Báo cáo khoa học: 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 pot

Eur. J. Biochem. 270, 2068–2081 (2003) Ó FEBS 2003

doi:10.1046/j.1432-1033.2003.03584.x

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
Damien Vanhoye, Francine Bruston, Pierre Nicolas and Mohamed Amiche
Laboratoire de Bioactivation des Peptides, Institut Jacques Monod, Paris, France

The dermal glands of frogs produce antimicrobial peptides
that protect the skin against noxious microorganisms and
assist in wound repair. The sequences of these peptides are
very dissimilar, both within and between species, so that the
5000 living anuran frogs may produce  100 000 different
antimicrobial peptides. The antimicrobial peptides of South
American hylid frogs are derived from precursors, the
preprodermaseptins, whose signal peptides and intervening
sequences are remarkably conserved, but their C-terminal
domains are markedly diverse, resulting in mature peptides
with different lengths, sequences and antimicrobial spectra.
We have used the extreme conservation in the preproregion
of preprodermaseptin transcripts to identify new members of
this family in Australian and South American hylids. All
these peptides are cationic, amphipathic and a-helical. They
killed a broad spectrum of microorganisms and acted in
synergy. 42 preprodermaseptin gene sequences from 10
species of hylid and ranin frogs were analyzed in the context
of their phylogeny and biogeography and of geophysical
models for the fragmentation of Gondwana to examine the

strategy that these frogs have evolved to generate an enormous array of peptide antibiotics. The hyperdivergence of
modern antimicrobial peptides and the number of peptides
per species result from repeated duplications of a  150million-year-old ancestral gene and accelerated mutations of
the mature peptide domain, probably involving a mutagenic,
error-prone, DNA polymerase similar to Escherichia coli
Pol V. The presence of antimicrobial peptides with such
different structures and spectra of action represents the
successful evolution of multidrug defense by providing frogs
with maximum protection against infectious microbes
and minimizing the chance of microorganisms developing
resistance to individual peptides. The hypermutation of the
antimicrobial domain by a targeted mutagenic polymerase
that can generate many sequence changes in a few steps may
have a selective survival value when frogs colonizing a new
ecological niche encounter different microbial predators.

Frogs and toads have developed a successful strategy for
surviving microbe-laden hostile environments. The skin
secretions of these animals not only produce huge amounts
of biologically active peptides that are very similar to
mammalian neuropeptides and hormones [1–7], they also
contain a rich arsenal of broad-spectrum, cytolytic antimicrobial peptides. These defend the naked skin against
noxious microorganisms and assist in wound repair [8–11].
The peptides are small, 10–50 amino acid residues long,
cationic, and act in a variety of ways, although disrupting
and permeabilizing the target cell membrane is the most
frequent [12–15]. This prevents a target organism from
developing resistance to the peptide. Hence, these peptides
have been recognized as potential therapeutic agents [16,17].

The sequences of these antimicrobial peptides differ
considerably from one amphibian to another [9,10]. The
skin of a frog may have 10–20 antimicrobial peptides of
differing sizes, sequences, charges, hydrophobicity, tridimensional structures and spectrum of action, and this
armament differs between frogs belonging to different
families, genera, species or even subspecies, so that no two
species have yet been found that have the same panoply of
peptide antibiotics [18]. This impressive divergence between
and within species means that there may be as many as
100 000 different peptides produced by the dermatous
glands of the  5000 anuran amphibians [19].
Advanced frogs (suborder: Neobatrachia) are by far the
most important source of antimicrobial peptides and tens of
peptide antibiotics have been found in only a few different
frog species [18]. These include peptides from European,
Asian and North American frogs of the genus Rana (family:
Ranidae; subfamily: Raninae) that are 10–47 residues long
and have a 6- to 9-membered disulfide-bridged, cyclic region
at their C-terminal end (Table 1). A total of 13 distinct
peptide families have been identified based on sequence
similarities [20,21]. These are the brevinin-1 and brevinin-2
families [22], the esculentins-1 and -2 [23,24], ranatuerins-1
and -2 [25], ranalexins [26,27], palustrins-1, -2 and -3 [2],
tigerinins [28] and the japonicins-1 and -2 [21]. The peptides

Correspondence to M. Amiche, Laboratoire de Bioactivation des
Peptides, Institut Jacques Monod, 2 Place Jussieu,
75251 Paris Cedex 05, France.
Fax: +33 1 44 27 59 94, Tel.: +33 1 44 27 69 52,
E-mail:

Abbreviations: DRP, dermaseptin-related peptide; MIC, minimal
inhibitory concentration; Ma, million years ago.
(Received 12 February 2003, revised 11 March 2003,
accepted 19 March 2003)

Keywords: antimicrobial peptides; frog skin; dermaseptin;
hypermutation; gene family.


Ó FEBS 2003

Evolution of amphibian antimicrobial peptides (Eur. J. Biochem. 270) 2069

Table 1. Origins and amino acid sequence of antimicrobial peptides for all frog species in this study. Cysteine residues in bold letters form a disulfide
bridge; a, amide.
Peptide

Amphibian
Family

Sub-family

Genus

Species

Name

Sequence

Hylidae

Phyllomedusinae

Phyllomedusa

bicolor

Dermatoxin
Phylloxin
DRS B1
DRS B2
DRS B3
DRS B4
DRS B6
PBN2
PBN1
DRP-AC1
DRP-AC2
DRP-AC3
DRP-AA11
DRP-AA2-5
DRP-AA3-1
DRP-AA3-3
DRP-AA3-4
DRP-AA3-6
DRP-PD1-5
DRP-PD2-2
DRP-PD3-3
DRP-PD3-6

DRP-PD3-7
Caerin 1.1
Caerin 1.11
Caerin 1.12
Caerin 1.13
Caerin 1.14
Caerin 1.15
Ranalexin
Brevinin-1 E
Brevinin-2 Ef
Esculentin 1B

SLGSFLKGVGTTLASVGKVVSDQFGKLLQAGQa
GWMSKIASGIGTFLSGMQQa

Agalychnis

callydryas

Agalychnis

annae

Pachymedusa

Pelodryadinae

Ranidae

Raninae

Litoria

Rana

dacnicolor

caerulea

catesbeiana
esculenta

rugosa
temporaria

pipiens

Gaegurin-4
Gaegurin-5
Temporin B
Temporin H
Temporin G
Brevinins ) 2Ta
Brevinins ) 2Tb
Ranatuerin-2P
Ranatuerin-2 Pa

in another family, the temporins, are short linear sequences
of 10–13 residues [29]. The South American hylid frogs of
the Phyllomedusinae subfamily (family: Hylidae) also

produce a rich array of linear a-helical antimicrobial
peptides that are 19–34 residues long (Table 1). They
include the dermaseptins B and dermaseptins S [30–33],
phylloxin [34] and dermatoxin [35] from frogs of the
Phyllomedusa genus and 24–33 residues peptides called
dermaseptin-related peptides DRP-AA and DRP-PD from
Agalychnis annae and Pachymedusa dacnicolor, respectively
[36]. Analysis of cDNA clones of antimicrobial peptides
from South American hylids [32,34–37] and Asian, European

AMWKDVLKKIGTVALHAGKAALGAVADTISQa
GLWSKIKEVGKEAAKAAAKAAGKAALGAVSEAVa
ALWKNMLKGIGKLAGQAALGAVKTLVGAE
ALWKDILKNVGKAAGKAVLNTVTDMVNQa
ALWKDILKNAGKAALNEINQLVNQa
GLVTSLIKGAGKLLGGLFGSVTGGQS
FLSLIPHIVSGVAALAKHLG
GLLSGILNTAGGLLGNLIGSLSNGES
GLLSGILNSAGGLLGNLIGSLSNGES
SVLSTITDMAKAAGRAALNAITGLVNQGEQ
SLGSFMKGVGKGLATVGKIVADQFGKLLEAGQG
GLVSGLLNTAGGLLGDLLGSLGSLSGGES
SLWSKIKEMAATAGKAALNAVTGMVNQGEQ
GMFTNMLKGIGKLAGQAALGAVKTLAGEQ
GMWGSLLKGVATVVKHVLPHALSSQQS
GMWSTIRNVGKSAAKAANLPAKAALGAISEAVGEQ
SLGSFMKGVGKGLATVGKIVADQFGKLLEAGKG
ALWKTLLKKVGKVAGKAVLNAVTNMANQNEQ
GMWSKIKNAGKAAAKASKKAAGKAALGAVSEALGEQ
GVVTDLLNTAGGLLGNLVGSLSGGER

