Cutting Edge: Feed-Forward Activation of Phospholipase Cγ2 via C2 Domain–Mediated Binding to SLP65

Engagement of the BCR and activation of spleen tyrosine kinase Syk triggers phosphorylation of the adaptor protein SLP65 (1) (or BLNK) (2). Phospho-SLP65 nucleates the assembly of downstream effectors into larger complexes to control a plethora of signaling pathways (3). For mobilization of the Ca²⁺ second messenger, SLP65 simultaneously recruits phospholipase Cγ2 (PLCγ2) and Bruton’s tyrosine kinase (Btk) to distinct phosphotyrosine (pTyr) residues by virtue of the enzymes’ Src homology (SH) 2 domains (4–6). This structural organization allows Btk to phosphorylate and activate PLCγ2. The model and its role for B cell activation is supported by mutational analyses and biochemical approaches in various experimental systems (7). However, SLP65 phosphorylation lasts longer than intracellular Ca²⁺ fluxes, suggesting that the activity of PLCγ2 or its inducible association to SLP65 is subject to further regulation. Moreover, the earliest and most prominent pTyr site of avian SLP65, tyrosine 138 (Y¹³⁸) (8), is not involved in SH2 domain binding (4) but is conserved between SLP65 orthologs and embedded within the consensus sequence EYΦDN (single-letter code for amino acids with Φ representing A, V, or I). A similar motif can be found in the T cell paralog SLP76 (3), albeit at a different position (i.e., Y¹⁷³ in murine SLP76) and without a negatively charged residue N terminal to the tyrosine. Phosphorylation of Y¹⁷³ in SLP76 by the Btk-related kinase Itk is pivotal for PLCγ1 activation in T lymphocytes and mast cells in vitro and in vivo (9). The mechanism of phospho-Y173 signaling remained elusive. In this article, we show that the corresponding EYΦDN motif of SLP65 critically contributes to B cell activation. The underlying mechanism is unusual. Unlike canonical pTyr–SH2 interactions, the phosphorylated EYΦDN motif provides a selective and Ca²⁺-regulated docking site for the C2 domain of PLCγ2. This finding uncovers a feed-forward activation loop in that binding to phosphorylated SLP65 is primed by the PLCγ2 SH2 domains and further stabilized by the C2 domain upon initial release of Ca²⁺. The transient nature of Ca²⁺ fluxes eventually limits the C2–pTyr interaction, which provides negative feedback regulation to PLCγ2 during later phases of B cell stimulation.

DT40 and Ramos B cells have been described previously (8, 10). Primary human B cells purified via the Isolation Kit II (Miltenyi) were stimulated with 10 μg/ml goat anti-human IgM (Southern Biotechnology). Abs recognizing the phospho-Y¹¹⁹ motif were raised against the peptide ¹¹³PFARGE(pY)IDNRS¹²⁴ (Perbio Science) and immobilized with AminoPlus Kit (Perbio Science). Abs against GFP or pTyr (4G10), and recombinant PLCγ were purchased from Roche and Millipore, respectively. Ratiometric Ca²⁺ concentrations were measured by flow cytometry (8, 10). Confocal laser-scanning microscopy was conducted on a Leica SP2 system.

Vectors encoding citrine- or STrEP-tagged SLP65 have been described previously (8, 10). The cDNA of rat PLCγ2 was inserted into pEGFP-N1 (BD Biosciences), and the PLCγ2-EGFP coding fragment was ligated into pcDNA3 (Invitrogen) containing a puromycin resistance cassette. The C2 mutant of PLCγ2 generated by overlap extension PCR lacks aa M¹⁰⁴³ to V¹²³⁷. Indicated point mutations were introduced by site-directed mutagenesis (Quickchange, Stratagene). Transduced or transfected cells were selected with puromycin at 1 mg/ml.

Stable isotope labeling with amino acids in cell culture (SILAC) as well as mass spectrometric identification and quantification of phospho–acceptor sites or SLP65 ligands have been described (8, 10). In brief, “heavy” SILAC medium contained ²D₄,¹²C₆,¹⁴N₂-lysine and [¹³C]₆,¹⁴N₄-arginine, whereas “light” medium contained ¹²C₆,¹⁴N₂-lysine and ¹²C₆,¹⁴N₄-arginine. Proteins bound to immobilized peptides [¹¹³PFARGE(p)YIDNRS¹²⁴] or STrEP-tagged SLP65 were identified on an Orbitrap Velos instrument. Specific interactors were represented by ≥3 unique peptides at a “heavy-to-light” ratio >5.

