Importance of the amino-acid composition of the shutter region of plasminogen activator inhibitor-1 for its transitions to latent and substrate forms Martin Hansen, Marta N. Busse and Peter A. Andreasen Laboratory of Cellular Protein Science, Department of Molecular and Structural Biology, University of Aarhus, Denmark The serpins are of general protein chemical interest due to their ability to undergo a large conformational change consisting of the insertion of the reactive centre loop (RCL) as strand 4 of the central b sheet A. To make space for the incoming RCL, the ‘shutter region’ opens by the b strands 3A and 5A sliding apart over the underlying a helix B. Loop insertion occurs during the formation of complexes of serpins with their target serine proteinases and during latency transition. This type of loop insertion is unique to plasminogen activator inhibitor-1 (PAI-1). We report here that amino-acid substitutions in a buried cluster of three residues forming a hydrogen bonding network in the shutter region drastically accelerate a PAI-1 latency transition; that the rate was in all cases normalized by the PAI-1 binding protein vitronectin; and that substitution of an adjacent b strand 5A Lys residue, believed to anchor b strand 5A to other secondary structural elements, had differential effects on the rates of latency transition in the absence and the presence of vitronectin, respectively. An overlapping, but not identical set of substitutions resulted in an increased tendency to substrate behaviour of PAI-1 at reaction with its target proteinases. These findings show that vitronectin regulates the movements of the RCL through conformation- al changes of the shutter region and b strand 5A, are in agreement with RCL insertion proceeding by different routes during latency transition and complex formation, and contribute to the biochemical basis for the potential use of PAI-1 as a therapeutic target in cancer and cardiovascular diseases. Keywords: cancer; extracellular proteolysis; fibrinolysis; proteinase inhibitors; serine proteinases. The serpins constitute a protein family of which the best characterized members are serine proteinase inhibitors, including antithrombin III, a 1 -antitrypsin, and plasminogen activator inhibitor-1 (PAI-1). The serpins are globular proteins consisting of nine a helices and three b sheets (reviewed in [1–3]). Serpins are of general protein chemical interest due to their ability to undergo a large confor- mational change with the insertion of the surface-exposed reactive centre loop (RCL) as strand 4 of the large central b sheet A as the main event (Fig. 1). The RCL insertion results in a considerable stabilization compared to the native serpin structure, and is often referred to as the stressed-to- relaxed transition (for a review, see [2]). This stabilization forms the basis for the mechanism behind the inhibitory function of serpins. After cleavage of the P 1 –P 1 0 peptide bond in the RCL, the active site serine of the proteinase remains attached to the carboxyl group of the P 1 residue by an ester bond [4–6]. The subsequent RCL insertion into b sheet A therefore results in an < 7-nm translocation of the proteinase from the position of its initial encounter with the RCL to the other pole of the serpin [7– 10]. The translocation results in distortion of the proteinase [11] and inactivation of the enzymatic machinery [10]. Delayed RCL insertion results in hydrolysis of the ester bond, the serpin thus behaving as an ordinary substrate [12]. The stabil- ization caused by RCL insertion also underlies the unique conversion of active PAI-1 to the latent state, in which the N-terminal part of the intact RCL is inserted as b strand 4A without cleavage of any peptide bonds, and the C-terminal part is stretched along the surface of the molecule [13] (Fig. 1). In order to make space for the incoming new strand during RCL insertion, a fragment of the structure consisting of b strands 1A, 2A, 3A, and a helix F (the small serpin fragment) must slide away from the rest of the structure (the large serpin fragment). During the b sheet opening, the region around a helices D and E forms a flexible joint, and b strands 3A and 5A slide apart in a shutter-like manner over the underlying a helix B [14]. The central part of b strands 3A and 5A and the N-terminal part of a helix B is therefore referred to as the shutter region [2]. By high resolution X-ray crystal structure analysis of the native form of the serpin plasminogen activator inhibitor-2 (PAI-2) and the P 1 –P 1 0 cleaved form of horse leukocyte elastase inhibitor, a buried Enzymes: Urokinase-type plasminogen activator (EC 3.4.21.73). Note: plasminogen activator inhibitor-1 and vitronectin have the NCBI accession numbers P05121 and P04004, respectively. Note: a website is available at http://www.mbio.aau.dk Correspondence to M. Hansen, Laboratory of Cellular Protein Science, Department of Molecular and Structural Biology, University of Aarhus, 10C Gustav Wieds Vej, 8000 Aarhus C, Denmark. Fax: þ 45 86123178, Tel.: þ 45 89425079, E-mail: mah@mbio.aau.dk (Received 16 July 2001, revised 5 October 2001, accepted 8 October 2001) Abbreviations: HEK293T, the human embryonic kidney cell line 293T; LMW-uPA, low M r uPA; PAI-1, plasminogen activator inhibitor-1; PAI-2, plasminogen activator inhibitor-2; RCL, reactive centre loop; S-2444, L-5-pyroglutamyl-glycyl-L-arginine-p-nitroaniline; uPA, urokinase-type plasminogen activator. Eur. J. Biochem. 268, 6274–6283 (2001) q FEBS 2001 cluster with a complicated hydrogen bonding network was seen to be present in the shutter region, although differently organized, in both the stressed and the relaxed confor- mations [15]. The network involves the side chains of the amino acids in positions 53 and 56 in a helix B, 186 in b strand 3A, and position 334 in b strand 5A (Fig. 1; the numbering of amino acids in PAI-1 is according to the a 1 -antitrypsin template numbering scheme [1,3]). Sequence alignments of 219 serpins showed that residue 53 is a Ser in 92% of the cases; residue 56 is a Ser in 74% of the cases; residue 186 an Asn in 87% of the cases; and residue 334 a His in 80% of the cases [3]. In addition, residue 54 is a Pro in 89% of the cases. The importance of the identity of the residues present in these and adjacent positions are supported by the clustering of disease-causing mutations in the shutter region [16,17]. PAI-1 differs from most other serpins with respect to the identity of the residues in the buried cluster in the shutter region, having a Gly in position 56 and a Gln in position 334 (Fig. 1). This composition of