Mapping structural determinants of biological activities in snake venom phospholipases A 2 by sequence analysis and site directed mutagenesis

Full text

Mapping structural determinants of biological activities
in snake venom phospholipases A
2
by sequence analysis
and site directed mutagenesis
Lucimara Chioato
a
, Richard J. Ward
b,
*
a
Department of Biochemistry and Immunology, FMRP-USP, Universidade de Sa
˜o Paulo, Brazil
b
Department of Chemistry, FFCLRP-USP, Universidade de Sa
˜o Paulo, Avenida Bandeirantes 3900, Monte Alegre,
CEP 14040-901, Ribeira
˜o Preto-SP, Brazil
Abstract
In addition to their catalytic activity, snake venom phospholipases A
2
(vPLA
2
) present remarkable diversity in their
biological effects. Sequence alignment analyses of functionally related PLA
2
are frequently used to predict the structural
determinants of these effects, and the predictions are subsequently evaluated by site directed mutagenesis experiments and
functional assays. In order to improve the predictive potential of computer-based analysis, a simple method for scanning amino
acid variation analysis (SAVANA) has been developed and included in the analysis of the lysine 49 PLA
2
myotoxins (Lys49-
PLA
2
). The SAVANA analysis identified positions in the C-terminal loop region of the protein, which were not identified using
previously available sequence analysis tools. Site directed mutagenesis experiments of bothropstoxin-I, a Lys49-PLA
2
isolated
from the venom of Bothrops jararacussu, reveals that these residues are exactly those involved in the determination of
myotoxic and membrane damaging activities. The SAVANA method has been used to analyse presynaptic neurotoxic and anti-
coagulant vPLA
2
s, and the predicted structural determinants of these activities are in excellent agreement with the available
results of site directed mutagenesis experiments. The positions of residues involved in the myotoxic and neurotoxic
determinants demonstrate significant overlap, suggesting that the multiple biological effects observed in many snake vPLA
2
s
are a consequence of superposed structural determinants on the protein surface.
q2004 Elsevier Ltd. All rights reserved.
Keywords: Lys49-PLA2; Myotoxin; Neurotoxin; Anti-coagulant; Bioinformatics
1. Introduction
The secreted group I/II phospholipases A
2
(PLA
2
,EC
3.1.1.4) are small (,14 kDa), stable enzymes that are
encountered in a wide variety of biological fluids and cells.
PLA
2
s catalyse the hydrolysis of the sn-2 ester bond of sn-3
glycerophospholipids to release lysophospholipids and fatty
acids (van Deenen and de Haas, 1963). The rate of
hydrolysis is considerably enhanced against the phospholi-
pid substrate in an aggregated form such as a micelle or a
bilayer, and the enzymatic mechanism by which this process
of interfacial activation occurs has become a paradigm for
interfacial catalysis (Berg et al., 2001). The products of
phospholipid hydrolysis may themselves be bioactive, or
may serve as precursors for the synthesis of other bioactive
compounds (Dennis, 1994; Dessen, 2000). Due to their
central role in many cellular processes, PLA
2
s from a
variety of sources have been extensively studied, not only to
understand the molecular bases of the catalytic mechanism
and interfacial binding (Scott and Sigler, 1994), but also
with a view to understanding their regulatory functions
within the cell (Murakami et al., 1997; Murakami and Kudo,
2001). Snake venoms are abundant sources of group I/II
PLA
2
, and although these venom derived PLA
2
s (vPLA
2
s)
show a high level degree of structural conservation with
mammalian secreted PLA
2
s, they present remarkable
diversity in terms of biological activities, and a given
vPLA
2
may demonstrate multiple biological effects.
0041-0101/$ - see front matter q2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.toxicon.2003.11.027
Toxicon 42 (2003) 869–883
www.elsevier.com/locate/toxicon
*Corresponding author. Tel.: þ55-16-602-3484; fax: þ55-16-
633-8151.
E-mail address: rjward@fmrp.usp.br (R.J. Ward).

It has been established that catalytic activity plays a key
role in certain pharmacological effects of vPLA
2
s(Chang
and Su, 1982; Dı
´az-Oreiro and Gutie
´rrez, 1997), however, it
is now well proven that many pharmacological effects are at
least partially independent of hydrolytic activity (Kini and
Evans, 1995; Rufini et al., 1996; Pa
´ramo et al., 1998; Soares
et al., 2001). The absence of a clear correlation between
catalysis and pharmacological activity together with the
diversity of biological effects raises the question as to the
structural bases of these biological functions. A growing
body of evidence suggests that these activities may be
mediated by interactions between vPLA
2
s and acceptors for
endogenous PLA
2
s on the membranes of the target cell
(Lambeau and Lazdunski, 1999; Hanasaki and Arita, 1999;
Valentin and Lambeau, 2000). The identification of
additional endogenous mammalian PLA
2
s(Six and Dennis,
2000), and the discovery of their protein acceptors in human
cells (Sribar et al., 2001; Higashino et al., 2002; Boilard
et al., 2003), has expanded the number of potential targets
and possible mechanisms of action for vPLA
2
s.
Acceptor mediated responses may explain many of the
observed effects of vPLA
2
s, however, other mechanisms of
PLA
2
/membrane interactions may also contribute to toxicity
of these proteins. For example, the role of quaternary
structure is highlighted by the high level of toxicity of the
crotoxin heterodimer in comparison with each individual
sub-unit (Habermann and Breithaupt, 1978; Faure et al.,
1993). Furthermore, it is noteworthy that two different types
of dimeric interaction (Brunie et al., 1985; Arni et al., 1995),
and a trimeric association (Fremont et al., 1993) have been
observed for vPLA
2
s in the crystalline state (Arni and Ward,
1996; Ward et al., 1998a). Although the effect of quaternary
structure on catalytic and biological activities remains
unclear, a mechanism involving quaternary structural
change in the Lys49-PLA
2
homodimer has been proposed
to explain the increased permeability of artificial mem-
branes after treatment with the protein, and this mechanism
is independent of both catalytic activity and acceptor
binding (da Silva Giotto et al., 1998).
In addition to their pharmacological functions, vPLA
2
s
also present a wide range of substrate preferences, which
must be determined by the specific surface topology of the
region of the protein that interacts with the membrane
surface. In group I/II PLA
2
s, this region is defined by a
highly conserved hydrophobic surface cleft that binds the
fatty acyl chain of the phospholipid substrate, together with
a surrounding ring of more variable polar and charged
residues, and which is called either the interfacial recog-
nition site (IRS; Pieterson et al., 1974) or the i-face
(Ramirez and Jain, 1991). Variation in the topology of the
i-face determines the specificity of phospholipid head-group
binding (Snitko et al., 1999; Yu et al., 2000), and may give
rise to additional and unexpected catalytic effects. For
example, the neutral PLA
2
from Naja n. atra venom exhibits
the phenomenon of ‘PC activation’ (Adamich and Dennis,
1978), in which the hydrolysis of phospholipids containing
ethanolamine head groups is enhanced in membranes that
include phosphatidycholine (PC). This effect is determined
by the binding of the PC head-group to a cluster of surface
residues close to the active site region in the i-face
(Lefkowitz et al., 1999).
These examples demonstrate the remarkable diversity
displayed by vPLA
2
s in their surface recognition properties.
This raises the question as to how the surface of small
proteins such as the group I/II PLA
2
s can determine such
functional complexity. In the group I/II PLA
2
s, between 35
and 40% of the residues are exposed to solvent at the surface
of the protein, therefore around 40 – 50 residues participate
in the definition of the physico-chemical properties of the
protein surface. Which of these residues participate in the
formation of the specific surface topologies that determine a
given biological function? In the face of such complexity,
what strategies can be employed to study these structural
determinants? Here we review amino acid sequence
comparison methods that have been applied to map the
structural determinants of vPLA
2
s. A sequence analysis of
the lysine 49 phospholipase A
2
(Lys49-PLA
2
) myotoxins is
presented which uses a simple method to evaluate amino
acid substitutions, and the results are discussed in the light
of recent site-directed mutagenesis studies. We have also
applied the method to study the structural determinants of
pre-synaptic neurotoxic and anticoagulant activities
observed in other vPLA
2
s. The method is robust and may
be applicable to the investigation of the surface properties of
other families of homologous proteins.
2. Lysine 49 phospholipase A
2
(Lys49-PLA
2
) myotoxins
Lys49-PLA
2
s have been identified as abundant com-
ponents of the venoms from New World Bothrops and
Agkistrodon snake species (Gutie
´rrez and Lomonte, 1995),
in the Asiatic Trimeresurus species (Liu et al., 1990) and
have been discovered more recently in additional New
World viperid species, although in lesser quantities (Tsai
et al., 2001). Not only is the distribution of the Lys49-PLA
2
s
more widespread than previously thought, but also the range
of known biological effects is broader. In addition to
myotoxic (Gutie
´rrez and Lomonte, 1995), cytotoxic
(Fletcher and Jiang, 1998) and edema inducing (Liu et al.,
1991; Landucci et al., 2000) effects, the Lys49-PLA
2
s show
pre-synaptic neurotoxicity (Dhillon et al., 1987), stimulate
the degranulation of mast cells (Landucci et al., 1998) and
directly influence leukocyte mobility (de Castro et al.,
2000).
The Lys49-PLA
2
s are characterized as a sub-family of
group IIA PLA
2
s (using the nomenclature of Six and
Dennis, 2000) in which the aspartic acid residue at position
49 (Asp49) in the active site region is substituted by lysine
(Lys49). In the Asp49-PLA
2
s, the carboxyl group oxygens
of Asp49 contribute to the coordination of the catalytically
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883870

