THE JOURNAL OF BIOLOGICAL CHEMI5TRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 14, Issue of April 8, pp. 10524-10528, 1994 Printed in U.S.A.

Multiple Forms of the Human Tyrosine Phosphatase RPTPa ISOZYMES AND DIFFERENCES IN GLYCOSYLATION*

(Received for publication, November 22, 1993, and in revised form, January 11, 1994)

Gunter DaumSO, Stephan Regenassfll, Jan Sap**, Joseph Schlessinger**, and Edmond H. FischerS From the $Department of Biochemistry SJ-70 and Wepartment of Pathology SM-30, University of Washington, Seattle, Washington 98195 and the **Department of Pharmacology, New York University Medical Center, New Yoik, New York 10016

Among all the receptor-linked protein-tyrosine-phos- phatase RPTPa clones described from mammalian tis- sues, one differed in that it encoded a 9-amino-acid in- sert 3 residues upstream from the transmembrane segment (Kaplan, R., Morse, B., Huebner, IC, Croce, C., Howk, R. Ravera, M., Ricca, G., Jaye, M., and Schless- inger, J. (1990) Prm. Nutl. A c d . Sci. U. S. A. 87, 700s 7004). Using the polymerase chain reaction technique, simultaneous expression of both isoforms was demon- strated in human T-cell and vascular smooth muscle li- braries, as well as in the A431 human epidermal cancer cell line. Following transient expression in COS-1 cells, each isoform gave rise to two proteins of 100 and 130 kDa, respectively. Endoglycosidase treatment showed that the 100-kDa species corresponded to a molecule ex- clusively glycosylated on N-residues, whereas the 130- kDa species contained both, N- and 0-linked carbohy- drates.' Pulse-chase experiments demonstrated that the smaller RPTPa protein is a precursor of the larger one. A high affinity antibody was generated that recognizes the immature protein only; however, both proteins can be detected by Western blot analysis after a simple chemical hydrolysis. Following Superose 12 chromatog- raphy, the 100- and 130-kDa species of RF'TPa emerged as 200- and 340-kDa proteins, respectively. Both species exhibited similar enzymatic activities as determined with a peptide substrate in immunoprecipitates.

Protein tyrosine phosphatases (PTPs)' consist of intracellu- lar, low molecular weight enzymes and transmembrane mol- ecules thought to act as receptors regulated by extracellular signals (reviewed in Refs. 1 and 2). Except for HPTPP ( 3 ) and the Drosophila enzyme DPTPlOD (4, 5), the cytosolic portions of all receptor PTPs contain two catalytic domains. In some instances, e.g. RPTPa (6) and the leukocyte common antigen CD45 (7), both domains exhibit enzymatic activity, whereas the second domain of others is said to be inactive (8). Receptor- linked PTPs exhibit high homology in their PTP domains but differ considerably in their extracellular segments: CD45, which plays a decisive role in T-cell receptor signaling, is ex- pressed in nucleated hematopoetic cells only and has a unique

* This work was supported by National Institutes of Health Grants DK0709 and GM42508, the Muscular Dystrophy Association of America, Sugen, and the Human Frontier Science Program. The costs of publication of this article were defrayed in part by the payment of page

in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. charges. This article must therefore be hereby marked "aduertisement"

8 To whom correspondence should be addressed. 11 Recipient of a fellowship from the Lichtenstein-Stiftung, Basel,

Switzerland.

PCR, polymerase chain reaction; Sf9 cells, Spodoptera frugiperda 9 The abbreviations used are: PTP, protein tyrosine phosphatase;

cells, PAGE, polyacrylamide gel electrophoresis.

