Proteome analysis of brown spider venom: Identification of loxnecrogin isoforms in Loxosceles gaucho...

REGULAR ARTICLE

Proteome analysis of brown spider venom: Identification

of loxnecrogin isoforms in Loxosceles gaucho venom

Leandro F. Machado1, Sabrina Laugesen2, Elvio D. Botelho1, Carlos A. O. Ricart1,Wagner Fontes1, Katia C. Barbaro3, Peter Roepstorff2 and Marcelo V. Sousa1

1 Brazilian Center for Protein Research, Department of Cell Biology, University ofBrasilia, Brasilia, Brazil

2 Protein Research Group, Department of Biochemistry and Molecular Biology,University of Southern Denmark, Odense, Denmark

3 Laboratory of Immunopathology, Butantan Institute, São Paulo, Brazil

Brown spiders of the Loxosceles genus are distributed worldwide. In Brazil, eight species are foundin Southern states, where the envenomation by Loxosceles venom (loxoscelism) is a health problem.The mechanism of the dermonecrotic action of Loxosceles venom is not totally understood. Twoisoforms of dermonecrotic toxins (loxnecrogins) from L. gaucho venom have been previously pu-rified, and showed sequence similarities to sphingomyelinase. Herein we employed a proteomicapproach to obtain a global view of the venom proteome, with a particular interest in the loxne-crogin isoforms’ pattern. Proteomic two-dimensional gel electrophoresis maps for L. gaucho,L. intermedia, and L. laeta venoms showed a major protein region (30–35 kDa, pI 3–10), where atleast eight loxnecrogin isoforms could be separated and identified. Their characterization used acombined approach composed of Edman chemical sequencing, matrix-assisted laser desorption/ionization-time of flight mass spectrometry, and electrospray ionization-quadropole-time offlight tandem mass spectrometry leading to the identification of sphingomyelinases D. Thevenom was also pre-fractionated by gel filtration on a Superose 12 fast protein liqiud chromatog-raphy column, followed by capillary liquid chromatography-mass spectrometry. Eleven possibleloxnecrogin isoforms around 30–32 kDa were detected. The identification of dermonecrotic toxinisoforms in L. gaucho venom is an important step towards understanding the physiopathology ofthe envenomation, leading to improvements in the immunotherapy of loxoscelism.

Received: September 1, 2004Accepted: October 28, 2004

Keywords:

Brown spider / Dermonecrotic toxin / Loxnecrogin isoforms / Loxosceles gaucho /Mass spectrometry / Sphingomyelinase D / Two-dimensional gel electrophoresis /Venom proteome

Proteomics 2005, 5, 2167–2176 2167

1 Introduction

The Loxosceles spiders, known as “brown spider,” are cosmo-politanly found either as native or exotic in many continents[1, 2]. There are eight species of Loxosceles in Brazil, wherethe accidental envenomation (loxoscelism) is considered a

health problem in some states. The medically importantspecies are L. gaucho, L. intermedia, and L. laeta, which arewidely distributed in Southern and Southeastern Brazilianstates.

The Loxosceles spider venom causes a characteristic der-monecrotic lesion, occasionally accompanied by systemicreactions [3, 4]. In some cases, the victim presents moresevere systemic reactions characterized by intravascular co-agulation and acute renal failure [5]. Differences in clinicalresponse can be associated to spider gender and age [6, 7],amount of venom injected into the victim, patient age [5],and the bite site [8].

Correspondence: Dr. Marcelo Valle de Sousa, Brazilian Center forProtein Research, Department of Cell Biology, University of Bra-silia, Brasilia DF, 70910-900, BrazilE-mail: [email protected]: 155-61-2734608

2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.de

DOI 10.1002/pmic.200401096

2168 L. F. Machado et al. Proteomics 2005, 5, 2167–2176

The mechanism of action of the venom is not completelyunderstood. The dermonecrotic lesion is due to the complexprocess initially characterized by the direct effect of thevenom on components of the cellular membrane, basalmembrane, and extracellular matrix. Besides endogenousmechanisms such as activation of the system complement,migration of polymorphonuclear leukocytes, and plateletaggregation also contribute to induce local tissue damage.Proteases, hydrolases, lipases, phosphatases, 5-ribonu-cleases, phosphohydrolases, sphingomyelinase D, and othercomponents have been described in Loxosceles venoms [2, 9].The spread factor during the intoxication was mainly attrib-uted to a hyaluronidase [2], and the noxious effects on extra-cellular matrix, to proteases [10–13].

