The Prostate 39:28–35 (1999) Characterization of the Enzymatic Activity of PSM: Comparison With Brain NAALADase Carol W. Tiffany, Rena G. Lapidus, Aviva Merion, David C. Calvin, and Barbara S. Slusher* Guilford Pharmaceuticals, Inc., Baltimore, Maryland BACKGROUND. The prostate cancer marker prostate-specific membrane antigen (PSM) is highly homologous to the brain enzyme N-acetylated alpha-linked acidic dipeptidase (NAALADase). NAALADase is known to cleave terminal carboxy glutamates from both the neuronal peptide N-acetylaspartylglutamate (NAAG) and folate polyglutamate. In this report, we compare the NAAG hydrolyzing activity of NAALADase and the prostate enzyme PSM. METHODS. Using a NAAG hydrolytic radioenzymatic assay, we compared the pharmacological and kinetic properties of the brain and prostate enzymes. RESULTS. Eight normal prostate tissues from different species exhibited NAAG hydrolyzing activity. Among 14 cancer cell lines examined, activity was observed in human LNCaP, PC-82, and rat Dunning G and AT-1 cells. Brain exhibited membrane-localized activity exclusively, while the prostate enzyme had activity in both membrane and cytosolic fractions. The only observed pharmacological difference was the sensitivity to their putative substrates, folate polyglutamate and NAAG. Kinetically, the soluble form of the prostate enzyme had two catalytic sites, while the membrane-bound form exhibited single site kinetics with a lower Vmax than the brain enzyme, which may suggest a less active hydrolase in the prostate. CONCLUSIONS. The brain enzyme NAALADase and the prostate enzyme PSM are remarkably similar. The importance of the differences in substrate specificities and kinetic parameters remains to be elucidated. Prostate 39:28–35, 1999. © 1999 Wiley-Liss, Inc. KEY WORDS: Dunning; folic acid; LNCaP; prostate cancer INTRODUCTION Recently, rat brain N-acetylated-alpha-linked acidic dipeptidase (NAALADase) was shown to have 86% DNA sequence homology to the prostate cancer marker, prostate-specific membrane antigen (PSM)  (reviewed by Heston  and Fair et al. ). PSM was cloned from the LNCaP prostate cancer cell line . It is a 110-kDa glycoprotein containing a short cytoplasmic region, a transmembrane domain, and large extracellular region. An alternatively spliced form of the protein, termed PSM⬘, has been cloned from normal prostate . PSM⬘ lacks 266 nucleotides of the amino terminus which comprises the ATG and the transmembrane domain, making it a cytosolic protein. PSM⬘ mRNA is primarily expressed in normal prostate tissue, while PSM mRNA is more prevalent in © 1999 Wiley-Liss, Inc. prostate cancers. In fact, PSM/PSM⬘ mRNA ratios have been reported to be 100 times greater in cancer than in normal prostate tissues, suggesting that these ratios may be utilized to indicate stage of disease . Also, the presence of PSM mRNA, not PSM⬘, as determined by in situ hybridization, may be indicative of hormone-refractory and metastatic disease . PSM has been reported to be measurable by RT-PCR and in some cases by Western blot analysis in the serum of patients with benign prostatic hyperplasia and prostate cancer [7,8]. Purified to homogeneity, rat brain NAALADase *Correspondence to: Barbara S. Slusher, Guilford Pharmaceuticals, Inc., 6611 Tributary St., Baltimore, MD 21224. E-mail: slusher b@ guilfordpharm.com Received 28 May 1998; Accepted 20 August 1998 Characterization of PSM Enzyme Activity has been extensively characterized and localized to brain, kidney, and testes [9–11]. In the brain, NAALADase is known to hydrolyze the neuronal dipeptide N-acetyl-aspartate-glutamate [NAAG]. The reaction product, glutamate, is implicated in neuronal transmission [12,13]. Brain-localized NAALADase is a membrane-bound, chloride-dependent zinc metalloproteinase which requires a divalent cation cofactor; it is inhibited by M amounts of quisqualic acid and phosphate and mM amounts of EDTA . PSM cDNA has been transiently transfected into two