with regard to
the ring size, the distance between positive charges and peptide backbone, the chirality of the
DPhe-Leu domain, and the nature of the charged groups. Apparently, the structure of PMBN
is highly specific for efficient perturbation of the outer membrane of Gram-negative bacteria
as well as for LPS binding. The present study further increases our understanding of the
complex PMBN-LPS and may, potentially, enable the design of compounds having enhanced
permeabilization potency of the Gram-negative outer membrane.
4
In tr od u ction
death. Neutralization of the devastating effects of LPS
is one of the main targets in combating endotoxicosis.5
In the past few years various therapeutic strategies and
novel agents for the treatment of LPS-mediated septic
shock have been evaluated, including anti-LPS antibod-
ies, LPS-neutralizing proteins (BPI and LBP), lipid A
antagonist (lipid X), polymyxin B (PMB), and polymyxin
The accelerated emergence of pathogenic bacteria
exhibiting a multiple pattern of resistance toward
conventional antibiotics is a major threat, especially in
severe infections such as septicemia where mortality is
1
high. The development of novel, versatile antibiotic
arsenal is thus most urgent. Optimal new antimicrobial
agents should withstand enzymatic modification and/
or degradation and be targeted toward a bacterial
component which is readily accessible and not likely to
undergo mutations.
5
-7
B-related synthetic peptides.
Lipid A is a target for
cationic proteins and peptides since it contains nega-
tively charged phosphorylated disaccharides linked, via
ester bonds, to long fatty acid chains which are inter-
8
calated in the bacterial outer membrane.
The bacterial endotoxin lipopolysaccharide (LPS) is
the major antigen of the outer membrane (OM) of Gram-
negative bacteria. It is composed of three major parts,
one of which is lipid A, a highly conserved hydrophobic
PMB is a naturally occurring cationic cyclic decapep-
9
,10
tide isolated from Bacillus polymyxa.
PMB is highly
bactericidal to Gram-negative bacteria and considered
one of the most efficient cell-permeabilizing com-
2
domain of LPS, in Gram-negative bacteria. LPS is
1
1
pounds. This capacity is due to its high-affinity binding
essential for survival of the bacteria through establish-
ing of an effective permeability barrier. Gram-negative
bacteria lacking lipid A have not yet been isolated, and
1
2
to lipid A. However, the therapeutic applications of
PMB are very limited because of its relatively high
1
3,14
toxicity.
Polymyxin B nonapeptide (PMBN) is a
2
the mutation is thought to be lethal. LPS is a predomi-
cyclic peptide obtained from PMB by proteolytic removal
of its terminal amino acyl residue (Scheme 1).15 PMBN
is an extremely poor antimicrobial compound,16 but it
is still capable of binding, like its parent compound, to
LPS, rendering Gram-negative bacteria susceptible to
3
nant inducer of sepsis. Uncontrolled perturbation of
immune cells by LPS followed by cytokine-mediated
damage to blood vessels and decrease in vascular
resistance frequently lead to collapse of organs and
1
7
various hydrophobic antibiotics. This ability of PMBN
to bind to bacteria with relatively high affinity and to
permeabilize their outer membrane is often referred to
as “sensitizing activity” and points at a novel therapeu-
*
To whom correspondence should be addressed. Tel: 972-8-
9
342505. Fax: 972-8-9344142. E-mail: mati.fridkin@weizmann.ac.il.
†
‡
#
The Weizmann Institute of Science.
Tel Aviv University.
Abbreviations: Agb, 2-amino-4-guanidinobutyric acid; CD, circular
18
tic direction. This approach is strengthened by the
dichroism; CFU, colon-forming unit; Dab, 2,4-diaminobutyric acid; Dap,
,3-diaminopropionic acid; DIEA, diisopropylethylamine; MIC, minimal
2
markedly reduced toxicity of the PMBN derivative as
inhibitory concentration; NMM, 4-methylmorpholine; LPS, lipopolysac-
charide; PMB, polymyxin B; PMBN, polymyxin B nonapeptide; PyBOP,
11
compared with its parental PMB molecule. In addition,
an enzyme capable of degrading the cyclic PMBN
peptide has not yet been reported.
(
benzotriazol-1-yloxy)trispyrolidinophosphonium hexafluorophosphate;
TES, triethylsilane; TFA, trifluoroacetic acid; TFE, trifluoroethanol.
1
0.1021/jm0000057 CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/13/2000

3
086 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 16
Tsubery et al.
Sch em e 1. Generation of PMBN from PMB by Enzymatic Processing
Sch em e 2. Synthetic Routes to sPMBN
The binding of PMBN to LPS on bacterial surface and
its membrane permeabilization activity (i.e., sensitiza-
tion) are thought to be consecutive and inseparable.
These capacities of the peptide can serve as a starting
point to novel anti-Gram-negative drugs. In the present
investigation we have performed structure-function
studies of PMBN pertaining to its antibacterial activity.
Our study indicates that the structural parameters
required for LPS binding and membrane permeabiliza-
tion are rather conserved in the PMBN parent molecule.
C-terminal carboxylic function of Thr was protected by
either the 2-chlorotrityl or the Wang polymeric carriers
(Scheme 2A,B, respectively). In route A, following
simultaneous cleavage of the linear peptide from the
resin and removal of the Mtt protection under mild
acidic conditions, cyclization was attempted in high
dilution (1 mg/mL) using four different reagents: (ben-
zotriazol-1-yloxy)trispyrolidinophosphonium hexafluo-
rophosphate (PyBOP), 2-(1H-benzotriazol-1-yl)-1,1,3,3-
tetramethyluronium hexafluorophosphate (HBTU), o-(7-
azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexa-
fluorophosphate (HATU) (all with HOBT in the presence
of NMM as a base), and N,N′-dicyclohexylcarbodiimide/
N-hydroxybenzotriazole (DCC/HOBT). The reactions
were monitored using analytical HPLC (see Experimen-
tal Section). The best results, as judged by yield and
purity, were obtained with PyBOP. Final deprotection
afforded the expected sPMBN, characterized by the
expected mass and amino acid content (Table 1) and by
HPLC identity with pPMBN obtained via proteolysis of
PMB. However, a major byproduct, presumably a result
Resu lts
P ep tid es Syn th esis. Three approaches were evalu-
ated for the synthesis of PMBN: two involved linear
chain assembly on a solid support following by solution
cyclization (Scheme 2A,B), whereas the third was based
on cyclization on the solid support (Scheme 2C). The
9
-fluorenylmethoxycarbonyl (Fmoc) group was used, as
3
a rule, for R-amino protection. γ-Amino groups of Dab
were protected by either 4-methyltrityl (Mtt) (Scheme
2
A) or tert-butyloxycarbonyl (tBoc) (Scheme 2B). The

Structure-Function Studies of PMBN
Ta ble 1. Peptide Analysis
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 16 3087
no.
