NOTCH1 Can Initiate NF-κB Activation via Cytosolic Interactions with Components of the T Cell Signalosome

Abstract

T cell stimulation requires the input and integration of external signals. Signaling through the T cell receptor (TCR) is known to induce formation of the membrane-tethered CBM complex, comprising CARMA1, BCL10, and MALT1, which is required for TCR-mediated NF-κB activation. TCR signaling has been shown to activate NOTCH proteins, transmembrane receptors also implicated in NF-κB activation. However, the link between TCR-mediated NOTCH signaling and early events leading to induction of NF-κB activity remains unclear. In this report, we demonstrate a novel cytosolic function for NOTCH1 and show that it is essential to CBM complex formation. Using a model of skin allograft rejection, we show in vivo that NOTCH1 acts in the same functional pathway as PKCθ, a T cell-specific kinase important for CBM assembly and classical NF-κB activation. We further demonstrate in vitro NOTCH1 associates physically with PKCθ and CARMA1 in the cytosol. Unexpectedly, when NOTCH1 expression was abrogated using RNAi approaches, interactions between CARMA1, BCL10, and MALT1 were lost. This failure in CBM assembly reduced inhibitor of kappa B alpha phosphorylation and diminished NF-κB-DNA binding. Finally, using a luciferase gene reporter assay, we show the intracellular domain of NOTCH1 can initiate robust NF-κB activity in stimulated T cells, even when NOTCH1 is excluded from the nucleus through modifications that restrict it to the cytoplasm or hold it tethered to the membrane. Collectively, these observations provide evidence that NOTCH1 may facilitate early events during T cell activation by nucleating the CBM complex and initiating NF-κB signaling.

Keywords: CARMA1; NF-κB; NOTCH1; PKCθ; T cell subject category: immunology; cytosolic; non-canonical; signal transduction.

Figure 1

NOTCH1 and PKCθ act in the same functional pathway to mediate allograft rejection. We monitored the influence on rejection kinetics of murine skin allografts by recipient mice deficient for various proteins. Grafts were visually scored for signs of rejection using a scale of 1–10, with 1 being completely rejected. An allograft was considered fully rejected when it was >80% necrotic. Graphs represent allograft score (y axis), with a score of “1” being fully rejected and a score of “10” being fully accepted versus time in days after skin grafting (x axis). Wild-type BALB/c skin was grafted onto (A) wild-type C57BL/6 mice (BALB/c → BL6), which served as a baseline for graft rejection for subsequent allograft combinations; (B) p50^null mice (BALB/c → p50^null); (C) PKCθ^null mice (BALB/c → PKCθ^null); (D) NOTCH1 conditional knock-out (N1KO) mice (BALB/c → N1KO); (E) wild-type C57BL/6 mice whose NOTCH signaling was abrogated by administering γ-secretase inhibitor (GSI LY411,575; BALB/c → WT + GSI); and (F) PKCθ^null mice whose NOTCH signaling was abrogated by administering γ-secretase inhibitor (GSI LY411,575) in chow (BALB/c → PKCθ^null + GSI). (G) Day to complete rejection was compared between different groups of recipient mice. For each animal grafted with BALB/c skin, an internal control of C57BL/6 skin was also grafted (BL6 → BL6) to monitor integrity of the grafting technique. Data represent the mean + SEM (n = 3–16 mice/group). ***P < 0.001; one-way ANOVA with Tukey’s post-test applied.

Figure 2

NOTCH1 associates with PKCθ in stimulated T cells. NOTCH1 and PKCθ co-localize and interact following T cell stimulation. (A) Purified CD4⁺ T cells from C57BL/6 mice were incubated for 30 min with Dynal beads pre-coated with anti-mouse CD3ε and anti-mouse CD28 on glass slides. Cells were fixed, quenched, permeabilized, blocked, and stained with antibodies to NOTCH1 (green) and PKCθ (red). Proteins were visualized using species-specific, fluorescently conjugated secondary antibodies. *Indicates bead. (B) Jurkat T cells were stimulated with plate-bound anti-human CD3ε and anti-human CD28 for the indicated time periods, and subjected to co-immunoprecipitation with anti-NOTCH1. Immunoprecipitates (upper panel) and 1/100 of input (lower panels) were immunoblotted with anti-PKCθ and anti-NOTCH1. Data are representative of at least three independent experiments.

