Two distinct tyrosine-based motifs enable the inhibitory receptor FcgammaRIIB to cooperatively recruit the inositol phosphatases SHIP1/2 and the adapters Grb2/Grap.

Two distinct tyrosine-based motifs enable the inhibitory receptor FcgammaRIIB to cooperatively recruit the inositol phosphatases SHIP1/2 and the adapters Grb2/Grap.. Abstract Fc γ RIIB are low-affinity receptors for IgG that contain an Immunoreceptor Tyrosine-based Inhibition Motif (ITIM) and that inhibit Immunoreceptor Tyrosine-based Activation Motif (ITAM)-dependent cell activation. When coaggregated with ITAM-bearing receptors, Fc γ RIIB become tyrosyl-phosphorylated and recruit the SH2 domain-containing phosphatidylinositol 5’-phosphatases SHIP1 and SHIP2, which mediate inhibition. The Fc γ RIIB ITIM was proposed to be necessary and sufficient for recruiting SHIP1/2. We show here that a second tyrosine-containing motif in the intracytoplasmic domain of Fc γ RIIB is required for SHIP1/2 to be coprecipitated with the receptor. This motif functions as a docking site for the SH2 domain-containing adapters Grb2 and Grap. These adapters interact via their C-terminal SH3 domain with SHIP1/2 to form a stable receptor-phosphatase-adapter trimolecular complex. Both Grb2 and Grap are required for an optimal coprecipitation of SHIP with Fc γ RIIB, but one adapter is sufficient for the phosphatase to detectably coprecipitate with the receptors. In addition to facilitating the recruitment of SHIPs, the second tyrosine-based motif may confer upon Fc γ RIIB the properties of scaffold proteins capable of altering the composition and the stability of signaling complexes generated following receptor engagement. S-Transferase; IL-2, Interleukin-2.


Introduction
FcγRIIB are low-affinity receptors for the Fc portion of IgG antibodies that are widely expressed by cells of hematopoietic origin (1). Their low affinity enables them to remain free in the presence of high concentrations of circulating IgGs and to bind immune complexes with a high avidity. They are unique among Fc Receptors (FcRs) in exhibiting inhibitory properties. Indeed, FcγRIIB were demonstrated to negatively regulate cell activation triggered by other FcRs in mast cells (2), by B Cell Receptors for antigen (BCR) in B cells (3,4) and by T Cell Receptors for antigen (TCR) in T cells (5), i.e. by all receptors containing Immunoreceptor Tyrosine-based Activation Motifs (ITAMs). FcγRIIB must be coaggregated with activating receptors via IgG immune complexes in order to exert their inhibitory effects (2). The in vivo relevance of the regulatory properties of FcγRIIB was ascertained in FcγRIIBdeficient mice. FcγRIIB -/mice were shown to mount enhanced antibody responses (6), to exhibit enhanced IgG-and IgE-induced anaphylactic reactions (7), to be hypersensitive to collagen-induced arthritis (8,9), and to develop spontaneous systemic lupus erythematosus in the C57BL/6 background (10). FcγRIIB are therefore likely to play major roles in the prevention of autoimmune diseases, allergies and other inflammatory diseases.
The regulatory properties of FcγRIIB were shown to depend on the presence of an Immunoreceptor Tyrosine-based Inhibitory Motif (ITIM) in their intracytoplasmic domain.
It is a general consensus that the FcγRIIB ITIM is both necessary and sufficient for inhibition. The conclusion that it is necessary was based on the pioneer work by Amigorena et al. who showed that a 13-aminoacid deletion, which was later understood to encompass the ITIM, abrogated inhibition in B cells (4). A point mutation of the ITIM tyrosine also abrogated FcγRIIB-dependent inhibition of mast cell and T cell activation (5), and abolished (28) or reduced (29) the calcium response in B cells. The conclusion that the ITIM is sufficient was based on works by Muta et al. who showed that a chimeric molecule whose intracytoplasmic domain contained the murine FcγRIIB ITIM retained inhibitory properties in B cells (28). More recently however, we found that a C-terminal deletion of the intracytoplasmic domain of FcγRIIB, which left the ITIM intact, prevented SHIP1 for being detectably coprecipitated, and reduced the inhibitory effect of FcγRIIB on BCR signaling (29). We show here that this C-terminal sequence contains a second tyrosine-based motif that mediates the recruitment, via their SH2 domain, of the adapter proteins Grb2 and Grap which interact, via their C-terminal SH3 domain, with SHIP1 and SHIP2, thus stabilizing the binding of these phosphatases to the FcγRIIB ITIM. Supporting a critical role of this trimolecular complex in vivo, we provide evidence that adapters are necessary for FcγRIIB to recruit phosphatases.
