Biochimica et Biophysica Acta, 1039 (1990) 81-89 81 Elsevier

BBAPRO 33653

Elegantin and Albolabrin purified peptides from viper venoms; homologies with the RGDS domain of fibrinogen

and von Willebrand factor

J a n i c e W i l l i a m s , B o g u s l a w R u c i n s k i , J o h n H o l t * a n d S t e f a n N i e w i a r o w s k i

Department of Physiology, Thrombosis Research Center and Macromolecular Analysis and Synthesis Laboratory, Temple University Medical School, Philadelphia, PA (U.S.A.)

(Received 5 October 1989) (Revised manuscript received 28 January 1990)

Key words: Disintegrin; Integrin; Glycoprotein I Ib-I l la ; Amino acid sequence; Recognition site; (Human platelet)

The RGD-containing peptides isolated from the venoms of the Viperidae constitute a new class of small cysteine-rich peptides of variable amino acid composition and biological activity (Huang, T.-F., et al. (1987) J. Biol. Chem. 262, 16157-16163; Gan, Z.R., et al. (1988) J. Biol. Chem 263, 19827-19832; Huang, T.-F., et al. (1989) Biochemistry 28, 661-668), which it is proposed by Gould et al. (unpublished data) that we call 'disintegrins'. These peptides bind to the glycoprotein IIb-Illa receptor on the platelet surface and inhibit aggregation induced by ADP, thrombin, platelet- activating factor and collagen. These peptides are also potent inhibitors of cell adhesion to fibrinogen (Knudsen, K.M., et al. (1988) Exp. Cell Res. 179, 42-49). We report the isolation of two further RGD-peptides from the venoms of Trimeserusus elegans and Trimeserusus albolabris, purified to homogeneity with high yield by a novel, rapid reverse-phase HPLC method. The primary structures of these two peptides were determined to be single polypeptide chains of 73 amino acids. Albolabrin differed from trigramin by eight residues whilst elegantin differed by 22 residues. The molecular mass of albolabrin calculated on the basis of amino acid sequence was 7574 Da and the pl similarly calculated was 4.27. The molecular mass of elegantin was calculated to be 7806 Da and the theoretical pl to be 4.69. RGD is maintained in the same position (51-53 AA) and all 12 cysteines are identical. Our data suggest that the presence of RGD, the conserved secondary and tertiary structure, are essential for the expression of biological activity by these peptides. Both peptides inhibited ADP-induced platelet aggregation. Extended homologies around the RGDS sequences in human von Willebrand Factor and bovine fibrinogen were found with both peptides.

Introduction

The glycoprotein (GP) l i b - I l i a complex on platelets [5,6] is a Ca2÷-dependent heterodimer which, on activated platelets, can bind one of four different ad- hesive proteins (i.e., fibrinogen, fibronectin, von Wil-

* Current address: Rorer Biotechnoiogy Inc., King of Prussia, PA 19406, U.S.A.

Abbreviations: One-letter codes have been used to represent the amino acids in this paper; vWF, von Willebrand Factor; GP, glyco- protein; PAF, platelet-activating factor; TFA, trifluoroacetic acid; BSA, bovine serum albumin; S-PE-E, S-(pyridylethyl)elegantin; PRP, platelet-rich plasma.

Correspondence: J. Williams, Protein Biochemistry Section, Throm- bosis Research Institute, Manresa Road, London SW3 6LR, U.K.

lebrand Factor (vWF) and vitronectin). The binding of fibrinogen primarily allows platelet aggregation, whereas fibronectin and vWF binding may also allow for the adhesion and spreading of platelets on the subendo- thelium. The GP l i b - I l i a complex is the most abun- dant platelet cell-surface protein; a normal platelet con- tains approx. 50000 of such complexes, about 70% of which are randomly distributed on the exposed platelet surface [7], while the rest are cryptic [8]. Glycoprotein lib is the subunit which binds Ca 2÷ [9,10], and putative Ca 2÷ binding sequences have been identified [11]. Calcium is essential for the integrity of the GP l i b - I l i a complex and thus for its interaction with fibrinogen [12,131.

Platelets stimulated by agonists such as ADP col- lagen or thrombin [14-17] express fibrinogen receptors associated with GP I Ib - I I Ia on their membrane surfaces. There is evidence that there are at least three platelet recognition sites in the fibrinogen molecule: (1)

0167-4838/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

82

the RGDF sequence of the Aa-chain 95-98; (2) the RCDS sequence of the Aa-chain 572-575; (3) the carboxyterminal dodecapeptide of the fibrinogen gamma chain [18-20]. The synthetic peptides corresponding to these putative fibrinogen binding sites block fibrinogen binding to platelets and ADP-induced platelet aggrega- tion at micromolar concentrations (4-100/LM).

Recently, trigramin, a 72 amino acid peptide purified from Tremeserurus gramineus snake venom, blocked platelet aggregation and binding of 125I-labelled fibrino- gen to ADP-stimulated platelets ( K i = 2 - 1 0 -8 M), 125I-labelled trigramin also bound to stimulated plate- lets with a K d value of 2.10 -8 M [1]. Binding of trigramin to platelets was blocked by monoclonal anti- bodies against GP l ib- I l ia and by high concentrations of RGDS or the fibrinogen gamma-chain pentade- capeptide. Both the presence of RGD and the sec- ondary structure of the molecule have been postulated as responsible for the biological activity of this peptide [1]. The primary structure of this peptide has been published [3], as has the primary structure of a related peptide from the venom of the viper Echis carinatus [2].

