Conversion of Plasminogen Activator Inhibitor-1 from Inhibitor to ...

THE JOURNAL OF BIOLCGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 30, Issue of July 29, pp. 19659-19664, 1994 Printed in U.S.A.

Conversion of Plasminogen Activator Inhibitor-1 from Inhibitor to Substrate by Point Mutations in the Reactive-site Loop*

(Received for publication, March 7, 1994, and in revised form, May 10, 1994)

Anne-Marie AudenaertSO, Isabelle JhockaertS, D6sir6 CollenO, and Paul J. DeclerckSln From the $Laboratory for Pharmaceutical Biology and Phytopharmacology and the §Center for Molecular and Vascular Biology, University of Leuven, K. U. Leuven, E. Van Evenstraat 4, B-3000 Leuven, Belgium

Plasminogen activator inhibitor-1 (PAI-l), the main physiological inhibitor of tissue-type plasminogen acti- vator (t-PA), may occur in three interconvertible confor- mations: active, latent, and substrate. To delineate spe- cific domains in the PAI-1 molecule responsible for its conformational flexibility and associated functional di- versity, four mutants of PAI-1 (with the amino acids at positions P,,, Plot P,, and P,, respectively, substituted with proline) were expressed in Escherichia coli, puri- fied, and characterized.

Wild-type PAI-1 (wtPAI-1) had a specific activity of 21 * 10% (mean * S.D., n = 3) of the theoretical maximum value. PAI-1-P,, (Ala + Pro at P,,), PAI-1-P,, (Ser + Pro at P,,), and PAI-1-P, (Thr + Pro at Pa) had specific activities of 0.06 * 0.03% (n = 3), 2.6 * 1.0% (n = 4), and 2.7 * 1.1% (n = 3), respectively ( p e 0.03 uersus wtPAI-1). PAI-1-P, (Val + Pro at P,) had a specific activity of 12 3.3% (n = 3) of the theoretical maximum value ( p = not significant uer- BUS wtPAI-1). SDS-polyacrylamide gel electrophoresis of mixtures of wtPAI-1 or PAI-1-P, with a 2-fold molar ex- cess of t-PA yielded a mixture of a covalent 110-kDa t-PA-PAI-1 complex (15-25%), nonreactive 45-kDa mate- rial (44-67%), and a 41-kDa band (1831%) representing cleaved PAI-1. PAI-1-P,,, PAI-1-P,,, and PAI-1-P, behaved as substrates, yielding predominantly the 41-kDa cleav- age product (8691%) and a small amount (9-15%) of non- reactive material. NH,-terminal amino acid sequencing revealed that cleavage occurred at the P,-P,’ bond (ArP6-MetS4’). Incubation of PAI-l-P,,, PAI-l-Plo, or PAI- l-P, with a 2-fold molar excess of urokinase-type plas- minogen activator, plasmin, or thrombin also primarily generated a 41-kDa cleavage product (6~-8W’). Incuba- tion of wtPAI-1 and PAI-1-P, at 37 “C resulted in a loss of inhibitory activity, whereas the substrate behavior of PAI-1-P,,, PAI-l-P,,, and PAI-1-P, remained unaltered. Treatment of the three substrate-like mutants with gua- nidinium C1 did not induce inhibitory activity. In con- clusion, point mutations at positions P,,, P,,, and P, yield PAI-1 variants with stable substrate properties, which may facilitate more detailed structure/function studies.

Plasminogen activator inhibitor-1 (PAI-11,’ a glycoprotein with an apparent molecular mass of 50 kDa (l), was first iden-

Grant CI1-CT920035 and Fund for Medical Scientific Research Project * This work was supported in part by European Community Project

Grant 3.0066.93. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

II Senior Research Associate of the National Fund for Scientific Re- search (Belgium). To whom correspondence should be addressed. “el.:

The abbreviations used are: PAI-1, plasminogen activator inhibi- tor-1; serpin, serine proteinase inhibitor; t-PA, tissue-type plasminogen activator; u-PA, urokinase-type plasminogen activator; wtPAI-1, wild-

32-16-283431; Fax: 32-16-283414.

tified in cell culture in 1977 (2). PAI-1 is a member of the serine proteinase inhibitor (serpin) superfamily (3-61, a group of more than 40 proteins with a sequence identity of -35% (7). PAI-1 inhibits tissue-type plasminogen activator (t-PA) and uroki- nase-type plasminogen activator (u-PA) with second-order as- sociation rate constants of - lo7 M-’ s-l(8). Of the various plas- minogen activator inhibitors present in plasma, PAI-1 appears to be the principal physiological inhibitor of t-PA in plasma (9). Elevated PAI-1 levels in plasma are associated with an in- creased risk of thromboembolic disease, suggesting a critical role of PAI-1 in the in vivo regulation of fibrinolysis (10). PAI-1 is synthesized as an active molecule that converts spontane- ously to a latent form that can be partially reactivated by protein denaturants such as sodium dodecyl sulfate, guani- dinium chloride, or urea (11). In addition, a third conformation with substrate properties has recently been identified (12-14).