extracellular domain containing 2 cysteine-rich repeats; it is heavily N- and 0-glycosylated (reviewed in Refs. 9 and 10). HPTPP contains only fibronectin type I11 repeats (31, whereas LAR (111, RPTPy (121, and RPTPk (13) have both immuno- globulin and fibronectin type I11 repeats, as found in cell adhe- sion molecules such as N-CAMS and fasciclin 11. Recent reports suggest a regulation of these phosphatases through homophilic interactions (14,151. The Drosophila homologue of L A R , DLAR, (16) is involved in neuronal differentiation together with two other receptor linked PTPs, DPTP99A and DPTPlOD (4, 5). HPTPS (RPTPP, 17, 18) and RPTPy (19) bear a globular seg- ment homologous to carbonic anhydrase. RPTPa (3,20-23) and HPTPE ( 3 ) differ from all other transmembrane PTPs in that they display particularly short extracellular domains of only 123 and 27 residues, respectively. Recent reports indicate that RPTPa plays a role in neuronal differentiation (24) and that its ectopic expression in fibroblasts leads to cell transformation and tumorigenesis (25). In both instances, it was suggested that the phosphatase-activated pp60""" by dephosphorylating its C-terminal autoinhibitory phosphorylation site.

Many of the receptor-linked PTPs exist in multiple forms differing in their extracellular segments. Human CD45 has five known isoforms resulting from alternative splicing of three exons close to the N terminus; they are differentially expressed in various lymphocyte subsets (reviewed in Ref. 10). A spliced form of RPTPP, designated as dvRPTPp (for deletion variant) lacks a 859-amino-acid segment located close to the carbonic anhydrase domain (18). Two variants of HPTPP were de- scribed, one lacking a single fibronectin-like I11 segment ( 3 ) . Two different clones encoding RPTPa were also reported (22). The proteins differ in a stretch of 9 residues situated in the external juxtamembrane segment. The reason for this struc- tural diversity is unknown. One may speculate, however, that different forms could be expressed in different tissues or during different stages of cell development.

This paper reports the expression pattern of both RPTPa isozymes in various tissues and that the enzyme is heavily glycosylated. A highly specific antibody was generated that rec- ognizes only a precursor protein which is exclusively N-glyco- sylated, whereas the mature phosphatase contains both N- and 0-linked carbohydrates.

EXPERIMENTAL PROCEDURES Materials-The T-cell library, the A431 cDNA, and the vascular

smooth muscle cell library were kind gifts from Drs. R. Seger and C. Giachelli, respectively. Oligonucleotides were synthesized on an Applied Biosystems DNA synthesizer at the Howard Hughes Medical Institute DNAfacility, University of Washington, Seattle. Restriction and nucleic acid modifying enzymes were from Stratagene, Boehringer Mannheim, and Bethesda Research Laboratories, Sequenase and DNA molecular weight markers from United States Biochemical Corp. X-ray Hyper- film-MP was purchased from Amersham Corp., and protein molecular weight markers were from Bio-Rad.

10524

Multiple Forms of Human RPTPa 10525

FIG. 1. Structure of EFPPa is@ forms. A, RPTPa consists of an extracel- lular domain, a transmembrane segment (I) and two PTP domains ( I I and III). Two forms have been described differing in a stretch of 9 residues (22). B, cDNA of hu- man RPTPa from nucleotide 943 to 1178 is shown (22, Genbank entry M34668). PCR with the sense primer 8s (943-960) and the antisense primer T (1152-1178) should yield DNA fragments of 236 and 209 base pairs, respectively, depending on the presence or absence of the insert se- quence.

G N S D S K D ~

B 1145

Buffers-Buffer AT consisted of IO mm imidazole HCI, pH 7.2, 5% glycerol (v/v), 0.1% 2-mercaptoethanol (vlv), 0.1% Triton X-100, 1 pepstatin A, 2 pglml leupeptin, and 20 kallikrein inhibitor unitdml aprotinin. Buffer ET consisted of buffer AT containing 2 m EDTA, 1 m benzamidine, and 1% Triton X-100. Buffer EPT was the same as buffer ET except that imidazole was replaced by 10 nm sodium phos- phate, pH 7.0.

