ARTICLE IN PRESS

0041-0101/$ - see

doi:10.1016/j.tox

�CorrespondiE-mail addre

Toxicon 50 (2007) 871–892

www.elsevier.com/locate/toxicon

Review

Understanding the molecular mechanism underlying thepresynaptic toxicity of secreted phospholipases A2

Joze Pungercar, Igor Krizaj�

Department of Molecular and Biomedical Sciences, Jozef Stefan Institute, SI-1000 Ljubljana, Slovenia

Received 15 May 2007; received in revised form 13 July 2007; accepted 20 July 2007

Available online 22 August 2007

Abstract

An important group of toxins, whose action at the molecular level is still a matter of debate, is secreted phospholipases

A2 (sPLA2s) endowed with presynaptic or b-neurotoxicity. The current belief is that these b-neurotoxins (b-ntxs) exert theirtoxicity primarily due to their extracellular enzymatic action on the plasma membrane of motoneurons at the

neuromuscular junction. However, the discovery of several extra- and intracellular proteins, with high binding affinity for

snake venom b-ntxs, has raised the question as to whether this explanation is adequate to account for all the observed

phenomena in the process of presynaptic toxicity. The purpose of this review is to critically examine the various published

studies, including the most recent results on internalization of a b-ntx into motor nerve terminals, in order to contribute to

a better understanding of the molecular mechanism of b-neurotoxicity. As a result, we propose that presynaptic

neurotoxicity of sPLA2s is a result of both extra- and intracellular actions of b-ntxs, involving enzymatic activity as well as

interaction of the toxins with intracellular proteins affecting the cycling of synaptic vesicles in the axon terminals of

vertebrate motoneurons.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Venom; Phospholipase A2; Presynaptic neurotoxicity; Molecular mechanism

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 872

2. Structural versatility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 872

3. Pathology and neurophysiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 872

4. Cellular models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 876

5. Structural determinants of b-neurotoxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 878

6. Enzymatic activity of b-ntxs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 879

7. Molecular targets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 880

7.1. Extracellular b-ntx-binding molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 880

7.2. Intracellular b-ntx-binding molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 882

8. Molecular models of b-ntxs action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883

front matter r 2007 Elsevier Ltd. All rights reserved.

icon.2007.07.025

ng author. Tel.: +386 1 477 36 26; fax: +386 1 477 39 84.

ss: igor.krizaj@ijs.si (I. Krizaj).

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892872

8.1. Action of b-ntxs on nerve cells is exclusively extracellular . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883

8.2. Action of b-ntxs on nerve cells is both extra- and intracellular . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 885

9. Perspectives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 887

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 888

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 888

1. Introduction

Animal, plant and microbial toxins constitute arich source of pharmacologically active compounds.This review is focused on the b-neurotoxins (b-ntxs),secreted phospholipases A2 (sPLA2s), that blockneuromuscular transmission in vertebrate skeletalmuscles. These toxins are most abundant in thevenom of snakes from the families Elapidae andViperidae, and are also found in some insectvenoms. In spite of many studies there are stillsome questions concerning their molecular mode ofaction that have to be answered before theseneurotoxins can be effectively used as biologicaltools or as prototypes of new therapeutic agents.

2. Structural versatility

Secreted PLA2s are enzymes that catalyze hydrolysisof the sn-2 ester bond in 1,2-diacyl-sn-3-phosphogly-cerides (EC 3.1.1.4). They are relatively small proteinsof 13–19kDa, very similar in structure irrespective ofwhether they originate from snake venoms or mam-malian cells. Secreted PLA2s comprise 17 knowndistinct groups or isoenzymes (Ho et al., 2001; Kudoand Murakami, 2002; Rouault et al., 2003); however,b-neurotoxicity has been detected only in members ofgroups IA, IIA and IIIA (Kini, 1997). Although theirpathophysiological action appears to be closely similar(see Section 3), b-ntxs can be composed of one, two,three or five subunits (Table 1). Most oligomeric b-ntxsinvolve non-covalent association of two or moreproteins, at least one of them being the enzymaticallyactive sPLA2. The exception to this rule has beenfound in the venoms of Bungarus snakes from whichseveral b-ntxs have been isolated, such as b-bungar-otoxins (b-Butxs). In b-Butxs an enzymatically activebasic sPLA2 (A subunit) is covalently connected to ahomolog of bovine pancreatic trypsin inhibitor (Bsubunit) by an inter-subunit disulfide bond.

3. Pathology and neurophysiology

As b-ntxs are usually injected into the peripheryof the snake’s prey, the likely target of these toxins

is the peripheral nervous system. The most probablecause of death is respiratory failure resulting fromthe blockade of neuromuscular transmission in therespiratory muscles. The paralysis of an experimen-tal animal is flaccid. Although skeletal muscles areinnervated by cholinergic synapses, b-ntxs do notact exclusively on cholinergic neurons (Harris,1985). Certain cholinergic neurons are not affectedby b-ntxs, e.g. from guinea pig ileum, whereas somenon-cholinergic neurons, such as hippocampal,cerebellar granule and cortical neurons from thecentral nervous system (CNS), which are allglutamatergic (i.e. pertaining to glutamate neuro-transmission), are sensitive to b-ntxs.

Under normal conditions b-ntxs do not pass theblood–brain barrier, so the most relevant andconvenient tissue to study the effect of b-ntxs ontransmitter release are isolated neuromuscular pre-parations that have been partially paralyzed bylowering the extracellular Ca2+ concentration tobelow 0.5mM. Under these conditions, smallchanges in transmitter release will be observeddirectly as changes in the amplitude of the twitchresponse. Indirectly stimulated isolated neuromus-cular preparations exhibit a complex responsefollowing exposure to b-ntxs. Despite inter- andintraspecies differences in the response to b-ntxs,three distinct phases are clearly visible in most cases.Typically, there is an initial transient inhibition ofevoked transmitter release (phase 1) followed byfacilitation of transmitter release (phase 2) and afinal phase during which there is a progressive fall inevoked release until transmission fails completely.Miniature end-plate potential (mepp) frequency issimilarly affected, although spontaneous releasetends to occur at a low frequency after the failureof evoked neuromuscular transmission. Impor-tantly, the mepp amplitude does not changesignificantly, indicating that synaptic vesicles (SVs)do not fuse extensively, if at all, inside the nerveterminal and that the acetylcholine (ACh) loadingapparatus is not impaired by the action of b-ntxs(Chang, 1985).

The first, transiently inhibitory phase of b-ntxsaction at the neuromuscular junction is not always

ARTICLE IN PRESS

Table 1

Representative neurotoxic sPLA2s and some of their characteristics

X -neurotoxin Venom Structural features Lethality in

mouse

(mg/kg)a

Binding proteins in

nerve tissuebReferencesc

AtxA Vipera ammodytes

ammodytes

Monomeric 21 (i.v.) M-sPLA2R, CaM,

PDI, 14-3-3 proteins

S: Ritonja and

Gubensek (1985)IIA sPLA2

L: Thouin et al. (1982)

B: Vardjan et al. (2001),

Sribar et al. (2001,

2003a, 2005)

Bee venom

sPLA2

Apis mellifera Monomeric 230 (i.c.) CaM S: Shipolini et al. (1974)

IIIA sPLA2 L: Lambeau et al. (1989)

B: Sribar et al. (2001)

Crotoxins Crotalus durissus

terrificus

Dimeric 60-240 (i.v.) Crocalbin, CaM S: Aird et al. (1986),

Faure et al. (1991, 1994)Subunit B—IIA sPLA2

L: Faure and Bon (1988)Subunit A—IIA sPLA2-likeB: Hseu et al. (1999),

Sribar et al. (2001)

X 1�5-Butx Bungarus multicinctus Dimeric 19-130 (i.p.) v.-d. K+ channel S: Kondo et al.

(1982a, b)Subunit A—IA sPLA2

L: Kondo et al. (1982b)S-S linked toB: Scott et al. (1990)subunit B—BPTI-like

Taipoxin Oxyuranus scutellatus Trimeric 2 (i.v.) M-sPLA2R, NP,

TCBP-49

S: Fohlman et al. (1977),

Lind (1982), Lind and

Eaker (1982)scutellatus Subunit a—IA sPLA2, toxic

L: Fohlman et al. (1979)Subunit b—IA sPLA2-like

B: Lambeau et al. (1990,

1994), Dodds et al.

(1995), Schlimgen et al.

(1995)

Subunit g—IB sPLA2,

glycosylated

Textilotoxin Pseudonaja textilis Pentameric 1 (i.v.) M-sPLA2R S: Pearson et al. (1993)

Subunit A—IA sPLA2, toxic L: Coulter et al. (1983)

Subunit B—IA sPLA2 B: Lambeau et al. (1990,

1994)Subunit C—IA sPLA2

Subunit D—two identical

S-S linked IB sPLA2s,

glycosylated

ai.v., intravenous; i.c. intracisternal; i.p., intraperitoneal.bCaM, calmodulin; NP, neuronal pentraxin; PDI, protein disulfide isomerase; TCBP-49, taipoxin-associated calcium-binding protein

49; M-sPLA2R, M-type sPLA2 receptor.cS, sequence; L, lethality; B, binding proteins.

J. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 873

observed. When it is present, the inhibitory phasestarts soon after the addition of a b-ntx and lasts forseveral minutes. The intensity of this phase varies,depending to some extent on the neuromuscularpreparation and the b-ntx used. Experimental dataindicate that this early inhibition is independent ofenzymatic activity of a b-ntx (Caratsch et al., 1981;Su and Chang, 1984; Prijatelj et al., 2006) and the

most frequent explanation offered to account forthis effect is that it is due to specific binding of theb-ntx to the presynaptic membrane. While Caratschet al. (1981) observed only little temperature influencein frog neuromuscular preparation, E.G. Rowan(personal communication) has shown that thisphase of the b-Butx effect is temperature dependentin mouse neuromuscular preparation. Recently, the

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892874

implication of phospholipase activity in this earlyphase also has been suggested (Caccin et al., 2006).

Facilitated neurotransmitter release in the secondphase reaches its peak in 10–20min depending onexperimental conditions. For instance, in mousephrenic nerve–diaphragm at 37 1C in low Ca2+

(0.38–0.50mM) Krebs solution, 10 mg/ml of AtxAelicited maximal release of ACh after 13min(Prijatelj et al., 2006). The significance of thephospholipase activity in this phase has not beenclarified. The experimental evidence shows that,while in mammalian systems (mouse, rat) inhibitionof the hydrolytic activity of a b-ntx by Sr2+, lowtemperature, chemical modification or mutagenesisdoes not have a significant effect on facilitation; infrog, lowering the enzymatic activity of a b-ntx alsoreduces the facilitatory effect (reviewed by Chang,1985; Prijatelj et al., 2006). One explanation for theaugmentation of neurosecretion from the motornerve terminal by a b-ntx in the second phase is thebinding of the b-ntx to or near voltage-dependentK+ channels, with concomitant inhibition of theinward K+ currents. The result of retarded repolar-ization of the axolemma would be an enhanced andprolonged exocytosis of the neurotransmitter. It isknown that b-ntxs influence terminal ionic currents,most probably K+ currents, and that the time of theinfluence matches exactly the time of facilitation(Dreyer and Penner, 1987; Rowan and Harvey,1988; Krizaj et al., 1995). However, direct bindingto K+ channels has been demonstrated only forb-Butx and, in that case, the connection betweenbinding to K+ channels and neurotoxicity wasobserved only in certain cells or species (seeSection 7). A second explanation for the increasedneurosecretion by b-ntxs was originally proposed byStrong et al. (1976) and resumed by Rigoni et al.(2005). They suggested that b-ntxs hydrolyzeplasma membrane phospholipids and form lysopho-spholipids and fatty acids (lysoPL and FA), which,due to their molecular geometry, synergisticallypromote the exocytosis of SVs filled with neuro-transmitter. However, according to this hypothesis,the second phase should depend entirely on theenzymatic activity of a b-ntx in mammalian systemsalso.

In the third phase, following facilitation, themuscle response begins to decline and transmissionfails completely in the next hour or two, dependingon the toxin, its concentration and other experi-mental conditions. For example, in the case of10 mg/ml of AtxA at 37 1C in low Ca2+

(0.38–0.50mM) Krebs solution, the mouse phrenicnerve–diaphragm is blocked in 90min (Prijateljet al., 2006). This last phase of the neuromuscularresponse to b-ntxs is strictly dependent on phos-pholipase activity. If the enzymatic activity of ab-ntx is partially or completely inhibited, theneuromuscular block is delayed or does not occurat all (Prijatelj et al., 2006). The time to transmissionfailure is concentration dependent. After the neu-romuscular transmission is completely blocked themuscle still remains fully responsive to directstimulation, and the resting membrane potential ofthe muscle fibers remains unchanged. The toxins donot influence the transmission of an action potentialalong the axon (reviewed in Chang, 1985).

There appears to be no correlation betweenphospholipase activity and b-neurotoxicity of b-ntxs (see, for example, Rigoni et al., 2004), but itshould be noted that enzymatic activity is usuallymeasured on synthetic substrates, rather than onpresynaptic membranes. Stimulation of the neuro-muscular preparation is not required for the fullexpression of toxicity (Prasarnpun et al., 2004);however, if we stimulate the neuromuscular pre-paration, then a weak correlation between thefrequency of stimulation and the time to reachcomplete transmission failure can be noticed (Cull-Candy et al., 1976; Chang and Su, 1982).

