Reversible Ser/Thr SHIP phosphorylation: A new paradigm in phosphoinositide signalling?

Targeting of SHIP1/2 phosphatases may be controlled by phosphorylation on Ser and Thr residues


  • William's Elong Edimo,

    1. Institut de Recherche Interdisciplinaire (IRIBHM), Université Libre de Bruxelles, Campus Erasme, Brussels, Belgium
    Search for more papers by this author
  • Veerle Janssens,

    1. Protein Phosphorylation & Proteomics Lab, Faculty of Medicine, Department of Cellular and Molecular Medicine, KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • Etienne Waelkens,

    1. Protein Phosphorylation & Proteomics Lab, Faculty of Medicine, Department of Cellular and Molecular Medicine, KU Leuven, Leuven, Belgium
    Search for more papers by this author
  • Christophe Erneux

    Corresponding author
    1. Institut de Recherche Interdisciplinaire (IRIBHM), Université Libre de Bruxelles, Campus Erasme, Brussels, Belgium
    • Institut de Recherche Interdisciplinaire (IRIBHM), Université Libre de Bruxelles, Campus Erasme, Brussels, Belgium.
    Search for more papers by this author


Phosphoinositide (PI) phosphatases such as the SH2 domain-containing inositol 5-phosphatases 1/2 (SHIP1 and 2) are important signalling enzymes in human physiopathology. SHIP1/2 interact with a large number of immune and growth factor receptors. Tyrosine phosphorylation of SHIP1/2 has been considered to be the determining regulatory modification. However, here we present a hypothesis, based on recent key publications, highlighting the determining role of Ser/Thr phosphorylation in regulating several key properties of SHIP1/2. Since a subunit of the Ser/Thr phosphatase PP2A has been shown to interact with SHIP2, a putative mechanism for reversing SHIP2 Ser/Thr phosphorylation can be anticipated. PI phosphatases are potential target molecules in human diseases, particularly, but not exclusively, in cancer and diabetes. Therefore, this novel regulatory mechanism deserves further attention in the hunt for discovering novel or complementary therapeutic strategies. This mechanism may be more broadly involved in regulating PI signalling in the case of synaptojanin1 or the phosphatase, tensin homolog, deleted on chromosome TEN.

Editor's suggested further reading in BioEssays: Pairing phosphoinositides with calcium ions in endolysosomal dynamics Abstract


INPP4A/B, inositol polyphosphate 4-phosphatases (types I and II); PI, phosphoinositide; PIPKI, phosphatidylinositol phosphate kinase; PP2A, protein phosphatase 2A; PtdIns(3,4)P2, phosphatidylinositol 3,4-bisphosphate; PtdIns(3,4,5)P3, phosphatidylinositol 3,4,5-trisphosphate; PtdIns(4,5)P2, phosphatidylinositol 4,5-bisphosphate; PTEN, phosphatase, tensin homolog, deleted on chromosome TEN; SAM, sterile α motif; SHIP1/2, SH2 domain-containing inositol 5-phosphatases 1 and 2; SYNJ1, synaptojanin1.


