Linked Acidic Dipeptidase and Dipeptidyl - The Journal of Biological ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 1999 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 274, No. 13, Issue of March 26, pp. 8470 –8483, 1999 Printed in U.S.A.

Isolation and Expression of Novel Human Glutamate Carboxypeptidases with N-Acetylated a-Linked Acidic Dipeptidase and Dipeptidyl Peptidase IV Activity* (Received for publication, September 11, 1998, and in revised form, November 6, 1998)

Menelas N. Pangalos‡§, Jean-Marc Neefs‡, Marijke Somers¶, Peter Verhasselt‡, Mariette Bekkers‡, Liesbet van der Helm‡, Erwin Fraiponts¶, David Ashtoni, and Robert D. Gordon‡ From the Janssen Research Foundation, Departments of ‡Biotechnology, ¶Biochemistry, and iNeuropsychopharmacology, B2340 Beerse, Belgium

Hydrolysis of the neuropeptide N-acetyl-L-aspartyl-Lglutamate (NAAG) by N-acetylated a-linked acidic dipeptidase (NAALADase) to release glutamate may be important in a number of neurodegenerative disorders in which excitotoxic mechanisms are implicated. The gene coding for human prostate-specific membrane antigen, a marker of prostatic carcinomas, and its rat homologue glutamate carboxypeptidase II have recently been shown to possess such NAALADase activity. In contrast, a closely related member of this gene family, rat ileal 100-kDa protein, possesses a dipeptidyl peptidase IV activity. Here, we describe the cloning of human ileal 100-kDa protein, which we have called a NAALADase“like” (NAALADase L) peptidase based on its sequence similarity to other members of this gene family, and its inability to hydrolyze NAAG in transient transfection experiments. Furthermore, we describe the cloning of a third novel member of this gene family, NAALADase II, which codes for a type II integral membrane protein and which we have localized to chromosome 11 by fluorescent in situ hybridization analysis. Transient transfection of NAALADase II cDNA confers both NAALADase and dipeptidyl peptidase IV activity to COS cells. Expression studies using reverse transcription-polymerase chain reaction and Northern blot hybridization show that NAALADase II is highly expressed in ovary and testis as well as within discrete brain areas.

The neuropeptide N-acetyl-L-aspartate-L-glutamate (NAAG)1 is expressed both in the central nervous system and in the periphery. NAAG has been localized to specific sub-populations of neurones

* 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. The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EBI Data Bank with accession number(s) HSA012370, AJ012370, HAS012371, AJ012371. § To whom correspondence should be addressed. SmithKline Beecham Pharmaceuticals, Neuropharmacology Research, Harlow, Essex CM19 5AW, U. K.; E-mail: [email protected]. 1 The abbreviations used are: NAAG, N-acetyl-L-aspartyl-L-(3,4)-glutamate; EST, expressed sequence tag; FISH, fluorescent in situ hybridization; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; NAALADase, N-acetylated a-linked acidic dipeptidase; ORF, open reading frame; DPP IV, dipeptidyl dipeptidase IV; bp, base pair(s); PCR, polymerase chain reaction; PSM, prostate-specific membrane antigen; QA, quisqualate; 59-RACE, 59-rapid amplification of cDNA ends; RT-PCR, reverse transcription-polymerase chain reaction; SDM, site-directed mutagenesis; AMC, 7-amino-4-methylcoumarin; mGluR, metabotropic glutamate receptor.

and is one of the most abundant peptides in brain, being present in millimolar concentrations in certain brain regions (for review see Ref. 1). Clarification of the physiological role of NAAG has been difficult because it is co-localized with glutamate, however, it does fulfill some of the criteria for a neurotransmitter. It is localized in synaptic vesicles and is released in Ca21-dependent manner from nerve terminals (1). In the central nervous system NAAG has been shown to act as a weak partial agonist at N-methyl-D-aspartate receptors but not at a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid or kainate receptors (2– 4). NAAG attenuates N-methyl-Daspartate- or glutamate-induced neurodegeneration, and its addition to hippocampal slices or mixed cortical cultures results in neuroprotection following addition of excitotoxins, via a mechanism distinct from the action of NAAG on the N-methylD-aspartate receptor (5). In this case it is postulated that NAAG, which is poorly transported or actively taken up, diffuses from the synaptic cleft and binds as an agonist to type II metabotropic glutamate receptors (mGluR), such as mGluR 3, leading to reduced glutaminergic neurotransmission (6, 7). This hypothesis is supported by studies using the mGluR 3 antagonist, ethyl glutamate, which eliminated the neuroprotective actions of NAAG (5). The N-acetylated a-linked acidic dipeptidase (NAALADase) cleavage of NAAG was first reported by Robinson et al. (8, 9) and shown to be sensitive to the synthetic glutamatergic agonist quisqualic acid (QA). Subsequently, a 94-kDa membrane glycoprotein purified from rat brain was shown to possess an NAALADase-type activity similar to that previously described from brain extracts (10). Antisera raised to this protein were used by Carter et al. (11) to screen rat brain expression libraries resulting in the cloning of a novel rat partial cDNA. The full-length coding sequence was cloned, and the protein it encoded was termed glutamate carboxypeptidase II (EC 3.4.17.21; also called NAAG peptidase; 12, 13). This rat gene sequence was 86% identical to the human prostate-specific membrane antigen (PSM), a cDNA of previously unknown function, which is highly expressed in prostate tumors (14). Transfection studies with PSM into NAALADase-negative cell lines conferred a NAAG-hydrolyzing activity to these cells that could be inhibited by the NAALADase inhibitor QA, demonstrating the first functional expression of a NAALADase cDNA clone (11). PSM has also been shown to possess a second enzymatic activity, that of a pteroyl-poly-g-glutamylcarboxypeptidase (15). Recently, a novel 100-kDa glycoprotein has been cloned, from rat ileal brush border membranes, which is homologous to but distinct from human PSM and rat glutamate carboxypeptidase II based on amino acid sequence. This protein possesses a dipeptidyl peptidase IV (DPP IV) type activity,

8470

This paper is available on line at http://www.jbc.org

Characterization of Novel NAALADases

8471

FIG. 1. Nucleotide sequence and amino acid sequence of human NAALADase II. The nucleotide and predicted one-letter code amino acid sequence are shown. The putative membrane spanning domain, deduced from hydrophilicity plots, is marked by a line. Potential N-linked glycosylation sites are in shaded squares. A putative Zn21 binding domain is marked within open boxes, and residues important in a putative a/b hydrolase catalytic site are within shaded circles. Base pairs are numbered in the right margin.

being able to hydrolyze Gly-Pro 7-amido-4-methylcoumarin, but has no reported NAALADase activity (16). For ease of understanding we have simplified the nomenclature used in this report to reflect the relatedness of these protein sequences. Human PSM and rat glutamate carboxypeptidase II (or rat NAAG peptidase) are termed human and rat NAALADase I, respectively. We have also termed ileal 100-kDa protein an NAALADase-“like” peptidase (NAALADase L) based on its sequence similarity to NAALADase I but lack of any functional NAALADase activity. In this study we describe the cloning, expression, and characterization of a third novel member of this enzyme family,

human NAALADase II, that is similar to but distinct from NAALADase I and NAALADase L. In addition we have identified, expressed, and characterized the full coding sequence of human NAALADase L. EXPERIMENTAL PROCEDURES

Sequence Similarity Searching for NAALADase L Molecules—Using the complete human (GenBankTM accession number M99487), rat (GenBankTM accession number RNU75973), and mouse (GenBankTM accession number AF026380) NAALADase I protein sequences, the complete rat NAALADase L (GenBankTM accession number AF009921) protein sequence and a partial human NAALADase L protein sequence (GenBankTM accession number AF010141) as query sequences, a Basic Local

8472


FIG. 2. Nucleotide sequence and amino acid sequence of human NAALADase L. The nucleotide and predicted single letter code amino acid sequence are shown. The putative membrane spanning domain, deduced from hydrophilicity plots, is marked by a line. Potential N-linked glycosylation sites are shaded. A putative Zn21 binding domain is marked within open boxes, and residues important in a putative a/b hydrolase catalytic site are within shaded circles. Base pairs are numbered in the right margin.