The neurotoxic effect is, however, temperaturedependent. While at 37 1C AtxA produced adistinctive triphasic effect in a mouse phrenicnerve–hemidiaphragm preparation, no such effecton muscle twitching was apparent at room tem-perature (23 1C), although the final block was stillobserved. On the contrary, taipoxin was able toelicit a clear triphasic effect also at this temperature(E.G. Rowan, personal communication). As there isclearly no correlation between enzymatic activityand toxicity of b-ntxs (e.g. taipoxin with specificenzymatic activity on phosphatidylcholine (PC) aslow as 0.4 mmol/min/mg produced a blockade ofmouse hemidiaphragm much more effectively thanAtxA with specific activity of 280 mmol/min/mg),which is even more evident at lower temperatures,the reason for such diverse actions of both toxinsmay lie in their difference in binding to theirreceptors on plasma membrane. A large variationbetween b-ntxs in the stability of their associationwith the presynaptic membrane was reported bySimpson et al. (1993). While binding to the mousephrenic nerve–hemidiaphragm was completely re-versible in the case of notexin, binding of crotoxin

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 875

was slightly reversible and that of b-Butx, taipoxinand textilotoxin practically irreversible. The com-parative study of AtxC and crotoxin gave similarresults, as the association of monomeric AtxC withpresynaptic membranes was completely reversible,the binding of dimeric crotoxin was much less(Delot and Bon, 1993; Krizaj et al., 1997). Theadditional subunits in oligomeric b-ntxs appear topromote their specific binding to the presynapticmembrane and enable toxins to act more efficiently.

Transmission failure did not occur if the exposuretime to b-Butx was less than 5min. Interestingly,there was no difference in the rate of transmissionfailure if the toxin was washed out after 10 or 20minof exposure or if the toxin was not washed out at all(Prasarnpun et al., 2004). This tells us that, after5–10min, either (1) the damage of the motor nerveterminal due to b-Butx is irreversible, which is notlikely to be the case since the addition of Sr2+ after15–30min abolished the blockade (Simpson et al.,1993), (2) b-Butx is internalized or (3) the toxin isirreversibly associated with the plasma membrane.

The experiments of Simpson et al. (1993) on themouse hemidiaphragm preparation with neutraliz-ing antibodies and Sr2+, an inhibitor of sPLA2

activity, revealed that, following association withthe plasma membrane, b-ntxs became inaccessibleto antibodies, but can still be affected by Sr2+. Sr2+

ions, in contrast to antibodies, can also enter theneuronal cytosol during depolarization throughvoltage-gated Ca2+ channels, so they could inhibitb-ntxs on the external as well as the internal side ofthe axolemma. Inhibition of b-ntxs action by Sr2+

up to 30min after the administration of the toxins,but not by the neutralizing antibodies, indicates thatafter 30min of action there are still not enoughhydrolytic products, lysoPL and FA, to achievethe blockade of neurotransmission. It is also clearthat the antibodies cannot neutralize the toxicity ofb-ntxs because they do not have access to them.When Sr2+ was added 30–60min after the exposureof the neuromuscular preparation to b-ntxs, theneutralizing effect was only partial. Addition ofSr2+ after 60min gave no protection at all, and thedamage to the nerve terminal was irreversible.

Ultrastructural or immunocytochemical analysesof the neuromuscular preparations treated withb-ntxs have been performed mainly on preparationsfollowing the complete block of neuromusculartransmission. Using different b-ntxs, the generalfeatures of the acute phase of intoxication ofneuromuscular junctions observed by electron

microscopy were a substantial reduction of SVswithin the nerve terminal, numerous O-shapedinvaginations of the axolemma coated with elec-tron-dense material (presumably clathrin), forma-tion of large vesicles, and swollen and damagedmitochondria, in some cases with complete loss ofthe internal structure of cristae. Schwann cellprocesses entering the synaptic gutter are alsoobserved, thus reducing the contact area betweenthe nerve terminal and muscle fiber (Cull-Candyet al., 1976; Chang and Su, 1982; Lee et al., 1984;Prasarnpun et al., 2004). Observations made over anextended period have shown that synaptic boutonsbecome detached from their anchoring counterpartson the postsynaptic membrane and undergo com-plete degeneration (Prasarnpun et al., 2004). Thetotal degeneration of the nerve terminal appears tooccur very suddenly as, prior to degeneration, thereis little evidence that the plasma membrane of thenerve terminal is damaged (Dixon and Harris, 1999;Kattah et al., 2002; Wei et al., 2003; Rigoni et al.,2005).

The most systematic morphological analysis ofthe neuromuscular junction intoxicated by b-ntxswas made by Dixon and Harris (1999) andPrasarnpun et al. (2004). They analyzed the effectsof b-Butx on mouse and rat phrenic nerve–hemi-diaphragms in vitro and in vivo, and found that SVdensity was substantially reduced in intoxicatednerve terminals, whereas mitochondria were swollenand floccular in appearance compared with controlpreparations. SNAP-25 and syntaxin, the twoplasma membrane proteins involved in the SNAREcomplex formation, remained in place in themajority of b-Butx-paralyzed nerve terminals, asrevealed by quantitative confocal microscopy. Avery important conclusion of the study by Prasarn-pun et al. (2004) was that the mean area of the nerveterminal boutons exposed to b-Butx was notsignificantly different from that of the nerveterminal boutons in control muscles. This was infact the first quantitative assessment of the influenceof b-ntxs on motor nerve terminal area. Otherreferences that are usually cited in support ofswelling of motor nerve terminals due to b-ntxs(e.g. Chen and Lee, 1970; Cull-Candy et al., 1976;Gopalakrishnakone and Hawgood, 1984; Dixonand Harris, 1999) do not describe this effect at all.In a recent work, Rigoni et al. (2005) also claim thatmotor nerve terminals enlarge due to the action ofb-ntxs; however, their conclusion is not based onmorphometry and it is also difficult to visualize it

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892876

from their electron micrographs as respectivecontrols have not been shown. Expansion of theplasma membrane of the motor nerve ending due tothe blockade of the SV endocytosis should be clearlyvisible, as demonstrated by Ringstad et al. (1999).Regarding the credibility of available results in thisrespect, it appears that motoneuron terminals differfrom those of the CNS neurons whose plasmamembrane clearly expands under the action ofb-ntxs (Rigoni et al., 2004, 2005; Bonanomi et al.,2005).

Mitochondria play an important role in cytosolicCa2+ homeostasis and ATP production. b-ntxscause the uncoupling of mitochondria (Howard,1975; Nicholls et al., 1985), which results in thebreakdown of energy-dependent concentration gra-dients. One of the effects is the movement of Ca2+

ions from mitochondria down to their concentra-tion gradient, resulting in an increase of thecytosolic Ca2+ level and, consequently, an increasein the frequency of spontaneous transmitter release.

The intoxicated nerve terminals contained 80%fewer SVs, as determined from transmission elec-tron micrographs and by quantitative fluorescencemicroscopy (Prasarnpun et al., 2004), in fullagreement with the findings of Chen and Lee(1970) and Abe et al. (1976). This suggests thatdepletion of SVs is primarily responsible for theacute onset of paralysis by b-ntxs. Substantialreduction of the fluorescence signal of the SVmarker synaptophysin in b-Butx-intoxicated nerveterminals is another major difference between theresults obtained on the motoneuron and on CNSneurons. In cultured CNS neurons, the signal forsynaptophysin I was not attenuated but just shiftedto bulges—swellings on the plasma membrane(Rigoni et al., 2004, 2005; Bonanomi et al., 2005).This further indicates that processes triggered byb-ntxs at neuromuscular junctions of intoxicatedanimals and in primary neuronal CNS cultures maynot be completely comparable, suggesting thatcaution is advisable in the extrapolation of results.

Reports on pathomorphological changes in pre-parations before the complete arrest of neuromus-cular communication due to b-ntxs are scarce. Forexample, Gopalakrishnakone and Hawgood (1984)reported that mouse neuromuscular junctionsshowed minimal morphological alterations prior tothe onset of clinical signs induced by crotoxin(30min). R.W. Dixon and J.B. Harris (personalcommunication) report that during the period ofmaximal potentiation of transmitter release (i.e.

phase 2) there is evidence of the migration of SVsand their accumulation around release sites on theaxolemma.

4. Cellular models

The neuromuscular junction is anatomically acomplex structure and hence an experimentalapproach at the molecular level is rather difficultand cumbersome. Different neuronal and non-neuronal cells have been used in an attempt toreplicate the response of cholinergic nerve terminalsto the b-ntxs. It is accepted that replication wouldnot be complete but it is always hoped that the basicmolecular events would be similar. Most experi-ments have used primary cell lines from CNS, suchas hippocampal, cerebellar granule and corticalneurons. In addition, PC12 and NSC34 cells havebeen also utilized. The majority of these cells arenon-cholinergic, which limits their use as models ofvertebrate motor nerve terminals. This is, however,not the case with a model very similar to actualmotoneurons, the NSC34 cell line. We will alsomention briefly one of the simplest eukaryotic cells,the yeast Saccharomyces cerevisiae, a primitiveeukaryotic organism, as a potential model forstudying certain cellular and toxic effects of b-ntxs.

In 2001, it was reported that b-Butx is a potentinducer of caspase-independent apoptosis in pri-mary rat hippocampal neurons but not in non-neuronal cells, astrocytes (Herkert et al., 2001).b-Butx was rapidly internalized into these CNScells after specific binding to voltage-dependentK+ channels and found preferentially localized tolysosomes. It appeared that hippocampal neuronsmight serve not only as a cellular model for studyingthe b-neurotoxicity of sPLA2s but also as an in vitromodel for certain neurodegenerative diseases(Shakhman et al., 2003). It was shown that AtxAis readily accumulated by hippocampal neurons andis finally translocated, probably through the cytosolwhere it may also be enzymatically active, to thenucleus of these cells (Petrovic et al., 2004). Otherdata suggest that the enzymatic activity of b-Butx isresponsible for the death of cerebellar granuleneurons, where caspase-3 activation is involved(Tseng and Lin-Shiau, 2003; Chen, 2005). Phospho-lipase A2 activity has also been shown to beimportant for exocytosis and neurotransmitterrelease in hippocampal neurons and PC12 cellstreated with either crotoxin subunit B or non-toxichuman group IIA sPLA2 (Wei et al., 2003). In both

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 877

cells, extracellular application of each sPLA2 pro-duced an immediate increase in neurotransmitterrelease, whereas its intracellular application led toan inhibition of neurotransmitter release. This indi-cates that the products of phospholipid hydrolysismay affect exocytosis differently, depending on theirorigin and location.

It has been demonstrated that taipoxin is rapidlytaken up by bovine chromaffin cells, causing releaseof catecholamines and fragmentation of F-actin(Neco et al., 2003). Although the internalization wasnot dependent on extracellular Ca2+, replacementof Ca2+ by Sr2+ during toxin incubation abolishedthe disassembly of F-actin.

The toxic effect of four elapid b-ntxs (notexin,b-Butx, taipoxin and textilotoxin) has been studiedon primary CNS cell cultures of hippocampal,cerebellar granule and cortical neurons, all beingglutamatergic (Rigoni et al., 2004). In all three typesof neurons, toxins induced a dose-dependentformation of distinct bulges on neurites, along withthe redistribution of two SV markers, synaptophy-sin I and VAMP2, to the bulges and the exposure ofthe luminal domain of synaptotagmin on the cellsurface. Additionally, partial fragmentation ofF-actin and neurofilaments was observed in neuronsbut not in the supporting cells, astrocytes. Applica-tion of taipoxin to hippocampal neurons showedthat the toxin disrupts the interaction betweensynaptophysin I and VAMP2, leading to massiveSV exocytosis, but not followed by appropriate SVmembrane retrieval (Bonanomi et al., 2005). Suchbulges were also observed when CNS cell cultureswere exposed to a mixture of lysoPL and FA,products of enzymatic activity of b-ntxs. As themixture of lysoPL and FA, similarly to variousb-ntxs, induced paralysis also on a mouse phrenicnerve–hemidiaphragm, this observation led to thehypothesis that the extracellular enzymatic action ofsPLA2 neurotoxins at the axolemma is sufficient fortheir neurotoxicity (Rigoni et al., 2005; Rossetto etal., 2006). Very recently, a significant influx ofextracellular Ca2+ was observed in cerebellargranule neurons and primary rat spinal motoneur-onal cells treated with different b-ntxs and with alysoPL and FA mixture (Rigoni et al., 2007).This result supports the belief that products ofsPLA2 activity (lysoPL and FA) released from theplasma membrane are the biochemical effectors ofb-neurotoxicity. In contrast, Rouault et al. (2006),using different mutants of OS2, a sPLA2 neurotoxinfrom the Australian taipan snake Oxyuranus S.

scutellatus very similar to notexin, demonstratedthat its catalytic activity is not of key importance forits central toxicity. Thus, OS2(H48Q), a 500-foldless enzymatically active mutant than wild-type (wt)OS2, showed only 16-fold lower central toxicity thanthe wt toxin.

