NBR1 is a novel Spred2 binding partner. (A) A schematic representation of Spred2 and NBR1 domains. The N-terminal EVH1 domain of Spred2 was used as bait in a yeast two-hybrid screen that identified NBR1. The shortest cDNA of NBR1 identified coded for the 131 C-terminal amino acids. CC, coiled-coil; hNBR1, human NBR1; ZZ, ZZ-type ZINC finger domain. (B) Endogenous NBR1 as well as its interacting protein P62 can be immunoprecipitated with myc-tagged Spred2 from 293T cell lysates, irrespective of FGF2 stimulation. IP and WCL samples were immunoblotted with the indicated antibodies (IB). (C) Endogenous NBR1 can also be immunoprecipitated with endogenous Spred2 from SH-SY5Y neuroblastoma cell lysates using an anti-Spred2 antibody, irrespective of FGF2 stimulation. Approximately 4 mg of cell lysate was used for each IP along with 15 µg of either anti-Spred2 or nonspecific rabbit IgG antibodies. (D) Both the EVH1 and SPRY domains of Spred2 are necessary for the interaction with NBR1 in vivo. Endogenous NBR1 can only be immunoprecipitated from 293T cell lysates with either wild type (WT) or the KBD-deleted mutant of Spred2 but not the EVH1 domain– or SPRY domain–deleted mutants. Note that in all sections, film exposure times for the NBR1 WCL blots were significantly more than for the NBR1 IP blots. TUB, tubulin.

Colocalization of Spred2 with NBR1 is dependent on both the EVH1 and SPRY domains. (A) NBR1-GFP colocalizes with WT myc-Spred2 in COS-7 cells. (B) NBR1-GFP does not colocalize with EVH1-deleted myc-Spred2. (C) NBR1-GFP colocalization with myc-Spred2 is not affected by deletion of the KBD. (D) NBR1-GFP colocalization with myc-Spred2 also depends on the presence of the SPRY domain, probably because of the mislocalization of SPRY-deleted Spred2 to the nucleus. (A–D) Images on the right are higher magnifications of the boxed areas on the left. Bars, 10 µm.

Inhibition of ERK1/2 activity by Spred2 is dependent on the interaction with NBR1. (A, left) 293T cells stimulated with FGF2 in the presence of WT or various deletion mutants of myc-Spred2 were analyzed by IB for ERK1/2 activity. Inhibition of ERK1/2 by myc-Spred2 was significantly reduced by the loss of both the EVH1 and SPRY domains but not KBD. (right) Densitometric analysis of pERK1/2 levels from left. The p-values were calculated using one-tailed nonpaired t tests (n = 6). NC, negative control. (B) NBR1 overexpression enhances Spred2-mediated ERK1/2 inhibition. (left) GFP or NBR1-GFP plus either empty vector– or myc-Spred2–cotransfected 293T cells were FGF2 stimulated. NBR1-GFP on its own did not affect pERK1/2 levels, but when coexpressed with myc-Spred2, it significantly enhanced the ability of Spred2 to reduce pERK1/2 levels. Threefold more NBR1-GFP (or GFP for controls) was transfected to ensure that the majority of myc-Spred2–expressing cells also expressed NBR1-GFP. (right) Densitometric analysis of the pERK1/2 levels from the left. The p-value was calculated using a one-tailed paired t test (n = 3). (C) Depletion of endogenous NBR1 impairs Spred2 activity. (left) 293T cells were double transfected (with 100 pmol and then 50 pmol per 10-cm well) with either nonsilencing control or NBR1-specific siRNA oligonucleotides. The Tet-inducible Spred2 construct was cotransfected with TR (1:6 ratio to ensure that every Spred-transfected cell also expressed TR) along with the oligonucleotides on the second transfection. Spred2 expression was induced 6 h before stimulation by 1 µg/ml Tet. Two independent NBR1 siRNA oligonucleotides were used (Oligo1 and -2). Spred2 inhibition of ERK1/2 activity was significantly reduced in both cases. (right) Densitometric analysis of the left. The p-value was calculated using a one-tailed paired t test (n = 5). (D) Ectopic expression of NBR1 can rescue NBR1 interaction with the EVH1-deleted Spred2 mutant in a cell type–specific manner. 293T or COS-7 cells overexpressing NBR1-GFP along with WT myc-Spred2, ΔEVH1 myc-Spred2, or empty vector as negative control were subjected to IP by myc antibody. Although, like endogenous NBR1, the interaction of ectopic NBR1 with Spred2 was EVH1 domain dependent in COS-7 cells, ectopic NBR1 could still interact with ΔEVH1 Spred2 in 293T cells. Note that the film exposure time for the NBR1 WCL blot was significantly more than for the NBR1 IP blot. (E) Rescuing the NBR1–Spred2 interaction also rescues the function of a noninteracting, otherwise nonfunctional ΔEVH1 Spred2 mutant. (left) 293T cells expressing either GFP or NBR1-GFP plus WT myc-Spred2, ΔEVH1 myc-Spred2, or empty vector as negative control were stimulated with FGF2 as shown. ΔEVH1 myc-Spred2 could not reduce ERK1/2 activity alone, but pERK1/2 reduction by the ΔEVH1 myc-Spred2 mutant was rescued back to WT levels when ectopic NBR1-GFP was also expressed to rescue the interaction. NBR1-GFP expression on its own did not reduce ERK1/2 activity. Threefold more NBR1-GFP (or GFP for controls) was transfected to ensure that the majority of Spred2-expressing cells also expressed NBR1-GFP. (right) Densitometric analysis of the left. The p-value was calculated using a one-tailed paired t test (n = 3). TUB, tubulin. (A–C and E) Error bars represent SEM.

