ARCHIVES
OF BIOCHEMISTRY
AND
RIOPHYSICS
Vol. 278, No. 2, May 1, pp. 319-325, 1990
The Amino Acid Sequence of a Myotoxic Phospholipase from the Venom of Bothrops asper Ivan I. Kaiser,*,l Jose Maria Gutierrez,?
Dorothy
Plummer,*
Steven D. Aird,*‘2 and George V. Odell$
“Department of Molecular Biology, University of Wyoming, Laramie, Wyoming 82071; tlnstituto Clodomiro Picado, Universidad de Costa Rica, San Jose, Costa Rica; and #Department of Biochemistry, Oklahoma State University, Stillwater, Oklahoma 74078
Received August 14, 1989, and in revised form December 5,1989
A myotoxic, basic phospholipase A2 (pl > 9.5) with anticoagulant activity has been purified from the venom of Bothrops asper, and its amino acid sequence determined by automated Edman degradation. It is distinct from the B. asper phospholipase A2 known as myotoxin I [Lomonte, B. and Gutierrez, J. M., 1989, Toxicon 27, 7251 but cross-reacts with myotoxin I rabbit antisera, suggesting that the proteins are closely related isoforms. To our knowledge, this is the first myotoxic phospholipase to be sequenced that lacks presynaptic neurotoxicity (iv LD5,, ^I 8 pg/g in mice). The protein appears to exist as a monomer, contains 122 amino acids, and fits with subgroup IIA of other sequenced phospholipase A2 molecules. Its primary sequence shows greatest identity with ammodytoxin B (67%), a phospholipase A2 presynaptic neurotoxin from Vipera ammodytes ammodytes venom. Hydropathy profiles of B. asper phospholipase and the ammodytoxins also show great similarities. In contrast, even though the amino acid sequence identities between B. asper phospholipase and the basic subunit of crotoxin remain high (64%), their hydropathy profiles differ substantially. Domains and residues that may be responsible for neurotoxicity are discussed. t) dew Academic Press,
Inc.
Bothrops asper represents one of the most hazardous snakes found in Central America. Its venom can induce extensive tissue damage, characterized by edema, hemorrhage, and myonecrosis (I). Basic phospholipases and phospholipase-like molecules isolated from the venom have been purified and shown to play an important role in promoting myonecrosis (2-4). They act in part by ’ To whom correspondence should he addressed. ’ Present address: Natural Product Sciences, Inc., 420 Chipeta Way, Suite 240, Salt Lake City, UT 84108. 000:1-9861/90 $3.00 Copyright ‘J11990 hy Academic Press, Inc. All rights oi’ reproduction in any form reserved.
affecting the integrity of the plasma membrane of muscle fibers [see Lomonte and Gutierrez (4) and references therein]. Unlike many snake venom basic phospholipases, the myotoxic phospholipase AZ’s from B. asper do not possess presynaptic neurotoxicity. We felt that some insight into structure-function relationships might be gained in sequencing a myotoxic, nonneurotoxic phospholipase, since we were unaware of any that had been sequenced. While the residues involved in the enzymatic activity of phospholipase AZ’s have been fairly well established (5), little is known about the domains responsible for neurotoxicity and myotoxicity. Sequence and hydropathy comparisons with the ammodytoxins isolated from Vipera ammodytes ammodytes venom indicate great similarity and are consistent with the conclusion of Gubensek and co-workers (6), who suggest that one of the neurotoxic domains may reside near the carboxy-terminus. MATERIALS
AND
METHODS
Crude, dried venom of I?. asper, obtained from adult specimens collected in the Atlantic region of Costa Rica, was kindly provided by the Instituto Clodomiro Picado, Costa Rica. Sephacryl S200, Mono S ionexchange column, and the FPLC’ system were purchased from Pharmacia. Reversed-phase chromatography was carried out on the FPLC system using either a Pharmacia C,/Cs 0.5 X 5-cm column for intact molecules or a Vydac 0.46 X 15 cm C,, column (5 wm) for peptide fractionation. SDS-PAGE (15%) gels were silver-stained as described (7). Phospholipase and lethality assays followed the methods of Aird and Kaiser (8, 9). Myotoxic phospholipase used for lethality assays was diluted in phosphate-buffered saline containing 0.1% bovine serum alhumin. Phospholipase assays were carried out in a pH-stat against I.tu-phosphatidylcholine in a reaction mixture containing Triton X-100 as described earlier (8). Isoelectric focusing gels were kindly run by Brett Lennon using procedures essentially as described (10). Myot.oxic
a Abbreviations used: FPLC, fast protein liquid SDS, sodium dodecyl sulfate; PAGE, polyacrylamide sis; PTH, phenylthiohydantoin; TFA, trifluoroacetic
chromatography; gel electrophoreacid. 319
320
KAISER
0.0 0
10
20
FRACTION
30
40
50
60
NUMBER (20ml)
FIG. 1. Gel filtration profile of 240 mg of crude B. asper venom on a 3 X 95-cm column of Sephacryl S200 superfine. Elution was carried out with 0.1 M sodium acetate (pH 4.0) at room temperature at a flow rate of 80 ml/h. Recovery of A,,, material was quantitative. The indicated fractions (34-38) were pooled, concentrated, and dialyzed against 50 mM sodium phosphate (pH 7.2), in preparation for subsequent cation exchange chromatography.
activity was studied by injecting 50 pg of the phospholipase im in the right gastrocnemius muscle of mice (16-18 g). Control mice received 0.1 ml physiologic saline solution. After 3 h mice were bled from the tail and the creatine kinase activity of plasma was quantitated according to the Sigma Technical Bulletin 520 (Sigma Chemical Co., St. Louis, MO). Creatine kinase activity was expressed in units per milliliter with 1 unit representing the phosphorylation of 1 nanomole of creatine per minute at 25°C. Anticoagulant activity on sheep plateletpoor plasma was determined as described (11). Immunodiffusion studies were carried out by using polyclonal anti-B. asper myotoxin I antisera (12). Sequencing employed an Applied Biosystems 470A protein sequencer with an Applied Biosystems Model 120A on-line analyzer for the analysis of the phenylthiohydantoin (PTH)amino acid derivatives (7). Sequence comparisons were made using PRONUC software on a micro VAX (VMS), with selected sequences added to the NBRF database. Alignments of protein sequences allowed for insertions and deletions with an optimization algorithm based on that of Needleman and Wunsch (13). The phospholipase was reduced and carboxyamidomethylated using a 12-fold excess of DTT over protein disulfide and a 5-fold excess of iodoacetamide over DTT (14). Fragmentation following reduction and alkylation was achieved by cleavage with (i) arginine-specific endoproteinase (Arg-C) from Boehringer-Mannheim, (ii) an endoproteinase that cleaves specifically at the N-terminal side of aspartate (Asp-N) from Boehringer-Mannheim, and (iii) cyanogen bromide, using a 50fold excess of CNBr over protein Met residues in 70% HCOOH and incubation at room temperature (ca. 22°C) for 21 h in the dark. Tryptophan was added as a scavenging reagent to protect protein Trp residues (15). Cleavage with Arg-C was carried out (6 U/mg protein) in 0.1 M ammonium acetate (pH 7.8) for 20 hat 37°C. Asp-N cleavage (1: 280 by weight) was in 50 mM sodium phosphate (pH 8.0) containing a final concentration of 0.5 M guanidine hydrochloride for 15 h at 37°C. Hydropathy profiles were generated using Beckman’s MicroGenie version 5.0 software, using the Kyte and Doolittle (16) option and a window of nine residues. RESULTS
AND
DISCUSSION
A myotoxic phospholipase AZ from B. asper was purified by a modification of the original procedure of Gutie-
ET AL.
