Structure - Activity studies of 14-helical antimicrobial β-peptides: Probing the relationship between conformational stability and antimicrobial potency

Article abstract of DOI:10.1021/ja0270423

Antimicrobial α-helical α-peptides are part of the host-defense mechanism of multicellular organisms and could find therapeutic use against bacteria that are resistant to conventional antibiotics. Recent work from Hamuro et al. has shown that oligomers of

Full text of DOI:10.1021/ja0270423

Published on Web 10/02/2002
Structure-Activity Studies of 14-Helical Antimicrobial
â-Peptides: Probing the Relationship between Conformational
Stability and Antimicrobial Potency
†
†
‡
,†
Tami L. Raguse, Emilie A. Porter, Bernard Weisblum, and Samuel H. Gellman*
Contribution from the Department of Chemistry and Department of Pharmacology,
UniVersity of Wisconsin, Madison, Wisconsin 53706
Received May 24, 2002
Abstract: Antimicrobial R-helical R-peptides are part of the host-defense mechanism of multicellular
organisms and could find therapeutic use against bacteria that are resistant to conventional antibiotics.
Recent work from Hamuro et al. has shown that oligomers of â-amino acids (“â-peptides”) that can adopt
an amphiphilic helix defined by 14-membered ring hydrogen bonds (“14-helix”) are active against Escherichia
coli [Hamuro, Y.; Schneider, J. P.; DeGrado, W. F. J. Am. Chem. Soc. 1999, 121, 12200-12201]. We
have created two series of cationic 9- and 10-residue amphiphilic â-peptides to probe the effect of 14-helix
stability on antimicrobial and hemolytic activity. 14-Helix stability within these series is modulated by varying
the proportions of rigid trans-2-aminocyclohexanecarboxylic acid (ACHC) residues and flexible acyclic
residues. We have previously shown that a high proportion of ACHC residues in short â-peptides encourages
14-helical structure in aqueous solution [Appella, D. H.; Barchi, J. J.; Durell, S. R.; Gellman, S. H. J. Am.
Chem. Soc. 1999, 121, 2309-2310]. Circular dichroism of the â-peptides described here reveals a broad
range of 14-helix population in aqueous buffer, but this variation in helical propensity does not lead to
significant changes in antibiotic activity against a set of four bacteria. Several of the 9-mers display antibiotic
activity comparable to that of a synthetic magainin derivative. Among these 9-mers, hemolytic activity
increases slightly with increasing 14-helical propensity, but all of the 9-mers are less hemolytic than the
magainin derivative. Previous studies with conventional peptides (R-amino acid residues) have provided
conflicting evidence on the relationship between helical propensity and antimicrobial activity. This uncertainty
has arisen because R-helix stability can be varied to only a limited extent among linear R-peptides without
modifying parameters important for antimicrobial activity (e.g., net charge or hydrophobicity); a much greater
range of helical stability is accessible with â-peptides. For example, it is very rare for a linear R-peptide to
display significant R-helix formation in aqueous solution and manifest antibacterial activity, while the linear
â-peptides described here range from fully unfolded to very highly folded in aqueous solution. This study
shows that â-peptides can be unique tools for analyzing relationships between conformational stability and
biological activity.
Introduction
be highly amphiphilic, with cationic side chains and hydrophobic
side chains segregated into distinct regions of the molecular
Multicellular organisms employ diverse, short peptides to
defend against microbial invaders.1-3 Host-defense peptides are
usually cationic, and they appear to act by permeabilizing
microbial membranes, although other mechanisms of action have
been proposed. Many of these peptides display specific second-
ary structures in the presence of target membranes. R-Helical
surface. The amphiphilicity is presumably related to the mode
of action: cationic charge directs the peptides to anionic
bacterial membranes, and hydrophobic side chains then interact
with the core of the lipid bilayer, ultimately compromising the
barrier function of the membrane. Enantiomers of natural host-
defense peptides often exert native biological effects, implying
that antimicrobial activity does not involve interaction between
these peptides and specific bacterial protein targets.
7,8
4
,5
conformations are adopted, for example, by magainins (from
6
frogs) and cecropins (from insects). The folded forms tend to
Host-defense peptides have been of interest from a therapeutic
perspective because their proposed mode of action is not
conducive to rapid development of bacterial resistance. These
*
To whom correspondence should be addressed. E-mail: gellman@
chem.wisc.edu.
†
9
Department of Chemistry.
Department of Pharmacology.
‡
(
(
(
1) Boman, H. G. Annu. ReV. Immunol. 1995, 13, 61-92.
(6) Steiner, H.; Hultmark, D.; Engstrom, A.; Bennich, H.; Boman, H. G. Nature
1981, 292, 246-248.
2) Nicolas, P.; Mor, A. Annu. ReV. Microbiol. 1995, 49, 277-304.
3) Hancock, R. E. W.; Chapple, D. S. Antimicrob. Agents Chemother. 1999,
(7) Wade, D.; Boman, A.; Wahlin, B.; Drain, C. M.; Andreu, D.; Boman, H.
G.; Merrifield, R. B. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 4761-4765.
(8) Bessalle, R.; Kapitkovsky, A.; Gorea, A.; Shalit, I.; Fridkin, M. FEBS Lett.
1990, 274, 151-155.
4
3, 1317-1323.
(
(
4) Zasloff, M. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 5449-5453.
5) Williams, R. W.; Starman, R.; Taylor, K. M. P.; Gable, K.; Beeler, T.;
Zasloff, M.; Covell, D. Biochemistry 1990, 29, 4490-4496.
(9) Hancock, R. E. W. Lancet 1997, 349, 418-422.
12774
9
J. AM. CHEM. SOC. 2002, 124, 12774-12785
10.1021/ja0270423 CCC: $22.00 © 2002 American Chemical Society

14-Helical Antimicrobial â-Peptides
A R T I C L E S
peptides or synthetic analogues may provide new tools in the
struggle against pathogenic strains that are resistant to conven-
tional antibiotics. This prospect has inspired many structure-
activity studies based on natural and designed antimicrobial
peptides. Tossi et al.10 have comprehensively reviewed R-helical
antimicrobial peptides. These authors note that the parameters
influencing antimicrobial activity include size, conformational
stability, net charge, net hydrophobicity, amphiphilicity, and the
widths of the hydrophobic and hydrophilic helix faces. Extrac-
tion of general principles from these structure-activity studies
can be challenging, however, because sequence alterations
frequently affect more than one physical parameter.
Figure 1. Schematic representation of the position of the â-amino acid
side chains looking down the 14-helical axis (three residues per turn).
Amphiphilic â-peptides 1-7 are represented by illustration A and non-
3
amphiphilic 8 is represented by B. Positively charged â -homolysine side
3
chains are designated by + and hydrophobic side chains of ACHC, â -
3
3
homoleucine, and â -homovaline are designated by H. â -Homotyrosine
side chains of 5-8 are not shown.
Amphiphilic helix-forming oligomers of â-amino acids (“â-
peptides”) that display varying degrees of antimicrobial activity
have been recently reported.1
1-16
Antimicrobial activity has been observed for both 14-
Earlier work showed that
helical1
1,13
and 12-helical
12,14,16
â-peptides. The 14-helix has
â-peptides can adopt several distinct helical conformations,
1
7-20
approximately three residues per turn; DeGrado and co-
depending upon residue substitution pattern.
â-Peptides
1
1,13
comprised of â-substituted residues21 or R-substituted residues
22
workers
prepared antimicrobial versions by linking hydro-
phobic-cationic-hydrophobic residue triads to one another (as
shown in Figure 1A). The resulting 14-helices have a polar
surface that comprises roughly one-third (120°) of the helix
adopt a “14-helix,” which contains 14-membered ring i f i-2
CdO‚‚‚H-N hydrogen bonds. (The familiar R-helix of con-
ventional peptides contains 13-membered ring i f i+4
CdO‚‚‚H-N hydrogen bonds.) The 14-helix is observed also
when â-peptide residues have a cyclohexyl backbone constraint
1
1,13
circumference.
The 12-helix, on the other hand, has ap-
proximately 2.5 residues per turn. Antimicrobial versions were
generated by linking cationic-hydrophobic-cationic-hydrophobic-
hydrophobic pentads (one pentad ) two helical turns), to give
2
3,24
(e.g., trans-2-aminocyclohexanecarboxylic acid (ACHC)).
â-Peptides with an alternating sequence of R- and â-substituted
residues display a “12/10/12-helix,” which contains both 12-
membered ring i f i+3 CdO‚‚‚H-N hydrogen bonds and 10-
1
(
2-helices with a polar surface that covers roughly two-fifths
1
2,14,16
ca. 144°) of the helix circumference.
Here, we examine structure-activity relationships among 14-
2
5,26
membered ring i f i-1 CdO‚‚‚H-N hydrogen bonds.
Use
helical â-peptides. We are particularly interested in the relation-
ship between conformational stability and biological activity.
â-Peptides are intriguing subjects from this perspective since
the 14-helical propensity of an individual residue can be
profoundly enhanced by switching from an acyclic backbone
to a cyclohexane-constrained backbone. We have reported a
homologous series of hexa-â-peptides in which the proportion
of cyclohexane-constrained and acyclic (â-substituted) residues
of â-amino acids with a cyclopentyl constraint, for example,
trans-2-aminocyclopentanecarboxylic acid (ACPC), leads to
formation of the “12-helix,” which contains exclusively 12-
2
7,28
membered ring i f i+3 CdO‚‚‚H-N hydrogen bonds.
Oligomers of â-amino acids constrained by cis-substituted
oxetane rings adopt a “10-helical” conformation, which contains
2
9
10-membered ring i f i-1 CdO‚‚‚H-N hydrogen bonds.
3
0
was varied. High population of the 14-helix was observed in
aqueous solution when four or more of the six residues were
cyclohexane-constrained, but little or no 14-helix could be
detected in the absence of cyclohexane-constrained residues.
The R-helical stability of conventional peptides cannot be
modulated to this large extent because no R-amino acid residues
have sufficient folding propensity to generate R-helices at the
hexamer length.
(
10) Tossi, A.; Sandri, L.; Giangaspero, A. Biopolymers 2000, 55, 4-30.
11) Hamuro, Y.; Schneider, J. P.; DeGrado, W. F. J. Am. Chem. Soc. 1999,
21, 12200-12201.
12) Porter, E. A.; Wang, X. F.; Lee, H. S.; Weisblum, B.; Gellman, S. H. Nature
(
1
(
2
000, 404, 565.
(
13) Liu, D.; DeGrado, W. F. J. Am. Chem. Soc. 2001, 123, 7553-7559.
(
14) Porter, E. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc. 2002, 124,
7
324-7330.
(
15) Arvidsson, P. I.; Frackenpohl, J.; Ryder, N. S.; Liechty, B.; Petersen, F.;
Zimmermann, H.; Camenisch, G. P.; Woessner, R.; Seebach, D. Chem-
BioChem 2001, 2, 771-773.
(
16) LePlae, P. R.; Fisk, J. D.; Porter, E. A.; Weisblum, B.; Gellman, S. H. J.
Am. Chem. Soc. 2002, 124, 6820-6821.
Several groups have evaluated the relationship between
R-helical stability and biological activity for antimicrobial
(17) Seebach, D.; Matthews, J. L. Chem. Commun. 1997, 2015-2022.
(18) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173-180.
(19) DeGrado, W. F.; Schneider, J. P.; Hamuro, Y. J. Pept. Res. 1999, 54, 206-
10
peptides, but the results have varied from system to system.
In almost every case it has been necessary to compare extents
of R-helix formation within a series in trifluoroethanol (TFE)/
water mixtures because the peptides do not fold detectably in
water; the antimicrobial studies, of course, have been conducted
in aqueous solution. Some groups have examined the effect of
2
17.
(
20) Gademann, K.; Hintermann, T.; Schreiber, J. V. Curr. Med. Chem. 1999,
6
, 905-925.
(
21) Seebach, D.; Overhand, M.; K u¨ hnle, F. N. M.; Martinoni, B.; Oberer, L.;
Hommel, U.; Widmer, H. HelV. Chim. Acta 1996, 79, 913-941.
22) Hintermann, T.; Seebach, D. Synlett 1997, 437-438.
(
(
23) Appella, D. H.; Christianson, L. A.; Karle, I. L.; Powell, D. R.; Gellman,
S. H. J. Am. Chem. Soc. 1999, 121, 6206-6212.
31
enhancing R-helical stability. Chen et al. showed that changing
(
24) Appella, D. H.; Christianson, L. A.; Karle, I. L.; Powell, D. R.; Gellman,
S. H. J. Am. Chem. Soc. 1996, 118, 13071-13072.
two glycines in magainin II to alanine led to enhanced R-helicity
in aqueous TFE and to greater antibiotic activity against several
(
25) Seebach, D.; Abele, S.; Gademann, K.; Guichard, G.; Hintermann, T.; Jaun,
B.; Matthews, J. L.; Schreiber, J. V.; Oberer, L.; Hommel, U.; Widmer, H.
HelV. Chim. Acta 1998, 81, 932-982.
(26) Seebach, D.; Gademann, K.; Schreiber, J. V.; Matthews, J. L.; Hintermann,
(29) Claridge, T. D. W.; Goodman, J. M.; Moreno, A.; Angus, D.; Barker, S.
F.; Taillefumier, C.; Watterson, M. P.; Fleet, G. W. J. Tetrahedron Lett.
2001, 42, 4251-4255.
T.; Jaun, B.; Oberer, L.; Hommel, U.; Widmer, H. HelV. Chim. Acta 1997,
8
0, 2033-2038.
(
27) Appella, D. H.; Christianson, L. A.; Klein, D. A.; Powell, D. R.; Huang,
X. L.; Barchi, J. J.; Gellman, S. H. Nature 1997, 387, 381-384.
28) Applequist, J.; Bode, K. A.; Appella, D. H.; Christianson, L. A.; Gellman,
S. H. J. Am. Chem. Soc. 1998, 120, 4891-4892.
(30) Appella, D. H.; Barchi, J. J.; Durell, S. R.; Gellman, S. H. J. Am. Chem.
Soc. 1999, 121, 2309-2310.
(
(31) Chen, H. C.; Brown, J. H.; Morell, J. L.; Huang, C. M. FEBS Lett. 1988,
236, 462-466.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002 12775