Sphingomyelinase D has been regarded as the mostimportant component of the venom [2, 14]. It is related to thedermonecrotic lesion caused by venoms of several Loxoscelesspecies [1, 15–18]. The mechanism of action of the dermo-necrotic enzyme is connected to a metal dependent process[19, 20].

Isoforms of the dermonecrotic toxin have been demon-strated for Loxosceles species [19–21]. For L. gaucho specifi-cally, we showed in a previous report the purification andcharacterization of two isoforms of dermonecrotic toxins(loxnecrogins) [22], but evidences occurred in favor of a lar-ger number of isoforms present in the venom proteome.Herein we employed a proteomic approach aimed at obtain-ing a global view of the venom proteome, with a particularinterest in describing the loxnecrogin isoforms’ pattern.

2 Material and methods

2.1 Chemicals

Pharmalyte (pH 3–10), Immobiline DryStrip 3–10L, acryl-amide, bis-acrylamide, ammonium persulfate, TEMED,CHAPS, silicone oil, Tris, glycine, iodoacetamide, DTT, andurea were from Amersham Biosciences (Uppsala, Sweden).Trypsin was obtained from Promega (Madison, WI, USA). a-CHCA was purchased from Bruker Daltonics (Karlsruhe,Germany). ACN, formic acid, methanol, TFA, and ammo-nium bicarbonate were of analytical grade or better fromSigma (St. Louis, MO, USA). Milli-Q water was used to pre-pare all the aqueous solutions.

2.2 Venom

Specimens of L. gaucho were collected in the state of SãoPaulo, Brazil. The species L. laeta and L. intermedia werecollected in the state from Santa Catarina and Paraná, Brazil,respectively. The spiders were kept in quarantine for a weekwithout food before venom collection, and submitted toelectrostimulation to obtain venom as previously described[23]. The venom pools were dried in a vacuum centrifuge. Apool of 11 000 animals of each species was used to obtain the

venom samples. The protein content was determined by theBCA Protein Assay Reagent Kit (Pierce, Rockford, IL, USA)using its microtiter plate protocol.

2.3 Venom sample preparation

Venom samples were solubilized in 2-DE sample buffer (7 M

urea, 2 M thiourea, 4% w/v CHAPS, 1% w/v DTT, 0.5% v/vPharmalyte 3–10) with protease inhibitors (100 mM PMSF,100 mM TLCK, 1 mM pepstatin A, 100 mM leupetin, 5 mM

EDTA, 50 mM TPCK) to a final protein concentration of1.0 mg mL21, aliquoted and stored at 2807C.

2.4 Two-dimensional electrophoresis

For each of the studied species (L. gaucho, L. intermedia, andL. laeta), 50 mg of the prepared venom samples were dilutedin 2-DE sample buffer (380 mL final volume), applied to18 cm IPG gel strips with linear range of pH 3–10 and 4–7(Amersham Biosciences), and incubated for 12 h to rehy-drate the strip. IEF was carried out at 207C on an IPGphorunit (Amersham Biosciences) at step-n-hold conditions:500 V for 1 h; 1000 V for 1 h; and 8000 V for 4 h at 75 mA/strip. After focusing, the proteins were submitted to reduc-tion and alkylation. The strips were soaked for 20 min inreduction solution (6 M urea, 30% glycerol, 2% SDS, and125 mM DTT) followed by 20 min in alkylation solution (6 M

urea, 30% glycerol, 2% SDS, and 125 mM iodoacetamide).The SDS-PAGE step was performed in 10%T polyacrylamidegels run on a Protean II system (Bio-Rad, Hercules, CA,USA) at 207C. Electrophoresis was carried out at constantcurrent of 15 mA per gel at 157C. Bromophenol blue wasused as front dye. For silver staining, the gels were incubatedin 12% TCA, 50% methanol for 1 h, and then in 5% aceticacid, 50% ethanol for 1 h. The gels were then washed twicefor 5 min in water, incubated in a solution containing 0.89%AgNO3, 0.3% NH3, and 0.079% NaOH for 20 min, washedfor 5 min in water and incubated in 0.016% w/v citric acidand 0.085% v/v formaldehyde solution until the spots werevisible. To stop the reaction a solution of 20% v/v ethanol,7% v/v acetic acid was used. The gels were stored in 1%acetic acid at 57C, before protein digestion.

