Identification and function of the high affinity binding sites for Ca2+ on the surface of platelets

Identification and function of the high affinity

binding sites for Ca2+ on the surface of

platelets.

L F Brass, S J Shattil

J Clin Invest.

1984;

73(3)

:626-632.

https://doi.org/10.1172/JCI111252

.

Extracellular Ca2+ is required for platelet aggregation and secretion in response to ADP or

epinephrine. Recently, we reported that the platelet surface contains two classes of high

affinity binding sites for extracellular Ca2+. To identify these sites and clarify their role in

platelet function, we have now (a) studied platelets congenitally deficient in surface

membrane glycoproteins and (b) examined the effect of removing surface-bound Ca2+ on

platelet responses to ADP and epinephrine. Unstimulated normal platelets contained

86,000 Ca2+-binding sites/platelet with a dissociation constant (Kd) of 9 nM and 389,000

sites with a Kd of 400 nM. In contrast, thrombasthenic platelets, which lack glycoproteins IIb

and IIIa, exhibited a 92% reduction in the number of higher affinity Ca2+-binding sites and a

63% reduction in the number of lower affinity sites. Bernard-Soulier platelets, which lack

glycoprotein Ib, were not deficient in Ca2+-binding sites. After stimulation with ADP, both

normal and thrombasthenic platelets developed approximately 138,000 new Ca2+-binding

sites/platelet (Kd = 400 nM), while the larger Bernard-Soulier platelets developed 216,000

new sites. These data suggest that IIb and IIIa represent the major Ca2+-binding

glycoproteins on unstimulated platelets, while neither these glycoproteins nor Ib represent

the new Ca2+-binding sites on stimulated platelets. Removal of Ca2+ from the platelet

surface inhibited platelet function. Despite the presence of 1 mM Mg2+, ADP- and

epinephrine-induced aggregation and […]

Research Article

Identification

and

Function

of

the High

Affinity Binding Sites

for

Ca2+

on

the

Surface

of

Platelets

Lawrence F. BrassandSanfordJ. Shaffil

Hematology-Oncology Section, DepartmentofMedicine and the

Coagulation and Thrombosis Laboratory, Departmentof Pathology and Laboratory Medicine, Universityof Pennsylvania

Schoolof Medicine, Philadelphia, Pennsylvania19104

Abstract.

Extracellular

Ca"

is

required for

platelet

aggregation

and

secretion

in response to ADP or

epinephrine. Recently,

we

reported

that the

platelet

sur-face contains

two

classes

of high

affinity binding sites

for

extracellular

Ca

2.

To

identify

these

sites

and

clarify their

role

in platelet

function,

we

have now (a)

studied

platelets

congenitally deficient in surface membrane glycoproteins

and

(b) examined

the

effect of removing surface-bound

Ca>2

on

platelet

responses to ADP

and

epinephrine.

Un-stimulated

normal

platelets contained 86,000

Ca2+-bind-ing sites/platelet with

a

dissociation

constant (Kd)

of

9

nM

and

389,000 sites

with

a Kd

of

400 nM. In contrast,

thrombasthenic platelets, which

lack

glycoproteins

lIb

and

1Ila,

exhibited

a

92%

reduction

in the number

of

higher affinity

Ca2+-binding

sites

and a

63%

reduction

in

the

number

of

lower

affinity

sites. Bernard-Soulier

plate-lets,

which

lack

glycoprotein

Tb,

were not

deficient

in

Ca2+-binding

sites.

After stimulation with ADP,

both

normal

and

thrombasthenic platelets developed

138,000

new

Ca2+-binding

sites/platelet

(Kd

=

₄₀₀

_nM),

while the

larger

Bernard-Soulier platelets developed

216,000

new

sites.

These data suggest that

Ilb

and

IIla

represent the

major

Ca2+-binding

glycoproteins

on

un-stimulated platelets, while neither

these

glycoproteins

nor

lb

represent the new

Ca2+-binding

sites

on

stimulated

platelets.

Portions of this work were presented at the Annual Meetingof the American Society ofHematology,Washington, DC, December, 1982 and the Federation of American Societies forExperimental Biology,

Chicago, IL, April, 1983, and was published in abstractform (1982.

Blood. 60:195a).

Receivedfor publication 31 May1983andin

revisedform

4November 1983.

Removal

of

Ca"

from

the platelet

surface inhibited

platelet function. Despite the presence of 1 mM

Mg2+,

ADP-

and

epinephrine-induced aggregation and

[14C]serotonin

release

were

markedly decreased at

free

Ca2+

concentrations < 7 nM, a value

similar

to the Kd

of the

higher affinity Ca2+-binding sites. Moreover,

ga-dolinium, a lanthanide that competed for these

Ca-2+

binding sites, also inhibited aggregation and serotonin

release. These studies

demonstrate,

therefore, that the

binding

of extracellular

Ca>2

to

glycoproteins

HIb/IIa on

unstimulated

platelets or to

additional

membrane

proteins

on

stimulated

platelets

is

necessary

for

maximal platelet

responses

to ADP and

epinephrine.

