Human 70-kDa SHP-1L Differs from 68-kDa SHP-1 in Its C-terminal Structure and Catalytic Activity*

The tyrosine phosphatase SHP-1 functions as a negative regulator in hematopoietic cell development, proliferation, and receptor-mediated cellular activation. In Jurkat T cells, a major 68-kDa band and a minor 70-kDa band were immunoprecipitated by a monoclonal antibody against the SHP-1 protein-tyrosine phosphatase domain, while an antibody against the SHP-1 C-terminal 19 amino acids recognized only the 68-kDa SHP-1. The SDS-gel-purified 70-kDa protein was subjected to tryptic mapping and microsequencing, which was followed by molecular cloning. It revealed that the 70-kDa protein, termed SHP-1L, is a C-terminal alternatively spliced form of SHP-1. SHP-1L is 29 amino acids longer than SHP-1, and its 66 C-terminal amino acids are different from SHP-1. The C terminus of SHP-1L contains a proline-rich motif PVPGPPVLSP, a potential Src homology 3 domain-binding site. In contrast to SHP-1, tyrosine phosphorylation of SHP-1L is not detected upon stimulation in Jurkat T cells. This is apparently due to the lack of a single in vivo tyrosine phosphorylation site, which only exists in the C terminus of SHP-1 (Y564). COS cell-expressed glutathione S-transferase-SHP-1L can dephosphorylate tyrosine-phosphorylated ZAP70. At pH 7.4, SHP-1L was shown to be more active than SHP-1 in the dephosphorylation of ZAP70. At pH 5.4, SHP-1L and SHP-1 exhibited similar catalytic activity. It is likely that these two isoforms play different roles in the regulation of hematopoietic cell signal transduction.

SHP-1 plays a critical biological role in the regulation of hematopoietic cell growth and development. Motheaten mice (me/me) and viable motheaten mice (me v /me v ) are natural genetic models of mammals lacking the expression of functional SHP-1 (13,14). These mice suffer from chronic macrophage and neutrophil activation, and abnormal B cell development, as well as T and B cell depletion and dysfunction. In me/me mice, hematopoietic cells hyper-proliferate in response to growth factor stimulation, which results in the enormous overexpansion of multiple hematopoietic cell lineages. The elevated proliferation and accumulation of myeloid cells appears to be responsible for the early death of the me/me mice (6,8,15).
The structures of SHP-1 and SHP-2 are similar (37,38). The catalytic domains are conserved in all PTPs. The tandem N-SH2 and C-SH2 domains of SHP-1 and SHP-2 provide the docking sites for phosphorylated tyrosine residues. Several reports have shown that a truncation of 35-49 amino acids from the C terminus of SHP-1 dramatically increases its catalytic activity by 20 -40-fold (39,40). Interestingly, truncation of 60 amino acids from the C terminus of SHP-1 had no affect on SHP-1 catalytic activity. It has been hypothesized that the entire C-terminal domain of SHP-1 contains two regions with the N-region 50 amino acid residues being critical for the activation of SHP-1, which is blocked by the C-region 30 -40 residues. On the other hand, the catalytic activity of SHP-1 and SHP-2 is known to be negatively regulated through an intramolecular interaction between the SH2 domain(s) and catalytic domain (37,38). It has been shown that the engagement of the SH2 domain with an exogenous phosphotyrosyl peptide or truncation of the SH2 domain(s) activates SHP-1 and SHP-2 (40 -44). It has also been shown that the catalytic activity of SHP-1 displays a distinctive pH dependence (40). Escherichia coli-expressed SHP-1 is nearly inactive at pH 7.4 as determined in vitro using PNPP as a substrate, but is active at pH 5.4 (40). This suggests that a more open conformation of pnitrophenyl phosphate SHP-1 may be formed at pH 5.4. The recently published crystal structure of the catalytic domain of SHP-1 suggests that SHP-1 needs to change from the half-open conformation to the open conformation before it can bind substrates (38). The full-length SHP-2 crystal structure reveals a "closed" domain arrangement in which the two SH2 domains contour around the catalytic domain with the N-terminal domain directly blocking the active site (37). It is not clear whether the C-terminal tail is involved in the folding of SHP-2, since the crystal structure of SHP-2 is missing 66 amino acids from the C terminus. It has been observed that, when the C-terminal 35 amino acids were truncated, SHP-1 was not further activated by phosphotyrosine peptide engagement. These data suggest that a C-terminal truncation alone is sufficient to activate the PTPs (40).
