401
Biochtreltca et Btophvstca Acta, 1088 ( 1991 ) 401 ~408 ~': 1991 Elsevier Science Publishers BN. 0167-4781/91/$03 50 A D O N I S 016747819100107Z
BBAEXP 92228
Analysis of cDNAs encoding the two subunits of crotoxin, a phospholipase A: neurotoxin from rattlesnake venom: the acidic non enzymatic subunit derives from a phospholipase A :-like precursor Christiane Bouchier ,.2 Jean-Claude Boulain i Cassian Bon ' and Andrg M6nez i / Sere'we de Baln'hin,m' des Prnt~mc~ du l.uhoratotre d'lng~;nterle th'~ Prot~;im's. ("EN Sa('/av. Gifsur Yt'ette and z Untt~ des Ventn~. Untt~; Assot't~;e Instttut PASTI?UR / I N S E R M U285, Instttut Pasteur, Paris (France)
{Received 3 September 1990)
Key words: eDNA; Precursor; Phospholipase As; Neurotoxin; (Snake)
We report gae sequences of three cDNAs encoding the two subunits (CA and CB) of crotoxin, a neurotoxic phospholipa,~ A 2 from the venom of the South-American rattlesnake Crotalus dur~ssus terrificus. CB is a basic and toxic phospholipase A , and CA is an acidic, non toxic and non enzymatic three chain containing protein which enhances the lethal potency of CB. Two ¢DNAs encoding precursors of CB isoforms have been isolated from a eDNA library prepared from one venom gland. Both precursors are made of the same 16 residues signal peptide followed by a polypeptide of 122 amino acid residues. The two mature sequences differ from each other at eight positions and are in good agreement with the previous polypeptide sequence reported for CB. in the case of CA, the eDNA encodes a signal peptide identical to those found in CB precursors, followed by a polypeptide of f22 amino acids clearly homologous to phospholipases A , and including three regions which correspond to the three chains of mature CA. This demonstrates that CA is generated from a phospholipase A z-like precursor, called pro-CA, by the removal of three peptides, leaving unchanged the molecule core cross-linked by disulfide bridges. The 5'-untranslated tracts of cDNAs encoding CA and CB are nearly identical and the 3'-untranslated tracts are very similar, suggesting 'hat the mRNAs encoding the two crotoxin subunits may result from the alternative splicing of a single gene or from the existence of a recent gene conversion. Data have been analysed in light of recent results on other phospholipases A z from difterent origins. Introduclion
Phospholipases A 2 (E.C. 3.1.1.4) are enzymes which specifically catalyze the calcium dependent hydrolysis of fatty acyi bonds at position 2 of 3-sn-phosphoglycerides [2]. They represent a widespread class of enzymes present in bacteria, plants and animals [3-6]. In vertebrates, they are particularly abundant in mammalian pancreas and snake venoms [6-71 where they were first recognized for their digestive function [6].
EMBL Data bank accession numbers of the nucleotide sequences reported in this paper are: CA X12606; CB1 XI2603 and CB2 XI6100. In a preliminary report, we described the partial nuclcotide sequences of two eDNA5 encoding crotoxin subunits [l]. Abbreviations: CA, ~,cidic suhunit of crotoxin, component-A; CB, basic subunit of crotoxin, component-B; SDS, sodium dodet'yl sulfate. Correspondence: A. Mt~nez, Service de Biochimie des Prot,lines, ('i:~N de Saclay, 91191 Gif sur Yvene, France.
A 2 are also present in a variety of tissues and cell types where they play important roles. e.g., the production of arachidonic acid for leukotriene and prostaglandin synthesis [81 and l-alkylglycerophosphorylcholine for platelet activating factor (PAFacether) synthesis [91. the remodeling of phospholipid molecular species by deacylation-reacylation cycles [10], the protection of membranes against peroxydation damages f i l l and immuno-inflammatory reactions [12]. Vertebrates phospholipases A 2 are single chain polypeptides of about 120 amino-acid residues, cross-linked by six or seven disulfide bridges which share a cot lmon overall folding pattern, as judged from both cry tallographic data [131 and circular dichroism analysi,~ [14]. However, they have been classified in two groups, on the basis of differences in primary structure [15,16]. Group I contains mammalian pancreatic phospholipases A 2 and phospholipases A 2 from Elapidae and ttydrophiidae venoms; these enzymes possess a dis,lfidc bridge between cysteines 11 and 77. Group 11 is Phospholipases
402 composed oi phospholipases A, from the venom of V:per:dae and ('rotahdae and non pancreatic mammalian phospholipases A 2 [17,181; they have an extension at the C()OH-terminus and their last amino acid residue is a half-cystine linked to Cys-50. Several cDNAs encoding snake venom phospholipase,; A: behmging to group I have been described previously 119] and two human gcne~ encoding pancreatic and non-pancreatic phospholil,ases A, have been cloned and sequenced
