Molecular cloning of a g-phospholipase A2 inhibitor

from Lachesis muta muta (the bushmaster snake)

C.L. Fortes-Dias*, C.J. Barcellos, M.I. Estevao-Costa

Centro de Pesquisa e Desenvolvimento, Lab. Biologia Molecular, Fundacao Ezequiel Dias (FUNED),

Rua Conde Pereira Carneiro 80, Belo Horizonte, MG 30510 010, Brazil

Received 16 December 2002; accepted 17 March 2003

Abstract

Several endogenous phospholipase A2 inhibitors (PLIs) have been purified from the blood plasma of a number of snake

species and are classified into three classes (a, b and g) according to their structure and specificity. In the present study we have

cloned transcripts of a protein homologous to CNF, a gPLI present in Crotalus durissus terrificus plasma, that is encoded in the

liver of Lachesis muta muta (the bushmaster snake), a species evolutionarily related to Crotalus. The cDNA sequences code for

two isoforms of a 200-residue protein including a 19-residue signal peptide followed by 181 amino acid residues in the mature

form and a putative N-linked carbohydrate site. The deduced primary structures and some properties of those new proteins were

compared to those of CNF. Multiple alignment was performed with the aminoacid sequences of all the gPLIs described so far

and this used in the construction of a phylogenetic tree.

q 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Phospholipase A2 inhibitors; g-Phospholipase A2 inhibitor; Phospholipase A2 inhibitorg; Phospholipase A2; Lachesis

1. Introduction

Phospholipases A2 (E.C. 3.1.1.4; PLA2) are commonly

occurring enzymes that catalyze specifically the hydrolysis

of the sn-2 acylester bond in 1,2-diacyl-3-sn-phosphogly-

cerides in a calcium dependent reaction. According to their

biochemical characteristics and cellular origin, these

enzymes have been classified as cytosolic (cPLA2) or

secretory (sPLA2) phospholipases A2 (Glaser et al., 1993).

Snake venoms are a well known source of sPLA2 which are

responsible for a wide variety of distinct pharmacological

effects such as neurotoxicity, cardiotoxicity, myonecrosis,

anticoagulant and platelet effects upon envenoming (Kini

and Evans, 1989).

In the last decade, purification of a growing number

of endogenous PLA2 inhibitors (PLIs) from the blood

plasma of snake species from all continents has been

reported (Dunn and Broady, 2001; Faure, 2000; For-

tes-Dias, 2002 for recent reviews). The main role

ascribed to PLIs has been a physiological protection of

snakes against an accidental leaking of venom from their

venom glands. In South American snakes, a PLI was

isolated and characterized from the blood plasma of the

tropical rattlesnake, Crotalus durissus terrificus (Crotali-

nae, Viperidae) (Fortes-Dias et al., 1991, 1994; Perales

et al., 1995). That protein, named CNF, is a member of a

large class of PLA2 inhibitors, recently grouped as gPLIs

(Ohkura et al., 1997).

CNF is a 160 kDa glycoprotein composed of 6–8

identical subunits of approximately 24 kDa with a broad

range of PLA2 specificity like other gPLIs. CNF acts by

replacing the subunit CA in the crotoxin complex (CACB)

from C. d. terrificus venom, inhibiting the PLA2 activity of

the subunit CB. Besides CB, CNF is able to inhibit basic and

acidic PLA2s from L. m. muta (Fortes-Dias et al., 1999) and

B. jararacussu (Fortes-Dias, unpublished data) venoms,

even more efficiently.

0041-0101/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.

doi:10.1016/S0041-0101(03)00073-4

Toxicon 41 (2003) 909–917

www.elsevier.com/locate/toxicon

* Corresponding author. Fax: þ55-31-3371-1753.

E-mail address: consuelo@funed.mg.gov.br (C.L. Fortes-Dias).

Evolutionary studies on the Viperidae family have

demonstrated that Lachesis, comprising a single species

Lachesis muta, is the genus most closely related to Crotalus

(Brattstrom, 1964). However, the venom from L. muta

venom is much more complex than that from C. d. terrificus,

being composed of a variety of biologically active

components including high levels of phospholipases A2

(Valiente et al., 1992; Fuly et al., 1993, 1997; Fortes-Dias

et al., 1999).

Taking into consideration: (1) The closer evolutionary

relationship between Crotalus and Lachesis; (2)

The capability of CNF to inhibit L. muta PLA2; (3) The

significant levels of PLA2 in L. muta venom and (4) The

important physiological role of PLIs in snakes, we decided

to search for the presence of a CNF-homologue transcript in

L. m. muta liver.

2. Materials and methods

L. m. muta liver. The liver of one adult specimen of

L. m. muta was removed immediately after its death from

natural causes. That snake had been captured in Costa

Rica, bred at the Dallas Zoo (Texas, USA) and donated

to the Herpetarium of Fundacao Ezequiel Dias (Belo

Horizonte, Brazil), where it was registered as Lmt 9501.

