Protease nexin 1 is a potent urinary plasminogen activator inhibitor in the presence of collagen type IV.

Protease nexin 1 (PN1) in solution forms inhibitory complexes with thrombin or urokinase, which have opposing effects on the blood coagulation cascade. An initial report provided data supporting the idea that PN1 target protease specificity is under the influence of collagen type IV (1). Although collagen type IV demonstrated no effect on the association rate between PN1 and thrombin, the study reported that the association rate between PN1 and urokinase was allosterically reduced 10-fold. This has led to the generally accepted idea that the primary role of PN1 in the brain is to act as a rapid thrombin inhibition and clearance mechanism during trauma and loss of vascular integrity. In studies to identify the structural determinants of PN1 that mediate the allosteric interaction with collagen type IV, we found that protease specificity was only affected after transient exposure of PN1 to acidic conditions that mimic the elution protocol from a monoclonal antibody column. Because PN1 used in previous studies was purified over a monoclonal antibody column, we propose that the allosteric regulation of PN1 target protease specificity by collagen type IV is a result of the purification protocol. We provide both biochemical and kinetic data to support this conclusion. This finding is significant because it implies that PN1 may play a much larger role in the modeling and remodeling of brain tissues during development and is not simply an extravasated thrombin clearance mechanism as previously suggested.

Protease nexin 1 (PN1) in solution forms inhibitory complexes with thrombin or urokinase, which have opposing effects on the blood coagulation cascade. An initial report provided data supporting the idea that PN1 target protease specificity is under the influence of collagen type IV (1). Although collagen type IV demonstrated no effect on the association rate between PN1 and thrombin, the study reported that the association rate between PN1 and urokinase was allosterically reduced 10-fold. This has led to the generally accepted idea that the primary role of PN1 in the brain is to act as a rapid thrombin inhibition and clearance mechanism during trauma and loss of vascular integrity. In studies to identify the structural determinants of PN1 that mediate the allosteric interaction with collagen type IV, we found that protease specificity was only affected after transient exposure of PN1 to acidic conditions that mimic the elution protocol from a monoclonal antibody column. Because PN1 used in previous studies was purified over a monoclonal antibody column, we propose that the allosteric regulation of PN1 target protease specificity by collagen type IV is a result of the purification protocol. We provide both biochemical and kinetic data to support this conclusion. This finding is significant because it implies that PN1 may play a much larger role in the modeling and remodeling of brain tissues during development and is not simply an extravasated thrombin clearance mechanism as previously suggested.
Protease nexin 1 (PN1) 1 is a 43-kDa member of the serine protease inhibitor (SERPIN) superfamily and has been shown to be a potent physiologic inhibitor of thrombin and urinary plasminogen activator (urokinase) (2,3). PN1 is synthesized by astrocytes, smooth muscle, endothelial cells, and fibroblasts (3,4) and is the only SERPIN found in physiologic quantities in the brain (5)(6)(7)(8)(9). Although most SERPINs are found in the plasma, PN1 is found primarily in tissues and has been shown to be associated with the extracellular matrix (10). Previous studies have shown that thrombin is able to induce apoptosis in astrocytes (11)(12)(13), and thus the localization of PN1 in the tissues surrounding the vasculature of the brain has led to a hypothesis that one of the primary physiological functions of PN1 is to act as a protective mechanism against thrombin that has escaped the vascular compartment during cerebrovascular trauma. This hypothesis has been supported by reports demonstrating that PN1 bound to collagen type IV is refractory to urokinase inhibition, thus leaving PN1 free to react with thrombin in the tissues. In the present report we show that this is due to the exposure of PN1 to acid during the purification protocol and that the protease inhibitory specificity of native PN1 is not altered by collagen type IV binding.
