The first phospholipase inhibitor from the serum of Vipera ammodytes ammodytes Jernej S ˇ ribar 1 , Lidija Kovac ˇ ic ˇ 1 , Petra Dras ˇ kovic ˇ 1 , Grazyna Faure 2 and Igor Kriz ˇ aj 1 1 Department of Molecular and Biomedical Sciences, Joz ˇ ef Stefan Institute, Ljubljana, Slovenia 2 Unite ´ d’Immunologie Structurale, Institut Pasteur, Paris, France In many snake venoms the major proportion of the toxic components are secretory phospholipases A 2 (sPLA 2 ). Snakes have long been known to be resistant to their own venom. One of the reasons for this is the presence of the so-called neutralizing factors in the blood of these animals. Several proteins have been iso- lated from snake sera that exhibit antihemorrhagic, antineurotoxic or antimyotoxic properties. Many have been characterized as sPLA 2 inhibitors (PLIs) and are found in both venomous and nonvenomous snake spe- cies [1–3]. Based on their structural characteristics, PLIs have been classified into three groups: a, b and c [4]. PLI-a are 75–120 kDa globular glycoproteins, consisting of three to six noncovalently bound 20–25 kDa subunits [4–7]. They all possess a sequence similar to the carbo- hydrate-recognition domain of Ca 2+ -dependent lectins. The binding site for sPLA 2 on this type of PLI has, however, been suggested to be distinct from the carbo- hydrate-binding site of the homologous Ca 2+ -depen- dent lectin [8]. This type of PLI has, to date, been found in species of snakes belonging to the Viperidae family where it specifically inhibits the acidic type IIA sPLA 2 molecules from the species’ own venom [1,2]. The first PLI-b was isolated from the blood of Agkistrodon blomhoffii siniticus [4]. It is a 160 kDa Keywords ammodytoxin; inhibitor; secretory phospholipase A 2 ; snake serum; Vipera ammodytes ammodytes Correspondence I. Kriz ˇ aj, Department of Molecular and Biomedical Sciences, Joz ˇ ef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia Fax: +386 1477 3984 Tel: +386 1477 3626 E-mail: igor.krizaj@ijs.si (Received 26 June 2007, revised 2 October 2007, accepted 4 October 2007) doi:10.1111/j.1742-4658.2007.06127.x Ammodytoxins are neurotoxic secretory phospholipase A 2 molecules, some of the most toxic components of the long-nosed viper (Vipera ammo- dytes ammodytes) venom. Envenomation by this and by closely related vipers is quite frequent in southern parts of Europe and serotherapy is used in the most severe cases. Because of occasional complications, alternative medical treatment of envenomation is needed. In the present study, ammo- dytoxin inhibitor was purified from the serum of V. a. ammodytes using two affinity procedures and a gel exclusion chromatography step. The ammodytoxin inhibitor from V. a. ammodytes serum consists of 23- and 25-kDa glycoproteins that form an oligomer, probably a tetramer, of about 100 kDa. N-terminal sequencing and immunological analysis revealed that both types of subunit are very similar to c-type secretory phospholipase A 2 inhibitors. The ammodytoxin inhibitor from V. a. ammodytes serum is a potent inhibitor of phospholipase activity and hence probably also the neurotoxicity of ammodytoxins. Discovery of the novel natural inhibitor of these potent secretory phospholipase A 2 toxins opens up prospects for the development of new types of small peptide inhibitors for use in regulating the physiological and pathological activities of secretory phospho- lipases A 2 . Abbreviations AIVAS, ammodytoxin inhibitor from Vipera ammodytes serum; Atx, ammodytoxin; AtxC, ammodytoxin C; CB, basic subunit of crotoxin; CICS, crotoxin inhibitor from Crotalus serum; CIM, convective interactive media; CNF, Crotalus neutralizing factor; ConA, concanavalin A; LL, lentil lectin; PLA 2 , phospholipase A 2 ; PLI, PLA 2 inhibitor; PVDF, poly(vinylidene difluoride); RU, response unit; SPR, surface plasmon resonance; VAAS, Vipera a. ammodytes serum; sPLA 2 , secretory phospholipase A 2 ; WGA, wheat germ agglutinin. FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS 6055 globular glycoprotein, comprising three identical sub- units of  50 kDa. The subunits contain nine leucine- rich repeats that are probably involved in sPLA 2 binding [9]. PLI-b is not present exclusively in snakes of the Viperidae family. This type of PLI has also been isolated from a nonvenomous snake Elaphe quadrivirg- ata. Unlike the PLI-b from A. b. siniticus, the trimer from E. quadrivirgata comprises two homologous types of 50 kDa subunits [10]. These two types of PLI were found to inhibit only some type II sPLA 2 molecules. The majority of characterized PLIs isolated from snake serum belong to the c-type group. They are 90– 130 kDa acidic glycoproteins, composed of three to six noncovalently bound 20–31 kDa subunits [2]. They are characterized by a three-finger motif consisting of a unique pattern of cysteine residues that is found in various proteins with different biological functions, from a serine protease inhibitor to an ion channel blocker. PLI-c molecues are present in the sera of snakes from the families Viperidae, Colubridae, Elapi- dae, Boiidae and Hydrophidae, where they appear to be less specific than a- and b-type PLIs because they are able to inhibit sPLA 2 from groups I, II and III [11]. Besides snake venom sPLA 2 molecules, which exert neurotoxic, myotoxic, cardiotoxic and other pharmaco- logical effects [12], a number of other types of phos- pholipase A 2 (PLA 2 ) have been discovered in