An orphan dermaseptin from frog skin reversibly assembles
to amyloid-like aggregates in a pH-dependent fashion
Ruth Go
¨
ßler-Scho
¨
fberger
1
,Gu
¨
nter Hesser
2
, Martin Muik
3
, Christian Wechselberger
4
and
Alexander Jilek
1
1 Institute of Organic Chemistry, Johannes Kepler University Linz, Austria
2 CSNA Center for Surface- and Nanoanalytics, Johannes Kepler University Linz, Austria
3 Institute of Biophysics, Johannes Kepler University Linz, Austria
4 Center for Biomedical Nanotechnology, Upper Austrian Research, Linz, Austria
Keywords
amphibian skin; amyloid; bioactive peptide;
cytotoxicity; self-assembly
Correspondence
A. Jilek, Institute of Organic Chemistry,
Johannes Kepler University Linz,
Altenbergerstrasse 69, A-4040 Linz, Austria
Fax: +43 732 2468 8747

(
MI:0407)bycircular dichroism (MI:0016)
l
MINT-7255686: Dermaseptin (uniprotkb:O93455) and Dermaseptin (uniprotkb:O93455) bind
(
MI:0407)bybiophysical (MI:0013)
l
MINT-7256439: Dermaseptin (uniprotkb:O93455) and Dermaseptin (uniprotkb:O93455) bind
(
MI:0407)byfluorescence microscopy (MI:0416)
l
MINT-7256449: Dermaseptin (uniprotkb:O93455) and Dermaseptin (uniprotkb:O93455) bind
(
MI:0407)byelectron microscopy (MI:0040)
l
MINT-7256430: Dermaseptin (uniprotkb:O93455) and Dermaseptin (uniprotkb:O93455) bind
(
MI:0407)byfluorescence technologies (MI:0051)
Abbreviations
aDrs, dermaseptin PD-3-7; LDH, lactate dehydrogenase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide; POPC,
1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine; SUV, small unilamellar vesicles; TEM, transmission electron microscopy; ThT, thioflavin T.
FEBS Journal 276 (2009) 5849–5859 ª 2009 The Authors Journal compilation ª 2009 FEBS 5849
Introduction
Amphibian skin glands are known to contain a vast
variety of biologically active peptides, such as neuro-
peptides, peptide hormones, opioid peptides and pep-
tide antibiotics [1,2]. Many of these may help the
individuals to interact with their environment, i.e. they
are targeted exogenously [3]. In the skin of Pachymedu-
sa dacnicolor, a tree frog from southern Mexico, the

Induction of secondary structure, an amphipathic helix
in particular, by phospholipid bilayers is a good indica-
tor of whether a peptide is membrane active. We used
1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC)
unilamellar vesicles and dodecylphosphocholine deter-
gent micelles to mimic the outer leaflet of eukaryotic cell
membranes, which is rich in phospholipids with the
zwitterionic phosphocholine headgroup. By contrast,
SDS micelles favour an initial interaction of cationic
antibacterial peptides by electrostatic attraction in a
similar manner to negatively charged bacterial cell
membranes [9,11]. The CD spectrum of soluble aDrs in
buffer is that of a random coil, with 30% extended
conformation and 5% turn structures independent of
pH (Fig. S1). Addition of POPC small unilamellar vesi-
cles (SUVs) resulted in an increase in the helical struc-
ture up to 2% (7% at pH 8), although the b-sheet
fraction remained largely unchanged. The presence of
SDS or dodecylphosphocholine may have led to an over-
estimated value for the helical fraction of up to 20%
(Fig. 1).
Antibacterial and cytotoxic activity of soluble aDrs
We tested the antibacterial activity of aDrs against the
gram-positive Bacillus subtilis and the gram-negative
Escherichia coli using an inhibition zone assay. Mono-
meric aDrs displayed no antibiotic action in the entire
concentration range up to 32 lm. Cytotoxic activity
against eukaryotic Spodoptera frugiperda (Sf9) insect
cells and NIH-3T3 mouse fibroblast cells was investigated
by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazo-