amino acids in positions 53/56/334 is present in only 5% of the serpins, for example PN-1, RASP-1, TSA2004, and the viral serpins SPI-1, M2L, and H14-B [3]. A few previous studies have addressed the importance of the shutter region for the movements of the RCL in PAI-1. Berkenpas et al. [18] demonstrated that Ser and Thr substitutions of Pro54 delayed latency transition. We showed that a Q334H substitution accelerated latency transition [19]. We also implicated the region of b strand 5A overlying the buried cluster in RCL movements by demonstrating that increased proteolytic susceptibility of the peptide bonds Gln331– Ala332, Ala332–Leu333, and Lys335–Val336 accom- panied a transition to substrate behaviour in detergent- containing buffers at low temperatures [20,21]; and that a K335A substitution potentiated activity-neutralization of PAI-1 by some monoclonal antibodies [22]. Substitutions of Lys335 in a 1 -antitrypsin, a 1 -antichymotrypsin, and anti- thrombin III resulted in an increased conformational stability and a decreased specific inhibitory activity [23,24]. Lys335, localized in b strand 5A, points outward from the hydrophobic core and is conserved in 66% of serpins, the remaining serpins having Gln (10%), Ala (5%), and Arg (5%) in this position [3]. In order to investigate the importance of the shutter region for the unique types of RCL insertion in PAI-1, we have now undertaken a number of substitutions in the shutter region and b strand 5A of PAI-1 and studied their effect on the transition to latent and substrate forms and on the stabilizing effect of vitronectin, a flexible joint region-binding a cofactor known to delay PAI-1 latency transition (reviewed in [25,26]). Both transitions to latent and substrate forms were strongly but differently influenced by the amino-acid composition of the shutter region. Surprisingly, we found that substitution of Lys335 to Ala affected the rate of latency transition differently in the absence and presence of vitronectin. Fig. 1. The buried cluster and Lys335 in the shutter region of PAI-1. The top panel shows ribbon diagrams of active (left) and latent PAI-1 (right). Secondary structure elements are indicated as follows: blue, b sheet A; red, a helix B; green, gate region; yellow, RCL and b strand 1C in active PAI-1 and RCL inserted as b strand 4A in latent PAI-1. The P 1 Arg is displayed as a stick. The lower panel shows the three-dimensional structure of the shutter region of active PAI-1 (left) and latent PAI-1 (right). The molecules were rotated < 908 around a horizontal axis compared to the top panel. The colour code for secondary structure elements are as in the top panel. Presented amino-acid residues are: green, shutter region residues Ser53, Gly56, and Gln334; grey, Asn186 in b strand 3A; yellow, Lys335; purple, potential interaction partners for Lys335, i.e. Glu294 in b strand 6A and the backbone of Asn171 in the a helix F/b strand 3A loop. Note: SWISSPDB VIEWER uses the same signature for a helices and the short 3 10 -helix found in the a helix F/b strand 3A loop of active PAI-1. q FEBS 2001 PAI-1 shutter region mutations (Eur. J. Biochem. 268) 6275 MATERIALS AND METHODS PAI-1 In order to generate recombinant wild-type and mutated PAI-1, PAI-1 cDNA [27] was cloned into the expression vector pcDNA3.1(–) (Invitrogen) by use of standard techniques. The generated expression plasmid was denoted pcDNA3.1( –)PAI-1. Relevant fragments of the PAI-1 sequence were transferred to the mutagenesis vector LITMUS 28 (New England Biolabs). Point mutations were introduced into the PAI-1 cDNA fragment inserted into LITMUS 28 by use of the PCR-based QuickChangee Site- Directed Mutagenesis kit (Stratagene). The mutagenesis primers were from DNA Technology (Aarhus, Denmark), had a melting point above 60 8C, and were designed with the desired mutation(s) in the middle of their sequence. After mutagenesis, the fragments were moved back into pcDNA3.1(–)PAI-1 by the use of unique restriction sites. All mutations were verified by DNA sequencing of both strands of the PCR produced fragment after transfer back to pcDNA3.1(– )PAI-1, by use of either the Thermo Sequenasee II dye terminator cycle sequencing kit (Amersham Pharmacia Biotech AB) or the ABI PRISMe dye terminator cycle sequencing ready reaction kit (PerkinElmer). Recombinant PAI-1 variants were expressed in human embryonic kidney 293T cells (phenotype 293tsA1609neo) [28], grown in Dulbecco’s modified Eagle’s medium, by transient transfection using the calcium/phosphate precipi- tation technique [28]. Briefly, 1 h prior to transfection, new medium with 10% fetal bovine serum and 25 m M chloroquine was added to cells grown to 90% confluence in a 15-cm culture dish. Transfection was carried out by mixing 30 mgDNA(H 2 O added to a total of 1752 mL), 248 mL2 M CaCl 2 , and 2 mL 42 mM Hepes, pH 7.05, 274 m M NaCl, 10 mM KCl, 1.5 mM Na 2 HPO 4 ,11mM D -(þ )-glucose. After 1–2 min, this mixture was added dropwise to the cell medium and carefully distributed. Fresh medium without fetal bovine serum and chloroquine was added after 9–11 h of incubation. The conditioned medium was harvested after 48 and 96 h. Nontransfected or mock transfected HEK293T cells were shown not to express either PAI-1 or uPA by standard ELISA with monoclonal and polyclonal antibodies as capture and detection antibodies, respectively. Recombinant PAI-1 variants were purified from serum-free conditioned medium of the transfected cells by immunoaffinity chromatography in one step [29,30]. After purification, the variants were dialysed against NaCl/P i (0.01 M NaH 2 PO 4 , pH 7.4, 0.14 M NaCl) and concentrated to < 1 mg·mL 21 . Other proteins and miscellaneous materials The following materials were purchased from the indicated sources: BSA (Sigma); media components for HEK293T culturing (Life Technology); Qiaquick gel extraction kit (Qiagen); Rapid DNA ligation kit (Boehringer Mannheim); restriction enzymes (New England Biolabs Inc.; or Amersham Pharmacia Biotech AB; or Boehringer Mannheim); L-5-oxopropyl-glycyl-L-arginine-p-nitro- anilide (S-2444, Chromogenix AB); SDS (Serva); human urokinase-type plasminogen activator (uPA; Wakamoto Pharmaceutical Co.); vitronectin (Becton Dickinson; or Haemochrom AB). All other chemicals and reagents were of the highest quality commercially available. Activation of latent PAI-1 Unless otherwise indicated, latent PAI-1 was converted to the active conformation by denaturation with 0.1% SDS for 1 h at room temperature and refolding by a . 