MicroRNA-19b restricts Wnt7b to the hem, which induces aspects of hippocampus development in the avian forebrain

The avian hippocampus, which is known to be functionally similar to its mammalian counterpart (Clayton, 1998; Macphail, 2002), has been extensively studied for more than half a century, and yet very little is known about the mechanisms regulating its formation. Despite the significant similarities in gene expression patterns between the developing avian and mammalian hippocampus (Gupta et al., 2012), there are stark morphological differences in the adult hippocampus between the two classes (Atoji et al., 2016; Gupta et al., 2012; Atoji and Wild, 2006). This raises a question about the degree of evolutionary conservation between the developmental mechanisms that shape the hippocampus of birds and mammals. To address this, we decided to investigate the development of the avian hippocampus, beginning from when it is first induced.

In mammals, the cortical hem, which is first specified in the dorsomedial forebrain (Grove et al., 1998), later induces the hippocampus in the adjacent forebrain region (Mangale et al., 2008; Yoshida, 2006). Thus, the positioning of the hem itself is crucial for determining the position of the hippocampus. Although a few signalling molecules, (Subramanian and Tole, 2009) as well as transcription factors such as Lhx2, Emx2 and FoxG1 (Bulchand et al., 2001; Hanashima et al., 2007; Monuki et al., 2001; Shimogori et al., 2004), have been implicated in positioning the cortical hem, a comprehensive understanding of this process is still lacking. Studies carried out in mice have shown that Wnt secreted by the hem (Grove et al., 1998) is necessary and sufficient for hippocampus induction (Lee et al., 2000; Machon et al., 2007). Taken together with the fact that the hem is also necessary and sufficient for hippocampus development (Mangale et al., 2008; Yoshida, 2006), it is likely that the Wnt signal emanating from the cortical hem is the hippocampus-inducing factor in mammals.

In the avian forebrain, a tissue akin to the mammalian cortical hem has been described previously (Garda et al., 2002); however, its function has never been identified. Thus, we started our investigation on the avian hippocampus by trying to answer the following questions: (1) are the molecular mechanisms responsible for hippocampus induction conserved between birds and mammals; and (2) how is the hem positioned in the dorsomedial forebrain in birds?

We obtained some evidence for a molecular player involved in positioning the hem from a previous study, where we had examined the expression pattern of a microRNA, miR-19b, in the developing chick forebrain. We observed that miR-19b is expressed in the entire neuroepithelium, except in the hem region located in the dorsomedial forebrain (Fig. 1B,H, Fig. S1) at Hamilton-Hamburger stage 24 (HH24) (Hamburger and Hamilton, 1951). As Wnt ligands secreted from the hem are responsible for hippocampus development in the mammalian forebrain (Lee et al., 2000; Machon et al., 2007), we screened for the Wnt ligands expressed in the avian forebrain hem. We found that the hem region expresses Wnt2b, Wnt5b, Wnt7b, Wnt8b and Wnt9a, and that Wnt9a is also expressed in the midline (Fig. 1C-G) (Garda et al., 2002). The mutually exclusive expression of miR-19b and Wnt ligands in the hem (Fig. 1B-H) led us to speculate that miR-19b is likely to regulate the expression of one or more Wnt ligands. miRNAs are known to interact with their target mRNAs in a sequence-dependent manner (Bartel, 2009). When we used a sequence analysis tool, TargetScan (Agarwal et al., 2015), as well as a manual method to examine the mRNA sequence of the Wnt ligands (Fig. S2A), we found that only Wnt7b has a highly conserved miR-19b-binding site in its mRNA sequence (Table S1 and Fig. S2A), indicating that it is likely to be regulated by miR-19b.

