ilable at ScienceDirect

Toxicon 54 (2009) 128–137

Contents lists ava

Toxicon

journal homepage: www.elsevier .com/locate/ toxicon

Cysteine-rich secretory proteins in snake venoms form high affinitycomplexes with human and porcine b-microseminoproteins

Karin Hansson 1, Margareta Kjellberg, Per Fernlund*

Department of Laboratory Medicine (Division of Clinical Chemistry), Lund University, University Hospital MAS, SE-205 02 Malmo, Sweden

a r t i c l e i n f o

Article history:Received 3 December 2008Received in revised form 12 March 2009Accepted 23 March 2009Available online 31 March 2009

Keywords:Cysteine-rich secretory proteinb-MicroseminoproteinProstate cancerProtein complexPSP94Seminal plasmaSnake venom

Abbreviations: CRISP, cysteine-rich secretory proMSP; MS, mass spectrometry; MSP, b-microsemnanoelectrospray ionization; pMSP, porcine MSP; Ptory protein of 94 amino acids; PSPBP, PSP94-bresponse unit; SDS-PAGE, sodium dodecyl sulfateelectrophoresis; SGP28, specific granule protein of 2

* Corresponding author. Tel.: þ46 40 331459; faxE-mail address: per.fernlund@med.lu.se (P. Fernl

1 Present address: Department of ImmunotechnoBiomedical Centre D13, SE-221 84 Lund, Sweden.

0041-0101/$ – see front matter � 2009 Elsevier Ltddoi:10.1016/j.toxicon.2009.03.023

a b s t r a c t

b-Microseminoprotein (MSP), a 10 kDa protein in human seminal plasma, binds humancysteine-rich secretory protein-3 (CRISP-3) with high affinity. CRISP-3 is a member of thefamily of CRISPs, which are widespread among animals. In this work we show that humanas well as porcine MSP binds catrin, latisemin, pseudecin, and triflin, which are CRISPspresent in the venoms of the snakes Crotalus atrox, Laticauda semifasciata, Pseudechisporphyriacus, and Trimeresurus flavoviridis, respectively. The CRISPs were purified from thevenoms by affinity chromatography on a human MSP column and their identities weresettled by gel electrophoresis and mass spectrometry. Their interactions with human andporcine MSPs were studied with size exclusion chromatography and surface plasmonresonance measurements. The binding affinities at 25 �C were between 10�10 M and10�7 M for most of the interactions, with higher affinities for the interactions with porcineMSP compared to human MSP and with Elapidae CRISPs compared to Viperidae CRISPs.The high affinities of the bindings in spite of the differences in amino acid sequencebetween the MSPs as well as between the CRISPs indicate that the binding is tolerant toamino acid sequence variation and raise the question how universal this cross-speciesreaction between MSPs and CRISPs is.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

b-Microseminoprotein (MSP) is a 10 kDa disulfide richprotein present in high concentration in human seminalplasma (Lilja and Abrahamsson, 1988). MSPs have beendescribed in many vertebrates and the amino acidsequence is highly variable (Lazure et al., 2001; Wanget al., 2005b). Human MSP (hMSP), which has also been

tein; hMSP, humaninoprotein; nanoESI,SP94, prostate secre-inding protein; RU,-polyacrylamide gel8 kDa.

: þ46 40 336286.und).logy, Lund University,

. All rights reserved.

referred to as PSP94, is produced in the prostate gland(Doctor et al., 1986) and has also been identified in severalother tissues including tracheal and gastric mucosa(Weiber et al., 1990). Being a major secretory product of theprostate gland, MSP has been studied especially withrespect to prostate cancer (see Reeves et al., 2006) and theinterest was intensified recently, when two genome-wideassociation studies reported that a locus on chromosome10, which includes the gene for MSP, is associated withsusceptibility to prostate cancer (Eeles et al., 2008; Thomaset al., 2008).

A number of functions of MSP have been suggested butnone supported by convincing experimental evidence(Lazure et al., 2001). A possible clue to the function was thediscovery that hMSP binds human cysteine-rich secretoryprotein-3 (CRISP-3) (Udby et al., 2005).

CRISP-3 belongs to a family of proteins described inmammals, snakes, lizards (see Roberts et al., 2007) and

K. Hansson et al. / Toxicon 54 (2009) 128–137 129

lamprey (Ito et al., 2007). They are distinguished by aminoacid sequence characteristics such as molecular sizes of 22–25 kDa (200–250 amino acids), a high degree of amino acidsequence similarity, and a highly conserved specific patternof 16 cysteines (Roberts et al., 2007).

In mammals four different CRISPs have been identified(see Roberts et al., 2007) and a recent study in the mousehas shown that the CRISPs are widely expressed in theorganism (Reddy et al., 2008). In the human there are threeCRISPs (Kratzschmar et al., 1996). Of these, CRISP-1 (acidicepididymal glycoprotein or AEG) and CRISP-2 (testisspecific protein 1 or TPX-1) are expressed predominantly inthe male reproductive tract and many studies indicatea function in sperm maturation and the fertilization process(Roberts et al., 2007; Gibbs et al., 2007). CRISP-3 was firstdemonstrated as an androgen-dependent transcript in themouse salivary gland (Mizuki and Kasahara,1992; Haendleret al., 1993) and its eight-exon gene was later characterized(Schwidetzky et al., 1995). The human CRISP-3 protein wasfirst isolated from neutrophilic granulocytes, then calledSGP28 (Kjeldsen et al., 1996) and later further characterizedby Udby et al. (2002b). CRISP-3 is released from activatedgranulocytes (Udby et al., 2002a) and it is also expressed inother places, e.g. in the salivary gland, pancreas and pros-tate (Kratzschmar et al., 1996; Udby et al., 2002b).

The function of human CRISP-3 is unknown but severalCRISPs have been shown to be ion-channel inhibitors. Thishas been shown for CRISP-2 from the mouse (Gibbset al., 2006) and also for many CRISPs present in snakevenoms: pseudechetoxin (Brown et al., 1999), pseudecin(Yamazaki et al., 2002a), ablomin, latisemin, triflin (Yama-zaki et al., 2002b) and natrin (Wang et al., 2005a, 2006).Supposing that the putative function of human CRISP-3 ismediated via ion-channel inhibiting effects and consid-ering the fact that human CRISP-3 is not only bound withhigh affinity by MSP in seminal plasma (Udby et al., 2005)but also by a1B-glycoprotein in blood plasma (Udby et al.,2004) it is natural to think that MSP and a1B-glycoproteinare important components of a system for protectionagainst injurious systemic effects of CRISPs. MSP–CRISPinteractions are not limited to the human. Recently, a newMSP (named SSP-2) was isolated from the blood of Tri-meresurus flavoviridis and demonstrated to bind triflin, i.e.the CRISP present in the snakes’ own venom (Aoki et al.,2007). SSP-2 has also been shown to form a complex withserotriflin, a new CRISP found in the blood of T. flavoviridis(Aoki et al., 2008).