Although cellular models provide an importantapproach to the understanding of different mole-cular mechanisms, the conclusions drawn fromusing such models in vitro should be taken carefullyand verified in vivo. In addition, different neuronaltypes may show different responses to toxins,especially those with sPLA2 activity. It has beendemonstrated that some sPLA2s that display noperipheral neurotoxicity are very toxic when in-jected intraventricularly directly into the (rat) brain,for example, neutral sPLA2 ammodytin I2 (AtnI2)from Vipera a. ammodytes (Gubensek et al., 1978)and acidic sPLA2 from Naja naja atra (Rapuanoet al., 1986). Neurotoxic effects were ascribed toenzymatic destruction of neuronal membranes.However, in the case of Naja n. atra, acidicsPLA2-specific binding sites were found in brainsynaptic membranes, suggesting a specific anddifferent molecular mechanism of sPLA2 neurotoxi-city in CNS from that in peripheral motoneurons.Furthermore, our recent results on a mousemotoneuronal cell line, the NSC34 cell line,intoxicated by AtxA are not in complete agreementwith the hypothesis presented above. It appearsinstead that the neurotoxin is not only cytotoxic tothese cells due to its extracellular (enzymatic)activity but also clearly internalized into the cells,as demonstrated by fluorescence and nanogoldlabeling, and rapidly translocated to the cytosoland certain intracellular organelles, including SVsand mitochondria (Jenko Praznikar et al., sub-mitted).

In recent years, the yeast Saccharomyces cerevi-

siae has also been utilized as a eukaryotic cellularmodel to study some phenomena in b-neurotoxicityand intracellular action of sPLA2s. Since AtxAinteracts with evolutionarily highly conserved pro-teins such as calmodulin and 14-3-3 proteins (seeSection 7), the molecular basis of certain effects ofAtxA may be similar in all eukaryotic cells once thesPLA2 enters the cell. The yeast cell has been used asa model system also, as it has a number ofadvantages over the more complex mammaliancells. Most importantly, it can easily be geneticallymanipulated and allows application of a wide rangeof whole-genome experimental approaches. Very

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892878

conveniently, in yeast there is no known sPLA2

homolog that might interfere with the AtxA effects.Studies showed that AtxA inhibited G2 cell-cyclearrest in yeast, which is regulated by 14-3-3 proteins(Petrovic et al., 2005), induced F-actin fragmenta-tion in tropomyosin-deficient yeast cells and de-creased the rate of endocytosis in the wild-type cells(Petrovic et al., submitted). The last two effects areindeed similar to the neurotoxic effects of AtxA inmammalian neurons.

5. Structural determinants of b-neurotoxicity

In spite of many attempts, structure–functionrelationships of snake venom sPLA2 toxins, includ-ing those with b-neurotoxic effects, have not beenresolved. As snake venom sPLA2s exhibit a widerange of pharmacological effects, a model wasproposed already in 1989 (Kini and Evans, 1989)to explain the tissue susceptibility to a particulartoxin. According to the model, specific ‘‘target sites’’on the cell surface are recognized by complementary‘‘pharmacological sites’’ on the toxin molecule. Theproposed target sites may be membrane lipids orproteins, whereas the pharmacological sites may bestructural determinants of neurotoxicity, myotoxi-city, anticoagulant activity and some others.

Early attempts to answer the question as to whysome sPLA2s are neurotoxic and others, structurallyvery similar, are not, were based mainly on a simplealignment of primary and, later, tertiary structuresof sPLA2s (Kini, 2003). The so-called ‘‘presynapticneurotoxic site’’ has been identified, or rather,proposed, by several groups to be on differentsurface regions of b-ntxs. Unfortunately, mostof these predictions have not been confirmedexperimentally. At least in the case of oligomericb-ntxs, it appears that the major role of non-sPLA2

or enzymatically inactive sPLA2 subunits is intargeting a b-ntx to specific binding sites (Bon,1997; Betzel et al., 1999). On the other hand, thesearch for the ‘‘presynaptic neurotoxic site’’ on themain sPLA2 subunit primarily responsible forneurotoxicity of oligomeric b-neurotoxins, as wellas on monomeric b-ntxs, has not been completelysuccessful.

Nevertheless, protein engineering of AtxA, one ofthe most thoroughly studied monomeric b-ntx, bysite-directed mutagenesis, has helped in identifyingat least some of the structural elements thatcontribute to presynaptic neurotoxicity. It has beenshown that the C-terminal region of Atxs is most

important, but not sufficient, for neurotoxicity(Prijatelj et al., 2002). In this region, the stretch ofamino acid residues 115–124 at the top of themolecule (Fig. 1) appears to be essential (Pungercaret al., 1999; Ivanovski et al., 2000). The hydro-phobic patch at the top is surrounded by mostlybasic residues and may represent a novel bindingsite for calmodulin (Prijatelj et al., 2003). The basiccharacter of b-ntxs is not essential for b-neurotoxi-city (Prijatelj et al., 2000), although certain basicresidues, even those in the b-structure region distantfrom the C-terminal region (Ivanovski et al., 2004),contribute to the neurotoxic effect of AtxA. Inaddition to the C-terminal region, Phe24 in theN-terminal region has been identified as veryimportant for the neurotoxicity of AtxA (Petanet al., 2002). Both Phe24 and the C-terminal regionthat contribute most to the high neurotoxicity ofAtxs overlap partially with the interfacial bindingsurface surrounding the entrance to the enzymeactive site of sPLA2s, responsible for interaction ofthe enzyme with the phospholipid bilayer duringmembrane degradation. Although most of themutations affecting the neurotoxicity of AtxA alsoinfluence enzymatic activity, no correlation has beenfound between the two effects, even when testing avariety of phospholipid substrates (Petan et al.,2005). Interestingly, a certain correlation was onlyobserved between neurotoxicity of AtxA and itsbinding affinity for calmodulin and R25 (Prijateljet al., 2003; Ivanovski et al., 2004). All the AtxAmutants with relatively high toxicity also show highbinding affinity for these two intracellular proteins,although the high affinity alone is not sufficient.This suggests that these interactions may beimportant at later stages of the neurotoxic process,probably after the presumed internalization of thetoxin.

According to the results on the extensive muta-genesis of AtxA, we suggest that there is unlikely tobe a distinct, ‘‘presynaptic neurotoxic site’’ on ab-ntx molecule, i.e. comparable to the enzyme activesite. It appears instead that different parts of thetoxin molecule are involved in different stages of thecomplex process of neurotoxicity that contribute tothe final outcome. In this view, structurally differentb-ntxs may have different surface regions that bindto different (extra- and intracellular) targets, butaffect the same process, recycling of SVs at theskeletal neuromuscular junction (see e.g. Rouaultet al., 2006). However, they all share sPLA2 activity,which is essential for the complete, irreversible

ARTICLE IN PRESS

Fig. 1. Structural model of ammodytoxin A (AtxA) showing the residues involved in b-neurotoxicity. The AtxA molecule is shown in two

orientations. Left, the front view, with the N-terminus (Ser1 in green) and interfacial binding surface surrounding the catalytic active site

(with His48 in red) facing the viewer. The C-terminal region folds over the top of the molecule and the b-structure region is on the right

lower corner. Right, the back view, where the molecule is rotated by about 1801 around the vertical axis (with the last, Cys133, residue

shown in green). The residues whose mutations led to more than 5-fold lower toxicity (i.e. 45-fold higher LD50) are shown in orange, and

those whose mutations resulted in less than 5-fold lower toxicity are shown in blue (Pungercar et al., 1999; Ivanovski et al., 2000, 2004;

Prijatelj et al., 2000, 2002, 2003; Petan et al., 2002, 2005).

J. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 879

blockade of neuromuscular transmission. For thesereasons, we see the term ‘‘presynaptic neurotoxicsite’’ as inappropriate.

6. Enzymatic activity of b-ntxs

Early results by Chang and Su (1982), on thechemical modification of His48 in the active sites ofthree b-ntxs, b-Butx, crotoxin and notexin, sug-gested that their sPLA2 activity is a necessarycondition for their presynaptic neurotoxicity. De-spite the large number of experiments since Changand Su’s pioneering work, no direct relationship hasbeen found between the degree of enzymatic activityand the b-neurotoxic potential of sPLA2s. Clearly,catalytic activity of b-ntxs alone is not sufficient for

their pathopharmacological action (Rosenberg,1997).

In order to study the role of catalytic activity ofAtxA in b-neurotoxicity in more detail, Prijateljet al. (2006) have recently produced three AtxAmutants, with a short N-terminal fusion peptide of12 or 5 amino acid residues, exhibiting only0.4–0.04% of wt activity. Interestingly, (neuro)toxi-cities of these mutants were just about an order ofmagnitude lower than that of AtxA, indicating thata minimal catalytic activity is sufficient for lethalityin mice and b-neurotoxic effects observed on aneuromuscular preparation. In line with theseresults is also a recent report on a neurotoxic groupIA sPLA2, OS2, where catalytic activity was foundas a minor factor determining the central neuro-toxicity of this molecule (Rouault et al., 2006).

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892880

A thorough study on the enzymatic properties ofAtxs (AtxA, B and C), their mutants and somehomologs was performed a short time ago (Petanet al., 2005). The apparent rates of phospholipidhydrolysis by the sPLA2s were determined on avariety of artificial vesicles containing anionic and/orzwitterionic phospholipids, as well as on biologicalmembranes. As shown before for b-Butx andcrotoxin (Radvanyi et al., 1987), Atxs display astrong preference for anionic vesicles. However,although they hydrolyze vesicles containing phos-phatidylglycerol or phosphatidylserine (PS) moreefficiently than those containing zwitterionic PC,they are quite efficient in hydrolyzing pure PCvesicles as well as PC-rich plasma membranes ofintact mammalian cells, such as HEK293. As themembrane-binding affinity of Atxs for PC-richsurfaces increases dramatically in the presence ofanionic phospholipids, this may have an importantinfluence on both localization of the b-ntxs to theirtarget membrane(s) and their enzymatic efficiency invivo. For example, it was recently confirmed thatthe intracellular layer of mammalian plasma mem-branes is rich in anionic PS (Okeley and Gelb,2004). Other groups of anionic phospholipids, suchas phosphatidic acid (PA) and phosphorylatedphosphatidylinositols (phospho-PIs), may also bevery important in providing a suitable hydrolyticenvironment within the cell, even for b-ntxs, forwhich only minimal enzymatic activity is needed forhigh neurotoxic effect (Buckland and Wilton, 2000;Kooijman et al., 2007). This altogether strengthensthe possibility of intracellular trafficking and actionof b-ntxs on different targets. In addition, whentightly bound to the membrane interface, the Ca2+

requirements of Atxs are in the micromolar range(Petan et al., 2005), opening up the possibility thatthese neurotoxins are enzymatically active even inthe subcellular compartments with low Ca2+ con-centration. Furthermore, AtxA has been found tobe enzymatically active in the cytosol of eukaryoticcells (Petrovic et al., 2005).

7. Molecular targets

b-Neurotoxicity is a process that depends essen-tially on phospholipase activity of a toxin, but manyyears ago it became evident that this is not aproperty that could explain the difference inpharmacological action of neurotoxic and non-neurotoxic sPLA2s. b-Neurotoxicity could not berelated to enzymatic specificity of sPLA2s or to their

enzymatic potency. Secreted PLA2s are enzymesthat bind to biological membranes and hydrolyzethem, as well as presynaptic membranes. Using Atx,two different specific binding sites on bovine andTorpedo marmorata synaptic membranes werediscovered, of high and low affinity (Krizaj et al.,1994, 1995, 1997). Non-neurotoxic sPLA2s wereable to bind only to the low-affinity binding sitewhich was not proteinaceous. The high-affinitysPLA2 receptor was a protein and specific forb-ntxs. A concept of two-step association of b-ntxswith the presynaptic membrane was introduced(Gubensek and Krizaj, 1997), very similar to thepresynaptic array receptor concept proposed laterfor Clostridium neurotoxins by Montecucco et al.(2004).

Identification of the so-called b-neurotoxicity-related receptor(s) on the presynaptic membrane ofmotoneurons is one of the main issues to be resolvedin order to clarify the molecular mechanism ofaction of presynaptically neurotoxic sPLA2s. How-ever, the study of molecular targets for b-ntxs isimportant to reveal not only the molecular basis ofb-neurotoxicity but also the molecular modes ofaction of their non-toxic sPLA2 homologs. Twelveout of the 17 known sPLA2 isoenzymes were foundin mammals (Kudo and Murakami, 2002; Rouaultet al., 2003), where they are involved in a variety ofphysiological and pathological activities, many ofwhich are the result of binding of sPLA2s to otherproteins (Kini, 1997; Hanasaki and Arita, 1999;Lambeau and Lazdunski, 1999). At least some ofthe proteins that are targeted by toxic sPLA2s alsobind endogenous sPLA2s, making the former veryuseful molecular tools in characterizing theseproteins and the processes in which they areinvolved.

7.1. Extracellular b-ntx-binding molecules

The neuronal or N-type sPLA2 receptor wasdiscovered using Oxyuranus s. scutellatus toxin 2(OS2) as a ligand (Lambeau et al., 1989). Binding ofOS2 on rat brain presynaptic membranes revealedtwo distinct populations of binding sites, both withvery high affinity (Kds in the pM range). Membraneproteins of 18–24, 36–51 and 85 kDa were detectedas binding sites for the toxin; however, despite manyattempts, the identity of these proteins has not beenreported. Some of the binding molecules that weredetected with other toxins (Atxs, taipoxin, crotoxin)in other nerve tissues (bovine brain, guinea pig

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 881

brain, Torpedo marmorata electric organ) probablyconstitute the N-type sPLA2 receptor as well. Usingcrotoxin, binding proteins of 85 and 65 kDa fromguinea pig and porcine brain were purified but notsequenced (Tzeng et al., 1996). From the compara-tive binding studies it is apparent that OS2 and Atxsdo not share all neuronal binding sites and that theso-called N-type sPLA2 receptor may consist ofdifferent receptor subpopulations (Krizaj et al.,1997; Lambeau et al., 1997). Based on the highcorrelation between affinity of the N-type sPLA2

receptor binding and lethality of sPLA2s, and thereceptor localization on the synaptic membranes,the N-type sPLA2 receptor remains the most likelycandidate for the b-neurotoxicity-specific receptoron nerve cells.

