Novel Function of Antihemorrhagic Factor HSF as an SSP-Binding Protein in Habu (Trimeresurus flavoviridis) Serum
Narumi Shioi1), Masaaki Narazaki1), and Shigeyuki Terada1)
（ReceivedMay31,2011）
Abstract
Small serum proteins (SSPs) (molecular mass, 6.5–10 kDa) are high-molecular forms in the serum of Habu (Trimeresurus flavoviridis). An SSP-1–binding protein was purified from SSP- deficient serum by affinity chromatography on an SSP-1–HiTrap column and was identified as HSF, a known metalloproteinase inhibitor that suppresses the hemorrhagic activity of several snake venom metalloproteinases (SVMPs). The binding of SSP-1 to HSF was non-covalent and reversible, and the complex formed from these proteins dissociated at acidic pH. The mean dissociation constant of the complex was 2.5 ± 1.1 × 10-7 M as determined by surface plasmon resonance analysis. HSF also bound all other SSPs (SSP-2–SSP-5) with equimolar stoichiometry.
SSP-1 had no effect on the inhibition of several SVMPs by HSF. Thus, HSF plays an additional role in Habu serum as a carrier of SSPs.
1） DepartmentofChemistry,FacultyofScience,FukuokaUniversity,8-19-1Nanakuma,Jonan-ku,Fukuoka,814-0180, Japan
Abbreviations:CRISP,cysteine-richsecretoryprotein;HSF,habuserumfactor;PSP94,prostaticsecretoryproteinof94 aminoacids;SSP,smallserumprotein;SVMP,snakevenommetalloproteinase.
Introduction
Smallserumproteins(SSPs)arelow-molecular- mass proteins first found in Habu (Trimeresurus flavoviridis) serum［1］.Atpresent,5homologs, namelySSP-1–SSP-5,havebeenisolated［2］.Structural analysis indicated that they belong to the PSP94 (prostatic secretory protein of 94 amino acids) family,whichischaracterizedbyalowmolecular mass of approximately 10 kDa and 10 strictly conserved cysteine residues that form 5 disulfide bonds［3-5］.TheprotonNMRspectraofpigand human PSP94s have shown that they areβ-sheet proteinscomposedof2domains(N-andC-terminal domains)［6,7］. Although SSP-1 and SSP-2 are composedofabout90aminoacids,includingthe10
conserved Cys residues, SSP-5 has only 8 Cys residues.Interestingly,SSP-3andSSP-4consistof only 60 amino acids, as they lack the C-terminal domain.
Inapreviouspaper,wereportedthatallSSPs exist in high-molecular-mass forms in serum［8］. When T. flavoviridis serum was analyzed by analyticalgelfiltration,alltheSSPswereelutedasa broad peak with apparent molecular masses of approximately60–120kDa.SinceSSPsdonotself- associate in physiological buffers, they are hypothesized to be present in a protein complex.
LikeSSPs,humanPSP94existsincomplexwitha specific protein (PSP94–binding protein) in the blood and with cysteine-rich secretory protein-3 (CRISP-3)inprostatefluid［10］.
InasearchforSSP-bindingproteinsinT.
flavoviridisserum,wepreviouslyisolatedanovel protein named serotriflin that shows significant sequencesimilaritytotriflin,aCRISPfamilyprotein, inT. flavoviridisvenom［11］.Althoughserotriflinwas isolatedasabindingproteincandidateforSSPs,it showednoaffinitytoanySSPexceptSSP-2［8］.
Here, we report the identification of a novel SSP-binding protein. On performing affinity chromatographyusinganimmobilizedSSP-1column, we isolated a 48-kDa candidate SSP-1–binding protein.Sequenceanalysisrevealedthisproteinto be HSF, a metalloproteinase inhibitor that was isolated as an antihemorrhagic factor from T.
flavoviridis serum［12］. Furthermore, analytical gel filtration proved that all SSPs can bind to HSF.
Thus,HSFplaysadualroleinHabuserum—asa hemorrhagicSVMPdefenseproteinandasanSSP carrier.
Experimental
Materials
Blood from T. flavoviridis from the Amami OshimaIslandswascollectedbydecapitation.The serumwasseparatedbycentrifugationandstored at -20ºC. HSF and 5 SSPs were purified from T.
flavoviridisserum［2,13］.Low–molecular-weightsnake venommetalloproteinases(SVMPs)corresponding to HR2a and HR2b, and high–molecular-weight hemorrhagicenzymescorrespondingtoHR1Aand HR1BwerepurifiedfromT. flavoviridisvenom［14,15］. AllotherreagentswerepurchasedfromWakoPure Chemicals(Osaka).
