Absence of Whisker-related pattern formation in mice with NMDA receptors lacking coincidence detection properties and calcium signaling

Abstract

Precise refinement of synaptic connectivity is the result of activity-dependent mechanisms in which coincidence-dependent calcium signaling by NMDA receptors (NMDARs) under control of the voltage-dependent Mg2+ block might play a special role. In the developing rodent trigeminal system, the pattern of synaptic connections between whisker-specific inputs and their target cells in the brainstem is refined to form functionally and morphologically distinct units (barrelettes). To test the role of NMDA receptor signaling in this process, we introduced the N598R mutation into the native NR1 gene. This leads to the expression of functional NMDARs that are Mg2+ insensitive and Ca2+ impermeable. Newborn mice expressing exclusively NR1 N598R-containing NMDARs do not show any whisker-related patterning in the brainstem, whereas the topographic projection of trigeminal afferents and gross brain morphology appear normal. Furthermore, the NR1 N598R mutation does not affect expression levels of NMDAR subunits and other important neurotransmitter receptors. Our results show that coincidence detection by, and/or Ca2+ permeability of, NMDARs is necessary for the development of somatotopic maps in the brainstem and suggest that highly specific signaling underlies synaptic refinement.

Fig. 1.

Generation of the NR1 N598R alleles.A, Partial organization of the murine NR1 locus. Wild-type locus (NR1⁺), the targeted locus after homologous recombination (NR1^Rneo), and the targeted locus after Cre-mediated excision of the neo cassette (NR1^R). Exons are depicted as boxes. Coding regions are in black, and the 3′UTR are ingray. Exon 15 harbors the N598R mutation after targeting. loxP elements are shown as filled triangles, and the neo selection marker is shown as anopen box. The 5′ and 3′ limits of the targeting construct are indicated (× symbols) on the NR1^Rneo and NR1^R alleles. Relevant EcoRV (E) restriction sites and DNA fragments, as well as localization of hybridizing probe B, are shown. PCR primers 1–4 (see Materials and Methods) for routine genotyping are depicted as horizontal arrows.B, Southern blot of EcoRV-digested genomic mouse tail DNA. Lanes 1–3 show offspring from mating I (NR1^+/Rneo × NR1^+/Rneo), lanes 4–7 show offspring from mating II (NR1^+/−/Cre^+/+ × NR1^+/Rneo), and lanes 8 and9 show offspring of mating III (NR1^+/− × NR1^+/−).Top, The membrane was hybridized with probe B to test for the different mutated NR1 alleles: wild-type (+symbols; 5 kb); targeted (R^neo; 6 kb); and targeted after Cre-mediated resolution of the neo cassette (R; 4.6 kb).Bottom, The membrane was stripped and hybridized with probe A to test for the null allele: wild-type (+symbols; 6.9 kb) and null (− symbols; 9.6 kb) (see Materials and Methods).

Fig. 2.

Analysis of NR1 expression in brains from newborn mice. A, Northern blots. Three hundred nanograms of poly(A⁺) RNA from different genotypes were probed with a cDNA fragment covering NR1 exon 15–17 and reprobed for actin as a control. Lanes 1–4 show NR1^+/+, NR1^R/+, NR1^R/−, and NR1^+/− littermates from mating II, and lanes 5 and 6 show NR1^−/− and NR1^+/+ littermates from mating III.B, Western blots. Immunoblot analysis of 50 μg of membrane protein from littermates as in A with an antibody binding to the C terminus of NR1. The membrane was reprobed for tubulin. Even prolonged autoradiographic exposures did not detect specific NR1 signals in NR1^−/− samples.

Fig. 3.

Western blot analysis of expression of glutamate and GABA_A receptor subunits. Membrane protein fractions (50 μg) from whole brains of newborn and adult mice were probed with antibodies to specific receptor subunits. Genotypes of newborn mice (P0) are from matings II and III. Adult wild-type (+/+) is shown as control (ND, not done). A, NMDA receptor subunits. Protein levels of NR2B were found not to be altered in mice expressing the N598R mutant NR1 subunit, whereas absence of the NR1 subunit lead to decreased expression levels of NR2B. NR2A and NR2C were both found to be low in newborn mice of all genotypes compared with higher protein levels in adult mice, suggesting that the NR1 N598R mutation does not affect the protein levels of NR2A and NR2C. NR2D protein levels were at comparable levels in newborn mice of all genotypes. B, AMPA receptor subunits.C, Metabotropic glutamate receptor. D, GABA_A receptor subunits. Expression levels of these subunits were at comparable levels in all NR1 genotypes, suggesting that neither the absence of NR1 nor the N598R point mutation interfere with their expression.

Fig. 4.