Transfectants were selected and cloned as described (33)(34)(35). Expression of receptors on clones remained stable. Several clones of each transfectant were used and gave similar results.

Indirect immunofluorescence.
To measure the expression of FcγRIIB, cells were incubated with 10 µg/ml 2.4G2 or without, washed and stained with 50 µg/ml FITC-labeled MAR at F(ab') 2 . To measure the expression of BCR, cells were incubated with 10 µg/ml R ab AC or without, washed and stained with 50 µg/ml FITC-labeled GAR ab F(ab') 2 . Fluorescence was analyzed with a FACScalibur (Becton Dickinson, Mountain View, CA). Whole cell lysate analysis. IIA1.6 transfectants were stimulated at 37°C for indicated times with 30 µg/ml intact or 20 µg/ml F(ab') 2 fragments of R at AM IgG, and lysed by 3 cycles of incubation for 1 min in liquid nitrogen followed by 1 min at 37°C in Lysis Buffer pH 8.0 (50 mM Tris pH8, 150 mM NaCl, 1% Tx100, 1 mM Na 3 VO 4 , 5 mM NaF, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 10 µg/ml pepstatin and 1mM PMSF). Proteins were quantitated using a Biorad protein assay (Hercules, CA) and 40 µg of proteins were treated as described in Western blot analysis.

The ITIM is necessary, but not sufficient, for FcγRIIB to recruit SHIP1.
Because we reported previously that a deletion of the 16 C-terminal aminoacids of the transfectants expressing wild-type (wt) FcγRIIB1 were used as positive controls (Fig. 1A). Wt and mutant FcγRIIB1 were coaggregated with BCR using intact Rabbit anti-Mouse Ig (R ab AM) IgG antibodies that can bind both to BCR via their Fab portions and to FcγRIIB1 via their Fc portion. FcγRIIB1 were immunoprecipitated, and immunoprecipitates were Western blotted with anti-FcγRIIB, anti-phosphotyrosine and anti-SHIP1 antibodies. Wt and mutant FcγRIIB1 became tyrosyl-phosphorylated following coaggregation with BCR. Compared to wt FcγRIIB1, FcγRIIB1 mutants were less phosphorylated. The coprecipitation of SHIP1 with phosphorylated wt FcγRIIB1 was lost not only in cells expressing FcγRIIB1 Y309G, as expected, but also in cells expressing FcγRIIB1 Y326F, although the ITIM remained intact in this mutant (Fig. 1B). As observed previously (29), the coprecipitation of SHIP1 and SHIP2, was also lost in cells expressing FcγRIIB1 Δ314 (Fig. S1).
This loss of a detectable coprecipitation of SHIP1 with FcγRIIB1 Y326F was correlated with a loss of inhibition of Erk activation. Erk phosphorylation, induced upon BCR aggregation, was indeed decreased upon coaggregation of BCR with wt FcγRIIB1, but not upon coaggregation of BCR with either FcγRIIB1 Y309G or FcγRIIB1 Y326F ( Fig. 2A).
Akt phosphorylation induced upon BCR aggregation, that was abolished upon coaggregation of BCR with wt FcγRIIB1, was partially inhibited upon coaggregation of BCR with FcγRIIB1 Y309G, FcγRIIB1 Y326F (Fig. 2B), or FcγRIIB1 Δ314 (Fig. S2A). In order to understand how FcγRIIB1 mutants could still inhibit Akt activation to some extent, we analyzed the colocalization of SHIP1 with wt or mutant FcγRIIB1 by confocal microscopy ( Fig. S2B and C). As previously observed (36), SHIP1 colocalized with BCR-wt FcγRIIB1 coaggregates in more than 80% cells. The colocalization of SHIP1 with either BCR-FcγRIIB1 Δ314 or BCR-FcγRIIB1 Y309G coaggregates was reduced, but it could still be observed in 40-50% of the cells.