Protein quantification
The concentration of pure samples was determinedusingaNanoVuespectrophotometer(GE Healthcare),andthemolarextinctioncoefficients werecalculatedat280nmforSSP-1(9,105M–1･cm–1), SSP-2 (10,595 M–1･cm–1), SSP-3 (8,855 M–1･cm–1), SSP-4 (12,865 M–1･cm–1), SSP-5 (13,450 M–1･cm–1), andHSF(23,670M–1･cm–1)［16］.
Measurement of proteolytic activity
Proteolyticactivitywasmeasuredinasolution containing5mMCaCl2and50mMTris-HCl(pH 8.5)usingfluoresceinisothiocyanate-labeledcasein
asthesubstrate .Theincreaseinfluorescencewas analyzedonanFP-550Aspectrofluorometer(Jasco) at520nmwithexcitationat490nm.
Electrophoresis
SDS-PAGEwascarriedouton12%gelsusing Laemmli’s method［18］. A prestained XL-ladder marker(Apro)wasusedasthemolecular-weight marker.Thegelwasstainedusing0.1%Coomassie brilliantblueR-250anddestainedwith10%acetic acid.
Preparation of the SSP-deficient serum
Habuserumwasfractionatedwithammonium sulfateintoP30–P60fractionsasdescribedpreviously［13］. Fraction P60 (500 mg) was dissolved in a 20 ml solutioncontaining50mMNaCland50mMcitrate buffer(pH3.5),andthesolutionwascentrifuged at 7,000 rpm for 20 min. The supernatants were transferredtoultrafiltrationtubeswithamolecular weightcut-offof30,000(UltracelYM-30,Millipore).
Thetubeswerecentrifugedat4,000×gfor30min, and the solution retained on the membrane was dialyzedagainstPBS.Theproteinconcentrationwas adjustedto20mg/ml.
Affinity chromatography
Anaffinityadsorbentwaspreparedbyreacting anN-hydroxy-succinimide-activatedHiTrapcolumn (1 ml, GE Healthcare Bio-Science) with purified SSP-1 (5.5 mg), according to the manufacturer’s instructions.TheSSP-deficientserum(1ml)was appliedtothecolumn,andthecolumnwaswashed with20mMphosphatebuffer(pH7.4)containing 0.15MNaClTheabsorbedmaterialsonthecolumn werethenelutedwith0.1MGly-HClbuffer(pH3.0) containing0.5MNaCland,and1-mlfractionswere collected.
Column chromatography
Analyticalgelfiltrationwascarriedoutusing aTSKgelG3000SWcolumn(0.75×30cm,Tosoh) equilibratedwitha50mMphosphatebuffer(pH7.0) containing0.2MNaClataflowrateof1.0ml/min.
Elution was monitored at 280 nm using a 807-IT integrator.Thecolumnwascalibratedusingalcohol dehydrogenase (150 kDa), bovine serum albumin (67kDa),ovalbumin(46kDa),andsoybeantrypsin
inhibitor(20.5kDa).Analyticalreverse-phaseHPLC wasperformedonaSepaxBio-C8column(0.46×25 cm,SepaxTechnologiesInc.)withalineargradient ofacetonitrilein0.1%trifluoroaceticacidataflow rateof1.0ml/min.Elutionwasmonitoredat220nm.
Sequence analysis
The amino acid sequences of proteins were determinedusinganautomaticproteinsequencer PPSQ21(Shimadzu).
Surface plasmon resonance analysis of binding kinetics
Kinetic measurements of the interaction between SSP and HSF were performed using a Biacore2000instrument(Biacore).Theflowcellsof CM5sensorchipswereactivatedwith100µlof0.2 M1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and0.05MN-hydroxysuccinimideataflowrateof10 µl/min.SSP-1ataconcentrationof4.0µMin10mM sodiumacetatebuffer(pH4.0)wasinjectedtoreach 350resonanceunits(RU).Theunreactedgroups wereblockedwith20µlof1Methanolamine(pH 8.5).Theanalytes,dilutedtovariousconcentrations inasolutioncontaining10mMHepes,0.15MNaCl, 3mMEDTA,and0.005%surfactantP20(pH7.4), wereinjectedfor90sduringtheassociationphase ataconstantflowrateof20µl/min.Thedissociation wassubsequentlyfollowedfor180satthesameflow rate.Thesurfaceofthesensorchipwasregenerated using a 10 mM Gly-HCl buffer (pH 2.0) after analytebinding.Thesensogramswerecorrectedby subtractionofthesignalfromthenegativecontrol surfaceandusedtocalculatetherateandaffinity constantsusingBIAevaluation4.1(BiacoreAB)and Origin5.0(Microcal).
