Báo cáo khoa học: Purification and characterization of three isoforms of chrysophsin, a novel antimicrobial peptide in the gills of the red sea bream, Chrysophrys major doc

The synthetic peptides displayed broad-spectrum bactericidal activity against Gram-negative and Gram-positive bacteria including Escherichia coli, Bacillus subtilis, and fish and crustace

Purification and characterization of three isoforms of chrysophsin,

a novel antimicrobial peptide in the gills of the red sea bream,

Chrysophrys major

Noriaki Iijima1, Norio Tanimoto1, Yohko Emoto1, Yohko Morita1, Kazumasa Uematsu2,

Tomoya Murakami3and Toshihiro Nakai4

1

Laboratory of Molecular Cell Biology,2Laboratory of Fish Physiology and4Laboratory of Fish Pathology, Graduate School of Biosphere Science, Hiroshima University, Japan;3Hiroshima Fisheries Experimental Station, Ondo, Aki, Japan

We report here the isolation of three isoforms of a novel

C-terminally amidated peptide from the gills of red sea

bream, Chrysophrys (Pagrus) major Peptide sequences

were determined by a combination of Edman

degrada-tion, MS and HPLC analysis of native and synthetic

peptides Three peptides, named chrysophsin-1,

chryso-phsin-2, and chrysophsin-3, consist of 25, 25, and 20

amino acids, respectively, and are highly cationic,

con-taining an unusual C-terminal RRRH sequence The

a-helical structures of the three chrysophsin peptides were

predicted from their secondary structures and were

confirmed by CD spectroscopy The synthetic peptides

displayed broad-spectrum bactericidal activity against

Gram-negative and Gram-positive bacteria including

Escherichia coli, Bacillus subtilis, and fish and crustacean pathogens The three peptides were also hemolytic Immunohistochemical analysis showed that chrysophsins were localized in certain epithelial cells lining the surface

of secondary lamellae and eosinophilic granule cell-like cells at the base of the secondary lamellae in red sea bream gills Their broad ranging bactericidal activities, combined with their localization in certain cells and eosinophilic granule cell-like cells in the gills, suggest that chrysophsins play a significant role in the innate defense system of red sea bream gills

Keywords: antimicrobial peptide; chrysophsin; gills; red sea bream; synthetic peptide

Antimicrobial peptides are widely distributed throughout

the animal and plant kingdoms [1] They display a broad

spectrum of antimicrobial activity against bacteria, yeast

and filamentous fungi, and are recognized as an essential

component in the first line of the host defense system [2,3]

There are three major sites by which bacteria enter fish:

the gills, gastrointestinal tract and skin [4] Therefore,

antibacterial substances are thought to exist at these sites to

prevent penetration of bacteria into the circulatory system

In fact, skin mucus, eggs and serum of fish contain a variety

of nonspecific defense substances, such as lysozyme,

com-plement, C-reactive protein, transferrin, lectin, and

antimi-crobial proteins [5–10] Furthermore, antimiantimi-crobial peptides

have been purified from fish skin mucus: pardaxin from the

moses sole fish Pardachirus marmoratus [11], pleurocidin

from the winter flounder Pleuronectes americanus [12] and parasin I from the catfish Parasilurus asotus [13], and the gene expression of pleurocidin-like antimicrobial peptides found in the skin and intestine of the winter flounder [14] The antimicrobial peptide, misgurin, has also been purified from the whole body of the loach Misgurnus anguillicau-datus [15] These antibacterial peptides showpotent antimicrobial activity against negative and Gram-positive bacteria and act as nonspecific defense substances in fish skin

Fish gills are constantly being flushed with water that may contain fish pathogens, but are covered with only a thin layer of protective mucus and are constructed of only

a single layer of fragile cells that separate the vascular system from the external environment Thus, they are a very important site of pathogen penetration Therefore, potent antimicrobial peptides can be expected to be found

in fish gills to prevent such penetration However, there is

a paucity of information on nonspecific defense systems in the gills An antimicrobial peptide has been identified in the gills of only one fish species, hybrid striped bass (Morone saxatilis· M chrysops) [16–18] In this study, we therefore tried to purify antimicrobial peptides from the gills of the red sea bream Chrysophrys (formerly Pagrus) major and found that novel antimicrobial peptides, chrysophsin-1, chrysophsin-2, and chrysophsin-3, are localized in the eosinophilic granule cell-like cells of the gills They exhibited potent bactericidal activity against Gram-negative and Gram-positive pathogens of fish and crustaceans

Correspondence to N Iijima, Laboratory of Molecular Cell Biology,

Graduate School of Biosphere Science, Hiroshima University,

1-4-4 Kagamiyama, Higashihiroshima 739-8528, Japan.

Fax: + 81 824 22 7059, Tel.: + 81 824 24 7949,

E-mail: noriiij@jpc.hiroshima-u.ac.jp

Abbreviations: EGC, eosinophilic granule cell; ESI-ITMS,

electro-spray ionization/ion trap mass spectrometry; MLC, minimal lethal

concentration; mcKLH, Imject maleimide-activated mariculture

keyhole limpet hemocyanin; NaCl/P i , 50 m M phosphate buffer

(pH 7.4) containing 0.14 M NaCl; Tris/NaCl, 20 m M Tris/HCl

(pH 7.4) containing 150 m M NaCl.