50-fold dilution in 0.1 M Tris, pH 8.1 (37 8C), containing either 1% BSA or 0.2% Triton X-100. Alternatively, latent PAI-1 was reactivated by denaturation with guanidinium chloride and refolded by dialysis against NaCl/P i . Assays for measuring specific inhibitory activity of PAI-1 The specific inhibitory activity of the reactivated PAI-1 variants was measured by titration against uPA in a direct peptidyl anilide assay at 37 8C, in the presence or absence of a slight excess of vitronectin over PAI-1 [30]. A twofold dilution series of PAI-1, with or without vitronectin, was made immediately after refolding, to avoid loss of activity due to fast latency transition. The dilution series of denatured and refolded PAI-1 (0–20 mg·mL 21 , 0–370 nM) were quickly (in less than 1 min) mixed with an equal volume (100 mL) of 0.25 mg·mL 21 (4.3 nM)uPA,0.1M Tris, pH 8.1, 1% BSA or 0.2% Triton X-100. The final concentrations of uPA was 0.125 mg·mL 21 (2.15 nM), of PAI-1 in the range 0–10 mg·mL 21 (0–185 nM), and of vitronectin in the range 0 –15 mg·mL 21 (0–200 nM). Upon completion of the uPA inhibition reaction (. 5 min), the remaining uPA activity in the reaction mixture was determined by use of L-5-oxopropyl-glycyl-L-arginine- p-nitroanilide (S-2444), a chromogenic peptidyl anilide substrate for uPA. The amount of active PAI-1, and thus the specific inhibitory activity, was calculated from the total amount of PAI-1 that had to be present to inhibit half of the uPA activity in the assays. PAI-1 latency transition assay Denatured and refolded PAI-1 wild-type and variants, in a concentration of 20 mg·mL 21 (370 nM), were incubated at 37 8C in 0.1 M Tris, pH 8.1, 1% BSA in the presence or absence of 30 mg·mL 21 (400 nM) vitronectin. Following incubation for different time periods, the specific inhibitory activity of PAI-1 wild-type and variants were determined as described above, and the functional half-lives of the variants were calculated. Analysis of functional behaviour of PAI-1 by reaction with low M r uPA (LMW-uPA) and SDS/PAGE PAI-1 portions (30 mg each) were denatured with 3 mL1% SDS, refolded by dilution to 1200 mL with 0.1 M Tris, pH 8.1, 1% BSA and incubated at 37 8C. At various time points, samples of 5 mg PAI-1 were mixed with 7.5 mg LMW-uPA and incubated for at least 2 min at 37 8C. BSA was then removed by the following procedure, performed at room temperature unless otherwise indicated: One-hundred micrograms of monoclonal murine anti-(PAI-1) IgG from hybridoma clone 2 [31], coupled to Sepharose-4B, was transferred to Ultrafreew-MC 0.22-mm filter units for 6276 M. Hansen et al. (Eur. J. Biochem. 268) q FEBS 2001 centrifugal filtration (Millipore, USA) and washed twice with 0.1 M Tris, pH 8.1. The sample (PAI-1, LMW-uPA, BSA) was then added and incubated for at least 30 min followed by four washes with 0.1 M Tris, 1 M NaCl, pH 8.1, and one wash with 0.1 M Tris, pH 8.1. PAI-1 was eluted by incubation with 400 mL3 M ammoniumthiocyanat (NH 4 SCN) for at least 30 min at 37 8C before centrifugation at 16 500 g for 10 min. The samples were precipitated with trichloroacetic acid, and subjected to 6–16% gradient SDS/PAGE. Determination of second order rate constants for the reaction between PAI-1 and uPA The second order rate constants were determined as described previously [32]. The calculation of the second order rate constants is based on the assumption that the concentration of active PAI-1 is unchanged during the assay. As most of these variants have significantly shorter functional half-lives (see below) than wild-type, the calculated second order rate constants for the variants were expected to be somewhat lower than their real values. Gel filtration Thirty-microgram portions of PAI-1 wild-type and variants were analysed by FPLC gel filtration on a Superdex 200 HR 10/30 column (Pharmacia) in 0.1 M Tris, pH 8.1, 0.5 M NaCl at 4 8C, using a flow rate of 0.3 or 0.4 mL·min 21 . The following marker proteins were used: BSA (M r 67 000), murine IgG (M r 150 000), and b-galactosidase (M r 540 000). Molecular graphics SWISSPDB VIEWER [33] was used to display the three- dimensional X-ray structure of active [34] and latent [13] PAI-1. Statistical analysis Data were evaluated by Student’s t-test. Fig. 2. Gel filtration of purified recombinant PAI-1 from HEK293T cells. As representative gel filtration profiles are shown those of PAI-1 wild-type, G56S/Q334H, and PAI-1 G56S/Q334H/K335A. All other variants also showed a single peak in the position expected for monomeric PAI-1. V 0 , void volume. The migration of the marker proteins BSA (M r 67 000), murine IgG (M r 150 000), and b-galactosidase (M r 540 000) are indicated by arrows above the profiles. Table 1. Specific inhibitory activity of PAI-1 variants towards uPA. The most common amino-acid composition of the buried polar cluster (positions 53/56/334) in serpins is S/S/H. The composition S/S/Q is identical to that of alaserpin, S/G/H is identical to that of CP-9, A/G/H is identical to that of heparin cofactor II, while the S/A/S composition is present in angiotensinogen [1,3]. The investigated residues according to the PAI-1 numbering (1Ser-Ala-Val-His-His-) are 37/40/324/325 [27]. Means ^ SD (numbers of assays are indicated). *, Significantly different from wild-type (P , 0.005). †, Significantly different from the value without vitronectin (P , 0.005). PAI-1 variant Composition of positions 53/56/334 PAI-1 activity (% of theoretical max) Vitronectin effect (fold increase)– Vitronectin þ Vitronectin Wild-type S/G/Q 87.1 ^ 22.2 (20) 113.7 ^ 28.0 (10) † 1.3 K335A S/G/Q 114.9 ^ 21.5 (5) 129.5 ^ 21.1 (3) 1.1 S53A A/G/Q 71.6 ^ 9.4 (8) 84.9 ^ 5.9 (3) 1.2 S53A/K335A A/G/Q 84.3 ^ 15.5 (4) 91.0 ^ 13.3 (3) 1.1 G56A S/A/Q 59.6 ^ 7.8 (6) * 56.1 ^ 6.3 (3) * 0.9 G56S S/S/Q 73.1 ^ 10.6 (5) 127.1 ^ 6.2 (3) † 1.7 G56S/K335A S/S/Q 120.0 ^ 1.6 (3) 129.6 ^ 2.5 (3) † 1.1 Q334A S/G/A 58.1 ^ 7.5 (6) * 73.1 ^ 5.7 (3) 1.3 Q334A/K335A S/G/A 62.8 ^ 4.7 (4) 57.8 ^ 5.3 (4) * 0.9 Q334H S/G/H 61.9 ^ 13.5 (6) 90.7 ^ 5.8 (3) 1.5 Q334H/K335A S/G/H 72.5 ^ 24.0 (6) 87.3 ^ 15.5 (3) 1.2 Q334S S/G/S 75.9 ^ 20.6 (7) 104.8 ^ 5.9 (3) 1.4 S53A/Q334H A/G/H 48.9 ^ 10.0 (9) * 74.5 ^ 9.1 (3) † 1.5 G56A/Q334S S/A/S 34.5 ^ 4.4 (7) * 23.8 ^ 2.1 (3) *† 0.7 G56S/Q334H S/S/H 64.7 ^ 14.2 (9) * 101.0 ^ 5.2 (3) † 1.6 G56S/Q334H/K335A S/S/H 83.8 ^ 14.2 (6) 81.6 ^ 9.8 (3) 1.0 q FEBS 2001 PAI-1 shutter region mutations (Eur. J. Biochem. 268) 6277 RESULTS Expression and purification of recombinant PAI-1 in HEK293T cells PAI-1 wild-type and the substitution variants were expressed in HEK293T cells and purified from their conditioned medium by immunoaffinity chromatography. The yields of purified protein were 1.4 –11.4 mg protein per L of conditioned medium. PAI-1 wild-type, K335A, S53A/ K335A, G56S/K335A, and Q334H/K335A were obtained in the greatest yields. All variant preparations were . 