JBC Biosymposia, Inc.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 QUICK SEARCH:   [advanced]


     


Originally published In Press as doi:10.1074/jbc.M110023200 on December 13, 2001

J. Biol. Chem., Vol. 277, Issue 9, 7598-7609, March 1, 2002
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
277/9/7598    most recent
M110023200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Mitchelmore, C.
Right arrow Articles by Jensen, N. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Mitchelmore, C.
Right arrow Articles by Jensen, N. A.
Social Bookmarking
 Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

Characterization of Two Novel Nuclear BTB/POZ Domain Zinc Finger Isoforms

ASSOCIATION WITH DIFFERENTIATION OF HIPPOCAMPAL NEURONS, CEREBELLAR GRANULE CELLS, AND MACROGLIA*

Cathy Mitchelmore, Karen M. Kjærulff, Hans C. Pedersen, Jakob V. Nielsen, Thomas E. Rasmussen, Mads F. Fisker, Bente FinsenDagger , Karen M. Pedersen, and Niels A. Jensen§

From the Laboratory of Mammalian Molecular Genetics, The Panum Institute 6.5, University of Copenhagen, Blegdamsvej 3, DK-2200 Copenhagen N and the Dagger  Department of Anatomy and Neurobiology, University of Odense, Winsløwparken 21, DK-5000 Odense C, Denmark

Received for publication, October 17, 2001, and in revised form, December 10, 2001

    ABSTRACT
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

BTB/POZ (broad complex tramtrack bric-a-brac/poxvirus and zinc finger) zinc finger factors are a class of nuclear DNA-binding proteins involved in development, chromatin remodeling, and cancer. However, BTB/POZ domain zinc finger factors linked to development of the mammalian cerebral cortex, cerebellum, and macroglia have not been described previously. We report here the isolation and characterization of two novel nuclear BTB/POZ domain zinc finger isoforms, designated HOFL and HOFS, that are specifically expressed in early hippocampal neurons, cerebellar granule cells, and gliogenic progenitors as well as in differentiated glia. During embryonic development of the murine cerebral cortex, HOF expression is restricted to the hippocampal subdivision. Expression coincides with early differentiation of presumptive CA1 and CA3 pyramidal neurons and dentate gyrus granule cells, with a sharp decline in expression at the CA1/subicular border. By using bromodeoxyuridine labeling and immunohistochemistry, we show that HOF expression coincides with immature non-dividing cells and is down-regulated in differentiated cells, suggesting a role for HOF in hippocampal neurogenesis. Consistent with the postulated role of the POZ domain as a site for protein-protein interactions, both HOF isoforms are able to dimerize. The HOF zinc fingers bind specifically to the binding site for the related promyelocytic leukemia zinc finger protein as well as to a newly identified DNA sequence.

    INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

The hippocampus is located at the medial-temporal edge of the neocortex and harbors neural circuitry that is crucial for cognitive functions such as learning and memory (1). Anatomically, it can be distinguished as two distinct subregions, the dentate gyrus (DG)1 and the hippocampus proper (or Ammon's horn) with its three subfields CA1, CA2, and CA3 (2). Lateral to the CA1 subfield lies the subiculum, pre-, and parasubiculum, which are regarded as transitional neocortical areas located between the hippocampus proper and the entorhinal cortex (3). During early development of the cerebral cortex, the hippocampal neuroepithelium is organized into distinct germinative zones that specify the development of CA pyramidal neurons, DG granule neurons, and fimbrial glial cells (4, 5). Newly born Ammon's horn pyramidal precursor neurons exhibit a delay in migration to the hippocampal plate (HP) in which they reside temporally in an inferior band referred to as the intermediate zone (6). After sojourn in the intermediate zone, presumptive pyramidal precursors settle in the HP plate in an "inside-out" sequence where first born neurons constitute deep layers and later generated neurons settle above in more superficial (external) locations. In contrast to morphogenesis of the neocortex where cortical plate neurons segregate radially into layers of functionally distinct neurons, pyramidal neurons of the Ammon's horn remain settled in a single homogenous layer.

Granule cell neurogenesis in the hippocampus is complex and involves at least two germinative zones. An early wave of migrating granule cell precursors originate in a distinct subdomain of the hippocampal neuroepithelium, designated the dentate notch, which consists of a mixture of post-mitotic and mitotic cells (7). The former cells settle in the dentate plate (DP) scaffold, where deposition of these pioneer granule cells begins in the suprapyramidal limb and proceeds into the infrapyramidal limb, establishing the suprapyramidal-to-infrapyramidal morphogenic gradient of DG scaffolding. The migrating mitotic granule cell precursors are destined for the tertiary germinative matrix in the hilar region, which constitutes an ectopic proliferative zone that persists into early postnatal development giving rise to the majority of post-mitotic granule cells that settle in the DP in the late embryonic and early postnatal period. Deposition of late born granule neurons in the DP follows an "outside-in" rule in which the youngest post-mitotic cells are located in deep layers next to the hilar region, and older cells are located in superficial layers close to the outer molecular layer (7). During the 1st month of postnatal development, the tertiary germinative matrix is replaced by the subgranular zone that forms at the border between the hilus and the granule cell layer and which is the major site of DG neurogenesis in adults.

As a first step toward identifying factors that might be involved in defining hippocampal territory during early development, we describe the identification and preliminary characterization of a novel BTB/POZ-zinc finger gene, designated HOF. Alternative splicing gives rise to two protein isoforms that are nuclear, can dimerize with each other, and bind specifically to DNA. In the developing cerebral cortex, HOF expression is confined to neurons of the presumptive hippocampus and is transiently up-regulated in immature post-mitotic neurons derived from the hippocampal neuroepithelium and dentate notch. Postnatally, HOF expression was observed in a subpopulation of subventrical zone progenitors, in glial cells, and in developing cerebellar granule cells.

    EXPERIMENTAL PROCEDURES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Cell Culture-- The semi-differentiated oligodendrocytic cell lines 6E12 and 2H5 were established from the spinal cord (8) and the brain,2 respectively, of transgenic mice expressing the SV40 large T-antigen in oligodendroglia (9). The 2H5 and 6E12 cell lines were cultured in Dulbecco's modified Eagle's medium supplemented with 2% fetal bovine serum as described previously (8). Cortical primary cells were prepared by dissociation of medial temporal lobes from embryonic day 18 mice, and cells were plated on poly-D-lysine-coated coverslips for 2 days in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 0.1 IU/liter insulin, 7.3 µM p-aminobenzoic acid, and 50 IU/liter penicillin.

Library Screening and Plasmid Constructions-- First-strand cDNA was prepared from poly(A)+ RNA from the 2H5 cell line. Hybridization of 400 ng of 2H5 cDNA with 13 µg of poly(A)+ RNA from the 6E12 cell line was followed by chemical cross-linking and labeling as described previously (10), to generate a subtracted single-stranded cDNA hybridization probe. Approximately 80,000 clones from a non-amplified 6E12 cell cDNA library (10) were screened by hybridization to the subtracted probe. From the primary screening, 60 clones were pooled and rehybridized to the same probe. Of these, 40 clones hybridized in a reproducible manner to the subtracted probe. Positive phage clones from the secondary screening were excised as pZL1 plasmid clones (Invitrogen). The cDNA insert of the clone encoding HOF was labeled by random priming and used to screen the 6E12 cDNA library or a mouse genomic DNA library in lambda  Fix II (Stratagene). DNA sequencing was performed using the dye terminator method and an automated sequencer (Amersham Biosciences). PCRs carried out on genomic DNA using Advantage 2 Polymerase (CLONTECH) and restriction digests of positive lambda  clones were cloned in pBluescript (Stratagene) or pCR4-TOPO (Invitrogen) for sequencing.