Whether MSP binds human CRISP-1 and -2 is not known;that knowledge must await the isolation or recombinantproduction of these proteins. CRISPs from snake venoms aremore easily available; several have been isolated and char-acterized (see Yamazaki and Morita, 2004).

Up to now, interaction between an MSP and a CRISP hasonly been observed in two species, i.e. in the human andthe snake T. flavoviridis. These interactions are both intra-species, i.e. they take place between MSPs and CRISPs fromthe same species. An unpublished observation that porcineMSP also could bind human CRISP-3 raised the questionhow general the reaction is between MSPs and CRISPs, andif mammal MSPs could bind CRISPs from species as diver-gent as reptiles. In this work we show that hMSP as well as

porcine MSP (pMSP) binds CRISPs present in the venomsfrom two Elapidae and two Viperidae snakes and providea thorough characterization of the interactions.

2. Materials and methods

2.1. Materials

Crude venoms from Crotalus atrox, Laticauda semi-fasciata, Pseudechis porphyriacus, and T. flavoviridis wereobtained as lyophilized powders (Sigma–Aldrich Corp., St.Louis, MO, USA). Human and porcine MSPs were purifiedfrom human and boar seminal plasma, respectively, withthe method described for pMSP (Fernlund et al., 1994). MSPaffinity columns were prepared by coupling about 8 mg ofhMSP to a 1 ml HiTrap NHS-activated HP column (GEHealthcare BioSciences AB, Uppsala, Sweden) according tothe instructions of the manufacturer. The coupling yields,calculated from the differences in concentrations of hMSP(determined by HPLC) in the coupling solution and thewash solution, were about 90%.

2.2. Purification of venom CRISPs

One hundred mg of lyophilized crude venom wasextracted at room temperature with gentle shaking for15 min in 5 ml of a 50 mM sodium phosphate buffer pH 7.4containing 0.1 M NaCl. After centrifugation at 10,000� g for30 min, the clear supernatant was passed through a Min-isart 0.20 mm filter (Sartorius AG, Goettingen, Germany)and then pumped through an hMSP-affinity column. Thecolumn was washed, first with 10–20 ml of a 50 mMsodium phosphate buffer pH 7.4 containing 0.1 M NaCl andthen with 10–20 ml of the same buffer but with 0.5 M NaCl.Finally the bound material was eluted with 0.2 M glycinepH 2.0. The flow rate was 0.25 ml/min throughout theexperiment. The absorbance of the effluent was continu-ously monitored at 280 nm with a 2238 Uvicord SII (LKB,Bromma, Sweden) and the part corresponding to the peakof absorbance was collected (usually about 2–3 ml), dilutedto 6 ml with 50 mM sodium phosphate buffer pH 7.4 con-taining 0.5 M NaCl, and concentrated to 500 ml in a Vivaspin6 ultra filtration spin column fitted with a 5 kDa cut offmembrane (Vivascience Sartorius AG, Goettingen,Germany). The concentrated fraction was subjected toa final purification step by chromatography on a Superdex75 10/300 GL column (GE Healthcare Biosciences AB),essentially as described for the analysis of complexformation (Section 2.4).

2.3. SDS-PAGE and protein identifications by massspectrometry

Sodium dodecyl sulfate-polyacrylamide gel electro-phoresis (SDS-PAGE) was carried out under reducingconditions according to Laemmli (1970) in a 4% (w/v)stacking, 12% (w/v) separating polyacrylamide gel. The gelwas stained with GelCode blue stain reagent (Pierce,Rockford, IL, USA). The molecular size marker, Low Molec-ular Weight-SDS, was obtained as a ready-made mixture(Amersham Biosciences AB, Uppsala, Sweden). For

K. Hansson et al. / Toxicon 54 (2009) 128–137130

identifications by mass spectrometry, proteins were sepa-rated by SDS-PAGE and silver stained using the methoddescribed previously (Shevchenko et al., 1996). The gelpieces containing the proteins were digested with trypsinand the generated fragments were extracted followinga protocol slightly modified from that of Shevchenkoet al. (1996). Briefly, the excised gel pieces were rinsedtwice with 50% acetonitrile (v/v) and dehydrated in 100%acetonitrile, followed by reduction of the cystines with10 mM dithiothreitol/0.1 M NH4HCO3 and subsequentS-carbamidomethylation of free cysteines with 55 mMiodoacetamide in 0.1 M NH4HCO3. For enzymatic digestionof the reduced and alkylated protein the sample was dis-solved in 50 mM NH4HCO3 buffer pH 7.9 and digested withtrypsin (substrate to enzyme ratio of 20:1). The reactionmixture was incubated on ice for 45 min followed by 37 �Cfor 18 h and then the reaction was terminated by freezing at�20 �C. For mass spectrometry analysis the peptidesgenerated were micropurified and concentrated on Zip-TipC18 columns (Millipore Corporation, Billerica, MA, USA)and eluted with 70% (v/v) aqueous methanol containing 1%(v/v) formic acid. Nanoelectrospray ionization (nanoESI)mass spectrometry experiments were carried out on an APIQSTAR Pulsar-i quadrapole/time-of-flight mass spectrom-eter (Applied Biosystems/MDS Sciex, Toronto, Canada)equipped with a nanoESI source (MDS Protana, Odense,Denmark). The samples were sprayed from a silver-coatedglass capillary (New Objective, Woburn, MA, USA). Spectrawere obtained in positive ion mode with an ion sprayvoltage of 850–1000 V. The acquisition and the de-convo-lution were performed on an AnalystQS Windows PC datasystem. Bioanalyst (version 1.0) software (Applied Bio-systems/MDS Sciex) was used to analyze the mass spectra.Parent ions (identified in a time-of-flight MS survey scan)for tandem mass spectrometry (MS/MS) analysis wereselected in Q1, and product ions were generated in Q2 usingN2 as collision gas and collision energies of 14–36 eV toobtain amino acid sequence data.

