Eur. J. Biochem. 210, 1071- 1077 (1992)

0FEBS 1992

Differences in the proteinase inhibition mechanism of human a2-macroglobulin and pregnancy zone protein Poul Erik Hyldgaard JENSEN and Torgny STTGBRAND Department of Medical Biochemistry and Biophysics, University of Umeri, Sweden (Received June 30/September 30, 1992)- EJB 92 0909

Different conformational states of human az-macroglobulin (azM) and pregnancy zone protein (PZP) were investigated following modifications of the functional sites, i. e. the ‘bait’ regions and the thiol esters, by use of chymotrypsin, methylamine and dinitrophenylthiocyanate. Gel electrophoresis, mAb (7HllD6 and a 1 : 1) and in vivo plasma clearance were used to describe different molecular slates in the proteinase inhibitors. In a 2 M ,in which the thiol ester is broken by binding ofmethylamine and the ‘trap’ is closed, cyanylation of the liberated thiol group from the thiol ester modulates reopening of the ‘trap’ and the ‘bait’ regions become available for cleavage again. The trapping of proteinases in the cyanylated derivative indicates that the trap functions as in native a,M. In contrast, cyanylation has no effect on proteinase-treated a 2 M . As demonstrated by binding to mAb, the methylamine and dinitrophenylthiocyanate-treated azM exposes the receptor-recognition site, but the derivative is not cleared from the circulation in mice. The trap is not functional in PZP. In native PZP and PZP treated with methylamine, the conformational states seem similar. The receptor-recognition sites are not exposed and removal from the circulation in vivo is not seen for these as for the PZP-chymotrypsin complex. Tetramers are only formed when proteinases can be covalently bound to the PZP. Conformational changes are not detected in PZP derivatives in which the thiol ester is treated with methylamine and dinitrophenylthiocyanate. The results suggest that the conformational changes in azM are generated by mechanisms different to these in PZP. The key structure gearing the conformational changes in a2M is the thiol ester, by which the events ‘trapping’ and exposure of the receptor-recognition site can be separated. In PZP, the crucial step for the conformational changes is the cleavage of the ‘bait’ region, since cleavage of the thiol ester does not lead to any detectable conformational changes by the methods used.

In humans, two a-macroglobulins appear, the serum protein a2 macroglobulin (a2M) and the pregnancy-associated protein, pregnancy zone protein (PZP). Both are composed of two identical subunits of 180 kDa, held together by interchain disulfide bridges in dimers [ l , 21. azM is a tetramer of two non-covalently bound dimers [3], while PZP is a dimer in the native state, but is transformed to tetramers upon complex formation with proteinases [4]. These macroglobulins are proteinase inhibitors that possess the unique mechanism of covalently binding proteinases without inhibiting the active site, but they sterically hinder the bound proteinases from Correspondence to T. Stigbrand, Department of Medical Biochemistry and Biophysics, University of Umefi, S-90187 Umeri, Sweden Abbreviations. a 2 M , a,-macroglobulin; a,M-MeNH,, a,-macroglobulin treated with methylamine: a,M-CT, a,-macroglobulin complexed with chymotrypsin; ct,M-(MeNH2 + D), a,-macroglobulin treatcd with methylamine and dinitrophenylthiocyanate; CT, CIchymotrypsin; dnpSCN, dinitrophenylthocyanate; MeNH2, methylamine; PZP, pregnancy zone protein; PZP-MeNH,, pregnancy zone protein treated with methylamine; PZP-CT, pregnancy zone protein complexed with chymotrypsin; PZP-(MeNH2+ D), pregnancy zone protein treated with methylamine and dinitrophenylthiocyanate. Enzymes. a-Chymotrypsin (EC 3.4.21.I); horse-radish peroxidase (EC 1.11.1.7).

being catalytically active against high-molecular-mass substrates [ I , 51. The mechanism causing proteinase inhibition is basically the same in aZM and PZP. Proteinases activate the inhibitor by a cleavage in the ‘bait’ region, a uniquely sensitive region in the middle of the 180-kDa subunit [3, 5, 61. This cleavage triggers conformational changes in the polypeptide chain of a,M and PZP, which make the unique thiol ester (an internal covalent bond between a Cys residue and a Glu residue) susceptible for nucleophilic attack by which the proteinase can be covalently cross-linked by the s-lysyl-y-glutamyl bonds [l, 7 -91. In azM, non-covalent ‘trapping’ of proteinascs can occur after the bait-region cleavage as well [lo]. Apparently PZP cannot trap proteinases and the proteinases are essentially only covalently bound to PZP [4.11, 121. Due to conformational changes, receptor-recognition sites are exposed in the inhibitors causing removal from the circulation [13, 141. The bait-region sequence is markedly different in PZP compared to a 2 M , and is, in fact, the region with the lowest similarity. This observation suggests complementary functions for the proteinase inhibition exerted by a z M and PZP [l]. The stoichiometry of proteinase inhibition is furthermore different since azM binds proteinases in a ratio of up to 2 mol proteinases/mol a z M , while PZP (tetramers) maximally binds proteinases in a molar ratio of 1 :1 [l, 41.