95% pure, as evaluated by SDS/PAGE, and migrated as a single sharp peak in the position expected for monomeric PAI-1 in gel filtration (Fig. 2). N-Terminal sequencing of the produced PAI-1 showed two distinct sequences in almost equal amounts, SAVHHPPS and VHHPPSYV, in agreement with the previously reported N-terminal heterogeneity of natural PAI-1 [27]. The purified recombinant PAI-1 was in the latent form, but could be reactivated by denaturation and refolding, either by SDS and refolding by dilution into a buffer with 1% BSA, or by guanidinium chloride and refolding by dialysis against NaCl/P i . In this study, PAI-1 was routinely reactivated using SDS, as some of the variants had a very fast latency transition (see below) and would therefore lose all activity during the dialysis used for refolding after guanidinium chloride denaturation. The specific inhibitory activities of most PAI-1 variants, when denatured with SDS and refolded in BSA-containing buffer, were 60– 80% of the theoretical maximum, and thus indistinguishable from recombinant PAI-1 wild-type (Table 1). However, the recombinant variants G56A, Q334A, S53A/Q334H, G56A/Q334S, and G56S/Q334H showed a small, but statistically significant (P , 0.005) reduction in specific inhibitory activity as compared to wild- type. All variants except the variant G56A/Q334S had a second order rate constant differing less than 2.5-fold from that of wild-type (data not shown). The second order rate constant for G56A/Q334S was 3.8-fold lower than that of the wild-type, but this can be ascribed to a fast decrease in inhibitory activity of this variant during the experiment (see below). Vitronectin caused a small, but statistically significant increase of the specific inhibitory activity of the PAI-1 wild-type and the variants G56S, S53A/Q334H, G56A/Q334S, G56S/Q334H, and G56S/K335A. Interest- ingly, the specific inhibitory activity of G56A/Q334S was slightly decreased by vitronectin. Latency transition of PAI-1 wild-type and PAI-1 variants in the absence and the presence of vitronectin To estimate the functional stability of the variants, their specific inhibitory activities were measured after different times of incubation at 37 8C. The rate of activity loss was determined in the absence or the presence of vitronectin. Representative examples are given in Fig. 3 and a summary of all experiments is given in Table 2. All variants with substitutions in the buried cluster, except S53A, had significantly reduced functional half-lives. The K335A substitution caused at most a slight (, 1.4-fold) delay of latency transition when introduced into wild-type, S53A, G56S, and Q334A, but caused considerable delays Fig. 3. Determination of functional stability of PAI-1 variants. The specific inhibitory activity of the indicated PAI-1 variants was determined after the indicated time periods of incubation at 37 8C in the absence (closed circles) or presence (open circles) of vitronectin. The relative specific inhibitory activity was plotted as a function of time. The functional half-lives of the shown variants in the absence and presence of vitronectin, respectively, in the experiments shown were: wild-type, 48.6 and 65.2 min; K335A, 78.9 and 41.8 min; S53A, 69.4 and 109.9 min; S53A/K335A, 99.2 and 43.5 min; G56S/Q334H, 5.3 and 59.8 min; G56S/Q334H/K335A, 26.5 and 35.1 min. A summary of all experiments is given in Table 2. 6278 M. Hansen et al. (Eur. J. Biochem. 268) q FEBS 2001 (2.9-fold and 4.6-fold, respectively) of the very unstable variants Q334H and G56S/Q334H. At the routinely used pH of 8.1, there is only a slight effect of vitronectin on the rate of latency transition of PAI-1 wild-type, and the use of this pH therefore allowed optimization of the difference in the effect of vitronectin on wild-type and the unstable variants. The presence of vitronectin during the incubations had a stabilizing effect on all PAI-1 variants with substitutions in the buried cluster of the shutter region. The most pronounced effect was observed on the most unstable variants, so that their functional half-lives in the presence of vitronectin increased towards that of PAI-1 wild-type (Table 2). Surprisingly, the stabilizing effect of vitronectin was abolished by the K335A substitution. None of the variants with this substitution had longer half-lives in the presence of vitronectin, and with most of them, vitronectin even accelerated the activity loss. Hence, while the K335A substitution had a stabilizing effect in the absence of vitronectin, it had a destabilizing effect in the presence of vitronectin. Importantly, the K335A substitution did not result in altered affinity of PAI-1 to vitronectin (T. Wind, & P.A. Andreasen, Department of Molecular and Structural Biology, Aarhus University, Denmark, personal communication). Although the assays were routinely performed in a buffer of 0.1 M Tris, pH 8.1, similar results were obtained with a buffer of 0.1 M Tris, pH 7.4. In addition, when examining the results obtained with SDS-activated PAI-1 vs. guanidinium chloride-activated PAI-1 with wild-type and two of the most stable variants (S53A and K335A), no distinguishable difference was observed. In order to ensure that the loss of activity during the incubations at 37 8C was due to latency transition, PAI-1, that had been incubated for various time periods at 37 8C, was reacted with an excess of LMW-uPA, and the reaction products were analysed by SDS/PAGE. Representative experiments are shown in Fig. 4. Nonincubated wild-type reacted to form the expected < 80 000-Da LMW-uPA– PAI-1 complex. A fraction of wild-type reacted in a substrate-manner, giving rise to the < 50 000-Da N-terminal fragment resulting from P 1 –P 1 0 cleavage, the < 4000-Da C-terminal fragment not being recovered by the gel system used here. During incubation at 37 8C, the amount of a PAI-1 form inert to reaction with LMW-uPA increased with time, and the complex formation and substrate reaction decreased, in agreement with an increasing fraction of PAI-1 being in the latent state. The same was true for the PAI-1 variants, except that the accumulation of inert PAI-1 occurred faster, in agreement with the activity measurements. It should also be noted that a fraction of all tested variants seemed to be present in a stable substrate form [30,35,36], the amount of which did not decrease during the incubations at 37 8C. On this basis, we concluded that the loss of activity during incubations at 37 8C was caused by latency transition, and that