Identification of pp68 as the Tyrosine-phosphorylated Form of SYNCRIP/NSAP1 A CYTOPLASMIC RNA-BINDING PROTEIN*

Recently we reported that osmotic shock increased the insulin-stimulated tyrosine phosphorylation of a 68-kDa RNA-binding protein in 3T3-L1 adipocytes (Hresko, R. C., and Mueckler, M. (2000) J. Biol. Chem . 275, 18114– 18120). In this present study we have identified, by MALDI mass spectrometry, pp68 as the tyrosine-phos-phorylated form of synaptotagmin-binding cytoplasmic RNA-interacting protein (SYNCRIP)/NSAP1, a newly discovered cytoplasmic RNA-binding protein. Both SYNCRIP and pp68 were enriched in free polysomes found in low density microsomes isolated from 3T3-L1 adipocytes. In vitro phosphorylation studies revealed that SYNCRIP, once extracted from low density microsomes, can be tyrosine phosphorylated using purified insulin receptor. Binding of RNA to SYNCRIP specifically inhibited its in vitro phosphorylation but had no effect on receptor autophosphorylation or on the ability of the receptor to phosphorylate a model substrate, RCM-ly-sozyme. These results raise the possibility that regulation of mRNA translation or stability by insulin may involve the phosphorylation of SYNCRIP. Insulin binding to specific cell surface receptors initiates multiple signaling cascades that lead to a variety of cellular events, including stimulation of glucose and fatty acid uptake, ion and amino acid transport, glycogenesis, lipogenesis, gene transcription, mRNA turnover, protein synthesis and degradation, with time of flight detection mass spectrometry on a Perseptive Biosystems (Foster City, CA) Voyager DE-PRO mass spectrometry work station. Peptide masses were analyzed with Protein Prospector (University of Califor- nia, San Francisco) using MS-Fit. This analysis gave an unequivocal protein identification based on a clear distinction of the first MOWSE score relative to all other unrelated ranked proteins. The matched peptide masses represented 21% of the amino acid sequence of the identified protein. This identification was later corroborated by nano-capillary electospray mass spectrometry on a Finnigan (San Jose, CA) LCQ deca ion trap mass spectrometer. of Antibodies— Full-length SYNCRIP/NSAP1 cDNA by polymerase chain reaction amplification using Blot (cid:1)

Insulin binding to specific cell surface receptors initiates multiple signaling cascades that lead to a variety of cellular events, including stimulation of glucose and fatty acid uptake, ion and amino acid transport, glycogenesis, lipogenesis, gene transcription, mRNA turnover, protein synthesis and degradation, and DNA synthesis (1). The receptor/hormone interaction results in the autophosphorylation and subsequent activation of the receptor's intrinsic tyrosine kinase that can then phosphorylate several known cellular substrates, such as the insulin receptor substrate proteins (2), SHC isoforms, Gab-1, Cbl, p60 dok , and adaptor protein containing a PH and SH 2 domain (3). Once phosphorylated these proteins serve as docking sites for various Src homology 2 domain-containing proteins that are adaptor proteins themselves, kinases, or phosphatases. The result is the initiation of multiple phosphorylation/dephosphorylation signaling pathways that lead to the various pleiotropic responses.
Recently we found that osmotic shock enhanced by 10-fold the insulin-stimulated tyrosine phosphorylation of a 68-kDa protein in 3T3-L1 adipocytes (4). Further characterization revealed that pp68 was a peripheral protein that resides in a detergent-insoluble fraction of the low-density microsomes (LDM) 1 by binding to RNA. Phosphorylation by insulin was maximal by 1 min and was saturated with 50 -100 nM insulin. Activation of the p42/44 and p38 MAP kinase pathways by osmotic shock did not affect pp68 phosphorylation. Based on the failure to immunoprecipitate pp68 with antibodies directed against known 60 -70 tyrosine-phosphorylated proteins and on its phosphorylation characteristics, we speculated that pp68 may be a novel cellular target that lies downstream of the insulin receptor whose ability to bind RNA may be crucial in its physiological function.
In the present study we have identified pp68 as the tyrosinephosphorylated form of SYNCRIP (synaptotagmin-binding cytoplasmic RNA-interacting protein), a recently cloned cytoplasmic 66-kDa RNA-binding protein that has been independently shown to interact with ubiquitous synaptotagmins (5) and be part of a multiprotein complex involved in cytoplasmic RNA turnover (6). We report that SYNCRIP can be phosphorylated in vitro by the insulin receptor tyrosine kinase and that its phosphorylation is regulated by its ability to bind RNA.

