The HiB5 neuronal precursor cell line was cultured by previously described methods (Renfranz et al., 1991). Cells were maintained in Dulbecco's modified Eagle's medium (Gibco) with 10% fetal bovine serum (Gibco) supplemented with 0.11 g/l sodium pyruvate, 3.7 g/l NaHCO₃, 0.29 g/l glutamine, 0.6 g/l penicillin, and 0.1 g/l streptomycin. Medium was changed every three days and cells were incubated at 33℃ in 5% CO₂. HiB5 cells proliferate at 33℃, the permissive temperature of the oncogenic tsA58 allele of the SV40 large T antigen.

HiB5 cell line was cultured in 100 mm dishes for RNA isolation and plated onto glass coverslips in 24-well plates for immunocytochemistry at 75~80% confluence. The cells were transiently transfected using the calcium-phosphate co-precipitation method (Pear et al., 1993). Two µg/ml of the vector encoding 2.5 kb of the mouse Foxg1 cDNA, which was subcloned into the pEGFP C2 (clontech) expression vector, was used for transfection. After transfection, cells were incubated in serum-free medium including transfection solution for 12 hr. For transient gene expression assay, cells were incubated in complete medium after removing calcium-phosphate containing medium for 24 hr.

R18F1 and HiB5 cells were plated onto glass coverslips at a density of 75~80% confluence in 24-well plates. The transfected cells were fixed in 4% paraformaldehyde and permeabilized with 0.1% TritonX-100. Coverslips were incubated with the primary antibody (anti-Foxg1, 1:1,500) for 3 hr, and then incubated for 1 hr with a goat anti-rabbit rhodamine-conjugated antibody (Leinco, 1:300). The stained coverslips were observed under fluorescence microscope (Olympus PROVIS AX-70, X200).

Total RNA was isolated from cultured HiB5 cell line using TRIzol (GibcoBRL), and cDNA was made using 1µg of RNA and AMV reverse-transcriptase (Promega). Primer sequences for Foxg1 were forward (5'-GGGCAACAACCACTCCTTCTCCAC-3') and reverse (5'-GACCCCTGATTTTGATGTGTGAAA-3'). The expected size of the Foxg1 product was 396 bp. PCR cycling conditions were 94℃ for 30 s, 65℃ for 30 s, and 72℃ for 90 s, for a total of 30 cycles. The PCR products were electrophoresis on the 1.5% agarose gel and stained with ethidium bromide.

Foxg1-overexpressing HiB5 cell line was used as sources of RNA for ODD. The techniques used for ODD have been described (Matz et al., 1997). In brief, total RNA used for differential display was further purified to remove DNA contamination using the MessageClean Kit (GenHunter Corp) and a commercially available kit (Boehringer Mannheim) was used for the synthesis of double-stranded cDNA from 1µg of total RNA, except the T-primer for first strand synthesis provided with the kit that should be substituted for non-extended ODD T-primer. cDNA species were discriminated by the length of fragment between polyA attachment site and the first occurrence of site for restrictase, RsaI, and then ligated to 5'-ends of full length cDNA digests with ODD adaptor. After the generation of amplified cDNA, the subsets of cDNA were produced by PCR with both a T-primer and an adaptor-specific primer extended by two arbitrary bases at their 3'-ends. The population of cDNA was subdivided into 192 subsets (as there exist 16 possible variants of adaptor-specific primer extension and 12 of T-primer), which were displayed on an ordinary 6% sequencing gel. The gel was transferred to 3 M filter paper and dried. By terminal labeling of extended adaptor-specific primer with γ-³³P[dATP], the displayed bands were detected on the exposed X-ray films.

The PCR products were excised from the dried sequencing gel and eluted by boiling in 20µl of H₂O for 10 min. The eluted cDNA fragments were subjected to PCR reamplification using the appropriate extended T-primers and adaptor-specific primers. The reamplified fragments of cDNA were fractionated on 2% agarose gels and extracted. One µl of the reamplified cDNA fragments directly applied by diffusion as targets on a nitrocellulose membrane and cross-linked by UV-illuminator. Two copies of the membrane were prepared and hybridized overnight with [α-³²P]dCTP-labeled cDNA probes reverse-transcribed from the original RNA preparation of either the GFP or the GFP-Foxg1 transfected cells. After hybridization, the membranes were exposed to X-ray films at -80℃. If differential expression was confirmed by reverse northern blot analysis, the band was cloned and sequenced as described below.

The differentially expressed cDNA fragments confirmed by reverse northern blotting were cloned into plasmid vector pCRII using the TOPO TA cloning system (Invitrogen). The resulting plasmids were confirmed by restriction enzyme digestion and performed automatic sequencing (KAIST BioMedical Research Center). Partial length sequences were compared to all previously reported gene sequences in GeneBank.