To construct the plasmid Koz-L, the open reading frame of HOFL was amplified with Pfu Turbo polymerase (Stratagene) using a 5' primer that introduced a NotI site and a Kozak consensus sequence before the first ATG (5'-GCGGCCGCCACCATG-3'), together with a 3' primer which introduced a SalI site after the stop codon, and the resulting product was cloned into pBluescript. The open reading frames for HOFS and HOFL were PCR-amplified and cloned into the SmaI and NotI sites of pGEX-6P-2 (Amersham Biosciences). For expression of GST-HOF/Fingers, a PCR product encoding amino acids 573-741 of HOFL was cloned into the EcoRI and SalI sites of pGEX-6P-2.

Protein Studies and Western Blot Analysis-- Whole cell extracts were prepared by homogenizing cells from a confluent 75-cm2 dish in 100 µl of RIPA buffer containing 450 mM NaCl and 0.2% v/v protease inhibitor mixture (Sigma). After a 30-min incubation on ice, the extracts were spun at 14,000 × g for 10 min at 4 °C, and the supernatant was recovered. Nuclear extracts were prepared from nuclear pellets, obtained after removing the cytoplasmic extract with hypotonic buffer, as above. Protein concentration was measured using Bio-Rad dye reagent, and 10 µg of each extract was loaded on a 4-12% Tris glycine gel (NOVEX). Western blots were probed with affinity-purified anti-HOF (1:5000 dilution) using Western Breeze (NOVEX). The HOF polyclonal antibody was raised in rabbits against a peptide epitope (CPAKFDQIEQFNDHMR) located in the C-terminal end of HOF. [35S]Methionine-labeled proteins, prepared using HOF-S, HOF-L, or Koz-L plasmids or vector alone with the T7 coupled Transcription/Translation system (Promega), were run in adjacent lanes and subjected to autoradiography.

Induction of GST fusion protein expression in Escherichia coli BL21 codon + RIL cells (Stratagene) and binding to glutathione-Sepharose beads were carried out using standard procedures (Amersham Biosciences). Beads with bound GST or GST fusion protein (2 µg) were incubated with 2.5 (HOFS) or 5 µl (HOFL) of [35S]methionine-labeled protein in 200 µl of BB binding buffer (20 mM HEPES, pH 7.4, 100 mM KCl, 5 mM EDTA, 5 mM EGTA, 10% glycerol (v/v), 0.4% Nonidet P-40, 1 mM dithiothreitol, and 15% BSA (w/v)) for 1 h at room temperature. The beads were washed twice in BB + 10% BSA, three times in BB + 5% BSA, and three times in BSA-free BB. Bound proteins were released by boiling in SDS sample buffer and resolved on a 4-12% Tris glycine gel (NOVEX) followed by autoradiography.

Yeast Two-hybrid System-- The open reading frames for HOFS and HOFL were PCR-amplified and cloned in frame with the DNA binding domain of GAL4 in pGBKT7 (CLONTECH). The regions encoding the POZ domains of HOFS (amino acids 1-127) and HOFL (amino acids 1-200) were PCR-amplified and cloned in frame with the activation domain of GAL4 in pGADT7 (CLONTECH). The pGBKT7 and pGADT7 derived plasmids were co-transformed into the yeast strain AH109 and maintained by growth in media lacking leucine and tryptophan. Reporter gene expression was tested by spotting ~500 cells from each co-transformant on selective media in the absence of histidine and adenine and in the presence of X-alpha -galactosidase (CLONTECH).

DNA Binding Studies-- A pool of double-stranded oligomers containing a random 20-mer in the middle was prepared as described (11). Binding reactions were carried out with 500 ng of nonradioactive oligomers and 2 µg of GST-HOF/Fingers protein bound to glutathione-Sepharose beads in a buffer containing 25 mM Tris-HCl, pH 7.9, 0.5 mM EDTA, 50 mM KCl, 5 mM MgCl2, 100 µM ZnCl2, 5% Ficoll 400 (w/v), 0.1% Nonidet P-40 (v/v), 1 mM dithiothreitol, 100 µg/ml BSA (w/v), and 200 ng of poly(dI-dC) (buffer A). After four rounds of selection and PCR-based amplification of recovered DNA sequences (11), the resulting DNA was radioactively labeled by incorporation of [alpha -32P]dCTP (3000 Ci/mmol) and used in electrophoretic mobility shift assay (EMSA) with GST or GST-HOF/Fingers protein. The retarded complex seen specifically with the fusion protein was cut off the gel, eluted, and cloned in the pCR4-TOPO vector (Invitrogen). A total of 25 clones were sequenced. Because no obvious consensus was derived, double-stranded oligonucleotides corresponding to the sequences shown in Fig. 2C and containing 5'-TCGA overhangs were synthesized.

Double-stranded oligonucleotides were labeled by filling in overhangs with [alpha -32P]dCTP using the Klenow fragment of DNA polymerase I. EMSA assays were carried out as previously described (12) with 1 µg GST-HOF/Fingers protein and 6.5 ng labeled PLZF probe in buffer A containing 1 µg poly(dI-dC) per reaction. Where indicated, a 10 min preincubation of protein with cold competitor double-stranded DNA (100 ng) was included. Samples were loaded onto a 6% polyacrylamide gel (NOVEX) with 0.5 × TBE running buffer (45 mM Tris borate, pH 8.3, 1 mM EDTA). Electrophoresis was carried out (100 V, 1-2 h), and the gel was dried and subjected to autoradiography.

Immunohistochemistry and in Situ Hybridization-- For immunohistochemistry, brains were removed from mice and immediately frozen in liquid N2. Cryosections (30 µm) were collected on gelatinized glass slides, fixed with methanol for 5 min, and preincubated for 15 min in phosphate-buffered saline containing 2% BSA to reduce background. Cells grown on coverslips were washed twice with phosphate-buffered saline before being subjected to immunohistochemistry.

Immunohistochemistry was carried out as described previously, using phosphate-buffered saline, 2% BSA as the antibody diluent (13). The primary antibodies used at 1:100 dilution are mouse 2',3'-cyclic nucleotide 3'-phosphodiesterase (CNCase; Neomarkers), mouse NeuN (Chemicon), rabbit glial fibrillary acidic protein (GFAC; Sigma), rabbit c-Myc (Santa Cruz Biotechnology), rabbit Oct-6 (a kind gift from Michael Wegner), and rabbit HOF antisera. For staining with sheep bromodeoxyuridine antisera (1:200 dilution, Research Diagnostics), the fixed sections were first treated for 30 s each with 2 M HCl and 0.1 M borate. Primary antibodies were detected with fluorescein isothiocyanate- or tetramethylrhodamine B isothiocyanate-conjugated swine anti-rabbit, 7-amino-4-methylcoumarin-3-acetic acid-conjugated goat anti-rabbit, fluorescein isothiocyanate-conjugated goat anti-mouse (DAKO), and donkey anti-sheep (Research Diagnostics Inc.) secondary antibodies at 1:80 dilution. In situ hybridization with an alkali phosphatase-conjugated oligonucleotide probe (5'-XTCAGCTGTCTTGGGTTTCTTCCGTTCTAGC-3', DNA Technology) was carried out as described previously (10).