LLGDLLGKTSKLVNDLTDTVGSIV
GLLSVLGSVAKHVLPHVVPVIAEHLa
GLFSVLGSVAKHVVPRVVPVIAEHLa
GLFGILGSVAKHVLPHVVPVIAEHSa
GLLSVLGSLKLIVPHVVPLIAEHLa
SVLGKSVAKHLPHVVPVIAEKTa
GLFGLAKGSVAKPHVVPVISQLVa
FLGGLIKIVPAMICAVTKKC
FLPLLAGLAANFLPKIFCKTRKC
GIMDTLKNLAKTAGKGALQSLVKMASCKLSGQC
GIFSKLAGKKLKNLLISGLKNVGKEVGMDVVRTGIDI
AGCKIKGEC
GILDTLKQFAKGVGKDLVKGAAQGVLSTVSCKLAKTC
DVEVEKRFLGALFKVASKVLPSVFCAITKKC
LLPIVGNLLKSLLa
LSPNLLKSLLa
FFPVIGRILNGILa
GILDTLKNLAKTAGKGILKSLVNTASCKLSGQC
GILDTLKHLAKTAGKGALQSLLNHASCKLSGQC
GLMDTVKNVAKNLAGHMLDKLKCKITGC
GFLSTVVKLATNVAGTVIDTIKCKVTGGCRK

and North American ranins [24,29,38,39] has indicated
that they are all derived from a single family of precursor
polypeptides with unique features [39]. Precursors belonging
to this family, designated the preprodermaseptin, have an
N-terminal preprosequence of approximately 50 residues
that is remarkably well conserved both within and between
species, while the C-terminal sequence corresponding to
antimicrobial peptides varies markedly. The conserved

preproregion comprises a 22 residue signal peptide and an
acidic propiece that ends in a typical prohormone processing signal, Lys-Arg. The pattern of conserved and variable
regions in skin antimicrobial peptide precursors is therefore


Ó FEBS 2003

2070 D. Vanhoye et al. (Eur. J. Biochem. 270)

the opposite of that of conventional secreted peptides,
suggesting that the conserved preproregion is important for
the biology of the expressing cell.
The unexpected similarity between the preproregions of
precursors that result in structurally diverse end-products
suggests that the corresponding genes all came from a
common ancestor. The genes encoding dermaseptins B
from Phyllomedusa bicolor and gaegurin-4 from Rana
rugosa have been cloned [40,41]. They have a two exon
coding structure, the first contains codons for the 22-residue
signal peptide and the first three residues of the acidic
propiece and the second exon encodes the remainder of the
acidic propiece plus the processing signal Lys-Arg and the
antimicrobial peptide progenitor sequence. As the conserved
preproregion is encoded by the same gene as the mature
peptide, it cannot have been added by post-transcriptional
events. The vast number of different peptides encoded by
this gene family reflects an unprecedented degree of gene
diversification similar to that of the gene families that
mediate interactions between organisms, such as immunoglobulins [42,43] or venom-derived toxins [44,45]. The
amphibian antimicrobial peptides are thus ideal for studying

the evolution of a large variable gene family.
We have used the remarkable degree of conservation in
the preproregion of the preprodermaseptin transcripts to
identify novel members of this family in Australian hylids
belonging to the genus, Litoria (subfamily: Pelodryadinae)
and in South American hylid frogs (subfamily: Phyllomedusinae). We have determined the activity spectra of the
predicted antimicrobial peptides. A combination of phylogenetic reconstruction, analysis of mutation rates and
geophysical models for the sequence of fragmentation of
Gondwana suggests that the hypervariability of antimicrobial peptides and the number of peptides per species reflect
the combination of speciation events, gene duplications,
targeted hypermutation and subsequent actions of diversifying selection directed by the coevolution of the cell
membrane of microbes and/or adaptation to the particular
microbial biota that these frogs encounter.

Materials and methods
Frog species
Specimens of Agalychnis callydryas, Pachymedusa dacnicolor
and Litoria caerulea were obtained from ÔLa Ferme TropicaleÕ (Paris, France). All procedures involving frogs adhered
to ARVO resolution of the use of animals in research and
the guidelines of INSERM ethical committee on animal
research. Specimens of Phyllomedusa bicolor were housed in
large wooden cages (120 · 90 · 90 cm), covered on three
sides by plastic mosquito net as described previously [3].
Phyllodendron, Potos and Dracena were used as perches,
and water bowls were provided for nocturnal baths. The
frogs were fed crickets. Relative humidity was maintained at
65% by a constantly operating humidifier. The temperature
was maintained at 25 ± 1 °C.
cDNA cloning procedure
One specimen of A. callydryas and L. caerulea were

anesthetized by immersion in ice-water and sacrificed by

pithing. The skin was removed on dry ice and a sample of
 180 mg of tissue was homogenized. Poly(A+) RNAs
were purified over an affinity oligo(dT) spin cellulose
column supplied by Invitrogen (Micro-FastTrack kit). The
cDNA was synthetized by RT-PCR, with 3¢ RACE
(Invitrogen) using a 5¢-primer (5¢-GGCTTCCCTGAA
GAAATCTC-3¢) corresponding to the nucleotide sequence
encoding the conserved N-terminus of the preproregion of
dermaseptin precursors [37] and a primer specific to the 3¢adaptator under the following conditions: 35 cycles of 94 °C
for 240 s, 56 °C for 45 s, 72 °C for 60 s and one cycle of
72 °C for 10 min. The PCR product was cloned in the
pGEMt-easy vector system (Promega) using standard
procedures [46] and used to transform competent JM 109
E. coli. After overnight incubation, the white positive
colonies were screened both with T7 (5¢-ATTATGC
TGAGTGATACCCGCT-3¢) and SP6 (5¢-ATTTAGGTG
ACACTATAGAATAC-3¢) primers. Amplification products of the expected sizes (400–500 base pairs) were
sequenced by the dideoxy chain terminator method. We
determined the sequence for the 5¢-end of the preprocaerins
with the cDNA as a template in RACE PCR with a sense
primer specific of 5¢-adaptator and an antisense specific
primer
as
follows:
5¢-GGATGCTCAACTCTT
TATTGACC-3¢ for caerin 1.1, 5¢-ATGACTTTATCCT
AAGGC-3¢ for caerin 1.11, 5¢-CTGAGTGAACAGCTA
TAACTG-3¢ for caerin 1.12, 5¢-GACTTTATCCTAA

GTGTTCAGC-3¢ for caerin 1.13, 5¢-TGTGGAAGGTG
TTTACTAATGG-3¢ for caerin 1.14, and 5¢-GAAGT
ACGTGCTTAGCAACGG-3¢ for caerin 1.15, and for
nested PCR: 5¢-ATAACTGGAACAACGTGTGG-3¢
for caerin 1.1, 5¢-CTAAGTGCTCAGCAATGACG-3¢
for caerin 1.11, 5¢-AGCATAACTGGAACGTGGG-3¢ for
caerin 1.12, 5¢-CAGCAATAAGTGGAACAACG-3¢ for
caerin 1.13, 5¢-GTGTTTAGCAACGGATTTACC-3¢
for caerin 1.14 and 5¢-AGCAACGGATCCTAGGA
CAC-3¢ for caerin 1.15. The temperature cycle used for
the RACE PCRs was: 94 °C for 240 s, 35 cycles at 94 °C for
40 s, 56 °C for 45 s, 72 °C for 60 s, and a final extension
step of 72 °C for 10 min. The PCR products were cloned
and sequenced as above. A similar approach was used to
clone PBN1 and PBN2 cDNAs from P. bicolor, and DRPAC1, 2 and 3 cDNAs from A. callydryas.
Solid phase peptide synthesis
Caerin 1.11, dermaseptin B2 and PBN2 were synthesized
using solid-phase FastMoc chemistry procedures on an
Applied Biosystems 433 A automated Peptide Synthesizer
(Applera, France). Fmoc-protected amino acids and resins
were from Senn Chemicals (Switzerland) and solvents from
Sds (France). The carboxylic acid terminal peptides were
prepared on a 4-benzyloxybenzyl alcohol resin (Wang PS
resin) substituted at 1.18 mmolỈg)1. Carboxamidated peptides were prepared on a 4-methylbenzhydrylamin Polystyren resin (Rink Amide MBHA PS resin) substituted at
0.81 mmolỈg)1. Synthesis was carried out using a doublecoupling protocol: Fmoc amino acids (10 molar excess) were
coupled for 30–60 min with 2-(1H-benzotriazol-1-yl)1,1,3,3-tetramethyluronium hexafluorophosphate 1-hydroxybenzotriazol in a solution of N,N-dimethylformamide
and diisopropylethylamine as activating agents with the


