Peptides 30 (2009) 2191–2199

Characterization of a new bioactive peptide from Potamotrygon gr. orbignyifreshwater stingray venom

Katia Conceicao a, Juliane M. Santos b, Fernanda M. Bruni a, Clecio F. Klitzke a, Elineide E. Marques b,Marcia H. Borges c, Robson L. Melo a, Jorge H. Fernandez d, Monica Lopes-Ferreira a,*a Laboratorio Especial de Toxinologia Aplicada, Instituto Butantan, Avenida Vital Brazil 1500, Sao Paulo, SP 05503-900, Brazilb Nucleo de Estudos do Ambiente, Universidade Federal de Tocantins, Tocantins, Brazilc FUNED – Fundacao Ezequiel Dias, Belo Horizonte, MG, Brazild CBB – Universidade Estadual Norte Fluminense – UENF, Campo dos Goytacazes, RJ, Brazil

A R T I C L E I N F O

Article history:

Received 8 June 2009

Received in revised form 30 July 2009

Accepted 4 August 2009

Available online 12 August 2009

Keywords:

Potamotrygon gr. orbignyi

Venom

HPLC

Mass spectrometry

Bioactive peptide

Inflammation

Molecular dynamics

Porflan

A B S T R A C T

Brazilian freshwater stingrays, Potamotrygon gr. orbigyni, are relatively common in the middle-western

regions of Brazil, where they are considered an important public health threat. In order to identify some of

their naturally occurring toxin peptides available in very low amounts, we combine analytical protocols

such as reversed-phase high-performance liquid chromatography (RP-HPLC), followed by a biological

microcirculatory screening and mass spectrometry analysis. Using this approach, one bioactive peptide

was identified and characterized, and two analogues were synthesized. The natural peptide named Porflan

has the primary structure ESIVRPPPVEAKVEETPE (MW 2006.09 Da) and has no similarity with any

bioactive peptide or protein found in public data banks. Bioassay protocols characterized peptides as

presenting potent activity in a microcirculatory environment. The primary sequences and bioassay results,

including interactions with the membrane phospholipids, suggest that these toxins are a new class of fish

toxins, directly involved in the inflammatory processes of a stingray sting.

� 2009 Elsevier Inc. All rights reserved.

Contents lists available at ScienceDirect

Peptides

journa l homepage: www.e lsev ier .com/ locate /pept ides

1. Introduction

Toxins have proved to be valuable tools for understandingmolecular mechanisms involved in physiological processes, andtheir study has been important in the development of new drugs.Moreover, knowledge of toxin function promotes understandingand treating victims injured by venomous animals. Fish venomcontains a diversity of toxins yet to be discovered, since so far,mainly proteinaceous toxins have been purified from fish venoms[4,6,7,20,26].

Among the venomous fish of medical importance found inBrazil’s freshwater, we call attention to the freshwater stingrays,Potamotrygon gr. orbigyni. These animals cause severe injuries inhumans [11] and are common in the middle-western regions ofBrazil’s freshwater environments. Our group showed that thevenom of P. gr. orbignyi can induce edematogenic and nociceptiveresponses, necrosis in mice, presence of enzymatic activity as wellan increasing the intensity of leukocyte rolling [21].

* Corresponding author. Tel.: +55 11 3726 1024.

E-mail address: lopesferreira@butantan.gov.br (M. Lopes-Ferreira).

0196-9781/$ – see front matter � 2009 Elsevier Inc. All rights reserved.

doi:10.1016/j.peptides.2009.08.004

Considering the fact that relatively little is known aboutbioactive peptides in fish venom, this work aimed to betterunderstand this venom and its compounds, purifying toxins toallow for pharmacological investigations. Accordingly, we report ahigh number of peptides from P. gr orbignyi venom and describedthe first bioactive peptide called Orpotrin [8].

