- Email: [email protected]
Characterization of Newly Identified Four Isoforms for a Putative Cytosolic Protein Tyrosine Phosphatase PTP36
Biochemical and Biophysical Research Communications 266, 523–531 (1999) Article ID bbrc.1999.1845, available online at http://www.idealibrary.com on
Characterization of Newly Identified Four Isoforms for a Putative Cytosolic Protein Tyrosine Phosphatase PTP36 Koji Aoyama, Tsukasa Matsuda, and Naohito Aoki 1 Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan
Received November 5, 1999
In the course of determining the expression profiles of protein tyrosine phosphatases in lactating mammary gland, we found the expression of an isoform for a putative cytosolic and cytoskeleton-associated protein tyrosine phosphatase PTP36. Further detailed RT-PCR and Northern blot analyses revealed the expression of several isoforms for PTP36 in a tissue-dependent manner. We have cloned the cDNAs encoding four truncated isoforms for PTP36 and designated PTP36-A, -B, -C, and -D, respectively. PTP36-A and -C had new sequences generated due to frameshift, whereas PTP36-B and -D were in-frame variants. Gly- and Glu-rich domains and a putative PTP domain were missing from PTP36-A, but the band 4.1 domain remained. PTP36-B retained the band 4.1 and PTP domains but lacked Pro-, Gly- and Glu-rich domains. Most domain structures were lacking in PTP36-C and -D. Interestingly, PTP36-C contained an incomplete band 4.1 domain, but the newly created sequence exhibited high homology to human nebulette, which was also suggested to associate with cytoskeletons. When transiently expressed in COS7 and HEK293 cells, not only the wild type but also all the isoforms were recovered in Triton X-100-insoluble cytoskeletonassociated fractions and this distribution was not affected by mechanical cell detachment and treatment with a kinase inhibitor staurosporine. Such cellular distribution of PTP36 was also observed in stable COS7 clones. Further studies using deletion mutants suggested that the first 30 amino acids as well as the band 4.1 domain of PTP36 were involved in association with Triton X-100 insoluble cytoskeletons. Tissue-dependent expression and deletion in domain structures might reThe nucleotide sequences reported in this paper have been submitted to the GenBank with Accession Numbers AF170902 (PTP36A), AF170903 (PTP36-B), AF170904 (PTP36-C), and AF170905 (PTP36-D). Abbreviations used: PTK, protein tyrosine kinase; PTP, protein tyrosine phosphatase; RT-PCR, reverse transcription-polymerase chain reaction; ECL, enhanced chemiluminescence; HA, hemagglutinin. 1 To whom correspondence should be addressed. Fax: 181-52-7894128. E-mail: [email protected].
flect the biological significance of the isoforms for PTP36 in certain physiological conditions. © 1999 Academic Press
Protein tyrosine phosphorylation has been shown to play critical roles in regulating fundamental cellular processes such as proliferation, differentiation, and development (1) and to be controlled by the balance of protein tyrosine kinase (PTK) and protein tyrosine phosphatase (PTP) (2). PTPs seem to be as diverse as PTKs and the family of PTP has been growing rapidly (3–5). About 80 PTPs have been identified, with at least 300 members estimated from genome sequencing (6). However, information concerning the function and regulation of most PTPs in cellular events are largely unknown. There have been reports describing isoforms for receptor-type PTPs, LAR/PTP delta/PTP sigma (7), CD45 (8), PTP alpha (9), PTP gamma (10), PTP-U2 (11) and cytosolic PTPs, Lyp (12), PTP1B (13), SHP-1 (14), SHP-2 (15), PTP STEP (16), PTP-BAS (17). Thus, alternatively spliced isoforms of PTPs may play important roles in certain cellular contexts. The isoforms for certain PTPs were shown to modulate their biological functions. For instance, 4 amino acid insertion into PTP catalytic domain resulted in reduction in PTP activity of SHP-2 (15). PTP STEP isoforms were predominantly expressed in brain and, therefore, suggested to function within the central nervous system (16). PTP-U2 long form, but not short form, was shown to enhance apoptosis of differentiated monoblastoid leukemia U937 cells (11). We also found that a short isoform for PTP-PEST was also expressed in mouse mammary gland and other tissues (18), though its biological significance remains unknown. We have paid much attention to identify PTPs which modulate certain biological signals triggered by lactogenic hormones (18) and involved in neural cell differentiation (19). In the course of the studies, an isoform in addition to wild type for PTP36/PTPD2/pez (20 –22), a putative cytosolic PTP was found in mouse mammary
523
0006-291X/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
gland as well as in kidney and lung. PTP36 possesses a band 4.1 domain which is found in many cytoskeletonassociated proteins and subfamilies of PTPs such as PTPH1 (23), PTPMEG (24), PTPD1/PTP-RL10/ rPTP2E (21, 25, 26), PTP-BAS/hPTP1E/PTPL1/FAP-1 (17, 27–29). Band 4.1 domains are responsible for targeting proteins to the cytoskeleton–plasma membrane interface (30). It has been, therefore, speculated that the band 4.1 domain-containing PTPs might regulate cell adhesion, formation of focal adhesions, and cytoskeletal organization. Actually, a recent report demonstrated that the overexpression of PTP36 in HeLa cells induced changes in cytoskeletons, focal adhesions, cell-substrate adhesions, and cell growth (31). In the present study, we report identification and characterization of four isoforms for PTP36. In transiently transfected COS7 and HEK293 cells, PTP36 wild type and all the isoforms were distributed in cytoskeleton-associated fractions. Such cellular distribution of PTP36 wild type was also observed in stable COS7 clones. Further studies using various deletion mutants of PTP36 revealed that not only band 4.1 domain but the first 30 amino acid residues of PTP36 was shown to be involved in association with cytoskeletons. MATERIALS AND METHODS Animals and reagents. Virgin female and male ddY mice (8week-old) were purchased from Japan SLC (Hamamatsu, Japan). Enzymes were obtained from TaKaRa (Kyoto, Japan). Anti-actin (AC-40), anti-vimentin (V9), and Staurosporine were from Sigma (St. Louis, MO). Anti-hemagglutinin (HA) antibody and anti-a-tubulin (C-20) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). RT-PCR, cDNA cloning and construction of expression plasmids. Total RNA was prepared from the mouse tissues with TRIZOL (GIBCO BRL, Gaithersburg, MD) according to the manufacturer’s instructions and stored at 280°C until use. First strand cDNA was synthesized using Advantage RT-for-PCR kit (Clontech) with random hexamer and then directly used for PCR amplification. Primers were designed based on the reported sequence (D31842) as follows: 59-TAT-CCA-CAT-ACCTTA-GTG-TG-39 (31–50), 59-ATT-GTG-CGA-TGC-CTA-TCA-AG-39 (1695–1714), 59-ATC-CCG-GAG-GAG-ATC-AAA-TG-39 (3697–3716). All the PCR products were cloned into a mammalian expression vector, pTargeT vector (Promega). The expression plasmids for PTP36 wild type and the four isoforms were HA-tagged at their N-terminal ends by PCR amplification with 59-CCA-CCA-TGT-ACC-CAT-ACG-ACG-TCCCAG-ACT-ACG-CTC-CTT-TCG-GCC-TGA-AGC-TCC-GCA-GG-39 and 59-ATC-CCG-GAG-GAG-ATC-AAA-TG-39. The expression plasmids for HA-tagged PTP36-C Dnebulette, PTP36 DPTP, and Band 4.1 were constructed by PCR amplification using antisense primers of 59-TCA-TGT-GGCCTC-CTG-CTG-AAG-CCG-39, 59-TCA-GTT-CTC-AGG-CAG-AGTGGC-GGT-39, and 59-CTA-AAT-CCC-CAT-GAA-GAA-GAT-GCC-39, respectively, with 59-CCA-CCA-TGT-ACC-CAT-ACG-ACG-TCCCAG-ACT-ACG-CTC-CTT-TCG-GCC-TGA-AGC-TCC-GCA-GG-39 as a sense primer. The expression plasmids for HA-tagged PTP36 D320 and Band 4.1 D30 were constructed by PCR amplification by following primer sets: 59-CCA-CCA-TGT-ACC-CAT-ACG-ACG-TCC-CAGACT-ACG-CTC-AGC-CCA-CCT-GGA-GCC-GGT-CC-39 and 59-ATCCCG-GAG-GAG-ATC-AAA-TG-39, and 59-CCA-CCA-TGT-ACC-CATACG-ACG-TCC-CAG-ACT-ACG-CTA-ATG-TCA-TCG-AGT-GCA-
