2Broad Institute of Massachusetts Institute of Technology and Harvard, Cambridge, Massachusetts 02142, and 3Massachusetts Institute of Technology Department of Brain and Cognitive Sciences and the Picower Institute for Learning and Memory, Cambridge, Massachusetts 02139
De novo protein synthesis is globally increased in Tsc2-deficient rat hippocampal neurons in an mTOR-dependent manner. A, Hippocampal neurons infected with lentivirus that knock down either firefly luciferase (sh-Luc, control) or rat Tsc2 gene (sh-Tsc2), or left uninfected as indicated. At days 12–16 d in vitro (DIV12-16), neurons were pretreated with anisomycin (20 μm for 1.5 h), DMSO or rapamycin (20 nm overnight), incubated with 5 μg/ml puromycin for 30 min immediately before harvesting. Incorporation of puromycin in newly synthesized protein was detected by Western blotting with an anti-puromycin antibody. Total protein loaded on the gel was visualized by staining the postblotting membranes with EZblue dye. B, Quantification of puromycin blotting signals normalized to that of sh-Luc-infected neurons with DMSO treatment. Bars represent means ± SEM (n = 5; 0.115 ± 0.024 for anisomycin-treated wild-type, 1.159 ± 0.054 for DMSO treated sh-Tsc2 knockdown, 0.797 ± 0.076 for rapamycin treated sh-Luc, and 0.711 ± 0.075 for rapamycin treated sh-Tsc2 knockdown). *p < 0.05 (Student's t tests). **p < 0.01 (Student's t tests). ns, Not significant.
Application of TRAP assay for neuronal translational profiling. A, Expression of GFP-L10a transgene is seen in nucleoli as well as throughout the cytoplasm and neurites of cultured hippocampal neurons with MAP2 counterstaining. No significant cell death was observed in eGFP-L10a expressing cultures. Scale bar, 20 μm. B, Representative images show coexpression of GFP-L10a and mCherry-positive sh-Luc or sh-Tsc2. Scale bar, 50 μm. C, Representative Western blots confirm the coexpression of two lentivectors as indicated. D, TRAP assay using anti-GFP immunoprecipitation (IP). Representative Western blots show purification of GFP-L10a versus GFP from hippocampal neuronal cultures. Total RNAs were recovered from IP-purified ribosomes and amplified before microarray hybridization. The BioAnalyzer 2000 image confirms purification of RNAs from GFP-L10a-transfected cultures, but not from GFP-transfected cultures.
TRAP microarray reveals upregulation of stress and activity responsive genes and downregulation of synapse-associated genes in Tsc2-deficient hippocampal neurons. A, Heatmap of differentially translated genes at FDR < 1e-5 (59 annotated) by comparing either sh-Luc versus sh-Tsc2, or sh-Luc versus sh-Tsc2 after rapamycin treatment. The candidates fall into three categories: (1) 17 mRNAs downregulated in sh-Tsc2 cultures compared with control cultures; (2) 27 mRNAs upregulated in sh-Tsc2 cultures and not suppressed by rapamycin treatment; and (3) 38 mRNAs upregulated in sh-Tsc2 cultures and then significantly suppressed by rapamycin treatment. B, Correlation plots from two independent TRAP experiments within sh-Luc, sh-Tsc2, or sh-Tsc2 knockdown after overnight treatment with rapamycin (20 nm). C, Scatter plot of fold changes between sh-Luc versus sh-Tsc2 knockdown cultures, and sh-Luc versus sh-Tsc2 cultures treated with rapamycin. Colored dots indicate mRNAs with significant changes at FDR < 1e-5. Red represents mRNAs that are differentially expressed in the sh-Tsc2 cultures and corrected by rapamycin treatment (based on twofold change cutoff). Green represents mRNAs that remain significantly altered in the Tsc2-deficient neurons regardless of rapamycin treatment (also based on a twofold change cutoff). Blue represents mRNAs that have altered expression merely due to rapamycin treatment.