A R T I C L E S
Raguse et al.
bacterial species. In contrast, Houston et al.32 found that
stabilizing the R-helical conformation of cecropin-melittin
hybrids by introducing lactam bridges between glutamic acid
and lysine side chains (i, i+4) led to a decrease in antimicrobial
activity, relative to noncyclized analogues. The cyclized peptides
displayed partial R-helix formation in aqueous solution, while
the noncyclized analogues displayed substantial R-helix forma-
tion only in aqueous TFE. These authors concluded that
excessive R-helix stability is detrimental to antimicrobial
activity.
Here we compare nine- and ten-residue â-peptides that form
flexible amphiphilic 14-helices with analogues that have been
rigidified by incorporation of the cyclohexyl-rigidified ACHC
3
residue. The residues replaced by ACHC, â -homovaline and
3
â -homoleucine, are expected to have hydrophobicities com-
parable to that of ACHC. The â-peptides containing ACHC
display at least partial 14-helix in aqueous solution, while the
â-peptides containing only acyclic residues do not; therefore,
these â-peptides allow us to examine the effect of very high
helical propensity on biological activity without resorting to side
chain bridging, a modification that is required for R-helix
Complementary studies have involved R-helix destabilization,
either by incorporation of D residues or by incorporation of
proline. Dathe et al.33 reported a systematic study based on a
designed 18-residue peptide containing only lysine, leucine, and
alanine that forms an amphiphilic R-helix. These workers
compared the all-L peptide to nine diastereomers in which
adjacent residue pairs had D configuration. The double-D
replacements diminished R-helical stability, with the greatest
effect from replacements at the center of the sequence. Interest-
ingly, the least stable diastereomer displayed only a 3-fold
decrease in R-helix population relative to the all-L isomer in
aqueous TFE. Double-D replacements at the termini had
relatively little effect on antimicrobial activity (Escherichia coli
and Staphylococcus epidermidis), but replacements at the center
caused significant diminution of activity, up to a 16-fold increase
3
8
stabilization among conventional peptides.
Results
Design. We based the sequences of â-peptides 1-4 on 1-ent,
11
reported by Hamuro et al. to have moderate activity against
E. coli. We expected that increasing the proportion of ACHC
residues from zero (â-peptide 2) to two-thirds (â-peptide 4)
3
1
in minimal inhibitory concentration (MIC). Chen et al. found
that introducing three D-alanine residues into a magainin
analogue increased MIC values by 10- to 100-fold relative to
the all-L diastereomer. In contrast, Shai and Oren found that
3
4
incorporating several D residues into the toxins pardaxin or
35
melittin had little effect on antimicrobial activity. Zhang et
3
6
al. compared several related peptides containing zero, one, or
two prolines. In general, antimicrobial activity decreased as the
number of prolines increased (variations in this trend may have
resulted from other sequence changes). These workers sum-
marized conflicting results obtained by others regarding the
effect of proline introduction or removal on antimicrobial activ-
ity. Overall, the available results from conventional peptide stud-
ies do not provide consistent conclusions regarding the relation-
ship between R-helical stability and antimicrobial activity.
Peptides that disrupt bacterial membranes cannot have
therapeutic utility unless they leave human cell membranes
intact. Human cell membrane susceptibility to physical disrup-
tion is generally evaluated via hemolysis (lysis of red blood
cells). In most of the studies cited above, R-helix destabilization
led to a decrease in hemolytic activity. In some cases, the
diminution of hemolytic activity was greater than the effect on
antibacterial activity, leading some workers to propose that
R-helical stability is more important for the former than the
3
3-35
latter.
Indeed, Shai and co-workers have reported am-
phiphilic peptides with features designed to minimize R-helicity
(
e.g., 33% D residues; N-to-C cyclization) that show significant
37
antimicrobial activity but little hemolytic activity.
(
32) Houston, M. E.; Kondejewski, L. H.; Karunaratne, D. N.; Gough, M.; Fidai,
S.; Hodges, R. S.; Hancock, R. E. W. J. Pept. Res. 1998, 52, 81-88.
33) Dathe, M.; Schumann, M.; Wieprecht, T.; Winkler, A.; Beyermann, M.;
Krause, E.; Matsuzaki, K.; Murase, O.; Bienert, M. Biochemistry 1996,
(
3
5, 12612-12622.
(34) Shai, Y.; Oren, Z. J. Biol. Chem. 1996, 271, 7305-7308.
(
(
35) Oren, Z.; Shai, Y. Biochemistry 1997, 36, 1826-1835.
36) Zhang, L.; Benz, R.; Hancock, R. E. W. Biochemistry 1999, 38, 8102-
8
111.
(
37) Oren, Z.; Shai, Y. Biochemistry 2000, 39, 6103-6114.
1
2776 J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002
9

