Fig 2: Notch1, ApoER2 and Dab1 colocalize in hippocampal neurons. (A) Representative image of fluorescent immunolabeling shows that Notch1 is localized in a subset of CA3 neurons, co-expressing Dab1 and ApoER2 (magnified insert). ApoER2 is highly expressed in all CA3 neurons. (B) Fluorescent immunolabeling shows the distribution of Notch1 and Dab1 as compared to Arc/Arg3.1 in CA3 neurons. (B′) Close up of a dendrite showing colocalization of the three proteins (white arrowheads). (C) Western blot analysis on immunoprecipitated samples from whole hippocampal lysate shows that Notch1 displays a strong interaction with Dab1 whereas Notch1 and ApoER2 interaction appears weaker in the preparation (n = 4 independent experiments). (D) Immunolabeling on immunoprecipitated samples from whole cortical lysates show that Notch1 and Dab1 interact. ApoER2 is bound to both Notch1 and Dab1 (n = 3 independent experiments). (E) Co-IP from synaptosomal membrane fractions reveals a stronger interaction between Notch1 and ApoER2, which is reduced in the N1cKO (n = 3 independent pulldowns). In (C–E) GAPDH is used as a loading control for the inputs and to detect contamination in the IP samples. Scale bar in (A) is 50 μm and (B) 25 μm.

Fig 3: Notch1 affects NMDAR composition and conductance. (A) Western blot on co-IPed synaptosomal fraction shows that Notch1 and NMDAR1 physically interact (n = 3 independent experiments). (B) Representative western blot on WT and N1cKO whole hippocampal lysates indicates that NR1 is visibly reduced in the N1cKO (n = 5 independent experiments). (C) Immunoblot on whole hippocampal lysate indicates a reduction in NR2B expression in the N1cKO as compared to WT. NR2A does not appear affected (n = 4 independent experiments). (D) Bar graph summarizing the changes in protein expression of NR1, NR2B and NR2A normalized against β-actin. NR1 expression is reduced by 66.1% in the N1cKO mice (n = 6 mice per genotype; Student’s t-test, p < 0.05), NR2B is decreased by 41.2% in the KOs (n = 5 mice per genotype; Student’s t-test, p < 0.05), whereas NR2A has a 30.3% reduction in KO mice but is not significantly different from WT (n = 5 mice per genotype; Student’s t-test, p = 0.20; n = 4 animals per genotype). Immunoblot analysis on synaptosomal fraction shows that AMPA receptor subunits, (E) GluR1 and (E′) GluR2, are not affected in N1cKO mice as compared to the WT littermates (n = 3–6 independent experiments). (F) Bar graph indicates no difference in the expression of GluR1 (n = 5 mice per genotype; Student’s t-test, p = 0.16) and GluR2 (n = 5 mice per genotype; Student’s t-test, p = 0.95) between WT and N1cKO mice. (G) Bar graph showing the fold changes in transcript expression of NR1 (Nr1), NR2B (Nr2b), NR2A (Nr2a) and GluR1 (Glur1). Only mRNA levels of NR1 (0.62 ± 0.02 vs. 1 ± 0.13, n = 3–4 animals per genotype; Student’s t-test, p < 0.05) and NR2B (0.75 ± 0.09 vs. 1 ± 0.06, n = 4 animals per genotype; Student’s t-test, p < 0.01) are decreased in KO hippocampi as compared to WT. (H,H′) Immunoblots on synaptosomal (P2) and soluble (S2) fractions. (H) NR1 is reduced, whereas GluR1 and PSD95 tagging at the synapse appear unaffected. GAPDH is used as a positive control for the cytosolic fractions. (H′) GluR2 expression is not changed at the synaptic level (n = 4 independent experiments). (I) NMDAR- and AMPAR-mediated evoked responses in CA1 pyramidal neurons. Scaled sample current traces recorded at −70 mV (gray traces) and +40 mV in the presence of CNQX (black traces) from WT and N1cKO mice. For easier comparison of NMDAR/AMPAR ratio, the current at +40 mV was scaled to the peak current at −70 mV of the same recording. (J) Bar graph shows that the mean NMDAR/AMPAR is significantly decreased in N1cKO mice (n = 13–15 neurons per genotype; Student’s t-test, p < 0.006). Data are averages ± SEM, *p < 0.05 and **p < 0.01.