Figure 3

NOTCH1 interacts with CARMA1 and BCL10 following stimulation with anti-CD3ε and anti-CD28. (A) Jurkat T cells were stimulated with plate-bound anti-human CD3ε and anti-human CD28 and subjected to sucrose gradient centrifugation. All fractions were analyzed by dot blot with anti-cholera toxin-conjugated HRP (upper panel). Pooled fractions were immunoblotted with indicated antibodies (lower panel). (B) Jurkat T cells were stimulated as in (A) for the indicated time periods, followed by co-immunoprecipitation with antibodies specific for NOTCH1, CARMA1, or BCL10. Eluates were subjected to immunoblotting with indicated antibodies. 1/100 of input (shown) was immunoblotted with anti-CARMA1, anti-BCL10, and anti-NOTCH1. Arrow heads beside IB: NOTCH1 represent 120 kDa, transmembrane form (upper arrow head) and 110 kDa intracellular form (lower arrow head) of NOTCH1. Data are representative of at least three independent experiments.

Figure 4

Cytosolic N1IC can interact with CARMA1 in a bifluorescence complementation assay. Jurkat T cells were transfected with various constructs of nuclear and/or cytosolic proteins and their physical association was assessed microscopically by their ability to reconstitute two halves of a yellow fluorescent protein reporter. Twenty-four hours after transfection, cells were plated onto glass bottom culture dishes, stimulated with anti-human CD3ε and anti-human CD28, and fluorescent images of live cells were captured using a Zeiss LSM 510 confocal microscope. (A) Nuclear proteins, cFos and cJun, known to interact in the nucleus were co-expressed as a control; (B) NOTCH1 with an additional nuclear localization signal, N1IC–NLS was co-expressed with nuclear cJun; (C) NOTCH1 with an additional nuclear export signal, N1IC–NES, was co-expressed with nuclear cJun; (D) NOTCH1 with an additional nuclear export signal, N1IC–NES was co-expressed with CARMA1, a cytosolic protein; (E) NOTCH1 with an additional nuclear localization signal, N1IC–NLS was CARMA1; (F) NOTCH1 with an additional nuclear export signal, N1IC–NES was expressed alone. Upper panel of all images represent fluorescent channel only; lower panel of all images represent merged fluorescent and DIC images. Data are representative of at least three separate experiments.

Figure 5

NOTCH1 deficiency prevents the formation of the CBM complex. (A) NOTCH1 expression was efficiently reduced when Jurkat T cells were incubated with viral supernatants. NOTCH1-knock-down Jurkat T cells, or mock-infected control cells, were stimulated with plate-bound anti-human CD3ε and anti-human CD28 for 1 h, harvested for co-immunoprecipitation with indicated antibodies, then subjected to immunoblotting with (B,C) anti-NOTCH1, anti-CARMA1, anti-BCL10, or anti-PKCθ; 1/100 of input was immunoblotted with indicated antibodies, or with (D) antibodies specific for NOTCH1, phosphorylated and total IκBα, phosphorylated and total ERK, and phosphorylated and total p38 MAPK. (E) NOTCH1-knock-down Jurkat T cells, or mock-infected control cells, were stimulated with plate-bound anti-human CD3ε and anti-human CD28 for the times indicated. EMSAs were performed on nuclear extracts using radiolabeled oligonucleotides containing the NF-κB binding sequence. Red filled diamond represents competition with cold probe. Data are representative of at least three independent experiments.

Figure 6

The RAM domain of NOTCH1 mediates its interaction with CARMA1. (A) GFP–N1IC mutant expression plasmids were transfected into 293T cells and sub-cellular localization of NOTCH1 was examined using fluorescence microscopy. (B) VSV–CARMA1 expression plasmids were transiently co-transfected with GFP empty vector as a control, or with GFP–N1IC mutant plasmids in 293T cells, as indicated. Transfected cells were harvested and cell lysates were co-immunoprecipitated with antibodies specific for GFP. Co-immunoprecipitated proteins were then immunoblotted with anti-VSV or anti-GFP. Ig heavy chain indicates immunoglobulin heavy chain.

Figure 7

NOTCH1 increases NF-κB activity through its direct interaction with CARMA1 in the cytoplasm. (A) 293T cells were transiently transfected with GFP-N1IC–NLS or GFP-N1IC–NES and sub-cellular localization was determined using fluorescence microscopy. (B) CARMA1 expression plasmids were transiently co-transfected with GFP empty vector as a control, or with GFP-N1IC–NLS or GFP-N1IC–NES in 293T cells. Cytoplasmic extracts from transfected cells were co-immunoprecipitated with anti-GFP. 1/100 of input (shown) was immunoblotted with anti-VSV and anti-GFP. (C) Vectors expressing cytosolic N1IC–NES, membrane-tethered N1ΔE–PM, and/or cytosolic CARMA1 were co-transfected with an NF-κB luciferase reporter construct into Jurkat T cells, stimulated with PMA and CaI for 3 h, then subjected to a luciferase reporter gene assay. Ig heavy chain indicates immunoglobulin heavy chain. Data are mean + SEM of at least three independent experiments.