14-Helical Antimicrobial â-Peptides
A R T I C L E S
would also increase the amount of 14-helical structure in
aqueous solution. The â-peptide 14-helix has approximately
three residues per turn, giving rise to a helical triad repeat in
which every third residue is on the same face of the helix.
â-Peptide 4 contains a cationic-hydrophobic-hydrophobic triad
repeat in place of the hydrophobic-cationic-hydrophobic triad
repeat found in 1-3 because we originally believed that
antimicrobial activity, this feature was retained in all subsequent
â-peptides. We compared 2 and 2-ent to determine whether the
absolute stereochemistry of â-peptides affects their biological
activity.
We used circular dichroism (CD) data obtained in aqueous
solution to compare the intrinsic 14-helix stabilities among our
â-peptides. For two sets of homologous â-peptides with
incrementally increasing backbone rigidification, 2-4 and 5-7,
14-helix stability was compared with both antimicrobial and
hemolytic activity. The importance of helical amphiphilicity on
these biological activities was determined by comparing two
highly rigidified â-peptide isomers, 7, which will form an
amphiphilic 14-helix, and 8, which will form a nonamphiphilic
39
coupling ACHC to the resin would be difficult. We have
recently evaluated the ability of â-peptides 5-8 to form small
1
4-helix (Figure 1). The R-helical R-peptides melittin and
8
,13,18
(Ala
)-magainin II amide (a synthetic derivative of natural
3
1
magainin II with enhanced antimicrobial activity ) were used
as controls in determinations of biological activity. Both
8
,13,18
(Ala
)-magainin II amide and melittin are active against a
broad spectrum of bacteria, but melittin is toxic toward both
6
,44,45
8,13,18
bacterial and mammalian cells,
while (Ala
)-magainin
4
II amide is selectively toxic toward bacterial cells.
Synthesis of â-Peptides. â-Peptides 1-8, 1-ent, and 2-ent
were synthesized from Fmoc-protected â-amino acid mono-
46,47
mers
using standard automated solid-phase Fmoc chemistry.
However, during the synthesis of â-peptide 3, the Fmoc
protecting group was not completely removed from the 6th and
7
th residues with the standard deprotection conditions, 20%
piperidine in DMF. Reversed-phase high-performance liquid
chromatography (RP-HPLC) and matrix-assisted laser desorp-
tion-ionization time-of-flight mass spectral (MALDI-TOF-MS)
analysis of the crude product showed that the Fmoc-protected
hexamer and heptamer were present along with the desired
4
8
nonamer. During the synthesis of an analogous â-peptide, we
encountered a similar problem: neither the use of the stronger
base DBU nor longer deprotection times with 20% piperidine
in DMF led to complete removal of the Fmoc protecting group
from the 5th residue of the â-peptide attached to the resin.
However, when the same peptide-resin was deprotected at 60
soluble aggregates in aqueous solution as an initial step toward
the creation of â-peptide tertiary structure.40 Since 5-7 were
designed to form amphiphilic 14-helices, we evaluated â-pep-
tides 5-7 and 8, a nonamphiphilic isomer of 7, for antimicrobial
activity as well. Most â-peptides we prepared were of opposite
absolute stereochemistry relative to 1-ent, since (R,R)-ACHC
°
C in 20% piperidine in NMP, the Fmoc protecting group was
completely removed. Therefore, deprotections were performed
at 60 °C after the 6th and 7th residues during the synthesis of
3
6
. The problematic deprotections of the 5th and 6th residues of
were also performed at 60 °C.
Circular Dichroism. Several recent studies show that CD
4
1
is more synthetically accessible than its enantiomer. All
â-peptides except 8 were designed to be amphiphilic when
adopting the 14-helical conformation, with hydrophobic residues
covering roughly two-thirds of the helix circumference and
49
data can provide insight on â-peptide folding, analogously to
the way CD illuminates secondary structure formation in
50,51
R-peptides.
The right-handed 14-helical conformation of
â-peptides is characterized by a maximum in molar ellipticity
3
mostly positively charged â -homolysine residues on the
(
42) Cuervo, J. H.; Rodriguez, B.; Houghten, R. A. Pept. Res. 1988, 1, 81-86.
43) Li, Z. Q.; Merrifield, R. B.; Boman, I. A.; Boman, H. G. FEBS Lett. 1988,
231, 299-302.
remaining third of the helix circumference (Figure 1).
We replaced the C-terminal carboxylic acid group of 1 with
a carboxamide in â-peptide 2 because analogous C-terminal
(
(
44) Blondelle, S. E.; Houghten, R. A. Biochemistry 1991, 30, 4671-4678.
45) Habermann, E. Science 1972, 177, 314-322.
(
capping enhances antimicrobial activity among conventional
(46) Guichard, G.; Abele, S.; Seebach, D. HelV. Chim. Acta 1998, 81, 187-
peptides.35,42,43
206.
Since the C-terminal amide gave much higher
(47) Abele, S.; Guichard, G.; Seebach, D. HelV. Chim. Acta 1998, 81, 2141-
2156.
(
38) Houston, M. E.; Gannon, C. L.; Kay, C. M.; Hodges, R. S. J. Pept. Sci.
(48) Seebach et al. reported that low purity of some crude â-peptides synthesized
1
995, 1, 274-282.
on solid support resulted from incomplete removal of the Fmoc protecting
47
(
39) Preliminary results in our laboratory had shown that coupling of ACHC to
the resin proceeded in low yield. However, we observed no incomplete
coupling of ACHC to the resin during the subsequent synthesis of 8.
40) Raguse, T. L.; Lai, J. R.; LePlae, P. R.; Gellman, S. H. Org. Lett. 2001, 3,
group. The use of DBU and longer deprotection times during the synthesis
of â-peptides composed entirely of acyclic residues produced favorable
results (Seebach, D.; Schreiber, J. V.; Arvidsson, P. I.; Frackenpohl, J. HelV.
Chim. Acta 2001, 84, 271-279).
(
(
3
963-3966.
41) Appella, D. H.; LePlae, P. R.; Raguse, T. L.; Gellman, S. H. J. Org. Chem.
000, 65, 4766-4769.
(49) Cheng, R. P.; Gellman, S. H.; DeGrado, W. F. Chem. ReV. 2001, 101,
3219-3232.
(50) Johnson, W. C. Proteins 1990, 7, 205-214.
2
J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002 12777

A R T I C L E S
Raguse et al.
Figure 3. Molar ellipticity of 200 µg/mL â-peptide 2 in solutions of varying
proportions of TFE and aqueous buffer at 25 °C.
range from completely unstructured to largely 14-helical in
aqueous solution.
We examined the CD spectra of â-peptides 1-8 in aqueous
TFE to gain further insight on the conformational propensities
of these molecules. These studies were undertaken because we
wanted to know whether the most flexible â-peptides (1, 2, and
5) would adopt 14-helical conformations, at least to some extent,
in the presence of the structure-promoting cosolvent TFE. As
noted above, R-helix induction by TFE is commonly observed
for antimicrobial R-peptides. The CD spectrum of flexible
â-peptide 2 was examined at various solvent compositions to
discover the TFE/aqueous buffer proportion that produces the
maximum amount of 14-helical structure. The data in Figure 3
show that â-peptide 2 displays maximum 14-helicity in g40
vol % TFE in aqueous buffer, as indicated by the molar
ellipticity at 214 nm. The plateau in 14-helix induction observed
for 1 above 40 vol % TFE matches a similar plateau in R-helix
induction observed for R-peptides in aqueous TFE.54 On the
basis of these results, we compared the CD spectra of â-peptides
Figure 2. Circular dichroism data for â-peptides 1-8 at 25 °C in 10 mM
aqueous TRIS, 150 mM NaCl, pH 7.2. Concentrations of â-peptides were
approximately 200 µg/mL (approximately 100 µM).
at approximately 215 nm, and 14-helical â-peptides often display
a minimum near 200 nm as well.52 All of the conformational
conclusions we deduce for our â-peptides are based on CD data
obtained in aqueous solution. In contrast, as noted above, CD-
based conformational analysis of R-helix-forming antimicrobial
R-peptides is typically carried out in water-TFE mixtures
because the peptides do not fold in aqueous solution. (It has
been suggested that water-TFE mixtures somehow mimic the
hydrophobic environment of cellular membranes,32 although the
physical basis for this hypothesis is unclear.)
1
-8 in 60 vol % TFE (Figure 4), a solvent mixture that we as-
sumed would provide maximal 14-helix induction in each case.
The CD spectra of â-peptides 1-4 and 6-8 in 60 vol %
TFE (Figure 4) display maxima in molar ellipticity at 214 nm
and exhibit the characteristic signature of the right-handed 14-
55
helix. In contrast, â-peptide 5 does not appear to be 14-helical
Figure 4B). The extent of 14-helix formation in 60 vol % TFE
(
is similar among 1-4 and 6-8, although there seems to be a
slight increase in 14-helicity as the proportion of preorganized
ACHC residues rises. The most conformationally stable â-pep-
tide, 8, has only a two-fold greater molar ellipticity at 214 nm
relative to flexible â-peptide 2, which appears to be least
structured in 60% TFE. For all three of the â-peptides containing
the maximum number of rigid cyclic residues, 4, 7, and 8, the
molar ellipticities at 214 nm in aqueous buffer (Figure 2) are
equal to or greater than those in 60 vol % TFE (Figure 4).
Therefore, we conclude that these three â-peptides are com-
pletely folded into the 14-helix, or nearly so, in aqueous buffer.
Antimicrobial Activity. Minimal inhibitory concentration
CD data for 1-8 in aqueous buffer are shown in Figure 2;
part A shows the effect of incremental backbone rigidification
among 2-4 while part B shows the analogous rigidification
series 5-7. The â-peptides with the largest proportion of cyclic
residues, roughly two-thirds, have the greatest amount of 14-
helical structure by CD (4, from the first rigidification series,
and isomer pair 7 and 8); as discussed below, we suspect that
these â-peptides approach 100% 14-helix population under these
conditions. The â-peptides lacking cyclic residues (1, 2, and 5)
do not exhibit a maximum at 214 nm, which suggests that these
5
3
molecules form little or no 14-helix in aqueous solution. The
â-peptides containing roughly one-third cyclic residues (3 and
(MIC) values for â-peptides 1-8 were determined against both
6
) seem to be partially folded into the 14-helix under these
conditions. Thus, within each rigidification series, the â-peptides
(
53) This lack of 14-helical structure in aqueous solution is not surprising. Cheng
3
et al. have concluded that â -peptides are less helix-prone than R-peptides
(
51) Yang, J. T.; Wu, C. S.; Martinez, H. M. Methods Enzymol. 1986, 130,
in aqueous solution (Cheng, R. P.; DeGrado, W. F. J. Am. Chem. Soc.
2001, 123, 5162-5163).
2
08-269.
(52) Seebach, D.; Ciceri, P. E.; Overhand, M.; Jaun, B.; Rigo, D.; Oberer, L.;
(54) Luo, P.; Baldwin, R. L. Biochemistry 1997, 36, 8413-8421.
(55) â-Peptide 2-ent, which is of the opposite absolute configuration, had a
minimum molar ellipticity at 214 nm (data not shown).
Hommel, U.; Amstutz, R.; Widmer, H. HelV. Chim. Acta 1996, 79, 2043-
2
066.
12778 J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002