CGC-TG-39 and 59-CTA-AAT-CCC-CAT-GAA-GAA-GAT-GCC-39. All the expression plasmids were confirmed by DNA sequencing. Northern blot analysis. Northern blot analysis was done as described (19). Briefly, 6 mg of poly(A) 1 RNA was separated on a 1.0% agarose/2.2 M formaldehyde gel and blotted onto a Hybond-N 1 membrane (Amersham Pharmacia Biotech). The membrane was hybridized with the 32P-labeled full-length PTP36 cDNA. The membrane was washed and analyzed by use of an image analyzer Model BAS 2500 (Fuji Film, Tokyo). Cell culture and transfection. COS7 cells and HEK293 cells were inoculated at a density of 2 3 10 5 cells/6 cm dish and grown overnight in DMEM containing 10% FCS. Expression plasmids were transfected into the cells by the modified calcium phosphate precipitation method (32). After incubation under 3% CO 2/97% air for 18 h, the transfected cells were washed with PBS twice and cultured in fresh DMEM containing 10% FCS for another 48 h under humidified 5% CO 2 and 95% air. Stable expression of PTP36 wild type in COS7 cells. COS7 cells were transfected as mentioned above. After 48 h, selection started in growth medium containing 500 mg/ml G418 (GIBCO BRL). Following 2 weeks selection, discrete colonies were subcloned and expanded. Cell lysis, cytoskeletal fractionation and Western blotting. The transfected cells on 6 cm dishes were lysed followed by cytoskeletal fractionation as described (33) with slight modifications. Cells were washed once with 2 ml of PBS and lysed in 200 ml of cytoskeleton stabilizing buffer (10 mM PIPES, pH 6.8, 250 mM sucrose, 3 mM MgCl 2, 120 mM KCl, 1 mM EGTA) supplemented with 0.15% Triton X-100, 1 mM PMSF, 10 mg/ml leupeptin, 1 mM Na 3VO 4, 25 mM sodium fluoride, and 5 mM sodium pyrophosphate for 10 min on ice. Following centrifugation at 14,000g for 10 min yielded the pellet, which contained the polymerized actin and the intermediate filaments including vimentin, and the supernatant, which contained the depolymerized actin and tubulin. Alternatively, cells were lysed in 200 ml of SDS-PAGE sample buffer (34). For the cell-detachment experiments, cells were washed once with 2 ml of PBS containing 5 mM EDTA, detached from cell culture dish by scraping into 1 ml of PBS containing 5 mM EDTA and incubated in suspension at 37°C for 30 min. After centrifugation, harvested cells were lysed in cytoskeleton stabilizing buffer as above. Treatment of cells with 50 nM staurosporine was done at 37°C for 30 min prior to cell lysis. Proteins in the cell lysates were separated by SDS-PAGE under reducing condition (34) followed by blotting onto nitrocellulose membranes (Hybond C1, Amersham Pharmacia Biotech). The membranes were blocked in NET buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.05% Triton X-100) containing 1% gelatin and the sequentially incubated with rabbit anti-HA antibody and peroxidaseconjugated goat anti rabbit IgG. The protein bands were visualized with an enhanced chemiluminescence (ECL) detection kit (Amersham Pharmacia Biotech).
RESULTS While surveying the expression profile of a variety of PTPs in mouse mammary gland, we found the amplification of an extra band with 1.1 kb in addition to an expected one (2.0 kb) for PTP36. Both bands were hybridized with a full length cDNA for wild type PTP36 (Fig. 1A). This extra band was also amplified from cDNAs of lung and kidney (Fig. 1A). Furthermore, Northern blot analysis revealed four distinct bands with 4.1, 1.9, 1.0 and 0.6 kb in kidney and the ones except for 0.6 kb in lung and mammary gland (Fig. 1B). These results strongly suggested the existence of the isoforms or highly related molecules for PTP36 and
524
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Identification of PTP36 isoforms. (A) Primer set spanning 1695-3716 nt of PTP36 [D31842, Sawada et al. (20)] was used for PCR amplification using each cDNA pool of kidney (lane 1), lung (lane 2), and mammary gland (lane 3). PCR products were separated on a 1.0% agarose gel followed by Southern blotting with a 32Pradiolabeled full-length PTP36 cDNA probe. (B) A Northern blot of 6 mg of poly(A) 1 RNA isolated from kidney (lane 1), lung (lane 2), and mammary gland (lane 3) was hybridized with a 32P-radiolabeled full-length PTP36 cDNA probe. (C) Primer set spanning 31-3716 nt of PTP36 was used for PCR amplification using each cDNA pool of brain (lane 1), thymus (lane 2), heart (lane 3), liver (lane 4), small intestine (lane 5), adrenal (lane 6), lung (lane 7), kidney (lane 8), mammary gland (lane 9), and testis (lane 10) and processed as above. Distinct seven positive bands are indicated. GAPDH amplification was done using primer set corresponding to 42-1059 of the cDNA sequence (Accession Number M32599).
prompted us to clone them. Among primer sets examined for RT-PCR amplification, the one spanning open reading frame of PTP36 generated totally 7 distinct bands (I through VII) which were hybridized with full length cDNA of PTP36 (Fig. 1C). The band I was detected in all the tissues examined and shown to be wild-type PTP36 by direct sequencing (data not shown). The bands II, III and IV were clearly detected in thymus, heart, adrenal, lung, kidney, and mammary gland, whereas the band V was detected only in kidney. More interestingly, the bands VI and VII, which seemed to be slightly different in migration on an agarose gel, were amplified from testis cDNA pool with the same primer set. The bands II and IV were cloned from cDNA pools of lung and mammary gland, the band V
cloned from that of kidney and the band VII cloned from that of testis, while others are currently being cloned. Sequencing analysis revealed that all the cloned cDNAs contained parts of PTP36 cDNA sequences and, therefore, the cDNA clones corresponding to the bands II, IV, V, and VII were designated PTP36-A, -C, -D, and -B, respectively. Alternative amino acid sequences at C-terminus were created for PTP36-A and -C, due to frameshift (Figs. 2A and 2B). On the other hand, PTP36-B and -D were shown to be in-frame trancated isoforms (Figs. 2A and 2B). PTP36 consists of several domain structures as shown in Fig. 1A. Gly- and Glurich domains and a putative PTP domain were missing from PTP36-A but remained the band 4.1 domain. PTP36-B retained the band 4.1 and PTP domains but lacked Pro-, Gly- and Glu-rich domains. Most domain structures were lacking in PTP36-C and -D. Interestingly, PTP36-C contained an incomplete band 4.1 domain, but the newly created sequence exhibited high homology to human cardiac muscle-specifc actinbinding protein nebulette (35). The wild type and all the isoforms for PTP36 were HA epitope-tagged at N-terminus and transiently expressed in COS7 cells. As shown in Fig. 2C, the bands with expected molecular weight (indicated in Fig. 2A) were detected with anti-HA antibody. To confirm the expression of all the isoforms cloned, RT-PCR amplification with targeted primer sets was carried out (Fig. 2D). Primer set I produced 2 bands corresponding to wild type and PTP36-A which was amplified from the thymus, heart, lung, kidney, mammary gland and testis cDNAs. A band with 320 bp corresponding to PTP36-B was amplified only from the testis cDNA by using primer set II, whereas a band corresponding to wild type was amplified from all the cDNAs. A single band corresponding to PTP36-C was amplified from the thymus, heart, lung, kidney, and mammary gland cDNAs by using primer set III. A band corresponding to PTP36-D was amplified only from the kidney cDNA with primer set IV. In the case of primer sets III and IV, extension time for PCR reaction was reduced and, therefore, larger bands were not amplified. These results clearly indicate tissuedependent expression of PTP36 isoforms. Recently, Ogata et al. reported that cellular distribution of PTP36 was regulated through phosphorylation and dephosphorylation on serine residues by cellsubstrate adhesion in 3T3 fibroblasts (36). To study that this is the case for the isoforms cloned, COS7 cells were transiently transfected with expression constructs for wild type and the four isoforms and lysed in the cytoskeleton stabilizing buffer in the absence or presence of phosphatase inhibitors (5 mM sodium pyrophosphate, 50 mM sodium fluoride). As clearly shown in Fig. 3A, all the isoforms as well as wild type PTP36 were distributed in Triton X-100 insoluble frac-
525
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Molecular structures and expression of PTP36 isoforms. (A) Schematic presentation of PTP36 isoforms. Numbers indicate amino acid residues of PTP36 [D31842, Sawada et al. (20)]. Band 4.1 domain, Pro-, Gly-, and Glu-rich domains, and a putative PTP domain are indicated. Newly created amino acid sequences for PTP36-A and -C are indicated by black bars. Nucleotide residue numbers lacked in isoforms and calculated molecular weight for wild type and isoforms are indicated in the parentheses on the left and on the right, respectively. (B) The newly created sequences for PTP36-A and -C are shown. The sequence of PTP36-C is aligned with that of human nebulette [NM_006393, Millevoi et al. (35)]. Identical amino acid residues are shaded and similar ones are indicated by 1 over the amino acid residues. Asterisks indicate stop codons. (C) The expression constructs for PTP36 wild type and isoforms were transiently transfected into COS7 cells and the cells were lysed with 200 ml of SDS-PAGE sample buffer. The equal cell equivalents were separated by SDS-PAGE and electrophoretically blotted onto nitrocellulose membranes. PTP36 wild type and isoforms were visualized with anti-HA antibody. (D) RT-PCR amplification of PTP36 wild type and four isoforms with four primer sets (I through IV) was done as indicated using each cDNA pool of brain (lane 1), thymus (lane 2), heart (lane 3), liver (lane 4), small intestine (lane 5), adrenal (lane 6), lung (lane 7), kidney (lane 8), mammary gland (lane 9), and testis (lane 10). Extension times of PCR amplification were 2 min and 1 min for primer sets I and II and primer sets III and IV, respectively. Numbers indicate nucleotide residues.