rrez et al. (1). Crude venom was initially fractionated on Sephacryl S200 with the phospholipase of interest eluting in the major peak (Fig. 1). Indicated tubes were pooled, concentrated by ultrafiltration (Amicon, 5 kDa pore size), and dialyzed against 50 mM sodium phosphate, pH 7.2, at 4°C. Dialyzed material was applied to a Mono S column equilibrated in the same buffer and eluted with a NaCl gradient (Fig. 2). The recovered peak which eluted at 13 min and its trailing shoulder, showed the highest levels of phospholipase activity. When aliquots of recovered fractions from cation-exchange were run on 15% SDS-PAGE, a single major band was observed for the 13-min fraction (Inset, Fig. 2). On ureaPAGE gels however, two distinct bands of similar staining intensity were seen (not shown). The 13-min fraction eluted as a single symmetrical component from a C,/C,, reversed phase column in 0.1% TFA (not shown) and all molecules present exhibited pl values greater than 9.5 on isoelectric focusing. This component represented about 7% of the crude venom. Purified material from the cation-exchange column had a catalytic activity of 750 pmol/mg min, as determined from the slope of a line resulting from a plot of initial activity vs concentration. The material showed linear kinetics similar to those observed with the basic subunit of crotoxin when 0.9
i-
// // / R/ A
0
, 0
IO
20
30
MINUTES FIG. 2. Cation-exchange FPLC of pooled fractions from the S200 run. About 2.5 mg of dialyzed, pooled material from Fig. 1 dissolved in 50 mM sodium phosphate (pH 7.2) was applied to a 0.5 X 5-cm Mono S column and eluted with a linear gradient of NaCl (- -), buffered with 50 mM sodium phosphate (pH 7.2). The indicated material eluting at 13 min was collected, dialyzed against deionized water, and lyophilized. Inset: SDS-PAGE of 0.5 and 1 pg of the indicated pooled material run on a 15% polyacrylamide gel and stained with silver.
MYOTOXIC
PHOSPHOLIPASE
20
Ser
Leu
Ile
Glu
Phe Ala
Lys
Met
Ile
10 Leu
Glu
30 Trp
Gly
Gly
Gln
Gly
Pro
Lys
Asp
Ala
40 Thr
Gln
321
Bothrops asper
FROM
Arg
Leu
Pro
Phe Pro Tyr
Tyr Thr Thr N-terminus
Asp Arg Cys Cys N terminus
Phe
Val
His
Cys
Glu
Thr
Lys
Tyr
Gly
Cys Tyr
Gly
Lys
Leu Ser
Cys
Gly
Asn
Cys
Lys
Cys
Glu
50 Asp
Cys
Tyr
RC- 1
Pro
60 Lys
80
70 Thr
Asp
Arg
Tyr
Ser
Tyr
Ser
Arg
Lys
Ser
Gly
Val
Ile
Ile
Cys
Gly
Glu
RC-2 _
Gly
Thr
Pro
Cys
Glu
Lys
Gln
Ile
Met
Ala
110 Tyr
Pro
Asp Leu/
Phe
_
lx-1 DN 1
Cys
Asp
90 Lys
Ala
_
Ala
Ala
Val
Cys
Phe Arg
Glu
Asn
100 Leu Arg
Thr
Tyr
Lys
Lys
Arg
Tyr
Leu Cys
w-2 --__
CB-1
DN-2
Lys
Lys
Pro
120 Ala Glul Asp Lys/Pro CB- 1
FIG. 3.
Cys
The amino acid sequence of a myotoxic
carried out on a phosphatidyl choline substrate in a Triton X-IO@-containing reaction mixture (8). It had an iv LD,,, value in female mice of ca. 8 lug/g (9). Material in the later eluting shoulder of the 13-min fraction in Fig. 2 also consisted primarily of phospholipase molecules. On SDS-PAGE one major fraction was seen which comigrated with the 13-min fraction and had a catalytic activity of 430 pmol/mg min, when measured under conditions as described above. This material represented about 2% of the crude venom protein. It seems likely that the entire 13-min peak is a collection of phospholipase molecules with slightly different primary sequences. This is suggested by the presence of two amino acids at positions 113, 120, and 121. If the entire peak, shoulder and all, had been pooled and sequenced, it is probable that additional variant sequences would have been found. We also determined the relationship of the myotoxic phospholipase sequenced here with other phospholipases from B. asper venom. The myotoxic phospholipase induced a creatine kinase increment in plasma, reaching values of 433 + 36 units/ml (n = 4) [control values = 43 + 9 units/ml (n = 4)]. See the Materials and Methods section for details. This increment is quantitatively similar to the one observed after B. asper myotoxin I injection (11). It also exerted an anticoagulant effect on sheep platelet-poor plasma, similar to that observed with myotoxin I (11). Our phospholipase showed cross-reactivity with polyclonal antibodies raised to myotoxin I, although immunodiffusion analysis with rabbit polyclonal antisera showed only partial identity between our phospholipase and B. asper Myotoxin I. Finally, PAGE of
phospholipase
from B. asper.
crude venom on 12% urea gels (17) indicated at least four discrete bands representing basic proteins, which roughly correspond to peaks 4 through 7 as separated on CM-Sephadex (1). Lomonte and Kahan (18) demonstrated that monoclonal antibodies against B. asper myotoxin I reacted with these four proteins, suggesting that they are isoforms of myotoxic phospholipase A2’s. It appears that the myotoxic phospholipase which we sequenced compares to peak 4 material as identified by Gutierrez et al. (l), and is slightly less basic than myotoxin I present in peak 5 (1,4). This is based on the mobility of the myotoxic phospholipase sequenced here in urea-PAGE gels (not shown), when coelectrophoresed with known samples recovered from CM-Sephadex (4, 18). Our phospholipase is clearly different from myotoxin I, but it is a related isoform. This explains the lack of agreement between selected amino acid compositions reported by Mebs and Samejima (2) and Lomonte and Gutierrez (4) for myotoxic phospholipases from B. asper, and that calculated from the primary sequence determination of B. asper phospholipase A2 reported here. Studies in progress by Gutierrez and co-workers on myotoxins from Bothrops venoms should provide greater insight into the relationship among these molecules. The primary sequence of the myotoxic phospholipase from B. asper contains 122 amino acids and is shown in Fig. 3. Sequencing was carried out on the reduced and carboxyamidomethylated protein by a combination of direct sequencing from the amino-terminal end and sequencing of fragments. These were generated by enzymatic digestion with Arg-C and Asp-N and chemical cleavage with CNBr, and subsequently fractionated by
322
KAISER
ET AL.
