A Key Role for the Phosphorylation of Ser440 by the Cyclic AMP-dependent Protein Kinase in Regulating the Activity of the Src Homology 2 Domain-containing Inositol 5′-Phosphatase (SHIP1)*

The Src homology 2 domain-containing inositol 5′-phosphatase 1 (SHIP1) dephosphorylates phosphatidylinositol 3,4,5-trisphosphate to phophatidylinositol 3,4-bisphosphate in hematopoietic cells to regulate multiple cell signaling pathways. SHIP1 can be phosphorylated by the cyclic AMP-dependent protein kinase (PKA), resulting in an increase in SHIP1 activity (Zhang, J., Walk, S. F., Ravichandran, K. S., and Garrison, J. C. (2009) J. Biol. Chem. 284, 20070–20078). Using a combination of approaches, we identified the serine residue regulating SHIP1 activity. After mass spectrometric identification of 17 serine and threonine residues on SHIP1 as being phosphorylated by PKA in vitro, studies with truncation mutants of SHIP1 narrowed the phosphorylation site to the catalytic region between residues 400 and 866. Of the two candidate phosphorylation sites located in this region (Ser440 and Ser774), only mutation of Ser440 to Ala abolished the ability of PKA to phosphorylate the purified, catalytic domain of SHIP1 (residues 401–866). Mutation of Ser440 to Ala in full-length SHIP1 abrogated the ability of PKA to increase the activity of SHIP1 in mammalian cells. Using flow cytometry, we found that the PKA activator, Sp-adenosine 3′,5′-cyclic monophosphorothioate triethylammonium salt hydrate (Sp-cAMPS) blunted the phosphorylation of Akt downstream of B cell antigen receptor engagement in SHIP1-null DT40 B lymphocytes expressing native mouse SHIP1. The inhibitory effect of Sp-cAMPS was absent in cells expressing the S440A mutant of SHIP1. These results suggest that activation of SHIP1 by PKA via phosphorylation on Ser440 is an important regulatory event in hematopoietic cells.

There are two major members of the SHIP family, SHIP1 and SHIP2, that share a highly homologous catalytic region with less similarity in their C-terminal regions (6). SHIP2 is widely expressed, whereas SHIP1 is only expressed in hematopoietic cells (7). SHIP1 was originally identified based on its ability to bind to cytoplasmic adaptor proteins, such as Shc, Dab-1, and Dok-3 (8). SHIP1 contains an N-terminal SH2 domain that leads to interactions with immune receptors, such as Fc␥RIIB and Fc⑀RI (9), a central inositol 5Ј-phosphatase domain, and two tyrosines within NPXY motifs in the C-terminal prolinerich region (10). Genetic manipulations of SHIP1 in mice have shown that SHIP1 plays important roles in functions of myeloid cells and B lymphocytes; for example, in mature neutrophils and mast cells, PtdIns-3,4,5-P 3 levels and the phosphorylation of Akt activity are significantly elevated in the absence of SHIP1 (11). Similarly, B lymphocytes lacking SHIP1 are insensitive to inhibitory signaling via the Fc␥RIIB receptor (9). Moreover, SHIP1 Ϫ/Ϫ mice have a decreased viability due to an increased infiltration of myeloid cells into the lungs, perhaps due to activation of migration pathways because of the elevated levels of PtdIns-3,4,5-P 3 (12).
Due to its importance in hematopoietic cells, the activity of SHIP1 is tightly regulated. One model for regulation of SHIP1 function envisions translocation of SHIP1 from the cytosol to the membrane (13). Upon stimulation by growth factors, cytokine receptors, or specific signaling receptors on lymphocytes, SHIP1 is recruited via its N-terminal SH2 domain to phosphorylated tyrosine residues in receptor kinases and degrades the elevated levels of PtdIns-3,4,5-P 3 near the activated receptor (14). It was also reported that the catalytic activity of SHIP1 can be allosterically enhanced by the binding of its lipid product phophatidylinositol 3,4-bisphosphate or a small molecule to its C2 domain (15,16). SHIP1 has also been found to undergo phosphorylation on two tyrosine residues by tyrosine kinases (14,17). However, it is not clear that this phosphorylation event can regulate SHIP1 activity (14). We have demonstrated that SHIP1 can be phosphorylated and activated by the cyclic AMPdependent protein kinase (PKA) both in vitro and in cells, highlighting a new mode of SHIP1 regulation (18).
Although the activation of SHIP1 by PKA is a potentially important regulatory event, the site(s) of phosphorylation within the SHIP1 molecule have not been identified. In the present study, through mass spectrometry and site-specific mutations, we identified and characterized the potential PKAmediated phosphorylation sites within SHIP1. We also created SHIP1 Ϫ/Ϫ DT40 cell lines stably expressing native and mutated SHIP1 molecules to evaluate the importance of selected phosphorylation sites in the regulation of SHIP1 by PKA. These results demonstrate that Ser 440 within the catalytic region of the molecule is the site phosphorylated by PKA to activate SHIP1 and that the phosphorylation event is functionally relevant in B lymphocytes.

EXPERIMENTAL PROCEDURES
Materials-Anti-SHIP antibodies (catalogue nos. SC-8425 (P1C1) to the C terminus and SC-14503 (D20) to the phosphatase domain) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The anti-FLAG M2 monoclonal antibody and anti-FLAG M2 affinity gel were purchased from Sigma. The FLAG peptide was synthesized by Bio Basic Inc. (Markham, Canada). The unlabeled phospho-Akt (Thr 308 ), the Alexa Fluor 488-conjugated phospho-Akt (Thr 308 ), the total Akt, and the Alexa Fluor 647-conjugated anti-FLAG antibodies were purchased from Cell Signaling Technology (Danvers, MA). The goat F(abЈ) 2 anti-mouse IgM and mouse anti-chicken IgM (M4) antibodies were obtained from Southern Biotech (Birmingham, AL). Phosflow Fix Buffer I and Perm Buffer III were purchased from BD Biosciences. The sources of all other reagents have been published (18).