Alignment Search Tool (17) search was performed on the WashU Merck expressed sequence tag (EST) data base and on a proprietary LifeSeqTM human EST data base. Five EST clones 4190746, 1547649, 3448872, 3608639, and 1333965 were ordered from Incyte Pharmaceuticals. The DNA insert of each clone was sequenced on both strands using Applied Biosystems prism BigDye Terminator Cycle sequencing kits and an Applied Biosystems 377XL sequencer (Perkin-Elmer).

NAALADase Cloning NAALADase I—Sequence data from human NAALADase I (GenBankTM accession number M99487) was used to design primers to amplify the complete coding sequence of NAALADase I by PCR. Primers used were NAALD1S2 (BamHI) 5 59-CCC GGATCC GAG ATG TGG ATT CTC CTT CAC GAA AC-39 and NAALD1AS2(XhoI) 5 59-CCC CTCGAG TTA GGC TAC TTC ACT CAA AGT CTC TGC -39 (restriction sites to be introduced are underlined). PCR amplification was per-

formed using Marathon-Ready human prostate cDNA (CLONTECH) and the Expand High Fidelity PCR system (Boehringer Mannheim) with primers NAALD1S1(BamHI) and NAALD1AS1(XhoI) according to the manufacturer’s instructions. All PCR reactions began with an initial denaturation step (94 °C for 5 min) prior to addition of enzyme, followed by 30 cycles of amplification (45 s at 94 °C, 1 min at 55 °C, and 1 min 48 s at 68 °C) and ended with a final extension step (7 min at 72 °C). A 2303-bp PCR fragment was cloned into pCR2.1 according to manufacturer’s instructions (Invitrogen). A full-length clone containing a single PCR-induced error, at position 1183, was corrected by sitedirected mutagenesis (SDM) using the QuickChange SDM Kit (Stratagene). Primers designed for the SDM reactions were NAALD1-SDMS1 5 59-CCC TCA GAG TGG AGC AGC TGT TGT TCA TGA AAT TGT GAG G-39 and NAALD1-SDM-AS1 5 59-CCT CAC AAT TTC ATG AAC AAC AGC TGC TCC ACT CTG AGG G-39. NAALADase L—Sequence data from partial human NAALADase L


8473

FIG. 3. Alternative splicing of NAALADase L. Amino acid sequence for NAALADase L is shown. Sites where putative DNA sequences are spliced out are marked by an arrow with the resulting (in-frame) amino acid deletions highlighted in bold italics. Sites of putative intronic DNA insertion are marked by triangles, with the intronic DNA sequence shown above. Resulting changes to the amino acid sequence are highlighted in bold italics. Amino acid residues are numbered in the right margin. (GenBankTM accession number AF10141) was used to design primers to amplify the 39 end of NAALADase L by PCR. First round PCR amplification was performed with NAALDLS1 5 59-GTT CTT CAA CAA GCT GCA GGA GCG -39 and NAALDLAS1(XhoI) 5 59- CCC CTCGAG CCG GAG TAA AGG GAG GGC TGA AG-39. Second round PCR amplification was performed with nested primers NAALDLS2 5 59-GGC GAC CTG AGC ATC TAC GAC AAC-39 and NAALDLAS2 (XhoI) 5 59- CCC CTCGAG TCC CCT CAG AGG TCA GCC ACA G-39. Cycling conditions were similar to those described above (30 cycles of 45 s at 94 °C, 1 min at 57 °C, and 1 min at 72 °C). PCR products derived from MarathonReady human small intestine cDNA were cloned into pCR2.1 and extended to the translation termination codon of NAALADase L. To obtain unknown 59 coding sequence for human NAALADase L, antisense primers were designed for the 59-rapid amplification of cDNA ends (59-RACE); antisense primers NAALDLAS5 5 59-CTG CAG CTT GTT GAA CTC TTC TGT G-39 and NAALDLAS6 5 59-CAA ACA CGA

TTG ATC TGC GAG GAC-39 were synthesized for amplifications using various human Marathon-ReadyTM cDNAs. Cycling conditions were as described previously. PCR products derived the small intestine cDNA amplifications were cloned into pCR2.1 and found to extend the coding sequence beyond the putative translation start codon and into part of the 59-untranslated region. To construct a full-length NAALADase L clone, primers were designed to introduce a unique restriction site (MunI) into the DNA sequence of NAALADase L without changing the amino acid sequence of the ORF. The first primer set was NAALDLS3 (EcoRV) 5 59-CGGATATCC GCA GGA TGC AGT GGA CGA AG-39 and NAALDLAS8 (MunI) 5 59-CAA ACA CAATTG ATC TGC GAG GAC GC-39 and the second primer set was NAALDLS8 (MunI) 5 59-GCG TCC TCG CAG ATCAATTGT GTT TG-39 and NAALDLAS1 (XhoI). The base change introducing the MunI site is marked in bold. PCR amplification was performed on the 39 end clone with primers NAALDLS3 (EcoRV) and

8474


FIG. 4. Alignment of the predicted protein sequences for human NAALADases I, II, and L. The amino acid sequences were aligned using the ClustalW alignment program. Amino acid residues identical to all three proteins are shaded in black. Amino acid residues identical to two of the proteins are shaded in gray. A putative Zn21 peptidase domain is highlighted between arrows and was identified by comparison to yeast and bacterial aminopeptidases. Putative residues involved the catalytic site of the a/b hydrolase fold family of proteins are marked by three arrows (nucleophile acid base). Amino acid residues are numbered in the right margin.

NAALDLAS8 (MunI) and on the 59 end clone with primers NAALDLS8 (MunI) and NAALDLAS1 (XhoI). PCR products were cloned into pCR2.1 and sequence-verified as described previously. It should also be noted that amplifications using NAALADase Lspecific primers resulted in a number of PCR products of unexpected size using different cDNAs. These PCR products were cloned and sequenced in order to identify possible splice variants. NAALADase II—Sequencing results from Incyte clone 3608639, derived from a human lung cDNA, suggested that this clone contained a DNA sequence spanning the complete coding region of a putative NAALADase-like molecule (NAALADase II) similar to but distinct from NAALADase I and NAALADase L. Sequential 59-RACE PCR performed with antisense primers NAALD2AS1 5 59-CTT TGA TGA TAG CGC ACA GAA GTG G-39 and NAALD2AS2 5 59-GGA AAG ATG CCA GCG CAG GAC-39 failed to identify any potential upstream initiation codons in amplifications using brain, fetal brain, prostate, small intestine, and colon human Marathon-Ready cDNAs.

