Rat Brain IV-Acetylated a-linked Acidic Dipeptidase Activity PURIFICATION AND IMMUNOLOGIC CHARACTERIZATION*

titer polyclonal dipeptidase immunoreactivity cortices.

In this report, we have solubilized NAALA dipeptidase activity from synaptosomal membranes with Triton X-100 and purified it to apparent homogeneity by sequential column chromatography on DEAE-Sepharose, CM-Sepharose, and lentil lectin-Sepharose.
This procedure resulted in a 720-fold purification with 1.6% yield. The purified enzyme migrated as a single silver-stained band on a sodium dodecyl sulfate gel with an apparent molecular weight of 94 kDa. Using an enzymatic stain to visualize NAALA dipeptidase activity within a gel matrix, we have confirmed that the 94-kDa band is, indeed, NAALA dipeptidase. The purified enzyme was characterized and found to be pharmacologically similar to NAALA dipeptidase activity described previously in synaptosomal membrane extracts.
Using the purified NAALA dipeptidase as antigen, we have raised specific and high titer polyclonal antibodies in guinea pig. Immunocytochemical studies show intense NAALA dipeptidase immunoreactivity in the cerebellar and renal cortices.
Electrophysiologic, lesion, and immunocytochemical studies suggest that the endogenous neuropeptide, N-acetyl-r.aspartyl-L-glutamate (NAAG),' may act as a neurotransmitter/neuromodulator in the central nervous system (Blakely and Coyle, 1989). Recently a quisqualate (Quis)-sensitive peptidase activity was identified in brain membranes which cleaves NAAG to N-acetyl-L-aspartate (NAA) and glutamate. In a manner analogous to the synaptic inactivation of acetylcholine (Cooper et al., 1986), it is hypothesized that this peptidase inactivates NAAG, and that the liberated glutamate is subsequently transported into synaptosomes by the previ- ously characterized sodium-dependent high-affinity glutamate uptake site (Blakely et al., 1986;Robinson et al., 1987). Alternatively, NAAG may function as a precursor to glutamate, shifting the primary role of this peptidase to regulating glutamate availability.
This peptidase has been characterized in rat synaptosomal membranes (Robinson et al., 1987;Blakely et al., 1988). In this crude membrane preparation, the peptidase demonstrates remarkably high apparent affinity for its putative substrate NAAG, with a K,,, = 540 nM. The enzyme is membrane-bound, stimulated by chloride ions, and inhibited by divalent metal chelators, suggesting that it is a metallopeptidase.
It is enriched in synaptic plasma membranes and is primarily localized to neural tissue and kidney. Comparison of its properties to those of other known endopeptidases, aminopeptidases, dipeptidases, and acyl amino acid-releasing enzymes suggests that it is a novel peptidase (Robinson et al., 1987;Blakely et al., 1988). Since it is possible that NAAG is not the sole substrate for this activity in uiuo, this peptidase was named N-acetylated a-linked acidic dipeptidase (NAALA dipeptidase) for its structural specificity for N-acetylated a-linked acidic dipeptides.
Recently, it has been demonstrated that [3H]NAAG is degraded by a pharmacologically similar enzyme in uiuo (Stauch et al., 1989). These data are consistent with a role for NAALA dipeptidase in the disposition of endogenous NAAG.
In this manuscript, we describe for the first time the solubilization of NAALA dipeptidase from rat membranes, its purification to apparent homogeneity, the characterization of the purified protein, the determination of its molecular weight, the production and characterization of anti-NAALA dipeptidase antibodies, and the localization of NAALA dipeptidase immunoreactivity in brain and kidney.

EXPERIMENTAL PROCEDURES AND RESULTS2
Chromatogruphic Purification of NAALA Dipeptiduse-Results of the purification of rat brain NAALA dipeptidase activity are summarized in Table I. The overall purification was 720-fold with 1.6% recovery, yielding 2 mg of highly purified NAALA dipeptidase from 500 whole rat brains. Details of the solubilization and chromatographic steps are found in the miniprint supplement. Analysis of Enzyme Homogeneity- Fig.  2 shows SDS polyacrylamide gel electrophoresis of pooled fractions at various stages in the purification.
After the lentil Lectin step, there was one major silver-stained protein band migrating at 94   (Wray et al., 1981).
kDa, and a minor (diffuse) band migrating between 54 and 66 kDa. The literature suggests that minor contaminating protein(s) are either mercaptoethanol artifacts (Guevara et al., 1982;Tasheva and Dessev, 1983) or skin keratins (Ochs, 1983). The staining intensity of this 94-kDa band was correlated with the amount of NAALA dipeptidase activity applied to the gel. Gel electrophoresis of fractions surrounding a NAALA dipeptidase peak of activity from DEAE-Sepharose, CM-Sepharose, and size exclusion columns (n = 4) demonstrated that, in all cases, the 94-kDa protein band was the only band observed whose staining density coincided with NAALA dipeptidase activity (data not shown).