Truncated SHIP1 Construction and Site-directed Mutagenesis-The cDNA encoding mouse SHIP1 was subcloned as described previously into the pCMVTag2C vector (Stratagene), which adds a FLAG epitope to the N terminus of the expressed protein (18). To allow expression of and one-step purification of truncated regions of the SHIP1 protein, cDNA constructs encoding truncated regions of SHIP1 were generated by PCR and subcloned into the pCMVTag2C vector. The resulting truncated proteins contained amino acids 401-866, 401-1190, or 1-866 of SHIP1. Full-length SHIP1 cDNAs encoding single or double point mutations of putative phosphorylation sites (see "Results") were generated with the QuikChange II XL sitedirected mutagenesis kit (Stratagene) in the pCMVTag2C vector. All cDNA constructs used to express the full-length, truncated, or mutant SHIP1 proteins were sequenced to ensure fidelity.
Cell Culture and Transfection-HEK-293 cells were cultured in DMEM supplemented with 10% fetal bovine serum and transfected with the plasmids indicated under "Results" using Lipofectamine 2000 (Invitrogen) at a 1:1 ratio (DNA/Lipofectamine, w/v) as described (18). DT40 cell suspensions (9) were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum, penicillin, streptomycin, 2 mM L-glutamine, 1% (v/v) chicken serum (Sigma), and 50 M 2-mercaptoethanol as described (18). Stable DT40 cell lines expressing full-length, wild type SHIP1 and SHIP1 carrying two single point mutations of putative phosphorylation sites at Ser 440 (S440A) or Ser 774 (S774A) were generated for these experiments as follows. Thirty g of linearized cDNA encoding wild type, S440A, or S774A SHIP1 was co-transfected by electroporation at 250 V and 960 microfarads into SHIP-deficient DT40 cells (9) in combination with 3 g of linearized pApuro plasmid DNA to provide a selectable puromycin marker. Following electroporation, cells were allowed to stabilize for 24 h in a T-flask, the medium was removed by centrifugation, and the cells were plated in 96-well plates in RPMI 1640 medium containing all of the above supplements plus 0.5 g/ml puromycin to select for cells expressing the SHIP1 proteins. The clonal DT40 cells stably expressing wild type SHIP1 or the two point mutants were expanded and lysed, the supernatant was run on an SDS gel, and the expressed SHIP1 proteins were identified by Western blot with anti-SHIP and anti-FLAG antibodies. The DT40 lines stably expressing native SHIP1 and the two mutants were expanded in suspension culture as above. The stable DT40 cell line expressing the truncated SH2-401-900 SHIP1 protein was cultured as described (19).
Purification of Recombinant SHIP1 Proteins-The FLAGtagged wild type, truncated, or phosphorylation site mutant SHIP1 proteins were overexpressed in confluent HEK-293 cells for 48 h, and the cells were solubilized in 0.5 ml of lysis buffer containing 25 mM Hepes, pH 7.4, 150 mM NaCl, 3 mM MgCl 2 , 1% (v/v) Triton X-100, and a mix of protease and phosphatase inhibitors. This mixture contained 200 nM microcystin, 10 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, 10 g/ml pepstatin, 100 g/ml benzamidine, and 100 g/ml pefabloc SC. The cell debris was removed by centrifugation at 20,000 ϫ g for 15 min at 4°C, and the recombinant proteins were purified by immunoaffinity chromatography using 40 l of anti-FLAG beads as described (18). To purify SHIP1 from the clonal DT40 cells stably expressing wild type or mutant SHIP1 proteins, 1 ϫ 10 8 cells were pelleted by centrifugation and solubilized in 0.5 ml of the above lysis buffer, and the proteins were purified by immunoaffinity chromatography. To determine the yield and purity of each recombinant protein, fractions collected from the FLAG column were loaded on an 8% SDS gel, separated by electrophoresis, and stained with Coomassie Blue (Simply Blue, Invitrogen). The gels were scanned with a Bio-Rad GS-800 densitometer, and the concentration of SHIP1 was determined from the optical density of the protein bands as compared with a standard curve of ␤-galactosidase ranging from 0 to 500 ng/lane.
In Vitro Phosphorylation of SHIP1-To phosphorylate the purified SHIP1 protein in vitro, wild type, truncated, or phosphorylation site mutants of SHIP1 were incubated with 30 units of PKA catalytic subunit in 10 l of buffer containing 20 mM Hepes, pH 7.4, 200 nM microcystin, 0.1 mg/ml BSA, 12.5 mM magnesium acetate, and 1.25 mM EGTA (phosphorylation buffer). The reaction was started by the addition of 1 l of 1.25 mM ATP containing 2.2 ϫ 10 6 dpm of [␥-32 P]ATP/tube and incubated for 0 -30 min at 30°C with frequent gentle mixing. Reactions were terminated by adding SDS sample buffer and heating to 95°C. Aliquots were run on an 8% SDS gel, stained with Coomassie Blue or silver, and dried, and an autoradiograph was prepared with Kodak X-Omat LS film.