Until now, only b-Butx has been demonstratedto bind to voltage-dependent K+ channels (Scottet al., 1990), a subpopulation of a-dendrotoxin(from the green mamba Dendroaspis angusticeps)binding K+ channels (Black and Dolly, 1986).b-Butx binds to these proteins by its B subunitwhich is a structural homolog of a-dendrotoxin, anddendrotoxin considerably increased the survival ofhippocampal neurons in the presence of b-Butx(Herkert et al., 2001). It is no surprise that otherb-ntxs that lack this structural element do not bindto voltage-dependent K+ channels.

The relation between binding of b-Butx tovoltage-dependent K+ channel and its neurotoxicityis not clear. It may be that voltage-dependent K+

channels are b-neurotoxicity-related receptors forb-Butx in some species, e.g. chicken and frog, butnot in others, e.g. mouse and rat. Dendrotoxin andp-bromophenacylated-b-Butx prevented the presy-naptic effect of wt b-Butx in chick biventer muscleand frog muscle preparations but not in those ofmouse and rat diaphragm (reviewed in Chang,1985). The high sensitivity of birds to b-Butx couldbe due to this binding. In the case of mouse and ratpreparations, some other b-neurotoxicity-relatedreceptors for b-Butx probably exist on the pre-synaptic membrane. Fathi et al. (2001) showed thatthe cloned mouse, rat or human voltage-dependentK+ channels, mKv1.1, rKv1.2, mKv1.3, hKv1.5and mKv3.1K, were not blocked by b-Butx. Theyshowed additionally that taipoxin, notexin, crotoxinand AtxA similarly did not affect the activity ofcloned potassium channels. These data are inagreement with the previous observations, atleast in mouse and rat, that the assumption thatthe facilitatory phase is due to the inhibition of

voltage-dependent K+ channels may be unsafe. Theobservation that b-Butx has several receptors on theplasma membrane is in agreement with the idea ofparallel pathways by which b-ntxs can undertaketheir neurotoxic action on neuronal cells (Krizajand Gubensek, 2000). From two observations, (1) inspite of the pathophysiologically similar action ofb-Butx and other b-ntxs (see Section 3), none ofthese b-ntxs is able to compete with the neurospe-cific binding of b-Butx and (2) mutations in the Asubunit of b-Butx affect the b-neurotoxicity ofb-Butx, the following conclusion may be drawn.The location, but not the nature of the presynapticreceptor for a b-ntx, is important for expressingb-neurotoxicity. In support of this, mutual poten-tiation of b-neurotoxicity between some pairs ofb-ntxs has been reported, but not between others(Chang and Su, 1980).

The M-type sPLA2 receptor (sPLA2R), which isone of the most thoroughly characterized sPLA2-binding proteins, was discovered in muscle cells(hence the name muscle or M-type receptor) usingneurotoxic sPLA2 OS2 from Oxyuranus s. scutella-

tus (Lambeau et al., 1990). The M-type sPLA2Rhas been shown to undergo constitutive endo-cytosis (Zvaritch et al., 1996), and binding of ansPLA2, toxic or non-toxic, to this receptor isone way by which sPLA2 can enter the cell. UsingAtxC-affinity chromatography, a neuronal isoformof the M-type sPLA2R was identified later (Vardjanet al., 2001). It was demonstrated that in pigtwo different forms of the M-type sPLA2R exist.As only one gene for the M-type sPLA2Rs hasbeen found in the porcine genome, the structuraldifferences between the two isoforms are eitherthe result of different post-transcriptional and/orpost-translational processing. It was presumedthat the neuronal form of the receptor may possesssome unique physiological properties besidesthose that have already been assigned to the non-neuronal M-type sPLA2R (Hanasaki, 2004). In thiscontext, a possible role of the neuronal M-typesPLA2R in b-neurotoxicity was inspected anddemonstrated that interaction of AtxA with theM-type sPLA2R on the presynaptic membrane maynot be essential for b-neurotoxicity of the toxin(Prijatelj et al., 2006). However, it must beemphasized that, in that study, receptor-bindingaffinities of wt and three short N-terminal fusions ofAtxA were determined on the M-type sPLA2Risolated from porcine brain and not, for example,from mouse (target) tissue.

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892882

Neuronal pentraxin (NP1) is a soluble, extra-cellular taipoxin-binding protein (Schlimgen et al.,1995). It was shown to bind to neuronal pentraxinreceptor (NPR) on the surface of neurons, alone orin a complex with taipoxin (Dodds et al., 1997;Kirkpatrick et al., 2000). NPR is normally ex-pressed only in neurons; however, it is alsoexpressed on the surface of small cell lung cancer(SCLC) cells (Poulsen et al., 2005), which isconsistent with the neuroendocrine features of thesecells. Interestingly, a number of SCLC cell linesshowed marked sensitivity to taipoxin at toxinconcentrations that left the control cell linesunaffected. This may indicate that NPR is one ofthe extracellular neuronal receptors important forthe internalization and b-neurotoxicity of sPLA2s.Recently, a new class of cytosolic neuronal pen-traxins, putative binding proteins of taipoxin, hasbeen identified (NPCD) (Chen and Bixby, 2005).These cytosolic NPCD isoforms are composed of aneuronal pentraxin domain (formerly thought to beexclusively extracellular) linked to a chromo domain(formerly thought to be exclusively nuclear). NPCDisoforms are mainly associated with the inner side ofthe plasma membrane in brain neurons and ratPC12 cells in vitro, but also present in cell bodies,processes and growth cones. Their involvement intaipoxin neurotoxicity is possible.

Binding molecules for sPLA2s that are alsopresent on the extracellular side of neuronal cellsare proteoglycans (Van Vactor et al., 2006). Forexample, a GPI-anchored proteoglycan glypican I, alipid raft-residing molecule, was found to bind andinternalize a group IIA sPLA2 in HEK293 cells(Murakami et al., 1999). However, no convincingevidence about the involvement of these sPLA2-binding molecules in b-neurotoxicity is available atthe moment.

As already mentioned in the beginning of thissection, non-protein-specific binding sites for b-ntxshave been detected on different presynaptic mem-branes (Krizaj et al., 1994, 1995, 1997). It is possiblethat, as in the case of Clostridium neurotoxins,the efficiency of b-ntxs action is enhanced bybinding first to a large number of low-affinity,non-protein-binding sites, followed by transloca-tion to the low-density but high-affinity b-neuro-toxicity-related protein receptors. Candidates arebeing sought among different polysialogangliosideswhich constitute the initial attachment points in thecase of botulinum toxins (BoNTs; Verderio et al.,2006).

7.2. Intracellular b-ntx-binding molecules

One of the intracellular binding proteins for Atxshas been shown to be calmodulin (Sribar et al.,2001), a very important and highly conserved Ca2+-binding protein which participates in the signalingpathways that regulate many physiological pro-cesses (Chin and Means, 2000). Calmodulin sharesstructural similarity with the sPLA2-binding pro-teins, taipoxin-associated calcium-binding proteinof 49 kDa (TCBP-49) (Dodds et al., 1995) andcrocalbin (Hseu et al., 1999), all of them being EF-hand Ca2+-binding proteins (Lewit-Bentley andRety, 2000). Calmodulin is implicated in differentmodes of membrane trafficking and may also beinvolved as one of the Ca2+-sensors in regulatedexocytosis (Chen et al., 1999; Quetglas et al., 2000,2002; De Haro et al., 2004). Calmodulin would betherefore a perfect target for the b-ntx, which isknown to interfere with membrane trafficking byblocking synaptic transmission. Studies on AtxAmutants showed that all mutants expressing highb-neurotoxicity also possessed high affinity forcalmodulin (Prijatelj et al., 2003). The opposite isnot the case, as the mutant AtxA(F24N), which was130-fold less toxic than wt AtxA, possessed the sameaffinity for calmodulin as the wt toxin (Petan et al.,2002). In the same study, it was demonstrated that acluster of the C-terminal residues in AtxA, Y115,I116, R118 and N119 (the YIRN cluster) issignificantly involved in binding to calmodulin.The same amino acid residues are also among themost critical residues for the neurotoxicity of AtxA(Ivanovski et al., 2000). Besides Atxs, agkistrodo-toxin, bee venom sPLA2 and crotoxin were found tobind calmodulin (Sribar et al., 2001). Yeast calmo-dulin, Cmd1p, does not bind AtxA. As the yeast cellcan normally live with vertebrate calmodulin repla-cing endogenous Cmd1p (Davis and Thorner,1989), and since calmodulin is implicated in theregulation of many conserved fundamental physio-logical processes in eukaryotic cells, this organismseemed ideal for studying potential effects of theassociation of heterologously expressed AtxA andcalmodulin. In contrast to expectations, no obviouseffects on cell growth or phenotype were observed(U. Petrovic, personal communication).

Using AtxA, two other highly conserved proteins,14-3-3g and e isoforms, were detected and identifiedas sPLA2-binding proteins (Sribar et al., 2003a). Noexclusivity in the binding of b-neurotoxic sPLA2 to14-3-3 proteins was found. For example, human

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 883

group IIA sPLA2 also binds to 14-3-3 proteins invitro. In addition to other functions, 14-3-3 proteinsbind to membranes (similarly to BAR domains) andcould serve in this way as adaptors for localizationof sPLA2s at specific parts of the membrane (Martinet al., 1994; Roth et al., 1994). Most of theexperiments on 14-3-3 proteins were performed onthe two yeast 14-3-3 homologs, Bmh1p and Bmh2p,which exhibit 72% and 73% amino acid identity,respectively, with the human e isoform. AtxA bindsto Bmh1p and Bmh2p with similar affinity as toporcine 14-3-3 proteins (Petrovic et al., 2005). It isinteresting, but probably not directly related toneurotoxicity, that AtxA, heterologously expressedin yeast, inhibited cell-cycle arrest specifically in theG2 phase, which is known to be regulated by 14-3-3proteins. This suggests a possible mechanism bywhich sPLA2s are able to induce two oppositeeffects, proliferation and apoptosis, in mammaliancells (Petrovic et al., 2005).

It is interesting that different b-ntxs have differentbinding proteins whose location in the cell is thesame. This suggests that binding different b-ntxs todifferent targets in the same cellular compartmentresults in a similar effect, which is in accordancewith the observation that the action of some pairs ofb-ntxs is additive (Chang and Su, 1980) and, in spiteof that, that their final pathological effects arepractically identical. The links between the bindingof crotoxin to crocalbin (Hseu et al., 1999), taipoxinto TCBP-49 (Dodds et al., 1995) and AtxA toprotein disulfide isomerase (Sribar et al., 2005), allproteins of the lumen of the endoplasmic reticulum,with b-neurotoxicity have not yet been demon-strated. However, it is tempting to speculate thatthese interactions are essential for the retention andconcentration of b-ntxs inside the lumen of en-doplasmic reticulum and for their translocationacross the endoplasmic reticulum membrane intothe cytosol (Sribar et al., 2005).

As in the case of calmodulin, all neurotoxic AtxAbind with high affinity to a 25 kDa receptor, R25,located in neuronal mitochondria (Vucemilo et al.,1998; Prijatelj et al., 2000; Sribar et al., 2003b).Interestingly, only AtxA and its mutants have beenfound to be able to bind to R25 so far, although itmust be noted that this observation is based only on125I-AtxC binding competition and not on directbinding assays. The AtxA(F24N) mutant is again anexception, being 130-fold less toxic than wt AtxAbut still possessing high affinity for R25 (Petan etal., 2002). Two other exceptions, with high affinity

for R25 but no toxicity, are chimeras of 96 N-terminal amino acids from AtnI2 and 25 C-terminalamino acids from AtxA, i.e. AtnI2/AtxA(K108N)and AtnI2(F24N)/AtxA(K108N). These chimeras,however, do not bind calmodulin (Prijatelj et al.,2002). In spite of many attempts, the identity ofmitochondrial R25 has remained elusive.

8. Molecular models of b-ntxs action

One of the major problems in understanding themechanism of b-neurotoxicity at the molecular levelis the site(s) to which b-ntxs localize to producetoxic effects. One view is that b-ntxs act on theexternal side of the plasma membrane; another isthat they are endocytosed, acting also on intracel-lular membranes and binding to certain targetswithin the nerve terminal. Currently, there are twomolecular models of b-neurotoxicity, both assumingexclusively extracellular action of b-ntxs on moto-neurons. Based on our studies and some results ofother authors, we suggest, however, that both extra-and intracellular actions of b-ntxs are needed toproduce an effective and complete neuromuscularblockade.