Activity Determinations of NAALADases Transiently Expressed in COS Cells Subcloning of NAALADases into Expression Vectors—NAALADase I, II, and L clones were subcloned into the cytomegalovirus promoterbased plasmid pcDNA3. NAALADase I/pCR2.1 was digested with

BamHI/XhoI to excise the complete NAALADase I sequence. NAALADase II/pINCYTE was digested with EcoRI to excise the complete NAALADase II sequence. NAALADase L-59/pCR2.1 was digested with EcoRV/MunI, and NAALADase L-39/pCR2.1 was digested with MunI/ XhoI to excise the two halves of complete NAALADase L sequence. All expression constructs subcloned into pcDNA3 were verified by full sequence analysis. Transient Transfection into COS Cells—COS cells were maintained in complete medium (defined minimal essential medium supplemented with 10% fetal calf serum, 13 non-essential amino acids, and a 13 streptomycin/penicillin/glutamine mix). Cell titer was determined in a Coulter counter, and cells were seeded in six-well plates at a density of 15,000 cells/cm2 and allowed to reach approximately 80% confluence. For each transfection 6 ml of FuGENE6 (Boehringer Mannheim, Germany) was added to 96 ml of serum-free medium and incubated for 5 min at 20 °C. This preparation was added to a second tube containing 1 mg of NAALADase/pcDNA3 DNA, mixed gently, and allowed to stand for 15 min at room temperature. The DNA/FuGENE6/serum-free medium mix was pipetted into a well containing 2 ml of fresh complete medium. Cells were incubated for 72 h in a 37 °C incubator before harvesting. Determination of Biological Activity of NAALADase Homologues— Transfected COS cell pellets were scraped with 50 mM Tris-HCl (pH

Characterization of Novel NAALADases 7.4), 0.1% Triton X-100 and vortexed. Homogenates were put through at least one freeze/thaw cycle in liquid N2 before assay. Assay were performed as described previously (9) for each NAALADase, using equivalent numbers of cells or equivalent protein content, using N-acetyl-Laspartyl-L-3,4-[3H]glutamate ([3H]NAAG). Assays were initiated by the addition of the membrane homogenates to the pre-warmed assay mixture and incubated at 37 °C for various times. Reactions were terminated by addition of ice-cold 250 mM potassium phosphate and loaded onto 4-cm anion exchange mini-columns (Bio-Rad AG1-X8). [3H]Glutamate was eluted off the column with 0.5 M formic acid and counted in a scintillation counter (Packard). Inhibition curves were performed under similar conditions with increasing concentrations of QA. Assays were performed on cell samples from at least three independent transient transfections. DPP IV activity was determined by fluorescent analysis (excitation at 335 nm and emission 450 nm) of the hydrolysis of Gly-Pro-AMC as described previously (16). Assays were initiated by the addition of cellular homogenates (135 mg of total protein per reaction) to the buffered substrate solution (100 mM in 150 mM glycine, pH 8.5) in a total reaction volume of 100 ml and followed for 40 min at 37 °C. Chromosomal Localization of NAALADases by Fluorescent in Situ Hybridization (FISH) Analysis—Chromosomal mapping studies were carried out by SeeDNA Biotech (Ontario, Canada) using FISH analysis. Slide Preparation—Lymphocytes isolated from human blood were cultured in a-minimal essential medium supplemented with 10% fetal calf serum and phytohemagglutinin at 37 °C for 68 –72 h. The lymphocyte cultures were treated with bromodeoxyuridine (0.18 mg/ml, Sigma) to synchronize the cell population. The synchronized cells were washed three times with serum-free medium to release the block and re-cultured at 37 °C for 6 h in Slide Preparation—Lymphocytes isolated from human blood were cultured in a-minimal essential medium supplemented with 10% fetal calf serum and phytohemagglutinin at 37 °C for 68 –72 h. The lymphocyte cultures were treated with bromodeoxyuridine with thymidine (2.5 mg/ml). Cells were harvested and slides were prepared using standard procedures including hypotonic treatment, fixations, and air drying. FISH Detection—NAALADase I (bp 1–2252) and NAALADase II (bp 1–2503) probes and a partial NAALADase L probe (bp 1220 –2276) were biotinylated with dATP for 1 h at 15 °C using the BioNick labeling system (Life Technologies, Inc.). The procedure for FISH detection was performed as described previously (17, 18).

Tissue Distribution of NAALADases NAALADase II Gene Expression by Northern Blot—Human multiple tissue Northern blots (CLONTECH) containing 2 mg of poly(A)1 RNA derived from non-neuronal tissues were hybridized in ExpressHyb hybridization solution (CLONTECH) for 2 h at 68 °C. A 546-bp NAALADase II fragment, isolated from NAALADase II/pcDNA3 by digestion with EcoRI and BglII, was radiolabeled using a Rapid Multiprime Labeling kit (Amersham Pharmacia Biotech) and [32P]dCTP (NEN Dupont). Unincorporated label was removed using a Microspin S-200 column (Amersham Pharmacia Biotech), and the denatured probe (specific activity 5 2.6 3 108 dpm/mg) was incubated overnight at 68 °C in ExpressHyb. Washes were performed at high stringency (55 °C in 0.13 sodium chloride sodium citrate, 0.1% SDS) and blots exposed to X-Omat AR Film (Eastman Kodak) for 2 days at 270 °C with two intensifying screens. NAALADase I, II, and L Gene Expression Analysis by RT-PCR— 0ligonucleotide primers were designed for the specific amplification of a PCR fragment for each NAALADase; NAALADase I primers were NAALD1S3 59-GGG AAA CAA ACA AAT TCA GCG GC-39 and NAALD1AS3 59-GTC AAA GTC CTG GAG TCT CTC ACT GAA C-39 yielding a 341-bp product; NAALADase II primers were NAALD2S4 59-CAC TAA GAA TAA GAA AAC AGA TAA GTA CAG C-39 and NAALD2AS4 59-GAT CAA CTT GTA TAA GTC GTT TAT GAA AAT CTG-39 yielding a 353-bp product; and NAALADase L primers were NAALDLS7 59-GAC CGG AGC AAG ACT TCA GCC AG-39 and NAALDLAS7 59-GTG TTG ATA TGC GTT GGC CCA AG-39 yielding a 330-bp product. Each primer set was tested for its ability to specifically amplify the desired NAALADase and not to cross-react in amplification reactions with the other two family members (data not shown). These primers sets were used for PCR amplifications on human multiple tissue cDNA panels (CLONTECH) normalized to the mRNA expression levels of six different housekeeping genes. Control PCR amplifications using GAPDH-specific primers (CLONTECH) were also used. Human cDNAs from carefully dissected brain regions, transformed prostate tumor cell lines, and a prostate tumor tissue sample were prepared from mRNA using an Expand Reverse Transcriptase kit (Boehringer