(4
I 60
15
MINUTES
30
45
MINUTES
0
A-- / I
/’ 40
20
60
MINUTES FIG. 4. Separation of fragments teinase Asp-N, and (C) CNBr.
of myotoxic
phospholipase
from B. asper following
reversed phase column chromatography. After sequencing the amino terminal end (l-53) and two additional large fragments generated by Arg-C digestion [RC-1, (43-68) and RC-2, (69-97)], it was possible to search the protein sequence database for a closely related phospholipase since the nonoverlapping fragment 69-97 could be tentatively positioned by the location of Cys residues, whose positions are invariant in any one class of phospholipases. The ammodytoxins, from V. ammodytes, showed the greatest identity with B. asper phospholipase A, using this partial sequence (l-97) and were used as a
treatment
with (A) endoproteinase
Arg-C, (B) endopro-
guide to suggest further cleavage strategy. With only seven aspartate residues present in the ammodytoxins, the possibility of generating only three or four large peptides, plus several smaller ones by the Asp-N enzyme, seemed likely. Fractionation of the Asp-N digest yielded a relatively simple profile (Fig. 4B), yielding two large peaks that appeared homogeneous. Fragment DN-1 corresponded to a peptide which started at residue Asp-62. Its sequencing was continued through residue 81, which gave the confirming overlap at position 68-69. Sequencing was electively terminated on this fragment after resi-
MYOTOXIC Species Bothrops asper Vipera ammodytes ammodytes Vipera ammodytes ammodytes Agkistrodon blomhoffi Crotalus durissus terrificus Agkistrodon piscivorus piscivorus
PHOSPHOLIPASE
Protein Myotoxic Phospholipase Ammodytoxin B Ammodytoxin A Mamushi Phospholipase A2 Crotoxin, basic subunit Phospholipase A2
65
0 95 AAAVCFRENLRTYKKRYMAYPDLIFLCKKPAE/DK/PC AAAI CFRKNLKTYNHI AAAI CFRKNLKTYNYI AAAICFRDNLKTYKKRYMAYPDI VAAECLRRSLSTYKYGYMFYPD AVAI CLRENLDTYNKKYKAYFK
Bothrops
323
asper
5 1 SLIEFAKMILEETKRLPF SLLEFGMMILGETGKNPL SLLEFGMMILGETGKNPL H L L 0 F R K HLLQFNKMIKFETRKNAI SVLELGKMILQETGKNAI
15
10
M
l
K
K
M
T
G
45
40
35
30 25 20 PYYTTVGCYCGWGGOGQPKDATDRCCFVHDCCYGKL TSYSFYGCYCGVGGKGTPKDATDRCCFVHDCCYGNL TSYSFYGCYCGVGGKGTPKDATDR fSYAFYGCYCGSGGRGKPKDATDRCCFVHDCCYEKV PFYAFYGCYCGWGGQGRPKDATDRCCFVHDCCYGKL TSYGSYGCNCGWGHRGQPKDATDRCCFVHKCCYKKL 60 55 SNCKPKTDRYSYSRKSGVI PDCSPKTDRYKYHRENGAl PDCSPKTDRYKYHRENGAI TGCKPKWDDYTYSWKNDGI AKCNTKWDIYRYSLKSGYI TDCNHKTDRYSYSWKNKAI
FROM
K
E
P
V
50
CCFVHDCCYGNL
75 80 ICGEGTPCEKQICECDK VCGKGTSCENRICECDR VCGKGTSCENRI VCGGDDPCKKEICECDR TCGKGTWCEEQI ICEEKNPCLKEMCECDK
70
5
10 YMYYPD YRNYPD
F F S L
90
65
CECDR CECDR
15
20
LCKKES LCKKES LCSSKS RCRGPS KCKLP-
E E E E D
K K K T T
c c C c c
FIG. 5. Comparison of the amino acid sequence of the myotoxic phospholipase from B. asper with ot,her snake venom phospholipases. Identical residues are highlighted. References for sequences 2 through 6 are Ritonja et al. (6), Ritonja and Gubensek (23), Forst et al. (24), Aird scores for the myotoxic phospholipase with itself and related et al. (19), and Maraganore and Heinrikson (25), respectively. Optimization proteins are shown below, followed by the percentage identity. I?. asper, phospholipase, 734, 100%; V. a. ammo&es, ammodytoxin B, ~2,677;; V. a. ammodytm, ammodytoxin A, 545, 66 Y’,; A. blomhofi, mamushi phospholipase, 536, 65%; C. d. terrificus, basic subunit of crotoxin, ,509, 64%; and A. p. piscicorus, phospholipase A,, 501.62%.
due 81, since we had sufficient overlap of the earlier two peptides (RC-1 and RC-2) from the Arg-C digestion. The second large, homogeneous-appearing peak DN-2, gave the fragment predicted to start with Asp-89, which continued through Pro-111. This peptide revealed the presence of a Met at position 108. Cleavage with CNBr and reversed phase column chromatography of the products yielded the 109-122 fragment (CB-l), which gave a convenient three amino acid overlap with the carboxy-terminus of the DN-2 peptide. This final fragment also revealed the presence of a mixture of two amino acids at three different positions. Position 113 contained Leu: Phe at a molar ratio of l.EXX1.00; position 120, Glu:Asp at a molar ratio of 1.14:l.OO; and position 121, Lys:Pro at a molar ratio of 1.33:1.00. These were the only sites observed to have multiple amino acids present. With the partial sequence available and the ammodytoxin sequence as a guide, it was possible to predict nearly all of the cleavage sites for the endoproteinase Asp-N and CNBr. On the basis of amino acid composition, expected peptide size, and what appeared to be pure peptides in peaks DN-1 and DN-2 in Fig. 4B, as well as peak CB-1
in Fig. 4C, we fortuitously selected these peaks for initial sequencing. Since they yielded the necessary sequence information, additional fragments were not sequenced from these runs. In other instances fragment identification and location could be made after sequencing three residues in from the N-terminus, at which time a decision was made on whether to continue or terminate further sequencing. Figure 4 illustrates elution profiles of peptides following enzymatic and chemical hydrolyses. In Fig. 4A, peptide RC-3 corresponded to a peptide starting at Cys-28, which we sequenced through Gly-34 and aborted. RC-4 contained two peptides which were sequenced simultaneously; one started at Ala-91 and terminated at Arg-97; the other started at Gly-25 and continued through Gln33, when sequencing was aborted. Peptides RC-1 and RC-2 ran from Cys-43 through Arg-68 and Lys-69 through Arg-97, respectively (see Fig. 4A). With the ArgC endoproteinase we observed cleavages on the carboxyside of Tyr and Lys in addition to Arg, as noted before (19). The Arg-C endoproteinase used was not sequencing grade, in contrast with the Asp-N, with which we ob-
324
KAISER
IA
t30
I 0
20
40
60
60
100
120
RESIDUE NUMBER FIG, 6. Hydropathy profiles of (A) B. asper myotoxin phospholipase, (B) ammodytoxin A (solid line, 1) and B (broken line, 2) from V. a. ammodytes, and (C) the basic subunit of crotoxin from the venom of C. d. terr$icus. A span of nine consecutive residues was used to determine the profiles.
served no unexpected cleavages. Figure 4B shows the isolation of peptides DN-1 (Asp-62 through Cys-81) and DN-2 (Asp-89 through Pro-111). Figure 4C represents the fragments generated by CNBr cleavage of the reduced and alkylated phospholipase, with peak CB-1 corresponding to the terminal peptide (Ala-109 through Cys-122). It is unclear why the CB-1 peptide obtained by CNBr cleavage gave a single peak on reversed phase chromatography, yet differed in 3 out of 14 residues. The five most similar sequences to B. asper phospholipase A,, in our data base, included phospholipases from viperid and crotalid venoms (Fig. 5). These had similarity scores ranging from 552 (67% amino acid identity) with ammodytoxin B from V. a. ammodytes (a true viper) to 501 (62% amino acid identity) with the phospholipase A, from Agkistrodon piscivorus piscivorus (a pit viper). Between these two were ammodytoxin A, a homolog of ammodytoxin B, the phospholipase from the pit viper Agkistrodon halys blomhofi from Japan, and the basic subunit of crotoxin from the South American rattlesnake Crotalus durissus terrificus. The next 4O-odd sequences with the highest similarity scores (507-210) were all phospholipases from either viperids, crotalids, or elapids.