mammals [13]. In addition to digesting phospholipids, these molecules are involved in many other physiologi- cal processes [14,15]. In some processes, such as exocytosis ⁄ endocytosis, inflammation, blood coagula- tion, ischaemia and antibacterial defence, the sPLA 2 molecule participates as an enzyme, whereas in other processes, like cell migration, cell proliferation and inhibition of blood coagulation, the sPLA 2 molecule acts as a ligand for different cellular receptors [16–18]. Obviously, strict regulation of the enzymatic and receptor-binding activities of sPLA 2 is crucial to avoid pathological conditions such as cancer, atherosclerosis and acute respiratory distress syndrome, as well as to neutralize the toxic activity of venom sPLA 2 . There is a great need for a wide range of new inhibitors of sPLA 2 that are highly specific. Given the similarity between some tissue receptors for sPLA 2 and PLI (e.g. between the M-type sPLA 2 receptor and PLI-a) the characterization of new snake PLIs may eventually lead to the discovery of novel tissue receptors for sPLA 2 , resulting in a better understanding of the patho- physiology of this group of enzymes. In this work we present the characterization of the first PLI from the serum of the most dangerous Euro- pean snake, Vipera ammodytes ammodytes. This PLI potently inhibits the enzymatic and therefore very probably also the neurotoxic activity of ammodytoxins (Atx), presynaptically toxic sPLA 2 and the main toxic components of this venom. Results and Discussion Detection of Atx inhibitor from V. a. ammodytes serum by surface plasmon resonance Crotoxin inhibitor from Crotalus serum (CICS) binds also to Atx [19], so we worked on the assumption that a structurally similar PLI that we named ammodytoxin inhibitor from Vipera ammodytes serum (AIVAS) exists in the serum of V. a. ammodytes (VAAS). This assumption was verified by immobilizing anti-CICS IgG on a CM-5 surface plasmon resonance (SPR) sen- sor chip. An aliquot of VAAS was injected over this chip, allowing the proteins to bind, followed by ammo- dytoxin C (AtxC). AtxC interaction with the antibody- retained proteins is shown in Fig. 1A. The result clearly demonstrated the presence of CICS-like rb VAAS AtxC rb VAAS CBa A B retention time (s) retention time (s) response (RU) ANTI-CICS ANTI-CICS response (RU) 200 100 90 80 70 60 50 40 30 20 10 –10 0 180 160 140 120 100 80 60 40 20 0 –50 0 50 100 150 200 –50 0 50 100 150 –20 Fig. 1. Detection of CICS-like molecules in VAAS. Rabbit anti-CICS IgG was covalently immobilized on an SPR CM-5 sensor chip. Serum was injected over the sensor chip, followed by sPLA 2 AtxC (A) or CB (B). Finally, running buffer (rb) was injected and the disso- ciation rate observed. Data were analyzed using the BIAEVALUATION 3.1 software (Biacore). RU, response unit. Ammodytoxin inhibitor from Vipera ammodytes serum J. S ˇ ribar et al. 6056 FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS molecules in VAAS, which bind AtxC. The bound AtxC was not dissociated by running buffer, indicating a strong interaction between AtxC and its binding protein(s) in the serum. In a control experiment, the crotoxin basic subunit (CB) was injected instead of AtxC (Fig. 1B). It was bound to CICS-like molecule(s) in VAAS, but the interaction was weaker than that of AtxC, as judged by its dissociation by running buffer. When CICS was attached via anti-CICS IgG on the SPR chip, however, the situation was reversed [19]. In this case, CB bound to CICS with much higher affinity than Atx, revealing the specificity of the sPLA 2 -bind- ing proteins for the homologous sPLA 2 toxins from their species’ own venom, consistent with the self-pro- tective function of serum sPLA 2 -binding molecules [3]. Affinity labelling of AIVAS Affinity labelling was used to obtain structural infor- mation on AIVAS. Chemical cross-linking of 125 I-labelled AtxC to VAAS resulted in three specific adducts of 114, 74 and 40 kDa apparent molecular mass (Fig. 2, lanes 1 and 2). The molecular mass of AtxC is 14 kDa, and, assuming the association of one molecule of AtxC with one molecule of AtxC-binding protein, the apparent molecular masses of Atx-binding proteins in the serum are 100, 60 and 26 kDa. Based on the probable structural relatedness between AIVAS and CICS (Fig. 1A), the species with apparent molecu- lar masses of 60 and 100 kDa probably correspond to a dimer and a tetramer of the basic subunit of about 26 kDa. For most PLIs, however, a 1 : 1 stoichiometry of binding of sPLA 2 to the PLI subunit has been reported. For example, PLI-a from Bothrops asper inhibited the venom myotoxic sPLA 2 on a 1 : 1 basis [20]. The same was shown for the PLI-b trimer from A. b. siniticus, which bound three molecules of sPLA 2 [9]. In the case of a c-type PLI, 1 : 1 stoichiometry of binding between CB, a basic sPLA 2 subunit of crotox- in, and the Crotalus neutralizing factor (CNF) subunit has been suggested [21], although other results indicate that only one molecule of CB is bound per CICS oligomer [22]. sPLA 2 -binding characteristics of AIVAS To check the sPLA 2 -binding profile of AIVAS, aliqu- ots of serum were affinity labelled in the presence of different types of sPLA 2 at 2 lm concentration (Fig. 2). Atx and ammodytin L, a myotoxic sPLA 2 homologue from V. a. ammodytes venom, were the most potent inhibitors of formation of the specific adducts between 125 I-labelled AtxC and AIVAS. Inter- estingly, a