PSM-negative human prostate cancer cell lines, PC-3 and Du145. Cell lysates from the two transfected cell lines and the PSM-positive LNCaP cell line hydrolytically liberate glutamate from NAAG, and their enzymatic activities appear to have similar sensitivities to those of known NAALADase inhibitors . More recently, Pinto et al.  stably transfected PSM into PC-3 cells and characterized it as a folate hydrolase with high affinity for folate polyglutamate, cleaving the glutamyl bond of the terminal carboxy glutamates. In this report, we pharmacologically and kinetically characterize the enzymatic activity of prostate PSM in both mammalian tissues and prostate cancer cell lines and compare it to rat brain NAALADase. MATERIALS AND METHODS Materials The human prostate cancer cell lines, Du145, PC-3, and NCI H660, were obtained from the American Type Culture Collection (Rockville, MD). The human prostate cancer line LNCaP was a gift from Dr. William Nelson at Johns Hopkins Medical School (Baltimore, MD). The human prostate cancer cell lines Tsu-PR1 and DuPRO, and the rat Dunning R2337 prostate cancer lines G, H, AT-1, AT-2, AT-3, MatLu, and MatLyLu, were a kind gift of Dr. John Isaacs at the Johns Hopkins Oncology Center (Baltimore, MD). The PC-82 human prostate xenograft, also a kind gift of Dr. Isaacs, is a serially transplanted human prostate xenograft derived from a primary human prostate carcinoma . Interestingly, this tumor maintains many of the important physiological properties (e.g., differentiation status and androgen sensitivity) of clinical prostate cancer when serially transplanted in nude mice. When passaged, the PC-82 tumor is assessed histologically to ensure that no changes have occurred. Cell growth media and reagents were purchased from Paragon Biotech (Baltimore, MD). Frozen rat brains as well as frozen rabbit, cat, and dog prostates were purchased from Pell-Freeze (Rogers, AK). Frozen guinea pig prostates were obtained from Rockland 29 (Gilbertsville, PA). Nude mice and Sprague Dawley rats used for in-house prostate dissection were purchased from Harlan Sprague Dawley (Indianapolis, IN). Human prostates collected postmortem from three Caucasian men aged 89, 70, and 62 were prepared as membrane and soluble fractions by Analytical Biological Services, Inc. (Wilmington, DE). Tissues were assayed separately and reported as mean ± SD. The chemiluminescent detection agent ECL was purchased from Pierce Chemical Co. (Rockford, IL), and Kodak X-ray film was purchased from Amersham Life Science, Inc. (Arlington Heights, IL). Radiolabelled NAAG was purchased from NEN™ Life Science Products (Boston, MA), and scintillation cocktail UltimaGold was from Packard Instrument Co. (Meridan, CT). Baker Ultra-Pure water was obtained from VWR Scientific Products (Bridgeport, NJ). Folate polyglutamates and methotrexate polyglutamates were purchased from Schirck’s Laboratories (Jona, Switzerland). All other chemicals were purchased from Sigma-Aldrich Corp. (St. Louis, MO). Tissue Preparations Prostate tissue. Tissue was prepared by homogenizing with a Polytron威 P10 (Brinkman, Westbury, NY) in 10 volumes (weight/volume) of ice-cold Baker UltraPure water, sonicating twice for 30 sec each with a point sonicator, and passing the homogenate through a 100-mesh sieve. The resultant homogeneous solution was centrifuged at 50,000g for 20 min. Supernatant was removed and adjusted to 50 mM TrisCl pH 7.4 at 37°C (soluble fraction). Pellets were homogenized and sonicated in half the original water volume, using 50 mM TrisCl buffer (membrane). All fractions were aliquoted and stored at −80°C until assayed. Rat brain. Synaptic membrane fractions from frozen rat brains were prepared as previously described . Briefly, brains were homogenized in 10 volumes (weight/volume) of 0.32 M sucrose and centrifuged at 800g for 10 min. The supernatant was removed and centrifuged at 20,000g for 20 min. The pellet was suspended in 10 volumes (weight/volume) of Baker UltraPure water using a Polytron and then centrifuged at 8,000g for 20 min. The supernatant and buffy coat were removed and centrifuged at 35,000g for 10 min. The resulting pellet was resuspended in 20 volumes of 50 mM TrisCl, pH 7.4, at 37°C, incubated for 30 min at 37°C, and then centrifuged at 35,000g for 10 min. After washing the pellet in 20 ml buffer and centrifuging twice more, the membranes were suspended in TrisCl buffer to a final concentration of approximately 5 mg/ ml protein. 