THE JOURNAL OF BIOLOGICAL. CHEM~~RY Vol. 265, No. 21, Issue of July 25, pp. 12623-12628, 1990 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

The Human ar2-Macroglobulin Receptor Contains High Affinity Calcium Binding Sites Important for Receptor Conformation and Ligand Recognition*

(Received for publication, March 5, 1990)

Stiren K. MoestrupS, Keld KaltoftB, Lars Sottrup-Jensenli, and Jergen Gliemanns From the Slnstitute of Physiology, Slnstitute of Human Genetics, and IlDepartment of Molecular Biology, University of Aarhus, DK-8000 Aarhus C, Denmark

The receptor for cuz-macroglobulin-proteinase com- plexes &MR) was purified recently, and its binding of ligand was shown to depend on calcium ions (Moestrup, S. K., and Gliemann, J. (1989) J. Biol. Chem. 264, 155’74-155’77). This paper shows that the 440-kDa human placental azMR is a cysteine-rich glycoprotein with high affinity calcium binding sites important for receptor conformation; and the relationship between Ca2+ concentration and receptor function is presented.

Autoradiography showed 46Ca2+ binding to the 440- kDa cr2MR blotted onto nitrocellulose from a sodium dodecyl sulfate-polyacrylamide gel. a2MR immobilized on nitrocellulose in the absence of sodium dodecyl sul- fate bound 46Ca2+ in the presence of 5 mM Mg”‘, and 2-3 PM unlabeled Ca2+ was required to displace half of the bound 46Ca2+. The calcium concentration depend- ence showed upward concave Scatchard plots, and the number of binding sites was estimated to be approxi- mately eight/Lv2MR molecule.

Binding of calcium did not change in the pH range 6.5-8.0 but decreased at lower pH values. Addition of Ca2+ to the medium was necessary for receptor binding of the cu2-macroglobulin-trypsin complex, and half of the maximal binding capacity was obtained with about 16 KM Ca2’ at pH 7.8. The requirement for calcium was increased at lower pH values, and half of the maximal lz61-a2M-trypsin binding was obtained with about 30- 40 /.IM Ca2+ at pH 7.0. Monoclonal antibodies were produced against arzMR, and one of them distinguished between the Ca’+-occupied and nonoccupied forms. Like Ca’+, Sr*+ and Ba2+ elicited ligand binding affinity and competed for binding with 46Ca2‘ in the order Ca2’ > Sr2+ > Ba*+. In conclusion, calcium ions bind specif- ically to a2MR with high affinity, and it is likely that several sites on the a2MR molecule have to be occupied to elicit the conformation recognizing the ligand.

The tetrameric glycoprotein in human plasma cy2-macro- globulin (a,M)’ and the homologous pregnancy zone protein

* This work was supported by grants from the Danish Biomem- brane Research Centre, the Danish Medical Research Council, the Nordic Insulin Foundation, the NOVO foundation, the Danish Can- cer Society, and Aarhus University Research Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adver- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

’ The abbreviations used are: OVUM, human a*-macroglobulin; ol,MR, a,-macroglobulin receptor; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; CHAPS, 3-[(3-cholami- dopropyl)dimethylammonio]-1-propanesulfonic acid; EGTA, [ethyl- enebis(oxyethylenenitrilo)]tetraacetic acid; Hepes, 4-(2-hydroxy- ethyl)-l-piperazinelthanesulfonic acid.

form stable complexes with a wide variety of proteinases. Each of the HO-kDa subunits in cvzM contains an exposed peptide stretch, the bait region, which is recognized and cleaved by several proteinases. This is followed by the cleavage of an internal thiol ester in each subunit and a conformational change leading to the exposure of a previously concealed receptor recognition site in each subunit. Incubation of native c+M with small nucleophiles such as methylamine also leads to cleavage of the internal thiol esters and exposure of the receptor recognition sites (for review, see Ref. 1).

Receptors for cw2M-proteinase complexes (referred to as (YOM receptors, cuzMR) have been characterized kinetically in fibro- blasts (2-4), macrophages (5-7), hepatocytes @-lo), and membranes derived from placental syncytiotrophoblasts (11). cvzMR from these cell types recognizes both azM- and preg- nancy zone protein-proteinase complexes (3, 4, 11). Bound complex is rapidly taken up by receptor-mediated endocytosis in hepatocytes from rodents (12, 13) and humans (lo), and uptake into these cells largely accounts for the rapid removal of ayM-proteinase complexes from the circulation, occurring with a half-time of 2-3 min in rodents (12, 13). It is not known whether cvnMR of the normal human placenta is able to internalize bound ligand rapidly or whether it serves as a point of attachment of the complexes to the surface of the syncytiotrophoblast (11).