the effects of the substitutions and of vitronectin on the functional half-lives was caused by changes in the rate of latency transition. Effect of shutter region substitutions on PAI-1 active to substrate transition Nonionic detergents induce substrate behaviour in glycosy- lated PAI-1 at 0 8C, but much less so at 37 8C [20,21], and induce substrate behaviour in nonglycosylated PAI-1 at 37 8C as well as at 0 8C [37]. Therefore, to study the effect of shutter region substitutions on the transition of PAI-1 to a substrate form, we replaced the BSA in the assay buffer with 0.2% Triton X-100. The variants with a S53A, Q334A, or Q334S substitution all had significantly reduced specific inhibitory activity in Triton X-100 containing buffer at 37 8C as compared to PAI-1 wild-type (Table 3). Analysis of Table 2. Stability of specific inhibitory activity of PAI-1 variants at 37 8C. Means, SDs, and numbers of experiments are indicated. *, Significantly different from wild-type (P , 0.005). †, Significantly different from the value without vitronectin (P , 0.005). ‡, Significantly different from the corresponding variants without the K335A substitution (P , 0.02). PAI-1 variant Functional half-lives (min) Vitronectin effect (fold increase)– Vitronectin þ Vitronectin Wild-type 54.7 ^ 13.5 (16) 63.4 ^ 11.6 (10) 1.2 K335A 76.9 ^ 11.6 (4) *‡ 35.3 ^ 6.7 (3) *†‡ 0.5 S53A 64.0 ^ 8.6 (5) 100.7 ^ 9.4 (3) *† 1.6 S53A/K335A 85.1 ^ 20.5 (4) * 32.0 ^ 10.0 (3) *†‡ 0.4 G56A 26.1 ^ 4.4 (4) * 36.5 ^ 6.8 (3) * 1.4 G56S 19.7 ^ 1.7 (4) * 54.9 ^ 4.1 (3) † 2.8 G56S/K335A 25.5 ^ 6.3 (3) * 12.3 ^ 2.1 (3) *‡ 0.5 Q334A 10.9 ^ 1.5 (4) * 39.9 ^ 10.2 (3) *† 3.7 Q334A/K335A 12.9 ^ 2.9 (3) * 14.1 ^ 0.9 (3) *‡ 1.1 Q334H 10.9 ^ 1.4 (4) * 52.3 ^ 6.0 (3) † 4.8 Q334H/K335A 32.1 ^ 2.7 (5) *‡ 29.4 ^ 1.2 (3) *‡ 0.9 Q334S 23.4 ^ 2.5 (5) * 75.2 ^ 15.8 (3) † 3.2 S53A/Q334H 18.5 ^ 3.2 (7) * 78.3 ^ 12.5 (3) † 4.2 G56A/Q334S 9.7 ^ 1.6 (4) * 72.9 ^ 17.7 (3) † 7.5 G56S/Q334H 6.1 ^ 1.5 (7) * 62.4 ^ 15.6 (4) † 10.2 G56S/Q334H/K335A 27.8 ^ 2.6 (5) *‡ 34.3 ^ 2.3 (3) *‡ 1.2 q FEBS 2001 PAI-1 shutter region mutations (Eur. J. Biochem. 268) 6279 the distribution of PAI-1 between different functional forms by the use of reaction with LMW-uPA and SDS/PAGE confirmed that the decreased specific inhibitory activity was caused by an increased tendency to substrate behaviour and not the generation of an inert form, as shown by the representative experiments shown in Fig. 5. The K335A substitution did not counteract the increased tendency to substrate behaviour (Table 3). DISCUSSION In this report, we show that the combination of amino acids in positions 53, 56, and 334 in the shutter region of PAI-1 is an important determinant of the latency transition rate. Except one, all tested deviations from the wild-type combination of amino acids in these positions resulted in an accelerated latency transition. Substitution of the Lys residue in position 335 counteracted the accelerating effect of some of the substitutions in positions 53, 56, and 334. The substitutions also had specific effects on the vitronectin and Triton X-100 induced changes in PAI-1 latency transition and specific inhibitory activity, respectively. Based on our observations we propose that the substitutions in positions 53, 56, and 334 affect the latency transition by changing the local conformation of these residues, including their hydrogen bonds. As there is no P 1 –P 1 0 cleavage during latency transition, strand insertion must imply the passage of the intact RCL through the ‘gate region’, which is situated between (a) the turn between b strands 3C and 4C (residues 204–219) and (b) the turn between b strands 3B and a helix G (residues 257–259) [13,38,39] (Fig. 1). Because of steric reasons, it is not very likely that the RCL can surround the turn between b strands 3C and 4C without having a completely stretched-out conformation. Only after the RCL has passed this turn can the final insertion into b sheet A proceed [39]. Considering that RCL insertion into b sheet A is several orders of magnitude faster during complex formation than during latency transition, it seems reasonable to presume that the passage of the RCL through the gate region is rate limiting for latency transition. This presumption is supported by the observation that substitutions of basic residues in the turn between b strands 3C and 4C with acidic residues accelerate latency transition [40,41]. On this basis, we reach the conclusion that the substitutions in the shutter region affect the rate of latency transition by affecting the rate of passage of the RCL through the gate region. Based on the amino- acid sequence of the RCL and b strand 5A being directly continuous, it may be proposed that movements of the RCL during passage through the gate region are coupled to movements of b strand 5A and therefore sensitive to the interactions of b strand 5Awith the underlying structure. An alternative, but with the presently available information, a less likely explanation is that passage of the RCL through the gate region is rapid and reversible, and that it is the b sheet A opening and the final insertion of RCL as b strand 4A that is rate limiting for latency transition. Two facts complicate the interpretation of our results on the basis of detailed structural considerations. First, the three-dimensional structure available for active PAI-1 [34,42] is that of a mutant with a strongly delayed latency transition. The stabilizing mutations may well have affected Fig. 4. Analysis of the functional behavior of PAI-1 by SDS/PAGE. PAI-1 wild-type or Q334H (25 mg·mL 21 ) were SDS-denatured and incubated in a buffer of 0.1 M Tris, 1% BSA, pH 8.1 at 37 8C. After the indicated time periods, samples corresponding to 5 mg PAI-1 were incubated with 7.5 mg LMW-uPA for at least 2 min at 37 8C at a PAI-1 concentration of 20 mg·mL 21 and an LMW-uPA concentration of 30 mg·mL 21 before inert PAI-1, reactive center-cleaved PAI-1, and LMW-uPA–PAI-1 complex were isolated from the reaction mixture by immunoaffinity chromatography, removing most of the BSA and most of the excess of LMW-uPA. The samples were then precipitated with trichloroacetic acid and subjected to SDS/PAGE in gradient gels with 6–16% polyacrylamide. The positions of LMW-uPA, reactive centre- cleaved PAI-1 (RCC PAI-1), native/inert PAI-1, BSA, and LMW-uPA- PAI-1 complex (complex) are indicated to the right. N, PAI-1 incubated for at least 3 h at 37 8C without LMW-uPA added. The apparently low fraction of nonincubated PAI-1 forming a complex is related to a somewhat higher tendency to substrate behavior at the high PAI-1 and LMW-uPA concentrations used in this assay [22]. Table 3. Effect of 0.2% Triton X-100 on specific inhibitory activity of PAI-1 at 37 8C. The specific inhibitory activity of each variant is given as a fraction of the specific inhibitory activity of the same variant in 1% BSA. Means, SDs, and numbers of experiments are indicated. *, Significantly different from wild-type (P , 0.005). PAI-1 variant Specific inhibitory activity Wild-type 0.87 ^ 0.12 (7) K335A 0.95 ^ 0.09 (3) S53A 0.20 ^ 0.02 (3)* S53A/K335A 0.17 ^ 0.02 (3)* G56A 0.65 ^ 0.11 (3) G56S 1.17 ^ 0.16 (3) G56S/K335A 1.06 ^ 0.03 (3) Q334A 0.12 ^ 0.01 (3)* Q334A/K335A 0.18 ^ 0.03 (3)* Q334H 0.61 ^ 0.08 (3)* Q334H/K335A 0.77 ^ 0.01 (3) Q334S 0.21 ^ 0.05 (3)* S53A/Q334H 0.13 ^ 0.01 (3)* G56A/Q334S 0.10 ^ 0.02 (3)* G56S/Q334H 0.45 ^ 0.03 (3)* G56S/Q334H/K335A 0.74 ^ 0.09 (3) 6280 M. Hansen et al. (Eur. J. Biochem. 