Active serpins exert their inhibitory action through forma- tion of a 1:l stoichiometric reversible complex with their target proteinase, followed by covalent binding between the hydroxyl group of the active-site serine residue of the proteinase and the carboxyl group of the PI residue at the reactive center (“bait region”) of the serpin (15). The Arg346-Met347 bond in PAI-1 has been identified as the PI-PI’ bond (16). Recent crystallographic data (17) have suggested that in latent PAI-1, the bait region (Pl-Pl’) and secondary binding sites are not accessible to the active site of the serine proteinases.

In general, serpins consist of three highly conserved p-sheets designated A, B, and C, whereas the conformation of the reac- tive-site loop comprising residues PI, to Plo’ is highly variable (18). The flexibility of the P,, to PI region of the reactive-site loop of serpins has been suggested to play an important role in their functional behavior. In the intact noninhibitory serpin ovalbumin, the reactive-site loop forms a protruding segment and an a-helical conformation, which is inappropriate for in- hibitory activity (191, whereas active serpins have a more flex- ible reactive-site loop (19,20). Cleavage of the reactive center of the latter results in the insertion of the peptide loop (residues PI, to P,) as a P-strand into the center of p-sheet A, yielding a thermodynamically stable conformation (7,21). In latent intact PAI-1, residues P,, to P, are incorporated as a central p-strand in sheet A, while residues P, to Plo’ form an extended loop on the surface of the molecule (17). In addition, several serpins (e.g. antithrombin I11 and a,-antitrypsin) can adopt an inactive thermodynamically stable conformation after exposure to mild denaturing conditions (201, which may also be due to insertion of the reactive-site loop into p-sheet A.

type PAI-1; PAGE, polyacrylamide gel electrophoresis; PAI-1-PI,, PAI-1 obtained by site-specific mutagenesis of Ala3s6, at position P to Pro; PAI-1-Pl0, PAI-1 obtained by site-specific mutagenesis of Se&g7, at po-

of ThP’p’,t position p,, to Pro; PAI-1-p,, PAI-1 obtained by site-specific sition P to Pro; PAI-1-P,, PAI-1 obtained by site-specific mutagenesis

mutagenesis of VaP4’, at position P,, to Pro.

19559

19560 Substrate Properties of PM-1 Mutants

To characterize further the role of the reactive-site loop of PAI-1 in its conformationaVfimctiona1 flexibility, we have con- structed and characterized four mutants of PAI-1 in which the amino acids at positions PI,, PI,, P,, and P,, respectively, were substituted with proline based on the hypothesis that this might decrease the flexibility of the reactive-site loop and hence convert labile active PAT-1 into a molecule with stable sub- strate properties.

EXPERIMENTAL PROCEDURES Materials-Restriction enzymes were purchased from Pharmacia

(Uppsala, Sweden). T4 DNA ligase, the Klenow fragment of Escherichia coli DNA polymerase I, and alkaline phosphatase were obtained from Boehringer Mannheim (Brussels, Belgium). The oligonucleotide-di- rected mutagenesis system (22) (the pMdc plasmids) was kindly pro- vided by Corvas (Gent, Belgium). The expression vector pIGE2O was kindly provided by Innogenetics (Gent, Belgium), together with the bacterial strains E. coli DHlh for cloning and E. coli MC1061 for ex- pression as well as the pAcI plasmid encoding the thermolabile repres- sor. M13K07 helper phage was purchased from Promega (Leiden, The Netherlands). The oligonucleotides for mutagenesis were from Pharmacia.

Most of the chemical reagents, including the protease inhibitors leu- peptin, phenylmethanesulfonyl fluoride, dithiothreitol, pepstatin, benz- amidine, and antipain, were from Sigma. Luria broth growth medium was purchased from Life Technologies, Inc. (Gent, Belgium). The chro- mogenic substrates 5-2444 and S-2403 were obtained from KabiVitrum (Stockholm, Sweden). t-PA was a kind gift from Genentech Inc. (South San Francisco, CA). Q-Sepharose, SP-Sepharose, and heparin-Sepha- rose were purchased from Pharmacia.

General DNA Techniques-Plasmid DNA was isolated using a QIA- GEN purification protocol (provided by Westburg N.V., Leusden, The Netherlands). Transformations of E. coli were performed using the cal- cium chloride procedure (23). DNA sequencing was performed using the dideoxy chain termination reaction method and the automated laser fluorescent A.L.F.TM (Pharmacia). Site-directed mutagenesis was per- formed using the pMdc system (22) in the repair-deficient E. coli strain WK6 mutS. Propagation of the pMdc plasmids or derivatives and preparation of single-stranded DNA were carried out in E. coli WK6. For expression, wtPAI-1 and PAI-1 mutants were cloned into pIGE20, a P, expression vector, and expressed in E. coli MC1061 after cotransfor- mation with pAcI. DNA manipulation techniques were carried out ac- cording to standard procedures (23) and following the instructions of the manufacturers.