peptide
MS found (calcd)
tR (min)a
amino acid analysisb
1
2
3
4
5
6
7
8
9
0
1
2
3
pPMBN
sPMBN
962.5613 (962.5661, C43H74N14O11)
962.5637 (962.5661, C43H74N14O11)
1103.7248 (1103.7305, C53H94N14O11)
1033.6538 (1033.6522, C48H84N14O11)
893.4931(893.4957, C38H64N14O11)
991.6041 (991.6052, C45H78N14O11)
1074.6491 (1074.6914, C51H90N14O11)
962.7011 (962.5661, C43H74N14O11)
962.6402 (962.5661, C43H74N14O11)
1018.6754 (1018.6288, C47H82N14O11)
1018.7112 (1018.6288, C47H82N14O11)
962.5788 (962.5661, C43H74N14O11)
1130.6498 (1130.6533, C47H82N22O11)
25.0
25.0
23.5
22.3
23.5
24.7
24.8
24.1
24.9
25.1
24.0
25.0
26.1
T (2) 1.91; F (1) 0.98; L (1) 1; X (5) 4.98
T (2) 1.87; F (1) 0.95; L (1) 1; X (5) 4.9
T (2) 1.85; F (1) 0.98; L (1) 1; K (5) 5.1
T (2) 1.81; F (1) 0.91; L (1) 1; O (5) 4.7
T (2) 1.85; F (1) 0.9; L (1) 1; B (5) 4.6
T (2) 1.75; F (1) 0.9; L (1) 1; X (4) 3.75; K (1) 0.98
T (2) 1.84; F (1) 0.9; L (1) 1; X (1) 0.9; K (4) 3.8
T (2) 1.91; F (1) 0.95; L (1) 1; X (5) 4.75
T (2) 1.9; F (1) 0.95; L (1) 1; X (5) 4.8
T (2) 1.8; F (1) 0.9; L (1) 1; X (3) 2.9; K (2) 1.9
T (2) 1.8; F (1) 0.9; L (1) 1; X (3) 2.9; K (2) 1.9
T (2) 1.86; F (1) 0.95; L (1) 1; X (5) 4.8
T (2) 1.88; F (1) 0.92; L (1) 1; X (1) 0.95
2
,3,4,7,8
[Lys
[Orn
[Dap
[Lys ]PMBN
[Lys
]PMBN
]PMBN
]PMBN
2
2
,3,4,7,8
,3,4,7,8
3
2
,4,7,8
]PMBN
4
9
[cyclo-Dab ,Thr ]PMBN
2
9
[cyclo-Dab ,Thr ]PMBN
2
,4
1
1
1
1
[Lys ]PMBN
7
,8
[Lys ]PMBN
5
[lPhe ]PMBN
2
,4,7,8
[Agb
]PMBN
a
tR, retention time on analytical RP-18 HPLC column (see Experimental Section). b Calculated values are in parentheses. X ) Dab,
T ) Thr, F ) D- or LPhe, K ) Lys, B ) Dap, O ) Orn.
Ta ble 2. Chemical Structures of PMBN and PMBN Analogues
length of alkyl chain of charged aaa
position of aab
no.
peptide
ring size (atoms)
peptide structure
2
3
4
7
8
1
2
3
4
5
6
7
8
9
0
1
2
3
4
pPMBN
sPMBN
23
23
25
24
22
25
23
20
26
23
23
23
23
23
T-X-cyclo[X-X-DF-L-X-X-T]
T-X-cyclo[X-X-DF-L-X-X-T]
T-K-cyclo[K-K-DF-L-K-K-T]
T-O-cyclo[O-O-DF-L-O-O-T]
T-B-cyclo[B-B-DF-L-B-B-T]
T-X-cyclo[K-X-DF-L-X-X-T]
T-K-cyclo[X-K-DF-L-K-K-T]
T-X-X-cyclo [X-DF-L-X-X-T]
T- cyclo[X-X-X-DF-L-X-X-T]
T-K-cyclo[X-K-DF-L-X-X-T]
T-X-cyclo[X-X-DF-L-K-K-T]
T-X-cyclo[X-X-F-L-X-X-T]
T-Z-cyclo[X-Z-DF-L-Z-Z-T]
6-methylheptanoyl/octanoyl-X-T-X-
cyclo[X-X-DF-L-X-X-T]
2
2
4
3
1
2
4
2
2
4
2
2
4
2
2
2
4
3
1
4
2
2
2
2
2
2
2
2
2
2
4
3
1
2
4
2
2
4
2
2
4
2
2
2
4
3
1
2
4
2
2
2
4
2
4
2
2
2
4
3
1
2
4
2
2
2
4
2
4
2
2
,3,4,7,8
[Lys
[Orn
[Dap
[Lys ]PMBN
[Lys
]PMBN
]PMBN
]PMBN
2
,3,4,7,8
2
,3,4,7,8
3
2
,4,7,8
]PMBN
4
9
[cyclo-Dab ,Thr ]PMBN
2
9
[cyclo-Dab ,Thr ]PMBN
2
,4
1
1
1
1
1
[Lys ]PMBN
7
,8
[Lys ]PMBN
5
[lPhe ]PMBN
2
,4,7,8
[Agb
PMB
]PMBN
a
Bold numbers represent amino acids that participate in the formation of the peptide ring. b Numbers are according to Scheme 1.
X ) Dab, T ) Thr, DF ) DPhe, K ) Lys, B ) Dap, O ) Orn, Z ) Agb.
of â-elimination reaction of the C-terminal Thr(tBu)
residue (i.e., R-amino crotonic acid derivative), ac-
companied the desired product. An alternative route
PyBOP was successful, the method was rather laborious
and not suited for fast and efficient synthesis of PMBN
analogues.
