Biochimica et Biophysica Acta 1476 (2000) 372^376
www.elsevier.com/locate/bba
Short sequence-paper
A common motif in proparts of Cnidarian toxins and nematocyst
collagens and its putative role1
Gregor Anderluh *, Zdravko Podlesek, Peter Mac­ek
Department of Biology, Biotechnical Faculty, University of Ljubljana, Vec­na pot 111, 1000 Ljubljana, Slovenia
Received 3 September 1999; received in revised form 15 October 1999; accepted 10 November 1999
Abstract
In Cnidarians, cnidoblast cells contain organelles called cnidocysts, which are believed to be the product of an extremely
complex regulated secretory pathway. When matured, these stinging organelles are capable of storing and delivering toxins.
We hypothesized that translated nematocyst proteins might comprise specific sequences serving as signals in sorting to the
organelle. A sodium channel neurotoxin from the sea anemone Actinia equina was cloned and the toxin precursor sequence
was compared to those of nematocyst collagens, pore-forming toxins and ion channel neurotoxins. It was found that all the
analyzed sequences possess a highly conserved stretch of nine amino acid residues ending with Lys-Arg N-terminally of the
mature region. ß 2000 Elsevier Science B.V. All rights reserved.
Keywords: Cnidaria; Pore-forming toxin; Collagen; Regulated sorting pathway; Sorting signal; Sodium channel neurotoxin
Cnidarians (Coelenterates), evolutionarily the most
primitive metazoans, evolved a special weaponry
called cnidocysts (nematocysts). These organelles
are produced in cnidocytes (nematocytes), and are
considered to be one of the most complex cellular
secretory products. A number of di¡erent toxins
have been isolated from them (for examples see [1^
4]) providing evidence for a widely accepted belief
that cnidocysts serve for storage and active delivery
of venom [2,3,5]. The mature organelle consists of a
collagen capsule, a long tubule connected with a
stylet, with a cover which seals the cyst. Upon proper
external stimulus, the cover opens, and the tubule
evaginates at extreme speed, releasing the cyst con-
* Corresponding author. Fax: +386-61-273-390;
E-mail: [email protected]
1
GenBank accession number AF130344.
tents into the victim [6,7]. In many respects, the formation of cnidocyst constituents and tra¤cking of
their proteins is considered as an example of a very
complex secretory apparatus. It is reasonable to anticipate that translated cnidocyst constituents and
toxins could have a common signal directing them
to a maturing cnidocyst. It is notable that all cloned
cnidarian toxins [8^10] and nematocyst collagens
[11,12] have a similar structural organization (Fig.
1A). Between the signal peptide and mature region
they contain a propart 9^17 residues long, always
terminating with Lys-Arg. The toxin propart is composed of mainly polar (16.7%) and charged amino
acid residues (57.5%). Collagen sequences have a
slightly shorter propart of 8^12 amino acid residues.
The composition favoring polar and charged residues, however, is preserved. The aim of this study
was to search for an eventual motif in common to
peptides and proteins of nematocyst origin.
0167-4838 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 7 - 4 8 3 8 ( 9 9 ) 0 0 2 3 7 - X
BBAPRO 30457 28-1-00
G. Anderluh et al. / Biochimica et Biophysica Acta 1476 (2000) 372^376
373
Fig. 1. The structural organization of cnidocyst proteins. (A) The structural organization of cnidarian toxins and nematocyst collagens. Cnidarian toxins are presented schematically. Black, the signal sequence ; gray, the propart; and white, the mature region. A bar
denotes a length of 10 amino acid residues. Clx1, calitoxin 1 from Calliactis parasitica [10] ; EqtII, equinatoxin II from Actinia equina
[9] ; HmK, inhibitor of the potassium channel from Heteractis magni¢ca [8]; AeNa, inhibitor of sodium channel from A. equina (this
work); N-Col, collagens from Hydra magnipappilata [11]; AdC1, collagen from Acropora donei [12]. (B) Nucleotide and amino acid sequences of A. equina sodium channel neurotoxin (AeNa). The signal peptide is underlined. The propart is doubly underlined. The protein sequence is numbered according to the presumed initiation methionine residue, the numbering of the mature toxin is in parentheses.