8475

FIG. 5. Phylogram of NAALADase I, II, and L. The human and rat sequences were used, and the alignments were performed with the ClustalW program. The tree was constructed using the GCG Distances program with standard parameters and the Growtree program with the UPGMA method. Mannheim) and normalized to the mRNA expression levels of three different housekeeping genes, GAPDH, clathrin, and actin. PCR reactions with NAALADase or GAPDH-specific primers were performed with Advantage Taq polymerase mix (95 °C for 30 s, 68 °C for 1 min 30 s). Upon completion of 25 cycles, the PCR machine was paused at 80 °C, and a 15-ml aliquot was removed from each PCR tube. Tubes were returned to the machine, and the cycling method was continued. Further aliquots were removed after 30 and 35 cycles. Samples were analyzed agarose by gel electrophoresis, and images of the ethidium bromide-stained gels were captured using an Eagle Eye II system (Stratagene). RESULTS

Molecular Cloning and Sequence Analysis of NAALADase II and NAALADase L NAALADase II—Incyte Pharmaceutical clones 1547649, 3448872, 3608639, and 1333965 contained sequences originating from a single gene similar to but not identical to NAALADase I or NAALADase L. Clone 3608639, from a lung carcinoma cDNA library, contained a DNA sequence with a 2223-bp ORF coding for a 740-amino acid residue protein, which we termed NAALADase II (Fig. 1). Analysis of this open reading frame predicted a calculated molecular mass of 83,590 kDa and isoelectric point of 8.53. The putative ATG translation start codon is in a favorable context for translation initiation (19), and no ATG codons were detected upstream. NAALADase II was predicted to be a type II integral membrane protein containing a hydrophobic membrane spanning domain extending from amino acid residues 8 –31 (20). There are also seven potential N-glycosylation sites (N 3 S/T) as indicated in Fig. 1. NAALADase L—Similarity searching of the LifeSeqTM data base (Incyte Pharmaceuticals, Palo Alto, CA) with the human, rat, and mouse NAALADase I sequences and with rat and partial human NAALADase L protein sequences yielded 13 EST sequences, some of which were overlapping, encoding for a novel protein sequence similar to NAALADase I. DNA obtained from six of the clones were sequenced. Incyte Pharmaceuticals clone 4190746, isolated from a cerebellar cDNA library, contained known sequences corresponding to human NAALADase L (16). However, because this DNA sequence also contained two segments of intronic sequence, it

8476


FIG. 6. Alignment of the NAALADase peptidase domains with related peptidases. Amino acid sequences were aligned using the standard settings of ClustalW alignment program. Similar amino acid residues conserved in proteins are shaded in black. Similar amino acid residues conserved in 80% of the proteins are shaded in dark gray. Similar amino acid residues conserved in 60 –79% of the proteins are shaded in light gray. Amino acid residues are numbered to the right. Putative residues involved in zinc binding are marked by asterisks. The base residue thought to be important in catalysis is marked by an arrow. Sequence names other than NAALADases correspond to sequence accession numbers in Swiss-Prot and SPTREMBL; Ape 3 yeast, Saccharomyces cervisiae aminopeptidase Y; P96152, Vibrio cholerae aminopeptidase; Ampx vibpr, Aeromonas proteolyitca aminopeptidase; Apx strgr, Streptomyces griseus aminopeptidase. Putative residues involved in zinc binding are marked by asterisks. A general base residue thought to be important in catalysis is marked by an arrow. Amino acid residues are numbered in the right margin.

was not suitable for further cloning experiments. PCR reactions were performed to amplify a PCR product containing the 39-half of the NAALADase L coding region from small intestine cDNA. To identify the remaining unknown human 59-NAALADase L sequence, 59-RACE PCR was performed on a number of cDNAs. Sequencing of the amplification products obtained from reactions using small intestine cDNA yielded a further 1344-bp fragment covering the complete coding sequence of NAALADase L. The full cDNA sequence contained an open reading frame of 2223 bp encoding a protein of 740 amino acid residues with a calculated molecular mass of 80,638 Da and isoelectric point of 5.26 (Fig. 2). The putative ATG translation start codon is in a favorable context for translation initiation (19) with no ATG codons detected upstream. Analysis of the human NAALADase L ORF suggests that it is a type II integral membrane protein containing a single hydrophobic membrane spanning domain extending from amino acid residues 6 –27, with lysine residues bordering either side of the potential membrane spanning domain (20). There are seven potential Nlinked glycosylation sites (N 3 S/T) as indicated in Fig. 2. The

predicted protein sequences of human NAALADase L were compared with that of rat using the alignment program Genedoc. Human NAALADase L protein sequence was 78% identical and 87% similar to rat NAALADase L. Alternative Splicing of NAALADase L—In the course of our cloning and RT-PCR gene expression analysis of NAALADase L, a number of amplified PCR products of unexpected size were observed, isolated, and sequenced in order to identify possible splice variants. We found both the splicing out of putative exon sequences as well as the presence of intronic sequences (as judged by the presence of G(T/A)G donor/acceptor sites) that were repeatedly amplified from our cDNA preparations (Fig. 3). When performing 59-RACE amplifications, deletions between bases 497– 619 and 903–1007 were identified in small intestine and colon that resulted in two in-frame amino acid residue deletions. In addition, PCR products have been obtained containing a 153-bp intron insertion at base 1094 resulting in an in-frame amino acid insertion of 51 amino acid residues. This insertion is most likely an intron as it has the consensus G(T/A)G donor acceptor sites at its 59 and 39 ends, respectively


8477

(21). In the PCR amplifications of the 39 end of NAALADase L, several variants were also identified from small intestine, colon, brain, and fetal brain. These consisted of either a deletion of bases 1525–1615 or a larger deletion between bases 1525 and 1697 (Fig. 3). Both these deletions resulted in frameshifts and the premature termination of the protein sequence. Finally, in every cDNA sample examined, two intronic sequences were found to be inserted at either base 1697 and/or at base 1870. Inclusion of one or both of these intronic sequences into the ORF of NAALADase L accounted for PCR products of unexpected size seen in RT-PCR experiments migrating at 420 and 500 bp. Introduction of one or both these sequences results in a frameshift and therefore in altered amino acid sequence and premature stop (Fig. 3). The predicted protein sequences of NAALADase I, II, and L were compared with each other using the alignment program BESTFIT (Genetics Computer Group Software, WI), and the percent identity and percent similarity between each pair of sequences were calculated by the Genedoc program. NAALADase I sequence was 67% identical (81% similar) to NAALADase II and 35% identical (54% similar) to NAALADase L, whereas NAALADase II and L sequences were 37% identical (54% similar). The three protein sequences for NAALADase I, II, and L were aligned using the ClustalW alignment program (EMBL, Heidelberg, Germany) and are shown in Fig. 4. A phylogram of NAALADase I, II, and L was constructed using the GCG Distances program with standard parameters and the Growtree program with the UPGMA method and is depicted in Fig. 5. From this phylogram it is clear that NAALADase I and II are the most closely related proteins. Two putative catalytic domains have previously been identified in rat NAALADase I and L sequences by comparison to other peptidases (12, 16). By using multiple sequence alignments of NAALADase I, II, and L, we have identified similar putative catalytic domains in human NAALADase II and L (Fig. 4 and 6). The first catalytic domain is related to bacterial and yeast Zn21-dependent peptidase domains (12), and the second catalytic domain is related to members of the a/b hydrolase fold family of proteins (16).

