Fringe glycosyltransferases differentially modulate Notch1 proteolysis induced by Delta1 and Jagged1

Abstract

Fringe O-fucose-beta1,3-N-acetylglucosaminyltransferases modulate Notch signaling by potentiating signaling induced by Delta-like ligands, while inhibiting signaling induced by Serrate/Jagged1 ligands. Based on binding studies, the differential effects of Drosophila fringe (DFng) on Notch signaling are thought to result from alterations in Notch glycosylation that enhance binding of Delta to Notch but reduce Serrate binding. Here, we report that expression of mammalian fringe proteins (Lunatic [LFng], Manic [MFng], or Radical [RFng] Fringe) increased Delta1 binding and activation of Notch1 signaling in 293T and NIH 3T3 cells. Although Jagged1-induced signaling was suppressed by LFng and MFng, RFng enhanced signaling induced by either Delta1 or Jagged1, underscoring the diversity of mammalian fringe glycosyltransferases in regulating signaling downstream of different ligand-receptor combinations. Interestingly, suppression of Jagged1-induced Notch1 signaling did not correlate with changes in Jagged1 binding as found for Delta1. Our data support the idea that fringe glycosylation increases Delta1 binding to potentiate signaling, but we propose that although fringe glycosylation does not reduce Jagged1 binding to Notch1, the resultant ligand-receptor interactions do not effectively promote Notch1 proteolysis required for activation of downstream signaling events.

Figure 1.

Mammalian Fng proteins have differential effects on ligand-induced Notch1 signaling. (A) 3T3 cells were cotransfected with Notch1 (N1) and either SEAP, LFng-AP (LFng), MFng-AP (MFng), or RFng-AP (RFng), along with CBF-luciferase reporter pGL3JH26 and Renilla luciferase reporter pRLTK constructs; cocultured with either L, D1, or J1 cells; and assayed for luciferase activity. Luciferase activity is expressed as percent fold activation reflecting normalized relative luciferase units (RLUs) induced by ligand-expressing cells over RLUs obtained with L cells (ligand-activated N1+SEAP RLUs are set to 100% [D1/L = 8.65 ± 0.7, J1/L = 9.54 ± 1.5), bar graph shows mean ± SD; ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, n = 6; result from three independent experiments, each experiment done in duplicate). (B) 293T cells were either mock transfected (Mock) or transfected with SEAP, LFng-AP (LFng), LFngADD-AP (LFADD), RFng-AP (RFng), or RFngADD-AP (RFADD) plasmids. Conditioned medium and whole cell lysates from transfected cells were collected 48 h later and assayed for alkaline phosphatase (AP) activity and expressed as absorbance at OD₄₀₅ normalized for protein concentration. (C) 3T3 cells were cotransfected with Notch1 (N1) and either SEAP, LFng-AP (LFng), or LFngADD-AP (LFADD) (left) or with SEAP, RFng-AP (RFng), or RFngADD-AP (RFADD) (right) along with CBF-luciferase reporter pGL3JH26 and Renilla luciferase reporter pRLTK constructs, cocultured with either L, D1, or J1 cells and assayed for luciferase activity. Luciferase activity is expressed as fold activation over L cells reflecting normalized RLUs induced by ligand-expressing cells over RLUs obtained with L cells. Bar graph shows mean ± SD, n = 4, RLU obtained with L cells arbitrarily = 1; representative result of two independent experiments.

Figure 2.

Effects of mammalian fringe proteins on binding of soluble DSL ligands to Notch1-expressing cells. 293T cells transfected with N1 and either SEAP, LFng, MFng, RFng, or LFADD were incubated with preclustered D1Fc (A) or preclustered J1Fc (B) in culture media not containing 0.05% NaN₃ as outlined in Materials and Methods (B). Binding of increasing concentrations of D1Fc or J1Fc to transfected cells was measured by flow cytometry and expressed as MFI ± SD, n = 4 (vertical axis) plotted against the concentration of D1Fc (0.1, 0.3, and 1 μg/ml, horizontal axis) or the concentration of J1Fc (0.1, 0.3, and 1 μg/ml, horizontal axis). (C) 293T cells transfected with vector or N1 and either LFng, MFng, RFng, or LFADD were incubated with 10 μg/ml unclustered J1Fc in binding buffer containing 0.05% NaN₃, and bound J1Fc was detected with anti-human Fc FITC-conjugated antibody and the binding detected is plotted as MFI. (D) 293T cells described in C were incubated with 10 μg/ml preclustered J1Fc, and binding detected is plotted as MFI. (E) 293T cells cotransfected with pHcRed, vector, or N1 and either SEAP, LFng, MFng, RFng, or LFADD were incubated with 1 μg/ml preclustered J1Fc or preclustered Fc control, and the MFI of pHcRed-transfected cells was detected and plotted. (F) 3T3 cells cotransfected with pHcRed, vector or N1 and either SEAP, LFng, MFng, RFng, or LFADD were incubated with 1 μg/ml preclustered J1Fc and the MFI of pHcRed-transfected cells was detected and plotted. (G) 3T3 cells transfected with N1 and either SEAP, LFng, MFng, RFng, or LFngADD along with the CSL-reporter construct were incubated with either preclustered Fc (0.3 μg/ml), preclustered D1Fc (0.3 μg/ml), or (H) preclustered J1Fc (0.3 μg/ml) for 24 h. Luciferase activities were assayed and normalized as fold activation over Fc (relative luciferase units [RLUs] induced by either D1Fc or J1Fc over RLUs obtained with Fc). Bar graph indicates mean ± SD, n = 4; ^***p <0.001, RLU obtained with Fc arbitrarily = 1; data from three independent experiments are shown.