(Received 2 August 2002, revised 29 November 2002,

accepted 9 December 2002)

Materials and methods

Chemicals

Fmoc-L-amino acids, Fmoc-L-amino acid resins and

TentaGel S (TGS)-RAM were purchased from Shimadzu

(Kyoto, Japan) Chemicals for peptide synthesis,

trifluoro-acetic acid, ethylmethylsulfide, ethanedithiol, thiophenol,

2-methylindole, thioanisole, phenol, and anisole were

obtained from commercial sources and were of the highest

purity available Lysyl endopeptidase, a silver staining II kit

and 2,2,2-trifluoroethanol were purchased from Wako Pure

Chemicals (Tokyo, Japan) A TSKgel G2000SW column

was obtained from Tosoh (Tokyo, Japan), an Inertsil C8-3

column from GL Science (Tokyo, Japan), and a Capcellpak

C18 column from Shiseido (Tokyo, Japan) SP-Sephadex

C-25 was purchased from Pharmacia Biotech (Uppsala,

Sweden), and PolySulfoethyl Aspartamide column was

from Poly LC Inc (Columbia, MD, USA) Melittin and

BSA were obtained from ICN Biomedicals Inc (Aurora,

OH, USA), Imject maleimide activated mariculture keyhole

limpet hemocyanin (mcKLH) and Imject Alum were from

Pierce (Rockford, IL, USA), Simple Stain MAX-PO

(Multi) and Simple Stain DAB solution were from Nichirei

(Tokyo, Japan), peroxidase-labeled affinity-purified

anti-body to rabbit IgG (H + L) was from KPL (Gaithersburg,

MD, USA), DC-protein assay kit was from Bio-Rad

(Hercules, CA, USA), and marine broth 2216 was from

DIFCO (Detroit, MI, USA)

Fish

Red sea bream weighing 1.05–1.14 kg (n¼ 11) were

cultured by a commercial supplier (Mihara, Hiroshima

Prefecture, Japan) After starvation for 1 day, the fish

were killed by stabbing the brain with a knife, and gill

filaments were immediately collected, frozen in liquid

nitrogen, and stored at)80 C until use Red sea bream

weighing 109–305 g (n¼ 5) were obtained from the

Hiroshima Fisheries experimental station (Ondo, Aki,

Japan) The gill arches including gill lamellae were

immediately fixed in Bouin’s solution and transported

on ice to the laboratory for immunohistochemical

exami-nation as described below

Purification of antimicrobial peptides from the gills

The purification procedure is summarized in Fig 1

Frozen gill filaments were crushed in liquid nitrogen

The resulting gill powder (12 g) was boiled in 120 mL

water for 10 min After cooling, extractions were

per-formed by adding 120 mL 2M HCl, 10% (v/v) formic

acid, 2% (w/v) NaCl, and 1% (v/v) trifluoroacetic acid

followed by vortex-mixing for 1–2 min The homogenate

was centrifuged, and the supernatant adjusted to pH 4 by

adding 1MTris; it was then filtered The resulting filtrate

(219 mL) was used as the acid extract and was applied to

a Sep-Pak C18 cartridge (Waters, Milford, MA, USA)

After a wash with 0.1% trifluoroacetic acid, the peptide

was eluted with 80% acetonitrile/0.1% trifluoroacetic

acid The dried eluate was dissolved in 1M acetic acid

and then adsorbed on SP-Sephadex C-25 resin Successive elution with 1M acetic acid, 2M pyridine and 2M pyridine/acetic acid (pH 5.0) afforded three respective fractions of SP-1, SP-2 and SP-3 The SP-3 fraction was further lyophilized and dissolved in 40% acetonitrile containing 0.1% trifluoroacetic acid An aliquot of the solution was loaded on a TSKgel G2000SW column (7.6· 600 mm) equipped with a Tosoh HPLC pump (CCPM) and a UV-VIS detector (UV-8010) and was eluted with 40% acetonitrile containing 0.1% trifluoro-acetic acid The SP-3 fraction was repeatedly injected, and fraction A, estimated to be less than 5 kDa, was pooled and subjected to RP-HPLC

The on-line HPLC separation was performed on a Hewlett-Packard HP1100 series HPLC system equipped with an auto sampler, thermostatically controlled column compartment, UV-VIS detector, and degasser Solvent A was 5% acetonitrile containing 0.1% trifluoroacetic acid, and solvent B was 80% acetonitrile containing 0.085% trifluoroacetic acid The freeze-dried fraction A was recon-stituted with solvent A and subjected to RP-HPLC on an Inertsil C8-3 column (4.6· 150 mm) The gradient was 0–2 min 0% solvent B, 2–5 min 0–20% solvent B, 5–55 min 20–47% solvent B, and 55–80 min 47–100% solvent B The effluents were separated into two directions, and the flow rate of one direction was adjusted to 0.6 mLÆmin)1 and that of the other to 0.2 mLÆmin)1 The effluent from one direction (0.6 mLÆmin)1) was fractionated (0.6 mL each fraction) The effluent from the other direction

Fig 1 Flow chart of the purification procedure of antimicrobial peptides from the gills of red sea bream.