Like many other SERPINs, the rate of PN1 inhibition is under the influence of allosteric effectors. Cell surface proteoglycans are allosteric effectors for many SERPINs, including heparin cofactor II, antithrombin III, and PN1 (2,14,15). Interestingly, in the presence of heparin, PN1 increases its inhibitory activity toward thrombin by a factor of 1000, making PN1 a more efficient inhibitor of thrombin than even antithrombin III (2). Along with the proteoglycan-mediated acceleration there is also evidence of protein-protein interactions modifying the rate of SERPIN inhibitory action. A well characterized example of this is plasminogen activator inhibitor 1, a SERPIN with inhibitory activity toward plasminogen activators. Plasminogen activator inhibitor 1 becomes an effective inhibitor of thrombin only in the presence of the plasma borne cell adhesion protein vitronectin, increasing its rate of thrombin inhibition by a factor of 200 (16).
Collagen type IV helps form the core of basement membranes, the sheet-like extracellular structure that surrounds tissues and organs (17), and has been reported to be an allosteric effector of PN1 (1,19). Whereas other forms of collagen are formed through interactions of their noncollagenous ends that result in long fibers with high tensile strength, collagen type IV forms a meshed, nonfibrillar network with interactions not only at the ends but also within the long collagenous triple helical region (17,18). PN1 has been shown to colocalize with a component of the basement membrane, fibronectin (10). Further, PN1 secreted from cultured fibroblasts has been shown to copurify with a 120-kDa band that has been identified as collagen type IV (1). Studies on PN1 binding to fibroblast-secreted extracellular matrix led to the discovery that PN1 has altered inhibitory activity toward target proteases when bound to the extracellular matrix (19). It was later determined that collagen type IV was the component of the extracellular matrix that was causing this target protease specificity change in PN1. In the absence of collagen type IV, PN1 was able to form inhibitory complexes with thrombin, urokinase, and plasmin; however, when collagen type IV was added at a concentration of 1 M, PN1 was no longer able to form complexes at the same rate with urokinase and plasmin (1,19).
The original purpose of this study was to identify the specific structural determinants in PN1 that mediate the alteration of its target protease specificity when in complex with collagen type IV. However, we had no success reproducing the alteration of PN1 target protease specificity through interaction with collagen type IV. What we found was that the collagen type IV allosteric regulation of PN1 target protease specificity was the result of exposing the SERPIN to acidic conditions that were required for the purification of PN1 from a monoclonal antibody column. Because all of the studies to date on the regulation of PN1 target protease specificity with collagen type IV have used PN1 that was purified on a monoclonal antibody column, we propose that the observed allosteric effect is the result of acid exposure of PN1 during the monoclonal antibody column purification protocol and thus may not be physiologically relevant. In all of our previous studies we have used a heparin-Sepharose affinity chromatography protocol to purify PN1. We show here that there is no regulation of PN1 target protease specificity by collagen type IV and provide strong evidence that the allosteric effect of collagen type IV is the result of exposure to nonphysiological conditions required for the elution of PN1 from a monoclonal antibody column. Thus, PN1 is still a potent urokinase inhibitor in the basement membrane when bound to collagen type IV. These data are very important because they re-open the possibility that PN1 may play a key regulatory role in the urokinase receptor-mediated activation of signaling pathways that control cell migration during tissue repair/remodeling.

EXPERIMENTAL PROCEDURES
Materials-The cell culture media and reagents were purchased from Irvine Scientific and JRH Scientific. The cell culture plastics were from Corning. Thrombin (3,000 NIH units/mg) high molecular mass urokinase (80,000 IU/mg), and plasmin (27 units/mg protein) were purchased from Calbiochem. Collagen type IV was purchased from Calbiochem and from Sigma (the Calbiochem samples were used in all of the experiments shown here). High trap heparin-Sepharose columns were from Amersham Biosciences. Enzymes used for molecular biology were from Fisher/Promega or Invitrogen. All of the other common laboratory supplies were purchased from either Irvine Scientific or Sigma.
Cell Culture and Extracellular Matrix Preparation-Human foreskin fibroblasts were grown and maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum as described previously (20 -22). The experimental cultures were seeded at 1.0 ϫ 10 5 cells/well into 24-well plates and grown to confluency in the presence of 50 g/ml ascorbic acid. Before use, the cells were transferred to serum-free medium for 48 h. For fixation of the cells and extracellular matrix, the cells were washed with PBS four times before incubation for 30 min in the presence of 0.25 M NH 4 OH at room temperature. The extracellular matrix was then rinsed five times with phosphate-buffered saline before use.