In eukaryotes, phosphatidylinositol is the precursor of seven phosphorylated derivatives (the phosphoinositides or PIs) that are critical molecules in signalling events 1. Phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P3) is the major product of class I PI 3-kinase, an enzyme known to be mutated in human cancers of the prostate, breast and endometrium. It is also critical in cell migration, proliferation and apoptosis 2. In pancreatic β cells, the same enzyme, class IA PI 3-kinase, controls insulin secretion 3. Therefore, the latter enzyme is also considered as a therapeutic target in patients with type 2 diabetes. Dephosphorylation of the different PIs can occur at the 3, 4 or 5 position of the inositol ring catalyzed by specific PI phosphatases that are just as important as the PI kinases 4. The inositol polyphosphate 5-phosphatases are enzymes that remove the 5-phosphate. As shown in Fig. 1, this family of enzymes contains ten members that are widely implicated in human diseases; for example, loss of function of OCRL results in the X-linked oculocerebrorenal Lowe syndrome (reviewed in ref. 5). The SH2 domain-containing inositol 5-phosphatases 1/2, i.e. SHIP1 (or SHIP) and SHIP2, belong to this family 4. They contain an N-terminal SH2 domain, a catalytic domain, NPXY motifs (that once phosphorylated bind proteins with a phosphotyrosine or SH2 domain) and C-terminal proline-rich regions with consensus sites for SH3 domain interactions 6. In addition, SHIP2 contains a unique sterile α motif (SAM) domain at its C-terminal end that is typically involved in protein:protein interactions (Fig. 1). SHIP1/2 are essentially (but not exclusively) PtdIns(3,4,5)P3 5-phosphatases that negatively control PtdIns(3,4,5)P3 levels in intact cells, producing phosphatidylinositol 3,4-bisphosphate (PtdIns(3,4)P2) as the reaction product (Fig. 2) 7. The presence of an SH2 domain that recognizes phosphotyrosine in the sequence of a PI phosphatase is of particular interest due to the importance of phosphotyrosine signalling in cell proliferation and differentiation 8.

Figure 1.

The human inositol polyphosphate 5-phosphatase family. The human inositol polyphosphate 5-phosphatase family consists of ten different isoenzymes. The sequence of each enzyme shows the presence of a relatively conserved phosphatase domain and numerous signalling motifs or domains involved in enzyme localization and protein:protein interactions: CAAX motif, GAP (RhoGTPase-activating protein), SKICH [SKIP (skeletal muscle- and kidney-enriched inositol phosphatase) carboxy homology], SAC1, SH2, PRD (proline-rich domain), SAM. The presence of SHIP1/2 NPXY sites, SHIP2 S132 phosphorylation and K315 ubiquitination sites are indicated. The data have been reviewed in refs. 1, 4, 7, 59 (reproduced from ref. 4 with permission from Springer).

Figure 2.

PtdIns(3,4,5)P3 phosphatases: PTEN and SHIP1/2. PtdIns(3,4,5)P3 can be dephosphorylated by a family of inositol 5-phosphatases including SHIP1/2, OCRL and SYNJ1/2. PtdIns(3,4)P2 can be dephosphorylated to PtdIns3P by INPP4A/B.

Discovery of SHIP1/2 in a complex of tyrosine phosphorylated proteins

Tyrosine phosphorylated protein complexes are known to occur in association with growth factor receptors and protein adaptors. SHIP1 was initially identified as a 145-kDa protein in a complex of tyrosine phosphorylated proteins associated with adaptor proteins Shc and Grb2 in B and T cells 9, 10. In contrast, SHIP2 was isolated by a molecular biology approach (using degenerate PCR based on sequence homology between inositol 5-phosphatases, reviewed in ref. 6). Independently, Wisniewski et al. 11 also purified SHIP2 as a 155-kDa tyrosine phosphorylated protein in p210bcr/abl-expressing haematopoietic cells. SHIP1/2 are therefore tyrosine phosphorylated proteins associated with immunoreceptors or growth factor receptors (e.g. SHIP2 was found associated to the epidermal growth factor (EGF) receptor 12). SHIP1/2 tyrosine phosphorylation could be achieved at NPXY motifs in the sequence of both proteins (Fig. 1). The ‘classical’ targeting mechanism of SHIP1/2 to the plasma membrane therefore involves the SH2 domain of SHIP and/or the ability of SHIP to be tyrosine phosphorylated and to interact with adaptors such as Shc, Dok and Grb2 proteins 13.