Figure 2.

Effects of mammalian fringe proteins on binding of soluble DSL ligands to Notch1-expressing cells. 293T cells transfected with N1 and either SEAP, LFng, MFng, RFng, or LFADD were incubated with preclustered D1Fc (A) or preclustered J1Fc (B) in culture media not containing 0.05% NaN₃ as outlined in Materials and Methods (B). Binding of increasing concentrations of D1Fc or J1Fc to transfected cells was measured by flow cytometry and expressed as MFI ± SD, n = 4 (vertical axis) plotted against the concentration of D1Fc (0.1, 0.3, and 1 μg/ml, horizontal axis) or the concentration of J1Fc (0.1, 0.3, and 1 μg/ml, horizontal axis). (C) 293T cells transfected with vector or N1 and either LFng, MFng, RFng, or LFADD were incubated with 10 μg/ml unclustered J1Fc in binding buffer containing 0.05% NaN₃, and bound J1Fc was detected with anti-human Fc FITC-conjugated antibody and the binding detected is plotted as MFI. (D) 293T cells described in C were incubated with 10 μg/ml preclustered J1Fc, and binding detected is plotted as MFI. (E) 293T cells cotransfected with pHcRed, vector, or N1 and either SEAP, LFng, MFng, RFng, or LFADD were incubated with 1 μg/ml preclustered J1Fc or preclustered Fc control, and the MFI of pHcRed-transfected cells was detected and plotted. (F) 3T3 cells cotransfected with pHcRed, vector or N1 and either SEAP, LFng, MFng, RFng, or LFADD were incubated with 1 μg/ml preclustered J1Fc and the MFI of pHcRed-transfected cells was detected and plotted. (G) 3T3 cells transfected with N1 and either SEAP, LFng, MFng, RFng, or LFngADD along with the CSL-reporter construct were incubated with either preclustered Fc (0.3 μg/ml), preclustered D1Fc (0.3 μg/ml), or (H) preclustered J1Fc (0.3 μg/ml) for 24 h. Luciferase activities were assayed and normalized as fold activation over Fc (relative luciferase units [RLUs] induced by either D1Fc or J1Fc over RLUs obtained with Fc). Bar graph indicates mean ± SD, n = 4; ^***p <0.001, RLU obtained with Fc arbitrarily = 1; data from three independent experiments are shown.

Figure 3.

Fringe proteins do not alter cell surface expression of N1, and activation of N1 signaling leads to losses in N1 cell surface expression. (A) 293T cells were cotransfected with N1 (pBOS-N1) and either SEAP, LFng, MFng, or RFng and biotinylated with Sulfo-NHS-Biotin (0.5 mg/ml) for 30 min at 4°C, and biotinylated proteins were isolated with SAV-immobilized beads. The SAV precipitates (lanes 1–4) and WCL (lanes 5–8) were analyzed by IB with 93-4 serum (Shawber et al., 1996). N1^FL, furin-cleaved membrane-associated form of N1 (N1™), and (^*) that may represent either intracellular N1 that is not posttranslationally modified or a protease fragment of N1^FL are indicated. (B) 293T cells were cotransfected with N1 and either SEAP, LFng, MFng, or RFng and cocultured overnight with parental L cells (lanes 1, 4, 7, and 10), D1 cells (lanes 2, 5, 8, and 11), or J1 cells (lanes 3, 6, 9, and 12). After biotinylation WCL were incubated with SAV-immobilized beads and analyzed by IB with 93-4 serum, anti-pan cadherin antibody (Pan-Cad, to control for biotinylated cell surface proteins pulled down by SAV) and anti-dynamin antibodies (α-Dyn, to detect any inadvertent biotinylation of intracellular proteins due to dead or leaky cells). (C) The immunoblot from B was scanned by Typhoon 9410 scanner (Amersham Biosciences) and analyzed by Imagequant software. The intensities of N1™ were quantified as integration volume obtained in each lane normalized to the relative integration volume of cell surface cadherin (Cad).