268) q FEBS 2001 the conformation of the shutter region. Second, the rate of strand insertion during latency transition may be affected not only through a change in the conformation of the active form, but also by a change in the conformation of a transition state with an unknown three-dimensional structure. Nevertheless, it seems reasonable to conclude that the hydrogen bonds from the side chain of S53A have little influence on the rate of latency transition, as the S53A substitution resulted only in a slightly decreased rate of latency transition. Also, the Q334A substitution, removing the hydrogen bonding ability of the side chain in position 334, resulted in a half-life similar to that of the Q334H substitution, indicating that the specific hydrogen bonds formed by Gln334, absent in Q334A and apparently not substituted by any possible hydrogen bonds formed by a His in this position, play a pivotal role. Furthermore, the accelerating effect of substitutions of Gly56 is likely to be caused by a slight reorganization of the region introduced by the larger side chains. The variant with the amino-acid combination Ser53/ Ser56/His334, identical to that of 63% of serpins [3], has the shortest half-life of all variants tested here, and in fact one of the shortest half-lives reported for a PAI-1 variant. Thus, the reason for the tendency of PAI-1 to undergo latency transition is to be sought outside the shutter region, in agreement with previous results [18,40,41]. On the other hand, if PAI-1 had possessed an amino-acid composition in the shutter region identical to that of most other serpins, its tendency to latency transition would have given it an extremely short half-life, probably incompatible with its physiological functions. Irrespective of the substitutions introduced into positions 53, 56, and 334, vitronectin brought the latency transition rate back to values close to that of PAI-1 wild-type. The most plausible explanation of this observation is that vitronectin, from its binding site in the flexible joint region [43], directs the movements of the RCL, almost totally over-ruling the effect of the local hydrogen bonding network of the residues in positions 53, 56, and 334. The K335A substitution delayed the latency transition when introduced in some of the variants, and most strongly when introduced into the very unstable variants Q334H and G56S/Q334H. The side chain of Lys335 points away from the buried cluster in positions 53/56/334 (Fig. 1). On the basis of the available three-dimensional structures, several intramolecular interactions of Lys335 may be suggested. The possible interactions include a connection to the loop between a helix F and b strand 3A by a hydrogen bond to the carbonyl oxygen atom of the backbone of Asn171 [19,22], by hydrophobic interactions with residues in that loop [23], or by participation in formation of a chloride binding site together with residues in that loop and Lys337 [44]. The possible interactions also include a salt bridge to Glu294 in b strand 6A (Fig. 1). It therefore seems likely that the constraints caused by the interactions of Lys335 contribute to maintaining the RCL in a state with a relatively facilitated passage through the gate region during latency transition, via an effect on the conformation of b strand 5A and of the buried cluster in positions 53, 56, and 334. In contrast, in the presence of vitronectin, the K335A substitution caused a twofold to fivefold acceleration of latency transition compared to wild-type. In fact, vitronectin did not delay latency transition of any of the variants harbouring the K335A substitution. On the basis of the opposite effects of the K335A substitution in the absence and presence of vitronectin, we propose that the conformational change of PAI-1 following the binding of vitronectin implicates a reorientation of the side chain of Lys335 relative to its surroundings, allowing it to make new contacts, concerted rearrangements of the shutter region and changes in the movements of the RCL. We observed an increased tendency to substrate behaviour in Triton X-100 at 37 8C in a set of mutations overlapping with, but not identical to that giving an increased rate of latency transition. Previously, Triton X-100 was found to induce substrate behaviour in nonglycosylated PAI-1 at 0 and 37 8C [37] and to induce substrate behaviour Fig. 5. Effect of Triton X-100 on the distribution of PAI-1 between different functional forms. The indicated concentrations of the indicated PAI-1 variants, were incubated in 0.1 M Tris, 0.2% Triton X-100, pH 8.1 at 37 8C with LMW-uPA at a concentration of 0.175 mg·mL 21 for at least 15 min prior to determination of the remaining LMM-uPA activity with S-2444. Inset: portions of 3 mg of the PAI-1 variants were diluted in 0.1 M Tris, 0.2% Triton X-100, pH 8.1 to 0.16 mg·mL 21 followed by incubation at 37 8C with 0.175 mg·mL 21 LMM-uPA for 15 min. The samples were precipitated with trichloroacetic acid and subjected to SDS/PAGE in gradient gels with 6–16% polyacrylamide. The positions of LMW-uPA, reactive centre-cleaved PAI-1 (RCC PAI-1), native/inert PAI-1, and LMW-uPA–PAI-1 complex are indicated to the left. q FEBS 2001 PAI-1 shutter region mutations (Eur. J. Biochem. 