RNA extraction. Total RNA was extracted from the

liver of L. m. muta using Trizolw (GibcoBRL, Rockville,

MD, USA) according to the manufacturer’s instructions.

Briefly, approximately 120 mg of frozen liver was

powdered in the presence of liquid nitrogen and to this

1.0 ml de Trizolw was added. After incubation for 5 min

at 30 8C, the mixture was extracted with 0.2 ml of

chloroform, followed by another incubation period under

the same conditions. Samples were centrifuged at 2900g

for 15 min at 4 8C. RNA in the aqueous phase was

precipitated by isopropanol and resuspended in DEPC-

treated water. RNA integrity was checked by electrophor-

esis on 1% agarose gel at 100 V/cm in 1X TBE buffer and

examined under UV light in the presence of ethidium

bromide.

cDNA synthesis. cDNA was synthesized from 3 mg of

RNA by RT-PCR with Not I (dT)18 primer in the presence of

dithiotreitol (DTT) according to the accompanying instruc-

tions (First-Strand cDNA Synthesis kit, Amersham Bios-

ciences, Upsalla, Sweden).

DNA amplification. Oligonucleotides were designed on

the basis of the published nucleotide sequence encoding for

CNF (Fortes-Dias et al., 1994) and used as primers for PCR

amplification of cDNA. The following pair of primers was

used: 30GGTCAACTTCTCCAGTCC50 (P1, non-coding

region) and 30TCAGAGGCTTGCCAATCTGATG50 (P2,

carboxy-terminal protein segment). Amplification con-

ditions were 2 min at 94 8C, 30 cycles of 15 s at 94 8C,

15 s at 55 8C, 30 s at 72 8C and an extension period of 7 min

at 72 8C (Perkin Elmer 2400 thermocycler). A negative

control was run with no DNA. The amplification products

were analysed by electrophoresis on 1.0% agarose gels in

1 £ TBE buffer and used as the insert for plasmid ligation.

DNA cloning and sequencing. Nine and 27 ng of the

amplification product were incubated overnight at 16 8C

with 20 fmol of the plasmid pCRw 2.1 (Invitrogen, USA)

for ligation. The recombinant plasmid DNA was used to

transform INVaF0 One Shot competent cells as described

in the TA Cloningw kit manual (Invitrogen Co. Carlsbad,

CA, USA). Presence of the insert in putative positive

clones was confirmed by PCR with specific primers

(P2/P1 as described above and P2/P3, where P3

corresponds to the amino-terminal segment in the signal

sequence of CNF). Two positive clones were grown in

liquid culture and the recombinant DNA was purified

using the Wizard Plus Minipreps DNA purification

system (Promega Co. Madison, WI, USA). The DNAs

were completely sequenced in an automated ABI Prism

310 Genetic Analyser (Perkin Elmer Applied Biosystems,

Foster City, CA, USA) using M13 forward and M13

reverse oligonucleotides as primers in the Big Dye

Terminator Cycle Sequencing Ready Reaction (Perkin

Elmer Applied Biosystems, Foster City, CA, USA).

Structural predictions, alignments and phylogenetic

tree. Deduced primary and predicted secondary structures

were obtained using the NTI Vector Suite software

(Informax Inc. Bethesda, MD, USA). Multiple protein

alignments were based on the Clustal W algorithm and the

phylogenetic tree was calculated according to the

Neighbor-joining algorithm (Saitou and Nei, 1987) with

Kimura’s correction using the same software. gPLIs

sequences were obtained from the literature (CgMIP-I

from Cerrophidion godmani, Lizano et al., 2000; Lati-

cauda semifasciata subunits A and B, Ohkura et al., 1999)

or imported from NCBI Data Bank under the following

accession numbers: Agkistrodon blomhoffii siniticus sub-

units A (BAA86970.1) and B (BAA86971.1); C. d.