14-Helical Antimicrobial â-Peptides
A R T I C L E S
Table 1. MIC Values for Each Peptide in µg/mLa
peptide
E. coli JM109b
6.3
50
200
12.5
B. subtilis BR151c
S. aureus 1206d
E. faecium A634e
(
Ala8,13,18)-magainin II amide
3.1
3.1
50
25-50
12.5
200
25
6.3
g200
12.5
melittin
1
2
1.6
6.3
2
3
4
5
6
7
8
-ent
12.5-25
1.6
0.8
0.8
1.6-3.1
0.8
0.8
100
6.3
3.1
6.3
6.3
12.5-25
12.5
3.1
6.3
12.5
3.1-6.3
6.3
3.1-6.3
6.3
>25f
>12.5
>200
6.3-12.5
12.5
>200
g
>200
a
Values are the median of at least two independent experiments with two replicates. MICs within a factor of two of each other are considered within
experimental uncertainty; therefore, the MICs of two peptides must differ by more than a factor of four for the antimicrobial activity to be considered
different beyond experimental uncertainty. Reference 59. Reference 56. d Reference 57. e Reference 58. Determination of the MIC was not possible
b
c
f
g
because the â-peptide was not soluble above 25 µg/mL. Determination of the MIC was not possible because the â-peptide was not soluble above 12.5
µg/mL.
5
6
57
Gram-positive (Bacillus subtilis, Staphylococcus aureus, and
58
59
Enterococcus faecium ) and Gram-negative (E. coli ) bacteria.
For two of the bacteria, S. aureus57 and E. faecium, the strains
used are clinical isolates that are resistant to penicillin and
vancomycin, respectively (among other antibiotics). The anti-
microbial R-helical R-peptides melittin
magainin II amide served as positive controls.
58
6,44,45
8,13,18
and (Ala
)-
4
Most of the â-peptides tested have antimicrobial activity
comparable to or more potent than that of the positive controls
8
,13,18
melittin and (Ala
)-magainin II amide (Table 1); the two
exceptions are 1 and 8. Potency against all four bacterial species
increases by one to two orders of magnitude when we replace
the C-terminal free carboxylic acid of 1 with a carboxamide in
2
. The antimicrobial activity of 2-ent is within experimental
uncertainty of the activity of its enantiomer, 2. â-Peptide 8,
which unlike 1-7 cannot form an amphiphilic 14-helix, is
inactive against three of the four strains up to the maximum
concentration tested. â-Peptide 8 displays weak activity against
5
6
B. subtilis, but amphiphilic isomer 7 is approximately two
orders of magnitude more potent against this species. The very
feeble activity of 8 relative to 7 indicates that the ability to adopt
an amphiphilic structure is very important for antimicrobial
action.
â-Peptides 2-4 have similar antimicrobial activities despite
their vastly different extents of 14-helical structure in aqueous
solution (Figure 2A). Likewise, â-peptides 5-7 range from no
1
4-helical structure to highly structured in aqueous solution
(Figure 2B), but the MIC values differ by less than a factor of
four among these three â-peptides. Thus, there appears to be
little or no relationship between the extent of 14-helix formation
in aqueous solution and antimicrobial activity for these two sets
of â-peptides.
Figure 4. Circular dichroism data for â-peptides 1-8 at 25 °C in 60%
TFE, 40% TRIS-buffered saline. (Note: The vertical scale in B is different
from the scale used in A and in Figure 2.) Concentrations of â-peptides
were approximately 200 µg/mL (approximately 100 µM).
Minimal bactericidal concentration (MBC) values were
determined for selected â-peptides and the control R-peptide
number of viable bacteria after treatment decreased gradually,
rather than abruptly, as the concentration of R- or â-peptide
increased. This behavior may be correlated with the fact that
these strains of S. aureus and E. faecium were obtained as
clinical isolates. There was very little difference among the
MBCs of â-peptides 2-4 for each bacterial species; therefore,
bactericidal potency appears to be independent of 14-helical
propensity, at least above the minimal propensity displayed by
2, the most flexible member of this series.
8,13,18
(
Ala
)-magainin II amide (Table 2). For each peptide tested,
59
56
the MBC against E. coli and B. subtilis was no higher than
four times the corresponding MIC. In contrast, most of the MBC
values against the resistant strains S. aureus57 and E. faecium
were greater than the corresponding MIC by at least a factor of
eight. Interestingly, with S. aureus57 and E. faecium, the
57
58
58
58
(
(
(
56) Young, F. E.; Smith, C.; Reilly, B. E. J. Bacteriol. 1969, 98, 1087-1097.
57) Weisblum, B.; Demohn, V. J. Bacteriol. 1969, 98, 447-452.
58) Nicas, T. I.; Wu, C. Y.; Hobbs, J. N.; Preston, D. A.; Allen, N. E.
Antimicrob. Agents Chemother. 1989, 33, 1121-1124.
Hemolytic Activity. The lysis of human red blood cells
(hRBC) by each â-peptide was used to estimate the ability of
(
59) Yanisch-Perron, C.; Vieira, J.; Messing, J. Gene 1985, 33, 103-119.
J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002 12779

A R T I C L E S
Raguse et al.
Table 2. MBC Values for Selected â-Peptides and Control Peptide (Ala8,13,18)-Magainin II Amide in µg/mL
peptide
E. coli JM109a
B. subtilis BR151b
S. aureus 1206c
E. faecium A634d
(
Ala8,13,18)-magainin II amide
6.3-12.5
12.5-25
12.5-25
6.3
6.3-12.5
6.3
1.6-3.1
1.6
>200
g25
50
>200
>100
>25
2
3
4
50
>100
a
Reference 59. b Reference 56. c Reference 57. d Reference 58.
Figure 6. Initial velocities of the cleavage of 4-methylumbelliferyl
â-galactoside (MUG) by â-galactosidase leaked from B. subtilis cells after
incubation with â-peptides or controls. Initial velocities are expected to be
proportional to peptide-induced leakage of functional â-galactosidase from
bacterial cells. Error bars represent the average of at least two measurements
(
the standard deviation.
Peptide-Induced â-Galactosidase Leakage from B. subtilis.
Peptide-induced permeabilization of B. subtilis BAU10260 was
examined by measuring E. coli â-galactosidase (â-gal) synthe-
sized by lacZ integrated into the chromosome at the amiE locus,
6
0
as described. The chromosomal insertion was previously
selected first on the basis of chloramphenicol resistance present
in the plasmid construct used for insertion of lacZ and then on
the basis of erythromycin resistance introduced on a second
plasmid carrying VanRS, positive regulators of the VanH
promoter driving lacZ. B. subtilis BAU102 cells produce a
high basal level of â-gal in the absence of induction of VanRS,
which makes the peptide-induced leakage assay possible. Each
initial velocity in Figure 6 is proportional to the concentration
of functional â-gal that escaped from the cells after incubation
of BAU102 with an R- or â-peptide at a concentration well
above that peptide’s MIC. The â-gal activities observed after
treatment with â-peptides 2-4 were indistinguishable from one
Figure 5. Hemolysis data for â-peptides 1-8 and R-peptide controls.
Curves are the average of at least two independent experiments with two
replicates.
these antimicrobial agents to permeabilize mammalian cell
membranes (Figure 5). It has been previously demonstrated that
6
0
3
1
8,13,18
melittin is highly hemolytic, whereas (Ala
)-magainin II
amide is much less hemolytic;6
tides served as controls. All â-peptides without N-terminal â -
homotyrosine (1-4) had hemolytic activity equal to or less than
)-magainin II amide, while â-peptides 5-7,
which contain an N-terminal â -homotyrosine and can form an
,44,45
therefore, these two R-pep-
3
8
,13,18
that of (Ala
3
8
,13,18
amphiphilic structure, were more hemolytic than (Ala
)-
3
another and from the control peptide melittin, within experi-
magainin II amide. The addition of a â -homotyrosine residue
to the N-terminus of 4, to generate 7, resulted in nearly an order
of magnitude increase in hemolytic activity; the concentrations
mental uncertainty. The other positive control, (Ala8
,13,18
)-
magainin II amide, led to â-gal release slightly greater than that
induced by â-peptides 2-4. There is substantial evidence that
3
of â-peptides 4 and 7 (without and with â -homotyrosine,
8
,13,18
(
Ala
)-magainin II amide and melittin derive their anti-
respectively) required to lyse 50% of the hRBC were 200-400
µg/mL and 25-50 µg/mL, respectively. In contrast to the results
for antimicrobial activity, hemolytic activity changed little upon
replacement of the free carboxyl C-terminus in 1 with a
C-terminal amide (â-peptide 2). As expected, enantiomeric
â-peptides 2 and 2-ent had very similar abilities to lyse hRBC.
Among â-peptides 2-4, the hemolytic activity appeared to
increase slightly with 14-helical propensity in aqueous solution,
but this same trend was not seen among 5-7. As expected,
nonamphiphilic 8 had negligible hemolytic activity, indicating
that the ability to form an amphiphilic structure is necessary
microbial activity from their ability to permeabilize bacterial
9
,10
membranes. The similar extents of â-gal leakage caused by
8
,13,18
melittin, (Ala
)-magainin II amide, and â-peptides 2-4
suggests that these â-peptides also act via permeabilization of
bacterial membranes.
Lipophilicity Determination via RP-HPLC. RP-HPLC has
6
1
been used to measure the lipophilicity of various molecules,
6
2
including R-helical R-peptides. We used this approach to
analyze our â-peptides. The percent acetonitrile required to elute
(
60) Ulijasz, A. T.; Grenader, A.; Weisblum, B. J. Bacteriol. 1996, 178, 6305-
(but not sufficient) for the lysis of hRBC.
6309.
12780 J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002