I20-221. A number of phospholipases A 2 from snake venoms are potent neurotoxins which act by blocking neuromuscular transmission at the presynaptic level. Some of them, like ammodytoxin A from Vipera ammodytes ammodvtes venom 123] or agkistrodotoxin from the venom of Agkistrodon hairs (Pallas) [24] (recently renamed Agkistrodon hlomhoffu hrevicaudus [25]), are single chain phospholipases A2, whereas others are composed of several polypeptides, at least one of them being a phospholipase A~ [26]. Crotoxin. the major toxic component of the venom of the South American rattlesnake Crotalus durisstts" terrificus [271 is created by the non covalent association of a basic and weakly toxic phospholipase A, called component-B (CB) with an acidic, non-toxic and non-er~zymatic component-A (CA). CA enhances the pharmacological action and therefore 'he lethal potency of CB [28,29] by preventing the no'~-specific ads(,~r,tion of CB to membrane structures other than its physJ,)logical target [30-32], CB has all the characteristics of a conventional phospholipase A 2 in terms of both primary structure and catalytic activity [33]. In contrast, CA is composed of three peptides (namely chains A, B and (') cross-linked by disulfide bridges that are similar in sequence to three non adjacent regi~_>~s of phospholipases A 2. Therefore, it has been suggested that CA might be generated by the proteolv6c cleavage of a phospholipase A,-like precursor [34], but so far no experimental evidence has confirmed this proposal. It has now been shown that the venom of ('rotalus durissus terrifteus contains several isoforms of crotoxin which differ only slightly from each other in their amino acid composition [35,361. The presen~ paper reports the cloning and sequence analysis of a cDNA encoding a C~, precursor and of two cDNAs encoding precursors of two CB isoforms. Our data unambiguously show that CA originates from the cleavage of a phospholipase A~-like polypeptide. The sequence organization of the cDNAs encoding the crotoxin subunits has been compared with those encoding other snake phospholipases A: [37-39] and the phylogenetic origin of the crotoxin heterodimer is discussed. Data presented in ',his paper can presumably be extended to the crotoxin-like compounds found in other rattlesnake venoms, i.e., mojave toxin from Crotal,~s scululatus st'ltlulutu~' and concolor toxin from Crotalus t'irtdis con, ,/or [40,41],
Materials and Methods
Mater:al~ lave ('rotahts durlssus terri/tcus snake (from SaoPaulo state, Brazil) was kindly donated by Mr. Y. I)oljansky (Latoxan, Rosans, France). Oligonucleotide probes were synthesized by Dr. J. Igolen (Unit~ de Chin,ie Organique, Institut Pasteur, Paris). [~sS]dCTP was provided by Amersham (Les Ulis, France). Sequencing kits were from United States Biochemical Corporation -USB- (Cleveland, OH, USA). Modification and restriction enzymes were from Gibeo/Bethesda REsearch Laboratories (Cergy Pontoise, France), Boehringer Mannheim (Mannheim, FRG) or Appligene (lllkirch. France). Chemicals were from Prolabo (Paris, France). Method~ Extraction of n:RNA and eDNA synthesis. Crotalus durissus terrificus snake was milked 2 days before killing by decapitation. The head was immediately frozen in liquid nitrogen and stored at - 8 0 ° C until used. A venom gland was isolated from the frozen snake head and dissected. Total RNAs were extracted according to a modification of the method of Zitomer and Hall [42]. In brief, the tissues were cut and crushed in p h e n o l / chloroform/isoamyl alcohol (50:48 : 2; v/v) solution; an equal volume of sodium dodecyl sulfate ('qDS) medium (50 mM Tris-HCI (pH 9.0), !00 mM Na(,L 10 mM EDTA, 5% SDS) was added and the mixture was homogenized with an Uhraturax apparatus. The aqueous phase was removed, extracted again with phenol/ chloroform/isoamyl alcohol (50:48: 2; v/v) and twice with chloroform/isoamyl alcohol (48:2; v/v). RNAs were precipitated overnight at - 2 0 ° C using 0.3 M sodium acetate and 2.5 vol. of absolute ethanol. Poly(A +) RNA were isolated by chromatography on oligtXdT) cellulose column [43]. About 8 p.g of purified poly(A ~ ) mRNA were obtained from a single venomous gland. Synthesis of cDNA was performed according to the "one tube' method described by Gerard et al. [44]. In a first step, mRNAs were hybridized with a synthetic oligo (dT) 12-18 primer and copied as complementary DNAs using Moioney murine ~eukeraia virus reverse transcriptase (M-MLV RT). The ~erc,nd strand eDNA was synthesized according to a modification of the method of Okayama and Berg [45] and Gubler and Hoffman [46], using RNAse H and DNA polymerase I. Double-strand cDNAs were dC-tailed using terminal transferase and inserted into the dG-tailed Pstl site of pBR322 plasmid [47]. The recombinant plasmids were used to transform Escherichia coli strain MCI061 [48], according to the m6thod of Hanahan [49]. Transformation efficiency was 2.5 • 1 0 4 clones/#g of tailed eDNA.