region demonstrated that maxima at 1619 and
1635 cm
)1
were most likely. These bands can be
attributed to b-sheet structures. A band at
 1685 cm
)1
, which is often observed in b-sheet-rich
proteins, is almost absent (< 1%) [13,14]. Bands at
1609 and 1659 cm
)1
were most likely to be contribu-
tions from side chain amides [15]. Other maxima are
present at 1648 and 1671 cm
)1
, which can probably
be assigned to random coils and sterically restricted
amide bonds such as those present in turns, respec-
tively. Fitting the Gauss peaks onto the band posi-
tions revealed a total contribution of b-sheet
structures of  80%. Third, transmission electron
microscopy (TEM) investigations revealed long
(> 1 lm) fibre bundles. Single fibres (protofilaments)
were found to have a uniform thickness of 5 nm.
Fibril formation is a reversible process dependent
on pH
When exposed to pH 8, the fibrils instantaneously
underwent a phase transition to the predominantly
monomeric random coil form of the peptide because
this could pass freely through the membrane (3.5 kDa

tron-micrographs of aDrs fibrils formed at
acidic pH. Scale bar, 200 nm.
R. Go
¨
ßler-Scho
¨
fberger et al. Self-assembly of anionic dermaseptin
FEBS Journal 276 (2009) 5849–5859 ª 2009 The Authors Journal compilation ª 2009 FEBS 5851
1 h after the addition of buffer. This may indicate that,
in a pH interval close to the transition point, a certain
degree of stacked b-sheet structure is temporarily
preserved.
After incubation at pH 6, TEM inspection of
aliquots of the sample revealed that the number of
amorphous aggregates was drastically reduced after
5 days. After 14 days, such aggregates were apparently
undetectable (data not shown). However, equilibrium
dialysis, independent of time, yielded 20 lm of soluble
aDrs (12–14 kDa membrane cut-off). Thus, only a
A
B
Fig. 3. Phase behaviour of aDrs as a function of pH. (A) Response
of aDrs to repeated pH shifts (arrows). The kinetics of 200 l
M
aDrs was measured by recording the change in fluorescence of
ThT at 482 nm over time. After a lag period of 45 h, fluorescence
increased linearly upon a decrease in pH to pH 2. (B) Effect of pH
on aDrs amyloid-like aggregates. Fluorescence of ThT bound to
aDrs aggregates derived from fibrils recorded at 482 nm as a func-
tion of pH after incubation for 1 h.

detection limits of negative-stain TEM and light scat-
tering. Nevertheless, the putatively aggregated character
was found not to hinder retransition to the amyloidous
state upon a decrease in pH (data not shown).
In the FTIR spectra recorded 5 min after the pH-
shift to 6.0, a minor peak ( 10%) at 1621 cm
)1
was
clearly seen (Fig. 4). This suggested that a significant
number of interchain hydrogen bonds are still present
in the aggregates. Interestingly, this fraction remained
unchanged over the following hour, whereas ThT
fluorescence decreased further by  40% over this time
(Fig. 4). Within this interval, fibrils are no longer
detectable by TEM. In FTIR spectra of the soluble
peptide, the 1621 cm
)1
band finally disappeared in
favour of band contributions at 1645 cm
)1
(random
coil) and  1658 cm
)1
. The latter band is within the
range of contributions from a helices. CD spectra,
however, do not support an a-helical content. These
spectra are virtually indistinguishable from those of
the soluble form.
Antibacterial and cytotoxic activity of
disassembling fibrils