30 Tiffany et al. Prostate cancer cell lines. Prostate cancer cell lines were maintained in 95% air, 5% CO2 at 37°C in a humidified incubator in RPMI plus 10% heat-inactivated fetal calf serum and 100 units of penicillin and 100 g/ml streptomycin. Dexamethasone (1 × 10−8 M) was added to the growth media for the all the Dunning cell lines except G. Cells were grown to confluency in t-75 flasks, scraped into ice-cold phosphatebuffered saline (PBS), and centrifuged at 1000g to obtain a pellet. The cell pellets were sonicated in 0.5 ml of 50 mM TrisCl pH 7.4 and then centrifuged at 100,000g for 10 min. Supernatants (soluble fraction) were removed to clean tubes, and pellets (membrane fraction) were resonicated in 0.25 ml of TrisCl buffer. All samples were stored at −80°C until use. TABLE I. Enzyme Activity in Mammalian Prostate Tissues* Prostate tissues Specific activity (pmol/min/mg protein) Mouse Rat Guinea pig Cat Rabbit Dog Human Rat brain Membrane Soluble 1.87 ± .38 0.25 ± .04 4.01 ± 2.2 1.01 ± .30 0.66 ± .02 0.66 ± .16 6.15 ± 1.28 5.18 ± .21 0.38 ± .04 ND 0.53 ± .22 0.64 ± .30 0.54 ± .02 0.38 ± .05 3.16 ± .32 ND *ND, not detectable. Antibody Production A polyclonal antibody, GP-02, was produced by Covance Research Products (Denver, PA) against the external amino acid sequence position 390–409, SGAAVVHEIVRSFGTLKKEG . The peptide was synthesized by Princeton Biomolecules (Columbus, OH), and 0.25 mg were injected subcutaneously dorsally into 2 New Zealand white female rabbits. The animals were boosted with 0.25 mg or 0.4 mg of antigen every 3 weeks. Sera were collected 10 days postinjection and analyzed by Western blotting. Visualization of immunopositive bands from LNCaP and rat brain lysates at the approximate size of previously published reports of PSM/NAALADase was considered a positive response [8–10]. Production bleed and exsanguination followed analysis of the sixth test bleed. Western Blot Analysis Total cellular proteins (150 g) were resolved on a 7.5% SDS-PAGE gel and proteins were electroblotted to nitrocellulose membranes at 325 mA for 1 hr. Immunoblot analysis with the PSM antibody GP-02 was performed using standard protocols. Briefly, the blot was incubated overnight at 4°C in TBS-TM buffer (100 mM Tris, pH 7.6, 0.7 M NaCl, 1% Tween-20, and 10% evaporated milk) and then for 2 hr with a 1:1,000 dilution of PSM Ab GP-02 in TBS-TM. The membrane was washed six times for 5 min each in TBS-T (TBSTM without the milk) and then incubated for 1 hr with a 1:25,000 dilution of a horseradish peroxidase goat anti-rabbit antiserum in TBS-TM. The blot was washed five more times, and then Western blot reactions were detected by a chemiluminescent-based photoblot system. Protein Assay Proteins were assayed using the BioRad DC (BioRad Laboratories, Inc., Hercules, CA) procedure. NAAG Hydrolyzing Assay NAAG hydrolysis was performed essentially as described elsewhere , with minor modifications. Briefly, 50 mM TrisCl, pH 7.4, at 37°C, 1 mM CoCl2, and 10–100 g of tissue protein were preincubated for 10 min at 37°C; 50 l of 0.6 M 3H-NAAG radiolabelled on the terminal glutamate were added to each tube, and the incubation continued for 15 min. The reaction was stopped with 1 ml of ice-cold 100 mM NaPO4, and the cleaved glutamate was separated from unreacted substrate by ion exchange chromatography. When the effect of CoCl2 or NaCl was tested, the tissue was extracted and the reactions performed in 50 mM Tris-acetate. Kinetic Analysis Enzyme