In this study, a novel bioactive peptide Porflan was identifiedand characterized from the venom glands that cover the stings of P.

gr. orbignyi. This peptide presented no similarity with any knownprotein or peptide and showed an increase in the number ofleukocyte rolling detected by intravital microscopy. In an effort todevelop new and verify low molecular mass peptides withinflammatory activity, we generated two synthetic peptideanalogues of Porflan, namely, Porflan-N and Porflan-C. Thesetwo decapeptides are basically composed of the same amino acidsequence as the N-terminal and C-terminal 10-residue fragment ofnative Porflan. Additionally, as an important step to investigate themembrane-targeting of Porflan, we investigated Porflan interac-tions with the membrane phospholipids. In silico modeling andsimulation of DPPC residues as bio-membrane is a simplified formof simulation largely used to access the physico-chemical proper-ties of small peptides on cytoplasmatic membranes and obtain

K. Conceicao et al. / Peptides 30 (2009) 2191–21992192

structural insights of these peptides in contact with biologicalmembranes. In our specific case, this experimentation providedmolecular details on peptide structural rearrangements, mem-brane insertion and interaction with the hydrophobic portion ofthe membrane. The mechanism action of Porflan and theiranalogues with respect to its membrane-perturbation effect willbe discussed.

2. Materials and methods

2.1. Animals

Groups of 4 Swiss mice weighing 18–22 g were used through-out. The animals, provided by Instituto Butantan, were inaccordance with the guidelines provided by the Brazilian Collegeof Animal Experimentation (no 550/08).

2.2. Reagents

Dithiothreitol (DTT), a-cyano-4-hydroxycinnamic acid (a-CHCA), sinapic acid, iodoacetamide (IAA), NaI, formic acid andcleavage reagents were purchased from Sigma–Aldrich (St. Louis,MO, USA). Phosphoric acid was purchased from Merck (Darmstadt,Germany). The H-Pro-2-ClTrt resin and all Fmoc l-amino acidswere purchased from Novabiochem (Nottingham, UK). All solventswere of analytical grade.

2.3. Collection of the venom

Specimens of P. gr. orbignyi were collected on the TocantinsRiver and Parana River both in the state of Tocantins, Brazil, andtransferred immediately to the laboratory to extract the venomThe epithelium that cover the sting obtained from the animalswere scratched and dissolved in deionized water and immediatelycentrifuged at 6000 � g for 15 min. The supernatant was pre-purified using a Centricon filter devices (Millipore Bedford, USA).

2.4. Venom and peptide purification

To perform the venom reversed-phase chromatography, anAkta binary HPLC system (Amersham Biosciences, Uppsala,Sweden) was used. The filtered supernatant venom was loadedin a ACE C18 column (4.6 mm � 250 mm, 5 mm) eluted with agradient formed by trifluoroacetic acid (TFA)/water (A – 1:1000) toTFA/acetonitrile (Merck-Darmstadt, Germany) (ACN)/water (B –1:900:100). The peptides were eluted at a constant flow rate of1.0 mL/min with a 10–80% gradient of solvent B over 40 min. Thecolumn eluates were monitored by their UV absorbance at 214 nm.For peptide purification, further additional chromatographic stepswere necessary, using the same C18 column with optimizedgradients.

2.5. Mass spectrometry

The fraction and peptides were analyzed on a matrix-assistedlaser desorption/ionization time-of-flight mass spectrometer(MALDI-TOF MS) Ettan MALDI-TOF/Pro system (Amersham Bios-ciences, Uppsala, Sweden) in reflectron mode calibrated with P14R([M+H]+ 1533.8582) and angiotensin II ([M+H]+ 1046.5423)(Sigma). MALDI-TOF spectra were obtained in the positive ionmode at an acceleration voltage of 20 kV. The matrix solutioncontained saturated a-cyano-4-hydroxycinnamic acid (Sigma) in50% acetonitrile/0.1% TFA. For the electrospray analysis, weperformed direct injection in a LC–MS Surveyor MSQ Plus (ThermoElectron, USA) under positive ionization mode. The needle andcone potential were set to 3.1 kV and 40 V, respectively. The

aqueous sample solutions (10 mL) were directly injected at a50 mL/min constant flow rate of acetonitrile H2O/0.1% formic acid(1:1). External calibration was performed with NaI (Sigma) over m/z 100–2000.

2.5.1. MS/MS analysis

Mass spectrometric peptide sequencing was carried out inpositive ionization mode on a Q-TOF Ultima API fitted with anelectrospray ion source (Micromass, Manchester, UK). Purifiedlyophilized peptide sample was dissolved into 50% acetonitrilecontaining 0.1% formic acid, and directly infused into theinstrument at 1 mL/min, constant flow rate. The instrumentcontrol and data acquisition were done in a MassLynx 4.1 datasystem (Waters, MA, USA) and experiments were performed byscanning from a mass-to-charge ratio (m/z) of 50–1800 using ascan time of 1 s. External calibration was performed withPhosphoric acid (Sigma). For the MS/MS analysis, collision energyranged from 18 to 45 eV and the precursor ions were selectedunder a 1 m/z window. The peptide sequence was analyzedmanually and by the software using the MS/MS spectra.