Tumour suppressor PTEN is critical in lowering the level of PtdIns(3,4,5)P3 and the Akt pathway, whereas SHIPs may play a dual role

The phosphatase, tensin homolog, deleted on chromosome TEN (PTEN) dephosphorylates PtdIns(3,4,5)P3. PTEN is one of the most frequently mutated tumour suppressors in human cancers 14. In contrast to SHIP1/2, PTEN is a 3-phosphatase and phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) is the reaction product (Fig. 2). Therefore, it is widely assumed that the PI 3-kinase/Akt pathway is controlled by PtdIns(3,4,5)P3 5-phosphatase (SHIP1/2) and 3-phosphatase (PTEN) 13. The direct protein target of PtdIns(3,4,5)P3 is Akt (also called protein kinase B), which binds PtdIns(3,4,5)P3 through its Pleckstrin homology (PH) domain and is critical in the control of cell survival, migration and metabolism 15. However, PtdIns(3,4)P2, the product of PtdIns(3,4,5)P3 5-phosphatase reaction, is also an activator of the Akt pathway in addition to other specific protein targets (reviewed in ref. 16). PtdIns(3,4)P2 is rapidly dephosphorylated to PtdIns3P, thereby preventing a massive increase in PtdIns(3,4)P2. Inositol polyphosphate 4-phosphatase type II (INPP4B), one of the enzymes that dephosphorylate PtdIns(3,4)P2, has now been identified as a tumour suppressor (Fig. 2) 17, 18. Thus, both PTEN and SHIP1/2 negatively control PtdIns(3,4,5)P3, whereas only SHIP1/2 phosphatases positively control PtdIns(3,4)P2, which is critical in tumorigenesis.

SHIP1/2 function: Lessons from knockout mice

SH2 domain-containing inositol 5-phosphatases 1 expression is restricted to haematopoietic cells, whereas SHIP2 is more ubiquitous. Studies in SHIP1-knockout mice have shown that SHIP1 is particularly implicated in immune responses, myeloid cell survival and platelet activation (reviewed in ref. 19). SHIP1-knockout mice develop a myeloproliferative disease 20. The literature also provides evidence suggesting that SHIP1 may act as a tumour suppressor during leukemogenesis and lymphomagenesis 13. SHIP2-knockout mice show resistance to obesity and an elevated basal metabolic rate when placed on a high-fat diet 21. Interestingly, SHIP2-null mice also show a distinctive truncated facial profile resulting from an abnormality in skeletal growth in that region (Table 1). In summary, the role of SHIP1 in immune responses and cancer is very well supported. In contrast, the role of SHIP2 is much more diversified: many reports now suggest that SHIP2 function does not pertain to a single molecular pathway but is involved in different aspects of cell signalling (Table 1). It is very likely that this diversity is linked to specific targeting of the enzyme to different cellular locations. It may also be associated with specific post-translational modifications of SHIP, such as phosphorylation.

Table 1. SHIP2 in the context of physiopathology, proliferation and cell differentiation
SHIP2, like PTEN, causes a potent cell cycle arrest in G1 in glioblastoma cells 28
The presence of SHIP2 gene mutations is associated with type 2 diabetes in rat and human 57
Insulin-induced phosphorylation of AKT in liver is impaired in wild-type SHIP2-expressing db/m mice 60
Absence of SHIP2 in mice confers resistance to diet-induced obesity 21
SHIP2-null mutant mice exhibit runting and skeletofacial anomalies 21
Aggressive squamous cell carcinoma (SCC) cells exhibit elevated levels of miR-205 associated with a concomitant reduction in SHIP2 levels 61
SHIP2 is part of the molecular interaction network of the tyrosine kinase Bcr-ABL that causes chronic myeloid leukaemia 62
SHIP2 limits the spread of FGF signalling in the Zebrafish embryo 63

Reversible Ser/Thr phosphorylation of SHIP could be critical in targeting the enzyme