Current–voltage relationship of NMDA-induced currents. NMDA-induced macroscopic currents in N598R mutant cells lack rectification compared with wild type. Current–voltage relationships were determined after application of 20 μm NMDA using whole-cell patch recordings from CA1 pyramidal-shaped cells in organotypic hippocampal slice cultures. RepresentativeI–V curves of genotypes NR1^+/+, NR1^R/+, and NR1^R/− are shown in the presence of 100 μm Mg²⁺(NR1^+/+) or 500 μmMg²⁺ (NR1^R/+, NR1^R/−) (black traces), without added Mg²⁺ (light gray traces), and in the presence of 20 μm APV (dark gray traces). Currents were normalized to the current at −30 mV. A decreased rectification in NR1^R/+ and a complete lack in NR1^R/− are clearly visible, even at an increased Mg²⁺ concentration of 500 μm. For a quantitative evaluation, see Table 1.

Fig. 5.

Neuroanatomy of NR1^+/− and NR1^R/− littermates. Nissl-stained sections of cerebellum (A, B), hippocampus (C, D), and brainstem (E,F), as well as CO-stained sections of brainstem (G, H) of newborn NR1^+/− (A, C,E, G) and NR1^R/−(B, D, F,H) littermates are shown. Both Nissl and CO staining showed unchanged gross CNS structure in NR1^R/− mice compared with control NR1^+/− mice. However, whisker-related patterns were present in NR1^+/− mice (G) and absent in NR1^R/− mice (H). Scale bar: A–D, 250 μm; E–H, 500 μm. Section thickness:A–D, 15 μm; E–H, 50 μm.amb, Nucleus ambiguus; CA, cornu ammonis;dg, dentate gyrus; hyp, hypoglossal nucleus; nVi, subnucleus interpolaris of trigeminal nucleus; io, inferior olive; D, dorsal;L, lateral.

Fig. 6.

Cytochrome oxidase-stained brainstem sections in newborn NR1 wild-type and mutant mice. A–L, CO-stained transversal sections of the BSTC are shown at the level of the nVc (A, D, G,J), nVi (B, E,H, K), and nVp (C,F, I, L) of NR1^+/− (A–C), NR1^−/− (D–F), and NR1^R/− (G–L) animals. In animals expressing only the wild-type NR1 subunit, whisker-related rows (a–e) segregating into individual barrelettes were consistently found in nVi (B). Emerging whisker-related patterns were usually present in nVc (A), whereas patterning in nVp (C) was found only in a small number of animals. In contrast, whisker-related patterns were never found in animals exclusively expressing the N598R mutant NR1 subunit (G–L) or lacking the NR1 subunit (D–F). Examples from two NR1^R/− animals are shown in G–I andJ–L. MV, Motor nucleus of V;amb, nucleus ambiguus. Scale bar, 200 μm.

Fig. 7.

Barrelette detection by immunohistochemistry for Tenascin-C. Transversal section through the BSTC of newborn mice were immunostained with an antibody against the extracellular matrix protein Tenascin-C. The immunopositive pattern reflects the boundaries between barrelettes in control NR1^+/− animals (A–C). The emerging whisker-related pattern is most distinct in nVi (B) in which pronounced horizontal and finer vertical boundaries can be seen.Arrowheads denote distinct boundaries delineating whisker-specific rows. No emergent pattern within the BSTC can be discerned in NR1^R/− mice (G–I) or in NR1^−/−mice (D–F). The contours of the individual BSTC nuclei remain visible, and the sizes of the individual nuclei appear comparable in all three genotypes. Scale bar, 200 μm.

Fig. 8.

Projection and arborization of single whisker-related trigeminal afferents. Low-magnification, light microscopy (A, B,E–H) images of coronal sections of newborn NR1^+/+ (A, E,G) and NR1^R/− (B,F, H) mice through the brainstem at the level of the nVc (A, B), nVi (E, F), and nVp (G,H) after DiI application to a single whisker, B1 or B2. Patches of DiI-labeled axonal arborization were centered around the expected topographic location in the trigeminal nuclei. The location (l) of the single whisker-related patch in nVc for the B1 whisker was 0.160 (NR1^+/+, n = 2) and 0.161 (NR1^R/−, n = 2) and for the B2 whisker was 0.239 (NR1^+/+ or NR1^+/−, n = 8) and 0.225 (NR1^R/−, n = 6). Z-projections of stacks containing a series of 66 high-magnification confocal images (C, D, insets ofA and B, respectively) gave comparable overall impressions of axonal arborizations in both genotypes. Scale bar: A, B, E–H, 100 μm;C, D, 20 μm. Arrows mark the trigeminal tract (A, B).Arrowheads indicate axon collaterals branching off the trigeminal tract (C, D). White line depicts section margin (G,H). D, Dorsal; L, lateral.