1072 Even though r 2 M and PZP seem to use the same mechan- Preparation of derivatives ism for proteinase inhibition, some reports have recently indiAll derivatives of a,M and PZP were prepared at room cated that the mechanism may be basically different [15- 171. temperature in 0.1 M Na2HP04, pH 8.0. a2M and PZP wcrc The thiol ester of a,M and PZP can be modified by treatment treated with a twofold excess of r-chymotrypsin (CT) for with the nucleophilic agent methylamine (MeNH,), which 3 min at room temperature before blocking the active site by binds covalently to the y-glutamyl residue. MeNH,-treated making the solution 20 mM in phenylmethylsulfonylfluoride PZP exposes the receptor-recognition site and competes with for at least 10 min. a2M and PZP were treated with 0.4 M MeNH,-treated azM for binding to the a2M receptor [13], MeNH2 for 1 h. 1 mM dnpSCN was used as the total concenbut immunochemical studies using monoclonal antibodies and tration at cyanylation of derivatives for at least 1 h. It is of partitioning in aqueous two-phase systems have demonstrated importance to note that the cyanql\ited derivatives are undifferences in conformational states between PZP treated with stable and should be tested within a I N hours of preparation. MeNH, (PZP-MeNH,) and PZP treated with chymotrypsin Non-dena turing PAGE and in vivo clearance studies demon(PZP-chymotrypsin) and a closer resemblance between PZP- strated the conversion of a,M treated with methylamine and MeNH, and native PZP conformers [15, 161. Studies of the dinitrophenylthiocyanate [a,M-(McNH, + D)] within hours conformational changes in aZMhave demonstrated very simi- from the slow to the ‘fast’ form and from a non-clearing to a lar forms of the inhibitor by treatment with MeNH, and clearing protein, respectively (not shown). proteinase. Basically, only studies employing scanning calorGel electrophoresis was performed by using mini-PROimetry have been able to detect differences between the two TEAN 2 electrophoresis cells from Bio-Rad. The gels were derivatives of a,M [17]. made according to Laemmli [26] as 5% non-denaturing polyMore details on the conformational changes have been acrylamide or 7.5% SDSipolyacrylamide gels. Reduction of demonstrated by modifications of x2M by cis-dichlorodi- samplcs was performed by addition of dithioerythritol to a ammine platinum (11). dithiobis(succinimidy1propionate) and final concentration of 50 mM for at least 1 h at room temperadinitrophenylthiocyanate(dnpSCN) [I 7 -231. It has been reture. ported that cis-dichlorodiammineplatinum (11) can block the closing of the trap by treatment with MeNH, or proteinase, but the receptor-recognition site can still be exposed. By re- Monoclonal antibodies moval of platinum from azM, the rest of the conformational mAb a1 :1 was generated against human x,M-MeNH2 changes occur [24]. Based on studies of the platinum-modified and purified by immuno-adsorbent techniques as described a2M treated with trypsin, it was concluded that the receptor- [15]. mAb 7HllD6 was a kind gift from D. Strickland, Red recognition-site exposure is dependent upon thiol ester-bond Cross, Bethesda, Maryland, USA, and generated against a z M cleavage 1221. Treatment of a,M with the homobifunctional [27]. mAb rl : 1 binds to a receptor-binding 18-kDa C-terminal cross-linker dithiobis(succinimidy1propionate) demonstrated fragment of a,M, produced by papain digestion according to that proteinase treatment results in a bait-region cleavage- Sottrup-Jensen [28], which is exposed by the conformational dependent conformational change, which causes activation changes induced in azM by treatment with MeNH, or proteinand cleavage of the thiol ester bonds. However, exposure of the ase (unpublished results). mAb 7H11D6 has been shown to receptor-recognition site still occurs upon MeNH, treatment bind to the exposed receptor-recognition site of x2M [29]. [23]. Studies on r,M have shown that the trap is not closed by treatment with MeNH, and dnpSCN, even though the receptor-recognition site is exposed [I 81. DnpSCN is a reagent Enzyme-linked immunosorbent assays capable of cyanylating the liberated thiol group of the thiol ELISA were performed by coating microtiter plates (96 ester [I8 -201. wells) with 10 pg/ml rabbit anti-(a,M/PZP) immunoglobulins We have studied and compared the conformational (DakoPatts) in 0.1 M NaHC03 overnight at 4°C. Ten-times changes of r z M and PZP by use of chymotrypsin, MeNH2, serially diluted samples of antigen (25 pg/ml) were added to dnpSCN, and two mAb directed against the conformational eich well. As dilution buffcr 20 mM sodium phosphate, pH state in which the receptor-recognition site is exposed. By use 7.4 and 150 mM NaCl (NaCl/Pi), with the addition of 0.05% of thcse, it was possible to study the relation between the Tween 20, was used. Incubation with antigen was performed different functional sites in E,M and PZP, i.e. the bait region, for 2 h at room temperature, then the plates were washed the thiol ester and the receptor-recognition site. We conclude before addition of mAb (5 pg/ml) for 2 h. After washing, that the crucial structure causing conformational changes in rabbit anti-mouse immunoglobulin, coupled to horse radish a2M is the thiol ester, and in PZP it is the cleavage of the bait peroxidase (diluted 1 : 5000 in NaCI/Pi, pH 7.4), was addcd region. and incubated for 2 h at room temperature. As substrates ophenylenediamine and H 2 0 2in sodium citrate/sodium phosphate, pH 5.0, were used. The Ah5,,was recorded by a Titertek Multiscan spectrophotometer. MATERIALS AND METHODS Zn vivo plasma-clearance studies Human a2M was purified as described by Imber and Pizzo Plasma-clearance studies were performed in mice by injec[25]. P7,P, a kind gift from Lars Sottrup-Jensen, Aarhus University, Denmark, was isolated from pooled late pregnancy tion of 150 - 200 pl ‘251-labelleda2M-derivative or PZP-deserum [l].DnpSCN was kindly provided by I. Bjork, Uppsala, rivative into the lateral tail vein. About 30 s after injection, a Sweden. Chymotrypsin, MeNH,, phenylmethylsulfonyl- blood sample of 25 p1 was taken from the retroorbital venous fluoride and other reagents were purchased from Sigma. plexus of each mouse and radioactivity determined in a ;’ Scphadex G-50 medium and PD-10 Sephadex 2SM were from counter. This value was used as the 100% value and subPharmacia, Sweden. Na’251 was froni Amersham Inter- sequent samples were collected and calculated as percentage of this value. All clearance studies were performed in triplicate. national, England.