403
CIo,ung and sequencing Colonies containing cDNAs encoding crotoxin subunits were screened using as probes two synthetic oligonucleotides which were selected to detect clones containing CA or CB subunits (see below). '"P-labelling of these probes was performed according to Ricca et al.
gland. Specific oligonudeotidic probes, established from the published protein sequences [33 341 were used to screen clones encoding CA and CB subunits of crotoxin. The probes corresponded to peptides located in the N-terminal part of the chain C of CA (GIn-Phe-SerPro-Glu-Asn) and in the N-terminal part of CB (LeuLeu-GIn-Phe-Asn) where a minimal similarity was observed betw~'en the two subunits. Positive clones were detected with a frequency of about 10% for each type of probe. A single Pstl restriction site was found within cDNAs encoding CA and CB, yielding two fragments after an appropriate cleavage. Fig. 1 shows the nucleotide sequences of the cDNAs encoding the precursors of CA and of two isoforms of CB, called CB1 and CB2. cDNAs encoding CA and CBI possess 736 and 725 base pairs, respectively. The eDNA encoding CB2 nossesses at least 626 base pairs, the smaller fragment containing less than 100 base pairs having not been isolated from agarose gel. The three cDNAs have the same overall organization. Firstly, they possess a large and similar 5'-untranslated region (nucleotides 1 to 198) differing in the three sequences by only one nucleotide (the sequence of CB1 has a single base deletion at position 158). Secondly, the open reading frames are characterized by a lower similarity (about 70%) and encode 138 amino acid polypeptides containing an identical signal peptide of 16 amino acid residues followed by 122 amino acids. Thirdly, the Y-untranslated regions of the cDNAs contain a consensus polyadenylation signal (AATAAA) upstream to
1501. Probe 1: CB (15-mer)
3'
Leu 2 Leu 3 G i n 4 Phe s Asn 6 GAG GAG GTT AAG TTG 5 '
T
Probe2: CA C-chain(18-mer) Glnl Phe 2 Set3 Pro4 Glu s Asn 6 3'
GTT
AAG
TCG
GGG
CTT
5'
TTG
T
Plasmid DNAs from positive clones were isolated and inserted cDNAs were subcloned into M l 3 m p l 8 and M13mp19 vectors [51]. They were sequenced, on both strands, according to the method ot Sanger et a l
[521. Results
Cloning, nucleotide sequence and overall organization of cDNAs encoding crotoxin subunits A eDNA library was constructed with n R N A s extracted L'om a single Crotalus durissus terrificus venom eDNA
C&
eDNA eDNA
CBI CB2
.
.
.
.
. .
. .
. .
. .
. .
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. .
. .
. .
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
.
.
.
.
. .
. .
. .
. .
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
~
.......................................................... .
.
.
.
.
.
.
.
.
AGG
.
.
.
.
.
.
.
.
.
.
.
.
.
GCT .
.
.
.
.
.
.
.
.
.
.
CTC .
.
.
.
.
TGG
.
.
.
.
.
.
.
.
.
.
.
.
.
ATA .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
CA CB1
TGG CCA C .....
CB1 CB2
CA
.
.
.
ACA
.
.
ACA
.
.
.
.
GAC
.
.
GTC
.
.
.
TAC
.A. TGG .., A .... g . . . . . . . . . . . . . . . .
ATT
TGT
GAG
TO.-" G A C
.
.
ACC
.
.
A'rG RAA ATT Gf'G GGG AGA A:;1 .A. TTT GAG A.A A.. .A AC . . . . . . . . . . . . . . . . . . . . . . AGC CT
.
.
.
GCC
CA CB1 CB2
CAG GAC GCC A . . . . . . . . .
.
GTG
.
.
.
.
.
.
.
.
G'TS T T G .
.
.
.
.
.
.
AI'T
GG'['
.
.
.
,C .... G .
.
.
.
.
.
.
CTG .
.
.
.
.
TAC
.
.
.
.
AGA
.
.
.
.
.
.
.
.
.
.
.
.
.
.
GAG
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
T
GAG .
.
.
.
.
.
.
~GG
.
.
.
.
.
.
.
;;A
.
.
.
.
.
.
.
.
(}AA AF'~ G T ¢
.
.
.
.
.
.
.
.
GAC
ACA ';
AAG
TTC
TCG
CCC
GAA
AAT
.A.
C..
GA.
TC.
CG .... AG .... . . . . . . . . . . A
CA
CAATCGTAGTI
CBI
• • "A ............. the
GTGTTACTCTATTATTCT
nucleotide deletions
CAG
GCA
CA
of
TGC
GCC
CB1 CB2
Nucleot|&
A f ~
have
GCA
GGG .
.
ATC
TGC
.
.
.
.
.
.
TTC
r'I;A A A T
.
';.,ph~,hp~c A: 1701+pho,,phohpa~,e A 2 from human ~,ynowal fluid [21 I, membrane associated phosphohpase A 2 from rat spleen 1711. CA, CB1 and t'l~2 from ('rotaht~ durt~*ui rerrtftczl.~' venom, ammodytoxin C from Vq)era ammodYte.~ ammot#tcs venom 1391 and phospholipases A z from Latwauda lattcaudata and Notethis scutatzt~ ~t'utat~tr venoms 138,6zlI. The sequences of potential signal peptides a n d / o r pro-peptides were compared. The sequence of the actual propeptide of rat pancreas phospholipase A z is underlined.