From the sea anemone Actinia equina L. three different types of nematocyst toxins have already been
isolated: very potent cytolysins, equinatoxins [5], and
potassium [13] and sodium channel neurotoxins [14],
but only equinatoxins were studied at cDNA level
[9,15,16]. For the purpose of present study, we
cloned an additional cDNA from A. equina cDNA
library coding for the nematocyst derived precursor
polypeptide, namely, sodium channel neurotoxin.
The cDNA library from a single speciman of sea
anemone A. equina [9] prepared in bacteriophage
Vgt11 was used for PCR ampli¢cation. The library
was ¢rst ampli¢ed with the oligonucleotides Vgt11F
(5P-GGTGGCGACGACTCCTGGAGCC-3P)
and
Vgt11R (5P-GACACCAGACCAACTGGTAATG3P) in order to reduce false positive results after secondary PCR ampli¢cation. The mixture for ampli¢-
cation contained appropriate bu¡er, 50 pmol of each
primer, and 2.5 U of Taq polymerase (New England
Biolabs). Twenty-¢ve cycles of ampli¢cation were
performed at the following conditions: denaturation,
92³C for 1 min; annealing, 55³C for 1 min; extension
at 72³C for 1 min. The smear between 300 and 1000
bp was excised, puri¢ed and used in the secondary
PCR using Vgt11R as a sense oligonucleotide and
degenerate AeNa (5P-CGGAATTCAYTGYTTRCARCARTC-3P) as an antisense oligonucleotide.
AeNa is complementary to the last ¢ve amino acid
residues of an inhibitor of the sodium channel from
the sea anemone A. equina [14] having a degeneracy
of 32. DNA was ampli¢ed under the same conditions. The fragment of approximately 350 bp was
excised from the gel, restricted with EcoRI, and
cloned into the pUC19. Three independent clones
BBAPRO 30457 28-1-00
374
G. Anderluh et al. / Biochimica et Biophysica Acta 1476 (2000) 372^376
Fig. 2. Alignment of proparts of cnidarian toxins and nematocyst collagens. Proparts of cnidocyst proteins are aligned according to
the cleavage site after Lys-Arg (large arrow). Proteins presented and used for analysis are three equinatoxins, two calitoxins, inhibitors
of sodium and potassium channels and collagens. The abbreviations are as in Fig. 1B. The length of the signal peptide, in parentheses
on the left of the propart, was determined using SignalP [36]. The cleavage after the signal peptide is denoted by the small arrow. An
asterisk denotes incomplete sequences at their N-terminal part. Using a MEME tool [19] a motif of nine amino acid residues was revealed (shaded). The numbers on the right represent the length of the propart and the length of the precursor proteins (bold, in parentheses). Accession numbers from GenBank (normal), SwissProt (italics), DDBJ (bold) or EMBL (bold italics) databases are on the
far right. In consensus line, s 60% conserved residues are shown. The Lys-Arg recognition site for subtilisin-like endopeptidases is
conserved in all sequences (boxed). Hydrophobic amino acid residues (h; A, F, G, I, L, M, V, W, Y, C, P) are presented in black,
polar amino acid residues (p; D, E, H, K, R, N, Q, S, T) are shown in black on light gray background, negatively charged amino
acid residues (3; D, E) or Asp (d) are shown in white on a dark gray background. Conservative replacements are black on a dark
gray background. Species names are as follows: Ae, Actinia equina; Cp, Calliactis parasitica; Hm, Heteractis magni¢ca; Hy, Hydra
magnipapillata; Ad, Acropora donei.
were sequenced in both directions using T7 DNA
polymerase (Pharmacia) and [35 S]dATPKS (Amersham). Notably, we could not amplify the inhibitor
for potassium channels [13] by the same procedure.
Antisense AeK oligonucleotide (5P-CGGAATTCARCAYTTNCCRCANGT-3P) was used in the combination with Vgt11F or Vgt11R for the secondary
PCR. AeK has a degeneracy of 128. This means
that inhibitor was either not expressed in this particular speciman, or was expressed in such low amounts
that it was not possible to amplify it with PCR.