EXPERIMENTAL PROCEDURES
Subcellular Fractionation of 3T3-L1 Adipocytes-3T3-L1 fibroblasts were grown to confluence and 48 h later subjected to differentiation as described previously (7). 3T3-L1 adipocytes were used 10 -14 days after differentiation. Cells were washed three times with phosphate-buffered saline (PBS) and incubated for at least 2 h to overnight in serum-free Dulbecco's modified Eagle's medium (DMEM). Adipocytes were then incubated in DMEM alone or DMEM supplemented with insulin or sorbitol. After the treatment, the cells were washed three times with ice-cold PBS, scraped in 2 ml per 10-cm dish of ice-cold HES (255 mM sucrose, 20 mM HEPES, pH 7.4, and 1 mM EDTA) containing 100 mM sodium fluoride, 10 mM sodium pyrophosphate, 1 mM sodium vanadate, and protease inhibitors, and then homogenized by passing the cells 10 times through a Yamato LSC homogenizer at a speed of 1200 rotations/ min at 4°C. Samples were centrifuged at 1000 ϫ g for 5 min to remove nuclei and unbroken cells. Subcellular fractionation was then carried out on the resulting supernatant essentially as described previously (8).
Purification of pp68 -3T3-L1 adipocytes were washed three times with PBS and then incubated for 16 h at 37°C with the same medium containing 1 mM Na 3 VO 4 . Cells were then treated with DMEM-supplemented with 600 mM sorbitol for 15 min at 37°C and then 100 nM insulin was added for an additional 5 min. LDM, prepared as described above, were incubated in ice-cold high salt buffer (50 mM HEPES, pH 7.4, 1 mM EDTA, 600 mM NaCl, 1 mM sodium vanadate, 100 mM sodium fluoride, 10 mM sodium pyrophosphate, and protease inhibitors) for 30 min at 4°C. Samples were then centrifuged for 1 h at 200,000 ϫ g. The salt concentration of the resulting supernatant was then adjusted to 150 mM NaCl. Tyrosine-phosphorylated proteins from 300 10-cm plates worth of high salt-extracted LDM were purified using an anti-phosphotyrosine antibody column as previously described (9). Proteins eluted with 3 mM phenyl phosphate were precipitated with 10% trichloroacetic acid (v/v, final concentration) and then purified by SDS-PAGE. pp68 was visualized using Coomassie Blue R-250 stain and then excised.
MALDI Mass Spectrometry-Excised pp68 gel pieces were subjected to trypsin digestion with Promega (Madison, WI) sequencing-grade trypsin using a standard procedure developed in the Protein and Nucleic Acid Chemistry Laboratory (PNACL) at Washington University (St. Louis, MO). The resulting mixture of tryptic peptides was analyzed by matrix-assisted laser desorption ionization (MALDI) with time of flight detection mass spectrometry on a Perseptive Biosystems (Foster City, CA) Voyager DE-PRO mass spectrometry work station. Peptide masses were analyzed with Protein Prospector (University of California, San Francisco) using MS-Fit. This analysis gave an unequivocal protein identification based on a clear distinction of the first MOWSE score relative to all other unrelated ranked proteins. The matched peptide masses represented 21% of the amino acid sequence of the identified protein. This identification was later corroborated by nanocapillary electospray mass spectrometry on a Finnigan (San Jose, CA) LCQ deca ion trap mass spectrometer.
Preparation of Antibodies-Full-length SYNCRIP/NSAP1 cDNA was obtained by polymerase chain reaction amplification using a mouse skeletal muscle Marathon Ready cDNA (CLONTECH) and then subcloned into Bluescript SKϩ. Antibodies were generated against the same two regions of SYNCRIP as previously described (5). Rabbits were immunized with either a glutathione S-transferase fusion protein corresponding to the NH 2 -terminal region of SYNCRIP (amino acids 1-170) or with a peptide (amino acids 140 -152) cross-linked to keyhole limpet hemocyanin via an artificial NH 2 -terminal cysteine residue. The peptide-specific antibody was used for immunoblots because this antibody recognized only the denatured form of the protein. The antibody directed against the NH 2 -terminal region was used in immunoprecipitation studies.
