Role of the N-terminal region of phospholipase A2 subunit of βl-bungarotoxin in the toxin-Ca2+ complex-formation

Role

of

the

N-terminal

region of phospholipase A2 subunit

of

fl-bungarotoxin

in the toxin-Ca2+

complex-formation

Sin-Tak CHU and Yee-Hsiung CHEN*

Institute of Biochemical Science, College of Science, National Taiwan University, and Institute ofBiological Chemistry, Academia Sinica, Taipei 10764, Taiwan

/,-Bungarotoxin

consists of a phospholipase A2 subunit and a non-phospholipase A2 subunit. Modification of

8,1-bungarotoxin with CNBr resulted in cleavage at Met-6 and Met-8 of its phospholipase A2 subunit. Analysis of the fluorescence dataofboth thetoxin-Ca2+complex at 300-350 nm and the toxin-Tb3+ complex at 450-650 nm showed the existence oftwobinding sites for both metal ionsonthedifferent domains of the toxin molecule.AtpH 7.6the association

constants for the high-affinity and low-affinity sites of the toxin-Ca2+ complex were determined to be 2.79-x103+0.21 x 103M-1 and 0.47x103+ 0.06x103 -1 respectively. For the toxin-Tb3+ complex the association constant for the high-affinity site was 2.95x103+0.43x103 M-1 and that for the low-affinity site was 0.11x10 +0.03x103M-1. Removal of the N-terminal octapeptide of the phospholipase A2 subunit from the toxin molecule caused disintegration of the low-affinity site but did not disrupt the high-affinity site. This might accompany a change in theconfigurationaround His-48 of the phospholipase A2 subunit. Between pH 6 and 8 the binding of metal ions tothe high-affinity site increased but that to the low-affinity site did not change with -increasing pH. The neurotoxicity andenzymic activity of the toxin were lost on removal of the low-affinity site.

INTRODUCTION

The presynaptically active toxins in the venoms of snakes belongingtoElapidae, Crotalidae and Viperidaearevery potent in blocking neuromuscular transmission (Chang, 1985).

fll-Bungarotoxin

(fi1-BuTX),

the main component of

f,-bungaro-toxins in the venom of the Taiwan banded krait (Bungarus multicinctus) (Chen et al., 1982), is a representative of this kind of toxin. It inducesfirstly facilitation and irreversible disruption ofacetylcholine release from the cholinergic synapses before the blockade of neuromuscular transmission. Since this action is very specific and unique, the toxin has been exploited as an important tool or probe in physiological, biochemical and pharmacological studies.

/l3-BuTX

consists oftwosubunits linkedbyadisulphide bond (Kondo etal., 1978). One is a phospholipase A2 (PLA2) (EC 3.1.1.4) subunit having 120 amino acid

residues,

which shows a remarkable degree of structural similarity to other vertebrate PLA2enzymes (Kini& Evans, 1987). The otheris a rather small subunit with 60 amino acid residues. Its role in the neurotoxic effect is currently unclear, despite the fact that it shows sequence similarity to dendrotoxin and toxins I and K from the venom of mambas, which are representativeofKunitz proteinase inhibitors (Dufton, 1985) and themselves show facilitation oftransmitter release, interference with the binding of native

fl1-BuTX

in chick muscle and enhancement of the presynapticeffect of crotoxin and notexin(Harvey &Karlsson, 1982).

,/%-BuTX

has anabsolute

requirement

forCa2+as acofactor forPLA2activity,which is believedtoplayanessential role in the neurotoxicity (Kelly etal., 1976)in spiteof the theorythat the lethal toxicity of

1i1-BuTX

is notnecessarily associated with its enzymic activity (Yang&Lee, 1986).

Establishing

thenatureof theCa2+-bindingdomains in the

/,3-BuTX

moleculeis

important

in elucidating the structure-function

relationship

ofthe

protein

molecule. The maingoalof the present workwastoestablishthe

role of the N-terminal region of the PLA2 subunit in the toxin-Ca2+ complex-formation.

EXPERIMENTAL Materials

Crude venom ofthe Taiwan banded krait was supplied by Chen Hsin Tong Chemical Co., Taipei, Taiwan. CM-Sephadex C-25 and Sephadex G-50 were obtained from Pharmacia, Uppsala, Sweden.