Protein Radioiodination-125 I-Thrombin and 125 I-urokinase were prepared as described previously using the Iodogen method (1,23). Specific activities for 125 I-thrombin ranged from 8,000 to 15,000 cpm/ng of protein. Specific activities for 125 I-urokinase ranged from 5,000 to 10,000 cpm/ng of protein.
PN1 Purification, Active Site Titration, and Kinetic Studies-The purification protocol for PN1 is described in detail elsewhere (24). Briefly, 72 h serum-free conditioned medium was obtained from cultures of human foreskin fibroblast cells grown on microcarrier beads. After the addition of protease inhibitors, the medium was concentrated from 2 liters to 200 ml and recirculated overnight at 4°C over a 5-ml heparin-Sepharose affinity column. Following a 50-ml wash with PBS, the column was eluted with a 100-ml 0.2-1.0 M linear gradient of NaCl. Fractions containing PN1, as determined by a complex formation assay with thrombin, were pooled and dialyzed exhaustively against PBS to equilibrate the final NaCl concentration at 0.15 M. Samples of the purified and dialyzed PN1 were titrated with active thrombin to determine the percentage of activity. 30 ng of thrombin were added to various amounts of PN1, which were then diluted to a final total volume of 80 l in PBS, pH 7.2, containing 0.1% bovine serum albumin. At the end of a 30-min incubation at 37°C, the reactions were chilled on ice for 15 min, followed by the addition of a 200-fold molar excess of Chromozym-Th. The reactions were returned to room temperature for 30 min to allow for color development as a measure of residual thrombin activity. The absorbance measurements were taken at 405 nm to quantify activity colorimetrically.
The kinetic measurements were done using a protocol previously described (2). Briefly, approximately a 25-fold molar excess PN1 (8 nM) was preincubated with urokinase (0.3 nM) in a final volume of 100 l of Tyrode's buffer. At the indicated times, the reaction was diluted with Tyrode's buffer to a final volume of 2 ml containing 100 M Z-Gly-Gly-Arg-AMC, a quenched fluorogenic substrate of urokinase. Urokinase that was not inactivated by PN1 during the preincubation is free to cleave the fluorogenic substrate, providing a quantitative measurement of the amount of active urokinase remaining. The change in relative fluorescence was monitored over a 480-s period with a sampling frequency of 20 s. The samples were excited at 350 nm, and emission was monitored at 450 nm. The k assoc rates were calculated using the following equation (3).
Transient Exposure of PN1 to Acidic Conditions-PN1 elution from a monoclonal antibody column transiently exposes it to a strong acid at pH 3.0 (1). To mimic these conditions, we exposed samples of PN1 purified on a heparin-Sepharose column to 100 mM glycine HCl, pH 3.0, for 1 min, followed by neutralization by the addition of 200 mM NaPO 3 , pH 7.4. Longer exposure of PN1 to these conditions (15-30 min) resulted in an 80% loss of PN1 inhibitory activity (data not shown). PN1 exposed to acid for 1 min was dialyzed into PBS and then evaluated using the linkage assay described above Collagen Type IV Inhibition Assay-The protocol for this assay is described in detail elsewhere (1) and is summarized here. Iodinated protease at 20 nM was incubated with 5 nM PN1 with no collagen type IV or in the presence of increasing concentrations of collagen type IV for 15 min. The reactions were terminated by the addition of SDS-PAGE protein sample buffer. The samples were analyzed by SDS-PAGE on 10% polyacrylamide gels. The gels were exposed to a Bio-Rad phosphorus imaging screen for 30 min, and the digitized image was prepared using a Bio-Rad GS-250 molecular imager.
