Arginine 260 of the Amino-terminal Domain of NR1 Subunit Is Critical for Tissue-type Plasminogen Activator-mediated Enhancement of N-Methyl-D-aspartate Receptor Signaling*

Tissue-type plasminogen activator (tPA) has been involved in both physiological and pathological glutamatergic-dependent processes, such as synaptic plasticity, seizure, trauma, and stroke. In a previous study, we have shown that the proteolytic activity of tPA enhances the N-methyl-d-aspartate (NMDA) receptor-mediated signaling in neurons (Nicole, O., Docagne, F., Ali, C., Margaill, I., Carmeliet, P., MacKenzie, E. T., Vivien, D., and Buisson, A. (2001) Nat. Med. 7, 59–64). Here, we show that tPA forms a direct complex with the amino-terminal domain (ATD) of the NR1 subunit of the NMDA receptor and cleaves this subunit at the arginine 260. Furthermore, point mutation analyses show that arginine 260 is necessary for both tPA-induced cleavage of the ATD of NR1 and tPA-induced potentiation of NMDA receptor signaling. Thus, tPA is the first binding protein described so far to interact with the ATD of NR1 and to modulate the NMDA receptor function.

Tissue-type plasminogen activator (tPA) 1 is one of the two mammalian serine proteases that activate the zymogen plasminogen into the broad spectrum serine protease plasmin. Apart from this primary function in the regulation of intravascular fibrinolysis, tPA has been implicated over the last decade in a variety of brain functions both during development and in adults. tPA is widely expressed in the developing brain, and it is prominently present in the hippocampus, cerebellum, and amygdala of the adult brain (2)(3)(4)(5) where both neurons and glial cells are sources of tPA (6,7). Furthermore, it is now well known that membrane depolarization induces a calcium-de-pendent secretion of tPA by neuronal cells (7)(8)(9). During ontogenesis, growing axons secrete tPA to modulate cell extracellular matrix interactions (2,10). In the adult brain, tPA has also been involved in processes such as synaptic plasticity and long term potentiation (LTP) (9,11). Indeed, mice deficient in tPA display a selective reduction in the late phase of LTP (L-LTP) (12,13), whereas transgenic mice overexpressing tPA have increased and prolonged L-LTP (14). Accordingly, a role for tPA has been established in learning and memory processes (12,14,15). Several mechanisms have been proposed to explain how tPA influences these processes. These include activation of extracellular proteolysis leading to a remodeling of the extracellular matrix and synaptic growth (9,16) or the cleavage of cell adhesion molecules (17). Recently, it has been reported that the binding of tPA to the cell surface low density lipoprotein receptor-related protein could play a key role in L-LTP (18).
tPA has also been shown to play an important role in acute and chronic brain pathologies such as seizure (19,20), ischemic brain injury (21,22), and multiple sclerosis (23,24). The hippocampal neurons of tPA-deficient mice are resistant to in vivo excitotoxin-induced degeneration (19). Similarly, tPA potentiates the excitotoxic lesion induced following an intrastriatal injection of NMDA (1,25). After cerebral ischemia, neuronal damage induced by middle cerebral artery occlusion is also reduced in tPA-deficient mice and exacerbated when exogenous tPA is administered (21). The idea that tPA contributes to the extent of the final damage induced by stroke has been later supported by other groups (26,27). Different mechanisms have been proposed to explain the deleterious effect of tPA in these processes, including enhanced microglial activation (28), increased laminin degradation (29), and the potentiation of NMDA receptor-dependent signaling (1). Here, we describe the molecular mechanism by which tPA modulates the NMDA receptor-evoked Ca 2ϩ influx. We show that tPA cleaves the amino-terminal domain (ATD) of the NR1 subunit of the NMDA receptor at arginine 260 and that this specific cleavage is a necessary event for tPA-induced potentiation of NMDA receptor signaling. This study constitutes an important step toward the understanding of how tPA influences several physiological and pathological processes involving the glutamatergic signaling.