Thus,

the requirement

for extracellular

Ca>2

during

platelet

activation by these

agonists

may actually represent a requirement for

surface-bound Ca>.

Introduction

Extracellular

Ca2+

is required forseveral ofthe reactionsthat take place onthe platelet surface when platelets are activated

byagonists, includingthebinding of fibrinogento membrane glycoproteins Ilb and lIIa (1, 2). Since extracellular

Ca2"

exists in rapid equilibrium with

Ca2`

bound to the platelet surface (3), atleast part of the requirement for extracellular

Ca2`

may

reflect

Ca2`

bound to the platelet membrane atsitesinvolved in thesurfacereactions.Previously,wedemonstrated thatthere

areaminimumoftwoclassesofhigh affinity,saturable binding sites for

Ca2"

onthesurfaceof intactplatelets (3).Stimulation

byADPorepinephrine producedadditionalCa2+-bindingsites,

eveninthe absence ofplatelet aggregation and secretion.

Thehighaffinity (dissociation constant

_[Kd]

< 10-6M)of

all of these bindingsites forCa2+ suggested thatthey may be

located on membrane proteins or glycoproteins rather than

membrane phospholipids(4). Accordingly, we performed the

present studies (a)to identifythe specific membrane proteins

that represent thehigh affinityCa2+-bindingsitesonunstimulated

platelets;(b)todeterminewhether these samemembrane pro-teinsareinvolvedinADP-stimulated

Ca"

binding;and(c)to

J. Clin. Invest.

©The AmericanSociety for ClinicalInvestigation,Inc.

examine the effect on platelet aggregation and secretion of

se-lectively removing

Ca2"

from its binding sites on the platelet surface. The results suggestthat therequirementforextracellular

Ca2"

duringADP-andepinephrine-induced platelet aggregation and secretion reflectsa requirementfor

Ca2"

bound tospecific membrane proteins.

Methods

Preparation of gel-filtered platelets. Platelet-rich plasma from normal

andthrombasthenic donorswasobtained by differentialcentrifugation

ofbloodanticoagulated with 13 mM sodium citrate and incubated for 30 min at room temperature with 1 mM aspirin to inhibit platelet

cyclooxygenase(3). Platelet-rich plasma from donors with the Bernard-Souliersyndrome wascollected after the erythrocytes had settled by

gravity for 3hin anair-tightplastic container. Plateletswereisolated from plasma by gel filtrationonSepharose 2B(PharmaciaFine Chemicals, Piscataway, NJ) usinganelutionbuffer that contained 137mMNaCl,

2.7 mM KCI, I mM MgCl2, 5.6mMglucose, I mg/ml bovineserum albumin(Sigma Chemical Co..St. Louis, MO),and 4 mMHepes,pH

7.40(3, 5). This bufferwaspreparedwithwaterthat had beendepleted

of calcium bypassagethrough Chelex 100 ionexchangeresin(Bio-Rad

Laboratories, Richmond, CA).Thefinal totalCa2+contentofthebuffer

measured by atomic absorption spectroscopy was 11 MM.This value

was not affected by the presence of gel-filtered platelets, and it was

included in the calculation of theCa24concentration ofall buffer solutions used inthis study. The final freeCa2+concentration measuredby Ca2+

electrodewas5MM.Normalplateletspreparedin thismanneraggregate

in response toADP when stirred with added fibrinogen, but do not

undergo the release reaction.Intheabsence ofstirringandfibrinogen,

aggregationdoesnot occur(3).

Plateletsfrom three donors with Glanzmann's thrombasthenia (L.M., M.M., and N.L.) andoneof thetwodonors withthe Bernard-Soulier syndrome (A.J., previously A.H.)were obtained through the courtesy ofDr.Margaret Johnson oftheWilmingtonMedicalCenter,Wilmington,

DE. Plateletsfrom the seconddonorwiththeBernard-Souliersyndrome

(R.O.)wereobtained with the assistance ofDr.ScottMurphy, Thomas Jefferson University, Philadelphia, PA.All fivepatients have classical

findingsandhistoriesfor their respective diseases and havebeendescribed previously by multiple investigators (1,6-11). Platelets fromallthree

of the thrombasthenic donors contained<7%of the normalamountof

glycoproteinslIb andIla (measured antigenically withananti-Ilb-Illa monoclonalantibody)and do notbind fibrinogen(1, 6).