We have recently purified a2MR from detergent-solubilized rat hepatic (14) and human placental (15) membranes as an approximately 440-kDa single-chain protein capable of bind- ing the ligand with an apparent Kd of about 400 PM. Ligand binding occurred only in the presence of Ca*‘, in agreement with previous results using intact cells (2, 9) and membrane preparations (11, 16). The purpose of the present work was to clarify the Ca” dependence of ligand binding. We show that (r2MR binds Ca*+ with high affinity, and we quantify the relationship between bound Ca2+ and conformation-depend- ent ligand binding activity.

EXPERIMENTAL PROCEDURES

Preparation of Human Placental olJfR--Membranes were pre- pared as described previously from placentas obtained after normal deliveries (11). In brief, pieces (l-2 g) of villous tissue were washed in ice-cold phosphate-buffered saline and homogenized at 0 “C in 250 mM sucrose, 10 mM Hepes, 5 mM EDTA, 0.1 mM phenylmethylsul- fonyl fluoride, pH 7.4. The 3,000 X g supernatant (5 min) was centrifuged at 48,000 X g for 40 min and homogenized in 140 mM NaCl, 0.6 mM CaCls, 10 mM sodium phosphate, 0.1 mM phenylmeth- ylsulfonyl fluoride, pH 7.8. The membranes were solubilized by re- suspension in 6 mg of protein/ml in the above buffer containing 0.5% CHAPS (Aldrich). The suspension was stirred for 10 min at 0°C followed by centrifugation for 40 min at 48,000 X g. The supernatant was recirculated for 16 h at 0 “C through an affinity column with Sepharose-immobilized methylamine-treated apM (14,151 using 8 mg

12623

12624 c-uP-Macroglobulin Receptor Calcium Binding Sites

of ligand per g of dry Sepharose. The column was washed with about 100 bed volumes of buffer containing 1% Triton X-100 or 0.8% octyl &glucoside. The protein (referred toas OI~MR preparation) was eluted in buffer containing 0.1% Triton X-100 or 0.8% octvl a-alucoside and 2 mM EGTA, pH 6.0, and stored at -20 “C in the presence of 35% glycerol. The preparation contained the 440-kDa (Y~MR and small amounts of 40-kDa and 70-90-kDa proteins, as described previously (14, 15).

Determination of Amino Acid Composition-Approximately 1 ml of a,MR dissolved in octyl P-glucoside-containing buffer (A2a0 = 0.150) was lyophilized, redissolved in 0.5 ml of 50% formic acid, and gel filtered on a Superose 12 column equilibrated and eluted with 50% formic acid (see Fig. 1). The fractions containing the 440-kDa receptor were combined, and 50-~1 aliquots were dried and hydrolyzed with 6 M HCl at 110 “C for 20 or 72 h. Some aliquots were also oxidized with performic acid (17) prior to the 20-h hydrolysis, or they were hydro- lyzed in the presence of 3,3’-dithiodipropionic acid (18). For deter- mination of tryptophan, samples were hydrolyzed with 1 M NaOH at 110 “C for 20 h. After acidification with acetic acid, the hydrolysates were analyzed using a Hewlett-Packard Amino Quant instrument operated according to the manufacturer’s instructions and on an instrument built and operated as described by Barkholt and Jensen (18). Bovine prothrombin fragment 1 containing 4 phenylalanine and 3 tryptophan residues (19) was used to determine the recovery of tryptophan. Details of the methodology are provided in the legend of Table I. Calculation of E:?o, icm from the content of tryptophan, tyrosine, and cysteine was performed according to Gill and von Hippel (20). For calculation of the carbohydrate content, we assumed that the fraction of total sugar provided by aminosugar (glucosamine) was 0.4 (21).

Preparation of Monoclonal Anti-a2MR Antibodies-Male BALB/c mice were immunized intraperitoneally four times at ll-day intervals with about 10 pg of LY~MR preparation in 100 ~1 of affinity column effluent. Three days after the last booster injection, spleen cells were fused with 5 x 10’ mouse myeloma (NS-1) cells and seeded according to standard procedures (22). Screening for positive wells employed a solid-phase enzyme-linked immunosorbent assay using microtiter plates with about 50 ng of a,MR per well and peroxidase-conjugated rabbit anti-mouse IgG (Dacopatts, Copenhagen, Denmark). Cells were cloned twice from positive wells by limiting dilution, and the class and subclass specificities of the monoclonal antibodies were determined using a commercial mouse typing kit (Bio-Rad 172-2055).