TbCI3

was purchased from Alfa Products, Beverley, MA, U.S.A. All other chemicals were of reagent grade. Preparation of181-BuTX andmodification with CNBr

/,-BuTX

was isolated from the crude venom on a

CM-Sephadex C-25 column(Chen etal., 1982)and purified further

on a Sephadex G-50 column (Lin et al., 1984). It was cleaved withCNBr according to the method of Gross & Witkop (1962). The reaction mixture was dialysed against water to remove reagent andthesmall peptide fragment.

Assayof PLA2activity and toxicity

/,3-BuTX

releasedthefatty acid

from

10,umolof egg phospha-tidylcholine suspended i'n 4ml of 5

mhM-deoxycholate

at37'C. The amount offatty acidwas

measured

byHItitration, which was carried out witha tadiometer- pHM8 standard pH-meter attached toaT80titrator andanAFIU80

autoburette.

One unit ofPLA2activity wasdefinedasthereleaseof1

,sequiv.

offatty acid from

phospholipid/min.

Eight albino mice

weighing

15-20 g were injected intraperi-toneally w-ith

fl1-BuTX

or

CNBr-,fl-BuTX.

The 50% lethal dosage (LD50) was determined according to a previously de-scribed method(Litchfield &

Wilcoxon,

1949).

Measurement ofc.d. andfluorescence

The concentration of

fll-BuTX

was determined from the absorbance at 280 nm, by

using

A1%m- 12.5

(Lin

etal.,

1984).

Abbreviations used:

8,/-BuTX, fll-bungarotoxin; CNBr-f,i-BuTX,

theCNBr-modified

/ll-bungarotoxin;

PLA2,

phospholipase A2.

S.-T. Chu and Y.-H. Chen

The c.d. was measured with a Jasco J-20 spectropolarimeter underconstantflushing with

N2

atroom temperature. The mean residue elipticity, [0], was estimated from the mean residue weight, which was calculated from the amino acid composition. The fluorescence intensity, expressed in arbitrary units, was measured at room temperature with a Hitachi F-4000 fluor-escencespectrophotometer. Both the excitation and the emission slit width were 10 nm. Raman emission due to the scattering of solvent was minimized by adjusting the intensity scale. In the scanning of an emission spectrum of Tb31 in the presence of protein in 0.02 M-Pipes buffer, the solution was freshly prepared. It took no more than 5min to scan a spectrum to prevent precipitation. F2 representedthe fluorescence intensity at wave-length A2 (nm) when the fluor was excited at wavewave-length A1 (nm). Analyses offluorescence data

The modified Scatchard plot (Epstein et al., 1974) was con-structedtoanalyse the fluorescence data of a complex formed by ametal ion and

fl,-BuTX

or

CNBr-/,%-BuTX:

IAFI/[L]free

=

(KL FOO)-(KL

IAFI)

(1)

where AFis the change in protein fluorescence on adding metal ion, L, and F,oo theprotein fluorescence in the absence of metal ion. In the analysis ofTb3+ fluorescence, AF is the change in Tb3+ fluorescence inthe presence ofprotein, and FOOthe fluorescence atinfiniteTb3+concentration. KL is the association constantof the complex concerned. Throughout the titration,

IAFI/[L]totai

was plotted against

IAFI,

since [L]1reewasclose to

[L]total.

RESULTS

Modification of

fi,-BuTX

with CNBr

There are two methionine residues (Met-6 and Met-8 ofthe PLA2 subunit) in the

#,/-BuTX

molecule. Cleavage ofthe toxin with CNBryieldedamodifiedtoxin,

CNBr-,83-BuTX,

whichwas devoid ofmethionineorhomoserineand itslactonederivative,as shown intheanalysis ofamino acidcomposition giveninTable 1. A comparison of the amount of each amino acid in the modified toxin with the corresponding amount in the native toxin revealed the removal ofa peptide fragment comprising

Asx2Glx1Phe

Ile1Leu1Met2

from/J-BuTX bythechemical mod-ification. The

octapeptide

turnsout tobetherather

hydrophobic

NLINFMEM ofthe N-terminal region of the PLA2 subunit. Apparently, all of 14 tyrosine and one

tryptophan (Trp-19

of PLA2 subunit)residues off1-BuTX remained'in

CNBr-/?1-BuTX.