Construction of PAI-1 Variants-A 1286-base pair EcoRI-Hind111 fragment containing the entire coding region of PAI-1 (24) was isolated from the plasmid pKK223#3-PAI-l; the EcoRI end was blunted; and the fragment was cloned into the BamHI (blunted)-Hind111 sites of the pMc plasmid to yield pMc-PAI-1. For in vitro site-directed mutagenesis, single-stranded DNA of this construct was prepared by transformation of pMc-PAI-1 in E. coli WK6 and infection of an overnight culture with helper phage M13K07. After 6 h of infection, cells were centrifuged, and single-stranded DNA was isolated from the supernatant by polyethyl- ene glycol precipitation and phenollchloroform extraction. Subse- quently, single-stranded pMc-PAI-1 was hybridized with the EcoRI- HindIII-digested pMa and with one of the following synthetic oligo- nucleotides to obtain the desired mutation.

Positive clones were grown in small volumes, and the mutations were checked by restriction enzyme analysis. Then, larger-scale DNA prepa- rations were made for further confirmation of the mutations by restric- tion analysis and by nucleotide sequencing using A.L.F.TM

Construction of Expression Plasmids-pMc-PAI-1, with the PAI-1 cDNA cloned into the BamHI (blunted)-Hind111 sites of pMc, was mu- tagenized using the oligonucleotide 5"ATGGTGCACCATGGATG- GATCCTGTTTCCT, creating BamHI-NcoI sites (yielding pMa-PAI-1). After digestion with NcoI (partial) and XbaI, the wtPAI-1 cDNA NcoI- XbaI fragment was cut out of pMa-PAI-1 and inserted in the pIGE20 polycloning site (pIGE20-PAI-1). After mutagenesis of pMc-PAI-1 into pMa-PAI-1-P,,, pMa-PAI-1-PI,, pMa-PAI-1-P,, and pMa-PAI-l-P,, Sad- XbaI fragments were recovered from the mutant pMa-PAI-1 constructs and substituted for the wild-type fragments in pIGE2O-PAI-1.

Expression and Purification of wtPM-1 and Variants-After cloning the mutant fragments in the pIGE20 expression vector, the pIGE2O- PAI-1 constructs were cotransformed with pAcI (encoding the heat- labile repressor) in E. coli MC1061. Transformed E. coli MC1061 cul- tures were grown overnight at 28 "C, diluted 1:100, and grown until A580nm= 0.2. Then, PAI-1 synthesis was induced at 42 "C for 4 h. Cells were centrifuged for 10 min at 4000 xg, and cell extracts were prepared as follows. Cell pellets were resuspended in 30 ml of buffer containing 25% sucrose, 15 nm Na,HPO,, and 0.08 M EDTA, pH 6.5. After addition of leupeptin (0.5 pg/ml), pepstatin A (0.7 pg/ml), benzamidine HC1 (1 mM), antipain (1 pg/ml), phenylmethanesulfonyl fluoride (0.05 mM), and dithiothreitol(1 mM), 16 ml of a solution containing 0.2% Nonidet P-40, 0.062 M EDTA, and 50 m~ Na,HPO,, pH 6.0, was added. After disrup- tion of the cells by sonication for 4 min using a macrotip probe (Branson sonifier Model 250) and subsequent centrifugation at 20,000 rpm for 20 min at 4 "C, the PAI-1-containing supernatant was collected and stored at -20 "C until further processing.

Purifications were performed a t 4 "C. The supernatant from the lysed cells was dialyzed overnight against buffer containing 0.15 M phosphate, 2 mM glutathione, and 0.01% Tween 80, pH 6.9, and loaded onto a Q-Sepharose fast flow column (2.6 x 40 cm; Pharmacia) previously equilibrated with the same buffer. The effluent was collected. The PAI- 1-containing fractions were pooled; dialyzed extensively against buffer containing 0.15 M phosphate, 2 mM glutathione, and 0.01% Tween 80, pH 6.0 (buffer A); and loaded onto an SP-Sepharose fast flow column (1.6 x 20 cm; Pharmacia) equilibrated with buffer A. The column was washed with 2 column volumes of buffer A containing 0.15 M sodium chloride, and the bound proteins were eluted using a sodium chloride gradient (0.15-1.2 M NaCl) in buffer A. PAI-1-containing fractions (elut- ing at a NaCl concentration of -0.7 M) were pooled, dialyzed against buffer A, and loaded onto a heparin-Sepharose fast flow column (1.0 x 20 cm; Pharmacia) equilibrated with buffer A. The column was washed with 2 volumes of buffer containing 0.15 M phosphate, 2 mM glutathione, and 0.01% Tween 80, pH 6.5 (buffer B); and bound proteins were eluted using a sodium chloride gradient (0-1.2 M NaC1) in buffer B. PAI-1- containing fractions (eluting at a sodium chloride concentration of - 1 M) were pooled and stored at -20 "C in aliquots until use.