ET AL.
A number of papers have attempted to identify particular regions of the phospholipase molecules with a particular function such as neurotoxicity or myotoxicity. Kini and Iwanaga (20) recently proposed that monomeric presynaptic neurotoxins have a hydrophobic region between residues 80 and 110 that may play a significant role in the neurotoxicity of phospholipase AZ’s,. B. asper phospholipase Az is nonneurotoxic, but highly myotoxic, and it shows modest hydrophobicity in this region. Interestingly, hydropathy profiles of ammodytoxins A and B (LDSO’s of 0.021 and 0.58 pg/g, respectively, in mice) are very similar to B. asper phospholipase along its entire sequence. From 80 to 104 these profiles are virtually identical, with slight quantitative differences between residues 105 and 109 (see Fig. 6). This is the only region where ammodytoxins A and B differ in sequence (positions 105,108, and 109, see Fig. 5) and led Ritonja et al. (6) to suggest that a hydrophobic residue in position 105 and a basic residue in position 108 may be responsible for the increased toxicity of A over B. The nonneurotoxic B. asper phospholipase contains Lys and Met at these same relative positions, which makes it more like the less toxic ammodytoxin B, consistent with Ritonja’s hypothesis (6) that at least one part of the site of toxicity of phospholipase molecules is near here. While there is 67% amino acid identity between ammodytoxin B and B. asper phospholipase, there is nearly the same identity of B. asper with the basic subunit of crotoxin (64%). Yet, their hydropathy profiles differ substantially (Fig. 6). In an allied paper, Kini and Iwanaga (21) attribute a cationic site (+OO+++OO+) in monomeric phospholipase AZ’s around residues 79-87 as being involved in myotoxic activity of presynaptic neurotoxic phospholipase A,‘s. We observed no such site in B. asper phospholipase in the 79-87 region, although there is a +OO+++OOO region between residues 101 and 109. Smaller myotoxins lacking enzymatic activity, such as myotoxin a found in Crotalus viridis viridis venom, do contain the cationic site +OO+++OO+ (21). Using the sequence numbering system of Tsai et al. (22), B. asper phospholipase has a $2 charge score for residues 65, 70, 72, 73, and 90 (of their numbering system). In viperid-crotalid enzymes, those phospholipase A,‘s having a charge score of +3 or f4 were neurotoxic, whereas those with a charge score of 0 + 1 were nontoxic. A monomeric, nonneurotoxic phospholipase A, from the elapid Bungarus multicinctus also had a score of f2. However, ammodytoxin A and ammodytoxin B are identical to each other through residue 104, yet differ in toxicity by 30-fold in mice. Because of this sequence identity, they obviously have identical charge scores (+4) and hydropathy over this region and only differ near the carboxy-terminus. So, at least with these two structural analogs, only the three amino acid changes at positions 105, 108, and 109 can be responsible for toxicity differ-
MYOTOXIC
PHOSPHOLIF
ences. Of the four domains identified by Tsai et al. (22), which they suggest should be more hydrophobic in neurotoxic molecules, only the amino-terminus region (residues l-7) was more hydrophobic in the ammodytoxins than in our B. asper phospholipase. In two of the domains (59-64 and 90-95) they were either similar or identical and in one (69-74) the B. asper molecule was nearly three times more hydrophobic than the corresponding ammodytoxin regions. Therefore, the conclusion of Tsai et al. (22), that “the cationic sites of the neurotoxic enzyme are usually associated with regions of higher hydrophobicity than the corresponding regions of the nonneurotoxic enzymes” does not hold with nonneurotoxic B. asper phospholipase. ACKNOWLEDGMENTS We thank Dr. William Kruggel, manager of the Biotechnology Sequencing Facility at the University of Wyoming for his expertise in sequence analysis and Brett Lennon for running isoelectric focusing gels. This work was supported in part by the U.S. Army Medical Research and Development Command under Contract No. DAMD 17. 89-C-9007.
REFERENCES 1. Gutierrez, J. M., Ownby, 22,115128.
C. I,., and Odell, G. V. (1984) Toxicon
2. Mebs, D., and Samejima, Y. (1986) Tonicon 24,161l168. 3. Gutierrez, ,J. M., Ownby, C. I,., and Odell, G. V. (1984) Exp. Mol. Pnthol. 40, 367-379. 4. Lomonte,
B., and Gutierrez,
,J. M. (1989) Toxicon 27, 725-733.
5. Waite, M. (1987) Handbook of Lipid Research Ed.), Vol. 5, Plenum Press, New York.
(Hanahan,
D. J.,
’ “7
FROM
Bothrops
asper
325
W., Turk, V., and Gubensek, F. (1986) 6. Ritonja, A., Machleidt, Biol. Chem. Hoppe+Seyler 367,919-923. 7. Da Silva, N. J., Aird, S. D., Seebart, C., and Kaiser, I. I. (1989) Toxicon 27,763-771. 8. Aird, S. D., and Kaiser, I. I. (1985) Toxicon 23,361-374. 9. Aird, S. D., and Kaiser, I. I. (1985) Toxicon 23,11l13. 10. Pharmacia Fine Chemicals. Isoelectric Focusing, Principles and Methods, 1982. 11. Gutierrez, J. M., Lomonte, B., Chaves, F., Moreno, E., and Cerdas, L. (1986) Comp. Riochem. Physiol. 84C, 1599164. 12. Lomonte, B., Moreno, E., and Gutierrez, J. M. (1987) Toxicon 25, 947-956. 13. Needleman, S. B., and Wunsch, C. D. (1970) J. Mol. Biol. 48,443453. 14. Aird, S. D., Kaiser, I. I., Lewis, R. V., and Kruggel, W. G. (1985) Biochemistry 24,7054-7058. 15. Allen, G. (1981) Sequencing of Proteins and Peptides, Elsevier, New York. 16. Kyte, J., and Doolittle, R. F. (1982) J. Mol. Bid. 157, 105-132. 17. Traub, P., Mizushima, S., Lowry, C. V., and Nomura, M. (1971) in Methods in Enzymology (Moldave, K., and Grossman, L., Eds.), Vol. 20, pp 391-407, Academic Press, San Diego. 18. Lomonte, B., and Kahan, L. (1988) Toxicon 26,675689. 19. Aird, S. D., Kaiser, I. I., Lewis, R. V., and Kruggel, W. G. (1986) Arch. Biochem. Biophys. 249,296-300. 20. Kini, R. M., and Iwanaga, S. (1986) Toxicon 24,527-541. 21. Kini, R. M., and Iwanaga, S. (1986) Toxicon 24,895&905. 22. Tsai, I. H., Liu, H. C., and Chang, T. (1987) Biochim. Biophys. Acta 916,94-99. 23. Ritonja, A., and Gubensek, F. (1985) Biochim. Riophys. Acta 828, 306-312. 24. Forst, S., Weiss, J., Blackburn, P., Frangione, B., Goni, F., and Elsbach, P. (1986) Biochemistry 25, 4309P4314. 25. Maraganore, J. M., and Heinrikson, R. L. (1986) J. Riol. Chem.
261,4797-4804.
