A) Lysates from 293 cells transfected with empty vector (Ctrl) or E5-Myc overexpressing vector (WT) were assayed for endogenous RhoA activity using the RBD pulldown assay and analyzed by immunoblotting with an antibody to RhoA (top). GTPγS lane is a positive control for inducing RhoA activity. Increased endogenous RhoA activity is demonstrated by presence of α-RhoA signal in RBD pulldown lanes. Input protein levels and α-Actin loading control are shown (Bottom).
B) Lysates from 293 cells transfected with empty vector (Ctrl), E5-Myc (WT) or LQR mutant of E5-Myc (LQR) were assessed for RhoA activity as measured by RBD assay described in (A). Input protein levels and α-Actin loading control are shown (Bottom).
C) Presence of E5 is critical for wild type levels of endogenous RhoA signaling in vivo. P3 mouse whole brain lysates from WT or E5^−/− (KO) littermates were subjected to RBD pulldown assays as described in (A). A representative immunoblot is shown (left). From three experiments, blinded to condition, the quantification of α-RhoA signal was normalized to input RhoA signal (Right). Error bars ±SEM; *p<0.05.
See also Figure S2.

A) E16 hippocampi from E5^+/− or E5^−/− mice were dissected and dissociated for culture. At DIV10 dissociated neurons were transfected with GFP. At DIV14 neurons were fixed, stained and, excitatory synapses were measured as described in methods. Error bars ±SEM; ***p < 0.005, ANOVA.
B) Organotypic slices from WT or E5^fl/fl mice were biolistically transfected with Cre-recombinase (Cre) and dendritic spines were quantified as described in methods. Representative images are shown (left). Error bars ±SEM; ***p < 0.005, KS test.
C) Quantification of mEPSC inter-event interval and amplitude from acute hippocampal brain slices prepared from P12-P14 WT or E5^−/− mice. Error bars are standard deviation of the mean; ***p<0.005, *p<0.05.
D) Hippocampi from three independent littermate pairs consisting of P12 WT and E5^−/− mice were prepared as described in methods for quantification of synapses, Synapsin1 and PSD-95 using array tomography. Error bars ±SEM; *p < 0.05, Mann-Whitney U-Test.
E) Increase in excitatory synapse number following loss of E5 requires EphB2 signaling. At DIV10, control plasmid (−) or EphB2KD plasmid (+) were co-expressed in dissociated mouse hippocampal neurons with GFP and either Ctrl-shRNA or E5-shRNA. At DIV14 excitatory synapses were measured as described in methods. Error bars ±SEM; **p < 0.01, ***p<0.005, ANOVA.
F) E5 can suppress an EphB2-mediated increase in excitatory synapse number. At DIV10, control plasmid (-) or EphB2-expressing plasmid (+) were co-expressed in dissociated mouse hippocampal neurons with GFP and either control (Ctrl) plasmid or E5-Myc plasmid. At DIV14 excitatory synapses were measured as described in methods. Error bars ±SEM; **p < 0.01, ***p<0.005, ANOVA.
See also Figures S4, and S1.

A) Dissociated mouse hippocampal neurons were incubated with pre-clustered Fc, Fc-EB1 or Fc-EA1 for 60 minutes, lysed, and immunoprecipitated with α-C-E5 followed by immunoblotting with α-N-E5. Immunoblot of input with α-pEph or α-Actin (loading control) are shown. Western is one representative image and quantification is of three separate experiments with samples normalized to α-Actin (left). Error bars ±SEM; *p<0.05. Right, dissociated mouse hippocampal neurons were transfected with GFP (gray) and stimulated with either pre-clustered Fc (Ctrl) or Fc-EB1 (EB1) for 30 minutes, followed by fixing and staining for endogenous E5 using α-N-E5 (Red). White rectangle outlines magnified dendritic region showing examples of E5 staining (right).
B) Lysates from 293 cells previously transfected with various combinations of overexpressing plasmids containing E5-Myc, E1-Myc and/or Flag-EphB2 were immunoblotted with α-Myc, αFlag, or α-Actin (loading control).
C) Lysates from 293 cells previously transfected with various combinations of overexpressing plasmids containing Flag-EphB2, Flag-EphB2KD and/or E5-Myc were immunoblotted with α-Myc, α-Flag, or α-Actin (loading control).
D) Lysates from 293 cells previously transfected with various combinations of overexpressing plasmids containing E5-Myc, E5-Y361F-Myc and/or Flag-EphB2 were immunoblotted with α-Myc, α-Flag, or α-Actin (loading control). Representative immunoblot is shown (top). From three independent experiments E5 levels were quantified and normalized to E5 expression in absence of EphB2-Flag (bottom). Error bars ±SEM; **p<0.01.
E) Dissociated mouse hippocampal neurons transfected with GFP (gray) were stimulated similar to (B) in the absence or presence of lactacystin and immunostained with α-N-E5. White rectangle outlines magnified dendritic region showing examples of E5 staining (right).
F) WT and E5^−/− brains were lysed and immunoprecipitated with α-C-E5 followed by immunoblotting with α-ub or α-N-E5.
See also Figure S6.