Schematic representation of the proposed mechanism for Spred. Via interacting with NBR1, Spred redirects activated receptors to the lysosomal degradation pathway. Down-regulation of downstream signaling can be selective, as those pathways like ERK1/2 that need specific endosomal compartments for efficient activation are affected the most from a shift in the balance between different trafficking routes (e.g., sorting/recycling versus lysosomal).

Endogenous NBR1 and Spred colocalize with each other. (A) The majority of myc-Spred2 in COS-7 cells exhibits a punctate cytoplasmic staining. (B) Endogenous SPRED2 also localizes to cytoplasmic punctae in SH-SY5Y neuroblastoma cells. (C) GFP-tagged NBR1 is localized to the limiting membrane of some vesicular structure in COS-7 cells. (D) Endogenous NBR1 similarly localizes to vesicular structures in COS-7 cells treated with BafA. (E) Endogenous NBR1 colocalizes with endogenous SPRED2 in SH-SY5Y cells treated with BafA. (D and E) Cells were subjected to 200 nM BafA for 16 h before analysis. (A–E) Images on the right (A–D) or insets (E) are higher magnifications of the boxed areas. Bars, 10 µm.

NBR1 is specifically localized to the limiting membrane of late endosomes. (A) No colocalization between NBR1-GFP and endogenous EEA1, the marker of early endosomes, is seen in COS-7 cells. (B) No colocalization is also observed between NBR1-GFP and RAB11, the marker of recycling endosomes. (C) A strong colocalization is detected between the late endosomal/lysosomal marker LAMP2 and NBR1-GFP. (D and E) NBR1-GFP also colocalizes strongly with the specific late endosomal marker MPR (D) but not with lysosomes that stain positive for LysoTracker red (E). (F) Endogenous NBR1 also colocalizes with the late endosomal/lysosomal marker LAMP2 in COS-7 cells treated with BafA. Cells were subjected to 200 nM BafA for 16 h before analysis. (A–F) Insets are higher magnifications of their associated pictures. Bars, 10 µm.

Spred2 targets activated receptors to the lysosomal degradation pathway in an EVH1 domain–dependent manner. (A) Spred2 expression results in degradation of ectopic active FGFR1. (left) Levels of ectopic active FGFR1 are significantly decreased in lysates of myc-Spred2– but not GFP control–cotransfected 293T cells, whereas those of ectopic TrfR1-GFP or downstream signaling components such as ERK, C-RAF, or GRB2 do not change. Fivefold more Spred2 (or GFP for controls) was transfected to ensure that the majority of receptor-expressing cells also coexpressed myc-Spred2. (right) Densitometric analysis of the left for TrfR1 and FGFR1. The p-value was calculated using a one-tailed paired t test (n = 3). (B) Spred2-mediated degradation of FGFR1 is via the lysosomal degradation pathway. (left) Treatment of 293T cells 1 h before transfection with BafA hinders myc-Spred2–mediated FGFR1 degradation. Fivefold more myc-Spred2 (or GFP for controls) was transfected to ensure that the majority of receptor-expressing cells also coexpressed myc-Spred2. Cells were analyzed 12 h after transfection. (right) Densitometric analysis of FGFR1 degradation by myc-Spred2 in BafA-treated versus untreated cells. The p-value was calculated using a one-tailed paired t test (n = 3). (C) Spred2-induced decrease in FGFR1 levels is EVH1 dependent. (left) Contrary to WT myc-Spred2, ΔEVH1 myc-Spred2 does not degrade FGFR1 in cotransfected 293T cells. Fivefold more myc-Spred2 or ΔEVH1 myc-Spred2 (or GFP for controls) was transfected to ensure that the majority of receptor-expressing cells also coexpressed Spred2. (right) Densitometric analysis of the left. The p-value was calculated using a one-tailed paired t test (n = 3). (D) Depletion of NBR1 impairs Spred2-mediated degradation of endogenous FGFR2. (left) 293T cells were double transfected (with 100 pmol and then 50 pmol per 10-cm well) with either nonsilencing control (NSC) or NBR1-specific siRNA oligonucleotides. The Tet-inducible Spred2 construct was cotransfected with TR (1:6 ratio to ensure that every Spred-transfected cell also expressed TR) along with the oligonucleotides on the second transfection. Spred2 expression was induced 6 h before stimulation by 1 µg/ml Tet. The experiment was performed with two independent NBR1 siRNA oligonucleotides (Oligo1 and -2). Spred2-mediated FGFR2 degradation was significantly reduced with both NBR1 siRNA oligonucleotides. (right) Densitometric analysis of the left. The p-value was calculated using a one-tailed paired t test (n = 4). (E) Spred2 alters receptor trafficking toward the lysosomal degradation pathway. In COS-7 cells transfected with FGFR2-GFP plus empty vector as control (negative control [NC]), the majority of endosomal receptors do not colocalize with LAMP2. In COS-7 cells transfected with FGFR2-GFP and WT myc-Spred2, most of the endosomal receptors colocalize with LAMP2 (+Spred2 FULL). However, deletion of the EVH1 domain results in the loss of this Spred2-induced LAMP2 colocalization (+Spred2 ΔEVH1). To identify Spred2-cotransfected cells, coverslips were also stained for Spred2 (blue) using an anti-Spred2 antibody (not depicted). Insets are higher magnification of their associated pictures. (F) The percentage of FGFR2-GFP colocalization with LAMP2 as in E was quantified and averaged. Measurements were from nine or more cells for each cotransfection from two separate experiments. The p-value was calculated using a one-tailed nonpaired t test (n = 9). (A–D and F) Error bars represent SEM. TUB, tubulin. Bars, 10 µm.