2.6. Edman degradation

To determine the amino acid sequence, HPLC purified samplesof the native peptide were subjected to Edman degradation using aShimadzu PPSQ-21 automated protein sequencer, followingmanufacturer’s instructions.

2.7. Peptide synthesis

In addition to the natural occurring peptide Porflan, twoanalogues were synthesized, Porflan-N (ESIVRPPPVE) and Porflan-C (VEAKVEETPE). The peptides were synthesized by solid-phasepeptide synthesis in a H-Pro-2-ClTrt (Novabiochem, Nottingham,UK) resin using the Fmoc strategy in an automated bench-topsimultaneous multiple solid-phase synthesizer (PSSM 8 systemfrom Shimadzu Co.) [1]. Couplings were performed with O-(Benzotriazol-1-yl)-N,N,N0,N0-tetramethyluronium tetrafluorobo-rate (TBTU)/1-hydroxybenzotriazole in N,N-dimethylformamidefor 60 min. The synthesized peptides were cleaved from the resinby adding a solution of TFA/thioanisole/1,2-ethanedithiol/phenol/water (82.5:5:2.5:5:5) at room temperature for 8 h. The peptideproducts removed from resin were diluted in water/acetonitrile(9:1) and lyophilized. Synthetic peptides were purified bypreparative RP-HPLC (Shimadzu LC10 VP Series, Kyoto, Japan)using a semi-preparative Shimadzu C18 column (Shim-pack Prep-ODS, 20 mm � 250 mm, 5 mm). A gradient 5–45% of acetonitrilecontaining TFA 0.1% (solvent B) in 35 min was applied, wheresolvent A was ultra-pure H2O containing TFA 0.1%. The purity andidentity of the peptides were confirmed by MALDI-TOF massspectrometry (as above described) and by analytical RP-HPLC intwo different solvent systems. The purified fractions containingpeptides were pooled and lyophilized.

2.8. Intravital microscopy

The dynamics of alterations in the microcirculatory networkwere determined using intravital microscopy by trans-illumina-tion of the mice cremaster muscle after topical application of thefraction and the investigated peptides dissolved in 20 mL of salinesterile. Administration of the same amount of sterile saline (20 mL)was used as a control. Before beginning the experiments, micewere injected with a muscle relaxant drug (0.4% Xilazin)(Coopazine1, Schering-Plough) and then anaesthetized with0.2 g/kg chloral hydrate and the cremaster muscle was exposedfor microscopic examination in situ as previously described [2,18].

K. Conceicao et al. / Peptides 30 (2009) 2191–2199 2193

The animals were maintained on a special designed boardcontrolled thermostatically at 37 8C, which included a transparentplatform on which the trans-illuminated tissue was placed. Afterstabilization of the microcirculation, the number of roller cells andadherent leukocytes in the post-capillary venules were counted10 min after peptide injection. The number of adherent leukocyteswas determined off-line during playback of videotaped images. Aleukocyte was considered to be adhering to the venularendothelium if it remained stationary for 30 s or longer. Thenumber of adherent leukocytes was quantified per 100 mm lengthof venule. A rolling leukocyte was defined as a white cell thatmoved slower than the stream of flowing erythrocytes. Thenumber of rolling leukocytes was quantified as the number ofwhite cells that passed a fixed point (25 mm window) on thecomputer monitor. Arteriolar diameter and velocity were alsomeasured in these experiments 10 min after peptide injection. Thestudy of the microvascular system of trans-illuminated tissue wasperformed using an optical microscope (Axiolab, Carl Zeiss,Oberkochen, Germany) coupled to a photographic camera (Coolpix990, Nikon, Melville, NY, USA) using a 10/025 longitudinal distanceobjective/numeric aperture and 1.6 optovar. All of the determina-tions mentioned above were performed in triplicate.