Western Blot Analysis-50 g of protein were subjected to SDS-PAGE and then transferred to nitrocellulose. Phosphotyrosine-phosphorylated proteins were detected using the monoclonal PY-20 antibody (Transduction Laboratories). SYNCRIP was detected using the peptidespecific antibody described above at a 1:500 dilution. 125 I-Labeled secondary antibodies (Amersham Biosciences) were visualized by autoradiography. Radioactive bands were quantified using a PhosphorImager SI analyzer (Amersham Biosciences).
Immunoprecipitation of SYNCRIP-0.5 mg of LDM were boiled for 5 min in 1% SDS. Samples were then diluted with 0.5 ml of buffer A (50 mM HEPES, pH 7.4, 1 mM EDTA, 150 mM NaCl, 1 mM sodium vanadate, 100 mM sodium fluoride, 10 mM sodium pyrophosphate, and protease inhibitors) containing Triton X-100 such that the final Triton X-100: SDS concentration was 10:1. After centrifugation for 10 min at 4°C in a microcentrifuge, the supernatant was incubated overnight with 15 l of SYNCRIP antibody. 50 l of protein A-Sepharose was then added for 2 h at 4°C to the reactions. After washing pellets four times with ice-cold buffer A containing 1% Triton X-100 and twice with buffer A without detergent, proteins were eluted with SDS-PAGE sample buffer. The immunoprecipitates were subjected to SDS-PAGE and analyzed by Western blot using PY-20 anti-phosphotyrosine antibodies as described above.
Isolation of Free Polysomes-Free polyribosomes were prepared using a modified procedure of Cardelli and Pitot (10). 5-10-cm dishes of 3T3-L1 adipocytes were washed three times with PBS and incubated for at least 2 h to overnight in serum-free DMEM. Cells were then incubated for 30 min in DMEM alone, or in DMEM supplemented with 100 nM insulin, 600 mM sorbitol, or a combination of insulin and sorbitol. Adipocytes were washed three times with ice-cold PBS and then three times with ice-cold 0.44 M STKM (sucrose molarity as designated, 50 mM Tris-HCl, 25 mM KCl, and 5 mM MgCl 2 , 1 mM sodium vanadate, pH 7.4) containing protease inhibitors. After removal of the buffer, cells were scraped and then disrupted using a ball homogenizer. A postmitochondrial supernatant (PMS) was prepared by centrifugation at 13,350 ϫ g av for 10 min at 4°C in a TLA-100.3 Beckman rotor. 2 M STKM was added to the PMS to a final sucrose concentration of 1.35 M. A discontinuous sucrose gradient was prepared in a 4-ml sealed polyallomer tube. 1.6 ml of PMS (1.35 M) was layered over 1.6 ml of 2 M STKM. 0.44 M STKM was then layered on top of the 1.35 M PMS. Samples were centrifuged at 266,000 ϫ g av for 4.5 h at 4°C in a NVT 90 Beckman rotor. Free polysomes were found as a clear pellet near the bottom of the tube.
In Vitro Phosphorylation of SYNCRIP Using Purified Insulin Receptor-Wheat germ agglutinin (WGA)-purified insulin receptor was prepared from 3T3-L1 adipocytes as previously described (11). In some experiments 0.5 g of cytoplasmic domain of the ␤-subunit of the insulin receptor (Calbiochem) were used in the in vitro reactions. SYN-CRIP was immunoprecipitated from 500 g of LDM using 15 l of the glutathione S-transferase fusion SYNCRIP antibody. 1 M Insulin was added or not for 20 min to an aliquot of wheat germ agglutinin-purified receptor (10 g of total protein in 0.1% Triton X-100). Receptors were autophosphorylated for 5 min at room temperature in 50 mM HEPES (pH 6.9), 100 mM NaCl, 0.1% Triton X-100, 5 mM manganese acetate, and 1 mM ATP. Detergent was omitted in the reactions that contained the cytoplasmic domain of the insulin receptor. The activated receptor was then added to the immunoprecipitated SYNCRIP, incubated for 15 min, and then quenched with the addition of EDTA (20 mM final) and SDS-PAGE sample buffer. Tyrosine-phosphorylated proteins were analyzed by Western blot analysis. In some reactions, RCM-lysozyme (Sigma) was used as a substrate.