(0.2 mLÆmin)1) was introduced into a mass spectrometer as

described below The resulting active fractions were

lyophi-lized and reconstituted in 5 mMKH2PO4/H3PO4(pH 3.0)

containing 25% acetonitrile, and further loaded on a

PolySulfoethyl Aspartamide column (4.6· 200 mm)

equipped with a Tosoh HPLC pump The column was

eluted with a linear gradient of KCl The gradient was

0–3 min 0–0.1M KCl, 3–43 min 0.1–0.4M KCl, and

43–48 min 0.4–1MKCl

Mass spectrometry

MS analysis of peptides was performed with a Finnigan

LCQ ion trap mass spectrometer (ThermoQuest, San Jose,

CA, USA) equipped with an electrospray ionization source

(ESI-ITMS) as described previously [19] The mass scale

was calibrated using Ultra-mark provided by the

manufac-turer Ions were detected and analyzed in the positive mode

on the basis of their m/z ratio

Solid-phase peptide synthesis

The protected peptide chain was assembled with a

Shimadzu peptide synthesizer (PSSM8) according to

standard Fmoc chemistry [20] Non-amidated peptides

were prepared with Fmoc-L-His (Trt)-resin, and amidated

peptides with TGS-RAM as described previously [21]

Magainin 2 [22] was prepared with Fmoc-L-Ser

(tBu)-resin, and peptide-Cys, to which cysteine was added at the

carboxy end, was prepared with Fmoc-L-Cys (Trt)-resin

At the end of the synthesis, the peptides were freed from

the resin by cleavage with cocktail A (82.5%

trifluoro-acetic acid, 3% ethylmethylsulfide, 5% H2O, 2.5%

ethanedithiol, 3% thiophenol, and 5 mgÆmL)1

2-methyl-indole) for peptides with Trp and Arg, cocktail B (82%

trifluoroacetic acid, 2% ethylmethylsulfide, 5% H2O, 5%

thioanisole, 3% ethanedithiol, and 2% phenol) for

peptides with Arg, or cocktail C (94% trifluoroacetic

acid, 5% anisole, and 1% ethanedithiol) for peptides

without Arg and Trp The resulting peptides were

precipitated with cold diethyl ether, lyophilized, and then

purified using a Tosoh HPLC apparatus, equipped with a

Capcellpak C18 column (10· 250 mm), eluted with a

linear gradient of acetonitrile/0.1% trifluoroacetic acid

Aliquots of peptides purified by RP-HPLC were

charac-terized with ESI-ITMS as described below

Lysyl endopeptidase digestion

The three purified peptides, P-1 (2.9 lg), P-2 (0.6 lg) and

P-3 (0.46 lg), dissolved in 20 lL 20 mMTris/HCl (pH 8.0)

were mixed with lysyl endopeptidase and incubated for

2–4 h at 37C The ratio of enzyme to peptide was 1 : 100

for P-1 and 1 : 5 for P-2 and P-3 After digestion, the

reaction was stopped by adding 0.1% trifluoroacetic acid at

a final volume of 0.1 mL

Amino-acid sequence analysis

The amino-acid sequence of the three purified peptides, P-1,

P-2 and P-3 (1.5–2.9 lg), was determined on a Hewlett–

Packard G1005A protein sequencing system (Palo Alto,

CA, USA) by analyzing the data calibrated with 10 pmol phenylthiohydantoin amino-acid standards

Tricine/SDS/PAGE The molecular mass of the sample was estimated by Tricine/ SDS/PAGE using a 16.5% separating gel, 10% spacer gel and 4% stacking gel in the presence of 2-mercaptoethanol [23], and the protein/peptide bands were stained with a silver staining II kit from Wako

Determination of peptide concentration The concentrations of the samples from purification steps were measured with the DC-protein assay kit using BSA as standard The concentration of purified and synthetic chrysophsin-1 and chrysophsin-2 was obtained from the

A280[24], and that of chrysophsin-3 with the DC-protein assay kit

Circular dichroism

CD spectra were recorded on a Jasco J-500CH instrument (Tokyo, Japan) at room temperature in a 1-mm path length cell Synthetic crysophsins 1, 2 and 3 were dissolved in

20 mM potassium phosphate buffer (pH 7.25) containing

8 mM EDTA or 20 mM potassium phosphate buffer (pH 7.25) containing 8 mM EDTA and 50% trifluoroeth-anol The concentration of the chrysophsins was 20 lM Curves were smoothed by the algorithm provided by Jasco, and data analysis was performed as described previously [25] CD measurements were reported as Q (degreesÆ

cm2Ædmol)1) The relative helix content was deduced as described by McLean et al [26] as follow s