nontoxic sPLA 2 , ammodytin I 2 from V. a. ammodytes venom, also blocked specific labelling very effectively, a feature not observed in the case of CICS, which does not inhibit phospholipase activity or bind nontoxic sPLA 2 [19]. Practically no inhibition was, however, observed with type II sPLA 2 molecules structurally related to Atx (i.e. agkistrodotoxin, crotox- in and human sPLA 2 -IIA), or with type I sPLA 2 mole- cules, such as notexin, Oxyuranus s. scutellatus sPLA 2 , b-bungarotoxin and taipoxin, and type III bee venom PLA 2 . AIVAS obviously very specifically binds V. a. ammodytes venom sPLA 2 , Atx and ammody- toxins, probably to protect the snake from the action of its own venom sPLA 2 [3]. The pH stability of AIVAS was determined by incu- bating aliquots of VAAS at different pH values for 30 min at room temperature. Following incubation, the pH was adjusted to 8.2 and affinity labelling with 125 I-labelled AtxC was performed (Fig. 3). While the intensity of the specific adduct at 40 kDa was slightly reduced only after exposure of VAAS to a pH of < 6.0, the intensities of the bands at 74 and 114 kDa were more pH dependent. Given the structural similar- ity between AIVAS and CICS, adducts at 74 and 114 kDa probably represent 125 I-AtxC-labelled associa- tions of two and four inhibitor subunits. We assume 123 86 44.6 31.4 M [kDa] T AxtC AxAt AnI 2 t AnL t Atxg huI -PLAIA 2 x notei n OS 2 crotoxin -B x ut ta oxinip bvPLA 2 Fig. 2. Detection of Atx-binding proteins in VAAS and analysis of their sPLA 2 -binding profile. Serum aliquots were affinity labelled with 125 I-labelled AtxC in the absence or presence of indicated sPLA 2 molecules at a concentration of 2 lM. The samples were analyzed by SDS ⁄ PAGE (10% acrylamide gel) under reducing condi- tions and the gel autoradiographed. The positions of specific adducts are indicated by arrows. Agtx, agkistrodotoxin; Atn, ammo- dytins;Atx, ammodytoxins; b-Butx, b-bungarotoxin; bvPLA 2 , bee venom PLA 2 ; huIIA-PLA 2 , human type IIA sPLA 2 ;OS 2 , Oxyuranus scutellatus sPLA 2 -2. J. S ˇ ribar et al. Ammodytoxin inhibitor from Vipera ammodytes serum FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS 6057 that the oligomerization of subunits is reduced by exposure to low pH, but not to the same extent as their AtxC-binding ability. As demonstrated in Fig. 3, the binding of 125 I- labelled AtxC to AIVAS is independent of the pres- ence of divalent cations. AIVAS is a glycoprotein PLI molecules isolated from snake sera are typically glycoproteins [2,11,23,24]. However, glycosylation of PLI does not appear to play a role in binding sPLA 2 [11]. To check whether AIVAS contains carbohydrate, aliquots of VAAS were incubated with different gel- immobilized lectins, lentil lectin (LL), wheat germ agglutinin (WGA) and concanavalin A (ConA), and analysed for AIVAS in the bound and breakthrough fractions. AIVAS bound to all three lectins tested and is thus a glycoprotein (data not shown). Binding was strongest to LL, which was therefore chosen for the first step in the isolation procedure. Purification of AIVAS In the initial step, VAAS was chromatographed on an LL-affinity column. The bound material was further fractionated by AtxC affinity chromatography. The attempt to purify biologically active AIVAS on an AtxC–Sepharose column [25] was unsuccessful owing to the strong binding of AIVAS on the column and because of the denaturing conditions (pH 2.8) needed for elution. Inactivation of other PLIs on low-pH elu- tion has also been reported by other authors [20,24]. Another chromatographic medium was chosen that would allow rapid analysis with the same ligand. AtxC was immobilized on convective interactive media (CIM) monolithic disks and, by rapid elution at pH 2.8 followed by immediate elevation of the pH of the samples to 7.4, inactivation of AIVAS was avoided. SDS ⁄ PAGE analysis showed protein bands with apparent molecular masses of 23–25 kDa and 90 kDa (Fig. 4). In an attempt to separate the 23 and 25 kDa pro- teins from a 90 kDa protein, fraction 8 from CIM– AtxC affinity chromatography (Fig. 4) was subjected to analytical gel-filtration on an FPLC Superdex HR 10 ⁄ 30 column. The material was eluted in two strongly overlapping peaks. The major peak corresponded to a molecular mass slightly below 200 kDa, and the minor peak corresponded to a mass slightly above 100 kDa (Fig. 5A). Fractions were analysed by SDS ⁄ PAGE under reducing conditions and the gels were silver stained (Fig. 5B). In fraction 7, besides a major protein band at 90 kDa, two proteins with apparent molecular masses of 23 and 25 kDa were detected. In fraction 8, however, only the two low-molecular-mass proteins were found, under both reducing and nonreducing conditions. These analyses, in addition, revealed that under nondenaturing conditions on the gel-filtration column, the low-molecular-mass proteins oligomerize, which is a common feature of PLIs [11]. 31.4 44.6 86 123 M [kDa] 5.0 pH: 5.5 6. 0 6.5 7. 0 7. 5 8.0 8. 5 9.0 EG AT EDTA Fig. 3. Effect of pH and divalent ions on the binding of AtxC to AIVAS. Aliquots of serum were exposed to different pH levels (5.0–9.0) at room temperature and then affinity labelled at pH 8.2 with 125 I-labelled AtxC. Two samples were affinity labelled in the presence of EGTA or EDTA to probe the dependency of binding on the presence of divalent cations. All samples were analyzed by SDS ⁄ PAGE (10% acrylamide gel) under reducing conditions and the gel was autoradiographed. The positions of specific adducts are indicated by arrows. 