Ó FEBS 2003

Evolution of amphibian antimicrobial peptides (Eur. J. Biochem. 270) 2071

addition of N-methylpyrrolidone. Capping with acetic
anhydride was performed at the end of each cycle.
Temporary N-Fmoc protecting groups were removed by
20% piperidine in N-methyl-2-pyrrolidone. Side chains were
protected with tert-butyloxycarbonyl (tBoc) for lysine and
tryptophane; O-tert-butyl ester (OtBu) for glutamic acid
and aspartic acid; trityl (Trt) for histidine, threonine,
glutamine and asparagine; O-tert-butyl ether (tBu) for
serine and 2,2,4,6,7-pentamethyldihydrobenzofurane-5sulfonyl (Pbf) for arginine. Cleavage of the peptidyl resin
and side chain deprotection were carried out in a mixture
composed of 95% trifluoroacetic acid, 2.5% triisopropylsilane and 2.5% water for 2 h at room temperature. The
resulting mixture was filtered to remove the resin and the
crude peptides were precipitated with ether at )20 °C. They
were recovered by centrifugation at 5000 g for 15 min at
4 °C, washed three times with cold ether, dried under a
stream of nitrogen, dissolved in 10% acetic acid and
lyophilized. The lyophilized crude peptides were purified by
reverse-phase HPLC on a Nucleosil C18 column (5 lm,
10 · 250 mm) eluted at 4 mLỈmin)1 with a 0–60% linear
gradient of acetonitrile in 0.07% trifluoroacetic acid/water
over 30 min. The homogeneity of the synthetic peptides was
assessed by MALDI-TOF mass spectrometry (Voyager DE
RP, Perseptive Biosystems) and analytical HPLC as
described previously [32].
Antimicrobial Assays
Gram-positive eubacteria (Aerococcus viridans, Bacillus
megaterium, Staphylococcus aureus and Staphylococcus

haemolyticus), Gram-negative eubacteria (E. coli B, Salmonella typhimurium, Salmonella enteritidis, Enterobacter cloacae, Klebsiella pneumoniae) and Saccaromyces cerevisiae
were cultured as described previously [34,35]. The minimal
inhibitory concentrations (MICs) of peptides were determined in 96-well microtitration plates by growing the
bacteria in the presence of twofold serial dilutions of
peptide. Aliquots (10 lL) of each serial dilution were
incubated with 100 lL of a suspension of a midlogarithmic
phase culture of bacteria at a starting A630 value of 0.001 in
Poor-Broth nutrient medium (1% bactotryptone, 0.5%
NaCl, w/v) or yeast/peptone/glucose for S. cerevisiae.
Inhibition of growth was assayed by measuring the A630
value after 16 h at 37 °C for bacteria growth and at 30 °C
for yeast. The minimal inhibitory concentration (MICs) was
defined as the lowest concentration of peptide that inhibited
the growth of 99% of the cells. Bacteria were incubated for
2 h with different concentrations of peptides and plated on
solid culture medium containing 1% noble agar to distinguish between bacteriostatic and bactericidal effects. The
plates were subsequently incubated and examined daily for
the formation of colonies. All assays were performed in
triplicate plus positive-controls without peptide and negative-controls with 0.7% formaldehyde.
Sequence analysis
The nucleotide sequences of cDNAs encoding 18 known
dermaseptin B, phylloxins, dermatoxins, and dermaseptinrelated peptides from the hylids A. annae, P. dacnicolor and
P. bicolor were obtained from GenBank, in addition to the

cDNAs encoding PBN1, PBN2, DRP-AC 1, 2 and 3 and
caerins identified in this study (Table 1). The nucleotide
sequences of the cDNAs encoding 13 known brevinins,
esculentins, gaegurins, ranalexins, temporins and ranatuerin
from the ranins, Rana catesbeiana, R. esculenta, R. rugosa,
R. temporaria and R. pipiens were also obtained from

GenBank (Table 1). We aligned the nucleotide sequences
of the antimicrobial peptide transcripts with CLUSTAL X [47]
and by eye. We first aligned the predicted amino acid
sequences of the different domains of the peptides and
nucleotide sequences of the 5¢- and 3¢-untranslated regions
(UTR) separately with CLUSTAL X. Then we added the
nucleotide sequences of the different regions and finally
adjusted the alignment manually. Molecular phylograms
from the alignments were determined with NeighborJoining from Kimura-two-parameters distances [48] using
PAUP [49]. Levels of support for branches were estimated
with bootstrapping methods (1000 replicates) also with PAUP.
Sequence groups were denoted based on the existence of
distinct clades and similarity of predicted amino acid
sequences. We interpreted the origins of the gene families
from the topologies of the phylogram; we assume that the
sequences represent distinct loci in the species sampled. We
estimated the proportion of synonymous substitutions per
synonymous sites (Ds) from the beginning of the preproregion to the last codon before the stop codon with method I
of Ina [50] in South American hylids, Australian hylids and
ranins to determine if different gene regions are subject to
different mutation rates. For comparaison, Jukes–Cantor
distances [51] were estimated for the 5¢- and 3¢-UTR using
the same program. The transversion/transition rate ratios
(Tv/Ts) were estimated by pairwise comparison of the
sequences from South American hylids, Australian hylids
and ranins transcripts by counting with JADIS the proportion of sites with transitional and transversional differences
between two sequences. Tv and Ts were calculated independently within the signal, propiece and mature domains.

Results
cDNA cloning of preprodermaseptins from Australian

and South American hylids
The dermal secretions of Australian tree frogs of the genus
Litoria (family Hylidae; subfamily: Pelodryadinae) all
contain broad-spectrum antimicrobial peptides, the caerins,
whose structures are very different from those of South
American hylid antimicrobial peptides [52–54]. Four caerin
subfamilies, caerins 1–4, have been identified to date, each
comprising several distinct peptides. The most widespread of these is caerin 1.1, which has the sequence,
GLLSVLGSVAKHVLPHVVPVIAEHL-amide (Table 1).
Nothing was known about the gene encoding these
peptides. 3¢-RACE analysis of skin mRNA from L. caerulea using a primer based on the conserved coding region of
preprodermaseptins revealed six different cDNAs (Fig. 1).
One of these cDNAs encoded caerin 1.1, while the remaining five cDNAs coded for new members of the caerin 1
family. The five predicted peptides, tentatively designated
caerins 1.11–1.15, were 22–25 residues long and their amino
acid sequence was 58–88% identical to caerin 1.1. They are
more distantly related to the members of the other caerin


2072 D. Vanhoye et al. (Eur. J. Biochem. 270)

Ó FEBS 2003

Fig. 1. Nucleic acid and deduced amino acid sequences of cDNAs encoding caerins from the skin of Litoria caerulea. (A) Nucleic acid and predicted
amino acid sequence of the cDNA encoding caerin 1.1. The predicted amino acid sequence of preprocaerin 1.1 is given in capital letters under the
nucleotide sequence. The amino acid sequence of mature caerin 1.1 is given in bold letters. A solid line is drawn under the amino acid sequence of the
signal peptide. Nucleotides are numbered positively from the 5¢- to 3¢-ends of the cDNA. Amino acids are numbered starting with position 1 in
the open reading frame. *Stop codon. The glycine residue at the end of the peptide progenitor sequence is involved in the formation of the C-terminal
amide of caerin 1.1 (52, 53). (B) The deduced amino acid sequences of cDNAs encoding caerins 1.1., 1.11, 1.12, 1.13, 1.14 and 1.15. A solid line
is drawn under the amino acid sequences of the signal peptides. Predicted amino acid sequences of mature caerins are given in bold letters.