Expression and Functional Activity of NAALADases

FIG. 7. Biological activity of NAALADases. A, transfection of NAALADase I and II but not NAALADase L confers NAAG hydrolyzing activity. Hydrolysis of 500 nM [3H]NAAG was assayed using homogenates prepared from equivalent numbers of COS cells transiently transfected with either NAALADase I, II, or L. Reactions proceeded for 60 min at 37 °C and were terminated with 250 mM ice-cold phosphate. Four transient transfection experiments were performed, and representative results from a single experiment are shown after subtraction of a protein-free blank. B, inhibition of NAALADase I (open circles) and NAALADase II (closed circles) activity by increasing concentrations of quisqualic acid expressed as percent of control (CTRL) (activity in the absence of any inhibitor). C, transfection of NAALADase I, II, and L confers DPP IV peptidase-like activity. Hydrolysis of the fluorogenic substrate Gly-Pro-AMC was assayed using COS cells transiently transfected with NAALADase I, II, or L. 135 mg of protein was used in each

To determine if the newly identified NAALADases had peptidase activity, mammalian expression constructs were transiently transfected into COS cells and cellular homogenates prepared. Expression of NAALADase I in COS cells was performed as a positive control to establish the working conditions of the assay, and homogenates from mock transfections were used in parallel as negative controls. Hydrolysis of [3H]NAAG by recombinant NAALADases, measured by elution of [3H]glutamate, occurred in a time-dependent manner. Activity was observed in homogenates from NAALADase I and II but not NAALADase L transfections (Fig. 7A). Addition of 30 mM quisqualic acid to the reaction inhibited this activity by over 50% after 60 min (data not shown). Inhibition curves with increasing concentrations of QA gave IC50 values of 1.2 3 1025 and 1.7 3 1025 M for NAALADase I and II, respectively (Fig. 7B). DPP IV-like activity was assayed using homogenates from COS cells transiently transfected with a particular NAALADase cDNA. Enhanced DPP IV-like activity was observed in all samples tested when compared with mock-transfected cells. NAALADase I and II showed the higher levels of DPP IV activity than NAALADase L in repeated experiments (Fig. 7C).

reaction and followed for 40 min at 37 °C. Fluorescence emitted at 450 nm was measured. Three transient transfection experiments were performed, and representative results from a single experiment are shown after subtraction of a protein-free blank.

8478


FIG. 8. Chromosomal localization of human NAALADases. Diagrams of FISH mapping results for NAALADase I (A), II (C), and L (E). In each diagram dots represent the double FISH signals detected on identified chromosomes. Examples of FISH mapping of NAALADase I (B), II (D), and L (F) are shown. Left panels show the FISH signals on the identified chromosome marked by a white arrow, and right panels show the same mitotic figure stained with DAPI to identify the chromosome.

Chromosomal Localization The complete coding sequence of NAALADase I was used as a probe for FISH analysis. Under the conditions used, the hybridization efficiency was approximately 71% for this probe (among 100 checked mitotic figures, 71 of them showed signals on one pair of the chromosomes). DAPI banding was used to identify the specific chromosome, and an assignment between the signal from the probe and the short arm of chromosome 11 was made. The detailed position to region p11.21 was further determined based upon summary data from 10 photographs. A weak hybridization signal was also detected in the region of 11q14.3 with low frequency. From the mapping data obtained, it was concluded that this weak signal was a result of crosshybridization to NAALADase II. Examples of the mapping results are presented in Fig. 8, A and B. For NAALADase II the hybridization efficiency was approximately 74%, and DAPI banding was used to identify the signal to human chromosome 11, region q14.3-q21 (Fig. 8, C and D). For NAALADase L the hybridization efficiency was approximately 71%, and DAPI

banding was used to identify the signal to the long arm of chromosome 11 region q12 (Fig. 8, E and F).

Tissue Distribution of NAALADase II as Determined by Northern Blot Northern blot analysis was performed on mRNA derived from different human tissues (Fig. 9). A NAALADase II-specific probe indicated the presence of transcripts in testis .. ovary, spleen . prostate gland, heart, and placenta with no signal observed in other tissues. In testis, four transcripts were represented. The most predominant transcript was of approximately 3.4 kb, consistent with the approximate expected size of a NAALADase II message. Two transcripts of 2.4 and 4.4 kb, respectively, and a weaker transcript of about 7.5 kb were also present. In the other tissues the 3.4-kb transcript was the only signal detected, apart from ovary where a weak 7.5-kb signal could also be seen. The precise nature of these transcripts awaits further elucidation but may be due to alternative splicing of the message.


8479

FIG. 9. Northern blot analysis of NAALADase II expression. Human multiple tissue Northern blots (CLONTECH) containing 2 mg per lane of poly(A)1 RNA were hybridized with a 32Plabeled NAALADase II probe at 68 °C and washed at high stringency (50 °C with 0.13 sodium chloride sodium citrate, 0.1% SDS). Autoradiographic exposure was for 2– 4 days at 270 °C with two intensifying screens. Molecular mass markers are indicated in base pairs on the left side of each autoradiogram.

Analysis of NAALADase Gene Expression by RT-PCR To examine the detailed tissue distribution of all NAALADases, PCR was performed on normalized cDNAs from 16 different tissues. Fig. 10A shows the results from PCR reactions performed with NAALADase I-specific primers, yielding amplification products of the expected size (;341 bp). The highest expression of NAALADase I appeared to be in prostate gland. Rank order of expression after 25 cycles was prostate .. liver and kidney . small intestine . brain and spleen, with no product amplification observed in the other tissues. At 30 cycles amplification products could be seen in most other tissues with the exception of muscle, blood, and thymus in which products could only be observed after 35 cycles of amplification. NAALADase II-specific primers yielded a 353-bp amplification product of the expected size (Fig. 10B). NAALADase II expression was highest in ovary, testis, and spleen with PCR products detected after 25 cycles of amplification. After 30 cycles amplification products could be detected from all tissue cDNAs apart from lung, muscle, blood, and thymus in which a product could only be seen after 35 cycles. These results are in good accord with the expression data obtained with the multiple tissue Northern blots. NAALADase L-specific primers yielded a 330-bp amplification product of the expected size, as well as two products migrating with slightly higher sizes of 420 and 500 bp (Fig. 10C). NAALADase L expression was highest in small intestine, spleen, and testis with PCR products detected after 25 cycles of amplification, whereas products in heart, ovary, colon, blood, and prostate could be seen after 30 cycles. Some amplification products following 35 cycles were observed in all tissues, with brain and muscle showing the lowest levels. The 420- and 500-bp bands were due to amplification of NAALADase L sequences containing one or two intronic sequences that were commonly found in all our amplification reactions (see above). Control amplification reactions using GAPDH-specific primers demonstrated comparable levels of amplification products for each cDNA (Fig. 10D). Comparison of the relative abundance between the four messages was also possible from these experiments, since the same cDNAs were used for each set of amplifications. Abundance of NAALADase I message was greater than NAALADase II which was greater than NAALADase L, as judged by the relative amount of amplification products detected at 25 and 30 cycles. PCR reactions using the same NAALADase primers as in the above experiments were performed on 13 different brain cDNAs normalized to the expression levels of three housekeeping genes. NAALADase I-specific amplification products were detected with highest levels in ventral striatum and brain stem