2
1
(Scheme 2B) involved cleavage from Wang resin of a
The conversion of pPMBN to [Agb2,4,7,8]PMBN (13)
resulted in good yield, using 3,5-dimethylpyrazole-1-
carboxamidine nitrate, while the R-amino function of
partially protected linear peptide, i.e., possessing free
3
9
γ-amino Dab and R-carboxyl Thr residues, followed by
cyclization using PyBOP/HOBT/NMM as reagents and
base (see Experimental Section). This procedure was
accomplished in less than 1 h. Complete peptide depro-
tection was achieved by catalytic hydrogenation. The
single product obtained was purified to homogeneity by
HPLC and characterized by amino acid analysis, mass
1
Thr was not modified.
Ra tion a le of An a logu es Design . Eleven PMBN
analogues were synthesized (Tables 1 and 2) via route
B (Scheme 2B) in an attempt to locate key structural
features and amino acid residues essential for mem-
brane disorganizing activity of the parent PMBN mol-
1
2,3,4,7,8
spectrometry, H NMR, and its identity with pPMBN
ecule. Thus, three analogous peptides, namely, [Lys
PMBN (3), [Orn
]-
2
,3,4,7,8
]PMBN (4), and [Dap2,3,4,7,8]PMBN
(
HPLC). All 11 synthetic peptides (2-13, Tables 1 and
) evaluated in this study were prepared according to
route B.
An additional synthetic route (Scheme 2C) was initi-
ated by coupling a dipeptide, i.e., Fmoc-Dab-Thr(tBu)-
2
(5), differing in the length of the alkyl chains of the
charged amino acids and in the size of the peptide ring,
were synthesized. The number of the side chain meth-
ylene groups varied from 2 in the original PMBN
molecule to 4, 3, and 1 in the corresponding analogues.
In addition, the ring size of PMBN was modified from
23 atoms in the original molecule to 25, 24, and 22
2
2
OtBu, to benzyl chloroformate resin. Following comple-
tion of chain assembly, the tBoc and tBu protections of
1
3
9
Thr , Dab , and Thr were removed by TFA. Cyclization
was then performed on the resin-bound peptide using
either PyBOP or HATU as reagents. Treatment with
bromotrimethylsilane (TMSBr)/TFA accomplished the
final cleavage and deprotection. Cyclization with PyBOP
led to the desired, fully identical PMBN product. Cy-
clization with HATU afforded, in addition, a byproduct
identified as a tetramethylguanidinium (Tmg) deriva-
tive of Dab.23 Although the cyclization reaction with
3
atoms in the above analogues. [Lys ]PMBN (6) and
2
,4,7,8
[Lys
]PMBN (7) were synthesized in an attempt to
differentiate between the above two factors, i.e., ring
size and length of the alkyl chain of the charged amino
3
acids. Thus, in [Lys ]PMBN (6) the ring size was
extended to 25 atoms by the addition of two methylene
groups. The alkyl chain length of the charged amino
2,4,7,8
acids was changed in [Lys
]PMBN (7) by replacing

3
088 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 16
Tsubery et al.
the Dab residues with Lys while the ring size was left
unchanged. Further modifications in the size of the
4
9
PMBN ring were performed in [cyclo-Dab ,Thr ]PMBN
3
(
8) by subtracting one charged residue (Dab ) from the
2
peptide ring while adding a charged residue (Dab ) in
cyclo-Dab ,Thr ]PMBN (9). Since there are two sepa-
2
9
[
rated positively charged sites at the PMBN structure,
i.e., positions 2,4 and positions 7,8, interspaced by DPhe-
Leu, [Lys2 ]PMBN (10) and [Lys ]PMBN (11) were
synthesized in order to reveal the significance of each
site. Thus, the lengths of the alkyl chains of the charged
residues were extended from 2 to 4 methylene groups.
The hydrophobic core of PMBN DPhe-Leu was modified
,4
7,8
5
by replacing DPhe with LPhe in [LPhe ]PMBN (12). This
substitution could possibly affect the amphiphilic nature
of the molecule. Finally, the γ-amino side chain function
of the Dab residues was modified to yield the corre-
sponding guanidino peptide [Agb2,4,7,8]PMBN (13) in
order to evaluate the effect of enhanced pK value of the
positively charged moiety, i.e., 10.79 of Dab and 12.48
of Agb, and the nature of the charged group.
Affin ity of P MBN a n d An a logu es to LP S. The
interaction of PMB with cell-bound and cell-free forms
of LPS can be quantified using the dansyl-PMB binding
F igu r e 1. Binding curve of dansyl-PMBN to E. coli LPS.
Increasing concentrations of dansyl-PMBN were added to E.
coli LPS solution (3 µg/mL). The augmented fluorescence was
measured 5 min after each addition at excitation and emission
wavelengths of 340 and 485 nm, respectively. Hill plot of
dansyl-PMBN binding (insert) was generated as described
assay.20 The method is based on enhanced dansyl
fluorescence upon the interaction of LPS with the
labeled peptide. Binding experiments with PMBN were
performed by modification of the above method. Thus,
we have synthesized and used dansyl-PMBN rather
than dansyl-PMB. Dansyl-PMBN was synthesized by
reacting the free N-terminal R-amino function of the
peptide with dansyl chloride prior to its cleavage from
the polymeric support. A plot of enhanced fluorescence
as a function of dansyl-PMBN added to LPS is il-
lustrated in Figure 1. A Hill plot (insert) of the data
suggests a cooperative peptide-LPS association. The
molar ratio peptide/LPS as determined from the Hill
slope curve (Hill number, n) was calculated to be 2.3.
According to calculations based on maximal binding
curve (not shown) with excess LPS (400 µg/mL), a rather
similar ratio of dansyl-PMBN/LPS (1.98) was obtained.
The plot intercept (S0.5), i.e., binding affinity, was found
to be 0.41 µM.
(Experimental Section), where BND and bnd are the fluores-
cence determined with 400 and 3 µg/mL LPS, respectively, for
the same dansyl-PMBN concentrations.
The ability of the peptides pPMBN, sPMBN, and the
synthetic analogues to displace dansyl-PMBN (0.55 µM)
from cell-free LPS (3 µg/mL) was evaluated. Several
representative patterns are illustrated in Figure 2. The
calculated values for maximal and 50% displacement,
Imax and IC50, respectively, are summarized in Table 3.