14-Helical Antimicrobial â-Peptides
A R T I C L E S
Table 3. Percent Solvent B (Acetonitrile) Required to Elute
â-Peptides from a C4 Analytical Column
R-peptides such as the magainins because both types of peptides
appear to have similar mechanisms of action. A large body of
evidence suggests that R-helical R-peptides derive their anti-
a
â-peptide
percent solvent B
1
2
37.0
35.5
35.5
37.5
37.8
39.4
41.4
39.1
25.7
microbial properties from their ability to permeabilize bacterial
64-66
membranes,
reported in a few systems.
although evidence to the contrary has been
2
3
4
5
6
7
8
-ent
3
6,67
Several alternative mechanisms
of membrane permeabilization have been proposed, including
6
8,69
70-73
the formation of barrel-stave
or toroidal
pores or
65
membrane disruption through a carpet mechanism. More
recently, other mechanisms of membrane disruption have been
proposed, such as the fusion of the cell wall and cell membrane
of Gram-negative bacteria. We find that â-peptides 2-4 have
nearly the same ability as (Ala
a
74
Experiments were performed as described in the Materials and Methods
Section. All values are the average of at least two independent experiments.
The standard deviation in all cases was less than 0.4% solvent B.
8,13,18
)-magainin II amide and
melittin to permeabilize the membrane of B. subtilis BAU102,
as measured by â-gal release (Figure 6). It is unlikely that these
â-peptides derive their potency from interactions with a single
chiral receptor, since enantiomeric â-peptides 2 and 2-ent have
nearly identical antimicrobial activities (Table 1).
each â-peptide from a C4 analytical column is listed in Table
3
. A significantly higher percent acetonitrile was required to
elute â-peptide 7, which can form an amphiphilic 14-helix, than
to elute isomeric 8, which cannot form an amphiphilic 14-helix.
This difference is expected, since 7 displays a much larger
continuous hydrophobic surface in its 14-helical conformation
than does 8, and the difference provides a benchmark for
â-peptides that differ significantly in lipophilicity. All other
variations in elution properties among our â-peptides are small
in comparison to 7 versus 8, indicating that, with the exception
of 8, all â-peptides we examined have very similar lipophilici-
ties. Differences are small among 2-4 and among 5-7, the
two series in which hydrophobic acyclic residues are incremen-
tally replaced by ACHC. As expected, enantiomeric â-peptides
Most antimicrobial R-helical R-peptides require the ability
1
0
to form an amphiphilic helix for activity, although Oren and
Shai have found that peptides containing one-third D residues
and end-to-end cyclization are active despite the low probability
3
7
of R-helix formation. A requirement for amphiphilic helix
formation among our â-peptides is indicated by the observation
that 8 is much less active than isomer 7. Our â-peptides probably
cannot span bacterial membranes (ca. 30 Å wide) because the
1
1
4-helices formed by nine- or ten-residue â-peptides such as
-8 should be only approximately 15 Å long. Therefore, it is
2
and 2-ent had identical retention times.
unlikely that these â-peptides permeabilize bacterial membranes
via the barrel-stave mechanism, unless the helices align end-
to-end during transmembrane pore formation. The barrel-stave
mechanism is also unlikely for 2-4 because these â-peptides
are more potent in lysing bacterial cells relative to human red
blood cells; cell selectivity is less common with the barrel-stave
mode of action than with the carpet or toroidal pore mecha-
Discussion
We have examined â-peptides related to 1-ent, previously
reported by Hamuro et al.,11 to evaluate systematically the
relationship between 14-helical stability and biological activity.
We were able to vary 14-helical propensity, as monitored by
CD in aqueous solution, while keeping constant other factors
believed to be important for determining biological activity, such
as size, net charge, net hydrophobicity, amphiphilicity, and the
1
0
nisms. Analysis of interactions between these â-peptides and
7
5
liposomes is underway.
1
0
Replacement of the C-terminal carboxylic acid in 1 with a
primary amide (2) results in a one to two orders of magnitude
increase in antimicrobial activity against all four strains of
bacteria. The higher potency of 2 most likely results from an
increase in the net positive charge of 2, relative to 1, which
enhances electrostatic attraction to the negatively charged
phospholipids in bacterial membranes. Likewise, the C-terminal
widths of the hydrophobic and hydrophilic helix faces. The
antimicrobial activity we observed for 1-ent itself (data not
shown) is similar to that reported.11 The hemolytic activity of
1
-ent (data not shown) is in agreement with trends in activity
3
3
reported for â-peptides containing â -homovaline-â -homo-
lysine-â -homoleucine triad repeats.
are also in line with those reported by Arvidsson et al. for a
similar nonamer containing a â -homoalanine-â -homolysine-
3
11,13,63
Our results for 1-ent
1
5
4
2
43
3
3
amide forms of the magainins, cecropin A, and a melittin
3
5
3
diastereomer were also found to be more active than the
â -homophenylalanine triad repeat.
We believe that the structure-activity relationships deter-
mined for 14-helical â-peptides are relevant to R-helical
(
(
(
64) Bechinger, B. Biochim. Biophys. Acta 1999, 1462, 157-183.
65) Oren, Z.; Shai, Y. Biopolymers 1998, 47, 451-463.
66) Huang, H. W. Biochemistry 2000, 39, 8347-8352.
(
(
(
61) Du, C. M.; Valko, K.; Bevan, C.; Reynolds, D.; Abraham, M. H. Anal.
(67) Zhang, L.; Rozek, A.; Hancock, R. E. W. J. Biol. Chem. 2001, 276, 35714-
Chem. 1998, 70, 4228-4234.
35722.
62) Wagschal, K.; Tripet, B.; Lavigne, P.; Mant, C.; Hodges, R. S. Protein
Sci. 1999, 8, 2312-2329.
(68) Christensen, B.; Fink, J.; Merrifield, R. B.; Mauzerall, D. Proc. Natl. Acad.
Sci. U.S.A. 1988, 85, 5072-5076.
3
3
63) Reference 13 states that the â-peptide H-(â -homovaline-â -homolysine-
(69) Merrifield, R. B.; Merrifield, E. L.; Juvvadi, P.; Andreu, D.; Boman, H.
G. Ciba Found Symp. 1994, 186, 5-20.
3
4
â -homoleucine) -OH reported originally in ref 11 had been contaminated
with the Fmoc-protected â-peptide dodecamer. Although the presence of
(70) Ludtke, S.; He, K.; Huang, H. Biochemistry 1995, 34, 16764-16769.
(71) Ludtke, S. J.; He, K.; Heller, W. T.; Harroun, T. A.; Yang, L.; Huang, H.
W. Biochemistry 1996, 35, 13723-13728.
the impurity did not significantly affect antimicrobial activity, elimination
1
1,13
3
of the impurity resulted in a 10-fold decrease in hemolytic activity.
We infer that the analogous â-peptide reported in ref 11, H-(â -
(72) Matsuzaki, K.; Murase, O.; Fujii, N.; Miyajima, K. Biochemistry 1996,
35, 11361-11368.
3
3
homovaline-â -homolysine-â -homoleucine)
3
-OH (1-ent, here), was also
3
contaminated with the Fmoc-protected impurity. In the series Fmoc-(â -
(73) Matsuzaki, K.; Murase, O.; Fujii, N.; Miyajima, K. Biochemistry 1995,
34, 6521-6526.
3
3
homovaline-â -homolysine-â -homoleucine)
n
-OH, the elimination of one
11
triad of residues resulted in a 20-fold increase in the HC50
.
By analogy,
(74) Liechty, A.; Chen, J.; Jain, M. K. Biochim. Biophys. Acta 2000, 1463, 55-
3
we expect the HC50 of a pure sample of the â-peptide H-(â -homovaline-
64.
3
3
â -homolysine-â -homoleucine)
n
-OH (where n ) 3) will be approximately
(75) Epand, R. F.; Raguse, T. L.; Gellman, S. H.; Epand, R. M. Manuscript in
preparation.
1
3
8
00 µM, 20-fold higher than that of n ) 4 (37 µM ).
J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002 12781