Baculovirus Protein Expression-The methods used to prepare and express recombinant forms of PN1 in a baculovirus-driven protein expression system have been described in detail elsewhere and are briefly summarized here (25). The gene for PN1 was subcloned using the following forward and reverse primers to engineer overhanging restriction enzyme sites, SmaI and EcoR1: forward, 5Ј-GAC CCC GGG ATG AAC TGG CAT CTC CCC-3Ј, and reverse 5Ј-GAC GAA TTC TCA TCA GGG TTT GTT TAT CTG CCC C-3Ј. The PCR product was digested using EcoRI and SmaI, ligated into the pVL1393 baculovirus shuttle vector, and cotransfected into Sf9 insect cells along with BaculoGold baculovirus. Recombinant viruses were purified by a single round of plaque purification. For protein expression, Sf9 cells grown in T175 flasks were infected at a multiplicity of infection of 10:1. Five to seven days later the medium was harvested, centrifuged to remove dead cells, and concentrated using an Amicon concentrator with a molecular mass cut-off of ϳ10 kDa. The PN1 sample was then dialyzed against PBS using dialysis tubing with a cut-off of 12 kDa and further concentrated within the dialysis tubing using Calbiochem Aquacide. The PN1 was finally brought to a concentration similar to the human foreskin fibroblast cell-purified sample, and its activity was assayed as described above.

PN1 Forms Complexes with Urokinase in the Presence of
Collagen Type IV-Previous work indicates that in the presence of 1 M collagen type IV, PN1 was able to form inhibitory complexes with thrombin but became refractory to formation of inhibitory complexes with either plasmin or urokinase (1). To verify this assay, PN1 (5 nM) was incubated with either 125 Ithrombin (20 nM) or 125 I-urokinase (20 nM) in the presence of increasing concentrations of collagen type IV, and the amount of complex formation was quantified as described under "Experimental Procedures." At all concentrations of collagen type IV, PN1 was able to form inhibitory complexes with thrombin, a result that is consistent with previous work (Fig. 1A). However, our results with urokinase disagreed with previous results (1). We show that using PN1 purified according to the protocol described under "Experimental Procedures," PN1 was able to form inhibitory complexes with urokinase with the same efficiency independent of the presence or absence of the collagen type IV (Fig. 1B). These results conflict with the data presented in the previous study, which demonstrate reduced ability of PN1 to form inhibitory complexes with urokinase in the presence of 1 M collagen type IV. The experiment was repeated using concentrations up to 100 M collagen type IV, but PN1 continued to form complexes with urokinase at the same levels as the control that had no collagen type IV (data not shown).
Fibroblast-secreted Extracellular Matrix Does Not Alter PN1 Target Protease Specificity-To determine whether the commercially available collagen type IV might be inactive and thus account for the disparity in results, we turned to a different protocol using fixed extracellular matrix preparations as a source of collagen type IV. Previous studies by the same group demonstrated that in the presence of collagen type IV-rich extracellular matrix derived from ascorbate-induced fibroblasts, PN1 was also refractory to urokinase and plasmin inhibition (19). Repeating these protocols exactly gave the same result we obtained with commercially available collagen type IV; no augmentation of PN1 protease target specificity (data not shown). We next pursued a third line of experimentation to resolve the discrepancy in data.
The Rate of PN1-Urokinase Complex Formation Is Not Changed by the Presence of Collagen Type IV-Previous work established that the rate of urokinase inhibition by PN1 in the absence of collagen type IV was 1.50 ϫ 10 Ϫ5 M Ϫ1 s Ϫ1 (1). In that study, the presence of 1.0 M collagen type IV caused a 10-fold decrease in the pseudo-second order rate constant (k assoc ) to ϳ1.50 ϫ 10 Ϫ4 M Ϫ1 s Ϫ1 . To examine the effect of collagen type IV on the rate of inhibitory complex formation, PN1 (5 nM) was incubated with 125 I-urokinase (20 nM) in the presence and absence of 1 M collagen type IV and assayed for complex formation as described under "Experimental Procedures." As shown in Fig. 2, we were unable to detect any difference in the formation of inhibitory complexes between PN1 and urokinase in the presence and absence of 1 M collagen type IV. At all time points, there are equal amounts of complexes between PN1 and 125 I-urokinase independent of the presence or absence of collagen type IV.