Mapping PKA-phosphorylated Serine/Threonine Residues in SHIP1 by Mass Spectrometry-The SHIP1 protein has 32 potential phosphorylation sites, as predicted by the KinasePhos computer program (20). In an effort to identify the majority of these sites via mass spectrometry (MS), 3 g of pure, wild type SHIP1 was phosphorylated with a large amount of PKA (10,000 units) for 60 min at 30°C in the above phosphorylation buffer without radiolabeling. A control reaction was run with 3 g of SHIP1, the PKA storage buffer (20 mM Hepes, pH 7.4, 100 mM KCl, 1 mM DTT, 10% glycerol), and no added kinase. Reactions were terminated by adding SDS sample buffer and heating to 95°C. The reaction mix was run on an 8% SDS-polyacrylamide gel, and the phosphorylated residues in SHIP1 were identified by the Biomolecular Research Facility at the University of Virginia using a Finnigan LTQ-FT MS system (21). Briefly, the gel pieces were washed and destained in 200 l of 50% methanol for 16 h and then dehydrated, rehydrated, and digested with trypsin overnight at 37°C. The tryptic peptides were extracted from the gel with two 30-l aliquots of 50% acetonitrile, 5% formic acid. These extracts were combined and evaporated to 15 l for MS analysis. The digest was analyzed using the MS/MS capability of the instrument acquiring full scan mass spectra to determine peptide molecular weights and product ion spectra to determine the amino acid sequence. The data were analyzed using the Sequest search algorithm against the mouse SHIP1 sequence. Spectra that were thought to contain phosphorylated peptides were manually examined and verified. To reduce the interference from keratin in the mass spectrometry analysis, special care was taken while purifying the SHIP1 protein and during the phosphorylation reaction by wearing gloves and by filtering all buffers through a 0.22-m filter.
Assay of SHIP1 or Phosphorylated SHIP1 Activity-The protocols for the assay of SHIP1 activity have been published (18). Briefly, the assay was carried out in a mixture of 20 mM Hepes, pH 7.4, 10 mM MgCl 2 and 100 M di-o-octanoylglyceryl-PtdIns-3,4,5-P 3 . The amount of inorganic phosphate released by dephosphorylating di-o-octanoylglyceryl-PtdIns-3,4,5-P 3 was determined using the Malachite Green reagent. To compare the activity of native SHIP1 with PKA-treated SHIP1, the purified wild type or mutant SHIP1 proteins were incubated with 30 units of PKA catalytic subunit in phosphorylation buffer for 10 min at 30°C, and the mixture was diluted 10-fold into the SHIP1 activity assay as described (18). Control reactions were treated identically but did not contain PKA.
Measurement of Akt Phosphorylation by Western Blotting-Suspension cultures of DT40 cells were washed and resuspended at 2 ϫ 10 7 cells/ml in serum-free RPMI 1640. The resuspended cells were aliquoted into separate tubes containing vehicle (H 2 O) or Sp-cAMPS and incubated for 15 min at 37°C (to phosphorylate SHIP1) prior to stimulation of the B cell antigen receptor (BCR). The BCR was activated by adding a secondary antibody (2 g/ml F(abЈ) 2 fragment of goat antimouse IgM) for 5 min before adding a mouse anti-chicken IgM (M4) antibody at 0.5 g/ml. After 0, 1, 3, and 5 min of stimulation, aliquots containing 1 ϫ 10 7 cells were withdrawn, pelleted in a microcentrifuge for 5 s, and solubilized in 100 l of lysis buffer as described (18). The mixture was centrifuged at 12,000 ϫ g for 10 min at 4°C. Ten l of the supernatant was resolved on a 12% SDS-polyacrylamide gel, transferred to nitrocellulose, and probed for phospho-Akt (Thr 308 ) and total AKT.
Analysis of Phospho-AKT in DT40 Cells by Flow Cytometry-Flow cytometry was used to select the DT40 cells expressing FLAG-tagged SHIP1 and accurately measure the phosphorylation state of Akt without potential artifacts due to cell lysis and Western blotting. The procedures used for Sp-cAMPS treatment and BCR stimulation of DT40 cells were the same as described above. Two min after BCR stimulation, 3 ϫ 10 6 cells were immediately pelleted and resuspended in 0.5 ml of prewarmed BD Phosflow Fix Buffer I. Tubes were incubated for 10 min in a 37°C water bath to fix the cells and then chilled for 1 min on ice. Cells were washed once with 1 ml of incubation buffer (0.5% BSA in PBS) and then permeabilized in 0.5 ml of BD Phosflow Perm Buffer III at 4°C for 1 h. After two washes with the incubation buffer, cells were incubated with fluorochrome-conjugated anti-FLAG and anti-phospho-Akt (Thr 308 ) at a 1:100 dilution in incubation buffer for 1 h at room temperature. Finally, cells were washed with 1 ml of incubation buffer and resuspended in 0.5 ml of PBS.
Samples were then analyzed on a BD FACSCanto II flow cytometer equipped with 488-and 633-nm lasers and emission filters for FITC (to detect the Alexa 488-conjugated phospho-Akt antibody) and allophycocyanin (to detect the Alexa 647conjugated anti-FLAG antibody). At least 10,000 events were collected for every sample. Gating on FLAG-positive cells was used to assess SHIP1 expression within three different DT40 stable lines that expressed FLAG-tagged WT or mutant SHIP1. To measure the phosphorylation state of Akt, the level phospho-Akt (Thr 308 ) in FLAG-positive cells was analyzed with FlowJo software (Tree Star, Inc.).
Data Presentation and Statistical Analysis-Averaged data are presented as means Ϯ S.E. Control SHIP1 activity presented in Figs. 4 and 6 is normalized to 100%. Statistical difference between two individual treatments was examined via unpaired t tests using Statview software (SAS Institute Inc.).

Identification of the Serine/Threonine Residues Phosphorylated in the SHIP1 Protein by Mass Spectrometry-
We previously demonstrated that SHIP1 is phosphorylated by PKA and that phosphorylation increased its catalytic activity (18). The SHIP1 protein has 32 serines or threonines within consensus phosphorylation sites that might be phosphorylated by PKA.