Site-directed Mutagenesis and Cloning-The full-len~h cDNA of hu- man FPFPa(+if containing the 27 nucleotide insert (22) was subcloned into the Sal1 site of M13 bacteriophage DNA (Boehringer M a ~ e i m ) . 'Ib obtain RPTPaf-if cDNA, site-directed mutagenesis was performed according to standard protocols using MI3 single-stranded DNA con- taining the RPTPa(+i) cDNA as template and an oligonucleotide de- signed to delete the 27 bases of interest (oligo-8m: 5 ' - A C ~ C C T C - C~CAGATGAGACACCAATT-3'). Mutations were confirmed by sequence analysis using Sequenase. Both cDNAs were cloned into the pSVL expression vector (Phannacia LKE Biotechnology Inc.) using the SmaI and BamHI restriction sites.

Polymerase Chain Reaction and Southern Blot Analysis-The reac- tions were carried out on a Perkin-Elmer DNA Cycler in a total volume of 20 pl containing template (controls: 10 ng of Bluescript KSII- con- taining the futl-length cDNA of RPTPa(+i) and (-i), respectively, 10'' plaque-forming units of each library, or 100 ng of A431 EDNA), 1 of primers (oligo(8s), 5'-ACTCTGACAATG%ACCA-3', oligo(T1, 5'- AGAGGACAWCCACCATCACCGCAAT-3'), 0.1 m nucleotides, 1 unit of Amplitaq polymerase in PCR buffer (both from Perkin-Elmer). Prior to adding the polymerase, the reaction mixture was incubated for 5 min at 95 "C, Thirty-eight cycles were run each carried out for 45 s at 95 "6, 1 min at 63 "C, and 1 min at 72 "6. The reaction was ended by a terminal elongation step at 72 "C for 5 min. DNA fragments were sepa- rated on a 3% agarose gel and blotted onto nitrocellulose (Schleicher & Schuell). The blot was prehybridized for 2 h at 68 "C in 6 x SSC (1 x SSC = 0.15 M NaCl, 0.015 M sodiumcitrate, pH 7.0),0.1% SDS, 5 x Denhardts (500 ml of 50 x Denhardts contains 5 g of Ficoll, 5 g of polyvinylpymol- idone, and 5 g of bovine serum albumin) and 0.1 mgiml herring sperm DNA. Hybridization with an 32P-labeled oligonucleotide designed to match the insert sequence (oligo(J1, 5'-TClTCTGTCCTTCGAGTCA- GAATTACC-3') was carried out overnight in the prehybridization solu- tion at 42 "C. The blot was washed three times for 5 min in 6 x SSC, 0.1% SDS and 0.1% NaPP, a t 70 "C and then subjected to autoradiog- raphy.

n'ssue Culture, Zbansfections, and Cell Extracts-COS1 cells (ATCC CRL1650) were maintained in high glucose Dulbecco's modified Eagle medium (Bio Whittaker) complemented with 10% fetal bovine serum (Life 'kchnologies, Inc.), 100 unitdd penicillin, and 100 llglmI strep- tomycin (both from JRH Biosciences) a t 37 'C in 5% CO,. Transfections were carried out using a DNA/Ca2+-phosphate precipitation method as described previously (26). For transfection of a 10-cm dish ofCOS-1 cells grown to 10-20% confluence, 10-30 pg of DNA were used. After 16 h, cells were repeatedly washed with PBS and further incubated for 48-46 h. Cells were then washed twice with PBS and released by incubation

'. .. ... ,.... -.. .. ......

for 20 min in 5 mi Ca2+/Me-free PBS containing 0.2 mg/ml EDTA. The cell suspension was spun 2 min at 1000 x g, and the collected cells were washed in PES, spun again, and finally resuspended in 0.1-0.2 mI of extraction buffer (buffers ET or EPT). The homogenate was kept for 20 min on ice and the insoluble material precipitated by a 2-min centrifu- gation at 1000 x g. The clear s u ~ ~ a t a n t is referred to as the cell extract. Protein ~ncentration was d e t e ~ n e d according to Bradford (27).