A R T I C L E S
Raguse et al.
corresponding carboxylic acids. C-terminal amidation of R-
helical structure in aqueous solution would affect lysis of hRBC
by helical peptides. We find that the amount of 14-helical
structure in aqueous solution is not related to the ability of
â-peptides to lyse hRBC. â-Peptides 5-7 have very similar
hemolytic activities (Figure 5) despite their very different 14-
helix populations in aqueous solution (Figure 2B). Hemolytic
activity appears to increase marginally with the amount of 14-
helical structure in aqueous solution among â-peptides 2-4,
but higher hemolytic activity in this series may also result from
very subtle increases in lipophilicity (Table 3). The addition of
peptides stabilizes the R-helical conformation through favorable
interactions with the helix dipole7
6,77
and the potential for an
additional hydrogen bond. In contrast, C-terminal amidation of
the â-peptides described here should destabilize the 14-helical
structure through unfavorable interactions with the helix dipole
(the helix dipole of 14-helical â-peptides runs from C-terminus
to N-terminus, in contrast to the N-to-C dipole of R-helical
R-peptides).
Several groups have evaluated the antimicrobial activities of
R-peptides that display variable extents of R-helix formation in
mixed aqueous/organic solvents, which are proposed to represent
3
a hydrophobic N-terminal â -homotyrosine increases hemolytic
activity by approximately an order of magnitude (4 versus 7).
32
11,13
membrane-like environments; these R-peptides are unstruc-
DeGrado and co-workers
have obtained similar results; they
tured in aqueous solution.3
1,33,34,36
In most cases, increased
reported a 100-fold increase in hemolytic activity upon addition
of a hydrophobic Fmoc protecting group to the amino-terminus
R-helical population in mixed aqueous/organic solvents cor-
relates with higher antimicrobial activity.3
1,33,36
However, Shai
3
3
3
of H-(â -homovaline-â -homolysine-â -homoleucine)4-OH.
Neither antimicrobial activity nor hemolytic activity increases
significantly among our â-peptides when 14-helical propensity
is substantially increased. In contrast, stronger affinity of peptide
and peptidomimetic ligands to specific proteins is often observed
when the ligand is rigidified to favor the binding conformation.78
Cell membranes appear to be the biologically relevant binding
partners of the antimicrobial â-peptides. The lack of correlation
between the extent of 14-helix formation in aqueous solution
and either antimicrobial or hemolytic activity suggests that the
loss of conformational entropy upon binding of unfolded
â-peptides to cell membranes is relatively modest.
and Oren noted that the introduction of three D-amino acid
residues into the toxin paradaxin had little effect on antimicrobial
activity, even though the extent of R-helix population in aqueous
TFE was greatly reduced by the D residues.34 In contrast,
Houston and co-workers determined that increasing R-helical
stability in aqueous solution through the formation of lactam
bridges decreases antimicrobial activity; in most cases, the
creation of the lactam bridge appeared, from CD analysis, to
3
2
have no effect on R-helical stability in aqueous TFE.
We have shown that increasing the 14-helical propensity of
â-peptides has little effect on antimicrobial activity. The amount
of 14-helical structure in aqueous solution (Figure 2) varies from
We have used the unique structural properties of â-peptides
to determine whether the stability of the biologically active
conformation influences antibacterial or hemolytic activity. We
conclude that variation over a broad range of helical propensity
has little effect on either activity. In contrast, some other workers
have concluded that helix propensity among R-peptides has a
∼
0% to ∼100% within the two series of â-peptides, 2-4 and
5
-7, but there is no consistent relationship between 14-helical
stability and MIC values (Table 1) outside the limits of
experimental uncertainty. In contrast, the identity of the C-
terminal group (1 versus 2) and the ability of the â-peptide to
form an amphiphilic helix as opposed to a nonamphiphilic helix
3
1-33,36
substantial effect on antimicrobial activity
and hemolytic
3
3-36
(7 versus 8) are crucial for activity.
activity. In most of those studies, structural analysis was
conducted in mixed aqueous/organic solvents,
3
1-36
At first glance, our results appear to contradict the conclusions
while we
of Houston et al., who proposed that stabilizing the R-helical
conformation among R-peptides leads to a decrease in anti-
were able to analyze â-peptide folding in water. We believe
that our results are relevant to the behavior of R-peptides because
of mechanistic similarities between these two peptide classes
(both permeabilize bacterial membranes). This study suggests
that â-peptides capable of mimicking other R-peptide functions
will prove useful for elucidating the dependence of these
functions on conformational stability.
3
2
bacterial activity. However, the range of helical stability in
aqueous solution among our â-peptides was much greater than
among the R-peptides examined by Houston et al. â-Peptide
structuring in aqueous solution varied from no detectable 14-
helical structure to almost entirely 14-helical. In contrast,
Houston et al. reported much smaller differences in R-helical
structure in aqueous solution: increases of 16 to 66% or 8 to
Materials and Methods
1
3
7% upon formation of a lactam bridge. The only bacteria
General Procedures. Proton nuclear magnetic resonance ( H NMR)
spectra were recorded in deuterated solvents on a Bruker AC-300 (300
MHz) spectrometer. Chemical shifts are reported in parts per million
common to both studies was E. coli. For Houston’s R-peptide
analogues with the largest increase in helical structure upon
formation of the lactam bridge (16 to 66%), the MIC against
E. coli decreased only by a factor of two for the bridged peptide,
which is within experimental uncertainty in our hands.
(ppm, δ) relative to tetramethylsilane (δ 0.00). Matrix-assisted laser
desorption-ionization time-of-flight mass spectra (MALDI-TOF-MS)
were obtained on a Bruker REFLEX II spectrometer with a 337-nm
laser using the R-cyano-4-hydroxycinnamic acid matrix. The instrument
Increased R-helical propensity in mixed aqueous/organic
solvents seems to correlate with increased lysis of hRBC by
antimicrobial R-peptides, especially when the barrel-stave
5
+
was calibrated to a standard mixture of leu -enkephalin (M + H
)
+
+
556.28), angiotensin I (M + H ) 1296.7), and neurotensin (M + H
) 1672.9). Optical rotation was measured using sodium light (D line,
589.3 nm).
3
3-35
mechanism is the mode of membrane permeabilization.
However, it was previously unknown how preformation of
2 2 2
Materials. CH Cl was distilled from CaH . Hexane was distilled.
DMF for manual solid-phase peptide synthesis was purchased from
Aldrich (HPLC grade) and stored over Dowex 50W-X8 ion-exchange
resin. Ether was anhydrous. The highest available grade of all other
(
76) Shoemaker, K. R.; Kim, P. S.; Brems, D. N.; Marqusee, S.; York, E. J.;
Chaiken, I. M.; Stewart, J. M.; Baldwin, R. L. Proc. Natl. Acad. Sci. U.S.A.
1
985, 82, 2349-2353.
(77) Shoemaker, K. R.; Kim, P. S.; York, E. J.; Stewart, J. M.; Baldwin, R. L.
Nature 1987, 326, 563-567.
(78) Babine, R. E.; Bender, S. L. Chem. ReV. 1997, 97, 1359-1472.
12782 J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002

14-Helical Antimicrobial â-Peptides
A R T I C L E S
solvents was purchased and used without further purification. Diiso-
propylethylamine (DIEA) was distilled from CaH . H was purchased
1.01 (m, 6H); FAB-MS m/z calcd for C22
388.1525, obsd 388.1517.
H23NO
4
Na (M + Na+)
2
O
2 2
from Mallinckrodt, and 4 N HCl in dioxane was obtained from Pierce.
Wang resin (p-benzyloxybenzyl alcohol resin; loading ) 0.75 mmol/
g) and 9-fluorenylmethyloxycarbonyl-N-hydroxysuccinimide (Fmoc-
OSu) were purchased from Novabiochem. Fmoc amide resin [4-(2′,4′-
dimethoxyphenyl-Fmoc-aminomethyl)phenoxyacetamido-ethyl resin;
polystyrene resin functionalized with a Knorr linker; loading ) 0.63
Synthesis of â-Peptides. General Procedures for Solid-Phase
Synthesis. â-Peptides were synthesized on Fmoc amide resin (25-µmol
scale) on an Applied Biosystems Model 432A (Synergy) automated
peptide synthesizer using standard Fmoc/t-Bu strategy with O-benzo-
triazol-1-yl-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU)
and 1-hydroxybenzotriazole (HOBt) as coupling reagents unless stated
otherwise. Two-hour coupling times were employed. The program
module controlling Fmoc deprotection was modified to extend the
deprotection time automatically, if necessary. Cleavage from the resin
and simultaneous deprotection of the side chain protecting groups were
accomplished with 2,2,2-trifluoroacetic acid (TFA) via one of the
procedures described below. After purification via RP-HPLC, each
â-peptide was lyophilized. Freshly lyophilized â-peptides were dis-
solved in Millipore water to create 2 mg/mL â-peptide solutions that
were stored at -78 °C.
Combined Automated/Manual Synthesis of Difficult Sequences.
It was necessary to perform certain deprotection steps manually during
the synthesis of â-peptides 3 and 6 to attain the high temperatures
necessary for complete removal of the Fmoc-protecting group. For
convenience, each coupling step between two high-temperature depro-
tections was performed manually. The peptide-resin (25 µmol) was
removed from the automated peptide synthesizer and washed with DMF
8
,13,18
mmol/g] was obtained from Applied Biosystems. (Ala
)-Magainin
II amide, melittin, 4-methylumbelliferyl â-D-galactoside (MUG), and
TRIS were purchased from Sigma. All other reagents were purchased
from Aldrich.
Analytical thin-layer chromatography (TLC) was carried out on
Whatman TLC plates precoated with silica gel 60 (250-µm layer
thickness). Solvent mixtures used for TLC are reported in v/v ratios.
Sterile Falcon 3075 microtiter 96-well plates were used for biological
assays. â-Peptides were purified by RP-HPLC on a Vydac C4 semiprep
column using a flow rate of 3 mL/min. Solvent A and solvent B for
RP-HPLC were 0.045% TFA in Millipore water and 0.036% TFA in
acetonitrile, respectively. â-Peptide purity was assessed using a linear
gradient of 5-95% solvent B over 57 min on a Vydac C4 analytical
column, monitoring at 220 nm. The purity of each â-peptide after RP-
HPLC was greater than 95%.
Synthesis of â-Amino Acid Monomers. Fmoc-protected acyclic
(
1
3 × 1 min), Et
2
O (3 × 1 min), CH
2
Cl
2
(3 × 1 min), and NMP (3 ×
4
6,47
â-amino acid monomers were synthesized as described previously.
min). The resin was transferred to a glass centrifuge tube and
Boc-trans-2-aminocyclohexanecarboxylic acid (Boc-ACHC-OMe)41
deprotected with 20% piperidine in NMP at 60 °C, with mixing via
nitrogen bubbling, for 1 h. After cooling to room temperature, the
mixture was transferred to a fritted reaction vessel. The resin was
was converted to Fmoc-ACHC-OH as described below.
Boc-ACHC-OH. Methanol (105 mL) and Millipore water (35 mL)
4
1
were added to Boc-ACHC-OMe (1.83 g, 7.12 mmol), followed by
LiOH‚H O (3.58 g, 85.4 mmol) and 26% H in water (4.66 mL,
5.6 mmol). The mixture was stirred at room temperature for 36 h. A
solution of Na SO (13.4 g, 106 mmol) in Millipore water (80 mL)
isolated by filtration and washed with NMP (3 × 1 min), Et
2
O (3 × 1
2
2 2
O
min), CH Cl
2
2
(3 × 1 min) and DMF (3 × 1 min). The next â-amino
3
acid was coupled by adding HOBt (150 µL of a 0.5 M solution in
DMF, 75 µmol), DIEA (300 µL of a 0.5 M solution in DMF, 150
2
3
was then added at 0 °C, and the mixture was stirred for 10 min. The
methanol was removed by rotary evaporation. The aqueous solution
was cooled to 0 °C and acidified with 3 M HCl until white solid
precipitated. This mixture was treated with ethyl acetate, which caused
the precipitate to dissolve, and the layers were separated. The aqueous
layer was further acidified to pH 2 with 3 M HCl and re-extracted
with the same ethyl acetate and then with four additional aliquots ethyl
2 2
µmol), CH Cl (450 µL), the appropriate Fmoc-â-amino acid (75 µmol
freshly dissolved in 600 µL DMF), and benzotriazol-1-yloxytripyrroli-
dinophosphonium hexafluorophosphate (PyBOP) (39.0 mg, 75 µmol
freshly dissolved in 600 µL DMF) to the peptide-resin. The mixture
was agitated for 8 h. The resin was isolated by filtration and washed
with DMF (3 × 1 min), Et
Remaining free amines were capped by adding a solution of acetic
anhydride (84 µL, 890 µmol) and DIEA (84 µL, 480 µmol) in CH Cl
2 mL) to the resin and agitating for 1 h. The resin was isolated by
filtration and washed with CH Cl O (3 × 1 min),
(3 × 1 min), Et
CH Cl
2
O (3 × 1 min), and CH
2
Cl
2
(3 × 1 min).
4
acetate. The organic extracts were combined, dried over MgSO ,
2
2
concentrated, and dried in vacuo to yield 1.68 g Boc-ACHC-OH (97%
(
1
yield) as a white solid. The H NMR spectrum was identical to the
2
2
2
2
3
published spectrum of Boc-ACHC-OH.
2
2
(3 × 1 min), and NMP (3 × 1 min). The resin was transferred
Fmoc-ACHC-OH. A solution of 4 N HCl in dioxane (40 mL) was
added to Boc-ACHC-OH (1.68 g, 6.91 mmol), and after a few min a
white solid precipitated from the clear solution. The mixture was stirred
at room temperature for 1 h, and then the solvent was removed under
to a glass centrifuge tube and deprotected with 20% piperidine in NMP
at 60 °C, with mixing via nitrogen bubbling, for 1 h. After cooling to
room temperature, the resin was washed with NMP (3 × 1 min), Et
(3 × 1 min), and CH Cl (3 × 1 min). Synthesis was continued on the
automated peptide synthesizer.
Cleavage from Resin and Side Chain Deprotection. Procedure
A. The peptide was cleaved from the resin with 95:5 TFA:H O (8 mL)
2
O
2
2
2 3
a stream of N . The white solid was dried under vacuum. CH CN (66
mL) and Millipore water (17 mL) were added, followed by DIEA (3.57
mL, 20.1 mmol) and Fmoc-OSu (2.26 g, 6.70 mmol). The mixture was
stirred at room temperature for 30 min or until TLC showed that the
Fmoc-OSu was almost completely consumed (Rf ) 0.45 in 1:1 hexane/
ethyl acetate). The pH was neutralized with aqueous 1 M HCl at 0 °C
2
for 3 h. The resin was removed via filtration through cotton and rinsed
with additional TFA. The combined filtrate was concentrated under a
stream of nitrogen.
and the CH
3
CN was removed by rotary evaporation. More 1 M HCl
Procedure B. The peptide was cleaved from the resin with 95:2.5:
was added at 0 °C until the pH was equal to 2 and white solid
precipitated. The white solid was collected via suction filtration and
washed with dilute HCl solution. The product was dissolved in ethyl
acetate and washed with 1 M HCl and with brine. The organic layer
2.5 TFA:ethanedithiol:H O (8 mL) for 3 h. The resin was removed via
2
filtration through glass wool and rinsed with additional TFA. The
combined filtrate was concentrated under a stream of nitrogen. A
minimal amount of methanol was added to dissolve the crude â-peptide,
was dried over MgSO
4
and concentrated in vacuo. The product was
and Et O (10 mL) was added to form a white precipitate. The mixture
2
recrystallized from hexane/ethyl acetate to afford 2.17 g Fmoc-ACHC-
was cooled in an acetone/dry ice bath for 5 min, centrifuged, and the
23
OH (86% yield) as a fluffy white solid: mp 200-201 °C; [R]
D
)
;
Et
mL) was added, the mixture was stirred with a spatula, cooled in an
acetone/dry ice bath for 5 min, and centrifuged, and the Et O was
decanted from the white pellet. Any Et O remaining with the white
solid was evaporated under a gentle stream of nitrogen.
2 2
O was decanted from the white pellet. A new portion of Et O (10
-
37.9 (c 0.5, acetone); IR (thin film) 3310 (NH), 1694 (CdO), cm-1
1
H NMR (CDCl
3
+ CD
3
OD, 300 MHz): δ 7.77 (d, J ) 7.3 Hz, 2H),
2
7
3
.62 (d, J ) 7.0 Hz, 2H), 7.44-7.28 (m, 4H), 4.50-4.18 (m, 3H),
.77-3.48 (m, 1H), 2.36-2.13 (m, 1H), 2.07-1.92 (m, 2H), 1.82-
2
J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002 12783