%helix ¼  100ðH222 þ 3000Þ=33000 whereQ222is the CD at 222 nm

Bactericidal assay Bactericidal activity was routinely tested using Bacillus subtilisATCC6633 or Escherichia coli WP-2 After growth

in tryptic soy broth at 37C to exponential phase, bacteria were washed twice with 0.85% NaCl and diluted in 50 mM Hepes/NaOH buffer (pH 7.4) to give  2 · 105 colony-forming unitsÆmL)1(CFUÆmL)1) Aliquots (25–100 lL) of the fractions at each purification step from the acid extract

of the gills were lyophilized and dissolved in 0.1 mL 1 mM citric acid/sodium citrate buffer (pH 4.0) Solutions were mixed with an equal volume of bacterial suspension and incubated at 37C for 60 min In a control experiment, the cells were incubated with the same solvent as used for the preparation of each fraction After appropriate dilution of the mixture with 50 mM Hepes/NaOH buffer (pH 7.4), 0.1 mL aliquots were spread on tryptic soy agar plates and incubated for 14–18 h to measure the number of colonies formed The bactericidal activity was expressed as killing (%), using the formula [27]:

Killingð%Þ ¼ ½ð1  CFU in the sampleÞ=

CFU in the control 100

Serial doubling dilutions of the three native and synthetic

chrysophsins were carried out following the protocol

described above, and the minimal lethal concentration

(MLC) against E coli and B subtilis was defined as the

lowest peptide concentration that caused 99% killing of

the bacteria [28] All assays were performed in duplicate

Fish and crustacean pathogens, Lactococcus garvieae

YT-3, Streptococcus iniae F-8502, Aeromonas hydrophila

ET-4, Edwardsiella tarda ET-82016, Vibrio anguillarum

ATCC 19264, Vibrio vulnificus ET-7617 and Pseudomonas

putidaATCC 12633 were grown in tryptic soy broth at 25C,

and Aeromonas salmonicida NCMB 1102 was grown in

tryptic soy broth at 20C Vibrio harveyi HUFP 9111 and

Vibrio penaeicidaKH-1 were grown in marine broth 2216 at

25C After growing, exponential phase bacterial cultures

w ere w ashed and diluted in 50 mM Hepes/NaOH buffer

(pH 7.4) containing 2% NaCl to give 2 · 105CFUÆmL)1

The bacterial suspensions (each 0.1 mL) were incubated at

20–25C for 1 h with equal volumes of twofold serial

dilutions of the three synthetic chrysophsins in 1 mMcitric

acid/sodium citrate buffer (pH 4.0) containing 2% NaCl

After 100-fold dilution of the mixture, 0.1 mL aliquots were

spread on tryptic soy agar plates and incubated at 20–25C

for 18–24 h Then, MLC was determined as described above

Hemolytic assay

Before use, freshly collected human blood was washed with

50 mMphosphate buffer (pH 7.4) containing 0.14MNaCl

(NaCl/Pi) until the supernatant was colorless A suspension

was made of 1% packed cells in NaCl/Picontaining 2%

glucose Synthetic chrysophsin 1, 2 or 3, melittin or synthetic

magainin 2 was dissolved in 50% dimethyl sulfoxide at a

concentration of 1 mMand was serially diluted with NaCl/

Pi From this suspension, 90 lL aliquots were incubated

with 10 lL synthetic chrysophsin 1, 2 or 3, melittin or

synthetic magainin 2 at different concentrations As a

positive control (100% lysis), a 0.1% solution of SDS was

used After incubation for 30 min at 37C, the sample

was centrifuged at 900 g for 10 min The supernatant was

diluted 20-fold with NaCl/Pi, and the absorbance was

determined at 405 nm in a Shimadzu UV-VIS

spectropho-tometer (UV mini 1240) To control for dimethyl sulfoxide

in the peptide, erythrocyte suspensions were incubated with

50% dimethyl sulfoxide

Preparation of anti-(chrysophsin 1) IgG

Synthetic chrysophsin-1-Cys (2.3 mg) purified by

RP-HPLC was dissolved in 60% dimethyl sulfoxide and

conjugated with 2.3 mg mcKLH according to the

manu-facturer’s directions The resulting

chrysophsin-1-Cys-mcKLH conjugate (chrysophsin-1-KLH) was dialyzed

against NaCl/Piand used as described below For initial

immunization, 600 lg chrysophsin-1-KLH in 500 lL

NaCl/Piwas mixed with an equal volume of Imject Alum

and administered to Japanese white rabbits by multiple

subcutaneous injections Ten, 21 and 33 days later, the

rabbit received 600 lg chrysophsin-1-KLH in 500 lL

NaCl/Pi by subcutaneous injection At 4 days after the

final injection, blood was collected, and the resulting

antiserum was stored at)80 C The titer of the antisera

was determined by ELISA using chrysophsin-1-Cys as an antigen and peroxidase-labeled affinity-purified antibody to rabbit IgG (H + L) as a secondary antibody

Chrysophsin-1-Toyopearl was used for the purification of anti-(chrysophsin-1) IgG Chrysophsin-1-Cys (1 mg) was coupled to Toyoperal AF-Epoxy-650M (1 g) according to the manufacturer’s protocol Anti-(chrysophsin-1) serum (3 mL) was filtered through a 0.45-lm filter (Dismic-13, Advantec, Tokyo, Japan) and applied to the chrysophsin-1-Toyopearl column The column was washed with 20 mM Tris/HCl (pH 7.4) containing 150 mM NaCl (Tris/NaCl) and 20 mM Tris/HCl (pH 7.5) containing 1M NaCl and 1% Triton X-100, respectively, and specific antibody was eluted from the column with 0.1M glycine/HCl (pH 2.5) The eluted antibody was immediately neutralized with 1M Tris and stored at)80 C