In this report, we described the identification of SHP-1L, a splice isoform of SHP-1. SHP-1L contains a distinctive C terminus of 66 amino acid residues and exhibits a high catalytic activity at physiological pH. Thus, we suggest that this isoform may, unlike SHP-1, function in resting cells.

EXPERIMENTAL PROCEDURES
Cells and Antibodies-The Jurkat T cell line J77, a variant of clone E6 -1 (ATCC) was cultured in RPMI 1640 medium, and COS cells were cultured in minimal essential medium, supplemented with 10% fetal calf serum at 37°C in a 5% CO 2 humidified atmosphere. Anti-phosphotyrosine (RC20), anti-SHP-1-PTP domain mAb, anti-SHP-2 mAb, and anti-ZAP70 mAb were purchased from Transduction Laboratories (Lexington, KY). Anti-SHP-1 C-terminal 19 AA, an anti-SHP-1 polyclonal antibody raised to the 19 C-terminal amino acid, was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). When not specified, the mAb against the SHP-1 PTP domain was used. Anti-CD3 antibody (OKT3) was prepared from a hybridoma obtained from the ATCC.
Immunoprecipitation and Immunoblot Analysis-Jurkat T cells (2 ϫ 10 7 ) were washed and resuspended in 1 ml of PBS. For CD3 stimulation, cells were incubated with OKT3 (2 g/ml) for 5 min on ice, crosslinked by rabbit anti-mouse IgG (5 g/ml) on ice for another 5 min, then incubated at 37°C for 3 min. For pervanadate stimulation, cells were incubated with 1 M pervanadate at 37°C for 3 min. Pervanadate was prepared by mixing 20 l of 1 M vanadate with 11 l of 30% H 2 O 2 and 69 l of double-distilled H 2 O, which was incubated at room temperature for 15 min before use. After washing with PBS, cells were lysed in 1 ml of Nonidet P-40 lysis buffer (1% Nonidet P-40, 150 mM NaCl, 20 mM Tris, pH 7.4, 0.5% sodium deoxycholate, 50 mM NaF, 1 mM PMSF, 1 g/ml leupeptin, 2 g/ml antipain) at 4°C for 30 min. The Nonidet P-40 lysate was centrifuged at 12,000 ϫ g for 15 min at 4°C. Immunoprecipitations were carried out at 4°C overnight or at room temperature for 4 h with protein A-Sepharose beads. The beads were washed twice with 0.1% Triton X-100/TBS and once with TBS. Proteins were eluted from the beads by boiling for 5 min in 50 l of Laemmli reducing SDS sample buffer. Proteins from about 10 7 cells were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membranes. Membrane were blocked with 3% bovine serum albumin/TBST and incubated with antibodies in TBST for 2 h at room temperature. Following four 15-min washes with TBST, the membranes were incubated with second antibody for 30 min, washed three times for 5 min each with TBST, and developed by ECL (Amersham Pharmacia Biotech).
Subcellular Fractionation of Jurkat T Cells-After washing in PBS, Jurkat cells (2 ϫ 10 8 ) were incubated in 2 ml of hypotonic buffer (42 mM KCl, 10 mM Hepes, pH 7.4, 5 mM MgCl 2 ) for 15 min at 4°C. The cells were passed through a 30-gauge needle 10 times. The extract was centrifuged at 250 ϫ g for 10 min to remove the nuclei and intact cells. The postnuclear supernatant was centrifuged at 150,000 ϫ g for 30 min at 4°C to separate the cytoplasm from the membrane fraction.
Isolation of SHP-1L and SHP-1 Protein from Jurkat T Cells-Pellets from 10 liters of Jurkat T cell culture were lysed in 50 ml of Nonidet P-40 lysis buffer. After centrifugation, the supernatants were mixed with 5 ml of ThioBond resin (PAO-agarose) pretreated according to manufacturer's instruction (Invitrogen, Carlsbad, CA). The mixture was incubated at 4°C for 1 h, then packed into a mini-column of 20 ml.