A) E5 and EphB2 are expressed in the CA1 region and dentate gyrus (DG) of the hippocampus at P12. Adjacent 100 nm mouse brain sections were stained for E5 or EphB2 using digoxigenin-labeled RNA probes to the anti-sense strand or sense strand as a control (top). Lower panels show nuclear staining with DAPI.
B) Immunoprecipitation with α-Flag from 293 cell lysates previously transfected with various combinations of overexpressing plasmids containing E1-Myc, E5-Myc, Flag-EphB2, and/or Flag-EphA4, followed by immunoblotting with α-Myc or α-Flag. Input protein levels shown (bottom).
C) Immunoprecipitation of mouse cortical lysates with IgG or α-C-E5, followed by immunoblotting with α-EphB2 or α-N-E5 (left). Input protein levels shown (right).
D) Immunoprecipitation of WT or E5^−/− mouse whole brain lysates with α-C-E5 followed by immunoblotting with α-EphB2. Input EphB2 levels shown (bottom).
E) Dissociated rat hippocampal neurons were stained using α-N-E5 (Blue) and α-EphB2 (Red). A representative image of overlapped EphB2 and E5 is shown (left). White rectangle outlines magnified dendritic region (right) showing examples of EphB2/E5 co-localization (arrows). In three independent experiments, quantification of overlapped EphB2/E5 puncta was determined at DIV2, DIV4 and DIV8 and is represented as percent of EphB2 overlapped with E5 (right). Error bars ±SEM; *p <0.05, non-significant (n.s.).
See also Figure S1.

A) 10 ng of E5-shRNA or Ctrl-shRNA was co-transfected with GFP into rat hippocampal neurons at DIV14. At DIV18 dendritic spines were measured as described in methods. Representative image illustrates dendritic spines. N indicates number of neurons assessed. Error bars ±SEM; **p < 0.01, ANOVA.
B) 10 ng or 20 ng of two different E5-shRNA or Ctrl-shRNA constructs was co-transfected with GFP into rat hippocampal neurons at DIV10. At DIV14 excitatory synapses were measured as described in methods. Representative image illustrates quantified synapse puncta (White). Error bars ±SEM; **p < 0.01, ***p<0.005, ANOVA.
C) DIV10 rat hippocampal neurons were co-transfected with GFP and increasing concentrations of E5-Myc or control plasmid. At DIV 14 excitatory synapses (gray bars) and exogenous E5 expression (blue bars) were measured as described in methods. Representative image illustrates localization of E5-Myc on transfected neuron (Red). Error bars ±SEM; **p < 0.01, ANOVA.
D) Neurons were transfected with E5-Myc (E5-WT) or E5-LQR-Myc (E5-LQR) and quantified as in (C). Error bars ±SEM; **p < 0.01, ANOVA.
E) Quantification of mEPSC inter-event interval and amplitude from hippocampal neurons transfected as in (B) with 20 ng of shRNA. Cumulative distribution plots, bar graphs and representative traces are shown. Error bars are standard deviation of the mean, ***p<0.005, *p<0.05.
See also Figures S3, and S1.