It is well known that mammalian pancreatic enzymes possess an N-terminal propeptide extension of seven residues which is cleaved by trypsin in the duodenum tractus. As shown in Fig. 2, such a propeptide does not exist in phospholipases A 2 from Viperidae venoms. Although, the N-terminal peptidic extensions of phospholipases A 2 from Elapidae venoms are large enough to comprise a signal peptide plus a potential N-terminal propcptide, there is, however, no similarity with the sequence of the propeptides of pancreatic phospholipases A 2 (Fig. 2).
the poly(A ~ ) tract. These Y-untranslated regions are highly similar (90%) in full-length cDNAs encoding CA and ('BI.
Signal .~equence of crotoxin subttnit precursors Thc deduced amino acid sequences of cDNAs encoding ('B1, CB2 and CA are presented in Figs. 2 to 4. The three precursors possess an identical signal peptide of 16 amino acid residues which differs by only three residues from that of the precursor of ammodytoxin C [39 I. a single neurotoxic phospholipase A 2 from the venom of Vipera ammodytes an:,nodvtes (Fig. 2). We also noticed some similarities with the slgt~al peptides of phospholipases A 2 from mammalian pancreas. In contrast, the signal sequences of mammalian phospholipases A 2 from a non pancreatic origin, as well as those of phospholipases A 2 from Elapidae venoms, are quite different (Fig. 2),
i CB
I
C8
~'
Sequences of the CBI and CB2 isoforms As shown in Fig. 3, the amino acid sequences deduced from cDNAs encoding CB1 and CB2 differ from each other by eight residues at positions 1, 18, 33, 65, 82, 105, 106, 117. This underlines the great similarity of the crotoxin isoforms, previously noticed by comparing
10
HL
LQF'NKN
S
.
.
20
I K F E T R
.
.
.
.
.
.
.
~N&
.
.
V
.
40
30
X P F Y A F .
.
¥ G C Y C G W G G R G R P K D A T .
.
.
.
.
.
.
.
O
.
.
.
.
.
.
.
-
-
-
G
-
Q
~t$ At,a et ~,I
198~
H
Agkklnodolo1~rl
.
N
.
.
.
.
.
.
.
.
.
.
.
.
F:
5O
-
-
I
G
.
.
.
.
.
. . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
R
V
N
Y
G
•
i00 V
A
.
.
.
~
E
.
.
C
L
.
.
.
R
~
S
.
.
.
. F
.
.
S
.
T
Y
K
.
H E
. N
¥ D
N
N
Y
.
H
F
R
S
.
Y
.
.
.
.
.
.
.
.
.
.
G
.
.
S
. K R
-
R
-
-
-
R -
Q
-
K
-
-
80
L K S G ¥ .
90
I T C G K G T W C E .
.
KE
II0
L
.
70 Y P Y S
. . . . . . . . . . . . . . . . . . . . . . . . .
.
6O
D R C C F V H D C C Y G K L A K C N T K H D I
.
.
.
.
.
.
.
. .
.
. .
.
.
EQI
. . N
.
CIZCDR
K
. . . . . . .
E
. . . . . . . . .
.
.
.
.
.
.
.
120 P
D
.
~
.
R
C
~
G
.
P
$
~
G
E T C
-
G i~
K
T
E
T
Ihg 3 Similarity ~ co Tp ment~B snform~ and agkislrodotoxin. Deduced polypeptlde sequences of CBI and CB2 hay, .'en aligned with those of Ihe mature CB reported by Aird et a l 1331 and agkistrodotoxin [24]. The sequences have been numbered from the first amino acid of the presumptive mature proteins. Dashes indicate that the amino at: t residue is identical to the immediately one above.
405 1
•
*
1o
•
Pro-CA
S ,(,V E F leT
CA mature
"
~ . . . . . . . . . .
Acidic rno||ve toxin
.
.
.
.
.
.
.
.
50 DRC
erIC
ore
CF
C F •
I
LTGC
HDCCYAK HDC
[,TGCN
CYA
;: ; , , , , , .
A l l
i
ei
'
.
.
.
K LT
~ ". .
.
.
.
.
D'/
P TT
GC
Y TY
RQID
DV
¥
RNSHDT
~ ~ o o ~
A A A Z C F I ~ S N N T yo / ~
30
* *
*
•
G QG
W P Q D A 5
Y C G &GGQGii
P Q DAS
80 G [
I VC
GE
D GI
I VCGGDD
40
*
Q DAS
90
O D PCG
T~
I C E C DK
D D P C G T Q I PCGTQ
C :
C D ~
1C~C
DK
120
o,,,..
~ o,,.. $ o P e
OpiF 5 p • l
~ ~ ~ o ~ o
* * *
Y C G AGG@GNP
pie E D G IC I V C G I < 'pEE
110
o, ;.>,.