72 50 24 17 LLE fr .5 .6fr .7fr fr. 8 fr. 9 .fr 10 100 130 M [kDa] Fig. 4. Separation of AIVAS on AtxC affinity chromatography. The eluate from LL–Sepharose chromatography was chromatographed on a CIM–AtxC affinity disk. The fractions were analyzed by SDS ⁄ PAGE (12.5% acrylamide gel) under reducing conditions and the gel was silver stained. In lane LLE, the composition of the frac- tion retained by LL and applied on the AtxC affinity chromatography is shown, and lanes fr.5 to fr.10 display the composition of frac- tions eluted from the toxin affinity chromatography. The positions of AIVAS are indicated by arrowheads. Ammodytoxin inhibitor from Vipera ammodytes serum J. S ˇ ribar et al. 6058 FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS The purified AIVAS retained the ability to bind AtxC, as demonstrated in an affinity labelling experi- ment of 125 I-labelled AtxC and the gel-filtration frac- tions 7 and 8 (Fig. 6). Fraction 8, which contained only the proteins of 23 and 25 kDa, gave three specific adducts of apparent mass 40, 74 and 114 kDa. AIVAS is therefore a tetramer that is noncovalently associated with subunits of 23 and 25 kDa. The ratio between the 23 and 25 kDa subunits in the complex is currently unknown. The results suggest that only one molecule of Atx is bound by the tetramer, as proposed also for the structurally similar CICS [22]. AIVAS is structurally related to c-type PLIs The proteins from fractions 7 and 8 were electroblot- ted to a poly(vinylidene difluoride) (PVDF) membrane and their N-terminal sequences were determined. The 25-kDa bands from each fraction contained two sequences – a major sequence (VAA_1) and a similar minor sequence (VAA_2). The ratio between the major and the minor sequences was 80 : 20. In the 23-kDa bands, only the VAA_1 sequence was present (Fig. 7A). Comparison of the sequences with protein sequence databases (Fig. 7B) showed that VAA_1 is very similar to the 25 kDa subunit of the c-type PLI from the serum of A. b. siniticus [4], to the CNF from Crotalus durissus terrificus and 23 ⁄ 25 kDa CICS from the same snake serum [21,22]. On the other hand, the VAA_2 sequence shows strong similarity to the 20 kDa subunit of the c-type PLI from the serum of A. b. siniticus [4] and the minor 23 kDa sequence of CICS [22]. The immunoblot in Fig. 5C confirms that the isolated AIVAS of 23 and 25 kDa indeed belong structurally to the c-type PLIs. Thus, both proteins were cross-reactive with anti-CICS IgG which, however, did not recognize the 90 kDa Atx-binding protein. The N-terminal amino acid sequence of the latter shows similarity to immunoglobulin heavy chain (E-V-Q-L-V-E-X-G-Q-D). This suggests that the snake can produce auto-antibodies against toxic components in its own venom as part of its self-pro- tection system, a feature observed here for the first time. A B absorbance (mAU) M (kDa) Fig. 5. Final purification of AIVAS and its partial characterization. (A) Fraction 8 from the AtxC-affinity purification step was chro- matographed on a Superdex HR 10 ⁄ 30 column. Numbers 1–18 indicate fractions eluted from the column. (B) Proteins in fractions 7, 8 and 9 were analyzed by SDS ⁄ PAGE (12.5% acrylamide gel) under reducing conditions and the gel was silver stained. Protein bands of apparent molecular weights 23, 25 and 90 kDa are indicated by arrows. (C) Fractions 7 and 8 were analyzed by SDS ⁄ PAGE under reducing conditions, transferred to a nitrocellulose membrane and immunostained using polyclonal anti- CICS IgG and the enhanced chemilumines- cence (ECL) detection system. Proteins of 23 and 25 kDa (arrows) are AIVAS structur- ally related to PLI-c. (D) AIVAS was degly- cosylated using protein N-glycosidase F and analysed by SDS ⁄ PAGE as described above. A decrease of the molecular mass revealed that both AIVAS subunits contain 3–4 kDa of N-linked carbohydrates. J. S ˇ ribar et al. Ammodytoxin inhibitor from Vipera ammodytes serum FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS 6059 The deglycosylation of AIVAS using protein N-gly- cosidase F resulted in a reduction of the molecular mass of both subunits by 3–4 kDa (Fig. 5D). Based on this, it can be concluded that both subunits are gly- cosylated and that the difference in molecular mass between them stems from the difference in composition of their polypeptide chains. The N-terminal sequence of the subunits is identical, so the 23 kDa subunit is probably the C-terminally truncated form of the 25 kDa subunit, the explanation which pertains also in the case of CICS subunits [22]. The extent of glycosyl- ation of AIVAS is similar to that found in homolo- gous c-type PLIs. For example, the molecular mass of the CNF subunit calculated from its amino acid sequence differs, by about 3 kDa, from that deter- mined experimentally, because of glycosylation [23]. Similarly, treatment of CgMIP-I from Cerrophidion (Bothrops) godmani serum with protein N-glycosi- dase F reduced the molecular mass of its subunit by approximately 3 kDa [24]. Inhibition of phospholipase activity by AIVAS Isolated AIVAS (fraction 8 in Fig. 5B) was incubated at room temperature with AtxC and then assayed for phospholipase activity. AIVAS, in a 2 : 1 molar ratio to AtxC, lowered the phospholipase activity of AtxC by more than 80%. At the same molar ratio, CICS reduced the