Supporting Information

Transcription

1 Supporting Information Nowakowski et al /pnas SI Materials and Methods Primary Antibodies. Primary antibodies used in this study were raised against BrdU (rat, 1:50; Abcam), cleaved caspase 3 (rabbit, 1:50; Cell Signaling), GFP (goat, 1:400; Abcam), HuC/D (mouse, 1:150; Sigma), Pax6 (rabbit, 1:300; Covance), Sox2 (rabbit, 1:3,000; Milipore), Sox9 (1:1,500; Millipore), and Tbr2 (rabbit, 1:100; Abcam). Constructs. The CAG-cre-IRES-EGFP construct containing the cytomegalovirus early enhancer element and chicken β-actin promoter (CAG) element, internal ribosomal entry site (IRES), and enhanced green fluorescent protein (EGFP) was provided by Anjen Chenn (Northwestern University, Chicago, IL). To overexpress mir-92b, we used a strong mammalian expression vector containing the CAG promoter element upstream of myristolated, destabilized enhanced GFP (myr-degfp) coding sequence (1). The genomic pre-mir-92b hairpin sequence (2) was amplified from the genomic DNA using the following primers: forward, ATACTCCCCAGAGCACTCCA; reverse, GCTCGGTA- GAGATCGAAAGC. The 3 UTR of the mouse Eomes coding the Tbr2 transcription factor was amplified from the genomic DNA by nested PCR using the Expand High Fidelity Plus PCR system (Roche) using the following primers: forward1, GTGGCGCTTATCAGAGGAAG; reverse1, GGATTAAGCCTGCAGAGCAC; subsequently, forward2, AAGATCAGCTTGCCAAGGAA; reverse2, GAAAG- GGGGCAGAAAGAAAC. The UTR was subcloned into psicheck-2 (Promega) and is referred to as the WT 3 UTR in this paper. Nucleotides of the Eomes mrna coding sequence were deleted using the QuickChangeII XL kit (Agilent) using the following primers: sense, CCTTCCACCTTTGATG- TATCCTGTTTTTCTCTGAGAGAAGG; antisense, CCTTCT- CTCAGAGAAAAACAGGATACATCAAAGGTGGAAGG. For sponge experiments, six copies of the mir92 binding site (GGAGGCCGGGAAAGTGCAATA) were ligated downstream of the destabilized D2EGFP coding sequence and then subcloned into the pcag-ires-egfp expression vector. The scrambled control sequence was published before (3). To generate scrambled sponge or mir-92b sponge expressing retroviruses, six copies of the sponge were subcloned into the PmeI site of the CAG-GFP (Addgene plasmid 16664) retroviral vector (4). Replication incompetent retroviruses were produced by transient cotransfection of Platinum GP cells (Cell Biolabs) with pseudotyping vector VSV-G and the retroviral expression vector. The supernatant was harvested h after transfection, centrifuged, and filtered through a 0.45-μm filter (Millipore). Viral particles were concentrated at 25,000 g for 2 h at 4 C, resuspended in PBS, and stored at 80 C. Final titer tested in 3T3 cells was found to be cfu/ml. Retrovirus injection used the same surgery as for in utero electroporation. Quantification of Immunopositive Cells in Sections. Three nonadjacent coronal sections through the central telencephalon were selected for quantifications of immunopositive cells. Optical sections through the dorso-lateral telencephalon containing electroporated cells were acquired at constant separation, and three to four optical sections through the center of the section were summed using the image processing software ImageJ. Counting ladders consisting of 200 (wide) 40-μm (deep) boxes, 10 for E14.5 and 12 for E15.5, were positioned in Adobe Photoshop with the base along the ventricular edge. Exceptions included counts of cleaved caspase-3 positive cells, where only a single 300-μm-wide box was used, spanning the entire thickness of the dorsal telencephalic wall (E15.5) or of the cortical plate (E18.5). Between 7 and 12 embryos from at least three surgeries were analyzed per genotype for each quantification. For postnatal sections of in utero electroporated Dicer1 fl mice interbred with the Rosa26RYFP mice (5), brains of all surviving mice were sectioned and analyzed for GFP expression. Only brains with GFP + cells in the dorso-lateral cortex were considered. Counting ladders consisting of μm boxes spanning the entire thickness of the cortical wall were positioned with the base along the pial edge. For mice injected with CAG-GFPmiR-92b or scrambled sponge expressing retroviruses, five nonadjacent 20-μm coronal sections were analyzed by counting all detectable GFP immunoreactive cells in the neocortex. Astrocyte identity was confirmed by Sox9 immunostaining and morphology. Borders between cortical layers were established based on nuclear counterstaining with DAPI. 1. Wu S, et al. (2002) Characterization of ubiquilin 1, an mtor-interacting protein. Biochim Biophys Acta 1542(1-3): Griffiths-Jones S (2004) The microrna registry. Nucleic Acids Res 32(Database issue): D109 D Gentner B, et al. (2009) Stable knockdown of microrna in vivo by lentiviral vectors. Nat Methods 6(1): Zhao C, Teng EM, Summers RG, Jr., Ming GL, Gage FH (2006) Distinct morphological stages of dentate granule neuron maturation in the adult mouse hippocampus. J Neurosci 26(1): Srinivas S, et al. (2001) Cre reporter strains produced by targeted insertion of EYFP and ECFP into the ROSA26 locus. BMC Dev Biol 1:4. 1of6