Immunohistochemistry Gill arches were fixed in Bouin’s solution for 24 h at room temperature After dehydration, tissues were embedded in paraffin, cut into 5-lm sections, and mounted on gelatin-coated glass slides Immunohistochemical staining was performed using Simple Stain MAX-PO (Multi) as a secondary antibody [29] Briefly, sections were deparaffi-nized in xylene and rehydrated in decreasing concentrations

of ethanol After a brief wash with NaCl/Pi, sections were exposed to 3% H2O2in 90% methanol to block endogenous peroxidase and washed with NaCl/Pi They were then blocked with 10% normal goat serum in NaCl/Pifor 2 h at room temperature, incubated with the

chrysophsin-1-speci-fic antibody (0.11 lgÆmL)1) in NaCl/Picontaining 1% BSA and 5 mM NaN3 overnight at room temperature, and washed 3 times with NaCl/Pi The sections were incubated with the secondary antibody for 30 min at room tempera-ture and washed with NaCl/Pi The color was developed in a Simple Stain DAB solution To determine the type of immunoreactive cells in the gills, neighboring 5-lm serial sections of the gills were stained with chrysophsin-1-specific antibody or hematoxylin and eosin As a negative control, chrysophsin-1-specific antibody preabsorbed with chryso-phsin-1 (mol ratio of antibody to chrysophsin-1¼

1 : 5000) was used as a primary antibody

Results

Peptide purification and primary structure After fractionation of the acid-extracted gill powder on SP-sephadex C-25, fraction SP-3 showed the most bactericidal activity (MLC 1.2 lgÆmL)1) and was further separated by gel-filtration HPLC (Fig 2) Bactericidal activities against

B subtiliswere detected in almost all of the fractions eluted from the column The molecular mass of fractions eluted between 48 and 81min, showing high antibacterial activity, was determined by Tricine/SDS/PAGE As this study focuses exclusively on bactericidal peptides of less than

5 kDa which were mainly detected in fraction A (data not shown), it was subjected to RP-HPLC followed by ESI-ITMS (Fig 3) The deduced mean ± SD molecular masses

of the three peaks, P1, P2 and P3, showing strong bactericidal activity were 2891.2 ± 0.2, 2919.3 ± 0.4 and

2285.7 ± 0.2 Da, respectively By RP-HPLC, P1, P2 and

P3 were pooled As a broad but single band was observed by

Tricine/SDS/PAGE (data not shown), aliquots of P1, P2

and P3 were directly sequenced by Edman degradation

The amino-acid sequences were then partially determined

(Table 1) As the underlined residues could not be

deter-mined, they were digested with lysyl endopeptidase, and the

resulting peptide fragments subjected to RP-HPLC

fol-lowed by ESI-ITMS From the difference between the

deduced mean molecular mass of the peptide fragment by ESI-ITMS and the theoretical mean molecular mass of the identified amino-acid sequence of the peptide fragment, the unidentified amino acid was determined as His For example, three peptide fragments, (1) FFGWLIK, (2) GAIXAGK, and (3) AIXGLIXRRRX, were released from P1 by lysyl endopeptidase The mean molecular mass (909.6 ± 0.2 Da) of fragment 1 determined by ESI-ITMS was very similar to that of fragment 1 calculated from the identified amino-acid sequence (910.10 Da) The difference

in the mean molecular mass of fragment 2 determined by ESI-ITMS (652.5 ± 0.0 Da) and that of fragment 2 calculated from the identified amino-acid sequence (515.61 Da) was 137.1 Da, which agreed with the average molecular mass of His (Table 1) On replacing the unknown amino acids of fragment 3 with His, the theoretical average molecular mass (1365.62 Da) of fragment 3 calculated from the amino-acid sequence was almost the same as that of fragment 3 deduced by ESI-ITMS (1364.8 Da) Finally, amino-acid sequences of P1, P2 and P3 were determined as follows P1, FFGWLIKGAIHAGKAIHGLIHRRRH; P2, FFGWLIRGAIHAGKAIHGLIHRRRH; P3, FIG LLISAGKAIHDLIRRRH The theoretical average molecular masses of P1 (2892.43 Da), P2 (2920.45 Da) and P3 (2286.76 Da) matched the deduced average mole-cular masses with a difference of 1 Da This discrepancy (1 Da) suggests amidation of the C-terminal histidine Therefore, elution profiles of synthetic nonamidated pep-tides named P1-COOH, P2-COOH and P3-COOH, and synthetic amidated peptides named P1-CONH2, P2-CONH2 and P3-CONH2 were superimposed on those of native P1, P2 and P3 by HPLC Native P1 and P3 were recognized as single peaks by RP-HPLC (data not shown) and also by anion-exchange HPLC (Fig 4A,C) P1-COOH and P1-CONH2 could not be separated by RP-HPLC; however, they were clearly separated into two peaks at retention times of 18 and 22 min (Fig 5A), and the mixture of native P1and P1-CONH2 showed a single peak on ion-exchange HPLC (Fig 5B) Thus, the C-terminal amino acid His of native P1 was found to be

Fig 2 Gel-filtration HPLCof the SP-3 fraction obtained from the acid

extract of red sea bream gills The SP-3 fraction eluted with 2 M

pyri-dine/acetic acid (pH 5.0) from SP-sephadex C-25 was loaded on a

TSKgel G2000SW column pre-equilibrated with 40% acetonitrile/

0.1% trifluoroacetic acid The flowrate w as 0.25 mLÆmin)1, and

absorbance was monitored at 220 nm (solid line) The fraction volume

was 0.5 mL The bactericidal activity against B subtilis of a 100-lL

aliquot of each fraction was measured and expressed as killing (%)

(open bar) as described in Materials and methods Fraction A,

indi-cated by the bar, was collected.