OF BIOCHEMISTRY
AND
RIOPHYSICS
Vol. 278, No. 2, May 1, pp. 319-325, 1990
The Amino Acid Sequence of a Myotoxic Phospholipase from the Venom of Bothrops asper Ivan I. Kaiser,*,l Jose Maria Gutierrez,?
Dorothy
Plummer,*
Steven D. Aird,*‘2 and George V. Odell$
“Department of Molecular Biology, University of Wyoming, Laramie, Wyoming 82071; tlnstituto Clodomiro Picado, Universidad de Costa Rica, San Jose, Costa Rica; and #Department of Biochemistry, Oklahoma State University, Stillwater, Oklahoma 74078
Received August 14, 1989, and in revised form December 5,1989
A myotoxic, basic phospholipase A2 (pl > 9.5) with anticoagulant activity has been purified from the venom of Bothrops asper, and its amino acid sequence determined by automated Edman degradation. It is distinct from the B. asper phospholipase A2 known as myotoxin I [Lomonte, B. and Gutierrez, J. M., 1989, Toxicon 27, 7251 but cross-reacts with myotoxin I rabbit antisera, suggesting that the proteins are closely related isoforms. To our knowledge, this is the first myotoxic phospholipase to be sequenced that lacks presynaptic neurotoxicity (iv LD5,, ^I 8 pg/g in mice). The protein appears to exist as a monomer, contains 122 amino acids, and fits with subgroup IIA of other sequenced phospholipase A2 molecules. Its primary sequence shows greatest identity with ammodytoxin B (67%), a phospholipase A2 presynaptic neurotoxin from Vipera ammodytes ammodytes venom. Hydropathy profiles of B. asper phospholipase and the ammodytoxins also show great similarities. In contrast, even though the amino acid sequence identities between B. asper phospholipase and the basic subunit of crotoxin remain high (64%), their hydropathy profiles differ substantially. Domains and residues that may be responsible for neurotoxicity are discussed. t) dew Academic Press,
Inc.
Bothrops asper represents one of the most hazardous snakes found in Central America. Its venom can induce extensive tissue damage, characterized by edema, hemorrhage, and myonecrosis (I). Basic phospholipases and phospholipase-like molecules isolated from the venom have been purified and shown to play an important role in promoting myonecrosis (2-4). They act in part by ’ To whom correspondence should he addressed. ’ Present address: Natural Product Sciences, Inc., 420 Chipeta Way, Suite 240, Salt Lake City, UT 84108. 000:1-9861/90 $3.00 Copyright ‘J11990 hy Academic Press, Inc. All rights oi’ reproduction in any form reserved.
affecting the integrity of the plasma membrane of muscle fibers [see Lomonte and Gutierrez (4) and references therein]. Unlike many snake venom basic phospholipases, the myotoxic phospholipase AZ’s from B. asper do not possess presynaptic neurotoxicity. We felt that some insight into structure-function relationships might be gained in sequencing a myotoxic, nonneurotoxic phospholipase, since we were unaware of any that had been sequenced. While the residues involved in the enzymatic activity of phospholipase AZ’s have been fairly well established (5), little is known about the domains responsible for neurotoxicity and myotoxicity. Sequence and hydropathy comparisons with the ammodytoxins isolated from Vipera ammodytes ammodytes venom indicate great similarity and are consistent with the conclusion of Gubensek and co-workers (6), who suggest that one of the neurotoxic domains may reside near the carboxy-terminus. MATERIALS
AND
METHODS
Crude, dried venom of I?. asper, obtained from adult specimens collected in the Atlantic region of Costa Rica, was kindly provided by the Instituto Clodomiro Picado, Costa Rica. Sephacryl S200, Mono S ionexchange column, and the FPLC’ system were purchased from Pharmacia. Reversed-phase chromatography was carried out on the FPLC system using either a Pharmacia C,/Cs 0.5 X 5-cm column for intact molecules or a Vydac 0.46 X 15 cm C,, column (5 wm) for peptide fractionation. SDS-PAGE (15%) gels were silver-stained as described (7). Phospholipase and lethality assays followed the methods of Aird and Kaiser (8, 9). Myotoxic phospholipase used for lethality assays was diluted in phosphate-buffered saline containing 0.1% bovine serum alhumin. Phospholipase assays were carried out in a pH-stat against I.tu-phosphatidylcholine in a reaction mixture containing Triton X-100 as described earlier (8). Isoelectric focusing gels were kindly run by Brett Lennon using procedures essentially as described (10). Myot.oxic
a Abbreviations used: FPLC, fast protein liquid SDS, sodium dodecyl sulfate; PAGE, polyacrylamide sis; PTH, phenylthiohydantoin; TFA, trifluoroacetic
chromatography; gel electrophoreacid. 319
320
KAISER
0.0 0
10
20
FRACTION
30
40
50
60
NUMBER (20ml)
FIG. 1. Gel filtration profile of 240 mg of crude B. asper venom on a 3 X 95-cm column of Sephacryl S200 superfine. Elution was carried out with 0.1 M sodium acetate (pH 4.0) at room temperature at a flow rate of 80 ml/h. Recovery of A,,, material was quantitative. The indicated fractions (34-38) were pooled, concentrated, and dialyzed against 50 mM sodium phosphate (pH 7.2), in preparation for subsequent cation exchange chromatography.
activity was studied by injecting 50 pg of the phospholipase im in the right gastrocnemius muscle of mice (16-18 g). Control mice received 0.1 ml physiologic saline solution. After 3 h mice were bled from the tail and the creatine kinase activity of plasma was quantitated according to the Sigma Technical Bulletin 520 (Sigma Chemical Co., St. Louis, MO). Creatine kinase activity was expressed in units per milliliter with 1 unit representing the phosphorylation of 1 nanomole of creatine per minute at 25°C. Anticoagulant activity on sheep plateletpoor plasma was determined as described (11). Immunodiffusion studies were carried out by using polyclonal anti-B. asper myotoxin I antisera (12). Sequencing employed an Applied Biosystems 470A protein sequencer with an Applied Biosystems Model 120A on-line analyzer for the analysis of the phenylthiohydantoin (PTH)amino acid derivatives (7). Sequence comparisons were made using PRONUC software on a micro VAX (VMS), with selected sequences added to the NBRF database. Alignments of protein sequences allowed for insertions and deletions with an optimization algorithm based on that of Needleman and Wunsch (13). The phospholipase was reduced and carboxyamidomethylated using a 12-fold excess of DTT over protein disulfide and a 5-fold excess of iodoacetamide over DTT (14). Fragmentation following reduction and alkylation was achieved by cleavage with (i) arginine-specific endoproteinase (Arg-C) from Boehringer-Mannheim, (ii) an endoproteinase that cleaves specifically at the N-terminal side of aspartate (Asp-N) from Boehringer-Mannheim, and (iii) cyanogen bromide, using a 50fold excess of CNBr over protein Met residues in 70% HCOOH and incubation at room temperature (ca. 22°C) for 21 h in the dark. Tryptophan was added as a scavenging reagent to protect protein Trp residues (15). Cleavage with Arg-C was carried out (6 U/mg protein) in 0.1 M ammonium acetate (pH 7.8) for 20 hat 37°C. Asp-N cleavage (1: 280 by weight) was in 50 mM sodium phosphate (pH 8.0) containing a final concentration of 0.5 M guanidine hydrochloride for 15 h at 37°C. Hydropathy profiles were generated using Beckman’s MicroGenie version 5.0 software, using the Kyte and Doolittle (16) option and a window of nine residues. RESULTS
AND
DISCUSSION
A myotoxic phospholipase AZ from B. asper was purified by a modification of the original procedure of Gutie-
ET AL.