The preferred substrate of SHIP1/2 is PtdIns(3,4,5)P3, often only detected in cells stimulated by the action of growth factors or insulin. SHIP1/2 in starved cells (i.e. in the absence of any extracellular signal) are essentially cytoplasmic proteins 7. Therefore, mechanisms targeting SHIP1/2 to the plasma membrane, or other parts of the cell, are crucial events in the determination of enzymatic function and substrate specificity. Both SHIP1 and SHIP2 have been recently reported to be present in the nucleus 22, 23. Our data in human astrocytoma cells indicate that SHIP2 phosphorylation on S132 is targeted to the nucleus and to nuclear speckles 22. The importance of nuclear PIs in regulating stress response, nuclear actin and PtdIns(4,5)P2 nuclear function(s) is now well established 24 and has been reviewed in the case of nuclear phospholipase C β1 (an enzyme that uses PtdIns(4,5)P2 as substrate) 25.

In this hypothesis article, we propose a new paradigm in SHIP control of activity and targeting: reversible Ser/Thr phosphorylation. This mechanism may be critical in coordinating SHIP1/2 properties. It could account for adjusting PI levels, targeting of the enzyme, or directing new functions in mitosis, proliferation or differentiation. It could also be important in determining SHIP substrate specificity. Moreover, Ser/Thr phosphorylation could be more broadly involved in regulating signalling events in the case of two other PI phosphatases, i.e. synaptojanin1 (SYNJ1) and PTEN.

SHIP1 and SHIP2 substrate specificity: More than one substrate in vivo?

The classical view of SHIP1/2 is that they dephosphorylate PtdIns(3,4,5)P3 at the plasma membrane. This is essentially observed in cells in which PI 3-kinase is stimulated by immunoreceptors, growth factors (e.g. EGF) or insulin. SHIP1/2 in cellular extracts or as purified enzymes have also been shown to recognize other substrates; in particular, SHIP2 also recognizes PtdIns(4,5)P2 26–28. The in vivo significance of this was assessed in intact cells in the endocytic pathway: PtdIns(4,5)P2 levels in COS-7 cells deficient for SHIP2 were upregulated as compared to control cells 29. The data suggest a role of SHIP2 in the control of PtdIns(4,5)P2 and endocytic clathrin-coated pit dynamics. Thus, the existence of multiple substrates of SHIP1/2 can be evidenced in vivo: both PtdIns(3,4,5)P3 and PtdIns(4,5)P2 can be negatively controlled by SHIP2. This does not imply that these dephosphorylation events occur at the same time, nor in the same cell model. It is of interest that other members of the inositol 5-phosphatase family, e.g. OCRL, INPP5E (Fig. 2), also display phosphatase activity for both PtdIns(3,4,5)P3 and PtdIns(4,5)P2.

Is SHIP1/2 tyrosine phosphorylation an index of enzymatic activation?

It is often assumed that SHIP1/2 tyrosine phosphorylation could be an index of PtdIns(3,4,5)P3 phosphatase activation 30, 31 but this remains controversial 22, 28. In human platelets stimulated by thrombin, there is a good correlation between SHIP1 tyrosine phosphorylation, translocation to the actin cytoskeletal fraction and PtdIns(3,4)P2 production 32. However, this does not necessarily imply activation of the enzyme phosphorylated on tyrosine; it could just as well reflect its correct localization close to its substrate PtdIns(3,4,5)P3. This initial step is required for effective production of PtdIns(3,4)P2 and control of platelet aggregation 33. In totipotent and haematopoietic stem cells, there is evidence for the presence of a 110-kDa form of SHIP1 (sSHIP). This enzyme lacks the SH2 domain of the 145-kDa SHIP1 isoform (present in mature cells and referred to as the complete isoform SHIP1). sSHIP is not tyrosine phosphorylated 34, 35. It is, however, present in a membrane fraction due to its interaction with Grb2. It thus appears that sSHIP could be active in signalling in the absence of tyrosine phosphorylation. A similar observation applies to SHIP2 in human astrocytoma cells: SHIP2 controls cell morphology, proliferation and the level of PtdIns(3,4,5)P3 in response to serum despite the fact that SHIP2 is not tyrosine phosphorylated 22. Thus, in these two cell models, tyrosine phosphorylation of SHIP1/2 is not the mechanism required to correctly localize and/or activate the enzyme at the plasma membrane.