A n t i ~ ~ i e s , Western Blot Analysis, and Im~un~precipi~a~ion- Peptide antibody Ab35 was described pre~ously (21). Antibody Ab9104 was obtained from a goat following repeated i m m ~ ~ t i o n s with puri- fied baculo~~s-expressed RPTPa 128). SDS-PAGE and Western blot- ting were performed as described (29, 30). Dilution of antibodies was 1500 for Ab35 (serum) and 1:3000 for Ab9104 (affinity purified with an antigen column). Bands were detectad with an alkaline phosphatase coupled secondary antibodies (Bio-Rad and Boehringer Mannheim). When the 130-kDa protein of RPTPa had to be recognized by Ab9104, blots were treated at 80 'C in 0.05 N R2S0, prior to incubation with the antibody to unmask the epitope (31, see "Results"). Immunoprecipita- tions were carried out with 5 p1 of Ab35 (serum) and 20 pl of protein A-Sepharose (5 mgiml in PES) (Sigma) overnight a t 4 "C. The beads were washed three times in TTBS (20 mM "is-HC1, pH 7.4,10% glycerol fvlv), 0.05% Tween 20, and 0.5 M NaCl).

Reglyco~lution of RPTPa-COS-1 cells transiently overexpressing either isozyme of RPTPa were extracted in buffer EPT. Forty pl of extract (corresponding to half the cells of a 10-cm dish) were incubated at 37 'C after addition of the following endoglycosidases (or buffer) as indicated: 100 milliunits of neuraminidase (Sigma) for 1 h, 300 milli- units of endoglycosidase F (Sigma) for 1 h, and 1.25 milliunits of endo- a-N-acetylgalactosaminidase (Genzyme) for 3 h. Reactions were stopped by adding SDS-PAGE lysis buffer.

Pulse-chase-Cells transiently expressing RPTPa were washed twice with methionine-free medium (MFM, Sigma D4655, supplemented with 4 m L-glutamine, 0.8 IILM L-leucine, 0.8 nm L-lysine, and 3.7 g of NaHCOgiter) and incubated for 30 min in 2 d l 0 - c m dish of MFM containing 0.4 mCi of [35S]methionine (ICN Tran%-Label). Cells were washed twice with MFM before the chase medium (10 ml of complete medium supplemented with 1 m methionine) was added. At indicated time points, cells were lysed by adding 1 ml of buffer ET, incubated for 30 min on ice, and harvested by scraping with a rubber policeman. The homogenate was cleared by centrifugation for 5 min at 1000 x g. RPTPor was immunoprecipitated from 500 pI of the extract corresponding to cells contained in half a 10-cm dish. Following SDS-PAGE (29), the gel was incubated for 15 min in the enhance solution (Amplify, Amersham Corp.), dried, and subjected to autoradiography using a preflashed film.

Superose 12 Chromatography-Protein (1.5 mg) from COS1 cells transiently expressing RPTPa was chromatographed on a Superose 12 column (Pharmacia, HR10/30) equilibrated in buffer AT containing 0.1 M NaCI. The flow rate was 0.5 mlimin and fractions of 0.5 ml were collected. Three-hundred pl of each fraction were trichloroacetic acid precipitated (10% final concentration) and subjected to SDS-PAGE (29)

10526 Multiple Forms of Human RPTPa followed by Western blot analysis (30). Immunoprecipitations were car- ried out with Ab35 using 0.2 ml of each fraction. The immunoprecipi- tates were assayed for tyrosine phosphates activity toward the ENDY(P)INASL peptide as described (32).

RESULTS

Existence of l lvo RPTPa Isozymes-The sequence of RPTPa was determined from various cDNA libraries (3, 20-23). A single clone from a human brainstem cDNA library, however, differed from the others in that it contained an insert of 27 nucleotides immediately preceding the transmembrane seg- ment of the enzyme (22, see Fig. LA ). The two isoforms will be referred to as RPTPa(+i) and ( 4 ) depending on the presence or absence of the insert. To investigate whether expression of ei- ther isoform was tissue specific, PCR was performed using as templates A431 cDNA and T-cell and vascular smooth muscle libraries. Primers were chosen to hybridize up- and down- stream from the putative insert sequence. Vectors bearing the cDNAs of RPTPa(+i) and ( 4 ) were included as positive controls. Fig. IB summarizes the configuration of primers used and the size of the expected PCR products. As shown in Fig. 2A, DNA fragments of both sizes were generated from the A431 cDNA and the two libraries, whereby the yield of the smaller fragment was considerably higher in all three instances. The presence of the insert sequence in the upper band was further demonstrated by hybridization of a Southern blot with a specific oligonucle-