Measurementofsurface-bound Ca2` onunstimulatedplatelets.Ca2+

bindingtothesurface of unstimulated plateletswas measuredas

pre-viously described (3). Briefly, 45CaC12(1-1,500 MM) and EGTA (500

MM)wereaddedtogel-filtered plateletstoobtainbufferedextracellular freeCa2+concentrationsthat rangedfrom1

0V

M to 1

0V

M and platelet

concentrationsthat rangedfrom 1.5-3.0X 108/ml(3, 12). After 10min at roomtemperature, 0.4-ml aliquots oftheplatelet suspension were

layered ontopof silicone oiland the platelets sedimented by

centrif-ugation for 2 min at 12,000 g in a microcentrifuge. The amount of

45Ca2+

intheplateletpellet was measured by scintillation counting and correctedfortrapped extracellular space using

[3Hjsorbitol

(New England

Nuclear, Boston, MA) (3). Surface-bound45Ca2+ was defined as the

differencebetweenthisvaluefor platelet-associated45Ca2'and the value

obtained when 3.5 mM EGTA was added to the platelet suspension

immediatelybeforesedimenting the platelets. The number and affinity

oftheCa2+-bindingsitesmeasured in this waywere calculated by

non-linearregression analysis using the general equation for ligand binding

toone or morenoninteracting classes of binding sites (3, 13).

Measurement ofsurface-boundCal' onADP-stimulatedplatelets.

Measurements of surface-bound Ca2" onplateletsstimulatedbyADP were made over a morerestricted range of free

Ca2"

concentrations (3 X 10-8 to 1 X l0-'M) using 50 MM EGTA tobuffer the free

Ca2"

concentration. Otherwise the procedurewas the same asused in the absenceofADP.Overthis limited range of

Ca2"

concentrations, binding occurred predominantly at asingle class of sites, and the data were

analyzedby the method of Scatchard (14).

Free Ca2" concentration. FreeCa2+ concentrations> I MM were

measuredwith aCa2`-specificelectrode(Orion Research, Inc., Cambridge, MA). Free Ca2+concentrations<1pMwerecalculatedby the method of Portzehl (12) using published associationconstants for EGTA that were corrected for pH(15). In addition, the apparent association constant forCa2+-EGTAwasconfirmeddirectly in the platelet incubation buffer (minusMg2` andalbumin) using the method ofBers(16).

Platelet aggregation and serotonin release. Platelets in plasmawere

labeled with ['4C~serotonin (3).Followinggelfiltration,imipramine (I

MM)andfibrinogen (0.1 mg/ml) wereadded and platelet aggregation

and ['4C]serotonin release in responseto ADPandepinephrinewere measuredatvarious freeCa2+ concentrations(3).In someof the studies, platelet aggregation and ['4C]serotonin releasewere measured in the presence of 10-100 MMGdCI3(ICNPharmaceuticals, Plainview, NY).

Inthis case, the pH of the gel filtration bufferwasdecreasedto6.9to

avoidformation of insoluble Gd3+ salts (17). The cyclic AMPcontent

ofplatelets incubated with 100gMGdCl3for 15 minatroomtemperature was measured as described previously (18).

Results

Ca2`

bindingtounstimulated thrombasthenicplatelets.

Surface-boundCa2+ wasmeasuredatfreeCa2+ concentrations ranging from 10-9M to10-3Musing aspirin-treated,gel-filtered platelets. The results obtained withplatelets from sevennormal donors aresummarizedinTable I and arecomparabletothosereported previously (3). Normal platelets had two classes of saturable binding sites for Ca2+ with an apparentKd of 9 and 400 nM, respectively. Nonsaturable

Ca2'

binding averaged 0.05% of the free Ca2+ concentration.

Thrombasthenic platelets,which are deficient in glycopro-teins lIb and I11a, demonstrated abnormal Ca2+ binding. For purposes ofcomparison, the combined data from the three thrombasthenic donors and the normal donors are shown in

Fig. 1. Theanalysisof each individual thrombasthenic donor is shownin Table I. Although the thrombasthenic platelets also contained two classes ofsaturableCa2+-binding sites with ap-parent dissociation constants similar to normal platelets, the

amount of

Ca2'

bound by each class of sites was markedly reduced. Ca2' binding to the higher affinity sites was reduced byan averageof 92%, while binding to the lower affinity sites was reduced by anaverage of 63%. These data suggest, but do

notprove,that the majority of both classes of saturable binding

sites for

Ca2"

on unstimulated normal platelets is associated with glycoproteins lIb andlI1a.

Table L Cit-bindingSites on Unstimulated

Thrombasthenic Platelets

Platelets Kd Sites/platelet Kd Sites/platelet

nM nM

Normal 9±2 86,000±11,000 400±100 389,000±35,000

Thrombasthenic 16 5,000 100 113,000

1 15,000 900 141,000

5 2,000 200 177,000

Platelets were isolated from plasma by gel filtration in buffer containing 137 mMNaCI, 2.7mMKCI, ImMMgCl2,5.6 mM glucose, 1 mg/mlbovine

serumalbuminand 4mMHepes, pH 7.40.4"CaC12and500pMEGTA were added and the platelets were incubated for 10 min at room temperature. The final totalcalcium concentration (including residual calcium in the gel-filtration

buffer)ranged from10-'to1.5 X 10-3M, giving freeCallconcentrations from

10-9to10-3M (Methods). The amount of surface-bound45Ca2' was measured

after sedimentingthe platelets through silicone oil and was analyzed for each

individualdonor by nonlinearregression(3, 13). The results shown for the nor-malplatelets are the mean±SEM for seven donors.

the methods usedtostudyunstimulatedplatelets. Platelets in-cubated at370C with EGTA or EDTA irreversibly losetheir

ability

toaggregateinresponse to ADPinatime-and

concen-tration-dependent manner (19, 20). To avoid this potential

problem, we first examined the ability ofADP to stimulate platelet aggregation after normal platelets had beenincubated for 15 min at room temperature atvarious concentrations of EGTA and

Ca2".