The affinity resin was then washed with 100 ml of washing buffer (150 mM NaCl, 100 mM Tris, pH 7.4). Thio-resin-bound protein was eluted with 10 ml of the above buffer containing 200 mM ␤-mercaptoethanol. Column eluates were dialyzed against TBS three times for 30 min each and adjusted to 0.5% Nonidet P-40. SHP-1 was then immunoprecipitated by adding 100 l of anti-SHP-1 (Transduction Laboratories), and 200 l of protein A beads. Immunoprecipitates were separated in 8% SDS-PAGE and briefly stained with Coomassie Blue before excising for microsequencing.
Tryptic Digestion, High Performance Liquid Chromatography (HPLC) Separation, and Microsequencing-All procedures were carried out by the Harvard Microchemistry Facility. After in gel reduction and S-carboxyamidomethylation, the band(s) were subjected to in gel tryptic digestion (Promega) and a single 10% aliquot was analyzed by mass spectrometry. Sequence information was determined by microcapillary (75 m ϫ 10-cm column, packed-in house) reverse-phase chromatography, coupled to the electrospray ionization source of a quadrupole ion trap mass spectrometer (Finnigan LC, San Jose, CA). The remainder (90%) of the peptide mixture was separated by microbore high performance liquid chromatography using a Zorbax C18 1.0 mm ϫ 150-mm reverse-phase column coupled to a Hewlett-Packard 1090 HPLC/1040 diode array detector. Optimum fractions were chosen based on differential UV absorbance at 205, 277, and 292 nm, peak symmetry, and resolution; then further screened for length and homogeneity by matrixassisted laser desorption time-of-flight mass spectrometry. Tryptic peptides were submitted for automated Edman degradation on a Perkin Elmer/Applied Biosystems 494A cLC protein sequencer (Foster City, CA).
Molecular Cloning of SHP-1L-A human Jurkat leukemia T-cell cDNA library from CLONTECH (HL5008b) was screened for SHP-IL cDNA, using the oligonucleotide gcttggagtctagtgcagggaccgtgg as the probe. Filters were hybridized at 52°C in QuikHyb solution (Stratagene), washed twice by 2 ϫ SSC, 0.1% SDS, and washed once with 0.2 ϫ SSC, 0.1% SDS for 30 min each at 55°C. From a total of 2 ϫ 10 5 phage plaques, 11 positive clones were obtained. The EcoRI insert was subcloned into pBluescript. Two clones containing 2.2 kilobases of fulllength cDNA were submitted for sequencing.

Construction of GST-SHP-1L Fusion Plasmid and Expression in COS Cells-To construct GST-SHP-1L
and GST-SHP-1 fusion genes, cDNAs encoding SHP-1 and SHP-1L were amplified by PCR with NotI sites in both 3Ј and 5Ј primers. The PCR products were subcloned in frame into the NotI site of pEBG. The fusion gene structures were then confirmed by sequencing. The pEBG fusion plasmids were transfected into COS cells by using the DEAE/dextran method. Cells were cultured for 48 h before harvest. To isolate the expressed GST fusion proteins, cells were washed with PBS, incubated with hypotonic buffer for 10 min at 4°C, and then removed from the culture dishes. These cells were further lysed by passing through the 30-gauge needle for 10 times. After 30 min of centrifugation at 15,000 ϫ g, the supernatants were added to glutathione-Sepharose beads to absorb GST fusion protein. The GST-SHP-1L and GST-SHP-1 proteins were eluted by 20 mM glutathione, dialyzed, concentrated, and quantitated by UV spectrometer.
PTP Activity of GST-SHP-1L and GST-SHP-1-To prepare tyrosinephosphorylated ZAP70, 200 ϫ 106 Jurkat T cells were treated with pervanadate, lysed, and immunoprecipitated with anti-ZAP70 mAb plus protein A beads. ZAP70 kinase/protein A-Sepharose beads were diluted with two volumes of Sepharose beads and divided into aliquots for dephosphorylation assay. To study the kinetics of dephosphorylation, purified GST-SHP-1L or GST-SHP-1 was mixed with ZAP70/ protein A beads in 25 l of pH 7.4 buffer (100 mM Hepes, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM ␤-mercaptoethanol (BME) or pH 5.4 buffer (100 mM sodium acetate, pH 5.4, 150 mM NaCl, 1 mM EDTA, 1 mM BME). The reaction was stopped by adding 1 ml of cold TBS. The beads were pelleted and boiled in SDS sample buffer. The results were obtained by anti-Tyr(P) (RC20) immunoblotting.