F igu r e 2. Displacement of dansyl-PMBN bound to LPS by
The analogue [Agb2,4,7,8]PMBN (13) was almost as
PMBN and PMBN analogues. Increasing concentrations of
potent as PMBN in displacing dansyl-PMBN from LPS,
2,4,7,8
PMBN peptides [pPMBN ([), sPMBN (4), [Agb
]PMBN
3
having IC50 of 3.8 and 2.5 µM, respectively. [Lys ]-
2,3,4,7,8
7,8
5
(9), [Dap
]PMBN (]), [Lys ]PMBN (O), and [lPhe ]-
2
9
5
PMBN, [cyclo-Dab ,Thr ]PMBN, and [lPhe ]PMBN, i.e.,
peptides 6, 9, and 12, respectively, were somewhat less
PMBN (b) were added to E. coli LPS solution (3 µg/mL) bound
to dansyl-PMBN (0.55 µM). The fluorescence inhibition was
measured 5 min after each addition at excitation and emission
wavelengths of 340 and 485 nm, respectively.
effective, whereas the other analogues were significantly
less active (Table 2). PMBN (1, 2) and [Agb2
,4,7,8
]PMBN
(
13) were able to inhibit 80% of fluorescence, while the
tions, and a minor shoulder at 210-222 nm. The CD
spectrum of sPMBN (2) was, as expected, almost identi-
cal to that of pPMBN (1). Similar patterns were obtain
for the two peptides in trifluoroethanol (TFE)/aqueous
buffer mixture (1:1; v/v) or in bulk TFE. The CD spectra
of all analogues suggest random conformations.
other analogues were capable of up to 64% inhibition
(Table 3).
Cir cu la r Dich r oism (CD) Mea su r em en ts. The CD
spectra of PMBN and analogues were measured at 0.1
mM peptide concentrations in 5 mM PB, pH 7.5.
pPMBN (1) has a maximal negative ellipticity at 200
nm, characteristic of an ensemble of random conforma-
An t im icr ob ia l a n d Sen sit izin g Act ivit y. The
PMBN analogues were tested in vitro for their antimi-

Structure-Function Studies of PMBN
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 16 3089
Ta ble 3. Sensitizing Activity and LPS Binding of PMBN Analoguesa
relative sensitization activity (%)a
E. coli K. pneumoniae
100 100
no.
peptide
IC50 (µM)b
Imax (%)c
1
2
3
4
5
6
7
8
9
pPMBN
sPMBN
2.5
3
35
25
30
8
30
15
8
14
12
8
80
80
60
62
64
60
45
45
60
64
62
64
80
100
96
4
93
3
2
,3,4,7,8
[Lys
[Orn
[Dap
]PMBN
]PMBN
]PMBN
2
,3,4,7,8
2
6
2
,3,4,7,8
11
17
5
5
18
3
3
[Lys ]PMBN
2
,4,7,8
[Lys
]PMBN
4
9
[cyclo-Dab ,Thr ]PMBN
[cyclo-Dab ,Thr ]PMBN
4
5
2
9
7
5
2
,4
1
1
1
1
1
0
1
2
3
4
[Lys ]PMBN
[Lys ]PMBN
9
40
5
13
26
9
7
,8
5
[lPhe ]PMBN
2
,4,7,8
[Agb
PMB
]PMBN
25
25
3.8
0.5
a
Relative sensitizing activity was determined as a percent of pPMBN (50 µg/mL) employed as a control. The MIC values of novobiocin
for E. coli and K. pneumoniae were 125 and 250 µg/mL, respectively, in the absence of peptides. pPMBN at 50 µg/mL reduced the MIC
to 1 and 4 µg/mL novobiocin, representing a sensitization activity of 125 and 62 µg/mL for E. coli and K. pneumoniae, respectively. The
b
numbers are the average values derived from 4-8 experiments. The concentration required to displace 50% of dansyl-PMBN (0.55 µM)
c
from maximal specific binding to E. coli LPS. Maximal displacement of dansyl-PMBN.
crobial activity toward clinical isolates of E. coli, K.
pneumoniae, and P. aeruginosa. As expected, PMBN
and its analogues lacked antimicrobial activity against
the first two bacterial species (i.e., MIC g 1000 µg/mL).
However, the MIC of pPMBN (1) and sPMBN (2) for P.
aeruginosa was 8 µg/mL. The susceptibility of P. aerugi-
nosa to PMBN was previously reported and could
seemingly be attributed to the relative high number of
negative charges on the P. aeruginosa outer mem-
brane. A further study focused on the peptides ability
to disorganize the outer membrane of Gram-negative
bacteria. This activity was evaluated in terms of the
ability to reduce the MIC values of novobiocin toward
E. coli and K. pneumoniae as a function of increased
drug penetration into the bacteria. Both sPMBN and
pPMBN exhibited considerable activity by reducing the
MIC of novobiocin by a factor of 2 orders of magnitude
toward E. coli (from 125 down to 1 µg/mL) and a rather
similar activity against K. pneumoniae (from 250 down
to 4 µg/mL). However, PMBN analogues exhibited
significantly reduced activity at identical concentrations
actions are attributed to the amphiphilic feature of
2
4,26
PMB.
This interaction is translated into major
related hydrophobic peptide-lipid contacts which are
properly directed and strengthened by interactions of
the positively charged side chains of the 2,4-diaminobu-
tyric acid residues of PMB with the negatively charged
phosphate groups of lipid A.
We aimed in the present study to examine the effect
of structural alterations - specifically in the peptide
ring size, distance of positive charges from the peptidic
backbone, and amphiphilic character on the binding of
PMBN to LPS and on its synergistic antimicrobial
activity, i.e., capacity for bacterial cell permeabilization.
Thus, the fluorescence binding-inhibition assay employ-
ing dansyl-PMBN and E. coli LPS enabled to calculate
the number of peptide molecule associated to each LPS
molecule. The results are correlated with a previous
study with dansyl-PMB showing a ratio of ca. 2:1.12 The
LPS-peptide binding assay also enabled to conclude
that none of the PMBN analogues exhibited higher
affinity to LPS than PMBN.
1
2
(
Table 3).
The sensitizing activity of the various PMBN ana-
logues is consistent with the LPS binding affinity. In
general, reduced affinity (IC50 and Imax) to LPS, com-
pared to PMBN, reduced the peptide-mediated pertur-
bation activity and abolished the growth-inhibition
capacity toward P. aeruginosa. The results obtained
(Table 3) suggest that the structural parameters of the
natural pPMBN peptide, namely, ring size of 23 atoms,
a space of 2 methylene groups between the positive
charge and the backbone, the hydrophobic order of
DPhe-Leu, and the nature of the charged group, are
optimal for bacterial sensitization toward hydrophobic
antimicrobial agents. However, other modifications of
the peptidic ring, including insertion of peptidomimetic
elements or suitable enhancement of structural con-
straints pertaining to the stabilization of the “active
conformation” of the peptide, may yield more effective
derivatives. These results could also indicate that the
assembly of the LPS molecule in the bacterial OM has
a specific array and order.