When the

Ca2"

concentration wasbuffered with 50

MM

EGTA,ADP-inducedplateletaggregationwasfully maintained atfree

Ca2t

concentrations atleast as low as 30

0

O _x

M,_

im 0

SOw

4

40

-a U,

0)

_n

1-nM.Therefore,

Ca2"

binding to ADP-stimulated platelets was

studied between30 nM and 10

AM

extracellular freeCa2'.The upperlimit of 10 MM was chosen to minimize nonsaturable

Ca2'

binding (Fig. 1). Overthisrangeof free Ca2'concentrations,

surface-bound

Ca2"

reflects binding ofextracellular Ca2' pre-dominantlyto the lower affinity class of binding sites

(Kd

=400 nM).This permittedthe data on a Scatchard graph to be analyzed

by linear regression (Fig. 2). Normal platelets stimulated by ADP developed an average of 138,000±23,000 newCa2'-binding sites withno change in theapparent

Kd

(Figs. 2and 3). This increase in Ca2+-binding sites afterADPis ingoodagreement with the valueof140,000 sites/platelet found in our previous studyusingaplateletincubationsystem notbuffered withEGTA

(3). The absence ofa change in the apparent K4 after ADP suggeststhat thenewCa2+-binding sitesalso have a

Kd

of -400 nM. Incontrast to the marked deficiency of

Ca2"-binding

sites on unstimulated thrombasthenic platelets, thrombasthenic plateletsstimulated withADP developed a normal number of new

Ca2t-binding

sites(Fig. 3). This suggests that the

ADP-induced

Ca2"-binding

sitesare notassociated with glycoprotein

lIb

or

Hila.

Ca2"

binding

to Bernard-Soulier

platelets.

To determine

whether the ADP-stimulated Ca2-bindingsites are associated

with glycoprotein

lb, Ca2"

bindingwasmeasuredusing

Bernard-Soulier platelets, which are deficient in this glycoprotein (2).

Unstimulated Bernard-Soulier platelets fromtwodonors

con-tainedtwice the number ofhigher and lower affinity

Ca2'-binding

sites foundonunstimulated normal platelets. Thismayberelated tothelarger size and increasedmembranesurfaceareaofthese

platelets. After stimulationby ADP, Bernard-Soulierplatelets

111 LLJ

a:

Li-z :D 0

m

LOG

[Free Ca**]

(M)

Figure1.

Ca2t-binding

sitesonunstimulated thrombasthenic

plate-lets.Gel-filteredplateletswere

prepared

from three donors with

thrombasthenia(opencircles)andsevennormal donors

(filled

cir-cles). Surface-bound

CWt

wasmeasuredatfreeCa2tconcentrations

from I nMto 1mM.Platelets from all donorswerestudiedacross thisentirerange of free

Ca2?

concentrations. Each data

point

repre-sentsthemeanoftriplicatedeterminations.Thesolidlineswere gen-eratedbynonlinearregression

analysis

of the

grouped

data.

200 400 600

Co2+

BOUND (SITES/PLATELETx

1000)

Figure 2. The effectofADPonCa2+bindingtonormalplatelets. CaC12and EGTA(50

1M)

wereaddedto

gel-filtered

plateletsto ob-tainfreeCa2+ concentrationsfrom 30 nMto 10pM.Atthesame

time,ADP(10

pM)

wasaddedtohalf of theplatelets. 10 minlater,

theamountofsurface-bound

Ca2t

on

platelets

stimulated

_by

ADP (opencircles)oroncontrolplatelets

(filled circles)

wasmeasured.

The resultsobtainedaredisplayedas aScatchardplot.This

_figure

il-lustrates the dataobtainedonplateletsfromasinglenormaldonor,

butit isrepresentativeof the data from all the donorsstudied. ADP

increased the number of

Ca'+-binding

sitesonthese normal

platelets

from 354,000sites/platelet

(K4

= 221 nM)to

534,000

sites/platelet

(Kd =243nM).

CONTROL THROMBASTHENIA BERNARO-SOULIE Figure3. ADP-induced a ₇

_Ca2"-binding

siteson nor-0I

0

200/ mal, thrombasthenic, and

Bernard-Soulier

_platelets.