tions, irrespective of the presence or absence of phosphatase inhibitors. Upon detachment by scraping (lanes 5 and 6) and treatment with a kinase inhibitor staurosporine (lanes 7 and 8), which were shown to cause dephosphorylation of PTP36 and selective distribution to Triton X-100 insoluble pellet (36), cellular distribution of wild type and isoforms for PTP36 was unchanged (Fig. 3A). In all the cases, the migration of the protein bands on SDS-PAGE was apparently the same. The blot for PTP36-D were sequentially reprobed with anti-actin, anti-vimentin, and anti-a-
tubulin antibodies. As expected, actin was detected in both soluble and insoluble fractions, whereas vimentin and a-tubulin distributed in the pellet and the supernatant, respectively (Fig. 3A). Essentially the same results were obtained for other blots (data not shown). When the transfected cells were directly lysed in SDS sample buffer, the migration of the protein bands was also the same irrespective of attachment or detachment of the cells to cell culture dishes (Fig. 3A), suggesting that cellular distribution of PTP36 in transiently transfected COS7 cells is not regulated by cell-
526
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2—Continued
substrate adhesion. To confirm this observation, the same experiments were done using another cell line HEK293. As shown in Fig. 3B, all the protein bands were again recovered in Triton X-100 insoluble fractions. One might suspect that much higher expression of PTP36 wild and isoforms in transiently transfected COS7 and HEK293 caused such cellular distribution. To exclude this possibility, stable COS7 cells expressing PTP36 wild type were established. The expression level of PTP36 in two independent clones (clones 4 and 5) was much lower than that of a transient transfectant but the cellular distribution was the same (Fig. 3C). Taken together, PTP36 wild type and isoforms were resistant to solubilization by Triton X-100, suggesting that they constitutively associated with cytoskeletons. Band 4.1 domains have been thought to be primarily responsible for the association with cytoskeletons (30). If this is true for PTP36, the band 4.1 domain-deficient isoforms, PTP36-C and -D should be recovered in the supernatant after solubilization with Triton X-100containing cytoskeleton stabilizing buffer. However, both PTP36-C and -D were recovered in the insoluble
fractions (Fig. 3), although they have incomplete and no band 4.1 domains, respectively (Fig. 2). This suggests that a segment other than the band 4.1 domain has responsibility for the association with cytoskeletons. Therefore, to investigate the possibility that amino acid residues outside the band 4.1 domain bind cytoskeletons, four deletion mutants of PTP36 were constructed: PTP36 DPTP, lacking entire PTP domain; PTP36 D320, lacking the first 320 amino acids; Band 4.1, containing only the first 239 amino acids; Band 4.1 D30, lacking the first 30 amino acids of the Band 4.1 (Fig. 4A), and transiently transfected into COS7 cells. Like PTP36 wild type, PTP36 DPTP was recovered in the Triton X-100 insoluble fractions irrespective of attachment and detachment of the cells to cell culture dishes (Fig. 4B). On the contrary, PTP36 D320 was detected both in Triton X-100 soluble and insoluble fractions of the attached and detached cells in nearly the same proportion (Fig. 4B). Both Band 4.1 and Band 4.1 D30 were detected in the pellet fractions (Fig. 4B), which was indistinguishable from those of the PTP36 wild type and PTP36-D (Fig. 4B). These results strongly suggest that PTP36 containing either the first 30 amino acids
527
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
or the band 4.1 domain distributes in cytoskeletonassociated fractions. In the case of PTP36-C, a deletion mutant lacking the newly created sequence due to frame-shifting, which exhibited similarity to human nebulette, was made and transiently transfected into COS7 cells. The mutant protein (PTP36-C Dnebulette) again distributed in Triton X-100 insoluble fractions of the attached and detached cells (Fig. 4B). Another mutant lacking the first 74 amino acids of PTP36-C could not be expressed (data not shown). A deletion mutant of PTP36-D containing only the first 30 amino acids or the part of the putative PTP domain could not be expressed in transfected cells (data not shown). DISCUSSION Most of cellular events are regulated by the balance of protein phosphorylation and dephosphorylation. Accordingly, many attempts were taken to elucidate the expression profile of PTKs and PTPs in certain cellular events such as differentiation and development. First, we identified an extra transcript of PTP36 in mammary gland by chance. This observation was extended in various tissues by RT-PCR and Northern blot analyses (Fig. 1). Northern blot analysis revealed several transcripts, suggesting that newly identified isoforms were not resulted from PCR errors, though it still remains which isoform corresponds to the transcript on the Northern blot. These four isoforms are possibly generated by alternative splicing of a single premature transcript of PTP36, because nucleotide sequences of these isoforms were completely identical to each corresponding segment of PTP36 cDNA. However, this must await genomic cloning of PTP36. Existence of several isoforms for PTP36 may suggest their distinct biological functions. PTP36-B and -D were predominantly expressed in testis and kidney,
FIG. 3. Cellular distribution of PTP36 wild type and isoforms. (A) COS7 cells transiently transfected with expression plasmids for PTP36 wild-type and isoforms were lysed in Triton X-100-containing cytoskeleton stabilizing buffer in the absence (lanes 3 and 4) or
presence (lanes 1, 2, 5– 8) of 5 mM sodium pyrophosphate and 25 mM sodium fluoride. Cells were detached from cell culture dishes and incubated in suspension at 37°C for 30 min (lanes 5 and 6) or treated with 50 nM staurosporine at 37°C for 30 min. The total lysates were separated by centrifugation yielding the pellet (P) and the supernatant (S) as described under Materials and Methods. The blot for PTP36-D was sequentially reprobed with anti-actin, anti-vimentin, and anti-a-tubulin and essentially the same results were obtained for other blots (data not shown). Alternatively, the attached (lane 9) or detached (lane 10) cells were lysed in SDS-PAGE sample buffer. The equal cell equivalents from the fractions were analyzed by Western blotting with anti-HA antibody. Positions of the molecular weight markers are indicated on the right. (B) HEK293 cells were transiently transfected with the expression plasmids and processed as above. Cellular distribution of the expressed proteins in attached (lanes 1 and 2) and detached (lanes 3 and 4) cells were visualized as above. (C) Stable COS7 cells expressing PTP36 (clones 4 and 5), a mock transfectant and a transient transfectant of COS7 cells were lysed and processed as above. The blot was reprobed with anti-actin antibody.