To determine whether miR-19b interacts with Wnt7b mRNA, we carried out an in vitro sensor assay. For this, a Sensor-Wnt7b-WT construct, consisting of a mCherry reporter fused with a miR-19b target region of Wnt7b mRNA, was co-transfected with another construct expressing miR-19b (pRmiR-19b) (Smith et al., 2009) (Fig. S2B) into HEK293T cells. This resulted in significant repression of mCherry fluorescence (Fig. S2G). This repression was abolished when pRmiR-19b was co-transfected with Sensor-Wnt7b-MUT (Fig. S2J), a construct containing a mutation in the conserved nucleotides in the miR-19b target region of Wnt7b mRNA (pRmiR-Empty with Sensor-Wnt7b-WT, 22.82±0.98 A.U.; pRmiR-19b with Sensor-Wnt7b-WT, 8.59±1.42 A.U.; pRmiR-19b with Sensor-Wnt7b-MUT, 16.25±0.64 A.U.; Fig. S2C-K,L). This indicated that miR-19b can directly interact with the Wnt7b mRNA in vitro.

To investigate whether miR-19b can regulate Wnt7b expression in vivo, we first performed gain-of-function experiments with miR-19b. The avian retrovirus based-construct RCAS-miR-19b (Petropoulos and Hughes, 1991) (Fig. 2A) was electroporated in the presumptive medial forebrain at HH10, followed by harvesting and analysis of the brains at HH24 (Fig. 2B). We observed that there was a loss of Wnt7b expression, in the brains where miR-19b was misexpressed in the hem region, in comparison with the control (Fig. 2C-F). Next, we performed loss-of-function experiments for miR-19b using a CRISPR/Cas9-based strategy (Cong et al., 2013) to knockout the miR-19b genomic locus. We designed a construct pX330-19b-gRNA (Cong et al., 2013) that expresses Cas9 protein along with a guide RNA (19b-gRNA) that targets the miR-19b genomic locus (Fig. 2G, Fig. S3A). Before using this construct for in vivo experiments, we tested its efficacy using the following methods: (1) in vitro Cas9 endonuclease assay (see Materials and Methods, Fig. S3); (2) miRNA in situ hybridization for miR-19b (Fig. S4); and (3) DNA sequencing of the miR-19b genomic locus (see Materials and Methods, Fig. S5). Once the pX330-19b-gRNA construct was found to be effective in targeting the miR-19b genomic locus, we co-electroporated this construct with the pCAG-GFP construct, in the dorsolateral forebrain at HH22. We harvested and analysed the brain at HH24 (Fig. 2H) and found that there was an ectopic expression of Wnt7b in the region expressing GFP along with 19b-gRNA in comparison with the control forebrain (Fig. 2I-L′, Fig. S6). Thus, the gain-of-function and the loss-of-function experiments indicated that miR-19b is both necessary and sufficient to restrict the expression of Wnt7b to the hem region in the developing chick forebrain.

Various components of Wnt signalling have been shown to be necessary and sufficient for hippocampus development in the mouse brain (Lee et al., 2000; Machon et al., 2007). Hence, to determine whether Wnt7b can function as an inducer of the hippocampus, we misexpressed Wnt7b in different ectopic locations (dorsolateral, ventromedial and ventrolateral regions) in the developing chick forebrain. We observed that the ectopic expression of Wnt7b in the ventrolateral forebrain (Fig. 3I,I′, red dashed box region) resulted in the induction of three well-established hippocampal markers, namely neuropilin 2 (Nrp2), Neurod1 and Prox1 (Gupta et al., 2012) (Fig. 3J-L′,K″,L″, Fig. S7O) in this region. However, misexpression of Wnt7b in the sub-ventricular zone (SVZ) of the ventromedial forebrain (Fig. 3I,I*) resulted in the induction of expression of Prox1 alone (Fig. 3I*,L*). In addition, the dorsolateral forebrain also responded to Wnt7b by inducing the expression of only two hippocampal markers: Nrp2 and Neurod1 (Fig. S8H,I). Hence, the ventrolateral and ventromedial region seems to be the most and the least competent, respectively, to express hippocampal markers in response to Wnt7b (Fig. S9). The cellular architecture of the ectopic hippocampal region can be visualized in the high-magnification views showing cells co-stained using DAPI and NeuroD1/ Prox1 antibody (Fig. 3K″,L″), and in cells co-stained using Prox1 antibody and phalloidin to demarcate the actin cytoskeleton (Fig. S7).