50/ Surface-bound

Ca2"

was

measured andanalyzedas

"I

oc ; * | / / shown in Fig. 2. The values

W

* * / v shownarethenew

Ca2+-(n _{* * t/binding sites stimulated by}

±SE

-eo~2~idn ie

e

50

* * 1// /,

ADPandrepresentthe

dif-ference between the

num-measured inthe presence andabsenceof ADP.The resultsfor the normal plateletsaremean

±SEMfor

eight

donors. The valuesshown for the thrombasthenic and Bernard-Soulier plateletsare the results from each individual donor.

developedanaverage of216,000newlower affinity Ca2+-binding

sites/platelet,anincreasethatwasalso greater than normal (Fig.

3). SinceBernard-Soulier plateletsaredeficient in glycoprotein

lb,these data suggest that thenewCa2+-bindingsites stimulated byADP are not associated with thisglycoprotein.

Effect of

Ca2"

bindingon

platelet

aggregationandserotonin

release. Both classes of

Ca2+-binding

sites on unstimulated

platelets exhibitapparentKdthat are several orders ofmagnitude lower than the 1 mM freeCa2+ concentration ofplasma. Pre-sumably,therefore, thesesitesarenormallyfullyoccupied.To

determinewhethersurface-boundCa2+is necessary for normal platelet function, two separate approaches were taken. First,

platelet aggregation and ['4C]serotonin release in response to

ADPandepinephrineweremeasuredatlowfree

Ca2"

concen-trations but in the presence of1 mM

Mg2+.

In contrast tothe

Ca2+-binding

studies, theplateletswere notpreincubated with aspirin andthey were stirred in the presence offibrinogen so

thatplatelet aggregation andserotonin release could take place.

At0.8 nM free Ca2+, 10

_,M

ADP stimulated onlypartial ag-gregation and <5% ['4C]serotonin release (Fig. 4). This reduction ofaggregation and serotonin release was not due to a direct

effect ofthe EGTA orofthe lowCa2+concentration on

fibrin-ogen, because fibrinogen incubatedat0.8 nMfree

Ca2`

forup

to2 h wasstill able subsequently to support aggregation. The

extentofADP-induced plateletaggregation and

['4C]serotonin

release increased with increasing

Ca2+,

and wasmaximal at 7 nMfreeCa2+ (Fig. 4). Similar results were obtained when epi-nephrine was used as the agonist. The reductions of ADP or

epinephrine-induced aggregation and['4C]serotonin release at

Ca2"

concentrations<7nMwerereversed by adding

Ca2'

back

totheplatelet mixture. These experiments suggest that the higher

affinity

Ca2+-binding

sites (Kd = _{9nM) on the platelet surface}

mustbe atleastpartially occupied by

Ca2+

formaximal platelet aggregation and secretion to occur in response to ADP and

epinephrine.

Mg2+,

which was present in these experiments, cannot substitute for

Ca2+

in this regard but it can support partial aggregation.

The second approach used to assessthe functional role of surface-bound

Ca2"

took advantage of the binding properties ofthelanthanide, Gd3+. Gd3+exhibits ahighaffinityfor Ca2+-binding sitesoncellmembranes (17).Inpreliminarystudies we

incubated gel-filtered platelets with '53GdC13 forup to 1 h.We found that the volume of distribution of '53Gd3+ centrifuged through silicone oil with plateletswasthe same asthe volume

ofthe trappedextracellular space measuredwith [3H]sorbitol. Thissuggests that, as with other cells (17), Gd3+ is unable to enterthe platelet. Thus,weused Gd3+todisplace surface-bound

45Ca2"

selectively. At 5

gM

free

Ca2"

and 1 mM

Mg2",

GdC13

caused adose-dependent decrease in both surface-bound

Ca2"

and ADP-induced platelet ['4C]serotonin release (Fig. 5). Gd3+ also decreased platelet aggregation, but to a lesser extent(Fig.

5).Theseinhibitory effects of Gd3+werereversed bytheaddition ofexcess

Ca2",

butnot

Mg2",

tothe incubationmixture.Another lanthanide, La3+,has been reported to inhibitplatelet function by increasingthe levels ofplatelet cyclicAMP(21). However, in two experiments, 100 $M GdCI3 had no effect on platelet cyclicAMPcontent(control=0.70 and 0.81pmolcyclicAMP/

108 platelets;GdCl3 =0.77 and 0.91 pmol).