Fig 4: Loss of Notch1 interferes with the Reelin cascade and Reelin-mediated synaptic potentiation. (A) Western blot analysis on synaptosomal (P2) and cytosolic (S2) fractions from the cortex of Reelin heterozygous mice (Reln+/−) and wildtype controls (Reln+/+) shows that total NICD levels are not affected in both compartments. Moreover, Dab1 levels are increased and NMDAR1 expression is unchanged in the haploinsufficient (Reln+/−) mice (n = 3 independent fractionations). (B) Representative immunoblot on whole hippocampal lysates show a reduction in Notch1, ApoER2 and Dab1 expression in the N1cKO as compared to WT. Reelin and Dab1 phosphorylation appear unchanged (n = 3 independent experiments). (C) Bar graph summarizing the optical density analysis of Notch1, ApoER2 and Dab1 bands normalized to β-actin. In N1cKO hippocampi, a 76.2% reduction is observed in Notch1 (n = 6 mice per genotype; Student’s t-test, p < 0.01), a 65.9% reduction in ApoER2 (Student’s t-test, p < 0.05) and 50.7% reduction in Dab1 (n = 5 mice per genotype; Student’s t-test, p < 0.01) as compared to WT controls. (D) Dab1 phosphorylation is not significantly reduced in N1cKO mice (n = 5 mice per genotype; Student’s t-test, p = 0.09) as compared to control littermates. (E) qPCR analysis shows the transcript expression of Notch1, Reelin, ApoER2 and Dab1 in wildtype and N1cKO hippocampi. Dab1 in N1cKO mice is decreased by 20% as compared to WT (n = 4 per genotype; Student’s t-test, p < 0.001). Reelin and ApoER2 transcripts do not change. Notch1 transcript is reduced by 65% in N1cKO mice (n = 4 per genotype; Student’s t-test, p < 0.01). (F,F′) Field EPSPs (fEPSPs) were elicited in CA1 stratum radiatum by Schaffer collateral stimulation. LTP was induced with TBS. Upper traces represent average fEPSP recorded during baseline (gray) and 50 min after TBS (black). (F) In WT, superfusion of Reelin-conditioned medium (Rln) 10 min prior to TBS (gray bar) induces and promotes LTP (•, n = 9 slices, 6 mice) as compared to control medium (ctr; °, n = 9 slices, 7 mice; Student’s t-test, p < 0.05). (F′) In N1cKO mice, LTP is impaired in both Rln and control conditions (□, n = 9 slices, 4 mice; ▪, n = 9 slices, 5 mice; p = 0.63). *p = < 0.05, Data are averages ± SEM.

Fig 5: Notch1 affects CREB phosphorylation and transcription. (A) Immunoblot on whole hippocampal lysates shows a substantial reduction in CamKII in the hippocampi of N1cKO as compared to WT. β-actin is used as a loading control (n = 3 independent experiments). (B) Quantification of CamKII protein expression between N1cKO and WT normalized to β-actin shows a reduction by 78% in the N1cKO hippocampi (n = 6 animals per genotype; Student’s t-test, p < 0.05). (C) Western blot on hippocampal lysates shows that ERK1/2 phosphorylation (pERK1/2) in N1cKO is reduced as compared to WT hippocampi. β-actin is used as a loading control (n = 3 independent experiments). (D) Quantification of ERK1/2 phosphorylation in N1cKO as compared to WT indicates a 76% decrease in the KO samples (n = 6 mice per genotype; Student’s t-test, p < 0.05). (E) Immunoblot on whole hippocampal lysates shows a reduction in CREB phosphorylation in N1cKO as compared to WT (n = 3 independent experiments). (F) Bar graph shows a near to significance 56% reduction in CREB phosphorylation in the N1cKO hippocampi as compared to WT (n = 4 animals per genotype; Student’s t-test, p = 0.069). (G) Graph summarizing transcripts expression of CREB targets, Egr1, Bdnf and c-fos, in hippocampi from cage control (CC) and environmental exposed (EE) WT and N1cKO mice (n = 3–4 mice per genotype; Student’s t-test, p < 0.05; Black bars compare WT CC with WT EE; gray bars compare WT EE with N1cKO EE). (H) Model of the proposed Notch1 non-canonical signaling regulations of Reelin, NMDA and CREB pathways described in this study. Elements putatively connecting Notch1/Reelin signaling to Notch1/Glutamaterigic signaling are dotted and are based on the existing literature. Data are averages ± SEM and *=p < 0.05; 0 signifies p = 0.5.