528
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Association of the first 30 amino acids as well as the band 4.1-like domain of PTP36 with cytoskeletons. (A) Schematic presentation of the various deletion mutants of PTP36 and PTP36-D. See legend to Fig. 1 for detailed domain structures. (B) COS7 cells were transiently transfected with the expression plasmids for the indicated mutants and PTP36-D. The attached (lanes 1 and 2) and detached (lanes 3 and 4) cells were lysed in Triton X-100-containing cytoskeleton stabilizing buffer in the presence of 5 mM sodium pyrophosphate and 25 mM sodium fluoride. The total lysates were separated by centrifugation yielding the pellet (P) and the supernatant (S) as described under Materials and Methods. The equal cell equivalents from the fractions were analyzed by Western blotting with anti-HA antibody. Positions of the molecular weight markers are indicated on the right.
respectively, among tissues examined (Figs. 1C and 2D). This suggests that these two isoforms of PTP36 are alternatively spliced in a tissue dependent manner and function with distinct roles in these tissues. In this context, another clone corresponding to the band VII expressed only in testis would be of interest. Judging from the RT-PCR and Northern blotting data, some other isoforms for PTP36 seem to be expressed. Isoforms for PTP-BAS, which is a member of another band 4.1-containing PTP subfamily, are also reported (17). Another member of the PTP subfamily, PTP-D1 seems to have a shorter transcript according to the Northern blot data (21). Taken together, the band 4.1-containing
PTPs might constitute a diverse family and function with a complex role. When PTP36 wild type and isoforms were transiently expressed, all of them were shown to distribute in Triton X-100 insoluble fraction (Fig. 3). Furthermore, scraping and staurosporine treatment did not affect such distribution and the migration on SDSPAGE (Fig. 3), which was inconsistent with the results reported by Ogata et al., where they showed that distribution of stably expressed PTP36 wild type in 3T3 cells could be regulated by cell-substrate adhesion and that the dephosphorylated one migrated faster on SDSPAGE and distributed in Triton X-100 insoluble pellet
529
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(31). The same authors reported that stably expressed PTP36 wild type in HeLa cells was distributed in Triton X-100 insoluble pellet (36). In transiently expressed COS7 (Fig. 3A), HEK293 (Fig. 3B) and CHO cells (data not shown), such cellular distribution of PTP36 wild type and four isoforms were observed, suggesting cell-type specific cellular distribution of PTP36. Such cellular distribution of PTP36 might not be due to extraordinarily higher expression in transiently transfected cells, because the same distribution was observed in stable COS7 transfectants which expressed much less amount of PTP36 (Fig. 3C). Most interestingly, PTP36-C and -D, which have incomplete and no band 4.1 domains, respectively, were also recovered in Triton X-100 insoluble fractions (Fig. 3). Further studies using various deletion mutants suggested that the first 30 amino acids of PTP36 were involved in association with Triton X-100 insoluble cytoskeletons independently of the band 4.1 domain. However, we can not exclude the possibility that unknown mechanisms to make PTP36-C and -D distribute in cytoskeleton-associated fractions exist, because even the mutant lacking the first 320 amino acids of PTP36 could not be completely solubilized and distribute in the supernatant fraction (Fig. 4B). Blast search program through NCBI database revealed that the 30 amino acid sequence exhibited high similarity to mouse PTP-RL10 (25), human PTPD1 (26), and rat PTP 2E (27) with 70% identity, all which contain band 4.1 domains and suggested to be highly related to PTP36 among the members of band 4.1 superfamily. It remains examined whether PTP-RL10/PTPD1/PTP 2E also distribute in Triton X-100 insoluble fraction upon transient expression, although transient expression of PTPD1 in HEK293 cells were reported but it was unclear whether the PTP is solubilized in Triton X-100 or not, because the lysis buffer was not specified (21). PTP36 consists of various domain structures such as band 4.1 domain at N-terminus, Pro-, Gly-, Glu-rich domains at central part and a putative PTP domain at C-terminus (Fig. 1A). We also found that PTP36 wildtype as well as all the isoforms did not exhibit PTP activity against exogenous substrates (data not shown). They may function as an adapter molecule rather than as phosphatase, as suggested by Ogata et al. (31). Lack in one or more domain structures in PTP36 isoforms might make it possible to elucidate biological functions of each domain and each isoform might play a role in a dominant negative fashion against wild type. Based on this assumption, we are currently investigating how these isoforms are involved in cellular events such as cell proliferation, differentiation, adhesion and mobility. In conclusion, we cloned four isoforms for PTP36 and showed that they were expressed in a tissue dependent manner. Cellular distribution of the isoforms as well as wild type for PTP36 strongly suggests their important
roles in the regulation of cytoskeletons and cell adhesion. REFERENCES 1. Ullrich, A., and Schlessinger, J. (1990) Cell 61, 203–212. 2. Yarden, Y., and Ullrich, A. (1988) Annu. Rev. Biochem. 57, 443– 478. 3. Denu, J. M., Stuckey, J. A., Saper, M. A., and Dixon, J. E. (1996) Cell 87, 361–364. 4. Tonks, N. K. (1996) Adv. Pharmacol. 36, 91–119. 5. Tonks, N. K., and Neel, B. G. (1996) Cell 87, 365–368. 6. Pennisi, E. (1998) Science 282, 1972–1974. 7. Pulido, R., Serra-Pages, C., Tang, M., and Streuli, M. (1995) Proc. Natl. Acad. Sci. USA 92, 11686 –11690. 8. Trowbridge, I. S., and Thomas, M. L. (1994) Annu. Rev. Immunol. 12, 85–116. 9. Daum, G., Regenass, S., Sap, J., Schlessinger, J., and Fischer, E. H. (1994) J. Biol. Chem. 269, 10524 –10528. 10. Shintani, T., Maeda, N., Nishiwaki, T., and Noda, M. (1997) Biochem. Biophys. Res. Commun. 230, 419 – 425. 11. Seimiya, H., and Tsuruo, T. (1998) J. Biol. Chem. 273, 21187– 21193. 12. Cohen, S., Dadi, H., Shaoul, E., Sharfe, N., and Roifman, C. M. (1999) Blood 93, 2013–2024. 13. Shifrin, V. I., and Neel, B. G. (1993) J. Biol. Chem. 268, 25376 – 25384. 14. Martin, A., Tsui, H. W., Shulman, M. J., Isenman, D., and Tsui, F. W. (1999) J. Biol. Chem. 274, 21725–21734. 15. Mei, L., Doherty, C. A., and Huganir, R. L. (1994) J. Biol. Chem. 269, 12254 –12262. 16. Bult, A., Zhao, F., Dirkx, R., Jr., Raghunathan, A., Solimena, M., and Lombroso, P. J. (1997) Eur. J. Cell. Biol. 72, 337–344. 17. Maekawa, K., Imagawa, N., Nagamatsu, M., and Harada, S. (1994) FEBS Lett. 337, 200 –206. 18. Aoki, N., Kawamura, M., Yamaguchi-Aoki, Y., Ohira, S., and Matsuda, T. (1999) J. Biochem. (Tokyo) 125, 669 – 675. 19. Aoki, N., Yamaguchi-Aoki, Y., and Ullrich, A. (1996) J. Biol. Chem. 271, 29422–29426. 20. Sawada, M., Ogata, M., Fujino, Y., and Hamaoka, T. (1994) Biochem. Biophys. Res. Commun. 203, 479 – 484. 21. Moller, N. P., Moller, K. B., Lammers, R., Kharitonenkov, A., Sures, I., and Ullrich, A. (1994) Proc. Natl. Acad. Sci. USA 91, 7477–7481. 22. Smith, A. L., Mitchell, P. J., Shipley, J., Gusterson, B. A., Rogers, M. V., and Crompton, M. R. (1995) Biochem. Biophys. Res. Commun. 209, 959 –965. 23. Yang, Q., and Tonks, N. K. (1991) Proc. Natl. Acad. Sci. USA 88, 5949 –5953. 24. Gu, M. X., York, J. D., Warshawsky, I., and Majerus, P. W. (1991) Proc. Natl. Acad. Sci. USA 88, 5867–5871. 25. Higashitsuji, H., Arii, S., Furutani, M., Imamura, M., Kaneko, Y., Takenawa, J., Nakayama, H., and Fujita, J. (1995) Oncogene 10, 407– 414. 26. L’Abbe, D., Banville, D., Tong, Y., Stocco, R., Masson, S., Ma, S., Fantus, G., and Shen, S. H. (1994) FEBS Lett. 356, 351–356. 27. Banville, D., Ahmad, S., Stocco, R., and Shen, S. H. (1994) J. Biol. Chem. 269, 22320 –22327. 28. Saras, J., Claesson-Welsh, L., Heldin, C. H., and Gonez, L. J. (1994) J. Biol. Chem. 269, 24082–24089. 29. Sato, T., Irie, S., Kitada, S., and Reed, J. C. (1995) Science 268, 411– 415.
530
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
30. Arpin, M., Algrain, M., and Louvard, D. (1994) Curr. Opin. Cell Biol. 6, 136 –141. 31. Ogata, M., Takada, T., Mori, Y, Oh-hora, M., Uchida, Y., Kosugi, A., Miyake, K.,and Hamaoka, T. (1999) J. Biol. Chem. 274, 12905–12909. 32. Chen C., and Okayama H. (1987) Mol. Cell. Biol. 7, 2745–2752. 33. Arai, M., and Cohen, J. A. (1994) J. Neurosci. Res. 38, 348 –357.
34. Laemmli, U. K. (1970) Nature 227, 680 – 685. 35. Millevoi, S., Trombitas, K., Kolmerer, B., Kostin, S., Schaper, J., Pelin, K., Granzier, H., and Labeit, S. (1998) J. Mol. Biol. 282, 111–123. 36. Ogata, M., Takada, T., Mori, Y, Uchida, Y., Miki, T., Okuyama, A., Kosugi, A., Sawada, M., Oh-hora, M., and Hamaoka, T. (1999) J. Biol. Chem. 274, 20717–20724.