Transient Acid Exposure of PN1 Confers Allosteric Regulation of Inhibitory Specificity by Collagen Type IV-In an effort to understand the disparity between our data and those previously published, we turned to any potential biochemical differences between the PN1 preparations used in our present studies and those of the published studies (1,19). It came to our attention that the PN1 used in the previous studies on target protease regulation by collagen type IV was purified using a monoclonal antibody column, whereas a heparin affinity column was used to purify PN1 for the present studies. During the elution of PN1 from the monoclonal antibody column, it was transiently exposed to a strong acid at pH 3.0. To test whether this acid exposure resulted in the altered protease specificity of PN1 in the presence of collagen type IV, we exposed a sample of PN1 purified on a heparin-Sepharose column to 100 mM glycine HCl, pH 3.0, for 1 min, followed by neutralization by the addition of 200 mM NaPO 3 , pH 7.4. Longer exposure of PN1 to these conditions (15-30 min) resulted in an 80% loss of PN1 inhibitory activity (data not shown). PN1 exposed to acid for 1 min was then evaluated using the linkage assay described above. We first conducted a control experiment to assess the ability of the acid-exposed PN1 to form complexes with 125 Ithrombin in the presence and absence of collagen type IV. As expected, the data in Fig. 3A demonstrate no significant difference in the quantity of 125 I -thrombin-PN1 complexes formed regardless of the concentration of collagen type IV present. This important control shows, however, that the mere exposure of PN1 to acid does not affect its ability to act as a SERPIN. The acid-exposed PN1 was also able to form complexes with 125 Iurokinase in the absence of collagen type IV but, importantly, became refractory to complex formation in the presence of increasing concentrations of collagen type IV (Fig. 3B). A comparison of the data in Figs. 3B and 1B clearly shows that the acid exposure of PN1 confers allosteric regulation by collagen IV at the level of protease-inhibitor complex formation.
Because we could now reproduce the allosteric regulation of PN1 by collagen type IV after transient acid exposure using a complex formation assay, we proceeded to a more quantitative assessment of the kinetics of the inhibitory reaction between PN1 and urokinase using a fluorometric substrate approach  2. Collagen type IV does not alter the formation of inhibitory complexes between PN1 and urokinase. PN1 (5 nM) was incubated with 125 I-urokinase (20 nM) in the absence or the presence of 1.0 M collagen type IV for the times indicated. After incubation, the reaction was stopped by the addition of reducing SDS-PAGE protein sample buffer. The samples were run on a 10% SDS-PAGE gel and dried onto Whatman filter paper before exposure to the Bio-Rad phosphorus imaging system. (26). We took this approach for two reasons. First, in the previous studies that demonstrated the allosteric regulation of PN1 by collagen type IV, a similar substrate cleavage assay was employed. Secondly, the suicide-substrate mechanism can be broken down into discrete steps as shown below (32).
The inhibition of the protease occurs as an early step in the reaction and corresponds to the measured k assoc . The completion of the reaction and thus the formation of the covalently linked complex, IP* cpx , which is measured in the linkage assay, occurs more slowly. Transient exposure of PN1 to acid allowed us to reproduce the results of the linkage assay from the previous report, but this could result from two different steps in the reaction mechanism: either at the step of association between the inhibitor and the protease or at the step of procession to the acyl-linked end product (2). At submicromolar concentrations of reactants, a simple equation (Equation 1) provides a good measure of the association rate between PN1 and its target proteases (1). Urokinase (0.3 nM) was incubated with PN1 (8.1 nM) for the indicated times at 37°C in the presence and absence of collagen type IV, and residual urokinase activity was assayed by using a quenched fluorogenic urokinase substrate, Z-Gly-Gly-Arg-AMC, as described under "Experimental Procedures." Substrate cleavage progression curves for urokinase preincubated with PN1 from 2 to 20 min were measured using the quenched fluorogenic urokinase substrate as described above. The k assoc rates for PN1 and urokinase were nearly identical in the absence and presence of collagen type IV, 2.13 ϫ 10 5 M Ϫ1 s Ϫ1 Ϯ 0.12 and 2.04 ϫ 10 5 M Ϫ1 s Ϫ1 Ϯ 0.16, respectively (progression curve data not shown; the k assoc values are in Table I). These results are in close agreement with previously published values (1, 3). The same experiment was done after PN1 was transiently exposed to acidic conditions. The k assoc rate in the absence of collagen type IV was not changed significantly, 2.16 ϫ 10 5 M Ϫ1 s Ϫ1 Ϯ 0.22. In contrast, the inclusion of collagen type IV in the reaction containing acid-exposed PN1 markedly slowed the k assoc rate by a factor of nearly 10, 2.83 ϫ 10 4 M Ϫ1 s Ϫ1 Ϯ 0.21 (progression curve data not shown; the k assoc values are in Table I). This magnitude of decrease is very close to the previously published report demonstrating the allosteric regulation of PN1 by collagen type IV (1).