essential Ca
2þ
co-factor, and the substitution by lysine at this
position results in steric hindrace of Ca
2þ
binding by the
1
-amino group of the Lys49. It has been suggested that
the loss of Ca
2þ
binding in the Lys49-PLA
2
s results in the
observed lack of hydrolytic activity against both synthetic
and natural phospholipids (Maraganore and Heinrikson,
1986; van den Bergh et al., 1988). Despite their lack of
catalytic activity, the Lys49-PLA
2
s demonstrate membrane
damaging activity via a Ca
2þ
-independent, non-hydrolytic
mechanism (Dı
´az et al., 1991; Rufini et al., 1992; de
Oliveira et al., 2001).
As new members of the Lys49-PLA
2
s sub-family are
discovered, the number of amino acid sequences of
deposited in the protein databases has steadily increased.
Due to the range of biological effects and the unusual
mechanism of membrane damage induced by these toxins,
the Lys49-PLA
2
family represents an interesting example to
study by protein surface mapping in order to identify the
structural determinants of biological activities. This strategy
involves, using a series of bioinformatic tools to analyze
amino acid sequences and to predict the residues that are
structural determinants for a given biological effect. These
residues are subsequently targeted by site directed mutagen-
esis experiments.
3. Amino acid sequence alignment analysis
3.1. Initial sequence alignment and visual comparison
Studies comparing a limited number of similar amino acid
sequences can be made using manual sequence alignment.
Apart from being time consuming, this simple method has the
principal disadvantage that the positions of amino acid
insertions and deletions are not reliably determined. This is a
major drawback, since a reliable sequence alignment is
critical for the correct interpretation of the results. With
increased access via internet to sequence alignment tools
such as CLUSTAL (Higgins et al., 1992; Thompson et al.,
1994), the automated alignment of greater numbers of amino
acid sequences has become widespread. It should be noted
that although automated sequence alignment programs speed
up the process of obtaining an initial ‘draft’ alignment, the
algorithms employed are not infallible, and where possible
alignments made by computer should always be checked
against superposed three-dimensional protein structures in
order to verify the positions of insertions and deletions. The
quality of a sequence alignment is crucial for subsequent
analyses, and the importance of a reliable sequence
alignment should be emphasized.
The most straightforward method to analyze a sequence
alignment is by visual evaluation, and several programs are
available on the internet that aid this type of analysis by
coloring each amino acid residue symbol according to its
physicochemical properties (Beitz, 2000). Key differences
in the active site and substrate binding cleft regions of the
Lys49-PLA
2
from Agkistrodon p. piscivorus were identified
using visual analysis of a manual sequence alignment
(Maraganore and Heinrikson, 1986). In a subsequent study
using similar techniques, sequence comparison of the
Lys49-PLA
2
myotoxin-II from Bothrops asper with a
limited set of other Lys49-PLA
2
confirmed these differences
between the Lys49 and Asp49-PLA
2
s, and in addition drew
attention to the cationic/aromatic residues which were
grouped in the C-terminal loop region of the proteins
(Francis et al., 1991). Based on a visual analysis of a
computer generated Lys49-PLA
2
sequence alignment, an
extended surface including polar and cationic residues
primarily on helix 3 was suggested to be involved in the
determination of myotoxic activity (Selistre de Araujo et al.,
1996). This analysis may be compared with a previous
prediction that residues 78– 87 in the b-wing region of the
PLA
2
s are determinants of the myotoxic effect (Kini and
Iwanaga, 1986a). Taken together, these studies highlight the
ambiguities that arise from attempts to identify the structural
determinants of biological activity by visual analysis, and
suggest the need for more sophisticated analyses to extract
useful information from the increasingly large datasets used
in automated sequence alignments.
3.2. Residue mapping using the ‘SequenceSpace’ analysis
Automated amino acid sequence alignments commonly
use progressive pairwise sequence comparison algorithms,
which results in the clustering of the most similar sequences
within the alignment (Feng and Doolittle, 1987). This
sequence pairing is based on the optimization of a numeric
value, or score, and although a comparison of all residues
contributes to the score value, the exact information as to
which residues are conserved within a particular sequence
pair is lost. This information is unfortunately the most
interesting for defining conserved sequence motifs in
protein families with common functions. To recover this
information, computational methods have been developed
which identify the specific residues that define sub-groups of
amino acid sequences within multiple sequence alignments
(Casari et al., 1995; Lichtarge et al., 1996; Andrade et al.,
1997). In these methods, all sequences in an alignment are
mapped as vectors in a multi-dimensional ‘SequenceSpace’.
Vectors pointing to conserved amino acid residues and
patterns in this sequence space become clustered, and these
clusters will therefore be comprised of specific amino acid
residues that ‘define’ sub-groups of proteins.
When applied to a data set of 72 PLA
2
sequences this
analysis yielded a list of 12 residues that were highly
specific to the Lys49-PLA
2
s sub-family (Ward, 1998b).
Surface mapping of these residues revealed that 10 out of
these 12 residues mapped to three amino acid clusters
located in the active site region, the hydrophobic fatty acyl
chain binding pocket and the tip of the b-wing. The cluster
of residues in the active site region includes Lys49, Leu32
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883 871

and Asn28 and has been the focus of a site directed
mutagenesis study using bothropstoxin-I (BthTx-I), a
Lys49-PLA
2
from the venom of Bothops jararacussu
(Ward et al., 2002). Recombinant BthTx-I shows no
detectable hydrolytic activity against natural phospholipid
substrates, and mutation of Lys49Asp did not restore
catalytic activity. This implies that the lack of catalytic
activity observed in the Lys49-PLA
2
is not simply a
consequence of the presence of the Lys49, and mutagenesis
of other residues in the active site region is currently
underway in order to investigate this result. The mutagen-
esis study of the active site region also included residues that
are highly conserved in the active site of the Asp49-PLA
2
s.
None of the active site mutants, including substitution of the
nucleophile His48 by glutamine, influenced the myotoxic
activity which provides strong evidence against the
involvement of phospholipid hydrolysis in the myotoxic
effect (Ward et al., 2002). The second residue cluster in the
hydrophobic and fatty acyl binding site of Lys49-PLA
2
sis
comprised of Leu5/Val102/Leu106, which substitute the
Phe5/Ala102/Phe106 triad found in the Asp49-PLA
2
s.
Although the total volume of the amino acid triad remains
virtually unchanged, the topology of the binding site region
is altered, and the consequence of these alterations is
currently unknown.
The substitutions in the active site and hydrophobic fatty
acyl chain binding pocket regions of the Lys49-PLA
2
s had
already been noted by visual comparison (Maraganore and
Heinrikson, 1986; Francis et al., 1991), however, the b-wing
cluster was not detected. Analysis of the crystal structures of
several Lys49-PLA
2
s demonstrated that the Glu12, Trp77
and Lys80 residues in the b-wing cluster participate in
intermolecular contacts resulting in the formation of a
homodimer (Arni et al., 1995; Arni and Ward, 1996). The
Glu12 and Lys80 from each monomer in the homodimer
form two salt-bridges, and this interaction is weakened at
reduced pH resulting in separation of the two monomers. In
the BthTx-I, the separation of the monomers is concomitant
with the reduction of the Ca
2þ
-independent membrane
damaging activity (de Oliveira et al., 2001).
The ‘SequenceSpace’ analysis is a useful tool for
identifying amino acid clusters that are specific to a protein
family, however, the question remains as to the significance
of the result. A drawback of the technique is that the
sequences become grouped based on increasingly subtle
similarities, and at a certain level the similarities due to the
high amino acid identity in PLA
2
s from related species mask
the similarities arising from shared biological function. In
order to minimize this effect, an amino acid sequence
selection procedure is used which eliminates sequences that
are considered to be highly similar, such as isoforms. This
sequence selection procedure is common practice in many
comparative studies, and unfortunately may result in the loss
of relevant information from the alignment under analysis.
Furthermore, those amino acids that may determine
common functional properties between otherwise dissimilar
proteins will have a decreased significance. For example,
the aromatic/cationic C-terminal loop cluster is conserved in
all Lys49-PLA
2
s, however, the ‘SequenceSpace’ analysis
identified only one cationic residue the C-terminal loop
region (Ward, 1998b), which is a consequence of this
aromatic/cationic motif being common to other families of
PLA
2
s.
3.3. Scanning amino acid variability analysis (SAVANA)
of sequence alignments
Although a powerful analytic tool, the ‘SequenceSpace’
program failed to extract subtle yet important sequence
information, and this has prompted us to develop a new
analysis specifically designed for alignments of highly
similar (.95% identity) amino acid sequences. The analysis
is simple and is derived from a comparison of the amino acid
sequences of isoforms with differing levels of a given
activity. Any differences in the sequences must be
responsible for the observed modulation of the effect of
interest, and surface mapping of these positions will
therefore indicate the location of the structural determinant
of the effect. Initially the amino acid sequence database is
searched in order to identify the most closely related
sequences to the protein of interest. Typically, an alignment
of the 8–10 most closely related sequences is performed,
and although the sequences are closely related, subsequent
manual adjustments may still be required.
The results of the database search and sequence
alignments of three vPLA
2
s with different biological
activities are presented in Fig. 1. These sequence alignments
serve as the input for the SAVANA program, which initially
derives the consensus amino acid sequence shown for each
sequence block in Fig. 1. Subsequently, the program scans
all positions in the alignment, identifies every amino acid
variation from the consensus sequence, and calculates a
position score based on a point accepted mutation matrix
(PAM, Dayhoff et al., 1978). We have chosen the PAM5
matrix since the sequences that are routinely included in the
analysis differ by less than 5%. The result of SAVANA is a
score for each position in the sequence alignment that is
directly related to the number and type of amino acid
substitutions found at that position. This score is a numerical
value that can be represented using a color scale in a solid
surface representation of a protein structure hosen from the
PLA
2
family that is under analysis. In this way the positions
and significance of dissimilarities between otherwise highly
similar amino acid sequences may be identified and mapped
onto the protein surface.
The SAVANA program treats the insertion or deletion
of an amino acid residue as a highly significant event
which is assigned a high score. In order to represent these
insertion events in a surface map, the 3D structure of the
protein with the longest amino acid sequence must therefore
be used. In the case of the Lys49-PLA
2
myotoxins, the
longest sequence is that of the myotoxin-II from Bothrops
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883872