Within the hippocampus, Prox1 expression is limited to a region close to the hem, whereas Nrp2 and Neurod1 are expressed in a broader domain (black or white boxed regions in Fig. 3E-G). This raises the possibility that Prox1 expression requires a higher amount of Wnt ligand secreted from the hem when compared with Nrp2 and Neurod1. To assess the dose dependence of Wnt ligand on the expression of hippocampal markers, we misexpressed Wnt7b by electroporating different concentrations of the RCAS-Wnt7b construct in the ventrolateral region at HH22, and harvested and analysed the brains at HH34 (Fig. 3B, red dashed box). The ventrolateral region was chosen as it is competent to express all three hippocampal markers: Nrp2, Neurod1 and Prox1. Based on the expression level of Wnt7b, assessed by mRNA in situ hybridization, we identified domains expressing relatively low and high amounts of Wnt7b in the ventrolateral forebrain (Fig. 3M,Q). We found that a low amount of Wnt7b can induce the expression of NeuroD1 alone; however, a relatively high amount can induce the expression of all three hippocampal markers: Nrp2, Neurod1 and Prox1 (Fig. 3N-P,R-T). Thus, it seems that Wnt7b acts in a dose-dependent manner for the induction of expression of hippocampal markers in the ectopic location in the developing forebrain. Notably, only a high amount of misexpression of Wnt ligand could induce Prox1 expression, which may be the reason why Prox1 is expressed closer to the hem.

Besides Wnt7b, other Wnt ligands are also expressed in the hem region, such as Wnt2b, Wnt5b, Wnt8b and Wnt9a (Fig. 1C-H). Hence, we tested whether Wnt7b is also capable of establishing the expression of other Wnt ligands in the hem. To ascertain this, we expressed a miR-19b-resistant version of Wnt7b (Wnt7b-19b-R) in the neuroepithelium adjacent to the hem region. This is because the endogenously expressed miR-19b in the neuroepithelium may suppress the misexpressed Wnt7b. We transduced this version of Wnt7b using a RCAS-Wnt7b-19b-R construct, by in ovo electroporation in the dorsolateral forebrain at HH22, and analysed the brains after 24 h. We found that the ectopic expression of Wnt7b-19b-R (Fig. S10E,E′) in the dorsolateral forebrain is capable of inducing the expression of two other canonical Wnt ligands: Wnt2b and Wnt9a (Fig. S10G,G′,I,I′) (Cho, 2006; Spater et al., 2006). Thus, Wnt7b likely enables the hem to secrete a few other canonical Wnt ligands, which may also pattern the adjacent forebrain region. To further determine whether canonical Wnt signalling is sufficient for inducing the hippocampal markers, we expressed the constitutively active form of β-catenin (CA-β-catenin) ectopically in the ventrolateral forebrain by co-electroporating the RCAS-CA-β-catenin (Funayama et al., 1995) and the RCAS-GFP constructs at HH22. We observed that, similar to Wnt7b, CA-β-catenin is also capable of inducing the expression of Nrp2, Neurod1 and Prox1 in the ventrolateral region at HH34 (Fig. 3U-X). Therefore, these data suggest that in both birds and mammals: (1) the hem secretes Wnt ligands (Garda et al., 2002; Grove et al., 1998); and (2) Wnt signalling is sufficient for the expression of hippocampal markers (Lee et al., 2000; Machon et al., 2007). Thus, it is likely that the function of the hem as an inducer of the hippocampus is evolutionarily conserved between these two classes.