Functional and Immunological Determination of PM-I-PAI-1 activ- ity was determined using the method described by Verheijen et al. (25) by adding a fixed amount of t-PA to the PAI-1 samples or, alternatively, using an immunofunctional method as described previously (26). t-PA was calibrated versus the international reference preparation for t-PA(N1BSC 86/670). All PAI-1 activity data are expressed as a percent- age of the theoretical maximum activity, i.e. -745,000 unitslmg, calcu-

PAI-l-P],: 5'- GT GGA TGA GGA CGG TAC CGT GCC AC-3' ( ~ 1 a ~ ~ ~ + Pro) PAI-l-Plo: 5'- AC AGC TGT GGA CGG GGA GGC CAG-3' (serS3' + Pro)

PAI-1-P8: 5'-GAC TAT GAC AGC CGG CGA TGA GGA GGC CA-3' (Thr'" + Pro)

PAI-1-P6: 5'-GCG GGC TGA GAC GAT CGG AGC TGT GGA TGA-3' (Val3" + Pro)

In addition, specific restriction sites were simultaneously created to allow confirmation of the desired mutation. This included KpnI, NaeI, and PvuI sites for PAI-1-PI,, PAI-1-P,, and PAI-1-P,, respectively.

Extension reactions were carried out with the Klenow fragment of DNA polymerase as described (23). After transformation of E. coli WK6 mutS and selection on ampicillin, colonies were grown on nitrocellulose membranes and denatured in situ, and DNA was hybridized overnight at room temperature using the respective radiolabeled "mutant" oligo- nucleotides (1.5 x 10' cpm [Y-~~PIATP used for labeling of 20 ng of oligonucleotide). Filters were washed at 42 "C using solutions contain- ing 0.1% SDS and 2 x SSC, 1 x SSC, 0.2 x SSC, or 0.1 x SSC (23).

lated on the basis of a specific activity for t-PA of 500,000 unitslmg and molecular masses of 67 and 45 kDa for t-PA and PAI-1, respectively. PAI-1 antigen was determined with an enzyme-linked immunosorbent assay as described previously (27).

Characterization of PAI-1 VariantsSamples of both wtPAI-1 and PAI-1 mutants were incubated either with catalytic amounts (5%) or with a 2-fold molar excess of t-PA, u-PA, plasmin, or thrombin. PAI-1 samples were diluted to a concentration of 0.2 mg/ml and incubated for 30 min or 2 h at 37 "C with the serine proteinases. The reaction was terminated by adding SDS (final concentration of 1%) and heating for 30 s at 100 "C. Reaction products were analyzed by SDS-polyacrylamide

Substrate Properties of PAI-1 Mutants 19561

- 97

-. 67

- 45

- 30

- 20

- 14

1 2 3 4 5 Mr(kD4 FIG. 1. SDS-PAGE of purified wtPAT-1 and PAT-1 mutants. Lane

I , PAI-1-PI,; lane 2, PAI-l-P,,; lane 3, PAI-1-P,; lane 4, PAI-1-P,; lane 5, wtPAI-1. Migration positions of molecular weight standards are indi- cated: phosphorylase b, M, 97,000; albumin, M, 67,000; ovalbumin, M, 45,000; carbonic anhydrase, M, 30,000; trypsin inhibitor, M, 20,100; and a-lactalbumin M, 14,400. The arrowhead indicates the migration posi- tion of intact PAI-I.

gel electrophoresis (PAGE) and subsequent densitometric scanning with the gel scan accessory supplied with the Beckman DU60 spectro- photometer. Reactivation of PAI-1 preparations was performed by incu- bating samples with 6 M guanidinium CI for 25 min a t 37 "C, followed by extensive dialysis a t 4 "C against buffer containing 0.1 M sodium ac- etate, pH 5.5.

Other Analytical Methods-The purity of the preparations was as- sessed by SDS-PAGE using 10-15% gradient gels under nonreducing conditions with the Pharmacia Phast system. Proteins were visualized by staining with Coomassie Brilliant Blue. PAI-1 protein in purified preparations was determined spectrophotometrically at 280 nm using an absorbance coefficient (A :Tm) of 10.

NH,-terminal sequence analysis (before and after reaction with t-PA) was kindly performed by Dr. J. Van Damme (Laboratory of Microbiol- ogy, Rega Institute, University of Leuven, Leuven, Belgium) using an Applied Biosystems Model 477A Protein Sequencer, with identification of phenylthiohydantoins by high performance liquid chromatography.

The statistical significance of differences was assessed using Stu- dent's t test; p values > 0.05 were considered to be not significant.

RESULTS

Expression and Purification of PM-1 Mutants-wtPAI-1 and PAI-1 variants, cloned into pIGE20, were expressed in E. coli MC1061 after cotransformation with pAcI. Out of 1 liter of cell culture, 1-3 mg of highly purified PAI-1 was obtained as de- scribed under "Experimental Procedures" (Fig. l).