families. Each precursor polypeptide had an extra-glycine
residue at the carboxyl terminus of its progenitor sequence,
indicating that C-terminal amidation is involved in production of the final peptide. The N-terminal regions of the
caerin precursors encompassing the signal peptides all
contained 22 residues and were superimposable with only
four exceptions (82% identical). The acidic propieces
contained 27 residues and the amino acid sequences were
92.5% identical. Lastly, the 5¢- and 3¢-UTR of the corresponding cDNAs were 84% and 77% identical, respectively.
A comparison of the amino acid sequences of the
preproregions of the six caerin precursors with those of
preprodermaseptins from South American hylid frogs
revealed that the signal peptides (95% identitical) and the
acidic propieces (96% identical) were highly conserved
(Fig. 2). This similarity also extended to the 5¢- and 3¢untranslated regions of the respective mRNAs (not shown).
Caerin 1.1 and related peptides from Australian hylids are
thus an unexpected addition to the structurally diverse
peptides encoded by genes belonging to the preprodermaseptin family.
A similar approach was used to identify new preprodermaseptin-related cDNA sequences in the skins of South
American hylid frogs (Fig. 2). Two of these sequences from
P. bicolor encoded novel putative peptides we have called
PBN1 and PBN2. PBN1 (FLSLIPHIVSGVAALAKHL)
and PBN2 (GLVTSLIKGAGKLLGGLFGSVTGGQS)
do not resemble any antimicrobial peptides identified to
date in the skin secretions of P. bicolor (Table 1). The other
three sequences from A. callidryas encoded peptides, called
DRP-AC 1, 2 and 3, that are structurally related, but not
identical to DRP-AAs from A. annae.

Secondary structure and antimicrobial activities

of predicted peptides
We selected caerin 1.11 and PBN2 to evaluate whether
within-species differences between antimicrobial peptides
reflect functional differentiation. Caerin 1.11 and caerin 1.1
differ only by three amino acid substitutions. In contrast,
the sequence of PBN2 is very different from that of other
P. bicolor antimicrobial peptides. As shown previously, the
common structural feature of linear cationic antimicrobial
peptides such as caerin 1.1 and dermaseptins S and B is the
adoption of a stable amphipathic a-helix upon binding to
the membrane surface [55,56]. The predicted secondary
structures of caerin 1.11 and PBN2 suggest that they can
assume an amphipathic a-helical structure, and therefore be
antimicrobial peptides (Fig. 3). According to Segrest et al.
[57], both peptides belong to the class L group of helices that
are highly positively charged with a narrow polar face and a
highly hydrophobic apolar face. The average charged polar
face subtended a mean radial angle of  200° for both
caerin 1.11 and PBN2. The circular dichroism spectra of
synthetic PBN2 and caerin 1.11 (not shown) had strong
minima at 200 nm in aqueous solution, reflecting the great
degree of unordered structure. Adding micellar concentrations of SDS to the aqueous solution greatly altered the
dichroic spectra of both peptides. Ellipticity decreased at
208 and 222 nm and increased at 192 nm, indicating the
stabilization of the a-helical structure (30% helix content for
caerin 1.11 and 42% for PBN2) in the membrane-mimetic
environment.
The antibacterial and cytotoxic activities of synthetic
caerin 1.11 and PBN2 were tested (Table 2). The

Ó FEBS 2003

Evolution of amphibian antimicrobial peptides (Eur. J. Biochem. 270) 2073

Fig. 2. Conserved preproregions and hypervariable antimicrobial domains of preprodermaseptins. (A) Diagram of preprodermaseptin cDNAs. The
coding region, including the signal peptide, the acidic propiece and the antimicrobial progenitor sequence is drawn as a rectangle. (B) Alignment of
the predicted amino acid sequences (single-letter code) of preprodermaseptin cDNAs obtained from hylid and ranin frogs, including the signal
peptide, the acidic propiece and the antimicrobial progenitor sequence. The predicted hydrophobic signal peptide includes the first 22 amino acid
residues, while the acidic propiece comprises 16–27 residues. Gaps (–) have been introduced to maximize sequence similarities. Identical (black
background) and similar (shaded background) amino acid residues are highlighted. Among the hylid sequences, DRS, dermaseptin B from
P. bicolor, DRP, dermaseptin-related peptide (appended with AA, AC or PD to indicate that the sequences were identified from A. annae,
A. callidryas and P. dacnicolor, respectively). Among the ranin sequences, temporins B, H and G and brevinins 2Ta and 2Tb are from Rana
temporaria, brevinins 1E and 2Ef and esculentin 1B from R. esculenta, ranalexin from R. catesbeiana, gaegurins 4 and 5 from R. rugosa, and
ranatuerin-2P and 2 Pa from R. pipiens. cDNAs encoding PBN1 and PBN2 (GenBank accession numbers: AY218784 and AY218783), DRP-AC
1, 2 and 3 (accession numbers: AY218775, AY218776 and AY218777) and caerins (accession numbers: AY218778-82 and AY218785) were
identified in this study.

corresponding values for dermaseptin B2 from P. bicolor
are shown for comparison. Although their primary structures are very different, PBN2 and dermaseptin B2 showed
overlapping antimicrobial spectra. They had broadspectrum antibacterial activities, inhibiting the growth of
Gram-positive bacteria, Gram-negative bacteria and yeast
with minimal inhibitory concentrations in the lM range. The
dose–response profiles showed sharp curves in which
0–100% inhibition was generated within a 1–2-fold peptide
dilution (not shown). Bacteria incubated overnight with
25 lM PBN2 or dermaseptin B2 produced no colony
forming units, indicating that the peptides are bactericidal.

A combination of PBN2 and dermaseptin B2 was also

dramatically synergistic, so that the mixture sometimes had
15-times greater antibiotic activity than the peptides separately (Table 2). Although their primary structures are very
similar, caerin 1.1 and caerin 1.11 differred unexpectedly in
their capacity to inhibit the growth of various bacteria. For
instance, whereas caerin 1.11 effectively inhibited E. coli
(MIC, 25 lM), caerin 1.1 was inactive [54]. Conversely,
caerin 1.1 effectively inhibited the proliferation of S. aureus
[54], while caerin 1.11 had no effect. Clearly, the withinspecies differences between members of the caerin 1 family
indicates that the peptides are functionally differentiated.


Ó FEBS 2003

2074 D. Vanhoye et al. (Eur. J. Biochem. 270)

Table 2. Inhibition of yeast and bacterial growth in vitro by caerin 1.11,
PBN2 and dermaseptin B2. The MIC is the minimal dose producing
100% inhibition of growth after incubation for 24 h in culture
medium. ND, not determined.
Peptide minimal inhibitory
concentration (lM)

Microorganism
Escherichia coli B
Salmonella typhimurium
Salmonella enteritidis
Enterobacter cloacae
Klebsiella pneumoniae
Aerococcus viridans
Bacillus megaterium

Staphylococcus aureus
Staphylococcus
haemolyticus
Saccaromyces cerevisiae
a

Caerin
1.11

DRS
B2

PBN2

DRS
B2 + 0.25 lM
PBN2

25
Ra
R
50
25
ND
R
R
50

1.5
3.1

3.1
3.1
1.5
ND
0.4
12.5
6.2

1.5
3.1
1.5
12.5
0.8
3.1
0.8
3.1
0.8

0.4
0.8
1.5
3.1
0.2
ND
0.4
0.8
0.8

ND

5.5

3.1

0.8

MIC > 100 lM.

ness of the apolar face, net charge and conformational
flexibility play a crucial role in modulating the biological
potency of linear helical peptide antibiotics.
Molecular phylogeny of preprodermaseptins
We examined the evolutionary relationships between preprodermaseptin cDNAs by constructing phylogenetic trees
from alignments of DNA and predicted protein sequences
(Fig. 2). The phylogenetic reconstruction shown in Fig. 4A
indicates that the 10 frog species from which the 42
preprodermaseptin sequences were obtained fall into two
distinct clusters. The nucleotide sequences of the antimicrobial peptides from the South American and Australian hylids
cluster separately from those from the ranins. The cluster of
hylid sequences is not very well resolved, but there were
several distinct clades: one clade of 12 dermaseptins B and
dermaseptin-related peptides was supported by a bootstrap

Fig. 3. Helical wheel plots (73) of (A) caerin 1.11 and (B) PBN2. Apolar
residues are in bold letters. The amino acid sequence of each peptide is
shown beneath the corresponding wheel plot. The predicted helical
domains are underlined.

