The role of the finger and growth factor domains in the clearance of tissue-type plasminogen activator by hepatocytes.

The relative contribution of the finger/growth factor domains of tissue-type plasminogen activator (t-PA) and of the other t-PA domains to the clearance of t-PA by hepatocytes was investigated. A recombinant finger/growth factor construct inhibited t-PA and t-PA/plasminogen activator inhibitor type-1 degradation with an IC of 1800 nM, whereas a t-PA mutant lacking the finger and growth factor domains inhibited degradation with an estimated IC of 1200 nM. In comparison the IC of t-PA was found to be approximately 10 nM. Clearance of t-PA by human or rat hepatoma cells was not inhibited by high concentrations of fucose (50 mM), which suggests that the fucose on Thr-61 is not involved in clearance by these cells. These results suggest that the binding of t-PA involves several low affinity binding sites located on distinct domains of the t-PA molecule.

The relative contribution of the finger/growth factor domains of tissue-type plasminogen activator (t-PA) and of the other t-PA domains to the clearance of t-PA by hepatocytes was investigated. A recombinant finger/ growth factor construct inhibited t-PA and t-PA/plasminogen activator inhibitor type-1 degradation with an IC 50 of 1800 nM, whereas a t-PA mutant lacking the finger and growth factor domains inhibited degradation with an estimated IC 50 of 1200 nM. In comparison the IC 50 of t-PA was found to be approximately 10 nM. Clearance of t-PA by human or rat hepatoma cells was not inhibited by high concentrations of fucose (50 mM), which suggests that the fucose on Thr-61 is not involved in clearance by these cells.
These results suggest that the binding of t-PA involves several low affinity binding sites located on distinct domains of the t-PA molecule.
Tissue-type plasminogen activator (t-PA) 1 is responsible for the degradation of intravascular fibrin deposits. Its plasma activity is regulated by the rate of its release from the vascular endothelium, its inhibition by plasminogen activator inhibitor type-1, and by its rapid hepatic clearance. The latter is mediated to a large extent by the low density lipoprotein receptor related protein (LRP) (1)(2)(3)(4). Free and PAI-1-complexed t-PA bind to LRP at or near the second cluster of eight complementtype cysteine-rich repeats (5). The region of t-PA mediating the interaction with LRP has not yet been identified with certainty. Deletion of the finger and growth factor domains leads to a reduced rate of plasma clearance, suggesting a role for one or both of these domains (6 -8). Also, mutation of Tyr-67 in the growth factor domain (9) affects t-PA clearance. However, some mutations or deletions in other parts of the t-PA molecule also lead to a diminished rate of clearance (10), suggesting that t-PA contains more than one receptor binding site or that the loss of binding to the clearance receptor after deletion of one domain is due to conformational modifications. A more precise approach to identify the receptor binding domain(s) would be to determine the binding affinity of individual domains or domain clusters.
The aim of the present work was to determine the relative contribution of the finger/growth factor domains and of the kringle 1/kringle 2/protease domains to the interaction of t-PA with its clearance receptor. To this end, the inhibitory effects of a t-PA mutant lacking the finger and growth factor domains (t-PA⌬FG) and of a recombinant finger/growth factor construct (FG) on the binding and/or degradation of free and PAI-1 complexed t-PA by hepatocytes were investigated. The results show that the finger and/or the growth factor domains interact with the clearance receptor and that additional binding sites may be located elsewhere on the t-PA molecule.

t-PA and t-PA Deletion Mutants
Recombinant t-PA (Actilyse®) was provided by Dr. J. Krause (Dr. K. Thomae GmbH, Biberach an der Riss, Federal Republic of Germany); dilutions of t-PA were made in 0.2 M L-arginin, 0.110 M phosphate, 0.01% Tween 80, pH 7.2. Chinese hamster ovary cells, stably expressing a t-PA mutant lacking the finger and growth factor domains (t-PA⌬FG), were provided by Dr. L. Nelles (Leuven, Belgium) (11). The mutant protein was produced and purified as described previously (2). A recombinant t-PA finger/growth factor construct (FG) produced in yeast, containing residues 1-91, was a gift from Dr. T. Dudgeon (British Bio-Technology, Oxford, United Kingdom); dilutions of FG were made in phosphate-buffered saline (composed of: NaCl (8 g/liter), KCl (0.2 g/liter), NaH 2 PO 4 .2 H 2 O (1.44 g/liter), KH 2 PO 4 (0.2 g/liter), pH 7.4). The FG protein contained no fucose on Thr-61 and was mutated at position 83 (Cys Ͼ Ser) to prevent dimerization of the protein. The secondary structure of FG has been described (12).