To determine if these residues in SHIP1 were indeed phosphorylated by PKA, pure, recombinant SHIP1 obtained from HEK-293 cells was incubated with an excess of PKA in vitro and digested with trypsin, and the resulting peptides were analyzed by mass spectrometry. Table 1 presents a summary of the peptides containing serine or threonine residues that were found to be phosphorylated in the SHIP1 protein using this technique. In total, 17 peptides containing serine or threonine residues were identified; eight were phosphorylated in the "native" protein as Phosphorylation of Ser 440 in SHIP1 Increases Its Activity NOVEMBER 5, 2010 • VOLUME 285 • NUMBER 45 purified from the HEK-293 cells (the control sample incubated without PKA), and nine additional sites were observed in SHIP1 following treatment with PKA. These nine peptides are indicated with an asterisk in Table 1. Although the phosphorylations of Tyr 917 and Tyr 1020 have been well characterized in SHIP1 function (22), none of the 17 residues presented in Table  1 have been identified previously. Fig. 1 presents the location of the 17 phosphorylated serines or threonines on a linear representation of the SHIP1 protein.
Note that most of the residues identified are clustered close to the N and C terminus of SHIP1 but that Ser 440 is near the N-terminal side of the conserved 5Ј-inositol phosphatase domain of SHIP1 (located between amino acids 408 and 701) (23). Interestingly, this domain is within the minimal catalytic region of SHIP1 (residues 401-866) (19), and the phosphopeptide containing Ser 440 (TRDDSADYIPHDIYVIGTQEDPLGEK) harbors a consensus PKA phosphorylation motif (24 -26) consisting of residues 437-440 (RDDS). In addition to Ser 440 , Ser 774 was also found to be phosphorylated following treatment of SHIP1 with PKA; this residue lies in the C-terminal side of the minimal catalytic region of SHIP1.
Phosphorylation of Truncated SHIP1-As shown in Fig. 1, the serine or threonine phosphorylation sites identified in SHIP1 by mass spectrometry tend to be distributed to the N-or C-terminal regions of the molecule. Previous work has demonstrated that fragments of the SHIP1 protein that contain the conserved 5Ј-inositol phosphatase domain (residues 401-866) have catalytic activity (19). Thus, to pursue which of the 17 phosphorylated residues might be important in the activation of SHIP1, we first generated N-terminal FLAG-tagged versions of fulllength SHIP1 and three truncation mutants of the molecule. These truncation mutants were designed to code for an active molecule by virtue of containing the catalytic region (residues 401-866) but also to contain either the N-terminal region (residues 1-866) or the C-terminal region (residues 401-1190) (see schematic in Fig. 1).
To determine if the truncation mutants could be phosphorylated in vitro, the four constructs were transiently expressed in HEK-293 cells, and the proteins were purified using anti-FLAG beads and incubated with 30 units of the catalytic subunit of PKA and [␥-32 P]ATP at 30°C for 10 min. Fig. 2A shows that all four SHIP1 proteins can be phosphorylated when equal amounts of SHIP1 protein were incubated with PKA. Note that, as expected, the full-length SHIP1 (left) was found to contain higher amounts of 32 P than the N-and C-terminal truncation mutants or the 401-866 SHIP1 catalytic domain. Fig. 2B shows that all four SHIP1 proteins maintained the ability to dephosphorylate PtdIns-3,4,5-P 3 . Consistent with our previous report (18), phosphorylation of full-length SHIP1 (residues 1-1190) with PKA increased its activity about 2-fold (left bars). More importantly, phosphorylation of the catalytic domain of SHIP1 (residues 401-866) increased its activity over 3-fold, whereas phosphorylation of either the N-terminal (residues 1-866) or C-terminal domains (residues 401-1190) of SHIP1 did not increase their activities (Fig. 2B). Also of importance is the observation that all three truncation mutants of SHIP1 had higher specific activities than the wild type protein.
The effect was particularly dramatic with the mutant missing its C-terminal region (residues 1-866), whose specific activity was 8 -10 times higher than that of full-length SHIP1 (hatched bars in Fig. 2B). The latter observation suggests an important role for the C-terminal region of SHIP1 in the regulation of its activity. Although not observed previously with SHIP1, this result is in keeping with recent studies on the related SHIP2 isoform showing that the C terminus and SH2 domains of the molecule confer an inhibitory effect to lower the basal activity of SHIP2 under unstimulated conditions (27).
Identification of the PKA Phosphorylation Site by Mutation of Serine/Threonine Residues-To search for the amino acid residue phosphorylated by PKA that activates SHIP1, we generated six single and two double point mutations of serines or threonines that were phosphorylated in the context of full-length SHIP1 protein (as shown FIGURE 1. Phosphorylation sites detected in SHIP1. The schematic diagram shows positions of phosphorylated serine or threonine residues as determined by mass spectrometry. The SH2 domain, inositol 5Ј-phosphatase domain (5-IPase), proline-rich sequences (dark bars), and the tyrosines in the C terminus are depicted. The phosphorylated serine or threonine residues found by mass spectrometry are indicated above the line. The known phosphorylated tyrosine residues are indicated below the line. The regions of the molecule encoded by the catalytic core (residues 401-866), the catalytic core and the C terminus (residues 401-1190), and the N terminus and the catalytic core (residues 1-866) are shown below the full-length molecule.