Discu ssion
We have evaluated here three routes for the synthesis
of PMBN. Route B (Scheme 2B) was found to be the
method of choice as determined by its convenient
chemical manipulations, good yield, and quality of
products. The various PMBN analogues described hereby
were prepared successfully using this approach. Route
C, i.e., on-polymer cyclization (Scheme 2C), seems,
however, to have the highest potential as a means for
the synthesis of cyclic peptides. Notably, the removal
of the side chain tBu protecting group from the C-
terminal threonine residue prior the cyclization step
(Scheme 2B,C) minimized substantially the formation
of the R-crotonic acid-related side product.
The association of PMB with the bacterial LPS,
mediated by the lipid A moiety, has been studied rather
intensively with regard to affinity, stoichiometry, coop-
erativity, and molecular features; yet, certain related
1
2,24,25
conflicts were detected.
orimetry, 2D H NMR, and molecular dynamics studies,
it appears that the intermolecular peptide-lipid inter-
Based on isothermal cal-
CD studies suggest, however, that PMBN as well as
its analogues have a random conformation. An active
architecture can, theoretically, be assumed upon as-
1

3
090 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 16
Tsubery et al.
sociation of PMBN and PMB, with their LPS target.
Recently, the solution structure of LPS-bound PMB was
elucidated as an envelope-like fold of the peptide ring
and as a â-turn type II′ for the free peptide.
However, both structures could not be observed by the
CD measurements.
Spectra-Physics SP8800 liquid chromatography system equipped
with an Applied Biosystems 757 variable wavelength absor-
bance detector. The column effluents were monitored by UV
absorbance at 220 nm. Following HPLC purification the
lyophilized peptides (>97% pure) were analyzed, after exhaus-
tive acid hydrolysis and precolumn reaction with 6-amino-
quinolyl N-hydroxysuccinimidylcarbamate (AQC), to ascertain
amino acid composition (Waters 2690 separations module,
Milford, MA). Yields were 35-40%. Molecular weights of
compounds were determined by FAB mass spectrometry on a
VG-high-resolution magnetic sector (Fisons, U.K.) with PEG
or PEG-Me as internal standard or by ESI mass spectrometery
on a VG-platform-II electrospray single quadropole mass
spectrometer (Micro Mass, U.K.). Proton magnetic resonance
spectra were recorded at 25 °C on a Bruker AVANCE 400
spectrometer (400 MHz). J values are given in hertz (Hz).
1
2
6,27
Con clu sion s
The structure of PMBN appears to be highly specific
for efficient perturbation of the outer membrane and
consequent sensitization of the Gram-negative bacteria
toward hydrophobic antibiotics as well as for LPS
binding. Moreover, it is not a mere platform for a set of
positive charges but rather a molecule with precise and
unique topographic features which are essential for
complementarity in LPS binding and subsequent per-
turbation of the Gram-negative outer membrane. In
light of this study, the use of PMBN as a scaffold for
antibiotic targeting should be considered.
2 2
sP MBN (2): H NMR (400 MHz, H O:D O, 9:1; pH ) 2.3) δ
0
6
.69 (d, 3H, J ) 5.9), 0.75 (d, 3H, J ) 5.87), 1.19 (d, 3H, J )
.24), 1.30 (d, 3H, J ) 6.4), 2.07 (m, 10H), 3.10 (m, 10H), 3.95
(d, 1H, J ) 5.0), 4.23 (m, 9H), 7.26 (d, 2H, J ) 6.9) 7.36 (m,
4H), 7.76 (t, 1H, J ) 5.4), 7.88 (d, 1H, J ) 7.6), 8.24 (d, 1H,
J ) 6.4), 8.69 (d, 1H, J ) 4.7), 8.74 (d, 1H, J ) 5.8), 8.98 (d,
1
H, J ) 6).
[
Lys2
,3,4,7,8
1
2 2
]P MBN (3): H NMR (400 MHz, H O:D O, 9:1;
Exp er im en ta l Section
pH ) 3.1) δ 0.82 (d, 3H, J ) 6.5), 0.88 (d, 3H, J ) 6.6), 1.23 (d,
6H, J ) 6.4), 1.33 (d, 6H, J ) 6.43), 1.51 (m, 16H), 1.73 (m,
11H), 1.88 (m, 6H), 2.90 (m, 3H), 3.05 (m, 10H), 3.21 (m, 4H),
3.92 (d, 1H, J ) 3.2), 4.16 (m, 3H), 4.26 (m, 6H), 4.36 (m, 4H),
7.32 (d, 3H, J ) 6.9), 7.40 (m, 4H), 7.52 (d, 1H, J ) 4.81), 7.58
(t, 1H, J ) 5.8), 8.13 (d, 1H, J ) 6.3), 8.33 (d, 1H, J ) 7.54),
8.43 (d, 1H, J ) 6.4), 8.63 (d, 2H, J ) 6.6), 8.70 (d, 1H, J )
7.5), 8.8 (d, 1H, J ) 6.9).