In this paper we describe the purification, complete primary sequence and comparative biological activity of two further peptides from Trimeresurus albolabris and Tr. elegans [21].

Materials and Methods

Reagents The lyophilised venoms of Tr. albolabris and Tr.

elegans were purchased from Latoxan (Rosans, France) and Sigma (MO, U.S.A.). Endoproteinases Lys-C and Asp-N were purchased from Boehringer Mannheim Bio- chemicals. Endoproteinase GIu-C (V8 proteinase) was from Pierce Chemical. Chymotrypsin was obtained from Worthington Biochemicals and carboxypeptidase Y was from Sigma. C-18 silica matrix columns were from Vydac and The Separations Group, CA and the Nucleosil C-18 column was from Phenomenex. All other reagents were analytical or sequencing grade.

Protein sequence analysis Protein sequence homology searches were performed

using the Sequence Analysis Software Package (Version 6 February 1989) from the Genetics Computer Group (GCG), U.W. Biotechnology Center (Madison, WI, U.S.A.).

Purification of proteins HPLC purifiation of elegantin and albolabrin, from

the whole venoms of Trimeresurus elegans and Tr. al- bolabris, was by single- or multiple-step reverse-phase chromatography on a wide-pore C-18 silica matrix col- umn (2.1 mm × 15 cm, Vydac). Active fractions were identified by testing for the inhibition of ADP-induced platelet aggregation in platelet-rich plasma (PRP).

20-40 mg of whole lyophilized venom protein was dissolved in 200-400 /~1 of 0.1% trifluoroacetic acid (TFA), stored on ice for 10 min, then centrifuged at 15 000 rpm for 2 min to remove any insoluble material. The supernatant was absorbed onto the C-18 reverse- phase column. For Tr. elegans venom (20 mg), the elution gradient used was 0.1% TFA containing: 0-15% acetonitrile (4% per rain) followed by 15-20% aceto- nitrile (0.25% per min) followed by 20-30% acetonitrile at 2% per min. For Tr. elegans purification was a single column procedure. For Tr. albolabris venom (40 mg) the gradient used was 0.1% TFA containing: 0-20% acetonitrile (2% per min) followed by 20-40% aceto- nitrile (0.65% per rain). Active peaks for Tr. albolabris, identified by testing for the inhibition of ADP-induced platelet aggregation, were freeze dried, pooled (in 0.1% TFA) and repurified to homogeneity on an identical C-18 column using a gradient containing 0.1% TFA: 0-40% acetonitrile (4% per min), followed by 40-55% acetonitrile at 0.5% per min.

Pyridylethylation of proteins Pyridylethylation of elegantin and albolabrin was

carried out by adding 1 /~1 of vinylpyridine to the reduced protein (50 #g in 99 #1 of 6 M guanidine hydrochloride, 4 mM EDTA, 0.1 M Tris-HC1 (pH 8.5) and 4 mM dithiothreitol). The reaction mixture was incubated for 2 h at 22 o C, in the dark under argon, first for reduction and then a further 2 h for pyridylethyla- tion. Modified protein was isolated free of reagents by reverse-phase HPLC, on a C-18 column (Vydac), in 0.1% trifluoroacetic acid with acetonitrile as organic modifier.

S-Methylation of proteins Dry protein (50 /~g) was dissolved in 200 /tl of a

solution containing 6 M guanidine hydrochloride, 0.06 M Tris-HC1 (pH 8.6), 5 mM EDTA and 3 mM dithio- threitol. After incubation in the dark for 1 h under argon, 2/~1 of methyl iodide was added. 20 min later the addition was repeated and the solution agitated periodi- cally over a 1 h period. Finally the S-methylated protein was isolated from the reaction mixture by reverse-phase HPLC on a C-18 silica matrix column.

Amino acid sequencing Automated NH2-terminal sequencing was performed

on a gas-phase sequencer (Applied Biosystems, Model 470A) coupled to an on-line HPLC for identification of PTH amino acids (Applied Biosystems, Model 120A). These instruments are operated routinely by the Macro- molecular Analysis and Synthesis Laboratory of the Temple University Health Sciences Center. Standard protocols of the manufacturer were followed with re- gard to both Edman degradation and separation of PTH-amino acids by HPLC. Cysteine was detected as S-(pyridylethyl)cysteine or S-methylcysteine.

Amino acid analysis Amino acid analysis was carried out by the Protein

Chemistry Facility of the Wistar Institute, under the direction of Dr. D.W. Speicher. Hydrolysis (Vapor phase 6 M HC1 with 1% phenol, 1 h at 160°C) was followed by a manual derivatization with phenylisothiocyanate and separation of amino acids by reverse-phase HPLC.

Generation of overlapping peptides by proteolytic cleavage S-Pyridylethyl or S-methyl proteins were incubated

at 37 °C (or 22°C where appropriate) in the presence of the following proteinases (final conditions are given): V8 proteinase, 6% (w/w) in 0.03 M sodium phosphate (pH 7.8), 5 mM EDTA, for 17 h; chymotrypsin 5% (w/w) in 0.2 M ammonium acetate (pH 7.2), for 5 h; endoproteinase Lys-C, 3% (w/w) in 0.1 M ammonium bicarbonate (pH 9), for 17 h; endoproteinase Asp-N, 3% (w/w) in 0.05 M sodium phosphate (pH 7.8), for 20 h. COOH-terminal sequencing with carboxypeptidase Y was carried out essentially as described by Huang et al. [3]. S-(Pyridylethyl)elegantin (S-PE-E) or S-(pyri- dylethyl)albolabrin (S-PE-A) was incubated with 5% (w/w) of carboxypeptidase Y in 0.1 M pyridine and 0.3 M acetic acid (pH 5.5), at 37 ° C. After 0, 10, 20 and 30 min, aliquots were withdrawn into pyrolyzed glass tubes, stored for the duration of the experiment on ice, heated to 100°C to inactivate the enzyme and dried under vacuum. Amino acids released were quantitated by amino acid analysis, using the recovery of 250 pmol of hydroxyproline to correct for losses. The exact con- centration of elegantin or albolabrin was determined by hydrolysis of an aliquot of the protein solution used for cleavage.