Figure 4.

Fringe proteins differentially modulate ligand-induced nuclear translocation of NICD. (A) Depicts Gal4-VP16 (GV) system to measure S3 nuclear localization in response to ligand-induced proteolysis of N1GVΔmyc (see text for details). (B) 3T3 cells were cotransfected with N1GVΔmyc and either SEAP, LFng, MFng, or RFng, along with Gal4-luciferase reporter pG5M2 and pRLTK reporter constructs and cocultured with either L, D1, or J1 cells followed by detection of luciferase activity. Luciferase activity is expressed as fold activation over L cells reflecting normalized relative luciferase units (RLUs) induced by ligand-expressing cells over RLUs obtained with L cells (ligand-activated N1+SEAP RLUs are set to 100% (D1/L = 2.08 ± 0.14, J1/L = 1.91 ± 0.06), bar graph shows mean ± SD; ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, n = 4; results from two independent experiments, each experiment done in duplicate).

Figure 5.

Fringe proteins differentially modulate ligand-induced proteolysis of N1. (A) Depicts the structure of N1 and the N1Δmyc constructs used to identify the furin-cleaved form of N1Δmyc (S1), ADAM-cleaved form of N1Δmyc (S2), and NICD form of N1Δmyc (S3). The ZEDN1Δmyc construct encodes a mutant protein starting at the S2 cleavage site of N1. The ZEDN1V/LΔmyc construct is ZEDN1Δmyc containing a point mutation (circle) in the S3 cleavage site (V1742 to L). The FCDN1Δmyc construct encodes a mutant protein that contains only the sequences C-terminal to N1 transmembrane domain. (B) 293T cells were transfected with either N1Δmyc, ZEDN1Δmyc, ZEDN1/LΔmyc, or FCDN1Δmyc, and whole cell lysates were analyzed by IB with 9E10. (C) FCDN1Δmyc 293T cells were either untreated (-) or treated (+) with λ phosphatase (λPPase; New England Biolabs, Beverly, MA) to confirm the phosphorylated form of S3 (S3^*). (D) 293T cells cotransfected with N1Δmyc and either SEAP or LFng were cocultured with parental L cells (lanes 1, 3, 5, and 7) or D1 cells (lanes 2, 4, 6, and 8) in the absence or presence of proteasome inhibitor MG132 (10 μM). Cells transfected with N1Δmyc+LFng were cultured in the absence of ligand cells as a control (lane 9). Whole cell lysates were IPd with 9E10 and analyzed by IB with 9E10 to identify S1, S2, S3, and S3^*. (E) 293T cells cotransfected with N1Δmyc and either SEAP or LFng cocultured with parental L cells (lanes 2 and 6), D1 cells (lanes 3 and 7), J1 cells (lanes 4 and 8), or no cells control (lanes 1 and 5) in the presence of MG132 (10 μM) were processed as described in D. The membrane was stripped and reprobed with N1 Val1744 antibody to further identify the NICD form generated by γ-secretase activity. (F) 293T cells cotransfected with N1Δmyc and either SEAP, MFng, or RFng were cocultured with parental L cells (lanes 1, 4, and 7), D1 cells (lanes 2, 5, and 8), or J1 cells (lanes 3, 6, and 9) and processed as described in E.

Figure 6.

Inhibition of γ-secretase activity indicates that LFng alters S2 cleavage. (A) 293T cells cotransfected with N1Δmyc and either SEAP or LFng were cocultured with parental L cells (lanes 1, 4, 7, and 10), D1 cells (lanes 2, 5, 8, and 11), or J1 cells (lanes 3, 6, 9, and 12), either in the absence (lanes 1–3 and 7–9) or presence (lanes 3–6 and 9–12) of the γ-secretase inhibitor DAPT (50 μM) and analyzed by IP with 9E10 followed by IB with 9E10. (B) The immunoblot from A was scanned by Typhoon 9410 scanner (Amersham Biosciences) and analyzed by Imagequant software. The intensities of the proteolytic fragments generated in the assay were quantified as integration volume of S3 or S2 obtained in each lane normalized to the relative integration volume of S1.

Figure 7.