8.1. Action of b-ntxs on nerve cells is exclusively

extracellular

Based mainly on studies on model primaryculture CNS cells (see Section 4), Montecucco andhis colleagues proposed that b-ntxs, followingspecific association with the presynaptic membrane,start to hydrolyze it (Rigoni et al., 2005). Theyassume binding of b-ntxs to SV release sites (i.e.active zones). The products of the phospholipaseactivity of b-ntxs, lysoPL and FA graduallyaccumulate on the extracellular leaflet of theplasma membrane. Due to their molecular geometry(lysoPL are inverted-cone shaped while FA arecone-shaped molecules) and rapid trans-bilayerequilibration of free FA but not lysoPL, whichremain on the outer side, they promote fusion of thehemifused SVs with the plasma membrane, henceincreasing the release of neurotransmitter. This ideawas originally proposed by Strong et al. (1976) andwas inspired by the observation that syntheticliposomes containing FA or cells in the presenceof lysophosphatides became very fusogenic (Pooleet al., 1970; Kantor and Prestegard, 1975). Theauthors suggested that, for the same topologicalreasons, the lipids generated by the enzymatic

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892884

activity of a b-ntx obstruct the opposite process, theendocytosis of SVs. As more and more lysoPL andFA are generated, several additional pathologicalchanges of nerve terminals occur. The axolemmabecomes more permeable to ions and FA partitionfrom the plasma membrane to intracellular orga-nelles. In this way, mitochondria uncouple, genera-tion of ATP is stopped and Ca2+ ions are releasedinto the cytosol, which all lead to further functionaland structural degeneration of the nerve terminal.

Supporting this hypothesis, lysophosphatidesalone or mixed with FA produce a very similarresponse to b-ntxs in a twitch-tension experiment onthe mouse phrenic nerve–hemidiaphragm prepara-tion (Caccin et al., 2006). Muscle was not affectedby lysoPL and FA, suggesting that the observedeffects were neurospecific (Rigoni et al., 2005;Caccin et al., 2006). The exact molecular basis ofaction remains, however, elusive and the interpreta-tion of the data needs to be cautious. For example,using methyl-b-cyclodextrin to extract cholesterolfrom the plasma membrane of motoneurons weobserved a synergistic effect on the b-neurotoxicityof taipoxin and AtxA at the neuromuscular junction(Vardjan, 2003) similar to that reported for textilo-toxin and a mixture of lysoPC and oleic acid byRigoni et al. (2005). It appears that any substancethat disturbs the integrity of the plasma membraneassists the action of b-ntxs (Burack and Biltonen,1994; Burack et al., 1997; Leidy et al., 2006).

This model presumes that, in addition to thethird, the second phase in the triphasic response ofthe neuromuscular preparation to a b-ntx alsodepends completely on the enzymatic activity ofthe toxin. However, the results with AtxA mutantsthat possessed very low enzymatic activity (down to2500-fold lower than the wt AtxA) showed that theystill express substantial neurotoxicity. The delay inthe onset of the facilitatory phase produced by thesemutants was very small compared with the delay intime to the final neuromuscular blockade, indicatingthat catalytic activity is only a minor factor in thestage of facilitated exocytosis but a critical one inthe last stage to neurotransmission arrest (Prijateljet al., 2006).

Rigoni et al. (2005) reported that pancreaticsPLA2 was able to induce neuromuscular blockadewhen added to the neuromuscular preparation ata concentration matching the PLA2 activity of15 nM textilotoxin, in line with their proposal ofsolely extracellular action of b-ntxs by enzymaticdegradation of the presynaptic membrane. However,

essentially non-toxic AtxA (Y115K/I116K/R118M/N119L) or AtnI2, which are more than 3-foldmore enzymatically active on PC (the main con-stituent of the outer leaflet of the plasma membrane)than the neurotoxic AtxA (Petan et al., 2005),did not produce any signs of neurotoxicity in amouse phrenic nerve–hemidiaphragm preparationat a concentration matching the lethal concentra-tion of AtxA (Vardjan, 2003). Similarly, Stronget al. (1976) did not observe any effect onneurotransmitter release in the rat phrenic nerve–diaphragm preparation if weakly neurotoxic Vipera

(Daboia) russelli sPLA2 was added to the prepara-tion at a concentration that exceeded, in itsphospholipase activity, the lethal concentration ofb-Butx by 3-fold. It is also difficult to explain by thishypothesis why AtxA does not induce cell death inC2C12 mouse myoblasts or HEK293 cells whoseplasma membranes are intensively degraded by thistoxin (Jenko Praznikar et al., submitted), especiallyas it induces, under comparable conditions, celldeath of the motoneuronal NSC34 cells. Byanalogy, one would expect that FA formed byAtxA at the plasma membrane of C2C12 orHEK293 cells quickly redistribute to the cytosolicside of the plasma membrane, dissociate into thecytosol, reach membranes of the intracellularorganelles and, for example, uncouple mitochondriaas proposed for neuronal cells. Even more difficultto explain just by the extracellular phospholipolyticparadigm is the fact that, in the rat phrenicnerve–hemidiaphragm blocked by b-Butx, some ofthe neighboring terminal boutons were destroyed,while the others remained intact (Prasarnpun et al.,2004).

Similarly to the first hypothesis, the second onepresumes that the action of b-ntxs is initiated at theSV release sites, i.e. active zones. The authorsdemonstrated that t-SNARE proteins, SNAP-25and syntaxin, are critically involved in the neuro-pathology of b-Butx (Prasarnpun et al., 2004). Thesecond hypothesis also differs from the first in theinterpretation of the disappearance of SVs fromnerve terminals. While the first hypothesis explainsthe disappearance of SVs by accelerated exocytosisnot adequately balanced by retrieval of the fusedSVs from the plasma membrane (Montecucco andRossetto, 2000; Rigoni et al., 2004), the second,consistent with the finding that terminal boutonsdo not swell, suggests that SVs are hydrolyzed byb-ntxs during exocytosis, when the inner face ofthe fused SV is exposed to the external space

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 885

(Prasarnpun et al., 2004). A combination ofenhanced exocytosis and SV hydrolysis wouldaccount for a progressive loss of SVs available forrelease, the failure of neuromuscular transmissionand an empty terminal bouton of unchangeddiameter. Besides morphometric analysis of normaland b-Butx-intoxicated neuromuscular preparationsobserved under the electron microscope, such ascenario in the intoxicated motor nerve terminalswas also envisaged from the substantial decrease ofthe fluorescence signal corresponding to the SVmarker synaptophysin (Prasarnpun et al., 2004).The fact that in primary CNS neurons the signal forsynaptophysin was not affected in its intensity byb-ntxs, but just relocated to swellings in the plasmamembrane (Rigoni et al., 2004, 2005; Bonanomiet al., 2005), indicates that the effects induced byb-ntxs at neuromuscular junctions and in CNSneurons may not be identical.

The two models presuming extracellular action ofb-ntxs also differ in the interpretation of facilitationof SV exocytosis by b-ntxs. Bonanomi et al. (2005)concluded that b-ntxs, following their binding to theactive zone, hydrolyze phospholipids in the externallayer of the presynaptic membrane. FA rapidlyequilibrate, by trans-bilayer movement, between thetwo layers of the presynaptic membrane. The resultis that lysoPL, which induce a positive curvature ofthe membrane, are present in trans, and FA, whichinduce a negative curvature, are present in cis, withrespect to the fusion site. Such a membraneconfiguration facilitates the transition of a hemi-fused to an SV fully fused with the presynapticmembrane. On the contrary, Prasarnpun et al.(2004) demonstrated that the b-Butx-evoked SVdepletion in the nerve terminal could be inhibited bymore than 60% by blocking voltage-dependentCa2+ channels with conotoxin o-MVIIC. Thismeans that entry of Ca2+ ions into the nerveterminal via voltage-dependent Ca2+ channels isalso important for effective depletion of SVs inphrenic nerve–hemidiaphragm preparations.

8.2. Action of b-ntxs on nerve cells is both extra- and

intracellular

Prasarnpun et al. (2004) used BoNT-C todetermine whether depletion of SVs from the motornerve ending induced by b-Butx depends onformation of a SNARE complex. They incubatedthe phrenic nerve–hemidiaphragm preparation withBoNT-C which cleaves specifically two t-SNAREs,

i.e. SNAP-25 and syntaxin. BoNT-C had no directeffect on SV density; moreover, it completelyblocked the depletion of SVs in preparationssubsequently exposed to b-Butx. This shows thatthe formation of a functional SNARE complex isnecessary for the disappearance of SVs from thenerve terminal caused by b-Butx. The loss of SVs onintoxication with b-Butx is therefore, at least in theinitial phase, a very specific process, and SNAREcomplex formation is essential for the normalexpression of b-neurotoxicity, at least in the caseof b-Butx. Since BoNT-C is a very specific proteasewhich is activated only when it arrives in the cytosolof the nerve cell (Montecucco et al., 1996), it is veryunlikely that the specific binding of b-Butx to itspresynaptic membrane receptor, and thus its accessto the plasma membrane, was affected by BoNT-Cpre-treatment of the preparation. Since voltage-dependent K+-channels, receptors for b-Butx, arelocated at the active zones on the plasma membraneand not in the lumen of SVs, b-Butx should still beable to bind specifically to the plasma membraneand to hydrolyze its outer leaflet, producinglysophosphatides and FA. However, as the SVcycling is arrested by BoNT-C, the tentativeinternalization of b-Butx into the lumen of thesevesicles and its intracellular action are prevented. Inthis way, it is also possible to interpret their resultsobtained with conotoxin: the inhibition of Ca2+

influx through voltage-gated Ca2+ channels mini-mizes the basal SV cycling, which consequentlyslows down the entry of b-Butx into the nerveterminal and its neurotoxic effect. Some b-neuro-toxic effects, such as distorted mitochondria andformation of large vacuoles, were observed anywayin the BoNT-C pre-treated nerve endings, but theywere very limited in appearance compared with theeffects of b-Butx on nerve endings having competentt-SNAREs (Prasarnpun et al., 2004; J.B. Harris,personal communication). These effects may indeedresult from the hydrolytic action of b-Butx on theextracellular leaflet of the plasma membrane, sub-sequent flip-flopping of the excess FA onto thecytosolic side of the plasma membrane (Kamp et al.,1995) and their translocation to other intracellularmembranes. The possibility remains that thefavored substrate for b-Butx enzymatic activity isthe luminal leaflet of the SV membrane, which is notavailable in this case.

At least for b-Butx, internalization into the nerveending can be envisaged to proceed via SVs that areretrieved from the plasma membrane after exocytosis,

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892886

but it is too early to generalize that all b-ntxsare using the same pathway, as not all of themshare the same binding sites on the plasmamembrane. For example, structurally very similarClostridium neurotoxins bind to different receptorson the presynaptic membrane of the nerve terminal,the result of which is diverse internalizationmechanisms—BoNTs enter the nerve cell viarecycling SVs but TeNTs enter via lipid rafts (Lalliet al., 2003).

It is clear that the mechanisms of internalizationof b-ntxs and BoNTs are different and that theV-type ATP-ase is not involved in the internaliza-tion/action of b-ntxs (Simpson et al., 1994) asproposed (Montecucco and Rossetto, 2000). b-Butxmay be released into the cytosol by hydrolyticdegradation of SVs, gaining in this way access to theplasma membrane from its cytosolic side and to themembranes of subcellular organelles such as mito-chondria and SVs. Degradation of SVs inside thenerve terminal would explain why the nerveterminal area of intoxicated motoneuron is notenlarged and can be experimentally verified bymeasuring the cytosolic concentration of AChduring the intoxication by b-ntx. Arguments againstthe possibility of enzymatic action of sPLA2s in thecytosolic environment, such as too low [Ca2+]i andinstability of multi-disulfide-bridged sPLA2s in thereductive milieu (Montecucco and Rossetto, 2000),have been convincingly rejected (Singer et al., 2002;Petrovic et al., 2004, 2005). The increase in [Ca2+]ithrough e.g. voltage-dependent Ca2+ channels andfrom internal Ca2+-stores, such as mitochondria,would lead to further increases in the phospholipaseactivity of internalized b-ntx, thus augmenting itstoxic effects. In accordance with this, the reductionof Ca2+ influx into the nerve ending by conotoxinreduced the effect of b-Butx (Prasarnpun et al.,2004), and complexing of intracytosolic Ca2+ by themembrane-permeable Ca2+-chelator BAPTA-AMhampered the facilitation of spontaneous synapticcurrents induced by b-Butx in Xenopus motoneur-on–myocyte co-cultures (Liou et al., 2006). Theresults of Liou et al. also suggested that b-Butx isable to elicit Ca2+ from intracellular stores.

The observation that b-Butx and AtxA inhibitedthe neuromuscular preparation that was not ex-ternally stimulated (Prasarnpun et al., 2004) is veryinteresting. The pathomorphology of the prepara-tion intoxicated by b-Butx in the absence of anexternal stimulation was identical to that of thepreparation electro-stimulated during the exposure

to toxin (J.B. Harris, personal communication).Muscle did not contract during these experiments,meaning that the toxins did not induce a thresholdrelease of the neurotransmitter from motor nerveterminals. It is difficult to explain such experimentalfacts by a massive secretion of SVs not adequatelybalanced by their retrieval from the plasma mem-brane, nor by their degradation on the plasmamembrane following the exocytosis, but easier bythe degradation of SVs by internalized b-ntxs insidethe nerve ending. The observation supporting thisview is that neurotoxic Atxs are much moreenzymatically active than non-toxic AtnI2 on acidicphospholipids which are abundantly present in theinner leaflet of the plasma membrane and thecytosolic side of subcellular organelles (Petan et al.,2005). In addition, we should not forget cytosolic b-ntx-binding proteins, which may specifically mod-ulate and/or localize the phospholipase activity ofinternalized toxins.