Four monoclonal antibodies (designated ABMR-1 through 4) of the IgG, class were characterized, and none of them inhibited binding of a*M-trypsin. A description of their binding to different epitopes will appear elsewhere (38).

Blotting Procedures-SDS-PAGE was performed according to Laemmli (23) as described previously (14). Transfer of proteins onto nitrocellulose was performed using a semidry electroblotter (JKA- Biotech, Copenhagen, Denmark) according to the suggestions of the manufacturer. Ca2+ binding was detected by the method of Marayama et al. (24). In brief, the nitrocellulose sheet was incubated for 15 min at 4 “C with 10 mM Tris, 5 mM MgCl*, pH 7.4. The sheet was washed three times in distilled water followed by autoradiography for 24-72 h using Hyperfilm (Amersham Corp.). Immunoblot analyses were performed using the peroxidase-conjugated rabbit anti-mouse IgG as the secondary antibody.

‘?a’+ and ‘251-a2M-trypsin Binding Assays-For Ca” binding, 2 ~1 of a2MR preparation was applied to nitrocellulose (Sartorius SM 12752) discs with a diameter of 4 mm. The discs were dried for 30 min at 4 “C, and 2 ~1 was applied to the other side of the disc now containing about 2 pmol of a,MR. The discs were washed three times in 60 mM KCl, 10 mM Tris, 10 mM glycine, 5 mM MgC12, pH 7.8 (if not stated otherwise) and incubated at 4 “C in 200 ~1 of the same buffer with 50 nCi of “Ca’+. The incubations were stopped by rapid washing in 2 x 5 ml of ice-cold buffer, and the discs were assayed for radioactivity in a liquid scintillation counter. Blank values deter- mined with discs not containing LY~MR showed association of 0.2% of the 45Ca2+ in the 200 ~1 of incubation buffer. This value was not different from that obtained in the presence of a saturating concen- tration of unlabeled Ca*+ (1 mM) and was subtracted from-all other values. Omission of Mt?’ from the buffer increased the blank value to a magnitude prohibiting the measurement of 45Ca*+ binding to n2MR. Analyses using atomic spectrophotometry showed that the deionized glass-distilled water used in these experiments contained less than 0.5 pM Ca’+.

Binding of iz51-cu2M-trypsin was performed using the same assay with the following modifications. About 0.2 pmol of cr2MR was applied

to the discs followed by soaking in buffer containing 1% Tween 20 to prevent nonspecific absorption of protein. The incubations were performed for 16 h, a time sufficient to achieve a steady state of binding (data not shown). Blank values (l.O-1.5% of the added tracer) were not different from the values obtained in the presence of a saturating concentration of unlabeled cupM-trypsin (140 nM) and were subtracted from all other values.

The following results were obtained in control experiments. rZ51- Labeled (u~MR applied to the nitrocellulose discs did not dissociate during the incubation and washing procedures. Soaking of the discs in Tween 20, necessary for obtaining reliable ““I-o12M-trypsin bind- ing, did not perturb Ca*+ binding. Previously, a*MR binding activity was measured in 140 mM NaCl, 10 mM sodium phosphate, 0.6 mM Ca’+, 4% bovine serum albumin, pH 8.0 (14, 15). Saturable binding of lz51-cu?M-trypsin in that buffer (with filters soaked in albumin before the addition of tracer) was not different from the results obtained when using the present buffer. However, Tween 20 was preferred since it gave lower blank values than albumin. No difference in binding activity was observed between (Y~MR eluted in buffer containing Triton X-100 or octyl P-glucoside. Storage at -20 “C in the presence of 35% glycerol was important for maintenance of the binding activity, and under these conditions cv*MR was stable for at least 6 weeks.

RESULTS

Amino Acid Composition--Fig. 1 shows the elution profile of the lyophilized cvzMR preparation in 50% formic acid. SDS- PAGE shows that the large peak (he I) contains only the 440-kDa azMR whereas the shoulder (lane 2) contains in addition the minor proteins present in the crzMR preparation (15). In several experiments the content of minor proteins was estimated from the elution profile and from gel scannings to be lo-15%.