/%-BuTX showedweak PLA2 activitybut' very potent

neuro-toxicity. Its enzymic activity in 10mM-CaCI2 at'37°C was 115 units/mg of toxin and its LD50 was

assayed

to

be'0.017 ,ug/g

bodywt.ofmice. On theother

hand,

CNBr-3,1-BuTX

showedno

enzymic activityandcausednolethalityeven atadoseashighas 2,ug/gbody wt.of mice.

Intrinsic fluorescences of thetwoproteins

The effect ofCaCl2 on the characteristics of intrinsic fluor-escence ofthearomatic residues in thetwo

protein

molecules in 0.02 M-PipesbufferatpH7.6was

compared.

Fig. l(a) shows the emission spectra when the

proteins

were

excitedat275nm.Both

l,1-BuTX

and

CNBr-/,%-BuTX

exhibited ashoulderat300-310nminadditiontoamain

peak

at339 nm, but

CNBr-,81-BuTX

produced

less fluorescence intensity

(Fig.

Ia).Theshoulder iscontributed

by

tyrosine

residues that donot transfer their resonance energy to

Trp-19,

and the

peak

arises fromTrp-19,which isexcited

directly

with

absorption

energyor

indirectly

through

resonanceenergy transferred from the

proxi-matetyrosine residue(s).

Apparently,

inboth

protein

molecules

the tyrosine residues contributing the shoulder are probably in thedomains separable from Trp- 19 and its neighbouring tyrosine residue(s) for the peak. Addition of CaC12 to the

f,l-BuTX

solution caused a considerable decrease is both F2305 and F239 (Fig. la). The Ca2+-induced change in F305 might reflect the characteristic of one Ca2+-binding site and that of F39 the

Table 1. Amino acidcompositionofCNBr-081-BuTX

Allvaluesareexpressedasmolar ratios withrespect toleucine. The values in parentheses give the amino acid composition from the primarystructureof

f,l-BiT.X.

Amino acid Composition

Asx* Thr Ser Glxt Pro Gly Ala Cys Val Met Ile Leu Tyr Phe Lys His Arg Trpt Homoserine+ homoserinelactone 20.5(22) 11.9(12) 4.3(4) 11.3(12) 6.9(7) 16.7(17) 11.2(11) -(20) 3.8(4) 0.2 (2) 8.1(9) 6.0(7) 14.2 (14) 4.9 (6) 12.7(13) 4.9(5) 13.9(14) 0.9(1) 0.0

* Fifteenaspartic acidand sevenasparagineresiduesfrom the primary

structureof

/1-BuTX.

t Sevenglutamic acid and five glutamine residues from the primary structureof

?1'-BuTX.

t Tryptophan content wasdeterminedbya u.v.-absorption method (Scoffone&Fontana, 1975).

300 330 360 300 330 360

Wavelength (nm)

Fig.1. Fluorescenceemissionspectra of

/1l-BuTX

and

CNBr-fl-BuTX

in 0.02M-PipesatpH

7.6-Theemissionspectrawerescannedwiththeexcitationwavelengthat

275nm (a) or295nm (b). Both proteins were at5

gM:

,

4,-BuTX alone; ----,

/3-BuTX

and 10mM-CaCl2; ---, CNBr-/%-BuTXalone; ---,

CNBr-,fl-BuTX

and 10mM-CaCl2.

1991 482

.0 2) C 0) C) 2) 0 '0 E 2 E a) 0) CD -2 'a x _s .-I '0 0 -0) E a) 0 t-C a, co 0 C 0 0 [Guanidiniumchloride] (M)

Fig. 2. Flexibility of tryptophan configuration and polypeptide backbone folding of Il,-BuTX and CNBr-,8,-BuTX in guanidinium chloride solution

Both proteins at5 aM wereused in thefluorescence study.

Fluor-escence was measured at 339nm (0) and c.d. was measured at

222nm(0): , 81-BuTX;----CNBr-fl,-BuTX. 2.0 1.5k-1.0e 0 5 6 7 8 9 pH

Fig.4. Enhancement of Th3+ fluorescence induced byeither /l,-BuTX or

CNBr-/l,-BuTXin0.02M Pipes bufferatpH 6-9

Fluorescence enhancement is represented byaratio ofTb3+ (1 mM)

fluorescence inthepresenceof each protein (5uM)tothe fluorescence in theabsence of protein:A, l,-BuTX; 0,CNBr-,81-BuTX.

>60K-40

20

450 500 550 600 650

Wavelength (nm)

Fig. 3. EmissionspectraofTb3+inthepresenceof/I,-BuTXor

CNBr-/)l-BuTX in 0.02

M-Pipes

at

pH

7.6

Thespectrawerescanned with theexcitationwavelengthat295nm.