Functional Behavior of wtPM-1 and PM-1 Mutants- wtPAI-1 had a specific activity of 21 2 10% (mean * S.D., n = 3) of the theoretical maximum value; PAI-1-PI,, PAI-l-P,,, PAI-1- P,, and PAI-1-P, had specific activities of 0.06 2 0.03% (n = 31, 2.6 2 1.0% ( n = 4), 2.7 2 1.1% (n = 3), and 12 2 3.3% (n = 3), respectively. PAI-1-PI,, PAT-1-PI,, and PAI-1-P, were signifi- cantly less active ( p < 0.03) than wtPAI-1.

Treatment of wtPAI-1 with guanidinium C1 resulted in an increase of the mean specific activity to 35% of the theoretical maximum value, whereas the low inhibitory activity of PAI-1- PI,, PAI-1-PI,, and PAI-1-P, did not change. Treatment of PAI- 1-P, with guanidinium C1 resulted in reactivation comparable to that observed for wtPAI-1 (Table I).

Characterization of Reaction Products Generated afrer Zncu- bation of wtPM-1 and PM-1 Mutants with Various Serine Proteinases-Incubation of wtPAI-1 with catalytic amounts of t-PA (5%) resulted in the neutralization of t-PA with the con- comitant formation of small amounts of a t-PA.PAI-1 complex; in addition, small amounts of a 41-kDa degradation product were formed, while the majority (>85%) remained intact (Fig. 2 A ) . In contrast, addition of catalytic amounts of t-PA to PAI- l-Pl,, PAT-1-PI,, and PAI-1-P, resulted in turnover of the t-PA

TARLF: I Specific activities of wtPM-1 and PAI-I mutants toward t-PA before

and after treatment with 6 M guanidium C1 Values are the means f S.D. (n = 3).

Specific activity"

Starting material After reactivation

Q

wtPAI-1 10 -c 1.6 35 2 10 PAI-l-Pl, <0.15 <0.15 PAI-l-P,, PAI-1-P,

3.2 2 1.0 2.6 2 0.9 2.2 2 0.2 2.8 2 0.5

PAI-1-P, 13 2 3.4 36 2 10

a Expressed as a percentage of the theoretical maximum value.

added and a virtually complete (82-91%) degradation to a 41- kDa derivative, indicative of substrate behavior of these mu- tants. Incubation of PAI-1-P, with catalytic amounts yielded a pattern similar to that obtained with wtPAI-1 and t-PA, i.e. formation of small amounts of a t-PA.PAI-1 complex and a 41-kDa degradation product, while the majority (85%) re- mained intact (Fig. 2A ).

As expected, in the presence of a 2-fold molar excess of t-PA, wtPAI-1 revealed the formation of t-PA.PAI-1 complexes (15 2 2%; mean 2 S.D., n = 41, small amounts of cleaved derivative (18 2 4%), and residual nonreactive PAI-1 (67 2 4%). Under these conditions, PAI-1-PI,, PAI-1-PI,, and PAI-1-P, were vir- tually completely degraded (86 2 3, 91 2 1, and 85 '' 2%, re- spectively; n = 2) (Fig. 2B), whereas PAI-1-P, yielded a pattern comparable to that obtained with wtPAI-1, i.e. 25 2 3% t-PA.PAI-1 complex (n = 2), 44 2 2% nonreactive PAI-1, and 31 2 1% cleaved derivative (Fig. 2B).

Incubation of wtPAI-1 with a 2-fold molar excess of u-PA re- vealed the formation of u-PA.PAI-1 complexes (10 2 4%; mean 2 S.D., n = 3), small amounts ofthe 41-kDa degradation product (17 2 l%), and residual nonreactive material (73 2 4%) (Fig. 3A). PAT-1-PI, and PAI-1-P, yielded predominantly a 41-kDa degra- dation product (86 2 4 and 84 2 5%, respectively; n = 3). Sur- prisingly, in contrast to the results obtained in the presence of t-PA, incubation of PAI-1-PI, with a 2-fold molar excess of u-PA resulted in the formation of u-PA.PAI-1 complexes (19 2 6%; n = 31, large amounts of the 41-kDa cleaved derivative (73 2 4%), and small amounts of residual nonreactive material (8 * 2%). Under these conditions, PAI-1-P, yielded no detectable u-PA.PAI-1 complex, but significant amounts of cleaved deriva- tive (49 * 3%; n = 3) and nonreactive material (49 2 3%) (Fig. 3A ).

Incubation of wtPAI-1 with a 2-fold molar excess of plasmin (Fig. 3B) yielded a pattern comparable to that obtained after incubation with t-PA. In addition, two low molecular mass cleavage products were generated, representing up to 40% of total PAI-1 protein. In the presence of plasmin, PAI-1-P,,, PAI- 1-PI,, and PAI-1-P, primarily yielded the 41-kDa cleavage prod- uct. PAI-1-P, revealed a pattern comparable to that of wtPAI-1 (Fig. 3B). Incubation of wtPAI-1 and PAI-1 variants with a 2-fold molar excess of thrombin revealed a pattern comparable to that observed in the presence of t-PA (Fig. 3C).