Table I shows the amino acid composition of the 440-kDa CQMR. The content of half-cystine is remarkably high. The only amino sugar was glucosamine, accounting for about 4.5% of the combined molar yield of amino acids. The carbohydrate

kDa

0.15

-1 I- -360

-200

- 67

2

5 10 15 20 25 Fraction no

FIG. 1. Gel filtration of (Y~MR. The a2MR preparation was eluted from the affinity column in 0.8% octyl P-glucoside (A280 = 0.150), lyophilized, redissolved in 50% formic acid, chromatographed on a 24-ml Superose I2 column, and eluted in that solvent. Fraction size was 0.5 ml. SDS-PAGE (5-20% polyacrylamide) shows that the first peak (peak I) contains only the 440-kDa (Y~MR (lane 1, fractions 15-17), whereas the second (peak 2) contains the proteins of lower molecular weight (lane 2, fractions 20 and 21). Molecular size markers were aaM-methylamine (360 kDa), rat al-inhibitor 3 (200 kDa), and bovine serum albumin (67 kDa).

cup-Macroglobulin Receptor Calcium Binding Sites 12625

TABLE I Amino acid composition and glucosamine content of c&R

Aliquots (50 ~1) of the combined fractions of the eluate containing pure receptor (fractions 15-17 shown in Fig. 1, A280 = 0.103) were hydrolyzed in 6 M HC1 for 20 or 72 h at 110 “C. The content was estimated to be 30.3 nmol (3.35 pg) of amino acid residues and 1.54 nmol (276 ng) of glucosamine (see “Experimental Procedures” and Footnotes a-f, below). Galactosamine was not detected. When assum- ing that glucosamine accounts for 40% of the carbohydrate mass, the total carbohydrate is 17% (w/w) (75 kDa), and the protein is 83% of the 440-kDa glycoprotein. Ei2.i E,” was calculated as 12.5 when the amino acid plus carbohydrate estimates (sum = 4.04 pg) were related to the A,,, (0.103 in 50 ~1) of the nonhydrolyzed material.

B

Residue mol % Residues/molecule

Asp + Asn Thr” Serb Glu + Gin Pro GUY Ala c SE.d “SIP Met Iled Leu Tyr Phe His LYS Arg Trp’

Sum

GlcNH,’

13.58 448 5.68 187 7.16 236 9.28 306 4.93 163 8.53 281 5.61 185 6.36 210 5.88 194 1.52 50 4.34 143 6.77 223 2.74 90 3.11 103 2.66 88 4.11 135 5.80 191 1.93 64

99.99 3297

148

a Corrected for an estimated hydrolysis loss of 5%. * Corrected for an estimated hydrolysis loss of 10%. ‘Determined as cystic acid or as a 3,3’-dithiodipropionic acid

adduct (Ref. 18). ’ Results of hydrolysis for 72 h. ‘Results of hydrolysis with 1 M NaOH (110 “C, 20 h). ’ Corrected for an estimated hydrolysis loss of 50%.

groups constituted about 17% of the mass of the receptor, leaving about 365 kDa for the peptide moiety (approximately 3300 residues). From the data shown in Table I, E:g, 1 cm was determined as 12.5 when using 50% formic acid as solvent (see the legend to Table I). At neutral pH and under nonde- naturing conditions E?& icm is probably 5-10% higher, i.e. about 13.5. E:& lem was calculated as 11.5 on the basis of the content of tryptophan, tyrosine, and cysteine (20). We decided to use the mean value of 12.5 since both methods for the determination of E:&, I Cm are subject to errors.

Binding of Ca’+-Fig. 2, panel A, shows a nitrocellulose blot of cr*M-methylamine for control (lane 1) and the a2MR prep- aration (lane 2). The proteins were blotted from a continuous 4% polyacrylamide gel, and the minor proteins in the azMR preparation therefore migrated with the front. Panel B shows the results of incubation of the nitrocellulose sheet with 45Ca2’ followed by washing and autoradiography. The blotted 440- kDa azMR bound 45Ca2’ whereas the minor proteins did not (lane 2), and no binding to cupM-methylamine was observed (lane I).