2 Fig. S1. Evidence for efficient, rapid depletion of Dicer1 following electroporation at E13.5. (A) Schema showing experimental approach. E13.5 Dicer1 fl/fl or Dicer1 +/fl embryos were electroporated with cre expression vector. On E14.5, tissue containing GFP + cells was dissected, and dissociated cells were either subjected to FACS to isolate GFP + cells for extraction of (B) RNA or (C P) plated and reacted with LNA riboprobes for mature mir9 or mir124. (B) Quantification of Dicer1 mrna levels relative to GAPDH levels 1 d after electroporation (n = 3). (C P) Levels of two brain-enriched mature mirnas, mir9 and mir124, are reduced in cells 1 d after electroporation. (C J) Examples show in situ hybridization staining for mir9 in dissociated cells. Electroporated Dicer1 fl/fl (C) and control Dicer1 fl/+ (D) cells were identified based on the expression of GFP (filled arrowheads). GFP-positive (GFP + ) cells were chosen at random and paired with nearby GFP-negative (GFP ) cells (C and D, open arrows). Among (E) Dicer1 fl/fl and (F) control Dicer1 fl/+ cells, grayscale image intensity profiles were determined for pairs of GFP + and GFP cells (boxes). (Scale bar, 10 μm.) (G J) Examples of image intensity profiles for the individual cells boxed in E and F:(G) Dicer1 fl/fl GFP +, (H) Dicer1 fl/fl GFP,(I) Dicer1 fl/+ GFP +, and (J) Dicer1 fl/+ GFP ; lower grayscale values correspond to more intense in situ staining and hence more mirna. To evaluate the amount of mirna staining in GFP + cells compared with GFP cells, values of total image intensity were calculated for each cell by integrating its image intensity profile (G J) according to the equation in K, and for each pair of GFP + and GFP cells, the total intensity values were subtracted according to the equation in L. (M P) Estimating the proportion of Dicer1 fl/fl cells that underwent depletion was done by examining the distributions of intensity differences. (M) In controls, because both mir124 and mir9 are expressed by all cells in the developing cortex with about 50% expressing them strongly (1, 2) (dark shading, cells expressing high level of mirna; lighter shadings, cells expressing moderate to low level of mirna), analysis of mirna in situ staining using the equation in L is expected to result in either a number close to zero (cases 1 and 4), a positive number (case 2), or a negative number (case 3). (N) If Dicer1 is efficiently depleted, little or no in situ staining is expected in GFP + cells (lightest shading), and so the differences should be less than zero in all cases. (O and P) Distributions of intensity differences among Dicer1 fl/+ (filled bars) and Dicer1 fl/fl (open bars) cell pairs for in situ staining for (O) mir124 and (P) mir9: in controls, data were normally distributed around zero, as predicted, and in Dicer1 fl/fl cells, distributions were shifted such that the vast majority of pairs (85 90%) showed differences of less than zero. These wholesale shifts in the distributions indicate rapid and efficient depletion of these mirnas from almost all GFP + Dicer1 fl/fl cells, as illustrated in N. Data represent analysis of results pooled from three experiments, each analyzing pairs of cells. 1. Landgraf P, et al. (2007) A mammalian microrna expression atlas based on small RNA library sequencing. Cell 129(7): Maiorano NA, Mallamaci A (2009) Promotion of embryonic cortico-cerebral neuronogenesis by mir-124. Neural Dev 4:40. 2of6