2.4. Analysis of complex formation by gel chromatography

Venom CRISP (1.0 nmol) was incubated at roomtemperature (22–24 �C) for 15 min with 0.5, 1.0, or 2.0 nmolof MSP in 100 ml of a 10 mM sodium phosphate buffer pH7.4 containing 0.5 M NaCl. The mixture was then analyzedby gel chromatography on a Superdex 75 10/300 GL column(GE Healthcare Biosciences AB) at room temperature usingthe same buffer delivered at a flow rate of 0.48 ml/min bya Waters 650E Advanced Protein Purification System(Waters Corporation, Milford, MA, USA). The 100 ml samplewas applied via a 500 ml sample loop. The effluent waspassed through a Waters 486 Tunable Absorbance Detectorand the absorbance was continuously monitored at280 nm.

2.5. Kinetic studies by surface plasmon resonance analysis

The interactions between CRISPs and MSPs wereexamined by surface plasmon resonance analysis on a Bia-core 2000 instrument (Biacore International AB, Uppsala,Sweden). Catrin, latisemin, pseudecin and triflin were

coupled to a carboxymethylated dextran matrix sensor chip(CM5) (Biacore International AB) to about 2000 responseunits (RU) using a concentration of 20 mg/ml in 10 mMsodium acetate pH 4.5. Porcine MSP was coupled to about1500 RU at 20 mg/ml in 10 mM sodium acetate pH 5.0 andhMSP to about 1500 RU at 200 mg/ml in 10 mM sodiumacetate pH 4.0. In the binding experiments the runningbuffer used was 10 mM Hepes pH 7.5, containing 0.5 MNaCl, 3 mM EDTA, and 0.005% TWEEN 20. The temperaturewas 25 �C, the flow rate 30 ml/min and the association anddissociation times were 180 s and 600 s, respectively.

Increasing concentrations of human and porcine MSPs(2.5–100 nM) and of the CRISPs (0.625–320 nM) wereinjected and the chips were regenerated with 2� 20 mlof 0.1 M glycine pH 10.0 containing 0.5 M NaCl, followedby 2� 20 ml of 0.2 M glycine pH 2.0 containing 0.5 MNaCl, at a flow rate of 10 ml/min. Data were evaluatedusing the Biaevaluation 3.0 software program (BiacoreInternational AB). The 1:1 Langmuir model (Aþ L 4 AL)was used for fitting the experimental data except forcatrin, for which the two-state (conformational change)model (Aþ L 4 AL 4 AL*) was used.

2.6. Determination of protein concentration

The concentration of MSPs and purified venom CRISPswas determined by measurement of the absorbance at280 nm and using the following theoretical extinctioncoefficients (M�1 cm�1): for hMSP 17585, pMSP 16095,catrin 47870, latisemin 47410, pseudecin 48900, and triflin52340. The extinction coefficients were derived from theamino acid sequences of the proteins using the ProtParamtool at the Swiss-Prot database (http://www.expasy.org/tools/protparam.html) (Gill and von Hippel, 1989).

3. Results

3.1. Isolation of CRISPs from snake venoms

From the earlier finding that hMSP binds with highaffinity to human CRISP-3 (Udby et al., 2005) and the factthat human CRISP-3, according to the amino acid sequence,seems to be related to CRISPs isolated from snake venoms(Fig. 1) it was decided to test whether MSP also binds thesesnake proteins. Venoms from four different snakes, twoViperidae (C. atrox and T. flavoviridis) and two Elapidae(L. semifasciata and P. porphyriacus), were selected for thestudy. Extracts of the crude venoms were applied to anaffinity column with coupled hMSP. After washing, thecolumn was eluted with a glycine buffer pH 2.0. The resultswere similar for all four snake species; here we show thatobtained for T. flavoviridis (Fig. 2a).

As a final purification step, the materials eluted from thehMSP-affinity column were subjected to gel chromatog-raphy on a Superdex 75 column. Also here the results weresimilar for all four species; that for T. flavoviridis is shown inFig. 2b. The isolated venom proteins were pure according toSDS-PAGE (Fig. 3). The migrations indicated molecular sizesof 25–29 kDa for the four proteins. The yields of the puri-fied proteins obtained with the combined affinity and gelchromatography procedure were 2.1, 10.5, 5.0, and 8.6 mg

Fig. 1. Comparison of the amino acid sequences of human CRISP-3 and CRISPs from the venoms of four species of snakes. Gaps have been inserted to allow properalignment of the sequences. Residues that are identical in all five proteins are shaded. The parts of the sequences that constitute the strands of the main b-pleatedsheet of pseudecin and triflin (for which 3D structures have been determined) are boxed.

K. Hansson et al. / Toxicon 54 (2009) 128–137 131

per g of crude venom for C. atrox, L. semifasciata, P. por-phyriacus, and T. flavoviridis, respectively.

3.2. Identification of CRISPs

The purified proteins were analyzed by nanoESI massspectrometry after trypsin cleavage (Fig. 4 and Table 1). Theanalysis identified more than two thirds of the amino acidsequence of each protein (84% for C. atrox, 82% for L. sem-ifasciata, 68% for P. porphyriacus, and 91% for T. flavoviridis).As expected, the results identified the four proteins ascatrin, latisemin, pseudecin, and triflin, all of which hadbeen purified and characterized earlier (Yamazakiet al., 2002a,b, 2003).

3.3. Demonstration of complex formation of the CRISPs withporcine and human MSPs

The isolated CRISPs were mixed in three different molarproportions with porcine or human MSP and, after a 15 minincubation, analyzed by gel chromatography on a Superdex75 column at room temperature. The results obtained withtriflin and pMSP, which were similar to those obtained withthe other three isolated CRISPs, are shown in Fig. 5: Incu-bation of triflin with half a molar amount of pMSP resultedin two peaks in the chromatogram (Fig. 5a), one elutingearlier than triflin and MSP, representing a triflin–MSPcomplex, and a second, eluting in the position for free tri-flin, representing non-complexed triflin. With equimolarproportions of triflin and MSP almost all material eluted inthe position for the complex (Fig. 5b) and little materialwas seen in the positions for free triflin or free MSP. Incu-bation of triflin with a double molar amount of MSPresulted in a peak corresponding to the triflin–MSPcomplex and a second peak eluting in the position for freeMSP (Fig. 5c). The chromatographies had to be performedin a buffer with high ionic strength (0.5 M NaCl) in order to

avoid retentions caused by interactions of the CRISPs withthe gel matrix. This interaction seemed to be especiallystrong for latisemin and pseudecin.