The following experiments were designed to quantify Ca2+ binding and the relationship between Ca2+ occupancy and a2M-trypsin binding. For this purpose it was important to use nondenaturing conditions (15) and to employ the same type of assay system for the two measurements. We therefore used the assay with nitrocellulose-immobilized a2MR preparation and disregarded the presence of the minor proteins since the

kDa

360 -

200 -

12 12

FIG. 2. SDS-PAGE and Ca2+ blot of a2MR. oc*M-methylamine (control, mass = 360 kDa) and CQMR were subjected to SDS-PAGE (4% polyacrylamide, nonreducing conditions) and blotted onto nitro- cellulose. The nitrocellulose strips were incubated with 45Ca*+ (lo6 cpm/ml), washed, dried, autoradiographed, and stained with Amido Black. Panel A shows the blotted. stained nroteins (lane 1. cu?M- methylamine; lane 2, o~zMR preparation). Panel B shows the ‘results of the autoradiography (lane 1, a2M-methylamine; lane 2, o,MR preparation with 45Ca2+ bound to the 440-kDa LY~MR).

0 min

I I I 0 4 6 12 16

min FIG. 3. Time course of Ca’+ binding to nitrocellulose-im-

mobilized a2MR. Nitrocellulose discs with 2.0 pmol of (Y,MR were incubated in 200 ~1 of buffer with 0.25 &i/ml (0.3 pM) “‘Ca*’ at 4 “C followed by washing in nonradioactive buffer. The points are the mean values of four replicates f 1 SD. The inset shows dissociation of bound Ca*+ into nonradioactive buffer alone (0) or with 2 mM EGTA (0).

440-kDa cvzMR accounts for nearly all Ca2+ (Fig. 2) and asM- trypsin (15) binding as well as S-90% of the absorbance at 280 nm (Fig. 1).

Fig. 3 shows that a plateau of binding of 0.3 PM 45Ca2’ to nitrocellulose-immobilized a2MR is achieved by about 15 min. The inset shows that dissociation of 45Ca2’ from the filter occurs quite slowly, leaving enough time for the washing inherent in the assay procedure. However, the rate of disso- ciation is markedly increased in the presence of EGTA.

Fig. 4 shows the concentration dependence of calcium bind- ing at pH 7.8. Scatchard plots were consistently upward concave as shown in the inset, suggesting multiple sites with different affinities. The concentration of added unlabeled Ca2+ giving half-maximal binding of 45Ca2+ was determined from this and four similar experiments as 1.8 f 0.3 pM (1 SD.). The number of calcium sites/cY2MR molecule was cal- culated from the intercept with the abscissa (see legend to Fig. 4) in the five experiments as 8 + 2.

Effect of Ca2’ and pH on a2M-trypsin Binding-Fig. 5 shows the dependence of ‘251-oc2M-trypsin binding on the concentra- tion of Ca2+ at pH 7.8. No binding was observed when Ca2+ was not added to the buffer, and 50-100 pM Ca2+ was required

12626 cu2-Macroglobulin Receptor Calcium Binding Sites

0.07 1

9 0.06

.3 0.05

;: d 0.04

; 0.03

1 i 0.02

0.01 I

O.OO;, 40 50 60 70 Free (PM)

FIG. 4. Concentration dependence of Ca2+ binding. Immobi- lized (u,MR (2.0 pmol/disc in 200 11) was incubated with 45Ca*+ at the indicated Ca” concentrations for 15 min at 4 “C. The inset shows the data plotted according to Scatchard. The curue is computed according to the assumption that Ca”+ binds to two independent sites with different affinities. The constants calculated according to this model are: I&, 0.62 PM; I&2, 12.71 PM; I&,,, 23.6 nM; I&, 57.2 nM, where the Kd values are the apparent affinity constants, and the R, values are the Cay+ site concentrations. Thus, the site concentration is 81 nM with 10 nM a,MR, and the receptor contains about eight sites/ molecule.

OYI 0 50 100 1000 2000

ca2+ (/Ad)

FIG. 5. Effect of Ca”’ on (Y~MR ligand recognition. Immobi- lized n?MR (0.2 pmol/disc) was incubated (200 ~1) for 16 h at 4 “C in buffer containing 20 pM “’ I-apM-trypsin (7’) and Ca’+ as indicated. The points are the mean values of four replicates + 1 S.D.

oc PH

FIG. 6. Effect of pH on binding of cuzM-trypsin (!I’) and Ca’+. Immobilized (u~MR was incubated at the indicated pH values with 20 pM ‘“‘I-cuzM-trypsin for 16 h either in the presence of 1 mM (0) or 32 pM (0) Ca*’ or with 0.25 &i/ml 45Ca2+ (A). The results are expressed as percent of the maximal “‘1-a2M-trypsin binding (pH 7.8) obtained at the given Ca*’ concentration or as percent of the maximal “CaL+ binding.

for maximal binding activity. About 16 pM Ca’+ was required for half-maximal activity (17.1 + 3.9 pM (1 SD.) in five experiments).