Altogether, these data indicate that the tyrosine contained in the C-terminal sequence of FcγRIIB1 contributes to the recruitment of SHIP1 and to FcγRIIB-dependent inhibition of Erk and Akt activation.

FcγRIIB contain a second tyrosine-based motif that binds the adapters Grb2/Grap in vitro and recruit these adapters in vivo.
The C-terminal tyrosine of FcγRIIB is within a consensus Grb2-binding site. Indeed, phosphorylated peptides corresponding to the 16 C-terminal aminoacids of FcγRIIB that were deleted in FcγRIIB1 Δ314 (pC-ter), but not the same non phosphorylated peptides (C-ter), precipitated Grb2 and Grap from a IIA1.6 cell lysate, but not the related adapters Nck, Nckβ or CrkL, that were all present in the lysate (Fig. 3A). pC-ter, but not C-ter, bound to GST fusion proteins containing Grb2 or Grap (Fig. 3B). Finally, pC-ter, but not C-ter, also bound to a GST fusion protein containing the SH2 domain of Grb2. This GST-Grb2 SH2 fusion protein failed to bind to a phosphorylated peptide corresponding to the FcγRIIB ITIM (pITIM). Conversely, a GST fusion protein containing the SH2 domain of SHIP1 bound to pITIM, but not to pC-ter (Fig. 3C). These data indicate that pC-ter can bind to the adapters Grb2 and Grap, but not to SHIP1. In vitro binding results from a direct interaction of pC-ter with the two adapters and, at least for Grb2, this interaction is via its SH2 domain.
Based on the above in vitro results, we investigated whether adapter molecules would coprecipitate with phosphorylated FcγRIIB in IIA1.6 cells. FcγRIIB1 phosphorylation was induced either by coaggregating the receptors with BCR using R at AM IgG antibodies or by treating cells with pervanadate. The coprecipitation of SHIP1 varied with the intensity of FcγRIIB1 phosphorylation. Neither Grb2 nor Grap coprecipitated with FcγRIIB1 in untreated cells. Grb2, but not Grap, detectably coprecipitated with FcγRIIB1 following coaggregation with BCR. Both Grb2 and Grap coprecipitated with FcγRIIB1 following pervanadate treatment (Fig. 3D). Phosphorylated FcγRIIB1 therefore recruit the adapters Grb2 and Grap in vivo.
In order to confirm these in vitro data, we examined the coprecipitation of SHIP1/2 with adapter proteins in IIA1.6 cells expressing wt FcγRIIB1. As detected by Western blotting with corresponding antibodies, Grb2, but not Grap, was precipitated by anti-Grb2 antibodies whereas Grap, but not Grb2, was precipitated by anti-Grap antibodies, and comparable amounts of each adapter were precipitated in all conditions. Small amounts of SHIP1 and SHIP2 coprecipitated with Grb2 in unstimulated cells. Higher amounts of both phosphatases coprecipitated with Grb2 following BCR aggregation and even higher amounts following the coaggregation of BCR with FcγRIIB1. Neither SHIP1 nor SHIP2 detectably coprecipitated with Grap in unstimulated cells. Minute amounts of SHIP2 coprecipitated with Grap following BCR aggregation and both SHIP1 and SHIP2 coprecipitated with Grap following the coaggregation of BCR with FcγRIIB1 (Fig. 4C).
Taken together, these results indicate that the two known SH2 domain-containing inositol 5'-phosphatases SHIP1 and SHIP2 can bind in vitro to Grb2 and can associate in vivo with Grb2 and Grap in B cells.

Two tyrosine-based motifs are required for FcγRIIB to recruit either SHIP1 or Grb2.