The results obtained in similar experiments with hMSPand the CRISPs (not shown) were principally identical tothose obtained with pMSP, although the apparent largersize of hMSP on gel chromatography made the interpreta-tion less obvious. Human MSP eluted from the Superdex 75column at a position corresponding almost to the size ofthe CRISPs, a phenomenon observed earlier (Udbyet al., 2005) and possibly due to a reversible dimerization ofthe free hMSP. However, the results were all consistentwith a 1:1 stoichiometry.

3.4. Characterization of the kinetics of the binding betweenthe CRISPs and MSPs

Surface plasmon resonance experiments were per-formed both with the MSPs and the snake CRISPs coupledto the CM5 sensor chip. Approximately the same dissocia-tion constants (kD) were obtained whether the CRISP or theMSP was coupled (Table 2). In order to avoid an unspecificinteraction between the CRISPs and the chip matrix, whichoccurred with the routine running buffer containing 0.1 MNaCl, a buffer containing 0.5 M NaCl was used. However,despite the high concentration of NaCl in the runningbuffer the interaction between pseudecin and MSP couldnot be determined with pseudecin in the fluid phase,probably due to an unspecific binding of pseudecin to thechip dextran surface.

For the interactions of the CRISPs with pMSP the highestaffinity (kD about 0.1 nM) was recorded for latisemin fol-lowed by pseudecin and triflin (kD about 1 nM) and catrin(kD about 10 nM) (Table 2). The affinities of the CRISPs forhMSP followed essentially the same order but they werelower than for pMSP (Table 2). For catrin, however, theaffinity was too low to be determined with the

50 10 15 60 65 70Elution volume (ml)

Ab

so

rb

an

ce at 280 n

m

0.4

0.0

0.8

1.2

1.6

0.20.0 0.4 0.6 0.8 1.0 1.2 1.4

Elution volume (fraction of column volume)

Ab

so

rb

an

ce at 280 n

m

0.4

0.0

0.8

1.2

1.6

a

b

Fig. 2. Isolation of CRISP from the venom of T. flavoviridis. (a) Affinitychromatography on an hMSP column. The material eluted with the glycinebuffer was collected as indicated by the bar. (b) Gel chromatography. Thecollected fraction from the affinity chromatography was subjected to chro-matography on a Superdex 75 column. Part of the eluant was collected asindicated by the bar.

K. Hansson et al. / Toxicon 54 (2009) 128–137132

experimental settings used; the kD was estimated to be>1 mM. All the binding data, except those for catrin, couldbe well fitted to the 1:1 Langmuir binding model. As anexample of the fittings attained, the binding of porcine andhuman MSPs to triflin is shown in Fig. 6. For catrin theexperimental data for the binding to pMSP did not fit a 1:1Langmuir model well but a good fitting was obtained usingthe two-state model, suggested in the Biaevaluation 3.0software program.

4. Discussion

The affinity chromatography on a column with coupledhMSP, a protein that previously only had been shown tobind a human CRISP, surprisingly worked also for theisolation of CRISPs from snakes. For each venom the chro-matography resulted in a seemingly pure, single compo-nent, which mass spectrometry identified as the CRISP thatearlier had been described for the respective venom. Themass spectrometry analysis covered a major part of thesequence of each venom protein and no data were at oddswith the published sequences: catrin (Yamazakiet al., 2003), latisemin (Yamazaki et al., 2002b), pseudecin(Yamazaki et al., 2002a) and triflin (Yamazaki et al., 2002b;Shikamoto et al., 2005).

The yields of CRISPs from crude venoms differedsomewhat from that obtained earlier. For pseudecin andtriflin they were almost the same as those obtained byYamazaki et al. (2002a,b) but the yield of catrin was onlyabout two thirds of that reported by Yamazaki et al. (2003),and that of latisemin was about double that reported earlier(Yamazaki et al., 2002b). Disregarding possible differencesin CRISP content in the starting materials, the lower yield ofcatrin could be explained by leakage from the affinitycolumn during the washing step caused by the compara-tively low affinity of catrin for hMSP. Latisemin on the otherhand had a very high affinity for hMSP making the affinitychromatography an efficient high yield purification.

Stan

dard

s

C.atrox

L.sem

ifascia

ta

P. porphyria

cu

s

T. fla

vovirid

is

latis

emin

pseu

deci

n

trifli

n

catri

n

66.045.030.0

20.1

97.0

14.4

Fig. 3. SDS-PAGE of snake venom extracts and the corresponding isolatedCRISPs. For each venom extract an amount corresponding to 40 mg of dryvenom and for each CRISP 2 mg was applied. The electrophoresis was per-formed under reducing conditions and the gel concentration was 12%. Themolecular masses (kDa) of the standards are indicated to the left.

Fig. 4. Amino acid sequences of catrin, latisemin, pseudecin, and triflin.The putative signal peptide sequences are shown in light grey and thepro-peptide sequence in dark grey. The underlined amino acid sequencesindicate tryptic peptide fragments identified from MS molecular massdetermination and the doubly underlined sequences denote the trypticpeptide fragments identified from MS/MS sequencing.

K. Hansson et al. / Toxicon 54 (2009) 128–137 133

Both in the gel chromatography experiments and theBiacore measurements an interaction of the CRISPs withthe gel matrix was noted when buffers with low saltconcentration were used. This interaction, which wasespecially strong for latisemin and pseudecin, manifesteditself as retentions on the gel chromatography columns(pseudecin did even not elute at all in a 0.1 M NaCl buffer)and abnormal binding curves in the Biacore experiments.This gel interaction was probably caused by the highisoelectric points of the proteins: latisemin has a theoret-ical pI of 8.6 and pseudecin 9.6. The remedy was to increasethe salt concentration to 0.5 M NaCl, which was usedthroughout all experiments in order to allow a comparisonof the results. Still, however, latisemin eluted later thanexpected from its molecular size, and the Biacoremeasurements could not be carried out with pseudecin asanalyte, i.e. in the fluid phase.

The purification of the venom CRISPs on the hMSP-affinity column indicated that hMSP has a broad specificity,not only binding a human CRISP but also CRISPs fromorganisms as different as reptiles. It was, therefore, ofinterest to test whether an MSP from another species alsobinds the venom CRISPs. Porcine MSP, which has a 51%amino acid sequence identity with hMSP, was thereforeincluded in the studies of the complex formation of MSPwith the snake CRISPs. The gel chromatography experi-ments showed that both MSPs formed stoichiometric 1:1complexes with all four snake CRISPs and the Biacoremeasurements revealed that the affinity of pMSP for thevenom CRISPs was even higher than that of hMSP.