Fig. 6 demonstrates the interrelationship between Ca*+ and

proton concentrations. At 1 mM Ca*+, binding of lZ51-a2M- trypsin increased with increasing pH, and half of the maximal binding was obtained at pH about 7.0. Binding of cYzM-trypsin was more pH sensitive at submaximal Ca*+ concentrations, and at 32 pM Ca*+, pH 7.4, was required to obtain half the binding observed at pH 7.8. On the other hand, binding of “%a*+ remained nearly unchanged in the pH range 8.2-6.5 and decreased at more acidic pH. It is concluded that both Ca*+ binding and deprotonation are required for the binding of LuzM-trypsin and that the requirement for deprotonation is increased at suboptimal Ca*+ concentrations.

Ca*+-dependent Reactivity with a Monoclonal Antibody- Three of the antibodies did not distinguish between the Ca*+- occupied and nonoccupied forms, and results with one of them are demonstrated in Fig. 7A together with an immunoblot showing reaction with the 440-kDa cu2MR. Fig. 7B shows that one of the antibodies distinguished between the two forms. Separate experiments (not demonstrated) showed identical gel filtration profiles (Sephacryl S-400, pH 7.8) of the (Y*MR preparation in the absence and presence of Ca”‘, indicating lack of Ca’+-induced oligomerization of the receptor. Thus, Ca2+ induces a conformational change in the a2MR monomer.

Specificity of the Ca*’ Binding Sites-Several divalent metal ions and La3’ were tested for the ability to elicit (u*MR ligand binding activity. Fig. 8 shows that in addition to Ca”, Sr*+ and Ba*+ elicited a2M-trypsin binding activity in the order Ca*+ > Sr2+ > Ba*+ whereas the other divalent ions shown in Fig. 8 and La3’ did not. Fig. 9 shows that Sr2+ and Ba*+ competed for binding of 45Ca2’ following the same order. In addition, La”+ was an effective competitor of Ca*+ binding even though it did not elicit Lu2M-trypsin binding activity. Thus, the two ions most closely related to Ca*+ both bind to

1.0

A 0.6

360-

Fj 06 zoo-

$ /

ay 0.4

0.2

0.0 b 0.1

ng’antib:dy ad%d Ii

FIG. 7. Monoclonal antibody distinguishing between the Ca2+ and non-Ca2+ forms of (x2MR. A shows enzyme-linked im- munosorbent assays in the absence (0) or presence (0) of 2 mM EGTA using antibody A2MR-2 and a Western blot demonstrating reaction with the 440-kDa cunMR. B shows the analogous data using antibody A2MR-3. The antibodies were diluted in RPM1 1640 con- taining 10% fetal calf serum, pH 7.6.

/- C&l’+

sr’+ LJf Ed+ k

/

-// I

:i

:!I

z

0. 0. 5. 4 0 1.00 0

cr2-Macroglobulin Receptor Calcium Binding Sites 12627

Moreover, increasing protonation increases the requirement for Ca’+. What does this mean in physiological terms? As a minimum, Ca’+ is necessary for the proper ligand binding conformation of CQMR. If the Ca2+ sites are on the extracel- lular side of the protein, the Ca2+ concentrations should always be sufficient for optimal ligand binding, even if only the sites with the lowest apparent affinity are important for the receptor conformation. However, a decrease in Ca*+, to- gether with decrease of pH, would cause a rapid and efficient removal of the protein ligand from a2MR. This probably happens in the endocytotic vesicle after internalization of the a*MR.cu*M-proteinase complex. It seems unlikely that the sites are facing the cytosolic side since at least 16 pM Ca*+ is required for half-maximal expression of ligand binding activ- ity. However, sites in or near the membrane-spanning region might be exposed to Ca2+ concentrations in the range of 5-50 PM and thus be regulatory.

Ion (InId)

FIG. 8. Effect of divalent ions and La” on azMR ligand binding activity. Immobilized a2MR was incubated with ‘251-012M- trypsin (2’) as explained in the legend to Fig. 5 and ions as indicated. 1; -some cases, ihe presence of an ion (e.g. Zn”) increased the association of Y-cuZM-trypsin to the nitrocellulose in the absence of receptor, and data were corrected accordingly.

i

Additions (mM) + 2

“5 z r-l Ca Sr sr Ba Ba La

Electrophysiological studies have shown that azM added extracellularly can induce spikes of hyperpolarization elicited by opening of Ca*+-dependent potassium channels (32). It is tempting to speculate that this phenomenon is somehow related to the Ca*+ binding of (Y~MR and internalization of the complex. In any case, the role of the Ca” binding sites, apart from keeping the protein in the proper conformation, is unclear, and information on sequence and structure is essen- tial for clarification.