Preparation of human platelet suspensions Platelet-rich plasma was prepared essentially as de-

scribed in Huang et al. [1]. In brief, blood collected in sodium citrate (pH 7.35), final concentration 0.35%, was centrifuged at room temperature for 10 min to collect platelet-rich plasma (PRP).

Protein estimations All protein estimations were performed using a mi-

croscale versions [22] of the Lowry method [23].

125I-labelling of proteins RadiolabeUing of proteins with 125I was carried out

with chloramine T using a modification of the method of Hunter as detailed by Rucinski et al. [24]. In brief, the following were added to 10/~g of protein dissolved in 40/xl of phosphate-buffered saline (pH 7.5): 20/~g of chloramine T (1 mg/ml); 200/~g of sodium bisulphate; 200 #g potassium iodide and 50/xl of 10% bovine serum albumin (BSA). The incubation time of the chloramine T with the protein was 25 s. For separation of labelled protein from free iodine a Sephadex G-25 column was used. Incorporation of 125I w a s 25-40% of the total added.

83

Immunoprecipitation with rabbit anti-elegantin polyclonal antibody

Immunoprecipitation of a25I-elegantin and 125I-al- bolabrin by rabbit anti-elegantin polyclonal antibody was performed as described previously by Rucinski et al. [24]. In brief, 50 /~1 of normal rabbit serum (NRS) was added: 100 /~1 of suitably diluted 125I-peptide (1001300); 100/~1 of suitably diluted anti-elegantin poly- clonal antibody; 300 /xl of Tris-saline buffer (pH 7.4) containing 0.1% BSA and 0.05% Triton X-100. Anti- elegantin antibody was diluted with Tris-buffered saline containing BSA 0.1% (pH 7.4) over the dilution range 1 : 10 to 1 : 2000. After mixing and overnight incubation at 4°C, 50 /xl of 2% bovine 7-globulin and 400 /~1 of saturated ammonium sulphate (22°C) were added at room temperature. After 20 min incubation at 22 ° C, the sample was centrifuged at 3000 rpm for 30 min in a Hill Scientific MC15 benchtop centrifuge. Supernatant and pellet were separated and counted.

Preparation of rabbit anti-elegantin polyclonal antibody Peptide was administered by multiple subcutaneous

injections, beginning with initial injections of a total of 50 #g of elegantin, in 1 ml Freunds complete adjuvant/ saline (1 : 1). The rabbit was boosted with an additional 50/~g of peptide three times, after three successive 10 day intervals.

R e s u l t s

HPLC purification of elegantin and albolabrin A single active fraction was obtained from a com-

plete profile of Tr. elegans venom (Fig. 1), identified by testing for the inhibition of ADP-induced platelet ag-

,<

20

0 20.0 30.0 40.0

Minutes

Fig. 1. HPLC purification of elegantin from the whole venom of Tr. elegans on a wide-pore C-18 silica matrix column (Vydac). Column 1: 40 mg of whole lyophylized venom protein was dissolved in 400 #l of 0.1% TFA, stored on ice for 10 min then centrifuged at 15000 rpm for 2 rain to remove any insoluble material. The gradient used was 0.1% TFA containing: 0-15% acetonitrile (4% per min) followed by 15-20% acetonitrile (0.25% per rain) followed by 20-30% acetonitrile (2% per min). The active fraction ( x ) was identified by testing of aliquots from the complete profile for their ability to inhibit ADP-induced platelet aggregation in PRP. Active protein recovered, represented

0.2% of the total protein applied.

o O6

o.4 o . 0 <

02

/ 0.0 , ,

0.0 lO.O

84

gregation. The fraction was found to contain a single NH2-terminal sequence and was judged pure by this criteria. Elegantin was purified in a single chromato- graphic step and was estimated to constitute 0.25% of the total venom protein applied.

A single active fraction was obtained from the com- plete HPLC profile of Tr. albolabris venom (Fig. 2), identified as above. This fraction was repurified on the same C-18 HPLC column (Fig. 3). The active fraction contained a single protein species as shown by N H 2- terminal analysis and the recovery represented approx. 0.5% of the initial venom protein from this source.

Sequencing of elegantin S-(Pyridylethyl)elegantin was used for most experi-

ments. For chymotryptic cleavage and COOH-terminal sequencing with carboxypeptidase Y S-methylelegantin was used. Automated NH2-terminal sequencing of the intact protein yielded a single sequence of 35 clear residues later found to be positions 3-37 (Fig. 4a). The bulk of the protein sequenced (90%) lacked two N H 2- terminal reisdues, which were found in a peptide ob- tained after proteolytic cleavage with endoproteinase Lys-C. Cleavage with V8 proteinase was undertaken with S-PE-E. Fractions obtained from either the Nucleosil C-18 column or the small bore Vydac C-18 column, gave only peptides with sequences near the N H 2- and COOH-termini (Fig. 4a). Cleavage with en- doproteinase Lys-C was carried out also using S-PE-E. All peaks collected from the C-18 Vydac column yielded sequence information consistent with the final sequence shown, not all of which are shown in Fig. 4a.