We observed that Wnt7b is capable of inducing hippocampal markers in an ectopic location in the forebrain at HH34 (Fig. 3J′,K′,L′). In addition to this, we had observed earlier that the loss of function of miR-19b permitted the expression of Wnt7b in an ectopic location in the forebrain at HH24 (Fig. 2L,L″). Hence, as an extension to the early stage loss-of-function experiment for miR-19b, we wanted to test whether brain regions in which miR-19b has been knocked out can also induce hippocampal markers in an ectopic location at a later stage. For this, we analysed ventrolateral forebrain regions at HH34 co-electroporated with pX330-19b-gRNA and pCAG-GFP constructs (Fig. 4A). We found that, similar to the observation made at HH24, there was ectopic expression of Wnt7b (Fig. 4H″) at this stage in the regions of the forebrain where miR-19b had been knocked out (Fig. 2L,L″). These data show that the effect of miR-19b knockout on the expression of Wnt7b was consistent between the early and late stages of forebrain development. Upon analysis we found that there was an induction of the expression of all three hippocampal markers, i.e. Nrp2, Neurod1 and Prox1 (Fig. 4I″-K″), in the ventrolateral forebrain region. This demonstrated that miR-19b is necessary for the appropriate positioning of the hippocampus through the inhibition of Wnt7b expression in all regions of the avian forebrain except the hem.

To gather additional evidence in support of the observed effect of the knockout of the genomic locus of miR-19b, we used an alternative loss-of-function strategy involving a miRNA sponge (Ebert et al., 2007). We generated a sponge construct for miR-19b (RCAS-Sponge-19b) containing 12 copies of the target sequence of miR-19b. Next, we determined its efficacy (Fig. S12). We then electroporated RCAS-Sponge-19b and RCAS-GFP, as a control, in the lateral forebrain at HH22 (Fig. S11A-D), and found ectopic expression of Wnt7b in the location where Sponge-19b is misexpressed, at HH34 (Fig. S11K). We also observed that there was an induction of expression of only two hippocampal markers, Nrp2 and NeuroD1 but not Prox1 (Fig. S11L-N), in the Sponge-19b-expressing forebrain regions. In comparison, CRISPR/Cas9-mediated loss of function of miR-19b leads to the expression of all three hippocampal markers: Nrp2, Neurod1 and Prox1 (Fig. 4I″,J″,K″). Keeping in mind that Wnt7b acts in a dose-dependent manner for inducing the expression of hippocampal markers (Fig. 3M-T), it seems that the differential amount of expression of Wnt7b between the two methods used to achieve loss of function of miR-19b could explain the differences in the number of hippocampal markers induced in each case. These results reinforce the importance of miR-19b in suppressing hippocampal fate in other regions of the forebrain.

Throughout this study we have demonstrated that Wnt7b is sufficient for the induction of hippocampal markers in the avian forebrain. Furthermore, it appears that the Wnt-mediated mechanism of hippocampal induction, by the hem tissue, is conserved between birds and mammals, despite the fact that they have morphologically distinct hippocampi. In addition, we have identified a microRNA, miR-19b, that is crucial for restricting the expression of Wnt7b to the hem region. To the best of our knowledge, this is the only example of a microRNA that defines the position of the hem and the hippocampus. Thus, miR-19b is likely to be an important player in the patterning of the avian forebrain.

Eggs of White Leghorn strain were procured from Ganesh Enterprises (Nankari), Indian Institute of Technology Kanpur (Kanpur, Uttar Pradesh, India) and Central Avian Research Institute (Bareilly, Uttar Pradesh, India). Eggs were incubated at 38°C in a humidified incubator until desired Hamilton Hamburger (HH) stages (Hamburger and Hamilton, 1951).

Human embryonic kidney (HEK-293T, ATCC CRL-3216) fibroblast cell lines were cultured in Dulbecco's Modified Eagle's Medium (DMEM) (Sigma, D7777) with 10% foetal bovine serum.

The desired HH stage chick embryo heads were decapitated and brains were dissected out and fixed overnight in 4% paraformaldehyde (PFA) at 4°C. Then they were washed with phosphate-buffered saline (PBS), followed by dehydration in increasing concentration of sucrose up to 30% then embedded in Polyfreeze (Sigma, P0091). Thereafter 12-16 µm sections were prepared using a cryostat (Leica CM 1850).