Identification of a 70-kDa Molecule in Jurkat T Cells Immunoprecipitated by
Anti-SHP-1-Jurkat T cell lysates were immunoprecipitated with a mAb against the PTP domain of human SHP-1. Two proteins with apparent molecular masses of 68 and 70 kDa were detected upon immunoblotting (Fig. 1A). When similar immunoprecipitations were carried out by using a polyclonal Ab raised against the C-terminal 19 amino acids of human SHP-1, only the major 68-kDa SHP-1 band was detected. This suggested that the 70-kDa protein might differ from the 68-kDa SHP-1 at the C terminus. To purify the 70-kDa protein, we adapted a novel approach using ThioBond resin originally designed for affinity purification of thioredoxin fusion proteins. The resin consists of an agarose-based support to which phenylarsine oxide (PAO) is covalently bound. PAO is a PTP inhibitor, which may oxidize and bind to the thiolate anion of the cysteine residue in the PTP-reactive center, blocking the formation of a phosphoryl-cysteine intermediate, a critical step in dephosphorylation (45). Based on the x-ray structure of Yersinia PTP co-crystallized with vanadate, a specific PTP inhibitor, the PTP inhibitors are able to covalently bind to the cysteine residue in the PTP-reactive center (46). To determine whether SH2 containing PTPs will bind to a PAO column, Nonidet P-40 lysates of Jurkat T cells were passed through the PAO column. Column-bound proteins were eluted by BME and analyzed by immunoblotting. Both SHP-1 and the 70-kDa protein were found to be in the column bound protein mixture by anti-SHP-1 immunoblotting (Fig. 1B). When re-blotted with anti-SHP-2, SHP-2 was determined to migrate slightly slower than the 70-kDa protein (Fig. 1B). The results indicated that SHP-1, the 70-kDa protein, and SHP-2 all covalently bound to a PAO column and were eluted by BME. The PAO affinity-purified protein mixture was further purified by anti-SHP-1 immunoprecipitation (Fig. 1C). When the relative amount of 70-kDa protein immunoprecipitated from the T cell lysates (Fig. 1A) was compared with that immunoprecipitated from the PAO column elutes (Fig. 1C), it was found that the 70-kDa protein appeared to be enriched by the PAO affinity purification, suggesting that the 70-kDa protein might have a higher affinity than SHP-1 for PAO. The binding of SHP-1 and the 70-kDa protein to PAO was not affected by stimulation of Jurkat T cells via CD3 cross-linking (Fig. 1C). To determine the intracellular localization of the 70-kDa protein, Jurkat T cells were fractionated into membrane and cytoplasm fractions, lysed, and purified on a PAO column, followed by SDS-PAGE and immunoblotting with anti-SHP-1. The results showed that both SHP-1 and the 70-kDa proteins appear predominantly in the cytosol (Fig. 1D).
The 70-kDa Protein Is a C-terminal Alternatively Spliced Form of SHP-1-To isolate the 70-kDa protein, lysates from 10 liters of cultured Jurkat T cells were PAO affinity-purified, then immunoprecipitated with anti-SHP-1. Both the 68-kDa SHP-1 and 70-kDa protein bands were excised from SDS-PAGE and subjected to trypsin digestion. The tryptic peptides were analyzed by HPLC (Fig. 2A). The HPLC profiles from 68-kDa SHP-1 and from the 70-kDa protein appeared to be quite similar, except that several peaks from the 70-kDa protein were distinctive. Similar results were also obtained by mass spectrometry (data not shown). Three peaks from HPLC and two from mass spectrometry were collected and microsequenced (Fig. 2B). The sequencing results showed that two distinctive peaks from the 70-kDa protein did not match the SHP-1 sequence, whereas the other three peaks from the 70-kDa protein matched different regions of SHP-1 (Fig. 3, A and B). These results indicated that the 70-kDa protein shared some common regions with SHP-1 but also contained unique regions not present in SHP-1 molecule. A GenBank search found that the DNA sequence encoding the unique region of the 70-kDa protein is located on human chromosome 12p12-p13 connected to exon 16 of SHP-1 and that the 70-kDa protein is generated by alterna- tively splicing and a reading frameshift in exon 16 of SHP-1 (Fig.  3, A and B) (2). We termed the 70-kDa protein SHP-1L, since its molecular mass appears larger than SHP-1.