to the C-terminus. BODIPY C1–Fl was chosen for this
purpose because of its moderate hydrophobicity and
its suitability for laser techniques. The labelled peptide
(aDrs*) was shown to have similar aggregation and
cytotoxic properties as unlabelled aDrs (Fig. S2). Con-
focal fluorescence microscopy revealed that both aggre-
gated and soluble aDrs* were found at the plasma
membrane in the early stages (1 h) and after 12 h were
intracellularly distributed within Sf9 target cells, but
did not enter the nucleus (Fig. S3). With the setup
used, we observed no autofluorescence of the cells
(data not shown). Leakage of cytoplasmic material
(blobs) was not observed at any time.
In order to investigate the progression of the poison-
ing effect of aDrs aggregates (12 lm) we also varied
A
B
Fig. 5. Cytotoxic effect of disassembling fibrillar aDrs aggregates
on Sf9 insect cells investigated by the MTT assay. (A) Cells were
treated with lyophilized fibrils resuspended in medium at pH 6.0 at
the indicated peptide concentration for 24 h. (B) Cells were treated
with ‘backward’ aggregates (15 l
M) after preincubation in medium
at pH 6.0 for the indicated intervals. ‘F’, soluble aDrs (15 l
M) was
incubated in medium at pH 6.0 for 1 week. S, soluble aDrs (15 l
M).
(A,B) Reported values are relative to control cells without peptide.
R. Go
¨

C-terminally amidated aDrs (aDrsa) (data not shown).
However, C-terminal amidation of peptide N4 (N4a)
fully stabilized the fibrils at neutral pH (Fig. S5). Dis-
assembly of amyloid-like aDrs is thus likely caused by
a cumulative repulsive effect of negative charges. Nota-
bly, the modifications in N4a also confer a potent
cytotoxic activity (R. Go
¨
ßler-Scho
¨
fberger and A. Jilek,
unpublished observations) which was not characterized
in detail, to the monomeric form.
Discussion
Most cationic antimicrobial peptides are unstructured
in solution. Folding into the active amphipathic con-
formation is triggered by binding to the target mem-
brane. In this study, we used vesicles and micelles to
see whether aDrs would also fold to the proposed
amphipathic helix. The CD spectra indicated a high
propensity for an extended conformation in solution
and only a low increase in helical secondary structure
upon addition of zwitterionic vesicles or even deter-
gents. This may not suffice for a lytic activity
(Fig. 1). Indeed, we could not detect any antibacterial
or significant cytotoxic activity of the monomer, sug-
gesting a function divergent from classical peptide
antibiotics.
Reversible self-assembly of aDrs is controlled
by pH

lian skin, which has a pH of 4.5–5.5, frog skin also
has a pH which may be essential for several key
defence functions [22]. As a consequence, maturation
of the secretory granules might be accompanied by
aggregation of aDrs. In turn, environmental pH would
trigger the dissociation of these aggregates after release
of the gland contents.
Table 1. Peptide sequences of aDrs and variants in single letter
code. C* denotes fluorophor coupled to cysteine via a thioether
linkage.
5101520
aDrs LLGDLLGQTSKLVNDLTDTVGSIV-OH
aDrs* C*-OH
aDrs a NH
2
N4 N N N
N4a N N N NH
2
Self-assembly of anionic dermaseptin R. Go
¨
ßler-Scho
¨
fberger et al.
5854 FEBS Journal 276 (2009) 5849–5859 ª 2009 The Authors Journal compilation ª 2009 FEBS
Metastable granular aggregates are formed
during disassembly
Within the transition interval (pH 5–6.5), metastable
amorphous ‘backward’ aggregates were transiently
formed during melting of the amyloid-like fibrils
(Fig. 4). The clearance or rearrangement of these gran-

mediated by the folded amphipathic monomer, which
is localized at the plasma membrane, where high
effective concentrations can be reached at the sites of
impact of the aggregates. We were interested in deli-
neating the fundamental biochemical mechanism
through which aDrs granular intermediates are cyto-
toxic to eukaryotic cells. Apparently, the peptide
quickly interacted with the cells and penetrated into
the cytoplasm independent of the aggregation state
(Fig. S3). Cellular damage, however, was induced only
by the aggregated form and became detectable after
3 h. During these stages, the integrity of the cell
membrane remained largely intact because no LDH
(M
r
, 140 kDa; diameter,  8 nm) was released from
the cells [27]. These results rule out a cytolytic
mechanism.
In some instances, antimicrobial peptides such as
apidaecin [28] are directed against intracellular targets
rather than against the cell membrane. However,
protofibrils, the prefibrillar ‘forward’ oligomers, are
suspected of being the effectors of cellular damage
associated with numerous protein deposition diseases
[29], including neurodegenerative disorders such as
Alzheimer’s or Parkinson’s disease [30], and amyloi-
doses [31,32]. The inherent cytotoxicity of protofibrils
is characterized by the intracellular localization of the
polypeptides, the intactness of the cell membrane, the
onset of apoptosis and mitochondrial damage [33,34].