Fig. 1. Amino acid sequence alignment of snake venom phospholipases A
2
.(a) Myotoxic Lys49 PLA2 s from Bothrops species. MjTX-I: myotoxin I Bothrops moojeni (Soares et al., 2000a);
MjTX-II: myotoxin II B. moojeni (Soares et al., 1998); PrTX-I: piratoxin-I B. pirajai (Toyama et al., 1998); PrTX-II: piratoxin-II B. pirajai (Toyama et al., 2000); BnSP-7: Lys49-PLA2 B.
neuwiedi (Soares et al., 2000b); BthTX-I: bothropstoxin-I B. jararacussu (Cintra et al., 1993); BaspII: myotoxin II B. asper (Francis et al., 1991); and M1-3-3: myotoxin II iosoform B. asper
(Pescatori et al., 1998). (b) Pre-synaptic neurotoxic PLA
2
s, PA2C_VIPAA from Vipera ammodytes ammodytes (Pungercar et al., 1989); PA2A_VIPAA from Vipera a. ammodytes (Pungercar
et al., 1991); PA2B_VIPAA from Vipera a. ammodytes (Kordis et al., 1990); PA28 DBRR from Daboia russellii russellii (Gowda et al., 1994); PA2L_VIPAA from Vipera a. ammodytes
(Pungercar et al., 1990); PA25_ECHOC from Echis ocellatus (Harrison et al., 2002); PA21_BOTAS from Bothrops asper (Kaiser et al., 1990); PA21_BOTJR from Bothrops jararacussu
(Moura-da-Silva et al., 1995); PA24_AGKHP from Gloydius halys (Pan et al., 1998) and PA2A_TRIMU from Protobothrops mucrosquamatus (Guo et al., 2001). (c) Anti-coagulant PLA
2
s.
PA2A_TRIFL from Trimeresurus flavoviridis (Yamaguchi et al., 2001); PA2B_TRIFL from T. flavoviridis(Yamaguchi et al., 2001); PA2X_TRIFL from T. flavoviridis (Kini et al., 1986);
PA2Q_TRIFL from T. flavoviridis (Chijiwa et al., 2003); PA2W_TRIFL from T. flavoviridis (Ogawa et al., 1992); PA2Y_TRIFL from T. flavoviridis (Chijiwa et al., 2003); PA29_AGKHP from
Gloydius halys pallas (Pan et al., 1998); PA24_AGKHP from Gloydius h. pallas (Pan et al., 1998); and PA21_AGKHA from Gloydius h. blomhoffi (Forst et al., 1986). The consensus sequence is
shown in the last line for each alignment block. The first sequence in each block was used as the query sequence in the database searches using BLAST (Altschul et al., 1990), and the penultimate
column shows the expectation values for each sequence.
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883 873

Fig. 2. Solid surface representations of SAVANA results. Each column presents the view of the chosen PLA
2
structure looking towards: (1) the C-
terminal loop region (left column), (2) the i-face (centre column) and (3) the b-wing region (right column). The first row presents ribbon
representations of the PLA
2
molecules in each of the three orientations, with the principal structural features labelled in thecase of the i-face view.
The second, third and fouth rows present solid surface representations of: myotoxin II, a Lys49-PLA
2
myotoxin from Bothrops moojeni (de
Azevedo et al., 1997); Ammodytoxin A, a pre-synaptic neurotoxin from Vipera a. ammodytes (computer generated molecular model—see text for
details); Basic PLA
2
, an anticoagulant PLA
2
from Gloydius (Agkistrodon) halys Pallas (Zhao et al., 1998). The surfaces are coloured according to
the position scores derived from the SAVANA of each corresponding sequence block shown in Fig. 1. Fully conserved positions are shown in dark
blue, the highest scored (i.e. the most significant) positions are shown in red, and intermediate scores lie on a colour scale between these two
extremes. All non-conserved positions are labelled in black, and those positions which have been targeted by site directed mutagenesis and which
influenced the biological activity are underlined. Those positions, which have been targeted by site directed mutagenesis and which did not
influence the biological activity are shown in grey. The figure was prepared using the Swiss-Pdb Viewer program (Guex and Peitsch, 1997).
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883874

moojeni, and the surface map representation of the
SAVANA analysis is presented in Fig. 2 using the crystal
structure of this protein (de Azevedo et al., 1997). High
scores were observed for residues in the C-terminal loop
region of the protein, which are therefore identified as
candidates for being the structural determinants of the
biological effects of Lys49-PLA
2
myotoxins. This predic-
tion is supported by the observation that a synthetic peptide
of residues 115–129 in the C-terminal region of myotoxin
II from Bothrops asper demonstrates biological activity,
although at reduced levels when compared to the whole
protein (Lomonte et al., 1994). When the prediction made
by SAVANA is correlated with results of site directed
mutagenesis experiments in BthTx-I, the result is quite
striking. The positions 117– 123 identified in the SAVANA
superpose exactly with the BthTx-I mutants that influence
the myotoxic activity in the C-terminal loop region
(Chioato et al., 2002). Furthermore, SAVANA identified
additional positions 115– 123 in the C-terminal loop region
which show a strong correlation with those positions that
influence membrane damaging activity as identified by site
directed mutagenesis experiments (Chioato et al., 2002).
The SAVANA analysis therefore successfully identified the
structural determinants involved in both the membrane
damaging and myotoxic effects. It should be noted,
however, that additional positions (Phe3, Asp62, Ser90,
Glu82, Lys20 and Ser76) are also identified in the analysis,
and these remain to be evaluated experimentally using site
directed mutagenesis. We are currently using a surface
walking strategy centered on the C-terminal loop region
that will encompass these additional positions.
4. Expression of recombinant vPLA
2
s
The value of sequence analysis lies in the reduction of
a vast range of possibilities to a manageable number of
site-directed mutagenesis experiments. In order to use
this strategy an efficient heterologous protein expression
system, together with a purification protocol, which
yields native protein, is essential. The successful
expression of group I pancreatic PLA
2
s as heterologous
proteins in Escherichia coli or yeast when coupled with
site-directed mutagenesis have proven to be powerful and
reliable tools in the study of the structural bases of basic
enzymatic properties such as catalysis, lipid specificity
and interfacial activation (Yuan and Tsai, 1999). The use
of these tools to study of the structural bases of
biological activities of vPLA
2
s depends on the develop-
ment of adequate expression systems, and as shown in
Table 1, the last decade has witnessed a steady increase
in the number and diversity of vPLA
2
s that have been
expressed and studied using site directed mutagenesis.
Venom PLA
2
s frequently express well in E. coli, and
are generally non-toxic to the host cells. However, in
most cases the major hurdle is the purification of the
recombinant material in the native conformation, since
without exception, group I/II PLA
2
s form inclusion
bodies when expressed in E. coli.Althougheasily
purified by centrifugation, the protein in inclusion bodies
must be solubilized with chemical denaturants and
refolded to their native conformation by slowly decreas-
ing the denaturant concentration. Furthermore, since
native group I/II PLA
2
s contain 6–8 disulphide bonds,
the protein refolding must be performed in the presence
of an oxidation/reduction buffer. Native recombinant
vPLA
2
s have been refolded from inclusion bodies in
high yields by changing the concentration of the
chemical denaturant using dialysis (see Table 1). More
recently, protein refolding protocols using commonly
available gel filtration resins have been reported (Batas
and Chaudhuri, 1996; 1999a,b) and these methods have
been successfully used to refold BthTx-I expressed as
inclusion bodies E. coli (Ward et al., 2001). Successful
attempts to circumvent the refolding problem in E. coli
through the use of fusion protein constructs have also
been reported (Liang et al., 1993; Hodgson et al., 1993;
Pan et al., 1994; Giuliani et al., 2001; Yang et al., 2003).
In these cases, although heterologous protein is expressed
in the refolded state, enzymatic or chemical treatment is
required to liberate the native PLA
2
from the fusion
protein. Although the use of eukaryotic cells for
the production of native correctly folded mammalian
class I/II PLA
2
s has been reported (Bekkers et al., 1991),
to date the neutral PLA
2
from Naja n. naja expressed in
Pichia pastoris is the only vPLA
2
produced using a
eukaryotic host cell system (Lefkowitz et al., 1999).
These expression systems have been used to produce site
directed mutants, which allows the correlation of amino
acid prediction studies with experimental surface map-
ping results.
5. Comparison of vPLA
2
protein surfaces
In the case of the Lys49-PLA
2
myotoxins, preliminary
surface mapping results have revealed that the structural
determinants of the myotoxic and membrane damaging
activities are localized in the C-terminal region of the
protein. Although the residues involved in each of these
activities localizes to the same region in the protein
structure, the specific residues involved in each activity
are clearly independent (Chioato et al., 2002). This provides
clear and direct evidence that specific biological effects are
determined by defined residue clusters on the protein
surface. The strategy used in the Lys49-PLA
2
surface
mapping project was alanine scanning mutagenesis, in
which all target residues in a given surface region are
substituted by alanine residues. This strategy has proven to
be successful in the mapping and study of protein surfaces
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883 875