The potential biological significance of this data is best appreciated by the V/V 0 plot shown in Fig. 4. This is a plot of the data used to obtain the k assoc values and shows the fraction of active urokinase remaining at times after the addition of PN1. The fraction of active urokinase is reduced by 50% in the presence of native PN1 and acid-exposed PN1 in as little as 7 min. In contrast, urokinase activity was still above 80% after 30 min of incubation with acid-exposed PN1 in the presence of collagen Type IV. Finally, these kinetic data also demonstrate that the effect of collagen type IV on the acid-exposed PN1 is manifest early in the mechanism of the inhibitory reaction, at the level of k assoc .
The Lack of PN1 Allosteric Regulation by Collagen Type IV Cannot Be Attributed to Heparin Exposure during Purification-Because the native PN1 used in the present study was purified by heparin-Sepharose affinity chromatography, we wanted to rule out the possibility that the absence of allosteric regulation by collagen type IV might be the explained by a heparin-induced "conformational lock" of the PN1 during purification. The approach we used was expression of PN1 in Sf9 insect cells to high levels, followed by concentration of the medium that contained the secreted PN1 as previously described (25). The rationale behind this approach was that the PN1 would not be exposed to either a heparin purification step or an acid exposure under these conditions. It should also be noted that the insect cell culture medium does not contain heparin or collagen type IV. Sf9 cells were induced to generate recombinant PN1 through baculovirus infection, and medium was collected and concentrated to use in the collagen type IV assays. Control experiments indicated that there was no significant biochemical or kinetic difference between the baculovirus-expressed PN1 samples and the fibroblast-secreted PN1 that was purified using the heparin-Sepharose column protocol (data not shown).
We first examined the effect of transient acidification on the baculovirus-expressed PN1 using the linkage assay. 125 I-Thrombin and 125 I-urokinase were incubated with transiently acidified baculovirus-expressed PN1 in the presence of increasing concentrations of collagen type IV (Fig. 5). As expected, PN1 was able to form inhibitory complexes with thrombin at the same quantitative level independent of the presence of collagen type IV (Fig. 5A). Interestingly, the acid-exposed PN1 became refractory to formation of inhibitory complexes with urokinase in the presence of increasing concentrations of collagen type IV (Fig. 5B). The magnitude of the collagen-induced decrease in linkage formation was nearly identical to that seen in Fig. 3B using PN1 purified on heparin-Sepharose. These data clearly rule out the possibility that purification on hepa- FIG. 3. PN1 that has been transiently exposed to acid demonstrates an altered target protease specificity in response to collagen type IV. A, PN1 and 125 I-thrombin; B, PN1 and 125 I-urokinase. PN1 was transiently exposed to acidic conditions as described. Acid-exposed PN1 and 125 I-protease were incubated for 15 min for inhibitory complex formation under the conditions listed. After incubation, the reaction was stopped by the addition of reducing SDS-PAGE protein sample buffer. The samples were run on a 10% SDS-PAGE gel and dried onto Whatman filter paper before exposure to the Bio-Rad rin-Sepharose accounts for the lack of allosteric regulation of PN1 by collagen type IV and strongly support our contention that the observed regulation is conferred by transient acid exposure.