268) 6281 in glycosylated PAI-1 at 0 8C, but not at 37 8C [20,21]. On the basis of our present findings, we propose that Triton X-100 acts by destabilizing the shutter region, this happening more readily with less perfect interactions between the side chains, resulting in a delay in strand insertion during reaction with the target proteinase. On the other hand, the Triton X-100-induced substrate behaviour did not seem to implicate the interactions of the Lys335 side chain, in contrast to antibody-induced substrate behaviour that was potentiated by the K335A substitution [19,22]. The observation of latency transition and complex formation being affected differently by mutations in the shutter region and b strand 5A is in agreement with RCL insertion following different routes in the two cases. PAI-1 is a potential target for antithrombotic [45] and anticancer therapy [46,47]. The biochemical mechanism of action of a few PAI-1 neutralisers has been characterized, including monoclonal antibodies and organochemical compounds. These compounds neutralize PAI-1 either by steric hindrance, by inducing conversion to the latent state, by inducing substrate behaviour, and/or by inducing conversion to inert polymers [48– 52]. The present results prompt further studies into the role of the shutter region and b strand 5A in PAI-1 in conformational changes leading to neutralization. ACKNOWLEDGEMENTS Dr Kees Rodenburg is thanked for fruitful discussions in the early phase of this work. Dr Claus Oxvig is acknowledged for providing the HEK293T cell line. This work was supported financially by the Danish Cancer Society, the Danish Research Agency, the Danish Heart Foundation, the NOVO-Nordisk Foundation, and the Danish Cancer Foundation. REFERENCES 1. Huber, R. & Carrell, R.W. (1989) Implications of the three- dimensional structure of a 1 -antitrypsin for structure and function of serpins. Biochemistry 28, 8951–8966. 2. Carrell, R.W. & Stein, P.E. (1996) The biostructural pathology of the serpins: critical function of sheet opening mechanism. Biol. Chem. Hoppe Seyler 377, 1–17. 3. Irving, J.A., Pike, R.N., Lesk, A.M. & Whisstock, J.C. (2000) Phylogeny of the serpin superfamily: implications of patterns of amino acid conservation for structure and function. Genome Res. 10, 1845–1864. 4. Lawrence, D.A., Ginsburg, D., Day, D.E., Berkenpas, M.B., Verhamme, I.M., Kvassman, J.O. & Shore, J.D. (1995) Serpin- protease complexes are trapped as stable acyl-enzyme inter- mediates. J. Biol. Chem. 270, 25309–25312. 5. Wilczynska, M., Fa, M., Ohlsson, P.I. & Ny, T. (1995) The inhibition mechanism of serpins. J. Biol. Chem. 270, 29652–29655. 6. Egelund, R., Rodenburg, K.W., Andreasen, P.A., Rasmussen, M.S., Guldberg, R.E. & Petersen, T.E. (1998) An ester bond linking a fragment of a serine proteinase to its serpin inhibitor. Biochemistry 37, 6375–6379. 7. Shore, J.D., Day, D.E., Francis-Chmura, A.M., Verhamme, I., Kvassman, J., Lawrence, D.A. & Ginsburg, D. (1995) A fluorescent probe study of plasminogen activator inhibitor-1. Evidence for reactive center loop insertion and its role in the inhibitory mechanism. J. Biol. Chem. 270, 5395–5398. 8. Stratikos, E. & Gettins, P.G. (1999) Formation of the covalent serpin-proteinase complex involves translocation of the proteinase by more than 70A ˚ and full insertion of the reactive center loop into b-sheet A. Proc. Natl Acad. Sci. USA 96, 4808–4813. 9. Fa, M., Bergstrom, F., Hagglof, P., Wilczynska, M., Johansson, L.B. & Ny, T. (2000) The structure of a serpin–protease complex revealed by intramolecular distance measurements using donor– donor energy migration and mapping of interaction sites. Struct. Fold Des. 8, 397–405. 10. Huntington, J.A., Read, R.J. & Carrell, R.W. (2000) Structure of a serpin–protease complex shows inhibition by deformation. Nature 407, 923–926. 11. Egelund, R., Petersen, T.E. & Andreasen, P.A. (2001) A serpin- induced extensive proteolytic susceptibility of urokinase-type plasminogen activator implicates distortion of the proteinase substrate-binding pocket and oxyanion hole in the serpin inhibitory mechanism. Eur. J. Biochem. 268, 673 – 685. 12. Lawrence, D.A., Olson, S.T., Muhammad, S., Day, D.E., Kvass- man, J.O., Ginsburg, D. & Shore, J.D. (2000) Partitioning of serpin- proteinase reactions between stable inhibition and substrate cleavage is regulated by the rate of serpin reactive center loop insertion into b-sheet A. J. Biol. Chem. 275, 5839–5844. 13. Mottonen, J., Strand, A., Symersky, J., Sweet, R.M., Danley, D.E., Geoghegan, K.F., Gerard, R.D. & Goldsmith, E.J. (1992) Structural basis of latency in plasminogen activator inhibitor-1. Nature 355, 270–273. 14. Stein, P. & Chothia, C. (1991) Serpin tertiary structure transformation. J. Mol. Biol. 221, 615–621. 15. Harrop, S.J., Jankova, L., Coles, M., Jardine, D., Whittaker, J.S., Gould, A.R., Meister, A., King, G.C., Mabbutt, B.C. & Curmi, P.M. (1999) The crystal structure of plasminogen activator inhibitor 2 at 2.0A ˚ resolution: implications for serpin function. Structure 7, 43–54. 16. Stein, P.E. & Carrell, R.W. (1995) What do dysfunctional serpins tell us about molecular mobility and disease? Nat. Struct. Biol. 2, 96–113. 17. Davis, R.L., Shrimpton, A.E., Holohan, P.D., Bradshaw, C., Feiglin, D., Collins, G.H., Sonderegger, P., Kinter, J., Becker, L.M., Lacbawan, F., Krasnewich, D., Muenke, M., Lawrence, D.A., Yerby, M.S., Shaw, C.M., Gooptu, B., Elliott, P.R., Finch, J.T., Carrell, R.W. & Lomas, D.A. (1999) Familial dementia caused by polymerization of mutant neuroserpin. Nature 401, 376 –379. 18. Berkenpas, M.B., Lawrence, D.A. & Ginsburg, D. (1995) Molecular evolution of plasminogen activator inhibitor-1 func- tional stability. EMBO J. 14, 2969–2977. 19. Kirkegaard, T., Jensen, S., Schousboe, S.L., Petersen, H.H., Egelund, R., Andreasen, P.A. & Rodenburg, K.W. (1999) Engineering of conformations of plasminogen activator inhibitor- 1. A crucial role of b-strand 5A residues in the transition of active form to latent and substrate forms. Eur. J. Biochem. 263, 577–586. 20. Kjøller, L., Martensen, P.M., Sottrup-Jensen, L., Justesen, J., Rodenburg, K.W. & Andreasen, P.A. (1996) Conformational changes of the reactive-centre loop and b-strand 5A accompany temperature-dependent inhibitor-substrate transition of plasmino- gen-activator inhibitor 1. Eur. J. Biochem. 241, 38–46. 21. Andreasen, P.A., Egelund, R., Jensen, S. & Rodenburg, K.W. (1999) Solvent effects on activity and conformation of plasminogen activator inhibitor-1. Thromb. Haemost. 81, 407–414. 22. Schousboe, S.L., Egelund, R., Kirkegaard, T., Preissner, K.T., Rodenburg, K.W. & Andreasen, P.A. (2000) Vitronectin and substitution of a b-strand 5A lysine residue potentiate activity- neutralization of PA inhibitor-1 by monoclonal antibodies against a-helix F. Thromb. Haemost. 83, 742–751. 23. Im, H. & Yu, M.H. (2000) Role of Lys335 in the metastability and function of inhibitory serpins. Protein Sci. 9, 934–941. 24. Im, H., Seo, E.J. & Yu, M.H. (1999) Metastability in the inhibitory mechanism of human a 1 -antitrypsin. J. Biol. Chem. 274, 11072–11077. 6282 M. Hansen et al. (Eur. J. Biochem. 