For micro RNA in situ hybridization, we used the established protocol (Thompson et al., 2007). The probe for miR-19b was synthesized wherein the 3′ end of the antisense sequence of miR-19b (Eurofins MWG Operon) was labelled with digoxigenin. This was detected using the established method (Trimarchi et al., 2007). For mRNA in situ hybridization, an established protocol was used to detect mRNAs in the tissue sections (Trimarchi et al., 2007). The antisense RNA probes were prepared from the following cDNA clones: (1) Wnt2b, Wnt5b, Wnt7b, Wnt8b and Wnt9a (gifts from Prof. Cliff Tabin, Harvard Medical School, Boston, MA, USA); (2) Nrp2 (ChEST739d12), this chicken EST (ChEST) clone was obtained from Source Bioscience.

The primary sequence of miR-19b was synthesized (Macrogen) and cloned in the pRmiR vector after BsaI digestion (Smith et al., 2009) to obtain pRmiR-19b.

The pRmiR-19b vector was digested with ClaI restriction enzyme and the insert GFP-miR-19b was cloned into RCAS BP(A) vector (Petropoulos and Hughes, 1991) at the ClaI site.

These two constructs were generated by ligating the putative miR-19b target region (see below) present in the mRNA of chick Wnt7b to obtain the sensor-Wnt7b-wild type (sensor-Wnt7b-WT) construct and by ligating the mutated target region (see below) to obtain sensor-Wnt7b-mutated (Sensor-Wnt7b-MUT) in the pCAG-mCherry vector (a gift from Prof. Constance Cepko, Harvard Medical School, Boston, MA, USA) between NotI and HindIII restriction sites. The chicken Wnt7b wild-type target region (NM_001037274.1; from 583rd nucleotide position to 634th nucleotide position) was 5′GAAATCAAAAAGAATGCACGAAGGCTGATGAACTTGCACAACAATGAGGCT3′ and the chicken Wnt7b mutated target region (in bold letters) was 5′GAAATCAAAAAGAATGCACGAAGGCTGATGAACTTCGTCAACAATGAGGCT3′.

To generate the Sponge-19b construct, first Sponge-19b-4x was generated by ligating four copies of the reverse complement sequence of miR-19b (in bold letters) with two nucleotide deletion (Ebert and Sharp, 2010): 5′TGCTGCCGtcagttttgcaccttttgcacaCCGGtcagttttgcaccttttgcacaCCGGtcagttttgcaccttttgcacaCCGGtcagttttgcaccttttgcacaCCGG3′. This was followed by cloning it in the pRmiR vector (Smith et al., 2009). This Sponge-19b-4x sequence was then chained three times in the same Sponge-19b-3x vector by ligation using SalI, BglII and BamHI restriction sites to obtain Sponge-19b-12x sequence in pSponge19b vector, with 12 copies of sponge sequence for sequestering miR-19b. Subsequently, we subcloned the Sponge-19b-12x sequence into the RCAS(A)BP vector (Petropoulos and Hughes, 1991) using ClaI restriction enzyme to generate the construct RCAS-Sponge-19b.

To test the efficacy of pSponge-19b, we have used the sensor assay with Sensor-Wnt7b constructs, in which HEK cells were co-transfected using Turbofect (Thermo Fisher Scientific Inc., R0533) as per manufacturer's protocol with the following constructs: (i) pRmiR-Empty vector and Sensor-Wnt7b-WT at 6:1 molar ratio, (ii) pRmiR-Empty vector: pRmiR-19b: Sensor-Wnt7b-WT at 3:3:1 molar ratio and (iii) pRmiR-19b:pSponge-19b: Sensor-Wnt7b-WT at 3:3:1 molar ratio. We found that Sponge-19b was able to rescue the repression of mCherry expression caused by miR-19b expression as shown in (Fig. S12A-I) and quantified as mentioned below: pRmiR-Empty and Sensor-Wnt7b-WT: 31.19±0.18 A.U., pRmiR-Empty, pRmiR-19b and Sensor-Wnt7b-WT: 15.7±0.71 A.U., Sponge-19b, pRmiR-19b and Sensor-Wnt7b-WT: 37.44±0.09 A.U.; Fig. S12J). This suggested that Sponge-19b possibly sequestered miR-19b and hence the expression of Sponge-19b can result in loss of function of miR-19b.