Biochtreltca et Btophvstca Acta, 1088 ( 1991 ) 401 ~408 ~': 1991 Elsevier Science Publishers BN. 0167-4781/91/$03 50 A D O N I S 016747819100107Z
BBAEXP 92228
Analysis of cDNAs encoding the two subunits of crotoxin, a phospholipase A: neurotoxin from rattlesnake venom: the acidic non enzymatic subunit derives from a phospholipase A :-like precursor Christiane Bouchier ,.2 Jean-Claude Boulain i Cassian Bon ' and Andrg M6nez i / Sere'we de Baln'hin,m' des Prnt~mc~ du l.uhoratotre d'lng~;nterle th'~ Prot~;im's. ("EN Sa('/av. Gifsur Yt'ette and z Untt~ des Ventn~. Untt~; Assot't~;e Instttut PASTI?UR / I N S E R M U285, Instttut Pasteur, Paris (France)
{Received 3 September 1990)
Key words: eDNA; Precursor; Phospholipase As; Neurotoxin; (Snake)
We report gae sequences of three cDNAs encoding the two subunits (CA and CB) of crotoxin, a neurotoxic phospholipa,~ A 2 from the venom of the South-American rattlesnake Crotalus dur~ssus terrificus. CB is a basic and toxic phospholipase A , and CA is an acidic, non toxic and non enzymatic three chain containing protein which enhances the lethal potency of CB. Two ¢DNAs encoding precursors of CB isoforms have been isolated from a eDNA library prepared from one venom gland. Both precursors are made of the same 16 residues signal peptide followed by a polypeptide of 122 amino acid residues. The two mature sequences differ from each other at eight positions and are in good agreement with the previous polypeptide sequence reported for CB. in the case of CA, the eDNA encodes a signal peptide identical to those found in CB precursors, followed by a polypeptide of f22 amino acids clearly homologous to phospholipases A , and including three regions which correspond to the three chains of mature CA. This demonstrates that CA is generated from a phospholipase A z-like precursor, called pro-CA, by the removal of three peptides, leaving unchanged the molecule core cross-linked by disulfide bridges. The 5'-untranslated tracts of cDNAs encoding CA and CB are nearly identical and the 3'-untranslated tracts are very similar, suggesting 'hat the mRNAs encoding the two crotoxin subunits may result from the alternative splicing of a single gene or from the existence of a recent gene conversion. Data have been analysed in light of recent results on other phospholipases A z from difterent origins. Introduclion
Phospholipases A 2 (E.C. 3.1.1.4) are enzymes which specifically catalyze the calcium dependent hydrolysis of fatty acyi bonds at position 2 of 3-sn-phosphoglycerides [2]. They represent a widespread class of enzymes present in bacteria, plants and animals [3-6]. In vertebrates, they are particularly abundant in mammalian pancreas and snake venoms [6-71 where they were first recognized for their digestive function [6].
EMBL Data bank accession numbers of the nucleotide sequences reported in this paper are: CA X12606; CB1 XI2603 and CB2 XI6100. In a preliminary report, we described the partial nuclcotide sequences of two eDNA5 encoding crotoxin subunits [l]. Abbreviations: CA, ~,cidic suhunit of crotoxin, component-A; CB, basic subunit of crotoxin, component-B; SDS, sodium dodet'yl sulfate. Correspondence: A. Mt~nez, Service de Biochimie des Prot,lines, ('i:~N de Saclay, 91191 Gif sur Yvene, France.
A 2 are also present in a variety of tissues and cell types where they play important roles. e.g., the production of arachidonic acid for leukotriene and prostaglandin synthesis [81 and l-alkylglycerophosphorylcholine for platelet activating factor (PAFacether) synthesis [91. the remodeling of phospholipid molecular species by deacylation-reacylation cycles [10], the protection of membranes against peroxydation damages f i l l and immuno-inflammatory reactions [12]. Vertebrates phospholipases A 2 are single chain polypeptides of about 120 amino-acid residues, cross-linked by six or seven disulfide bridges which share a cot lmon overall folding pattern, as judged from both cry tallographic data [131 and circular dichroism analysi,~ [14]. However, they have been classified in two groups, on the basis of differences in primary structure [15,16]. Group I contains mammalian pancreatic phospholipases A 2 and phospholipases A 2 from Elapidae and ttydrophiidae venoms; these enzymes possess a dis,lfidc bridge between cysteines 11 and 77. Group 11 is Phospholipases
402 composed oi phospholipases A, from the venom of V:per:dae and ('rotahdae and non pancreatic mammalian phospholipases A 2 [17,181; they have an extension at the C()OH-terminus and their last amino acid residue is a half-cystine linked to Cys-50. Several cDNAs encoding snake venom phospholipase,; A: behmging to group I have been described previously 119] and two human gcne~ encoding pancreatic and non-pancreatic phospholil,ases A, have been cloned and sequenced