terrificus (CNF, U08289); Elaphe quadrivirgata subunits

A (BAA83078.1) and B (BAA83079.1); Naja naja

kaouthia subunits A (JC2393) and B (JC2394); N. ater

(NAI) subunit A isoforms 1A (AF211161.1), 2A

(AF211162.1) 3A (AF211163.1) and subunit B

(AF211155.1); Notechis scutatus a chain isoforms i

(CAB56615.1), isoform ii (CAB56616.1), isoform iii

(CAB56617.1) and b chain (CAB56618.1); Oxyuranus

microlepidotus (OMI) subunit A isoforms 1 (AF211167.1)

and 2 (AF211168.1) and subunit B isoforms 1

(AF211157.1) and 2 (AF211158.1); Oxyuranus scutellatus

(OSI) subunit A isoform 1 (AF211164.1), isoform 2

(AF211165.1) and subunit B (AF211156.1); Pseudonaja

textilis (PTI) subunits A (AF211166.1) and B isoforms 1

(AF211159.1) and 2 (AF211160.1); Python reticulatus

(PIP, AF232771); Trimeresurus flavoviridis

(AB003472.1).

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917910

Fig. 1. Alignment of the nucleotide sequences for Lmm1 and Lmm2 transcripts from L. m. muta liver and CNF from C. d. terrificus. The coding

region is in uppercase and the non-coding region in lowercase letters. Initiation and termination codes are in bold and nucleotide substitutions

are shaded in black. The nucleotide sequences used as primers are underlined.

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917 911

3. Results

Amplification of cDNA from L. m. muta liver with CNF

primers resulted in an amplicon with about 640 bp (data not

shown), which was cloned in a plasmid vector. After

checking for the presence of the insert in the recombinant

plasmid DNA, two clones (Lmm1 and Lmm2) were selected

for DNA purification and nucleotide sequencing. Compari-

son of these sequences with CNF cDNA showed 7 and 24

nucleotide (nt) substitutions in Lmm1 and Lmm2, respect-

ively, including one in the non-coding region (Fig. 1).

Calculated percentages of Lmm1 and Lmm2 (nt1-623)

identities to CNF cDNA (nt 79–701) are 98.9 and 96.1%,

respectively. The 30 end segment of 22 nucleotides,

corresponding to one of the primers used in DNA

amplification, was not computed in these calculations.

Translation of the open reading frames in Lmm1 and

Lmm2 indicated two proteins named LNF1 and LNF2,

respectively. The aminoacid sequences show typical

secretion-type signal sequences of 19 amino acids, with

one amino acid substitution (Y/S217) in LNF2 (Fig. 2). The

predicted mature protein isoforms are composed of 181

amino acid residues with 4 and 14 aminoacid substitutions

in LNF1 and LNF2, respectively, compared to CNF, most of

them (78%) maintaining the polarity character of the

residue. Evaluation of the residue substitutions in the

predicted secondary structures indicated that they would

give rise to little or no change, mainly due to the presence of

many unpredicted segments (Fig. 3). Only one residue

substitution was consistent in both LNF clones: L/P/P80

(Fig. 2).

The putative N-glycosylation site described for CNF

(N157) is present in both isoforms as well as the 16 half-

cystines in tandem. Determination of molecular weight (Da)

and pI from the deduced amino acids composition revealed

minor changes: 20,025.19/5.74, 20,073.06/5.51 and

20,057.19/5.55 for LNF1, LNF2 and CNF, respectively.

The deduced amino acid sequences of mature LNF

isoforms were aligned with the gPLIs described so far

(Fig. 4). Besides 15 out of the 16 cysteines, there are other

Fig. 2. Global alignment of the deduced amino acid sequences of LNF1 and LNF2 from L. m. muta and the primary structure of CNF from C. d.

terrificus. Identical amino acid residues are indicated by asterisks ( p ). The signal peptide is underlined and conserved half cystines are in bold.

The carboxy-terminal residues coded by the nucleotide sequence used as primer are in bold.

Fig. 3. Comparison of secondary structure predictions of LNFs and CNF. H, helix; E, strand; –, no prediction.

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917912

Fig. 4. Multiple alignment of gPLIs (Clustal W algorithm). Strictly conserved residues are shaded in black; conserved residues in most

members, taking CNF as reference, are shaded in gray. The two additional half cystines (C104 and C112), exclusive for subunits B, and the

C150/152 are marked in bold.

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917 913

Fig. 4 (continued )

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917914

residues strictly conserved in all members: D25, S40, S51,

L56, P100, G104, K144 and G145. The 16th half-cystine is

positioned as either the 150 or 152 residue.

A phylogenetic tree was constructed based on the

multiple alignment, where at least three main clusters

were defined (Fig. 5).

4. Discussion

Two transcripts corresponding to CNF-homologues

(LNFs) have been isolated from the liver of a single L. m.

muta snake. Each clone isolated coded for a distinct isoform,

which contrasts with C. d. terrificus, where a single nt

substitution was found after complete sequencing of eight

clones (Fortes-Dias et al., 1994). The presence of isoforms,

however, seems to be a common feature in gPLIs, especially

for heteromeric inhibitors such as those described in the

Elapidae family (Dunn and Broady, 2001).