Table 1
Snake venom phospholipases A
2
expressed in E. coli
Phospholipase A
2
Species Activity Comments Referenece
Acidic PLA
2
Naja naja naja Non-toxic Expression in E. coli, refolded from inclusion
bodies
Kelley et al., 1992
Mutation of positive charge reduces manoalide
binding
Bianco et al., 1995
Acidic PLA
2
Naja naja naja Non-toxic Expression of native protein in Pichia pastoris Lefkowitz et al.,
1999
Mutant D23N reduced ‘PC activation’ effect
Acidic PLA
2
Naja naja naja Non-Toxic Expression of synthetic gene in E.coli, refolded
from inclusion bodies.
Sumandea et al., 1999
Mutation of aromatic residues in IRS influences
binding to lipid surfaces.
Acidic PLA
2
(and isoforms)
Naja naja atra Non-toxic Expression in E. coli, His-tag fusion protein
refolded from inclusion bodies.
Pan et al., 1994
Active site mutagenesis abolished catalytic
activity in two isoforms (PLA
2–1
, and PLA
2-2
).
Pan et al., 1998
PLA
2
Bungarus multicinctus
(Taiwan banded krait)
Non-toxic Expression in E. coli, refolded from inclusion
bodies
Chang et al., 1996a
Chang et al., 1997
Ammodytoxin A Vipera ammodytes
ammodytes
Presynaptic
neurotoxin
Expression in E. coli as a fusion protein Liang et al., 1993
F24 mutant show decreased neurotoxicity Petan et al., 2002
Position 124 important for neurotoxicity Pungercar et al., 1999
Mutagenesis in the C-terminal loop region
influences receptor binding and neurotoxicity
Prijatelj et al., 2000,
2002
Ivanovski et al., 2000
Trimucrotoxin Trimeresurus
mucrosquamatus
Presynaptic
neurotoxin
Expression in E. coli, His-tag fusion protein
refolded from inclusion bodies.
Tsai and Wang, 1998
- Mutation of residues in N-terminal helix
reduced activity
b-bungarotoxin Bungarus multicinctus
(Taiwan banded krait)
Presynaptic
neurotoxin
Expression in E. coli, refolded from inclusion
bodies.
Kuo et al., 1995;
Chang et al., 1996b
(A-chain) Alternative engineered disulphides had no effect
on catalytic activity
Chang et al., 1996c
APP-D-49 Agkistrodon piscivorus
piscivorus
Anti-coagulant Expression in E. coli, refolded from inclusion
bodies
Mutation N1S had no effect on catalytic activity. Lathrop et al., 1992
Cationic residues in N-terminal and active site
regions mediate protein binding to the lipid
interface.
Han et al., 1997
Acidic/basic PLA
2
(ABPLA
2
)
Agkistrodon halys pallas Anti-coagulant Expression in E. coli, refolded from inclusion
bodies
Liu et al., 1999
Acidic PLA
2
Agkistrodon halys Pallas Anti-coagulant Expression in E. coli, refolded from inclusion
bodies
Pan et al., 1998
(APLA
2
) Mutation of residues in N- and C-termini
reduced activity
Liu et al., 2001
Acidic-PLA
2
(APLA
2
isoform)
Agkistrodon halys Pallas Anti-coagulant Expression in E. coli, refolded from inclusion
bodies
Zhong et al., 2001
PLA
2
-9 Lapemis hardwickii Anti-coagulant Expression in E. coli as a fusion protein Yang et al., 2003
Active site mutants do not affect anti-coagulant
activity.
Notechis 11’2 Notechis scutatus scutatus Myotoxin Expression in E. coli as a fusion protein Hodgson et al., 1993
Mutant M8L had no effect on catalysis or
myotoxicity
ACL myotoxin Agkistrodon contortrix
laticinctus
Myotoxin Expression in E. coli as a fusion protein Giuliani et al., 2001
(Continued on next page)
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883876

involved in protein/protein interactions (Bogan and Thorn,
1998; Raffa, 2002) and preliminary results have shown that
this strategy is equally successful in the study of the
myotoxic and membrane damaging activities of the Lys49-
PLA
2
(Chioato et al., 2002).
We have extended the SAVANA analysis to predict
residues that may be determinants of pre-synaptic
neurotoxicity using the ammodytoxin A from Vipera a.
ammodytes (Pungercar et al., 1991) as the reference
sequence for the database searches (see Fig. 1). Pre-
synaptic neurotoxic PLA
2
s have been extensively studied,
and the influence of site-directed mutants on the
neurotoxic activity of ammodytoxin A has been reported
(Pungercar et al., 1999; Prijatelj et al., 2000; 2002;
Ivanovski et al., 2000; Petan et al., 2002). The three-
dimensional structure of ammodytoxin A has not been
experimentally determined, therefore the surface mapping
representation of the SAVANA analysis of this protein
is presented in Fig. 2 using a molecular model of
ammodytoxin A (SWISSPROT Acc. No: PA2A_VIPAA,
Pungercar et al., 1991a) as generated by the program
MODELLER (Sali et al., 1995) using the structure of the
presynaptic neurotoxic PLA
2
from Daboia russelli pul-
chella (PDB code 1FB2, Chandra et al., 2001) as the
protein structure template.
In contrast to the Lys49-PLA
2
myotoxins, the presyn-
aptic neurotoxinic PLA
2
s show a more extensive biologi-
cally active surface that extends from the C-terminal loop
region through the calcium binding loop and short helical
turn and includes positions in the N-terminal helix and
b-wing (see Fig. 2). A previous prediction based on the
visual analysis of a manual amino acid sequence
alignment suggested that residues 80 – 110 on helix 3
were involved in the structural determinant of neurotoxi-
city (Kini and Iwanaga, 1986b; Tsai et al., 1987).
However, the current analysis clearly demonstrates that
the protein surface involved in this activity is more
extensive, and comparison of those regions of the
ammodytoxin A molecule identified in the SAVANA
with the results site directed mutagenesis experiments
reveals a strong correlation. For example, position 24 is
located on the short helix (Fig. 2) scores highly in
SAVANA, and has been demonstrated by site directed
mutagenesis to be a critical residue involved in neurotoxic
activity (Petan et al., 2002). Furthermore, many of the
positions identified by SAVANA and mapped by site
directed mutagenesis are located on the i-face of the
protein, and not only modulate neurotoxicity but also may
result in alterations in the catalytic properties of these
PLA
2
s(Prijatelj et al., 2000, 2002; Ivanovski et al., 2000;
Petan et al., 2002), perhaps as a consequence of the
modification of interfacial binding properties. Finally,
comparison of the surface maps in Fig. 2 reveals that the
C-terminal loop region is a structural determinant of both
presynaptic neurotoxic and myotoxic activities. This is a
highly significant result, since it suggests that a shared
structural motif may explain the mixed neurotoxic and
myotoxic activities observed in many of vPLA
2
s.
Site directed mutagenesis has recently been used to
probe the structural determinants of the anti-coagulant effect
induced by vPLA
2
s, and we have compared the available
mutagenesis data with the results of SAVANA. The
sequence of the basic PLA
2
from Gloydius halys Pallas
(previously Agkistrodon halys Pallas;Pan et al., 1998) was
used to search the amino acid database, and Fig. 2 presents
the surface mapping representation of the analysis using the
structure of the basic anti-coagulant PLA
2
from Gloydius
halys Pallas (PDB code; 1B4W, Zhao et al., 1998,
SWISSPROT; PA24_AGKHP, Pan et al., 1998) as the
protein structure template. The results reveal that a cluster of
positions between 71 and 95 at the base of the beta wing and
residues 115–133 in the C-terminal loop region may be
important structural determinants of the anticoagulant
activity.
A previous visual analysis of a manual amino acid
sequence alignment predicted that residues at 55 – 77 at the
end of helix 2 and the first strand of the b-wing define a
region that is the anticoagulant determinant (Kini and
Evans, 1987), and this prediction therefore correlates well
with the results of the current computer-based analysis.
Furthermore, the results of site directed mutagenesis
experiments are in good agreement with these predictions,
which have identified positions 72 and 77 as significant
determinants of activity (positions 67 and 70 using the
numbering scheme adopted by the authors (Zhong et al.,
2001), together with residues in the C-terminal loop region
Table 1 (continued)
Phospholipase A
2
Species Activity Comments Referenece
(Lys49-PLA
2
)
Bothropstoxin-I Bothrops jararacussu Myotoxin Expression in E. coli, refolded from inclusion
bodies.
Ward et al., 2001
(Lys49-PLA
2
) Active site mutants do not influence myotoxic
activity
Ward et al., 2002
Mutagenesis in the C-terminal loop region
influences myotoxic activity
Chioato et al., 2002
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883 877

