Regulation of RasGRP1 by B cell antigen receptor requires cooperativity between three domains controlling translocation to the plasma membrane

Abstract

RasGRP1 is a Ras-activating exchange factor that is positively regulated by translocation to membranes. RasGRP1 contains a diacylglycerol-binding C1 domain, and it has been assumed that this domain is entirely responsible for RasGRP1 translocation. We found that the C1 domain can contribute to plasma membrane-targeted translocation of RasGRP1 induced by ligation of the B cell antigen receptor (BCR). However, this reflects cooperativity of the C1 domain with the previously unrecognized Plasma membrane Targeter (PT) domain, which is sufficient and essential for plasma membrane targeting of RasGRP1. The adjacent suppressor of PT (SuPT) domain attenuates the plasma membrane-targeting activity of the PT domain, thus preventing constitutive plasma membrane localization of RasGRP1. By binding to diacylglycerol generated by BCR-coupled phospholipase Cgamma2, the C1 domain counteracts the SuPT domain and enables efficient RasGRP1 translocation to the plasma membrane. In fibroblasts, the PT domain is inactive as a plasma membrane targeter, and the C1 domain specifies constitutive targeting of RasGRP1 to internal membranes where it can be activated and trigger oncogenic transformation. Selective use of the C1, PT, and SuPT domains may contribute to the differential targeting of RasGRP1 to the plasma membrane versus internal membranes, which has been observed in lymphocytes and other cell types.

Figure 1.

Domain structures of GFP-tagged RasGRP1 proteins used in this study. GEF, guanine nucleotide exchange domain that catalyzes GTP loading of Ras GTPases. EF, EF-hands. PT and SuPT are the plasma membrane targeting regulatory domains identified in this study. The black circle represents prenylation.

Figure 2.

Plasma membrane translocation and activation of RasGRP1 in response to BCR ligation. (A) Unstimulated DT40 cells expressing GFP-tagged RasGRP1 (RG1) were stained with either ER Tracker to mark ER or anti-GM130 to mark Golgi membranes, as described in Materials and Methods. Individual cells showing fluorescence from the GFP-tagged RG1 and either ER Tracker or GM130 staining are shown. (B) DT40 cells transduced with GFP alone or RG1 were untreated (nil) or treated with 5 μg/ml for 15 min. Cells were then fixed and photographed by fluorescence microscopy. The single cells shown are representative of the typical appearance of the population of cells. Histograms of fluorescence intensities along the indicated segments are shown to the top right of each cell image. The edge of the cell is indicated by the vertical line on the segment. The percentages of cells in each population with detectable GFP at the plasma membrane are listed to the bottom right of each cell image. (C) DT40 cells expressing GFP-RBDL were untreated (nil) or treated with 5 μg/ml anti-IgM for 15 min. Cells were then fixed and photographed by fluorescence microscopy. The cells are displayed as described for B. (D) DT40 cells transduced with a retroviral vector expressing HA epitope-tagged wild-type N-Ras, and untransduced (nil) or transduced with RG1, were stimulated with anti-IgM for the indicated times. Activated Ras GTPases were purified by Raf-RBD chromatography and detected by Western blot, by using anti-HA to detect the transduced N-Ras and an anti-Ras antibody to detect endogenous K-Ras and H-Ras. (E) DT40 cells expressing either GFP as a control, RG1, or RG1 with a point mutation that prevents Ras binding (RG1-GEFμ) were treated with anti-IgM for the indicated times, and levels of phosphorylated ERK2 were quantified by Western blot. The numbers below the P-ERK blot are relative quantities of each band. The expression levels of transduced RasGRP1 protein in each sample, determined by Western blot with an anti-GFP antibody, are shown in the lower blot.

Figure 3.