The phylogenetic tree derived from the multiple

alignment of the gPLIs indicates three main clusters. In

Fig. 5, the middle and lower clusters group the subunits B

and A, respectively, of heteromeric inhibitors from non-

viperid snakes. The upper cluster comprises inhibitors

belonging to viperid snakes of the subfamily Crotalinae.

At present there is no gPLI DNA or protein sequence from

any species in the subfamily Viperinae available. Within

that cluster, C. d. terrificus seems closer to L. m. muta than

to C. godmani, A. b. siniticus or T. flavoviridis. This order is

in agreement to the observation of Lizano et al. (2000) that

the viperid snakes from the American continent are closer to

each other than to Asian species in the same subfamily

(Crotalinae). It is worth noting that on one side, A. b.

siniticus PLI is the only heteromeric inhibitor in that cluster

of homomeric complexes. On the other side, the homomeric

Fig. 5. Phylogenetic tree of the endogenous gPLIs protein sequences from snakes constructed by the neighbor-joining (NJ) method. Names of

species are followed by the abbreviation (in brackets) ascribed to each inhibitor in the original description.

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917 915

PIP from P. reticulatus (non-venomous snake, Boidae) is

part of a cluster of heteromeric elapid and colubrid

members.

In the alignment of LNFs with other gPLIs, there are two

striking differences involving cysteine residues. On one

hand, one half-cystine (C150) is not strictly conserved for all

members. It does occupy that position in the homomeric

inhibitors and in all subunits A of heteromeric inhibitors, but

is displaced downstream by two residues in the subunits

B. On the other hand, subunits B have two additional half-

cystine, numbered C104 and C112 in AbsPLIB, totalizing 18

cysteine residues instead of the 16 present in the subunits A

and homomeric inhibitors. The question of quaternary

structure in PLIs is a point of contention. Ohkura et al.

(1999), responsible for the isolation and characterization of

an important number of PLIs, reported the belief that all

venomous snakes, including Elapidae and Crotalinae, have

g-PLIs in their sera, which would be generally composed of

two subunits A and B. They claimed the existence of a

second subunit B even for T. flavoviridis PLI-I, originally

described as a homomer (Nobuhisa et al., 1997), after

modifications in the purification procedure (Ohkura et al.,

1999). A databank search for the DNA or amino acid

sequence of that subunit, however, was unsuccessful.

Therefore, the sequence used in the alignment herein

corresponds to the amino acid sequence of PLI-I (Nobuhisa

et al., 1997).

For LNF, no combination of primers (coding and non-

coding regions) ever resulted in the amplification of more

than one transcript by RT-PCR. Unfortunately, we did not

have access to L. m. muta blood plasma to test for the

presence of circulating LNF. Most studies with L. muta

species have been impaired by difficulties in field capture

and the failure of the snake to breed away from its natural

environment. In the case of CNF, we also found no evidence

yet for a second subunit, either by RT-PCR of the snake liver

with a series of primers or by direct purification of the

inhibitor by affinity chromatography with a mixture of CB

isoforms. A single subunit has been described also for the

homomeric gPLI from C. godmani by Lizano et al. (2000)

who even proposed the splitting of gPLIs into two

subclasses, according to their quaternary structure, as well

as other differences in their specificity and level of

glycosylation. Although the number of g inhibitors

described in American snakes is limited, the phylogenetic

clustering of homomeric PLIs is in accordance to their

proposal. Isolation of other PLIs, mainly from American

snakes, is in progress and hopefully will help to further

clarify this question.

Acknowledgements

We are grateful to G.A. Cotta, biologist from the

Herpetarium of Fundacao Ezequiel Dias, for providing

the snake used in our study. DNA sequencing was performed

by one of us (Fortes-Dias) in a short-term stay (FAPEMIG

CAM 90403/00) at the Department d’Ingenierie et d’Etudes

des Proteines (CEA, Saclay) and we thank Pr. A. Menez and

Dr F. Ducancel for the support and suggestions. We are also

indebted with Dr M. Richardson for kindly reviewing the

language. This work was supported by the Fundacao de

Amparo a Pesquisa do Estado de Minas Gerais (FAPEMIG),

grants CBB 909/98 and EDT 24000/01, including a fellow-

ship to C.J. Barcellos, our undergraduate student.

References

Brattstrom, B.H., 1964. Evolution of the pit vipers. Trans. San

Diego Soc. Nat. Hist. 13, 185–268.