RESULTS
Identification of pp68 -pp68 is a 68-kDa RNA-binding protein that resides in the LDM of 3T3-L1 adipocytes whose insulin-stimulated phosphorylation is dramatically increased with hypertonic stress (4). Based on the elimination of known tyrosine-phosphorylated proteins and on the time and concentration dependence of phosphorylation, we previously concluded that pp68 may be a novel substrate of the insulin receptor tyrosine kinase whose ability to bind RNA may dictate its physiological function. To identify this protein, pp68 was purified to homogeneity from 300 10-cm dishes of 3T3-L1 adipocytes using phosphotyrosine affinity chromatography and SDS-PAGE as described under "Experimental Procedures." Purified pp68 was excised from Coomassie Blue-stained polyacrylamide gels, trypsinized, and then analyzed by MALDI mass spectrometry. pp68 was unequivocally identified as being one of three highly homologous RNA-binding proteins, SYNCRIP, NSAP1, or heterogeneous nuclear ribonucleoprotein R (hnRNP R). The peptide sequences of SYNCRIP, which represent 21% of the protein, identified by MALDI based on matched peptide masses are shown in Table I. SYNCRIP is a recently cloned cytoplasmic 66-kDa mouse RNA-binding protein that was reported to interact with ubiquitous synaptotagmins (5). NSAP1 was identified from a two-hybrid screen as a cellular interactor of NS1, the major nonstructural parovirus protein (12). 99% amino acid identity between NSAP1 and SYNCRIP indicates that NSAP1 is the human homologue of SYNCRIP. hnRNP R, an 82-kDa protein localized in the nucleoplasm, was identified using autoimmune antibodies (13). SYNCRIP is 81.2% identical to hnRNP R but lacks ϳ70 carboxyl-terminal amino acids that contain a nuclear localization motif (5). Based on similarity in molecular mass and the cytoplasmic localization of pp68, we conclude that pp68 is the tyrosine-phosphorylated form of SYNCRIP/NSAP1 and not that of hnRNP R. Both SYNCRIP/ NSAP1 and hnRNP R contain three sets of RNA recognition motifs with well conserved RNP-1 and RNP-2 submotifs (5). In addition, both possess COOH termini enriched in arginine and glycine residues that have seven Arg-Gly-Gly (RGG) repeats. RGG boxes, another type of RNA-binding motif (14,15), are also involved in protein-protein interactions (15). Interspersed between several of the RGG repeats were regions enriched in tyrosine and acidic residues that contain three copies of the tetrapeptide YYGY. These tyrosine-rich regions contain several consensus sites for tyrosine phosphorylation in which acidic residues are found on the NH 2 -terminal side of the tyrosine (2,11).
To rule out the possibility that SYNCRIP was an abundant contaminating protein in the pp68 preparation, we immunoprecipitated SYNCRIP from LDM that were first boiled in SDS to disrupt intermolecular interactions, and then analyzed SYN-CRIP for phosphotyrosine content by Western blot analysis. 3T3-L1 adipocytes were treated with sorbitol and insulin as described in the legend to Fig. 1. A phosphotyrosine Western blot of isolated LDM (Fig. 1A) revealed a 68-kDa protein that was weakly phosphorylated on tyrosine in response to insulin alone, but whose phosphorylation was greatly enhanced when cells were treated with both insulin and sorbitol. The phosphotyrosine Western blot of immunoprecipitated SYNCRIP (Fig.  1B) gave exactly the same phosphotyrosine pattern observed for pp68 in total LDM. These results indicate that pp68 is the tyrosine-phosphorylated form of SYNCRIP. The same amount of SYNCRIP was immunoprecipitated under all conditions (Fig. 1C), indicating that the amount of SYNCRIP in the LDM fraction did not change with either sorbitol or insulin treatment.
Subcellular Localization of SYNCRIP and pp68 -Previously we demonstrated that pp68 resides in a detergent-insoluble fraction of LDM by binding to RNA (4). To determine the extraction properties of non-tyrosine-phosphorylated SYN-CRIP, LDM from basal cells were treated with 1% Triton X-100 or RNase A. The soluble and insoluble fractions were isolated after centrifugation for 1 h at 200,000 ϫ g. Western blots of the different fractions using SYNCRIP antibodies (Fig. 2) revealed that like pp68, nearly all of the non-phosphorylated SYNCRIP was found in the detergent-insoluble fraction of the LDM and almost all could be released from the LDM with RNase A treatment.