2.5 Western blotting

The proteins of L. gaucho venom (200 mg) were first fraction-ated by 2-DE as described in Section 2.4. 2-DE separatedproteins were electroblotted to an NC membrane [24] usinga Multiphor II (Amersham Biosciences) electroblottingapparatus, and transferred for 70 min at 1.0 mA cm22. TheNC membrane was then incubated with anti-loxnecroginmAb MoALg1 (1.5 mg mL21, diluted 1:1000) [25]. Theimmunoreactive proteins were detected using peroxidase-labeled anti-mouse IgG (1:500) (Sigma) followed by devel-


Proteomics 2005, 5, 2167–2176 Animal Proteomics 2169

opment with 0.05% 4-chloro-1-naphthol (Merck, Darmstadt,Germany) in 15% v/v methanol, in the presence of 0.03%v/v H2O2.

2.6 Peptide mass mapping

Selected protein spots from L. gaucho venom 2-DE gel wereexcised, and digested with trypsin [26]. The tryptic peptideswere extracted twice with 40 mL of ACN/water/TFA(66:33:0.1) solution for 20 min with the aid of a sonicatorapparatus. The protein digest was dried in a vacuum cen-trifuge and stored at 2207C prior to use. The protein digestswere solubilized in 4 mL 0.1% TFA, followed by microscaleconcentration and desalting using C18 Zip-Tips (Millipore,Bedford, MA, USA). Peptides were eluted directly onto aMALDI-TOF probe using 1 mL of 50% ACN, 0.1% TFA so-lution containing matrix (a-CHCA 20 mg mL21). MS wasperformed using a Reflex IV (Bruker Daltonics, Karlsruhe,Germany) mass spectrometer in positive reflector mode.Mass spectra were processed using XMASS and Biotoolssoftware (Bruker Daltonics). Spectra were internally cali-brated using trypsin autolysis products (m/z 842.509 andm/z 2211.104). Protein identification was performed usingMASCOT (http://www.matrixscience.com) at 50 ppm masstolerance against NCBI (nonredundant) and Swiss-Protdatabases.

2.7 Tandem mass spectrometry experiments

Peptide mixtures derived from the gel digests were enrichedand purified on a custom-made POROS R2 nano-columnperformed as described [27] with some modifications. A col-umn consisting of 100–300 nL of POROS R2 resin (AppliedBiosystems, Framingham, MA, USA) was packed in a con-stricted GeLoader tip (Eppendorf, Hamburg, Germany). A1.25 mL syringe was used to force liquid through the columnby applying gentle air pressure. The column was equilibratedwith 20 mL 5% formic acid, and then the analyte solution wasadded. The column was washed with 20 mL of 5% formicacid, and the bound peptides subsequently eluted with 1 mLof 50% methanol/5% formic acid v/v into a pre-coated bo-rosilicate nanoelectrospray needle (Protana Engineering,Odense, Denmark). Samples were analyzed by nanoelec-trospray MS/MS using a quadrupole TOF instrument(Micromass, Manchester, UK). All spectra were obtained inthe positive ion mode. The data acquisition was performedwith a Mass Lynx PC data system (version 3.5 for Win-dows NT). For all experiments, a sodium iodide solution(2 mg mL21 in 50% propan-2-ol) was used for the TOF cali-bration.

2.8 Edman chemical sequencing

Proteins from L. gaucho venom were electroblotted ontoPVDF membrane (Applied Biosystems) using transfer buffer(48 mM Tris, 39 mM glycine, 0.037% SDS, 20% methanol).

Before the electroblotting, the membrane was soaked inmethanol for 30 s and incubated in transfer buffer for10 min. Protein transfer was carried out using a Multiphor II(Amersham Biosciences) for 1 h and 15 min at constant200 mA and 500 V. The PVDF membrane was washed fourtimes (30 s) in water before CBB R-250 staining. The spotswere excised and loaded on a 477A-120A automated proteinsequencer (Applied Biosystems) with some modifications[28] for N-terminal sequence determination. Similarity sear-ches were performed using BLAST (http://www.ncbi.nlm.-nih.gov/BLAST/).

2.9 Gel filtration

The L. gaucho venom was pre-fractionated using an FPLCSuperose 12 column (Amersham Biosciences) equilibratedwith 50 mM Tris, pH 8.0. A venom sample of 200 mL(1 mg mL21) was applied on the column, and eluted with thesame buffer in 120 min at 0.5 mL min21 flow rate. The ab-sorbance was monitored at 280 nm. The fractions obtainedwere dried in a vacuum centrifuge and stored at 2207C.