A quantitative evaluation of the kinetics of inhibition was next done using the fluorogenic substrate assay described earlier. As expected by the results of the linkage assay, the allosteric regulation of the acidified PN1 by collagen type IV occurs at the level of the protease inhibitor association rate. The calculated k assoc rate of 2.2 ϫ 10 5 M Ϫ1 s Ϫ1 Ϯ 0.14 for baculovirus PN1 transiently exposed to acid (Fig. 6A) is in good agreement with the k assoc rate obtained for native PN1 affinity purified using heparin (Table I). Similarly, the reduction of the k assoc rate for the transiently acidified PN1 in the presence of collagen type IV (Fig. 6B), 2.75 ϫ 10 4 M Ϫ1 s Ϫ1 Ϯ 0.49, is in close agreement the k assoc rate obtained for heparin purified PN1 in TABLE I Rate of association (k assoc ) between PN1 and urokinase Summary of kinetic data between PN1 and urokinase using a quenched fluorogenic substrate of urokinase, Z-Gly-Gly-Arg-AMC. The rate of association (k assoc ) was determined as described under "Experimental Procedures." Reported is the k assoc between PN1 and urokinase in the absence and presence of 1 M collagen type IV. Samples of PN1 that were transiently exposed to acidic conditions are indicated.  4. Collagen type IV alters the k assoc rate of interaction between acid-exposed PN1 and urokinase. The progression curves for the cleavage of the quenched fluorescent substrate specific for urokinase were obtained as described under "Results." The curves were obtained for the various combinations of PN1 and urokinase preincubated from 2 to 30 min prior to the addition of the quenched substrate as follows: PN1 and urokinase only, PN1 and urokinase in the presence of 1 M collagen type IV, acid-exposed PN1 and urokinase, and acidexposed PN1 and urokinase in the presence of 1 M collagen type IV. The concentrations of PN1 (7.6 nM) and urokinase (0.3 nM) were identical in all of the reactions. At the end of the preincubation, the reactions were diluted to 2 ml with 100 M quenched fluorogenic urokinase substrate. The change in relative fluorescence was monitored for each of the preincubation time periods over a 480-s monitoring phase. The progression curves were used to calculate the k assoc values that are discussed in the text and reported in Table I. The data has been plotted as a ratio of initial velocity at each of the preincubation times (V) to the initial velocity in the absence of the inhibitor (V 0 ). q, native PN1, no collagen; E, native PN1, 1 M collagen type IV; , acid-exposed PN1, no collagen; ƒ, acid-exposed PN1, 1 M collagen type IV. the presence of collagen type IV after transient acid exposure (Table I). Also included in Table I are the k assoc rates for urokinase and untreated baculovirus-expressed PN1 for comparison. The progression curves used to obtain the values are not shown. DISCUSSION The physiological role of PN1 in biological processes that are driven by proteases depends on three primary factors: protease specificity, protease affinity, and the site of localization. Protease specificity and affinity are closely related and have the additional complication that they may be augmented by other biological molecules depending on the location. A clear example of this is the inhibition of thrombin by PN1, which is accelerated to a diffusion-limited rate in the presence of heparin found in high concentrations in the extracellular matrix (2). What remains unclear, however, is what percentage of PN1 in a tissue is bound to heparin versus that which is free in solution. It is also not clear whether heparin is the only extracellular matrix molecule that may sequester PN1 and augment its activity. It was originally reported that the specificity of PN1 was allosterically regulated by binding to components in the extracellular matrix of human fibroblasts (19). In those studies it was shown that matrix-bound PN1 was a potent thrombin inhibitor but became refractory as an inhibitor of urokinase and plasmin (19). Subsequent studies identified collagen type IV in the matrix as the molecule responsible for the allosteric regulation of PN1 (1). As previously reported, this apparent regulation is manifest in a lowering of the k assoc between PN1 and urokinase from 1.5 ϫ 10 5 M Ϫ1 s Ϫ1 to 1.5 ϫ 10 4 M Ϫ1 s Ϫ1 (1). This is a very important concept as to how the biological functions of PN1 are perceived. This apparent loss of inhibitory activity toward urokinase and plasmin in the extracellular matrix would suggest a less important role for PN1 in the regulation of cell migration, in matrix turnover, and in the overall process of tissue development and remodeling where urokinase and plasmin are involved. The results of the present studies demonstrate that in fact the inhibitory activity of PN1 toward urokinase is not altered by binding to collagen type IV under normal physiological conditions and reaffirm the potentially important role for PN1 in processes where the proteolytic activity of urokinase is involved.