Next, we fractionated the adipocytes to determine which intracellular compartments were enriched in SYNCRIP. 3T3-L1 adipocytes were treated for 30 min with DMEM alone, or in DMEM supplemented with 100 nM insulin, 600 mM sorbitol, or a combination of insulin and sorbitol and then were fractionated by differential centrifugation as described under "Experimental Procedures." Western blot analyses of the different fractions were carried out using both SYNCRIP and phosphotyrosine antibodies. Because the different treatments had no effect on the subcellular distribution of SYNCRIP and pp68, only the fractionation results using sorbitol-and insulintreated cells are shown (Fig. 3). SYNCRIP was found highly enriched in LDM and to a lesser extent in high density microsomes (HDM) and the nucleus. The ϳ80-kDa protein recognized by the SYNCRIP antibodies in the nuclear fraction is most likely hnRNP R. The phosphotyrosine blot showed that pp68 was highly enriched in LDM and somewhat in HDM.
Based on its subcellular localization and on its affinity for RNA, we tested whether SYNCRIP was associated with polysomes. Polysomes are mRNA molecules associated with multiple ribosomes actively synthesizing nascent proteins. These RNA-protein complexes are membrane-bound in HDM (16) and free in solution in LDM. 2 Free polysomes were isolated from basal, insulin, sorbitol, or sorbitol and insulin-treated adipocytes using high speed centrifugation through a 2 M sucrose cushion as described under "Experimental Procedures" and then analyzed for SYNCRIP protein by Western blot analysis 2 R. C. Hresko and M. Mueckler, unpublished observation.  1. pp68 is the phosphorylated form of SYNCRIP/NSAP1. 3T3-L1 adipocytes were incubated in DMEM alone (B), DMEM supplemented with 100 nM insulin for 5 or 30 min (I(5) and I (30), respectively), in DMEM containing 600 mM sorbitol for 20 min (S(20)), in DMEM containing 600 mM sorbitol for 15 min followed by an additional 5 min with 100 nM insulin (S(20)I(5)), or in DMEM containing sorbitol and insulin for 30 min (SϩI (30)). LDM were then prepared from each of the treated cells. A, LDM were separated by SDS-PAGE (50 g of protein) and then immunoblotted using an anti-phosphotyrosine antibody. B, 500 g of LDM from each of the treated cells were boiled in 1% SDS, then diluted in 1% Triton X-100 and then immunoprecipitated with a SYNCRIP-specific antibody. Western blot analysis on the immunoprecipitates was carried out using an anti-phosphotyrosine antibody. C, immunoblot shown in panel B was reprobed with an antibody directed against SYNCRIP. (Fig. 4A). SYNCRIP was enriched about 7.5-fold in free polysomes isolated from basal or insulin-treated cells compared with that in total lysates. There was approximately half the amount of SYNCRIP found in free polysomes from sorbitol Ϯ insulin-treated cells. This is consistent with the fact that hypertonic stress is known to dissociate ribosomes from free polysomes (17,18). SYNCRIP/NSAP1 has been reported to be part of a multiprotein complex that includes both poly(A)-binding protein and poly(A)-binding protein-interacting protein (PAIP-1) (6). Like SYNCRIP, PAIP-1 was enriched in free polysomes (5.1-fold) isolated from basal or insulin-treated cells (Fig.  4A). Sorbitol treatment dramatically reduced the amount of PAIP-1 in the free polysomes. Our attempts to co-immunoprecipitate PAIP-1 and SYNCRIP in our system, however, were unsuccessful (data not shown). The ribosomal protein S6 was enriched 58-fold in free polysomes, illustrating the extent of purity of our free polysome preparation. The quantity of S6 protein in free polysomes did not change with sorbitol treatment, indicating that when free polysomes are dissociated by osmotic shock ribosomes can still be pelleted through a 2 M sucrose cushion, whereas proteins that associate with mRNA like SYNCRIP and PAIP-1 cannot. To determine whether pp68 was also found in free polysomes, SYNCRIP was immunoprecipitated from free polysomes isolated from each of the treated cells and then analyzed for phosphotyrosine content by Western blot analysis (Fig. 4B). Like SYNCRIP, pp68 was also associated with free polysomes. The relative amounts of SYN-CRIP immunoprecipitated in each of the reactions are shown for comparison (Fig. 4B, bottom panel).