the same peptide exerted antimicrobial activity [40].
In conclusion, there is now emerging evidence that
amyloid-like aggregates and their intermediate states
could be an integrative component of the innate
immune system, for which amphibian skin is an
approved model system [41–43]. Remarkably, the
noncytotoxic aDrs from this source exhibits a reversible
R. Go
¨
ßler-Scho
¨
fberger et al. Self-assembly of anionic dermaseptin
FEBS Journal 276 (2009) 5849–5859 ª 2009 The Authors Journal compilation ª 2009 FEBS 5855
pH-dependent self-assembly to amyloid in vitro.A
natural defence strategy could involve such an amyloid
deposit, from which a temporarily cytotoxic agent could
be quickly formed triggered by an increase in pH.
Experimental procedures
Peptide synthesis and purification
All peptides were prepared using solid-phase peptide synth-
esis and were provided by Peptide Specialty Laboratories
(Heidelberg, Germany). BODIPY C1-FLÒ was from Invi-
trogen (Carlsbad, CA, USA). The peptides were purified
by RP-HPLC using a 0.1%(v ⁄ v) HCl ⁄ acetonitrile gradient
solvent system.
SUVs
A solution of 20 mgÆmL
)1
POPC (Avanti Polar Lipids, Ala-
baster, AL, USA) in water was sonicated for 10 min under

Lyophilized aDrs fibrils were incubated with a filtrated
Congo Red stock solution (saturated Congo Red, NaCl in
80% ethanol) [47], centrifuged and the resuspended pellet
was examined for birefringence under crossed polarisators
with a light microscope (Olympus) using a DPlanApo 20
objective.
TEM
Peptide samples were spotted on carbon-coated formvar-cov-
ered copper grids (Ager Scientific, Stansted, UK), negatively
stained with 2% (w ⁄ v) uranyl acetate (Electron Microscope
Sciences, Hatfield, PA, USA) or NanoVan (Nanoprobes,
Yaphank, NY, USA) and examined with a Jeol 2010 electron
microscope (Tokyo, Japan) operated at 100 kV.
FTIR
Hydrogen exchange of the peptides was performed by
repeated dissolution in D
2
O and lyophilization at a neutral
pD. Peptide samples were dissolved in DCl ⁄ D
2
O to a final
concentration of 2 mm and allowed to aggregate.
All spectra were collected at room temperature on a Bruker
(Billerica, MA, USA) spectrometer (model Tensor 27) as pre-
viously described [48]. The amide I’ region (1600–1700 cm
)1
)
of the sample spectrum was examined [49]. Buffers in D
2
O

Self-assembly of anionic dermaseptin R. Go
¨
ßler-Scho
¨
fberger et al.
5856 FEBS Journal 276 (2009) 5849–5859 ª 2009 The Authors Journal compilation ª 2009 FEBS
Liquid growth inhibition assay [51]
Preinocula of E. coli were prepared in LB medium (pH 6.5
or 5.5) at 37 °C and diluted to a D
600
of 0.0001 (1 : 10
6
).
Peptides were added to the bacteria suspension to final con-
centrations of 0, 4or16lm and incubated at 37 °C for
2 h. Bacteria were plated out, incubated overnight at 37 °C
and the colonies counted.
Cell culture
Sf9 cells were cultured in Insect Express SF9-S2 medium
(PAA cell culture company, Pasching, Austria) adjusted to
pH 6.0 at 27 °C. NIH-3T3 cells (mouse fibroblasts) were
cultured in Dulbecco’s modified Eagle’s medium containing
10% bovine calf serum in a 5% CO
2
humidified atmo-
sphere at 37 °C.
MTT inhibition assay [52]
Cells were seeded at a density of 4 · 10
3
cellsÆwell