To generate these constructs, we analysed miR-19b DNA sequence (NR_031404.1) using the crispr.mit.edu web tool and obtained possible gRNA sequences. Among these gRNA sequences, one was selected from within miR-19b locus. Additional nucleotide sequences, i.e. CACC and AAAC, were then added to the 5′ end of the top and bottom strand, respectively, to generate ends cohesive with the linearized pX330-gRNA vector (Cong et al., 2013) digested with BbsI restriction enzyme. Top-strand (5′CACCGACAGCAGAGTATCATACAGC3′) and bottom-strand (5′AAACGCTGTATGATACTCTGCTGTC3′) DNA sequences of 19b-gRNA were annealed and ligated with the pX330 construct (Cong et al., 2013).

The chick mir-19b genomic locus was targeted using gRNA (the miR-19b-5p sequence is written in lowercase letters and the miR-19b-3p sequence is written in bold letters) to cause double-strand break at the site marked with a solidus: 5′CACTGTTCTCTGGTTagttttgcaggtttgcatccagct/GTATGATACTCTGCTGTGCAAATCCATGCAAAACTGACTGTGGCAGTG3′

To generate this construct, we mutated the coding sequence (CDS) of Wnt7b such that it alters the miR-19b-binding site in the CDS region of Wnt7b without changing the amino acid sequence of Wnt7b protein. This CDS was the first cloned in a pSlax21 shuttle vector at the NcoI and EcoRI sites. Finally, this fragment was sub-cloned from the pSlax21 construct to the RCAS construct (Petropoulos and Hughes, 1991) at the ClaI restriction enzyme site.

The RCAS-CA-β-Catenin construct was a gift from Prof. Clifford Tabin (Harvard Medical School, Boston, MA, USA) and pCAG-GFP construct was a gift from Prof. Constance Cepko.

To verify that miR-19b guide RNA expressed using pX330-19b-gRNA construct can knock out the miR-19b genomic locus, an in vitro Cas9 assay was performed. In this assay, target DNA, which is a 761 bp (NC_006088.5:149439866-149440626) genomic region harbouring the miR-19b genomic locus, was amplified using a PCR method with following sets of primers: forward, 5′GTGCTTACAGTGCAGGTAGTGATA3′; reverse, 5′GACAAGTGCAATACCACAGAAACG3′. To obtain miR-19b guide RNA, in vitro transcription was performed. For in vitro transcription, the DNA template was amplified from the pX330-19b-gRNA construct using PCR with the following primers (T7 RNA polymerase in bold letters): forward, TTAATACGACTCACTATAGGACAGCAGAGTATCATACAGC; reverse, AAAAGCACCGACTCGGTGCC. In vitro transcription was performed using the Life Technologies/MEGAshortscript T7 Transcription Kit (Ambion, AM 1354). To test the activity of miR-19b, a guide RNA endonuclease reaction was then set up with Cas9 enzyme (PNA Bio – CP01) in a 15 μl reaction with 1 μg of Cas9, ∼300-400 ng of single guide RNA with 230 ng of target DNA.

To verify that the guide RNA against miR-19b is functional and can target the genomic locus of miR-19b, the miR-19b guide RNA-expressing construct (pX330-19b-gRNA) was co-electroporated with pCAG-GFP in the HH22 chick forebrain and harvested after 24 h. The GFP-expressing brain region was dissected out using fine forceps. Total genomic DNA was extracted from the tissue using Trizol (Sigma-Aldrich, #93289) according to the manufacturer's protocol. The miR-19b genomic locus was then amplified from this mixture of total genomic DNA using following a set of primers in polymerase chain reaction using Phusion Polymerase (NEB, #M0530) according to the manufacturer's protocol: forward, TTAGGTTTCCTACATCAATGAG; reverse, AACTGATGGTGGCCTGTTATAA. This amplified the miR-19b genomic locus with flanking regions, i.e. a 508 bp DNA fragment. The mixed pool of 508 bp DNA fragments obtained was analysed by Illumina Sequencing systems. The Contig(s) were analysed using NCBI blast and aligned with the sequence of the miR-19b genomic locus.