The Biacore measurements showed that, except for thebinding of catrin to hMSP, the affinities (expressed as kD)were between 10�10 and 10�7 M with higher affinities forthe interactions with pMSP compared to hMSP and withElapidae CRISPs (latisemin and pseudecin) compared toViperidae CRISPs (catrin and triflin). Of the four venomCRISPs, latisemin had the highest affinity for binding tohMSP (about 3�10�9 M) but this is still about 50 timeslower than for the binding of human CRISP-3 to hMSP(6.5�10�11 M) (Udby et al., 2005) and not even the bindingof latisemin to pMSP (9�10�11 M) had such a high affinity.

Catrin showed a weaker binding to the MSPs than thethree other CRISPs and the interaction probably involveda more complicated mechanism of binding than a simple1:1 Langmuir adsorption. A more detailed study is requiredto reveal the mechanism of interaction. The apparent kD forthe binding of catrin to hMSP could not be measured withprecision within the range of ligand concentrations used inthe Biacore experiments. The apparent kD was estimated tobe above 1 mM, an affinity that apparently was sufficient forisolation of the protein on an hMSP column. However, asmentioned above, catrin probably leaked from the affinitycolumn causing a yield lower than expected.

The fact that two MSPs, as different in amino acidsequence as the human and the porcine forms, bind CRISPsfrom species as different as humans and snakes suggeststhat the interaction between MSP and CRISPs occurs viaconserved structural elements. A typical CRISP is a two-domain protein consisting of a larger N-terminal domainand a smaller C-terminal cysteine-rich domain, the twoconnected via a hinge region (Roberts et al., 2007). Thedistal part of the C-terminal domain of CRISPs has a simi-larity with toxin proteins found in non-vertebrates (Robertset al., 2007) and the C-terminal domain is thought to beresponsible for the ion-channel inhibiting activity of theCRISPs. This has been explicitly shown for CRISP-2 from themouse (Gibbs et al., 2006). The N-terminal domain containsthe highly conserved CAP domain (also named PR-1 or SCPdomain) present in many different proteins from organismsas divergent as mammals, reptiles, insects, plants, and fungi(Roberts et al., 2007; Gibbs et al., 2008). The CAP domain isalso present in the PSP94-binding protein (PSPBP), a 70/95-kDa, glycosylated human plasma protein that has beenshown to bind hMSP (Reeves et al., 2005). PSPBP does notcontain any other part with sequence similarities in theCRISPs indicating that the CAP containing N-terminaldomain of the CRISPs is responsible for the binding of MSP.A support for this model has recently been obtained by

Table 1Theoretical and observed monoisotopic masses of peptide fragments derived from tryptic digestion of purified triflin.

Tryptic fragment Sequence Theoretical molecularmass (Da)

Observed molecularmass (Da)

Sequenced by MS/MS

20–29 NVDFDSESPR 1164.51 1164.52a, b þ30–45 KPEIQNEIIDLHNSLR 1918.02 1918.06b, c, d þ47–57 SVNPTASNMLK 1160.59 1160.60a, b þ58–70 MEWYPEAAANAER 1536.67 1536.69a, b þ71–74 WAYR 594.29 594.29a

75–82 CIESHSSR 974.42 974.51a, b

92–104 CGENIYMATYPAK 1516.67 1516.69a þ105–117 WTDIIHAWHGEYK 1654.79 1654.83b, c, d þ121–142 YGVGAVPSDAVIGHYTQIVWYK 2422.23 2422.29b, c

146–157 AGCAAAYCPSSK 1241.52 1241.54a, b þ158–175 YSYFYVCQYCPAGNIIGK 2201.99 2202.01b, c þ176–181 TATPYK 679.35 679.38b þ182–204 SGPPCGDCPSDCDNGLCTNPCTR 2595.95 2596.02b, c þ205–217 ENEFTNCDSLVQK 1582.69 1582.73a, b, c þ218–226 SSCQDNYMK 1131.43 1131.47a, b þ229–238 CPASCFCQNK 1270.49 1270.52a, b þ

The protein was reduced with DTT and alkylated with iodoacetic acid prior to endoproteinase digestion. The numbering of the tryptic fragments is accordingto the coding region of triflin.

a Calculated from þ1 ion.b Calculated from þ2 ion.c Calculated from þ3 ion.d Calculated from þ4 ion.

K. Hansson et al. / Toxicon 54 (2009) 128–137134

NMR studies of the complex formation of hMSP with bothnative human CRISP-3 and a recombinant N-terminaldomain of CRISP-3 that contains the CAP domain but lacksthe C-terminal domain (Ghasriani et al., 2009).

The N-terminal (CAP containing) domain of the CRISPsis dominated by a b-pleated sheet surrounded by a-helices.These structure elements are very well conserved between

Elution volume (fract

0.20.0 0.4 0.6 0.8 0.20.0 0.4

Ab

so

rb

an

ce at 280 n

m

0.02

0.04

0.06

0.08

a

Fig. 5. Analytical gel chromatography of triflin in complex with pMSP. Triflin (1(a); 1.0 nmol (b); 2.0 nmol (c). Each mixture was then subjected to chromatography owas continuously monitored (continuous curve). The broken (triflin) and dotted (p

all CRISPs for which a structure has been determined (Guoet al., 2005; Shikamoto et al., 2005; Wang et al., 2005a,2006; Suzuki et al., 2008). MSP is also a b-pleated sheetprotein (Ghasriani et al., 2006) and the model of the hMSP–CRISP-3 complex shows that the binding is between onelateral side of the main b-pleated sheet of the CRISP to onelateral side of the b-pleated sheet of the MSP

ion of column volume)

0.6 0.8 0.20.0 0.4 0.6 0.8 1.0

b c

.0 nmol) was incubated with three different amounts of pMSP: 0.5 nmoln a Superdex 75 10/300 GL column. The absorbance at 280 nm of the effluentMSP) curves are overlays showing the results with the components alone.