Is SHIP1/2 tyrosine phosphorylation a signal for other cellular functions unrelated to control of activity? This is very likely; for example, SHIP1 tyrosine phosphorylation acts as a signal for its ubiquitination and proteasomal degradation 13. Tyrosine phosphorylation is required here for the adjustment of protein levels of SHIP1, e.g. in bone-derived macrophages in response to IL-4 36. In this instance, phosphorylation would be a mechanism to switch off SHIP1-mediated PtdIns(3,4,5)P3/Akt control of cell survival.

SYNJ1 and SHIP1 phosphorylation on Ser residues

The PI 5-phosphatase SYNJ1 regulates normal PtdIns(4,5)P2 balance at synapses and is required for neurotransmission 37. It participates with dynamin in synaptic vesicle recycling. This enzyme, concentrated in nerve terminals, dephosphorylates PtdIns(4,5)P2 in endocytic membranes, thereby facilitating clathrin uncoating. It was the first example of a PI 5-phosphatase controlled by Ser/Thr phosphorylation catalyzed by cyclin-dependent kinase 5 38.

Proteomic studies found in databases (e.g. PhosphoSitePlus: indicate that SHIP1/2 can be phosphorylated not only on Tyr but also on Ser or Thr residues 39. Does the Ser/Thr phosphorylation of SHIP1/2 influence PI phosphatase activity? Starting with the purified enzyme, Zhang et al. identified S440 as a phosphorylation site of SHIP1 in the catalytic region of the enzyme, providing evidence that phosphorylation of this residue increases the PtdIns(3,4,5)P3 5-phosphatase activity in B cells 40. The data suggest a role of G protein-coupled receptors and protein kinase A activation of SHIP1 via a phosphorylation mechanism. Thus, SHIP1 phosphatase activity can be controlled by Ser phosphorylation.

SHIP2 phosphorylation on T958 and S132 and other phosphosites

The first evidence of SHIP2 phosphorylation on Ser/Thr was provided in platelet-derived growth factor (PDGF)-stimulated 3T3-L1 preadipocytes overexpressing SHIP2 41. Phospho SHIP2 T958 was identified in response to PDGF and this phosphorylation event may decrease SHIP2 tyrosine phosphorylation and interaction with a major adaptor protein Shc. In a more recent study, in 32P-labelled COS-7 cells overexpressing SHIP2, cells were stimulated by growth factors and insulin. SHIP2 was found to incorporate 32P under all conditions (thus reflecting its phosphorylation, Fig. 3). The incorporation of 32P on Ser, Thr and Tyr residues assessed by autoradiography was compared with SHIP2 phosphorylation on tyrosine determined by Western blotting. It was observed that EGF stimulated phosphorylation on the tyrosine residues of SHIP2 more potently than insulin (Fig. 3). Thus, the fraction of tyrosine phosphorylated-SHIP2 depends on the agonist used to stimulate the PI 3-kinase. Eight phosphosites on Tyr, Ser and Thr residues have been identified in COS-7 cells overexpressing SHIP2, and three (S132, T1254 and S1258) in human astrocytoma cells at endogenous levels of SHIP2 (Fig. 3).

Figure 3.

Evidence of SHIP2 phosphorylation on Ser/Thr and Tyr residues. A: SHIP2 transfected COS-7 cells were maintained for 24 hours in serum-free phosphate-depleted medium and incubated for 4 hours in the presence of 500 µCi 32P. SHIP2 was immunoprecipitated, and phosphorylated SHIP2 was identified by autoradiography and immunodetection. Tyrosine phosphorylated SHIP2 was detected using a 4G10 (anti-phosphotyrosine) monoclonal antibody. Cells were treated for 10 minutes with various stimuli, including okadaic acid (OA), a vanadate stable analogue bpv(Phen), insulin, EGF and PDGF. B and C: SHIP2 phosphosites were identified by mass spectroscopy (reproduced from ref. 22 with permission; © the Biochemical Society).