Syn th esis of sP MBN a n d An a logu es 2-12. All protected
amino acids, coupling reagents and polymers were obtained
from Nova Biochemicals (Laufelfingen, Switzerland) or from
Bachem (Bubendorf, Switzerland). Synthesis grade solvents
were obtained from Labscan (Dublin, Ireland). Linear peptide
chains were assembled by conventional solid-phase synthesis,
using an ABIMED AMS-422 automated solid-phase multiple
peptide synthesizer (Langenfeld, Germany). The Fmoc strategy
was employed throughout the peptide chain assembly19 fol-
lowing the company’s commercial protocols. Synthesis was
initiated using Fmoc-Thr(tBu)-Wang resin (0.7 mmol/g) and
performed on a 25-µmol scale. Side chain amino protecting
groups for 2,4-diaminobutyric acid (Dab), lysine (Lys), orni-
thine (Orn), and 2,3-diaminopropionic acid (Dap) were tert-
butyloxycarbonyl (tBoc) and benzyloxycarbonyl (Cbz). Fmoc-
Thr(tBu)-OH was employed as the final building unit. Coupling
was achieved, as a rule, using two successive reactions with 4
equiv of PyBOP as a coupling reagent and 8 equiv of NMM,
all dissolved in DMF. The fully protected peptide-bound resin
was treated with piperidine (20% in DMF) for 20 min, then
washed (DMF), and the free N-terminus amino moiety was
reacted with 4 equiv of Cbz-OSu and 4 equiv of DIEA in DMF
for 3 h. The fully protected peptide-bound resin was then
treated with TFA/water/TES (95:2.5:2.5, v/v/v) for 1 h at room
temperature. The cleavage mixture was cooled to 4 °C, and
the partially protected linear peptides were precipitated with
ice-cold di-tert-butyl methyl ether/petroleum ether (30-40 °C)
2,3,4,7,8
1
[
Da p
2 2
]P MBN (5): H NMR (400 MHz, H O:D O, 9:1;
pH ) 2.38) δ 0.87 (d, 3H, J ) 6.5), 0.92 (d, 3H, J ) 6.6), 1.22
(
1
4
d, 1H, J ) 6.0), 1.38 (d, 3H, J ) 6.5), 1.56 (m, 2H), 2.26 (s,
H), 3.07 (m, 4H), 3.34 (m, 2H), 3.56 (m, 2H), 3.67 (m, 4H),
.09 (d, 1H, J ) 4.6), 4.16 (m, 4H), 4.37 (m, 3H), 4.49 (d, 1H,
J ) 6.8), 7.30 (d, 3H, J ) 7.2), 7.41 (m, 4H), 7.49 (d, 1H, J )
6
9
.1), 7.93 (t, 1H, J ) 6.4), 8.32 (d, 1H, J ) 6.5), 8.71 (m, 3H),
.35 (d, 1H, J ) 7.3).
[
cyclo-Da b ,Th r ]P MBN (9): 1H NMR (400 MHz, H
2
9
2
O:
D
2
O, 9:1; pH ) 2.35) δ 0.79 (d, 3H, J ) 6.4), 0.86 (d, 3H, J )
6
.6), 1.21 (d, 3H, J ) 6.5), 1.31 (t, 2H, J ) 7.3), 1.36 (d, 3H,
J ) 6.5), 1.54 (m, 2H), 2.03 (m, 4H), 3.18 (m, 10H), 3.44 (m,
1
7
7
H), 3.95 (d, 1H, J ) 6.1), 7.31 (d, 3H, J ) 6.9), 7.40 (m, 4H),
.95 (d, 1H, J ) 7.5), 8.11 (t, 1H, J ) 5.6), 8.45 (d, 1H, J )
.1), 8.54 (d, 1H, J ) 6.0) 8.57 (d, 1H, J ) 6.7), 8.65 (d, 1H,
J ) 6.9), 8.68 (d, 1H, J ) 6.4), 8.80 (d, 1H, J ) 7.2), 8.96 (d,
1
H, J ) 6.1).
p P MBN (1). pPMBN was prepared by proteolysis of PMB
with papain or ficin as described elsewhere11 (PMB, papain
and ficin were purchased from Sigma Chemical Co., St. Louis,
MO). Crude product was purified (>98%) by HPLC and
(1:3, v/v) and centrifuged. The pellet was washed with the
same mixture, dissolved in water/acetonitrile (2:3, v/v) and
lyophilized. Cyclization was then performed in DMF at peptide
concentration of 1 mM, using PyBOP/HOBT/NMM (4:4:8,
equiv) as reagents for 2 h at room temperature (yield, >90%
according to analytical HPLC). The reaction mixture was
concentrated in high vacuum and the peptidic cyclic product
was precipitated by treatment with water. Final deprotection,
i.e., removal of Cbz, was achieved by catalytic hydrogenation
1
characterized as described above (yield 60%): H NMR (400
MHz, H
2 2
O:D O, 9:1; pH ) 2.3) δ 0.70 (d, 3H, J ) 5.8), 0.77 (d,
3
H, J ) 5.8), 1.20 (d, 3H, J ) 6.3), 1.32 (d, 3H, J ) 6.4), 2.06
(m, 10H), 3.13 (m, 10H), 3.95 (d, 1H, J ) 5.5), 4.23 (m, 9H),
7.26 (d, 2H, J ) 6.7), 7.36 (m, 4H), 7.78 (t, 1H, J ) 5.6, 5.9),
8.24 (d, 1H, J ) 6.2), 8.49 (d, 1H, J ) 7.8), 8.69 (d, 1H, J )
4.7), 8.74 (d, 1H, J ) 5.8), 8.98 (d, 1H, J ) 6.1).
,4,7,8
[Agb2
]P MBN (13). pPMBN (40 mg, 41.5 µmol) was
(Pd/C) in acetic acid/methanol/water (5:4:1, v/v/v).
Rever sed -P h a se HP LC a n d An a lyses. The crude syn-
dissolved in water (40 mL) and 3,5-dimethylpyrazole-1-car-
boxamidine nitrate (125 mg, 0.62 mmol; Sigma) was added.
The pH of the solution was adjusted to 10 with sodium
carbonate solution (1 M) and the reaction mixture was stirred
overnight at room temperature. Progress of guanidation was
followed using analytical HPLC. The major product was
purified to homogeneity using semipreparative HPLC as
thetic peptides were purified by using a prepacked LichroCart
RP-18 column (250 × 10 mm; 7-µm bead size) employing a
binary gradient formed from 0.1% TFA in water (solution A)
and 0.1% TFA in 75% acetonitrile in water (solution B), eluted
at t ) 0 min, B ) 0%; t ) 48 min, B ) 60%; t ) 60 min, B )
1
00% at a flow rate of 5 mL/min. For purity evaluation,
analytical reversed-phase HPLC was performed using a pre-
packed Lichrospher-100 RP-18 column (250 × 4 mm, 5-µm
bead size) and the following binary gradient: t ) 0 min, B )
described above, characterized by amino acid analysis and
1
mass spectrometry: yield, 82%; H NMR (400 MHz, H
2
O:D
2
O,
9:1; pH ) 2.30) δ 0.71 (s, 3H), 0.79 (s, 4H), 1.24 (d, 3H, J )
6.4), 1.31 (t, 2H, J ) 7.3), 1.35 (d, 3H, J ) 6.4), 1.5 (m, 2H),
2.03 (m, 9H), 3.10 (m, 3H), 3.23 (m, 2H), 3.36 (m, 6H), 3.97 (d,
1
0%; t ) 40 min, B ) 60%; t ) 50 min, B ) 100% at a flow
rate of 0.8 mL/min. Separations were performed using a

Structure-Function Studies of PMBN
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 16 3091
1
7
H, J ) 5.8), 4.25 (m, 6H), 4.55 (m, 2H), 6.8 (broad s, 12H),
.15 (t, 1H, J ) 5.3), 7.34 (m, 9H), 7.73 (t, 1H, J ) 5.5), 7.90
were dissolved in 5 mM phosphate buffer (PB), pH 7.2, at a
final concentration of 0.1 mM. The CD of pPMBN and sPMBN
were evaluated, as well, in a mixture of TFE/aqueous buffer
(1:1, v/v) or in bulk TFE. A baseline was recorded and
subtracted after each spectrum. Ellipticity is reported as the
(
8
d, 1H, J ) 7.3), 8.22 (d, 1H, J ) 6.2), 8.49 (d, 1H, J ) 7.6),
.66 (m, 2H), 8.76 (d, 1H, J ) 5.4), 8.96 (d, 1H, J ) 7.0).