Dunn, R.D., Broady, K.W., 2001. Snake inhibitors of phospho-

lipases A2 enzymes. Biochim. Biophys. Acta 1533, 29–37.

Faure, G., 2000. Natural inhibitors of toxic phospholipases A2.

Biochimie 82, 833–840.

Fortes-Dias, C.L., 2002. Endogenous inhibitors of snake venom

phospholipases A2 in the blood plasma of snakes. Toxicon 40,

481–484.

Fortes-Dias, C.L., Fonseca, B.C.B., Diniz, C.R., Kochva, E., 1991.

Purification and properties of an antivenom factor from the

plasma of the South American rattlesnake (Crotalus durissus

terrificus). Toxicon 29, 997–1008.

Fortes-Dias, C.L., Jannotti, M.L.D., Franco, F.J.L., Magalhaes, A.,

Diniz, C.R., 1999. Studies on the specificity of CNF, a

phospholipase A2 inhibitor isolated from the blood plasma of

the South American rattlesnake (Crotalus durissus terrificus).

I. Interaction with PLA2 from Lachesis muta muta snake venom.

Toxicon 37, 1747–1759.

Fortes-Dias, C.L., Lin, Y., Ewell, J., Diniz, C.R., Liu, T.-Y., 1994.

A phospholipase A2 inhibitor from the plasma of the South

American rattlesnake (Crotalus durissus terrificus). J. Biol.

Chem. 269, 15646–15651.

Fuly, A.L., Francischetti, I.M., Zingali, R.B., Carlini, C.R., 1993.

Partial purification and physicochemical properties of phospho-

lipases A2 from the venom of the bushmaster snake (Lachesis

muta). Braz. J. Biol. Med. Res. 26, 459–463.

Fuly, A.L., Machado, O.L.T., Alves, E.W., Carlini, C.R., 1997.

Mechanism of inhibitory action on platelet activation of a

phospholipase A2 isolated from Lachesis muta (bushmaster)

snake venom. Thromb. Haemost. 78, 1372–1380.

Glaser, K.B., Mobilio, D., Chang, J.Y., Senko, N., 1993.

Phospholipase A2 enzymes: regulation and inhibition. Trends

Pharmacol. Sci. 14, 92–98.

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

logical effects of snake venom phospholipases A2. Toxicon 27,

613–635.

Lizano, S., Angulo, Y., Lomonte, B., Fox, J.W., Lambeau, G.,

Lazdunski, M., Gutierrez, J.M., 2000. Two phospholipase A2

inhibitors from the plasma of Cerrophidion (Bothrops) godmani

which selectively inhibit two different group-II phospholipase

A2 myotoxins from its own venom: isolation, molecular cloning

and biological properties. Biochem. J. 346, 631–639.

Nobuhisa, I., Inamasu, S., Nakai, M., Tatsui, A., Mimori, T.,

Ogawa, T., Shimogashi, Y., Fukumaki, Y., Hattori, S., Kihara,

H., Ohno, M., 1997. Characterization and evolution of a gene

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917916

encoding a Trimeresurus flavoviridis serum protein that inhibits

basic phospholipases A2 isozymes in the snake’s venom. Eur.

J. Biochem. 249, 838–845.

Ohkura, N., Kitahara, Y., Inoue, S., Ikeda, K., Hayashi, K., 1999.

Isolation and amino acid sequence of a phospholipase A2

inhibitor from the blood plasma of the sea krait, Laticauda

semifasciata. J. Biochem. 125, 375–382.

Ohkura, N., Okuraha, H., Inoue, S., Ikeda, K., Hayashi, K., 1997.

Purification and characterization of three distinct types of

phospholipase A2 inhibitors from the blood plasma of the

Chinese mamushi, Agksitrodon blomhoffii siniticus. Biochem. J.

325, 527–531.

Perales, J., Villela, C., Domont, G.B., Choumet, V., Saliou, B.,

Moussatche, H., Bon, C., Faure, G., 1995. Molecular

structure and mechanism of action of the crotoxin inhibitor

from Crotalus durissus terrificus serum. Eur. J. Biochem.

227, 19–26.

Saitou, N., Nei, M., 1987. The neighbor-joining method for

reconstructing phylogenetic trees. Mol. Biol. Evol. 4,

406–425.

Valiente, C., Moreno, E., Sittenfeld, A., Lomonte, B., Gutierrez,

J.M., 1992. An electrophoretic study on phospholipase A2

isoenzymes in the venoms of Central American Crotaline

snakes. Toxicon 30, 815–823.

C.L. Fortes-Dias et al. / Toxicon 41 (2003) 909–917 917