Results
Identification of the SSP-1 binding protein in T.
flavoviridis serum
InordertoidentifytheSSP-1–bindingprotein, purified SSP-1 immobilized to a HiTrap column wasusedforaffinitychromatography.Theserum from T. flavoviridis was first filtered through an ultrafiltrationmembranewithanominalmolecular weightcut-offof30,000inacidicconditionstoremove SSPs. The SSP-deficient serum was then applied
tothecolumn.Thecolumnwaswashedwiththe startingbuffer,theabsorbedmaterialswereeluted withabufferwithapHof3.0(datanotshown).
Theelutedproteinwasnothomogenous,andSDS- PAGEshowed2bandsat48kDaand43kDa(Fig.
1,lane1intheinset).Theaffinity-purifiedsample wasthensubjectedtoreverse-phaseHPLConaC8 column(Fig.1).Themajorpeakcorrespondedto aproteinwithamolecularmassof48kDa,which wasconsistentwiththeSDS-PAGEresults(Fig.1, lane2intheinset).Theexactmassofthemajor proteinwas48,026.3Da.Theaminoacidsequence revealed by sequence analysis—SQVRGDLEXD DKEAKNWADD—identicalto20HSFN-terminal residues,anantihemorrhagicfactorisolatedfromT.
flavoviridisserum［12,19］.
Binding analysis between HSF and SSP-1
InordertoconfirmthatHSFisanSSP-1carrier Fig.1.Purification of the affinity-purified serum fractionbyreverse-phaseHPLConacolumn of SepaxBio-C8 column (0.46 × 25 cm).
Acetonitrilegradientof20–50%isshownina dashedline.Fractionsshownbyabarwere pooled.SDS-PAGEanalysisofHPLC-purified protein is shown in the inset. M, molecular weightmarkers;lane1,SSP-1–HiTrap-purified material;lane2,HPLC-purifiedproteininthe mainpeak.
proteininT. flavoviridisserum,weanalyzedthe bindingbetweenHSFandSSP-1byanalyticalgel filtration.HSFandSSP-1wereelutedat7.28and9.98 min,respectively(Fig.2A).Themolecularweight ofHSFwasestimatedtobeapproximately100,000 by using a calibration curve; this indicates that HSFisdimeric.ThemixturesofHSFandvarious amountsofSSP-1in50mMphosphatebuffer(pH 7.0) containing 0.2 M NaCl were analyzed by gel filtration.WhenthemolarratioofSSP-1/HSFwas less than 1, the SSP-1 peak was not present, but ashoulderwasobservedat7.11min.Whenmore than1equivalentofSSP-1wasmixedwithHSF,the 7.11-minpeakwasprominent,andapeakappeared at9.98minbecauseofexcessSSP-1.Whentheeluate containingthe7.11-minpeakshowninFig.2Awas recoveredanddirectlysubjectedtoreverse-phase HPLC, both SSP-1 and HSF were detected (Fig.
2B). Thus, the 7.11-min peak was attributable to SSP-1andHSFbinding.Themolecularweightofthe complexwasestimatedtobeapproximately120,000 on the basis of a calibration curve. This value agreedwellwiththecomplexformationbetweenan HSFdimerand2moleculesofSSP-1,indicatinga stoichiometryof1:1onthebasisofsubunits.
Kinetics of complex formation between HSF and SSP-1
TheSSP-1–HSFcomplexisstableanddoesnot dissociateduringgelfiltration,asdescribedabove.
TheaffinitybetweenSSP-1andHSFwasfurther examinedbycontinuousmonitoringofassociation and dissociation by means of surface plasmon resonance technology. The affinity was evaluated
asillustratedinFig.2C.Sensogramsrevealedthat HSFhasahighaffinityforSSP-1,asindicatedby alowdissociationrate.Globalevaluationusingthe 1:1Langmuirbindingmodelyieldedkineticbinding parameters (Table 1). The association rate (ka) of SSP-1 for HSF was considerably higher than the dissociation rate (kd). The calculated mean dissociationconstant(KD)was2.5×10-7M(SD1.1
×10-7),suggestingthattheinteractionbetweenHSF andSSP-1ishighlyspecific.