Binding of RNA Inhibits the in Vitro Phosphorylation of SYNCRIP by the Insulin Receptor Tyrosine Kinase-Our initial study characterizing the insulin-stimulated tyrosine phosphorylation of pp68 suggested that SYNCRIP could be a direct substrate of the insulin receptor tyrosine kinase (4). To test this possibility, SYNCRIP was immunoprecipitated from LDM prepared from basal or sorbitol-treated cells and used as a substrate for the insulin receptor in an in vitro phosphorylation reaction (Fig. 5A). No detectable tyrosine phosphorylation of SYNCRIP was observed. Extraction of SYNCRIP from the LDM by RNase treatment or with 600 mM NaCl prior to its immunoprecipitation resulted in a robust tyrosine phosphorylation of SYNCRIP in the in vitro reaction. No difference in phosphorylation was observed between SYNCRIP purified from basal or sorbitol-treated cells under these conditions. SYNCRIP was not phosphorylated by the insulin receptor if SYNCRIP was immunoprecipitated from cytosol prepared from either basal or sorbitol-treated cells. The relative amounts of SYNCRIP immunoprecipitated from each of the LDM samples were identical, although the amount purified from the cytosol FIG. 2. SYNCRIP associates with a Triton X-100 (TX-100)-insoluble fraction of the LDM by binding to RNA. LDM were prepared from basal 3T3-L1 adipocytes. 75 g of LDM were left either untreated, were solubilized in 1% Triton X-100 for 30 min at 4°C, or were incubated in 0.4 mg/ml RNase A for 2 h on ice. Soluble (S) and insoluble (P) materials were isolated after centrifugation for 1 h at 200,000 ϫ g. The distributions of SYNCRIP for each of the treatments was determined by Western blot analysis using a SYNCRIP-specific antibody. The distribution of pp68 isolated from sorbitol and insulin-treated cells is shown for comparison. P-TYR, phosphotyrosine. and then homogenized. Free polysomes (POLY) were isolated from a PMS by centrifugation through a 2 M sucrose cushion as described under "Experimental Procedures." A, 20 g of whole cell lysate (WC), PMS, and free polysomes (POLY) were analyzed by Western blot analysis using SYNCRIP, PAIP-1, and S6 antibodies. B, 200 g of free polysomes isolated from each of the treated cells were boiled in SDS for 5 min and then diluted with 1% Triton X-100. Samples were then immunoprecipitated using a SYNCRIP-specific antibody and analyzed for phosphotyrosine content by Western blot analysis. The relative amounts of SYNCRIP protein immunoprecipitated in each of the reactions were determined by reprobing the phosphotyrosine (P-TYR) blot using a SYNCRIP-specific antibody. AB represents a mock immunoprecipitation reaction containing the SYNCRIP antibody without the polysome sample.
was significantly less than that purified from LDM (Fig. 5B).
Based on the result that extracting SYNCRIP from the LDM with RNase treatment increased its phosphorylation by the insulin receptor, we tested whether the addition of RNA to the in vitro phosphorylation reaction could specifically inhibit SYNCRIP phosphorylation. LDM was first treated with 600 mM NaCl to extract SYNCRIP from LDM. After the insoluble material was removed by centrifugation and the salt concentration diluted to 150 mM NaCl, SYNCRIP was immunoprecipitated and used as a substrate in the in vitro phosphorylation reaction. The addition of poly(A) RNA completely abrogated insulin-stimulated tyrosine phosphorylation of SYNCRIP (Fig.  6). In contrast, autophosphorylation of the insulin receptor and its ability to phosphorylate a model substrate, RCM-lysozyme, in an insulin-dependent manner was not significantly affected by poly(A) RNA. DISCUSSION In the present study we have identified pp68 as being the tyrosine-phosphorylated form of SYNCRIP/NSAP1 by MALDI mass spectrometry. SYNCRIP is a cytoplasmic RNA-binding protein that has been reported to interact with ubiquitous synaptotagmins (5). Although the significance of this association is not known, the authors (5) of this particular study proposed that non-neuronal synaptotagmins could be involved in organelle-based mRNA transport through its interaction with SYNCRIP. Our results, however, showed that SYNCRIP and pp68 were predominantly localized to LDM through their RNA binding (Fig. 2) and not through their interaction with an integral membrane protein such as synaptotagmin. Therefore, we decided not to focus our attention on the association between SYNCRIP and synaptotagmins in our system.