1073

Fig. 1. The effect of dnpSCN on a,M-derivatives. Non-denaturing Laemmli PAGE (5%) of a 2 M derivatives. The derivatives were produced as described in Materials and Methods. Lane 1, native r,M; lane 2, a,M-D; lane 3, a2M-MeNH2; lane 4, a,M-(MeNH,+D); lanc 5, s12M-McNH2 + D.

Fig. 2. SDS/PAGE (7.5%) of a2M. clzM derivatives under non-rcducing (lanes 1 - 5) and reducing conditions (lanes 6 - 10). Lane 1, native a z M ; lane 2, r2M-MeNH2 treated with CT, then dnpSCN; lane 3, a2M-CT; lanc 4, x2M-McNH, treated with CT; lane 5, mZM(MeNH, + D) treated with CT. Lanes 6- 10 present the same derivatives as in lanes 1 - 5 in the same order, but reduced with 50 mM dithioerythritol. The molecular masses (kDa) are shown on the right.

Other methods

Iodination with lZsIwas performed using iodobeads from Pierce. Removal of excess MeNH,, DnpSCN, Na'251, phenylmethylsulfonyl fluoride or CT was performed by gel filtration or by dialysis.

RESULTS Native human azM migrates during electrophoresis in non-denaturing gels in a state commonly referred to as the slow form (Fig. 1). The conformational changes in xzM, caused by reaction with MeNH, or proteinase, can be detected as an increase in the electrophoretic mobility of the inhibitor (Fig. 1 ) . This form is referred to as the fast form of azM [27, 281. Samples of a2M treatcd with MeNH, and dnpSCN migrate as the slow form (Fig. 1). Furthermore, a,M can be :treated with MeNH, to give the fast form, then partially be converted back to the slow form by cyanylation with dnpSCN ([18-201, Fig. 1). Different derivatives of azM and PZP were generated for the studies by PAGE. The different derivatives were native azM, a,M treated with dnpSCN (a2M D), MeNH2-treated a2M (azM-MeNH2),a,M treated with MeNH, and dnpSCN (a2M-(MeNH, D), MeNH,-treated azM, then dnpSCN treated (a,M-MeNH, + D), MeNH, and dnpSCN-treated a,M followed by CT treatment [x2M-(MeNHz D) + CT], CT-treated azM (a,M-CT), dnpSCN-treated and CT-treated a 2 M [azM-(CT D)], CT-treated a2M-MeNH, (xzMMeNH2 CT), CT-treated a2M-MeNH2 treated with dnpSCN [(x2M-MeNH2 + CT) dnpSCN], native PZP, MeNH2-treated PZP (PZP-MeNH,), MeNH,-treated and dnpSCN-treated PZP [(PZP-(MeNH, D)], MeNH2-treated and dnpSCN-treated PZP followed by CT treatment [PZP(MeNHz D) CTJ, CT-treated PZP (PZP-CT), dnpSCNtreated and CT-treated PZP [PZP-(CT + D)]. As seen in Fig. 2, the a2M-CT derivative forms mainly tetrameric complexes in SDS/PAGE, in which a2M and a2MMeNH, are dimeric. This indicates that the two dimers in azM-C?' are held together by the covalent binding of CT. Reduction of these complexes yielded the amino-terminal fragment and different relatively weak bands of x2M carboxyterminal fragment (with or without covalently bound proteinase) of the bait-region cleavage products. azM-(MeNH2 D) showed the same dimer and subunit bands in SDS/

+

+

+

+

+

+