531
Characterization of Newly Identified Four Isoforms for a Putative Cytosolic Protein Tyrosine Phosphatase PTP36 Koji Aoyama, Tsukasa Matsuda, and Naohito Aoki 1 Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8601, Japan
Received November 5, 1999
In the course of determining the expression profiles of protein tyrosine phosphatases in lactating mammary gland, we found the expression of an isoform for a putative cytosolic and cytoskeleton-associated protein tyrosine phosphatase PTP36. Further detailed RT-PCR and Northern blot analyses revealed the expression of several isoforms for PTP36 in a tissue-dependent manner. We have cloned the cDNAs encoding four truncated isoforms for PTP36 and designated PTP36-A, -B, -C, and -D, respectively. PTP36-A and -C had new sequences generated due to frameshift, whereas PTP36-B and -D were in-frame variants. Gly- and Glu-rich domains and a putative PTP domain were missing from PTP36-A, but the band 4.1 domain remained. PTP36-B retained the band 4.1 and PTP domains but lacked Pro-, Gly- and Glu-rich domains. Most domain structures were lacking in PTP36-C and -D. Interestingly, PTP36-C contained an incomplete band 4.1 domain, but the newly created sequence exhibited high homology to human nebulette, which was also suggested to associate with cytoskeletons. When transiently expressed in COS7 and HEK293 cells, not only the wild type but also all the isoforms were recovered in Triton X-100-insoluble cytoskeletonassociated fractions and this distribution was not affected by mechanical cell detachment and treatment with a kinase inhibitor staurosporine. Such cellular distribution of PTP36 was also observed in stable COS7 clones. Further studies using deletion mutants suggested that the first 30 amino acids as well as the band 4.1 domain of PTP36 were involved in association with Triton X-100 insoluble cytoskeletons. Tissue-dependent expression and deletion in domain structures might reThe nucleotide sequences reported in this paper have been submitted to the GenBank with Accession Numbers AF170902 (PTP36A), AF170903 (PTP36-B), AF170904 (PTP36-C), and AF170905 (PTP36-D). Abbreviations used: PTK, protein tyrosine kinase; PTP, protein tyrosine phosphatase; RT-PCR, reverse transcription-polymerase chain reaction; ECL, enhanced chemiluminescence; HA, hemagglutinin. 1 To whom correspondence should be addressed. Fax: 181-52-7894128. E-mail: [email protected].
flect the biological significance of the isoforms for PTP36 in certain physiological conditions. © 1999 Academic Press
Protein tyrosine phosphorylation has been shown to play critical roles in regulating fundamental cellular processes such as proliferation, differentiation, and development (1) and to be controlled by the balance of protein tyrosine kinase (PTK) and protein tyrosine phosphatase (PTP) (2). PTPs seem to be as diverse as PTKs and the family of PTP has been growing rapidly (3–5). About 80 PTPs have been identified, with at least 300 members estimated from genome sequencing (6). However, information concerning the function and regulation of most PTPs in cellular events are largely unknown. There have been reports describing isoforms for receptor-type PTPs, LAR/PTP delta/PTP sigma (7), CD45 (8), PTP alpha (9), PTP gamma (10), PTP-U2 (11) and cytosolic PTPs, Lyp (12), PTP1B (13), SHP-1 (14), SHP-2 (15), PTP STEP (16), PTP-BAS (17). Thus, alternatively spliced isoforms of PTPs may play important roles in certain cellular contexts. The isoforms for certain PTPs were shown to modulate their biological functions. For instance, 4 amino acid insertion into PTP catalytic domain resulted in reduction in PTP activity of SHP-2 (15). PTP STEP isoforms were predominantly expressed in brain and, therefore, suggested to function within the central nervous system (16). PTP-U2 long form, but not short form, was shown to enhance apoptosis of differentiated monoblastoid leukemia U937 cells (11). We also found that a short isoform for PTP-PEST was also expressed in mouse mammary gland and other tissues (18), though its biological significance remains unknown. We have paid much attention to identify PTPs which modulate certain biological signals triggered by lactogenic hormones (18) and involved in neural cell differentiation (19). In the course of the studies, an isoform in addition to wild type for PTP36/PTPD2/pez (20 –22), a putative cytosolic PTP was found in mouse mammary
523
0006-291X/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
gland as well as in kidney and lung. PTP36 possesses a band 4.1 domain which is found in many cytoskeletonassociated proteins and subfamilies of PTPs such as PTPH1 (23), PTPMEG (24), PTPD1/PTP-RL10/ rPTP2E (21, 25, 26), PTP-BAS/hPTP1E/PTPL1/FAP-1 (17, 27–29). Band 4.1 domains are responsible for targeting proteins to the cytoskeleton–plasma membrane interface (30). It has been, therefore, speculated that the band 4.1 domain-containing PTPs might regulate cell adhesion, formation of focal adhesions, and cytoskeletal organization. Actually, a recent report demonstrated that the overexpression of PTP36 in HeLa cells induced changes in cytoskeletons, focal adhesions, cell-substrate adhesions, and cell growth (31). In the present study, we report identification and characterization of four isoforms for PTP36. In transiently transfected COS7 and HEK293 cells, PTP36 wild type and all the isoforms were distributed in cytoskeleton-associated fractions. Such cellular distribution of PTP36 wild type was also observed in stable COS7 clones. Further studies using various deletion mutants of PTP36 revealed that not only band 4.1 domain but the first 30 amino acid residues of PTP36 was shown to be involved in association with cytoskeletons. MATERIALS AND METHODS Animals and reagents. Virgin female and male ddY mice (8week-old) were purchased from Japan SLC (Hamamatsu, Japan). Enzymes were obtained from TaKaRa (Kyoto, Japan). Anti-actin (AC-40), anti-vimentin (V9), and Staurosporine were from Sigma (St. Louis, MO). Anti-hemagglutinin (HA) antibody and anti-a-tubulin (C-20) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). RT-PCR, cDNA cloning and construction of expression plasmids. Total RNA was prepared from the mouse tissues with TRIZOL (GIBCO BRL, Gaithersburg, MD) according to the manufacturer’s instructions and stored at 280°C until use. First strand cDNA was synthesized using Advantage RT-for-PCR kit (Clontech) with random hexamer and then directly used for PCR amplification. Primers were designed based on the reported sequence (D31842) as follows: 59-TAT-CCA-CAT-ACCTTA-GTG-TG-39 (31–50), 59-ATT-GTG-CGA-TGC-CTA-TCA-AG-39 (1695–1714), 59-ATC-CCG-GAG-GAG-ATC-AAA-TG-39 (3697–3716). All the PCR products were cloned into a mammalian expression vector, pTargeT vector (Promega). The expression plasmids for PTP36 wild type and the four isoforms were HA-tagged at their N-terminal ends by PCR amplification with 59-CCA-CCA-TGT-ACC-CAT-ACG-ACG-TCCCAG-ACT-ACG-CTC-CTT-TCG-GCC-TGA-AGC-TCC-GCA-GG-39 and 59-ATC-CCG-GAG-GAG-ATC-AAA-TG-39. The expression plasmids for HA-tagged PTP36-C Dnebulette, PTP36 DPTP, and Band 4.1 were constructed by PCR amplification using antisense primers of 59-TCA-TGT-GGCCTC-CTG-CTG-AAG-CCG-39, 59-TCA-GTT-CTC-AGG-CAG-AGTGGC-GGT-39, and 59-CTA-AAT-CCC-CAT-GAA-GAA-GAT-GCC-39, respectively, with 59-CCA-CCA-TGT-ACC-CAT-ACG-ACG-TCCCAG-ACT-ACG-CTC-CTT-TCG-GCC-TGA-AGC-TCC-GCA-GG-39 as a sense primer. The expression plasmids for HA-tagged PTP36 D320 and Band 4.1 D30 were constructed by PCR amplification by following primer sets: 59-CCA-CCA-TGT-ACC-CAT-ACG-ACG-TCC-CAGACT-ACG-CTC-AGC-CCA-CCT-GGA-GCC-GGT-CC-39 and 59-ATCCCG-GAG-GAG-ATC-AAA-TG-39, and 59-CCA-CCA-TGT-ACC-CATACG-ACG-TCC-CAG-ACT-ACG-CTA-ATG-TCA-TCG-AGT-GCA-
