Figure 1.
Isolation of a mouse line with behavioral abnormalities.
A) Mutant mice fall frequently and display an inability to right themselves easily. B) Determination of hereditary pattern by IVF. Squares and circles indicate male and female, respectively. Black-fill indicates mice, possibly homozygote of the mutated allele, with behavior abnormalities. Half-filled and open symbols are presumed to be heterozygote and WT with normal behavior. Numbers indicate number of offsprings. C) Hind heel footprints for normal and mutant mice. D) Frequent limb-clasping displayed by older mice (30-week old).
Figure 2.
Cerebellar atrophy in the ts3 mutant mice revealed by Micro-MRI and histological analysis.
A) Brain images acquired by micro-MRI (left and middle panels). Horizontal (left) and sagittal (middle) images of cerebella indicate cerebellar atrophy in the ts3 mutant (lower panels) and WT control (upper panels). The right panels show H&E-stained cerebellar sagittal sections of the ts3 mutant (lower) and WT control (upper). Scale bar indicates 1 mm. B, C) Cerebellar size of ts3 mutants and WT controls. First, the lengths of major and minor axes (mm) of the whole cerebellum were measured by micro-MRI (B, upper panels). They were consequently multiplied and compared between two genotypes (C, upper, n = 6, unit: mm2). To examine significant anterior folial atrophy in ts3 cerebella, we also measured the length of the folia (shown in B, lower) and compared in ts3 and WT mice (C, lower, n = 6, unit: mm). *P<0.05, **P<0.01.
Figure 3.
Normal morphology for spinal cord and sciatic nerves.
A) No significant abnormality in the sciatic nerve (arrow) was detected in ts3 mutants by micro-MRI. B) Toluidine blue (left) and H&E staining (right) shows normal morphology of sciatic nerves in both genotypes. Scale bar: 25 µm. C) Micro-MRI sagittal images of brain stem and spinal cord in ts3 and WT mice. No abnormalities were detectable in ts3. Ce: cerebellum, MO: medulla oblongata, SC: spinal cord. D) Toluidine blue and H&E-stained spinal cord of ts3 and WT mice. Spinal cord morphology was normal in both genotypes, and no significant difference was observed. Scale bar indicates 100 µm.
Figure 4.
Morphological abnormality of Purkinje cells in ts3 cerebella.
A) Kluver-Barrera's stain was performed on ts3 and WT cerebella. Number and morphology of Purkinje cells were normal in ts3 mutants. In contrast, however, GC numbers were reduced, and the size of the nuclei was significant smaller than WT. Scale bars, 50 µm (upper panel); 20 µm (lower panel). B) Immunohistochemisty was performed on cerebella in ts3 and WT mice using anti-calbindin antibody as a Purkinje cell marker. Purkinje cell morphology became progressively altered commensurate with mouse age. Note that dendrites of ts3 Purkinje cells at 30 weeks of age look significantly thicker than their WT counterparts. Dendrite branching was also abnormal in ts3 mutants. Scale bar, 20 µm. C) Significant reduction in granule cell density (cell number/mm3) in old ts3 mutants (30 weeks old). D) The PC dendrite is significantly thicker in ts3 mutants at 30 weeks of age (n = 70 each, averaged diameters of the thickest regions). *P<0.05, ***P<0.001.
Figure 5.
ts3 purkinje cells demonstrate a characteristic of senescent cell as observed by electronmicroscopy.
Electron photomicrographs of PC cells in 30-week-old WT (A, B) and ts3 (C, D, E). Lipofuscin accumulation in cell bodies (arrowheads in B and C), representing senescent postmitotic neurons, was observed in ts3, but not in WT PC cells. Scale bars, A, B, 5 µm; C, D, 2 µm; E, 0.5 µm.
Figure 6.
Abnormal morphology of dendrites and spines of the Purkinje cells in ts3 mutants.