    RESULTS
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Cloning and Characterization of Two HOF Isoforms-- The cDNA encoding the novel BTB/POZ domain factor, HOF, was isolated by a subtractive cloning technique originally aimed at identifying novel mRNAs expressed in brain oligodendroglia (see "Experimental Procedures"). The HOF cDNA was used to rescreen an oligodendrocyte cDNA library, resulting in the identification of several alternately spliced cDNAs that yield two protein products, HOFL and HOFS. The HOF-L1 cDNA is 2990 bp long and contains one long open reading frame from nucleotides 359-2584, encoding HOFL which has a deduced length of 741 amino acids (Fig. 1A). The HOF-S1 cDNA is 2760 bp long and contains one long open reading frame of 668 amino acids from nucleotides 370-2376, encoding HOFS. The predicted HOFS translation product is identical to amino acids 74-741 of HOFL but lacks the N-terminal 73 amino acids of HOFL (Fig. 1A). The N and C termini of both HOF protein isoforms have homology with BTB/POZ domain zinc finger proteins (Fig. 1A). In addition, the HOFL and HOFS translation products are 97% identical to human homologs (GenBankTM accession numbers AAG28340 and CAB43377, respectively).


View larger version (78K):
[in this window]
[in a new window]
 
Fig. 1.   Cloning and characterization of two HOF isoforms. A, schematic representation of the two protein isoforms of HOF deduced from cDNAs isolated from an oligodendrocyte cDNA library. The N-terminal region of both HOF isoforms contains a POZ domain and the C-terminal region contains five C2H2 zinc fingers (ZnF). HOFS lacks the N-terminal 73 amino acids present in HOFL but is otherwise identical to HOFL. B, in vitro coupled transcription/translation (TNT) of HOF-S1 cDNA results in a single major protein product, whereas translation of HOF-L1 cDNA results in a protein of the same size as HOFS as well as a larger product. Translation of two cDNA templates with an optimized Kozak consensus around the first ATG of HOFL (Koz-L) resulted in improved expression of the long isoform. Western blotting of 6E12 and 2H5 cell extracts with a polyclonal antibody against HOF resulted in two bands, which were also present in 6E12 nuclear extracts (NE) but not in HeLa cell extracts. The two bands identified by Western blotting co-migrate with in vitro translated HOFL and HOFS when run on the same protein gel (data not shown). C, comparison of the HOF BTB/POZ domain with homologous BTB/POZ domain proteins. Identical residues are highlighted with dark gray and similar residues with light gray boxes. D, in vitro [35S]methionine-labeled HOFS or HOFL was incubated with the indicated GST fusion proteins bound to glutathione-Sepharose beads, and the bound protein was eluted and analyzed on a protein gel. E, the AH109 yeast strain was co-transformed with plasmids encoding the indicated DBD and activation domain (AD) fusion constructs and maintained in the absence of leucine and tryptophan. Approximately 500 cells from each co-transformant were plated on selective plates to test for reporter gene expression. Interaction between the two hybrid proteins leads to yeast growth in the absence of histidine and adenine and blue coloring on X-alpha -galactosidase (X-alpha -gal). F, schematic representation of the DBD and AD fusion constructs used in E.

In vitro transcription-coupled translation of HOF-S1 cDNA gave a single major translation product, whereas translation of HOF-L1 cDNA resulted in a protein of the same size as HOFS as well as a larger product (Fig. 1B). We hypothesized that the suboptimal translational context of the HOFL start codon (TGACAAATGC) compared with the consensus derived by Kozak (GCC(A/G)CCATGG), might result in translational initiation of both HOFL and HOFS isoforms from the HOF-L1 cDNA template due to leaky ribosomal scanning (14), resulting in expression of two products. When the region around the start site of HOFL was mutated to GCCACCATGC, thus closer to the Kozak consensus, translation of HOFL was indeed improved (Fig. 1B, Koz-L) and the relative amount of the long form was improved from 43 to 68%, supporting the hypothesis that the short HOFS isoform is generated in addition to HOFL from the HOF-L1 transcript by a leaky scanning mechanism (Fig. 1B). Improved translation from the Koz-L construct relative to HOF-L1 may also result from the removal of the 5'-untranslated region of HOF-L1.

To detect endogenous HOF proteins, a polyclonal antibody was raised against a peptide located in the C-terminal end of HOF, whose sequence was identical with that of the human homolog. Because HOF was originally cloned from a 6E12 cell cDNA library, we examined the oligodendrocytic cell lines 2H5 and 6E12 for HOF expression by Western blotting. As shown in Fig. 1B, Western blotting identified two major bands in these cell lines, which co-migrated with the HOFL and HOFS bands obtained by in vitro translation (Fig. 1B and data not shown). The same two bands were also evident in nuclear extract from 6E12 cells, indicating that both HOF isoforms are present in the nucleus (Fig. 1B). The HOF antibody was shown to be specific because no bands were observed in Western blots of HeLa cell extracts (Fig. 1B). Interestingly, in protein extracts from the 2H5 and 6E12 cells there appears to be more protein corresponding to the small HOF isoform than the large (Fig. 1B).

The N-terminal region of HOF contains a BTB/POZ domain that has about 36% overall amino acid homology to the BTB/POZ domains of BCL-6, a potential oncogene involved in chromosomal translocations in a third of diffuse large cell lymphomas (15); BAZF, which is a BCL-6 homologous gene of unknown function (16); PLZF, whose chromosomal translocation is involved in acute promyelocytic leukemia (17); and Miz-1, a c-Myc-interacting protein containing 13 C2H2 zinc fingers (18) (Fig. 1C).

BTB/POZ-zinc finger factors are often found in huge multimeric complexes within the cell nucleus, and the BTB/POZ domain is thought to represent the major protein-protein interaction interface on these proteins (19). Hence, physical interaction between HOFS and HOFL proteins was investigated in vitro by GST pull-down experiments. GST fusion proteins were generated with full-length HOFS and HOFL proteins (Fig. 1D). Approximately equal amounts of each of these fusion proteins as well as GST alone were used in pull-down assays with 35S-labeled in vitro translated full-length HOF. As shown in Fig. 1D, full-length HOFS and HOFL can interact strongly in both homotypic and heterotypic interactions. In further experiments designed to localize the interaction interface, the yeast two-hybrid system was used to test for an interaction between full-length HOFS or HOFL and either the short or the long BTB/POZ domains (Fig. 1E). Yeast growth on selective plates lacking histidine and adenine was observed when the activation domain (AD) fusion protein with full-length HOF protein was expressed together with the DNA-binding domain (DBD) fusion protein of either POZ domain. We conclude that both HOF isoforms can engage in homo- and heterotypic protein interactions and that the BTB/POZ domain seems to constitute a major interaction interface.

The C-terminal region of HOF contains five C2H2 zinc finger motifs, where the fifth and most C-terminal zinc finger is isolated from the proximal zinc fingers by a 32-amino acid linker (Fig. 1A). Because four zinc fingers are clustered and connected by a conserved stretch of seven amino acids (the His-Cys link), HOF can be assigned to the Krüppel-like subfamily of zinc fingers (20). Notably, this region of HOF shares 55% similarity (41% identity) to zinc fingers 4-7 of PLZF (17), indicating that the four proximal zinc fingers of HOF may bind a related DNA sequence to PLZF (Fig. 2A). To test this hypothesis, the DNA-binding site for PLZF (11) was used in an EMSA. As can be seen in Fig. 2B, a GST fusion protein with HOF zinc fingers (GST-HOF/Fingers), but not GST alone, gives a retarded band in EMSA analysis when the PLZF-binding site is used as radioactively labeled DNA probe (compare 1st and 2nd lanes). This binding is essentially competed away by a 15-fold excess of cold competitor DNA for PLZF (3rd lane) or S63 (7th lane), a DNA sequence isolated by a PCR-based binding site selection (11). There is little or no competition by the other DNA oligonucleotides tested (lanes 4-6). In conclusion, a GST fusion protein with the five zinc fingers of HOF binds specifically to both PLZF and S63 sequences, which contain an identical 5-bp sequence (Fig. 2C, underlined).