(Liu et al., 2001). As is the case in the pre-synaptic
neurotoxins, several of the positions identified as antic-
oagulant determinants are located on the i-face of the
protein, and site directed mutagenesis in these regions
reduce both the catalytic and anticoagulant activities (Liu
et al., 2001). However, these studies show no clear
correlation between lipid hydrolysis and anticoagulant
potency, which suggests a more subtle interplay between
the two activities. Indeed, the regulation of prothrombinase
activity by human secreted group IIA PLA
2
involves both
catalytic and non-catalytic mechanisms, and position 56 is a
structural determinant for the non-catalytic mechanism
(Inada et al., 1994).
6. Interpretation of surface mapping predictions
Comparative analyses of amino acid sequences make the
inherent assumption that the PLA
2
s used in the sequence
alignment not only show the same biological function (e.g.
myotoxicity), but also have the same mechanism of action
(e.g. association with the same acceptor). However, the
correlation of structure with function may not be as
straightforward as is it might first appear. For example,
myotoxic PLA
2
s as a group include toxins which act at the
pre-synaptic membraneof the neuromuscular junction (Fatehi
et al., 1994) that may be distinct from the Lys49-PLA
2
mechanism of action, and which may be have a discrete
structural determinant. Furthermore, the Lys49-PLA
2
myo-
toxins demonstrate various additional biological effects (see
Section 2), and so caution should be exercised when assigning
a surface feature as a structural determinant of a given
biological function. Ina final example, neurotoxic PLA
2
sasa
group have been shown to target diverse membrane acceptors
(Krizaj et al., 1994; 1997), which raises the question as to
whether a unique ‘neurotoxic’ structural determinant exists.
These considerations pose challenges for the reliable
interpretation of comparative analyses of amino acid
sequences that are selected and grouped according to
biological function, and raise the question as to how can
information with respect to the structural determinants of a
given function be derived from grouped protein sequences.
The answer to this problem lies in the careful selection of
sequences, and the use of adequate bioinformatics tools to
search andanalyze the protein sequence databases. Predictions
derived from these studies may serve to focus site-directed
mutagenesis experiments on specific residues or to indicate
regions on the protein surface that are of potential interest. In
the preceeding sections we have described how Lys49-PLA
2
amino acid sequences may be analyzed, which highlights the
strengths and weaknesses of computer based analyses. A
sequence analysis can be used to make predictions, but the
ultimate proof of the involvement of a surface region as a
determinant of biological activity is experimental evidence
derived from site directed mutagenesis.
7. Conclusions and perspectives
Computer-based structural prediction in conjunction
with site directed mutagenesis is proving to be a powerful
combination for the mapping of protein surfaces. Although
no single computer program has proven to be adequate to
extract all the information from a sequence alignment, the
use of several programs in conjunction, each of which
analyzes the same data in a different way, is proving to be an
effective strategy for sequence analysis. Comparisons of
computer predictions with site-directed mutagenesis studies
provide insights as to the structural bases of pharmacologi-
cal effects of vPLA
2
s, and highlight some basic concepts
that are useful to understand the relation between structure
and function of these proteins.
The weight of evidence suggests that a given pharma-
cological activity is determined by a defined surface region
on the vPLA
2
, which may reflect the conservation of a
specific protein surface topology for a specific receptor. This
is in accord with the concept of a surface ‘hot-spot’, which is
a localized region on a protein surface that determines a
specific protein/protein interaction (Bogan and Thorn,
1998). Studies with vPLA
2
isoforms indicate that amino
acid substitutions within the surface regions of likely to alter
the affinity for these receptors, and thereby modulate the
activity of any given PLA
2
. Since the surface topology of a
protein determines the structural basis of a given biological
activity, overlap between the surface regions that determine
specific activities will give rise to PLA
2
s with mixed
pharmacological properties. Since PLA
2
s are relatively
small proteins, the surface area is limited and therefore
overlap between structural determinants of different phar-
macological activities is likely to be a frequent event, and
may be the structural basis that underlies the rich diversity
of vPLA
2
effects. SAVANA results indicate a significant
variation in the areas of the protein surfaces that contribute
to the structural determinants of biological activities. This
may reflect the intrinsic variations of the interactions of
different vPLA
2
s with their specific receptors, and is in
accord with the wide range of buried protein surface areas
observed for different protein/protein interactions (Jones
and Thornton, 1996).
The study of vPLA
2
s has found applications in the
production and improvement of anti-venoms, and computer
prediction together with surface mapping may provide a
useful strategy for the selection or improvement of venom
components that are used as antigens in this process. More
recently, the focus of vPLA
2
research has shifted to the
identification and characterization of protein/ligand inter-
actions, and the study of the mechanism of action of vPLA
2
s
has led to the identification of many cell surface proteins
which act as acceptors for endogenous secreted mammalian
PLA
2
, providing important insights as to the molecular and
cellular biology of this important class of regulatory
proteins. Furthermore, in recent years the complexity of
the regulation of vPLA
2
activity through interaction with
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883878

specific protein inhibitors has become apparent, and the
techniques described here may be applied to study the
structural bases of the interaction between PLA
2
s and these
novel inhibitors. Such insights are fundamental in the design
of novel PLA
2
inhibitors with improved specificities.
Acknowledgements
We are grateful to Dr Jero
ˆnimo Ruiz de Conceic¸a
˜o for
help with the protein modeling. The financial support of
FAPESP genome program (SmolBNet Proc. No. 01/07537-
2), FAPESP Proc. No. 02/04367-1 and CNPq is acknowl-
edged. LC is the recipient of FAPESP doctorate fellowship
01/00279-8.
References
Adamich, M., Dennis, E.A., 1978. Specificity reversal in phospho-
lipase A
2
hydrolysis of lipid mixtures. Biochem. Biophys. Res.
Commun. 80, 424.
Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J.,
1990. Basic local alignment search tool. J. Mol. Biol. 215 (3),
403– 410.
Andrade, M.A., Casari, G., Sander, C., Valencia, A., 1997.
Classification of protein families and detection of the determi-
nant residues with an improved self-organizing map. Biol.
Cybern. 76 (6), 441–450.
Arni, R.K., Ward, R.J., 1996. Phospholipase A
2
–a structural
review. Toxicon 34 (8), 827–841.
Arni, R.K., Ward, R.J., Cintra, A.C., Giglio, J.R., 1995. Crystal-
lization and preliminary diffraction data of bothropstoxin I
isolated from the venom of Bothrops jararacussu. Toxicon 33
(3), 383– 386.
de Azevedo, W.F., Ward, R.J., Lombardi, F.R., Giglio, J.R., Fontes,
M.R.M., 1997. Crystal structure of myotoxin-II: a myotoxic
phospholipase A
2
homologue from Bothrops moojeni venom.
Prot. Peptide Lett. 4, 329–334.
Batas, B., Chaudhuri, J.B., 1996. Improving protein refolding
yields by minimizing aggregation. Ann. NY Acad. Sci. 782,
495– 505.
Batas, B., Chaudhuri, J.B., 1999a. Inclusion body purification and
protein refolding using microfiltration and size exclusion
chromatography. J. Biotechnol. 68 (2–3), 149– 158.
Batas, B., Chaudhuri, J.B., 1999b. Considerations of sample
application and elution during size-exclusion chromatography-
based protein refolding. J. Chromatogr. A 864 (2), 229 – 236.
Beitz, E., 2000. TEXshade: shading and labeling of multiple
sequence alignments using LATEX2 epsilon. Bioinformatics 16
(2), 135– 139.
Bekkers, A.C., Franken, P.A., van den Bergh, C.J., Verbakel, J.M.,
Verheij, H.M., de Haas, G.H., 1991. The use of genetic
engineering to obtain efficient production of porcine pancreatic
phospholipase A
2
by Saccharomyces cerevisiae. Biochim.
Biophys. Acta. 1089 (3), 345 – 351.
Berg, O.G., Gelb, M.H., Tsai, M.D., Jain, M.K., 2001. Interfacial
enzymology: the secreted phospholipases A
2
-paradigm. Chem.
Rev. 101, 2613–2653.
van den Bergh, C.J., Slotboom, A.J., Verheij, H.M., de Haas, G.H.,
1988. The role of aspartic acid-49 in the active site of
phospholipase A
2
. A site-specific mutagenesis study of porcine
pancreatic phospholipase A
2
and the rationale of the enzymatic
activity of [lysine49]phospholipase A
2
from Agkistrodon
piscivorus piscivorus venom. Eur. J. Biochem. 176 (2),
353–357.
Bianco, I.D., Kelley, M.J., Crowl, R.M., Dennis, E.A., 1995.
Identification of two specific lysines responsible for the
inhibition of phospholipase A
2
by manoalide. Biochim.
Biophys. Acta 1250 (2), 197–203.
Bogan, A.A., Thorn, K.S., 1998. Anatomy of hot spots in protein
interfaces. J. Mol. Biol. 280 (1), 1 –9.
Boilard, E., Bourgoin, S.G., Bernatchez, C., Surette, M.E., 2003.
Identification of an autoantigen on the surface of apoptotic
human T-cells as a new protein interacting with inflamma-
tory group IIA phospholipase A
2
. Blood (Epub ahead of
print).
Brunie, S., Bolin, J., Gerwith, D., Sigler, P.B., 1985. The refined
crystal structure of dimeric phospholipase A
2
at 2.5 A
˚. Access to
a shielded catalytic site. J. Biol. Chem. 260, 9742 – 9749.
Casari, G., Sander, C., Valencia, A., 1995. A method to predict
functional residues in proteins. Nat. Struct. Biol. 2 (2),
171–178.
de Castro, R.C., Landucci, E.C., Toyama, M.H., Giglio, J.R.,
Marangoni, S., De Nucci, G., Antunes, E., 2000. Leucocyte
recruitment induced by type II phospholipases A(2) into the rat
pleural cavity. Toxicon 38 (12), 1773–1785.
Chandra, V., Kaur, P., Jasti, J., Betzel, C., Singh, T.P., 2001.
Regulation of catalytic function by molecular association:
structure of phospholipase A
2
from Daboia russelli pulchella
(DPLA2) at 1.9 A resolution. Acta Cryst. D Biol. Crystal 57 (Pt
12), 1793– 1798.
Chang, C.C., Su, M.J., 1982. Presynaptic toxicity of the histidine-
modified, phospholipase A
2
-inactive, beta-bungarotoxin, cro-
toxin and notexin. Toxicon 20 (5), 895 – 905.
Chang, L.S., Wu, P.F., Chang, C.C., 1996a. cDNA sequence
analysis and mutagenesis studies on the A chain of beta-
bungarotoxin from Taiwan banded krait. J. Prot. Chem. 15 (8),
755–761.
Chang, L.S., Wu, P.F., Chang, C.C., 1996b. Expression of Taiwan
banded krait phospholipase A
2
in Escherichia coli, a fully active
enzyme generated by hydrolyzing with aminopeptidase. Bio-
chem. Biophys. Res. Commun. 225 (3), 990 – 996.
Chang, L.S., Wu, P.F., Chang, C.C., 1996c. cDNA sequence
analysis and expression of the a chain of beta-bungarotoxin
from Bungarus multicinctus (Taiwan banded krait). Biochem.
Biophys. Res. Commun. 221 (2), 328 –332.
Chang, L.S., Chang, C.C., Wu, P.F., 1997. The structural and
functional essentiality of the N-terminal alpha-helix in the
phospholipase A
2
of the Taiwan banded krait. Biochem. Mol.
Biol. Int. 41 (6), 1247–1253.
Chijiwa, T., Yamaguchi, Y., Ogawa, T., Deshimaru, M., Nobuhisa,
I., Nakashima, K., Oda-Ueda, N., Fukumaki, Y., Hattori, S.,
Ohno, M., 2003. Interisland evolution of Trimeresurus flavovir-
idis venom phospholipase A(2) isozymes. J. Mol. Evol. 56 (3),
286–293.
Chioato, L., de Oiveira, A.H.C., Ruller, R., Sa
´, J.M., Ward, R.J.,
2002. Distinct sites for myotoxic and membrane-damaging
activities in the C-terminal region of a Lys49 phospholipase A
2
.
Biochem. J. 366, 971–976.
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883 879