To isolate the cDNA coding SHP-1L, a Jurkat cDNA library was screened by using an oligonucleotide probe specific for the SHP-1L unique sequence. Two identical clones were identified and DNA sequenced (GenBank accession no. AF178946). The cloned SHP-1L cDNA is identical to human SHP-1 cDNA except for a 144-nucleotide insertion between nucleotides 1785 and 1786 generated by alternative splicing. Sequencing revealed the alternative splice sites of exon 16 (Fig. 3B). Due to the reading frameshift, the amino acid sequence encoded by exon 16 of SHP-1L is entirely different from the sequence encoded by exon 16 of SHP-1 (Fig. 3, B and C). The cDNA predicted full-length SHP-1L amino acid sequence is shown in Fig. 3C. The open reading frame of the SHP-1L cDNA encodes a protein of 624 amino acids, which is 29 amino acids longer than SHP-1. The N-terminal 558 amino acids of SHP-1L are identical to SHP-1, and the C-terminal 66 amino acids are distinct. The SHP-1L cDNA predicts a PTP containing two contiguous SH2 domains and a PTP domain identical to that of SHP-1. There is no sequence homology between the C terminus of SHP-1 and SHP-1L. The C terminus of SHP-1 is lysine-rich, whereas SHP-1L is arginine-rich. The last three amino acids of SHP-1L are RRK, similar to the KRK of SHP-1. Within the unique C-terminal region of SHP-1L, there exists a proline-rich motif, PVPGPPVLSP, constituting a potential SH3 binding site, whereas it lacks the tyrosine phosphorylation site, ED-VYE, a potential SH2 domain binding motif that exists in the C terminus of SHP-1 (47).

SHP-1L Is Predominantly Expressed in Human
Hematopoietic Tissues-To study the cellular expression of SHP-1L, RT-PCR analysis was carried out using two sets of oligo primers. The PCR amplification using SHP-1L-specific primers (primer 1L) will generate a 365-bp PCR product specific for SHP-1L, while SHP-1 primer (primer I) will generate a 308-bp PCR product for SHP-1 and a 452-bp PCR product for SHP-1L (Fig.  4A). The PCR amplification of human Jurkat T cells and Daudi B cells generated products with predicted sizes for SHP-1 and SHP-1L (Fig. 4B). The ratio of PCR products between SHP-1 and SHP-1L was approximately 5 to 1. The identities of some larger PCR products generated by the SHP-1L-specific primers are un- known. No SHP-1L-specific PCR product was generated by SHP-1L-specific primer (1L) from murine LSTRA T cells and WEHI B cells. Using SHP-1 primer (I), PCR amplification from murine cell lines generated two main products: a small 308-bp product, which is most likely from SHP-1; and another 500-bp product, which is larger than the PCR product from human SHP-1L. These results suggest that the alternative splicing of murine SHP-1 mRNA may be different from that of human SHP-1.
To determine the distribution of SHP-1L in various human and murine tissues, PCR amplification was conducted using the SHP-1L-specific primers. The PCR product of 452 bp (bottom band) was generated from human peripheral blood leukocyte, spleen, and thymus tissues but not from human colon, ovary, prostate, small intestine, and testis tissues (Fig. 4C). Consistent with the results of PCR amplification using the murine cell lines, SHP-1L-specific PCR product was not amplified from various murine tissues including thymus, bone marrow, spleen, liver, and kidney. Similar results were obtained using tissues from normal mice and from motheaten mice (data not shown).