It has been shown in astrocytoma cells that phospho SHIP2 on S132 concentrates in nuclear speckles 22. In the same cells, phospho SHIP2 S132 immunoreactivity shows an overlap with PtdIns(4,5)P2 staining in speckles. The data thus suggest that nuclear SHIP2 may control nuclear PtdIns(4,5)P2 levels 22. This could have several interesting consequences as nuclear PIs, particularly PtdIns(4,5)P2 play important roles in signalling. For example, type I phosphatidylinositol phosphate kinases (PIPKIs) synthesize PtdIns(4,5)P2 by using PtdIns4P as substrate. One of the isoforms, PIPKIα, targets to membrane ruffles and nuclear speckles, where it interacts with a nuclear speckle targeted PIPKIα-regulated poly(A) polymerase 42. This protein complex triggers a mechanism that regulates gene expression under the control of nuclear PtdIns(4,5)P2 42. Thus, it is possible that this mechanism of selective control of mRNAs is controlled by nuclear phospho SHIP2 S132 and perhaps also by other PtdIns(4,5)P2 phosphatase(s).

Phosphorylation of SHIP2 on S132 also has an impact on its stability, as phospho SHIP2 appears less stable than the non-phosphorylated enzyme 22. It is noteworthy that PTEN stability is also controlled by phosphorylation (by casein kinase II and glycogen synthase kinase 3) on Ser/Thr residues within its unstructured C-terminal end and this could have a major influence on tumour progression 43. Likewise, Plk3, a member of the Polo-like kinase family (involved in regulating the cell cycle checkpoint control in response to genotoxic stresses), phosphorylates PTEN at two C-terminal residues, resulting in an increased stability of this protein 44.

In summary, SHIP2 phosphorylation on Ser/Thr residues has been identified in proteomic screening, and at the endogenous level in astrocytoma cells. Phospho SHIP2 S132 (which could be confirmed and visualized using a phospho-specific antibody) is enriched in the nucleus, particularly in nuclear speckles, as compared to total SHIP2. Phospho SHIP2 S132 in speckles may control the levels of PtdIns(4,5)P2 and be involved in the control of gene expression or protein stability. Other Ser/Thr SHIP2 phosphorylation sites have also been determined 22, so far with unknown regulatory functions. In addition to possible functions in targeting and stability, it is possible that SHIP2 phosphorylation could also play a role in mitosis as suggested from a large-scale phosphoproteome analysis and the involvement of proline-directed kinases 45.

Regulating SHIP2 Ser/Thr phosphorylation: SHIP2 interaction with PP2A

SHIP2 S132 phosphorylation is within a consensus phosphorylation site for casein kinase II 46, suggesting that this protein kinase may be responsible for phosphorylating SHIP2. This hypothesis is being tested by ongoing work in our laboratory. Recently, the Ser/Thr phosphatase PP2A PR130/B″α1 subunit has been shown to bind SHIP2 (constitutively) in several cell lines 47. As casein kinase II shows high constitutive activity, it is tempting to speculate that PR130 interaction with SHIP2 would provide a mechanism for reversing SHIP2 Ser/Thr phosphorylation. In this respect, two Ca2+-dependent PP2A activity-regulating mechanisms have been described for this PP2A subunit class 48–50, which may have the potential to tilt the kinase-phosphatase balance in a particular direction, eventually determining the net phospho Ser/Thr status of SHIP2. Loss of function experiments of PR130 and manipulation of intracellular Ca2+ levels may allow researchers to address this question. It is noteworthy that PR130-PP2A also interacts with ectopically expressed SHIP1 47, and a B56-type (also called PR61/B′-type) PP2A subunit was identified as part of the PTEN interactome in lymphoma cells 51, suggesting that PP2A may be more broadly involved in controlling Ser/Thr phosphorylation of all SHIP1/2 and perhaps even PTEN. This is of particular relevance in light of the well-described tumour suppressive functions of these protein phosphatases 52.