An experimental finding that favors the inhibitionof SV endocytosis by b-neurotoxic sPLA2s from thecytosolic side of the plasma membrane was obtainedon perforated PC12 cells (Schmidt et al., 1999). Theinhibition of phospholipase A2 activity specificallyon the cytosolic side of the plasma membraneeffectively increased the formation of synaptic-likemicrovesicles in this system, while its amplificationresulted in decreased vesicle formation, i.e. reducedendocytosis. The disappearance of SVs from thenerve ending is surely assisted also by their impairedendocytotic retrieval from the plasma membranebut, as the motor nerve terminal area does notsignificantly expand (Prasarnpun et al., 2004), thiscannot be the main reason.

Strong support for the intracellular action ofb-ntxs comes from the discovery of high-affinitybinding proteins in the cytosol (Sribar et al., 2001,2003a) and mitochondria of nerve cells (Vucemiloet al., 1998; Sribar et al., 2003b), although a firmconnection between the binding to these proteinsand b-neurotoxicity has not yet been established.The endogenous group IIA sPLA2 has recently beenfound to localize to mitochondria (Macchioni et al.,2004), which strengthens the physiological relevanceof cytosolic and mitochondrial sPLA2-bindingproteins.

Using yeast experimental systems (Tong et al.,2001) it was demonstrated that AtxA interactsgenetically with some proteins that are involved inendocytosis, such as amphiphysin and AP-180,which is an indication of specific localization of

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 887

the toxin at the neck of the closing vesicle (Petrovicet al., submitted). Further, Bmh2, the yeast 14-3-3homolog, is specifically located at the neck of theclosing vesicle in yeast. Binding of AtxA to Bmh2 atthis particular, highly curved region of the plasmamembrane can result in localized higher-than-else-where phospholipase activity, which effectivelyblocks the process of SV formation (Schmidt et al.,1999). In agreement with this, the expression of themature form AtxA in the yeast cytosol led to theinhibition of endocytosis, indicated by measuringthe uptake of Lucifer yellow dye (Petrovic et al.,submitted).

Finally, the results on internalization of AtxAinto the intact motoneuronal NSC34 cells furthersupport the internalization paradigm. In less than5min, the photo-reactive derivative of the toxin,sulfo-SBED-AtxC (Kovacic et al., 2007) labeled itsbinding proteins, calmodulin and 14-3-3 proteins, inthe cytosol of these cells (see Section 4).

Fig. 2. Possible mode of action of b-neurotoxic sPLA2s (b-ntxs) at thepresynaptic membrane of the motor neuron (green) where the exo/endo

the plasma membrane, they start to hydrolyze its outer leaflet, producing

alters the fusogenic properties of the plasma membrane, which participa

by means of recycling synaptic vesicles and, by hydrolyzing them,

internalization pathways, such as receptor-mediated endocytosis are,

proteins (blue, yellow) implicated in exo/endocytosis (5), and hydrolyze t

subcellular organelles (6). The result is uncoupling of mitochondria, wit

leading to inhibition of protein phosphorylation, degradation of syn

cytoskeleton. The Ca2+ concentration in the cytosol is further increased

activity of b-ntxs and leading to complete degeneration of the nerve ter

Based on the above arguments, we believe thatb-neurotoxicity is fully expressed only if sPLA2s acton both sides of the neuronal plasma membrane andon the intracellular membranes, and that hydrolysisof the plasma membrane from just the external sideis not sufficient (Fig. 2). The hypothesis that wepropose accounts for all pathophysiological andmorphological observations derived from intoxi-cated neuromuscular junctions, also explainingthose experimental results that could not beadequately described by the two models presentedin Section 8.1.

9. Perspectives

The exact molecular mechanism of action ofpresynaptically neurotoxic sPLA2s is apparently stillfar from being completely understood. Some im-portant questions that are not convincingly answeredare the following. Do b-ntxs act extracellularly,

neuromuscular junction. b-ntxs bind to specific receptors on the

cytosis of synaptic vesicles takes place (1). Tightly associated with

lysophospholipids and fatty acids (2). Changed lipid composition

tes in facilitation of exocytosis. b-ntxs may enter the nerve ending

chew their way into the cytosol (3). Other possible/alternative

however, not excluded (4). In the cytosol, b-ntxs interact with

he inner side of the plasma membrane as well as the membranes of

h the loss of Ca2+ homeostasis and the arrest of ATP production

aptic vesicles, arrest of endocytosis and fragmentation of the

as the plasma membrane becomes leaky, promoting the hydrolytic

minal. The mitochondrial receptor for Atxs, R25, is coloured red.

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892888

intracellularly or from both sides on the motoneur-on? Do all b-ntxs act on motoneurons in the sameway, i.e. is the generalization justified? Is the modeof action of a particular b-ntx on the motoneuronidentical to its action on some other susceptibleneuronal cells?

To answer these questions:

(1)

A detailed picture of particular steps in themolecular mechanism of action of b-ntxs usingdifferent neuronal systems should be obtained.Only in this way, it will be possible todiscriminate between the primary and secondaryeffects of b-ntxs at the neuromuscular junction.We believe that the use of weak b-ntxs or weaklyneurotoxic b-ntx mutants is advantageous overthe use of very potent b-ntxs in obtaining betterresolution.

(2)

The receptor(s) in presynaptic membranesof different susceptible neurons responsible forb-neurotoxicity should be identified.

(3)

The measurement of enzymatic activity of b-ntxson intracellular membranes should be per-formed (in vitro and in vivo). The influence ofthe cytosolic b-ntx-binding proteins on thephospholipase activity of b-ntxs, regarding therate, specificity and cellular localization, shouldbe assessed.

Acknowledgments

This work was supported by several grants fromthe Slovenian Ministry of Higher Education,Science and Technology. We sincerely thank ourcolleagues from the group, Dr. John B. Harris,Dr. Edward G. Rowan and Dr. Roger H. Pain fora critical reading of the manuscript. We thankDr. Jernej Sribar and Uros Logonder for technicalassistance.

References

Abe, T., Limbrick, A.R., Miledi, R., 1976. Acute muscle

denervation induced by b-bungarotoxin. Proc. R. Soc.

London Ser. B 149, 545–553.

Aird, S.D., Kaiser, I.I., Lewis, R.V., Kruggel, W.G., 1986. A

complete amino acid sequence for the basic subunit of

crotoxin. Arch. Biochem. Biophys. 249, 296–300.

Betzel, C., Genov, N., Rajashankar, K.R., Singh, T.P., 1999.

Modulation of phospholipase A2 activity generated by

molecular evolution. Cell. Mol. Life Sci. 56, 384–397.

Black, A.R., Dolly, J.O., 1986. Two acceptor sub-types for

dendrotoxin in chick synaptic membranes distinguishable by

b-bungarotoxin. Eur. J. Biochem. 156, 609–617.

Bon, C., 1997. Multicomponent neurotoxic phospholipases A2.

In: Kini, R.M. (Ed.), Venom Phospholipase A2 Enzymes:

Structure, Function and Mechanism. Wiley, Chichester,

pp. 269–285.

Bonanomi, D., Pennuto, M., Rigoni, M., Rossetto, O., Mon-

tecucco, C., Valtorta, F., 2005. Taipoxin induces SV

exocytosis and disrupts the interaction of synaptophysin I

with VAMP2. Mol. Pharmacol. 67, 1901–1908.

Buckland, A.G., Wilton, D.C., 2000. Anionic phospholipids,

interfacial binding and the regulation of cell functions.

Biochim. Biophys. Acta 1483, 199–216.

Burack, W.R., Biltonen, R.L., 1994. Lipid bilayer heterogeneities

and modulation of phospholipase A2 activity. Chem. Phys.

Lipids 73, 209–222.

Burack, W.R., Dibble, A.R., Allietta, M.M., Biltonen, R.L.,

1997. Changes in vesicle morphology induced by lateral phase

separation modulate phospholipase A2 activity. Biochemistry

36, 10551–10557.

Caccin, P., Rigoni, M., Bisceglie, A., Rossetto, O., Montecucco,

C., 2006. Reversible skeletal neuromuscular paralysis

induced by different lysophospholipids. FEBS Lett. 580,

6317–6321.

Caratsch, C.G., Maranda, B., Miledi, R., Strong, P.N., 1981.

A further study of the phospholipase-independent action of

b-bungarotoxin at frog end-plates. J. Physiol. 319, 179–191.

Chang, C.C., 1985. Neurotoxins with phospholipase A2 activity

in snake venoms. Proc. Natl. Sci. Counc. Repub. China B 9,

126–142.

Chang, C.C., Su, M.J., 1980. Mutual potentiation, at nerve

terminals, between toxins from snake venoms which contain

phospholipase A activity: b-bungarotoxin, crotoxin, taipoxin.Toxicon 18, 641–648.

Chang, C.C., Su, M.J., 1982. Presynaptic toxicity of the histidine-

modified, phospholipase A2-inactive, b-bungarotoxin, crotox-in and notexin. Toxicon 20, 895–905.

Chen, B., Bixby, J.L., 2005. Neuronal pentraxin with chromo

domain (NPCD) is a novel class of protein expressed in

multiple neuronal domains. J. Comp. Neurol. 481, 391–402.

Chen, I.L., Lee, C.Y., 1970. Ultrastructural changes in the motor

nerve terminals caused by b-bungarotoxin. Virchows Arch. B

Cell Pathol. 6, 318–325.

Chen, Y.A., Duvvuri, V., Schulman, H., Scheller, R.H., 1999.

Calmodulin and protein kinase C increase Ca2+-stimulated

secretion by modulating membrane-attached exocytic ma-

chinery. J. Biol. Chem. 274, 24476–26469.

Chen, Y.J., 2005. Phospholipase A2 activity of b-bungarotoxin is

essential for induction of cytotoxicity on cerebellar granule

neurons. J. Neurobiol. 64, 213–223.

Chin, D., Means, A.R., 2000. Calmodulin: a prototypical calcium

sensor. Trends Cell Biol. 10, 322–328.

Coulter, A., Harris, R., Broad, A., Sutherland, S., Sparrow, L.,

Misconi, L., Hamilton, R., Rubira, M., 1983. The isolation

and some properties of the major neurotoxic component from

the venom of the common or Eastern Australian brown snake

(Pseudonaja textilis). Toxicon 21 (Suppl. 3), 81–84.

Cull-Candy, S.G., Fohlman, J., Gustavsson, D., Lullmann-

Rauch, R., Thesleff, S., 1976. The effects of taipoxin and

notexin on the function and fine structure of the murine

neuromuscular junction. Neuroscience 1, 175–180.

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 889

Davis, T.N., Thorner, J., 1989. Vertebrate and yeast calmodulin,

despite significant sequence divergence, are functionally

interchangeable. Proc. Natl. Acad. Sci. USA 86, 7909–7913.

De Haro, L., Ferracci, G., Opi, S., Iborra, C., Quetglas, S.,

Miquelis, R., Leveque, C., Seagar, M., 2004. Ca2+/calmodu-

lin transfers the membrane-proximal lipid-binding domain of

the v-SNARE synaptobrevin from cis to trans bilayers. Proc.

Natl. Acad. Sci. USA 101, 1578–1583.

Delot, E., Bon, C., 1993. Model for the interaction of crotoxin, a

phospholipase A2 neurotoxin, with presynaptic membranes.

Biochemistry 32, 10708–10713.

Dixon, R.W., Harris, J.B., 1999. Nerve terminal damage by

b-bungarotoxin. Its clinical significance. Am. J. Pathol. 154,

447–455.

Dodds, D., Schlimgen, A.K., Lu, S.-Y., Perin, M.S., 1995. Novel

reticular calcium binding protein is purified on taipoxin

columns. J. Neurochem. 64, 2339–2344.

Dodds, D.C., Omeis, I.A., Cushman, S.J., Helms, J.A., Perin,

M.S., 1997. Neuronal pentraxin receptor, a novel putative

integral membrane pentraxin that interacts with neuronal

pentraxin 1 and 2 and taipoxin-associated calcium-binding

protein 49. J. Biol. Chem. 272, 21488–21494.

Dreyer, F., Penner, R., 1987. The actions of presynaptic snake

toxins on membrane currents of mouse motor nerve

terminals. J. Physiol. 386, 455–463.

Fathi, H.B., Rowan, E.G., Harvey, A.L., 2001. The facilitatory

actions of snake venom phospholipase A2 neurotoxins at the

neuromuscular junction are not mediated through voltage-

gated K+ channels. Toxicon 39, 1871–1882.

Faure, G., Bon, C., 1988. Crotoxin, a phospholipase A2

neurotoxin from the South American rattlesnake Crotalus

durissus terrificus: purification of several isoforms and

comparison of their molecular structure and of their

biological activities. Biochemistry 27, 730–738.

Faure, G., Guillaume, J.L., Camoin, L., Saliou, B., Bon, C.,

1991. Multiplicity of acidic subunit isoforms of crotoxin, the

phospholipase A2 neurotoxin from Crotalus durissus terrificus

venom, results from posttranslational modifications. Bio-

chemistry 30, 8074–8083.

Faure, G., Choumet, V., Bouchier, C., Camoin, L., Guillaume,

J.L., Monegier, B., Vuilhorgne, M., Bon, C., 1994. The origin

of the diversity of crotoxin isoforms in the venom of Crotalus

durissus terrificus. Eur. J. Biochem. 223, 161–164.