Chymotryptic cleavage was undertaken with the aim of finding an additional overlap peptide containing

1.0

0.8

0 Q6

~ o4 L

0.2

OD ~ 0.0

100

• ~ 4 0 8 "4

- I

-( 20 . i

- i

~o 10.0 200 3 0 0 4 0 0

Minutes

Fig. 2. HPLC purification of albolabrin from the whole venom of Tr. albolabris. Column 1: wide-pore C-18 silica matrix (Vydac). 20 mg of whole iyophilized venom dissolved in 0.1% TFA, treated as for Tr. elegans. The gradient was 0.1% TFA containing 0-20% acetonitrile (2% per rain) followed by 20-40% aeetonitrile (0.65% per min). A single active peak ( x ) was identified by testing for the inhibition of ADP-induced platelet aggregation by fibrinogen. This substantially purified fraction (×) was lyophilized and subjected to further purifi-

cation.

I.C

0.~

0 Q6

8 c

i i i • , , i i i L , , t L i , , i i i , i ,

Minutes

100

80

60 ~ b

40

20

~ l A I 0 30.0 40.0

Fig. 3. HPLC repurification of albolabrin from the active fraction collected from column I (Fig. 2). Column 2: wide-pore C-18 silica matrix (Vydac). Two identical fractions from two first columns were pooled and repurified using the following gradient in 0.1% TFA:0- 40% acetonitrile (4% per min) followed by 40-55% acetonitrile (0.5% per min). Platelet inhibitory activity was found to be associated with

the major peak ( × ).

residues 41-44. The digest of S-methylelegantin yielded a number of fractions (CHT-1 to -11) by HPLC (Vydac) which were sequenced. All sequences were as expected, except that four previously undetected COOH-terminal residues were found (Fig. 4a). However, no sequences containing residues 41-44 were identified. A sample of S-PE-E was therefore digested with endoproteinase Asp-N. Among the peptides found (Asp-N-1 to -8) was the overlap fragment Asp-N-8 (Fig. 4a). The COOH- terminal sequence was further confirmed by cleavage of S-methylelegantin with carboxypeptidase Y. The release of amino acids were consistent (results not shown) with the results of C-terminal sequencing of peptide frag- ments.

Sequencing of albolabrin Fig. 4b shows the completed sequence of albolabrin

and the overlapping peptides generated. All experiments were carried out with S-(pyridylethyl)albolabrin. Auto- mated NH2-terminal sequencing of the intact protein yielded a single sequence comprising the first 29 re- sidues (Fig. 4b). The first 11 residues were also con- firmed on a separate occasion. To obtain further infor- mauon, protein was cleaved with V8 proteinase. The fractionated material (Supelco wide-pore C-8) yielded numerous peaks which were sequenced until the entire molecule had been covered (Fig. 4b). In order to pro- vide overlapping peptides, cleavage was undertaken with chymotrypsin. Five fractions (of nine) were collected and sequenced (CHT-1 to CHT-6, Fig. 4b). Cleavages with the V8 proteinase were at Asp or Glu as expected, except for one at Ser-63. An anomalous cleavage by V8 proteinase also occurred in this region of trigramin, namely at Gly-65 [3]. The COOH-terminus was con- firmed essentialy as described for elegantin, the release of amino acids agreeing well (results not shown) with

I0 20 30 40 50 60 70

EAGEECDCGSPENPCCDAATCKLRPGAQCADGLCCDQCRFKKKRTICRRARGDNPDDRCTGQSADCPRNGLYS

GEECDCGSPENPCCDAATCKLRPGAQCADGLCCDQ... .........................................................................

CDCGSPENPCC... QSADCPRNGLY... V8 -4 AATCKLRPGAQCAD -5

.........................................................................

LRPGAQCADGLCCDQCRFK LYS-C -8 RTICRRARGDNPDDRCTGQSADCPR -4

EAGEECDC... -3 GEECDCGSPENPCCDAATCK -2

.........................................................................

SADCPRNGLY CHT -ii .........................................................................

DQCRFKKKRTICRRARG ASP-N -8 DAATCKLRPGAQCA -7

DCPRNGLYS -6 GEEC -I

DCGSPENPCC DGLCC -5 .........................................................................

S-PE-EL

85

i0 20 30 40 50 60 70

EAGEDCDCGSPANP6CDAATCKLLPGAQCGEGLC6DQCSFMKKGTICRRARGDD£DDYCNGISAGCPRNPLHA

EAGEDCDCGSPANPCCDAATCKLLPGAQC... .........................................................................

MKKGTIC CHT -I EAGEDCD... CCDQCSF CHT -2 EAGEECDCGSPENPC... RRARGDDLDDY SAGCPRNPL CHT -3

CNGISAGCPRNPL CHT -5 KLLPGAQCGEGL CHT -6

.........................................................................

CGSPANPCCD V8 -2 AATCKLLPGAQCGE V8 -5

GLCCDQCSFMKKG... V8 -9 QCSFMKKGTICRRARGDDLDDYCNGIS... V8 -8

AGCPRNPLHA V8 -4 .........................................................................