Stability of wtPM-1 and PM-1 Mutants-wtPAI-1 and PAI-1 mutants were incubated at 37 "C, and PAI-1 activity was de- termined after various time intervals. Incubation of wtPAI-l or PAI-1-P, for 24 h a t 37 "C resulted in a significant loss of the inhibitory activity (Table 111, while the low specific activity of PAI-1-PI, and PAI-1-P, remained relatively constant. Analysis of the reaction products after incubation of the treated (24 h, 37 "C) samples with t-PA revealed that wtPAI-1 and PAI-1-P, were primarily converted to a latent inactive form, while the substrate behavior of PAI-1-PI,, PAI-l-P,,, and PAI-1-P, re- mained unaltered (Fig. 4).

Substrate Properties of PM-1 Mutants 19562

A L

1

B I

I

1

A

- 97 t

- 67

- 45

- 30 + I

> ."I "

- 97 +

- 67

- 45

- 30 I

- 20

- 14

i

FIG. 2. SDS-PAGE of wtPAI-1 and PAI-1 mutants after addition of catalytic amounts (A) and 2-fold molar excess ( B ) of t-PA. Lune 1, PAI-1-P,,; lane 2, PAI-1-P,,; lane 3, PAI-1-P,; lane 4, PAI-1-P,; lane 5, wtPAI-1. Migration positions of molecular weight standards are indi- cated (see legend to Fig. 1). The closed arrowhead indicates the migra- tion position of intact PAI-1. The open arrowhead indicates the migra- tion position of the t-PA.PAI-1 complex.

Determination of Cleavage Site in PM-1-P,, PAI-1-P,, and PM-1-Ps-NH,-terminal sequence analysis of 200 pmol of un- treated wtPAI-1, PAI-1-P,,, PAI-1-P,,, PAI-1-P,, and PAI-1-P, revealed the known NH,-terminal sequence of PAI-1 (i.e. Val- His-His-Pro-Pro). NH,-terminal sequence analysis of 200 pmol of PAI-1-P,,, PAI-1-PI,, and PAI-1-P, aRer incubation with cata- lytic amounts of t-PA (5% molar ratio) revealed the sequence of intact PAI-1, i.e. Val-His-His-Pro-Pro, and, in addition, the gen- eration of equimolar amounts of a new sequence, Met-Ala-Pro- Glu-Glu-Ile, corresponding to the sequence of wtPAI-1 starting at Met347. These results indicate that cleavage of the mutants occurs at the Arpfi-Met3'7 bond in the P,-PI' position.

DISCUSSION PAI-1, a member of the family of serine proteinase inhibitors,

is synthesized as an active form that spontaneously inactivates to a latent form. Recently, the occurrence of yet another con- formation that reacts as a substrate toward t-PA and u-PA has been reported (12-14). The latent as well as the substrate con- formation can be partially reactivated with guanidinium C1, which suggests that the three forms are interconvertible through conformational transitions (12). Several studies (19- 21) have suggested that the flexibility of the reactive-site loop of serpins may contribute significantly to their functional be- havior. In addition, several natural mutants of serpins have been described in which substitutions of "small" to bulky amino acids at P,, or P,, (28-30) or insertion of an alanine (31) in the reactive-site region (P,, to P,) produced substrate-like mol-

3 4 5 Mr(kDa)

C

- 97

- 97

- 67

-45

- 30

- 20

- 14

3 4 5 Mr (kDa)

J

97

67

FIG. 3. SDS-PAGE of wtPAI-1 and PAI-1 mutants after addition of 2-fold molar excesses of urokinase (A), plasmin ( B ) , and thrombin ( C ) . Lane 1, PAI-1-P,,; lane 2, PAI-1-P,,; lane 3, PAI-1-P,; lune 4, PAI-1-P,; lane 5, wtPAI-I. Migration positions of molecular weight standards are indicated (see legend to Fig. 1). The closed arrow- head indicates the migration position of intact PAI-1. The open arrow- head indicates the migration position of the u-PA.PAI-1 complex. The arrow indicates the migration position of the proteinase.

ecules. Because of the known conformational instability of ac- tive PAI-1, the importance of the conformation of the reactive- site loop, and the ability of proline to affect the conformation as well as the mobility of a protein strand, we have characterized four mutants of PAI-1 in which the amino acids a t positions P,,, P,,, P, and P,, respectively, were substituted with proline.

Functional analysis revealed that wtPAI-1 and PAI-1-P, had comparable specific activities and spontaneously converted to an inactive form that could be partially reactivated with gua- nidinium C1. The nature of the nonreactivable forms is cur-

Substrate Properties of PAI-1 Mutants 19563

TABLE I1 Specific activities of wtPAI-1 and PAI-1 mutants after incubation at 37 "C

Values are the means e S.D. ( n = 3) of activities expressed as a percentage of the theoretical maximum value. PAI-1-P,, was not included because of its low specific activity.