after 30 cycles. After 35 cycles NAALADase I-specific amplification products could clearly be detected in all brain areas studied (Fig. 11A). NAALADase II-specific primers yielded a 353-bp amplification product of the expected size (Fig. 11B). Amplification products were observed after 30 cycles in striatum, parietal cortex, and ventral striatum with lower levels of amplification product detected in hippocampus, brain stem, putamen, and superior colliculus. After 35 cycles the presence of NAALADase II-specific products could be detected in all cDNAs apart from inferior colliculus. NAALADase L-specific primers yielded a 330-bp amplification product of the expected size, as well as a product migrating at a higher size of 500 bp (Fig. 11C). Amplification of the 500-bp product was observed after 35 cycles in brain stem, amygdala, thalamus, ventral striatum, and to a lesser extent in striatum and hippocampus, whereas the expected 330-bp product was only seen in brain stem and ventral striatum. Control amplification reactions using GAPDH-specific primers demonstrated comparable levels of amplification products for each cDNA apart from brain stem which yielded relatively more GAPDH-specific product (Fig. 11D). Overall expression of NAALADase L appears to be lower in these brain areas relative to NAALADase I and II. Finally, NAALADase expression was investigated in cDNAs prepared from either prostate tumor cell lines or prostate tumor tissue that had been normalized against three different housekeeping genes. NAALADase I expression was highest in LNCaP and prostate tumor (Fig. 12A). Amplification products were also detected in PC-3 cDNA after 30 and 35 cycles but not in DU145 cDNA. NAALADase II expression was higher in LNCaP than prostate tumor. A faint amplification product could also be detected in PC-3 after 35 cycles but not in DU145 cDNA (Fig. 12B). The 330-bp NAALADase L product was detected in highest amounts in cDNA from prostate tumor and less in PC-3 and DU145 samples after 35 cycles (Fig. 12C). Interestingly, in all samples apart from prostate tumor, the higher 500-bp amplification product could be detected. Representative amplifications with GAPDH primers are also shown (Fig. 12D). DISCUSSION

In this report we describe the cloning of two novel human NAALADase peptidases that represent an expansion of the glutamate carboxypeptidase II gene family and which we have called NAALADase II and NAALADase L. Both peptidases contain a single hydrophobic region which is a likely membrane spanning a short intracellular domain and a large globular extracellular domain, typical of type II integral membrane

8480


FIG. 10. RT-PCR analysis of NAALADase expression in different tissues. PCR amplifications with primers specific for human NAALADase I (A), II (B), L (C), or GAPDH (D) were performed on normalized human MTC cDNAsTM for 25, 30, and 35 cycles. Images from ethidium bromide-stained agarose gels were captured using an EagleEye system and EagleSight software (Stratagene). Positive control amplifications were performed using the appropriate NAALADase DNA construct. Negative control amplifications contained reaction mix, enzyme, and no DNA template. Positive and negative control reactions were performed for 35 cycles. Arrows highlight a specific size in base pairs as determined using a 100-bp ladder.

proteins and common among membrane-bound hydrolases (23). Carboxypeptidase activity of NAALADase I has been demonstrated against two classes of substrate in vitro, a-linked acidic peptides such as NAAG or a-glutamylglutamate and g-linked peptides such as folyl-poly-g-glutamate or g-glutamylglutamate. In brain preparations, the QA-sensitive a-linked hydrolysis of NAAG was shown to be increased by divalent cations and inhibited by divalent metal chelators or general metalloprotease inhibitors (9, 24). In this study we have transiently expressed NAALADase I, II, and L in COS cells, and we have shown that NAALADase I and II but not NAALADase L confer NAALADase activity to the cell homogenates comparable to that described for lysed synaptosomal membranes (9). Alignment of NAALADase I with zinc aminopeptidases from yeast and bacteria identified a domain of conserved sequences involved in the coordination of two zinc ions located in the catalytic site (12, 25). This catalytic site was identified from the three-dimensional crystal structure of the Aeromonas proteo-

lytica zinc aminopeptidase, in which residues His379, Asp389, Glu427, Asp455, and His555 are thought to be involved in the binding of two zinc ions, and Glu426 is proposed to be a base residue important in catalysis (26, 27). Alignment of the newly identified NAALADases with these aminopeptidases shows that the putative zinc binding domains, including the five residues important in Zn21 binding, are highly conserved, suggesting that they may share the same catalytic domain (Fig. 4 and Fig. 6). Shneider and colleagues (16) have suggested that rat NAALADase L may also be a member of the a/b hydrolase fold family because of its DPP IV and acylaminoacylpeptidaselike activity, although sequence homology alignments have shown these protein sequences to be clearly distinct. However, an hypothesized catalytic site arrangement of nucleophilic, acidic, and basic residues (Ser623, Asp663, and His686) is conserved and found in NAALADase I, II, and L, downstream of the predicted zinc binding domain (Fig. 4). In support of this, transient transfection experiments with each NAALADase


8481

FIG. 11. RT-PCR analysis of NAALADase expression in brain areas. PCR amplifications for 25, 30, and 35 cycles, with primers specific for human NAALADase I (A), II (B), L (C) or GAPDH (D), were performed on normalized human cDNAs prepared from dissected brain areas. Images from ethidium bromidestained agarose gels were captured using an EagleEye system and EagleSight software (Stratagene). Results from 30 and 35 cycles of amplification are shown. Positive control amplifications were performed using the appropriate NAALADase DNA construct. Negative control amplifications contained reaction mix and enzyme with no DNA template. All control reactions were performed for 35 cycles. Arrows highlight a specific size in base pairs as determined from the 100-bp ladder.

cDNA did consistently confer enhanced DPP IV activity to cellular homogenates as demonstrated by their ability to cleave the substrate Gly-Pro-AMC. DPP IV activity is an abundantly expressed serine peptidase activity. The physiological role of such an enzymatic activity is not clear, but it may play a role in the regulation of various biologically active peptides such as collagen, neuropeptide Y, and growth hormone releasing factor both in the intestine and in other organs (16). With abundant NAALADase expression in a number of different tissues, it will be interesting to identify which of these or other biologically important peptides can be cleaved by these peptidases. The second g-linked enzymatic activity of NAALADase I to peptides such as pteroyl-g-glutamate (folate hydrolase activity) has been found at high levels in the brush border of human small intestine (15, 28). In addition, carcinoma cells transfected with NAALADase I show increased folate hydrolase activity and the ability to progressively liberate glutamate from methotrexate triglutamate by hydrolysis of g-glutamyl linkages (15). A correlation has been observed between increased pteroyl hydrolase activity and methotrexate resistance in tumors (29), suggesting that modulating NAALADase activity may be useful in developing improved or novel cancer treat-

ments. It will be interesting to see if human NAALADase II or L with their common secondary structure have a dual peptidase activity similar to that of NAALADase I. Analysis of NAALADase L sequences using different PCR primers sets revealed the presence of multiple splice variants and isoforms. Changes in sequence due to alternative splicing may affect levels of glycosylation and more importantly the conformation and activity of the protein. Furthermore, inclusion of three different intronic sequences identified in amplification reactions from numerous cDNAs results in either the in-frame addition of a proline-rich (51-amino acid residue) sequence close to the putative zinc binding domain or in frameshifts and premature termination of the peptidase. These premature terminations result in either partial deletion of the zinc binding domain or in elimination of the predicted nucleophile acid base arrangement of the putative a/b hydrolase catalytic site (see Fig. 3). Finally, insertion of two intronic DNA sequences (at bp 1697 and 1870), resulting in frameshifts and premature protein termination, was identified in all cDNAs studied suggesting that these are not artifactual in nature due to contaminating genomic DNA, for example. It is possible that expression of these sequences may be used to regulate the