The original goal of the present study was to identify the domain in PN1 that mediates binding to collagen type IV and to understand the mechanism underlying the allosteric regulation. Our inability to reproduce the original results demonstrating the allosteric regulation of PN1 by collagen type IV led to a re-examination of every component of the experimental system including protein sources and purification procedures. The PN1 used in the previous studies on PN1-collagen interactions was affinity purified using a monoclonal antibody column and eluted with transient exposure to low pH (1,19). The PN1 used in the present studies, however, was purified by heparin-Sepharose chromatography and eluted with 0.6 M NaCl (24). Three independent lines of evidence obtained in the present studies demonstrate that the transient exposure of PN1 to acid pH is required for allosteric regulation by collagen type IV. First, a 1-min acid exposure of PN1 purified on heparin-Sepharose, which was unaffected by the addition of collagen type IV prior to acid exposure, conferred the ability of the PN1 to be allosterically regulated by collagen type IV. This was determined using the same linkage assay previously described, and the collagen dose-response curves were nearly identical to those previously published (1). Second, in kinetic assays using a quenched fluorogenic peptide substrate, we determined quantitatively that the effect of the collagen type IV on the inhibition of urokinase by PN1 was due to a decrease in the association rate. This was also shown in the previous studies, and the rate constants we determined were nearly identical to theirs (1). Finally, and perhaps most importantly, we show that purification of PN1 using heparin-Sepharose does not play a role in the PN1-collagen type IV interactions. To remove the possibility that exposure to immobilized heparin during our purification protocol imposes a "conformational lock" on PN1 that is removed by acid exposure, we obtained the same data in experiments using recombinantly expressed PN1 that had never been subjected to purification. These three lines of evidence provide compelling support for the idea that indeed it is the acid exposure of PN1 that confers susceptibility to collagen type IV allosteric regulation.
This study does not rule out the possibility that the allosteric regulation of PN1 by collagen type IV might occur under specific sets of circumstance in vivo. PN1 is abundantly produced by kidney tubule epithelial cells, where it is exposed to the acid content of urine (27). In addition, little is known about extracellular changes in pH during ischemia in neuronal tissue FIG. 6. Acid-exposed baculovirus PN1 shows a 10-fold reduced k assoc rate with urokinase in the presence of 1 M collagen type IV. A, PN1 and urokinase; B, PN1 and urokinase in the presence of 1 M collagen type IV. Baculovirus PN1 was expressed and processed as described under "Experimental Procedures." The PN1 samples were then transiently exposed to acidic conditions to mimic the elution protocol of a monoclonal antibody column. Acid-exposed PN1 (7.6 nM) and urokinase (0.3 nM) were preincubated for the times indicated in the absence or presence of collagen type IV. At the end of the preincubation, the reaction was diluted to 2 ml with 100 M quenched fluorogenic urokinase substrate. Reported is the change in relative fluorescence versus time for each of the preincubation time periods over a 480-s monitoring phase. The preincubation times for 6A/6B are: 0/0 min (q), where PN1 is also very abundant (3,4). It may also be that the exposure of PN1 to acidic conditions mediates a conformational change in PN1 normally mediated by another component of the extracellular matrix, although no such components have been described. The present studies do suggest that under most physiological circumstances PN1 is a potent inhibitor of urokinase and plasmin and not subject to allosteric regulation by collagen type IV. Because PN1 is known to be concentrated at the cell surface by binding to heparin sulfate proteoglycans and the action of urokinase bound to the urokinase receptor is directly tied to cell migration in certain cell types (28 -31), the role of PN1 in the regulation of this process could be crucial. PN1 should once again be considered as a potentially important factor in cellular processes in which the proteolytic activity of urokinase and plasmin play a role.