Fohlman, J., Lind, P., Eaker, D., 1977. Taipoxin, an extremely

potent presynaptic snake venom neurotoxin. Elucidation of

the primary structure of the acidic carbohydrate-containing

taipoxin-subunit, a prophospholipase homolog. FEBS Lett.

84, 367–371.

Fohlman, J., Eaker, D., Dowdall, M.J., Lullmann-Rauch, R.,

Sjodin, T., Leander, S., 1979. Chemical modification of

taipoxin and the consequences for phospholipase activity,

pathophysiology, and inhibition of high-affinity choline

uptake. Eur. J. Biochem. 94, 531–540.

Gopalakrishnakone, P., Hawgood, B.J., 1984. Morphological

changes induced by crotoxin in murine nerve and neuromus-

cular junction. Toxicon 22, 791–804.

Gubensek, F., Krizaj, I., 1997. Ammodytoxins (Vipera

ammodytes ammodytes). In: Rappuoli, R., Montecucco, C.

(Eds.), Guidebook to Protein Toxins and their Use

in Cell Biology. Oxford University Press, New York,

pp. 224–226.

Gubensek, F., Ritonja, A., Sek, S., Zupan, J., 1978. Phospho-

lipases A2 from Vipera ammodytes venom. In: Rosenberg, P.

(Ed.), Toxins: Animal, Plant and Microbial. Pergamon Press,

Oxford, pp. 135–139.

Hanasaki, K., 2004. Mammalian phospholipase A2: phospholi-

pase A2 receptor. Biol. Pharm. Bull. 27, 1165–1167.

Hanasaki, K., Arita, H., 1999. Biological and pathological

functions of phospholipase A2 receptor. Arch. Biochem.

Biophys. 372, 215–223.

Harris, J.B., 1985. Phospholipases in snake venoms and their

effects on nerve and muscle. Toxicon 31, 79–102.

Herkert, M., Shakhman, O., Schweins, E., Becker, C.M., 2001.

b-bungarotoxin is a potent inducer of apoptosis in cultured

rat neurons by receptor-mediated internalization. Eur. J.

Neurosci. 14, 821–828.

Ho, I.C., Arm, J.P., Bingham III, C.O., Choi, A., Austen, K.F.,

Glimcher, L.H., 2001. A novel group of phospholipase A2s

preferentially expressed in type 2 helper T cells. J. Biol. Chem.

276, 18321–18326.

Howard, B.D., 1975. Effects of b-bungarotoxin on mitochondrial

respiration are caused by associated phospholipase A activity.

Biochem. Biophys. Res. Commun. 67, 58–65.

Hseu, M.J., Yen, C.H., Tzeng, M.C., 1999. Crocalbin: a new

calcium-binding protein that is also a binding protein for

crotoxin, a neurotoxic phospholipase A2. FEBS Lett. 445,

440–444.

Ivanovski, G., Copic, A., Krizaj, I., Gubensek, F., Pungercar, J.,

2000. The amino acid region 115–119 of ammodytoxins plays

an important role in neurotoxicity. Biochem. Biophys. Res.

Commun. 276, 1229–1234.

Ivanovski, G., Petan, T., Krizaj, I., Gelb, M.H., Gubensek, F.,

Pungercar, J., 2004. Basic amino acid residues in the

b-structure region contribute, but not critically, to presynaptic

neurotoxicity of ammodytoxin A. Biochim. Biophys. Acta

1702, 217–225.

Kamp, F., Zakim, D., Zhang, F., Noy, N., Hamilton, J.A., 1995.

Fatty acid flip-flop in phospholipid bilayers is extremely fast.

Biochemistry 34, 11928–11937.

Kantor, H.L., Prestegard, J.H., 1975. Fusion of fatty acid

containing lecithin vesicles. Biochemistry 14, 1790–1795.

Kattah, L.R., Ferraz, V., Matos Santoro, M., Ribeiro da Silva

Camargos, E., Ribeiro Diniz, C., De Lima, M.E., 2002.

Analysis of fatty acids released by crotoxin in rat brain

synaptosomes. Toxicon 40, 43–49.

Kini, R.M., 1997. Venom Phospholipase A2 Enzymes: Structure,

Function and Mechanism. Wiley, Chichester.

Kini, R.M., 2003. Excitement ahead: structure, function and

mechanism of snake venom phospholipase A2 enzymes.

Toxicon 42, 827–840.

Kini, R.M., Evans, H.J., 1989. A model to explain the

pharmacological effects of snake venom phospholipases A2.

Toxicon 27, 613–635.

Kirkpatrick, L.L., Matzuk, M.M., Dodds, D.C., Perin, M.S.,

2000. Biochemical interactions of the neuronal pentraxins.

Neuronal pentraxin (NP) receptor binds to taipoxin and

taipoxin-associated calcium-binding protein 49 via NP1 and

NP2. J. Biol. Chem. 275, 17786–17792.

Kondo, K., Toda, H., Narita, K., Lee, C.Y., 1982a. Amino acid

sequence of b2-bungarotoxin from Bungarus multicinctus

venom. The amino acid substitutions in the B chains.

J. Biochem. 91, 1519–1530.

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892890

Kondo, K., Toda, H., Narita, K., Lee, C.Y., 1982b. Amino

acid sequences of three b-bungarotoxins (b3-, b4-, and

b5- bungarotoxins) from Bungarus multicinctus venom.

Amino acid substitutions in the A chains. J. Biochem. 91,

1531–1548.

Kooijman, E.E., Tieleman, D.P., Testerink, C., Munnik, T.,

Rijkers, D.T.S., Burger, K.N.J., de Kruijff, B., 2007. An

electrostatic/hydrogen bond switch as the basis for the specific

interaction of phosphatidic acid with proteins. J. Biol. Chem.

282, 11356–11364.

Kovacic, L., Sribar, J., Krizaj, I., 2007. A new photoprobe for

studying biological activities of secreted phospholipases A2.

Bioorg. Chem. 35, 295–305.

Krizaj, I., Gubensek, F., 2000. Neuronal receptors for phospho-

lipases A2 and b-neurotoxicity. Biochimie 82, 807–814.

Krizaj, I., Dolly, J.O., Gubensek, F., 1994. Identification of the

neuronal acceptor in bovine cortex for ammodytoxin C, a

presynaptically neurotoxic phospholipase A2. Biochemistry

33, 13938–13945.

Krizaj, I., Rowan, E.G., Gubensek, F., 1995. Ammodytoxin A

acceptor in bovine brain synaptic membranes. Toxicon 33,

437–449.

Krizaj, I., Faure, G., Gubensek, F., Bon, C., 1997. Neurotoxic

phospholipases A2 ammodytoxin and crotoxin bind to

distinct high-affinity protein acceptors in Torpedo marmorata

electric organ. Biochemistry 36, 2779–2787.

Kudo, I., Murakami, M., 2002. Phospholipase A2 enzymes.

Prostaglandins Other Lipid Mediat 68–69, 3–58.

Lalli, G., Bohnert, S., Deinhardt, K., Verastegui, C., Schiavo, G.,

2003. The journey of tetanus and botulinum neurotoxins in

neurons. Trends Microbiol. 11, 431–437.

Lambeau, G., Lazdunski, M., 1999. Receptors for a growing

family of secreted phospholipases A2. Trends Pharmacol. Sci.

20, 162–170.

Lambeau, G., Barhanin, J., Schweitz, H., Qar, J., Lazdunski, M.,

1989. Identification and properties of very high affinity brain

membrane-binding sites for a neurotoxic phospholipase from

the taipan venom. J. Biol. Chem. 264, 11503–11510.

Lambeau, G., Schmid-Alliana, A., Lazdunski, M., Barhanin, J.,

1990. Identification and purification of a very high affinity

binding protein for toxic phospholipases A2 in skeletal

muscle. J. Biol. Chem. 265, 9526–9532.

Lambeau, G., Ancian, P., Barhanin, J., Lazdunski, M., 1994.

Cloning and expression of a membrane receptor for secretory

phospholipases A2. J. Biol. Chem. 269, 1575–1578.

Lambeau, G., Cupillard, L., Lazdunski, M., 1997. Membrane

receptors for venom phospholipases A2. In: Kini, R.M. (Ed.),

Venom Phospholipase A2 Enzymes: Structure, Function and

Mechanism. Wiley, Chichester, pp. 389–412.

Lee, C.Y., Tsai, M.C., Chen, Y.M., Ritonja, A., Gubensek, F.,

1984. Mode of neuromuscular blocking action of phospho-

lipases A2 from Vipera ammodytes ammodytes venom. Arch.

Int. Pharmacodyn. 268, 313–324.

Leidy, C., Linderoth, L., Andresen, T.L., Mouritsen, O.G.,

JØrgensen, K., Peters, G.H., 2006. Domain-induced activa-

tion of human phospholipase A2 type IIA: local versus global

lipid composition. Biophys. J. 90, 3165–3175.

Lewit-Bentley, A., Rety, S., 2000. EF-hand calcium-binding

proteins. Curr. Opin. Struct. Biol. 10, 637–643.

Lind, P., 1982. Amino-acid sequence of the b1 isosubunit of

taipoxin, an extremely potent presynaptic neurotoxin from

the Australian snake taipan (Oxyuranus s. scutellatus). Eur. J.

Biochem. 128, 71–75.

Lind, P., Eaker, D., 1982. Amino-acid sequence of the a-subunitof taipoxin, an extremely potent presynaptic neurotoxin from

the Australian snake taipan (Oxyuranus s. scutellatus). Eur. J.

Biochem. 124, 441–447.

Liou, J.C., Kang, K.H., Chang, L.S., Ho, S.Y., 2006. Mechanism

of b-bungarotoxin in facilitating spontaneous transmitter

release at neuromuscular synapse. Neuropharmacology 51,

671–680.

Macchioni, L., Corazzi, L., Nardicchi, V., Mannucci, R., Arcuri,

C., Porcellati, S., Sposini, T., Donato, R., Goracci, G., 2004.

Rat brain cortex mitochondria release group II secretory

phospholipase A2 under reduced membrane potential. J. Biol.

Chem. 279, 37860–37869.

Martin, H., Rostas, J., Patel, Y., Aitken, A., 1994. Subcellular

localisation of 14-3-3 isoforms in rat brain using specific

antibodies. J. Neurochem. 63, 2259–2265.

Montecucco, C., Rossetto, O., 2000. How do presynaptic PLA2

neurotoxins block nerve terminals? Trends Biochem. Sci. 25,

266–270.

Montecucco, C., Schiavo, G., Tugnoli, V., de Grandis, D., 1996.

Botulinum neurotoxins: mechanism of action and therapeutic

applications. Mol. Med. Today 2, 418–424.

Montecucco, C., Rossetto, O., Schiavo, G., 2004. Presynaptic

receptor arrays for clostridial neurotoxins. Trends Microbiol.

12, 442–446.

Murakami, M., Kambe, T., Shimbara, S., Yamamoto, S.,

Kuwata, H., Kudo, I., 1999. Functional association of type

IIA secretory phospholipase A2 with the glycosylphosphati-

dylinositol-anchored heparan sulfate proteoglycan in the

cyclooxygenase-2-mediated delayed prostanoid-biosynthetic

pathway. J. Biol. Chem. 274, 29927–29936.

Neco, P., Rossetto, O., Gil, A., Montecucco, C., Gutierrez, L.M.,

2003. Taipoxin induces F-actin fragmentation and enhances

release of catecholamines in bovine chromaffin cells. J.

Neurochem. 85, 329–337.

Nicholls, D., Snelling, R., Dolly, J.O., 1985. Bioenergetic actions

of b-bungarotoxin, dendrotoxin and bee-venom phospholi-

pase A2 on guinea-pig synaptosomes. Biochem. J. 229,

653–662.

Okeley, N.M., Gelb, M.H., 2004. A designed probe for acidic

phospholipids reveals the unique enriched anionic character

of the cytosolic face of the mammalian plasma membrane. J.

Biol. Chem. 279, 21833–21840.

Pearson, J.A., Tyler, M.I., Retson, K.V., Howden, M.E., 1993.

Studies on the subunit structure of textilotoxin, a potent

presynaptic neurotoxin from the venom of the Australian

common brown snake (Pseudonaja textilis). 3. The complete

amino-acid sequences of all the subunits. Biochim. Biophys.

Acta 1161, 223–229.

Poulsen, T.T., Pedersen, N., Perin, M.S., Hansen, C.K., Poulsen,

H.S., 2005. Specific sensitivity of small cell lung cancer cell

lines to the snake venom toxin taipoxin. Lung Cancer 50,

329–337.

Petan, T., Krizaj, I., Gubensek, F., Pungercar, J., 2002.

Phenylalanine-24 in the N-terminal region of ammodytoxins

is important for both enzymic activity and presynaptic

neurotoxicity. Biochem. J. 363, 353–358.

Petan, T., Krizaj, I., Gelb, M.H., Pungercar, J., 2005. Ammo-

dytoxins, potent presynaptic neurotoxins, are also highly

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892 891

efficient phospholipase A2 enzymes. Biochemistry 44,

12535–12545.

Petrovic, U., Sribar, J., Paris, A., Rupnik, M., Krzan, M.,

Vardjan, N., Gubensek, F., Zorec, R., Krizaj, I., 2004.

Ammodytoxin, a neurotoxic secreted phospholipase A2, can

act in the cytosol of the nerve cell. Biochem. Biophys. Res.