View larger version (34K):
[in this window]
[in a new window]
 
Fig. 2.   Characterization of HOF DNA binding activity. A, comparison of the four clustered zinc fingers of HOF with fingers 4-7 of PLZF, with which they share the greatest similarity (55%). The Cys and His residues of the zinc finger motif are underlined. Identical residues are highlighted with dark gray and similar residues with light gray boxes. The arrow indicates the position of the intron in the corresponding HOF genomic sequence. B, EMSA analysis using GST or a GST fusion protein with the zinc fingers of HOF (GST-HOF/Fingers) and a radioactive PLZF probe. A 15-fold excess of cold competitor was included as indicated. The position of the DNA-protein complex is indicated on the left. C, the sequences of the top strand for each double-stranded oligonucleotide and a summary of the competition results in B are shown. Both PLZF and S63 oligonucleotides compete well for binding and contain an identical 5-p sequence (underlined).

In addition to HOF-L1 and HOF-S1, six additional alternative splice forms of HOF were identified by screening the oligodendrocyte cDNA library (Fig. 3A). All HOF-L cDNAs are ~3000 nucleotides long and contain one long open reading frame identical to HOFL (Fig. 1A). The HOF-S cDNAs are about 2700 bp long and contain one long open reading frame identical to HOFS (Fig. 1A). To understand how these splice variants are generated, the mouse genomic locus for HOF was characterized (Fig. 3B). Of the 11 exons that we identified, only exons II and V-VII were present in all splice forms. There are four alternative sequences in the most 5' ends of the oligodendrocyte-derived cDNAs, which are derived from exons ID-1 and ID-2 spliced together or exons IA-C individually (Fig. 3). Expressed sequence tags in the GenBankTM data base suggest further complexity in the 5'-untranslated region, but none of these give rise to new protein isoforms. The human genomic locus corresponding to mouse exons ID-1 to VII covers 313 kb, although if expressed sequence tags are taken into account, the locus exceeds 700 kb. The mRNAs coding for HOFS lack exon IV, which contains the translation initiation codon for HOFL, and utilize instead a downstream in-frame ATG codon in exon VI, corresponding to nucleotides 578-580 in HOF-L1 (Fig. 3B). The intron between exons VI and VII is located at nucleotides 2162-2163 in HOF-L1, which corresponds to the linker region between the first and second zinc finger (Fig. 2A, arrow). Thus, the BTB/POZ domain and the first zinc finger is encoded by exon VI and the four C-terminal zinc fingers by exon VII.


View larger version (25K):
[in this window]
[in a new window]
 
Fig. 3.   Alternative splice forms and genomic structure of HOF. A, schematic representation of the alternative HOF splice forms isolated by screening an oligodendrocyte cDNA library with HOF cDNA. The start codon is indicated by an arrow and the stop codon by an asterisk. Exons are depicted as boxes, and filled boxes are coding. B, genomic structure of the HOF locus in mice, drawn to scale. The dotted lines indicate regions that have been sequenced on genomic DNA. Double slanted lines indicate introns of undetermined length.

HOF Expression Coincides with Hippocampal Neurogenesis-- To identify HOF expression in the developing brain, we performed immunohistochemistry (Fig. 4, A-F) and in situ hybridization (Fig. 4, G-J) on sections of mouse brains. This procedure showed that during late embryonic and postnatal development of the cerebral cortex, HOF expression is mainly confined to presumptive CA pyramidal and DG granule neurons of the hippocampal subdivision (Fig. 4, B inset, and G). Generation of post-mitotic Ammon's horn precursor neurons from progenitor cells in the hippocampal neuroepithelium begins around embryonic day 11 (E11), and the first CA field-specific differentiation markers identified in these cells appear around E15 (21). At E16.5, HOF is expressed in a broad band of neurons constituting the intermediate zone (or inferior band) of the hippocampal anlagen (Fig. 4A). At E18.5, HOF appears in both migrating and settled pyramidal precursor neurons of the inferior band and the HP, respectively (Fig. 4B). The first granule cell precursors migrate out from the DG neuroepithelium around the dentate notch at E17. From E18.5 and until birth, numerous HOF-expressing cells were observed in the migratory path from the dentate neuroepithelium to the DG, and at least a subpopulation of these HOF-positive cells appears to be the so-called scaffolding pioneer granule cells that settle first in the suprapyramidal limb of the dentate anlagen, establishing the suprapyramidal-to-infrapyramidal morphogenic gradient (Fig. 4, B and C). During the end of the first postnatal week, HOF expression was still pronounced in hippocampal pyramidal and granule neurons (Fig. 4D). During postnatal development, there is a clear gradient of HOF expression in which highest levels occur in the newest born cells (Fig. 4, D, arrow, and H), and there is a marked decline of HOF expression in mature neurons expressing the neuronal differentiation marker NeuN. A general decline in HOF expression occurred in pyramidal neurons of Ammons horn during the 2nd postnatal week and persisted into adulthood (Fig. 4, E, F, I, and J). A similar decline in HOF expression occurred in the DG except for recently formed granule cells that settle in a band along the hilar side of the DG (Fig. 4, E, F, and I). The indusium griseum is a thin layer of gray matter on the dorsal surface of the corpus callosum, which is regarded as a rudimentary (or vestigial) portion of the hippocampus. As shown in Fig. 4M, HOF expression was pronounced in indusium griseum cells at E18.5.


View larger version (66K):
[in this window]
[in a new window]
 
Fig. 4.   Localization of HOF in the developing cerebral cortex. Horizontal sections of mouse brains of E16.5 (A), E18.5 (B), P0.5 (C and G), P5 (D), P7 (H), P14 (I), P15 (E), P28 (F), and adult (J) mice were subjected to immunohistochemistry using HOF-specific polyclonal antiserum (A-F) and to in situ hybridization with an alkaline phosphatase-labeled probe complementary to HOF mRNA (G-J). A and B, HOF expression was confined to the archicortical neuroepithelium lining the lateral ventricle, the hippocampal intermediate zone, and the expanding hippocampal plate. B, upper inset, saggital brain section showing HOF expression is confined to the hippocampal territory (circle) but not the adjacent neocortex. B, lower inset, numerous HOF-expressing progenitors leave the dentate notch (arrow) to migrate to the dentate gyrus. C and G, in newborn mice, HOF expression appears in migrating cells of the hippocampal intermediate zone and in cells along the dentate gyrus migratory path, in presumptive pyramidal cells settled in the CA1 and CA3 subregions of Ammon's horn as well as in granule cells settled in the external blade of the dentate gyrus. D and H, during the 1st postnatal week, HOF expression is pronounced in settled neurons in Ammon's horn and dentate gyrus. D, inset, double immunostaining for the mature neuronal marker NeuN (green) and HOF (red) shows that HOF is expressed at high levels in latest born neurons in both the pyramidal cell layer (arrow) and the granule cell layer. E and I, at the end of the 2nd postnatal week, a decline in HOF expression was revealed in pyramidal neurons of the Ammon's horn, whereas high levels of expression occurred in newborn granule neurons lining the hilus (arrow). E, inset, double NeuN (green) and HOF (red) immunostaining showing high HOF expression in less mature granule neurons (arrow). F and J, after the 1st month of postnatal life, a modest HOF expression appeared in Ammon's horn neurons and in granule cells of dentate gyrus. K and L, immunofluorescent staining of HOF and Oct-6, respectively, in horizontal sections of E18.5 hippocampi showing that the two proteins are co-expressed in migrating and settled CA1 cells (circle). L, inset, in contrast to HOF, Oct-6 is also expressed at high levels in the cortical plate of the neocortex. M, expression of HOF (red) in the indusium griseum of E18.5 mice. N-P, to identify actively dividing cells in the hippocampus, E18.5 mice were analyzed by immunohistochemistry after 3 h administration of BrdUrd. Note the nuclear staining of HOF (red) and the lack of co-staining with BrdUrd (green). The abbreviations used are as follows: IZ, intermediate zone; LV, lateral ventricle; HP, hippocampal plate; DN, dentate notch; Hipp, hippocampus; SVZ, subventricular zone; CP, cortical plate; CC, corpus callosum.