Further characterization of the localization of SYNCRIP revealed that both SYNCRIP and its phosphorylated form, pp68, were found enriched in free polysomes (Fig. 4). mRNAs are translated in circular form. Recent evidence indicates that the association between the cap structure (m 7 GpppX, where X represents any nucleotide) of the 5Ј terminus and the 3Ј poly(A) tail is required for efficient translation initiation (19,20). The cap, which directs the translation machinery to the 5Ј end, binds to the eIF4F complex. eIF4F is a three-component com-plex comprising the cap-binding protein eIF4E, the RNA helicase eIF4A, and eIF4G, a large scaffolding protein which interacts with ribosome-associated eIF3, eIF4E, and eIF4A. The actual linkage between the two termini in mammalian cells is thought to be through PAIP-1, a protein that simultaneously interacts with poly(A)-binding protein and eIF4A (21). Loss of the 3Ј poly(A) tail greatly reduces the rate of translation (20,22). In addition to promoting efficient translation initiation, the interaction between the two termini is important in maintaining the integrity of an mRNA (19). Proteins that bind to the poly(A) tail and the 5Ј cap are thought to protect mRNA from deadenylation and decapping (23), events which precede the 5Ј to 3Ј-exonucleolytic degradation of the mRNAs (24). Short halflive mRNAs contain adenylate, uridylate-rich (AU-rich) instability elements (25) in their 3Ј-untranslated regions, as well as regions of instability within the protein-coding region (23) that regulate the removal of the poly(A) tail. SYNCRIP/NSAP1 along with Unr, a purine-rich RNA-binding protein, poly(A)binding protein, PAIP-1, and hnRNP D, an AU-rich elementbinding protein have been reported to form a multiprotein complex that binds to the major protein-coding region determinant of instability of c-fos (6). It has been postulated that formation of this complex would bridge the poly(A) tail and the major protein-coding region determinant of instability and perhaps stabilize the mRNA and that transit of the ribosome might disrupt or reorganize the complex and lead to the deadenylation and subsequent degradation of the mRNA. Contrary to this published report (6), we were unable to coimmunoprecipitate SYNCRIP from an RNase A-treated lysate using an antibody directed against PAIP-1 even though we could demonstrate efficient PAIP-1 immunoprecipitation (data not shown). We did, however co-localize SYNCRIP and PAIP-1 to polysomes, suggesting that these proteins may associate with the same mRNA but in an indirect manner at least in our system.
In our initial study (4), we proposed that p68 (SYNCRIP) could be a direct substrate of the insulin receptor tyrosine FIG. 5. SYNCRIP can be directly phosphorylated by the insulin receptor once it has been released from the LDM. LDM and CYT were isolated from basal (B) or sorbitol-treated (S) adipocytes. SYNCRIP was extracted from 500 g of LDM by treating with 0.4 mg/ml RNase A for 2 h on ice or with 600 mM NaCl for 30 min. Samples were centrifuged for 1 h at 200,000 ϫ g to remove insoluble material. SYNCRIP was immunoprecipitated from 1 mg of cytosol, 500 g of intact LDM, or once it was removed from 500 g of LDM using RNase or high salt treatment as described above. Immunoisolated SYNCRIP was phosphorylated using the cytoplasmic domain of the ␤-subunit of the insulin receptor as described under "Experimental Procedures." A, reactions were analyzed using phosphotyrosine Western blot analysis. B, blot shown in panel A was reprobed using a SYNCRIP-specific antibody.