activity of AtxC by only 45% [19], provid- ing further evidence for the preference of serum PLIs for their own venom sPLA 2 molecules and in agree- ment with an auto-envenomation protection role for PLIs. Interestingly, the inhibition of crotoxin subunit CB by CICS was very similar to that of AIVAS on AtxC. At 1 : 1 and 2 : 1 molar ratios of CICS to CB, the enzymatic activity of CB was reduced to 37 and 16%, respectively, of the starting activity [19], whereas in the case of AIVAS and AtxC, at the same ratios, the phospholipase activity of AtxC was reduced to 38 and 19%, respectively (Fig. 8). This indicates that structur- ally, and thus also in terms of the sPLA 2 inhibition mechanism, CICS and AIVAS are closely related. AIVAS was probed also for its ability to inhibit the phospholipase activity of the main structural types of sPLA 2 . As presented in Fig. 8B, besides AtxC, AIVAS inhibited efficiently only ammodytin I 2 , a nontoxic sPLA 2 from V. a. ammodytes venom, as AtxC also a monomeric type II sPLA 2 . A dimeric type II sPLA 2 (crotoxin), type I sPLA 2 molecules (b-bungarotoxin and taipoxin) and a type III bee venom PLA 2 were not significantly inhibited. The sPLA 2 -inhibition spec- tra of AIVAS is in good agreement with the affinity labelling competition results (Fig. 2) where also only 123 86 44.6 31.4 M [kDa] LLE(T ) LL )E(C fr.7 fr.8 Fig. 6. 125 I-Labelled AtxC affinity labelling of purified AIVAS. Aliqu- ots of LL-bound material (LLE), and fractions 7 and 8 from gel chro- matography, were affinity labelled with 125 I-labelled AtxC. The control experiment LLE (C) contained a 100-fold molar excess of native AtxC over the radioactively labelled AtxC. The samples were analyzed by SDS ⁄ PAGE (12.5% acrylamide gel) under reducing con- ditions and the gel was autoradiographed. Three specific adducts were formed, corresponding to molecular masses of 114, 74 and 40 kDa (arrows), as in the case of VAAS labelling in Fig. 2. A B Fig. 7. Sequence analysis of AIVAS. AIVAS was electrotransferred to a PVDF membrane and N-terminally sequenced. (A) The 23 kDa band gave exclusively VAA_1 sequence, which was also the main sequence in the 25 kDa band. In addition, the 25 kDa band con- tained a sequence of low intensity (VAA_2), homologous to VAA_1. (B) Protein sequence database search, using the FASTA algorithm, revealed the highest similarity of VAA_1 and VAA_2 with snake serum c-type PLIs. PLI_20 and PLI_25 are the 20- and 25-kDa subunits of the c-type PLI from Agkistrodon blomhoffii siniticus [4], whereas CNF, CICS_23 ⁄ 25 and CICS_23minor are c-type PLIs from Crotalus durissus terrificus [21,22]. A dot represents a position where no amino acid residue could be assigned. Ammodytoxin inhibitor from Vipera ammodytes serum J. S ˇ ribar et al. 6060 FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS some of the type II sPLA 2 s (AtxA and ammodytin I 2 ) efficiently inhibited the 125 I-AtxC labelling of AIVAS. Preferential inhibition of type II sPLA 2 molecules qualifies AIVAS as a subclass II c-PLI [11]. The neuroprotective effect of CICS has been sug- gested to be the result of its action as a false soluble receptor for presynaptically acting neurotoxins, pre- venting them from binding to their receptors on pre- synaptic membranes [22]. Besides specific binding to neuronal receptors, the neurotoxic action of sPLA 2 molecules also depends on their phospholipase activity [17]. Inhibition of the phospholipase activity of an sPLA 2 neurotoxin leads to its neutralization as a toxin, and therefore AIVAS is expected to act as a natural inhibitor of sPLA 2 neurotoxicity, similarly to CICS. In summary, we have purified biologically active sPLA 2 inhibitor AIVAS from the serum of V. a. am- modytes. AIVAS is structurally related to c-type sPLA 2 inhibitors. It is a noncovalent tetramer of 23 and 25 kDa glycosylated subunits. AIVAS is a potent inhibitor of Atx phospholipase activity, also suggesting its action as a natural inhibitor of their neurotoxic action. Studies of the structural basis of the interaction between AIVAS and Atx will aid in designing novel peptide inhibitors of the phospholipase and hence neu- rotoxic action of sPLA 2 molecules. Experimental procedures V. a. ammodytes serum VAAS was obtained from the Institute of Immunology Inc., Zagreb, Croatia. Blood from V. a. ammodytes was col- lected by heart puncture, incubated for 30 min at 30 °C, centrifuged at 3260 g for 20 min at 4 °C and the serum dec- anted and stored at )20 °C. Surface plasmon resonance SPR experiments were performed at 25 °C, using a Biacore Ò 2000 system (Biacore AB, Uppsala, Sweden). The running and dilution buffer in all experiments was 10 mm Hepes, pH 7.4, 150 mm NaCl, 5 mm CaCl 2 and 0.005% (w ⁄ v) P20 (Pharmacia, Uppsala, Sweden). After purification on CICS- affinity chromatography, polyclonal rabbit anti-CICS IgG was covalently immobilized on a CM-5 sensor chip, as pre- viously described [19]. The SPR signal for immobilized rab- bit anti-CICS IgG was found to be 2275 RU and 3422 RU, corresponding to a protein surface concentration of 2275 pgÆmm )2 and 3422 pgÆmm )2 . To detect the presence of CICS-like molecules in VAAS, 15 lL of diluted serum (pro- tein concentration 0.25 mgÆmL )1 ) was injected for 1 min at a flow rate of 20 lLÆmin )1 . Immediately following that, sPLA 2 , AtxC or CB a (one of the isoforms of CB), at 1– 6 lgÆmL, was injected for another 1 min followed by injec- tion of