Human embryonic kidney fibroblast 293T cells at 60-70% confluency were transfected using Turbofect (Thermo Fisher Scientific, R0533) according to the manufacturer's protocol, with pRmiR vector(s) and respective sensor vector(s) in a 6:1 molar ratio. GFP and mCherry fluorescence were observed after 72 h.

After incubating the eggs for 24 h, 2-3 ml of albumin was taken out in order to lower the developing chick embryo. A window was then created on the top of the egg and DNA constructs (2 µg/µl) with 0.1% of Fast Green (Sigma-Aldrich, 44715) were added to one of the forebrain vesicles. Platinum hockey stick electrodes (Nepa Gene, CUY611P3-1) were positioned on this forebrain vesicle such that the positive electrode was placed over the dorsal forebrain while the negative electrode was placed below the ventral side of the forebrain. Five electric pulses of 13 V each for embryos at HH22 and 10 V each for embryos at HH10 were applied for the 50 ms duration with a gap of 950 ms, using a square wave pulse electroporator (Nepa Gene, CUY21SC). Sterile PBS containing penicillin and streptomycin was layered over the embryos and the egg was sealed with tape. Eggs were put back into the egg incubator until the desired HH stage of development and then harvested for tissue section preparation.

We used an established protocol for immunostaining. Briefly, after a 5 min fixation in 4% PFA, three washes with PBS were carried out followed by 1 h of blocking with 5% of heat-inactivated goat serum (HINGS) in PBS. After that, the primary antibody at the required dilution was layered on the slide in the blocking reagent and incubated overnight at 4°C. The primary antibodies were used at the following concentrations: anti-NeuroD1 antibody, 1:1500 (Abcam, ab60704); and anti-Prox1 antibody, 1:300 (Sigma, P7124). Then slides were washed three times in PBS for 5 min each and the secondary antibody (1:250) was layered on the slides followed by incubation for 1 h at room temperature. The secondary antibodies used were as follows: Alexa Fluor 594-conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories, 115-585-003) and Alexa Fluor 594-conjugated anti-rabbit IgG (Jackson ImmunoResearch Laboratories, 111-585-045).

We performed phalloidin staining using Alexa Fluor 594-conjugated phalloidin (Thermo Fisher Scientific, A12381) at 1:10 dilution as per the manufacturer's protocol.

A Leica stereomicroscope (DM500B) equipped with a DFC500 camera was used to capture RNA in situ hybridization images and fluorescent images. Phalloidin-stained sections were imaged using a Leica confocal microscope (SP5).

All experiments were repeated at least three times; for animals, n=1 means experiments performed on one animal; for cell culture, n=1 means experiment performed in one dish. We included only those samples where we observed transfection or electroporation assessed by GFP and/or mCherry expression. Fluorescence intensity was quantified using ImageJ software (imagej.nih.gov/ij/). We were blinded to the groups of controls and the test during the analysis of quantification. Graphs were plotted using Origin Pro 9.1 software (www.originlab.com/) and to test the statistical significance of the data, an unpaired Student's t-test was carried out using Graph Pad Prism 7.0 (www.graphpad.com/quickcalcs/ttest1.cfm).

We acknowledge Prof. Amitabha Bandhopadhyay and Prof. Shubha Tole for valuable suggestions on designing experiments, as well as for critical comments on the manuscript. We are grateful to Dr Thamarai Lingam for help with imaging phalloidin-stained brain sections. We are thankful to Ms Archita Mishra for providing CA-β-catenin-electroporated brain samples. We are also thankful to Mr Naresh Gupta for technical support.