CGC-TG-39 and 59-CTA-AAT-CCC-CAT-GAA-GAA-GAT-GCC-39. All the expression plasmids were confirmed by DNA sequencing. Northern blot analysis. Northern blot analysis was done as described (19). Briefly, 6 mg of poly(A) 1 RNA was separated on a 1.0% agarose/2.2 M formaldehyde gel and blotted onto a Hybond-N 1 membrane (Amersham Pharmacia Biotech). The membrane was hybridized with the 32P-labeled full-length PTP36 cDNA. The membrane was washed and analyzed by use of an image analyzer Model BAS 2500 (Fuji Film, Tokyo). Cell culture and transfection. COS7 cells and HEK293 cells were inoculated at a density of 2 3 10 5 cells/6 cm dish and grown overnight in DMEM containing 10% FCS. Expression plasmids were transfected into the cells by the modified calcium phosphate precipitation method (32). After incubation under 3% CO 2/97% air for 18 h, the transfected cells were washed with PBS twice and cultured in fresh DMEM containing 10% FCS for another 48 h under humidified 5% CO 2 and 95% air. Stable expression of PTP36 wild type in COS7 cells. COS7 cells were transfected as mentioned above. After 48 h, selection started in growth medium containing 500 mg/ml G418 (GIBCO BRL). Following 2 weeks selection, discrete colonies were subcloned and expanded. Cell lysis, cytoskeletal fractionation and Western blotting. The transfected cells on 6 cm dishes were lysed followed by cytoskeletal fractionation as described (33) with slight modifications. Cells were washed once with 2 ml of PBS and lysed in 200 ml of cytoskeleton stabilizing buffer (10 mM PIPES, pH 6.8, 250 mM sucrose, 3 mM MgCl 2, 120 mM KCl, 1 mM EGTA) supplemented with 0.15% Triton X-100, 1 mM PMSF, 10 mg/ml leupeptin, 1 mM Na 3VO 4, 25 mM sodium fluoride, and 5 mM sodium pyrophosphate for 10 min on ice. Following centrifugation at 14,000g for 10 min yielded the pellet, which contained the polymerized actin and the intermediate filaments including vimentin, and the supernatant, which contained the depolymerized actin and tubulin. Alternatively, cells were lysed in 200 ml of SDS-PAGE sample buffer (34). For the cell-detachment experiments, cells were washed once with 2 ml of PBS containing 5 mM EDTA, detached from cell culture dish by scraping into 1 ml of PBS containing 5 mM EDTA and incubated in suspension at 37°C for 30 min. After centrifugation, harvested cells were lysed in cytoskeleton stabilizing buffer as above. Treatment of cells with 50 nM staurosporine was done at 37°C for 30 min prior to cell lysis. Proteins in the cell lysates were separated by SDS-PAGE under reducing condition (34) followed by blotting onto nitrocellulose membranes (Hybond C1, Amersham Pharmacia Biotech). The membranes were blocked in NET buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM EDTA, 0.05% Triton X-100) containing 1% gelatin and the sequentially incubated with rabbit anti-HA antibody and peroxidaseconjugated goat anti rabbit IgG. The protein bands were visualized with an enhanced chemiluminescence (ECL) detection kit (Amersham Pharmacia Biotech).
RESULTS While surveying the expression profile of a variety of PTPs in mouse mammary gland, we found the amplification of an extra band with 1.1 kb in addition to an expected one (2.0 kb) for PTP36. Both bands were hybridized with a full length cDNA for wild type PTP36 (Fig. 1A). This extra band was also amplified from cDNAs of lung and kidney (Fig. 1A). Furthermore, Northern blot analysis revealed four distinct bands with 4.1, 1.9, 1.0 and 0.6 kb in kidney and the ones except for 0.6 kb in lung and mammary gland (Fig. 1B). These results strongly suggested the existence of the isoforms or highly related molecules for PTP36 and
524
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Identification of PTP36 isoforms. (A) Primer set spanning 1695-3716 nt of PTP36 [D31842, Sawada et al. (20)] was used for PCR amplification using each cDNA pool of kidney (lane 1), lung (lane 2), and mammary gland (lane 3). PCR products were separated on a 1.0% agarose gel followed by Southern blotting with a 32Pradiolabeled full-length PTP36 cDNA probe. (B) A Northern blot of 6 mg of poly(A) 1 RNA isolated from kidney (lane 1), lung (lane 2), and mammary gland (lane 3) was hybridized with a 32P-radiolabeled full-length PTP36 cDNA probe. (C) Primer set spanning 31-3716 nt of PTP36 was used for PCR amplification using each cDNA pool of brain (lane 1), thymus (lane 2), heart (lane 3), liver (lane 4), small intestine (lane 5), adrenal (lane 6), lung (lane 7), kidney (lane 8), mammary gland (lane 9), and testis (lane 10) and processed as above. Distinct seven positive bands are indicated. GAPDH amplification was done using primer set corresponding to 42-1059 of the cDNA sequence (Accession Number M32599).
prompted us to clone them. Among primer sets examined for RT-PCR amplification, the one spanning open reading frame of PTP36 generated totally 7 distinct bands (I through VII) which were hybridized with full length cDNA of PTP36 (Fig. 1C). The band I was detected in all the tissues examined and shown to be wild-type PTP36 by direct sequencing (data not shown). The bands II, III and IV were clearly detected in thymus, heart, adrenal, lung, kidney, and mammary gland, whereas the band V was detected only in kidney. More interestingly, the bands VI and VII, which seemed to be slightly different in migration on an agarose gel, were amplified from testis cDNA pool with the same primer set. The bands II and IV were cloned from cDNA pools of lung and mammary gland, the band V
cloned from that of kidney and the band VII cloned from that of testis, while others are currently being cloned. Sequencing analysis revealed that all the cloned cDNAs contained parts of PTP36 cDNA sequences and, therefore, the cDNA clones corresponding to the bands II, IV, V, and VII were designated PTP36-A, -C, -D, and -B, respectively. Alternative amino acid sequences at C-terminus were created for PTP36-A and -C, due to frameshift (Figs. 2A and 2B). On the other hand, PTP36-B and -D were shown to be in-frame trancated isoforms (Figs. 2A and 2B). PTP36 consists of several domain structures as shown in Fig. 1A. Gly- and Glurich domains and a putative PTP domain were missing from PTP36-A but remained the band 4.1 domain. PTP36-B retained the band 4.1 and PTP domains but lacked Pro-, Gly- and Glu-rich domains. Most domain structures were lacking in PTP36-C and -D. Interestingly, PTP36-C contained an incomplete band 4.1 domain, but the newly created sequence exhibited high homology to human cardiac muscle-specifc actinbinding protein nebulette (35). The wild type and all the isoforms for PTP36 were HA epitope-tagged at N-terminus and transiently expressed in COS7 cells. As shown in Fig. 2C, the bands with expected molecular weight (indicated in Fig. 2A) were detected with anti-HA antibody. To confirm the expression of all the isoforms cloned, RT-PCR amplification with targeted primer sets was carried out (Fig. 2D). Primer set I produced 2 bands corresponding to wild type and PTP36-A which was amplified from the thymus, heart, lung, kidney, mammary gland and testis cDNAs. A band with 320 bp corresponding to PTP36-B was amplified only from the testis cDNA by using primer set II, whereas a band corresponding to wild type was amplified from all the cDNAs. A single band corresponding to PTP36-C was amplified from the thymus, heart, lung, kidney, and mammary gland cDNAs by using primer set III. A band corresponding to PTP36-D was amplified only from the kidney cDNA with primer set IV. In the case of primer sets III and IV, extension time for PCR reaction was reduced and, therefore, larger bands were not amplified. These results clearly indicate tissuedependent expression of PTP36 isoforms. Recently, Ogata et al. reported that cellular distribution of PTP36 was regulated through phosphorylation and dephosphorylation on serine residues by cellsubstrate adhesion in 3T3 fibroblasts (36). To study that this is the case for the isoforms cloned, COS7 cells were transiently transfected with expression constructs for wild type and the four isoforms and lysed in the cytoskeleton stabilizing buffer in the absence or presence of phosphatase inhibitors (5 mM sodium pyrophosphate, 50 mM sodium fluoride). As clearly shown in Fig. 3A, all the isoforms as well as wild type PTP36 were distributed in Triton X-100 insoluble frac-
525
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Molecular structures and expression of PTP36 isoforms. (A) Schematic presentation of PTP36 isoforms. Numbers indicate amino acid residues of PTP36 [D31842, Sawada et al. (20)]. Band 4.1 domain, Pro-, Gly-, and Glu-rich domains, and a putative PTP domain are indicated. Newly created amino acid sequences for PTP36-A and -C are indicated by black bars. Nucleotide residue numbers lacked in isoforms and calculated molecular weight for wild type and isoforms are indicated in the parentheses on the left and on the right, respectively. (B) The newly created sequences for PTP36-A and -C are shown. The sequence of PTP36-C is aligned with that of human nebulette [NM_006393, Millevoi et al. (35)]. Identical amino acid residues are shaded and similar ones are indicated by 1 over the amino acid residues. Asterisks indicate stop codons. (C) The expression constructs for PTP36 wild type and isoforms were transiently transfected into COS7 cells and the cells were lysed with 200 ml of SDS-PAGE sample buffer. The equal cell equivalents were separated by SDS-PAGE and electrophoretically blotted onto nitrocellulose membranes. PTP36 wild type and isoforms were visualized with anti-HA antibody. (D) RT-PCR amplification of PTP36 wild type and four isoforms with four primer sets (I through IV) was done as indicated using each cDNA pool of brain (lane 1), thymus (lane 2), heart (lane 3), liver (lane 4), small intestine (lane 5), adrenal (lane 6), lung (lane 7), kidney (lane 8), mammary gland (lane 9), and testis (lane 10). Extension times of PCR amplification were 2 min and 1 min for primer sets I and II and primer sets III and IV, respectively. Numbers indicate nucleotide residues.