bP 1057

770 612 495

392 335/341/345

297l291

210

162

A B

1 2 3 4 5 1 2 3 4 5

FIG. 2. Both isoforms of RPTPa are expressed simultaneously. A, separation of PCR generated fragments on agarose corresponding to RPTPa(+i) and (-i), respectively; the reaction mixtures contained the following templates: vector controls, containing RPTPa(-i) (lane 1 ) and

4 ) , and A431 cDNA (lane 5). B, autoradiography of a Southern blot of (+i) (lane 2 ) , vascular smooth muscle library (lane 3 ) , T-cell library (lane

the same gel as shown in A probed with oligo(J) matching the insert sequence.

kDa

FIG. 3. Deglycosylation of RPTPa. Extracts of COS-I cells transiently ex- pressing RPTPa(-i) were incubated vari- ous endoglycosidases as indicated fol- lowed by SDS-PAGE and Western blot analysis. Blots were screened with Ab35 and Ab9104, respectively. Arrows indicate the peak fraction of thyroglobulin (670 kDa), immunoglobulin G (158 kDa), and ovalbumin (44 kDa).

otide probe (Fig. 2 B ) . The data show that both isozymes are simultaneously expressed in these tissues and the cell line, re- spectively, with RPTPa(-i) being the predominant form.

Characterization of Ab35 and Ab9104-By Western blotting, Ab9104 was approximately 10 times more sensitive thanAb35, detecting 1 ng of baculovirus expressed RPTPa as compared with 10 ng (data not shown). As demonstrated below, however, Ab9104 recognized the major 130 kDa species of RPTPa from COS-1 cells only after the blots had been treated with sulfuric acid (see under "Experimental Procedures") which releases sialic acid (31) and unmasks the epitope. Interestingly, the same results were obtained with a rabbit antibody raised against the baculovirus expressed phosphatase (data not shown), indicating that both antibodies are directed against the same epitope.

Expression of RPTPa(+i) and (4) in COS-1 Cells-Both RPTPa isoforms were transiently expressed in COS-1 cells us- ing the pSVL vector. The cDNA encoding for RPTPa(-i) was obtained by deleting the insert sequences from the (+i) species by site-directed mutagenesis. All experiments described below were performed with extracts of COS-1 cells expressing the (+i) or ( 4 ) isoform of RPTPa. No difference was observed between the two isozymes with regard to size, in vitro enzymatic activity or glycosylation pattern. Therefore, no distinction will be made between the two isoforms in the following sections.

Glycosylation ofRPTPa-The results are summarized in Fig. 3. Expression of RpTPa in COS-1 cells yielded two proteins. In Western blots of cell extracts, Ab35 detected a material of 130 kDa, whereas Ab9104 recognized a 100-kDa protein only (lanes 1 and 6). Since both proteins were generated by the same cDNA, it seemed likely that the difference in size was due to a translational modification such as glycosylation. To investigate this possibility, cell extracts of COS-1 cells were incubated with different endoglycosidases prior to Western blot analysis. While treatment with neuraminidase did not result in a visible change in the size of either protein band, i t now enabled Ab9104 to recognize both the 100- and 130-kDa species (lanes 2 and 7). Removal of N-linked carbohydrates by N-glycosidase F resulted in a slight decrease in apparent molecular mass of both proteins. Under these conditions, however, Ab9401 only detected the smaller species (lanes 3 and 8). Again, when neuraminidase was added, both were recognized by Ab9104 (lanes 4 and 9). Further removal of 0-glycosylated residues by endo-a-N-acetylgalactosaminidase led to the appearance of many bands corresponding to decreasing molecular masses down to -85 kDa which correlates well to the mass of the