Discussion

The plasmamembrane of humanplatelets contains high affinity Ca2+-binding sites that can be characterized by equilibrium binding studies performedwithintactplatelets atsubmicromolar extracellular free Ca2+ concentrations. Approaching in this

manner,wepreviously found that normal plateletshave at least

100,

-a

O a) 80O

0

o

C 60

C (

0

< 40

0'

0.

tO 100 1000 10,000

Free Cao2+ (nM)

Figure4. Theeffect of the extracellular freeCa2` concentrationon

plateletresponses to ADP. Aggregation and['4C]serotoninrelease in response to 10 uM ADP were measured at 37°C in the presence of

0.1 mg/mlfibrinogen and 1 mMMgCI2.The ADPwas added

imme-diately after addition of500 uM EGTA andsufficientCaC12 to give theinitialfreeCa2+concentrationsindicated. Theresults shown are the percentageof platelet aggregation(open circles)and

['4C]sero-tonin release (filled circles)measured 3 min after the addition of ADPand represent the mean±SEM of three experiments. Percent

W ₄₀₀

-i w

300

I-I

0 Z x 200U

D

-~0

m

+ 100

a 0

z z

0

I-LU 0 V)

LWW

CO)-J

30

20

10

0 80

z 0

I-O w (D

(D

al

60

40

A

B

C

20 [

o _{0 20 40 60 80 100}

TOTALGd3*

_(pM)

Figure 5. The effect ofGd3+on surface-boundCa2",platelet

aggrega-tion and ['4C]serotoninrelease. Gel-filtered platelets were prepared at pH 6.9. (A) Aspirin-treated platelets were incubated with 45CaC12

(5 MM free Ca2+)and 10-100MMGdCl3.After 10

min,

surface-boundCa2+was measured. Using platelets not preincubated with as-pirin, platelet

['4C]serotonin

release (B), and aggregation (C) in re-sponse to 50

MM

ADP were measured. Data represent the mean

±SEM ofthree experiments.

twoclassesof saturable

Ca2+-binding

sitesand

develop

additional

Ca2+-binding

sites after stimulation with ADP or

epinephrine

(3).Inthese studieswehaveused abnormal

platelets

withdefined deficiencies of plasma membrane

glycoproteins

to

help

localize the saturable

Ca2+-binding

sites.We have also

explored

the role

of these sites in

platelet

function

by observing

the effect of

removingordisplacing

Ca2'

from theplatelet surfaceon

platelet

aggregation andsecretion.

The studies performed with abnormal

platelets

showthat unstimulated

thrombasthenic platelets,

which are deficient in the membrane

glycoproteins

Ilb and IIIa, had

markedly

fewer

high affinity

Ca2+-binding

sitesthan normal. These data may

explain the observations ofPeerschke et al. (22), who found thatthrombasthenic platelets have areduced

uptake

of

extra-cellular

45Ca2".

However,their studies didnot

specifically

quan-titatesurface-bound

Ca(2. Despite

the

deficiency

of

Ca2+-binding

sitesonunstimulated thrombasthenic

platelets,

ADP-stimulated

thrombasthenic platelets had normal numbers ofnew

Ca2+-bindingsites.

Similarly,

Bernard-Soulier

platelets,

whichare de-ficient in membraneglycoprotein lb,demonstratedno

deficiency

of ADP-stimulated Ca2+-binding sites. Taken together, these

studiessuggestthat most of the saturable binding sites forCa2+

onunstimulatednormal platelets are associated with

glycopro-teins Ilband IIIa. Confirmation of this suggestion must await

direct

Ca2`-binding

studies with isolated platelet membrane

gly-coproteins.In contrast to the unstimulated platelet, glycoproteins Ilbandlila,as well as glycoprotein Tb,can beexcluded as the

locationfor the new

Ca2+-binding

sites that appear on the surface

of platelets stimulatedwith ADP.

Theseconclusions are dependent upon adequate definition

ofthe membrane abnormalities present in thrombasthenic and Bernard-Soulier platelets. It is generally accepted that a deficiency

in glycoproteinsIlb andIlIais themajor change in

thrombas-thenicplatelet membranes; however, other as yet poorly

char-acterizedchangesin membrane glycoproteins have been detected

by two-dimensionalgel electrophoresis (23). Nonetheless, the

identification ofglycoproteins Ilb and lIaas amajorbinding

site for extracellular

Ca2+

is supported by other observations. When glycoproteins Ilb andIllaareisolated from platelet

mem-branes,theyforma

Ca2+-dependent

complex (24, 25). In stim-ulated platelets, this complex appears to mediate fibrinogen

binding, a process that also requires divalent cations (1). In contrast, glycoprotein

Tb,

which bindsvonWillebrand factor in

thepresenceofristocetin,does not appear to bind extracellular

Ca2'

and has no known structural or functional requirement

for

Ca2+

(2). Similarly, glycoprotein V,which may also be de-creased inBernard-Soulierplatelets, appears to beinvolvedin

thrombin-induced platelet activation, a process that does not require extracellular

Ca2+

(2, 26).