the running buffer to observe the dissociation rate. At the end of each run, the binding capacity of the sensor chip was regenerated for 30 s with 20 mm HCl. Data were analyzed using the biaevaluation 3.1 software (Biacore AB) after subtracting control signals obtained from the injection of sPLA 2 molecules on the control flow cell. Affinity labelling To 40 lL of serum (0.15 mgÆmL )1 of total protein), 10 lL of 50 mm Hepes, pH 8.2, containing 150 mm NaCl and 0 20 40 60 80 100 120 0 0,5 1 1,5 2 2,5 molar ratio (SABP:AtxC) relative enzymatic activity A B Fig. 8. Determination of the capacity of AIVAS to inhibit phospholi- pase activity. (A) Isolated AIVAS was incubated at room tempera- ture for 30 min with a fixed concentration of AtxC at different molar ratios and the remaining phospholipase activity was deter- mined. The two-fold molar excess of AIVAS over AtxC lowered the enzymatic activity of AtxC by approximately 81%. (B) AIVAS was incubated with representatives of different structural types of sPLA 2 at designated molar ratios and then the remaining phospholi- pase activity was measured. The results shown represent the per- centage of the full activity of an sPLA 2 (in the absence of AIVAS) which remained following incubation of the enzyme with AIVAS. Values are expressed as the mean ± SD of duplicates. AtnI 2 , am- modytin I 2 ; AtxC, ammodytoxin C; b-Butx, b-bungarotoxin; bvPLA 2 , bee venom PLA 2 . J. S ˇ ribar et al. Ammodytoxin inhibitor from Vipera ammodytes serum FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS 6061 2mm CaCl 2 (buffer A) with the addition of 0.625% (w ⁄ v) Triton X-100 and 12.5 lLof 125 I-labelled AtxC to give a final concentration of 10 nm were added and incubated for 30 min at room temperature in the presence or absence of an unlabelled competitor (2 lm final concentration). Where indicated, the Ca 2+ in buffer A was replaced with either 1mm EGTA or 1 mm EDTA. Disuccinimidyl suberate (Pierce, Rockford, IL, USA) was added (100 lm final con- centration), and, after 5 min, the cross-linking reaction was stopped by the addition of SDS ⁄ PAGE sample buffer. Samples were analyzed by SDS ⁄ PAGE and the gels were dried and autoradiographed at )70 °C using Kodak X-Omat AR films (Kodak, Rochester, NY, USA) [26]. To test the effect of pH on AIVAS, aliquots of VAAS were diluted 30-fold in buffers with pH ranging from 5 to 9 and incubated for 30 min at room temperature, followed by a 10-fold dilution in buffer A to adjust the pH back to 8.2. Affinity labelling was then performed as described above. Lectin-affinity chromatography One-hundred-microlitre samples of LL–, WGA– and ConA– Sepharose (Pharmacia) were separately equilibrated in buf- fer A. Five-hundred microlitres of VAAS (1.5 mgÆmL )1 ) was added to each gel suspension and incubated for 4 h at 4 °C with moderate agitation. Following a short centrifuga- tion at 13 100 g for 30 s at 4 °C, the supernatant was removed and the gels were washed with 5 mL of buffer A. To elute bound proteins, the gels were agitated moderately for 1 h at 4 °C in 500 lL of the following elution buffers: 0.5 m N-acetylglucosamine in buffer A for WGA–Sepharose and 0.5 m Me-a-d-mannopyranoside in buffer A for ConA– Sepharose and LL–Sepharose. Following a short centrifuga- tion at 13 100 g for 30 s at 4 °C, the supernatants were removed and the aliquots were analyzed by 10% (w ⁄ v) or 12.5% (w ⁄ v) SDS ⁄ PAGE [27] under reducing conditions [0.5% (w ⁄ v) SDS, 50 mm dithiothreitol, 10% (v ⁄ v) glycerol, 30 mm Tris ⁄ HCl, pH 6.8], followed by silver staining [28]. As the initial step of the isolation of AIVAS from VAAS, 6 mL of VAAS (3.6 mgÆmL )1 total protein) was incubated with 3 mL of LL–Sepharose and the bound proteins were eluted with 6 mL of the elution buffer under the conditions described above. CIM–AtxC affinity chromatography A CIM Epoxy Disc (BIA Separations, Ljubljana, Slovenia) was washed on an A ¨ KTA FPLC apparatus (Amersham Pharmacia Biotech, Uppsala, Sweden) with buffer B (0.5 m Na 2 HPO 4 , pH 8.2) at 1 mLÆmin )1 until the absorbance at 280 nm (A 280 ) was zero. Two millilitres of AtxC (2 mgÆmL )1 in buffer B) was loaded at room temperature on the disk at a flow rate of 0.02 mLÆmin )1 and the break- through was collected. The disk was then immersed in this breakthrough and incubated for 24 h at room temperature. Following the incubation, the disk was washed again with 5 volumes of buffer B and the remaining reactive groups were blocked by loading 2 mL of 1 m ethanolamine in buf- fer B at room temperature at a flow rate of 0.02 mLÆmin )1 . The disk was again immersed in the resulting breakthrough for 24 h at room temperature and finally washed at 1 mLÆ min )1 with 5 volumes of buffer B containing 1 m NaCl, and then with 5 volumes of water, and stored in 20% (v ⁄ v) ethanol until use. Routinely, about 50% of the applied AtxC was bound to the disk. To isolate AIVAS, the CIM–AtxC disk was first equili- brated in 10 mL of buffer A at a flow rate of 0.1 mL Æmin )1 . Two millilitres of the eluate from the LL-affinity chroma- tography step was loaded at a flow rate of 0.02 mLÆmin )1 , and the unbound proteins were washed away with 10 mL of buffer A at a flow rate of 0.1 mLÆmin )1 . Bound proteins were eluted with 0.1 m glycine ⁄ HCl, pH 2.8, containing 150 mm NaCl and 2 mm CaCl 2 . Fractions of 0.5 mL were collected directly into 0.2 mL of 0.5 m Tris ⁄ HCl, pH 8.0, containing 150 mm NaCl, and then immediately