268) q FEBS 2001 25. Preissner, K.T. & Jenne, D. (1991) Structure of vitronectin and its biological role in haemostasis. Thromb. Haemost. 66, 123–132. 26. Deng, G., Royle, G., Seiffert, D. & Loskutoff, D.J. (1995) The PAI- 1/vitronectin interaction: two cats in a bag? Thromb. Haemost. 74, 66–70. 27. Andreasen, P.A., Riccio, A., Welinder, K.G., Douglas, R., Sartorio, R., Nielsen, L.S., Oppenheimer, C., Blasi, F. & Danø, K. (1986) Plasminogen activator inhibitor type-1: reactive center and amino- terminal heterogeneity determined by protein and cDNA sequen- cing. FEBS Lett. 209, 213–218. 28. DuBridge, R.B., Tang, P., Hsia, H.C., Leong, P.M., Miller, J.H. & Calos, M.P. (1987) Analysis of mutation in human cells by using an Epstein–Barr virus shuttle system. Mol. Cell. Biol. 7, 379–387. 29. Munch, M., Heegaard, C., Jensen, P.H. & Andreasen, P.A. (1991) Type-1 inhibitor of plasminogen activators. Distinction between latent, activated and reactive centre-cleaved forms with thermal stability and monoclonal antibodies. FEBS Lett. 295, 102–106. 30. Munch, M., Heegaard, C.W. & Andreasen, P.A. (1993) Interconversions between active, inert and substrate forms of denatured/refolded type-1 plasminogen activator inhibitor. Bio- chim. Biophys. Acta 1202, 29–37. 31. Nielsen, L.S., Hansen, J.G., Andreasen, P.A., Skriver, L., Danø, K. & Zeuthen, J. (1983) Monoclonal antibody to human 66,000 molecular wieght plasminogen activator from melanoma cells. Specific enzyme inhibition and one-step affinity purification. EMBO J. 2, 115–119. 32. Petersen, H.H., Hansen, M., Schousboe, S.L. & P.A.A. (2001) Localisation of epitopes of monoclonal anti-urokinase-type plasminogen activator antibodies. Relationship to antibody effects on molecular interactions of the enzyme. Eur. J. Biochem. 268, 4430–4439. 33. Guex, N. & Peitsch, M.C. (1997) SWISS-MODEL and the Swiss- PdbViewer: an environment for comparative protein modeling. Electrophoresis 18, 2714–2723. 34. Nar, H., Bauer, M., Stassen, J.M., Lang, D., Gils, A. & Declerck, P.J. (2000) Plasminogen activator inhibitor 1. Structure of the native serpin, comparison to its other conformers and implications for serpin inactivation. J. Mol. Biol. 297, 683–695. 35. Declerck, P.J., De Mol, M., Vaughan, D.E. & Collen, D. (1992) Identification of a conformationally distinct form of plasminogen activator inhibitor-1, acting as a noninhibitory substrate for tissue- type plasminogen activator. J. Biol. Chem. 267, 11693–11696. 36. Urano, T., Strandberg, L., Johansson, L.B. & Ny, T. (1992) A substrate-like form of plasminogen-activator-inhibitor type 1. Conversions between different forms by sodium dodecyl sulphate. Eur. J. Biochem. 209, 985–992. 37. Gils, A. & Declerck, P.J. (1998) Modulation of plasminogen activator inhibitor 1 by Triton X-100-identification of two consecutive conformational transitions. Thromb. Haemost. 80, 286–291. 38. Aertgeerts, K., De Bondt, H.L., De Ranter, C.J. & Declerck, P.J. (1995) Mechanisms contributing to the conformational and functional flexibility of plasminogen activator inhibitor-1. Nat. Struct. Biol. 2, 891–897. 39. Kru ¨ ger, P., Verheyden, S., Declerck, P.J. & Engelborghs, Y. (2001) Extending the capabilities of targeted molecular dynamics: simulation of a large conformational transition in plasminogen activator inhibitor 1. Protein Sci. 10, 798–808. 40. Gils, A., Lu, J., Aertgeerts, K., Knockaert, I. & Declerck, P.J. (1997) Identification of positively charged residues contributing to the stability of plasminogen activator inhibitor 1. FEBS Lett. 415, 192–195. 41. Vleugels, N., Leys, J., Knockaert, I. & Declerck, P.J. (2000) Effect of stabilizing versus destabilizing interactions on plasminogen activator inhibitor-1. Thromb. Haemost. 84, 871–875. 42. Sharp, A.M., Stein, P.E., Pannu, N.S., Carrell, R.W., Berkenpas, M.B., Ginsburg, D., Lawrence, D.A. & Read, R.J. (1999) The active conformation of plasminogen activator inhibitor 1, a target for drugs to control fibrinolysis and cell adhesion. Structure 7, 111–118. 43. Lawrence, D.A., Berkenpas, M.B., Palaniappan, S. & Ginsburg, D. (1994) Localization of vitronectin binding domain in plasminogen activator inhibitor-1. J. Biol. Chem. 269, 15223–15228. 44. Stout, T.J., Graham, H., Buckley, D.I. & Matthews, D.J. (2000) Structures of active and latent PAI-1: a possible stabilizing role for chloride ions. Biochemistry 39, 8460–8469. 45. Juhan-Vague, I. & Alessi, M.C. (1996) Fibrinolysis and risk of coronary artery disease. Fibrinolysis 10, 127–136. 46. Andreasen, P.A., Kjøller, L., Christensen, L. & Duffy, M.J. (1997) The urokinase-type plasminogen activator system in cancer metastasis: a review. Int. J. Cancer 72, 1–22. 47. Andreasen, P.A., Egelund, R. & Petersen, H.H. (2000) The plasminogen activation system in tumor growth, invasion, and metastasis. Cell. Mol. Life Sci. 57, 25–40. 48. Debrock, S. & Declerck, P.J. (1997) Neutralization of plasminogen activator inhibitor-1 inhibitory properties: identification of two different mechanisms. Biochim. Biophys. Acta 1337, 257–266. 49. Bjo ¨ rquist, P., Ehnebom, J., Inghardt, T., Hansson, L., Lindberg, M., Linschoten, M., Stro ¨ mqvist, M. & Deinum, J. (1998) Identification of the binding site for a low-molecular-weight inhibitor of plasminogen activator inhibitor type 1 by site-directed mutagen- esis. Biochemistry 37, 1227–1234. 50. Friederich, P.W., Levi, M., Biemond, B.J., Charlton, P., Templeton, D., van Zonneveld, A.J., Bevan, P., Pannekoek, H. & ten Cate, J.W. (1997) Novel low-molecular-weight inhibitor of PAI-1 (XR5118) promotes endogenous fibrinolysis and reduces postthrombolysis thrombus growth in rabbits. Circulation 96, 916 – 921. 51. Wind, T., Jensen, M.A. & Andreasen, P.A. (2001) Epitope mapping for four monoclonal antibodies against human plasminogen activator inhibitor type-1: implications for antibody-mediated PAI-1-neutralization and vitronectin-binding. Eur. J. Biochem. 268, 1095–1106. 52. Egelund, R., Einholm, A.P., Pedersen, K.E., Nielsen, R.W., Christensen, A., Deinum, J. & Andreasen, P.A. (2001) A regulatory hydrophobic area in the flexible joint region of plasminogen activator inhibitor-1, defined with fluorescent activity-neutralizing ligands. Ligand-induced serpin polymerization. J. Biol. Chem. 276, 13077–13086. q FEBS 2001 PAI-1 shutter region mutations (Eur. J. Biochem. 268) 6283 . Importance of the amino-acid composition of the shutter region of plasminogen activator inhibitor-1 for its transitions to latent and substrate forms Martin. (1992) Identification of a conformationally distinct form of plasminogen activator inhibitor-1, acting as a noninhibitory substrate for tissue- type plasminogen activator.