Interestingly, SYNJ1 can be phosphorylated by cyclin-dependent kinase 5 on S1144 and dephosphorylated by calcineurin (also called PP2B) 38. Phosphorylation decreases the activity and reduces the interaction of SYNJ1 with endofilin and amphiphysin in pull-down experiments. Thus, the data point to a cycle of Ser/Thr phosphorylation/dephosphorylation similar to that proposed for SHIP2.

Interactome of SHIP1/2 and its influence on function and signalling

In addition to harbouring lipid phosphatase catalytic activity, SHIP1/2 are also docking proteins for a large number of interactors, i.e. receptors, adaptors, cytoskeletal or focal adhesion proteins, protein kinases and protein phosphatases (Table 2). As shown for PTEN, SHIP1/2, independently of its phosphatase activity, could be acting as a scaffold protein. For example, SHIP1 appears to be required for platelet contractility and thrombus organization. In this model, SHIP1 may act on the actin cytoskeleton organization both via its docking properties and via its PtdIns(3,4,5)P3 5-phosphatase activity 53.

Table 2. SHIP2-interacting proteins
ReceptorsAdaptors/regulatorsCytoskeletal proteinsProtein phosphatases/PI phosphatase
  1. JIP-1, JNK-interacting protein 1; PTP1B, protein-tyrosine phosphatase 1B; CD2AP, CD2-associated protein; SH3YL1, SH3 domain-containing Ysc84-like 1; PDGFR, platelet-derived growth factor receptor; EGFR, epidermal growth factor receptor; M-CSFR, macrophage colony-stimulating factor receptor; cMET, hepatocyte growth factor (HGF/SF) receptor; EphrinA2R, ephrinA2 receptor; ARAP-3, Arf GAP, Rho GAP, ankyrin repeat and PH domains.

EGFR 12Shc 69c-Cbl 75PTP1B 81
PDGFR 28p130Cas 70Vinexin 76PR130 47
FcγRIIB 64p85α 62Filamin 77SHIP1 82
FcγRIIA 65APS 71Actin 77 
M-CSFR 66DOK-1 72Intersectin1 78 
cMET 67JIP-1 73ARAP-3 79 
ABL 62CD2AP 74CIN-85 80 
EphrinA2R 68SH3YL1 54Lamin A/C 22 

Could phosphorylation on Ser/Thr of SHIP1/2 be an important molecular switch in protein:protein interaction? This possibility is very likely as multiple SHIP1/2 phosphosites are found in close proximity to the proline-rich sequences mediating interaction with other proteins: SH3 domain-containing Ysc84-like 1 (SH3YL1) is a protein that regulates dorsal ruffle formation, cell polarity and internalization of cell surface receptors 54. It interacts with the proline-rich sequence of SHIP2 located N-terminally in the sequence (residues 139–144 of human SHIP2). It is possible that nearby phosphorylation of S132 could modulate this interaction. This could be related to the extent of SHIP2 S132 phosphorylation, which does vary in different cells and tissues 22. This hypothesis could imply a role of phospho SHIP2 in ruffle formation.

In human astrocytoma cells, SHIP2 interactors also include nuclear proteins such as the nucleoskeleton protein lamin A/C, which plays multiple roles in signalling, nuclear shape and chromatin remodelling. As SHIP2 is part of the lamin interactome, it is possible that S132 phosphorylation could influence lamin nuclear function and perhaps related laminopathies 55.