Cortical Cultures-Neuronal cortical cultures were prepared from fetal mice (embryonic day [15][16]. Dissociated cortical cells were resuspended in Dulbecco's modified Eagle's medium supplemented with 5% fetal bovine serum, 5% horse serum, and 2 mM glutamine and plated in 24-well dishes previously coated with poly-D-lysine and laminin. After 3 days in vitro, the cells were exposed to 10 M Ara-C to inhibit glial proliferation. Cultures were used after 14 days in vitro (30).
Immunoblotting-Protein samples were resolved on SDS-polyacrylamide gel and transferred onto a polyvinylidene difluoride membrane. Membranes were blocked with 5% dried milk in Tris-buffered saline containing 0.05% Tween 20 and incubated with primary antibodies. After incubation with the corresponding biotinylated secondary antibody and peroxidase-conjugated streptavidin reagent, proteins were visualized with an enhanced chemiluminescence ECL Plus immunoblotting detection system (PerkinElmer Life Sciences).
Calcium Videomicroscopy Analysis-Transfected HEK-293 cells were loaded in the presence of a HEPES-buffered saline solution containing 5 M fura-2/AM plus 0.1% pluronic F-127 (Molecular Probes, Leiden, the Netherlands) (30 min, 37°C) and incubated for an additional 30-min period in a HEPES-buffered saline solution. Experiments were performed at room temperature on the stage of a Nikon Eclipse inverted microscope equipped with a 75 W Xenon lamp and a Nikon 40ϫ, 1.3 numerical aperture epifluorescence oil immersion objective. Fura-2 (excitation 340 and 380 nm, emission 510 nm) ratio images were acquired with a CCD camera (Princeton Instrument, Trenton, New Jersey) and digitized (256 ϫ 512) using Metafluor 4.11 software (Universal Imaging Corporation, Cherter, Pennsylvania).

Construction of His-tagged Recombinant ATD/Leucine-Isoleucine-Valine-binding Protein(LIVBP)-like Domain-
The region of the NR1 subunit encoding amino acids 19 -371 for NR1-1a or 19 -389 for NR1-1b corresponding to the ATD was amplified from the full-length rat NR1-1a or NR1-1b cDNA, respectively, by using the upstream primer 5Ј-CGGGATCCCGCGCCGCCTGCGAC-3Ј generating a BamHI restriction site and the downstream primer 5Ј-ATGGGTACCATTGTAGAT-GCCCAC-3Ј containing an internal KpnI restriction site. PCR products were digested and inserted in pQE100-Double Tag vector (Qiagen), which encodes for His 6 at the amino terminus of the insert. Recombinant proteins were purified from inclusion bodies of isopropyl 1-thio-␤-D-galactopyranoside-induced bacterial cultures (Escherichia coli, M15 strain) on a nickel affinity matrix as described by the manufacturer (Qiagen).
Site-directed Mutagenesis-Mutagenesis of either recombinant NR1-ATD (R260A and R217A) or full-length NR1-1a (R260A) was performed by using QuikChange® XL site-directed mutagenesis kit purchased from VWR International France (Fontenay-sous-Bois, France). All mutations were confirmed using an automated sequence analysis.