Binding of HSF to other SSPs
Sequence similarity between the 5 SSPs suggestedthatHSFcouldalsobindtootherSSPs.
WhenaslightexcessofeachofSSP-2,SSP-3,SSP-4, orSSP-5wasmixedwithHSF,the7.28-minpeak correspondingtoHSFdisappearedandanewpeak, attributable to the SSP-HSF complex, appeared at7.12,7.15,7.13,and7.06min,respectively(Fig.
3).TheseresultsrevealedthatalloftheSSPsare capableofbindingtoHSF.Thebindingpotencyof someSSPstoHSFwasalsodeterminedbysurface plasmonresonanceusingaBiacore2000(datanot shown).AffinitiesbetweenHSFandotherSSPswere similar to that between HSF and SSP-1 complex (Table1).
Effect of HSF–SSP-1 complex on the proteolytic activity of SVMPs
Like HSF, SSP-1 inhibits brevilysin H6, an SVMPisolatedfromGloydius blomhoffi brevicaudus venom［20］. However, the inhibition was weak and thebindingofSSP-1toH6wasnotstoichiometric.
To explore the reason behind complex formation
Table1.SurfaceplasmonresonanceanalysesofthebindingofHSFtoimmobilizedSSPs.
Ligand Analyte ka(1/Ms) kd(1/s) KD(M)
SSP-1 HSF 2.0 ± 1.1 × 103 5.3 ± 3.2 × 10-4 2.5 ± 1.1 × 10-7(1.4–4.2 × 10-7) SSP-2 HSF 1.8 ± 0.9 × 103 4.6 ± 2.0 × 10-4 2.6 ± 0.3 × 10-7(2.4–2.8 × 10-7) SSP-3 HSF 2.2 × 103 4.6 × 10-4 2.0 × 10-7
SSP-5 HSF 0.76 × 103 3.5 × 10-4 4.6 × 10-7
Thekineticsoftheinteractionbetween4SSPsandHSFwasevaluatedwithaBiacore2000instrument.Ligands wereimmobilizedonasurfacethatwaschallengedwithdifferentconcentrationsofHSF,andassociationand dissociationrates(kaandkd)weremeasured.Theaveragevaluesof2independentexperimentsareshownfor SSP-1andSSP-2.Therangeoftheequilibriumdissociationconstants(KD)aregiveninparentheses.
0.4
Fig.2.BindingofSSP-1toHSF.(A)Analyticalgelfiltrationanalysis.AconstantamountofHSF(90μM)was addedtovaryingamountsofSSP-1andthemixtureswereappliedtoaTSKgelG3000SW(0.75×30cm) columnina50mMphosphatebuffer(pH7.0)containing0.2MNaCl.Arrowsindicatetheretentiontimesof molecular-weightmarkerproteins.Retentiontimesofmajorpeaksareshowninparentheses.(B)Analysis ofSSP-1–HSFcomplexbyreverse-phaseHPLConacolumnofSepaxBio-C8column(0.46 ×25cm).(C) Surface plasmon resonance analyses of the binding between SSP-1 and HSF. Representative data of equilibriumbindingareshown.SeriallydilutedHSF(from500nMto31.5nM)wereinjectedataflowrate of20μl/minthroughflowcellswithSSP-1immobilizedtothesensorchipsurface.
betweenSSP-1andHSF,weinvestigatedtheeffect ofSSP-1–HSFcomplexontheproteolyticactivity ofseveralSVMPs.SSP-1alonedidnotinhibitthese enzymesatall(datanotshown).AsshowninFig.4, HSFinhibitedtheactivitiesofseveralSVMPssuchas HR1A,HR1B,HR2a,andHR2b,buttheinhibitionwas notinfluencedbythepresenceofSSP-1.
Discussion
Since SSPs are low-molecular-mass proteins, theymaybepresentinahigh-molecular-massform to prevent excretion in the urine. Therefore, the presenceofabindingproteinwithahighmolecular mass was expected. Serotriflin was isolated from T. flavoviridisserumasacandidateSSP–binding protein［8］.Structuralanalysisrevealedthatserotriflin
isamemberoftheCRISPfamily.TheCRISPfamily ofproteinsshowsawidephylogeneticdistribution, rangingfromplantstomammals［21］.Indeed,SSP-2 forms a noncovalent complex with serotriflin and triflin［1,8］.Recently,wehavefoundthatSSP-5also binds to serotriflin and triflin (unpublished). As discussed in a previous paper［1］, SSP-2 seems to preventself-injurybyaccidentalleakingofvenom intothebloodsincetriflinpossessesneurotoxin-like activity［11］.