2.9. Molecular simulation experiments and system set-up

To study the conformational space of Porflan, moleculardynamics simulations were carried out with the program Gromacs3.3.3 [17] on a quad-core Linux workstation using the GROMOS9653a5 force field [22]. Initial molecular model of Porflan in extendedconformation was obtained using a Swiss-PDB Viewer [10] andsolvated in 12.000 SPC/E water molecules, representing more thanthree solvation layers in periodic boundary conditions [3].Standard protonation states (7.0 pH) for all residues were usedand Na+ and Cl� ions were used (10 mM) for system neutralization.Simulation systems were submitted to a steepest-decent (SD)energy minimization, and minimized structures were used inunrestrained molecular dynamic for 10 ns with berendsen-typetemperature (310 K) and pressure (1 atm) coupling in a NVT, usingthe PME method of electrostatic treatment [9,12]. Productiondynamics were carried out in the same conditions for 200–225 ns.

Fig. 1. HPLC chromatogram of P. gr. orbignyi venom. The peptide elution was conducted u

(B – 1:900:100) at a constant flow rate of 1.0 mL/min with a 10–80% gradient of solve

In Porflan–DPPC membrane interaction studies, a systemformed by Porflan peptide and DPPC bilayer (128 molecules).System set-up, solvation, protonation states and all dynamicparameters were as described above, but production dynamicswere carried out for 25–50 ns.

Cluster analysis of the obtained trajectory was performedusing g_cluster tool with a 2.5 A cutoff. The most stableclusters in the MD simulation were selected for an unrestrainedmultiple step conjugate-gradient (CG) minimization process(0.1 kJ mol�1 nm�1) for further analysis. A GROMACS 3.3.3 packagewas used in all plot and matrix calculations.

2.10. Statistical analysis

All results were presented as means � SEM of at least fouranimals in each group. Differences among data were determined byOne Way Analysis of Variance (ANOVA) followed by Dunnett’s test.Differences between two means were determined using unpairedStudent’s t-test. Data were considered significant at p < 0.05.

3. Results

3.1. Purification and characterization of venom peptide

The venom from P. gr. orbignyi was purified and the fractionswere manually collected. The chromatographic separation ofvenom by RP-HPLC presents several fractions. Excluding the firstfraction, which contained the Orpotrin peptide [8], all fractionswere analyzed by intravital microscopy, and the fraction eluted atRt 19.12 min (Fig. 1) presented a significant effect on micro-circulation. After one additional chromatography step, the peptidewas pure.

The purification of the peptide was analyzed by MALDI-TOFmass spectrometry, and the isotopic distribution of chargesrevealed a monoisotopic mass of 2006.091 m/z (Fig. 2A). By ESI-MS, we confirmed the presence of a single peptide as observed inFig. 2B, where 1003.81 m/z is in agreement with the doublecharged peptide. We therefore processed the sample in order toassess its cysteine content. The 2006 Da peptide was reduced withDTT and alkylated IAA, desalted and subjected to MALDI-TOF-MS

sing trifluoroacetic acid (TFA)/water (A – 1:1000) and TFA/Acetonitrile (ACN)/water

nt B over 40 min. UV absorbance was monitored at 214 nm. Arrow: Porflan.

Fig. 2. (A) MALDI-TOF mass spectrum of Porflan (2006.091 m/z), presenting the monoisotopic mass of the peptide; (B) ESI-MS spectrum of double charged peptide

(1003.81 m/z).

K. Conceicao et al. / Peptides 30 (2009) 2191–21992194

Fig. 3. Representative MS/MS spectra of the peptide (Porflan), in a Q-TOF Ultima API (Micromass). This profile corresponds to the b and y series. The peptide sequence using

one-letter code following the b and y series orientation is shown on the top of the graph.

K. Conceicao et al. / Peptides 30 (2009) 2191–2199 2195

analysis. Presenting no alteration in molecular mass clearlydemonstrated that this molecule has no cysteine residue (datanot shown).

3.2. Porflan sequencing

The selected active peptide was purified and submitted tosequencing. The purified peptide was individually selected for MS/MS analyses and fragmented by collision with argon (CID), yieldingan ion spectrum as shown in Fig. 3. The MS/MS acquired spectrawere analyzed by the BioLynx – PepSeq software module ofMassLynx 4.1 and manually verified for accuracy in the evaluationof daughter ion attributions and amino acid sequence interpreta-tion. The peptide was fully sequenced and identified as one novelpeptide whose amino acid sequence is ESIVRPPPVEAKVEETPE. Thesearch for similarities and/or homologies with other knownpolypeptides was performed with BLAST from the NCBI (http://www.ncbi.nlm.nih.gov/), which revealed no similarity with anyknown toxic peptide or protein. Since this peptide constitutes anovel toxin described for first time in animal venom, we namedthem by taking into consideration the scientific name for stingraysas Porflan. Moreover, Edman degradation was performed toconclude the deduced amino acid sequence obtained by massspectrometry. Porflan and analogues were chemically synthesized,analyzed by RP-HPLC and MALDI mass spectrometry (Table 1).