determine background lysis, cells treated with 1% Triton
X-100 were used to determine total LDH (100% lysis).
Determination of reactive oxygen species
The generation of intracellular reactive oxygen species in Sf9
cells was examined by oxidation of the dye 2¢,7¢-dichloro-
fluorescin diacetate (Invitrogen). Cells were seeded in 24-well
plates and allowed to reach 50% confluence. The cells were
loaded with 5 lm 2¢,7¢-dichlorofluorescin diacetate and
incubated in the presence of 15 lm aggregated peptide for
differing time intervals. Reactive oxygen species levels were
detected by measuring the fluorescence of the oxidized dye
with a plate reader (Zenyth 3100 Multimode Spectrofluori-
meter, Anthos Mikrosysteme GmbH, Krefeld, Germany)
with excitation at 485 nm and emission at 535 nm.
Laser scanning fluorescence microscopy
Sf9 cells were seeded into an AttoFluor chamber (Invitro-
gen). Fibril suspension or soluble aDrs was added to the
cells to a final concentration of 15 lm. After 3 and 20 h of
incubation, a QLC100 Real-Time Confocal System (Visi-
Tech Int., Sunderland, UK) was used for recording fluores-
cence images. A Photometrics CoolSNAPHQ monochrome
camera (Roper Scientific, Sarasota, FL, USA) and a dual
port adapter (dichroic: 505lp; emission filter: 535 ⁄ 50;
Chroma Technology Corp., Rockingham, VT, USA) were
connected in conjunction with an argon ion multiwave-
length (514 nm) laser (Spectra Physics, Mountain View,
CA, USA). This system was attached to an Axiovert 200M
microscope (Zeiss, Oberkochen, Germany).
Acknowledgements
We thank Christa Mollay for helpful suggestions and

characterization, pharmacological activity and cloning
of precursor cDNA. Regul Pept 117, 25–32.
R. Go
¨
ßler-Scho
¨
fberger et al. Self-assembly of anionic dermaseptin
FEBS Journal 276 (2009) 5849–5859 ª 2009 The Authors Journal compilation ª 2009 FEBS 5857
6 Wechselberger C (1998) Cloning of cDNAs encoding
new peptides of the dermaseptin-family. Biochim
Biophys Acta 1388, 279–283.
7 Vanhoye D, Bruston F, Nicolas P & Amiche M (2003)
Antimicrobial peptides from hylid and ranin frogs
originated from a 150-million-year-old ancestral
precursor with a conserved signal peptide but a
hypermutable antimicrobial domain. Eur J Biochem
270, 2068–2081.
8 Amiche M, Ladram A & Nicolas P (2008) A consistent
nomenclature of antimicrobial peptides isolated from
frogs of the subfamily Phyllomedusinae. Peptides, 29,
2074–2082.
9 Shai Y (2002) Mode of action of membrane active anti-
microbial peptides. Biopolymers 66, 236–248.
10 Bechinger B & Lohner K (2006) Detergent-like actions
of linear amphipathic cationic antimicrobial peptides.
Biochim Biophys Acta 1758, 1529–1539.
11 Matsuzaki K (1999) Why and how are peptide–lipid
interactions utilized for self-defense? Magainins and
tachyplesins as archetypes. Biochim Biophys Acta 1462,
1–10.