Although these data provide compelling evidence that the 94-kDa band is NAALA dipeptidase, it is still possible that NAALA dipeptidase is not represented by any band on the gel. Therefore, a specific enzymatic activity stain was devised to visualize NAALA dipeptidase activity within a polyacrylamide gel (Sugiura et al., 1977; see "Experimental Procedures"). Since a protein separated by an SDS gel (e.g. denatured enzyme) is highly unlikely to exhibit enzymatic activity, partially purified NAALA dipeptidase was electrophoretically separated on a nondenaturing gel and then stained for activity. Only one band was identified having NAALA dipeptidase activity (Fig. 3A). Since it is not possible to accurately determine molecular weight on a nondenaturing gel, this active band was excised, homogenized, and subjected to SDS-polyacrylamide gel electrophoresis (Fig. 3B,   % five bands were revealed with this procedure, all bands, except for a 94-kDa protein, were attributed to the staining components used to visualize NAALA dipeptidase activity in the nondenaturing gel (e.g. glutamate dehydrogenase, iodonitrotetrazolium violet, phenazine methosulfate; Fig. 3B, right lane). These data suggest that the 94-kDa band is NAALA dipeptidase.
Properties of the Purified NAALA Dipeptidase-As was observed for activity characterized in lysed synaptosomal membranes (Robinson et aZ., 1987), purified NAALA dipeptidase was potently inhibited by quisqualate with 50% inhibition at 0.48 pM (Table II). Peptidase activity was also inhibited by phosphate and EGTA; cobalt strongly stimulated activity. Purified NAALA dipeptidase showed a high apparent affinity for NAAG hydrolysis with a K,,, of 140 nM (mean of two determinations within 10%).
Structure-activity relationships of purified NAALA dipeptidase were examined using peptide analogs of NAAG and compared with what was previously determined using a lysed synaptosomal membrane preparation (Robinson et al., 1987). All peptides examined were used at their previously reported IC-,o concentrations.
The results are summarized in Table II was used to raise polyclonal antibodies in guinea pigs (see "Experimental Procedures"). Enzyme-linked immunosorbent assay, immunoprecipitation assay, and Western blotting demonstrated that the anti-NAALA dipeptidase antisera was specific and of high titer. The titer, determined by enzyme-linked immunosorbent assay (see "Experimental Procedures"), was approximately l:lO,OOO (data not shown). Zmmunoprecipitation-The immune serum precipitated NAALA dipeptidase activity from solubilized brain membranes; preimmune serum did not precipitate activity (Fig. 4). Immune serum partially inhibited NAALA dipeptidase activity; therefore, at antibody dilutions less than l:lOO,OOO total activity using the immune serum was less than activity in the preimmune control. Western Blotting-An SDS gel loaded with crude homogenates from rat heart, brain, intestine, liver, kidney, pancreas, spleen, and testis was transferred to nitrocellulose and probed with anti-NAALA dipeptidase immune serum as described under "Experimental Procedures." The immune serum recognized a single 94-kDa band in brain, kidney, and testis only, consistent with the localization of NAALA dipeptidase activity (Fig. 5). No staining was detected using preimmune serum. Zmmunocytochemistry-Immunocytochemical experiments revealed NAALA dipeptidase-positive staining in the molecular and granule cell layers of the cerebellar cortex; the Purkinje cell layer was devoid of immunoreactivity ( Fig. 6,  top). In the rat kidney, NAALA dipeptidase-positive staining was detected in the proximal tubules and glomeruli of the renal cortex; the distal tubules were essentially devoid of immunoreactivity (Fig. 6, bottom). No immunostaining was revealed using preimmune serum, even at lo-fold higher concentrations.

DISCUSSION
Brain NAALA dipeptidase activity was solubilized with Triton X-100 and sequentially purified with ion-exchange and lentil lectin affinity chromatography (Fig. 1). DEAE-Sepharose resolved NAALA dipeptidase activity into two peaks; however, both peaks were pharmacologically and kinetically similar. Since DEAE peak I was the predominant species (>85% of eluted activity), it was employed for further purification. NAALA dipeptidase activity (peak I) did not interact with DEAE-Sepharose at pH 7.9 and bound to CM-Sepharose at this same pH, suggesting that this protein has an unusually high isoelectric point (PI); most proteins have p1 values below pH 7.4 (Righetti and Caravaggio, 1976). Chromatofocusing chromatography confirmed this finding, revealing a p1 for NAALA dipeptidase of approximately pH 9.0. Following lentil lectin chromatography, the purified preparation showed one major silver-stained band on SDS-polyacrylamide gel electrophoresis migrating at 94 kDa and a minor (diffuse) band migrating between 54 and 66 kDa. The literature suggests that this minor broad band is either a mercaptoethanol artifact (Guevara et aZ., 1982;Tasheva and Dessev, 1983) or skin keratins (Ochs, 1983). In fact, we have seen this diffuse band in lanes run with sample buffer alone.