Tb3+wasat1mMandeachproteinwasat5msm: *---, Tb3+ alone; ,Tb3+and f,-BuTX;---,Tb3+andCNBr-/,%-BuTX.

characteristic of the otherCa2+-bindingsite. Theformerwaslost but the latter remained when the octapeptide from the N-terminus of the PLA2 subunit was removed from the toxin

molecule, as shown by the fact that addition of CaCl2 to the

CNBr-/,%-BuTX

solution diminished F339 considerably but had

noeffecton F30 (Fig. la).

Thetwoproteinsshowed similar emission spectraarisingfrom excitation of tryptophan at 295nm (Fig. lb). Both proteins

exhibitedapeakat339nmwithaslightdifferencein fluorescence intensity ateach wavelength. Thepeakshifted to346nmwhen

,/3-BuTX

wasin 6M-guanidinium chlorideand to 349nmwhen

the protein disulphide bonds were broken with dithiotheritol (resultsnotshown).Similarbehaviour occurred withthepeakof

CNBr-,8,-BuTX.

Apparently, the sole tryptophan residue of

bothproteins isrestricted into aconfigurationthat differs from

that of free tryptophan in aqueous solution. The F339 of both proteins was diminished by addition of CaCl2 to the protein solution.

We compared further the change in both tryptophan

fluor-escence-and [01222 of the two proteins in 0-6 M-guanidinium

chloride(Fig. 2). The

/,3-BuTX

fluorescence changed slightly at

at concentrations less than 1.5 M-guanidinium chloride but increasedmarkedly from 2M-to4M-guanidinium chloride and

levelled off at 5-6 M-guanidinium chloride. In contrast, the

CNBr-,31-BuTX

fluorescence increased gradually in 0-3

M-guanidinium chloride and levelled off at the concentrations higher than 3 M-guanidinium chloride.

Both proteinsgavedouble minimaat220-222 and 208-210nm

in the c.d. between 200 and 250nm, indicating the existence of a helical structure in the protein molecules. The magnitudes

of [01222 and [01209 of

/,3-BuTX

were -6700 and -8400

degrees cm2 dmol-V respectively, and those of

CNBr-/,-BuTX

were -5200and -6200 degrees-cm2 dmol-1 respectively. The helix-coil transition as represented by the change in [01222

occurred in 4-6M-and 3-6M-guanidiniumchloride for

#,/-BuTX

and CNBr-fl1-BuTX respectively. The structure of the two proteins in 6M-guanidinium chloride became an unordered

form that gave an average mean residue ellipiticity of 1000

degrees cm2 dmol-1. On the basis of a simple method for calculating the helicalcontent ofaprotein from its c.d. spectra

(Chen etal., 1972, 1974), the difference between the helical contents of the two proteins could be implied in a ratio of

([01222-

1000)CNBr-fl-BuTX

to ([01222

l000)flBUTX9

where [01222 iS

measured in the absence ofguanidiniumchloride. The ratiowas

estimatedtobe0.8, revealingthat 80% of the helicalstructureof

/l%-BuTX

remains in

CNBr-/%-BuTX.

Enhancement ofTb3+fluorescence bythe two proteins

The emission spectrum ofTb3+ at pH7.6 showed a distinct quartet between 450 and 650nm (Fig. 3). Both

/,1-BuTX

and

CNBr-fl1-BuTXwereabletoenhance theTb3+fluroescence(Fig. 3).

The enhancement ofTb3+fluorescencebythetwoproteinsat pH 6-9 was compared (Fig. 4). The fluorescence enhancement

was represented interms ofthe ratio ofF545 in thepresence of

II IJ

S.-T. Chu and Y.-H. Chen

Table 2. Association constants forthe formation of the complex between

/1l-BuTX

and metal ion

The associationconstants were determined by the two methods discussed in the text.

Ka,

andKa2are the association constants for the high-affinity and low-affinity sites respectively.

10-3 xCa2'bindingKa (M) I0- xTb3+binding Ka(M)

/l-BuTX CNBr-fl1-BuTX

#1-BuTX

CNBr-fl-BuTX

pH K_al K_Ka2 al Ka a2aK K K 8.0 3.60+0.49 7.6 2.79+0.21 6.8 0.63 +0.05 6.5 0.26+0.05 0.45+0.07 0.47+0.06 0.46 +0.06 0.49+0.06 0.56+0.06 0.12+0.04

* Theaveragevalue determined from thetwomodified Scatchard plots shown in Fig. 5.

t It could not be determined precisely.