Enzymatic Assay-Recombinant tPA (29 nM) was incubated in the presence of a tPA-specific fluorogenic substrate (5 M) (Spectrozyme® A, map of the generated construct allowing the production of a recombinant form of the ATD of the NR1 subunit tagged with His 6 at its amino terminus. B, immunoblotting was performed from the recombinant ATD-NR1 previously incubated with tPA (20 and 100 g/ml) for 2 h at 37°C. Immunoblots were revealed with an antibody raised against His 6 (WB 6xHIS). XF444) and in the presence of either tPA-Stop® (10 nM) or recombinant ATD of NR1 (225 nM). The reaction was carried out at 25°C in 100 mM Hepes (pH 8.0) containing 150 mM NaCl, and 0.01% Tween 80 in a total volume of 100 l. The amidolytic activity of tPA was measured as the change in fluorescence emission at 440 nm (excitation at 360 nm). The apparent inhibition constant, K i Ј, was determined as described by Petersen et al. (32). The data sets on the inhibition of proteases were analyzed in terms of the equation where V i and V 0 are the inhibited and uninhibited rates, respectively, and [I] 0 is the total concentration of inhibitor. K i values were obtained by correcting for the effect of substrate (S) with the equation MALDI-TOF Analysis-Proteins in Coomassie-stained gels were subjected to acetonitrile washing and reductive alkylation by iodoacetamide in ammonium carbonate (0.1 M) (30 min in the dark). Then, in-gel trypsin digestion (Roche Applied Science, EC 3.421.4) was allowed as described by Shevchenko et al. (33). Peptides were resuspended in 20 l of formic acid 1%, desalted by using Zip Tip C-18 (Millipore), and eluted with 50 and 80% acetonitrile. The desalted peptide mixture was dried and dissolved in 3 l of formic acid 1%. The matrix used was a saturated solution of 2,5-dihydroxybenzoic acid in trifluoroacetic acid 0.1%. The sample and the matrix (1:1, v/v) were loaded on the target using the dried droplet method. MALDI-TOF spectra of the peptides were obtained with a Voyager-DE STR Biospectrometry Work station mass spectrometer (PE Biosystems Inc.). Analyses were performed in positive ion reflector mode, with an accelerating voltage of 20,000 V, a delayed extraction of 200 ns, and ϳ250 scans were averaged. For subsequent data processing, the Data Explorer software (PE Biosystems Inc.) was used. Spectra obtained for the whole protein were calibrated externally using the [MϩH]ϩ ion from Des-Arg bradykinin peptide (904.4681 Da) and ACTH peptide (2465.1989 Da). The trypsin autoproteolysis products (fragment-(132-142), 1153.57 Da and fragment-(56 -75), 2163.06 Da) were used as the external calibration standard. A mass deviation of 0.1 Da was allowed in the data base searches.
Homology Modeling-The sequence of the ATD of NR1 has been aligned with an alignment deduced from the structural superposition of the extracellular domain of the metabotropic glutamate receptor 1 (1EWT, 1EWK) (34) and four other proteins with a LIVBP fold (the atrial natriuretic peptide clearance receptor (1DP4), the amide receptor (1PEA), the ribose-binding protein (2DRI), and the leucine/isoleucine/valine-binding protein (2LIV)). The alignment and model have been built and optimized iteratively using Model Tool Server (35) and Modeler (36).

RESULTS
tPA Potentiates NMDA Receptor-mediated Ca 2ϩ Influx in HEK-293 Cells Expressing NMDA Receptors Containing Either NR1-1a or NR1-1b-We have previously shown that tPA en-hances the NMDA-evoked Ca 2ϩ influx in cultured cortical neurons (1,7). To investigate the molecular mechanism through which tPA is able to modulate the NMDA receptor signaling, we developed a heterologous expression system of functional NMDA receptors. We transiently transfected HEK-293 cells with the cDNA encoding for NR1-1a and NR2A subunits of the NMDA receptor. The expression of NMDA receptor subunits was confirmed by immunocytochemistry and immunoblotting analyses (data not shown). As shown in Fig. 1, in HEK-293 cells expressing NMDA receptors, tPA is able to potentiate NMDA receptor signaling, as it does in cortical neurons (1). The mRNA encoding for the NR1 subunit exhibits an alternative splicing in its 5Ј-region, generating isoforms characterized by the absence (a forms) or the presence (b forms) of a 21-residue insert (N1 cassette) encoded by exon 5 (37). Because the exon 5 splicing is known to influence receptor properties (38 -41), we have assessed its potential implication in tPA effect. We found that tPA increases NMDA-induced Ca 2ϩ influx to the same extent in HEK-293 cells exhibiting NR1-1b/NR2A or NR1-1a/ NR2A receptors (36.87 Ϯ 10.9%, n ϭ 32 and 31.37 Ϯ 5.4%, n ϭ 56, respectively, mean Ϯ S.E.) (Fig. 1, A-C). Hence, the N1 cassette does not influence the tPA-induced potentiation of NMDA receptor signaling.