Fig 3 RP-HPLCof fraction A obtained by gel-filtration HPLC Fraction A obtained by gel-filtration HPLC was loaded on an Inertsil C8-3 column and eluted with a linear gradient of acetonitrile in aqueous trifluoroacetic acid (dotted line) The flow rate was 0.8 mLÆmin)1, and the absorbance was monitored at 220 nm (solid line) The effluent was separated into two directions, and the effluent from one direction was collected The bactericidal activity of each fraction was expressed as killing (%) (open bar) as described in Materials and methods The effluent from the other direction was directly introduced into an electrospray ionization/mass spectrometer (ESI/ITMS).

amidated Native P3 was also found to be amidated

(Figs 4C and 5F) Native P2 was detected as a single peak

by RP-HPLC (data not shown) However, native P2

separated into two peaks, P2-1 and P2-2 (Fig 4B), and

P2-2 was coeluted with P2-CONH2 on anion-exchange

HPLC (Fig 5C,D) P2-1 and P2-2 were separated by

anion-exchange HPLC and further analyzed by RP-HPLC

followed by ESI-ITMS As the mean molecular mass of

2 (2919.3 ± 0.4 Da) was identical with that of

P2-CONH2, P2-2 was also an amidated peptide Two peptides,

2925 ± 0.7 Da and 2907 ± 0.3 Da, were detected in the

P2-1 fraction From these results, the primary structures of three novel bactericidal peptides, P1, P2-2 and P3, were determined, and we decided to call them chrysophsin-1, chrysophsin-2 and chrysophsin-3 based on the genus of red sea bream, C major An alignment of these three peptides with other antimicrobial peptides is shown in Fig 6 Chrysophsin-1, chrysophsin-2 and chrysophsin-3 are C-terminally amidated, 25, 25 and 20 amino acids in length, and rich in cationic residues (9/25, 9/25 and 6/20, respect-ively) Chrysophsin-2 corresponds to the chrysophsin-1 isoform differing by a single residue at position 7 (lysine or arginine) Interestingly, the characteristic C-terminal cat-ionic tetrapeptide, RRRH, is conserved in chrysophsins, in addition to the C-terminal amidation

Secondary structure of chrysophsins Schiffer–Edmunson helical wheel modeling was used to predict hydrophobic and hydrophilic regions within the secondary structure of chrysophsin-1, chrysophsin-2 and chrysophsin-3 (Fig 7) All three had an amphipathic a-helix conformation, indicating hydrophilic and hydro-phobic residues on opposite sides of the N-terminal 21 residues for chrysophsin-1 and chrysophsin-2, and 18 residues for chrysophsin-3 This conformation was con-firmed by CD spectroscopy of the three synthetic chry-sophsin peptides in the absence or presence of 50% trifluoroethanol (Fig 8) In the absence of 50% trifluoro-ethanol, the spectra of all three chrysophsins are typical

of a disordered structure However, their ellipiticities decreased at both 208 and 222 nm and increased at

190 nm, indicating stabilization of the a-helix structure (87%, 88% and 81% helix content in chrysophsin-1, chrysophsin-2 and chrysophsin-3, respectively) in the presence of 50% trifluoroethanol

Bactericidal spectrum, salt sensitivity and hemolytic activity of chrysophsins

As the bactericidal activities of the three native and synthetic chrysophsins were found to be almost equal against

B subtilis and E coli (Table 2), bactericidal activity against fish and crustacean pathogens was investigated with the synthetic chrysophsins (Table 3) All three were active against Gram-positive bacteria (MLC < 10 lM) Most of the Gram-negative pathogens were sensitive to less than 40 lM, except for A hydrophila and E tarda

Table 1 Mass measurement of peptide fragments of P1, P2 and P3.

Partial sequence

ESI-ITMS

M r Peptide fragment

ESI-ITMS

M r

Calc.

M r His (·)

PI FFGWLIKGAIXAGKAIXGLIXRRRX 2891.2 FFGWLIK 909.6 910.10

GAIXAGK 652.5 515.61 137.1 AIXGLIXRRRX 1364.8 954.20 3 ·137.1 P2 FFGWLIRGAIXAGKAIXGLIXRRRX 2919.3 FFGWLIRGAIXAGK 1573.0 1435.71 137.1

AIXGLIXRRRX 1364.8 954.20 3 ·137.1 P3 FIGLLISAGKAIXDLIRRRX 2285.7 FIGLLISAGK 1017.8 1018.26 –

AIXDLIRRRX 1285.3 1012.24 2 · 137.1

Fig 4 Cation-exchange HPLC of P1, P2 and P3 Aliquots of P1 (A),

P2 (B) and P3 (C) were loaded on a PolySulfoethyl Aspartamide

col-umn and eluted with a linear gradient of KCl in 5 m M KH 2 PO 4 /H 3 PO 4

(pH 3.0)/25% acetonitrile at a flowrate of 0.8 mLÆmin)1(dotted line).