Labeling of Proteins
t-PA was iodinated using the Iodogen method (Pierce) to specific activities of 3-4 Ci/g (1 Ci/15 pmol). Complexes of labeled t-PA with PAI-1 (provided by Dr. T. Reilly, DuPont Merck Pharmaceutical Co.) were prepared as described previously (2).

Binding and Degradation Assays on Rat and Human Hepatoma Cells
Novikoff Cells-The binding and degradation assays on rat hepatoma cells were performed as described previously (2).
HepG2 Cells in Suspension-Confluent cells were washed with HEPES-buffered saline (21 mM HEPES, 137 mM NaCl, 5 mM KCl, 0.76 mM Na 2 HPO 4 , 5.55 mM glucose, pH 7.4) and detached by 20-min incubation at room temperature with the same buffer containing 5 mM EDTA. Cells were harvested, washed twice, and resuspended (at 4 ϫ 10 6 cells/ml) in HEPES-buffered saline containing 3 mM CaCl 2 , 1 mM * This work was supported by a grant from the Swiss National Fund for Scientific Research (31-40889.94) and by the Foundation for Research on Atherosclerosis and Thrombosis. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Binding to Human Liver Membranes
The effect of competitors on the binding of 125 I-t-PA to liver membranes was determined as described previously (13).

Analysis of Inhibition Curves
The concentration of competitor needed to block ligand degradation by 50% (IC 50 ) was calculated from the inhibition curves. The data points were fitted to the equation "y ϭ {(100 Ϫ b)/((C/IC 50 ) ϩ 1)} ϩ b" using the Ultrafit program (Biosoft, Cambridge, United Kingdom), where y ϭ degradation in the presence of competitor divided by degradation in the absence of competitor ϫ 100%, C ϭ concentration of competitor, and b ϭ noninhibitable degradation (on average 8.7%).