2.10 Capillary liquid chromatography with mass

spectrometry detection (cLC-MS)

Gel filtration fractions were submitted to cLC-MS. A electro-spray triple quadrupole mass spectrometer model API 300(Perkin Elmer-Sciex, Ontario, Canada) was coupled to anApplied Biosystems 140C microgradient HPLC. The frac-tions (3 mg) were dissolved in 40 mL of 0.1% formic acid,from which 2 mL were applied directly into a home-madecapillary column (150 mm id 6 25 cm length) packed withVydac C18 resin at 5 mm particle diameter. The elution wascarried out by a 0–100% ACN gradient in 0.1% formic acid ata splitter regulated flow rate of 25 mL min21. The eluent flowfrom the column was directed to the microionspray probe ofthe mass spectrometer, which was run under positive mode.

3 Results

3.1 Proteome maps of L. gaucho, L. intermedia, and

L. laeta venoms

A comparative proteomic analysis was carried out withvenoms of L. gaucho, L. intermedia, and L. laeta. For the threeproteomes, a principal region of protein spots at 30–35 kDarange was observed on 2-DE gels (Fig. 1). Venom proteins inthe same mass range were previously described as respon-sible for dermonecrosis [15, 18, 23].

For the L. gaucho venom, at least eight main spots(LOXN 1–8) around 30–35 kDa and pI 4–9 were detectedwhen using a pH gradient from 3 to 10 (Fig. 1a). When a pHgradient from 4 to 7 was used, spot LOXN8 could be betterresolved, although not yet totally separated in individualspots (not shown). The 2-D map of L. intermedia venom



Figure 1. Proteome maps of Loxosceles spider venoms. 2-DEwas performed using IPG 3–10 L strips and 10% SDS-PAGE. (a)L. gaucho; (b) L. intermedia, and (c) L. laeta. The maps show anumber of intense protein spots on 30–35 kDa mass range. Thearrows on L. gaucho map indicate protein spots of loxnecroginisoforms (LOXN1–8) with respective pIs (lower inset). LOXN 7and 8 were immunodetected by anti-L. gaucho loxnecroginmonoclonal antiboby (MoALg1) (upper inset).

showed most spots between pI 6 and 10, but more intenseones were concentrated at neutral pI (Fig. 1b). For L. laeta, its2-D map displayed the 30–35 kDa protein region rather shift-ed to acidic pIs (Fig. 1c), with a minor group remaining at

basic pIs. The proteome maps of all Loxosceles species alsopresented a high mass region between 45 and 94 kDa(Fig. 1). The low molecular mass region (14–25 kDa) exhib-ited few proteins (Fig. 1). Nevertheless, L. intermedia pro-teome exhibited a little more of these low molecular massproteins than the other species.

3.2 Loxnecrogin immunodetection

In a first approach to find dermonecrotic toxins (loxnecro-gins) present in the venom proteome, 2-DE (pH 3–10 and 4–7) of L. gaucho venom was carried out followed by immuno-blotting using a mAb (MoALg1) produced against the der-monecrotic component (35 kDa) of the venom. Only twospots were immunolabeled corresponding to LOXN 7 and 8(Fig. 1a, inset). As several other protein spots were present inthe same molecular mass range of the imunnodetected pro-teins, MS was then chosen for further characterization andidentification of the separated proteins.

3.3 Loxnecrogin isoforms’ identification

Taking the proteome map from L. gaucho venom as ourreference, spots 1–7 (LOXN 1–7) were submitted to PMF byMALDI-TOF MS. LOXN 8 was not analyzed since it had notbeen completely resolved by 2-DE. Although no positiveidentification could be achieved by MASCOT searches, massspectra of the digests presented some similar m/z regions(Fig. 2) suggesting that we might be dealing with either iso-forms or very similar proteins. Based on visual inspection ofthe peptides mass maps a higher level of similarity wasfound between LOXN1 and LOXN2, between LOXN3 andLOXN4 and between LOXN6 and LOXN7, whereas the pep-tide map of LOXN5 showed lower similarity to the others.