I20-221. A number of phospholipases A 2 from snake venoms are potent neurotoxins which act by blocking neuromuscular transmission at the presynaptic level. Some of them, like ammodytoxin A from Vipera ammodytes ammodvtes venom 123] or agkistrodotoxin from the venom of Agkistrodon hairs (Pallas) [24] (recently renamed Agkistrodon hlomhoffu hrevicaudus [25]), are single chain phospholipases A2, whereas others are composed of several polypeptides, at least one of them being a phospholipase A~ [26]. Crotoxin. the major toxic component of the venom of the South American rattlesnake Crotalus durisstts" terrificus [271 is created by the non covalent association of a basic and weakly toxic phospholipase A, called component-B (CB) with an acidic, non-toxic and non-er~zymatic component-A (CA). CA enhances the pharmacological action and therefore 'he lethal potency of CB [28,29] by preventing the no'~-specific ads(,~r,tion of CB to membrane structures other than its physJ,)logical target [30-32], CB has all the characteristics of a conventional phospholipase A 2 in terms of both primary structure and catalytic activity [33]. In contrast, CA is composed of three peptides (namely chains A, B and (') cross-linked by disulfide bridges that are similar in sequence to three non adjacent regi~_>~s of phospholipases A 2. Therefore, it has been suggested that CA might be generated by the proteolv6c cleavage of a phospholipase A,-like precursor [34], but so far no experimental evidence has confirmed this proposal. It has now been shown that the venom of ('rotalus durissus terrifteus contains several isoforms of crotoxin which differ only slightly from each other in their amino acid composition [35,361. The presen~ paper reports the cloning and sequence analysis of a cDNA encoding a C~, precursor and of two cDNAs encoding precursors of two CB isoforms. Our data unambiguously show that CA originates from the cleavage of a phospholipase A~-like polypeptide. The sequence organization of the cDNAs encoding the crotoxin subunits has been compared with those encoding other snake phospholipases A: [37-39] and the phylogenetic origin of the crotoxin heterodimer is discussed. Data presented in ',his paper can presumably be extended to the crotoxin-like compounds found in other rattlesnake venoms, i.e., mojave toxin from Crotal,~s scululatus st'ltlulutu~' and concolor toxin from Crotalus t'irtdis con, ,/or [40,41],
Materials and Methods
Mater:al~ lave ('rotahts durlssus terri/tcus snake (from SaoPaulo state, Brazil) was kindly donated by Mr. Y. I)oljansky (Latoxan, Rosans, France). Oligonucleotide probes were synthesized by Dr. J. Igolen (Unit~ de Chin,ie Organique, Institut Pasteur, Paris). [~sS]dCTP was provided by Amersham (Les Ulis, France). Sequencing kits were from United States Biochemical Corporation -USB- (Cleveland, OH, USA). Modification and restriction enzymes were from Gibeo/Bethesda REsearch Laboratories (Cergy Pontoise, France), Boehringer Mannheim (Mannheim, FRG) or Appligene (lllkirch. France). Chemicals were from Prolabo (Paris, France). Method~ Extraction of n:RNA and eDNA synthesis. Crotalus durissus terrificus snake was milked 2 days before killing by decapitation. The head was immediately frozen in liquid nitrogen and stored at - 8 0 ° C until used. A venom gland was isolated from the frozen snake head and dissected. Total RNAs were extracted according to a modification of the method of Zitomer and Hall [42]. In brief, the tissues were cut and crushed in p h e n o l / chloroform/isoamyl alcohol (50:48 : 2; v/v) solution; an equal volume of sodium dodecyl sulfate ('qDS) medium (50 mM Tris-HCI (pH 9.0), !00 mM Na(,L 10 mM EDTA, 5% SDS) was added and the mixture was homogenized with an Uhraturax apparatus. The aqueous phase was removed, extracted again with phenol/ chloroform/isoamyl alcohol (50:48: 2; v/v) and twice with chloroform/isoamyl alcohol (48:2; v/v). RNAs were precipitated overnight at - 2 0 ° C using 0.3 M sodium acetate and 2.5 vol. of absolute ethanol. Poly(A +) RNA were isolated by chromatography on oligtXdT) cellulose column [43]. About 8 p.g of purified poly(A ~ ) mRNA were obtained from a single venomous gland. Synthesis of cDNA was performed according to the "one tube' method described by Gerard et al. [44]. In a first step, mRNAs were hybridized with a synthetic oligo (dT) 12-18 primer and copied as complementary DNAs using Moioney murine ~eukeraia virus reverse transcriptase (M-MLV RT). The ~erc,nd strand eDNA was synthesized according to a modification of the method of Okayama and Berg [45] and Gubler and Hoffman [46], using RNAse H and DNA polymerase I. Double-strand cDNAs were dC-tailed using terminal transferase and inserted into the dG-tailed Pstl site of pBR322 plasmid [47]. The recombinant plasmids were used to transform Escherichia coli strain MCI061 [48], according to the m6thod of Hanahan [49]. Transformation efficiency was 2.5 • 1 0 4 clones/#g of tailed eDNA.