A) Dissociated mouse hippocampal neurons were stimulated with either α-Fc IgG (Ctrl) or pre-clustered Fc-EB1 for 15 minutes. Neuronal lysates were immunoprecipitated with α-N-E5, followed by immunoblotting for pan-phosphotyrosine (α-pTyr) or E5 with α-N-E5. EB1 stimulation was determined by immunoblotting neuronal lysates for phospho-Eph (pEph). Input protein levels and α-Actin loading control are shown (Bottom).
B) E5-Y361 is a conserved residue with E1-Y87 (Sahin et al., 2005).
C) Immunoprecipitation with α-Myc from 293 cell lysates previously transfected with various combinations of overexpressing plasmids containing E5-Myc, E5 (Y361F)-Myc and/or EphB2-Flag, followed by immunoblotting with α-pTyr, α-Myc, α-pY361 or α-Flag. Input EphB2 levels are shown (bottom).
D) Neurons were treated and lysates prepared as in part A followed by immunoblotting with α-pY361 or α-N-E5. Representative immunoblot with input phospho-Eph (pEph) levels is shown (top). Quantification of three independent experiments is shown as a percent increase in pY361 over Ctrl stimulation (bottom). Error bars ±SEM; *p<0.05.
E) Dissociated rat hippocampal neurons were transfected with GFP (gray) and stimulated as in part A, followed by fixing and staining for endogenous phosphorylated E5 using α-pY361 (Red). Representative image shown (left). White rectangle outlines magnified dendritic region showing examples of phospho-E5 staining (left bottom). Four independent experiments were imaged and analyzed for pY361 (bar graph). Error bars ±SEM; *p<0.05.
F) Lysates from 293 cells transfected with empty vector (-) or increasing concentrations of E5-Myc with or without Flag-EphB2 were assessed for endogenous RhoA activity by RBD assay (previously described). GTPγS lane is a positive control for inducing RhoA. Input protein levels and α-Actin loading control are shown (Bottom).
G) WT and EphB2^−/− (B2^−/−) brain lysates were immunoblotted with α-EphB2, α-N-E5, α-Actin, or α-pY361 according to methods (left). Quantification of α-N-E5 or α-pY361 signal from three independent experiments is normalized to α-Actin and represented as fold change compared to wild type. Error bars ±SEM; *p<0.05.
See also Figure S5.

A) Immunoprecipitation with α-HA from 293 cell lysates previously transfected with various combinations of plasmids containing E1-Myc, E5-Myc, HA-DNUbe3A, HA-MEF2A, HA-Cbl-b, and/or HA-Ube3A, followed by immunoblotting with α-HA or α-Myc. Input protein levels and αActin loading control are shown (Bottom).
B) Hippocampal mouse neurons were co-transfected with GFP and control, HA-DNUbe3A or Ube3A-shRNA at DIV10. At DIV14, neurons were incubated with clustered Fc (-) or Fc-EB1 (+) for 30 minutes. Neurons were fixed and stained for E5 with α-N-E5 and quantified according to methods. Quantification is of E5 staining intensity normalized to Fc control. Error bars ±SEM; **p < 0.01, ANOVA.
C) Ube3A wild type and maternal-deficient (Ube3A^m−/p+) mouse brains were lysed and immunoblotted with α-N-E5, α-EphB2, α-MEF2, α-Actin (loading control), or α-Ube3A (top). Samples were normalized to α-Actin and quantified as described in methods (bottom). Error bars ±SEM; *p<0.05, Mann-Whitney.
D) Brain lysates from WT and Ube3A^m−/p+ were collected and treated similar to E, immunoprecipitated with α-C-E5 and immunoblotted with α-N-E5 and α-ub. Input protein levels are shown (right).
E) Neurons from WT and Ube3A^m−/p+ mice were dissociated, cultured and transfected with GFP at DIV10. At DIV14, neurons were incubated with pre-clustered Fc or Fc-EB1 for 30 minutes. Neurons were fixed and stained for E5 with α-N-E5 and quantified according to methods. Error bars ±SEM; **p<0.01.
See also Figure S7.