The protein expression of SHP-1L was detected in all human hematopoietic cell lines examined, including Jurkat T cells, Daudi B cells, and HL60 myeloid cells. However, it was not detected in all murine cell lines examined, including LSTRA T cells, WEHI B cells, and BAF-3 myeloid cells, by anti-SHP-1 immunoprecipitation and immunoblot analysis. SHP-1 protein was detected in all six cell lines described above (Fig. 4D). It is very interesting that, in addition to SHP-1 and SHP-1L, several other proteins of varying sizes were co-immunoprecipitated with anti-SHP-1 from both human and murine hematopoietic cell lysates. These co-immunuprecipitated proteins were detected by immunoblotting with anti-SHP-1 mAb. Thus, both PCR amplification and SHP-1 immunoprecipitation suggest the presence of additional isoforms or family members of SHP-1/SHP-1L with distinctive distribution in human and murine tissues.
SHP-1L Is Not Tyrosine-phosphorylated by Pervanadate Treatment in Jurkat T Cells-We previously reported that SHP-1 can be tyrosine-phosphorylated at Tyr-564 by CD3 cross-linking in BYDP but not in Jurkat T cells (47). Tyr-564 is absent in SHP-1L (Fig. 3C). Four additional tyrosine resides including Tyr-536, an in vitro tyrosine phosphorylation site of SHP-1, are still present in the C-terminal 129 amino acids of SHP-1L. To determine whether SHP-1L is tyrosine-phosphorylated upon stimulation, Jurkat T cells were stimulated either by CD3 cross-linking or by pervanadate. After immunoprecipitation, an anti-phosphotyrosine blot indicated that SHP-1 was not tyrosine-phosphorylated when Jurkat T cells were stimulated by CD3 cross-linking, consistent with our previous report (data not shown) (47). However, SHP-1 was strongly tyrosylphosphorylated in Jurkat T cells when treated with 10 M pervanadate (Fig. 5). In contrast, SHP-1L was not tyrosinephosphorylated in Jurkat T cells either by CD3 cross-linking or by pervanadate treatment. Anti-SHP-1 blots indicated that both SHP-1L and SHP-1 were immunoprecipitated. Since Tyr-564 is the only tyrosine residue missing in SHP-1L, the results strongly suggest that Tyr-564 is the site phosphorylated in vivo by pervanadate treatment of Jurkat T cells.
COS-expressed GST-SHP-1L Is More Active at pH 7.4 than SHP-1 in the Dephosphorylation of ZAP70 -GST-SHP-1L and GST-SHP-1 fusion plasmids were constructed using the vector pEBG, and then transfected into COS cells. COS cells expressing GST fusion proteins were precipitated by glutathione-Sepharose beads, eluted, and analyzed by immunoblotting with anti-SHP-1 Abs (Fig. 6A). GST-SHP-1L appears as a 97-kDa protein, which is slightly larger than GST-SHP-1. Tyrosinephosphorylated ZAP-70 was used as a substrate in a dephosphorylation assay. Its interaction with and dephosphorylation by SHP-1 have been reported previously (16). Immunoprecipitated ZAP70 was mixed with GST-SHP-1L (0.25 mg/ml). After incubation at 37°C, dephosphorylation of ZAP70 was measured by an anti-phosphotyrosine immunoblot. Phosphorylation of ZAP70 was found to decrease rapidly during the course of incubation (Fig. 5B). This decrease in anti-phosphotyrosine FIG. 4. Cellular distribution of SHP-1L in human and murine tissues. A, schematic map of primers used in PCR amplification for SHP-1 and SHP-1L. We used the same forward primer 5Ј-ctccaccaagggcctggac-3Ј and two different reverse primers: 1L (for SHP-1L), 5Јggtgcaggtcaggagacagcacaggg-3Ј; I (for both SHP-1 and SHP-1L), 5Ј-ctcctccctcttgttcttagtgtgc-3Ј. B, RT-PCR amplification of total RNAs from various human and murine cell lines. Total RNA was prepared from 2 ϫ 10 7 cultured cells using TRIzol reagent (Life Technologies, Inc.). The reverse transcription was done using the Superscript preamplification system following the manufacturer's instructions (Life Technologies, Inc.). The PCR condition was as follows: 30 s at 94°C, 1 min at 65°C, and 1 min at 72°C for 30 cycles. The DNA marker used is the 1-kilobase DNA ladder (Life Technologies, Inc.; catalog no 15615-016). C, PCR amplification of the first strand cDNA from