tions, irrespective of the presence or absence of phosphatase inhibitors. Upon detachment by scraping (lanes 5 and 6) and treatment with a kinase inhibitor staurosporine (lanes 7 and 8), which were shown to cause dephosphorylation of PTP36 and selective distribution to Triton X-100 insoluble pellet (36), cellular distribution of wild type and isoforms for PTP36 was unchanged (Fig. 3A). In all the cases, the migration of the protein bands on SDS-PAGE was apparently the same. The blot for PTP36-D were sequentially reprobed with anti-actin, anti-vimentin, and anti-a-
tubulin antibodies. As expected, actin was detected in both soluble and insoluble fractions, whereas vimentin and a-tubulin distributed in the pellet and the supernatant, respectively (Fig. 3A). Essentially the same results were obtained for other blots (data not shown). When the transfected cells were directly lysed in SDS sample buffer, the migration of the protein bands was also the same irrespective of attachment or detachment of the cells to cell culture dishes (Fig. 3A), suggesting that cellular distribution of PTP36 in transiently transfected COS7 cells is not regulated by cell-
526
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2—Continued
substrate adhesion. To confirm this observation, the same experiments were done using another cell line HEK293. As shown in Fig. 3B, all the protein bands were again recovered in Triton X-100 insoluble fractions. One might suspect that much higher expression of PTP36 wild and isoforms in transiently transfected COS7 and HEK293 caused such cellular distribution. To exclude this possibility, stable COS7 cells expressing PTP36 wild type were established. The expression level of PTP36 in two independent clones (clones 4 and 5) was much lower than that of a transient transfectant but the cellular distribution was the same (Fig. 3C). Taken together, PTP36 wild type and isoforms were resistant to solubilization by Triton X-100, suggesting that they constitutively associated with cytoskeletons. Band 4.1 domains have been thought to be primarily responsible for the association with cytoskeletons (30). If this is true for PTP36, the band 4.1 domain-deficient isoforms, PTP36-C and -D should be recovered in the supernatant after solubilization with Triton X-100containing cytoskeleton stabilizing buffer. However, both PTP36-C and -D were recovered in the insoluble
fractions (Fig. 3), although they have incomplete and no band 4.1 domains, respectively (Fig. 2). This suggests that a segment other than the band 4.1 domain has responsibility for the association with cytoskeletons. Therefore, to investigate the possibility that amino acid residues outside the band 4.1 domain bind cytoskeletons, four deletion mutants of PTP36 were constructed: PTP36 DPTP, lacking entire PTP domain; PTP36 D320, lacking the first 320 amino acids; Band 4.1, containing only the first 239 amino acids; Band 4.1 D30, lacking the first 30 amino acids of the Band 4.1 (Fig. 4A), and transiently transfected into COS7 cells. Like PTP36 wild type, PTP36 DPTP was recovered in the Triton X-100 insoluble fractions irrespective of attachment and detachment of the cells to cell culture dishes (Fig. 4B). On the contrary, PTP36 D320 was detected both in Triton X-100 soluble and insoluble fractions of the attached and detached cells in nearly the same proportion (Fig. 4B). Both Band 4.1 and Band 4.1 D30 were detected in the pellet fractions (Fig. 4B), which was indistinguishable from those of the PTP36 wild type and PTP36-D (Fig. 4B). These results strongly suggest that PTP36 containing either the first 30 amino acids
527
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
or the band 4.1 domain distributes in cytoskeletonassociated fractions. In the case of PTP36-C, a deletion mutant lacking the newly created sequence due to frame-shifting, which exhibited similarity to human nebulette, was made and transiently transfected into COS7 cells. The mutant protein (PTP36-C Dnebulette) again distributed in Triton X-100 insoluble fractions of the attached and detached cells (Fig. 4B). Another mutant lacking the first 74 amino acids of PTP36-C could not be expressed (data not shown). A deletion mutant of PTP36-D containing only the first 30 amino acids or the part of the putative PTP domain could not be expressed in transfected cells (data not shown). DISCUSSION Most of cellular events are regulated by the balance of protein phosphorylation and dephosphorylation. Accordingly, many attempts were taken to elucidate the expression profile of PTKs and PTPs in certain cellular events such as differentiation and development. First, we identified an extra transcript of PTP36 in mammary gland by chance. This observation was extended in various tissues by RT-PCR and Northern blot analyses (Fig. 1). Northern blot analysis revealed several transcripts, suggesting that newly identified isoforms were not resulted from PCR errors, though it still remains which isoform corresponds to the transcript on the Northern blot. These four isoforms are possibly generated by alternative splicing of a single premature transcript of PTP36, because nucleotide sequences of these isoforms were completely identical to each corresponding segment of PTP36 cDNA. However, this must await genomic cloning of PTP36. Existence of several isoforms for PTP36 may suggest their distinct biological functions. PTP36-B and -D were predominantly expressed in testis and kidney,
FIG. 3. Cellular distribution of PTP36 wild type and isoforms. (A) COS7 cells transiently transfected with expression plasmids for PTP36 wild-type and isoforms were lysed in Triton X-100-containing cytoskeleton stabilizing buffer in the absence (lanes 3 and 4) or
presence (lanes 1, 2, 5– 8) of 5 mM sodium pyrophosphate and 25 mM sodium fluoride. Cells were detached from cell culture dishes and incubated in suspension at 37°C for 30 min (lanes 5 and 6) or treated with 50 nM staurosporine at 37°C for 30 min. The total lysates were separated by centrifugation yielding the pellet (P) and the supernatant (S) as described under Materials and Methods. The blot for PTP36-D was sequentially reprobed with anti-actin, anti-vimentin, and anti-a-tubulin and essentially the same results were obtained for other blots (data not shown). Alternatively, the attached (lane 9) or detached (lane 10) cells were lysed in SDS-PAGE sample buffer. The equal cell equivalents from the fractions were analyzed by Western blotting with anti-HA antibody. Positions of the molecular weight markers are indicated on the right. (B) HEK293 cells were transiently transfected with the expression plasmids and processed as above. Cellular distribution of the expressed proteins in attached (lanes 1 and 2) and detached (lanes 3 and 4) cells were visualized as above. (C) Stable COS7 cells expressing PTP36 (clones 4 and 5), a mock transfectant and a transient transfectant of COS7 cells were lysed and processed as above. The blot was reprobed with anti-actin antibody.
528
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 4. Association of the first 30 amino acids as well as the band 4.1-like domain of PTP36 with cytoskeletons. (A) Schematic presentation of the various deletion mutants of PTP36 and PTP36-D. See legend to Fig. 1 for detailed domain structures. (B) COS7 cells were transiently transfected with the expression plasmids for the indicated mutants and PTP36-D. The attached (lanes 1 and 2) and detached (lanes 3 and 4) cells were lysed in Triton X-100-containing cytoskeleton stabilizing buffer in the presence of 5 mM sodium pyrophosphate and 25 mM sodium fluoride. The total lysates were separated by centrifugation yielding the pellet (P) and the supernatant (S) as described under Materials and Methods. The equal cell equivalents from the fractions were analyzed by Western blotting with anti-HA antibody. Positions of the molecular weight markers are indicated on the right.