tPA Induces a Single Cleavage of NR1 in Its ATD-To further investigate the mechanism of action of tPA on NMDA receptor signaling, we have examined whether tPA, per se, can cleave NR1. We have shown previously that the treatment of membrane preparations of cultured cortical neurons with tPA leads to the appearance of a cleaved form of NR1, with a molecular mass reduced to ϳ25 kDa, recognized by an antibody raised against the carboxyl terminus of NR1 (1). This suggests that the amino-terminal portion of the NR1 subunit is the region in which the cleavage occurs. Hence, to determine where tPA cleaves the amino terminus of NR1, we have generated a construct encoding for an amino terminus His-tagged corresponding to the first domain of NR1-1a (amino acid residues 19 -371 of NR1, 39875 Da), termed ATD or LIVBP-like domain ( Fig. 2A). Incubation of recombinant NR1-ATD with tPA shows that tPA is able to cleave NR1-ATD, in a dose-dependent manner, generating a cleaved fragment of 28 kDa detected by immunoblotting revealed with an antibody raised against the His-tag (Fig. 2B). In addition, tPA-Stop®, an inhibitor of tPA proteolytic activity, prevents the tPA-induced cleavage of both the full-length NR1 (of neuronal membrane preparations from cultured cortical neurons) and the recombinant NR1-ATD (Fig.  3A). Finally, as determined by the digestion pattern of the recombinant NR1-ATD generated by tPA or plasmin, we observed that although tPA induces a single cleavage of the ATD of NR1 subunit, plasmin leads to a complete degradation of the ATD (Fig. 3B).
The ATD of NR1 Is a Substrate of tPA-The next question was to determine whether the ATD of NR1 could be a direct substrate for tPA. To address this question, we have used the recombinant ATD of NR1 to compete the ability of tPA to cleave a specific fluorogenic substrate (Spectrozyme®, XF444). Our data showed that the recombinant ATD of NR1 is able to compete with the tPA-specific substrate with a K i of 0.234 Ϯ 0.097 M (Table I). Control experiments were performed in  3. Cleavage of the ATD of the NMDA receptor NR1 subunit is prevented by a synthetic tPA inhibitor, tPA-Stop®, and occurs independently of the plasminogen/plasmin system. A, immunoblots were performed from membrane preparations of mouse cultured cortical neurons and recombinant ATD-NR1 incubated in the presence of tPA alone or with the tPA inhibitor tPA-Stop®. Immunoblots were revealed with an antibody raised against either NR1 or His 6 , respectively. B, immunoblots were performed from the recombinant ATD-NR1 incubated with either tPA (20 g/ml) or plasmin at increasing concentrations. Immunoblots were revealed with an antibody raised against either NR1 or His 6 , respectively. parallel in the presence of the tPA inhibitor tPA-Stop® with a K i of 0.046 Ϯ 0.015 M (Table I).