UnlikeSSP-2andSSP-5,however,otherSSPs could not effectively bind to serotriflin. Since all the SSPs do not aggregate in aqueous buffers but exist in high-molecular-mass forms in the serum, another binding protein may be present.
Affinity chromatography of SSP-deficient serum onanSSP-1–HiTrapcolumnrevealedthatHSF,an Fig.3.BindinganalysisofHSFtoSSP-2(A),SSP-3(B),SSP-4(C),andSSP-5(D).Anequalamountofeachof the4SSPs(100μM)wasaddedto5μlofHSF(90μM).Themixtureswereanalyzedbygelfiltration HPLConacolumnofTSKgelG3000SW(0.75×30cm)in50mMphosphatebuffer(pH7.0)containing0.2 MNaClataflowrateof1.0ml/min.RetentiontimesofSSPsandSSP-HSFcomplexesareshownin parentheses.
antihemorrhagicfactor,maybetheSSP-1carrier protein. Furthermore, the analytical gel filtration proved that HSF could bind to all the SSPs, and thatHSFispresentasa100-kDahomodimer(Figs.
2and3).BJ46a,anHSFproteinhomologisolated fromtheserumofBothrops jararaca,alsoformsa homodimer［22］.ThemolecularweightoftheSSP-HSF complexsuggestedthatthestoichiometryofHSFand SSPinthecomplexis1:1onthebasisofsubunits.
Thisisincontrasttothe2:1stoichiometrybetween humanPSP94andCRISP-3.Preliminaryexperiments showedthatMSF,anotherhomologofHSFinG.
blomhoffi brevicaudusserum［23］,couldalsobindtoall theSSPsinaneutralbuffer(datanotshown).
ThecomplexesbetweenSSPsandtheirbinding proteinscoulddissociateinacidicsolutionswithpH below4［1］.HSFandSSPsareconsiderablystableat extremepHvaluesaswellashightemperatures［1,13］. ThehighstabilityofSSPsmayberesponsiblefor theirtightconformationasestimatedfromthe3D structureofhumanandporcinePSP94proteins［7］. TheacidstabilityofbothHSFandSSPsfacilitated theBiacoreexperiments;repeatedwashingofthe sensor chip conjugating the SSPs with a buffer atpH2didnotleadtoadecreaseinthebinding potencytoHSF.HSFboundtoSSPswithdissociation constantsof2.0×10–7–6.6×10–7M.Theaffinitywas approximately10timeslowerthanthatofSSP-2to triflinandserotriflin［8］.Thissuggeststhatserotriflin isabetterSSP-2carrierproteinthanHSF.Sincethe serumconcentrationofHSFismuchhigherthan thatofserotriflin,SSP-2seemstobepresentasa complex with either HSF or serotriflin. However, apreliminaryexperimentrevealedthattheSSP-2–
serotriflincomplexcouldfurtherbindHSF,andvice versa,toformaternarycomplex(datanotshown).
ThisindicatesthatSSP-2canbindCRISPandHSFat 2distinctregions.
Although SSP-1 is the major component amongthe5SSPs,itsphysiologicalfunctionisstill speculative. It weakly inhibits brevilysin H6［20］
but not other SVMPs. The significance of SSP-1–
HSFcomplexformationisnotyetclear.SSP-1may playaroleintheantihemorrhagicactivityofHSF as a regulator of its inhibitory activity by either enhancement or suppression. For example, SSP-1 can help the binding of HSF to the P-I class of SVMPs,whicharelesseffectivetargetsofHSF［19］.
AnotherpossibilityistheinhibitionofHSFactivity onphysiologicallyimportantmetalloproteinasessuch as matrix metalloproteinases that are present in thetissuesandbloodinsnakes.IfHSFcaninhibit theproteolyticactivityoftheseenzymes,itwillnot befavorablefortheanimal.SSP-1maycounteror weakenthisdeleteriouseffectofHSF.