0.

x

0.5

0 2 4 6 8

IAFI

Fig. 5. Modified Scatchard plot for the binding ofCa2" toeitherf1l-BuTX

or

CNBr-811-BuTX

AtpH 7.6 androomtemperature,eachproteinat5 Mwastitrated

with thesamesolution in the presence of0.1M-CaCl2. The

fluor-escencedatawereanalysed by usingeqn.(1) with linear-regression fitting. The correlation coefficientwascalculatedto bemorethan

0.95 for eachcurve.*, F30 for,l1-BuTX;0,F...forfl-BuTX;A,

F...forfl1-BuTX;0,F...forCNBr-,fl-BuTX;A,F...for

CNBr-/l-BuTX.

proteintothat in the absence ofprotein. 83-BuTXenhancedTb3+ fluorescence slightly atpH 6-7; thiswas followedby amarked increased at pH7-8 and then a levelling offatpH8-9(Fig. 4). The midpoint of the change was at pH7.5.

CNBr-,/3-BuTX

increased the Tb3+ fluorescence gradually below pH6.7 and enhanced the fluorescence markedly from pH6.7 to pH 7.6, givingthemidpointat pH 7.0,but the enhancement diminished gradually from pH7.6 to 9.0 showing the midpoint at pH 7.9 (Fig. 4).

Characterization of the

Ca2l-binding

siteson

f6l-BuTX

The associationconstants of thecomplex-formation are

sum-marized in Table 2 for comparison. The fluorescence data obtained from adding Ca2+ to the

/,?-BuTX

or the

CNBr-,81-BuTX solutionwere analysed usingeqn. (1).AtpH7.6alinear

curve wasobtained in the modified Scatchardplotfor each of the

five cases shown inFig. 5. Inthe /%-BuTX solution, the linear

curvefor the effect of Ca2+onF339wasnearly parallelwith that for

the effectofCa2" onF339(Fig. 5).These twocurvesappearedto

characterizethesameCa2+-bindingsite. The association constant

for thisaffinity sitewasestimatedtobe 2.79x 103+0.21x 103 M-1.

AnotherCa2l-binding sitewascharacterized by the effect of Ca2+

onF305(Fig. 5). The associationconstantfor thisaffinity sitewas

calculatedtobe 0.47x103+0.06x 103 M-1. Apparently,/?-BuTX

has two distinct Ca2+-binding sites. At pH 6-8 the binding of Ca2+tothelow-affinitysitewaspH-independent whereas thatto

thehigh-affinity sitewaspH-dependent: the lower thepHvalue, theweaker the affinity (Table 2).

In the

CNBr-,f1-BuTX

solution, only one Ca2"-binding site

could be determined fromthe effect of Ca2+onF339orF339 (Fig.

5), but no Ca2+-binding site could be detected by the effect of

Ca2+onF305.As for thehigh-affinity siteof

81-BuTX,

the binding

ofCa2+ to

CNBr-,81-BuTX

was weaker at the lower pH value. The associationconstantfor the

CNBr-,81-BuTX-Ca2+

complex

was determined to be 0.56x 103+ 0.06x 103M-1 at pH 7.6 and 0.12x103+0.04x103M- atpH6.5 from themodified Scatchard plot.

Tb3+, atervalentlanthanide ion, has been demonstratedtobe

an excellent probe for studying the binding domain of Ca2+ in protein molecules (Epsteinetal., 1974; Wangetal., 1981). Our previous work showed that the fluorescence studies on the

bindingofTb3+to

,l1-BuTX

arerelevanttoCa2+binding (Chu & Chen, 1989). At pH7.6 /31-BuTX or

CNBr-,/1-BuTX,

both at

5.0JaM,was titrated withthe sameprotein concentration in the

presenceof

TbCl3.

The fluorescence datawereanalysed by using

eqn. (1). A concave curve was obtained for

fl,1-BuTX

in the

modified Scatchard plot (Fig. 6), revealing the existence of multiple binding sites of Tb3+ on the 181-BuTX molecule.