The data presented here showed that despite very
different primary structures, caerin 1.11 and PBN2 belong

to the class of cationic, amphipathic a-helical antimicrobial
peptides which interact with and disrupt cell membrane. CD
and antimicrobial assays showed that the helical contents of
PBN2 and caerin 1.11 are not correlated with antibacterial
potency. These observations suggest that additional parameters, namely hydrophobicity, hydrophobic moment, bulky-

Fig. 4. Molecular phylogeny of preprodermaseptins. (A) Neighborjoining tree constructed from Kimura two parameters distances computed from comparison of entire preprodermaseptin cDNA sequences
(including 3¢- and 5¢-untranslated regions) obtained from hylid and
ranid frogs. Bootstrap values from 1000 replicates greater than 50%
are indicated on branches. The distance scale is drawn below the tree.
Maximum parsimony, maximum likelihood and LogDet analyses
yielded the same ordinal phylogeny. Phylogram is midpoint rooted. (B)
Paleogeographic reconstruction of the fragmentation of Gondwana
during the late Jurassic/early Cretaceous period [64]. Land areas are
shaded. Ancestors of Australian hylids are believed to have crossed
Antartica to Australia from South America  150–130 Ma. Ancestors
of Asian, European and North American ranins evolved on isolated
India between 150 and 65 Ma and colonized the Laurasian land mass
after India collided with Asia. Abbreviations; AF, Africa; IND, India;
AUS, Australia; SA, South America; ANT, Antartica. Reconstruction
map from />

Ó FEBS 2003

Evolution of amphibian antimicrobial peptides (Eur. J. Biochem. 270) 2075

values of 81% (hylid clade 1) while a second clade of 17
dermaseptin-related peptides included the caerins from
Litoria (hylid clade 2). Within hylid clade 2, the sequences

from Litoria clustered together monophyletically and were
more closely related to PBN1 from P. bicolor. The average
Kimura 2-parameter pairwise distances were 0.34 for the


2076 D. Vanhoye et al. (Eur. J. Biochem. 270)

Ó FEBS 2003

sequences from South American hylids, 0.32 for those from
ranins, and 0.08 for the sequences from L. caerulea. The
average pairwise distances were much greater between each
cluster of sequences (0.70 between South American hylids
and ranins, and 0.72 between ranins and L. caerulea). Thus,
despite considerable variations between the sequences of the
mature antimicrobial peptides, phylogenetic reconstruction
using the complete sequence of the preprodermaseptin
transcript produced a tree topology that agreed with the
traditional classification of neobatrachian frogs [19,58]. The
divergence of the antimicrobial peptides and their evolutionary relationships would never have been apparent without
the strong conservation of the precursor preproregion.
The molecular phylogram shows that the genes encoding
the antimicrobial peptides in Hylidae and Ranidae arose
from a common ancestral locus that subsequently diversified by several rounds of duplication and subsequent
divergence of loci. Most of the duplication events predated
the radiations of ranins and of South American hylids and
occurred before cladogenesis in a species ancestral to all
these species, i.e., the sequences do not cluster according to
species and are more closely related between than within
species. In contrast, the phylogenetic tree suggests that gene

duplications in L. caerulea occured after the divergence
of South American and Australian hylids.
Accelerated mutation of the antimicrobial peptide
domain of preprodermaseptins
The strikingly greater variability in the antimicrobial
peptide progenitor sequences of the precursors compared
to the highly conserved preproregions may indicate different
mutation rates in the corresponding regions of the genes.
We tested this hypothesis by measuring the rates of
nucleotide synonymous (silent) substitutions in the three
domains (signal peptide, acidic propiece and antimicrobial
peptide progenitor sequence) of the translated regions of the
preprodermaseptin sequences in South American and
Australian hylids and in ranins. As synonymous substitutions are apparently neutral, the fixation rates can be
considered to be proportional to the mutation rates and the
number of synonymous substitutions per synonymous sites,
Ds, to be an adequate representation of the mutation rate.
The nucleotide substitution rates in the 5¢- and 3¢-untranslated regions were calculated using the Jukes–Cantor one
parameter model to be consistent with the Ds estimation.
The average Ds in the mature antimicrobial domains were
more than two to five times greater than Ds estimated from
the signal peptide domain or the Jukes–Cantor distance
estimated from the untranslated regions (Fig. 5A). and the
substitution rates for the mature peptide domain were
certainly underestimated because of multiple substitutions
per site (saturation) in the mature region and because
additions and deletions were ignored in the calculation. In
all cases, the length of the signal peptide was constant while
many additions and deletions were needed for optimal
alignment of the antimicrobial peptide progenitor sequences

(Fig. 2).
The different mutation rates of the different regions in the
transcripts of the preprodermaseptins is also indicated by
the unrooted trees shown in Fig. 6. Genes were segregated
in clearly defined branches according to the classical

Fig. 5. Accelerated mutation of the antimicrobial peptide domain of
preprodermaseptins. (A) Average (± SD) pairwise Jukes–Cantor distances (5¢- and 3¢-untranslated regions) and proportions of synonymous substitutions per synonymous site (D) among the signal peptide,
acidic propiece and antimicrobial peptide progenitor domains estimated from the nucleotide sequences of preprodermaseptins from
South American and Australian hylids and from ranins (Fig. 4A). (B)
Average (± SD) ratios of nucleotide transversions to transitions (Tv/
Ts) calculated for the signal peptide, acidic propiece and antimicrobial
peptide progenitor domains of preprodermaseptins from South
American and Australian hylids and from ranins. Tv/Ts ratios determined by Maor-Shoshani et al. [62] for DNA polymerases Pol III and
Pol V are shown for comparison.

phylogeny of species when using either the 5¢-UTR, the
signal peptide sequences or the acidic propiece sequences. In
contrast, transcripts started to diverge directly from the
origin with mixed branches when using the antimicrobial
progenitor sequences. Synonymous substitutions should
accumulate at similar rates in different regions of a gene.
Deviations from this behavior may be due to difference in
codon usage, or differences in mutation rate across the gene.
As no codon bias was detected, our data suggest that a
mechanism is operating that leads to very different mutation
rates in adjacent regions of these small genes.
Molecular signatures of mutagenic polymerases
targeted to the antimicrobial peptide domain
Whereas diversifying (positive) selection contributes to the

accelerated evolution of antimicrobial peptides [59], recent
studies have suggested that hypervariability in specific gene
regions may result from the actions of targeted mutagenic,


Ó FEBS 2003

Evolution of amphibian antimicrobial peptides (Eur. J. Biochem. 270) 2077

Fig. 6. Unrooted neighbor-joining trees constructed using Kimura two parameters distances computed from comparison of specific regions of preprodermaseptin cDNA sequences.

error-prone, DNA polymerases similar to DNA polymerase
V from E. coli [44,60,61]. Molecular signatures of the SOSinducible polymerase V are low processivity and a strong
bias for transversions over transitions [62]. We examined the
transversion/transition (Tv/Ts) ratios in alignments of
preprodermaseptin transcripts from hylid and ranid frogs.
The Tv/Ts ratios in the different regions of the transcripts
were clearly different (Fig. 5B). It increased from the signal
peptide to the acidic propiece, to the antimicrobial peptide
progenitor sequence, with a twofold bias for Tv over Ts in
the progenitor sequence. This value was similar to that

predicted for random substitions and corresponded to the
in vitro measured transversion bias of DNA Pol V reported
by Maor-Shoshani et al. [62]. These results suggest that a
targeted mutagenic process involving a DNA Pol V-like
enzyme has operated in hylids and ranins within the peptide
progenitor sequence of antimicrobial peptide loci, but not in
the signal peptide and acidic propiece domains. Thus
speciation events, gene duplications, targeted hypermutations and the subsequent actions of diversifying selection

have all contributed to the evolution and diversification of
this large family of hypervariable genes.