+

+ +

+

polyacrylamide gels, respectively, as native a2M, whether or not the samples were reduced. CT treatment of x2M-MeNH,, followed by treatment with dnpSCN, gave a slow form in nondenaturing gels. No bait-region cleavage fragments of this derivative were demonstrated in the reduced SDS/polyacrylamide gels, indicating that CT cannot cleave the bait region in the closed form of a2M-MeNHz (Fig. 2). Treatment of x2M-(MeNH,+D) with CT gave the fast form and baitregion cleavage fragments were detected, which proves the bait region to be available for CT in this derivative. PZP is a dimer in the native state. Tetramer complexes are seen only after CT treatment of PZP (Fig. 3), indicating that CT covalently binds the two dimers together. When PZPMeNH, is treated with CT, the two fragments by the baitregion cleavage are recognized in reduced SDS/polyacrylamide gels without CT binding (Fig. 3). This confirms that MeNH, is indeed binding to the thiol ester and blocks CT binding. In contrast to a2M-MeNH,, the bait region is still susceptible for cleavage by CT in PZP-MeNH,. Furthermore, PZP-MeNH, treated with CT shows no tetramer formation, which indicates that it is the cross-linking of CT to the dirners which generates the tetrameric form. PZP-CT yields the amino-terminal fragment and different weak bands of the carboxy-terminal fragment (with or without CT bound) produced by bait-region cleavage. No difference in mobility between PZP, PZP-MeNH, and PZP-(MeNH,+D) was observed in SDS/PAGE. The trap mechanism exerted by PZP and x2M was further analyzed (Fig. 4). By incubation of a2M-(MeNH2+ D) and PZP-MeNH, with CT ('"I labelled), it was demonstrated by non-denaturing PAGE and by autoradiography, that PZPMeNH, was not capable of trapping proteinases, in contrast to a,M-(MeNH2 +D). Furthermore, PZP-MeNH, and a2M(MeNH, D) were analyzed after incubation with i251labelled CT, inactivated by phenylmethylsulfonyl fluoride, and with active non-labelled CT. In these experiments, trapping or binding of the inactivated '251-labelled CT was detected in a2M-(MeNH2+ D); but not in PZP. These results dernonstratc that the trap is not in function in PZP. The results suggest that only active proteinase can be bound in PZP and only when the thiol esters are intact upon bait-region cleavage, can the proteinase be bound.

+

1074

2

0

Fig.3. SDS/PAGE (7.5%) of PZP and PZP derivatives in (a) nonreduced samples and in (b) reduced samples. (a) Lane 1, PZP (which represents PZP-MeNH2 and PZP-(MeNH, D) as well, because of the similarity); lane 2, PZP-(MeNH,+D) treated with CT; lane 3, PZP-MeNH, treated with CT, then dnpSCN; lane 4, PZP-(CT+ D); lane 5, PZP-CT. (b). Lane 1, PZP (representing PZP-MeNH, and PZP-(MeNH, D) as above); lane 2, PZP-MeNH, treated with CT; lane 3, PZP-CT; lane 4, PZP-(CT+D). The molecular masses are shown on the right.