kinetics were examined using hot saturation of substrate, with 3H-NAAG concentrations from 10 nM to 2 M. Saturation isotherms and EadieHofstee graphs of kinetic data were generated in the Prism™ program from GraphPad (San Diego, CA). Statistical Analysis Statistical analysis was performed using a one-way ANOVA in the Origin graphics program from Microcal™ (Northampton, MA). RESULTS Prostate Tissues As shown in Table I, human prostate and each animal prostate assayed had measurable amounts of NAAG hydrolyzing activity. Significant amounts of activity were observed in mouse, guinea pig, and human prostate tissue, with specific activities of 1.87, Characterization of PSM Enzyme Activity TABLE II. Enzyme Activity in Prostate Cancer Cell Lines* Specific activity (pmol/min/mg) Prostate cell line Human I LNCaP PC-82 DuPRO NCI H660 Du145 PC-3 TsuPR1 Rat Dunning G H At-1 At-2 At-3 MatLu MatLyLu Rat brain Membrane Soluble 12.5 ± 2.8 58.0 ± 5.2 ND ND ND ND ND 4.3 ± 1.3 21.9 ± 1.3 ND ND ND ND ND 0.11 ± .04 ND 0.19 ± .02 ND ND ND ND 5.2 ± 0.2 ND ND ND ND ND ND ND ND *ND, not detectable. 4.01, and 6.15 pmol/min/mg, respectively, in the membrane fraction, while cat, rabbit, dog, and rat had activities of 1.0 pmol/min/mg or less. Only the human tissue displayed appreciable enzyme activity in the soluble fraction, 3.16 pmol/min/mg, while the enzymes from the other species had activities less than 1.0 pmol/min/mg. Activity was undetectable in cytosol from rat prostate. Prostate Cancer Cell Lines Four of the 14 prostate cancer cell lines analyzed displayed NAAG hydrolyzing activity (Table II). The human cell line LNCaP and the PC-82 xenograft had high specific activities in the membrane fractions (12.5 and 58.0 pmol/min/mg, respectively) and soluble fractions (4.3 and 21.9 pmol/min/mg, respectively). The human prostate cancer cell lines DuPRO, Du145, PC-3, and Tsu-PR1 displayed no detectable enzymatic activity in either the membrane or cytosolic fractions. Rat Dunning G and AT-1 cells exhibited low activity levels in the membrane fraction (0.11 and 0.19 pmol/ min/mg) and no detectable activity in the soluble fraction. Rat Dunning lines H, AT-2, AT-3, MatLu, and MatLyLu had no detectable hydrolyzing activity. 31 brain and human prostate cancer cell lines. The tissues and cell lines containing enzyme activity (i.e., rat brain, LNCaP, and Dunning G and AT-1 cells) exhibited positive bands ranging from 90–98 kDa, as shown in Figure 1. The human and rat cell lines which were negative for enzymatic activity (i.e., DuPRO, Du145, PC-3, Tsu-PR1, AT-2, AT-3, MatLu, and MatLyLu) had no detectable immunopositive proteins in this range as measured by Western blot analysis. Specifically, the immunopositive protein in rat brain was 92 kDa, in human prostate cancer 98 kDa, and in rat Dunning prostate cancer 90 kDa. A separate immunoreactive band at 105 kDa was apparent in three of the rat Dunning lines and is currently under investigation. Pharmacological Characterization of PSM Enzyme Activity The PSM enzyme activity observed in prostate cancer cell lines had similar sensitivities to those of three structurally distinct NAALADase inhibitors including NaPO4, quisqualate, and 2-(phosphonomethyl)pentanedioic acid (2-PMPA) . Enzyme activity in PC82 tissue, Dunning G, AT-1, and LNCaP cells was inhibited by NaPO4, quisqualate, and 2-PMPA, with Ki values ranging from 170–312 M, 1.5–6.3 M, and 0.6– 4.6 nM, respectively (Table III). Because LNCaP cells exhibited significant enzyme activity and were easily grown in cell culture, we used LNCaP cells as a prototype to fully characterize the enzymatic activity of the membrane and soluble forms of PSM. A list of inhibitors and cofactors that were evaluated is provided in Table IV. The prostate enzyme, like brain NAALADase, was found to be chloride- and cobalt-simulated (2-fold and 10-fold stimulation at 10 mM, respectively). Both tissue sources exhibited inhibition of enzyme activity by quisqualate (Ki, 3–5 M), NAAG-related