3.3. Porflan increases leukocyte rolling

To investigate the potential of the peptide fraction, Porflan andtheir analogues in leukocyte rolling and other activities observed in

Table 1Molecular mass of the natural and analogues of Porflan.

Peptides Sequence

Porflan (Natural) ESIVRPPPVEAKVEETPE

Porflan-N ESIVRPPPVE

Porflan-C VEAKVEETPE

a Observed by MALDI-MS.

the microcirculatory environment, the intravital microscopy of thecremaster muscle of mice was performed. Topical application ofthe fractions, peptides (diluted in 20 mL saline), and controls werecarried out for up to 30 min (Fig. 4). A few rolling leukocytes wereobserved in the microcirculation of control mice (dark bars). Afterthe first 10 min of topical application of 20 mM, the syntheticPorflan peptide induced a significant increase in the number ofrolling leukocytes in the post-capillary venule, and this valueremained elevated up to 20 min of the experimental period (Fig. 4Aand B). At 30 min, slight decrease in the number of leukocyte couldbe seen, in which some leukocytes adhered to the endothelium(Fig. 4A – inset). Topical application of 20 mM Porflan-N alsoproduced an increase in the number of rolling leukocytes in post-capillary venules where at 20 and 30 min of experiment the valuesobserved were higher compared to Porflan. Neither of the peptideschanged the vascular permeability or capillary diameter (data notshown). Interestingly, Porflan-C did not induce an increase in theleukocyte rolling but exhibited toxic activity. Immediately aftertopic application of the peptide (1 min), the presence of abundantthrombi of different sizes was observed in the venules (Fig. 4B), andafter 5 min, stasis was observed in post-capillary venules (data notshown). No alteration in arteriolar or capillary diameter wasobserved at any time after administration of the peptides.

3.4. Simulation of Porflan structure fluctuation in aqueous solution

Fluctuation of the Porflan structure was studied using moleculardynamics simulation in aqueous solution. A simulation system wascomposed of a Porflan molecule solvated in 12.000 water molecules.The initial 10–20 ns of dynamics were used for system equilibration,

Mw (observed)a Mw (theoretical)

2006.091 2006.04

1122.54 1122.61

1130.41 1130.56

Fig. 4. Intravital micrographs of cremaster muscle. (A) Representative graph showing leukocyte rolling in different times. Insets: leukocyte venular rolling after 30 min of

Porflan administration (arrow); leukocytes adherent to endothelium after 30 min of Porflan-N (brackets); (B) intravital micrographs of cremaster muscle before (Normal) and

after 10 min of topical application of 20 ml (20 mM) of Porflan, Porflan-N and Porflan-C. Photographs were obtained from digitized images on the computer monitor.

K. Conceicao et al. / Peptides 30 (2009) 2191–21992196

and fluctuation of the peptide structure was analyzed by clusteranalysis over the 190–200 ns of the production dynamics trajectory(Fig. 5A). After the 100 ns simulation, a stable structure fluctuationwas detected for 45–50 ns, and the most representative structure ofthis cluster (Fig. 5A and C) was used as the most stable structure ofPorflan peptide in solution (Fig. 5A and B). A central hydrophobic corewas noted in this structure surrounded by hydrophilic and dynamicN- and C-terminal residues. The central disposition of compacted Proresidues surrounded by hydrophobic residues (I V R P P P V sequence)adopts an open a-helix-like rigid structure (Fig. 5B and C). Anotherimportant characteristic was the high mobility of the negative-charged C-terminal portion of the peptide (E E T P E sequence).

Despite the presence of a ‘‘stable’’ structure in 35% of theobtained trajectory, the large conformational space explored byPorflan in solution indicates a very dynamic structure in solution.All these data together suggest an amphypathic Porflan structurewith more dynamic C-terminal electronegative residues, despitethe hydrophobic and rigid central sequence of proline residues.