In view of the absence of conclusive identification of themain 30–35 kDa proteins in L. gaucho venom by PMF, wethen turned to MS/MS sequencing. Figure 3a shows the ESI-MS/MS spectrum of a doubly charged peak at m/z 777.936 inthe ESI-MS spectrum of the digest from LOXN1, corre-sponding to the peak at m/z 1553.7 in the MALDI-MS spec-trum of the same digest. Direct search of the raw MS/MSusing MASCOT did not result in any significant match.Therefore, de novo sequencing was performed allowing nearcomplete assignment of the peaks. The amino acid sequenceobtained was (L\I)(L\I)SSYW(Q\K)DGNS. The sequence wasused to search non-redundant sequences using BLASTsearch for short, nearly exact matches identifying a sphingo-myelinase-like protein from L. Laeta (Q81912, TrEMBL/AAM21154, NCBI). The peptide sequence was placed fromposition 140 to 150 on the protein sequence (Fig. 4). Thealignment of the tryptic peptide sequence with the proteinsequence allowed the confirmation of the isobaric aminoacids in the MS/MS spectrum. Similarly, LOXN2 was alsoidentified as the same sphingomyelinase-like proteinL. Laeta (Q81912, TrEMBL/AAM21154, NCBI) based on MS/MS sequence of a same peptide at m/z 777.936.



Figure 2. PMF of loxnecroginisoforms (LOXN1–7). Protein inspots were digested with tryp-sin, cleaned up through ZipTips,and submitted to MALDI-TOFanalysis. Searches based onPMF did not result in the identi-fication of the proteins.

Also the MS/MS spectra derived from the peptide peak at(M 1 2H)21 = 833.013 from LOXN1 and LOXN2 (data notshown) were submitted to search using MASCOT leading tothe same sphingomyelinase-like protein (Q81912, TrEMBL/AAM21154, NCBI) from L. laeta with a MOWSE score of 88.The resulting sequence was LGVDAIMTNYPEDVK. Theregion matched residues from 266 to 280 along the proteinsequence (Fig. 4).

For LOXN 4, the MS/MS sequence LATYEDNPWETFKobtained for a doubly charged peptide at m/z 872.085 in theESI-MS spectrum (1744.4 in the MALDI-MS spectrum)(Fig. 3b) allowed a BLAST identification of another sphingo-myelinase-like protein (Q81913, TrEMBL/AAM21155,NCBI) from L. laeta. That is the C-terminal peptide (Fig. 4).Again, the LOXN4 peptide sequence RYDMSGNDALGDVK,resulting from a doubly charged peptide peak at m/z 843.04(data not shown), provided a match to the same L. laetasphingomyelinase-like protein upon BLAST search.

Altogether, the MS/MS identifications indicated that wemight be dealing with sphingomyelinases carrying dermo-necrotic activity, which were previously purified and char-acterized in our group and called loxnecrogins [22]. To gatherfurther evidence that all the eight main 30–35 kDa proteinspots corresponded to loxnecrogin isoforms, we determinedthe N-terminal amino acid sequences by Edman degrada-tion. All of the protein spots provided clear sequences,including LOXN8, which seemed to be a mixture of unre-solved, very close isoforms.

All of them showed similarity upon sequence alignmentwith protein sequences described as responsible for dermo-necrotic action on Loxosceles venom (Fig. 4), despite someamino acid residue variations. The first N-terminal residuewas assigned as A for all but three loxnecrogin isoforms (V forLOXN2, D for LOXN5 and G for LOXN6) as well as for thewhole family of previously sequenced dermonecrotic toxins.The second residue is D for all the eight loxnecrogin isoforms,but G was sometimes present in other proteins of the family.The third residue was very variable. The fourth residue was Rfor all the eight isoforms and basic (R or K) for the wholefamily of dermonecrotic toxins, which was also true for thefifth residue. The sixth residue showed P in all of the proteinsof the family, except in LOXN5 and 6, which presented Q. ForLOXN4, there is a missing I at the sixth position.