dialysed into buffer A. The fractions were analysed by SDS ⁄ PAGE as described previously [27,28]. Gel filtration chromatography Gel filtration chromatography was performed on an A ¨ KTA FPLC apparatus (Amersham Pharmacia Biotech). The frac- tion from the CIM–AtxC affinity chromatography step containing AIVAS was dialysed against buffer A and applied to a Superdex 75 HR 10 ⁄ 30 column (Amersham Pharmacia Biotech) previously calibrated with the follow- ing molecular mass standards: Dextran blue (2000 kDa); BSA (65 kDa); egg albumin (45 kDa); chymotrypsinogen (25 kDa); cytochrome c (12.3 kDa); aprotinin (6.5 kDa); and l-tyrosine (0.1 kDa). Proteins were eluted at 0.5 mLÆ min )1 with buffer A and 1-mL fractions were collected, concentrated in Centricon YM-10 concentrators (Millipore, Billerica, MA, USA) and analysed by SDS ⁄ PAGE, as described previously [27,28]. In SDS ⁄ PAGE analysis under nonreducing conditions, dithiothreitol was omitted from the sample buffer. Western blotting Samples were run on SDS ⁄ PAGE [12.5% (w ⁄ v) polyacryl- amide gels]. The gel was soaked in the blotting buffer: 25 mm Tris ⁄ HCl, pH 8.3, 192 mm glycine, 0.1% (w ⁄ v) SDS, 20% (v ⁄ v) methanol. A semidry blotter (Biometra, Go ¨ ttingen, Germany) was used for 15 min (2 W per 100 cm 2 of gel) at room temperature to transfer the pro- teins from the gel to a nitrocellulose membrane (Costar, Cambridge, MA, USA) for subsequent immunodetection or to a PVDF membrane (Bio-Rad, Hercules, CA, USA) for N-terminal sequence analysis. Ammodytoxin inhibitor from Vipera ammodytes serum J. S ˇ ribar et al. 6062 FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS Deglycosylation Deglycosylation was performed essentially as described pre- viously [23]. To 0.1 lg of AIVAS in 18 lLof50mm Na 2 HPO 4 , pH 7.5, 3 lL (3 U) of protein N-glycosidase F (Roche, Mannheim, Germany) was added and incubated at 37 °C for 24 h. SDS ⁄ PAGE analysis, as explained above, followed. In the control experiment, no protein N-glycosi- dase F was added to the reaction mixture. Immunodetection The nitrocellulose membrane was incubated with rabbit polyclonal anti-CICS IgG diluted 1 : 1000. The position of specific proteins on the membrane was revealed using the BM chemiluminescence western blotting kit (Roche Molec- ular Biochemicals, Mannheim, Germany) following the manufacturer’s instructions. N-terminal sequence analysis Proteins were detected on the PVDF membrane using Coo- massie Brilliant Blue R250. Protein bands were excised and analysed by automated Edman N-terminal sequencing anal- ysis on an Applied Biosystems Procise 492 A protein- sequencing system (Applied Biosystems, Foster City, CA, USA). Inhibition of phospholipase activity Isolated AIVAS was incubated in 50 l Lof50mm KCl, 1mm CaCl 2 ,50mm Tris ⁄ HCl, pH 7.4, at room tempera- ture for 30 min with 72 nm AtxC, at different molar ratios. In the same way, AIVAS was incubated also with the repre- sentatives of other structural types of sPLA 2 molecules – 0.4 and 2 molar parts of AIVAS were mixed with 1 part of an sPLA 2 . In a 96-well plate, a 1 lL aliquot of the incuba- tion mixture was added to 200 lL of 0.09% (w ⁄ v) BSA (fatty-acid free) in 50 mm KCl, 1 mm CaCl 2 ,50mm Tris ⁄ HCl, pH 7.4, followed by 100 lL of a substrate, lipid vesicles of 1-hexadecanoyl-2-(1-pyrenedecanoyl)-sn-glycero- 3-phosphoglycerol (4.2 l m) (Invitrogen, Carlsbad, CA, USA) in the same buffer. Phospholipase activity was mea- sured using a modified method of Radvanyi et al. [29]. Flu- orescence was monitored in 10 kinetic cycles on a SAFIRE microplate monochromator reader (Tecan, Salzburg, Aus- tria) using the following parameters: excitation wavelength 342 nm, emission wavelength 395 nm, number of flashes 10 and integration time 40 ls. To determine the enzymatic activity of the sample, the slopes of the curves were calcu- lated, the background fluorescence subtracted and the resulting value compared with that of the sample with known enzymatic activity. To determine the background fluorescence only AIVAS, without an sPLA 2 , was present. To establish the full enzymatic activity of particular types of sPLA 2 (100% value) only the sPLA 2 , without AIVAS, was added to the reaction mixture. Acknowledgements This work was supported by grant P1-0207-0106 from the Slovenian Ministry of Higher Education, Science and Technology and by the PROTEUS program from the same Slovenian Ministry and the French Ministry of Foreign Affairs. We thank Maja Lang Balija from the Institute of Immunology, Croatia, for kindly pro- viding us with Vipera a. ammodytes serum; Adrijana Leonardi from Jozˇ ef Stefan Institute, for performing the protein N-terminal sequence analyses; and BIA Separations, Slovenia, for their kind gift of CIM Epoxy Discs. We sincerely thank Dr Roger H. Pain for critical reading of the manuscript. References 1 Dunn RD & Broady KW (2001) Snake inhibitors of phospholipase A 2 enzymes. Biochim Biophys Acta 1533, 29–37. 2 Faure G (2000) Natural inhibitors of toxic phospho- lipases A 2 . Biochimie 82, 833–840. 3 Fortes-Dias CL (2002) Endogenous inhibitors of snake venom phospholipases A 2 in the blood plasma of snakes. Toxicon 40, 481–484. 4 Ohkura N, Okuhara H, Inoue S, Ikeda K & Hayashi K (1997) Purification and characterization of three distinct types of phospholipase A 2 inhibitors from the blood plasma of the Chinese mamushi, Agkistrodon blomhoffii siniticus. Biochem J 325, 527–531. 