Mouse brains (P0) were fixed at 4℃ in 4% paraformaldehyde for 4 hr, immersed in 20% sucrose overnight and embedded in OCT (Tissue-Tek). Sections were cut at 10µm on a cryostat. Nonradioactive in situ hybridization was performed as described by Gradwohl (Gradwohl et al., 1996). Antisense RNA probe from partial fragment of Mss4 were prepared by in vitro transcription of the linearized DNA templete in the presence of digoxigenin-11-UTP (Boehringer Mannheim). The substrate for the chromogenic reaction of alkaline phosphatase was BM Purple (Boehringer Mannheim). Stained sections were mounted and photographed with Olympus PROVIS AX-70.

It is known that the Foxg1 is highly expressed in the developing mammalian forebrain. Homozygous null Foxg1 mutants die at birth, and have a dramatic reduction in the size of the cerebral hemispheres and multiple developmental anomalies of the eyes (Tao and Lai, 1992; Xuan et al., 1995; Huh et al., 1999). To study the developmental mechanism of Foxg1 in the hippocampus through the regulation of the downstream genes, we used neuronal precursor cell line which was originated in the hippocampus. As the first step to examine the molecular mechanisms of Foxg1, HiB5 cell line was transfected with Foxg1 subcloned into the pEGFP C2. Total RNAs from the transfected cells were reverse-transcribed into cDNAs and the cDNAs were used in PCR reaction. Foxg1 was weakly expressed in the normal state but increased in the transfected cells (Fig. 1A). According to the results of immunocytochemistry using specific antibody, Foxg1-expressing cells were identical with the GFP-expressing cells under the fluorescence microscopy. Especially, Foxg1 expressed in the transfected cells was localized in the nucleus (Fig. 1B). These results indicate that mRNA and protein level of Foxg1 were simultaneously increased in the Foxg1-transfected cells.

We compared the profiles of gene expression between the GFP- and GFP-Foxg1-transfected cell line by ordered differential display, using 192 sets of primer pairs as described previously (Matz et al., 1997). ODD often shows false positive bands because of differences in the efficiency of reverse transcription or PCR among each sample. To overcome this problem, we duplicated each sample and remove contaminating DNA in total RNA using DNase I. Double-stranded cDNAs synthesized from isolated RNAs were used to prepare representative pools of 3' cDNA fragment from polyA-trac to the first occurrence of RsaI recognition site, and 3' cDNA fragments were amplified (Fig. 2A). Adapter-ligated 3'-end cDNA fragments were subjected to ODD. For each sample, 192 different combinations of primer sets made of 12 extended T-primers and 16 adaptor-specific primers were used for PCR amplification to generate differential displays. PCR products were separated by PAGE and visualized on the exposed X-ray films. The vast majority of the PCR products were common in all of the samples (Fig. 2B). A total of 98 differentially displayed bands designated as 1 to 100 were excised and recovered from the dried gels, and re-amplified using the corresponding primer sets. All of cDNAs were re-amplified with sizes between 100 ~500 bp.

Those cDNA fragments were recovered from the gel and reamplified by PCR with the same primer sets. To confirm the expression patterns observed in the sequencing gels, RNAs from the GFP- and GFP-Foxg1-transfected cell line was radioactively labeled and used as probes for reverse-northern blot analysis of all 100 reamplified cDNA fragments. Seventeen percent (17 out of 100) of these reamplified cDNA fragments were certified to the original observed displayed patterns (Fig. 3). For internal control, cyclophilin was used. Reverse-northern blot analysis confirmed that these clones corresponded to mRNA that was differentially expressed in response to Foxg1.

Differentially expressed cDNA fragments were subcloned into pCRII cloning vector and sequenced automatically using M13 primer. The unique sequence of the differentially expressed cDNA fragments were compared to GeneBank entries using BLASTn (Altschul et al., 1997). All 17 confirmed cDNA fragments were successfully cloned into TA cloning vectors for further analysis. The above cloned cDNA fragments were subjected to DNA sequencing. All sequences of 17 cDNA fragments were flanked by sequences derived from the extended T-primers and adaptor-specific primers used in PCR amplification. The features of each clone are summarized in Table 1. Homology searches against GeneBank database revealed that nine fragments are not homologous to any known genes and the other eight fragments contain partial cDNA sequences for known genes (Table 1).

Using non-radioactive in-situ hybridization, Mss4 mRNA level was examined in the heterozygote and homozygote Foxg1 mutant mouse brain. In the Foxg1(+/-) mouse hippocampus, the specific Mss4 signal was detected (arrows in Fig. 4A). However, in the Foxg1(-/-) mutant hippocampus, the specific Mss4 massage was not detected in the presumptive hippocampal region (arrows in Fig. 4), suggesting that Mss4 expression might be dependent on the normal expression of Foxg1 during development of hippocampus.