3.4 Loxnecrogin isoforms’ analysis by cLC-MS

As a further way to see loxnecrogin isoforms, we submittedL. gaucho venom fractions to cLC-MS. Pre-fractionation ofthe venom was carried by gel filtration using an FPLCSuperose 12 column. Seven fractions named F1-F7 were col-lected (Fig. 5a). Fraction F1 showed no detectable massesupon cLC-MS, while F2, F3, and F4 (Fig. 5b) displayed atleast 11 proteins in the mass range from 31 to 33 kDa(Fig. 5b–e) suggesting that the number of loxnecrogin iso-forms might exceed the ones observed by 2-DE, which wereeight in number. In F2 only, eight possible different isoforms



Figure 3. MS/MS identification of some loxnecrogin isoforms. (a) (M 1 2H)21 = 777.936 m/z tryptic peptide fromLOXN1 and LOXN2 generated the sequence tag (L/I)(L/I)SSYW(Q/K)DGNS matched a sphingomyelinase-like pro-tein from L. laeta. (b) MS tag of (M 1 2H)21 = 872.085 m/z from LOXN4 produced the sequence LATYEDNPWETFKfrom a sphingomyelinase-like protein from L. laeta.



Figure 4. Alignment of dermonecrotic toxin sequences from Loxosceles genus. The protein sequences and cDNA sequences were alignedusing the CLUSTALW (1.82) program [32]. The sequences obtained in this work were aligned manually. Access numbers are either from theNCBI database (A38107 [33]; AAB50794, AAB50795, and AAB50796 [18]; AAM21154, AAM21155, and AAM21156 [21]; AAP44735 (Binford,G. J. et al., unpublished data); AAQ16123 [30]; AAP97091 and AAP97092 [31]) or from the Swiss-Prot database (P83045 and P83046 [20]).Loxnecrogin A partial sequence was from [22].

(31 439.5; 31 446.0; 31 620.8; 3 1627.5; 31 729.0; 31 501.0;and 32 578.5 Da) were detected. Moreover, in F3 and F4 therewere one (31 884.5 Da) and two possible isoforms (31 616.0and 31 875.0 Da), respectively. Two loxnecrogin isoformswere purified and characterized previously showing molecu-lar masses of 31 444 Da and 32 626 Da [22].

4 Discussion

Proteome maps were obtained for L. gaucho, L. intermedia,and L. laeta venons, presenting many protein spots in the 30–35 kDa mass range of the dermonecrotic loxnecrogins(Fig. 1). At least eight loxnecrogin isoforms were separated



Figure 5. FPLC and cLC-MS analysis of L. gaucho venom. Thevenom was first separated into seven fractions (F1–7) using anFPLC Superose 12 column (a). Each fraction was submitted tocLC-MS (C18 column, 150 mm id by 25 cm lengh) connected to anelectrospray mass spectrometer (b–e). Eleven different proteinswere detected in the mass range of dermonecrotic toxins.



by 2-DE of the L. gaucho venom, and two of them (LOXNs 7and 8) were immunodetected by anti-loxnecrogin mAb(Fig. 1). Combined use of MS de novo peptide sequencingand N-terminal chemical sequencing analyses followed byBLAST searches identified all proteins in the eight spots assphingomyelinase toxins (Figs. 3, 4), which are correlated tothe dermonecrotic action caused by brown spider venoms. BycLC-MS, 11 molecular species were detected in the massrange (31–33 kDa) expected for loxnecrogins (Fig. 5). Arecent report showed a 2-DE profile of L. intermedia venom[29] that resembles the one presented in this work.

Evidence for the existence of sphingomyelinase isoformsin a North American brown spider (L. reclusa) venom wasfirst reported long ago [19]. Concerning Brazilian brown spi-ders, two isoforms of L. intermedia dermonecrotic sphingo-myelinase have been purified [20]. A gene (LiD1) belongingthe dermonecrotic family of L. intermedia was then cloned,sequenced, and expressed [30]. Immunochemical studies onthat recombinant protein indicated that it might be one ofthe larger family of isoforms. Indeed, three other isoforms ofdermonecrotic sphingomyelinases of L. intermedia wererecently cloned and sequenced [31] as well as three fromL. laeta [21]. Finally, two dermonecrotic toxin isoformsnamed loxnecrogin A and B were purified from the venom ofL. gaucho, and partially sequenced showing similarity toLiD1 [22].

A recent report showed the use of MS in the analysis ofLoxosceles venoms [1]. Venoms from L. laeta and nine otherNorth American and African Loxosceles species together withHaplogyne spiders were analyzed by SELDI-MS in that study.Whilst Loxosceles venons showed proteins ranging from 31 to32 kDa, the resolution of the technique was not sufficient todetermine neither the number nor the exact masses of theisoforms.