respectively, among tissues examined (Figs. 1C and 2D). This suggests that these two isoforms of PTP36 are alternatively spliced in a tissue dependent manner and function with distinct roles in these tissues. In this context, another clone corresponding to the band VII expressed only in testis would be of interest. Judging from the RT-PCR and Northern blotting data, some other isoforms for PTP36 seem to be expressed. Isoforms for PTP-BAS, which is a member of another band 4.1-containing PTP subfamily, are also reported (17). Another member of the PTP subfamily, PTP-D1 seems to have a shorter transcript according to the Northern blot data (21). Taken together, the band 4.1-containing
PTPs might constitute a diverse family and function with a complex role. When PTP36 wild type and isoforms were transiently expressed, all of them were shown to distribute in Triton X-100 insoluble fraction (Fig. 3). Furthermore, scraping and staurosporine treatment did not affect such distribution and the migration on SDSPAGE (Fig. 3), which was inconsistent with the results reported by Ogata et al., where they showed that distribution of stably expressed PTP36 wild type in 3T3 cells could be regulated by cell-substrate adhesion and that the dephosphorylated one migrated faster on SDSPAGE and distributed in Triton X-100 insoluble pellet
529
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
(31). The same authors reported that stably expressed PTP36 wild type in HeLa cells was distributed in Triton X-100 insoluble pellet (36). In transiently expressed COS7 (Fig. 3A), HEK293 (Fig. 3B) and CHO cells (data not shown), such cellular distribution of PTP36 wild type and four isoforms were observed, suggesting cell-type specific cellular distribution of PTP36. Such cellular distribution of PTP36 might not be due to extraordinarily higher expression in transiently transfected cells, because the same distribution was observed in stable COS7 transfectants which expressed much less amount of PTP36 (Fig. 3C). Most interestingly, PTP36-C and -D, which have incomplete and no band 4.1 domains, respectively, were also recovered in Triton X-100 insoluble fractions (Fig. 3). Further studies using various deletion mutants suggested that the first 30 amino acids of PTP36 were involved in association with Triton X-100 insoluble cytoskeletons independently of the band 4.1 domain. However, we can not exclude the possibility that unknown mechanisms to make PTP36-C and -D distribute in cytoskeleton-associated fractions exist, because even the mutant lacking the first 320 amino acids of PTP36 could not be completely solubilized and distribute in the supernatant fraction (Fig. 4B). Blast search program through NCBI database revealed that the 30 amino acid sequence exhibited high similarity to mouse PTP-RL10 (25), human PTPD1 (26), and rat PTP 2E (27) with 70% identity, all which contain band 4.1 domains and suggested to be highly related to PTP36 among the members of band 4.1 superfamily. It remains examined whether PTP-RL10/PTPD1/PTP 2E also distribute in Triton X-100 insoluble fraction upon transient expression, although transient expression of PTPD1 in HEK293 cells were reported but it was unclear whether the PTP is solubilized in Triton X-100 or not, because the lysis buffer was not specified (21). PTP36 consists of various domain structures such as band 4.1 domain at N-terminus, Pro-, Gly-, Glu-rich domains at central part and a putative PTP domain at C-terminus (Fig. 1A). We also found that PTP36 wildtype as well as all the isoforms did not exhibit PTP activity against exogenous substrates (data not shown). They may function as an adapter molecule rather than as phosphatase, as suggested by Ogata et al. (31). Lack in one or more domain structures in PTP36 isoforms might make it possible to elucidate biological functions of each domain and each isoform might play a role in a dominant negative fashion against wild type. Based on this assumption, we are currently investigating how these isoforms are involved in cellular events such as cell proliferation, differentiation, adhesion and mobility. In conclusion, we cloned four isoforms for PTP36 and showed that they were expressed in a tissue dependent manner. Cellular distribution of the isoforms as well as wild type for PTP36 strongly suggests their important
roles in the regulation of cytoskeletons and cell adhesion. REFERENCES 1. Ullrich, A., and Schlessinger, J. (1990) Cell 61, 203–212. 2. Yarden, Y., and Ullrich, A. (1988) Annu. Rev. Biochem. 57, 443– 478. 3. Denu, J. M., Stuckey, J. A., Saper, M. A., and Dixon, J. E. (1996) Cell 87, 361–364. 4. Tonks, N. K. (1996) Adv. Pharmacol. 36, 91–119. 5. Tonks, N. K., and Neel, B. G. (1996) Cell 87, 365–368. 6. Pennisi, E. (1998) Science 282, 1972–1974. 7. Pulido, R., Serra-Pages, C., Tang, M., and Streuli, M. (1995) Proc. Natl. Acad. Sci. USA 92, 11686 –11690. 8. Trowbridge, I. S., and Thomas, M. L. (1994) Annu. Rev. Immunol. 12, 85–116. 9. Daum, G., Regenass, S., Sap, J., Schlessinger, J., and Fischer, E. H. (1994) J. Biol. Chem. 269, 10524 –10528. 10. Shintani, T., Maeda, N., Nishiwaki, T., and Noda, M. (1997) Biochem. Biophys. Res. Commun. 230, 419 – 425. 11. Seimiya, H., and Tsuruo, T. (1998) J. Biol. Chem. 273, 21187– 21193. 12. Cohen, S., Dadi, H., Shaoul, E., Sharfe, N., and Roifman, C. M. (1999) Blood 93, 2013–2024. 13. Shifrin, V. I., and Neel, B. G. (1993) J. Biol. Chem. 268, 25376 – 25384. 14. Martin, A., Tsui, H. W., Shulman, M. J., Isenman, D., and Tsui, F. W. (1999) J. Biol. Chem. 274, 21725–21734. 15. Mei, L., Doherty, C. A., and Huganir, R. L. (1994) J. Biol. Chem. 269, 12254 –12262. 16. Bult, A., Zhao, F., Dirkx, R., Jr., Raghunathan, A., Solimena, M., and Lombroso, P. J. (1997) Eur. J. Cell. Biol. 72, 337–344. 17. Maekawa, K., Imagawa, N., Nagamatsu, M., and Harada, S. (1994) FEBS Lett. 337, 200 –206. 18. Aoki, N., Kawamura, M., Yamaguchi-Aoki, Y., Ohira, S., and Matsuda, T. (1999) J. Biochem. (Tokyo) 125, 669 – 675. 19. Aoki, N., Yamaguchi-Aoki, Y., and Ullrich, A. (1996) J. Biol. Chem. 271, 29422–29426. 20. Sawada, M., Ogata, M., Fujino, Y., and Hamaoka, T. (1994) Biochem. Biophys. Res. Commun. 203, 479 – 484. 21. Moller, N. P., Moller, K. B., Lammers, R., Kharitonenkov, A., Sures, I., and Ullrich, A. (1994) Proc. Natl. Acad. Sci. USA 91, 7477–7481. 22. Smith, A. L., Mitchell, P. J., Shipley, J., Gusterson, B. A., Rogers, M. V., and Crompton, M. R. (1995) Biochem. Biophys. Res. Commun. 209, 959 –965. 23. Yang, Q., and Tonks, N. K. (1991) Proc. Natl. Acad. Sci. USA 88, 5949 –5953. 24. Gu, M. X., York, J. D., Warshawsky, I., and Majerus, P. W. (1991) Proc. Natl. Acad. Sci. USA 88, 5867–5871. 25. Higashitsuji, H., Arii, S., Furutani, M., Imamura, M., Kaneko, Y., Takenawa, J., Nakayama, H., and Fujita, J. (1995) Oncogene 10, 407– 414. 26. L’Abbe, D., Banville, D., Tong, Y., Stocco, R., Masson, S., Ma, S., Fantus, G., and Shen, S. H. (1994) FEBS Lett. 356, 351–356. 27. Banville, D., Ahmad, S., Stocco, R., and Shen, S. H. (1994) J. Biol. Chem. 269, 22320 –22327. 28. Saras, J., Claesson-Welsh, L., Heldin, C. H., and Gonez, L. J. (1994) J. Biol. Chem. 269, 24082–24089. 29. Sato, T., Irie, S., Kitada, S., and Reed, J. C. (1995) Science 268, 411– 415.
530
Vol. 266, No. 2, 1999
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
30. Arpin, M., Algrain, M., and Louvard, D. (1994) Curr. Opin. Cell Biol. 6, 136 –141. 31. Ogata, M., Takada, T., Mori, Y, Oh-hora, M., Uchida, Y., Kosugi, A., Miyake, K.,and Hamaoka, T. (1999) J. Biol. Chem. 274, 12905–12909. 32. Chen C., and Okayama H. (1987) Mol. Cell. Biol. 7, 2745–2752. 33. Arai, M., and Cohen, J. A. (1994) J. Neurosci. Res. 38, 348 –357.
34. Laemmli, U. K. (1970) Nature 227, 680 – 685. 35. Millevoi, S., Trombitas, K., Kolmerer, B., Kostin, S., Schaper, J., Pelin, K., Granzier, H., and Labeit, S. (1998) J. Mol. Biol. 282, 111–123. 36. Ogata, M., Takada, T., Mori, Y, Uchida, Y., Miki, T., Okuyama, A., Kosugi, A., Sawada, M., Oh-hora, M., and Hamaoka, T. (1999) J. Biol. Chem. 274, 20717–20724.
531