Since there are -44,000 glycoprotein Ilb-Illacomplexes/

platelet(1, 6),the numberof

Ca2+

ionsbound to each complex may be estimated using the data in Table I. Normal platelets have an average of86,000 higheraffinity and 389,000 lower

affinity

Ca2+-binding

sites. Thrombasthenic platelets have an

averageof 7,000 higher affinity and 144,000 lower affinity

sites/

platelet. Assuming that the differences between the normal platelets and the thrombasthenic platelets are due

exclusively

to the absence of

glycoproteins

IIb and

lIla,

normal platelets containapproximatelytwo

Ca2+-binding

sitesperIlb-Illa

com-plex withanaverageKdof9 nM andsix

Ca2+-binding

sitesper

Ilb-IIa complexwithaKdof 400nM.Recently,

Fujimura

and Phillips(24),using isolated lIband

Ila,

demonstratedthatthe

Ca2+

concentration

required

for formation ofthe

glycoprotein

Ilb-Illa

complex is -1 ,uM. This

Ca2+

concentration is similar

to theK ofthe lower

affinity

Ca2+-binding

sites thatweobserved

onintactplatelets. However, the

relationship

between the

Ca2+

requiredforformation ofthe

I1b/IIla

complex withinthe

platelet

membrane and the

Ca2+

thatwe have found to be bound to

glycoproteins IIb and

lIIa

using

intact

platelets

remains to be

explored.

Using two separate

approaches,

we found that

removing

Ca2+

from the

platelet

surfacedecreasedthe

ability

of

platelets

aggregationand

_{['4Clserotonin}

release weremarkedlydecreased atfree

Ca2"

concentrations<7 nM. This free

Ca2"

concentration

is similartothe apparentKdof thehigher affinity

Ca2"-binding

sites onthe platelet surface andsuggests that these sites must

be at leastpartially occupiedfor normalplateletfunction.

Fur-thermore,thelanthanide,

Gd3",

wasable todisplace

Ca2"

from its specificbinding siteson theplateletsurface, resultingin an

inhibitionofplateletaggregationandserotonin release. Thefactthat I mM

Mg2"

alonewasunabletosupportfull

aggregation andserotonin release but could do so if at least 7

nMfree

Ca2"

wassimultaneouslypresent suggests thatdivalent

cation binding sites on platelets are heterogeneous (20, 27). Although notspecifically addressed by the presentstudies, the

data are consistent with the presence ofdivalent cationbinding

sites on platelets distinct from the high affinity binding sites

specific forCa2+. These sites would be capableofbindingeither

Ca2+ or

Mg2'

at millimolar concentrations of these cations.

Occupationof such sitesbyeitherCa2+orMg2+may berequired for partial platelet aggregation. In contrast, the high affinity

bindingsites defined in the presentstudyarespecificfor Ca2 , and occupancyofthese byCa2'appearstoberequiredfor both serotonin release and maximal platelet aggregationinresponse

toADPorepinephrine.

Howmight thebindingof extracellularCa2+tospecific high affinitysitesonglycoproteinsIlb, Illaorother membrane pro-teins be involved in platelet activation? On the one hand, the highaffinitybindingof extracellular

Ca2+

toIlb orIlamaybe

aprerequisite for theseglycoproteins to bindfibrinogen max-imally and thereforesupport maximalaggregation (1).On the otherhand, thebinding ofextracellularCa2+may also influence the platelet secretory responsedirectly, independent of itseffect

onplatelet aggregation.Forexample,anumberofsurface

mem-brane proteins believed to be involved in receptor-mediated

plateletactivation areCa2'dependent(28, 29). However,further studies will be necessary toclarifythetopographyof these mem-braneproteins and determine whether they are capable ofbinding extracellular

Ca2'

during ADP- or epinephrine-induced platelet activation.

Acknowledgments

Theauthorswishto expresstheir gratitude to Betty Belmonte and Judy Turnbull fortheir expert technical assistance and to Janet Waxman for

typingthemanuscript.

Thisstudy was supported in part by grants HL29018 and HL26523

fromtheNational Institutes of Health, by the American Cancer Society (institutional research grant IN135),and byBiomedical Research support grantS07-RR-05415-21 awarded by the Biomedical Research Support

GrantProgram, Division of Research Resources, National Institutes of Health.

References

1.Bennett, J. S., and G. Vilaire. 1979. Exposure of platelet fibrinogen receptors byADP and epinephrine.J. Clin.Invest. 64:1393-1401.

2.Berndt, M. C., and D. R. Phillips. 1981. Platelet membrane pro-teins:composition and receptorfunction. In Platelets in Biologyand Pathology. J. L. Gordon, editor. Elsevier/North-Holland Biomedical Press, Amsterdam. 2:43-75.

3. Brass, L. F., and S.J.Shattil. 1982.Changes in surface-bound

and exchangeable calcium during platelet activation. J. Biol. Chem. 257:14000-14005.

4.Kretsinger,R. H. 1979.The informational role of calcium in the

cytosol.Adv. Cyclic Nucleotide Res. 11: 1-26.

5.Tangen,O., H. J. Berman, and P. Marfey. 1971. Gel filtration:

a newtechnique for separation of blood platelets from plasma.Thromb.

Diath. Haemorrh. 25:268-278.

6.Bennett, J. S., J. A. Hoxie, S. F. Leitman, G. Vilaire, and D.B.