403
CIo,ung and sequencing Colonies containing cDNAs encoding crotoxin subunits were screened using as probes two synthetic oligonucleotides which were selected to detect clones containing CA or CB subunits (see below). '"P-labelling of these probes was performed according to Ricca et al.
gland. Specific oligonudeotidic probes, established from the published protein sequences [33 341 were used to screen clones encoding CA and CB subunits of crotoxin. The probes corresponded to peptides located in the N-terminal part of the chain C of CA (GIn-Phe-SerPro-Glu-Asn) and in the N-terminal part of CB (LeuLeu-GIn-Phe-Asn) where a minimal similarity was observed betw~'en the two subunits. Positive clones were detected with a frequency of about 10% for each type of probe. A single Pstl restriction site was found within cDNAs encoding CA and CB, yielding two fragments after an appropriate cleavage. Fig. 1 shows the nucleotide sequences of the cDNAs encoding the precursors of CA and of two isoforms of CB, called CB1 and CB2. cDNAs encoding CA and CBI possess 736 and 725 base pairs, respectively. The eDNA encoding CB2 nossesses at least 626 base pairs, the smaller fragment containing less than 100 base pairs having not been isolated from agarose gel. The three cDNAs have the same overall organization. Firstly, they possess a large and similar 5'-untranslated region (nucleotides 1 to 198) differing in the three sequences by only one nucleotide (the sequence of CB1 has a single base deletion at position 158). Secondly, the open reading frames are characterized by a lower similarity (about 70%) and encode 138 amino acid polypeptides containing an identical signal peptide of 16 amino acid residues followed by 122 amino acids. Thirdly, the Y-untranslated regions of the cDNAs contain a consensus polyadenylation signal (AATAAA) upstream to
1501. Probe 1: CB (15-mer)
3'
Leu 2 Leu 3 G i n 4 Phe s Asn 6 GAG GAG GTT AAG TTG 5 '
T
Probe2: CA C-chain(18-mer) Glnl Phe 2 Set3 Pro4 Glu s Asn 6 3'
GTT
AAG
TCG
GGG
CTT
5'
TTG
T
Plasmid DNAs from positive clones were isolated and inserted cDNAs were subcloned into M l 3 m p l 8 and M13mp19 vectors [51]. They were sequenced, on both strands, according to the method ot Sanger et a l
[521. Results
Cloning, nucleotide sequence and overall organization of cDNAs encoding crotoxin subunits A eDNA library was constructed with n R N A s extracted L'om a single Crotalus durissus terrificus venom eDNA
C&
eDNA eDNA
CBI CB2
.
.
.
.
. .
. .
. .
. .
. .
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. .
. .
. .
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. .
.
.
.
.
.
.
.
.
.
.
.
. .
. .
. .
. .
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
~
.......................................................... .
.
.
.
.
.
.
.
.
AGG
.
.
.
.
.
.
.
.
.
.
.
.
.
GCT .
.
.
.
.
.
.
.
.
.
.
CTC .
.
.
.
.
TGG
.
.
.
.
.
.
.
.
.
.
.
.
.
ATA .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
CA CB1
TGG CCA C .....
CB1 CB2
CA
.
.
.
ACA
.
.
ACA
.
.
.
.
GAC
.
.
GTC
.
.
.
TAC
.A. TGG .., A .... g . . . . . . . . . . . . . . . .
ATT
TGT
GAG
TO.-" G A C
.
.
ACC
.
.
A'rG RAA ATT Gf'G GGG AGA A:;1 .A. TTT GAG A.A A.. .A AC . . . . . . . . . . . . . . . . . . . . . . AGC CT
.
.
.
GCC
CA CB1 CB2
CAG GAC GCC A . . . . . . . . .
.
GTG
.
.
.
.
.
.
.
.
G'TS T T G .
.
.
.
.
.
.
AI'T
GG'['
.
.
.
,C .... G .
.
.
.
.
.
.
CTG .
.
.
.
.
TAC
.
.
.
.
AGA
.
.
.
.
.
.
.
.
.
.
.
.
.
.
GAG
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
T
GAG .
.
.
.
.
.
.
~GG
.
.
.
.
.
.
.
;;A
.
.
.
.
.
.
.
.
(}AA AF'~ G T ¢
.
.
.
.
.
.
.
.
GAC
ACA ';
AAG
TTC
TCG
CCC
GAA
AAT
.A.
C..
GA.
TC.
CG .... AG .... . . . . . . . . . . A
CA
CAATCGTAGTI
CBI
• • "A ............. the
GTGTTACTCTATTATTCT
nucleotide deletions
CAG
GCA
CA
of
TGC
GCC
CB1 CB2
Nucleot|&
A f ~
have
GCA
GGG .
.
ATC
TGC
.
.
.
.
.
.
TTC
r'I;A A A T
.
';.,ph~,hp~c A: 1701+pho,,phohpa~,e A 2 from human ~,ynowal fluid [21 I, membrane associated phosphohpase A 2 from rat spleen 1711. CA, CB1 and t'l~2 from ('rotaht~ durt~*ui rerrtftczl.~' venom, ammodytoxin C from Vq)era ammodYte.~ ammot#tcs venom 1391 and phospholipases A z from Latwauda lattcaudata and Notethis scutatzt~ ~t'utat~tr venoms 138,6zlI. The sequences of potential signal peptides a n d / o r pro-peptides were compared. The sequence of the actual propeptide of rat pancreas phospholipase A z is underlined.