peptides (Ki, 27–38 M), 2-PMPA (Ki, 1–5 nM), and EDTA (Ki, 600–1,000 M). With regard to substrates, both rat brain NAALADase and the two forms of PSM showed nM sensitivities to methotrexate triglutamate, folate pentaglutamate, and NAAG; however, there were two statistically significant differences. First, the brain enzyme was less sensitive to folate pentaglutamate as compared to the membrane-bound prostate enzyme (884 nM vs. 196 nM, P < 0.04) and the soluble prostate enzyme (884 nM vs. 174 nM, P < 0.04). Secondly, the membrane form of PSM was more sensitive than the soluble form of the enzyme to NAAG (0.67 M vs. 1.69 M, P < 0.02), although neither was statistically different from the brain enzyme (1.09 M). Western Blot Analysis Kinetic Analysis Immunopositive proteins were detected by antibody GP-02 on Western blot using extracts from rat The kinetic analysis of LNCaP PSM enzyme activity is shown in Figure 2. Rat brain enzyme displayed 32 Tiffany et al. Fig. 1. Western blot analysis of PSM antibody GP-02 in rat brain and human and rat prostate cancer cell lines. The antibody recognized immunoreactive proteins in rat brain (92 kDa), LNCaP cells (98 kDa), and rat Dunning G and AT-1 cells (90 kDa), consistent with the molecular weight range previously reported for PSM/NAALADase. An unexplained 105-kDa immunoreactive positive protein was also found in AT-1, MatLu, and MatLyLu and is currently under investigation. TABLE III. Comparison of Enzyme Activity in Prostate Cancer Cell Lines and Rat Brain Kivalues Tissue source Rat brain LNCaP Dunning G Dunning AT-1 PC-82 Membrane Membrane soluble Membrane Membrane Membrane soluble NaPO4 Quisqualate 2-PMPA 170 M ± 12 253 M ± 22 298 M ± 35 210 M ± 60 312 M ± 54 175 M ± 27 171 M ± 12 6.3 M ± 2.0 3.6 M ± 1.4 5.8 M ± 2.1 2.3 M ± 1.0 1.7 M ± 0.9 2.2 M ± 0.9 1.5 M ± 0.5 1.1 nM ± 0.2 1.9 nM ± 0.4 2.4 nM ± 0.6 3.7 nM ± 1.2 4.6 nM ± 2.0 1.0 nM ± 0.5 0.6 nM ± 0.2 single-site kinetics with a Km of 0.5 M and Vmax at 200 pmol/min/mg (Fig 2A); these data are similar to previously reported values . LNCaP membrane PSM also displayed single-site kinetics, with a Km of 0.2 M and a Vmax of 80 pmol/min/mg (Fig. 2B). Interestingly, the soluble prostate enzyme exhibited two enzyme sites with Km values of 0.2 M and 1.2 M and Vmax values of 40 and 125 pmol/min/mg, respectively (Fig. 2C). DISCUSSION The prostate cancer marker prostate-specific membrane antigen (PSM) was recently cloned and reported to have significant DNA homology (86%) to the rat brain carboxypeptidase termed NAALADase . NAALADase was identified over 15 years ago and has been extensively characterized. It is a type II integral membrane protein and a zinc metalloproteinase which exhibits Cl− dependence, Co2+ stimulation, and inhibition by M phosphate, M quisqualate, and nM 2-PMPA. It is localized in the brain, kidney, prostate, and brush border membrane of the intestine [9–11,13]. In this paper, we pharmacologically and kinetically characterized the ability of the prostate enzyme to hydrolyze NAAG and compared this enzyme activity to that of rat brain NAALADase. Normal prostate tissues from eight mammalian species examined displayed NAAG hydrolyzing in both membrane and cytosolic fractions. The only exception was rat prostate, which exhibited no detectable activity in the cytosol. Of the malignant cell lines examined, only the human LNCaP, PC-82, and rat Dunning G and AT-1 cell lines had measurable enzyme activity; only the human lines possessed both membrane and soluble forms of the enzyme. These Characterization of PSM Enzyme Activity 33 TABLE IV. Pharmacological Comparison of Enzymatic Activity of LNCaP Cells and Rat Brain Kivalues (M) LNCaP Inhibitors Reducing agents Dithiothreitol Chelators EDTA EGTA Membrane Soluble Rat brain membrane >1,000 >1,000 >1,000 667 ± 110 847 ± 250 Peptides NAAG Asp-glu Glu-glu Glutamate NAA 0.670 ± 0.110** 32.1 ± 11.0 38.6 ± 13.7 132 ± 33.0 >1,000 Chemotherapeutics Methotrexate Triglutamate Methotrexate 