Discussion

thereby its recruited signaling complex. This emerged from results of a model system established to study the FcyRIIb-SHIP complexes using combined sequences of the tyrosine phosphorylated motif of SHIP and the FcyRIIb p-ITIM. Binding of this bisphosphopeptide to SHP-2 was analysed and compared with that of the monophosphopeptides. P-ITIM peptide bound SFIP-2 and SHIP SH2 domains at nanomolar range, (K

=104 nM and 40 nM, respectively), while the apparent affinity constant for SHP-1 was 2 mM. This is in agreement with previous findings, suggesting that SHP-1 binds to p-ITIM with the lowest K

as compared to SHP-2, and SHIP, furthermore, with data showing that SHP-1 is dispensable for FcyRIIb dependent B cell inhibition (33, 36 ). Since p-ITIM bound complexes contain tyrosine phosphorylated proteins, including SHIP and She, we supposed that these molecules may also interact with the SFI2 domains of SHP-2. Indeed, the SHIP and She phosphopeptides bound to SHP-2 SH2 domains with a 60-300 nM and to SHIP SH2 with 14-116 nM affinity constants. Interestingly, SHIP phosphopeptides also bound to the autologous SH2, suggesting that phosphorylated SHIP might inhibit the interactions of its own SH2 domain.

In spite of the high affinity binding of recombinant SHP-2 SH2 domains to the biosensor coated with SHIP and She phosphopeptides, the same molecules did not bind native SHP-2 from the cell lysates suggesting that tyrosine phosphorylated motifs of SHIP and She cannot bind to the free C-terminal SFI2. Flowever, they may react with the N-terminal SH2 domain of SHP-2, which becomes accessible in the native molecule only after the occupation of the C-terminal SFI2 domain by the phosphopeptide ligand, such as p-ITIM.

Crystallographic data have shown that the activation of SHP-2 is initiated by the phosphopeptide binding of C SH2, localizing the second tyrosine containing motif in the proximity of N-terminal SI-I2 domain (37). These phosphorylated ligands then would achieve

15

a sufficient local concentration to compete for the binding to N SH2, thus liberating the catalytically active site. Recent data have also shown that the two SH2 domains of SHP-2 are widely spaced and oppositely oriented, thus a ligand with a spacer larger than 50 intervening amino acids between the two tyrosines would be suitable for the bidental binding (38). The latter authors concluded that the two activating phosphotyrosines do not have to be located on the same protein. Thus, we expected that following the binding of p-ITIM to the C-terminal SH2 domain of SHP-2, an other phosphotyrosine containing protein(s) in the same p-ITIM associated complex maybe in a position to react with the N-terminal domain and activate the phosphatase. The results of the SHP-2 activity assays have shown that indeed, this is the case:

phosphatase activity of affinity purified SHP-2 was approximately 10 times higher when measured using the bisphosphopeptide, as compared to the monophosphopeptide substrates.

On the basis of these results, we suggest that the FcyRIIb p-ITIM bound multimolecular complex, containing multiple tyrosine phosphorylated motifs, is able to activate SHP-2 via interacting sequentially with both the C- and the N-terminal SH2 domains. Thus FcyRIIb-SHIP complexes may switch the adaptor function of SHP-2 into a catalytically active phosphatase function.

The adaptor protein She when phosphorylated binds Grb2 via its SH2 domain and translocates it to the cell membrane, where Grb2 interacting with SOS, induces GDP-GTP exchange of ras (39). Monitoring the molecules isolated from the cell lysate samples by the biotinylated bisphosphopeptide, we have found that SHIP-p-ITIM precipitated a significant amount of Grb2. Binding of Grb2 to SHIP-p-ITIM was inhibited by the N-terminal monophosphopeptide of She, and SHIP as well. Since the N-terminal phosphopeptide of SHIP bound Grb2, while only the C-terminal SHIP phosphopeptide bound She, we suggest that SHIP-Grb2 as well as Shc-Grb2 complexes may interact with p-ITIM.