Ligand-binding is required for fringe modulation of N1 ADAM cleavage. (A) 293T cells cotransfected with N1Δmyc and either SEAP or LFng were treated with EDTA for 15 min at room temperature followed by DMEM + 10% fetal bovine serum and incubation at 37°C for 30 min. Cells were lysed, and WCL were IPd with 9E10 and analyzed by IB with 9E10 and FCDN1Δmyc (FCDN1) served to identify S3. (B) 3T3 cells cotransfected with N1Δmyc and either SEAP or LFng and reporter constructs were treated with increasing amounts of EDTA (0.05–0.5 mM, horizontal axis) as described in A, except that the incubation in DMEM + 10% fetal bovine serum was lengthened to 8 h. Luciferase activity is expressed as fold activation over “minus EDTA,” reflecting relative luciferase units (RLUs) induced by EDTA treatment over RLUs obtained in the absence of treatment (line graph shows mean ± SD, n = 4; representative result of two independent experiments). (C) 293T cells cotransfected with N1Δmyc and either SEAP or LFng were treated with the phorbol ester TPA (100 nM) for the indicated time periods and processed as described in Figure 5D. (D) Parental L cells or cells stably expressing a C-terminal hemagglutinin (HA)-tagged form of D1 (D1-HA) were metabolically labeled with [³⁵S]methionine and treated with TPA (100 nM). One plate of D1-HA cells contained BB3103 (5 μM) throughout the starvation and labeling period (lane 5). Conditioned medium from labeled cells was collected and IPd with 148G antibody (raised against the extracellular domain of D1; DiSibio and Weinmaster, unpublished data) to isolate and identify shed D1 extracellular domain (D1 ^ECD). (E) D1-HA cells were either treated with ethanol (lanes 1–3) or 100 nM TPA (lanes 4–6) for 2 h in the absence or presence of 10 μM BB3103 (lanes 2 and 5) or 30 μM DAPT (lanes 3 and 6). Whole cell lysates were harvested and analyzed by IB with 12CA5 that recognizes the C-terminal HA tag on D1^FL, TACE-cleaved form of D1 (D1^TM1 and D1^TM2), and ICD form of D1 (D1^ICD) are indicated. A protein band that nonspecifically reacts with 12CA5 is indicated with an asterisk.

Figure 8.

N1 associates with TACE but not Kuzbanian, and expression of LFng does not alter the N1–TACE complex in response to ligands. (A) 293T cells were transfected with N1Δmyc and increasing amounts of TACE (0, 0.5, and 2 μg, top) or Kuzbanian (0, 0.5, and 2 μg, bottom). The transfected cells were either untreated (-) or treated (+) with metalloprotease inhibitor BB-94, and WCL were IPd with 9E10 and IB with anti-TACE antibody (Chemicon International) or anti-Kuzbanian antibody (Chemicon International) and then stripped and reprobed with 9E10. The proform (TACE^p) and mature form (TACE^m) of TACE, and the proform (Kuz^p) and mature form (Kuz^m) of Kuzbanian are indicated. The furin-cleaved transmembrane form of N1 is indicated as N1™. (B) 293T cells either mock transfected (-) or transfected with N1Δmyc (+, 1 μg/60-mm dish) were either untreated (-) or treated (+) with metalloprotease inhibitor BB-94, and WCL were IPd with anti-TACE antibody and IB with 9E10. (C) 293T cells transfected with N1, TACEea (TACE mutant in which the catalytic Glu-406 was mutated to Ala), and either SEAP or LFng were cocultured with parental L (lanes 1 and 4), D1 (lanes 2 and 5), or J1 (lanes 3 and 6) cells, and WCLs were either IPd with anti-TACE antibody and analyzed by IB with 9E10 (top) to reveal the amount of N1 associated with TACE or IB with 9E10 to reveal the N1 proteolytic cleavages induced by ligand (bottom). The membrane was stripped and reprobed with anti-TACE antibody (middle).

Figure 9.

Fringe glycosylation differentially modulates Notch1 proteolysis induced by Delta1 and Jagged1. (A) N1 O-fucosylation (circles) allows Delta/Jagged binding and sustained interactions during ligand internalization into the ligand-presenting cell, which function to mechanically dissociate the noncovalent N1 heterodimer. Once the NECD is removed from the N1 cell, the remaining membrane-bound N1 (N1™) is efficiently cleaved at S2 and S3 to release NICD and effect downstream signaling. (B) Fringe glycosylation of N1 (triangles) enhances Delta binding to Notch to promote NECD dissociation and transendocytosis that increase S2 and S3 cleavage and downstream signaling. (C) Jagged binds fringe glycosylated N1 but the interactions with Notch are not sufficiently strong to sustain the “pulling” force that is necessary to remove NECD and thus N1 proteolysis and signaling are suppressed.