+

4

1:10 Dilutions

8

6

Fig.5. ELISA of mAb 7HllD6 binding to a2M derivatives, The mAb 7H11D6 raised against a2M and binding to the receptor-recognition site in a,M, was tested against the different ct2M derivatives as described in Materials and Methods. Microtiter plates were coated with rabbit anti-a2M immunoglobulins overnight at 4°C. clzM derivatives were added as ten-times serially diluted samples for 2 h at room temperature. Then mAb 7 H l l D 6 was incubated for 2 h at room temperature and horse-radish peroxidase conjugated rabbit antimouse immunoglobulins were added and incubated for 2 h at room temperature. As substrate, o-phenylendiamine and HzOz in sodium citrate/sodium phosphate, pH 5.0, were used. Thc colour intensity wasmeasured at A450. (0) a z M ; ( 0 )cr2M+D; ( 0 )r,M+MeNH,; (0) a2M+(MeNH2+D); (.)(a,M M e N H 2 ) ~ D ;( 3 ) a 2 M + (MeNH2+L))+CT; (A)a2M-MeNH2+CT+D; ( A ) x,M+ (CT +D); (B) a 2 M + C T ; (+) r,M+MeNH,fCT.

+

+

0.00

Fig. 4. A study of the trapping of CT by a Z Mand PZP-derivatives. The figure shows an autoradiography of a 5% non-denaturing PAGE according to Laemmli. The ct2M- and PZP derivatives were incubated with a 2-M excess of '251-labelled CT at room temperature for 3 min before addition of phenylmethylsulfonylfluoride for 10 min. Then. the complexes were incubated for 1 h with the sample buffer before PAGE. Lanes 1 and 2, ct2M-(MeNH2 D) and PZP-MeNH2, respectively, incubated with '251-labelled CT. Lanes 3 and 4, a2M(MeNH, + D) and PZP-MeNH,, respectively, incubated at the same time with unlabelled active CT and 'ZSI-labelled CT blocked with phenylmethylsulfonyl fluoride.

+

Two mAb (7HllD6 and a1 :1) were tested for binding to the different derivatives of a,M and PZP. ELISA of the a2M derivatives using mAb 7Hll D6 or a1 : 1 both showed binding with high affinity to all the fast forms of azM and the slow form of cr2M-(MeNH2 D), while low binding affinity was demonstrated for native a,M. Only the ELISA using mAb 7H11D6 is shown here (Fig. 5). The ELISA experiments demonstrate that the mAb specific for the conformational changes

+

0

2

4

6

8

1 :10 Dilutions

Fig. 6. ELISA of mAb 7HllD6 binding to PZP derivatives. Incubations were performed as described in Fig. 5 , but using rabbit anti-PZP immunoglobulins instead of rabbit anti-a2M immunoglobulins. (m) PZP; ( 8 )PZP-MeNH,; ( 0 )PZP-(MeNH,+D); ( 3 )PZP-(MeNH, +D)+ CT; (B) (PZP + MeNH,) + CT+ D; (U) PZP-(CT + D); (A) PZP+ CT; ( A ) (PZP + MeNH2)+ CT.

related to the receptor-recognition site bind with as high affinity to the open form of a2M-(MeNH2+D) as to the closed forms of r2M-MeNH2 and a2M-CT. This indicates that the receptor-recognition site is exposed in a,M-(MeNH, + D) even though the trap is open and accessible for proteinase. ELISA of the PZP derivatives using mAb 7HllD6 or crl : 1 both demonstrated relativcly high affinity for the PZP derivatives treated by CT and low affinity or no binding for native PZP and PZP-MeNH, (with or without dnpSCN). Only the ELISA using mAb 7 H 1 1 D 6 is shown (Fig. 6). This indicates that the receptor-recognition site is not exposed in

1075

I

L+

2

20'

0

2

6

4

i

a

OJ I

2

0

Time (min)

Fig. 7. In-vivo plasma clearance experiments of &zM-derivatives. In each experiment approximately 1,5 x lo6 cpm of ['2sI]rzM derivative was injected into a mouse. ( U ) x2M; (@) x,M-MeNF12; (El)a z M CT; (0) a2M-(MeNH2+ D).