1,020 ± 150 872 ± 125 775 ± 264 411 ± 48.0 1.69 ± 0.360 27.0 ± 6.00 34.4 ± 7.30 132 ± 28.0 >1,000 1.09 ± 0.170 32.5 ± 9.0 22.0 ± 7.50 125 ± 63.0 >1,000 0.196 ± 0.059 30.3 ± 6.70 0.170 ± 0.008 38.2 ± 2.60 0.190 ± 0.046 42.0 ± 2.0 Folate derivatives Folate pentaglutamate Folic acid 0.196 ± 0.029* 22.7 ± 8.30 0.174 ± 0.007* 25.1 ± 4.50 0.884 ± 0.209 27.7 ± 4.70 Miscellaneous Quisqualate 2-PMPA 3.60 ± 1.40 0.002 ± 0.0004 5.80 ± 2.10 0.002 ± 0.0006 5.50 ± 1.70 0.001 ± 0.0002 1 2.31 ± 0.4 stim@ 10 mM 13.2 ± 4.1 stim@ 10 mM 1.9 ± 0.2 stim@ 10 mM 8.9 ± 1.7 stim@ 10 mM 1.3 ± 0.2 stim@ 10 mM 3.0 ± 0.7 stim@ 10 mM Activators Metal ions NaCl CoCl2 *P 艋 .04 vs. rat brain. **P = .02 vs. supernatant. 1 = stimulation of enzymatic activity. data are in agreement with previously published reports that PSM is expressed in normal and prostate cancer tissues . Additionally, it has been reported that increased PSM gene expression is observed in prostate cancer [5,6]. We corroborate these results by demonstrating an increase in enzymatic activity in two human prostate cancer cell lines as compared to normal human prostate tissues. Using a newly developed antibody to PSM, GP-02, we performed Western blot analysis on rat brain tissue and prostate cancer cell lines. Consistent with the enzymatic activity reported, only rat brain, LNCaP, and Dunning G and AT-1 cell lines displayed reactive bands in the size range previously reported for PSM/ NAALADase [8–10,18]. Our antibody reacted with a 92-kDa rat brain protein, a 98-kDa species in the LNCaP cells, and a 90-kDa species in the rat Dunning cell lines. The varying molecular size of PSM may be due to different posttranslational modifications, as has previously been described . It has been reported that the 750-amino-acid PSM protein before posttranslational modification is 84 kDa, but with the addition of N-linked sugar complexes the molecular weight is 110 kDa . A larger molecular weight protein that was reactive to GP-02 was also identified in the Dunning AT-1, MatLu, and MatLyLu cell lines. These cell lines did not have PSM enzyme activity but it is possible that they contain a nonfunctional form of the enzyme or a protein highly homologous to PSM. This is currently under investigation. We used the LNCaP cell line to pharmacologically characterize the prostate enzyme and compare it to rat brain NAALADase. The two enzymes were found to be remarkably similar in their sensitivities to metal ions, EGTA, PO4, quisqualate, the potent NAALADase inhibitor 2-PMPA , and the chemotherapeutic agent methotrexate and its polyglutamated form, methotrexate-triglutamate. The similarities observed 34 Tiffany et al. Fig. 2. Saturation isotherm of NAAG hydrolyzing activity in (A) rat brain membranes (Km, 0.5 µM; Vmax, 200 pmol/min/mg), (B) LNCaP membranes (Km, 0.2 µM; Vmax, 80 pmol/min/mg), and (C) LNCaP supernatants (Km1, 0.2 µM; Vmax1, 40 pmol/min/mg; Km2, 1.2 µM; Vmax2, 125 pmol/min/mg). Analysis was performed with 3 H-NAAG concentrations from 10 nM to 2 µM. Graphs are representative of independent experiments performed in duplicate. are remarkable because the enzymes compared were from two tissue sources (brain vs. prostate) and two different species (human vs. rat). The two enzymes were divergent only with regard to substrate specificity. LNCaP enzyme activity, both membrane and cytosolic, had a statistically significant higher affinity for folate pentaglutamate than brain NAALADase. These Ki values support the theory of different functional roles of brain and prostate enzymes. Specifically, the brain enzyme may be primarily NAAG-hydrolyzing, while prostate PSM may hydrolyze folate polyglutamate as part of its functional role. Pinto et al.  hypothesized that PSM carboxypeptidase activity is necessary for folate polyglutamate hydrolysis. Since folic acid is necessary for cell growth and polyglutamated folic acid is not transported into the cell, a glutamate-hydrolyzing enzyme would be necessary to