Figure 6.

Nrx HS Modification Is Required for its Function in Vivo

(A) Transgenic overexpression of Dnrx, but not DnrxΔHS induces synaptic overgrowth at Drosophila larval muscle 6/7 neuromuscular junctions.

(B and C) Dnrx mutants showed reduced numbers of synaptic boutons (postsynaptic Discs large in green and neuronal horse radish peroxidase in magenta). Bouton number was rescued by neuronal transgenic expression of Dnrx but not DnrxΔHS.

(D) Larval crawling speed was reduced in dnrx mutants and rescued by neuronal Dnrx but not DnrxΔHS.

(A–D) ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001 by ANOVA and Bonferroni tests; n = 15–20 (A), 17–31 (B), or 27–37 (D). Full genotypes are as follows. Con = w¹¹¹⁸. Dnrx overexpression = Elav-Gal4; UAS-Dnrx. DnrxΔHS overexpression = Elav-Gal4; UAS-DnrxΔHS. dnrx²⁷³/dnrx²⁴¹ no rescue = Elav-Gal4; dnrx²⁷³/dnrx²⁴¹. dnrx²⁷³/dnrx²⁴¹ Dnrx rescue = Elav-Gal4; UAS-Dnrx; dnrx²⁷³/dnrx²⁴¹. dnrx²⁷³/dnrx²⁴¹ DnrxΔHS rescue = Elav-Gal4; UAS-DnrxΔHS; dnrx²⁷³/dnrx²⁴¹.

(E) Nrxn1ΔHS mouse brain homogenate shows a shift in Nrx molecular weight by western blot.

(F and G) Nrxn1ΔHS mice have reduced neonatal survival indicated by reduced litter size (F). Postnatal survival rates from P0 to P15 are also reduced (G). ^∗∗p < 0.01 and ^∗∗∗∗p < 0.0001 by Mann-Whitney test, n = 18–26 litters from homozygous breeding.

(H) In situ hybridization revealed high expression of Nrx1 and Nrx2 but very low expression of Nrx3 in dentate gyrus granule cells. Mossy fiber synapses (red) from these granule cells to thorny excrescence spines on CA3 pyramidal neurons were studied in panels (I)–(S).

(I-K) The frequency and amplitude of mEPSCs in CA3 pyramidal cells was reduced in slices from Nrxn1ΔHS mice relative to WT. ^∗∗p < 0.01 and ^∗∗∗p < 0.001 by Mann-Whitney test, n = 20 cells from 4 mice each. Inter-event interval and amplitude distributions were significantly different p < 0.0001 by Kolmogorov-Smirnov test.

(L and M) Paired pulse ratios of mossy fiber CA3 synapses were elevated in Nrxn1ΔHS mice relative to WT. Sample traces are for 40 ms interstimulus interval, with the lower Nrxn1ΔHS traces scaled to compare with WT. Genotype p < 0.0001 by two way RM ANOVA with ^∗p < 0.05 by Bonferroni’s test, n = 11 cells from 4 mice each.

(N and O) The integrated intensity of punctate VGlut1 excitatory presynaptic marker was reduced in stratum lucidum, the CA3 region of mossy fiber inputs. NeuN costains CA3 cell bodies. ^∗p < 0.05 by t test from 3 mice each averaging data from 7 sections per mouse.

(P–S) Thorny excrescence (TE) spines were imaged from fluorescent fills of CA3 cells. Total TE area per dendrite shaft area was reduced in Nrxn1ΔHS mice relative to WT. ^∗∗∗∗p < 0.0001 by Mann-Whitney test, n = 37–38 dendrites from 3 mice each; p < 0.0001 by Kolmogorov-Smirnov test of TE area per shaft area distributions. The number of TE spines per dendrite area was also reduced. ^∗p < 0.05 by Mann-Whitney test, n = 37–38 dendrites from 3 mice each.

Error bars represent SEM. Scale bars: (C) 100 μm, (H) 500 μm, (N) 10 μm, (P) 5 μm. See also Figure S6 and Table S3.