3.5. Molecular dynamic simulations of Porflan interactions with DPPC

membranes

As the Porflan structure suggests amphypathic characteristics,molecular simulations were carried out to study the interactions of

Porflan and their analogues with a DPPC bilayer (128 molecules ofDPPC, Fig. 6). In these experiments an almost immediateinteraction was observed between the negative-charged C-terminal sequence of the peptide with the hydrophilic portionof the DPPC membrane (composed of choline groups, phosphateions and glycerol). The electrostatic interaction of the carboxylicgroups of Porflan Glu residues with the amine groups from DPPCoccurred since the first nanosecond of dynamic simulation andpeptide was in total interaction with the membrane of DPPC at thefifth ns (Fig. 6A). After 10 ns of simulation, the Porflan structurewas stable and close to that observed in water solution,maintaining the arrangement of the central hydrophobic core asan open a-helix and the negative-charged C-terminal residues intight contact with the DPPC choline groups. The same results wereobtained for the simulations of N-terminal portion of the Porflan(Porflan-N) and for the C-terminal portion of the (Porflan-C)structure (Fig. 6B). Therefore, the complete interaction of the allpeptides with the hydrophilic layer of the membrane occurs after10 ns of simulation, when the peptide adopts a stable conforma-tion to the end of simulations.

The results of our simulation experiments clearly demonstratethat only physic-chemical properties of Porflan and Porflan-C alonedo not assist these peptides in crossing biological membranes, andthe presence of cell active transport and extracellular domains

Fig. 5. Simulation of Porflan structure in solution. (A) Most probable structure of Porflan peptide in solution and (B) representation of electrostatic and hydrophobic properties

on the peptide surface; (C) fluctuation of the peptide structure analyzed on the 225 ns of dynamics simulation trajectory. Structure types were represented in color code and

the most populated cluster of structures was highlighted on the square. (For interpretation of the references to color in this figure legend, the reader is referred to the web

version of the article.)

K. Conceicao et al. / Peptides 30 (2009) 2191–2199 2197

appears to be an important target for physiological activity of thesepeptides.

4. Discussion

Certain species of venomous fish are endemic to most ocean andfreshwater environments of Brazil, frequently causing injuries tohumans, and the stingray P. gr. orbignyi is an example of such fish.In this paper, we demonstrated the existence of another novelbioactive peptide, named Porflan in the venom of P. gr orbignyi, andthe role of two analogues.

Porflan and Porflan-N increase the number of rolling leukocytesin post-capillary venules. In mice, the venom of P. gr. orbignyi

induced a striking augmentation of leukocyte rolling and adherentcells in the endothelium of cremaster [21], demonstrating the pro-inflammatory effects of the crude P. gr. orbignyi venom. Themechanism by which the venom induces the recruitment ofleukocytes has not been determined, and now may be associatedwith presence of Porflan and other toxins mediating this event.Zarbock and Ley [27] describes that the first contact of neutrophilswith the endothelium is mediated by selectins and their counter-receptors, followed by rolling of neutrophils along the endothelialwall of post-capillary venules and integrin-mediated arrest [5,16].During leukocyte interaction with the endothelium, leukocytesreceive signals from P-selectin glycoprotein ligand (PSGL)-1, L-selectin, G-protein-coupled receptors (GPCRs), and integrins[15,25]. While rolling, neutrophils receive different inflammatorysignals that can activate several pathways including the expressionof integrins, and leading to adhesion and extravasation of rollingleukocytes [24], this last event was not observed in ourexperiments. Our results support the hypothesis that Porflan

and Porflan-N may up-regulate the expression of importantmolecules such as P- and L-selectins in endothelium for leukocyterecruitment and rolling.

Porflan-N (1ESIVRPPPVE10) induced immediate inflammatoryreactions that reached maximum intensities after 20–30 min,demonstrating the most powerful pro-inflammatory ability andincreasing the number of leukocytes. In contrast, the novel Porflaninduced a smaller increase in the number of rolling leukocytescompared to Porflan-N. The above data suggest that Porflan andPorflan-N can induce and modulate the expression of molecules onthe surface of the endothelial cells. However, for Porflan-C(9VEAKVEETPE18), the result was the induction of fibrin depositscharacterizing the release of thrombi and stasis of venules. In thiscase the observed effects can be assigned exclusively to thealteration of only one parameter since preliminary studies with anew analogue, with the replacement of three amino acids in thesequence, did not present this toxic activity.