Absorbance was monitored at 220 nm (solid line).

(MLC > 40 lM) P putida was sensitive to chrysophsin-1

and chrysophsin-3, but not to chrysophsin-2

(MLC > 40 lM)

To determine whether the bactericidal action of

chry-sophsins is dependent on salt, NaCl concentration was

varied with the chrysophsin concentration kept fixed

(Fig 9) Chrysophsin-1 and chrysophsin-2 were bactericidal

up to 0.32MNaCl, but chrysophsin-3 was effective only up

to 0.16MNaCl

Chrysophsins were hemolytic for human red blood cells, but they were less hemolytic than melittin and more hemolytic than magainin (Fig 10)

Fig 5 Cation-exchange HPLC of native P1, P2-2 and P3, and their synthetic peptides Mixtures of synthetic P1-COOH and P1-CONH 2 (A), native P1 and P1-CONH 2 (B), P2-COOH and P2-CONH 2 (C), native P2 and P2-CONH 2 (D), P3-COOH and P3-CONH 2 (E), and native P3 and P3-CONH 2 (F) were loaded on a PolySulfoethyl Aspartamide column and eluted with a linear gradient of KCl in 5 m M KH 2 PO 4 /H 3 PO 4 (pH 3.0) containing 25% acetonitrile at a flowrate of 0.8 mLÆmin)1(dotted line) Absorbance was monitored at 220 nm (solid line).

Fig 6 Comparison of the amino-acid sequence of the chrysophsins with other antimicrobial peptides Alignment of the mature amino-acid sequences

of chrysophsin-1, chrysophsin-2 and chrysophsin-3, misgurin [15], pleurocidin [12] and pleurocidin-like antimicrobial peptides/WF2-4 [14], piscidin

1 (sb-moronecidin), 2 (wb-moronecidin) and 3 [16,17] and melittin [22] Gaps have been introduced to maximize sequence similarities Identical amino-acid residues are shaded, and basic amino-acid residues are shown in bold.

Localization of chrysophsin-1 in the gills of red sea

bream

Immunohistochemical staining showed that chrysophsin-1

was localized in the gills of red sea bream as two distinct cell

populations The first type of immunopositive cells were

found at the base of the secondary lamellae (Fig 11A,

arrows in Fig 11C and Fig 11G), and were located

adjacent to the blood capillaries in the connective tissue

(Fig 11D,H) In the immunopositive cells, cytoplasmic

granules were immunostained (arrows in Fig 11E,I) and

were eosinophilic in nature (arrows in Fig 11F) The second

type of immunoreactive cell was an epithelial cell on the secondary lamellae (Fig 11A and arrowheads in Fig 11C)

No immunoreactivity was observed in any sections stained with chrysophsin-1-specific antibody preabsorbed with chrysophsin-1 (Fig 11B)

Discussion

Here we report the isolation of chrysophsin-1, chryso-phsin-2 and chrysophsin-3, novel peptides of 25, 25 and

20 residues with bactericidal and hemolytic activity, from the gills of the red sea bream, C major They are highly

Fig 7 Schiffer–Edmunson helical wheel diagram demonstrating probable amphipathic a-helical conformation of chrysophsin-1, chrysophsin-2 and chrysophsin-3 Shaded gray indicates hydrophobic amino acids Residue numbers starting from the N-terminus are shown.

cationic peptides without cysteine (Fig 6) Searches of

sequence databases show70% identity between

chryso-phsin-1 and the mature peptide sequence predicted from

the nucleotide sequence of winter flounder pleurocidin-like

genomic clone (WF3), which has not yet been purified as

a mature peptide [14] However, chrysophsin-1 shows low

identity (24–36%) with other fish antimicrobial peptides,

such as piscidins [16], moronecidins [17] and pleurocidin

[12,30] The C-terminal amino acid was amidated in all

three chrysophsins, similarly to those in marine animals, such as solitary ascidians (styelin D) [31] and hybrid striped bass (Morone saxatilis x M chrysops) (moroneci-din/piscidin) [16,17] It has been proposed that amphi-pathic a-helical peptides showantimicrobial activity by interacting electrostatically with the anionic bacterial membrane, adopting an amphipathic a-helical conforma-tion that allows them to insert the hydrophobic face into the lipid bilayers and form a pore [1,8,22] The amphi-pathic a-helical structure of chrysophsins was predicted by the Schiffer–Edmundson wheel analysis (Fig 7) All three chrysophsins form an a-helical structure (> 80% helical content) in the structure-forming solvent, trifluoroethanol, but not in phosphate buffer (Fig 8) This suggests that random-coiled chrysophsins in the water environment will form amphipathic a-helical conformation after contacting with the bacterial membrane Thus, chrysophsins will showantimicrobial activity in a similar way to other previously studied amphipathic a-helical antimicrobial peptides Interestingly, the identity in amino-acid sequence between chrysophsin and misgurin is low (16%), but chrysophsins and misgurin have strongly cationic tetra-peptide sequences, RRRH and RRRK, at the C-terminus Bee venom melittin also has a C-terminal cationic tetrapeptide sequence, KRKRQQ, and was found to form a nonhelical hydrophilic domain that allows elec-trostatic binding with the polar head group of negatively charged phospholipids [32–34] Thus, the C-terminal RRRH of chrysophsins could form a nonhelical hydro-philic domain similarly to bee venom melittin