2078 D. Vanhoye et al. (Eur. J. Biochem. 270)

Discussion
Given the low dispersal abilities of amphibians over salty
environments, tectonic movments and changes in sea-level
have been of importance in shaping the distribution of
lineages of ranid and hylid frogs. As emphasized by Savage
[63], the historical biogeography of neobatrachian frogs is
associated mainly with Gondwanaland. Thus, the current
biogeographic distributions of ranids and hylids together
with tectonic events during the fragmentation of Gondwana
can be used to provide a temporal framework for the origins
and evolution of the preprodermaseptin genes.
The supercontinent Gondwana began to break up about
150 Ma when India/Madagascar separated from Africa
(Fig. 4B) [64]. India drifted north-eastward during the
ensuing 100 Myr, and finally collided with Asia 65–56 Ma.
Antartica/Australia became disjoined from South America
around 150 Ma. They were reconnected by a narrow land
bridge between 140 and 130 Ma, after which there was an
archipelago until 110 Ma. Australia separated from Antartica about 60 Ma. Africa and South America were
substantially separated around 125 Ma and became completely detached before 100 Ma. The many endemic and
very diverse species of ranid frogs in Africa, Madagascar
and India [19,65], is usually interpreted as indicating that
they originated on these land masses before Africa and
India/Madagascar separated. Furthermore, thorough analyses of the phylogenetic relationships of ranids based on

mitochondrial and nuclear DNA sequences [65–67] has
demonstrated that the speciation events giving rise to the
Indian ranid lineages took place  130–150 Ma when India
had separated from Africa. Indian Raninae, including the
genus Rana, originated and began to diversify on the
drifting Indian block between 130 and 65 Ma, spread from
India after its collision with Asia 65–56 Ma and radiated
into more than 200 species in Asia, Europe and North
America [19,67]. Biogeographic and molecular evidence
indicates that South America is the site from which the
Hylidae radiated [19,65]. Among the four subfamilies of
hylids, the Phyllomedusinae, the Hemipractinae and the
Hylinae are predominantly South American, while the
Pelodryadinae, including the genus Litoria, are restricted to
Australia and New Guinea. These data suggest that
ancestors of the Australian hylids originated in South
America and probably reached Australia via the connection
with Antartica and South America [66]. Evidence for the
dispersal of land vertebrates from South America to
Australia via Antartica also comes from fossils and
molecular data on marsupials [68]. Thus, the paleogeographic reconstruction shown in Fig. 4B suggests that the
immediate ancestor of Australian Pelodryadinae must have
been present in Australia before Australia and Antartica
separated 60 Ma, and in Antartica well before the last
known connection between Antartica and South America
150–130 Ma. Once South America and Antartica were
separated, the hylid fauna in Australia and South America
diversified independently. This historical reconstruction
shows that the gene family encoding antimicrobial peptides
from South American and Australian hylids and from

Indian, Eurasian and North American ranins arose before
the isolation of India and South America from Africa in a
pan-Gondwanan common ancestor of these species. Given

Ó FEBS 2003

that Archeobatrachia (Ôarchaic frogsÕ) and Neobatrachia
diverged about 200 Ma [69], and that preprodermaseptins
have not been detected in archaic frogs, the present data
indicates maximum and minimum origination dates of
approximately 200–150 Ma.
Our study provides evidence that several different mechanisms appear to have contributed to the diversity of
modern antimicrobial peptide loci since the Jurassic period.
These include several gene duplications, most of them
predating the divergence of species, and diversifying (positive) selection. We have shown previously that members of
the preprodermaseptin gene family have been subject to
positive selection within the acidic propiece and the
antimicrobial peptide domain of hylids and only within
the antimicrobial peptide domain of ranins [59]. The third
mechanism is an accelerated rate of nucleotide substitution
that is focused on the antimicrobial peptide domain in
preprodermaseptin genes. This mechanism, that has yet to
be proved, may involve a mutagenic DNA polymerase with
a strong transversion bias, such as DNA Pol V from E. coli
which can generate multiple changes in the sequence in a
single step. Antimicrobial peptides are synthesized in the
multinucleated cells of the dermatous glands of the skin and
large amounts are stored in the secretory granules of these
glands [70,71]. The glands may release their content onto the
skin surface by a holocrine mechanism involving the rupture

of the plasma membrane and the extrusion of the granules
through a duct opening to the surface. Skin antimicrobial
peptides do not participate in the general metabolism and
physiology of the frog producing them. As no deleterious
effects are expected when a new peptide variant rapidly
emerges by multiple changes in a current sequence, the new
sequence may bypass the actions of neutral and negative
(purifying) selections. Thus, the combination of targeted
hypermutations to generate great variation plus the subsequent action of positive selection may explain both the
hypervariation and large number of antimicrobial peptides
per species.
Although the organization of eukaryotic signal peptides
is evolutionary conserved, the sequences are not similar. The
evolutionary pressure that results in the conservation of the
signal peptide and, albeit to a lesser extent, the acidic
propiece in preprodermaseptins is especially striking in view
of the extreme variations in the contiguous antimicrobial
peptide domain and the very ancient history of the gene
family. This suggests that these conserved elements have
important functions. Conticello et al. have suggested that
conserved elements in an otherwise hypermutable DNA
sequence might be protected from mutagenesis by specifically bound macromolecules that serve to prime DNA
Pol V-like polymerase in the vicinity [60]. By stopping
normal DNA replication, DNA-bound macromolecules
may then create single-strand gaps in the replicating strand,
thus recruiting mutagenic polymerase as part of the damage
response to the lesion. This hypothetical scenario is
attractive in that it provides a plausible explanation for
the conservation of the signal peptide sequences for millions
of years although it lies in the precursors of very different

antimicrobial peptides produced by very distantly related
species of frog.
Finally, the underlying biology of preprodermaseptin
diversification within the hylid and ranin frogs invites some


Ó FEBS 2003

Evolution of amphibian antimicrobial peptides (Eur. J. Biochem. 270) 2079

speculation. Each ranin or hylid frog species produces its
own set of preprodermaseptin-derived antimicrobial peptides. Some of these peptides differ by only a few amino acid
substitutions or deletions and have similar biochemical
characteristics. Other peptides have widely different
sequences and physicochemical properties. The presence
in frog skin of numerous antimicrobial peptides, acting
separately or in concert, may have a selective survival value
in habitats laden with microorganisms. As shown in this and
other studies [52–55,70], a comparison of antimicrobial
peptides issued from Phyllomedusa or from Litoria revealed
that most of them kill a broad spectrum of microorganisms
but differ widely in their capacity to kill the various agents.
Even peptides with very similar structures, such as caerin 1.1
and caerin 1.11 or dermaseptins S [9], target specific
microorganisms. Without direct functional characterization, it would have been difficult to predict the different
spectra of these peptides. Moreover, these peptides act in
synergy, with the mixture having up to a 10- to 100-fold
greater antibiotic activity than the peptides separately
(Table 2) [72]. Hence, the hypervariability of skin antimicrobial peptides and the proposed mechanisms of diversification could be part of a strategy for providing frogs with
a maximum protection against a wide range of infectious

microorganisms. Also, these antimicrobial peptides with
such diverse structures and spectrum of action can be
viewed as the successful evolution of a multidrug defence
system, which minimizes the chance of microorganisms
developing resistance to individual peptides.
Comparisons of species has shown that different frog
species have sets of homologous but different peptides that
have diversified in a species-specific manner. South American
and Australian hylids, as well as Indian, European, Asian
and North American ranins, have different patterns of
distribution with respect to geography, climate, vegetation
and habitats (aquatic, semiaquatic, terrestrial, arboreal,
torrential, rocky), some of them showing very unusual and
extreme adaptations [19]. The diversification of antimicrobial
peptides between species could thus be part of an optimum
evolutionary strategy developed during the radiations of
these frog species when microbial predators changed very
rapidly with shifts to new ecological niches. Focal hypermutation of the C-terminal antimicrobial-coding region of
preprodermaseptin genes might have evolved as a way of
increasing genetic diversity and so accelerating the adaptation of frogs to noxious microbial fauna.

Acknowledgements
This work was partly supported by the CNRS and the ÔProgramme de
Recherche Fondamentale en Microbiologie et Maladies Infectieuses et
Parasitaires (PRFMMIP)Õ.

References
1. Lazarus, L.H., Bryant, S.D., Cooper, F.S. & Salvadori, S.
(1999) What peptides these deltorphins be. Prog. Neurobiol. 57,
377–420.

2. Basir, Y.J., Knoop, F.C., Dulka, J. & Conlon, M.J. (2000) Multiple antimicrobial peptides and peptides related to bradykinin and
neuromedin N isolated from skin secretions of the pickerel frog,
Rana palustris. Biochim. Biophys. Acta 1543, 95–105.