Porflan sequence shows no similarity with any bioactivepeptide or protein found in public databanks but presents a PPPsequence on the N-terminal part. Molecular simulation experi-ments showed that this particular PPP sequence confer conforma-tional rigidity and an exposed hydrophobic surface in the peptide(Figs. 5 and 6). Molecular dynamics of Porflan in solution and incontact with the DPPC bilayer, as well as CD results (data notshown), shows that, polyproline sequence tend to adopt a rigidopen a-helix conformation. These results suggest that thepresence of a stable helical structure at the N-terminus is necessarybut not sufficient condition for enhancing bioactivity. The PPP helixis an unusual structure presenting an easily accessible hydro-phobic surface, as well as a good hydrogen-bonding site [14].In biological membranes proline-rich motifs are particularly

Fig. 6. Molecular simulation experiments of Porflan (A) and Porflan-N and Porflan-C (B) in contact with DPPC bilayer. Water box was represented in gray and all the other

residues were represented in CPK code and thin (DPPC) and thick (peptides) sticks. Progressive snapshots were tacked from the obtained trajectories at 1, 5, 10 and 15 or 20 ns

of dynamics using VMD (www.ks.uiuc.edu/Research/vmd). At the bottom, the relative density of each compound in the simulating system was calculated over the trajectory

and represented in graphical mode, and representing the movement of the peptide center of mass from the water layer to the DPPC bilayer over the simulation trajectories.

K. Conceicao et al. / Peptides 30 (2009) 2191–21992198

recognized by several domains such as SH3 and WW [19] link cellsignaling to the membrane cytoskeleton [13]. As molecularsimulations in the peptide–DPPC system suggest, Porflan andPorflan-N may use active carriers to interact with proline-richrecognizing domains in the cytoplasmatic part of the membrane.Preliminary molecular modeling shows that Porflan-N peptidesmay interact with WW domains, showing excellent rates for ‘‘insilico’’ calculated DG of complex formation (data not shown)compared with other canonical motifs recognized by this domain[23]. Nevertheless, considering these molecular simulation results,the proper characterization of these carriers or domain targets is inthe central scope of our future investigations.

Moreover, considering these findings, it is clear that thesevenoms represent a vast source of novel peptide toxins that mayprove useful in research. In this paper, we have demonstrated theexistence of a novel pro-inflammatory peptide, Porflan in thevenom of P. gr orbignyi, and the role of two analogues. Porflan waspotent in promoting leukocyte recruitment and adhesion in themicrocirculatory environment, and this activity should be parti-cularly connected to the N-terminus sequence and polyprolinesequence of this peptide. The biological roles of pro-inflammatorysubstances such Porflan in the venom of this stingray have not yetbeen clarified. However, it is probable that Porflan has served thefish well throughout their long history. The unique structure of

K. Conceicao et al. / Peptides 30 (2009) 2191–2199 2199

Porflan and their analogues as well proper dose response assays areimportant and may help to contribute to the understanding of thestructural/functional relationships that have evolved in theproduction of this peptide.

Acknowledgments

We are grateful to NEAMB (Nucleus of Environmental Studies)for supporting the fieldwork and Central Analitica IQ-USP. Thiswork was supported by the Fundacao de Amparo a Pesquisa doEstado de Sao Paulo (FAPESP), CNPq and E-26/110.806/2008FAPERJ grant.

References

[1] Atherton E, Sheppard RC. Solid phase peptide synthesis—a practical approach.Oxford: IRL Press; 1989.

[2] Baez S. An open cremaster muscle preparation for the study of blood vessels byin vivo microscopy. Microvasc Res 1973;5:384–96.

[3] Berendsen HJC, Grigera JR, Straatsman TP. The missing term in effective pairpotentials. J Chem Phys 1987;91:6269–71.

[4] Breton P, Delamanche I, Buee J, Goudey-Perriere F, Perriere C. Evidence for aneurotoxic activity in crude venom of the stonefish (Synanceia verrucosa). J NatToxins 2002;11:305–13.

[5] Butcher EC. Leukocyte-endothelial cell recognition: three (or more) steps tospecificity and diversity. Cell 1991;67:1033–6.

[6] Church JE, Hodgson WC. Dose-dependent cardiovascular and neuromusculareffects of stonefish (Synanceja trachynis) venom. Toxicon 2000;38:391–407.

[7] Church JE, Hodgson WC. The pharmacological activity of fish venoms. Toxicon2002;40:1083–93.

[8] Conceicao K, Konno K, Melo RL, Marques EE, Hiruma-Lima CA, Lima C, et al.Orpotrin: a novel vasoconstrictor peptide from the venom of the Brazilianstingray Potamotrygon gr. orbignyi. Peptides 2006;27:3039–46.