TABLE 1
Phosphopeptides detected by mass spectrometry in pure, mouse SHIP1 treated with PKA Pure, recombinant mouse SHIP1 was phosphorylated with PKA in vitro, and the reaction mixture was separated on an SDS-polyacrylamide gel. The SHIP1 bands were removed, digested with trypsin, and analyzed by MS and MS/MS. Fifteen phosphopeptides were found to contain phosphorylated serine or threonine residues. Two of the peptides contained two phosphorylated residues. The peptide masses (MH ϩ ) are listed in the table, and the phosphorylated residues are in boldface type and underlined. An asterisk indicates that the phosphopeptide was only detected in PKA-treated SHIP1. The amino acid coverage for PKA-phosphorylated SHIP1 and control was above 80%. See "Experimental Procedures" for details.  Fig. 1, these residues are located throughout the SHIP1 molecule, including the SH2 domain, the 5Ј-phosphatase domain, and the C-terminal region of SHIP1. The eight mutant SHIP1 proteins were expressed in HEK-293 cells and immunopurified with a FLAG antibody column, and the mutant proteins were incubated with 30 units of PKA for 10 min with [␥-32 P]ATP. All of the mutant SHIP1 proteins were highly expressed and efficiently purified. Fig. 3 (A and the  upper part of B) shows that 30 units of PKA was able to phosphorylate all eight of the mutants although the double mutant, S30A/S36A, and the single mutant, T1091A, were phosphorylated less well than native SHIP1 or other SHIP1 mutants. Given that the 401-866 truncated mutant of SHIP1 was markedly activated by PKA (Fig. 2B) and contained only two phosphorylated residues (Ser 440 and Ser 774 ), we made additional point mutations in the truncated 401-866 protein. Single mutations to Ala were made at Ser 440 or Ser 774 along with the double Ser 440 and Ser 774 mutant. These proteins were purified and tested in the in vitro kinase assay with PKA. The lower part of Fig. 3B demonstrates that PKA can phosphorylate the unmodified 401-866 SHIP1 fragment and the fragment harboring the S774A mutation. However, PKA did not phosphorylate S440A or the S440A/S774A double mutant of the truncated SHIP1 molecule, even after 30 min of incubation. It is noteworthy that S440A and the double mutant S440A/S774A run slower than the S774A and the unmodified 401-866 SHIP1 fragment (Fig.  3B, bottom). These experiments indicate that Ser 440 is the most likely residue within SHIP1 that, when phosphorylated by PKA, increases the catalytic activity of the inositol phosphatase.

Peptide
Effect of Mutations at Ser 440 and Ser 774 on the Activation of SHIP1 by PKA-The observation that only the activity of the 401-866 fragment of SHIP1 was increased by phosphorylation suggests that the phosphatase domain in SHIP1 may be an important target for regulation by PKA. The functional importance of modifications of the residues in the phosphatase domain is supported by the finding that naturally occurring mutations of Arg or Lys residues in the catalytic domain of INPP5E (type IV 5Ј-phosphatase) in patients with Joubert syndrome result in impaired 5Ј-phosphatase activity. Examination of fibroblasts isolated from these patients found altered ratios of phosphatidylinositol lipids, suggesting elevated phosphatidylinositol 1,4,5-trisphosphate levels. When cDNAs encoding the mutant INPP5E proteins were transfected into HEK-293T cells, the cells displayed a larger than normal phosphorylation of Akt following application of PDGF (28). Both of these results point to the hypothesis that modification of the amino acid residues in the catalytic domain may alter the activity of the  5Ј-phosphatase with direct consequences on signaling. Therefore, we mutated either Ser 440 or Ser 774 to Ala in the full-length protein and investigated the functional impact of these mutations on the phosphorylation and activation of SHIP1 both in vitro and in cells.
The wild type and full-length SHIP1 harboring the S440A or S774A mutations were expressed in HEK293 cells and purified, and the 5Ј-phosphatase activity of the recombinant proteins was measured following incubation with 30 units of PKA for 10 min. Fig. 4A shows that, as expected, the activity of the wild type protein is increased following phosphorylation with PKA (left bars); however, the protein harboring the S440A mutation was not activated following incubation with PKA (middle bars). The protein with the S774A mutation did show a small but significant increase in activity following incubation with PKA (right bars). To determine how mutations of S440A or S774A affect the activation of the protein by PKA in intact cells, experiments analogous to those shown in Fig. 4A were performed by cotransfecting HEK-293 cells with cDNA encoding wild type or mutant SHIP1 proteins and a plasmid expressing the active PKA catalytic subunit, the proteins were purified, and their 5Ј-phosphatase activity was measured. Expression of an exoge-nous PKA catalytic subunit stimulated the activity of the wild type and S774A SHIP1 activity about 1.4-fold. In contrast, cotransfection of S440A SHIP1 with PKA did not stimulate the activity of the protein purified from the cells (Fig. 4B). As above, these results indicate that Ser 440 may be the most likely residue to regulate the activity of SHIP1 via phosphorylation.
Ser 440 in SHIP1 Is Functionally Relevant in B Lymphocytes-The above data suggested that Ser 440 is an important regulatory site in SHIP1. We chose DT40 B lymphocytes to determine whether the phosphorylation of Ser 440 (or Ser 774 ) regulates the function of SHIP1 and inositol lipid metabolism in a cellular model. PtdIns-3,4,5-P 3 levels are markedly stimulated in DT40 cells by engagement of the BCR, and the activity of SHIP1 plays a key role in opposing the increase in PtdIns-3,4,5-P 3 caused by BCR stimulation (2,18). Importantly, the DT40 cell line has a high rate of homologous recombination activity, and variant DT40 lines with specific proteins genetically knocked out are easily created (9,19). The availability of SHIP1 Ϫ/Ϫ DT40 cells (9) and the ease of developing new stable lines expressing modified SHIP1 proteins (19) makes them an especially attractive cell system to study how the SHIP1 protein is regulated by phosphorylation. The first SHIP Ϫ/Ϫ DT40 line we examined expressed a truncated version of SHIP1 containing only the N-terminal SH2 domain and the 401-900 minimal catalytic region of SHIP1 (19). This 80-kDa protein contains two of the important functional domains of SHIP1, the SH2 sequence needed for translocation to the membrane and an active catalytic region that includes the Ser 440 phosphorylation site. It was reasoned that the activity of this truncated protein would respond to phosphorylation by PKA, but it had no other domains to complicate the response (19).