´
3. Seon, A.A., Pierre, T.N., Redeker, V., Lacombe, C., Delfour, A.,
Nicolas, P. & Amiche, M. (2000) Isolation, structure, synthesis,
and activity of a new member of the calcitonin gene-related peptide family from frog skin and molecular cloning of its precursor.
J. Biol. Chem. 275, 5934–5940.
4. Salmon, A.L., Cross, L.J., Irvine, A.E., Lappin, T.R., Dathe, M.,
Krauser, G., Canning, P., Thim, L., Beyermann, M., Rothemund,
S., Bienert, M. & Shaw, C. (2001) Peptide leucine arginine, a
potent immunomodulatory peptide isolated and structurally
characterized from the skin of the Northern Leopard frog, Rana
pipiens. J. Biol. Chem. 276, 10145–10152.
5. Severini, C., Improta, G., Falconieri-Erspamer, V., Salvadori, S.
& Erspamer, V. (2002) The tachykinin peptide family. Pharmacol.
Rev. 54, 285–322.
6. Chen, T., Orr, D.F., Bjourson, A.J., McLean, S., O’Rourke, M.,
Hirst, D.G., Rao, P. & Shaw, C. (2002) Novel bradykinins and
their precursor cDNAs from European yellow-bellied toad
(Bombina variegata) skin. Eur. J. Biochem. 269, 4693–4700.
7. Doyle, J., Llewellyn, L.E., Brinkworth, C.S., Bowie, J.H., Wegener,
K.L., Rozek, T., Wabnitz, P.A., Wallace, J.C. & Tyler, M.J.
(2002) Amphibian peptides that inhibit neuronal nitric oxide
synthase. Isolation of lesuerin from the skin secretion of the
Australian Stony Creek frog Litoria lesueuri. Eur. J. Biochem. 269,
100–109.
8. Zasloff, M. (1987) Magainins, a class of antimicrobial peptides
from Xenopus skin: isolation, characterization of two active

forms, and partial cDNA sequence of a precursor. Proc. Natl
Acad. Sci. 84, 5449–5453.
9. Nicolas, P. & Mor, A. (1995) Peptides as weapons against
microorganisms in the chemical defense system of vertebrates.
Annu. Rev. Microbiol. 49, 277–304.
10. Simmaco, M., Mignogna, G. & Barra, D. (1998) Antimicrobial
peptides from amphibian skin: what do they tell us? Biopolymers
47, 435–450.
11. Zasloff, M. (2002) Antimicrobial peptides of multicellular organisms. Nature 415, 389–395.
12. Hwang, P.M. & Vogel, H.J. (1998) Structure-function relationships of antimicrobial peptides. Biochem. Cell. Biol. 76, 235–246.
13. 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.
14. Shai, Y. (2002) Mode of action of membrane active antimicrobial
peptides. Biopolymers 66, 236–248.
15. Maddox, M.W. & Longo, M.L. (2002) A Monte Carlo study of
peptide insertion into lipid bilayers: equilibrium conformations
and insertion mechanisms. Biophys. J. 82, 244–263.
16. Jacob, L. & Zasloff, M. (1994) Potential therapeutic applications
of magainins and other antimicrobial agents of animal origin. Ciba
Found. Symp 186, 197–223.
17. Zasloff, M. (2002) Antimicrobial peptides in health and disease.
New Engl. J. Med. 347, 1151–1160.
18. Tossi, A. (2003) A data base with more than 400 entries. Available
at />19. Duellman, W.E. & Trueb, L. (1994) Biology of Amphibians. John.
Hopkins University Press, London.
20. Goraya, J., Wang, Y., Li, Z., O’Flaherty, M., Knoop, F.C., Platz,
J.E. & Conlon, M.J. (2000) Peptides with antimicrobial activity
from four different families isolated from the skins of the North
American frogs Rana luteiventris, Rana berlandieri and Rana
pipiens. Eur. J. Biochem. 267, 894–900.

21. Isaacson, T., Soto, A.M., Iwanuro, S., Knoop, F.C. & Conlon,
J.M. (2000) Antimicrobial peptides with atypical structural features from the skin of the Japanese brown frog Rana japonica.
Peptides 23, 419–425.
22. Morikawa, N., Hagiwara, K. & Nakajima, T. (1992) Brevinin-1
and -2, unique antimicrobial peptides from the skin of the frog,


2080 D. Vanhoye et al. (Eur. J. Biochem. 270)

23.

24.

25.

26.

27.

28.

29.

30.

31.
32.

33.

34.

35.

36.

37.

38.

Rana brevipoda porsa. Biochem. Biophys. Res. Commun. 189,
184–190.
Simmaco, M., Mignogna, G., Barra, D. & Bossa, F. (1993) Novel
antimicrobial peptides from skin secretion of the European frog
Rana esculenta. FEBS Lett. 324, 159–161.
Simmaco, M., Mignogna, G., Barra, D. & Bossa, F. (1994)
Antimicrobial peptides from skin secretions of Rana esculenta.
Molecular cloning of cDNAs encoding esculentin and brevinins
and isolation of new active peptides. J. Biol. Chem. 269,
1956–1961.
Goraya, J., Knoop, F.C. & Conlon, M.J. (1998) Ranatuerins:
antimicrobial peptides isolated from the skin of the American
bullfrog, Rana catesbeiana. Biochem. Biophys. Res. Commun. 250,
598–592.
Clark, D.P., Durel, S., Maloy, W.L. & Zasloff, M. (1994)
Ranalexin, a novel antimicrobial peptide from bullfrog (Rana
catesbeiana) skin, structurally related to the bacterial antibiotic,
polymyxin. J. Biol. Chem. 269, 10849–10855.
Halverson, T., Basir, Y.J., Knoop, F.C. & Conlon, M.J. (2000)
Purification and characterization of antimicrobial peptides from

the skin of the North American green frog Rana clamitans. Peptides 21, 469–476.
Sai, P.S., Jagannadham, V.J., Vairamani, M., Raju, N.P., Devi,
A.A., Nagaraj, R. & Sitaram, N. (2001) Tigerinins: novel antimicrobial peptides from the Indian frog Rana tigerina. J. Biol.
Chem. 276, 2701–2707.
Simmaco, M., Mignogna, G., Cannofeni, S., Miele, R., Mangoni,
M.L. & Barra, D. (1996) Temporins, antimicrobial peptides from
the European red frog Rana temporaria. Eur. J. Biochem. 242,
788–792.
Mor, A., Nguyen, V.H., Delfour, A., Migliore, D. & Nicolas, P.
(1991) Isolation, amino acid sequence, and synthesis of dermaseptin, a novel antimicrobial peptide of amphibian skin. Biochemistry 30, 8824–8830.
Mor, A. & Nicolas, P. (1994) Isolation and structure of novel
defensive peptides from frog skin. Eur. J. Biochem. 219, 145–154.
Charpentier, S., Amiche, M., Mester, J., Vouille, V., Le Caer, J.P.,
Delfour, A. & Nicolas, P. (1998) Structure, synthesis, and molecular cloning of dermaseptins B, a family of skin peptide antibiotics. J. Biol. Chem. 273, 14690–14697.
Brand, G.D., Leite, J.R., Silva, L.P., Albuquerque, S., Prates,
M.V., Azevedo, R.B., Carregaro, V., Da Silva, J.S., Sa, V.C.,
Brandao, R.A. & Bloch, C. Jr (2002) Dermaseptins from Phyllomedusa oreades and Phyllomedusa distincta. Anti-Trypanosoma
cruzi activity without cytotoxicity to mammalian cells. J. Biol.
Chem. 277, 49332–49340.
´
Seon, A., Amiche, M. & Nicolas, P. (2000) Isolation of dermatoxin from frog skin, an antibacterial peptide encoded by a novel
member of the dermaseptin genes family. Eur. J. Biochem. 267,
370–378.
´
Amiche, M., Seon, A., Wroblewski, H. & Nicolas, P. (2000) Isolation of dermatoxin from frog skin, an antibacterial peptide
encoded by a novel member of the dermaseptin genes family. Eur.
J. Biochem. 267, 4583–4592.
Wechselberger, C. (1998) Cloning of cDNAs encoding new
peptides of the dermaseptin-family. Biochim. Biophys. Acta 1388,
279–283.

Amiche, M., Ducancel, F., Mor, A., Boulain, J.C., Menez, A. &
Nicolas, P. (1994) Precursors of vertebrate peptide antibiotics
dermaseptin b and adenoregulin have extensive sequence identities
with precursors of opioid peptides dermorphin, dermenkephalin,
and deltorphins. J. Biol. Chem. 269, 17847–17852.
Park, J.M., Lee, J.Y., Moon, H.M. & Lee, B.J. (1995) Molecular
cloning of cDNAs encoding precursors of frog skin antimicrobial
peptides from Rana rugosa. Biochim. Biophys. Acta 1264, 23–30.