Ab 35 Ab 9104 kDa

180 - 110 - 80 - 65 - 55 - 38 -

1 2 3 4 5 6 7 8 9 1 0

- 180

- 110

- 80 - 65 - 55

- 38

neuraminidase - + - + + - + - + + N-glycosidase F

endo+N-acetyl- galactosaminidase

- - + + + " + + +

Multiple Forms of Human RPTPa! 10527

protein deducted from its amino acid composition. The decline of size was accompanied by a significant increase in staining of these bands by Ab9104 (lanes 5 and IO). These data are con- sistent with the view that the extracellular segment of RPTPa contains a powerful epitope that is masked by 0-glycosylation.

The 100-kDa RPTPa Is a Precursor of the 130-kDa Species-To test the possibility that the 100-kDa species of RPTPa is a precursor of the 130-kDa protein, COS-1 cells tran- siently expressing the phosphatase were pulse-labeled with ["Slmethionine and extracted a t different time points after chasing the label. The enzyme was immunoprecipitated with Ab35 and analyzed by autoradiography following SDS-PAGE. As shown in Fig. 4, 10 min after the chase the label was only present in the smaller species of the phosphatase, whereas 30 min later it appeared in the 130-kDa protein. After 2 h the label was exclusively found in the larger species.

Apparent Molecular Mass and Enzymatic Actiuity-Extracts of COS-1 cells transiently expressing RPTPa were fractionated on Superose 12, and the phosphatase was immunoprecipitated from each fraction with Ab35. An aliquot of the immunobeads was assayed with the ENDIY(P)INASL peptide substrate, and the remainder was subjected to SDS-PAGE followed by West- ern blot analysis usingAb9104. The results are shown in Fig. 5. Both the 100- and 130-kDa species of the phosphatase emerged from the column at an apparent molecular mass higher than expected from SDS-PAGE, with the peak fractions correspond- ing to molecular masses of 200,000 and 340,000, respectively.

kDa

130 100

chase 0 10 30 60 120 min FIG. 4. The 100 kDa RPTPa is a precursor of the 130-kDa spe-

cies. COS-1 cells transiently expressing RPTPa(-i) were labeled with [35Slmethionine. After addition of the cold amino acid, cells were ex- tracted at various times as indicated. RPTPa was immunoprecipitated with Ab35 and subjected to SDS-PAGE. The gel was analyzed by auto- radiography.

FIG. 5. Superose 12 chromatogra- phy of RF'TPa. The Superose 12 chroma- tography of extracts from COS-1 cells transiently expressing RPTPa(-i) was performed as described under "Experi- mental Procedures." Following immuno- precipitation, PTP activity was deter-

Western blotting. mined and the protein was analyzed by

1000

0

An estimation of the amount of phosphatase based on Western blot staining led to the conclusion that both RPTPa proteins exhibited the same in vitro enzymatic activity.

DISCUSSION

Many receptor-linked tyrosine phosphatases exist in mul- tiple forms that differ in their extracellular domain, originating either from alternative splicing, as shown for CD45 (reviewed in Ref. 101, or from proteolytic cleavage of the enzyme as re- ported for LAR (32) and RPTPk (13). The two RPTPa isozymes cloned from human cDNA libraries are also likely the result of alternative splicing, although this has not been confirmed since the genomic sequence remains to be determined.

The consequences of the heterogeneity of RPTPa isozymes are unknown. Transient expression of both isozymes in COS-1 cells revealed no obvious difference with regard to enzymatic activity measured under in vitro conditions, size and glycosyla- tion pattern. One may speculate, however, that the addition of 9 amino acid residues close to the transmembrane segment of the phosphatase might change its conformation in such a way to alter its specificity for soluble ligands or other interacting molecules. Tissue specific or developmental specific expression of each isozyme would indicate a specific function. Although the PCR data show that both isozymes are present simultaneously in various tissues or cell lines, it is possible that different re- sults are obtained with other tissues or cells in different devel- opmental stages. Therefore, these studies should be extended, especially to the brain since the RPTPa(+i) isoform was origi- nally cloned from a brain stem library (22).