Commun. 324, 981–985.

Petrovic, U., Sribar, J., Matis, M., Anderluh, G., Peter-Katalinic,

J., Krizaj, I., Gubensek, F., 2005. Ammodytoxin, a secretory

phospholipase A2, inhibits G2 cell cycle arrest in yeast

Saccharomyces cerevisiae. Biochem. J. 391, 383–388.

Poole, A.R., Howell, J.I., Lucy, J.A., 1970. Lysolecithin and cell

fusion. Nature 227, 810–813.

Prasarnpun, S., Walsh, J., Harris, J.B., 2004. b-bungarotoxin-induced depletion of SVs at the mammalian neuromuscular

junction. Neuropharmacology 47, 304–314.

Prijatelj, P., Copic, A., Krizaj, I., Gubensek, F., Pungercar, J.,

2000. Charge reversal of ammodytoxin A, a phospholipase

A2-toxin, does not abolish its neurotoxicity. Biochem. J. 352,

251–255.

Prijatelj, P., Krizaj, I., Kralj, B., Gubensek, F., Pungercar, J.,

2002. The C-terminal region of ammodytoxin is important

but not sufficient for neurotoxicity. Eur. J. Biochem. 269,

5759–5764.

Prijatelj, P., Sribar, J., Ivanovski, G., Krizaj, I., Gubensek, F.,

Pungercar, J., 2003. Identification of a novel binding site for

calmodulin in ammodytoxin A, a neurotoxic group IIA

phospholipase A2. Eur. J. Biochem. 270, 3018–3025.

Prijatelj, P., Vardjan, N., Rowan, E.G., Krizaj, I., Pungercar, J.,

2006. Binding to the high-affinity M-type receptor for secreted

phospholipases A2 is not obligatory for the presynaptic

neurotoxicity of ammodytoxin A. Biochimie 88, 1425–1433.

Pungercar, J., Krizaj, I., Liang, N.-S., Gubensek, F., 1999. An

aromatic, but not a basic, residue is involved in the toxicity of

group-II phospholipase A2 neurotoxins. Biochem. J. 341,

139–145.

Quetglas, S., Leveque, ., Miquelis, R., Sato, K., Seagar, M., 2000.

Ca2+-dependent regulation of synaptic SNARE complex

assembly via a calmodulin- and phospholipid-binding domain

of synaptobrevin. Proc. Natl. Acad. Sci. USA 97, 9695–9700.

Quetglas, S., Iborra, C., Sasakawa, N., De Haro, L., Kumakura,

K., Sato, K., Leveque, C., Seagar, M., 2002. Calmodulin and

lipid binding to synaptobrevin regulates calcium-dependent

exocytosis. EMBO J. 21, 3970–3979.

Radvanyi, F., Saliou, B., Bon, C., Strong, P.N., 1987.

The interaction between the presynaptic phospholipase

neurotoxins b-bungarotoxin and crotoxin and mixed deter-

gent-phosphatidylcholine micelles. A comparison with non-

neurotoxic snake venom phospholipases A2. J. Biol. Chem.

262, 8966–8974.

Rapuano, B.E., Yang, C.C., Rosenberg, P., 1986. The relation-

ship between high-affinity noncatalytic binding of snake

venom phospholipases A2 to brain synaptic plasma mem-

branes and their central lethal potencies. Biochim. Biophys.

Acta 856, 457–470.

Rigoni, M., Schiavo, G., Weston, A.E., Caccin, P., Allegrini, F.,

Pennuto, M., Valtorta, F., Montecucco, C., Rossetto, O.,

2004. Snake presynaptic neurotoxins with phospholipase A2

activity induce punctate swellings of neurites and exocytosis

of SVs. J. Cell Sci. 117, 3561–3570.

Rigoni, M., Caccin, P., Gschmeissner, S., Koster, G., Postle,

A.D., Rossetto, O., Schiavo, G., Montecucco, C., 2005.

Equivalent effects of snake PLA2 neurotoxins and lysopho-

spholipid–fatty acid mixtures. Science 310, 1678–1680.

Rigoni, M., Pizzo, P., Schiavo, G., Weston, A.E., Zatti, G.,

Caccin, P., Rossetto, O., Pozzan, T., Montecucco, C., 2007.

Calcium influx and mitochondrial alterations at synapses

exposed to snake neurotoxins or their phospholipid hydrolysis

products. J. Biol. Chem. 282, 11238–11245.

Ringstad, N., Gad, H., Low, P., Di Paolo, G., Brodin, L.,

Shupliakov, O., De Camilli, P., 1999. Endophilin/SH3p4 is

required for the transition from early to late stages in clathrin-

mediated synaptic vesicle endocytosis. Neuron 24, 143–154.

Ritonja, A., Gubensek, F., 1985. Ammodytoxin A, a highly lethal

phospholipase A2 from Vipera ammodytes ammodytes venom.

Biochim. Biophys. Acta 828, 306–312.

Rosenberg, P., 1997. Pitfalls to avoid in the study of correlation

between enzymatic activity and pharmacological properties of

phospholipase A2 enzymes. In: Kini, R.M. (Ed.), Venom

Phospholipase A2 Enzymes: Structure, Function and Me-

chanism. Wiley, Chichester, pp. 155–183.

Rossetto, O., Morbiato, L., Caccin, P., Rigoni, M., Montecucco,

C., 2006. Presynaptic enzymatic neurotoxins. J. Neurochem.

97, 1534–1545.

Roth, D., Morgan, A., Martin, H., Jones, D., Martens, G.J.M.,

Aitken, A., Burgoyne, R.D., 1994. Characterisation of 14-3-3

proteins in adrenal chromaffin cells and demonstration of

isoform-specific phospholipid binding. Biochem. J. 301,

305–310.

Rouault, M., Bollinger, J.G., Lazdunski, M., Gelb, M.H.,

Lambeau, G., 2003. Novel mammalian group XII secreted

phospholipase A2 lacking enzymatic activity. Biochemistry

42, 11494–11503.

Rouault, M., Rash, L.D., Escoubas, P., Boilard, E., Bollinger, J.,

Lomonte, B., Maurin, T., Guillaume, C., Canaan, S.,

Deregnaucourt, C., Schrevel, J., Doglio, A., Gutierrez, J.M.,

Lazdunski, M., Gelb, M.G., Lambeau, G., 2006. Neurotoxi-

city and other pharmacological activities of the snake venom

phospholipase A2 OS2: the N-terminal region is more

important than enzymatic activity. Biochemistry 45,

5800–5816.

Rowan, E.G., Harvey, A.L., 1988. Potassium channel blocking

actions of b-bungarotoxin and related toxins on mouse and

frog motor nerve terminals. Br. J. Pharmacol. 94, 839–847.

Schlimgen, A.K., Helms, J.A., Vogel, H., Perin, M.S., 1995.

Neuronal pentraxin, a secreted protein with homology to

acute phase proteins of the immune system. Neuron 14,

519–526.

Schmidt, A., Wolde, M., Thiele, C., Fest, W., Kratzin, H.,

Podtelejnikovk, A.V., Witkek, W., Huttner, W.B., Soling, H.-

D., 1999. Endophilin I mediates synaptic vesicle formation by

transfer of arachidonate to lysophosphatidic acid. Nature

401, 133–141.

Scott, V.E.S., Parcej, D.N., Keen, J.N., Findlay, J.B.C., Dolly,

J.O., 1990. a-Dendrotoxin acceptor from bovine brain is a K+

channel protein. J. Biol. Chem. 265, 20094–20097.

Shakhman, O., Herkert, M., Rose, C., Humeny, A., Becker,

C.M., 2003. Induction by b-bungarotoxin of apoptosis in

cultured hippocampal neurons is mediated by Ca2+-depen-

dent formation of reactive oxygen species. J. Neurochem. 87,

598–608.

Shipolini, R.A., Callewaert, G.L., Cottrell, R.C., Vernon, C.A.,

1974. The amino-acid sequence and carbohydrate content of

ARTICLE IN PRESSJ. Pungercar, I. Krizaj / Toxicon 50 (2007) 871–892892

phospholipase A2 from bee venom. Eur. J. Biochem. 48,

465–476.

Simpson, L.L., Lautenslager, G.T., Kaiser, I.I., Middlebrook,

J.L., 1993. Identification of the site at which phospholipase A2

neurotoxins localize to produce their neuromuscular blocking

effects. Toxicon 31, 13–26.

Simpson, L.L., Coffield, J.A., Bakry, N., 1994. Inhibition of

vacuolar adenosine triphosphatase antagonizes the effects of

clostridial neurotoxins but not phospholipase A2 neurotoxins.

J. Pharmacol. Exp. Ther. 269, 256–262.

Singer, A.G., Ghomashchi, F., Le Calvez, C., Bollinger, J.,

Bezzine, S., Rouault, M., Sadilek, M., Lazdunski, M.,

Lambeau, G., Gelb, M.H., 2002. Interfacial kinetic and

binding properties of the complete set of human and mouse

groups I, II, V, X, and XII secreted phospholipases A2.

J. Biol. Chem. 277, 48535–48549.

Strong, P.N., Goerke, J., Oberg, S.G., Kelly, R.B., 1976.

b-Bungarotoxin, a pre-synaptic toxin with enzymatic activity.

Proc. Natl. Acad. Sci. USA 73, 178–182.

Su, M.J., Chang, C.C., 1984. Presynaptic effects of snake venom

toxins which have phospholipase A2 activity (b-bungarotoxin,taipoxin, crotoxin). Toxicon 22, 631–640.

Sribar, J., Copic, A., Paris, A., Sherman, N.E., Gubensek, F.,

Fox, J.W., Krizaj, I., 2001. A high-affinity acceptor for

phospholipase A2 with neurotoxic activity is a calmodulin.

J. Biol. Chem. 276, 12493–12496.

Sribar, J., Sherman, N.E., Prijatelj, P., Faure, G., Gubensek, F.,

Fox, J.W., Aitken, A., Pungercar, J., Krizaj, I., 2003a. The

neurotoxic phospholipase A2 associates, through a non-

phosphorylated binding motif, with 14-3-3 protein g and eisoforms. Biochem. Biophys. Res. Commun. 302, 691–696.

Sribar, J., Copic, A., Poljsak-Prijatelj, M., Kuret, J., Logonder,

U., Gubensek, F., Krizaj, I., 2003b. R25 is an intracellular

membrane receptor for a snake venom secretory phospholi-

pase A2. FEBS Lett. 553, 309–314.

Sribar, J., Anderluh, G., Fox, J.W., Krizaj, I., 2005. Protein

disulphide isomerase binds ammodytoxin strongly: possible

implications for toxin trafficking. Biochem. Biophys. Res.

Commun. 329, 741–745.

Thouin, L.G., Ritonja, A., Gubensek, F., Russell, F.E., 1982.

Neuromuscular and lethal effects of phospholipase A from

Vipera ammodytes venom. Toxicon 20, 1051–1058.

Tong, A.H.Y., Evangelista, M., Parsons, A.B., Xu, H., Bader,

G.D., Page, N., Robinson, M., Raghibizadeh, S., Hogue,

C.W.V., Bussey, H., Andrews, B., Tyers, M., Boone, C., 2001.

Systematic genetic analysis with ordered arrays of yeast

deletion mutants. Science 294, 2364–2368.

Tseng, W.P., Lin-Shiau, S.Y., 2003. Activation of NMDA

receptor partly involved in b-bungarotoxin-induced neuro-

toxicity in cultured primary neurons. Neurochem. Int. 42,

333–344.

Tzeng, M.C., Yen, C.H., Tsai, M.D., 1996. Binding proteins

on synaptic membranes for certain phospholipases

A2 with presynaptic toxicity. Adv. Exp. Med. Biol. 391,

271–278.

Van Vactor, D., Wall, D.P., Johnson, K.G., 2006. Heparan

sulfate proteoglycans and the emergence of neuronal con-

nectivity. Curr. Opin. Neurobiol. 16, 40–51.

Vardjan, N., 2003. Ph.D. Thesis. University of Ljubljana,

Slovenia.

Vardjan, N., Sherman, N.E., Pungercar, J., Fox, J.W., Gubensek,

F., Krizaj, I., 2001. High molecular mass receptors for

ammodytoxin in pig are tissue-specific isoforms of M-type

phospholipase A2 receptor. Biochem. Biophys. Res. Com-

mun. 289, 143–149.

Verderio, C., Rossetto, O., Grumelli, C., Frassoni, C., Mon-

tecucco, C., Matteoli, M., 2006. Entering neurons: botulinum

toxins and SV recycling. EMBO Rep. 7, 995–999.

Vucemilo, N., Copic, A., Gubensek, F., Krizaj, I., 1998.

Identification of a new high affinity binding protein

for neurotoxic phospholipases A2. Biochem. Biophys. Res.

Commun. 251, 209–212.

Wei, S., Ong, W.Y., Thwin, M.M., Fong, C.W., Farooqui, A.A.,

Gopalakrishnakone, P., Hong, W., 2003. Group IIA secretory

phospholipase A2 stimulates exocytosis and neurotransmitter

release in pheochromocytoma-12 cells and cultured rat

hippocampal neurons. Neuroscience 121, 891–898.

Zvaritch, E., Lambeau, G., Lazdunski, M., 1996. Endocytic

properties of the M-type 180-kDa receptor for secretory

phospholipases A2. J. Biol. Chem. 271, 250–257.