Cines. 1983. Inhibition of fibrinogen bindingtostimulated human

plate-lets by amonoclonalantibody. Proc. Nat!. Acad. Sci. USA.

80:2417-2421.

7.Howard, L.,S.Shulman, S.Sadanandan, and S. Karpatkin. 1982.

Crossedimmunoelectrophoresis of human platelet membranes. J. Biol.

Chem. 257:8331-8336.

8. Peterson, D., and B. Wehring. 1981. Isoelectric characteristics and surface radioiodination of normal and thrombasthenic platelet

membraneglycoproteins. Thromb. Res. 22:53-65.

9. Peerschke, E., M. B. Zucker, R. Grant, J. Egan, and M. M. Johnson. 1980. Correlation between fibrinogen bindingtohuman plate-letsandplateletaggregability. Blood. 55:841-847.

10. Walsh,P., D.C.B.Mills,F.Pareti, G. Stewart,D.MacFarlane,

M. M.Johnson,and J.Egan. 1975.Hereditary giant platelet syndrome.

Br.J. Haematol. 29:639-654.

11.DeMarco,L., and S.Shapiro. 1981. Properties of human

asialo-FactorVIII. J. Clin.Invest. 68:321-328.

12. Portzehl, H., P. C.Caldwell, and J.C. Ruegg. 1964. The

de-pendenceof contractionandrelaxation of muscle fiber from the crab Maia squinado on the internal concentration of free calcium ions. Biochim. Biophys.Acta. 79:581-591.

13.Munson, P. J., and D. Rodbard. 1980. LIGAND: A versatile computerized approach for characterization ofligand-bindingsystems.

Anal. Biochem. 107:220-239.

14.Scatchard, G. 1949. Theattractions of proteins for small molecules

andions.Ann.NYAcad. Sci. 51:660-672.

15.Sillen, L. G., and A. E. Martell. 1971. StabilityConstants of Metal-IonComplexes. Special PublicationNo.25,TheChemical Society, BurlingtonHouse, London.

16. Bers, D. M.1982.Asimple method forthe accuratedetermination offree calciuminCa-EGTA solutions.Am.J.Physiol. 242:C404-C408.

17.DosRemedios, C. G. 1981. Lanthanideion probesof

calcium-binding sitesoncellular membranes.CellCalcium. 2:29-51.

18. Shattil, S.J., J. A. Montgomery, and P. K. Chiang. 1982. The

effect of pharmacologic inhibition ofphospholipid methylation on human

platelet function.Blood. 59:906-912.

19. Zucker, M. B., and R. A. Grant. 1978. Nonreversible loss of

platelet aggregabilityinduced by calcium deprivation. Blood. 52:505-514.

20. Scrutton, M. C., and C. M. Egan. 1979. Divalent cation

re-quirements for aggregation of human blood platelets and the role of

anticoagulant. Thromb. Res. 14:713-727.

21. Best, L. C., E. A. Bone, P. B. B. Jones, and R. G. G. Russell. 1980. Lanthanum stimulates the accumulation ofcyclic AMP and inhibits

secretion and thromboxane formation in human platelets. Biochim.

22. Peerschke, E., R. Grant, and M. Zucker. 1980. Decreased

as-sociation of 45calcium with platelets unabletoaggregatedueto throm-basthenia or prolongedcalciumdeprivation.Br.J.Haematol. 46:247-256.

23. McGregor, J. L., K. J. Clemetson, E. James, A.Capitanio, T. Greenland, E. F. Luscher, and M. Dechavanne. 1981. Glycoproteins of platelet membranesfrom Glanzmann's thrombasthenia. Eur. J. Biochem.

116:379-388.

24. Fujimura, K., and D. Phillips. 1982. Calcium ion regulation of Ilb-III complex formation within platelet plasma membrane proteins. Circulation. 66(Suppl. 2):II- 174.

25. Jennings, L. K., and D. R. Phillips. 1982. Purification of gly-coproteins 1Ib and III from human platelet plasma membranes and characterization of a calcium-dependentglycoprotein Ilb-III complex. J.Biol. Chem. 257:10458-10466.

26.Clemetson, K.,J. L.McGregor,E.James,M.Dechavanne, and

E. F. Luscher. 1982. Characterization of theplatelet membrane gly-coproteinabnormalities in Bernard-Souliersyndromeandcomparison

with normal by surface-labeling techniquesandhigh resolution two-dimensionalgel electrophoresis.J. Clin. Invest. 70:304-311.

27. Lages, B., and H. J. Weiss. 1981. Dependence of humanplatelet

functional responsesondivalent cations: aggregation and secretion in heparin- and hirudin-anticoagulated platelet-rich plasma and effects of chelating agents. Thromb. Haemostasis. 45:173-179.

28.Feinstein, M. B. 1982. The role of calmodulin in hemostasis. Prog. Hemostasis Thromb. 6:25-61.