RESULTS
Previously we observed that t-PA⌬FG at 100 nM had little or no effect on the binding to and degradation by rat hepatoma cells of free and PAI-1 complexed 125 I-t-PA (2). This suggested binding via the finger or growth factor domain. We therefore investigated the effect of a recombinant finger/growth factor construct on the degradation of 125 I-t-PA and 125 I-t-PA/PAI-1 by Novikoff cells. Results indicated that FG is able to completely block degradation of free 125 I-t-PA (Fig. 1A), but its IC 50 was 2 orders of magnitude higher than that for recombinant t-PA (apparent IC 50 of 1800 nM versus 10.9 nM). The degrada-tion of 125 I-t-PA/PAI-1 was fully inhibited by t-PA, with an apparent IC 50 of 21.4 nM, and partially by FG with an apparent IC 50 of 1500 nM and an uninhibitable part of 125 I-t-PA/PAI-1 degradation of 26% (Fig. 1B).
FG inhibited the binding of 125 I-t-PA to liver membranes prepared from normal human liver, with an apparent IC 50 of 2400 nM and an uninhibitable part of binding of 15%, whereas t-PA inhibited binding by 50% at a concentration of 23 nM (Fig.  2). Receptor-associated protein (RAP), an inhibitor of the binding of t-PA to its hepatic clearance receptor, LRP (2, 4), also fully inhibited t-PA binding to human liver membranes (IC 50 ϭ 6 nM).
The low affinity of the FG construct led us to re-evaluate the effect on t-PA degradation by Novikoff cells of t-PA⌬FG concentrations up to 1 M, the maximal concentration at which t-PA⌬FG was soluble at the experimental conditions. At 1 M a partial inhibition of t-PA degradation was observed. Curve fitting suggested an IC 50 of 1200 nM (Fig. 3).
To determine whether the finger/growth factor part of t-PA could cooperate with the remainder of the t-PA molecule (t-PA⌬FG) in binding to the clearance receptor, we performed clearance inhibition experiments in the presence of both molecules. At concentrations of 350 nM, the inhibitory effect of the combination of t-PA⌬FG and FG was only slightly greater than that of either competitor alone (Table I).
One study observed that 50 mM fucose blocked the binding and degradation of t-PA by HepG2 cells in suspension (14), suggesting that the fucose on Thr-61 is involved. As the FG construct used in the present study lacks this fucose, we studied the effect of fucose (up to 50 mM) on t-PA binding and degradation by rat hepatoma cells; no effect was observed (Fig.  4). We also studied the effect of different concentrations of D-fucose on t-PA degradation by adherent HepG2 cells and by HepG2 cells in suspension. At 50 mM D-fucose, the highest concentration which was not cytotoxic, we observed only a modest decrease in t-PA degradation: 88% of control values for HepG2 cells in suspension and 89% for adherent HepG2 cells at 4-h incubation. DISCUSSION The present study was undertaken to identify the domains on the t-PA molecule that interact with its clearance receptor on rat hepatoma cells and on human liver membranes. The strong inhibitory effect of RAP on t-PA binding to human liver membranes suggests that LRP is the principal specific binding site of t-PA on liver membranes. The hepatoma cell model and the human liver membrane model thus appear to address the same receptor system.
For the competition experiments two complementary constructs were used. The FG construct was chosen because its secondary structure is known (12). The t-PA⌬FG construct represents the remainder of the t-PA molecule. The FG construct completely inhibited t-PA degradation by Novikoff cells and binding to human liver membranes, but at concentrations 2 orders of magnitude higher than for t-PA (degradation by Novikoff cells: IC 50 of 1800 for FG versus 10.9 nM for t-PA; binding to liver membranes: 2400 nM versus 23 nM for t-PA). Similarly most of t-PA/PAI-1 degradation by Novikoff cells could be inhibited by FG (IC 50 of 1500 nM versus 21.4 nM for t-PA), but part of t-PA/PAI-1 degradation was not inhibitable by FG, which suggests the presence of binding sites on the t-PA/PAI-1 complex not involving the FG domains. Taken together, these results, as well as our previous observation that monoclonal antibodies to the growth factor domain inhibit t-PA clearance by hepatoma cells (13), provide clear evidence that the finger and/or the growth factor domains interact with the t-PA clearance receptor. However, the poor affinity of the finger/growth factor construct suggests that: 1) other domains of t-PA contribute to binding (see below), 2) the interaction of finger and growth factor domains with other t-PA domains (15) is important for binding of these domains to the clearance receptor, or 3) essential posttranslational modifications were not made in yeast.
The consistent inhibitory effect of high concentrations of t-PA⌬FG on t-PA degradation suggests that this part of the molecule contains a low affinity binding site. The results, however, should be interpreted with caution. For a precise determination of IC 50 it is essential to employ t-PA⌬FG concentrations well above the estimated IC 50 of 1200 nM. However, under the conditions of in vitro degradation of t-PA by hepatoma cells, precipitation of t-PA⌬FG was observed above 1 M. Thus, the IC 50 of 1200 nM should be considered a preliminary estimate rather than a definitive value. Our previous observation that monoclonal antibodies to the kringle 2 domain interfered with t-PA clearance (13) is in agreement with the hypothesis that other domains contribute to the binding to the clearance receptor. We observed no cooperative interaction between FG and t-PA⌬FG, which suggests that these t-PA fragments do not interact with each other under the experimental conditions. Recently Hajjar and Reynolds (14) reported that the fucose group on the growth factor domain mediates binding of t-PA to human HepG2 hepatoma cells. However, we observed no inhibitory effect of fucose, even at high concentrations. This suggests that the low affinity of the nonfucosylated FG domain construct is not due to the lack of fucose on Thr-61. These results are in agreement with the observation that t-PA mutants, in which Thr-61 was mutated, efficiently inhibited the degradation of t-PA by human smooth muscle cells (16).
In conclusion, our results provide clear evidence in favor of a role of the finger and/or the growth factor domains in the interaction of t-PA with its clearance receptor. Results also suggest the presence of other binding sites located on the remainder of the t-PA molecule. The low affinity of the different binding sites may complicate their precise identification.