Table 2Surface plasmon resonance analysis of the interactions between MSPs and CRISPs from snake venoms.a

Analyte/ligand ka (M�1 s�1) kd (s�1) kD (M)b

Catrin/pMSPb 1.35 (1.33–1.36)� 105 1.07 (1.06–1.08)� 10�2 0.13 (0.13–0.13)� 10�7

pMSP/catrinb 2.00 (1.96–2.05)� 105 0.80 (0.78–0.83)� 10�2 0.50 (0.47–0.53)� 10�8

Latisemin/pMSP 2.39 (2.23–2.51)� 105 2.30 (1.46–3.05)� 10�5 0.96 (0.60–1.22)� 10�10

pMSP/latisemin 0.98 (0.90–1.04)� 106 8.07 (7.09–8.69)� 10�5 0.82 (0.79–0.84)� 10�10

pMSP/pseudecin 1.64 (1.33–1.79)� 105 8.06 (7.88–8.41)� 10�5 0.50 (0.44–0.59)� 10�9

Triflin/pMSP 1.36 (1.35–1.37)� 105 1.48 (1.47–1.48)� 10�4 1.09 (1.07–1.10)� 10�9

pMSP/triflin 1.14 (1.06–1.27)� 105 1.37 (1.29–1.45)� 10�4 1.20 (1.14–1.24)� 10�9

Latisemin/hMSP 0.79 (0.63–0.91)� 105 2.21 (2.19–2.26)� 10�4 2.86 (2.42–3.47)� 10�9

hMSP/latisemin 2.12 (1.90–2.26)� 105 8.35 (8.28–8.42)� 10�4 3.95 (2.68–4.41)� 10�9

hMSP/pseudecin 1.13 (1.12–1.13)� 105 1.96 (1.93–1.98)� 10�4 1.76 (1.71–1.80)� 10�9

Triflin/hMSP 0.97 (0.93–1.00)� 104 5.28 (5.23–5.34)� 10�3 0.54 (0.52–0.56)� 10�7

hMSP/triflin 0.55 (0.53–0.57)� 105 8.62 (8.55–8.71)� 10�3 1.56 (1.52–1.62)� 10�7

a Ligand (underlined) is the protein bound to the sensor chip. The constants were derived by fitting 1:1 Langmuir model to the experimental data exceptfor the interactions between catrin and pMSP, for which a two-state model was used. Data in the table are the mean values of three experiments with therange in parenthesis.

b For catrin kD is the apparent dissociation constant.

K. Hansson et al. / Toxicon 54 (2009) 128–137 135

(Ghasriani et al., 2009). This model of the complex is sup-ported by the strong binding of MSP to the four venomCRISPs observed in this work. The lateral side of theb-pleated sheet of the CRISPs that binds to the MSP isb-strand 4 (Ghasriani et al., 2009) and as can be seen inFig. 1, these parts of the sequences are very similar in all fiveCRISPs (b-strand 3, which is even more conserved, is thecentral strand of the b-pleated sheet).

Both MSPs and CRISPs seem to be present in mostvertebrates; MSP has been found in the amphioxus (Wanget al., 2005b), a chordate, and CRISP in the lamprey (Ito

0

100

200

300

Sign

al (R

U)

Sign

al (R

U)

40 nM

30 nM

20 nM15 nM

10 nM

5 nM2.5 nM

0 100 200 300 400 500 600 700 8000

25

50

100

120

Sign

al (R

U)

Sign

al (R

U)

Time (s)Time (s)

100 nM

80 nM

60 nM

40 nM

20 nM

10 nM5 nM

a

b

Fig. 6. Surface plasmon resonance analysis of the binding of porcine (a) andhuman (b) MSPs to immobilized triflin. Dot symbols represent experimentaltime points of adsorption and desorption. Solid curves are the associationsand dissociations calculated from the derived kinetic constants. Theconcentrations of the MSPs in the running buffer are indicated to the right.

et al., 2007), a primitive vertebrate. The function is knownneither for MSPs nor for the CRISPs. Intraspecies interac-tions between MSPs and CRISPs have been demonstratedboth in the human and in a reptile, indicating that theseinteractions are well conserved and, therefore, probablyphysiologically significant. As mentioned, several CRISPshave been shown to be ion-channel inhibitors, both inmammal and snake. Ion-channels are involved innumerous cell functions and it is not improbable that thephysiological functions of CRISPs are mediated via theirion-channel inhibiting activities. Supposing that theseactivities normally should take place locally in theorganism, e.g. by acting on contractility or permeability ofcapillaries in an inflammatory focus, it is conceivable thatfactors, such as MSP or a1B-glycoprotein, are present toimpede systemic effects. Further work is needed to showhow common CRISP–MSP interactions are in other verte-brates and also if homologues of a1B-glycoprotein arepresent and can bind CRISPs.

In this work we have shown that high affinity bindingcan also occur between species, i.e. between MSPs frommammals and CRISPs from reptiles. However, the CRISPsinvestigated were all from such specialized organs asvenom glands. A well-known strategy for poisonousorganisms is to produce venom factors that act on physio-logic systems in the victim in an exaggerated fashion.Examples are snake venoms that act on the coagulationsystem causing fibrinolysis with bleedings or coagulationwith extensive systemic thrombosis. Therefore, the highaffinity of human and porcine MSPs for the CRISPs in thesnake venoms could be merely an adaptation of the MSPsto neutralize a venom poison. Whether the MSP–CRISPinteractions are as general as our studies might indicateremains to be tested with less specialized CRISPs. For suchstudies the affinity chromatography on an MSP columndescribed in this work could be a helpful aid.

To conclude, MSP and CRISPs seem to be evolutionaryconserved, at least in vertebrates, and binding betweenMSPs and CRISPs has been reported earlier in two species.This work has shown that MSP–CRISP interactions can takeplace also across species boundaries. How general thisphenomenon is, and what physiological function it mightserve, remain to be investigated.

K. Hansson et al. / Toxicon 54 (2009) 128–137136

Acknowledgements

We are grateful to Johan Stenflo, Per Simonsson, andJohan Malm, Department of Laboratory Medicine, MalmoUniversity Hospital, for making laboratory space andresources available to P.F. The mass spectrometer waspurchased with grants from the Knut and Alice WallenbergFoundation.

Conflict of interest

The authors declare that there are no conflicts ofinterest.

References

Aoki, N., Sakiyama, A., Deshimaru, M., Terada, S., 2007. Identification ofnovel serum proteins in a Japanese viper: homologs of mammalianPSP94. Biochem. Biophys. Res. Commun. 359, 330–334.

Aoki, N., Sakiyama, A., Kuroki, K., Maenaka, K., Kohda, D., Deshimaru, M.,Terada, S., 2008. Serotriflin, a CRISP family protein with bindingaffinity for small serum protein-2 in snake serum. Biochim. Biophys.Acta 1784, 621–628.