rrez et al. (1). Crude venom was initially fractionated on Sephacryl S200 with the phospholipase of interest eluting in the major peak (Fig. 1). Indicated tubes were pooled, concentrated by ultrafiltration (Amicon, 5 kDa pore size), and dialyzed against 50 mM sodium phosphate, pH 7.2, at 4°C. Dialyzed material was applied to a Mono S column equilibrated in the same buffer and eluted with a NaCl gradient (Fig. 2). The recovered peak which eluted at 13 min and its trailing shoulder, showed the highest levels of phospholipase activity. When aliquots of recovered fractions from cation-exchange were run on 15% SDS-PAGE, a single major band was observed for the 13-min fraction (Inset, Fig. 2). On ureaPAGE gels however, two distinct bands of similar staining intensity were seen (not shown). The 13-min fraction eluted as a single symmetrical component from a C,/C,, reversed phase column in 0.1% TFA (not shown) and all molecules present exhibited pl values greater than 9.5 on isoelectric focusing. This component represented about 7% of the crude venom. Purified material from the cation-exchange column had a catalytic activity of 750 pmol/mg min, as determined from the slope of a line resulting from a plot of initial activity vs concentration. The material showed linear kinetics similar to those observed with the basic subunit of crotoxin when 0.9
i-
// // / R/ A
0
, 0
IO
20
30
MINUTES FIG. 2. Cation-exchange FPLC of pooled fractions from the S200 run. About 2.5 mg of dialyzed, pooled material from Fig. 1 dissolved in 50 mM sodium phosphate (pH 7.2) was applied to a 0.5 X 5-cm Mono S column and eluted with a linear gradient of NaCl (- -), buffered with 50 mM sodium phosphate (pH 7.2). The indicated material eluting at 13 min was collected, dialyzed against deionized water, and lyophilized. Inset: SDS-PAGE of 0.5 and 1 pg of the indicated pooled material run on a 15% polyacrylamide gel and stained with silver.
MYOTOXIC
PHOSPHOLIPASE
20
Ser
Leu
Ile
Glu
Phe Ala
Lys
Met
Ile
10 Leu
Glu
30 Trp
Gly
Gly
Gln
Gly
Pro
Lys
Asp
Ala
40 Thr
Gln
321
Bothrops asper
FROM
Arg
Leu
Pro
Phe Pro Tyr
Tyr Thr Thr N-terminus
Asp Arg Cys Cys N terminus
Phe
Val
His
Cys
Glu
Thr
Lys
Tyr
Gly
Cys Tyr
Gly
Lys
Leu Ser
Cys
Gly
Asn
Cys
Lys
Cys
Glu
50 Asp
Cys
Tyr
RC- 1
Pro
60 Lys
80
70 Thr
Asp
Arg
Tyr
Ser
Tyr
Ser
Arg
Lys
Ser
Gly
Val
Ile
Ile
Cys
Gly
Glu
RC-2 _
Gly
Thr
Pro
Cys
Glu
Lys
Gln
Ile
Met
Ala
110 Tyr
Pro
Asp Leu/
Phe
_
lx-1 DN 1
Cys
Asp
90 Lys
Ala
_
Ala
Ala
Val
Cys
Phe Arg
Glu
Asn
100 Leu Arg
Thr
Tyr
Lys
Lys
Arg
Tyr
Leu Cys
w-2 --__
CB-1
DN-2
Lys
Lys
Pro
120 Ala Glul Asp Lys/Pro CB- 1
FIG. 3.
Cys
The amino acid sequence of a myotoxic
carried out on a phosphatidyl choline substrate in a Triton X-IO@-containing reaction mixture (8). It had an iv LD,,, value in female mice of ca. 8 lug/g (9). Material in the later eluting shoulder of the 13-min fraction in Fig. 2 also consisted primarily of phospholipase molecules. On SDS-PAGE one major fraction was seen which comigrated with the 13-min fraction and had a catalytic activity of 430 pmol/mg min, when measured under conditions as described above. This material represented about 2% of the crude venom protein. It seems likely that the entire 13-min peak is a collection of phospholipase molecules with slightly different primary sequences. This is suggested by the presence of two amino acids at positions 113, 120, and 121. If the entire peak, shoulder and all, had been pooled and sequenced, it is probable that additional variant sequences would have been found. We also determined the relationship of the myotoxic phospholipase sequenced here with other phospholipases from B. asper venom. The myotoxic phospholipase induced a creatine kinase increment in plasma, reaching values of 433 + 36 units/ml (n = 4) [control values = 43 + 9 units/ml (n = 4)]. See the Materials and Methods section for details. This increment is quantitatively similar to the one observed after B. asper myotoxin I injection (11). It also exerted an anticoagulant effect on sheep platelet-poor plasma, similar to that observed with myotoxin I (11). Our phospholipase showed cross-reactivity with polyclonal antibodies raised to myotoxin I, although immunodiffusion analysis with rabbit polyclonal antisera showed only partial identity between our phospholipase and B. asper Myotoxin I. Finally, PAGE of
phospholipase
from B. asper.
crude venom on 12% urea gels (17) indicated at least four discrete bands representing basic proteins, which roughly correspond to peaks 4 through 7 as separated on CM-Sephadex (1). Lomonte and Kahan (18) demonstrated that monoclonal antibodies against B. asper myotoxin I reacted with these four proteins, suggesting that they are isoforms of myotoxic phospholipase A2’s. It appears that the myotoxic phospholipase which we sequenced compares to peak 4 material as identified by Gutierrez et al. (l), and is slightly less basic than myotoxin I present in peak 5 (1,4). This is based on the mobility of the myotoxic phospholipase sequenced here in urea-PAGE gels (not shown), when coelectrophoresed with known samples recovered from CM-Sephadex (4, 18). Our phospholipase is clearly different from myotoxin I, but it is a related isoform. This explains the lack of agreement between selected amino acid compositions reported by Mebs and Samejima (2) and Lomonte and Gutierrez (4) for myotoxic phospholipases from B. asper, and that calculated from the primary sequence determination of B. asper phospholipase A2 reported here. Studies in progress by Gutierrez and co-workers on myotoxins from Bothrops venoms should provide greater insight into the relationship among these molecules. The primary sequence of the myotoxic phospholipase from B. asper contains 122 amino acids and is shown in Fig. 3. Sequencing was carried out on the reduced and carboxyamidomethylated protein by a combination of direct sequencing from the amino-terminal end and sequencing of fragments. These were generated by enzymatic digestion with Arg-C and Asp-N and chemical cleavage with CNBr, and subsequently fractionated by
322
KAISER
ET AL.