S-PE-AL

Fig. 4. Amino acid sequence of elegantin (a) and albolabrin (b). Residues 1-37 of S-PE-EL (a) or residues 1-29 of S-PE-AL (b) were determined by Edman degradation of the intact molecule. Confirmation and extension of these results was obtained by analysis of fragments produced by proteolytic cleavage by: chymotrypsin (CHT), S. aureus V8 proteinase (V8) and endoproteinases Lys-C (LYS-C) and Asp-N (ASP-N). Residues 1 and 2 of S-(pyridylethyl)elegantin were confirmed by sequencing of Lys-C-3 produced by Lys-C proteolysis. The ratio of Lys-C-3 to Lys-C-4 was 1 : 9. S-PE-EL denotes S-(pyridylethyl)elegantin; S-PE-AL denotes S-(pyridylethyl)albolabrin. Absence of dots indicate that no new residue was detected in at least two further cycles of Edman degradation. All peptide sequences (some not shown) were consistent with the structure presented.

other C-terminal peptides sequenced. The molecular mass of alb61abrin, calculated on the basis of the amino acid sequence, was 7574 Da and the pI similarly calcu- lated was 4.27. The molecular weight of elegantin was calculated to be 7806 Da and the theoretical p I to be 4.69.

Homology with trigramin and echistatin As can be seen in Fig. 5, the cysteine sequence is

completely conserved in albolabrin, elegantin and tri- gramin. Compared to trigramin, the degree of homology for albolabrin was 90% and for elegantin it was 69%. When conservatively substituted amino acids are taken

Alb.

Trg.

Elg.

EAGEDCDCGSPANPCCDAATCKLLPGAQCGEGLCCDQCSFMKKGTICRRARGDDLDDYCNGISAGCPRNPLHA

I IEE V I R F

EAGEECDCGSPENPCCDAATCKLRPGAQCADGLCCDQCRFKKKRTICRRARGDNPDDRCTGQSADCPRNGLYS

Ech, ECESGPCCRNCKF LKEGT I CKRARGDDMDDYCNGKTCDCPRNPHKGPAT

Fig. 5. Comparative amino acid sequences of albolabrin, elegantin, trigramin and echistatin. The sequence of albolabrin (Alb) was compared with

trigramin (Trg) and elegantin (Elg). For trigramin only those amino acids which differ from albolabrin are indicated. For echistatin (Ech) the

shorter sequence has been aligned and compared relative to the ROD (Arg-Gly-Asp) sequence in the longer peptides. * indicates conservative

amino acid substitutions (codons); and * * indicates nonconservative substitutions.

86

Albolabrin

H-vWF

31 EGLCCDQCSFMKKGTICRRARGDDLDDYCNGISA.GCPRNP 70 ## ##* # * * *### * # **# ##

1724 EGECCGRCLPSACEVVTGSPRGDSQSSWKSVGSQWASPENP 1764

Elegantin

H-vWF

31 DGLCCDQCRFKKKRTICRRARGDNPDDR.CTGQSADCPRNG 70 *# ##* # * *### * * # *# #

1724 EGECCGRCLPSACEVVTGSPRGDSQSSWKSVGSQWASPENP 1764

Echistatin

H-vWF

31 SGPCCRNCKFLKEGTICKRARGDDMDDY.CNGKTCDCPRNP 70 # ## # * * *### * * # *# ##

1724 EGECCGRCLPSACEVVTGSPRGDSQSSWKSVGSQWASPENP 1764

Albolabrin 34 CCDQCSFMKKGTIC..RRARGDDLDDYCNGISAG 72 # # *# *#* *#### , , #, *#

Fn-A-B 48 CPSGCRMMVLETFGGDGHARGDSVSQ.GTGLAPG 80

Elegantin 31 CCDQCRFKKKRTIC..RRARGDNPDDRCTGQSAD 65 # ##* #* *#### * ## **

Fn-A-B 48 CPSGCRMMVLETFGGDGHARGDSVSQ.GTGLAPG 80

Fn-A-B 64 GHARGDSVSQGTGLA 78 # *#### # ***

H-vWF 1741 GSPRGDSQSSWKSVG 1755

Fig. 6. Homology between albolabrin or elegantin and von Willebrand Factor (H-vWF) or the A-chain of bovine fibrinogen (Fn-A-B). Searches

were conducted using the GCG protein database using the Wordsearch, Segments and Bestfit programmes. The sequences 31-70 of aibolabrin and elegantin, and the equivalent sequence in echistatin were compared with H-vWF 1724-1764 by the Bestfit programme. Similarly, the amino acid sequence containing RGDS of Fn-A-B (48-80) was aligned with albolabrin and elegantin sequences 34-66. A short region of high homology was found by comparing H-vWF (1726-1764) and Fn-A-B (48-80). # , indicates identical amino acids and *, indicates conservatively substituted

amino acids (codons).

into account, these figures rise to 95 and 79%, respec- tively. The RGD sequence is positioned identically be- tween cysteines in each case. Albolabrin has only eight amino acid substitutions compared with trigramin (Fig. 4). Elegantin has 17 amino acid changes in the C-termi- nal portion which contains the RGD sequence. The echistatin sequence is also shown in Fig. 5 for compari- son [2]. Cysteines are identical on the overlapping se- quence except for an additional cysteine substituted at position 58. Comparing the overlapping sequence in echistatin with the other three peptides, shows 15 amino acid substitutions and an overall degree of identity of 67%.

Homology with other proteins The elegantin and albolabrin sequences were com-

pared to sequences of known peptides by the Protein Identification Resource, National Biomedical Research Foundation, Georgetown University Medical Center, Washington D.C. Using the Mutation Data Matrix, with cysteine scoring six in order to allow the non-cy- steine amino acid sequence of the molecule to be com- pared, no significant overall sequence homologies were found, nor were any homologies containing the RGD

sequence. However, the tetrapeptide sequence LLPG, present in albolabrin (LRPG in elegantin, LIPG in trigramin), was found to be present in both human and bovine fibronectin (1295-1303) and in chicken fibronectin at position 59-67 (TNLLPGTEY). The se- quence LNPG also exists in the cell-binding domain of bovine fibronectin: TNLNPGTEY (168-176).