various human tissues. The cDNA was purchased from CLONTECH. The PCR condition is the same as in B, using 3Ј primer 1L. D, SHP-1 immunoprecipitation and immunoblot analysis of proteins from various hematopoietic cells. mAb specific for the PTP domain of SHP-1 was used in both immunoprecipitation and immunoblotting. blots was not due to the lost of ZAP70 protein, as shown by anti-ZAP-70 immunostaining of the same blot (Fig. 6B). SHP-1 has been reported to have low activity in vitro at pH 7.4 and is 20 -30-fold more active at pH 5.4 (40). We sought to determine whether SHP-1L is also more active at pH 5.4 than at pH 7.4. The results were compared with that of SHP-1 (Fig. 6C). We were surprised to find that the kinetics of ZAP70 dephosphorylation by SHP-1L was similar at pH 5.4 and at pH 7.4, while the dephosphorylation of ZAP70 by SHP-1 was found to be much faster at pH 5.4 than at pH 7.4. In other words, SHP-1L and SHP-1 have a comparable catalytic activity at pH 5.4, but SHP-1L is more active than SHP-1 at pH 7.4. These data suggest that SHP-1L may be constitutively active in vivo, and therefore may play a different role from SHP-1 in the negative regulation of hematopoietic signal transduction pathways. DISCUSSION Here we report the identification, isolation, and cloning of a human SHP-1 isoform that we term SHP-1L. Analysis of SHP-1L cDNA and the genomic DNA of SHP-1 reveals that SHP-1L is generated by alternative splicing of SHP-1 in exon 16 within chromosome 12p12-13p locus (Fig. 3). The cloned SHP-1L cDNA predicts a protein of 614 amino acids, which is 29 amino acids longer than SHP-1, in agreement with the 70-kDa molecular mass of SHP-1L as determined by SDS-PAGE (Fig. 1). The N-terminal 548 amino acid residues of SHP-1L are identical to that of SHP-1, while the C-terminal 66 amino acid residues of SHP-1L are unique. Both SHP-1 and SHP-1L are cytosolic proteins (Fig. 1). SHP-1L is detected predominantly in human hematopoietic tissues and cell lines. Therefore, SHP-1L is a cytoplasmic protein-tyrosine phosphatase expressed predominantly in human hematopoietic tissues.
Based on its cDNA structure, SHP-1L contains two contiguous SH2 domains and a PTP domain identical to SHP-1. This predicts that SHP-1L possesses PTP activity similar to SHP-1. The C-terminal 66 amino acids of SHP-1L are distinct and contain a potential SH3 binding motif PVPGPPVLSP (Fig. 3). It is known that the truncation of 35-49 amino acids from the C terminus of SHP-1 increased the catalytic activity of SHP-1 by 20 -40-fold (39,40). This is the region where SHP-1 and SHP-1L diverge, suggesting that the catalytic activity of SHP-1L and SHP-1 may be regulated differentially. Indeed, we found that SHP-1L is more active than SHP-1 at physiological pH and that the catalytic activity of SHP-1L is unaffected by pH change, in contrast to SHP-1 activity (Fig. 6). Whereas the mechanism of such regulation is unknown, it is possible that the C-terminal region affects the folding between the SH2 domain and the catalytic domain. However, SHP-1 activity is unaffected by a larger truncation of 60 amino acids from the C terminus (40). In trying to reconcile this observation, it has been hypothesized that the C-terminal tail of SHP-1 may be further divided into a N-region activation motif and a C-region inhibitory motif. The C terminus of SHP-1L contains the same putative N-region motif of SHP-1 but a different C-region motif. This may explain why the catalytic activity of SHP-1L and SHP-1 are regulated differently.
The catalytic mechanism of PTPs involves the formation of a phosphoryl-cysteine intermediate. The active cysteine is also the target site of PTP inhibitors (45,46). Our results showed that SHP-1L was relatively enriched compared with SHP-1 after PAO affinity column purification, suggesting that SHP-1 has a higher affinity than SHP-1 toward the protein phosphatase inhibitor PAO. This is consistent with the observation that SHP-1L is more active than SHP-1 at pH 7.4, and implies that SHP-1L has a different conformation from that of SHP-1.