Dan syl-P MBN. Labeling of PMBN with dansyl was achieved
2
-1
mean residue ellipticity [Θ] in deg cm dmol . [Θ] ) [Θ]obs
MRW/10lc), where [Θ]obs is the ellipticity measured in mdeg,
MRW is the mean residue molecular weight of the peptide
molecular weight divided by the number of peptide bonds), c
×
by the following method: resin-bound linear peptide (25 µmol)
see above) was treated with piperidine (20% v/v, in DMF) in
(
(
order to remove the Fmoc protecting group from the N-
terminal R-amino function of the threonine. The resin-bound
peptide was then reacted with dansyl chloride (100 µmol, 27
mg; Sigma) and NMM (200 µmol, 22 µL) in DMF for 2 h. The
labeled peptide was cleaved off the resin, cyclized, purified by
semipreparative HPLC (>98%), and characterized as described
above: yield, 40%; m/z 1197.6 (calcd 1197.2); Thr, 1.1; Dab,
(
is the concentration of the sample in mg/mL, and l is the
optical path length of the cell in cm.
Su p p or tin g In for m a tion Ava ila ble: CD spectra of com-
pounds 1, 2, 5, 12, and 13 are available free of charge via the
Internet at http://pubs.acs.org.
4
.9; DPhe, 1; Leu, 1; t
Deter m in a tion of Min im a l In h ibitor y Con cen tr a tion
MIC). Clinical isolates of E. coli, K. pneumoniae and P.
R
, 33.54 min.
(
Refer en ces
1
8
aeruginosa, obtained as described elsewhere, were employed.
(
1) Stone, R. Search for Sepsis Drugs Goes On Despite Past Failures.
The Gram-negative bacteria were grown on nutrient agar
plates (Difco Laboratories, Detroit, MI) and kept at 4 °C. An
overnight culture in isotonic sensitest broth (ISB, Oxoid) was
Science 1994, 264, 365-367.
(2) Rietschel, E. T.; Kirikae, T.; Schade, F. U.; Mamat, U.; Schmidt,
G.; Loppnow, H.; Ulmer, AJ .; Zahringer, U.; Seydel, U.; Di
Padova, F.; Schreier, M.; Brade, H. Bacterial Endotoxin: Mo-
lecular Relationships of Structure to Activity and Function.
FASEB J . 1994, 8, 217-225.
5
adjusted to 1 × 10 CFU/mL and inoculated into microtiter
plate wells containing each 100 µL of a serial 2-fold dilution
(1000-0.5 µg/mL) of the tested antibiotics in ISB. MIC was
(3) Rietschel, E. T.; Brade, H. Bacterial Endotoxins Sci. Am. 1992,
defined as the lowest concentration at which there was no
visible bacterial growth after incubation for 20 h at 37 °C. The
results are reported for 4-8 separate tests that varied by no
more than one dilution.
267, 26-33.
(4) Karima, R.; Matsumoto, S.; Higashi, H.; Matsushima, K. The
Molecular Pathogenesis of Endotoxic Shock and Organ Failure.
Mol. Med. Today 1999, 5, 123-132.
(
5) Lynn, W. A.; Cohen, J . Adjunctive Therapy for Septic Shock: A
Review of Experimental Approaches. Clin. Infect. Dis. 1995, 20,
143-158.
6) Lynn, W. A.; Golenbock, D. T. Lipopolysaccharide Antagonists.
Immunol. Today 1992, 13, 271-276.
7) Pollack, M.; Ohl, C. A. B. Endotoxin-Based Molecular Strategies
for the Prevention and Treatment of Gram-Negative Sepsis and
Septic Shock. Curr. Top. Microbiol. Immunol. 1996, 216, 275-
297.
8) Vaara, M. Agents that Increase the Permeability of the Outer
Membrane. Microbiol. Rev. 1992, 56, 395-411.
9) Ainsworth, G. C.; Brown, A. M.; Brownlee, G. ‘Aerosporin’, an
Antibiotic Produced by Bacillus aerosporus Greer. Nature 1947,
160, 263.
10) Benedict, R. G.; Langlykke, A. F. Antibiotic Activity of Bacillus
polymyxa. J . Bactriol. 1947, 54, 24-25.
11) Danner, R. L.; J oiner, K. A.; Rubin, M.; Patterson, W. H.;
J ohnson, N.; Ayers, K. M.; Parrillo, J . E. Purification, Toxicity,
and Antiendotoxin Activity of Polymyxin B Nonapeptide. Anti-
microb. Agents Chemother. 1989, 33, 1428-1434.
12) Moore, R. A.; Bates, N. C.; Hancock, R. E. Interaction of
Polycationic Antibiotics with Pseudomonas aeruginosa Lipo-
polysaccharide and Lipid A Studied by Using Dansyl-Polymyxin.
Antimicrob. Agents Chemother. 1986, 29, 496-500.
13) Carig, W. A.; Turner, J . H.; Kunin, C. M. Prevention of the
Generalized Shwartzman Reaction and Endotoxin Lethality by
Polymyxin B Localized in Tissues. Infect. Immunity 1974, 10,
5
Sen sitizin g Activity. Bacterial suspension (10 µL, 1 × 10
CFU) was inoculated into microtiter plate wells containing 100
(
µL of a serial 2-fold dilution (1000-0.5 µg/mL) of novobiocin
(Sigma) in ISB. To each well was added 10 µL of the test
(
peptide to a final concentration of 50 µg/mL. The fold decrease
in MIC for novobiocin comparing wells with and without test
peptides was calculated and designated as sensitizing activity.