Using the Align programme, elegantin and al- bolabrin (residues 7-33) showed 13 and 12 identities, respectively, out of a possible 26 matches with residues 1555-1581 of human vWF. This compares with 12 identities shown previously for trigramin over the same sequence [3]. Residues 1555-1581 of human vWF are located in the putative GP l i b / I l i a binding C-terminus of this protein 1365-2050 [25,26], which also contains RGD at position 1744-1746.

Further searches were made by us, using the Se- quence Analysis Software Package (GCG), with sections of the C-terminal portion (31-70) of either protein, which established two further homologies of greater interest; the first with human von Willebrand Factor between residues 1724 and 1764 (Fig. 6). Bestfit of albolabrin with this segment (Fig. 5) showed 47.5% similarity, whilst elegantin and echistatin showed 42.5%

87

80

g 60

i

40 [3

° \ "~ 2(3 D .

¢

O I I I I I I I I I I I I I I I I I I I I I I I I I I

10 100 1000 10000 Ant ibody (log. scale)

Fig. 7. Immunoprecipitation of 12SI-elegantin and 12SI-albolabrin by rabbit anti-elegantin polyclonal antibody. Incubation mixture: 50/xl of normal rabbit serum; 100 /~l ]25I-peptide; 100 /~l rabbit anti- elegantin antibody; 300/xl Tris-saline buffer (TBS) (pH 7.4) contain- ing 0.1% BSA and 0.05% Triton X-100. Incubation was overnight at 4°C. Next, 50 /xl of 2% bovine y-globulin and 400 /*1 of saturated ammonium sulphate were added. After 20 min at 22°C the sample was centrifuged at 3000 rpm for 30 min; the supernatant and pellet were counted. Nonspecific binding was 4.5%, maximal specific bind- ing was 50% for elegantin and 1.5% for albolabrin at an antibody

dilution of 1 : 1000. ~ , 125I-elegantin and El, 125I-albolabrin.

similarity. The ratio given for each was 0.415, 0.517 and 0.345, respectively. The second homology was found with sequence 31-72 and the bovine fibrinogen Aet- chain [27] between residues 48 and 80 of a fragment (New: A05294). This homology also contains an aligned RGDS sequence (Fig. 6). Percent similarities for al- bolabrin and elegantin were 58.1 and 51.6, respectively; with ratios of 0.487 and 0.619. When the human fibrinogen Aa-chain was examined by the Bestfit pro- gramme, no extended homology beyond the RGDS

4C

0

~ 30"- - • •

~ 2o ~ o o ~

_ w

I I I I I I I I I I I I I t I I I I i i i I i l l l l i , , , , , , J 1 10 100 1000

Protein conc (log. scale, ng/ rn l )

Fig. 8. Displacement of 1251-elegantin binding to anti-elegantin anti- body by elegantin and albolabrin. The antibody was diluted 1:2000 with Tris-buffered saline containing BSA 0.1% (pH 7.4). The follow- ing mixture was prepared: 50 lal of normal rabbit serum; 100/~l of 12SI-elegantin; 100/d of serially diluted albolabrin or elegantin (range 3-10/xg/ml); 100/~l anti-elegantin antibody diluted 1 : 2000; 200/~l of Tris-saline buffer (pH 7.4) containing 0.1% BSA and 0.05% Triton X-100. The remaining procedure is as described in Fig. 7. Maximum specific binding at this antibody dilution was 44.5%, nonspecific binding was 5.5%. Competition with: o , elegantin and O, albolabrin.

100

8O

6O

4O

20!

0 10 20

E c h A I

I00 200 5(X) I000 nrnolor (log scale)

d e t e r m i n e d at 10 #M ADP

Fig. 9. Inhibition of ADP-induced (10 /~M) platelet aggregation by fibrinogen in platelet-rich plasma (PRP) by albolabrin (A) elegantin (e) and echistatin (o). PRP was prepared as described in Materials and Methods. At least two inhibition curves for different disintegrins were determined within the same experiment, n = 4-6. IC50 is indi-

cated by the arrows.

sequence was found. When the bovine Aet-chain frag- ment 49-80 [27] was cross-matched by the Bestfit pro- gramme (GCG) to the human vWF segment 1726-1764, a 15 residue homology with 66.7% similarity and a ratio of 0.687 was identified (Fig. 6).

Immunoprecipitation of 125I-elegantin and ~ 25I-albolabrin by rabbit anti-elegantin polyclonal antibody

The results of these experiments can be seen in Figs. 7 and 8. At an antibody dilution of 1 : 500, albolabrin showed only 6% immunoprecipitation, whereas elegan- tin showed 50%. At antibody dilutions of 1:1000 and 1:2000 albolabrin showed no cross-reactivity, whereas elegantin showed 42 and 25% immunoprecipitation, re- spectively (Fig. 7). Using ]25I-elegantin, the binding of unlabelled albolabrin and elegantin to the rabbit anti- elegantin antibody was compared. No competition was found between albolabrin and mSI-elegantin at an anti- body dilution of 1:2000, whereas elegantin showed competition over the range used (Fig. 8).

Inhibition of ADP-induced platelet aggregation in plate- let-rich plasma by albolabrin, elegantin and echistatin

The results of these experiments can be seen in Fig. 9. The IC5o values obtained with PRP at a concentration of 10 ttM ADP were 220 nM for albolabrin, 136 nM for elegantin and 56 nM for echistatin. This compares with a published value of 200 nM for trigramin [ll.