(4
I 60
15
MINUTES
30
45
MINUTES
0
A-- / I
/’ 40
20
60
MINUTES FIG. 4. Separation of fragments teinase Asp-N, and (C) CNBr.
of myotoxic
phospholipase
from B. asper following
reversed phase column chromatography. After sequencing the amino terminal end (l-53) and two additional large fragments generated by Arg-C digestion [RC-1, (43-68) and RC-2, (69-97)], it was possible to search the protein sequence database for a closely related phospholipase since the nonoverlapping fragment 69-97 could be tentatively positioned by the location of Cys residues, whose positions are invariant in any one class of phospholipases. The ammodytoxins, from V. ammodytes, showed the greatest identity with B. asper phospholipase A, using this partial sequence (l-97) and were used as a
treatment
with (A) endoproteinase
Arg-C, (B) endopro-
guide to suggest further cleavage strategy. With only seven aspartate residues present in the ammodytoxins, the possibility of generating only three or four large peptides, plus several smaller ones by the Asp-N enzyme, seemed likely. Fractionation of the Asp-N digest yielded a relatively simple profile (Fig. 4B), yielding two large peaks that appeared homogeneous. Fragment DN-1 corresponded to a peptide which started at residue Asp-62. Its sequencing was continued through residue 81, which gave the confirming overlap at position 68-69. Sequencing was electively terminated on this fragment after resi-
MYOTOXIC Species Bothrops asper Vipera ammodytes ammodytes Vipera ammodytes ammodytes Agkistrodon blomhoffi Crotalus durissus terrificus Agkistrodon piscivorus piscivorus
PHOSPHOLIPASE
Protein Myotoxic Phospholipase Ammodytoxin B Ammodytoxin A Mamushi Phospholipase A2 Crotoxin, basic subunit Phospholipase A2
65
0 95 AAAVCFRENLRTYKKRYMAYPDLIFLCKKPAE/DK/PC AAAI CFRKNLKTYNHI AAAI CFRKNLKTYNYI AAAICFRDNLKTYKKRYMAYPDI VAAECLRRSLSTYKYGYMFYPD AVAI CLRENLDTYNKKYKAYFK
Bothrops
323
asper
5 1 SLIEFAKMILEETKRLPF SLLEFGMMILGETGKNPL SLLEFGMMILGETGKNPL H L L 0 F R K HLLQFNKMIKFETRKNAI SVLELGKMILQETGKNAI
15
10
M
l
K
K
M
T
G
45
40
35
30 25 20 PYYTTVGCYCGWGGOGQPKDATDRCCFVHDCCYGKL TSYSFYGCYCGVGGKGTPKDATDRCCFVHDCCYGNL TSYSFYGCYCGVGGKGTPKDATDR fSYAFYGCYCGSGGRGKPKDATDRCCFVHDCCYEKV PFYAFYGCYCGWGGQGRPKDATDRCCFVHDCCYGKL TSYGSYGCNCGWGHRGQPKDATDRCCFVHKCCYKKL 60 55 SNCKPKTDRYSYSRKSGVI PDCSPKTDRYKYHRENGAl PDCSPKTDRYKYHRENGAI TGCKPKWDDYTYSWKNDGI AKCNTKWDIYRYSLKSGYI TDCNHKTDRYSYSWKNKAI
FROM
K
E
P
V
50
CCFVHDCCYGNL
75 80 ICGEGTPCEKQICECDK VCGKGTSCENRICECDR VCGKGTSCENRI VCGGDDPCKKEICECDR TCGKGTWCEEQI ICEEKNPCLKEMCECDK
70
5
10 YMYYPD YRNYPD
F F S L
90
65
CECDR CECDR
15
20
LCKKES LCKKES LCSSKS RCRGPS KCKLP-
E E E E D
K K K T T
c c C c c
FIG. 5. Comparison of the amino acid sequence of the myotoxic phospholipase from B. asper with ot,her snake venom phospholipases. Identical residues are highlighted. References for sequences 2 through 6 are Ritonja et al. (6), Ritonja and Gubensek (23), Forst et al. (24), Aird scores for the myotoxic phospholipase with itself and related et al. (19), and Maraganore and Heinrikson (25), respectively. Optimization proteins are shown below, followed by the percentage identity. I?. asper, phospholipase, 734, 100%; V. a. ammo&es, ammodytoxin B, ~2,677;; V. a. ammodytm, ammodytoxin A, 545, 66 Y’,; A. blomhofi, mamushi phospholipase, 536, 65%; C. d. terrificus, basic subunit of crotoxin, ,509, 64%; and A. p. piscicorus, phospholipase A,, 501.62%.
due 81, since we had sufficient overlap of the earlier two peptides (RC-1 and RC-2) from the Arg-C digestion. The second large, homogeneous-appearing peak DN-2, gave the fragment predicted to start with Asp-89, which continued through Pro-111. This peptide revealed the presence of a Met at position 108. Cleavage with CNBr and reversed phase column chromatography of the products yielded the 109-122 fragment (CB-l), which gave a convenient three amino acid overlap with the carboxy-terminus of the DN-2 peptide. This final fragment also revealed the presence of a mixture of two amino acids at three different positions. Position 113 contained Leu: Phe at a molar ratio of l.EXX1.00; position 120, Glu:Asp at a molar ratio of 1.14:l.OO; and position 121, Lys:Pro at a molar ratio of 1.33:1.00. These were the only sites observed to have multiple amino acids present. With the partial sequence available and the ammodytoxin sequence as a guide, it was possible to predict nearly all of the cleavage sites for the endoproteinase Asp-N and CNBr. On the basis of amino acid composition, expected peptide size, and what appeared to be pure peptides in peaks DN-1 and DN-2 in Fig. 4B, as well as peak CB-1
in Fig. 4C, we fortuitously selected these peaks for initial sequencing. Since they yielded the necessary sequence information, additional fragments were not sequenced from these runs. In other instances fragment identification and location could be made after sequencing three residues in from the N-terminus, at which time a decision was made on whether to continue or terminate further sequencing. Figure 4 illustrates elution profiles of peptides following enzymatic and chemical hydrolyses. In Fig. 4A, peptide RC-3 corresponded to a peptide starting at Cys-28, which we sequenced through Gly-34 and aborted. RC-4 contained two peptides which were sequenced simultaneously; one started at Ala-91 and terminated at Arg-97; the other started at Gly-25 and continued through Gln33, when sequencing was aborted. Peptides RC-1 and RC-2 ran from Cys-43 through Arg-68 and Lys-69 through Arg-97, respectively (see Fig. 4A). With the ArgC endoproteinase we observed cleavages on the carboxyside of Tyr and Lys in addition to Arg, as noted before (19). The Arg-C endoproteinase used was not sequencing grade, in contrast with the Asp-N, with which we ob-
324
KAISER
IA
t30
I 0
20
40
60
60
100
120
RESIDUE NUMBER FIG, 6. Hydropathy profiles of (A) B. asper myotoxin phospholipase, (B) ammodytoxin A (solid line, 1) and B (broken line, 2) from V. a. ammodytes, and (C) the basic subunit of crotoxin from the venom of C. d. terr$icus. A span of nine consecutive residues was used to determine the profiles.
served no unexpected cleavages. Figure 4B shows the isolation of peptides DN-1 (Asp-62 through Cys-81) and DN-2 (Asp-89 through Pro-111). Figure 4C represents the fragments generated by CNBr cleavage of the reduced and alkylated phospholipase, with peak CB-1 corresponding to the terminal peptide (Ala-109 through Cys-122). It is unclear why the CB-1 peptide obtained by CNBr cleavage gave a single peak on reversed phase chromatography, yet differed in 3 out of 14 residues. The five most similar sequences to B. asper phospholipase A,, in our data base, included phospholipases from viperid and crotalid venoms (Fig. 5). These had similarity scores ranging from 552 (67% amino acid identity) with ammodytoxin B from V. a. ammodytes (a true viper) to 501 (62% amino acid identity) with the phospholipase A, from Agkistrodon piscivorus piscivorus (a pit viper). Between these two were ammodytoxin A, a homolog of ammodytoxin B, the phospholipase from the pit viper Agkistrodon halys blomhofi from Japan, and the basic subunit of crotoxin from the South American rattlesnake Crotalus durissus terrificus. The next 4O-odd sequences with the highest similarity scores (507-210) were all phospholipases from either viperids, crotalids, or elapids.