Time after incubation

Oh 4 h 8 h 24 h

WtPAI-1 PAI-l-P1, PAI-1-P, PAI-1 P,

20 e 9 1.9 e 0.4 1.5 e 0.4 5.7 ? 0.5

10 f 5.8 1.7 e 0.4 1.7 e 0.3 4.4 e 0.9 -

C - 97

- 67

- 45

- 30

b

- 20

- 14

1 2 3 4 5 Mr(kDa) FIG. 4. SDS-PAGE of wtPAI-1 and PAI-1 mutants after incuba-

tion for 24 h at 37 "C and subsequent addition of t-PA. Lane 1 , PAI-1-PI,; lane 2, PAI-1-P,,; lane 3, PAI-l-P,; lane 4, PAI-l-P,; lune 5, wtPAI-1. Migration positions of molecular weight standards are indi- cated (see legend to Fig. 1). The closed arrowhead indicates the migra- tion position of intact PAI-1. The open arrowhead indicates the migra- tion position of the t-PA.PAI-1 complex.

rently unknown, but may be related to an improper folding process either during biosynthesis or during the reactivation procedure (32) and/or to the occurrence of the substrate form (12). Incubation of PAI-1-P, with t-PA yielded similar reaction products as with wtPAI-1, corresponding to latent, active, and substrate forms (12). Introduction of proline at position P, or PI,, however, yielded PAI-1 variants with a significantly lower specific activity, whereas substitution with proline at position PI, (PAI-l-P12) produced a virtually inactive variant. The latter PAI-1 mutants behaved as stable substrates toward various serine proteinases and could not be converted to an inhibitory form with guanidinium C1. NH,-terminal sequence analysis revealed that these substrate-like mutants, like the wild-type form of substrate PAI-1 (121, are cleaved at the P,-P,' peptide bond of PAI-1 (16). Thus, introduction of a proline at position P,, PI,, or PI, yields PAI-1 variants with stable substrate prop- erties, most likely as a result of reduced flexibility of the reac- tive-site loop, as was previously suggested for ovalbumin (19). It is also worth noting that incubation of PAI-1-PI, with u-PA, in contrast to incubation with t-PA (Fig. 3A versus Fig. 2B, lane 21, results in the formation of complexes; on the other hand, PAI-1-P, does not form stable complexes with u-PA in contrast to its behavior with t-PA (Fig. 3A versus Fig. 2 B , lane 4) . These observations are in agreement with the finding that the specific inhibitory activity of PAI-1-P,, is 3-4-fold higher toward u-PA than toward t-PA, while that of PAI-1-P, is 3040-fold lower toward u-PA than toward t-PA (data not shown).

A number of PAT-1 mutants have recently been produced to study the secondary binding site interaction between PAI-1 and t-PA (33) and to evaluate the influence of mutations near the reactive-center bond (P3-P3') (34-37). I t appeared that these reactive-center residues are of critical importance for the speci- ficity and activity of PAI-1. Our current data demonstrate that substitution at P, (and, to a lesser extent, at PI,) markedly

5.5 = 2.7 1.6 f 0.6 1.6 2 0.4 2.9 f 0.2

1.5 1.3 1.2 e 0.01 1.4 e 0.5 1.0 e 0.1

affects the specificity of this serpin. In contrast to previous findings that suggested that the PI, to P,' region of PAI-1 did not influence its stability (38), our present data indicate that the PI, to P, region in PAI-1 does play a major role in the functional properties and stability of the molecule.

In conclusion, this study demonstrates that residues P,, to P, of PAT-1 play a crucial role in its functional behavior and con- formational stability. To our knowledge, this is the first report of stable functional (Le. as substrate) mutants of PAI-1. There- fore, these stable mutants and their reaction products may constitute useful tools for the elucidation of the molecular de- terminants involved in the diverse conformational functional properties of PAI-1.

Acknowledgments-We are grateful to Dr. J. Van Damme and P. Proost (Laboratory of Microbiology, Rega Institute, University of Leuven) for the amino-terminal amino acid sequence analysis and to R. Vleugels for technical assistance.

REFERENCES 1. Van Mourik, J. A., Lawrence, D. A,, and Loskutoff, D. J. (1984) J. Biol. Chem.

2. Loskutoff, D. J., and Edgington, T. E. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 259, 14914-14921

3903-3997 3. Pannekoek, H., Veerman, H., Lambers, H., Diergaarde, P., Venveij, C. L., Van

4. Ny, T., Sawdey, M., Lawrence, D., Milan, J. L., and Loskutoff, D. J. (1986) Proc. Zonneveld, A. J.. and Van Mourik, J. A. (1986) EMBO J. 5,2539-2544

5. Ginsberg, D., Zeheb, R., Young,A. Y., Rafferty, U. M.,Andreasen, P.A.. Nielsen, Natl. Acad. Sci. U. S. A. 83, 6776-6780

L., Dana, K., Lebo, R. V., and Gelehrter, T. D. (1986) J. Clin. Inuest. 78, 1673-1680

6. Andreasen, P. A,, Riccio,A., Welinder, K. G., Douglas, R., Sartoria, R., Nielsen, L. S., Oppenheimer, C., Blasi, F., and Dana, K. (1986) FEES Lett. 209, 213-218

.~..