Our present cLC-MS results demonstrated that at least 11possible loxnecrogin isoforms were detected in L. gauchovenom proteome (Fig. 5), of which eight major isoformswere identified by MS/MS and N-terminal sequencing asbelonging to the sphingomyelinase family, after separationby 2-DE (Fig. 1). Spot 8 (LOXN8) seemed to contain morethan one protein species. Several other spots in the sameregion were present, but did not allow identification due tolow abundance. Thus the number of 11 isoforms determinedby cLC-MS may represent a more accurate picture.

Some of the protein spots in the 31–33 kDa range on the2-DE map (Fig. 1) might be metalloproteinases since severalof such activities in the 28–35 kDa range on 1-DE zymogramgels have been determined [10]. High molecular mass serine-proteases (85–95 kDa) were also previously detected by 1-DEzymograms [12]. Indeed, our 2-DE gels revealed high molec-ular protein spots at 85–95 kDa at acidic pIs for the threespecies studied (Fig. 1).

Differences in toxicity among L. gaucho, L. intermedia,and L. laeta venons have been reported [15], and could per-haps be associated with variations in the distribution of Lox-osceles spider loxnecrogin isoforms along the pI gradient

(Fig. 1). In relation to L. gaucho, the venom proteome ofL. intermedia presented a larger number of basic isoforms,whilst L. laeta displayed a rather acidic major group of iso-forms, with a minor basic group. Probably, brown spiderspossess a multi-gene family of loxnecrogins as indicated withthe cloning and sequencing of three L. laeta dermonecroticcDNA toxins [21]. Post-translational modifications might alsobe a source of loxnecrogin micro-heterogeneity.

The variation in the pI of loxnecrogins may reflect sub-stitutions of amino acid residues on the surface of the pro-tein structure. Comparison of loxnecrogin isoform peptidemaps separated by 2-DE showed that no peptide peaks werecommon to all of them (Fig. 2). Close pairwise similaritycould however be observed between LOXNs 1 and 2,LOXNs 3 and 4, and LOXNs 6 and 7. These structural changescan affect the recognition of antibodies, since the mono-clonal antiboby (MoALg1) produced against L. gaucho loxne-crogin failed to recognize and neutralize dermonecrotic andlethal activities of L. intermedia and L. laeta venoms [25].

Loxnecrogin A and B, previously purified and character-ized [22], were most likely present among the isoformsdetected by cLC-MS. A protein with molecular mass of31 446 Da may be assigned to loxnecrogin A (31 444 Da) andone of 31 627.5 Da to loxnecrogin B (31 626 Da).

The reason for the existence of such a number of iso-forms related to a single toxic activity in a venom proteomemay reside in an evolutionary adaptation for the spider toobtain maximum success in both killing its prey anddefending itself from predators. The purification and char-acterization of all loxnecrogin isoforms and identification ofother toxins in the proteomes of Loxosceles spiders will beneeded to understand the physiopathology of the enveno-mation that could lead to improvements in the immu-notherapy of loxoscelism.

The authors thank Nuno Domingues for technical assistanceon protein chemical sequencing. The mAb (MoALg1) was kindlyprovided by Patrícia Guilherme and Dr. Irene Fernandes. Thiswork was supported by CNPq (Brazil) research grant PADCT620.545/98-4. LFM was a recipient of a CNPq studentshipthrough the Molecular Pathology Post-Graduation Program atthe University of Brasilia.

5 References

[1] Binford, G. J., Wells, M. A., Comp. Biochem. Physiol. B. Bio-chem. Mol. Biol. 2003, 135, 25–33.

[2] Futrell, J. M., Am. J. Med. Sci. 1992, 304, 261–267.

[3] Malaque, C. M. S., Valencia, J. E. C., Cardoso, J. L. C., Franca,F. O. S. et al., Rev. Inst. Med. Trop. S. Paulo 2002, 44, 139–143.

[4] Mota, I., Barbaro, K. C., J. Toxinol. – Toxin Rev. 1995, 14, 401–422.

[5] Sezerino, U. M., Zannin, M., Coelho, L. K., Goncalves Jr. et al.,Trans. R. Soc. Trop. Med. Hyg. 1998, 92, 546–548.



[6] Oliveira, K. C., Goncalves de Andrade, R. M., Giusti, A. L.,Dias da Silva, W., Tambourgi, D. V., Toxicon. 1999, 37, 217–221.

[7] Andrade, R. M. G., De Oliveira, K. C., Giusti, A. L., Dias daSilva, W., Tambourgi, D. V., Toxicon. 1999, 37, 627–632.

