Table S1. Protein identified using tandem mass spectrometry (MS/MS) of extract from p-2 Dicer-deficient and control neural stem cells.

Table S2. The reference list after ‘protein score’ threshold cutoff of >25 was applied to the protein list in Table S1. Proteins with >twofold higher level in Dicer-deficient neural stem cells (KO bias) were annotated in red. Proteins with>twofold higher level in control neural stem cells (WT bias) were annotated in blue. Proteins with similar level were annotated in black.

Table S3. Protein list of biological processes revealed by enrichment analyses for Gene Ontology by DAVID of proteins expressed >twofold higher in control neural stem cells (WT bias).

Table S4. Protein list of biological processes revealed by enrichment analyses for Gene Ontology by DAVID of proteins expressed >twofold higher in Dicer-deficient neural stem cells (KO bias).

Table S5. Protein list of biological processes revealed by enrichment analyses for Gene Ontology by DAVID of proteins similarly expressed in control and Dicer-deficient neural stem cells.

Table S6. List of proteins involved in the EGF signaling pathway. Although some of the genes involved the EGF pathways were upregulated by >twofold in either the Dicer-deficient or control NSCs, the majority of genes were similarly expressed in Dicer-deficient and control NSCs.

Table S7. List of proteins involved in the FGF signaling pathway.

Table S8. List of proteins involved in the VEGF signaling pathway.

Table S9. List of proteins involved in the cell-death pathway. The ratio of pro-cell death against pro-survival genes is about 0.7, 1.0 and 2.5 in the case of similar expressed protein, WT bias and KO bias, suggesting the Dicer-deficient neural stem cells are more sensitive to cell death due to increased pro-cell death proteins and the lack of protective pro-survival proteins.

Fig. S1. Dicer cortical deletion results in smaller cortices. (A) The breeding strategy to generate Dicer conditional knockout mice using Emx1-Cre mice. (B) The Emx1-Cre line was bred with the Rosa26-LacZ reporter line. LacZ expression demonstrated cortical-specific activities of the Emx1-Cre at E12.5 and P1. (C) At P5 and P14, the cerebral cortices (arrows) were greatly reduced in the Dicer knockout (Ko) as compared to controls (Ctrl). The cerebellum was not affected in Dicer Ko brains (arrowheads). (D) In coronal sections of P5 Dicer Ko cortices with Nissl staining, the cortical wall (arrows) was significantly thinner than controls. The hippocampus (arrowheads) was not detectable in the Dicer Ko.

Fig. S2. Normal neural progenitor development in Dicer knockout (Ko) cortices at E13.5. (A-D) The expression of neural progenitor markers was normal in E13.5 Dicer Ko cortices compared to controls (Ctrl), as detected by in situ hybridization for Hes5 (A), Ngn2 (B) and β-catenin (β-Cat) (C) and immunohistochemistry for Pax6 (D). (E) More dead cells (arrowheads) were detected in E13.5 Dicer Ko cortices than in controls. (F) Mitotic cells, detected by PH3 (green), were normal in Dicer Ko cortices. (G-J) No significant changes in the cell-cycle length of E13.5 Dicer Ko progenitors, compared to controls. The progenitor cells (Ki-67⁺, green) were labeled with BrdU (red) after a 30 minute pulse. Fields: n=6. At least 3 Dicer Ko and 3 control animals were used for statistical analyses.

Fig. S3. Dicer cortical deletion reduces cortical neural progenitors. (A) In coronal sections of E15.5 cortices, neural progenitors in the ventricular zone (VZ) and subventricular zone (SVZ), labeled by Hes5, Ngn2 and β-catenin (β-Cat), were reduced in Dicer^Cre/flox (Ko) compared to controls (Ctrl), as detected by in situ hybridization. (B,C) The percentage (%) of radial glial cells in the VZ (solid line), labeled by Pax6 (red), was greatly reduced in E15.5 Dicer Ko cortices. Fields: n=6; **P<0.005. (D,E) The percentage (%) of neural progenitors in the VZ and SVZ (solid line), labeled by Ki-67 (green), was significantly reduced in Dicer Ko cortices. Fields: n=6; **P<0.002. (F) Many dead cells (green) were detected in E15.5 cortices of Dicer Ko, compared to Ctrl, as detected by a TUNEL assay. (G) The percentage (%) of TUNEL-positive cells per field was significantly increased in Dicer Ko mice. Fields: n=6; *P<0.05; **P<0.001. (H,I) Decreased mitotic cells, detected by PH3 (green), in Dicer Ko cortices. Fields: n≥4; **P<0.001. (J,K) The percentage (%) of intermediate progenitors, labeled with Tbr2 (green) and BrdU (red) (30 minute pulse), in the SVZ was greatly reduced in E15.5 Dicer Ko cortices, compared to Ctrl. Fields: n=6; *P<0.05; **P<0.001. At least 3 Dicer^Cre/flox and 3 control animals were used for all statistical analyses.

Fig. S4. Dicer-deficient neural stem cells (NSCs) from passage-0 (p-0) displayed abnormal differentiation. NSCs that formed neurospheres were directly cultured under the differentiation condition without mitogens. NSCs collected from Dicer knockout (Ko) cortices can give rise to cells that express the neuronal markers Tuj1 and Map2, and the astrocyte marker GFAP. However, their morphology was abnormal, as shown with shorter neurites and processes than controls (Ctrl). The oligodendrocyte marker O4 was undetectable due to cell death in Dicer Ko NSCs. Days of in vitro (div) culture are labeled.

Fig. S5. Enrichment analysis of protein expression profiles of neural stem cells (NSC). The proteins identified for (A) WT bias, (B) KO bias, and (C) similar level from the tandem mass spectrometry proteomic analyses were performed with DAVID for Gene Ontology (GO) biological processes. The categories for unknown biological processes or molecular function were omitted in the results. Only categories with at least 10 genes were displayed in the pie chart. The number of genes and enrichment P-value were annotated behind each category shown. Asterisk denotes categories found to be enriched P<0.05 relative to the whole reference list. (A) There appeared to be an enrichment of genes involved in cell-cycle processes in control NSC (WT bias). (B) There appeared to be an enrichment of a more complex repertoire of biological processes in the Dicer-deficient NSCs (KO bias), which includes processes like signal transduction and molecular transport. (C) There appeared to be enrichment of biological processes involved in metabolic regulation for proteins with similar expression levels between Dicer-deficient and control NSCs.

Fig. S6. Responses to mitogens were not significantly affected in Dicer-deficient neural stem cells. Genes involved in the EGF, FGF and VEGF pathway were collated and compared to the reference list and broken down into WT bias, KO bias or similar level. The number of genes identified and the overall percentage were annotated. Although some of the genes involved in these pathway were upregulated by >twofold in either the Dicer-deficient or control NSCs, the majority of genes appeared to be expressed at similar level. Furthermore, no significant difference of numbers of affected gene hits in Dicer-deficient NSCs was detected.

Fig. S7. The complete networks of proteins affected by Dicer deletion in neural stem cells (NSCs). Using Ingenuity Pathways Analysis, several hub proteins upregulated in Dicer-deficient NSCs were identified, such as FMR1/FMRP, and hub proteins upregulated in control NSCs were found, for instance TGFβR2 and SOD-1.