FIG. 9. Inhibition of 4aCaz* by divalent ions and La3+. Im- mobilized azMR was incubated for 15 min with %a’+, as explained in the legend to Fig. 3, and ions as indicated. None of the ions changed the association of ‘%a*+ to the nitrocellulose in the absence of cu*MR.

(-Y*MR and induce the conformational change necessary for cw2M-trypsin binding.

DISCUSSION

The results show that (r2MR is a cysteine-rich glycoprotein containing multiple binding sites for Ca*+. Moreover, occu- pancy with Ca*+ or related ions induces a conformational change necessary for binding of azM-proteinase complex.

The upward concave Scatchard plot shows that the Ca*+

Little is known about Ca*+ binding to other mammalian cell membrane receptor proteins. The hepatic asialo-glycopro- tein receptor is reported to bind Ca*+ only when occupied with the protein ligand (33) whereas others have reported Ca*+ binding (Kd = 350 FM) to the unoccupied receptor (34). A related chicken hepatic lectin requires l-2 mM Ca2+ for half- maximal ligand binding activity at pH 7.8 (35). The low density lipoprotein receptor requires Ca*+ for binding of the protein ligand (36), but no direct information on Ca*+ binding seems available. The recently cloned low density lipoprotein receptor-related protein, with possible growth-modulating ef- fects, binds 45Ca2+ when blotted onto nitrocellulose (37), but no information is available on the affinity or location of the sites.

binding is complex. The total number of sites may be under- estimated from such curved plots, and eight Ca*+ binding sites/cuzMR molecule should therefore be taken as a minimum estimate. The finding that the Ca2+ binding curve can be described adequately within the experimental errors by as- suming two separate classes of sites (see the legend to Fig. 4) does not exclude other models including negative cooperativ-

Thus, as compared with other mammalian cell membrane receptors, a2MR has high affinity Ca2+ binding sites, and Ca*+ occupancy is as a minimum essential for the ligand binding conformation of the receptor protein. These properties may provide an interesting model system for studying the effects of Ca2+ binding on protein conformation.

ity among homogeneous sites. The overall affinity (Kd = 2 PM) is similar to that of sites

in classical Ca2+ binding proteins such as calmodulin and tropomyosin C. Calmodulin binds four Ca*+, probably with pairwise different affinities and mutual interactions (25, 26). Sr*+ interacts with the same sites, although with lower affin- ity, and can elicit activation of calmodulin-activated enzymes (27). Tropomyosin C from skeletal muscle also binds four Ca*+, and two of the sites (Kd = 3 PM) are thought to be regulatory (28, 29). Recently, a single Ca” binding site (Kd = 2 PM) in the galactose chemoreceptor protein from Escherichia coli was described in detail (30). The structural basis for Ca*+ bicding to these as well as most other proteins with high affinity sites is the EF-hand motif (31). This may also apply

Acknowkdgment-We thank Dr. Karen G. Welinder, Institute of Biochemical Genetics, University of Copenhagen, for performing the analysis of nonoxidized hydrolysates.

REFERENCES

1. Sottrup-Jensen, L. (1989) J. Biol. Chem. 264,11539-11542 2. Van Leuven, F., Cassiman, J.-J., and Van den Berghe, H. (1979)

J. Biol. Chem. 254, X55-5160 3. Gliemann, J., Moestrup, S. K., Jensen, P. H., Sottrup-Jensen, L.,

Andersen, H. B., Petersen, C. M., and Sonne, 0. (1986) Biochim. Biophys. Acta 883, 400-406

4. van Leuven, F., Cassiman, J.-J., and Van den Berghe, H. (1986) J. Biol. Chem. 261, 16622-16625

5. Kaplan, J., and Nielsen, M. L. (1979) J. Biol. Chem. 254, 7323- 7328

to (Y~MR although the question obviously cannot be answered before information on the amino acid sequence is available.

The results show that efficient binding of cr2M-trypsin occurs only when a large fraction of the Ca2’ sites is occupied.