16

The Grb2 binding docking/scaffolding protein, Gabl was also suggested earlier to play role in the Grb2 translocation to the inner leaflet of the cell membrane (10). When tyrosine phosphorylated upon BCR clustering, Gabl translocates to the plasma membrane and recruits SHP2, Grb2, SHIP, She, and PI3-K to the vicinity of their substrates, regulating thereby the activity of these proteins (10,40). We detected Gabl in samples isolated by p-ITIM, SHIP-p-ITIM, N-terminal SHIP and She peptides and also identified Gabl in immunoisolated FcyRIIb samples obtained from BCR activated and BCR- FcyRII co-clustered cells. Since Gabl has no SH2 or PTB domains the results indicate that these interactions may be mediated by the free SH2 domain of She, SHP-2 or the p85 subunit of PI3-K.

Gab family docking proteins are known as a potential substrates of SHP-2 (40). When tyrosine phosphorylation of the Gabl was monitored after an extended time of BCR-FcyRIIb co-clustering, a marked decrease in tyrosine phosphorylation of a 110 kDa protein, corresponding to Gabl was observed. This result indicates that Gabl is dephosphorylated, at

t

least partially in BCR-FcyRIIb co-clustered B cells. In order to control whether SHP-2 is indeed responsible for dephosphorylating Gabl, we have employed synthetic peptides corresponding of three tyrosine phosphorylated motifs of Gabl as possible substrates of affinity isolated SHP-2. All phosphopeptides bound SHP-2, though to a different extent, and one of them containing the tyrosine 626, was completely dephosphorylated by SHP-2 suggesting that this peptide efficiently activated the phosphatase.

The tyrosine phosphorylated state of Gabl is a pre-requisite for binding SH2 domain containing proteins. PI3-K binds to Gabl in BCR activated cells via the SH2 domains of the p85 regulatory subunit (40). Membrane targeting was shown to be sufficient for PI3-K activation and triggering multiple signal transduction pathways (19, 41). Gabl is recruited to the cell membrane via its PFI domain binding specifically to PIP

, a product of PI3-K. Thus

17

PI3-K apparently functions as both an upstream modulator and a downstream effector of Gabl (42). We have shown here that PI3-K binding to Gabl is diminished in the BCR-FcyRII co-clustered cells. Interfering with the PI3-K translocation to the cell membrane by dephosphorylating Gabl may result in the inactivation of PI3-K. Thus PIP

level in the cell membrane might be downregulated, resulting in the abolished binding of Gabl PH domain, and the release of Gabl from the membrane. Moreover, since Gabl regulates ras activation via Shc-Grb2-SOS complexes, disruption of the Gabl bound signaling complex may also lead to the inhibition of the ras pathway. We may conclude that Gabl dephosphorylation might be responsible for the previously observed inhibition of both the PI3-K- and the ras-dependent activation pathways in FcyRIIb-BCR co-clustered cells (19,20,43).

The level of PIP

in the cell membrane is regulated by two enzymes: SHIP, which hydrolyses PIP

to yield phosphatidylinositol 3,4 phosphate, and PI3-K, which phosphorylates phosphatidylinositol (4, 5) bisphosphate at position 3, resulting PIP

(34, 44).

Therefore both the recruitment of SHIP to the cell membrane as well as the inactivation of PI3-K may result in a decreased level of PIP

, consequently, an insufficient binding of both Btk and Gabl PH domains. Thus, taken together, our data suggest that SHIP has a double role in course of regulating B cell activation. Beside dephosphorylating PIP

, SHIP co-operates with the phosphorylated FcyRIIb and by activating SHP-2 promotes dephosphorylation of Gabl. This results in the dissociation of Gabl bound signaling molecules, leading to the inactivation of PI3-K, and a reduced Btk and PLCy activity.

18