Ó FEBS 2003
´
39. Amiche, M., Seon, A., Pierre, T.N. & Nicolas, P. (1999) The
dermaseptin precursors: a protein family with a common preproregion and a variable C-terminal antimicrobial domain. FEBS
Lett. 456, 352–356.
40. Vouille, V., Amiche, M. & Nicolas, P. (1997) Structure of genes
for dermaseptins B, antimicrobial peptides from frog skin.
Exon 1-encoded prepropeptide is conserved in genes for peptides of highly different structures and activities. FEBS Lett. 414,
27–32.
41. Kwon, S.Y., Carlson, B.A., Park, J.M. & Lee, B.J. (2000) Structural organization and expression of the gaegurin 4 gene of Rana
rugosa. Biochim. Biophys. Acta. 1492, 185–190.
42. Ota, T., Sitnikova, T. & Nei, M. (2000) Evolution of vertebrate
immunoglobulin variable gene segments. Curr. Top. Microbiol.
Immunol. 248, 221–245.
43. Hughes, A.L. (1997) Rapid evolution of immunoglobulin superfamily C2 domains expressed in immune system cells. Mol. Biol.
Evol. 14, 1–5.
44. Olivera, B.M., Walker, C., Cartier, G.E., Hooper, D., Santos,
A.D., Schoenfeld, R., Shetty, R., Watkins, P., Bandyopadhyay, P.
& Hillyard, D.R. (1999) Speciation of cone snails and
interspecific hyperdivergence of their venom peptides. Potential
evolutionary significance of introns. Ann. NY Acad. Sci. 870,

223–237.
45. Ohno, M., Menez, R., Ogawa, T., Danse, J.M., Shimohigashi, Y.,
Fromen, C., Ducancel, F., Zinn-Justin, S., Du Le, M.H., Boulain,
J.C., Tamiya, T. & Menez, A. (1998) Molecular evolution of snake
toxins: is the functional diversity of snake toxins associated with a
mechanism of accelerated evolution? Prog. Nucl Acid. Res. Mol.
Biol. 59, 307–364.
46. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular
cloning. A Laboratory Manual, 2nd edn. Cold Spring Harbor
laboratory, Cold Spring Harbor, NY.
47. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. &
Higgins, D.G. (1997) The CLUSTAL–windows interface: flexible
strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24, 4876–4882.
48. Kimura, M. (1980) A simple method for estimating evolutionary
rates of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16, 111–120.
49. Swofford, D.L. (1999) PAUP: Phylogenetic Analysis Using Parsimony and Other Methods, Version 4. Sinauer Associates, Sunderland, Mass.
50. Ina, Y. (1995) New methods for estimating the numbers of synonymous and nonsynonymous substitutions. J. Mol. Evol. 40,
190–226.
51. Jukes, T.H. & Cantor, C.R. (1969) Evolution of protein molecules., pp. 21–132. In Mammalian Protein Metabolism (H. N.
Munro. ed.). Academic Press, New York.
52. Bowie, J.H., Tyler, M.J., Chia, C.S.B., Carver, J.A. & Wallace,
J.C. (1998) The peptide arsenals of Australian Tree Frogs of the
genus Litoria. Chem. Australia. 65, 45–48.
53. Steinborner, S.T., Waugh, R.J., Bowie, J.H., Wallace, J.C., Tyler,
M.J. & Ramsay, S.L. (1997) New caerin antibacterial peptides
from the skin glands of the Australian tree frog Litoria xanthomera. J. Peptide Sci. 3, 181–185.
54. Rozek, T., Waugh, R.J., Steinborner, S.T., Bowie, J.H., Tyler,
M.J. & Wallace, J.C. (1998) The maculatin peptides from the skin
glands of the tree frog Litoria genimaculata: a comparison of the
structures and antibacterial activities of maculatin 1.1 and Caerin

1.1. J. Peptide Sci. 4, 111–115.
55. Wong, H., Bowie, J.H. & Carver, J.A. (1997) The solution structure and activity of caerin 1.1, an antimicrobial peptide from the
Australian green tree frog, Litoria splendida. Eur. J. Biochem. 247,
545–557.


Ó FEBS 2003

Evolution of amphibian antimicrobial peptides (Eur. J. Biochem. 270) 2081

56. Shai, Y. (1999) Mechanism of the binding, insertion and destabilization of phospholipid bilayer membranes by alpha-helical
antimicrobial and cell non-selective membrane-lytic peptides.
Biochim. Biophys. Acta. 1462, 55–70.
57. Segrest, J.P., De Loof, H., Dohlman, J.G., Brouillette, C.G. &
Anantharamaiah, G.M. (1990) Amphipathic helix motif: classes
and properties. Proteins 8, 103–117.
58. Frost, D.R. (2002) Amphibian Species of the World: an online
reference. Available at />amphibia/index.html.
59. Duda, T.F., Vanhoye, D. & Nicolas, P. (2002) Roles of diversifying selection and coordinated evolution in the evolution of
amphibian antimicrobial peptides. Mol. Biol. Evol. 19, 858–864.
60. Conticello, S.G., Gilad, Y., Avidan, N., Ben-Asher, E., Levy, Z. &
Fainzilber, M. (2001) Mechanisms for evolving hypervariability:
the case of conopeptides. Mol. Biol. Evol. 18, 120–131.
61. Radman, M. (1999) Enzymes of evolutionary change. Nature 401,
866–869.
62. Maor-shoshani, A., Reuven, N.B., Tomer, G. & Livneh, Z. (2000)
Highly mutagenic replication by DNA polymerase V (UmuC)
provides a mechanistic basis for SOS untargeted mutagenesis.
Proc. Natl Acad. Sci. 97, 565–570.
63. Savage, J. (1973) The geographic distribution of frogs: patterns

and predictions. In Evolutionary Biology of the Anurans (J.L. Vial,
ed.), pp. 352–445. University Missouri Press, Columbia.
64. Smith, A.G., Smith, D.G. & Funnell, B.M. (1994) Atlas of
Mezozoic and Cenozoic Coastlines. Cambridge University Press,
Cambridge.
65. Feller, A.E. & Hedges, S.B. (1998) Molecular evidence for the early
history of living amphibians. Mol. Phylogenetics Evol. 9, 509–516.

66. Ruvinsky, I. & Maxson, L.R. (1996) Phylogenetic relationships
among bufonoid frogs (Anura: Neobatrachia) inferred from
mitochondrial DNA sequences. Mol. Phylogenetics Evol. 5,
533547.
67. Bossuyt, F. & Milinkovitch, M.C. (2001) Amphibians as indicators
of early tertiary Ôout-of-IndiaÕ dispersal of vertebrates. Science 292,
93–95.
68. Springer, M.S., Westerman, M., Kavanagh, J.R., Burk, A.,
Woodburne, M.O., Kao, D.J. & Krajewski, C. (1998) The origin
of the Australasian marsupial fauna and the phylogenetic affinities
of the enigmatic monito del monte and marsupial mole. Proc. R.
Soc. Lond. B Biol. Sci. 265, 2381–2386.
69. Pethiyagoda, R. & Manamendra-Arachchi, K. (1998) Evaluating
Sri Lankas amphibian diversity. Occas. Pap. Wildlife Heritage
Trust, 2, 1.
70. Delfino, G., Brizzi, R., Alvarez, B.B. & Kracke-Berndorff, R.
(1998) Serous cutaneous glands in Phyllomedusa hypochondrialis
(Anura, Hylidae): secretory patterns during ontogenesis. Tissue
Cell. 30, 30–40.
71. Lacombe, C., Cifuentes-Dias, C., Dunia, I., Auber-Thomay, M.,
Nicolas, P. & Amiche, M. (2000) Peptide secretion in the cutaneous glands of South American tree frog Phyllomedusa bicolor:
an ultrastructural study. Eur. J. Cell. Biol. 79, 631–641.

72. Mor, A., Hani, K. & Nicolas, P. (1994) The vertebrate
peptide antibiotics dermaseptins have overlapping structural features but target specific microorganisms. J. Biol. Chem. 269,
31635–31640.
73. Schiffer, M. & Edmundson, A. (1967) Use of helical wheels to
represent the structures of proteins and to identify segments with
helical potential. Biophys. J. 7, 121–135.