Immunoblotting results presented in Fig. 5A demonstrate that the truncated, SH2-401-900 SHIP1 protein was properly expressed in the SHIP Ϫ/Ϫ DT40 cells as a 80-kDa protein. Raising PtdIns-3,4,5-P 3 levels in these cells by ligation of the BCR caused a 2-fold increase in the phosphorylation of Akt on Thr 308 over a 5-min time period (Fig. 5B). Pretreatment of the cells with 100 M Sp-cAMPS for 15 min markedly reduced the BCR-induced phosphorylation of Akt at the 1 and 3 min time points, suggesting that the SHIP1 protein was activated. This result is consistent with the in vitro result shown in Fig. 2B with pure proteins and implies that the 401-900 catalytic region of SHIP1 is modified via phosphorylation with PKA.
We then generated new stable lines expressing wild type, S440A, or S774A mouse SHIP1 proteins. Between two and eight clones were obtained for each stable line, and the expression level of SHIP1 was monitored using an anti-SHIP1 antibody. Fig. 6A presents a Western blot showing high expression of SHIP1 in the three DT40 lines used in the flow cytometry experiments described below. Fig. 6B provides quantitative data on the expression of wild type SHIP1 and the two mutants as measured by flow cytometry using a fluorescent antibody against the FLAG epitope on the protein. Fig. 6B shows that 92.4 -96% of the cells in each of the three new lines express SHIP1. Importantly, the shape of the peaks suggests that most cells in each population express similar amounts of SHIP1. We then used each of the three lines to measure the phosphorylation of Akt following BCR engagement as an index of PtdIns-

. Phosphorylation of Ser 440 in SHIP1 is required for activation of SHIP1 by PKA in vitro and in intact HEK-293 cells.
A, the activity of the purified wild type protein, the full-length S440A mutant, and the S774A mutant SHIP1 protein were compared using the Malachite Green assay following treatment with 30 units of PKA for 10 min. Activity was stimulated only with the wild type and S774A mutant SHIP1 proteins. NS, not significant; ***, p Ͻ 0.001, n ϭ 8. B, plasmids expressing the wild type or mutant SHIP1 proteins were co-transfected into HEK-293 cells with or without a plasmid expressing the PKA catalytic subunit and incubated for 40 h. The recombinant SHIP1 proteins were immunopurified, and the activity of 10 ng of SHIP1 protein was measured using the Malachite Green assay. Transfection of the catalytic subunit only increased the activity of the wild type and S774A mutant SHIP1 proteins. NS, not significant; **, p Ͻ 0.01; ***, p Ͻ 0.001, n ϭ 5. The rate of PtdIns-3,4,5-P 3 hydrolysis for control samples ranged between 1.4 and 1.9 pmol of P i /min/ng of SHIP1 and was normalized to 1.0. Error bars, S.E.
Using flow cytometry, we measured the phospho-Akt levels in SHIP Ϫ/Ϫ cells expressing the wild type mouse SHIP1 protein.
The cells were stimulated with the BCR antibody for 2 min because the experiments performed with immunoblotting (Fig.  5B) show a nearly maximal level of phospho-Akt activation and a nearly complete inhibition of the response by activation of PKA with Sp-cAMPS at that time point. Note that in cells expressing wild type SHIP1, there is a clear increase (right peak shift) in phospho-Akt fluorescence at 2 min following BCR engagement (compare the bottom two peaks in the upper panel of Fig. 6C). The quantitative analysis of the increase in fluorescent intensity is shown in the bottom panel of Fig. 6C. BCR activation in cells expressing native SHIP1 caused a 3.4-fold increase in the mean fluorescence intensity (MFI) of the phospho-Akt (Thr 308 ) antibody at the 2 min time point (solid bar in the bottom panel of Fig. 6C). When these cells were pretreated with Sp-cAMPS for 15 min, the increase in MFI (phospho-Akt) due to BCR engagement was only about 1.6-fold (hatched bars). This result establishes that Sp-cAMPS-dependent phosphory-lation of SHIP1 activates the enzyme and leads to the expected result of lowered PtdIns-3,4,5-P 3 levels.
We next examined the phospho-Akt levels in SHIP1 Ϫ/Ϫ DT40 cells expressing the full-length mouse SHIP1 molecule harboring specific point mutations (S440A or S774A) using similar protocols. Fig. 6D presents an experiment performed with cells expressing the full-length S440A point mutation of SHIP1. The increase in phospho-Akt (peak shift) caused by BCR engagement was similar to that observed in cells expressing the wild type protein, but pretreatment of the cells with Sp-cAMPS did not inhibit the phospho-Akt response (peak shift) to BCR engagement. The quantitative analysis of the changes in fluorescent intensity is shown in the bottom panel of Fig. 6D. The level of phospho-Akt was increased about 3.9-fold following BCR engagement (solid bar), but the increase was still about 3.6-fold after treatment with Sp-cAMPS (hatched bars). There is no significant difference in the increase in MFI (increase in phospho-Akt levels) between the control and Sp-cAMPS-treated cells (p ϭ 0.08). Fig. 6E presents an analogous experiment performed with DT40 cells expressing the S774A point mutation. As observed with cells expressing the wild type protein, BCR engagement caused a marked increase in phospho-Akt (a 3.2-fold increase in MFI), and pretreatment of the cells with Sp-cAMPS markedly blunted the response (only about a 1.7-fold increase in MFI). This result is consistent with the hypothesis that Ser 774 is not essential for the regulation of SHIP1 via phosphorylation by PKA. Taken together, these results suggest that activation of SHIP1 via phosphorylation of Ser 440 is an important regulatory event in hematopoietic cells.