[8] Bernstein, B., Ehrlich, F., J. Emerg. Med. 1986, 4, 457–462.

[9] Rash, L. D., Hodgson, W. C., Toxicon. 2002, 40, 225–254.

[10] Feitosa, L., Gremski, W., Veiga, S. S., Elias, M. C. et al., Tox-icon. 1998, 36, 1039–1051.

[11] Veiga, S. S., Feitosa, L., dos Santos, V. L., de Souza, G. A. etal., Histochem. J. 2000, 32, 397–408.

[12] Veiga, S. S., da Silveira, R. B., Dreyfus, J. L., Haoach, J. et al.,Toxicon. 2000, 38, 825–839.

[13] Veiga, S. S., Zanetti, V. C., Braz, A., Mangili, O. C., Gremski,W., Braz. J. Med. Biol. Res. 2001, 34, 843–850.

[14] Forrester, L. J., Barrett, J. T., Campbell, B. J., Arch. Biochem.Biophys. 1978, 187, 355–365.

[15] Barbaro, K. C., Ferreira, M. L., Cardoso, D. F., Eickstedt, V. R.,Mota, I., Braz. J. Med. Biol. Res. 1996b, 29, 1491–1497.

[16] Gomez, H. F., Miller, M. J., Waggener, M. W., Lankford, H. A.,Warren, J. S., Toxicon. 2001, 39, 817–824.

[17] Young, A. R., Pincus, S. J., Toxicon. 2001, 39, 391–400.

[18] Barbaro, K. C., Sousa, M. V., Morhy, L., Eickstedt, V. R., Mota,I., J. Protein. Chem. 1996, 15, 337–343.

[19] Kurpiewski, G., Forrester, L. J., Barrett, J. T., Campbell, B. J.,Biochim. Biophys. Acta 1981, 678, 467–476.

[20] Tambourgi, D. V., Magnoli, F. C., van den Berg, C. W., Mor-gan, B. P. et al., Biochem. Biophys. Res. Commun. 1998, 251,366–373.

[21] Fernandes Pedrosa, M. F., Junqueira de Azevedo, I. L. M.,Goncalves-de-Andrade, R. M., van den Berg, C. W. et al.,Biochem. Biophys. Res. Commun. 2002, 298, 638–645.

[22] Cunha, R. B., Barbaro, K. C., Muramatsu, D., Portaro, F. C. etal., J. Protein Chem. 2003, 22, 135–146.

[23] Barbaro, K. C., Cardoso, J. L., Eickstedt, V. R., Mota, I., Tox-icon. 1992, 30, 331–338.

[24] Towbin, H., Staehelin, T., Gordon, J., Proc. Natl. Acad. Sci.USA 1979, 76, 4350–4354.

[25] Guilherme, P., Fernandes, I., Barbaro, K. C., Toxicon. 2001,39, 1333–1342.

[26] Shevchenko, A., Wilm, M., Vorm, O., Jensen, O. N. et al.,Biochem. Soc. Trans. 1996, 24, 893–896.

[27] Gobom, J., Nordhoff, E., Mirgorodskaya, E., Ekman, R.,Roepstorff, P., J. Mass Spectrom. 1999, 34, 105–116.

[28] Fontes, W., Cunha, R. B., Sousa, M. V., Morhy, L., Anal. Bio-chem. 1998, 258, 259–267.

[29] Luciano, M. N., da Silva, P. H., Chaim, O. M., dos Santos, V. L.et al., J. Histochem. Cytochem. 2004, 52, 455–467.

[30] Kalapothakis, E., Araujo, S. C., de Castro, C. S., Mendes, T.M. et al., Toxicon. 2002, 40, 1691–1699.

[31] Tambourgi, D. V., de F. Fernandes Pedrosa, M., van den Berg,C. W., Goncalves-de-Andrade, R. M. et al., Mol. Immunol.2004, 41, 831–840.

[32] Thompson, J. D., Higgins, D. G., Gibson, T. J., Nucleic AcidsRes. 1994, 22, 4673–4680.

[33] Cisar, R. C., Fox, J. W., Geres, C. R., Toxicon. 1989, 27, 37–38.


Proteome analysis of brown spider venom: Identification of loxnecrogin isoforms in Loxosceles gaucho...

Documents

Transcript of Proteome analysis of brown spider venom: Identification of loxnecrogin isoforms in Loxosceles gaucho...