The nucleotide and derived amino acid sequences
of the 316-bp fragment obtained by PCR are presented in Fig. 1B. An open reading frame of 246bp codes for a precursor protein of 82 amino acid
residues. A potential translation start site is found at
positions 68^70, with highly conserved adenosine at
positions 33 and +4 for the initiation of translation
in vertebrates [17]. The initiation start provides
translation of a typical signal peptide of 19 residues
[18]. A further insertion of nine mostly negatively
charged amino acid residues is found between the
signal peptide and the mature peptide. It ends with
a pair of basic residues Lys-Arg at positions 27^28, a
potential proteolytic cleavage site. The deduced sequence of a 54 amino acid residues long mature protein matches the already published one obtained by
protein sequencing [14].
In order to search for a motif within the precursors of nematocyst toxins and collagens, a MEME
tool [19] was used. Twelve precursor sequences of
pore-forming proteins, equinatoxins, inhibitors of
ionic channels, calitoxins, HmK and AeNa and ¢ve
collagens were used as an input data set. The most
conserved sequence of 9 residues DEDEDIEKR was
recognized in all proteins located N-terminally to the
mature region (Fig. 2). The nine amino acid residues
long motif ends with the highly conserved sequence
Glu-Lys-Arg (similar to the same part around the
dibasic recognition sequence discovered by Block
Maker [20]). The E-values for the alignment of motifs and cnidarian sequences were from 6.7e-3 to
9.7e-9, indicating statistically signi¢cant results (usually E 6 0.001). This was the only motif common to
all 12 proteins. None was recognized in 12 sequences
if only mature parts were used as an input, further
indicating that the motif must reside in the propart.
Two sets of randomized sequences of the same length
BBAPRO 30457 28-1-00
G. Anderluh et al. / Biochimica et Biophysica Acta 1476 (2000) 372^376
were used for the input as a control. In both cases,
no motif common to all 12 sequences was recognized. Similarly, no motif was discovered if shu¥ed
sequences of our set were used as an input for
MEME. Terminal Lys-Arg are conserved in all sequences. Dibasic sites are usually present before mature part of prohormones as a cleavage point for
subtilisin like proteases, most commonly PC1 and
PC2, in secretory granules. Such processing enzymes
also exist in cnidarians: i.e. a subtilisin-like prohormone convertase has been cloned from Hydra attenuata [21].
In general, proparts have various functions in the
precursors of proteins. Most important are assistance
in folding, prevention of activity, and tra¤cking a
precursor to di¡erent locations within the cell. The
role of the propart in folding is not likely for cnidarian toxins since it was possible to obtain fully active
toxins by expressing them in bacterial hosts as mature proteins without a propart [8,9,22]. Furthermore, fully active cnidarian potassium channel neurotoxins were chemically synthesized and properly
folded [23,24]. The second possibility, that the propart prevents activity of toxins within cnidocytes is
unlikely. Cnidarian toxins exhibit positive net charge
and proparts might neutralize such a positive charge,
disabling their interaction with lipid membranes.
This was suggested for an acidic propart present in
antibacterial peptides [25]. However, once the toxins
arrive to the cnidocyst, this propart is no longer
needed, since the interaction of toxins with the lipid
membrane is prevented by a collagen wall. Furthermore, it has been proposed by some authors [8,10]
that toxins are stored as inactive precursors, being
processed prior to release after the proper stimulus.
However, as toxins are released in less than 3 ms [7],
it is unlikely that any modi¢cation could occur at
this stage. They are more likely processed during
cyst maturation. In addition, Meinardi et al. [26]
proved that cytolysins from the sea anemone Phymactis clematis, belonging to the same family of lysins as equinatoxins, are inactive against the membranes of sea anemones. It was shown that the
P. clematis membrane contains phosphonosphingolipids but not sphingomyelin which is considered a
speci¢c acceptor for that type of toxins [5]. Therefore, we rather suggest, that the propart may have a
role in the regulated secretory pathway (RSP), i.e.
375
the targeting of toxin precursors to secretory membranous structures involved in cnidocyst formation.
In RSP, proteins are stored in secretory vesicles
and released from the cell after a proper stimulus
[27]. Typical examples of proteins following the
RSP include hormones and neuropeptides [28,29].