Effect of Sp-cAMPS on the Activity of SHIP1 Purified from DT40 Cells-To address how phosphorylation of SHIP1 with the endogenous DT40 cell PKA affects its enzymatic activity, SHIP1 was immnunopurified from the DT40 cell lines overexpressing the wild type and S440A point mutants of SHIP1. The cells were treated with Sp-cAMPS for 15 min, and the proteins were purified and assayed for their ability to dephosphorylate di-o-octanoylglyceryl-PtdIns-3,4,5-P 3 in vitro. As shown in Fig.  7, the activity of native SHIP1 purified from DT40 cells was increased about 1.5-fold following treatment with Sp-cAMPS. The activity of the S440A mutant of SHIP1 was not increased by treatment of the cells with Sp-cAMPS. The right two bars present experiments performed with a second, S440A-expressing clone (Clone 2), which provided identical results. Collectively, the results shown in Figs. 6 and 7 strongly support a role for Ser 440 as a relevant PKA phosphorylation site on SHIP1 in cells.

DISCUSSION
The two major isoforms of the SH2 domain containing phosphatidylinositol phosphatase, SHIP1 and SHIP2, are important enzymes that lower the level of the PtdIns-3,4,5-P 3 signal (5, 6). SHIP1 is highly expressed in hematopoietic cells and is an important regulator of the phosphatidylinositol 3-kinase/AKT pathway in these cells (5,6). We recently demonstrated that the activity of SHIP1 could be stimulated by phosphorylation with PKA (18); this finding provides a new mode of SHIP1 regulation by G protein-coupled receptors that raise cyclic AMP. Because of the biological and therapeutic significance of SHIP1 (5), identification of the phosphorylation sites in SHIP1 is an important step toward understanding the molecular mechanism of SHIP1 activation. In the present study, we have identified and characterized the amino acid residues responsible for the regulation of SHIP1 by PKA. Using a combination of HPLC and MS/MS analysis of the purified protein phosphorylated in vitro by PKA, a phosphorylation map containing 17 potential phosphorylation sites was established for SHIP1. Further mutational analysis of full-length and truncation mutants of SHIP1 showed that Ser 440 is the site phosphorylated by PKA that increases its activity.
The initial step in determining which of the 17 serine or threonine residues might be important in the activation of SHIP1 was to measure the effect of PKA phosphorylation on three active but truncated SHIP1 proteins, designed to contain the known regulatory regions of the molecule. One protein contained the catalytic domain (residues 401-866) flanked by the N-terminal region (residues 1-866), another contained the catalytic domain flanked by the C-terminal region (residues 401-1190), and the third contained only the catalytic core of the protein (residues 401-866). All three truncated SHIP1 molecules were phosphorylated in vitro by PKA ( Fig. 2A), but only the catalytic domain of the molecule (amino acid residues 401-866) was activated significantly by phosphorylation (Fig. 2B). This result suggests that this region is of primary importance for activation of SHIP1 by PKA. Using in vitro studies with pure proteins, cell-based approaches, and point mutations of potential serine residues, we identified Ser 440 in the catalytic domain as the amino acid responsible for most of the functional activation of SHIP1 by PKA.  This conclusion is based on six lines of evidence. First, the phosphopeptide containing Ser 440 that was identified by mass spectrometry (TRDDSADYIPHDIYVIGTQEDPLGEK) harbors the RXX(S/T) consensus motif for phosphorylation by PKA (24 -26). It is noteworthy that the KTRDDSAD sequence is conserved among all five known SHIP1 homologues in Homo sapiens, Rattus norvegicus, Mus musculus, Xenopus laevis, and Salmo salar (Fig. 8). Second, mutation of Ser 440 to Ala in the truncated, catalytic domain of the molecule (residues 401-866) abolished the phosphorylation of SHIP1 by PKA in vitro (Fig.  3B). Third, mutation of Ser 440 to Ala in the full-length molecule (residues 1-1190) greatly reduced the ability of PKA to activate SHIP1 in vitro (Fig. 4A). Fourth, overexpression of the PKA catalytic subunit with the full-length wild type SHIP1 protein in HEK-293 cells results in an activated SHIP1 (Fig. 4B); the fulllength protein containing an S440A point mutation is not activated in this protocol. Fifth, flow cytometry analysis demonstrates that transfection of a full-length S440A mutant protein into SHIP Ϫ/Ϫ DT40 cells yields a cell in which the phospho-Akt response is not inhibited by activation of PKA (Fig. 6D). This finding suggests that the phosphorylation of SHIP1 is a physiologically relevant process in hematopoietic cells. Finally, when SHIP Ϫ/Ϫ DT40 cells expressing the wild type SHIP1 protein are treated with Sp-cAMPS to activate PKA, the enzyme is activated. If a full-length S440A point mutation is expressed in the cells, treatment with Sp-cAMPS does not activate SHIP1 (Fig.  7). Taken together, these data identify Ser 440 as the site in the molecule that is phosphorylated by PKA to activate SHIP1.
The phosphorylation site identified, Ser 440 , is not in one of the domains in SHIP1 previously thought to be involved in its regulation (6). The major regulatory domains of SHIP1 include an SH2 domain in its N-terminal region, a C2 domain located at the C-terminal end of its phosphatase domain, and a C terminus containing a proline-rich region and two NPXY amino acid sequences. These domains have been shown to act as important docking sites for signaling molecules, adaptor proteins, allosteric ligands, and immunoreceptors (8 -10, 15, 16). The SH2 domain mediates interaction with tyrosine-phosphorylated proteins, the immunoreceptor tyrosine-based inhibitory motifs, and immunoreceptor tyrosine-based activation motifs (8,9). The C2 domain is the binding site for potential allosteric activators, such as phophatidylinositol 3,4-bisphosphate, and novel small molecules, such as AQX-MN100 (15,16). The proline-rich region of the C terminus of SHIP1 interacts with the SH3 domain of Grb2 or PLC␥ following activation of the immunoreceptor tyrosine-based activation motif-containing Fc⑀RI in RBL-2H3 cells (29). The two NPXY sequences, when phosphorylated following stimulation with antibodies, bind proteins such as Shc and Dok2, which have phosphotyrosine binding domains (17,30). No phosphorylation or direct regulation of the central catalytic core (the 5Ј-phosphatase region itself) has been previously reported. The current model for SHIP1 regulation in various physiological situations is primarily based on the translocation of SHIP1 from the cytosol to the membrane following activation of receptor tyrosine kinases (growth factors, cytokine receptors, or immunoreceptors). In this translocation model, no significant change has been detected in the 5Ј-phosphatase activity of SHIP1 (13). Our results indicate that SHIP1 activity can be regulated by PKA phosphorylation on Ser 440 in the catalytic core and highlight a new mode of SHIP1 regulation.