tPA Cleaves the ATD of NR1 at the Arginine 260 -Next, to identify the exact location of the cleavage site, we analyzed both native (39 kDa) and tPA-cleaved (28 kDa) recombinant NR1-ATD using MALDI-TOF analysis. As summarized in Fig.  4, MALDI-TOF analysis allowed us to identify the putative cleavage site of the NR1-1a and NR1-1b subunits by tPA as the arginine in position 260 (Arg-260) and arginine 281 respectively (Fig. 4A). Sequence analyses revealed that the tPA- FIG. 4. Mass spectrometric analysis of both native and tPA-cleaved ATD of the NR1 subunit. A, recombinant ATD-NR1-1a and ATD-NR1-1b were incubated in the presence of tPA (20 g/ml, 2 h, 37°C) prior to SDS-PAGE and Coomassie staining. Mass spectrometry analyses of both full-length ATD-NR1-1a and ATD-NR1-1b and tPA-cleaved ATD-NR1-1a and ATD-NR1-1b were performed after in-gel digest procedure (tryptic digestion) as described under "Experimental Procedures." Peptides identified from corresponding full-length ATD-NR1 and tPA-cleaved ATD-NR1 are summarized in the Table I. The comparison of the peptidic profiles obtained between full-length ATD and tPA-cleaved ATD allows the identification of the tPA cleavage site. Peptide EISGNALR (bold characters) corresponds to the carboxyl-terminal peptide identified following cleavage induced by tPA. ISGNALRYAPDG was designed as the peptide containing the tPA cleavage site Arg-260 (recapitulative of three individual experiments for each isoform of NR1 subunit). B, ribbon-type representation of the calcium chain of the homology model of the ATD-NR1 constructed using the metabotropic glutamate receptor 1 structure showing the two lobes surrounding a hydrophobic pocket. The position of the arginine 260, in a loop overhanging the pocket domain is shown in ball-and-stick representation. C, alignment of ATD-NR1 and metabotropic glutamate receptor 1 sequences used to build the homology model, the resulting secondary structure of ATD-NR1 is shown above. Graphics were created using MOLSCRIPT (55) and rendered by RASTER3D (56), and alignment representations were made using ESPript software (57). cleavage region "ISGNALRYAPDG" is highly conserved in NR1 subunits of NMDA receptors, whatever the species analyzed, and not found in other NMDA receptor subunits (data not shown). The analysis of the homology model of the ATD of the NR1 subunit (Fig. 4, B and C) shows that the Arg-260 is located close to the entry of a hydrophobic pocket for which no ligand has been described so far.
Mutation of the Arg-260 Prevents both tPA-induced Cleavage of NR1-ATD and Potentiation of NMDA-receptor Signaling-To validate mass spectrometry analyses, we performed a mutation of the Arg-260 (and Arg-217 as a negative control) of the NR1-ATD recombinant protein into alanine and tested the ability of tPA to cleave these proteins. Although tPA cleaved both wildtype and control R217A proteins, it failed to cleave the protein mutated at the arginine in position 260 (Fig. 5A). Moreover, wild-type NR1-ATD inhibited the ability of tPA to cleave its specific fluorogenic substrate (40%, p Ͻ 0.01). In contrast NR1R260A-ATD had no effect on tPA activity (Fig. 5B). Next, to determine whether a point mutation of this arginine at the position 260 could prevent the tPA-induced potentiation of NMDA receptor-mediated signaling, we transiently co-transfected HEK-293 cells with NR2A in the presence of either wild-type or R260A mutated NR1-1a, and performed calcium videomicroscopy experiments. When co-transfected with NR2A, both wild-type and R260A mutated NR1-1a subunits displayed the same responses to increasing concentrations of NMDA (Fig. 6, A and B). In contrast, although the addition of exogenous tPA potentiated NMDA-induced Ca 2ϩ influx in HEK-293 cells co-transfected with wild-type NR1-1a and NR2A, tPA failed to potentiate NMDA receptor signaling in transfected cells containing the mutated NR1-1a R260A subunit and NR2A (27.57 Ϯ 3.7%, n ϭ 102 and 2.49 Ϯ 1.7%, n ϭ 110, respectively) (Fig. 7, A and B). DISCUSSION Fast excitatory neurotransmission in the mammalian central nervous system is mediated by ionotropic glutamate-gated receptors. When overstimulated, ionotropic glutamate-gated receptors, particularly the NMDA subtype, cause excitotoxicity. We have shown previously that the serine protease tPA potentiates NMDA receptor signaling, which might be of particular relevance in several physiological and pathological brain conditions (1). However, whether this effect is the direct consequence of the ability of tPA to bind to and then cleave the NMDA receptor NR1 subunit remained to be demonstrated. Here, we demonstrated that tPA cleaves NR1 within a particular extracellular region, the ATD, in which we identified the arginine in position 260 as a critical residue for the tPA-induced potentiation of NMDA signaling.