In addition to migrating post-mitotic cells in the inferior band and settled cells in the HP, HOF is also expressed in cells scattered in the ventricular and subventricular zone of the hippocampal neuroepithelium at E18.5 (Fig. 4K). These cells may be newly born post-mitotic cells at their departure from the subventricular zone. Alternatively, they may represent a precursor cell population that is still dividing. In addition, the stream of HOF-expressing cells leaving the dentate notch may be newly born post-mitotic cells migrating to the DG to form the granule cell layer as well as mitotic precursor cells migrating to the DG to establish the tertiary germinative matrix (7). To test these various possibilities, the pattern of BrdUrd labeling was examined in E18.5 embryos after administration of BrdUrd 3 h before killing. This short time frame ensures that only actively dividing or very recent post-mitotic cell populations were labeled. Using BrdUrd labeling and immunohistochemical localization, we were unable to detect BrdUrd and HOF double-labeled cells in the hippocampal subventricular zone (data not shown) or along the DG migratory path (Fig. 4, N-P). Thus, HOF is unlikely to be expressed in the nuclei of dividing precursor cells in the developing hippocampus. Taken together, the analyses described above show that HOF is transiently up-regulated in immature post-mitotic neurons derived from the hippocampal neuroepithelium.

Previous studies have shown that cells in the hippocampal neuroepithelium and CA fields express specific molecular markers during early cortical development prior to the arrival of hippocampal afferents, and overt morphological differences become apparent between CA1 and CA3 neurons (22, 23). Thus it is highly likely that hippocampal neurons are specified early, presumably through the influence of patterning centers (21, 24). One of the earliest markers of CA1 field specification in the hippocampus is the POU homeodomain transcription factor Oct-6, the RNA of which appears in cortical plate neurons and in presumptive CA1 neurons as early as E15.5 (22). In line with this, we detected Oct-6 protein in immature CA1 neurons in the inferior band and the hippocampal plate at E18.5 as well as in developing pyramidal neurons of the cortical plate but not in cells of the CA3 field and DG (Fig. 4L). Thus, HOF and Oct-6 proteins exhibit a distinct co-expression domain in presumptive CA1 neurons during cortical neurogenesis (Fig. 4, K and L).

HOF Is Transiently Expressed during Granule Cell Development in the Cerebellum-- During postnatal brain development, HOF expression occurs transiently in the granule cell layer of the cerebellum (Fig. 5). Cerebellar granule cells are born postnatally in the external granular layer (EGL), and these post-mitotic precursor cells migrate inward through the molecular layer along the processes of radial Bergman glia to settle in the internal granule cell layer (IGL). In the mouse, the majority of cerebellar granule cells are generated during the first 3 weeks of postnatal development (25, 26). By using immunohistochemistry, we investigated the pattern of HOF expression during cerebellar granule cell neurogenesis. At P5 and P10, the EGL has expanded beneath the pial surface, and only a small number of granule cells have completed their migration along Bergman glia and initiated expression of the mature neuronal marker NeuN (Fig. 5, A and B). HOF expression appears diffuse in the EGL and is not confined to nuclei. On the contrary, a distinct nuclear HOF staining is apparent in immature granule cells residing transiently in a layer in the upper aspect of the IGL (Fig. 5, A-C). These cells may correspond to the population of brain-specific lipid-binding protein expressing immature granule cells described previously (25). HOF expression in the IGL seems to coincide with immature granule cells that appear not to express the mature neuronal marker NeuN (Fig. 5B). The EGL is a transitory structure, which in the mouse ceases to exist by the end of the 3rd postnatal week at which time numerous granule cells populate the expanding IGL (26). At P15, HOF expression in the molecular layer is restricted to a narrow band of cells lining the pial surface as well as to scattered cells along the migratory path to the IGL (Fig. 5D). In addition, numerous presumptive granule cell precursors express HOF in the IGL, and only a minor portion of these appear to co-express NeuN (Fig. 5E). During the following week, HOF expression in the IGL and molecular layer declined and was confined to only a few scattered cells (Fig. 5F). Taken together, this preliminary expression analysis indicates that HOF is transiently up-regulated in the nuclei of migrating granule cell precursors and that expression of the gene is extinguished as granule neurons mature in the IGL.


View larger version (95K):
[in this window]
[in a new window]
 
Fig. 5.   HOF expression in cerebellar granule cells. Immunofluorescent detection of HOF (red) in horizontal sections of mouse cerebellum from P5 to P28. A-C, during the 1st 2 postnatal weeks, HOF immunostaining was revealed in the nuclei of immature granule neurons migrating in the molecular layer, sojourning in a transient layer in the upper aspect of the IGL (arrow) as well as in scattered cells in deeper layers of the IGL. Note that HOF expression in the EGL is not confined to nuclei, and cells expressing HOF (red) in the IGL do not co-express the mature neuronal marker NeuN (green). D and E, after the 2nd postnatal week, the EGL has ceased to exist, and HOF expression (red) is confined to numerous cells that have settled in the IGL. Note that only a minor portion of the cells seems to co-express NeuN (green). F, at late stages of cerebellar development (P28), expression of HOF in the granule cell layer has declined, and staining is confined to a few scattered cells. The abbreviations used are as follows: wm, white matter; egl, external granule cell layer; igl, internal granule cell layer; ml, molecular layer; gcl, granule cell layer.