FIG. 6. Binding to RNA inhibits the tyrosine phosphorylation of SYNCRIP by the insulin receptor. SYNCRIP was extracted from LDM by incubating the LDM in 600 mM NaCl for 30 min. After removing the insoluble material by centrifugation, SYNCRIP was immunoisolated using a SYNCRIP-specific antibody and then incubated for 30 min in the presence (ϩRNA) or absence (CONTROL) of 3 mg/ml poly(A) RNA. Samples were phosphorylated with wheat germ agglutinin-purified insulin receptor as described under "Experimental Procedures." Reactions were analyzed using phosphotyrosine Western blot analysis (upper panel). Control phosphorylation reactions using the substrate RCM-lysozyme (10 M final concentration) were carried out under identical conditions (lower panel). kinase. In vitro phosphorylation studies revealed that SYN-CRIP, once extracted from the LDM, could be directly phosphorylated using purified insulin receptor, whereas SYNCRIP isolated from cytosol was poorly phosphorylated (Fig. 5). Binding of RNA specifically inhibited insulin-stimulated SYNCRIP phosphorylation in vitro but had no effect on receptor autophosphorylation or on the ability of the receptor to phosphorylate RCM-lysozyme (Fig. 6). Previously we reported that both osmotic shock and arsenite treatment greatly enhanced the tyrosine phosphorylation of SYNCRIP in 3T3-L1 adipocytes (4). It is also known that osmotic stress and arsenite treatment rapidly blocks protein synthesis at the initiation step leading to the formation of 80S monoribosomes that are free of mRNA (17,18). The dramatic reduction by sorbitol treatment in the amount of SYNCRIP and PAIP-1 but not ribosomal S6 protein that could be pelleted through a 2 M sucrose cushion (Fig. 4) confirmed this observation. Our working hypothesis based on the phosphorylation results is the following. SYNCRIP can be phosphorylated by the insulin receptor when it is not bound to RNA. Osmotic shock and arsenite treatment leads to the dissociation of free polysomes and increases the amount of RNAfree SYNCRIP in the cell. Our failure to observe a difference in tyrosine phosphorylation between SYNCRIP samples isolated from cytosol prepared from either basal or sorbitol-treated cells most likely indicates that during the homogenization/subcellular fractionation procedure using sucrose-containing buffers, free SYNCRIP can reassociate with RNA.
SYNCRIP is not the first reported hnRNP protein phosphorylated in response to extracellular signals. hnRNP A1 is predominantly localized in the nucleus but can shuttle between the nucleus and cytoplasm (26). Osmotic shock and UVC irradiation induces the serine phosphorylation of hnRNP A1 via the p38 MAP kinase pathway and causes a redistribution of hnRNP A1 to the cytoplasm (27). hnRNP K, a protein that can recruit RNA, DNA, Vav, transcriptional repressors, and inducible kinases (28) has been shown to be tyrosine-phosphorylated in both cell culture and in mouse liver treated with H 2 O 2 and vanadate (29). Protein K can be phosphorylated in vitro with protein kinase C (30), extracellular signal-regulated kinase (31), casein kinase II (32), and Src kinase (29). More recently, insulin has been shown to induce the tyrosine phosphorylation of hnRNP K in hepatocytes (33). Based on its phosphorylation characteristics and on its diverse set of molecular interactions, protein K might serve as a docking site to recruit signaling molecules involved in insulin-directed nucleic acid processes.
The significance of SYNCRIP phosphorylation is presently not known due in part to the lack of a clear functional role for SYNCRIP within the cell. hnRNPs are important in mRNA metabolism both inside the nucleus and in the cytoplasm (34). Based on subcellular fractionation, SYNCRIP was enriched in both LDM and nuclear fractions (Fig. 3) and therefore like hnRNP A1 may shuttle between the nucleus and cytoplasm. In the cytoplasm where pp68 is found, hnRNPs are known to regulate mRNA localization, mRNA translation, and mRNA turnover (34). Similar to other hnRNPs, SYNCRIP has been implicated in several different aspects of mRNA metabolism such as mRNA vesicular transport (5) and regulating c-fos mRNA turnover (6). The functional roles for SYNCRIP in these reports, however, are more speculative than conclusive. Both SYNCRIP and pp68 are predominantly localized to LDM by binding RNA (Fig. 2) and although SYNCRIP can bind preferentially to poly(A) resins in vitro, the actual sequence of RNA it binds in vivo is not known. In addition to possessing three sets of RNA recognition motifs, SYNCRIP has seven RGG boxes, another type of RNA-binding motif that is also involved in protein-protein interactions (15). Interspersed between the RGG boxes are several consensus-type sites for tyrosine phosphorylation. RNA-binding proteins not only bind RNA, they can interact with other RNA-binding proteins to form multiprotein complexes as has been previously reported for SYN-CRIP (6). Phosphorylation, therefore, could either affect the affinity of SYNCRIP to bind a specific sequence of RNA or modify its ability to associate with other binding proteins.
Insulin is known to regulate both mRNA translation and stability (35). Protein synthesis is globally induced severalfold with insulin, whereas some individual mRNAs are regulated more dramatically. Insulin can destabilize mRNAs encoding some proteins including: PEPCK, GLUT4, and glycogen synthase, whereas stabilize other messages such as for GLUT1, glycerol-3-phospate dehydrogenase, malic enzyme, and phosphorylase (35). Similar to the recent findings for hnRNP K, it is possible that the phosphorylation of SYNCRIP could be involved in the regulation of mRNA translation or mRNA turnover by insulin.