In addition to the existence of isozymes, RPTPa is present in two distinct species. One contains only N-linked carbohydrates, and the other contains both N- and 0-linked carbohydrates. The difference in glycosylation results in a remarkable differ- ence in size of 100 uersus 130 kDa, as determined by SDS- PAGE. The discovery of the smaller species in mammalian cells was enabled by the high affinity, polyclonal antibody Ab9104 that only recognized the 100-kDa protein when standard meth- ods of Western blotting and immunoprecipitation were used. This antibody was raised against purified material from St9 cells following infection with recombinant baculovirus. Since these insect cells are not capable of 0-glycosylation, only the smaller species was produced (28). The initially surprising fact that Ab9104 does not detect the 130-kDa protein can be ex-

. - 130

lo0 PI

10528 Multiple Forms of Human RPTPa plained with the existence of an epitope in the extracellular domain of the molecule which is masked by 0-glycosylation when the protein is expressed in mammalian cells. Indeed, a series of experiments with glycosidases revealed that the epitope for Ab9104 can be freed by cleavage of sialic acid, and more efficiently, by removal of 0-linked carbohydrates.

Pulse-chase experiments clearly showed that the 100-kDa species is a precursor of the 130-kDa protein. However, the question is still open, whether the smaller species has a distinct function in the cell. Since N-glycosylation occurs in the endo- plasmatic reticulum, the exclusively N-glycosylated phospha- tase is likely to localize in the same membrane system known to contain at least two other tyrosine phosphatases, the non- receptor, intracellular enzymes PTPlB (34) and TCPTP.2 The glycosylation of RPTPa seems to be a highly specific reaction and, therefore, of biological relevance, since we never detected any other %labeled species of RPTPa except the two men- tioned above. We assume that this translational modification contributes largely to the specificity for ligand binding.

Interestingly, both the 100- and 130-kDa species emerge from Superose 12 columns in fractions corresponding to an approximate molecular mass 2-fold higher than expected. One explanation for this chromatographic behavior could be the presence of phosphatase dimers, as recently suggested for CD45 (35). Efforts to resolve this question are currently in progress.

the A431 cDNA and Dr. C. Giachelli for the vascular smooth muscle cell Acknowledgments-We thank Dr. R. Seger for the T-cell library and

library.

REFERENCES

1. Fischer, E. H., Charbonneau, H., and Tonks, N. K. (1991) Science 253,401406 2. Charbonneau, H., and Tonks, N. K. (1992) Annu. Reu. Cell Biol. 8,463493 3. Krueger, N. X., Streuli, M., and Saito, H. (1990) EMBO J. 9,3241-3252 4. Yang, X., Seow, K. T., Bahri, S. M., Oon, S. H., and Chia, W. (1991) Cell 67,

5. Tian, S.-S., Tsoulfas, P., and Zinn, K. (1991) Cell 67, 675485 6. Wang, Y., and Pallen, C. J. (1991) EMBO J. 10, 3231-3237 7. Tan, X., Stover, D. R., and Walsh, K. A. (1993) J. Biol. Chem. 268,68354838 8. Streuli, M., Krueger, N. X., Thai, T., Tang, M., and Saito, H. (1990) EMBO J.

661473

9, 2399-2407

J. Lorenzen, manuscript in prepara t ion .

9. Koretzky, G. A. (1993) F M E E J. 7, 420-426 10. Hathock, K. S., Hirano, H., and Hodes, R. (1993) Zmmunol. Res. 12, 21-36 11. Streuli, M., Krueger, N. X., Hall, L. R., Schlossman, S. F., and Saito, H. (1988)

J. Exp. Med. 168, 1523-1530 12. Gebbink, M. E G. B., van Etten, I., Hateboer, G., Suijkerbuijk, R., Beijersber-

gen, R. L., Geurts van Kessel, A,, and Moolenaar, W. H. (1991) FEES Lett. 290, 123-130