To determine the respective contributions of the two FcγRIIB1 motifs in the binding of adapter-phosphatase complexes, we constructed an in vitro model of the intracytoplasmic domain of FcγRIIB. The C-ter peptide, phosphorylated or not, and the ITIM peptide, phosphorylated or not, were mixed in variable proportions and a constant amount of the mixture was used to coat agarose beads. These were used to precipitate SHIP1 and Grb2 from IIA1.6 cell lysate (Fig. 5A). pITIM alone, but not ITIM, precipitated SHIP1 and a small amount of Grb2. Conversely, pC-ter alone, but not C-ter, precipitated Grb2 and a small amount of SHIP1. The amount of SHIP1 precipitated by pITIM-coated beads decreased when beads were coated with decreasing amounts of pITIM and increasing amounts of C-ter (left panel), but not when beads were coated with decreasing amounts of pITIM and increasing amounts of pC-ter (right panel). Likewise, the amount of Grb2 precipitated by pC-ter-coated beads decreased when beads were coated with decreasing amounts of pC-ter and increasing amounts of ITIM (middle panel), but it increased when beads were coated with decreasing amounts of pC-ter and increasing amounts of pITIM (right panel). These results indicate that, when present on the same beads, pC-ter could enhance the in vitro binding of SHIP1 to pITIM and that, conversely, pITIM could enhance the in vitro binding of Grb2 to pC-ter.
In order to validate these in vitro observations in vivo, we examined whether the FcγRIIB ITIM contributes to the recruitment of Grb2, as the FcγRIIB C-terminal motif does for the recruitment of SHIP1, following the coaggregation of FcγRIIB1 with BCR in IIA1.6 transfectants. Both Grb2 and SHIP1 failed to coprecipitate not only with phosphorylated FcγRIIB1 Y326F, but also with phosphorylated FcγRIIB1 Y309G (Fig. 5B). Both the ITIM and the C-terminal motifs are therefore necessary for FcγRIIB to cooperatively recruit SHIP1 and Grb2, as a phosphatase-adapter complex.

Grb2 or Grap is required for FcγRIIB to recruit SHIP.
To investigate the respective roles of the two adapters Grb2 and Grap in this cooperative binding, we used the same in vitro model as in Fig. 5A with cell lysates from the chicken B cells DT40. These were wt cells, Grb2-deficient cells, Grap-deficient cells or Grb2and Grap-deficient cells (30). Beads coated with pITIM and pC-ter required lower amounts of pITIM to precipitate SHIP from wt DT40 cell lysate (Fig. 6, right panel) than beads coated with pITIM and C-ter (Fig. 6, left panel). The same was observed in lysate from Grb2deficient cells and in lysate from Grap-deficient cells, but not in lysate from Grb2-and Grapdeficient cells where comparable amounts of SHIP were precipitated by beads coated with pITIM and pC-ter or with pITIM and C-ter (Fig. 6). Either Grap or Grb2 is therefore necessary and sufficient to support the binding of SHIP to the FcγRIIB ITIM.
To confirm these in vitro data, wt and the three deficient DT40 cells were stably transfected with wt FcγRIIB1 (Fig. 7A). FcγRIIB1 were coagregated with the DT40 BCR by Rabbit anti-Chicken Ig (R ab AC) IgG antibodies. In wt DT40 cells, SHIP coprecipitated with FcγRIIB1 upon coaggregation with BCR. Coprecipitation was lost in Grb2/Grap-doubly deficient cells (Fig. 7B). Coprecipitation was retained in Grb2-or Grap-single deficient cells, albeit in lower amount than in wt cells (Fig. 7C). One adapter is therefore necessary and sufficient for FcγRIIB1 to recruit SHIP but both are required for an optimal recruitment.

Discussion
We show here 1) that, in addition to the ITIM, the intracytoplasmic domain of FcγRIIB contains a second tyrosine-based motif that recruits the SH2 domain-containing adapters Grb2 and Grap, 2) that these adapters interact with the inositol-phosphatases SHIP1 and SHIP2 via their C-terminal SH3 domain, 3) that the two tyrosine-based motifs each contribute to the recruitment of both SHIP1 and Grb2 by FcγRIIB, and 4) that adapters are necessary for FcγRIIB to recruit SHIP1.  (28). Inhibition of BCR-mediated IL-2 secretion by this chimera was, however, half that induced by wt FcγRIIB1, and the authors suggested that other sequences, in FcγRIIB1, might be required to maximize inhibition. One also notices that a TCRζ tyrosine residue was present in the construction, in addition to the ITIM tyrosine, as well as another two prolines and four serines, that could potentially recruit cytosolic molecules. The conclusion that the FcγRIIB ITIM is sufficient to account for the inhibitory properties of the receptor may therefore not be as firmly established than it is usually accepted.