In Jurkat T cells, SHP-1 becomes tyrosine-phosphorylated upon pervanadate treatment but not by CD3 cross-linking (Fig.  5). SHP-1L lacks a potential tyrosyl phosphorylation site at Tyr-564 and is not tyrosine-phosphorylated by either CD3 cross-linking or pervanadate treatment. Previously, it has been demonstrated that Tyr-564 is the only site phosphorylated in vivo in a T cell hybridoma by co-cross-linking of CD3 with CD4 or CD8, and that the second potential tyrosine phosphorylation site Tyr-536 could only be tyrosine-phosphorylated in vitro by Lck (47). SHP-1L, containing Tyr-536, does not become tyrosylphosphorylated by pervanadate treatment, suggesting that Tyr-564 of SHP-1 is also the only site to be tyrosyl-phosphorylated in T cells by pervanadate treatment. Recently, we re- ported that the tyrosine phosphorylation of CD3-and several protein kinases Lck, Fyn, ZAP70, and Syk was induced by inhibition of PTPs (48). Here we provide an additional example that inhibition of PTPs triggers SHP-1 tyrosine phosphorylation.
Numerous proteins involved in the regulation of cellular signaling contain SH2 and SH3 domains, which serve as noncatalytic modules that regulate cellular processes by the formation of protein-protein interactions (6 -12). The difference between SHP-1L and SHP-1 at their C terminus implies that they may interact with different proteins. The phosphorylation of Tyr-564 could mediate the interaction between SHP-1 and other SH2 containing signaling molecules. The surrounding sequence of Tyr-564 is quite similar to that seen in other signaling molecules such as the ITAM motifs of CD3 and CD3⑀ in which acidic (Asp, Glu) and nonpolar (Leu, Val) amino acids are C-terminal to the phosphotyrosine. Recently a 30/32-kDa protein was shown to associate with the C terminus of SHP-1 (49). In SHP-2, tyrosinephosphorylated Tyr-542 and Tyr-580 in the C-terminal tail can bind to Grb2 and the SH2 domain of SHIP (50,51). In SHP-1L, this region contains a proline-rich, potential SH3 domain binding motif PVPGPPVLSP with a consensus sequence similar to the motif PLPPLPXP preferred by the SH3 domains of Src and Lyn (52). SHP-2 also contains a potential C-terminal SH3 binding motif PLPPCTPTPP. The motifs found in SHP-1L and SHP-2 are similar in that both stretch for 10 amino acids and contain five or six proline residues, one leucine residue, and seven or eight hydrophobic residues. It is possible that the SH3 binding motif contributes to SHP signaling and/or specificity.
Evidence from cDNA and genomic sequencing suggest that multiple isoforms of SH2-containing PTPs may be generated by alternative splicing of RNA transcripts. For example, a nonhematopoietic form of SHP-1 cDNA predicts a protein containing M-L-S-R-G at its N terminus (5). This differs from the N-terminal sequence M-V-R of hematopoietic SHP-1/SHP-1L. These two different forms were reported to be generated by alternative usage of promoters and exon skipping (53). Multiple isoforms of SHP-2 cDNA have also been reported (5). A splice variant of SHP-2, found by cDNA cloning, occurs at amino acids 549 -593 (5). The expression of multiple isoforms of CD45, a transmembrane leukocyte PTP, is highly regulated in cellular differentiation and activation, which implies that the pattern of isoforms expressed, is functionally important (54). Our results demonstrate the presence of a novel isoform of SHP-1 in human hematopoietic tissues but not in murine tissues. This raises the question whether murine tissues have an isoform like SHP-1 that may have similar biological function to human SHP-1.Our results from RT-PCR amplification, anti-SHP immunoprecipitation, and immunoblotting all suggested that SHP-1 could be alternatively spliced in murine hematopoietic tissues, too (Fig. 4). A 66-kDa protein was detected by anti-SHP-1 immunoprecipitation and immunoblotting in several human and murine cell lines (Fig. 4D). Our preliminary experiments indicated that this 66-kDa protein was not detected by immunoblotting using the antibody against the Cterminal 19 amino acids of SHP-1 (data not shown). Therefore, this 66-kDa protein could be a C-terminal alternatively spliced form of SHP-1 present in both human and murine tissues. For the first time, our report provides the direct evidence that SHP-1 is alternatively spliced in C terminus. Further investigation of the alternative splicing of SHP-1 may help us to understand how the catalytic activity of SHP-1 is regulated in hematopoietic tissues.