[9] Essman U, Perela L, Berkowitz ML, Darden T, Lee H, Pedersen LG. A smoothparticle mesh Ewald method. J Chem Phys 1995;103:8577–92.

[10] Guex N, Peitsch MC. Swiss-Model and the Swiss-PDB Viewer: an environmentfor comparative protein modeling. Electrophoresis 1997;18:2714–23.

[11] Haddad Jr V, Neto DG, de Paula Neto JB, de Luna Marques FP, Barbaro KC.Freshwater stingrays: study of epidemiologic, clinic and therapeutic aspectsbased on 84 envenomings in humans and some enzymatic activities of thevenom. Toxicon 2004;43:287–94.

[12] Hess B, Bekker H, Berendsen HJ, Fraaije JGEM. LINCS: a linear constraint solverfor molecular simulations. J Comput Chem 1997;18:1463–72.

[13] Ilsley JL, Sudol IM, Winder SJ. The WW domain: linking cell signaling to themembrane cytoskeleton. Cell Signal 2002;14:183–9.

[14] Kay BK, Williamson MP, Sudol M. The importance of being proline: theinteraction of proline-rich motifs in signaling proteins with their cognatedomains. FASEB J 2000;14:231–41.

[15] Ley K, Laudanna C, Cybulsky MI, Nourshargh S. Getting to the site of inflamma-tion: the leukocyte adhesion cascade updated. Nat Rev Immunol 2007;7:678–89.

[16] Ley K, Zarbock A. Hold on to your endothelium: postarrest steps of theleukocyte adhesion cascade. Immunity 2006;25:185–7.

[17] Lindahl E, Hess B, Van der Spoel D. GROMACS 3.0: a package for molecularsimulation and trajectory analysis. J Mol Model 2001;7:306–17.

[18] Lomonte B, Lundgren J, Johansson B, Bagge U. The dynamics of local tissuedamage induced by Bothrops asper snake venom and myotoxin II on the mousecremaster muscle: an intravital and electron microscopic study. Toxicon1994;34:41–55.

[19] Macias MJ, Weisner S, Sudol M. WW and SH3 domains, two different scaffoldsto recognize proline-rich ligands. FEBS Lett 2002;513:30–7.

[20] Magalhaes GS, Lopes-Ferreira M, Junqueira-De-Azevedo IL, Spencer PJ, AraujoMS, Portaro FC, et al. Natterins, a new class of proteins with kininogenaseactivity characterized from Thalassophryne nattereri fish venom. Biochimie2005;87:687–99.

[21] Magalhaes KW, Lima C, Piran-Soares AA, Marques EE, Hiruma-Lima CA, Lopes-Ferreira M. Biological and biochemical properties of the Brazilian Potamo-trygon stingrays: Potamotrygon cf. scobina and Potamotrygon gr. orbignyi.Toxicon 2006;47:575–83.

[22] Oostenbrink C, Villa A, Mark AE, van Gunsteren WF. A biomolecular force fieldbased on the free enthalpy of hydration and solvation: the GROMOS force-fieldparameter sets 53A5 and 53A600 . J Comput Chem 2004;25:1656–76.

[23] Pires JR, Taha-Nejad F, Toepert F, Ast T, Hoffmuller U, Schneider-Mergener J,et al. Solution structures of the YAP65 WW domain and the variant L30 K incomplex with the peptides GTPPPPYTVG, N-(n-octyl)-GPPPY and PLPPY andthe application of peptide libraries reveal a minimal binding epitope. J Mol Biol2001;314:1147–56.

[24] Rioli V, Prezoto BC, Konno K, Melo RL, Klitzke CF, Ferro ES, et al. A novelbradykinin potentiating peptide isolated from Bothrops jararacussu venomusing catalytically inactive oligopeptidase EP24.15. FEBS J 2008;275:2442–54.

[25] Springer TA. Traffic signals for lymphocyte recirculation and leukocyte emi-gration: the multistep paradigm. Cell 1994;76:301–14.

[26] Ueda M, Suzuki T, Honma H, Nagai Y, Nagashima K, Shiomi K. Purification,properties and cDNA cloning of neoverrucotoxin (neoVTX), a hemolytic lethalfactor from the stonefish Synanceia verrucosa venom. Biochim Biophys Acta2006;1760:1713–22.

[27] Zarbock A, Ley K. Mechanisms and consequences of neutrophil interactionwith the endothelium. Am J Pathol 2008;172:1–7.