A model of SHIP2 targeting to the plasma membrane or to the nucleus

A model of SHIP2 regulation by phosphorylation is proposed in Fig. 4. SHIP2 phosphorylation on S132 occurs in the cytoplasm. This allows the phosphorylated and the non-phosphorylated forms of SHIP2 to interact with a series of different proteins. SHIP2, surrounded by a network of proteins, would then be translocated to the plasma membrane where it could control the levels of PtdIns(3,4,5)P3. SHIP2 phosphorylated on S132 would be preferentially translocated to nuclear speckles, where it could control the levels of PtdIns(4,5)P2 and nuclear properties. This simplified model does not take into account the tyrosine phosphorylation of SHIP2 (clearly observed in some cells in response to EGF 6), which could also shift an equilibrium between proteins, initiating or influencing targeting to, for example the plasma membrane.

Figure 4.

A putative model of SHIP2 targeting in astrocytoma cells. It is proposed that SHIP2 is phosphorylated to phospho SHIP2 (pSHIP2) S132 in the cytoplasm. This induces a conformational change that influences the SHIP2 (and pSHIP2) interactome. In a second step, targeting could occur at the plasma membrane or in the nucleus. In astrocytoma cells, pSHIP2 S132 immunoreactivity is never detected at the plasma membrane. In contrast, it is found in the nucleus and nuclear speckles where it could dephosphorylate PtdIns(4,5)P2. PR130 (PP2A PR130/B″α1) can be found in the cytoplasm, the nucleus and at the cell membrane (only transiently, upon EGF stimulation 47). It is not known whether PR130-PP2A associates with SHIP2 in its unphosphorylated form or whether Ca2+ is involved in the interaction. The association between pSHIP2 S132 and nuclear proteins (i.e. lamin A/C and PR130-PP2A) may be critical in targeting SHIP2 to the nucleus 22.


SH2 domain-containing inositol 5-phosphatases 1/2 are important signalling enzymes implicated in human diseases, particularly, but not exclusively, cancer and diabetes 13, 56, 57. Therefore, the control of the SHIP1/2 gene and SHIP1/2 protein expression and activity are important mechanisms to consider. In addition, the control of SHIP1/2 subcellular localization has not yet been solved and this could be another step in regulation. This is particularly interesting as SHIP1/2 have been shown to be present not only in cytoplasmic and perinuclear regions in the absence of stimulation but also in the nucleus. The relocation of a minor fraction of SHIP2 to the plasma membrane close to PtdIns(3,4,5)P3 only occurs in the presence of EGF, PDGF or insulin. In the case of PTEN, myosin Va controls PTEN transport to the plasma membrane 58. The myosin Va:PTEN interaction depends on the phosphorylation status of PTEN on Ser/Thr residues mediated by casein kinase II and/or glycogen synthase kinase 3. We hypothesize that the phosphorylation status of SHIP1/2 also regulates targeting mechanisms probably through the interaction with key proteins. As for PTEN, phosphorylation of SHIP1/2 may be required for these interactions. The fact that phospho SHIP2 S132 is found in the nucleus, in nuclear speckles, and interacts with nuclear proteins such as lamin A/C provides a framework in support of this hypothesis. We propose that Ser/Thr phosphorylation of SHIP1/2, as opposed to Tyr phosphorylation, is particularly important in targeting mechanisms.

It is of interest that one subunit of PP2A PR130/B″α1 interacts with SHIP2, suggesting that Ser/Thr phosphorylation of SHIP2 can be reversed. Finally, the fact that PR130 has three nuclear localization signals and can be seen in the nucleus could explain why part of the SHIP2 is recovered in the nucleus. In future, it will be important to study the roles of the different SHIP1/2 phosphosites that have so far not been investigated and explore the consequences of SHIP1/2 targeting to the nucleus on gene expression, proliferation or differentiation.


This work was supported by a grant from the Fonds de la Recherche Scientifique Médicale (FRSM) and from the Interuniversity Attraction Poles Program (P6/28) – Belgium State – Belgian Science Policy to C.E., E.W. and V.J., W.E. is supported by a Télévie fellowship. We would like to thank Dr. Gilbert Vassart for reading the manuscript.