8482


FIG. 12. RT-PCR analysis of NAALADase expression in prostate tumor cell lines. PCR amplifications with primers specific for human NAALADase I (A), II (B), L (C), or GAPDH (D) were performed on normalized cDNAs prepared from prostate tumor cell lines or from prostate tumor tissue. Images from ethidium bromide-stained agarose gels were captured using an EagleEye system and EagleSight software (Stratagene). Results from 25, 30, and 35 cycles of amplification are shown. Positive control amplifications were performed using the appropriate NAALADase DNA construct. Negative control amplifications contained reaction mix and enzyme with no DNA template. All control reactions were performed for 35 cycles.

levels of active protein. In addition, analysis of NAALADase II tissue distribution by Northern hybridization revealed a number of differently sized transcripts, suggesting the presence of multiple isoforms. It remains to be determined whether these are due to physiologically relevant smaller transcripts or due to problems with the integrity of the mRNA. Future studies using different splice variants will help us understand the structure/ activity relationship of NAALADase L and NAALADase II, and help to identify which domains are important for enzymatic activity. Interestingly, a splice variant of NAALADase I, lacking the first 40 amino acids including the membrane spanning domain, has been observed at decreased ratios relative to its

full-length transcript in malignant prostate tissues, suggesting that expression of this alternative splice variant may correlate with tumor progression (30). Further work will determine if NAALADase II or NAALADase L is involved in oncogenesis or if alternative splicing of these peptidases has any role to play in tumor progression. Using FISH analysis to determine the chromosomal localization of NAALADase I, II, and L, we observed a signal on chromosome 11, at p11.21 for NAALADase I, between q14.3q21 for NAALADase II and q12 for NAALADase L. In our studies, a NAALADase I probe revealed two hybridization signals, one hybridizing at 11p11.21 and another weakly at 11q14.3. This is similar to the results obtained by Leek et al. (31) who observed two hybridization signals at 11p11.2 and 11q13.5. Having localized NAALADase II to 11q14.3-q21, it is clear that the second signal is due to cross-hybridization of the NAALADase I probe with the NAALADase II locus. Chromosome 11 contains a number of genetic disease loci in these regions, including vitreoretinopathy (11q13-q23), xeroderma pigmentosa (11q12-q13), atopy (11q12-q13), and perhaps more interestingly, a tumor suppression locus (11p11.2-p11.13) involved in rat prostate carcinoma. Introduction of this portion of the chromosome into highly metastatic rat prostatic cells was able to suppress cancer metastases without suppression of the in vivo growth rate or tumorigenicity of the cells (32). Since it has been shown that NAALADase I expression increases with decreasing androgen levels, it is possible that current prostate cancer treatments involving androgen level reduction (e.g. orchidectomy) may work at least in part through alteration of NAALADase expression (33). The in vivo enzymatic activity of NAALADase I may be similar to that of NAALADase II and L, so it is conceivable that these enzymes may also have a role to play in tumor suppression. Interestingly, region 11q13-q23 has also been identified as a region with tumor suppressor activity using tumorigenic HeLa/fibroblast hybrids (34). In addition, in a systematic analysis of primary cervical carcinomas, region 11q22-q24 was shown to contain tumor suppressor activity (35). These latter two tumor-suppressing regions on the long arm of chromosome 11 cover the gene loci of NAALADase L and NAALADase II. It should be noted that mapping of these tumor-suppressing activities to these three chromosomal regions in no way establishes that the identified NAALADases are capable of having any tumor or metastasis-suppressing activity. Furthermore, it has been suggested that expression of NAALADase I is related to an increase in cancerous phenotype, suggesting an oncogenic rather than a tumor-suppressing role for NAALADase I (33). Given the fact that NAALADase II and NAALADase L also appear to be expressed in a number of prostate tumor cell lines as well as in prostate tumor tissue, it will be interesting to see if these two genes will be implicated in prostate tumor development. In this study, NAALADases II has been shown to be able to hydrolyze [3H]NAAG to N-acetylaspartate and glutamate as has been previously shown with NAALADase I. Given the localization of these NAALADases in prostate and ovary, as well as other peripheral tissues, it is quite possible that these enzymes may modulate local extracellular glutamate levels in these tissues. For example it is known that substantial amounts of glutamate are present in seminal fluid. As endogenous levels of NAAG are high in brain, the catabolism of NAAG by NAALADases to glutamate and N-acetylaspartate may in theory be a rich source of glutamate. Whether in fact the inhibitory action of NAAG or its excitatory metabolite glutamate is the active species at particular synapses will be dependent upon the receptors present and factors regulating the expression of different NAALADases. However, aberrant

Characterization of Novel NAALADases catabolism of NAAG by NAALADases to release excessive levels of glutamate may well result in activation of numerous glutamate receptor subtypes, resulting in an excessive positive modulation of glutamatergic neurotransmission and excitotoxicity. The possible importance of careful regulation of NAALADase activity therefore becomes apparent if deleterious excitotoxic effects are to avoided. Indeed abnormal levels of NAAG or NAALADase activity have been suggested for a number of disorders including schizophrenia (36), ALS (37), Alzheimer’s disease (38, 39), seizure disorders (40, 41), and stroke (42). It is clear from our studies that the NAALADase genes are differentially expressed within discrete brain areas at the mRNA level, but whether there are any disease-specific changes in NAALADase gene expression remains to be determined. In addition, molecules able to modulate the activity of specific NAALADases may be potentially useful in treating ischemiainduced neurodegeneration or other neurodegenerative disorders involving abnormalities in glutamate neurotransmission, such as Alzheimer’s disease, schizophrenia, or amyotrophic lateral sclerosis (43). In vitro at least, the NAALADase (carboxypeptidase) inhibitor, 2-(phosphonomethyl)pentanedoic acid, inhibited toxicity induced by the carboxypeptidase cleavage of folic acid hexaglutamate (44). In summary we have identified and characterized two novel human peptidases NAALADase II and NAALADase L. With several biological roles suggested for NAALADases, including regulation of glutamatergic neurotransmission and prostate tumor progression, it will be interesting to see what role individual members of this newly expanded protein family will play within these systems. Further work in the understanding of the relative biological importance of each of these proteins and elucidation of possible physiological substrates will help researchers to identify more specific, small molecule NAALADase inhibitors which may be of use in a number of clinically important disorders. Acknowledgments—We thank Professor R. C. A. Pearson for providing brain RNA samples and Dr. B.-J. Van Der Leede for providing prostate tumor RNA samples. We also thank Joerg Sprengel for valuable advice in bioinformatics and sequence analysis. Petra De Wilde and Nathalie Delcroix provided technical assistance for the sequencing reactions and Annemie Heylen helped with preparation of the figures. REFERENCES 1. Coyle, J. T. (1997) Neurobiol. Dis.4, 231–238 2. Valivullah, H. M., Lancaster, J., Sweetnam, P. M., and Neale, J. H. (1994) J. Neurochem. 63, 1714 –1719 3. Puttfarcken, P. S., Handen, J. S., Montgomery, D. T., Coyle, J. T., and Werling, L. L. (1993) J. Pharmacol. Exp. Ther. 266, 796 – 803 4. Sekiguchi, M., Wada, K., and Wenthold, R. J. (1992) FEBS Lett. 311, 285–289 5. Bruno, V., Wroblewska, B., Wroblewski, J. T., Fiore, L., and Nicoletti, F. (1998) Neuroscience 85, 751–757 6. Wroblewska, B., Wroblewski, J. T., Saab, O. H., and Neale, J. H. (1993) J. Neurochem. 61, 943–948 7. Wroblewska, B., Wroblewski, J. T., Pshenichkin, S., Surin, A., Sullivan, S. E., and Neale, J. H. (1997) J. Neurochem. 69, 174 –181 8. Robinson, M. B., Blakely, R. D., and Coyle, J. T. (1986) Eur. J. Biochem. 130, 345–347