Discussion

The RGD sequence is a cell-recognizing domain ex- isting in a variety of adhesive proteins such as fibrino- gen, vWF, fibronectin and vitronectin [28]. It is well established that RGD-containing peptides can inhibit both fibrinogen [19,20] and vWF binding [29] to the

88

GPIIb- I I Ia complex. The RGD sequence can take very different conformations in different proteins [30,31]. Ruoslahti and Piershbacher have suggested that the RGD sequence may serve as a shared binding site, whereas specificity may be generated by a second bind- ing site unique to each protein ligand. Alternatively, the specificity could reside in the conformation of the RGD tripeptide modulated by surrounding sequences [28]. Theoretical studies by Reed et al. [29] have also shown that the sequence RGD is generally found to be associ- ated with a series of probable beta-bends, which would result in a highly ordered and unusual structure. It has already been demonstrated that a certain conformation of trigramin is a prerequisite for its binding toward this complex, since reduction of the trigramin molecule re- suits in a loss of ability to inhibit fibrinogen [1] and vWF binding to activated platelets [3].

We have described, a rapid novel method of purifica- tion of the RGD peptides albolabrin and elegantin from whole lyophylized viper venoms. Whereas previous methods involved several chromatographic steps, the present method allows for purification to homogeneity of active peptide using one or two rapid reverse-phase HPLC steps. This greatly reduces the possibility of proteolytic cleavage of the peptide during its separation from venom proteinases and significantly improves pro- tein yield. This approach has already been used to provide milligram amounts of material for in vitro studies (Musial et al., unpublished data).

The amino acid sequences of albolabrin and elegan- tin described here show that all twelve cysteines are conserved for peptides from the same genus, Tri- meresurus (also Tr. flavoviridis, unpublished results, Rucinski et al.). Furthermore, the positioning of the RGD sequence between cysteines is maintained in all peptides, including echistatin. However, substitution of the amino acids adjacent to the RGD sequence is com- mon (Fig. 5). Whereas albolabrin and trigramin have only one substitution between cysteines 47 and 59, elegantin has three additional substitutions converting RGDDLDDYC into RGDNPDDRC. These substitu- tions could be expected to contribute to the different values of IC50 for inhibition of platelet aggregation by ADP between these peptides (albolabrin 220 nM, elegantin 136 nM). However, echistatin shows high ho- mology with albolabrin around the RGD sequence, yet is four times as potent as albolabrin (Fig. 5). Albolabrin and trigramin, which differ by only seven amino acids, have very similar IC50 values for inhibition of platelet aggregation, 220 and 200 nM [1], respectively, as is expected. This suggests that both proteins are folded via the same cystine bridges. Clearly amino acid substitu- tions adjacent to the RGD sequence are not the only factor influencing the potency of these RGD peptides. To put this in perspective, the most potent tetrapeptide described to date, RGDF, was shown to be up to

10-times more potent that RGDS with an ICs0 value of 4-10/~M [30]. However, reduction of the cystine bridges of the disintegrins effectively reduces the activity of the peptides several hundred-fold, as has been demon- strated for trigramin [1] and echistatin [2].

The sequence homologies found between von Wil- lebrand Factor, albolabrin and elegantin as illustrated in Fig. 6 show three groups of homologies. Firstly, the RGD sequence, which is known to be involved in the binding of vWF to GPIIb- I I Ia [32]. A homology centered around three cysteines which define the N- terminal portion of the protein alignment and lastly the sequence SPENP which is followed by a cysteine in the VWF sequence. Interestingly, besides the homology shown, elegantin contains the N-terminal sequence CGSPENPCC, whilst albolabrin (and trigramin) con- tains GSPANP. Echistatin, of course, does not contain this latter sequence (Fig. 5), but exhibits a similar degree of homology over residues 31-70 as elegantin and albolabrin, with which it shares the PRNP motif (Fig. 6). It should thus be possible to begin to evaluate the significance of the SPENP sequence by comparing the ability of these three proteins to inhibit vWF bind- ing to platelet GP IIb-IIIa.

The homology of albolabrin and elegantin to the RGDS region in the Aa-chain of bovine but not human fibrinogen is intriguing (Fig. 6). While there is extensive initial homology between the bovine Aa-chain (3-54) and the human Aa-chaain (19-70), there is little ho- mology seen in the RGDS-region between the human (71-118) and bovine Aa-chains (55-100). There is also evidence that RGD may not be the only recognition sequence in the disintegrins. The removal of PRNP from echistatin has been found to reduce its ability to inhibit platelet aggregation in the presence of fibrinogen 4-fold (Gould et al., unpublished data). SPRNP is found in the fibrinogen Aa-chain at position 266-270 and is part of a surface epitope defined by monoclonal anti- body 9C3 [33]. It is interesting to speculate that SPXNP could be a modified recognition sequence, of common origin, for human fibrinogen and human vWF.

The experiments with anti-elegantin antibody (Fig. 8) show that there is surprisingly little immunological cross-reactivity between albolabrin and elegantin de- spite a high degree of amino acid identity (74%). This suggests that the major antigenic epitopes could be in the C-terminal portion of these molecules, since the N-terminus is highly conserved. Since all cysteines are identical the probability is that the molecules are folded identically [34]. However, these amino acid substitutions clearly have a much smaller effect on potency, as de- termined to date, than the secondary and tertiary struc- tures of these molecules. These observations are of potential significance in the design and possible use of the disintegrins as antithrombotic drugs.