A R T I C L E S
Raguse et al.
â-Peptide 1. Fmoc-â -homoleucine-OH44 was manually loaded onto
3
synthesizer, except that the program module controlling the DMF rinse
cycles after the 6th, 7th, and 8th couplings was modified to extend the
rinsing times automatically if needed to allow the conductivity level
to return to baseline. Cleavage of the â-peptide and simultaneous side
chain deprotection were performed as described in Procedure A. The
crude â-peptide was dissolved in the HPLC A solvent and purified by
RP-HPLC employing a linear gradient from 27% to 49% solvent B
the resin via the procedure given below. Wang resin (33.3 mg, 25.0
µmol) was washed with DMF (3 × 1 min), Et
2
O (3 × 1 min), and
3
2
CH Cl
2
(3 × 1 min). Fmoc-â -homoleucine-OH (55.1 mg, 150 µmol)
was weighed into a vial and dissolved in CH
carbodiimide (11.7 µL, 75 µmol) and 4-(dimethylamino)pyridine (9.2
mg, 75 µmol) were added to the CH Cl solution. The CH Cl solution
was then transferred to the vessel containing the washed resin with
CH Cl (1.2 mL total CH Cl ) and the mixture was agitated at room
2 2
Cl . Diisopropyl
2
2
2
2
111 13 9
over 45 min. MALDI-TOF-MS m/e calcd for (C63H N O ) (M)
+
+
2
2
2
2
1193.9, found 1194.9 (M + H ), 1216.8 (M + Na ), 1232.8 (M +
K ).
+
temperature for 48 h. The resin was isolated by filtration and washed
with DMF (4 × 1 min). Trimethylacetic anhydride (51 µL, 250 µmol),
pyridine (20 µL, 250 µmol), and DMF (2 mL) were combined in a
separate vial and added to the resin to cap any unreacted free hydroxyl
groups. The mixture was agitated at room temperature for 1 h. The
resin was isolated by filtration and washed with DMF (4 × 1 min),
â-Peptide 5. The synthesis was performed as described under
General Procedures for Solid-Phase Synthesis on the automated peptide
synthesizer. Cleavage of the â-peptide and simultaneous side chain
deprotection were performed as described in Procedure B. The crude
â-peptide was dissolved in 10:1 HPLC A solvent:methanol with
sonication and purified by RP-HPLC employing a linear gradient from
30% to 42.5% solvent B over 25 min. MALDI-TOF-MS m/e calcd for
methanol (4 × 1 min), and CH
2
Cl
2
(4 × 1 min). The resin was dried
in vacuo. The synthesis was then continued as described under General
Procedures for Solid-Phase Synthesis on the automated peptide
synthesizer. The â-peptide was cleaved from the resin with 50% TFA
+
(C H N O ) (M) 1383.0, found 1384.2 (M + H ), 1406.3 (M +
73
134 14 11
+
+
Na ), 1422.2 (M + K ).
in CH
2
Cl
2
(12 mL) for 1.5 h. The resin was removed via filtration
â-Peptide 6. The coupling of residues 1-5 was performed as
described under General Procedures for Solid-Phase Synthesis on the
automated peptide synthesizer, except that the program module
controlling the DMF rinse cycles after each coupling was modified to
extend the rinsing time automatically if needed to allow the conductivity
level to return to baseline. Fmoc-deprotection of the 5th residue was
repeated manually at 60 °C, and coupling and high-temperature
deprotection of the 6th residue were also performed manually as
described under Combined Automated/Manual Synthesis of Difficult
Sequences. Synthesis of residues 7-10 was performed as described
under General Procedures for Solid-Phase Synthesis on the automated
peptide synthesizer, except that the final ACHC residue was double-
coupled, and the program module controlling the DMF rinse cycles
after each coupling was modified to extend the rinsing time automati-
cally if needed to allow the conductivity level to return to baseline.
Cleavage of the â-peptide and simultaneous side chain deprotection
were performed as described in Procedure B. The crude â-peptide was
dissolved in the HPLC A solvent and purified by RP-HPLC employing
a linear gradient from 30% to 50% solvent B over 40 min. MALDI-
through cotton and rinsed with additional 50% TFA in CH
2
Cl . The
2
combined filtrate was concentrated under a stream of nitrogen. The
crude â-peptide was dissolved in the HPLC A solvent and purified by
RP-HPLC employing a linear gradient from 25% to 45% solvent B
over 40 min. MALDI-TOF-MS m/e calcd for (C60
H
116
N
12
O10) (M)
+
+
1
164.9, found 1165.7 (M + H ), 1187.7 (M + Na ), 1203.7 (M +
+
+
+
K ), 1209.7 (M + 2Na - H ).
â-Peptide 1-ent. Synthesis and purification were performed as
described for 1. MALDI-TOF-MS m/e calcd for (C60
H
116
N
12
O10) (M)
+
+
1
164.9, found 1165.8 (M + H ), 1187.7 (M + Na ), 1203.7 (M +
+
+
+
K ), 1209.7 (M + 2Na - H ).
â-Peptide 2. The synthesis was performed as described under
General Procedures for Solid-Phase Synthesis on the automated peptide
synthesizer. Cleavage of the â-peptide and simultaneous side chain
deprotection were performed as described under Procedure A. The crude
â-peptide was dissolved in the HPLC A solvent and purified by RP-
HPLC employing a linear gradient from 26% to 41% solvent B over
3
0 min. MALDI-TOF-MS m/e calcd for (C60
H
117
N
13
O
9
) (M) 1163.9,
+
+
+
found 1164.7 (M + H ), 1186.8 (M + Na ), 1202.7 (M + K ).
â-Peptide 2-ent. Synthesis and purification were performed as
TOF-MS m/e calcd for (C73
H
128
N
14
O
11) (M) 1377.0, found 1377.9 (M
+
+
+
+ H ), 1399.9 (M + Na ), 1415.9 (M + K ).
described for 2. MALDI-TOF-MS m/e calcd for (C60
H
117
N
13
O
9
) (M)
â-Peptide 7. The synthesis was performed as described under
General Procedures for Solid-Phase Synthesis on the automated peptide
synthesizer, except that the program module controlling the DMF rinse
cycles after each coupling was modified to extend the rinsing time
automatically if needed to allow the conductivity level to return to
baseline. The final three ACHC residues were double-coupled. Cleavage
of the â-peptide and simultaneous side chain deprotection were
performed as described in Procedure B. The crude â-peptide was
dissolved in the HPLC A solvent and purified by RP-HPLC employing
a linear gradient from 30% to 42.5% solvent B over 25 min. MALDI-
+
+
1
K ).
163.9, found 1164.9 (M + H ), 1186.9 (M + Na ), 1202.8 (M +
+
â-Peptide 3. The coupling of residues 1-6 was performed as
described under General Procedures for Solid-Phase Synthesis on the
automated peptide synthesizer. The program module controlling the
DMF rinse cycle after the 6th coupling was modified to extend the
rinsing time automatically if needed to allow the conductivity level to
return to baseline. Fmoc-deprotection of the 6th residue was repeated
manually at 60 °C, and coupling and high-temperature deprotection of
the 7th residue were also performed manually as described under
Combined Automated/Manual Synthesis of Difficult Sequences. Syn-
thesis of residues 8 and 9 were performed as described under General
Procedures for Solid-Phase Synthesis on the automated peptide
synthesizer, except that the final ACHC residue was double-coupled,
and the program module controlling the DMF rinse cycle after each
coupling was modified to extend the rinsing time automatically if
needed to allow the conductivity level to return to baseline. Cleavage
of the â-peptide and simultaneous side chain deprotection were
performed as described in Procedure A. The crude â-peptide was
dissolved in the HPLC A solvent and purified by RP-HPLC employing
a linear gradient from 28% to 43% solvent B over 30 min. MALDI-
TOF-MS m/e calcd for (C73
+ H ), 1393.7 (M + Na ), 1409.7 (M + K ).
H
122
N
14
O
11) (M) 1370.9, found 1371.7 (M
+
+
+
â-Peptide 8. The synthesis was performed as described under
General Procedures for Solid-Phase Synthesis on the automated peptide
synthesizer, except that each ACHC residue was double-coupled, and
the program module controlling the DMF rinse cycle after each coupling
was modified to extend the rinsing time automatically if needed to allow
the conductivity level to return to baseline. Cleavage of the â-peptide
and simultaneous side chain deprotection were performed as described
in Procedure B. The crude â-peptide was dissolved in the HPLC A
solvent and purified by RP-HPLC employing a linear gradient from
17% to 30% solvent B over 25 min. MALDI-TOF-MS m/e calcd for
+
TOF-MS m/e calcd for (C63
H
117
N
13
O
9
) (M) 1199.9, found 1200.8 (M
(C73
H
122
N
14
O
11) (M) 1370.9, found 1371.7 (M + H ), 1393.7 (M +
+
+
+
+
+
+
H ), 1222.8 (M + Na ), 1238.8 (M + K ).
Na ), 1409.7 (M + K ).
â-Peptide 4. The synthesis was performed as described under
General Procedures for Solid-Phase Synthesis on the automated peptide
Circular Dichroism. A 2 mg/mL stock solution of each â-peptide
in Millipore water (concentration determined by mass of freshly
1
2784 J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002
9