A) Representative ultra-high voltage electron microscopy (UHVEM) images of Purkinje cell dendrites in ts3 mutant and WT mice (2000x). The arrow in the lower panel indicates isthmic portions of dendrites observed in ts3 mutants. Note the significant difference in morphology between ts3 and WT cerebella. Whereas the bulbous shape of dendritic spines is observed in WT controls, dendritic spines in ts3 mutants are smaller and irregularly-shaped, as indicated by arrowheads. Scale bar indicates 5 µm. B) Dendritic spine density (number of spine/µm2) was comparable in ts3 and WT controls. C) Morphology of the dendritic spines is significantly different in ts3 mutants when compared to that of WT controls. Whereas stubby or mushroom-shaped spines, which are characteristic of normal and mature PCs, are commonly observed in WT cerebella, thin (filopodia-like or headed) spines, representing a deficit in spine maturation, are predominant in ts3 mutants. Double-headed spines, including branched spines, were rarely observed in either genotype. “Unknown” indicates spines that could not be categorized from HVEM images by gross observation due to high density of dendritic spines in PCs. The χ2 test indicated that the difference in dendritic morphology between two genotypes is statistically significant (p<0.01).
Figure 7.
Distal expansion of CF territory in ts3 mutants.
(A-N) Immunofluorescence for calbindin (A, C, green in E, F, blue in K, M), VGluT2 (B, D, red in E, F), VGluT1 (G, I), VIAAT (H, J), parvalbumin (red in K, M) and GFAP (L, N) in control (A, B, E, G, H, K, L) and the ts3 mutants (C, D, F, I, J, M, N). Note that VGluT2-positive CF terminals are markedly increased in number and distributed throughout the molecular layer in ts3 mutants (E, F). Scale bars, A, 1 mm; E, G, K, 20 µm.
Figure 8.
Free spines and mismatched PF-PC synapses in ts3 mutant cerebella.
(A-E) Conventional electron microscopy in control (A, B) and ts3 mutant (C-E). A) Asterisks indicate PC spines contacting with PF terminals. B) Higher magnification image of a normal spine in WT. C) Letters f and m indicate free spines and mismatched PF–PC synapses, respectively. High-power images of free spine (D) and mismatched synapse (E). Note that PSD, as indicated between two arrowheads, is well-matched with the synaptic terminal in normal control (B). In contrast, the free spine and mismatched synapse do not have any contact with PF synaptic terminal (D) or has a significantly longer PSD (E), respectively. Scale bars, 500 nm. F) Percentage of normal synapses, mismatched synapses, and free spines in ts3 and WT cerebella. Majority of spines observed in ts3 cerebella were either mismatched or free spines. ***P<0.001.
Figure 9.
Severe reduction of GluD2 and Cbln1 immunoreactivities in ts3 mutants.
(A-L) Immunohistochemistry for GluD2 (A-D), Cbln1 (E-H) and Car8 (I-L) in control (A, B, E, F, I, J) and ts3 mutants (C, D, G, H, K, L). Scale bars, A, E, I, 500 µm; B, F, J, 20 µm.
Figure 10.
Identification of the ts3 mutation in the grid2 gene.
A) RT-PCR was performed on mRNA extracted from either ts3 or WT cerebella. A band that was roughly 1 kb shorter was detected from the mutant cDNA, suggesting a large deletion in the grid2 DNA of ts3 mice. B) Proteins were extracted from the ts3 and WT cerebella and Western blotting was performed using anti-GluD2 and β-actin antibodies. The GluD2 protein was not detected in ts3 mutant cerebella. C) Schematic representation of the deletion in ts3 genomic DNA (a) and protein (b). a) The 6 exons (exon 3 to 8) are missing, and only portions of intron 2 and 9 were detected by genomic PCR (DNA sequences between the two Xs in the WT are missing in the ts3 mutants). b) The deletion in the grid2 DNA results in protein truncation (from amino acids 83 to 145) and also a frame-shift mutation. The resulting gene product in ts3 mutants is 124 amino acids in length. D) Sequencing analyses confirmed that whereas exon 4 (enclosed sequences in the upper line) follows exon 3 sequences in the WT cDNA, exons 9 (shaded sequences in light gray) and 10 (shaded in dark gray) follow exon 3 in grid2ts3. The deletion in the grid2ts3 gene causes a frame-shift mutation from amino acid 83 (C to W) and following amino acids in the GluD2ts3 protein and terminates the protein in a truncated, 124 amino acid form due to a termination codon (tga, *).