The C1 domain promotes but is insufficient for BCR-induced plasma membrane translocation of RasGRP1. DT40 cells expressing the indicated proteins were untreated (nil) or treated with anti-IgM 5 μg/ml anti-IgM for 15 min. Cells were prepared for fluorescence microscopy, analyzed, and displayed as described for Figure 2B. In this and subsequent figures, the percentages of cells in each population with detectable GFP at the plasma membrane are indicated at the bottom right of each cell image as % pm⁺. (A) DT40 cells expressing RG1 or RG1ΔC1. For the anti-IgM-stimulations, a cell representative of those showing plasma membrane localization of RG1ΔC1 is shown. (B) PLCγ2-deficient DT40 cells expressing RG1 were treated as described for A and prepared for fluorescence microscopy, analyzed, and displayed as described for Figure 2B. For the anti-IgM-stimulation, a cell representative of those showing plasma membrane localization of RG1 is shown. (C) DT40 cells expressing the isolated C1 domain of RasGRP1 or the isolated C1b domain of PKCδ, each with an N-terminal GFP tag. PMA treatment (500 ng/ml) was for 5 min. (D) DT40 cells expressing the tandem C1a + C1b domains of PKCε with an N-terminal GFP tag.

Figure 4.

The region C-terminal to the C1 domain contains the PT and SuPT domains that control BCR-induced translocation of RasGRP1 to the plasma membrane. (A and B) DT40 cells expressing the indicated RasGRP1 constructs were untreated (nil) or treated with anti-IgM for 15 min. Cells were prepared for fluorescence microscopy, analyzed, and displayed as described for Figures 2 and 3. (B) Cells representative of those showing plasma membrane localization of the RasGRP1 constructs are shown. (C) Sequence of the C-terminal region of murine RasGRP1, from amino acid 571 to the C terminus (GenBank accession no. NP 035376). Bases conserved between murine and chicken RasGRP1 are in bold. The potential leucine zipper is underlined. The arrows show the N-terminal boundaries of the indicated RasGRP1 constructs, and the regions deleted in RG1ΔSuPT and RG1ΔPT. (D) DT40 cells expressing RG1 or RG1ΔPT were untreated (nil) or treated with anti-IgM for 15 min. Cells were prepared for fluorescence microscopy, analyzed, and displayed as described for Figures 2 and 3. (E) DT40 cells expressing RG1 or RG1ΔSuPT were untreated (nil) or treated with anti-IgM for 15 min. Cells were prepared for fluorescence microscopy, analyzed, and displayed as described for Figures 2 and 3. In the experiment shown in the bottom panels, BCR stimulation was relatively weak, resulting in only a low level of plasma membrane translocation of RG1.

Figure 5.

PT domain-mediated plasma membrane localization of RasGRP1 in murine B and T cell lines. (A) WEHI-231 B cells expressing GFP or the indicated RasGRP1 constructs were untreated (nil) or treated with 5 μg/ml anti-IgM for 15 min. Cells were prepared for fluorescence microscopy, analyzed, and displayed as described for Figures 2 and 3. (B) DO11.10 T cells expressing GFP or the indicated RasGRP1 constructs were untreated (nil) or treated with 10 μg/ml anti-CD3ε plus 10 μg/ml anti-CD28 for 15 min. Cells were prepared for fluorescence microscopy, analyzed, and displayed as described for Figures 2 and 3. For the anti-CD3ε + α-CD28 treated cells expressing RG1, the displayed cell is representative of those showing detectable plasma membrane localization.

Figure 6.

Plasma membrane targeting by the PT domain occurs independently of PLCγ2, whereas the ability of the C1 domain to counteract the SuPT domain requires PLCγ2. PLCγ2-deficient DT40 cells expressing the indicated RasGRP1 constructs were untreated (nil) or treated with anti-IgM for 15 min. The cells were then prepared for fluorescence microscopy, analyzed, and displayed as described for Figures 2 and 3. Cells representative of those showing plasma membrane localization of the RasGRP1 constructs are shown.

Figure 7.