internalize folic acid. Given that folic acid metabolism plays a large role in cell division, folate polyglutamate hydrolysis, and thus PSM enzymatic action, may play an important role in the progression of prostate cancer. However, the high expression of functional PSM activity present in normal prostate tissue from postmortem subjects suggests that there is a role for PSM enzymatic action in normal prostate tissue as well. It may be reasonable to propose that the liberation of glutamate acts as a signal in prostate tissue metabolism, similar to the action of glutamate as a transmitter in neuronal tissue. In fact, Heston et al. recently reported the presence of ionotropic glutamate receptors in prostate tissue . Rat brain NAALADase also differs from PSM and PSM⬘ in its enzyme kinetics. The rat brain enzyme displayed single-site kinetics, with Km of 0.5 M and Vmax of 200 pmol/min/mg, as previously described . The prostate membrane PSM kinetics were also single-site; the Km was 0.2 M and the Vmax was 80 pmol/min/mg, possibly suggesting that the prostate enzyme may be a less active hydrolase than the brain enzyme. With regard to the prostate soluble enzyme, PSM⬘, we report for the first time that the soluble enzyme had two catalytic sites. The first site was comparable to membrane PSM (Km = 0.2 M and Vmax = 40 pmol/min/mg), while the second site appeared more similar to the brain enzyme (Km = 1.2 M and Vmax = 125 pmol/min/mg). The significance of this difference and the relationship, if any, between the prostate-localized peptidase and brain NAALADase remain to be discovered. It is also important to note that the soluble form of the enzyme in LNCaP cells displayed nearly identical Ki values to those of the membrane enzyme in LNCaP cells. We believe it is the cytosolic form of PSM or PSM⬘. However, this has not been determined conclusively, and we admit the possibility that the activity may be due to an as-yet uncharacterized enzyme or variant of PSM. It is noteworthy that other metalloproteinases, such as angiotensin I converting enzyme (ACE) and neutral endopeptidase (enkephalase), were originally thought to be brain-specific and were later discovered to have widespread tissue distribution. Both ACE  and neutral endopeptidase (NEP) , like PSM, are abun- Characterization of PSM Enzyme Activity dant in prostate tissue. Interestingly, NEP activity is associated with androgen-dependent prostate cancer cell lines, and loss of NEP activity can be correlated with both androgen-insensitivity and metastatic prostate cancer cells . 8. CONCLUSIONS PSM displayed NAAG hydrolyzing activity in all normal prostate tissues examined and several prostate cancer cell lines. Pharmacologically, PSM enzyme activity was remarkably similar to that of the rat brain enzyme NAALADase. Ions, chelators, and known inhibitors and cofactors of brain NAALADase had similar effects on the prostate enzyme. Although the two enzymes were found to be highly homologous, several differences were discovered. First, unlike the brain enzyme which is exclusively present in the membrane fraction, the prostate enzyme has both membrane and cytosolic forms, both of which display functional enzyme activity. Second, the enzymes differ pharmacologically in regard to substrates. The prostate enzyme exhibited a higher affinity for folate polyglutamate than the brain enzyme, which may suggest unique functional roles. Last, variations in kinetic patterns were also noted. The brain enzyme had a lower Km for NAAG and a higher Vmax than the prostate enzyme; these may also relate to differences in their function. The apparent role of the metalloproteinase PSM remains to be elucidated. REFERENCES 1. Carter RF, Feldman AR, Coyle JT. Prostate-specific membrane antigen is a hydrolase with substrate and pharmacologic characteristics of a neuropeptidase. Proc Natl Acad Sci USA 1996; 93:749–753. 2. Heston WDW. Characterization and glutamyl preferring carboxypeptidase function of prostate-specific membrane antigen: a novel folate hydrolase. Urology [Suppl]1997;49:104–112. 