+

8

10

Time (min)

Fig. 8. In-vivo plasma clearance experiments of PZP-derivatives. In each experiment, approximately 1.5 x lo6 cpm of 1251-PZPwas injected into each mouse. PZP; (@) PZY-MeNHz; ( 0 )PZP-CT; (0) PZP-MeNH2-CT.

(m)

-( PZP treated with MeNH, and that bait-region cleavage is required for the exposure. The biological properties of some of the a2M and PZP derivatives were studied by in- vivo clearance of the complexes in mice. Native azM and a2M-(MeNH, + D) are cleared very slowly, if cleared at all from the circulation, in contrast to a,M-MeNH, and a,M-CT, which are cleared with a half-life of 2 - 4 min (Fig. 7). Since the mAb demonstrate the exposure of the receptor-recognition site in a,M-(MeNH, Dj, but remain in circulation in the in-vivo studies, it indicates that the open form of a2M-(MeNH2+D) is not recognized by the receptor even though the receptor-recognition site is exposed. As seen in Fig. 8, the clearance of PZP and PZP-MeNH, is significantly slower than PZP-CT, which has a half-life within the published range of the half-life for a2M fast forms. The slow disappearance of PZP and PZP-MeNH2 may reflect binding of these two forms to cell surfaces, as suggested previously [30].

6

4

10

PZP

U

RRS

c

ti

s

Fig. 9. A schematic drawing of the human a-macroglobulin dimer including the functionally important sites, i.e. the bait regions, the thiol esters, and the receptor-recognition sites (RRS). The subunits demonstrate differences in the regulation of the conformational changes in azM and PZP, respectively. In azM, the critical key is the cleavage of the thiol ester and in PZP it is the cleavage of the bait region.

as a dimer (Fig. 3). Only when CT is covalent bound to PZP

are the tetramer complexes produced. PZP-MeNH, is not affected by modification by dnpSCN and trapping of proteinases do not occur in PZP as in a,M. Studies of the receptor-recognition sites in a,M-CT and a2M-MeN€12demonstrate that these sites become exposed in a2M by cleavage of the thiol ester. The reopening of the a2MDISCUSSION MeNH2 by cyanylation binds the receptor-recognition-sitecl,M-MeNH2 demonstrates the fast mobility similar to specific mAb (Fig. 5). This derivative is, surprisingly, not a,M-CT in non-denaturing PAGE (Fig. 1). While CT can cleared from the circulation in mice, which indicates that the cleave the bait regions and covalently bind two dimers to structure of the molecule is important for receptor recognition. tetramers, MeNH, only cleave the thiol esters and the deriva- PZP-MeNH, is neither cleared in mice nor recognized by the tive can be separated as dimers by SDSjPAGE (Fig. 2). mAb, but when treated with CT it is recognized by the mAb Cyanylation of a,M-CT by dnpSCN demonstrated no detect- and cleared from the circulation as well as PZP-CT (Figs 6 able changes in PAGE, but a2M-MeNH2 treated with and 8). The significant difference in the in-vivo clearance experdnpSCN returns to a slow form, in which the bait regions become available for cleavage by CT. a2M-MeNH2is resistant iments of the a2M-(MeNHzt- D) derivative versus the cr2Mto bait-region cleavage by CT. Cyanylated r2M-MeNH2, MeNH, and a2M-CT derivatives (Fig. 7) indicates that the treated with CT, generates the fast form with the bait region exposure of the receptor recognition site, identified by cleaved, but CT is not covalently bound. It is demonstrated the mAb 7H11 D6, is not enough for receptor recognition. A that CT in this derivative can be non-covalently trapped possible explanation may be that the closing of the trap generates a certain structure, which may involve two sites in the (Fig. 4). These results suggest that the bait region is only available same molecule, to give the high-affinity binding to the recepfor the proteinase in the open trap structure and that the tor. cleavage of the bait region is irreversible closing the trap Even though a2M and PZP basically work the same way forming a fast form. MeNH2 can generate a fast a,M form as after attack by proteinases, which become bound and inhibitwell, but this structure can be changed to a slow form with an ed by the conformational rearrangements, caused by baitopen trap by cyanylation of the liberated thiol group from the region cleavage, it is now demonstrated that the critical conthiol ester. PZP-MeNH, is a dimer like the native PZP and formational step is different in a2M and PZP (Fig. 9j. In a z M this derivative is still accessible for bait-region cleavage by it is the cleavage of the thiol ester and in PZP it is the cleavage CT, but even after bait-region cleavage, the derivative remains of the bait region, which regulate the gross conformational