HOF Is an Early Marker for Astrocytes in the Subventricular Zone and Rostral Migratory Stream-- The subventricular zone (SVZ) of the lateral ventricle generates macroglia and olfactory bulb interneurons during postnatal brain development (27, 28). The latter cells migrate several millimeters along the rostral migratory stream (RMS) to the olfactory bulb, where they settle in the granule cell layer and glomerular layer (29). As shown in Fig. 6, A and B, HOF expression appeared in cells lining the entire SVZ implying that HOF may be an early marker for postnatal gliogenesis and/or neurogenesis. In addition, a pronounced HOF expression appeared in a subpopulation of cells in the anterior SVZ (aSVZ, Fig. 6C). Stem cells in this region are considered by some authors to be a homogenous population giving rise to mainly olfactory bulb granule neurons and periglomerular neurons (30, 31). Thus, HOF may be an early marker for these specialized neurons. Moreover, numerous astroglia are present in the RMS and the olfactory bulb during postnatal development, but little attention has been put to the origin of these non-neuronal cells (32, 33). By using immunohistochemistry, we decided to determine the fate of HOF-positive cells migrating in the RMS. At P5, a subpopulation of cells in the anterior SVZ express HOF (Fig. 6C). As was expected for undifferentiated SVZ cells, these HOF-positive cells do not express the astroglial differentiation marker GFAP nor the neuronal differentiation marker NeuN (data not shown). At P15, expression of HOF is pronounced in cells migrating in the RMS and in cells scattered throughout the olfactory bulb (Fig. 6, D-F and H). In the olfactory bulb glomerular layer, HOF expressing cells are found mainly at the periphery of individual glomeruli with a few positive cells at the center (Fig. 6F). Immature interneurons migrating in the RMS do not express NeuN, as opposed to mature neurons in the granule cell layer that express the neuronal differentiation marker (Fig. 6G). In the granule cell layer, there is a clear non-overlapping expression of NeuN and HOF (Fig. 6, I and K), indicating that HOF is either expressed in non-neuronal cells in this layer or in immature interneurons. However, the scattered appearance of HOF-positive cells in most layers of the olfactory bulb and the observation that only a subpopulation of cells in the RMS express HOF (Fig. 6J) strongly imply that the protein is expressed in astroglia. To test this possibility, we performed double immunohistochemistry with HOF and GFAP antibodies. As shown in Fig. 6L, the majority of HOF-positive cells in the RMS and granule cell layer express the astroglial marker GFAP. In addition to the olfactory bulb, HOF is also a marker for astroglia in other brain locations including the cerebral cortex and cerebellum (Fig. 7, A and B). Taken together, these results suggest that HOF is an early SVZ marker for glial precursor cells. Furthermore, our data indicate that the anterior SVZ is a significant source of olfactory bulb astrocytes migrating in the RMS.


View larger version (83K):
[in this window]
[in a new window]
 
Fig. 6.   HOF expression in subventricular zone progenitor cells. A, horizontal section of neonatal mouse olfactory bulb showing weak to moderate levels of HOF immunostaining of cells in the rostral migration stream (RMS) and glomerular layer. A, inset, horizontal section of lateral ventricle subjected to in situ hybridization with an alkaline phosphatase-labeled probe complementary to HOF. Note the strong staining of cells at the anterior (arrow) and posterior regions of the subventricular zone. B and C, horizontal sections of the lateral ventricle of P5 mouse brains, stained with an HOF-specific antibody (red), showing pronounced HOF expression in subventricular zone cells at the most anterior-posterior levels. Note that HOF (red) seems to define a subpopulation of cells (DAPI) in the anterior subventricular zone (C). D and E, saggital sections of P15 mouse brains subjected to HOF (red) immunostaining. Note the numerous HOF-positive cells migrating along the vertical limb of the RMS. F, horizontal section through the olfactory bulb of P15 animal subjected to HOF (red) immunostaining. Note that HOF-expressing cells are not confined to a single layer but are scattered throughout most regions of the bulb. G-I and K, horizontal sections through the olfactory bulb of P15 animal subjected to NeuN (green) and HOF (red) immunostaining. G, olfactory bulb granule cells express the mature neuronal marker NeuN. H, numerous cells in the horizontal limb of the RMS express HOF as well as cells scattered in the granule cell layer. I and K, cells expressing HOF (red) in the granule cell layer do not co-express NeuN (green). Note that specific NeuN staining is nuclear; the streaky green staining is due to nonspecific binding. J, saggital section of the olfactory bulb of P15 mouse showing co-staining with DAPI (blue) and HOF (red). Note that HOF expression is confined to a subpopulation of cells migrating in the horizontal limb of the RMS. L, horizontal section of mouse olfactory bulb of P15. Double immunostaining for GFAP (green) and HOF (red) demonstrates that the majority of HOF-positive cells in the RMS and GCL express the astroglial marker GFAP. The abbreviations used are as follows: OB, olfactory bulb; RMS, rostral migratory stream; GL, glomerular layer; Str, striatum; aSVZ, anterior subventricular zone; LV, lateral ventricle; CC, corpus callosum; GCL, granule cell layer; MCL, mitral cell layer; EPL, external plexiform layer; ONL, outer olfactory nerve fiber layer; DAPI, 4,6-diamidino-2-phenylindole.


View larger version (60K):
[in this window]
[in a new window]
 
Fig. 7.   HOF is expressed in macroglia. A and B, double immunostaining of HOF (red) and GFAP (green) in astroglia. A, HOF-positive astroglia in primary cultures of cortical cells established from the medial temporal lobe of E18 mouse embryos co-express GFAP (green). B, in the white matter (wm) of the cerebellum only a subpopulation of HOF-expressing glial cells express GFAP. C, HOF expression in the cerebellar white matter coincides with that of CNPase (green), a marker for myelin-forming oligodendroglia. D-F, co-staining of the fornix with DAPI (blue) and HOF (red) demonstrates that HOF-expressing cells in white matter tracts are present in long arrays characteristic of oligodendroglia. G and H, nuclear HOF expression is observed in 2H5 and 6E12 oligodendrocytic cells, respectively. Many, but not all, 6E12 oligodendrocyte cells exhibit a characteristic speckled nuclear staining (arrows). I, double immunostaining of 6E12 cells incubated for 2 h in the presence of BrdUrd with HOF- (red) and BrdUrd (green)-specific antibodies. HOF expression was occasionally detected in DNA-replicating cells (yellow, arrows). J-L, to investigate whether HOF expression coincides with proliferation of gliogenic progenitor cells, the pattern of HOF expression was revealed in the gliogenic subventricular zone of the fimbria in E18.5 mice killed 3 h after administration of BrdUrd. Note the lack of cells exhibiting double immunostaining for HOF (red) and BrdUrd (green). DAPI, 4,6-diamidino-2-phenylindole.

HOF Is Expressed in Oligodendroglia and in Reactive Astroglia during Brain Pathology-- As described above, a pronounced HOF expression was revealed in the postnatal subventricular zone of the lateral ventricle. This zone harbors numerous glial precursor cells, which give rise to both astrocytes and oligodendrocytes (27). HOF expression is prominent in white matter tracts (Fig. 7, C and D), which are enriched in glial cells and do not contain neuronal cell bodies. However, only a subpopulation of HOF-expressing cells in white matter expresses the astroglial marker GFAP (Fig. 7B), indicating that HOF is also expressed in oligodendrocytic cells in vivo. The domain of HOF expression coincides with that of CNPase, an oligodendrocytic marker (Fig. 7C). A notable feature of HOF-expressing glial cells in white matter tracts such as the fornix is that they are localized in long arrays characteristic of oligodendroglia (Fig. 7, D-F). Moreover, Western blots show that both HOF isoforms are expressed in two oligodendrocytic cell lines (Fig. 1B), and immunocytochemistry with the HOF antibody revealed uniform nuclear staining of HOF protein in 2H5 oligodendrocytic cells (Fig. 7G). However, staining for HOF in 6E12 oligodendrocytic cells showed a speckled nuclear staining pattern in many but not all cells (Fig. 7H), implying that HOF is concentrated in nuclear subdomains in these cells. Interestingly, several other BTB/POZ domain proteins have been found to display speckled nuclear localization, which presumably is related to the ability of these proteins to interface in higher molecular weight complexes (34).

In 6E12 cells, expression of HOF was detected in replicating cells labeled for 2 h with BrdUrd prior to staining (Fig. 7I). To test the possibility that HOF is also expressed during DNA replication of gliogenic stem cells, the pattern of BrdUrd labeling was revealed in progenitor cells in the fimbrial subventricular zone in E18.5 embryos after administration of BrdUrd 3 h before killing. This portion of the subventricular zone generates glial stem cells populating the fimbria/fornix during hippocampal development (4). As shown in Fig. 7, J-L, we were unable to detect HOF expression in BrdUrd-labeled stem cells in the fimbrial subventricular zone. Thus, we conclude that HOF expression coincides with post-mitotic development of glial cells in this area.