Translational versus transcriptional regulation of the TRAP candidate genes. A, Fold change at ribosome-associated mRNA (“translating transcripts”) level for indicated genes in sh-Tsc2 cultures, sh-Tsc2 cultures treated with rapamycin compared with sh-Luc hippocampal cultures. B, Fold change at transcriptional (“total transcripts”) level for the same genes. *p < 0.05 (two-way ANOVA). **p < 0.01 (two-way ANOVA). ***p < 0.001 (two-way ANOVA). ns, Not significant. C, Representative Western blot image showing elevated protein levels of ATF3 and UCP2 in sh-Tsc2 hippocampal cultures and differential suppression by rapamycin treatment for overnight or 7 d. D, Densitometry quantification of ATF3 fold change normalized to tubulin (vehicle n = 13, Rapa O/N n = 7, Rapa 7d n = 4). Error bars indicate mean ± SEM. **p < 0.01 (two-way ANOVA). E, Increased protein levels of ATF3 (1.49 ± 0.09 (SD) fold) in Tsc1-deficient mouse hippocampus versus wild-type controls (CC: Syn1-Cre+Tsc1cc; WW: Syn1-Cre+/Tsc1ww; postnatal day 21). F, Densitometry quantification of ATF3 fold change normalized to GAPDH. *p < 0.05 (Student's t test). n = 3 per condition.
Increased levels of ATF3 and gelsolin are activity-independent in the Tsc2-deficient neurons. A, Representative Western blot of control and Tsc2-deficient neurons untreated or treated with 1 μm TTX for 24 h. S6 phosphorylation at Ser240/244 (p-S6) indicates mTORC1 activation. Decreased CREB phosphorylation at Ser133 confirms reduced synaptic activity in the TTX-treated neurons. B, C, Densitometry quantification of ATF3 (B) and gelsolin (C) shown as fold change. Values were normalized to total Akt level (n = 6). Error bars indicate mean ± SEM (two-way ANOVA). n.s., Statistically not significant.
Acute rapamycin treatment rescues the reduction of dendritic spine length but not of spine density in Tsc2-deficient neurons. A, Hippocampal neurons were transfected with lentivector-based sh-Luc versus sh-Tsc2. Representative images show a dramatic decrease in the dendritic spine density as well as increased soma size in Tsc2-deficient neurons. Scale bars, 20 μm. B, High-magnification images showing the spine density and morphology in sh-Luc versus sh-Tsc2-transfected neurons. MAP2 stains dendrites. Scale bars, 5 μm. C, Quantification of spine densities in nontreated hippocampal neuronal cultures or cultures treated with vehicle versus rapamycin. Acute rapamycin treatment failed to rescue the deficits in spine density in Tsc2-deficient neurons compared with control cultures. ***p < 0.001 (Student's t test). D, Rapamycin rescued the spine length that decreases in Tsc2-deficient hippocampal neurons. **p < 0.01 (Student's t test). E, No significant differences in spine width between sh-Luc versus sh-Tsc2 cultures were detected, regardless of treatments with vehicle or rapamycin. F, The densities of PSD-95 puncta per 100 μm in sh-Luc versus sh-Tsc2 cultures are 18.38 ± 2.12 and 9.22 ± 2.27 (mean ± SEM). *p < 0.05 (Student's t test).
Blocking ATF3 induction reverses the spine density deficit in Tsc2-deficient neurons. A, Representative Western blots show the induction of ATF3 and gelsolin in Tsc2-deficient neurons. shRNA-mediated Atf3 knockdown suppressed protein levels of ATF3 as well as gelsolin. B, Representative images show expression of mCherry positive sh-Tsc2 (red) and/or GFP positive sh-Atf3 (green), and/or hATF3-FLAG (blue). Coexpression of sh-Atf3 increases spine density in Tsc2-deficient neurons. Overexpression of human FLAG-tagged ATF3 (resistant to Atf3 shRNA) under this condition reverses this phenotype by reducing spine density. Scale bars, 5 μm. C, Quantitation of spine density in sh-Tsc2 alone, sh-Atf3 alone, sh-Tsc2 and sh-Atf3, and sh-Tsc2 and sh-Atf3 plus ATF3-FLAG-expressing neurons at DIV 16. Spine densities per 100 μm in sh-Tsc2, sh-Atf3, and sh-Tsc2 and sh-Atf3, and sh-Tsc2 and sh-Atf3 plus ATF3-FLAG-expressing cultures are 8.64 ± 2.24, 35.17 ± 7.67, 44.15 ± 5.49, and 20.04 ± 1.84 (mean ± SEM). **p < 0.01 (one-way ANOVA with Tukey's multiple-comparison test). ***p < 0.001 (one-way ANOVA with Tukey's multiple-comparison test).