Brown, R.L., Haley, T.L., West, K.A., Crabb, J.W., 1999. Pseudechetoxin:a peptide blocker of cyclic nucleotide-gated ion channels. Proc. Natl.Acad. Sci. U.S.A. 96, 754–759.

Doctor, V.M., Sheth, A.R., Simha, M.M., Arbatti, N.J., Aaveri, J.P., Sheth, N.A.,1986. Studies on immunocytochemical localization of inhibin-likematerial in human prostatic tissue: comparison of its distribution innormal, benign and malignant prostates. Br. J. Cancer 53, 547–554.

Eeles, R.A., Kote-Jarai, Z., Giles, G.G., Olama, A.A., Guy, M., Jugurnauth, S.K.,Mulholland, S., Leongamornlert, D.A., Edwards, S.M., Morrison, J.,Field, H.I., Southey, M.C., Severi, G., Donovan, J.L., Hamdy, F.C.,Dearnaley, D.P., Muir, K.R., Smith, C., Bagnato, M., Ardern-Jones, A.T.,Hall, A.L., O’Brien, L.T., Gehr-Swain, B.N., Wilkinson, R.A., Cox, A.,Lewis, S., Brown, P.M., Jhavar, S.G., Tymrakiewicz, M., Lophatananon, A.,Bryant, S.L., Horwich, A., Huddart, R.A., Khoo, V.S., Parker, C.C.,Woodhouse, C.J., Thompson, A., Christmas, T., Ogden, C., Fisher, C.,Jamieson, C., Cooper, C.S., English, D.R., Hopper, J.L., Neal, D.E.,Easton, D.F., 2008. Multiple newly identified loci associated withprostate cancer susceptibility. Nat. Genet. 40, 316–321.

Fernlund, P., Granberg, L.B., Roepstorff, P., 1994. Amino acid sequence ofbeta-microseminoprotein from porcine seminal plasma. Arch. Bio-chem. Biophys. 309, 70–76.

Ghasriani, H., Teilum, K., Johnsson, Y., Fernlund, P., Drakenberg, T., 2006.Solution structures of human and porcine beta-microseminoprotein.J. Mol. Biol. 362, 502–515.

Ghasriani, H., Fernlund, P., Udby, L., Drakenberg, T., 2009. A model of thecomplex between human beta-microseminoprotein and CRISP-3based on NMR data. Biochem. Biophys. Res. Commun. 378, 235–239.

Gibbs, G.M., Scanlon, M.J., Swarbrick, J., Curtis, S., Gallant, E., Dulhunty, A.F.,O’Bryan, M.K., 2006. The cysteine-rich secretory protein domain ofTpx-1 is related to ion channel toxins and regulates ryanodine receptorCa2þ signaling. J. Biol. Chem. 281, 4156–4163.

Gibbs, G.M., Bianco, D.M., Jamsai, D., Herlihy, A., Ristevski, S., Aitken, R.J.,Kretser, D.M., O’Bryan, M.K., 2007. Cysteine-rich secretory protein 2binds to mitogen-activated protein kinase kinase kinase 11 in mousesperm. Biol. Reprod. 77, 108–114.

Gibbs, G.M., Roelants, K., O’Bryan, M.K., 2008. The Cap superfamily:cysteine-rich secretory proteins, antigen 5 and pathogenesis-related1 proteins – roles in reproduction, cancer and immune defense.Endocr. Rev. 29, 865–897.

Gill, S.C., von Hippel, P.H., 1989. Calculation of protein extinction coeffi-cients from amino acid sequence data. Anal. Biochem. 182, 319–326.

Guo, M., Teng, M., Niu, L., Liu, Q., Huang, Q., Hao, Q., 2005. Crystalstructure of the cysteine-rich secretory protein stecrisp reveals thatthe cysteine-rich domain has a Kþ channel inhibitor-like fold. J. Biol.Chem. 280, 12405–12412.

Haendler, B., Kratzschmar, J., Theuring, F., Schleuning, W.D., 1993. Tran-scripts for cysteine-rich secretory protein-1 (CRISP-1; DE/AEG) andthe novel related CRISP-3 are expressed under androgen control inthe mouse salivary gland. Endocrinology 133, 192–198.

Ito, N., Mita, M., Takahashi, Y., Matsushima, A., Watanabe, Y.G., Hirano, S.,Odani, S., 2007. Novel cysteine-rich secretory protein in the buccal

gland secretion of the parasitic lamprey, Lethenteron japonicum. Bio-chem. Biophys. Res. Commun. 358, 35–40.

Kjeldsen, L., Cowland, J.B., Johnsen, A.H., Borregaard, N., 1996. SGP28,a novel matrix glycoprotein in specific granules of human neutrophilswith similarity to a human testis-specific gene product and a rodentsperm-coating glycoprotein. FEBS Lett. 380, 246–250.

Kratzschmar, J., Haendler, B., Eberspaecher, U., Roosterman, D., Donner, P.,Schleuning, W.D., 1996. The human cysteine-rich secretory protein(CRISP) family. Primary structure and tissue distribution of CRISP-1,CRISP-2 and CRISP-3. Eur. J. Biochem 236, 827–836.

Laemmli, U.K., 1970. Cleavage of structural proteins during the assemblyof the head of bacteriophage T4. Nature 227, 680–685.

Lazure, C., Villemure, M., Gauthier, D., Naude, R.J., Mbikay, M., 2001.Characterization of ostrich (Struthio camelus) beta-micro-seminoprotein (MSP): identification of homologous sequences in ESTdatabases and analysis of their evolution during speciation. ProteinSci. 10, 2207–2218.

Lilja, H., Abrahamsson, P.A., 1988. Three predominant proteins secreted bythe human prostate gland. Prostate 12, 29–38.

Mizuki, N., Kasahara, M., 1992. Mouse submandibular glands express anandrogen-regulated transcript encoding an acidic epididymal glyco-protein-like molecule. Mol. Cell. Endocrinol. 89, 25–32.

Reddy, T., Gibbs, G.M., Merriner, D.J., Kerr, J.B., O’Bryan, M.K., 2008.Cysteine-rich secretory proteins are not exclusively expressed in themale reproductive tract. Dev. Dyn. 237, 3313–3323.

Reeves, J.R., Xuan, J.W., Arfanis, K., Morin, C., Garde, S.V., Ruiz, M.T.,Wisniewski, J., Panchal, C., Tanner, J.E., 2005. Identification, purifica-tion and characterization of a novel human blood protein withbinding affinity for prostate secretory protein of 94 amino acids.Biochem. J. 385, 105–114.