6. Petersen, C. M., Ejlersen, E., Hansen, P. W., and Gliemann, J. (1987) &and. J. Clirz. Inuest. 47,55-61

7. Moestrup, S. K., Christensen, E. I., Sottrup-Jensen, L., and Gliemann, J. (1987) Biochim. Biophys. Acta 930,297-303

8. Gliemann, J., Larsen, T. R., and Sottrup-Jensen, L. (1983) Biochim. Biophys. Acta 756, 230-237

12628 az-Macroglobulin Receptor Calcium Binding Sites

9.

10.

Gliemann, J., and Davidsen, 0. (1986) Biochim. Biophys. Acta 885,49-57

11.

12.

13.

Petersen, C. M., Christiansen, B. S., Jensen, P. H., Moestrup, S. K., Gliemann, J., Sottrup-Jensen, L., and Ingerslev, J. (1988) Eur. J. Clin. Znuest. 18, 184-190

Jensen, P. H., Moestrup, S. K., Sottrup-Jensen, L., Petersen, C. M., and Gliemann, J. (1988) Placenta 9, 463-477

Davidsen, O., Christensen, E. I., and Gliemann, J. (1985) Biochim. Biophys. Acta 846,85-92

Feldman, S. R., Rosenberg, M. R., Ney, K. A., Michalopoulos, G., and Pizza, S. V. (1985) Biochem. Biophys. Res. Commun. 128, 795-802

14.

15.

16.

17. 18.

19.

Moestrup, S. K., and Gliemann, J. (1989) J. Biol. Chem. 264, 15574-15577

Jensen, P. H., Moestrup, S. K., and Gliemann, J. (1989) FEBS Z&t. 255,275-280

Gliemann, J., Davidsen, O., and Moestrup, S. K. (1989) B&him. Biophys. Acta 980, 326-332

Hirs, C. H. W. (1967) Methods Enzymol. 11,59-62 Barkholt, V., and Jensen, A. L. (1989) Anal. Biochem. 177,318-

322 Magnusson, S., Petersen, T. E., Sottrup-Jensen, L., and Claeys,

H. (1975) in Proteases and Biobgicol Control (Reich, E., and Rifkin, D., eds) pp. 123-149, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

20. Gill, S. C., and von Hippel, P. H. (1989) Anal. Biochem. 182, 319-326

21. Kornfeld, R., and Kornfeld, S. (1985) Annu. Reu. Biochem. 54, 631-664

22. Kohler, G., and Milstein, L. (1975) Nature 256, 495-497 23. Laemmli, U. K. (1970) Nature 227,680-685

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Marayama, K., Mibawa, T., and Ebashi, S. (1984) J. Biochem. 95,511-519

Burger, D., Cox, J. A., Comte, M., and Stein, E. A. (1984) Biochemistry 23,1966-1971

Wang, C. L. (1985) Biochem. Biophys. Res. Commun. 130, 126- 130

Cox, J. A., Malnoe, A., and Stein, E. A. (1981) J. Biol. Chem. 256,3218-3222

Johnson, J. D., Charlton, S. C., and Potter, J. D. (1979) J. Biol. Chem.254,3497-3502

Putkey, J. A., Sweeney, H. L., and Campell, S. T. (1989) J. Biol. Chem.264,12370-12378

Vyas, M. N., Jacobson, B. L., and Quiocho, F. A. (1989) J. Biol. Chem. 264,20817-20821

Kretsinger, R. H., and Nockolds, C. E. (1973) J. Biol. Ch.em. 248, 3313-3326

Dixon, S. J., and Aubin, J. E. (1987) J. Cell. Physiol. 132, 215- 225

Blomhoff, R., Trolleshaug, H., and Berg, T. (1982) J. Biol. Chem. 257,7456-7459

Andersen, T. T., Freytag, J. W., and Hill, R. L. (1982) J. Biol. Chem. 257,8036-8041

Loeb, J. A., and Drickamer, K. (1988) J. Biol. &em. 263, 9752- 9760

van Driel, I. R., Goldstein, J. L., Siidhof, T. C., and Brown, M. S. (1987) J. Biol. Chem. 262, 17443-17449

Herz, J., Hamann, U., Rogne, S., Myklebost, O., Gausepohl, H., and Stanley, K. K. (1988) EMBO J. 7,4119-4127

Moestrup, S. K., Kaltoft, K., Petersen, C. M., Pedersen, S., Gliemann, J., and Christensen, E. I. (1990) Exp. Cell Res., in press