In conclusion the results discussed here indicate that

the study of the disintegrins can offer new insights into the contribution of amino acid sequence and conforma- tion to receptor specificity. As such, these peptides have unique promise as high-affinity RGD probes, to in- vestigate the interactions of cell surface integrins.

Acknowledgements

J.W. would like to thank Dr. Bradford Jameson for very useful help with the GCG Sequence Analysis Software Package, and Wei-Qui Lu for technical assis- tance with the figures. These studies were supported by HL-15226 and HL-36579 from N.I.H.

References

1 Huang, T.-F., Holt, J.C., Lukasiewicz, H. and Niewiarowski, S. (1987) J. Biol. Chem. 262, 16157-16163.

2 Gan, Z.R., Gould, R.J., Jacobs, J.W., Friedman, P.A. and Polo- koff, M.A. (1988) J. Biol. Chem. 263, 19827-19832.

3 Huang, T.-F., Holt, J.C., Kirby, E.P. and Niewiarowski, S. (1989) Biochemistry 28, 661-666.

4 Knudsen, K.M., Tuszynski, G.P., Huang, T.-F. and Niewiarowski, S. (1988) Exp. Cell Res. 179, 42-49.

5 Nurden, A.T., George, J.N. and Phillips, D.R. (1985) Biochemistry of Platelets, pp. 159-224, Academic Press, New York.

6 Phillips, D.R., Charo, I.F., Parise, V. and Fitzgerald, L.A. (1988) Blood 71, 831-843.

7 Isenberg, W.M., McEver, R.P., Phillips, D.R., Shuman, M.A. and Bainton, D.F. (1987) J. Cell Biol. 104, 1655-1662.

8 Woods, V.L., Jr., Wolff, L.E. and Keller, D.M. (1986) J. Biol. Chem. 261, 15242-15246.

9 Fujimura, K., Fujimoto, T., Kuramoto, A. and Phillips, D.R. (1987) in Platelet Activation, pp. 87-95, Academic Press, New York.

10 Charo, I.F., Fitzgerald, L.A., Steiner, B., Rail, S.C., Jr. and Be- keart, L.S. (1986) Proc. Natl. Acad. Sci. USA 83, 8351-8364.

11 Poncz, M., Eisman, R.E., Heidenreich, R., Silver, S.M., Vilaire, G., Surrey, S., Schwartz, E. and Bennett, J.S. (1987) J. Biol. Chem. 262, 8476-8482.

89

12 Jennings, L.K. and Phillips, D.R. (1982) J. Biol. Chem. 257, 10458-10466.

13 Brass, L.F., Shattil, S.J., Kunicki, T.J. and Bennett, J.S. (1985) J. Biol. Chem. 260, 7875-7881.

14 Marguerie, G.A., Plow, E.F. and Edgington, T.S. (1979) J. Biol. Chem. 5357-5363.

15 Bennett, J.S. and Villaire, G.J. (1979) J. Clin. Invest. 64, 1393-1401. 16 Hawiger, J., Parkinson, S. and Timmons, S. (1980) Nature 283,

195-197. 17 Plow, E.F. and Marguerie, G.A. (1980) Blood 56, 553-555. 18 Kloczewiak, M., Timmons, S., Lukas, T.J. and Hawiger, J. (1984)

Biochemistry 23, 1767-1774. 19 Gartner, T.K. and Bennett, J.S. (1985) J. Biol. Chem. 260, 11891-

11894. 20 Plow, E.F., Pierschbacher, M.D., Ruoslahti, E., Marguerie, G.A.

and Ginsberg, M.H. (1985) Proc. Natl. Sci. USA 82, 8057-8061. 21 Williams, J.A., Rucinski, B.R., Holt, J. and Nierwiarowski, S.

(1989) FASEB J. 3, A310, 487. 22 Williams, J.A., Villaroya, H. and Petek, F. (1978) Biochem. J. 175,

1069-1077. 23 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J.

(1951) J. Biol. Chem. 193, 265-268. 24 Rucinski, B., Niewiarowski, S., James, P., Daniel, D.A. and

Budzynski, A.Z. (1979) Blood 53, 47-62. 25 Girma, J.P., Chopek, M.W., Titani, K. and Davie, E.W. (1986)

Biochemistry 25, 3156-3163. 26 Titani, K., Kumar, S., Takio, K., Ericsson, L.H. et al. (1986)

Biochemistry 25, 3171-3184. 27 Henchen, A., Lottspeich, F., Topfer-Petersen, E., Kehl, M. and

Timpl, R. (1980) Protides Biol. Fluids 28, 47-50. 28 Ruoslahti, E. and Pierschbacher, M.D. (1987) Science 238, 491-

497. 29 Reed, J., Hull, W., Lieth, C.-W., Kubler, D., Suhai, S. and Kinzel,

V. (1988) Eur. J. Biochem. 178, 141-154. 30 Williams, S. and Gralnick, H. (1987) Thromb. Res. 46, 457-471. 31 Plow, E.F., Pierschbacher, M.D., Ruoslahti, E., Marguerie, G. and

Ginsberg, M.H. (1987) Blood 70, 110-115. 32 Berliner, S., Niiya, K., Roberts, J.R., Houghten, R.A. and Ruggeri,

Z.M. (1988) J. Biol. Chem. 263, 7500-7505. 33 Cierniewski, C.S. and Budzinski, A. (1989) Thromb. Haemostasis

62, 72. 34 Low, B. (1979) Handbook Exp. Pharm. 52, 213-257.