Reeves, J.R., Dulude, H., Panchal, C., Daigneault, L., Ramnani, D.M., 2006.Prognostic value of prostate secretory protein of 94 amino acids andits binding protein after radical prostatectomy. Clin. Cancer Res. 12,6018–6022.

Roberts, K.P., Johnston, D.S., Nolan, M.A., Wooters, J.L., Waxmonsky, N.C.,Piehl, L.B., Ensrud-Bowlin, K.M., Hamilton, D.W., 2007. Structure andfunction of epididymal protein cysteine-rich secretory protein-1.Asian J. Androl. 9, 508–514.

Schwidetzky, U., Haendler, B., Schleuning, W.D., 1995. Isolation andcharacterization of the androgen-dependent mouse cysteine-richsecretory protein-3 (CRISP-3) gene. Biochem. J. 309 (Pt 3), 831–836.

Shevchenko, A., Wilm, M., Vorm, O., Mann, M., 1996. Mass spectrometricsequencing of proteins silver-stained polyacrylamide gels. Anal.Chem. 68, 850–858.

Shikamoto, Y., Suto, K., Yamazaki, Y., Morita, T., Mizuno, H., 2005. Crystalstructure of a CRISP family Ca2þ-channel blocker derived from snakevenom. J. Mol. Biol. 350, 735–743.

Suzuki, N., Yamazaki, Y., Brown, R.L., Fujimoto, Z., Morita, T., Mizuno, H.,2008. Structures of pseudechetoxin and pseudecin, two snake-venomcysteine-rich secretory proteins that target cyclic nucleotide-gatedion channels: implications for movement of the C-terminal cysteine-rich domain. Acta Crystallogr. D. Biol. Crystallogr 64, 1034–1042.

Thomas, G., Jacobs, K.B., Yeager, M., Kraft, P., Wacholder, S., Orr, N., Yu, K.,Chatterjee, N., Welch, R., Hutchinson, A., Crenshaw, A., Cancel-Tassin, G., Staats, B.J., Wang, Z., Gonzalez-Bosquet, J., Fang, J., Deng, X.,Berndt, S.I., Calle, E.E., Feigelson, H.S., Thun, M.J., Rodriguez, C.,Albanes, D., Virtamo, J., Weinstein, S., Schumacher, F.R.,Giovannucci, E., Willett, W.C., Cussenot, O., Valeri, A., Andriole, G.L.,Crawford, E.D., Tucker, M., Gerhard, D.S., Fraumeni Jr., J.F., Hoover, R.,Hayes, R.B., Hunter, D.J., Chanock, S.J., 2008. Multiple loci identified ina genome-wide association study of prostate cancer. Nat. Genet. 40,310–315.

Udby, L., Calafat, J., Sorensen, O.E., Borregaard, N., Kjeldsen, L., 2002a.Identification of human cysteine-rich secretory protein 3 (CRISP-3) asa matrix protein in a subset of peroxidase-negative granules ofneutrophils and in the granules of eosinophils. J. Leukoc. Biol. 72,462–469.

Udby, L., Cowland, J.B., Johnsen, A.H., Sorensen, O.E., Borregaard, N.,Kjeldsen, L., 2002b. An ELISA for SGP28/CRISP-3, a cysteine-richsecretory protein in human neutrophils, plasma, and exocrinesecretions. J. Immunol. Methods 263, 43–55.

Udby, L., Sorensen, O.E., Pass, J., Johnsen, A.H., Behrendt, N., Borregaard, N.,Kjeldsen, L., 2004. Cysteine-rich secretory protein 3 is a ligandof alpha1B-glycoprotein in human plasma. Biochemistry 43,12877–12886.

Udby, L., Lundwall, A., Johnsen, A.H., Fernlund, P., Valtonen-Andre, C.,Blom, A.M., Lilja, H., Borregaard, N., Kjeldsen, L., Bjartell, A., 2005.beta-Microseminoprotein binds CRISP-3 in human seminal plasma.Biochem. Biophys. Res. Commun. 333, 555–561.

K. Hansson et al. / Toxicon 54 (2009) 128–137 137

Wang, J., Shen, B., Guo, M., Lou, X., Duan, Y., Cheng, X.P., Teng, M., Niu, L.,Liu, Q., Huang, Q., Hao, Q., 2005a. Blocking effect and crystal structureof natrin toxin, a cysteine-rich secretory protein from Naja atravenom that targets the BK(Ca) channel. Biochemistry 44,10145–10152.

Wang, Y., Zhang, S., Liu, Z., Li, H., Wang, L., 2005b. Identification andexpression of amphioxus beta-microseminoprotein (MSP)-like geneencoding an ancient and rapidly evolving protein in chordates.Comp. Biochem. Physiol. B Biochem. Mol. Biol. 142, 251–257.

Wang, F., Li, H., Liu, M.N., Song, H., Han, H.M., Wang, Q.L., Yin, C.C.,Zhou, Y.C., Qi, Z., Shu, Y.Y., Lin, Z.J., Jiang, T., 2006. Structural andfunctional analysis of natrin, a venom protein that targets various ionchannels. Biochem. Biophys. Res. Commun. 351, 443–448.

Weiber, H., Andersson, C., Murne, A., Rannevik, G., Lindstrom, C.,Lilja, H., Fernlund, P., 1990. Beta microseminoprotein is not

a prostate-specific protein. Its identification in mucous glands andsecretions. Am. J. Pathol. 137, 593–603.

Yamazaki, Y., Brown, R.L., Morita, T., 2002a. Purification and cloning oftoxins from elapid venoms that target cyclic nucleotide-gated ionchannels. Biochemistry 41, 11331–11337.

Yamazaki, Y., Koike, H., Sugiyama, Y., Motoyoshi, K., Wada, T.,Hishinuma, S., Mita, M., Morita, T., 2002b. Cloning and characteriza-tion of novel snake venom proteins that block smooth musclecontraction. Eur. J. Biochem. 269, 2708–2715.

Yamazaki, Y., Hyodo, F., Morita, T., 2003. Wide distribution of cysteine-rich secretory proteins in snake venoms: isolation and cloning ofnovel snake venom cysteine-rich secretory proteins. Arch. Biochem.Biophys. 412, 133–141.

Yamazaki, Y., Morita, T., 2004. Structure and function of snake venomcysteine-rich secretory proteins. Toxicon 44, 227–231.