Cintra, A.C.O., Marangoni, S., Oliveira, B., Giglio, J.R., 1993.
Bothropstoxin-I: amino acid sequence and function. J. Prot.
Chem. 12, 57–64.
van Deenen, L.L.M., de Haas, G.H., 1963. The substrate
specificity of phospholipase A
2
. Biochem. Biophys. Acta
70, 538–553.
Dayhoff, M.O., Schawartz, R.M., Arcutt, 1978. A model of
evolutionary change in proteins. In: Dayhoff, M.O., (Ed.), tlas
of Protein Sequence and Structure, National Biomedical
Research Foundation, Washington, DC, pp. 345– 352.
Dennis, E.A., 1994. Diversity of group types, regulation, and
function of phospholipase A
2
. J. Biol. Chem. 269 (18),
13057– 13060.
Dessen, A., 2000. Phospholipase A
2
enzymes: structural diversity in
lipid messenger metabolism. Structure, r15– r22.
Dhillon, D.S., Condrea, E., Maraganore, J.M., Heinrikson, R.L.,
Benjamin, S., Rosenberg, P., 1987. Comparison of enzymatic
and pharmacological activities of lysine-49 and aspartate-49
phospholipases A
2
from Agkistrodon piscivorus piscivorus
snake venom. Biochem. Pharmacol. 36, 1723–1730.
Dı
´az-Oreiro, C., Gutie
´rrez, J.M., 1997. Chemical modification of
histidine and lysine residues of myotoxic phospholipases A
2
isolated from Bothrops asper and Bothrops godmani snake
venoms: effects on enzymatic and pharmacological properties.
Toxicon 35 (2), 241–252.
Dı
´az, C., Gutie
´rrez, J.M., Lomonte, B., Gene
´, J.A., 1991. The effect
of myotoxins isolated from Bothrops snake venoms on multi-
lamellar liposomes: relationship to phospholipase A
2
, antic-
oagulant and myotoxic activities. Biochem. Biophys. Acta
1070, 455–460.
Fatehi, M., Rowan, E.G., Harvey, A.L., Harris, J.B., 1994. The
effects of five phospholipases A
2
from venom of king brown
snake. Pseudechis australis, on nerve and muscle. Toxicon 32
(12), 1559– 1572.
Faure, G., Harvey, A.L., Thomson, E., Saliou, B., Radvanyi, F.,
Bon, C., 1993. Comparison of crotoxin isoforms reveals that
stability of complex plays a major role in its pharmacological
action. Eur. J. Biochem. 214, 491–496.
Feng, D.F., Doolittle, R.F., 1987. Progressive sequence alignment
as a prerequisite to correct phylogenetic trees. J. Mol. Evol. 25
(4), 351– 360.
Fletcher, J.E., Jiang, M.S., 1998. Lys49 phospholipase A
2
myotoxins lyse cell cultures by two distinct mechanisms.
Toxicon 36 (11), 1549–1555.
Forst, S., Weiss, J., Blackburn, P., Frangione, B., Goni, F., Elsbach,
P., 1986. Amino acid sequence of a basic Agkistrodon halys
blomhoffii phospholipase A
2
. Possible role of NH
2
-terminal
lysines in action on phospholipids of Escherichia coli.
Biochemistry 25 (15), 4309– 4314.
Francis, B., Gutie
´rrez, J.M., Lomonte, B., Kaiser, I.I., 1991.
Myotoxin II from Bothrops asper (Terciopelo) venom is a
lysine-49 phospholipase A
2
. Arch. Biochem. Biophys. 284,
352– 359.
Fremont, D.H., Anderson, D.H., Wilson, I.A., Dennis, E.A., Xuong,
N.H., 1993. Crystal structure of phospholipase A
2
from Indian
cobra reveals a trimeric association. Proc. Natl Acad. Sci. USA
90 (1), 342–346.
Giuliani, C.D., Iemma, M.R., Bondioli, A.C., Souza, D.H., Ferreira,
L.L., Amaral, A.C., Salvini, T.F., Selistre-de-Araujo, H.S.,
2001. Expression of an active recombinant lysine 49
phospholipase A(2) myotoxin as a fusion protein in bacteria.
Toxicon 39 (10), 1595–1600.
Gowda, V.T., Schmidt, J., Middlebrook, J.L., 1994. Primary
sequence determination of the most basic myonecrotic phos-
pholipase A
2
from the venom of Vipera russelli. Toxicon 32 (6),
665–673.
Guex, N., Peitsch, M.C., 1997. SWISS-MODEL and the Swiss-
PdbViewer: an environment for comparative protein modeling.
Electrophoresis 18 (15), 2714–2723.
Guo, Y.W., Chen, C.T., Chang, T.Y., 2001. Sequence Direct
Submission, Accession number Q90W39,.
Gutie
´rrez, J.M., Lomonte, B., 1995. Phospholipase A
2
myotoxins
from Bothrops snake venoms. Toxicon 33 (11), 1405 – 1424.
Habermann, E., Breithaupt, H., 1978. Mini-review. The crotoxin
complex—an example of biochemical and pharmacological
protein complementation. Toxicon 16 (1), 19–30.
Han, S.K., Yoon, E.T., Scott, D.L., Sigler, P.B., Cho, W., 1997.
Structural aspects of interfacial adsorption. A crystallographic
and site-directed mutagenesis study of the phospholipase A
2
from the venom of Agkistrodon piscivorus piscivorus. J. Biol.
Chem. 272 (6), 3573–3582.
Hanasaki, K., Arita, H., 1999. Biological and pathological functions
of phospholipase A(2) receptor. Arch. Biochem. Biophys. 372
(2), 215– 223.
Harrison, R.A., Hasson, S., Bharati, K., 2002. Sequence direct
submission. Acession number, P59171.
Higashino, K, 2002. Identification of a soluble form of phospho-
lipase A
2
receptor as a circulation endogenous inhibitor for
secretorry phospholipase A
2
. J. Biol. Chem. 277 (16),
13583–13588.
Higgins, D.G., Bleasby, A.J., Fuchs, R., 1992. CLUSTAL V:
improved software for multiple sequence alignment. CABIOS 8
(2), 189– 191.
Hodgson, D., Gasparini, S., Drevet, P., Ducancel, F., Bouet, F.,
Boulain, J.C., Harris, J.B., Menez, A., 1993. Production of
recombinant notechis II’2L, an enzymatically active mutant
of a phospholipase A
2
from Notechis scutatus scutatus
venom, as directly generated by cleavage of a fusion protein
produced in Escherichia coli. Eur. J. Biochem. 212 (2),
441–446.
Inada, M., Crowl, R.M., Bekkers, A.C., Verheij, H., Weiss, J., 1994.
Determinants of the inhibitory action of purified 14-kDa
phospholipase A
2
on cell-free prothrombinase complex.
J. Biol. Chem. 269 (42), 26338 –26343.
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.
Comm. 276, 1229–1234.
Jones, S., Thornton, J.M., 1996. Principles of protein – protein
interactions. Proc. Natl. Acad. Sci. USA 93, 13 – 20.
Kaiser, I.I., Gutie
´rrez, J.M., Plummer, D., Aird, S.D., Odell, G.V.,
1990. The amino acid sequence of a myotoxic phospholipase
from the venom of Bothrops asper. Arch. Biochem. Biophys.
278 (2), 319–325.
Kelley, M.J., Crowl, R.M., Dennis, E.A., 1992. Renaturation of
cobra venom phospholipase A
2
expressed from a synthetic
gene in Escherichia coli. Biochim. Biophys. Acta 1118 (2),
107–115.
Kini, R.M., Evans, H.J., 1987. Structure-function relationships of
phospholipases. The anticoagulant region of phospholipases A
2
.
J. Biol. Chem. 262 (30), 14402 –14407.
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883880