By having demonstrated HOF expression in macroglia, we next investigated whether the protein is expressed in glial cells during brain pathology. In these experiments, we analyzed brains from transgenic mice in which c-Myc is ectopically expressed in myelin-forming oligodendroglia under control of a myelin basic protein promoter (Fig. 8A). Affected transgenic mice develop a hypomyelinating phenotype characterized by untimely programmed cell death of oligodendroglia (13, 35). Immunostaining of white matter tracts with CNPase reveals multiple hypomyelinated lesions in c-Myc transgenic mice compared with control mice (Fig. 8, B and C). These hypomyelinated lesions are populated by numerous GFAP-positive astrogliotic cells, which are enlarged relative to those present in the fornix of control mice (Fig. 8, E and H). To test whether HOF is expressed in these activated astrocytes, we have performed triple staining of the fornix from c-Myc transgenic mice and controls (Fig. 8, F and I). HOF expression is pronounced in GFAP-positive reactive astrocytes in lesioned areas (Fig. 8F, circle, purple). As mentioned above, HOF is also expressed in normal astrocytes in control animals (Fig. 8I).


View larger version (113K):
[in this window]
[in a new window]
 
Fig. 8.   HOF is expressed in reactive astrocytes. Immunostaining of c-Myc, NeuN, CNPase, HOF, and GFAP in brain sections from myelin basic protein/c-Myc transgenic mice (A and B, and D-F) and control mice (C and G-I) at P20. A, double c-Myc (red) and NeuN (green) staining showing c-Myc transgene expression in fimbrial oligodendroglia (arrow). B, CNPase immunostaining of fornix showing pronounced hypomyelination (circle). C, CNPase staining of fornix of age-matched control mouse. D, expression of HOF (blue) in lesioned fornix of myelin basic protein/c-Myc transgenic mouse. E, double immunostaining for CNPase (green) and GFAP (red) showing concentration of reactive astroglia in lesioned area (circle). F, triple immunostaining for CNPase (green), GFAP (red), and HOF (blue) showing pronounced HOF expression in reactive astroglia (pink) in hypomyelinated area (circle). G-I, HOF (blue) is co-expressed in GFAP positive astrocytes (red) in the fornix of control mice.


    DISCUSSION
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

Here we identify two isoforms of a new murine nuclear BTB/POZ domain zinc finger factor, designated HOF, that is expressed in hippocampal neurons, cerebellar granule neurons, and macroglia. Nuclear BTB/POZ domain factors have been implicated in regulation of chromatin conformation associated with cellular differentiation, including the establishment of accessible (i.e. euchromatin) or inaccessible (i.e. heterochromatin) domains for classical transcription factors (34). HOF expression coincides with early post-mitotic development of the three principal hippocampal cell types, presumptive CA1 and CA3 pyramidal neurons of Ammon's horn and presumptive granule neurons of the dentate gyrus, and a sharp decline in expression appears at the border between the CA1 region and the subiculum. Although several genes are known to be expressed in subdomains within the developing hippocampus (21, 23), HOF seems to be the first example of a gene whose expression is strictly confined to neurons of the hippocampal territory as opposed to the adjacent neocortical and transitional areas. Thus, HOF is a candidate regulatory factor for hippocampal cell fate in the medial temporal lobe. Furthermore, the observation that HOF is expressed in all three major neuronal cell types argues for a coordination function for the gene in hippocampal development. The finding that HOF and the POU homeodomain transcription factor Oct-6 are co-expressed in migrating and settled CA1 neurons implies that HOF could function with such subregion-restricted factors to specify neuronal subtype identities in the hippocampus. Alternatively, HOF could exhibit a more global role during hippocampal development such as a function in morphogenesis of the single cell-layered stratum pyramidal and granule cell layer or in coordination of hippocampal axon pathfinding.

In addition to the dentate gyrus, HOF was also expressed during postnatal granule cell development in the cerebellum. Based on spatial and temporal expression patterns of molecular markers, Heintz and co-workers (25) defined four stages of development of cerebellar granule neurons. The first stage is granule cell neurogenesis, which is localized to proliferating cells in the superficial layer of the EGL. The second stage is initiation of differentiation of young post-mitotic cells in the deeper aspect of the EGL and extension of T-shaped parallel fibers. The third stage of granule cell development ranges from axon extension and migration along radial Bergmann glial processes in the molecular layer to their settling in deeper layers in the IGL. This stage is also characterized by the presence of a transient layer of immature granule cells in the upper aspect of the IGL, at the level of the Purkinje cell layer and the end of the radial glial trail. The transient layer presumably marks a transition from glial to non-glial guided migration and is characterized by expression of a brain-specific lipid-binding protein in both radial Bergmann glia and granule cells (25). The fourth stage is characterized by terminal differentiation of settled granule cells in the IGL. This mature stage involves establishment of connections between Purkinje cell dendrites and parallel granule cell axons in the molecular layer as well as connections between afferent mossy fiber axons and granule cell dendrites in the IGL.

Moderate levels of HOF occur in EGL cells during the first 2 weeks of postnatal development, but localization of the protein is not confined to nuclei. On the contrary, a distinct nuclear staining was revealed in immature granule cells which (i) migrate in the molecular layer, (ii) sojourn in a transient layer in the upper aspect of the IGL, and (iii) settle in deeper layers of the IGL. After the 2nd postnatal week, expression of the gene was confined to immature granule cells of the IGL but was down-regulated in mature granule cells. This expression pattern corresponds to stage three of granule cell development described above indicating that HOF may play roles in granule cell migration and axon growth but less in processes associated with granule cell proliferation (stage one) and refinement of synaptic connections (stage four).

The postnatal subventricular zone lining the lateral ventricle is thought to derive from the subventricular zone of the lateral ganglionic eminence (28). The latter generates tangentially migrating cortical interneurons during embryonic neurogenesis (36, 37), whereas the former is a major source of immature neurons migrating along the RMS to the olfactory bulb, where they differentiate into granule and periglomerular interneurons. In addition to olfactory bulb interneurons, the postnatal subventricular zone also harbors gliogenic stem cells, which generate astroglia and oligodendroglia. The anterior subventricular zone was considered to be restricted to generation of olfactory bulb interneurons, whereas more caudal levels were enriched in gliogenic progenitors (27, 30). Recent studies have shown that neuronal progenitors can be isolated from all rostro-caudal levels (38) and that the subventricular zone is a continuum of migrating neurons and astroglia arranged as chains in which some migrate tangentially from more caudal levels to join the RMS (28, 39). Moreover, although numerous astrocytes migrate along the RMS to the olfactory bulb, little attention has been put into the origin of these cells.

We find HOF expression at all rostro-caudal levels of the subventricular zone, and this expression coincides with immature cells that do not express astrocyte or neuronal late differentiation markers. In addition, high levels of HOF were revealed in a major portion of cells lining the anterior subventricular zone (Fig. 4C) indicating that this region is not a homogenous cell mass but is a patchwork of HOF-expressing and non-expressing cells. A similar patchwork of HOF-expressing and non-expressing cells appeared in the horizontal limb of the RMS. In the RMS, the great majority of these HOF-positive cells appeared to be astrocytes as they express the glial marker GFAP. Whether HOF expression in the anterior subventricular zone occurs in a progenitor common to astroglia and olfactory neurons or in cells already restricted to a glial fate remains to be determined. Furthermore, HOF expression was also persistent in mature astroglia in other locations including the striatum, the cerebral cortex, and cerebellum as well as in white matter oligodendroglia indicating that HOF-expressing subventricular zone progenitors give rise to at least two types of mature glial cells. Whatever the case, this expression pattern suggests that HOF is a candidate marker for gliogenic progenitor cells in this region of the subventricular zone, including more caudal regions, and that the RMS may be a significant source of astroglia populating the olfactory bulb during postnatal development.