8483

9. Robinson, M. B., Blakely, R. D., Couto, R., and Coyle, J. T. (1987) J. Biol. Chem. 262, 14498 –14506 10. Slusher, B. S., Robinson, M. B., Tsai, G., Simmons, M. L., Richards, S. S., and Coyle, J. T. (1990) J. Biol. Chem. 265, 21297–21301 11. Carter, R. E., Feldman, A. R., and Coyle, J. T. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 749 –753 12. Bzdega, T., Turi, T., Wroblewska, B., She, D., Chung, H. S., Kim, H., and Neale, J. H. (1997) J. Neurochem. 69, 2270 –2277 13. Luthi-Carter, R., Berger, U. V., Barczak, A. K., Enna, M., and Coyle, J. T. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 3215–3220 14. Israeli, R. S., Powell, C. T., Fair, W. R., and Heston, W. D. (1993) Cancer Res. 53, 227–230 15. Pinto, J. T., Suffoletto, B. P., Berzin, T. M., Qiao, C. H., Lin, S., Tong, P. T., May, F., Mukherjee, B., and Heston, W. D. (1996) Clin. Cancer Res. 2, 1445–1451 16. Shneider, B. L., Thevananther, S., Moyer, M. S., Walters, H. C., Rinaldo, P., Devarajan, P., Sun, A. Q., Dawson, P. A., and Ananthanarayanan, M. (1997) J. Biol. Chem. 272, 31006 –31015 17. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) J. Mol. Biol. 215, 403– 410 18. Heng, H. H., and Tsui, L.-C. (1993) Chromosoma 102, 325–332 19. Heng, H. H., Squire, J., and Tsui, L.-C. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 9509 –9513 20. Kozak, M. (1989) J. Cell Biol. 108, 229 –241 21. Kyte, J., and Doolittle, R. F. (1982) J. Mol. Biol. 157, 105–132 22. Breathnach, R., and Chambon, P. (1981) Annu. Rev. Biochem. 50, 349 –383 23. Goosens, F., De Meester, I., Vanhoof, G., Hendriks, D., Vriend, G., and Scharpe, S. (1995) Eur. J. Biochem. 233, 432– 441 24. Kenny, A. J., and Maroux, S. (1982) Physiol. Rev. 62, 91–128 25. Blakely, R. D., Robinson, M. B., Thompson, R. C., and Coyle, J. T. (1988) J. Neurochem. 50, 1200 –1209 26. Rawlings, N. D., and Barrett, A. J. (1997) Biochim. Biophys. Acta 1339, 247–252 27. Chevrier, B., Schalk, C., D’Orchymont, H., Rondeau, J. M., Moras, D., and Tarnus, C. (1994) Structure 2, 283–291 28. Chevrier, B., D’Orchymont, H., Schalk, C., Tarnus, C., and Moras, D. (1996) Eur. J. Biochem. 237, 393–398 29. Chandler, C. J., Wang, T. T., and Halsted, C. H. (1996) J. Biol. Chem. 261, 928 –933 30. Banerjee, D., Ercikan-Abali, E., Waltham, M., Schnieders, B., Hochhauser, D., Lei, W. W., Fan, J., Gorlick, R., Goker, E., and Bertino, J. R. (1995) Acta Biochim. Pol. 42, 457– 464 31. Su, S. L., Huang, I. P., Fair, W. R., Powell, C. T., and Heston, W. D. (1995) Cancer Res. 55, 1441–1443 32. Leek, J., Lench, N., Maraj, B., Bailey, A., Carr, I. M., Andersen, S., Cross, J., Whelan, P., MacLennan, K. A., Meredith, D. M., and Markham, A. F. (1995) Br. J. Cancer 72, 583–588 33. Ichikawa, T., Ichikawa, Y., Dong, J., Hawkins, A. L., Griffin, C. A., Isaacs, W. B., Oshimura, M., Barrett, J. C., and Isaacs, J. T. (1992) Cancer Res. 52, 3486 –3490 34. Israeli, R. S., Powell, C. T., Corr, J. G., Fair, W. R., and Heston, W. D. (1994) Cancer Res. 54, 1807–1811 35. Misra, B. C., and Srivatsan, E. S. (1989) Am. J. Hum. Genet. 45, 565–577 36. Hampton, G. M., Penny, L. A., Baergen, R. N., Larson, A., Brewer, C., Liao, S., Busby-Earle, R. M., Williams, A. W., Steel, C. M., Bird, C. C., Stanbridge, E. J., and Evans, G. A. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 6953– 6957 37. Tsai, G., Passani, L. A., Slusher, B. S., Carter, R., Baer, L., Kleinman, J. E., and Coyle, J. T. (1995) Arch. Gen. Psychiatry 52, 829 – 836 38. Tsai, G. C., Stauch-Slusher, B., Sim, L., Hedreen, J. C., Rothstein, J. D., Kuncl, R., and Coyle, J. T. (1991) Brain Res. 556, 151–156 39. Passani, L. A., Vonsattel, J. P., Carter, R. E., and Coyle, J. T. (1997) Mol. Chem. Neuropathol. 31, 97–118 40. Jaarsma, D., Veenma-van der Duin, L., and Korf, J. (1994) J. Neurol. Sci. 127, 230 –233 41. Meyerhoff, J. L., Koller, K. J., Walczak, D. D., and Coyle, J. T. (1985) Brain Res. 346, 392–396 42. Meyerhoff, J. L., Carter, R. E., Yourick, D. L., Slusher, B. S., and Coyle, J. T. (1992) Epilepsy Res. 9, 163–172 43. Sager, T. N., Laursen, H., and Hansen, A. J. (1995) J. Cereb. Blood Flow Metab. 15, 639 – 646 44. Passani, L. A., Vonsattel, J. P., and Coyle, J. T. (1997) Brain Res. 772, 9 –22 45. Slusher, S., Thomas, A. T., Jackson, P. F., Tiffany, C. W., and Suzdak, P. D. (1997) Soc. Neurosci. Abstr. 898.7