To demonstrate that the 94-kDa protein represented NAALA dipeptidase, activity applied to an SDS gel was correlated with protein staining intensity of this band. In all gels, the 94-kDa band was the only band observed whose staining density coincided with the amount of applied NAALA dipeptidase activity. Furthermore, NAALA dipeptidase activity was visualized directly in a nondenaturing gel (Fig. 3A).
NAALA dipeptidase activity demonstrated a low degree of migration into the nondenaturing gel, consistent with its high isoelectric point. This resulting activity band was excised from the nondenaturing gel and run on an SDS gel. Protein staining again revealed a single unique band at 94 kDa (Fig. 3B). Together, these data strongly suggest that this 94-kDa band is NAALA dipeptidase.
Size exclusion chromatography of the purified and semipurified protein show that NAALA dipeptidase migrates consistent with a molecular mass of 225 kDa, although larger species were occasionally observed. Both protein and activity gels demonstrate that peptidase has a denatured molecular mass of approximately 94 kDa. Therefore, it is possible the NAALA dipeptidase is a dimer composed of two identical subunits; alternatively, the larger species identified with size exclusion chromatography may represent protein-detergent complexes.
Properties of the purified protein were similar to activity previously characterized in lysed synaptosomal membranes (Robinson et al., 1987), demonstrating that these properties are due to direct interaction with NAALA dipeptidase and are not indirectly mediated by other proteins present in the membrane preparation.
The potent inhibition of peptidase activity by quisqualate suggests that some of its actions, which were previously attributed to interaction with a subclass of glutamate receptors , may be due to inhibition of NAALA dipeptidase. EGTA sensitivity and cobalt stimulation support initial data suggesting that NAALA dipeptidase was a metallopeptidase.
Similar to activity in lysed synaptosomal membranes, purified NAALA dipeptidase displayed structure specificity for N-acetylated 01linked acidic dipeptides (Table II). Finally, the purified NAALA dipeptidase displayed a remarkably high apparent affinity for its putative substrate, NAAG, with a K,,, of 140 nM.
The availability of purified protein permitted the production of anti-NAALA dipeptidase antisera. The results presented in this study demonstrate that the polyclonal antibodies raised in guinea pig are remarkably selective, of high titer, and capable of recognizing both native and denatured NAALA dipeptidase.
Western analysis of gels loaded with crude brain homogenates revealed that the antisera exclusively recognized the 94-kDa band. Besides brain NAALA dipeptidase, the antibodies cross-reacted with kidney and testis NAALA dipeptidase; no immunoreactivity was observed in regions devoid of NAALA dipeptidase activity (Fig. 5). The antisera inhibited NAALA dipeptidase activity, although not completely (70% inhibition at 1:lOO dilution), and was capable of precipitating NAALA dipeptidase activity from a crude brain extract (Fig. 4).
Using this selective and specific antisera, NAALA dipeptidase immunoreactivity was localized to the glomeruli and proximal tubules of the kidney cortex (Fig. 5B). This localization is consistent with micropunch analysis of NAALA dipeptidase activity,3 which found that the vast majority of NAALA dipeptidase activity was localized to the renal cortex (kidney cortex = 166 pmol/mg/min versus kidney medulla = 10 pmol/ mg/min).
Interestingly, other brain peptidases, angiotensin converting enzyme and enkephalinase (Schulz, 1988, Taut, 1988 have also been localized in the glomeruli and proximal tubules of the kidney, areas where their putative brain substrates are not found. In the neural tissue, NAALA dipeptidase immunoreactivity was found in areas reported previously to contain NAAG (Blakely and Coyle, 1989).
NAALA dipeptidase is a novel enzymatic activity involved in NAAG hydrolysis.
In this study, we have solubilized and purified rat brain NAALA dipeptidase to apparent homogeneity, developed specific anti-NAALA dipeptidase antiserum, and have begun to map its distribution in rat brain and kidney. We anticipate using the antiserum to fully determine its renal and neuronal localization and to screen a cDNA library to obtain the NAALA dipeptidase clone.