In Drosophila, several BTB/POZ domain zinc finger transcription factors have been implicated in nervous system development, including the involvement of the tramtrack (ttk) gene in glial cell, sensory organ, photoreceptor, and cone cell development (5, 40-45); the longitudinal lacking (lola) gene in axon growth and path finding (46); the abrupt (ab) gene in the formation of neuromuscular connections and proper muscle attachments (47); and the fruitless (fru) gene, which is expressed in a subset of neurons in the development of proper male sexual behavior (48-51). A number of mammalian BTB/POZ-zinc finger factors are expressed in the central nervous system (52-57), but the functional significance of this expression is not clear. In addition, the BTB/POZ domain factors BCL-6/LAZ3 and PLZF have been implicated in human malignancies involving chromosomal translocations (15, 58, 59). The importance of BTB/POZ zinc finger factors in cell type specification in the Drosophila nervous system implies that such factors also play important roles in specifying neuronal and glial cell lineages in the mammalian central nervous system. Thus, identification of novel mammalian BTB/POZ factors is an important first step in understanding their role in brain development. The cloning of the HOF gene provides an opportunity to inactivate the gene by homologous recombination. This could reveal definitive evidence regarding the role of this gene product in mammalian brain development.

    ACKNOWLEDGEMENT

We thank Malene B. Hansen for help with establishment of primary cortical cultures.

    FOOTNOTES

* This work was supported by grants from the Lundbeck Foundation, the Danish Cancer Society, Løvens Kemiske Fabriks Foundation, the Michelsen Foundation, and the NOVO-Nordisk Foundation.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EBI Data Bank with accession number(s) AF185576 and AF194030.

§ To whom correspondence should be addressed. Tel.: 45-35327722; Fax: 45-35327701; E-mail: naj@imbg.ku.dk.

Published, JBC Papers in Press, December 13, 2001, DOI 10.1074/jbc.M110023200

2 N. A. Jensen, unpublished data.

    ABBREVIATIONS

The abbreviations used are: DG, dentate gyrus; BTB/POZ, broad complex tramtrack bric-a-brac/poxvirus and zinc finger; HP, hippocampal plate; DP, dentate plate; GST, glutathione S-transferase; EMSA, electrophoretic mobility shift assay; DBD, DNA-binding domain; AD, activation domain; PLZF, promyelocytic leukemia zinc finger protein; EGL, external granular cell layer; IGL, internal granule cell layer; SVZ, subventricular zone; RMS, rostral migratory stream; BrdUrd, bromodeoxyuridine; BSA, bovine serum albumin; AD, activation domain.

    REFERENCES
TOP
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
REFERENCES

1. Lisman, J. E. (1999) Neuron 22, 233-242[CrossRef][Medline] [Order article via Infotrieve]
2. Amaral, D. G., and Witter, M. P. (1995) in The Rat Nervous System (Paxinos, G., ed) , pp. 443-493, Academic Press, San Diego
3. Levitt, P., Barbe, M. F., and Eagleson, K. L. (1997) Annu. Rev. Neurosci. 20, 1-24[CrossRef][Medline] [Order article via Infotrieve]
4. Altman, J., and Bayer, S. A. (1990) J. Comp. Neurol. 301, 325-342[CrossRef][Medline] [Order article via Infotrieve]
5. Guo, M., Bier, E., Jan, L. Y., and Jan, Y. N. (1995) Neuron 14, 913-925[CrossRef][Medline] [Order article via Infotrieve]
6. Altman, J., and Bayer, S. A. (1990) J. Comp. Neurol. 301, 343-364[CrossRef][Medline] [Order article via Infotrieve]
7. Altman, J., and Bayer, S. A. (1990) J. Comp. Neurol. 301, 365-381[CrossRef][Medline] [Order article via Infotrieve]
8. Jensen, N. A., Smith, G. M., Garvey, J. S., Shine, H. D., and Hood, L. (1993) J. Neurosci. Res. 35, 288-296[CrossRef][Medline] [Order article via Infotrieve]
9. Jensen, N. A., Smith, G. M., Shine, H. D., Garvey, J. S., and Hood, L. (1993) J. Neurosci. Res. 34, 257-264[CrossRef][Medline] [Order article via Infotrieve]
10. Pedersen, K. M., Finsen, B., Celis, J. E., and Jensen, N. A. (1999) Electrophoresis 20, 249-255[CrossRef][Medline] [Order article via Infotrieve]
11. Li, J. Y., English, M. A., Ball, H. J., Yeyati, P. L., Waxman, S., and Licht, J. D. (1997) J. Biol. Chem. 272, 22447-22455[Abstract/Free Full Text]
12. Olsen, J., Laustsen, L., Karnstrom, U., Sjostrom, H., and Noren, O. (1991) J. Biol. Chem. 266, 18089-18096[Abstract/Free Full Text]
13. Jensen, N. A., West, M. J., and Celis, J. E. (1999) J. Biol. Chem. 274, 29921-29926[Abstract/Free Full Text]
14. Kozak, M. (1999) Gene (Amst.) 234, 187-208[CrossRef][Medline] [Order article via Infotrieve]
15. Ye, B. H., Lista, F., Lo, Coco, F., Knowles, D. M., Offit, K., Chaganti, R. S., and Dalla-Favera, R. (1993) Science 262, 747-750[Abstract/Free Full Text]
16. Okabe, S., Fukuda, T., Ishibashi, K., Kojima, S., Okada, S., Hatano, M., Ebara, M., Saisho, H., and Tokuhisa, T. (1998) Mol. Cell. Biol. 18, 4235-4244[Abstract/Free Full Text]
17. Chen, Z., Guidez, F., Rousselot, P., Agadir, A., Chen, S. J., Wang, Z. Y., Degos, L., Zelent, A., Waxman, S., and Chomienne, C. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 1178-1182[Abstract/Free Full Text]
18. Peukert, K., Staller, P., Schneider, A., Carmichael, G., Hanel, F., and Eilers, M. (1997) EMBO J. 16, 5672-5686[CrossRef][Medline] [Order article via Infotrieve]
19. Bardwell, V. J., and Treisman, R. (1994) Genes Dev. 8, 1664-1677[Abstract/Free Full Text]
20. Turner, J., and Crossley, M. (1999) Trends Biochem. Sci. 24, 236-241[CrossRef][Medline] [Order article via Infotrieve]
21. Grove, E. A., and Tole, S. (1999) Cereb. Cortex 9, 551-561[Abstract/Free Full Text]
22. Tole, S., Christian, C., and Grove, E. A. (1997) Development 124, 4959-4970[Abstract]
23. Pleasure, S. J., Collins, A. E., and Lowenstein, D. H. (2000) J. Neurosci. 20, 6095-6105[Abstract/Free Full Text]
24. Rubenstein, J. L., and Rakic, P. (1999) Cereb. Cortex 9, 521-523[Free Full Text]
25. Hatten, M. E., and Heintz, N. (1995) Annu. Rev. Neurosci. 18, 385-408[Medline] [Order article via Infotrieve]
26. Goldowitz, D., and Hamre, K. (1998) Trends Neurosci. <