Activation of RasGRP1 by BCR can occur at the plasma membrane independently of the C1 domain, or at internal membranes independently of the PT domain. (A) The C1 domain is required for RasGRP1 activation by BCR. DT40 cells transduced with GFP as a control or with the indicated RasGRP1 constructs were treated with anti-IgM for the indicated times. Levels of phosphorylated ERK2 were quantified by Western blot, with relative amounts indicated below each sample. Transduced RasGRP1 protein, detected by Western blot with anti-GFP, is shown in the lower blot. (B) RasGRP1 translocates to the plasma membrane in response to BCR ligation when both the C1 and SuPT domains are absent. DT40 cells expressing RasGRP1 with deletion of the C1 and SuPT domains were untreated (nil) or treated with anti-IgM for 15 min. The cells were then prepared for fluorescence microscopy, analyzed, and displayed as described for Figures 2 and 3. (C) RasGRP1 can be activated by BCR when both the C1 and SuPT domains are absent. DT40 cells transduced with GFP as a control or with the indicated RasGRP1 constructs were treated with anti-IgM for the indicated times. Levels of phosphorylated ERK2 were quantified by Western blot, with relative amounts indicated below each sample. Transduced RasGRP1 protein, detected by Western blot with anti-GFP, is shown in the lower blot. (D) ERK2 activation by RG1ΔC-term and RG1ΔC1+SuPT in response to BCR ligation, relative to ERK2 activation by RG1. P-ERK2 levels were measured in DT40 cells expressing RG1, RG1ΔC-term, or RG1ΔC1+SuPT and stimulated with αIgM for the indicated times. The P-ERK2 levels induced by RG1ΔC-term or RG1ΔC1+SuPT are displayed relative to the P-ERK2 levels induced by RG1, calculated as described in Materials and Methods. Results are from four experiments with RG1ΔC-term and three experiments with RG1ΔC1+SuPT. Bars show SD. The p values are for comparison to hypothetical mean of 0 (no activation of ERK2 by the RasGRP1 mutants). (E) RasGRP1 can be activated in the absence of the PT domain. DT40 cells transduced with GFP as a control or with the indicated RasGRP1 constructs were treated with anti-IgM for the indicated times. Levels of phosphorylated ERK2 were quantified by Western blot, with relative amounts indicated below each sample. Transduced RasGRP1 protein, detected by Western blot with anti-GFP, is shown in the lower blot. (F) Membrane localization of RasGRP1 by prenylation. DT40 cells expressing prenylated RasGRP1 were untreated (nil) or treated with anti-IgM for 15 min. The cells were then prepared for fluorescence microscopy, analyzed and displayed as described for Figures 2 and 3. (G) RasGRP1 can be activated by BCR when the C1 and PT domains are absent, if membrane localization is provided by prenylation. DT40 cells transduced with GFP as a control or with the indicated RasGRP1 constructs were treated with anti-IgM for the indicated times. Levels of phosphorylated ERK2 were quantified by Western blot, with relative amounts indicated below each sample. Transduced RasGRP1 protein, detected by Western blot with anti-GFP, is shown in the lower blot.

Figure 8.

The PT domain does not contribute to membrane targeting of RasGRP1 in fibroblasts, and it is not required for activation of RasGRP1 as an oncogene. (A) NIH 3T3 cells expressing RG1 were stained with either ER Tracker to mark ER or anti-GM130 to mark Golgi membranes, as described in Materials and Methods. Individual cells showing fluorescence from GFP-tagged RG1 and either ER Tracker or GM130 staining are shown. (B) The top pictures show localization of the indicated RasGRP1 constructs in transduced NIH 3T3 cells. Subconfluent cells were fixed and photographed by fluorescence microscopy. Representative cells are shown. The RG1/pren construct serves to highlight the plasma membrane. The bottom pictures are low-magnification views of NIH 3T3 cells cultured for 5 d after confluence. High refractility, elongation, high cell density, and loss of contact inhibition are indicative of oncogenic transformation. (C) Localization of the isolated C1 domain and PT (C-termΔ2) domain of RasGRP1 in NIH 3T3 cells. Subconfluent cells were fixed and photographed by fluorescence microscopy. Representative cells are shown.

Figure 9.

Proposed mechanism for BCR-induced translocation of RasGRP1 to the plasma membrane. (A) In unstimulated cells, the C1 domain of RasGRP1 is attracted predominantly to internal membranes, via DAG binding. A hypothetical ligand for the PT domain is present at the plasma membrane but at functionally low levels. Due to the suppressive action of the SuPT domain, binding of RasGRP1 to the PT ligand is minimal, relative to its binding to internal membranes via the C1 domain. (B) BCR ligation induces high levels of functional PT ligand. Binding of this ligand to the PT domain, partially counteracted by the suppressive effect of the SuPT domain, provides an interaction site at the plasma membrane. Once attracted to the plasma membrane by this interaction, RasGRP1 can acquire an additional interaction via binding of its C1 domain to the local pool of DAG generated at the plasma membrane by BCR-coupled PLCγ2. The combination of these two interactions is sufficient to confer stable binding of RasGRP1 at the plasma membrane, despite the persistent capability of the C1 domain to bind to internal membranes.