It is well known that mammalian pancreatic enzymes possess an N-terminal propeptide extension of seven residues which is cleaved by trypsin in the duodenum tractus. As shown in Fig. 2, such a propeptide does not exist in phospholipases A 2 from Viperidae venoms. Although, the N-terminal peptidic extensions of phospholipases A 2 from Elapidae venoms are large enough to comprise a signal peptide plus a potential N-terminal propcptide, there is, however, no similarity with the sequence of the propeptides of pancreatic phospholipases A 2 (Fig. 2).
the poly(A ~ ) tract. These Y-untranslated regions are highly similar (90%) in full-length cDNAs encoding CA and ('BI.
Signal .~equence of crotoxin subttnit precursors Thc deduced amino acid sequences of cDNAs encoding ('B1, CB2 and CA are presented in Figs. 2 to 4. The three precursors possess an identical signal peptide of 16 amino acid residues which differs by only three residues from that of the precursor of ammodytoxin C [39 I. a single neurotoxic phospholipase A 2 from the venom of Vipera ammodytes an:,nodvtes (Fig. 2). We also noticed some similarities with the slgt~al peptides of phospholipases A 2 from mammalian pancreas. In contrast, the signal sequences of mammalian phospholipases A 2 from a non pancreatic origin, as well as those of phospholipases A 2 from Elapidae venoms, are quite different (Fig. 2),
i CB
I
C8
~'
Sequences of the CBI and CB2 isoforms As shown in Fig. 3, the amino acid sequences deduced from cDNAs encoding CB1 and CB2 differ from each other by eight residues at positions 1, 18, 33, 65, 82, 105, 106, 117. This underlines the great similarity of the crotoxin isoforms, previously noticed by comparing
10
HL
LQF'NKN
S
.
.
20
I K F E T R
.
.
.
.
.
.
.
~N&
.
.
V
.
40
30
X P F Y A F .
.
¥ G C Y C G W G G R G R P K D A T .
.
.
.
.
.
.
.
O
.
.
.
.
.
.
.
-
-
-
G
-
Q
~t$ At,a et ~,I
198~
H
Agkklnodolo1~rl
.
N
.
.
.
.
.
.
.
.
.
.
.
.
F:
5O
-
-
I
G
.
.
.
.
.
. . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
R
V
N
Y
G
•
i00 V
A
.
.
.
~
E
.
.
C
L
.
.
.
R
~
S
.
.
.
. F
.
.
S
.
T
Y
K
.
H E
. N
¥ D
N
N
Y
.
H
F
R
S
.
Y
.
.
.
.
.
.
.
.
.
.
G
.
.
S
. K R
-
R
-
-
-
R -
Q
-
K
-
-
80
L K S G ¥ .
90
I T C G K G T W C E .
.
KE
II0
L
.
70 Y P Y S
. . . . . . . . . . . . . . . . . . . . . . . . .
.
6O
D R C C F V H D C C Y G K L A K C N T K H D I
.
.
.
.
.
.
.
. .
.
. .
.
.
EQI
. . N
.
CIZCDR
K
. . . . . . .
E
. . . . . . . . .
.
.
.
.
.
.
.
120 P
D
.
~
.
R
C
~
G
.
P
$
~
G
E T C
-
G i~
K
T
E
T
Ihg 3 Similarity ~ co Tp ment~B snform~ and agkislrodotoxin. Deduced polypeptlde sequences of CBI and CB2 hay, .'en aligned with those of Ihe mature CB reported by Aird et a l 1331 and agkistrodotoxin [24]. The sequences have been numbered from the first amino acid of the presumptive mature proteins. Dashes indicate that the amino at: t residue is identical to the immediately one above.
405 1
•
*
1o
•
Pro-CA
S ,(,V E F leT
CA mature
"
~ . . . . . . . . . .
Acidic rno||ve toxin
.
.
.
.
.
.
.
.
50 DRC
erIC
ore
CF
C F •
I
LTGC
HDCCYAK HDC
[,TGCN
CYA
;: ; , , , , , .
A l l
i
ei
'
.
.
.
K LT
~ ". .
.
.
.
.
D'/
P TT
GC
Y TY
RQID
DV
¥
RNSHDT
~ ~ o o ~
A A A Z C F I ~ S N N T yo / ~
30
* *
*
•
G QG
W P Q D A 5
Y C G &GGQGii
P Q DAS
80 G [
I VC
GE
D GI
I VCGGDD
40
*
Q DAS
90
O D PCG
T~
I C E C DK
D D P C G T Q I PCGTQ
C :
C D ~
1C~C
DK
120
o,,,..
~ o,,.. $ o P e
OpiF 5 p • l
~ ~ ~ o ~ o
* * *
Y C G AGG@GNP
pie E D G IC I V C G I < 'pEE
110
o, ;.>,.