The inhibitory properties of FcγRIIB could be accounted for by the ability of the receptor to recruit SHIP1 (23)(24)(25). A role of SHIP2 was also suggested in FcγRIIB-dependent negative regulation of LPS-activated B cells (27). We show here that cytosolic molecules other than the two phosphatases are recruited by phosphorylated FcγRIIB1. These are the two adapter molecules Grb2 and Grap, that bind to the C-terminal motif via their SH2 domain.
The intracytoplasmic domain of FcγRIIB therefore contains two tyrosine-based motifs that bind specifically the SH2 domain of SHIP1 and the SH2 domain of Grb2 respectively. The recruitment of phosphatases, however, required an intact adapter-binding motif and, conversely, the recruitment of adapters required an intact phosphatase-binding motif. These observations could be explained by a cooperative binding of phosphatases and adapters to FcγRIIB1. Supporting this possibility, we found that Grb2 could interact with SHIP1 and SHIP2 via its C-terminal SH3 domain, and that SHIP1 (38) and SHIP2 coprecipitated with Grb2 and Grap in IIA1.6 cells. Coprecipitation of phosphatases with adapters was enhanced following BCR aggregation and further enhanced following coaggregation of BCR with FcγRIIB1. Since the interactions between SH3 domains and proline-rich sequences are not inducible per se, this suggests that phosphotyrosine-dependent interactions may stabilize phosphotyrosine-independent interactions when adapters and phosphatases are brought in proximity within signaling complexes. Conversely, phosphotyrosine-independent interactions may stabilize phosphotyrosine-dependent interactions. Using a model of the FcγRIIB1 intracytoplasmic domain, in which peptides containing the two SH2 domain-binding sites were bound to the same beads, we indeed found that, when phosphorylated, the C-terminal peptide enhanced the binding of SHIP1 to the phosphorylated ITIM peptide, and that conversely, when phosphorylated, the ITIM peptide enhanced the binding of Grb2 to the phosphorylated C-teminal peptide. This reciprocal enhancement of phosphatase and adapter binding suggests that the recruitment of SHIP1 and Grb2 by FcγRIIB1 involves cooperative binding within a trimolecular complex composed of the phosphorylated receptor, the phosphatase and the adapter. This conclusion may not be restricted to the interactions of FcγRIIB1, SHIP1 and Grb2. Indeed, molecules that contain two SH2 domains require the cooperative binding of these two domains to two sequences containing phosphorylated tyrosines in order to be recruited in vivo. Thus, the recruitment of the protein tyrosine kinases ZAP-70 and Syk (39,40), or of the tyrosine phosphatase SHP-1 (41), requires the conservation of their two SH2 domains and the conservation of the two tyrosines of ITAMs in immunoreceptors (42) or of the two ITIMs in Killer cell Inhibitory Receptors (33,43) respectively. Moreover, molecules that contain a single SH2 domain were found to require the cooperation of other SH2 domaincontaining molecules in order to be recruited (44). We wish therefore to propose that one SH2 domain alone may not be sufficient to enable stable interactions between signaling molecules.
Stable interactions between FcγRIIB1 and SHIP1 can be operationally defined as enabling the coprecipitation of the phosphatase with the receptor. Based on our results, such an interaction would require the two SH2-binding motifs in FcγRIIB1 and adapter molecules. These could either be sequestered from nearby signaling complexes or/and contribute to FcγRIIB-derived signals (46). Grb2 associates with a variety of molecules via its N-terminal SH3 domain and, interestingly, Grap associates with only some molecules among Grb2 partners (45). It was recently reported that Ras-dependent T cell proliferation and IL-2 production were enhanced in Grap-deficient mice (47), suggesting that Grap itself could mediate negative regulation. When recruited by FcγRIIB, adapters may thus reinforce inhibition. Finally, our work provides evidence that FcγRIIB may have a more complex function than simply recruiting SHIP. They indeed appear to function as scaffold proteins that modulate the composition of signaling complexes generated by immunoreceptors with which they are coengaged.