All are synthesized as long precursors, usually composed of a signal peptide, propart, and mature protein. They are processed during transportation to the
secretory vesicles, ¢rst by signal peptidases and later
with subtilisin-like endopeptidases. How proteins are
sorted to RSP is not well understood. In general,
signals for tra¤cking proteins to di¡erent cell compartments are diverse and include short, conserved
sequences of amino acid residues, amphiphilic helices
or surface patches. A speci¢c signal common to all
types of proteins which follow the RSP is not known.
Our suggestion on the putative biological role of
the nematocyst motif is contrasted by the fact that
no signal composed of linear amino acid sequence
has yet been determined for proteins following RSP
despite attempts to identify such a motif from
aligned sequences of various prohormones [30]. Instead, a surface patch within the proregion has been
proposed to be the signal for binding to carboxipeptidase E (CPE), a sorting receptor for POMC, proinsulin, and proenkephalin [31]. This 24-residue surface
patch is located after the signal sequence, and consists of highly conserved two charged and two hydrophobic amino acid residues [32]. It is folded into the
right conformation with two disul¢de bonds. A similar structure of a N-terminally disul¢de bridged loop
of 22 amino acid residues has been recently proposed
to be responsible for the sorting of chromogranin B
to secretory vesicles [33]. Considering the proteins
analyzed here, only N-Col4 possesses one Cys residue within the proregion, thus the needed bridge
cannot be formed. Furthermore, none of the analyzed proparts is long enough to form an appropriate
patch.
Similar precursor organization was also found in
frog antimicrobial peptides of the magainin and dermaseptin family. They form a large heterogeneous
family produced as precursors with a highly conserved signal peptide and a negatively charged propart [34]. They are stored in secretory granules within
poisonous glands and are released after mechanical
stimulus [35]. An acidic propart is usually longer, but
BBAPRO 30457 28-1-00
376
G. Anderluh et al. / Biochimica et Biophysica Acta 1476 (2000) 372^376
exhibits similar sequence around dibasic site (not
shown).
To conclude, we determined the nucleotide sequence of sodium channel neurotoxin from the sea
anemone A. equina. The analysis of its propart and
those from other nematocyst toxins and collagens
revealed a conserved stretch of nine residues similar
to that found previously in precursors of antimicrobial peptides. We suggest that the motif may have a
role in tra¤cking to RSP, which, however, calls for
experimental con¢rmation.
This work was supported by a grant from the
Ministry of Science and Technology of Slovenia.
The authors would like to thank to Prof. Dr. Franc
Gubens­ek, Joz­ef Stefan Institute, Ljubljana, Slovenia, for enabling us to determine the DNA sequence
of sodium channel inhibitor.
References
[1] R.S. Norton, Toxicon 29 (1991) 1051^1084.
[2] G. Rottini, L. Gusmani, E. Parovel, M. Avian, P. Patriarca,
Toxicon 33 (1995) 315^326.
[3] A. Lotan, L. Fishman, E. Zlotkin, J. Exp. Zool. 275 (1996)
444^451.
[4] M. Klug, J. Weber, P. Tardent, Toxicon 27 (1989) 325^339.
[5] P. Mac­ek, FEMS Microbiol. Immunol. 105 (1992) 121^130.
[6] A. Lotan, L. Fishman, Y. Loya, E. Zlotkin, Nature 375
(1995) 456^456.
[7] P. Tardent, BioEssays 17 (1995) 351^362.
[8] G.S. Gendeh, L.C. Young, C.L.C. De Medeiros, K. Jeyaseelan, A.L. Harvey, M.C.M. Chung, Biochemistry 36 (1997)
11461^11471.
[9] G. Anderluh, J. Pungerc­ar, B. Sítrukelj, P. Mac­ek, F. Gubens­ek, Biochem. Biophys. Res. Commun. 220 (1996) 437^
442.
[10] A. Spagnuolo, L. Zanetti, L. Cariello, R. Piccoli, Gene 138
(1994) 187^191.