7. Huber, R., and Carrell, R. W. (1989) Biochemistry 28,89514966 8. Sprengers, E. D., and Kluft, C. (1987) Blood 69,381-387 9. Kruithof, E. K. 0.. Tran-Thang, C., Ransijn, A,, and Bachmann, F. (1984) Blood

10. Kruithof, E. K. 0. (1988) Fibrinolysis 2, Suppl. 2, 59-70 11. Hekman, C. M., and Loskutoff, D. J. (1985) J. Biol. Chem. 260,11581-11587 12. Declerck, P. J., De Mol, M., Vaughan, D. E., and Collen, D. (1992) J. Biol.

64,907-913

Chem. 267, 11693-11696 13. Urano, T., Strandberg, L., Johansson, L. B., and Ny, T. (1992) Eur. J . Biochem.

2n9.98~-992 ~~~

14. Munch, M., Heegaard. C. W., and Andreasen, P. A. (1993) Biochim. Biophys.

15. Laskowski, M., and Kato, I. (1980)Annu. Rev. Biochem. 49,593-626 16. Lindahl, T. L., Ohlsson, P. I., and Wiman. B. (1990) Biochem. J . 265,109-113 17. Mottonen, J., Strand,A., Symersky, J., Sweet, R. M., Danley, D. E., Geoghegan,

18. Sprang, S. R. (1992) 7 h d s Biochem. Sci. 17,49-50 19. Stein, E. P., Leslie, A. G. W., Finch, J. T., Turnell, W. G., McLaughlin, P. J., and

20. Carrell, R. W., Evans, D. L., and Stein, P. (1991) Nature 353,576-578 Carrell, R. W. (1990) Nature 347,99-I02

21. Stein, P., and Chothia, C. (1991) J . Mol. Biol. 221,615-621 22. Stanssens, P., Opsomer, C., McKeown, Y., Kramer, W., Zabeau, M., and Fritz,

23. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A M. J. (1989) Nucleic Acids Res. 17,4441-4454

Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring

, - - - ~~~

Acta 1202,29-37

K. F., Gerard, R. D., and Goldsmith, E. J. (1992) Nature 356, 270-273

Harbor, N%

J. Biochem. 175,531-540

. - . .

24. Alessi, M. C., Declerck, P. J., De Mol, M., Nelles, L., and Collen, D. (1988) Eur:

25. Verheijen, J. H., Chang, G. T. G., and Kluft, C. (1984) Thromb. Haemostasis 51,

26. Declerck, P. J.. Verstreken, M.. and Collen, D. (1988) Fibrinolysis 2, Suppl. 2,

27. Declerck, P. J., Alessi, M. C., Verstreken, M., Kruithof, E. K. 0.. Juhan-Vague,

28. Levy, N. J., Ramesh, N.. Cicardi, M., Harrison, R. A,, and Davis, A. E. (1990)

29. Skriver, K., Wikoff, W. R., Patston, P. A., Tausk, F., Schapira, M., Kap1an.A. P..

392-395

77-78

I., and Collen, D. (1988) Blood 71, 220-225

Proc. Natl. Acad. Sci. U. S. A. 87, 265-268

19564 Substrate Properties of PM-1 Mutants

30. Perry, D. J., and Carrell, R. W. (1989) Mol. Biol. & Med. 6, 239-243 31. Holmes, W. E., Lijnen, H. R., Nelles, L., KluR, C., Nieuwenhuis, H. K, Rijken,

35. Sherman, P. M., Lawrence, D. A,, Yang, A. Y., Vandenberg, E. T., Paielli, D., Ohlson, S. T., Shore, J. D., and Ginsburg, D. (1992) J. Bid. Chern. 267,

and Bock, S. C. (1991) J. Biol. Chem. 266,9216-9221 Chem. 265, 18379-18385

D. C., and Collen, D. (1987) Science 258,209-211 758S7595

G. D. (1993) Biochim. BioDhvs. Acta 1202. 221-229 37. Ehrlich. H. J.. Gebbink. R. K. Keiier. J.. Linders. M.. Preissner. K. T.. and 32. Vaughan, D. E., Declerck, P. J., Reilly, T. M., Park, K., Collen, D., and Fasman, 36. York, J. D., Li, P., and Gardell, S. J. (1991) J. Biol. Chem. 266,8495-8500

33. Madison,'E. L.; Goldsmith, E , i, GerarKR. D.yGeth;ng, M. J. H., and Bassel- Pannekoek,' H. (199Oj J. SioZ. C/&. 265, 13029-13035

34. Shubeita, H. E., Cottey, T. L., Franke, A. Z., and Gerard, R. D. (1990) J. B i d . 266,20293-20301

~~ , ,

Duby, R. S. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 353043533 38. Lawrence, D. A,, Strandberg, L., Ericson, J., and Ny, T. (1990) J. Biol. Chem.