Cal-culationonthe basis of the existence oftwodistinctCa2+-binding siteson

,8-BuTX

asprovedabovegavethe associationconstant of thehigh-affinitysitetobe2.95x 103+0.43x 103M-1and that

of thelow-affinitysitetobe0.11 x103+0.03x103M-1. Thesetwo values had thesameorder ofmagnitude asthose determined for

the

fi,-BuTX-Ca2+

complex. Binding of Tb3+tothe low-affinity sitewasprominent onlyathighmolar ratio ofTb3+to

#,/-BuTX,

under which condition the enhancement of Tb3+ fluorescencewas so small that the determination for the associationconstantwas

lessaccurate. On the otherhand,alinearcurve wasobtained for

CNBr-/,3-BuTX

in the modified Scatchard plot, indicating a

single bindingsite of Tb2+onthe

CNBr-pl-BuTX

molecule(Fig. 6). The association constant was estimated to be 0.67x 103+

0.07x 103 M-1. At pH 6.5, the binding strength ofTb3+ to the

high-affinity sites of both

f,i-BuTX

and

CNBr-,81-BuTX

de-creased to a great extent. The association constant for the high-affinity site on

/,3-BuTX

decreased to 0.30xl03+ 0.04x

103M-1 and that for the

CNBr-/,%-BuTX-Tb3+

complex was

downto0.21x 103+0.05x103M-1.Estimation ofthe association

2.95 +0.43 0.11 +0.03 0.30+0.04 N.D.t

0.67 + 0.07 0.21 +0.05 484

r--n I-x 0 0 20 40 AF(arbitraryunits) 60

Fig.6.Modified Scatchard plot forthebindingofTh3+to either

1l1-BuTX

orCNBr-L-BuTX

At pH7.6androomtemperature,each protein at 5,CMwastitrated with thesamesolution in the presence of 0.1M-ThCI3. The fluor-escence data were fitted to eqn. (1) with linear regression. The correlation coefficients of thecurvesfor thehigh-affinitysite and the low-affinity siteon

fl1-BuTX

(-)were0.97 and0.94 respectively. The correlation coefficient of the curve for the siteon

CNBr-,l-BuTX(0)was0.98.

conformations of the two molecules are very similar (Dijkstra etal., 1978, 1983; Keith etal., 1981). One Ca2+-binding loop and sixsegments in both proteins, which comprises more than 83 % of the total amino acid residues of themolecules, have a nearly superimposable backbone (Renetseder etal., 1985). These six similar segments may be taken as asingle rigidunit torepresent the 'homologouscore' structurethatmaybeglobally preserved in vertebrate PLA2 enzymes. The PLA2 subunit of

fl1-BuTX

contains highly conserved amino acid residues and sequences characteristic of all PLA2 molecules sequenced to date (van Scharrenberg etal., 1982; Dijkstra etal., 1984; Achari etal., 1987).Intheabsence ofan X-ray structurefor

f,l-BuTX,

wehave superimposedthe six segments of similar sequence and the Ca2+-binding loop predicted infl-BuTX on to the main-chain con-formation of bovinepancreaticPLA2 and C. atrox venom PLA2 and made some educated guesses. The plausible three-dimen-sional network (Fig. 7) together with the spectral properties shown in this work shed some light on several important structural features.

Thec.d. characteristicsshown in Fig. 2 suggest a stable tertiary structure of

fl1-BuTX

in a segmental sense. The N-terminal region (residues 1-12)ofthe PLA2familyis a segment of similar sequence in helicalconformation (Dijkstra et al., 1978), which is believed to be part of the so-called 'interface recognition site' thatrequires thepresenceofCa2+in the interaction process with the micellarlipid/water interface(van Dam-Mieras etal.,1975). The activeconformation of PLA2 for the aggregated substrate is preserved in a hydrogen-bonded network (Renetseder etal., 1985), whichin the

1,1-BuTX

PLA2 subunit may involve invariant His-48, Tyr-52, Pro-63, Tyr-68, Asp-94 and the N-terminal region (Fig. 8). EitherTyr-52 or Tyr-68 is distant from Trp-19 (Fig. 7). Considering the fluorescence characteristics shown in thiswork, wesuspectthe participation of thephenolic group of Tyr-52and/orTyr-68in the low-affinitysite. His-48 isrestricted and itsimidazole group mayhavean abnormal

pK.

at 7.5 (Fig. constants at pH 6.5 was less accurate, since the Tb3+

fluor-escenceenhanced by theproteins at pH 6.5 was less thanthat at pH7.6. Underthe circumstances, a precise determination ofthe assocition constant forthe low-affinity site was not feasible.