SSP-1didnotaffectHSF-inhibitionofSVMPs (Fig.4).ThisindicatesthatHSFfunctionsmerely asacarrierofSSP-1inT. flavoviridisserum.Itis alsolikelythatthetargetmoleculeoftheSSP-1–HSF
Fig.4.EffectofSSP-1ontheHSFinhibitoryactivity to several SVMPs. (A) Inhibitory effect on HR1A(circles,15μg/ml)andHR1B(triangles, 15μg/ml). (B) Inhibitory effect on HR2a (circles, 20μg/ml) and HR2b (triangles, 10 μg/ml). Open and closed symbols show the resultsintheabsenceorpresenceofSSP-1(8 μg/ml).
complexisnotSVMP.Sincemanyproteinsinthe serumofcrotalidsnakestargetthetoxicsubstances intheirownvenom,SSP-1mayinteractwithsome venomousproteinsinthefreestateorinanHSF- boundform.
References
［ 1 ］ N. Aoki, A. Sakiyama, M. Deshimaru, S.
Terada,Biochem. Biophys. Res. Commun.359 330-334(2007).
［ 2 ］ N.Aoki,H.Matsuo,M.Deshimaru,S.Terada, Gene4267-14(2008).
［ 3 ］ N.G.Seidah,N.J.Arbatti,J.Rochemont,A.
R.Sheth,M.Chretien,FEBS Lett.175349-355 (1984).
［ 4 ］ M.Mäkinen,C.Valtonen-Andre,A.Lundwall, Eur. J. Biochem.264407-414(1999).
［ 5 ］ C. Lazure, M. Villemure, D. Gauthier, R. J.
Naude,M.Mbikay,Protein Sci.102207-2218 (2001).
［ 6 ］ I.Wang,Y.C.Lou,K.P.Wu,S.H.Wu,W.C.
Chang,C.Chen,J. Mol. Biol.3461071-1082 (2005).
［ 7 ］ H. Ghasriani, K. Teilum, Y. Johnsson, P.
Fernlund, T. Drakenberg, J. Mol. Biol.362 502-515(2006).
［ 8 ］ N.Aoki,A.Sakiyama,K.Kuroki,K.Maenaka, D.Kohda,M.Deshimaru,S.Terada,Biochim.
Biophys. Acta1784621-628(2008).
［ 9 ］ J.R.Reeves,J.W.Xuan,K.Arfanis,C.Morin, S. V. Garde, M. T. Ruiz, J. Wisniewski, C.
Panchal,J.E.Tanner,Biochem. J.385105-114 (2005).
［10］ L. Udby, A. Lundwall, A. H. Johnsen, P.
Fernlund,C.Valtonen-Andre,A.M.Blom,,H.
Lilja,N.Borregaard,L.Kjeldsen,A.Bjartell, Biochem. Biophys. Res. Commun.333555-561 (2005).
［11］ Y. Yamazaki, H. Koike, Y. Sugiyama, K.
Motoyoshi,T.Wada,S.Hishinuma,M.Mita,T.
Morita,Eur. J. Biochem.2692708-2715(2002).
［12］ Y. Yamakawa, T. Omori-Satoh, J. Biochem.
112583-589(1992).
［13］ M.Deshimaru,C.Tanaka,A.Tokunaga,M.
Goto,S.Terada,Fukuoka Univ. Sci. Rep.33 45-53(2003).
［14］ T.Takahashi,A.Ohsaka,.Biochim. Biophys.
Acta20765-75(1970).
［15］ T.Omori-Satoh,S.Sadahiro,Biochim. Biophys.
Acta580392-404(1979).
［16］ S.C.Gill,P.H.vonHippel,Anal Biochem.182 319-326(1989).
［17］ S.S.Twining,Anal. Biochem.14330-34(1984).
［18］ U.K.Laemmli,Nature227680-685(1970).
［19］ M. Deshimaru, C. Tanaka, F. Kazuya, A.
Narumi, S. Terada, S. Hattori, M. Ohno, Toxicon46937-945(2005).
［20］ S. Fujimura, K. Oshikawa, S. Terada, E.
Kimoto,J. Biochem.128167-173(2000).
［21］ Y. Yamazaki, F. Hyodo, T. Morita, Arch.
Biochem. Biophys.412133-141(2003).
［22］ R.H.Valente,B.Dragulev,J.Perales,J.W.
Fox, G. B. Domont, Eur. J. Biochem.268 3042-3052(2001).
［23］ N. Aoki, K. Tsutsumi, M. Deshimaru, S.
Terada,Toxicon51251-261(2007).