1076

STATE 1

STATE 2

STATE 3

Fig. 10. A model for the proteinase inhibiting mechanism of PZP. The black circle represents a proteinase and the elongated structure represents PZP dimers. State 1, native PZP dimers present with proteinases. State 2 , demonstrates the binding of a proteinase to a PZP dimer, and state 3 is the tetrameric PZP-proteinase complex.

changes. In PZP, no trapping of the proteinase or exposure of the receptor-recognition site is detected by modifications of the thiol ester. This implies that the triggers of both conformational changes and proteinase binding in PZP are positioned at the same site. It is important to note that it is the cleavage of the bait region, which is the essential trigger of the proteinase binding in a,M [31]. In aZM,it is possible to separate the conformational changes related to the trapping mechanism and thc cxposure of the receptor-recognition site, respectively, as reported hy others [18 -21, 231. This indicates that only the conformational changes leading to the exposure of the receptor-recognition site and the activation of the thiol ester, essential for the proteinase binding, occur in PZP. These results suggest that the conformational changes in PZP are regulated in the same way as in bovine and rat a 2 M , while the conformational changes in human a z M are regulated as in rat a , M , as judged by the studies of these a-macroglobulins [32- 341. The evolution of at least two different ways of regulating the conformational changes may explain some differences in the biological function of a,M and PZP in humans. The same regulatory difference in a-macroglobulins within a single species as now demonstrated in humans may be revealed in other animals. These and earlier results lead us to suggest a model for the mechanism of proteinase inhibition in PZP (Fig. 10). When a PZP dimcr is attacked by a proteinase (state 1) the ‘bait’ region is cleaved and the proteinase becomes covalently bound to the PZP dimer (state 2). As n o trapping occurs in the molecule, the proteinase is not sterically hindered in cleaving other molecules, enabling the PZP-dimer -proteinase complex to cleave another bait region in another PZP dimer, which becomes covalently bound as well (state 3). This formation of tetrameric PZP demonstrates that a 1 : 1 molar complex of proteinase to PZP tetramer is most likely to appear. In a recent publication [35], it is suggested that an intermediate PZP-(tctramer)-CT complex, with intact thiol esters, is formed and that the activated state of the PZP dimer shows extreme affinity for another PZP dimer. Our results do not indicate this, since proteinase binding is necessary for the formation of the tetrameric complex. According to our studies, PZP-MeNH, treated with CT does not result in formation of PZP-tetramer complexes. This suggests that the CT is t h e key molecule in the tetrameric complex formation, but a change in affinity of the PZP-dimer - CT complex for binding another PZP dimer cannot be excluded. A change in the catalytic properties ofthe a,M-bound proteinase towards chromogenic substrates has been demonstrated [ 3 6 ] . The catalytic properties can be greatly modified by a conformational change involving the structure of the charge-relay system, which enhances the nucleophilicity of the active serine residue as well as the oxyanion hole involved in the transition state at catalysis [36]. Taking into consideration the similarity in the covalent binding of proteinase to PZP as to xzM, we suggest that the

same change in catalytic properties of the proteinase may occur upon binding to PZP. This may suggest a higher rate of tetramer-complex formation between a PZP-(dimer) - CT and a second PZP dimer than in the binding of a free CT to a PZP dimer. This new model is in linc with suggestions presented by Christensen et al. [35]in that the proteinase forms a complex with a PZP dimer before the PZP-(tetramer)- proteinase complex is formed, but our model suggests that the PZP-(dimer) proteinase complex is formed with the proteinase covalently bound, which is not in line with the model proposed by Christensen et al., in which the proteinase is not covalently bound before the tetramer complex is formed. Skilful technical assistance was provided by Berit Nilsson and Helen Genberg. I‘he investigation was supported by grants from the Swedish Medical Research Council and the Medical Faculty, University of Umei, Swcden.