5 Kogaki H, Inoue S, Ikeda K, Samejima Y, Omori-Satoh T & Hamaguchi K (1989) Isolation and fundamental properties of a phospholipase A 2 inhibitor from the blood plasma of Trimeresurus flavoviridis. J Biochem (Tokyo) 106, 966–971. 6 Inoue S, Kogaki H, Ikeda K, Samejima Y & Omori- Satoh T (1991) Amino acid sequences of the two sub- units of a phospholipase A 2 inhibitor from the blood plasma of Trimeresurus flavoviridis. Sequence homolo- gies with pulmonary surfactant apoprotein and animal lectins. J Biol Chem 266, 1001–1007. 7 Ohkura N, Inoue S, Ikeda K & Hayashi K (1993) Isola- tion and amino acid sequence of a phospholipase A 2 inhibitor from the blood plasma of Agkistrodon blomh- offii siniticus. J Biochem (Tokyo) 113, 413–419. 8 Okumura K, Ohno A, Nishida M, Hayashi K, Ikeda K & Inoue S (2005) Mapping the region of the alpha-type phospholipase A 2 inhibitor responsible for its inhibitory activity. J Biol Chem 280, 37651–37659. J. S ˇ ribar et al. Ammodytoxin inhibitor from Vipera ammodytes serum FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS 6063 9 Okumura K, Ohkura N, Inoue S, Ikeda K & Hayashi K (1998) A novel phospholipase A 2 inhibitor with leu- cine-rich repeats from the blood plasma of Agkistrodon blomhoffii siniticus. Sequence homologies with human leucine-rich alpha2-glycoprotein. J Biol Chem 273 , 19469–19475. 10 Okumura K, Inoue S, Ikeda K & Hayashi K (2002) Identification of beta-type phospholipase A 2 inhibitor in a nonvenomous snake, Elaphe quadrivirgata. Arch Biochem Biophys 408, 124–130. 11 Lizano S, Domont G & Perales J (2003) Natural phos- pholipase A 2 myotoxin inhibitor proteins from snakes, mammals and plants. Toxicon 42, 963–977. 12 Kini RM (1997) Venom Phospholipase A 2 Enzymes: Structure, Function and Mechanism. John Wiley & Sons, Chichester. 13 Schaloske RH & Dennis EA (2006) The phospholipase A 2 superfamily and its group numbering system. Biochim Biophys Acta 1761, 1246–1259. 14 Kudo I & Murakami M (2002) Phospholipase A 2 enzymes. Prostaglandins Other Lipid Med 68–69, 3–58. 15 Murakami M & Kudo I (2004) Secretory phospholipase A 2 . Biol Pharm Bull 27, 1158–1164. 16 Hanasaki K & Arita H (2002) Phospholipase A 2 recep- tor: a regulator of biological functions of secretory phospholipase A 2 . Prostaglandins Other Lipid Med 68– 69, 71–82. 17 Krizˇ aj I & Gubens ˇ ek F (2000) Neuronal receptors for phospholipases A 2 and beta-neurotoxicity. Biochimie 82, 807–814. 18 Lambeau G & Lazdunski M (1999) Receptors for a growing family of secreted phospholipases A 2 . Trends Pharmacol Sci 20, 162–170. 19 Faure G, Villela C, Perales J & Bon C (2000) Interac- tion of the neurotoxic and nontoxic secretory phospho- lipases A 2 with the crotoxin inhibitor from Crotalus serum. Eur J Biochem 267, 4799–4808. 20 Lizano S, Lomonte B, Fox JW & Gutierrez JM (1997) Biochemical characterization and pharmacological prop- erties of a phospholipase A 2 myotoxin inhibitor from the plasma of the snake Bothrops asper. Biochem J 326, 853–859. 21 Fortes-Dias CL, Lin Y, Ewell J, Diniz CR & Liu TY (1994) A phospholipase A 2 inhibitor from the plasma of the South American rattlesnake (Crotalus durissus terrif- icus). Protein structure, genomic structure, and mecha- nism of action. J Biol Chem 269, 15646–15651. 22 Perales J, Villela C, Domont GB, Choumet V, Saliou B, Moussatche H, Bon C & Faure G (1995) Molecular structure and mechanism of action of the crotoxin inhibitor from Crotalus durissus terrificus serum. Eur J Biochem 227, 19–26. 23 Dos Santos RM, Oliveira LC, Esteva ˜ o-Costa MI, de Lima ME, Santoro MM & Fortes-Dias CL (2005) Inhi- bition of crotoxin binding to synaptosomes by a recep- tor-like protein from Crotalus durissus terrificus (the South American rattlesnake). Biochim Biophys Acta 1717, 27–33. 24 Lizano S, Angulo Y, Lomonte B, Fox JW, Lambeau G, Lazdunski M & Gutierrez JM (2000) Two phospho- lipase A 2 inhibitors from the plasma of Cerrophidion (Bothrops) godmani which selectively inhibit two differ- ent group-II phospholipase A 2 myotoxins from its own venom: isolation, molecular cloning and biological prop- erties. Biochem J 346, 631–639. 25 S ˇ ribar J, C ˇ opic ˇ A, Paris ˇ A, Sherman NE, Gubens ˇ ek F, Fox JW & Krizˇ aj I (2001) A high affinity acceptor for phospholipase A 2 with neurotoxic activity is a calmodu- lin. J Biol Chem 276, 12493–12496. 26 C ˇ opic ˇ A, Vuc ˇ emilo N, Gubens ˇ ek F & Krizˇ aj I (1999) Identification and purification of a novel receptor for secretory phospholipase A 2 in porcine cerebral cortex. J Biol Chem 274, 26315–26320. 27 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 28 Merril CR, Goldman D, Sedman SA & Ebert MH (1981) Ultrasensitive stain for proteins in polyacryl- amide gels shows regional variation in cerebrospinal fluid proteins. Science 211, 1437–1438. 29 Radvanyi F, Jordan L, Russo-Marie F & Bon C (1989) A sensitive and continuous fluorometric assay for phos- pholipase A 2 using pyrene-labeled phospholipids in the presence of serum albumin. Anal Biochem 177, 103–109. Ammodytoxin inhibitor from Vipera ammodytes serum J. S ˇ ribar et al. 6064 FEBS Journal 274 (2007) 6055–6064 ª 2007 The Authors Journal compilation ª 2007 FEBS . understanding of the patho- physiology of this group of enzymes. In this work we present the characterization of the first PLI from the serum of the most dangerous. The first phospholipase inhibitor from the serum of Vipera ammodytes ammodytes Jernej S ˇ ribar 1 , Lidija Kovac ˇ ic ˇ 1 ,