DISCUSSION

Previously, the multiple Ca2'-binding sites on the

/1-BuTX

molecule were independently revealed by a Scatchard plot obtained by a binding assay with the use ofradioactive Ca2+ (Abe etal., 1977) or by the effect ofCa2+ on the absorption spectrum of

/,1-BuTX

(Yang & Lee, 1986). However, these two approachesarelimitedtoexploring the configurationaround the Ca2+-bindingdomain. For onething, Ca2+lacksuseful spectro-scopic properties that can be employed in probing the nature ofitsbindingsite. Foranother,theScatchardplotfor each of the two approaches gives no information abouteach

Ca2+-binding

domain. Thesetwo

problems

are overcome toacertainextent

by

the analyticalmethods used in the present work,which enables us todefine the twodistinct

Ca2+-binding

siteson the different domains ofthe

/l%-BuTX

molecule. This isuseful in

elucidating

the structuralfeaturesof the

l1-BuTX-Ca2+

complex-formation.

Our results indicateclearlythat: (1)there are twodistinct

Ca2+-binding

siteson the different domains of

fl1-BuTX; (2)

the N-terminal region of the PLA2 subunit plays anessential role in maintainingtheintegrityof thelow-affinity site,whichseems to benecessary for theneurotoxicityand

enzymic activity

towards the aggregatedsubstrate.

Acomparison of thetertiary structures of bovine

pancreatic

PLA2 and Crotalus atrox PLA2 shows that the main-chain

Fig.7.Superimposition of theprimary structureofPLA2 subunitof

81-BuTX on tothe main conformation of bovine pancreatic PLA2 (Dijkstraetal.,1978)

The broken line indicates the main points of variance between pancreatic and Crotalus atrox PLA2 enzymes. The amino acid residuespredictedfor eachanalogoussegment andthe

Ca2+-binding

S.-T.Chu and Y.-H. Chen Tyr-52 I. H

0

- . I Asp-94 H 0 Tyr-68 NH2 Asn-1 -.Asn 4 N H ,1 H H". * H \ 'O~ Cz Pro-63

Fig. 8. Schematicrepresentationof a proposed hydrogen-bonded network in the

/8,-BuTX

Thenetworkis drawn according to that described by Dijkstraetal. (1981).

4). The PKa shifts to a normal value at pH 7.0 (Fig. 4), as the restriction exerted on His-48 may be released by removal of the N-terminal region, which does not contact directly with His-48. This causes disintegration of the low-affinity site, which may be notaffected by the charged status of His-48 in native toxin (Table 2).

Thehigh-affinity site resides in a 'Ca2+-binding loop' running fromresidue 25 and 42, which are highly conserved glycine-rich sequencescharacteristic ofPLA2 molecules (Dijkstra et al., 1981). An oxygen cagethat sequesters the metal ion is stabilized by this loop (Dijkstra et al., 1981). In thefl-BuTXmolecule the oxygen cage may be formed in part fromthe three main-chain oxygen atoms atTyr-28, Gly-30 and Gly-32 and the carboxylate group ofAsp-49. The phenolic group of Tyr-28 is excluded as a Ca2+ ligandonthe basis of theconclusion drawn from an n.m.r. study onpig and bovine PLA2 molecules (Fisher et al., 1989). This is reflected in the fact that the high-affinity site is not characterized by the effect of Ca2+on F"'.In

CNBr-f,l-BuTX

only the high-affinity site isapparently stabilized byagroupwith PKa 7.9, as suggestedfrom thecharacteristics shown in Fig. 4.Thischarged groupmightbe ascribed to the a-amino group ofArg-1 of the non-PLA2 subunit, which might interact with the N-terminal region of the PLA2 subunit. The Ca2+-binding loop may not be so rigid that binding ofCa2+ to the high-affinity site causes a

change in local conformation around Trp-19 (Fig. lb). Tyr-22 andTyr-106, both of whichare close toTrp-19, are separated from the two Ca2+-binding sites (Fig. 7). Resonance energy

transfer from these two tyrosine residues to Trp- 19 accounts partly for F3.

This work waspartially supported by the National Science Council, Taipei, Taiwan (Grant NSC 79-0412-B-001-07). Some of the work describedinthis paperformspartof a dissertation submitted by S.-T. C. in partial fulfilment for the requirement of the degree of D.Sc. at the National Taiwan University. We thank ProfessorMu-Chin Tzeng for hishelpful discussions andreading ofthe manuscript.