13. Jiang, Y.-P., Wang, H., D'Eustachio, P., Musacchio, J. M., Schlessinger, J., and

14. Gebbink, M. F. B. G., Zondag, G . C. M., Wubbolts, R. W., Beijersbergen, R. L., Sap, J. (1993) Mol. Cell. B i d . 13, 2942-2951

15. Brady-Kalnay, S. M., Flint, A. J., and Tonks, N. K. (1993) J. Cell B i d . 122,961- van Etten, I . , Moolenaar, W. H. (1993) J. Biol. Chem. 268, 16101-16104

972 16. Streuli, M., Krueger, N. X., Tsai, A. Y. M., and Saito, H. (1989) Proc. Natl. Acad.

17. Krueger, N. X., and Saito, H. (1992) Proc. Natl. Acad. Sci. U. S . A. 89, 7417- Sci. U. S. A. 86, 8698-8702

7421 18. Levy, J. B., Canoll, P. D., Silvennoinen, O., Barnea, G., Morse, B., Honegger, A.

M., Huang, J.-T., Cannizzaro, L.A., Park, S.-H., Dmck, T., Huebner, K., Sap, J., Ehrlich, M., Musacchio, J. M., and Schlessinger, J. (1993) J. Biol. Chem. 268, 10573-10581

19. Barnea, G., Silvennoinen, O., Shaanan, B., Honegger, A. M., Canoll, P. D., D'Eustachio, P., Morse, B., Levy, J. B., Laforgia, S., Huehner, K., Huang, J.-T., Cannizzaro, L. A,, Park, S.-H., Druck, T., Huebner, K., Musacchio, J.

20. Matthews, R. J. , Cahir, E. D., and Thomas, M. L. (1990) Proc. Natl. Acad. Sci. M., Sap, J., and Schlessinger, J. (1993) Mol. Cell. B i d . 13, 1497-1506

21. Sap, J., D'Eustachio, P., Givol, D., and Schlessinger, J. (199O)Proc. Natl. Acad. U. S. A. 87, 444&4448

Sci. U. S. A. 87, 6112-6116 22. Kaplan, R., Morse, B., Huebner, K., Croce, C., Howk, R., Ravera, M., Ricca, G.,

Jaye, M., and Schlessinger, J. (1990) Proc. Natl. Acad. Sci. U. S. A. 87,7000- 7004

23. Jirik, E R., Janzen, N. M., Melhado, I. G., and Harder, K. W. (1990) FEBSLett. 273,239-242

24. den Hertog, J., Pals, C. E. G. M., Peppelenbosch, M. P., Tertoolen, L. G. J., de Laat, S. W., and Kruijer, W. (1993) EMBO J. 12,37893798

25. Zheng, X. M., Wang, Y., and Pallen, C. J. (1992) Nature 359.336339 26. Chen, C., and Okayama, H. (1987) Mol. Cell. B i d . 7, 2745-2752 27. Bradford, M. M. (1976)Anol. Biochem. 72, 248-254 28. Daum, G., Zander, N. F., Morse, B., Hurwitz, D., Schlessinger, J., and Fischer,

29. Laemmli, U. K. (1970) Nature 227,680-685 30. Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. U. S. A.

76,435M354 31. Kijimoto-Ochiai, S., Katagiri, Y. U., and Ochiai, H. (1985)AnaL. Biochem. 147,

222-229 32. Daum, G., Solca, F., Diltz, C. D., Zhao, Z., Cool, D. E., and Fischer, E. H. (1993)

And. Biochem. 211,5&54 33. Streuli, M., Krueger, N. X., Ariniello, P. D., Tang, M., Munro, J. M., Blattler,

W. A,, Adler, D. A,, Disteche, C. M., and Saito, H. (1992) EMBO J. 11,

34. Frangioni, J. V., Beahm, P. H., Shifrin, V., Jost, C. A., and Neel, B. G. (1992) 897- 907

35. Desai, D. M., Sap, J., Schlessinger, J., and Weiss, A. (1993) Cell 73, 541-554 Cell 68,545-560

E. H. (1991) J. B i d . Chem. 266, 12211-12215