3 Fig. S2. No evidence of increased apoptosis following the loss of functional Dicer. (A D) Images of coronal sections containing (A, B, and B ) Dicer1 +/ and (C and D) Dicer1 / cells at (A and C) E15.5 and (B and D) E18.5 showing immunostaining for GFP and the marker of apoptosis, cleaved caspase 3 (arrows). B shows an image of a positive control immunostaining for cleaved caspase 3 in the corpus callosum, where high levels of apoptosis have been reported (1), from the brain in B. (Scale bars, 100 μm.) (E) Enumeration of density of cells double-positive for GFP and cleaved caspase 3 in the dorsal telencephalon (dtel) at E15.5 or the cortical plate (CP) at E18.5 shows no difference between genotypes (n = 6). (F) To further test whether loss of functional Dicer results in abnormal levels of cell death, we performed the following experiment: Dicer1 fl/fl and control Dicer1 fl/+ embryos carrying the R26RYFP reporter allele were electroporated with cre-expression plasmid at E13.5 and developed until P14. Cells derived from electroporated radial glia were found reproducibly in the lateral cortex (red box). (G and H) Examples of coronal sections of cortex show Dicer1 +/ and Dicer1 / GFP/YFP cells. (Scale bar, 50 μm.) (I) The contribution of GFP/YFP cells to superficial cortical layers (I III) and deep cortical layers (IV VI) was higher in Dicer1 fl/fl than in Dicer1 fl/+ mice. Data represent means ± SEM from six animals from three surgeries with five sections analyzed per animal; borders between cortical layers were identified with DAPI nuclear counterstain. *P < 0.05, Tukey s test. 1. Niquille M, et al. (2009) Transient neuronal populations are required to guide callosal axons: a role for semaphorin 3C. PLoS Biol 7(10):e of6

4 Fig. S3. Comparison of mir-92b and the response element in the 3 UTR of Tbr2. (A) Multiple sequence alignment of sections of Tbr2 mrnas from 16 mammalian species containing the response element to mir-92b (highlighted in yellow). (B) Alignment of precursor mir-92b sequences from nine mammalian species including rodents and primates (MirBase accession numbers next to mirna name). Mature mmu-mir-92b is highlighted in yellow. Most of the nucleotides are identical in all species (*). Mouse pre-mir-92b differs from other species at a few residues (highlighted in green). (C) Predicted secondary structure of the mouse pre-mir-92b hairpin shows that residues that are different between the mouse and other mammals are not part of the mature mirna sequence and are not involved in base-pairing. All sequence alignments were performed using ClustalW. 4of6

5 Fig. S4. Mature mir-92b interacts with the 3 UTR of Tbr2. Dual luciferase assay in HEK-293 cells demonstrates that mir-92b can inhibit the expression of Renilla luciferase by interacting with the WT 3 UTR sequence of the Tbr2 mrna but not when nucleotides are mutated (n = 3 experiments). 5of6

6 Fig. S5. Results of prolonged inhibition of mir-92b. Following intraventricular injection of retrovirus expressing GFP-miR-92b sponge or GFP-scrambled sponge at E12.5, mice were allowed to develop until P14. (A A ) Example images of brains injected with GFP-scrambled sponge stained for (A) GFPand(A ) overlay image with nuclear counterstain DAPI. Oblique arrows indicate GFP expressing neurons and horizontal arrows indicate astrocytes, whose identities were confirmed by staining for astrocyte marker Sox9 (A A ; astrocytes accounted for 24 45% of GFP-positive cells in these control brains) (Scale bar, 50 μm.) (B B ) Example images of brain sections from mice injected with retrovirus expressing mir-92b sponge showing (B) GFP immunostaining and (B ) overlay image with nuclear counterstain DAPI. The distributions were shifted toward superficial layers, and the vast majority had neuronal morphologies, with 7 17% of cells identified as astrocytes on the basis of morphology and Sox9 staining. (Scale bar in A, A, B, and B, 100 μm.) (C) Quantification of the relative proportions of superficial and deep layer neurons in brains injected with retrovirus expressing scrambled sponge (brains 1 and 2) or mir-92b sponge (brains 3 5). Values for each brain represent mean ± SEM of all neurons counted in five nonadjacent 20-μm sections. ANOVA indicated a statistically significant difference in the relative proportions of superficial and deep layer neurons between treatment groups. Post hoc Student t tests indicated that all differences between scrambled and mir-92b sponge-treated brains were significant (P < 0.05). 6of6