ET AL.
A number of papers have attempted to identify particular regions of the phospholipase molecules with a particular function such as neurotoxicity or myotoxicity. Kini and Iwanaga (20) recently proposed that monomeric presynaptic neurotoxins have a hydrophobic region between residues 80 and 110 that may play a significant role in the neurotoxicity of phospholipase AZ’s,. B. asper phospholipase Az is nonneurotoxic, but highly myotoxic, and it shows modest hydrophobicity in this region. Interestingly, hydropathy profiles of ammodytoxins A and B (LDSO’s of 0.021 and 0.58 pg/g, respectively, in mice) are very similar to B. asper phospholipase along its entire sequence. From 80 to 104 these profiles are virtually identical, with slight quantitative differences between residues 105 and 109 (see Fig. 6). This is the only region where ammodytoxins A and B differ in sequence (positions 105,108, and 109, see Fig. 5) and led Ritonja et al. (6) to suggest that a hydrophobic residue in position 105 and a basic residue in position 108 may be responsible for the increased toxicity of A over B. The nonneurotoxic B. asper phospholipase contains Lys and Met at these same relative positions, which makes it more like the less toxic ammodytoxin B, consistent with Ritonja’s hypothesis (6) that at least one part of the site of toxicity of phospholipase molecules is near here. While there is 67% amino acid identity between ammodytoxin B and B. asper phospholipase, there is nearly the same identity of B. asper with the basic subunit of crotoxin (64%). Yet, their hydropathy profiles differ substantially (Fig. 6). In an allied paper, Kini and Iwanaga (21) attribute a cationic site (+OO+++OO+) in monomeric phospholipase AZ’s around residues 79-87 as being involved in myotoxic activity of presynaptic neurotoxic phospholipase A,‘s. We observed no such site in B. asper phospholipase in the 79-87 region, although there is a +OO+++OOO region between residues 101 and 109. Smaller myotoxins lacking enzymatic activity, such as myotoxin a found in Crotalus viridis viridis venom, do contain the cationic site +OO+++OO+ (21). Using the sequence numbering system of Tsai et al. (22), B. asper phospholipase has a $2 charge score for residues 65, 70, 72, 73, and 90 (of their numbering system). In viperid-crotalid enzymes, those phospholipase A,‘s having a charge score of +3 or f4 were neurotoxic, whereas those with a charge score of 0 + 1 were nontoxic. A monomeric, nonneurotoxic phospholipase A, from the elapid Bungarus multicinctus also had a score of f2. However, ammodytoxin A and ammodytoxin B are identical to each other through residue 104, yet differ in toxicity by 30-fold in mice. Because of this sequence identity, they obviously have identical charge scores (+4) and hydropathy over this region and only differ near the carboxy-terminus. So, at least with these two structural analogs, only the three amino acid changes at positions 105, 108, and 109 can be responsible for toxicity differ-
MYOTOXIC
PHOSPHOLIF
ences. Of the four domains identified by Tsai et al. (22), which they suggest should be more hydrophobic in neurotoxic molecules, only the amino-terminus region (residues l-7) was more hydrophobic in the ammodytoxins than in our B. asper phospholipase. In two of the domains (59-64 and 90-95) they were either similar or identical and in one (69-74) the B. asper molecule was nearly three times more hydrophobic than the corresponding ammodytoxin regions. Therefore, the conclusion of Tsai et al. (22), that “the cationic sites of the neurotoxic enzyme are usually associated with regions of higher hydrophobicity than the corresponding regions of the nonneurotoxic enzymes” does not hold with nonneurotoxic B. asper phospholipase. ACKNOWLEDGMENTS We thank Dr. William Kruggel, manager of the Biotechnology Sequencing Facility at the University of Wyoming for his expertise in sequence analysis and Brett Lennon for running isoelectric focusing gels. This work was supported in part by the U.S. Army Medical Research and Development Command under Contract No. DAMD 17. 89-C-9007.
REFERENCES 1. Gutierrez, J. M., Ownby, 22,115128.
C. I,., and Odell, G. V. (1984) Toxicon
2. Mebs, D., and Samejima, Y. (1986) Tonicon 24,161l168. 3. Gutierrez, ,J. M., Ownby, C. I,., and Odell, G. V. (1984) Exp. Mol. Pnthol. 40, 367-379. 4. Lomonte,
B., and Gutierrez,
,J. M. (1989) Toxicon 27, 725-733.
5. Waite, M. (1987) Handbook of Lipid Research Ed.), Vol. 5, Plenum Press, New York.
(Hanahan,
D. J.,
’ “7
FROM
Bothrops
asper
325
W., Turk, V., and Gubensek, F. (1986) 6. Ritonja, A., Machleidt, Biol. Chem. Hoppe+Seyler 367,919-923. 7. Da Silva, N. J., Aird, S. D., Seebart, C., and Kaiser, I. I. (1989) Toxicon 27,763-771. 8. Aird, S. D., and Kaiser, I. I. (1985) Toxicon 23,361-374. 9. Aird, S. D., and Kaiser, I. I. (1985) Toxicon 23,11l13. 10. Pharmacia Fine Chemicals. Isoelectric Focusing, Principles and Methods, 1982. 11. Gutierrez, J. M., Lomonte, B., Chaves, F., Moreno, E., and Cerdas, L. (1986) Comp. Riochem. Physiol. 84C, 1599164. 12. Lomonte, B., Moreno, E., and Gutierrez, J. M. (1987) Toxicon 25, 947-956. 13. Needleman, S. B., and Wunsch, C. D. (1970) J. Mol. Biol. 48,443453. 14. Aird, S. D., Kaiser, I. I., Lewis, R. V., and Kruggel, W. G. (1985) Biochemistry 24,7054-7058. 15. Allen, G. (1981) Sequencing of Proteins and Peptides, Elsevier, New York. 16. Kyte, J., and Doolittle, R. F. (1982) J. Mol. Bid. 157, 105-132. 17. Traub, P., Mizushima, S., Lowry, C. V., and Nomura, M. (1971) in Methods in Enzymology (Moldave, K., and Grossman, L., Eds.), Vol. 20, pp 391-407, Academic Press, San Diego. 18. Lomonte, B., and Kahan, L. (1988) Toxicon 26,675689. 19. Aird, S. D., Kaiser, I. I., Lewis, R. V., and Kruggel, W. G. (1986) Arch. Biochem. Biophys. 249,296-300. 20. Kini, R. M., and Iwanaga, S. (1986) Toxicon 24,527-541. 21. Kini, R. M., and Iwanaga, S. (1986) Toxicon 24,895&905. 22. Tsai, I. H., Liu, H. C., and Chang, T. (1987) Biochim. Biophys. Acta 916,94-99. 23. Ritonja, A., and Gubensek, F. (1985) Biochim. Riophys. Acta 828, 306-312. 24. Forst, S., Weiss, J., Blackburn, P., Frangione, B., Goni, F., and Elsbach, P. (1986) Biochemistry 25, 4309P4314. 25. Maraganore, J. M., and Heinrikson, R. L. (1986) J. Riol. Chem.
261,4797-4804.