First, by using transfected HEK cells, we show that tPA potentiates NMDA signaling indifferently in receptors containing either NR1a or NR1b splice variants. Both isoforms differ by the absence or presence of a amino-terminal N1 cassette, known to modulate some properties of the NMDA receptor (38 -41). Our results thus demonstrated that the N1 cassette is not involved in the potentiating activity of tPA on NMDA receptor signaling. We also demonstrated that tPA can interact with and cleave the ATD of both NR1a and b subunits. The ATD of NR1 also LIVBP-like domain is a conserved domain found in all metabotropic and ionotropic glutamate receptor subunits described so far. Mutagenesis analyses within the ATD of NMDA receptor subunits have suggested that this domain might regulate the functional features of this receptor, by controlling subunit-subunit interactions (42) and/or altering allosteric modulations (43)(44)(45). Our present data provide a novel function for the ATD of NR1. Indeed, we showed that this ATD of NR1 is a binding site for tPA and that the cleavage of its arginine 260 in NR1-a (arginine 281 for NR1-b) is a necessary event for tPA-induced potentiation of NMDA signaling.
As previously suggested (46,47) whether tPA could form a direct complex with NR1 or through the participation of its classical substrate plasminogen remained to be determined. Although the presence of locally synthesized tPA in the central nervous system has been well characterized (3,5) that of plasminogen is still a matter of debate (3,5,6,48,49). Hence, we can not exclude the idea that tPA could cleave the NMDA receptor via the activation of plasminogen into plasmin. Indeed, activated plasminogen has been shown to lead to a complete degradation of the NR1 subunit of NMDA isolated from brain lysates (47). Accordingly, we showed that plasmin leads to a complete degradation of the ATD of NR1. However, this proteolytic pattern differs from that of tPA, which induces a single cleavage at the Arg-260. Thus, these data demonstrated that tPA-induced cleavage of NR1 is not mediated by plasmin, which is in agreement with our previous observations that in contrast to tPA, plasmin does not influence NMDA receptormediated neuronal death (1).
Another important issue is to know whether the endogenous concentrations of tPA are high enough to lead to the cleavage of the NR1 subunit of the NMDA receptor and subsequent potentiation of NMDA-mediated signaling in in vivo situations. In a previous study, we have shown that in cultured cortical neurons, NMDA receptor activation is sufficient to promote both tPA release and subsequent cleavage of the NR1 subunit of NMDA receptor. In addition, this NMDA-dependent cleavage of NR1 is inhibited by exogenous plasminogen activator inhibitor-1 and does not occur in tPA-deficient cortical neurons (1). Additionally, we have demonstrated that the blockade of the proteolytic activity of endogenous tPA reduces NMDA-induced excitotoxic death (1, 30). Nevertheless, the concentration of tPA not only in the brain parenchyma but also at the synaptic cleft remain actually undetermined.
In physiological conditions, an impairment or increase of L-LTP has been observed in tPA-deficient or tPA-overexpressing mice, respectively (12)(13)(14). In addition to hippocampal functions, tPA has also been involved in phenomena such as cerebellar motor learning (15,50), visual cortex plasticity (51), and fear conditioning (52). During the last several years, several reports have also demonstrated a connection between tPA and brain disorders. tPA has been involved in many pathological processes including stroke (21,25), seizure (19,20,53), and multiple sclerosis (23,24). For instance, tPA-deficient mice display a high resistance to neuronal damages induced through either excitotoxic paradigms (19,24) or experimental ischemia (21,25). Conversely, plasminogen activator inhibitor-1-deficient mice show an increased volume of infarct following cerebral ischemia (25). In addition, mice overexpressing neuroserpin, another tPA inhibitor, have a reduced volume of lesion following ischemic brain injury (54). Here, we described a molecular mechanism through which tPA can influence the glutamatergic signaling. Although this mechanism might not explain all of the functions of tPA in the brain, it highlights the importance of this protease in physiological and pathological situations. Hence, these data should represent the basis of the development of new therapeutic strategies targeting the interaction between tPA and NMDA receptors.