We previously determined the phosphorylation dynamics of avian SLP65 in DT40 chicken B cells and identified Y¹³⁸ within a conserved EYΦDN sequence (see Supplemental Fig. 1A) as major and immediate early pTyr site (8). We now repeated these studies with primary human B cells. The mass spectrometric profiles in Fig. 1A show that the corresponding site of human SLP65, Y¹¹⁹, was likewise robustly phosphorylated early after BCR stimulation. To determine the upstream kinase, we generated phospho–site-specific Abs (anti-pY¹¹⁹) and expressed human SLP65 or, as control, a Y¹¹⁹F variant in wild-type DT40 B cells or mutant cells that lacked Lyn, Syk, or Btk. Anti-pY¹¹⁹ immunoblotting detected human SLP65 from activated but not from resting wild-type B cells (Fig. 1B, lanes 1, 2). The Y¹¹⁹F variant was not detected at all (lanes 3, 4). Inducible phosphorylation of Y¹¹⁹ was also observed in Lyn-deficient B cells (lanes 5–8). Notably, no other Src family kinase is expressed in these cells. The absence of Syk abrogated Y¹¹⁹ phosphorylation (lanes 9–12), suggesting that Syk phosphorylates Y¹¹⁹ either directly or through the activation of the downstream kinase Btk. In support of the former possibility, inducible phosphorylation of wild-type SLP65 was still observed in btk^−/− B cells (lanes 13–16). Furthermore, Y¹¹⁹ was also phosphorylated in SLP65 variants that lacked the Btk-binding site (Y⁹⁶F) or that harbor Y¹¹⁹ as the only tyrosine residue (Fig. 1C). These results strongly suggest that Y¹¹⁹ is a direct substrate of Syk in cultured and primary human B cells. The phosphorylation reaction is independent of other pTyr. By contrast, the corresponding site of SLP76 becomes phosphorylated by Btk-related Itk in a hierarchical manner only upon preceding phosphorylations on other sites by Syk-related ZAP70 (9). The independence of Y¹¹⁹ phosphorylation of a “primer kinase” is most likely caused by the N-terminal glutamate, which is absent in SLP76. Negatively charged amino acids guide tyrosine kinases to a C-terminally located substrate site (11). The 4G10 anti-pTyr Ab weakly recognizes phospho-Y¹¹⁹ (data not shown) so that it was missed in an earlier study (4).

The signaling function of the phospho-Y¹¹⁹ was assessed by flow cytometric recording of BCR-induced Ca²⁺ fluxes. Primary murine B cells from wild-type or SLP65-deficient mice that were retrovirally transduced to express wild-type human SLP65 responded more effectively than Y¹¹⁹F-expressing transductants (Fig. 1D, black and gray curves, respectively). The Ca²⁺ profile of the latter B cells was similar to that of slp65^−/− control cells (red curve). Likewise, SLP65-deficient DT40 B cells that were reconstituted with wild-type versions of human or avian SLP65 showed markedly increased Ca²⁺ fluxes compared with those B cells that expressed the species-matched Y-to-F variant (Supplemental Fig. 1B). Hence, SLP65 orthologs that cannot be phosphorylated at the EYΦDN motif fail to support a proper Ca²⁺ response in B cells from different species. This deficit translated into compromised activation of JNK and ERK as determined by phospho–site-specific immunoblottings (Supplemental Fig. 1C, 1D). Next, we expressed SLP65 variants as citrine fusion proteins to directly monitor the impact of Y¹¹⁹ phosphorylation on SLP65 recruitment from the cytosol to the activated BCR by confocal laser-scanning microscopy. This process is a critical step for SLP65 phosphorylation and signaling (10). The images of Fig. 1E show that both wild-type and Y¹¹⁹F mutant SLP65 translocated equally well to the plasma membrane on BCR ligation. Furthermore, Syk became activated with the same efficacy in both types of B cells as well as in slp65^−/− mutants (Supplemental Fig. 1C, 1D). Our results reveal that the major pTyr site of SLP65 in avian and human orthologs is dispensable for the subcellular navigation of SLP65, but mandatory for proper activation of BCR signal cascades.