3. Fair WR, Israeli RS, Heston WDW. Prostate-specific membrane antigen. Prostate 1997;32:140–148. 4. Israeli RS, Powell CT, Fair WR, Heston WDW. Molecular cloning of a complementary DNA encoding a prostate-specific membrane antigen. Cancer Res 1993;53:227–230. 5. Su SL, Huang IP, Fair WR, Powell CT, Heston WDW. Alternatively spliced variants of prostate-specific membrane antigen RNA; ratio of expression as potential measure of progress. Cancer Res 1995;55:1441–1443. 6. Kawakami M, Nakayama J. Enhanced expression of prostatespecific membrane antigen gene in prostate cancer a revealed by in situ hybridization. Cancer Res 1997;57:2321–2324. 7. Loric S, Dumas F, Eschwege P, Blanchet P, Benoit G, Jardin A, 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 35 Lacour B. Enhanced detection of hematogenous circulating prostate cells in patients with prostate adenocarcinoma by using nested reverse transcription polymerase chain reaction assay based on prostate-specific membrane antigen. Clin Chem 1995; 41:1698–1704. Rochon YP, Horoszewicz JS, Boynton AL, Holmes EH, Barron RJ III, Erickson SJ, Kenny GM, Murphy GP. Western blot assay for prostate-specific membrane antigen in serum of prostate cancer patients. Prostate 1994;25:219–223. Slusher BS, Robinson MB, Tsai G, Simmons ML, Richards SS, Coyle JT. Rat brain N-acetylated alpha-linked acidic dipeptidase activity. J Biol Chem 1990;265:21297–21301. Robinson MB, Blakely RD, Couto R, Coyle JT. Hydrolysis of the brain dipeptide N-acetyl-aspartyl-L-glutamate. J Biol Chem 1987;262:14498–14506. Slusher BS, Tsai G, Yoo G, Coyle JT. Immunocytochemical localization of the N-acetyl-aspartyl-glutamate hydrolyzing enzyme N-acetylated alpha-linked acidic dipeptidase [NAALADase]. J Comp Neurol 1992;315:217–229. Hollmann M, Heinemann S. Cloned glutamate receptors. Annu Rev Neurosci 1994;17:31–108. Stauch BL, Robinson MB, Forloni G, Tsai G, Coyle JT. The effects of N-acetylated alpha-linked acidic dipeptidase [NAALADase] inhibitors on [3H]NAAG catabolism in vivo. Neurosci Lett 1989; 100:295–300. Pinto JT, Suffoletto BP, Berzin TM, Qiao CH, Lin S, Tong WP, May F, Mukherjee B, Heston WDW. Prostate-specific membrane antigen: a novel folate hydrolase in human prostatic carcinoma cells. Clin Cancer Res 1996;2:1445–1451. Hoehn W, Schroeder FH, Riemann JF, Joebsis AC, Hermanek P. Human prostatic adenocarcinoma: some characteristics of a serially transplantable line in nude mice (PC-82). Prostate 1980;1: 95–104. Jackson PF, Cole DK, Slusher BS, Stetz SL, Ross LE, Donzanti BA, Trainor DA. Design, synthesis, and biological activity of a potent inhibitor of the neuropeptidase N-acetylated ␣-linked acidic dipeptidase. J Med Chem 1996;39:619–622. Silver DA, Pellicer I, Fair WR, Heston WDW, Cordon-Cardo C. Prostate-specific membrane antigen expression in normal and malignant tissues. Clin Cancer Res 1997;3:81–85. Holmes EH, Greene TG, Tino WT, Boynton AL, Aldape HC, Misrock SL, Murphy GP. Analysis of glycosylation of prostate specific membrane antigen derived from LNCaP cells, prostatic carcinoma tumors, and serum from prostate cancer patients. Prostate [Suppl] 1996;7:25–29. Heston WDW, Silver DA, Pellicer I, Fair WR, Cordon-Cardo C. Ionotrophic glutamate receptor distribution in prostate tissue. J Urol 1996;155:513. Corvol P, Michaud A, Soubrier F, Williams TA. Recent advances in knowledge of the structure and function of the angiotensin I converting enzyme. J Hypertens 1995;13:53–60. Krongrad A, Atochina E, Ryan JW, Roos BA. Endopeptidase 24.11 activity in the human prostate cancer cell lines LNCaP and PPC-1. Urol Res 1997;25:113–116. Papandreou CN, Usmani B, Geng Y, Bogenreider T, Freeman R, Wilk S, Finstad CL, Reuter VE, Powell CT, Scheinberg D, Magill C, Scher HI, Albino AP, Nanus DM. Neutral endopeptidase 24.11 loss in metastatic human prostate cancer contributes to androgen-independent progression. Nat Med 1998;4:50–57.