All three chrysophsins showed bactericidal activity against Gram-negative and Gram-positive pathogens in the presence of 0.34MNaCl, except for A hydrophila and

E tarda The results indicate that chrysophsins show broad-spectrum bactericidal activity against pathogenic bacteria up to 0.34M NaCl The bactericidal activity of chrysophsin-1 and chrysophsin-2 at a concentration of 0.5 lM was retained at physiological NaCl concentrations against E coli, but was lost at 0.64M NaCl, similarly to winter flounder pleurocidin: the bactericidal activity of pleurocidin (5.6 lM) against E coli was also lost at 0.625M NaCl [12] On the other hand, hybrid striped bass wb-moronecidin retained bacteriostatic activity against Staphylococcus aureus even in the presence of 1.28M NaCl; however, it remains unclear whether wb-moronecidin is bactericidal or not in 1.28M NaCl [17] The net charge of chrysophsin-1 (pI 12.64) and chrysophsin-2 (pI 12.79) is slightly higher than that of wb-moronecidin (piscidin 2) (pI 12.60) In addition, the C-terminal amino acids of chrysophsins and wb-moro-necidin were both amidated These findings may indicate that the difference in salt tolerance between chrysophsins and wb-moronecidin is partly due to the difference in bacteria, Gram-negative E coli and Gram-positive

S aureus used in the experiment Antimicrobial peptides must pass the lipopolysaccharide-rich external leaflet of the outer membrane to interact with the inner membrane

of Gram-negative bacteria such as E coli; however, they can directly interact with the anionic cytoplasmic mem-brane of Gram-positive bacteria such as S aureus It is necessary to compare the bacteriostatic and bactericidal activities of chrysophsins against Gram-positive bacteria,

Fig 8 CD spectrum for synthetic chrysophsins The spectra of

syn-thetic chrysophsin-1 (A), chrysophsin-2 (B) and chrysophsin-3 (C)

were obtained in 20 m M potassium phosphate buffer/8 m M EDTA,

pH 7.25, in the presence (solid line) or absence (dotted line) of 50%

(v/v) trifluoroethanol.

in addition to the Gram-negative bacteria Chrysophsins

showed hemolytic activity against human erythrocytes, in

addition to bactericidal activity (Fig 10); however, they

are less hemolytic than melittin, a cytotoxic peptide, but

more hemolytic than magainin It is still necessary to

examine whether chrysophsins are cytotoxic to the gill

cells of red sea bream or not Chrysophsin-3 was the least hemolytic and bactericidal This may correlate with the lower isoelectric point (pI 12.16) than that of chrysophsin-1 and chrysophsin-2

We used polyclonal antibody against chrysophsin-1 as a primary antibody for immunohistochemistry (Fig 11)

Table 3 Bactericidal activity (minimal lethal concentration) of synthetic chrysophsin-1, -2 and -3 against pathogenic bacteria Minimal lethal con-centrations of peptide are given in l M and are the concentration of substance required necessary to kill  99% of the bacteria Strains were considered resistant (R) when their growth was not inhibited by peptide up to 40 lM.

Strains of pathogenic bacteria Chrysophsin-1 Chrysophsin-2 Chrysophsin-3 Gram-positive bacteria

Grain-negative bacteria

Table 2 Bactericidal activity (minimal lethal concentration) of native and synthetic chrysophsins compared with that of magainin-2 Minimal lethal concentrations of peptide are given in l M and are the concentration of substance required necessary to kill  99% of the bacteria.

Chrysophsin-1 Chrysophsin-2 Chrysophsin-3 Strains of bacteria Native Synthetic Native Synthetic Native Synthetic Magainin2 Bacillus subtilis ATCC 6633 0.25 0.125 0.25 0.25 0.25 0.25 0.985 Escherichia coli WT-2 0.25 0.25 0.25 0.25 0.25 0.25 0.985

Fig 9 Effect of NaCl on bactericidal activity of chrysophsins

Bacte-ricidal activity of synthetic chrysophsin-1 (s), chrysophsin-2 (h) and

chrysophsin-3 (n) at 0.5 l M was determined with various

concentra-tions of NaCl ranging from 0 to 1.28

Fig 10 Hemolytic activity of chrysophsins Synthetic chrysophsin-1 (s), chrysophsin-2 (h) and chrysophsin-3 (n), synthetic magainin 2 (j), an antimicrobial peptide from the aquatic frog Xenopus laevis and melittin (d), a peptide from bee venom cytotoxic to human erythro-cytes, were incubated with a 1% suspension of washed human eryth-rocytes for 30 min at 37 C.