[11] E.M. Kurz, T. Holstein, B.M. Petrti, J. Engel, C.N. David,
J. Cell Biol. 115 (1991) 1159^1169.
[12] W. Wang, M. Omori, T. Hayashibara, K. Shimoike, M.
Hatta, T. Sugiyama, T. Fujisawa, Gene 152 (1995) 195^200.
[13] S. Minagawa, M. Ishida, Y. Nagashima, K. Shiomi, FEBS
Lett. 427 (1998) 149^151.
[14] X.Y. Lin, M. Ishida, Y. Nagashima, K. Shiomi, Toxicon 34
(1996) 57^65.
[15] J. Pungerc­ar, G. Anderluh, P. Mac­ek, F. Gubens­ek, B. Sítrukelj, Biochim. Biophys. Acta 1341 (1997) 105^107.
[16] G. Anderluh, I. Kriz­aj, B. Sítrukelj, F. Gubens­ek, P. Mac­ek,
J. Pungerc­ar, Toxicon 37 (1999) 1391^1401.
[17] D.R. Cavener, S.C. Ray, Nucleic Acids Res. 19 (1991) 3185^
3192.
[18] G. Von Heijne, Eur. J. Biochem. 133 (1983) 17^21.
[19] T.L. Bailey, M. Gribskov, Bioinformatics 14 (1998) 48^54.
[20] S. Heniko¡, J.G. Heniko¡, W.J. Alford, S. Pietrokovski,
Gene 163 (1995) 17^26.
[21] S.J. Chan, A.A.J. Oliva, J. LaMendola, A. Grens, H. Bode,
D.F. Steiner, Proc. Natl. Acad. Sci. USA 89 (1992) 6678^
6682.
[22] M.J. Gallagher, K.M. Blumenthal, J. Biol. Chem. 267 (1992)
13958^13963.
[23] M.W. Pennington, M.E. Byrnes, I. Zaydenberg, I. Khaytin,
J. De Chastonay, D.S. Krafte, R. Hill, V.M. Mahnir, W.A.
Volberg, W. Gorczyca, W.R. Kem, Int. J. Pept. Protein Res.
46 (1995) 354^358.
[24] M. Dauplais, A. Lecoq, J. Song, J. Cotton, N. Jamin, B.
Gilquin, C. Roumestand, C. Vita, C.L.C. De Medeiros,
E.G. Rowan, A. Harvey, A. Menez, J. Biol. Chem. 272
(1997) 4302^4309.
[25] R.L. Gallo, K.M. Huttner, J. Invest. Dermatol. 111 (1998)
739^743.
[26] E. Meinardi, M. Florin-Christensen, G. Paratcha, J.M. Azcurra, J. Florin-Christensen, Biochem. Biophys. Res. Commun. 216 (1995) 348^354.
[27] T.L. Burgess, R.B. Kelly, Annu. Rev. Cell Biol. 3 (1987)
243^293.
[28] M. Rholam, P. Nicolas, P. Cohen, FEBS Lett. 207 (1986) 1^
6.
[29] P.A. Halban, J.-C. Irminger, Biochem. J. 299 (1994) 1^18.
[30] J.S. Kizer, A. Tropsha, Biochem. Biophys. Res. Commun.
174 (1991) 586^592.
[31] D.R. Cool, E. Normant, F. Sheng, C.-H. Chen, L. Pannel,
Y. Zhang, Y. Peng Loh, Cell 88 (1997) 73^83.
[32] C.-F. Zhang, C.R. Snell, Y. Peng Loh, Mol. Endocrinol. 13
(1999) 527^536.
[33] M.M. Glombik, A. Kro«mer, T. Salm, W.B. Huttner, H.H.
Gerdes, EMBO J. 18 (1999) 1059^1070.
[34] C.L. Bevins, M. Zaslo¡, Annu. Rev. Biochem. 59 (1990)
395^414.
[35] K.S. Moore, C.L. Bevins, M.M. Brasseur, N. Tomassini, K.
Turner, H. Eck, M. Zaslo¡, J. Biol. Chem. 266 (1991)
19851^19857.
[36] H. Nielsen, J. Engelbrecht, S. Brunak, G. Von Heijne, Protein Eng. 10 (1997) 1^6.
BBAPRO 30457 28-1-00