Kini, R.M., Evans, H.J., 1995. The role of enzymatic activity in
inhibition of the extrinsic tenase complex by phospholipase A
2
isoenzymes from Naja nigricollis venom. Toxicon 33 (12),
1585– 1590.
Kini, R.M., Iwanaga, S., 1986a. Structure – function relationships of
phospholipases II: Charge density distribution and the myo-
toxicity of presynaptically neurotoxic phospholipases. Toxicon
24 (9), 895–905.
Kini, R.M., Iwanaga, S., 1986b. Structure –function relationships of
phospholipases I: Prediction of presynaptic neurotoxicity.
Toxicon 24 (6), 527–541.
Kini, R.M., Kawabata, S.I., Iwanaga, S., 1986. Comparison of
amino terminal region of three isoenzymes of phospholipases
A
2
(TFV PL-Ia, TFV PL-Ib, TFV PL-X) from Trimeresurus
flavoviridis (habu snake) venom and the complete amino acid
sequence of the basic phospholipase, TFV PL-X. Toxicon 24
(11-12), 1117– 1129.
Kordis, D., Pungercar, J., Strukelj, B., Liang, N.S., Gubensek, F.,
1990. Sequence of the cDNA coding for ammodytoxin B.
Nucleic Acids Res. 18 (13), 4016.
Krizaj, I., Dolly, J.O., Gubensek, F., 1994. Identification of the
neuronal acceptor in bovine cortex for ammodytoxin C, a
presynaptically neurotoxic phospholipase A
2
. Biochemistry 33
(46), 13938– 13945.
Krizaj, I., Faure, G., Gubensek, F., Bon, C., 1997. Neurotoxic
phospholipases A
2
ammodytoxin and crotoxin bind to distinct
high-affinity protein acceptors in Torpedo marmorata electric
organ. Biochemistry 36 (10), 2779–2787.
Kuo, K.W., Chen, Y.C., Chang, C.C., 1995. cDNA sequence
analysis and expression of alpha-bungarotoxin from Taiwan
banded krait Bungarus multicinctus. Biochem. Biophys. Res.
Commun. 216 (3), 1088–1094.
Lambeau, G., Lazdunski, M., 1999. Receptors for a growing family
of secreted phospholipases A
2
. Tr. Pharm. Sci. 20 (4), 162 –170.
Landucci, E.C., Castro, R.C., Pereira, M.F., Cintra, A.C., Giglio,
J.R., Marangoni, S., Oliveira, B., Cirino, G., Antunes, E., De
Nucci, G., 1998. Mast cell degranulation induced by two
phospholipase A
2
homologues: dissociation between enzymatic
and biological activities. Eur. J. Pharmacol. 343 (2 –3), 257 – 263.
Landucci, E.C., de Castro, R.C., Toyama, M., Giglio, J.R.,
Marangoni, S., De Nucci, G., Antunes, E., 2000. Inflammatory
oedema induced by the Lys-49 phospholipase A(2) homologue
piratoxin-I in the rat and rabbit. Effect of polyanions and p-
bromophenacylbromide. Biochem. Pharmacol. 59 (10),
1289– 1294.
Lathrop, B.K., Burack, W.R., Biltonen, R.L., Rule, G.S., 1992.
Expression of a group II phospholipase A
2
from the venom of
Agkistrodon piscivorus piscivorus in Escherichia coli: recovery
and renaturation from bacterial inclusion bodies. Prot. Expr.
Purif. 3 (6), 512–517.
Lefkowitz, L.J., Deems, R.A., Dennis, E.A., 1999. Expression of
group IA phospholipase A
2
in Pichia pastoris: identification of a
phosphatidylcholine activator site using site-directed mutagen-
esis. Biochemistry 38 (43), 14174–14184.
Liang, N.S., Pungercar, J., Krizaj, I., Strukelj, B., Gubensek, F.,
1993. Expression of fully active ammodytoxin A, a potent
presynaptically neurotoxic phospholipase A
2
,inEscherichia
coli. FEBS Lett. 334 (1), 55 – 59.
Lichtarge, O., Bourne, H.R., Cohen, F.E., 1996. An evolutionary
trace method defines binding surfaces common to protein
families. J. Mol. Biol. 257 (2), 342 – 358.
Liu, S.Y., Yoshizumi, K., Oda, N., Ohno, M., Tokunaga, F.,
Iwanaga, S., Kihara, H., 1990. Purification and amino acid
sequence of basic protein II, a lysine-49-phospholipase A
2
with
low activity, from Trimeresurus flavoviridis venom. J. Biochem.
(Tokyo) 107 (3), 400–408.
Liu, C.S., Chen, J.M., Chang, C.H., Chen, S.W., Teng, C.M., Tsai,
I.H., 1991. The amino acid sequence and properties of an
edema-inducing Lys-49 phospholipase A
2
homolog from the
venom of Trimeresurus mucrosquamatus. Biochim. Biophys.
Acta 1077 (3), 362–370.
Liu, X., Pan, H., Yang, G., Wu, X., Zhou, Y., 1999. Cloning,
expression and biochemical characterization of a basic – acidic
hybrid phospholipase A
2
-II from Agkistrodon halys Pallas.
Biochim. Biophys. Acta 1431 (1), 157–165.
Liu, X., Wu, X., Zhou, Y., 2001. Identification of key residues
responsible for enzymatic and platelet-aggregation-inhibiting
activities of acidic phospholipase A
2
from Agkistrodon halys
Pallas. J. Nat. Toxins 10 (1), 43 – 55.
Lomonte, B., Moreno, E., Tarkowski, A., Hanson, L.A., Maccar-
ama, M., 1994. Neutralizing interaction between heparin and
myotoxin II, a Lys49 phospholipase A2 from Bothrops asper
snake venom. Identification of a heparin-binding and cytolytic
toxin region by the use of synthetic peptides and molecular
modelling. J. Biol. Chem. 269, 29867–29873.
Maraganore, J.M., Heinrikson, R.L., 1986. The lysine-49 phospho-
lipase A
2
from the venom of Agkistrodon piscivorus piscivorus.
Relation of structure and function to other phospholipases A
2
.
J. Biol. Chem. 261 (11), 4797 –4804.
Moura-da-Silva, A.M., Paine, M.J., Diniz, M.R., Theakston, R.D.,
Crampton, J.M., 1995. The molecular cloning of a phospho-
lipase A
2
from Bothrops jararacussu snake venom: evolution of
venom group II phospholipases A
2
may imply gene dupli-
cations. J. Mol. Evol. 41 (2), 174–179.
Murakami, M., Kudo, I., 2001. Diversity and regulatory functions of
mammalian secretory phospholipases A
2
. Adv. Immunol. 77,
163–194.
Murakami, M., Nakatani, Y., Atsumi, G., Inoue, K., Kudo, I., 1997.
Regulatory functions of phospholipase A
2
. Crit. Rev. Immunol.
17 (3–4), 225– 283.
Ogawa, T., Oda, N., Nakashima, K., Sasaki, H., Hattori, M., Sakaki,
Y., Kihara, H., Ohno, M., 1992. Unusually high conservation of
untranslated sequences in cDNAs for Trimeresurus flavoviridis
phospholipase A
2
isozymes. Proc. Natl. Acad. Sci. USA 89 (18),
8557–8561.
de Oliveira, A.H., Giglio, J.R., Andria
˜o-Escarso, S.H., Ward, R.J.,
2001. The effect of resonance energy homotransfer on the
intrinsic tryptophan fluorescence emission of the bothropstoxin-
I dimer. Biochem. Biophys. Res. Comm. 284 (4), 1011 – 1015.
Pan, F.M., Chao, S.C., Wu, S.H., Chang, W.C., Chiou, S.H., 1998.
Characterization of phospholipase A
2
(PLA
2
) from Taiwan
Cobra: isoenzymes and their site-directed mutants. Biochem.
Biophys. Res. Comm. 250 (1), 154–160.
Pan, F.M., Yeh, M.S., Chang, W.C., Hung, C.C., Chiou, S.H., 1994.
Sequence analysis and expression of phospholipase A
2
from
Taiwan cobra. Biochem. Biophys. Res. Comm. 199 (2),
969–976.
Pa
´ramo, L., Lomonte, B., Pizarro-Cerda
´, J., Bengoechea, J.A.,
Gorvel, J.P., Moreno, P., 1998. Bactericidal activity of Lys-49
and Asp-49 myotoxic phospholipase A
2
from Bothrops asper
snake venom: synthetic Lys49 myotoxin II-(115 –129)-peptide
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883 881