The relative sensitizing activity of the test peptide was
calculated as percent of the sensitizing activity of 50 µg/mL
pPMBN.
(
(
Bin d in g of P ep tid es to LP S. The fluorescence of dansyl-
PMBN bound to E. coli LPS was measured using a MC200
monochromator (SLM Aminco, SLM Instruments, Inc.) set at
an excitation wavelength of 340 nm and an emission wave-
(
(
2
0
length of 485 nm. To a quartz cuvette containing LPS
solution (2 mL, 3 µg/mL) in N-[2-hydroxyethyl]piperazine-N′-
(
[
2-ethanesulfonic acid] buffer (HEPES; 5 mM, pH 7.2) was
-6
-3
added 5 or 10 µL of dansyl-PMBN solution (1 × 10 -1 × 10
M) at 5-min intervals for up to 1 h until a plateau in the
fluorescence intensity was reached (i.e., saturation of dansyl-
PMBN binding to LPS). The amount of dansyl-PMBN bound
to LPS at saturation was calculated as described elsewhere.12
Briefly, a binding curve of dansyl-PMBN to excess LPS (400
µg/mL) was plotted and referred to as maximal binding (Fmax).
The amount of bound dansyl-PMBN was calculated from:
(
2
87-292.
(
14) Rifkind, D. Prevention by Polymyxin B of Endotoxin Lethality
in Mice. J . Bacteriol. 1967, 93, 1463-1464.
(
15) Chihara, S.; Tobita, T.; Yahata, M.; Ito, A.; Koyana, Y. Enzymatic
Degradation of Colistin Isolation and Identification of R-N-acetyl
R,γ-Diaminobutyric Acid and Colistin Nonapeptide. Agric. Biol.
Chem. 1973, 37, 2455-2463.
[
dansyl-PMBN] ) (Fexp/Fmax) × [dansyl-PMBNtotal], where Fexp
is the fluorescence obtained with 3 µg/mL LPS.
To determine the binding of a given peptide to LPS, a
(
16) Duwe, A. K.; Rupar, C. A.; Horsman, G. B.; Vas, S. I. In Vitro
Cytotoxicity and Antibiotic Activity of Polymyxin B Nonapeptide.
Antimicrob. Agents Chemother. 1986, 30, 340-341.
displacement assay was performed in which 5 or 10 µL (1 ×
-
5
-3
1
0
-1 × 10 M) of the tested peptide at desired concentra-
tions were added at 5-min intervals to a preequilibrated
(17) Vaara, M.; Vaara, T. Polycations Sensitize Enteric Bacteria to
-
7
Antibiotics. Antimicrob. Agents Chemother. 1983, 24, 107-113.
mixture of LPS solution (2 mL, 3 µg/mL, ∼2 × 10 M) in
HEPES buffer (5 mM, pH 7.2) and dansyl-PMBN (0.55 µM).
The fluorescence intensity was recorded after each time
interval of adding the tested peptide. Each experiment was
repeated 2-3 times. The percent inhibition of fluorescence
intensity was plotted as a function of the peptide concentration
from which the concentration required for maximal (Imax) and
(
18) Ofek, I.; Cohen, S.; Rahmani, R.; Kabha, K.; Herzig, Y.; Rubin-
stein, E. Antibacterial Synergism of Polymyxin B Nonapeptide
and Hydrophobic Antibiotics in Experimental Gram-Negative
Infections in Mice. Antimicrob. Agents Chemother. 1994, 38,
374-377.
(
19) Atheron, E.; Sheppard, R. C. Solid-Phase Peptide Synthesis: A
Practical Approach; IRL Press: Oxford, England, 1989.
20) Fidai, S.; Farmer, S. W.; Hancock, R. E. Interaction of Cationic
Peptides with Bacterial Membranes. In Methods in Molecular
Biology; Shafer, W. M., Ed.; Humana Press: Totowa, NJ , 1997;
pp 187-204.
21) Chino, N.; Kimura, T.; Sakakibara S. A Side Reaction Observed
During Segment Condensation Reactions at the C-Trminal
Benzyl-Threonyl Residue. In Proceedings of the 1st Akabori
Conference; Wunsch, E., Ed.; Max Planck Institute for Biochem-
istry: Martinsried, Germany, 1985; pp 3-6.
(
5
0% (IC50) displacement of the dansyl-PMBN from LPS was
derived.
Cir cu la r Dich r oism (CD) Stu d ies. CD spectra were
recorded on an Aviv-202 circular dichroism spectrometer
Lakewood, NJ ). Duplicate scans over a wavelength range of
(
(
1
0
90-250 nm were taken at a chart speed of 12 nm/min in a
.1-cm path length quartz cell at room temperature. Peptides

3
092 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 16
Tsubery et al.
(
22) Sklyarov, Yu. L.; Shashkova, I. V. Synthesis of the Glycine
Analogue of Gramicidin S on a Polymyric Carrier. Zh. Obshch.
Khimii 1969, 39, 2778-2783.
(26) Bhattacharjya, S.; David, S. A.; Mathan, V. I.; Balaram, P.
Polymyxin B Nonapeptide: Conformations in Water and in
Lipopolysaccharide-Bound State Determined by Two-Dimen-
sional NMR and Molecular Dynamics. Biopolymers 1997, 41,
251-265.
(27) Pristovsek, P.; Kidric, J .; Solution Structure of Polymyxins B
and E and Effect of Binding to Lipopolysaccharide: An NMR
and Molecular Modeling Study. J . Med. Chem. 1999, 42,
4604-4613.
(
23) Story, S. C.; Aldrich J . V. Side-Product Formation During
Cyclization with HBTU on a Solid Support. Int. J . Pept. Protein
Res. 1994, 43, 292-296.
(
24) Srimal, S.; Surolia, N.; Balasubramanian, S.; Surolia, A. Titra-
tion Calorimetric Studies to Elucidate the Specificity of the
Interactions of Polymyxin B with Lipopolysaccharides and Lipid
A. Biochem. J . 1996, 315, 679-686.
(25) Schindler, M.; Osborn, M. J . Interaction of Divalent Cations and
Polymyxin B with Lipopolysaccharide. Biochemistry 1979, 18,
4
425-4430.
J M0000057