REFERENCES

Abe, T. A., Alema, S. & Miledi, R. (1977) Eur. J. Biochem. 80, 1-12 Achari,A., Scott, D., Barlow, P.,Vidal,J.C.,Otwinowski,Z.,Brunie,S.

& Sigler, P. B. (1987) Cold Spring Harbor Symp. Quant. Biol. 52, 441-452

Chang, C. C. (1985)Proc.Natl. Acad.Sci. Counc. BR.O.C. 9, 126-142 Chen, Y.H., Chan, K. H. & Yang, J. T. (1974) Biochemistry 13,

3350-3359

Chen, Y.H., Yang, J. T. & Martinez, H. M. (1972) Biochemistry 11, 4120-4131

Chen, Y.H., Tai, J.C., Huang, W.J., Lai, M.Z., Hung, M.C., Lai, M. D.& Yang, J. T. (1982)Biochemistry21, 2592-2600

Chu, S.T.&Chen,Y. H.(1989) Biochem.J.262,773-779

Dijkstra, B.W., Drenth, J., Kalk, K.H. &Vandermaelen, P. J.(1978) J. Mol. Biol.124,53-60

Dijkstra,B.W.,Kalk,K.H.,Hol,W.G.J.&Drenth,J.(1981)J.Mol. Biol. 147, 97-123

Dijkstra,B.W., Renetseter,R.,Kalk,K.H.,Hol,W.G.J.&Drenth,J. (1983)J. Mol. Biol.168, 163-179

Dijkstra,B.W.,Kalk,K.H.,Drenth, J., de Hass,G.H.,Egmond,M. R. &Slotboom,A. J. (1984)Biochemistry 23,2759-2766

Dufton,M. J.(1985)Eur. J.Biochem. 153, 647-654

Epstein, M.,Levitzki, A.&Reuben,J.(1974)Biochemistry 13,1777-1782 Fisher, J.,Primrose,W.U.,Roberts, G. C. K., Dekker, N., Boelens, R.,

Kaptein,R.&Slotboom,A. J.(1989) Biochemistry 28, 5939-5946 Gross,E.&Witkop, B.(1962)J.Biol. Chem.237, 1856-1860 Harvey, A.L.&Karlsson, E.(1982) Br. J.Pharmacol.77, 153-161 Keith, C.,Feldman,D.S.,Deganello, S., Glick, J., Ward,K.B.,Johnes,

E.0. &Sigler, P. B.(1981)J. Biol.Chem. 256,8602-8607

Kelly,R.B.,Oberg, S. G., Strong,P. N.&Wagner,G. M.(1976) Cold Spring Harbor Symp. Quant. Biol. 40, 117-125

Kini, R.M. &Evans,H.J.(1987)J. Biol.Chem.262,14402-14407 Kondo, K., Narita, K. & Lee, C.Y. (1978) J. Biochem. (Tokyo) 83,

91-99

Lin, W.Z., Chu, S.T. &Chen, Y. H. (1984) Proc. Natl. Acad. Sci. Counc. BR.O.C.8, 113-118

Litchfield, J. T. & Wilcoxon, F. (1949) J. Pharmacol. Exp. Ther. 96, 99-113

Renetseder, R.,Branie, S., Dijkstra, B.W., Drenth, J. &Sigler, P. B. (1985)J.Biol.Chem.260, 11627-11634

Scoffone,E.&Fontana,A.(1975)in Protein SequenceDetermination, 2nd edn. (Neeleman, S.B.,ed.),p. 167,Springer-Verlag,New York

van Dam-Mieras, M.C.E., Slotboom, A.J., Pieterson, W.A. & deHass, G.H.(1975) Biochemistry 20, 1584-1591

van Scharrenburg, G. J.M., Puijk, W.C., Egmond, M.R., van der Schaft,P.H., deHass, G.H. &Slotboom, A. J.(1982) Biochemistry 21,1345-1352

Wang, C.-L.A.,Leavis,P.C.,Horrocks,W. DeW.&Gergely,J.(1981) Biochemistry 20,2439-2444

Yang,C. C.&Lee,H. J.(1986)J.Protein Chem. 5, 15-28

1991

His-48

N N

N H

Received 17 December 1990/19March 1991;accepted2April 1991 486