To decipher the mechanism of how Y¹¹⁹ phosphorylation boosts SLP65 signaling, we used a 2-fold mass spectrometric approach. First, we screened for putative ligands of phospho-Y¹¹⁹–encompassing peptides in vitro. Second, we conducted a differential interactome analysis of wild-type versus Y¹¹⁹F mutant SLP65 in vivo (“reverse proteomics”) (10). For peptide-based screening of ligands, Ramos B cells were metabolically labeled in “heavy” or “light” SILAC culture medium. Subsequently, lysates of the heavily labeled cells were subjected to affinity purification with the immobilized SLP65 phosphopeptide ¹¹³PFARGE(pY)IDNRS¹²⁴, whereas the unphosphorylated counterpart was incubated with lysates of lightly labeled cells. Purified proteins were identified and relatively quantified by liquid chromatography–coupled tandem mass spectrometry (LC-MS/MS). For evaluation of the results, we determined the enrichment factor with which a given protein was selectively purified via the phosphorylated SLP65 peptide by calculating the ratio of the heavily versus the lightly labeled peptide species from that ligand. Then, the individual ratios were plotted against the ligand’s identification intensity that is represented by the number of different peptides derived from an individual protein (Fig. 2A). Except for one ligand, all of the obtained proteins showed a “heavy-to-light” ratio ≈1, demonstrating that they bound the unphosphorylated and phosphorylated SLP65 peptide with the same intensity irrespectively of how many peptides from those ligands were identified. The exception was PLCγ2, which was up to 20-fold enriched by the phosphorylated form of the Y¹¹⁹ peptide. This strongly suggested PLCγ2 to be an exquisite ligand of phospho-Y¹¹⁹. To explore that possibility in vivo, and moreover in an unbiased manner, we asked whether the loss of Y¹¹⁹ phosphorylation affects the overall composition of the SLP65 interactome in stimulated B cells. We therefore quantitatively compared the interactomes of wild-type and Y¹¹⁹F mutant SLP65 by SILAC-based mass spectrometry (“reverse proteomics”) (10). The amount of a given ligand that was purified with the Y¹¹⁹F variant was normalized to that obtained with wild-type SLP65 (Fig. 2B). Consistent with our peptide screening approach, only PLCγ2 showed an altered SLP65 association ratio that was decreased by ∼50% for the Y¹¹⁹F mutant. Fig. 2C shows the functional consequence of the Y¹¹⁹F exchange for BCR-induced PLCγ2 phosphorylation. It was barely detectable in B cells expressing the Y¹¹⁹F variant. In summary, the major pTyr site of SLP65 controls the recruitment and subsequent activation of a single effector protein, namely, PLCγ2.

Putative PLCγ2 docking modules for phospho-Y¹¹⁹ are the lipase’s two SH2 domains. However, their consensus binding motif (11) markedly differs from EYΦDN. In fact, they recognize three YXXP type phosphorylation sites in SLP65 (4). We therefore asked whether phospho-Y¹¹⁹ binds other PLCγ2 modules such as the C2 domain, which has been implicated in BCR-induced Ca²⁺ fluxes (12). C2 domains were described first as conserved region number two in classical protein kinase C isoforms, and are classical lipid-binding moieties. However, the C2 domains of protein kinase Cδ and θ can bind pTyr-based peptides (13, 14). We used the immobilized SLP65 peptides to purify proteins from DT40 transfectants expressing either wild-type PLCγ2 or a truncated version that lacks the C2 domain (PLCγ2ΔC2). As control, we used plcγ2^−/− parental cells. Fig. 3A shows that wild-type PLCγ2, but not PLCγ2ΔC2, complexed with the phosphorylated SLP65 peptide (upper panel, lanes 3, 4 and lanes 5, 6, respectively). Binding strongly increased in the presence of Ca²⁺ (upper panel, lanes 3, 4), which may induce conformational changes that tighten the interaction as reported for the homologous C2 domain of PLCδ (15). No PLCγ2 signal was detected in the negative control (upper panel, lanes 1, 2) or in the material that was purified with the unphosphorylated peptide (middle panel). Incubation of the peptides with different concentrations of recombinantly expressed PLCγ2 showed that the binding is direct (Fig. 3B, upper panel). In this approach, the positive influence of Ca²⁺ ions was obvious at PLCγ2 concentrations <100 pg/ml, which is in accordance with physiological conditions. Interestingly, no binding was observed between SLP65 peptides and the γ1 isoform of PLC (Fig. 3B, lower panel). Not only does the latter result demonstrate the specificity of the experimental assay, it also confirms our proteomic results that indicated a highly restricted function of the phospho-Y¹¹⁹, namely, to promote the recruitment of PLCγ2. Furthermore, B cells that express PLCγ2ΔC hardly mobilized Ca²⁺ upon BCR ligation (Fig. 3C), which is consistent with the almost blunted Ca²⁺ response of Y¹¹⁹F-expressing cells (see earlier). Collectively, the data reported in this article uncover a specific and unusual interaction between phospho-Y¹¹⁹ and the C2 domain of PLCγ2. This is mandatory to stabilize the SH2-initiated complex formation with SLP65 for proper Ca²⁺ flux. The Ca²⁺ sensitivity of the C2 interaction has two consequences. Early after BCR ligation, an increase of cytosolic Ca²⁺ boosts SLP65/PLCγ2 complex formation, whereas the later adjustment of Ca²⁺ to baseline concentrations weakens that interaction, which may contribute signal termination independently of phosphatases. Hence balanced PLCγ2 activation is controlled by a single phosphorylation event in SLP65. This ensures a rapid B cell response on the one hand, and a transient action of PLCγ2 for limited B cell stimulation on the other hand.

We thank Ines Heine and Tabea Testa for excellent assistance and Dr. Michael Reth for providing slp65^−/− mice.

The authors have no financial conflicts of interest.