14-Helical Antimicrobial â-Peptides
A R T I C L E S
lyophilized â-peptide as its TFA salt, assuming one TFA counterion
per amine) was combined with aqueous TRIS-buffered saline to give
aqueous solutions of 0.2 mg/mL â-peptide in 10 mM TRIS, 150 mM
NaCl, pH 7.2. Alternatively, each 2 mg/mL stock solution of â-peptide
in Millipore water was combined with aqueous TRIS-buffered saline
and TFE to give solutions of 0.2 mg/mL â-peptide, 4.4 mM TRIS, and
Hemolysis Assay. Human erythrocytes (hRBC) were collected and
stored refrigerated in a Becton Dickinson vacu-taner containing the
anticoagulant EDTA for e1 day. The hRBC were collected by
centrifugation and washed with TRIS-buffered saline (10 mM TRIS,
150 mM NaCl, pH 7.2) three times or until the supernatant was clear.
The cells were then diluted with TRIS-buffered saline to a final
concentration of 0.25% (v/v) and stored on ice for e1 h. Two-fold
serial dilutions with TRIS-buffered saline were performed in duplicate
for each peptide in a 96-well plate to a final volume of 20 µL in each
well. An aliquot (80 µL) of the 0.25% hRBC suspension was added to
each well. Wells containing 20 µL TRIS-buffered saline with 80 µL
hRBC suspension served as a negative control. The plate was incubated
at 37 °C for 1 h and then centrifuged for 5 min at 3500 rpm. An aliquot
(50 µL) of the supernatant from each well was transferred to a new
well of a 96-well plate containing 50 µL water, and the OD at 415 nm
was measured. Percent hemolysis was calculated as [(OD415 peptide -
OD415 buffer)/ (OD415 complete hemolysis - OD415 buffer)] × 100,
where complete hemolysis was defined as the average hemolysis of
all wells containing a final concentration of melittin ranging from 50
to 400 µg/mL.
6
7 mM NaCl in 60% aqueous TFE. The final concentration of
â-peptides 5-8 in each aqueous TRIS-buffered saline solution was
determined by the UV absorbance of that solution (UV absorbance was
measured immediately after obtaining the CD spectrum), whereas the
final concentration of â-peptide in each 60% TFE solution was
calculated from the UV absorbance of the 2 mg/mL stock solution.
-
1
We assume the extinction coefficient of each â-peptide is 1420 cm
M
-1
79
at 275 nm, the extinction coefficient of R-tyrosine. The
concentration determined by mass was within 80% of the concentration
determined by UV absorbance for each â-peptide solution. Circular
dichroism spectra were obtained on an Aviv 202SF spectrometer at
room temperature with 1-mm path length cells and 10-s averaging times.
The CD signal of the corresponding buffer solution was subtracted from
the CD spectrum of each â-peptide solution. Data were converted to
2
-1
ellipticity (deg cm dmol ) according to the equation:
Peptide-Induced â-Galactosidase Leakage from B. subtilis. B.
subtilis BAU10260 was grown in trypticase soy broth with 10 µg/mL
erythromycin and 34 µg/mL chloramphenicol to ensure retention of
VanRS and of lacZ, respectively. The cells were grown at 37 °C to an
absorbance of 0.6 at 660 nm. Growth of the bacteria to the same
absorbance each time was required to ensure reproducibility. The cells
were collected by centrifugation and resuspended in fresh medium to
remove residual â-galactosidase (â-gal). Aliquots (10 µL) of peptide
stock solutions (200 µg/mL) were added to wells in a sterile 96-well
plate. Bacterial suspension (90 µL) was then added to the wells to give
a total volume of 100 µL (final peptide concentration: 20 µg/mL).
The final peptide concentration used in this assay was above the MIC
in all cases. The plates were incubated at room temperature for 1 h to
allow for the release of â-gal. After the incubation period, the plate
was centrifuged for 5 min at 5000 rpm to remove all cells and cellular
debris. An aliquot (80 µL) of the supernatant was removed and placed
in a separate well. The enzymatically cleavable molecule 4-methyl-
umbelliferyl â-galactoside (MUG) was used as a fluorescent indicator
of â-gal. An aliquot of MUG in DMSO (20 µL, 0.4 mg/mL) was added
to the well, and fluorescence was monitored over time. Initial velocities
of the enzymatic reaction were obtained from the linear plot of
fluorescence versus time. Water without peptide was used as a negative
control.
[
Θ] ) ψM /100lc
r
r
where ψ is the CD signal in degrees, M is the molecular weight divided
by the number of residues, l is the path length in decimeters, and c is
the concentration in grams per milliliter.
TFE Titration. Aliquots (40 µL each) of a 2 mg/mL stock solution
of 2 were combined with 360 µL TRIS-buffered saline (11.1 mM TRIS
+
167 mM NaCl, pH 7.2) and TFE in varying proportions. The CD
spectra of the resulting 0.2 mg/mL solutions of 2 were obtained as
described above.
Minimal Inhibitory Concentration Determination. The bacteria
used for these experiments were Escherichia coli JM109, Bacillus
subtilis BR151, Enterococcus faecium A634 (vancomycin resistant),
and Staphylococcus aureus 1206 (penicillin resistant).57 Cells were
grown overnight in brain-heart infusion (BHI) medium (37 g Difco
brain-heart infusion dissolved in 1 L water) at 37 °C to a final
59
56
58
8
concentration of approximately 10 colony forming units/mL. Cells were
diluted to a concentration of 10 colony forming units/mL with sterile
7
BHI medium, determined by the absorbance, and then diluted 20-fold
with sterile BHI medium. Two-fold serial dilutions with sterile BHI
medium were performed in duplicate for each peptide in a sterile 96-
well plate to a final volume of 50 µL in each well. An aliquot (50 µL)
of the cell suspension in BHI medium was added to each well, and the
plate was incubated at 37 °C for 6 h. The absorbance at 590 nm was
monitored. The MIC was defined as the concentration of peptide
required for complete inhibition of growth (no change in the absorbance
at 590 nm).
RP-HPLC Assay. Each â-peptide solution (50 µL, 0.33 mg/mL
â-peptide in a solution of 0.038% TFA in water) was injected onto a
Vydac C4 analytical column (0.46 × 25 cm, 5 µ particles). A linear
gradient of 25-50% solvent B over 50 min or 20-50% B over 60
min was employed.
Minimal Bactericidal Concentration Determination. The MIC
assay was performed as described above. After incubation for 6 h at
Acknowledgment. This work was supported by the NIH
GM56414). T.L.R. was supported in part by an NSF pre-
3
7 °C and determination of the MIC, selected wells with complete
(
inhibition of growth were diluted 100-fold with sterile BHI medium.
Plating was performed by spreading 100 µL of each of these solutions
on a solidified mixture of 2% bacteriological agar in BHI medium and
incubating at 37 °C overnight. After incubation, the number of colonies
on each plate was counted and compared with a control plate containing
doctoral fellowship, and E.A.P. was supported in part by an
NIH Biotechnology training grant. CD data were obtained at
the NSF-supported Biophysics Instrumentation Facility at UW-
Madison. We thank Prof. Daniel H. Rich for advice on peptide
synthesis, Dr. Gary Case for special assistance with the peptide
synthesizer, and Prof. David Andes for advice on MBC
determinations.
3
approximately 10 colony forming units of the same strain of bacteria
without peptide.
(
79) Creighton, T. E. Proteins: Structures and Molecular Principles, 2nd ed.;
W. H. Freeman and Company: New York, 1993; p 14.
JA0270423
J. AM. CHEM. SOC.
9
VOL. 124, NO. 43, 2002 12785