Although Ser 440 may be a primary phosphorylation site involved in SHIP1 regulation, the possibility remains that other serine/threonine residues (Ser 30 , Ser 36 , Thr 1091 , Ser 1113 , and Ser 1160 ) identified by mass spectrometry may also be involved in the overall regulation. As shown in Fig, 2B, two of the truncated constructs (mutants 401-1190 and 1-866) were not activated by phosphorylation with PKA despite the presence of Ser 440 . Perhaps more importantly, the SHIP1 molecule lacking its C terminus (truncation mutant 1-866) is activated about 8 -10-fold (see Fig. 2B). In this regard, it is most interesting that truncation of the homologous SHIP2 molecule by removal of the C-terminal region (truncation mutant 1-822) has also been shown to activate the enzyme (27). The finding that the C-terminal region of SHIP2 seems to inhibit its catalytic activity is very similar to our finding with an analogous mutant of SHIP1 (Fig. 2B). Thus, besides directly affecting the catalytic activity of the enzyme, Ser 440 phosphorylation could direct a conformational change in the protein, which relieves an inhibitory effect of the molecule's C-terminal tail. The N terminus may also play a role in regulating the activity of SHIP1 (see Fig. 2B), but more information is needed to clarify this possibility. Taken together, the data suggest that phosphorylation of Ser 440 initiates events that activate the catalytic domain and may also relieve an inhibitory constraint by the C terminus of the molecule.
SHIP2 is a 142-kDa protein that contains a catalytic core structure similar to SHIP1 (7). Unlike the case with SHIP1, the phosphorylation of certain tyrosines in SHIP2 can enhance its 5Ј-phosphatase activity as measured in immunoprecipitates (27,31). For example, if HeLa cells are transfected with SHIP2 and treated with EGF or H 2 O 2 , there is a marked increase in the tyrosine phosphorylation state of SHIP2 and a 2-fold increase in the activity of the enzyme (27,31). Mutational analysis of tyrosines in SHIP2 showed that Tyr 986 /Tyr 987 and Tyr 1135 were the most likely sites phosphorylated in response to EGF (27). Importantly, there are three potential regulatory serine/threonine residues (Thr 958 , Ser 1003 , and Ser 982 ) in the C-terminal region of SHIP2 (32)(33)(34). These three residues are located near Tyr 986 , a well known tyrosine phosphorylation site in the NPXY motif of the molecule. Therefore, it has been suggested that the phosphorylation of Thr 958 , Ser 1003 , or Ser 982 could affect the affinity of Tyr 986 for its tyrosine kinase (32). Although there are a number of serine or threonine residues in the corresponding NPXY region of SHIP1 (see Fig. 1), based on our mutational analysis, they do not appear to be involved in the regulation of SHIP1 activity.

Phosphorylation of Ser 440 in SHIP1 Increases Its Activity
There is recent genetic and structural evidence that modification of amino acid residues in the catalytic region of the inositol phosphatase can markedly alter their activity. Genetic studies with the type IV 5Ј-phosphatase (INPP5E), a lipid phosphatase that hydrolyzes phosphate at the 5Ј-position from PtdIns-3,5-P 2 , phosphatidylinositol 4,5-bisphosphate, and PtdIns-3,4,5-P 3 , hint at how regulation of 5Ј-phosphatase activity could be modified by alteration of the residues in the 5Ј-phosphatase catalytic domain (28). INPP5E is a highly active 5Ј-phosphatase of PtdIns-3,4,5-P 3 that is widely expressed and found in brain, testis, breast, hemopoietic, and other cell types (35). Patients with human ciliary diseases were found to have single-nucleotide polymorphisms in the INPP5E gene in the region of the molecule corresponding to the catalytic core of SHIP1 (28). These single-nucleotide polymorphisms mutated arginine or lysine residues centered in the phosphatase domain and greatly impaired the 5Ј-phosphatase activity of the enzyme. Study of primary fibroblasts cultured from individual patients showed altered cellular phosphatidylinositol lipid ratios and transfection of the mutated proteins into HEK-293 cells resulted in phenotypes with high basal levels of phospho-Akt and an exaggerated response to PDGF (28). The authors modeled these mutations on the crystal structure of synaptojanin, a yeast 5Ј-inositol phosphatase (36), and found the residues to be clustered in the catalytic core of the molecule near the binding pocket for PtdIns-3,4,5-P 3 (28). These data suggest that modification of the amino acids in the 5Ј-phosphatase catalytic region of SHIP1 is capable of inducing meaningful changes in the enzyme's activity and provide additional support for our hypothesis that phosphorylation of Ser 440 in the catalytic core of SHIP1 is important for regulation of PtdIns-3,4,5-P 3 signaling.
In conclusion, although tyrosine phosphorylation and activation of SHIP1 and SHIP2 is a highly studied area, there are few comprehensive reports on the regulation of SHIP1 or SHIP2 by serine/threonine kinases. PKA is the best characterized member of the serine/threonine protein kinase family, and phosphorylation of its substrates is important for regulating many cellular processes. Our previous observation that phosphorylation of SHIP1 with PKA markedly activates its function provided insight into an important new mode of SHIP1 regulation. In the present study, we have identified the Ser 440 residue in SHIP1 as a major PKA phosphorylation site with functional implications. These results outline a new role for G protein-coupled receptors and PKA in regulating PtdIns-3,4,5-P 3 levels and inflammatory signals via the activation of SHIP1.