tropic phases. J Am Chem Soc 125, 9619–9628.
21 Spannhof L (1953) 54) Zur Genese, Morphologie und
Physiologie der Hautdru
¨
sen bei Xenopus laevis. Wiss Z
Humboldt Universita
¨
t Berlin 3, 295–305.
22 Lillywhite HB (2006) Water relations of tetrapod inte-
gument.
J Exp Biol 209, 202–226.
23 Krebs MR, Bromley EH & Donald AM (2005) The
binding of thioflavin-T to amyloid fibrils: localisation
and implications. J Struct Biol 149, 30–37.
24 Feder R, Dagan A & Mor A (2000) Structure–
activity relationship study of antimicrobial dermaseptin
S4 showing the consequences of peptide oligomeri-
zation on selective cytotoxicity. J Biol Chem 275 ,
4230–4238.
25 Sal-Man N, Oren Z & Shai Y (2002) Preassembly of
membrane-active peptides is an important factor in their
selectivity toward target cells. Biochemistry 41, 11921–
11930.
26 Kustanovich I, Shalev DE, Mikhlin M, Gaidukov L &
Mor A (2002) Structural requirements for potent versus
selective cytotoxicity for antimicrobial dermaseptin S4
derivatives. J Biol Chem 277, 16941–16951.
27 Korzeniewski C & Callewaert DM (1983) An enzyme-
release assay for natural cytotoxicity. J Immunol
Methods 64, 313–320.

beta-protein-(1-40) by monosialoganglioside GM1, a
neuronal membrane component. J Mol Biol 371, 481–
489.
Self-assembly of anionic dermaseptin R. Go
¨
ßler-Scho
¨
fberger et al.
5858 FEBS Journal 276 (2009) 5849–5859 ª 2009 The Authors Journal compilation ª 2009 FEBS
37 Fowler DM, Koulov AV, Balch WE & Kelly JW (2007)
Functional amyloid – from bacteria to humans. Trends
Biochem Sci 32, 217–224.
38 Zhao H, Mattila JP, Holopainen JM & Kinnunen PK
(2001) Comparison of the membrane association of
two antimicrobial peptides, magainin 2 and indolicidin.
Biophys J 81, 2979–2991.
39 Sood R, Domanov Y, Pietiainen M, Kontinen VP &
Kinnunen PK (2008) Binding of LL-37 to model bio-
membranes: insight into target vs host cell recognition.
Biochim Biophys Acta 1778, 983–996.
40 Auvynet C, El Amri C, Lacombe C, Bruston F,
Bourdais J, Nicolas P & Rosenstein Y(2008) Structural
requirements for antimicrobial versus chemoattractant
activities for dermaseptin S9. FEBS J 275, 4134–4151.
41 Simmaco M, Mangoni ML, Boman A, Barra D &
Boman HG (1998) Experimental infections of Rana
esculenta with Aeromonas hydrophila: a molecular
mechanism for the control of the normal flora. Scand J
Immunol 48, 357–363.
42 Boman HG (2000) Innate immunity and the normal

three inducible bactericidal proteins from hemolymph
of immunized pupae of Hyalophora cecropia. Eur J Bio-
chem 106, 7–16.
51 Bulet P, Dimarcq JL, Hetru C, Lagueux M, Charlet M,
Hegy G, Van Dorsselaer A & Hoffmann JA (1993) A
novel inducible antibacterial peptide of Drosophila
carries an O-glycosylated substitution. J Biol Chem 268,
14893–14897.
52 Bucciantini M, Giannoni E, Chiti F, Baroni F, Formigli
L, Zurdo J, Taddei N, Ramponi G, Dobson CM &
Stefani M (2002) Inherent toxicity of aggregates
implies a common mechanism for protein misfolding
diseases. Nature 416, 507–511.
Supporting information
The following supplementary material is available:
Fig. S1. CD spectra of soluble aDrs (100 lm) at var-
ious pH.
Fig. S2. aDrs* has similar self-aggregation and cyto-
toxic properties to unlabelled aDrs.
Fig. S3. Progression of the poisoning effect of aDrs
aggregates on Sf9 insect cells.
Fig. S4. No reactive oxygen species were induced by
aDrs aggregates.
Fig. S5. Amyloid-like characteristics of N4a aggregates
at neutral pH.
This supplementary material can be found in the
online article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and