REFERENCES 1. Sand, 0.. Folkersen, J., Westergaard; J. G. & Sottrup-Jensen, L. (1985) J . B i d . Chem. 260, 15723-15735. 2. Jensen, P. E. H. & Sottrup-Jensen, L. (1986) J . Biol. Chem. 261, 15863 - 15869. 3. Harpel, P. C . (1973) J . EXJJ. Med. 238, 508-521. 4. Christensen, U.; Simonsen, M.; Harrit. N. & Sottrup-Jensen, L. (1989) Biochemistry 28, 9324-9331. 5. Barrett, A. J. & Starkey, P. S. (1973) Biochem. J . 133, 709-724. 6. Sottrup-Jensen, L., Sand, O., Kristensen, L. & Fey, G. H. (1989) J . Bid. Chem. 264, 15781 -15789. 7. Sottrup-Jensen, L., Petersen, T. E. & Magnusson, S. (1981) FEBS Lett. 128, 123 - 126. 8. Sottrup-Jensen, L., Petersen, T. E. & Magnusson, S. (1981) FEBS Lett. 128. 127- 132. 9. Salvesen, G. S.,Sayers, C. A . & Rarrett, A. J. (1981) Biochem. J . 195,453-461. 10. Salvcsen, G. S.& Barrett, A. J. (19x0) Biochrm. J . 187, 695-701. 11. Sottrup-Jensen, L. & Birkedal-Hansen, H. (1989) J . Bid. Chem. 264, 393 -401. 12. Gettins, P. & Sottrup-Jensen, L. (1990) J . Biol. Clzem. 265, 72687272. 13. Van Leuven, F.: Cassiman, J.-J. & Van den Berghe, H. (1986) J . Biol. Chem. 261, 16622-16625. 14. Gliemann, J.: Moestrup, S., Jensen, P. H., Sottrup-Jensen, L., Andersen. H. B., Petersen, C. M. & Sonne, 0. (1986) Biochim. Biophys. Acta 883. 400-406. 15. Carlsson-Bostedt. L., Moestrup, S. K., Gliemann, J., SottrupJensen, L. & Stigbrand, T. (1988) J . Biol. Chcm. 263, 67386741. 16. Birkcnmeier, G., Carlsson-Bostedt, L., Shanbhag, V.; Kriegel. T., Kopperschlager, G., Sottrup-Jensen. L. & Stigbrand, T. (1989) Eur. J . Biochcm. 183, 239-243. 17. Cummings, H., Pizzo, S. V., Strickland, D. K . & Castellino, F. J. (1984) Biophys. J . 45. 721 -724. 18. Van Leuven, F., Maryncn, P., Cassirnan, J.-J. & Van den Bcrghc. H. (1982) Biochem. J. 203, 405-411. 19. Bjork, I. (1985) Biochem. J . 232, 451 -457. 20. Cunningham, L. W., Crews, R. C. & Gellins, P. (1990) Biochcmistry29, 1638-1643. 21. Gonias, S. L. & Pizzo, S. V. (1983) Ann. N Y . Acad. Sci. 421,457471. 22. Roche, P. A,, Jensen, P. E. H. & Pizzo, S. V. (1988) Biochemistry 27,759 - 764. 23. Roche, P. A,, Moncino, M. D. & Pizzo, S. V. (1989) Biochemistry 28,7629 - 7636. 24. Gonias, S. L. & Pizzo, S. V. (1983) Biochemistry 22, 536-546. 25. Imber, M. J.&Pizzo, S. V. (1981) J . Biol. Chem. 256. 8134-8139. 26. Laemmli, U. K. (1970) Nature 227, 680-6685,

1077 27. Strickland, D. K., Steiner. J. P., Migliorini, M. & Battey. F. D. (1988) Biochemistry 27. 1458-1466. 28. Sottrup-Jensen. L., Gliemann, J. Van Leuven, F. (1986) EEBS Lett. 205. 20- 24. 29. Isaacs, I. J., Steiner, J. P., Roche, P. A., Pizza, s. v. ti Strickland, D. K. (1988) J . B i d . Chem. 263, 6709-6714. 30. Chemnitz, J., Hau, Svendsen, P.. Folkersen. J., Westergaard, J. G. & Christensen, B. C . (1982) Bibl. Anat. 22: 87-92. 31. Christensen, u. Sottrup-Jensen, L., (1984) Biochemistry 23, 6619 - 6626. J.3

32. Feldman, S . R., Gonias, S. L., Ney, K. A,, Pratt, C. W. & Pizzo, S . V. (1984) J . Biol. Chem. 259,4458-4462. 33. Strickland, D. K., Steiner, J. P., Feldman, S . R. & Pizzo, S. V. (1 984) Biochemistry 23, 6679 - 6685. 34. Gonias, s. L., Balber, A. E., Hubbard, W. J. & Pizzo. S . V. (1983) Biochern. J . 209, 99 - 105. 35, Christensen, U,, Sottrup-Jensen, L, & Harrit, N, (1991) Biochim, Biophys. Acta Protein Slruct. Mol. Enzymol. 1076, 91 -96. 36, Martini,J.-L. & Pochon, F, (1989) Biochimie 71, 325-332,