identifies its bactericidal region. Eur. J. Biochem. 253 (2),
452– 461.
Pan, H., Liu, X.L., Ou-Yang, L.L., Yang, G.Z., Zhou, Y.C., Li, Z.P.,
Wu, X.F., 1998. Diversity of cDNAs encoding phospholipase
A
2
from Agkistrodon halys Pallas venom, and its expression in
E. coli. Toxicon 36 (8), 1155–1163.
Pescatori, M., Grasso, A., Rufini, S., 1998. Molecular cloning of a
K49 PLA2-like myotoxin from the snake Bothrops asper, Direct
Submission. Accession number AF109911_1,.
Petan, T., Krizaj, I., Gubensek, F., Pungercar, J., 2002. Phenyl-
alanine-24 in the N-terminal region of ammodytoxins is
important for both enzymatic activity and presynaptic toxicity.
Biochem. J. 363, 353–358.
Pieterson, W.A., Vidal, J.C., Volwerk, J.J., de Haas, G.H., 1974.
Zymogen-catalyzed hydrolysis of monomeric substrates and the
presence of a recognition site for lipid –water interfaces in
phospholipase A
2
. Biochemistry 13 (7), 1455– 1460.
Prijatelj, P., Copic, A., Krizaj, I., Gubensek, F., Pungercar, J., 2000.
Charge reversal of ammodytoxin A, a phospholipase A
2
-toxin,
does not abolish its neurotoxicity. Biochem. J. 352, 251 – 255.
Prijatelj, P., Krizaj, I., Krajl, B., Gubensek, F., Pungercar, J., 2002.
The C-terminal region of ammodytoxins is important but not
sufficient for neurotoxicity. Eur. J. Biochem. 269, 5759– 5764.
Pungercar, J., Kordis, D., Jerala, R., Trstenjak-Prebanda, M.,
Dolinar, M., Curin-Serbec, V., Komel, R., Gubensek, F., 1989.
Amino acid sequence of ammodytoxin C as deduced from
cDNA. Nucl. Acids Res. 17 (11), 4367.
Pungercar, J., Liang, N.S., Strukelj, B., Gubensek, F., 1990.
Nucleotide sequence of a cDNA encoding ammodytin L.
Nucl. Acids Res. 18 (15), 4601.
Pungercar, J., Kordis, D., Strukelj, B., Liang, N.S., Gubensek, F.,
1991. Cloning and nucleotide sequence of a cDNA encoding
ammodytoxin A, the most toxic phospholipase A
2
from the
venom of long-nosed viper (Vipera ammodytes). Toxicon 29
(2), 269– 273.
Pungercar, J., Krizaj, I., Liang, N.S., Gubensek, F., 1999. An
aromatic, but not a basic, residue involved in the toxicity of
group-II phospholipase A
2
neurotoxins. Biochem. J. 341,
139– 145.
Ramirez, F., Jain, M.K., 1991. Phospholipase A
2
at the bilayer
interface. Proteins 9 (4), 229–239.
Raffa, R., 2002. Harmonic mean relationship between affinity for
wold-type receptors and alanine-scan mutants. J. Theor. Biol.
218, 207–214.
Rufini, S., Cesaroni, P., Desideri, R.F., Gubensek, F., Gutie
´rrez,
J.M., Luly, P., Massoud, R., Morero, R., Pedersen, J.Z., 1992.
Calcium ion independent membrane leakage induced by
phospholipase-like myotoxins. Biochemistry 31, 12424– 12430.
Rufini, S., Cesaroni, M.P., Balestro, N., Luly, P., 1996.
Proliferative effect of ammodytin L from the venom of
Vipera ammodytes on 208F rat fibroblasts in culture.
Biochem. J. 320 (Pt 2), 467 – 472.
Sali, A., Potterton, L., Yuan, F., van Vlijmen, H., Karplus, M., 1995.
Evaluation of comparative protein modeling by MODELLER.
Proteins 23 (3), 318–326.
Scott, D.L., Sigler, P.B., 1994. Structure and catalytic mechanism of
secretory phospholipases A
2
. Adv. Protein Chem. 45, 53–58.
Selistre de Araujo, H.S., White, S.P., Ownby, C.L., 1996.
Sequence analysis of Lys49 phospholipase A
2
myotoxins: a
highly conserved class of proteins. Toxicon 34 (11/12),
1237– 1242.
da Silva Giotto, M.T., Garratt, R.C., Oliva, G., Mascarenhas, Y.P.,
Giglio, J.R., Cintra, A.C., de Azevedo, W.F., Arni, R.K., Ward,
R.J., 1998. Crystallographic and spectroscopic characterization
of a molecular hinge: conformational changes in bothropstoxin
I, a dimeric Lys49-phospholipase A
2
homologue. Proteins 30
(4), 442– 454.
Six, D.A., Dennis, E.A., 2000. The expanding superfamily of
phospholipase A
2
enzymes: classification and characterization.
Biochim. Biophys. Acta 1488, 1 –19.
Snitko, Y., Han, S.K., Lee, B.I., Cho, W., 1999. Differential
interfacial and substrate binding modes of mammalian pancrea-
tic phospholipases A
2
: a comparison among human, bovine, and
porcine enzymes. Biochemistry 38 (24), 7803–7810.
Soares, A.M., Rodrigues, V.M., Homsi-Brandeburgo, M.I.,
Toyama, M.H., Lombardi, F.R., Arni, R.K., Giglio, J.R., 1998.
A rapid procedure for the isolation of the Lys-49 myotoxin II
from Bothrops moojeni (caissaca) venom: biochemical charac-
terization, crystallization, myotoxic and edematogenic activity.
Toxicon 36 (3), 503–514.
Soares, A.M., Andria
˜o-Escarso, S.H., Angulo, Y., Lomonte, B.,
Gutie
´rrez, J.M., Marangoni, S., Toyama, M.H., Arni, R.K.,
Giglio, J.R., 2000a. Structural and functional characterization of
myotoxin I, a Lys49 phospholipase A
2
homologue from
Bothrops moojeni (Caissaca) snake venom. Arch. Biochem.
Biophys. 373, 7–15.
Soares, A.M., Andriao-Escarso, S.H., Bortoleto, R.K., Rodrigues-
Simioni, L., Arni, R.K., Gutie
´rrez, J.M., Giglio, J.R., 2001.
Dissociation of enzymatic and pharmacological properties of
piratoxins-I and -III, two myotoxic phospholipases A
2
from
Bothrops pirajai snake venom. Arch. Biochem. Biophys. 387
(2), 188– 196.
Soares, A.M., Guerra-Sa
´, R., Borja-Oliveira, C.R., Rodrigues,
V.M., Rodrigues-Simioni, L., Rodrigues, V., Fontes, M.R.M.,
Lomonte, B., Gutie
´rrez, J.M., Giglio, J.R., 2000b. Structural and
functional characterization of BnSP-7, a Lys49 myotoxic
phospholipase A
2
homologue from Bothrops neuwiedi pau-
loensis venom. Arch. Biochem. Biophys. 378, 201 – 209.
Sribar, J., Copic, A., Paris, A., Sherman, N.E., Gubensek, F., Fox,
J.W., Krizaj, I., 2001. A high affinity acceptor for phospholipase
A
2
with neurotoxic activity is a calmodulin. J. Biol. Chem. 276
(16), 12493–12496.
Sumandea, M., Das, S., Sumandea, C., Cho, W., 1999. Roles of
aromatic residues in high interfacial activity of Naja naja atra
phospholipase A
2
. Biochemistry 38 (49), 16290 –16297.
Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W:
improving the sensitivity of progressive multiple sequence
alignment through sequence weighting, position specific gap
penalties and weight matrix choice. Nucleic Acids Res. 22,
4673–4680.
Toyama, M.H., Soares, A.M., Vieira, C.A., Novello, J.C., Oliveira,
B., Giglio, J.R., Marangoni, S., 1998. Amino acid sequence of
piratoxin-I, a myotoxin from Bothrops pirajai snake venom and
its biological activity after alkylation with p-bromophenacyl
bromide. J. Prot. Chem. 17, 713 – 718.
Toyama, M.H., Soares, A.M., Novello, J.C., Oliveira, B., Giglio,
J.R., Marangoni, S., 2000. Amino acid sequence of piratoxin-II,
a Lys49 phospholipase A
2
homologue, from Bothrops pirajai
venom. Biochimie 82, 245–250.
Tsai, I.H., Liu, H.C., Chang, T., 1987. Toxicity domain in
presynaptically toxic phospholipases A
2
of snakes venom.
Biochim. Biophys. Acta. 6 (1), 94– 99.
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883882

Tsai, I.H., Wang, Y.M., 1998. Effect of site directed mutagenesis on
the activity of recombinant trimucrotoxin, a neurotoxic
phospholipase from Trimeresurus mucrosquamatus venom.
Toxicon 36 (11), 1591–1597.
Tsai, I.H., Chen, Y.H., Wang, Y.M., Tu, M.C., Tu, A.T., 2001.
Purification, sequencing, and phylogenetic analyses of novel
Lys-49 phospholipases A(2) from the venoms of rattlesnakes
and other pit vipers. Arch. Biochem. Biophys. 394 (2),
236– 244.
Valentin, E., Lambeau, G., 2000. Increasing molecular diversity of
secreted phospholipases A(2) and their receptors and binding
proteins. Biochim. Biophys. Acta 1488 (1-2), 59 – 70.
Ward, R.J., de Azevedo, W.F. Jr, Arni, R.K., 1998a. At the
interface: crystal structures of phospholipases A
2
. Toxicon 36
(11), 1623– 1633.
Ward, R.J., Rodrigues Alves, A., Rugiero Neto, J., Arni, R.K.,
Casari, J., 1998b. A SequenceSpace analysis of Lys49
phospholipases A
2
: clues towards identification of residues
involved in a novel mechanism of membrane damage and in
myotoxicity. Prot. Engng. 11, 285–294.
Ward, R.J., de Oliveira, A.H., Bortoleto, R.K., Rosa, J.C., Faca,
V.M., Greene, L.J., 2001. Refolding and purification of
Bothropstoxin-I, a Lys49-phospholipase A
2
homologue,
expressed as inclusion bodies in Escherichia coli. Prot. Expr.
Purif. 21 (1), 134–140.
Ward, R.J., Chioato, L., de Oliveira, A.H., Ruller, R., Sa, J.M.,
2002. Active-site mutagenesis of a Lys49-phospholipase A
2
:
biological and membrane-disrupting activities in the absence of
catalysis. Biochem. J. 362, 89–96.
Yamaguchi, Y., Shimohigashi, Y., Chijiwa, T., Nakai, M., Ogawa,
T., Hattori, S., Ohno, M., 2001. Characterization, amino acid
sequence and evolution of edema-inducing, basic phospholipase
A
2
from Trimeresurus flavoviridis venom. Toxicon 39 (7),
1069–1076.
Yang, W.L., Peng, L.S., Zhong, X.F., Wei, J.W., Jiang, X.Y., Ye,
L.T., Zou, L., Tu, H.B., Wu, W.Y., Xu, A., 2003. Functional
expression and characterization of a recombinant phospholipase
A
2
from sea snake Lapemis hardwickii as a soluble protein in E.
coli. Toxicon 41 (6), 713 –721.
Yu, B.Z., Poi, M.J., Ramagopal, U.A., Jain, R., Ramakumar, S.,
Berg, O.G., Tsai, M.D., Sekar, K., Jain, M.K., 2000. Structural
basis of the anionic interface preference and kcat* activation of
pancreatic phospholipase A
2
. Biochemistry 39 (40),
12312–12323.
Yuan, C., Tsai, M., 1999. Pancreatic phospholipase A(2): new
views on old issues. Biochim. Biophys. Acta 1441 (2-3),
215–222.
Zhao, K., Song, S., Lin, Z., Zhou, Y., 1998. Structure of a basic
phospholipase A
2
from Agkistrodon halys Pallas at 2.13 A
˚
resolution. Acta Cryst. D Biol. Cryst. 54 (Pt 4), 510–521.
Zhong, X., Liu, J., Wu, X., Zhou, Y., 2001. Expression, purification
and biochemical characterization of a recombinant phospho-
lipase A
2
, with anticoagulant activity from Agkistrodon halys
Pallas. J. Nat. Toxins 10 (1), 17 – 25.
L. Chioato, R.J. Ward / Toxicon 42 (2003) 869–883 883