Transcription Factors Zic1 and Zic2 Bind and Transactivate the Apolipoprotein E Gene Promoter*

We have used the yeast one-hybrid system to identify transcription factors that bind to specific sequences in proximal regions of the apolipoprotein E gene promoter. The sequence between 2 163 and 2 124, that has been previously defined as a functional promoter element, was used as a bait to screen a human brain cDNA library. Ten cDNA clones that encoded portions of the human Zic1 (five clones) and Zic2 (five clones) transcription factors were isolated. Electrophoretic mobility shift assays confirmed the presence of a binding site for Zic1 and Zic2 in the 2 136/ 2 125 region. Displacement of binding with oligonucleotides derived from adjacent sequences within the APOE promoter revealed the existence of two additional Zic-binding sequences in this promoter. These sequences were identified by electrophoretic mobility shift assays and mutational analysis in regions 2 65/ 2 54 and 2 185/ 2 174. Cotransfection of Zic1 and Zic2 expression vector and different APOE pro-moter-luciferase reporter constructs in U87 glioblastoma cell line showed that the three binding sites partially contributed to the trans-stimulation of the luciferase reporter. Ectopic expression of Zic1 and Zic2 in U87 YM4271. APOE1-pLacZi, APOE1-pHIS-1. Transformants leu 2 a dual reporter host yeast library yeast a MATCHMAKER human brain cDNA library constructed in the pGAD10 vector (CLONTECH) the LiAc/polyethylene Approximately 5 3 10 4 transformants were plated per 150-mm dish containing his 2 leu 2 minimal selective medium supplemented with 15 m M 3-aminotriazole. Approximately 2.5 3 10 6 cDNA plasmids were screened. Based on large colony size and rapid growth, a total of 19 positive clones were selected. These clones were tested for b -galactosid- ase activities using the filter replica method. Ten clones showed strong blue color compared with the host strain. Plasmids were recovered from His 1 /LacZ 1 colonies by transformation into DH5 a cells. The plasmids were partially sequenced and the nucleotide sequences were compared with sequences in the GenBank y /EMBL data bases using the Fasta program. Purification of Recombinant His-Zic1 and His-Zic2— To produce his- tidine-tagged Zic1 and Zic2 proteins (His-Zic1 and His-Zic2), we prepared bacterial expression constructs for both proteins. First, the Zic- encoding cDNA inserts from the pGAD10 cloning vector, isolated after the one-hybrid procedure, transferred into the Bam HI/ Bgl II sites of the prokaryotic expression vector pTrcHisA Escherichia BL21 strain transformed with the expression constructs and grown in LB culture medium containing 50 m g/ml ampicillin and 70 m g/ml chloramphenicol. Expression induced addition of 1 m M isopropyl- b -thiogalactoside for 5 h. Bacte-ria collected and lysed sonication

We have used the yeast one-hybrid system to identify transcription factors that bind to specific sequences in proximal regions of the apolipoprotein E gene promoter. The sequence between ؊163 and ؊124, that has been previously defined as a functional promoter element, was used as a bait to screen a human brain cDNA library. Ten cDNA clones that encoded portions of the human Zic1 (five clones) and Zic2 (five clones) transcription factors were isolated. Electrophoretic mobility shift assays confirmed the presence of a binding site for Zic1 and Zic2 in the ؊136/؊125 region. Displacement of binding with oligonucleotides derived from adjacent sequences within the APOE promoter revealed the existence of two additional Zic-binding sequences in this promoter. These sequences were identified by electrophoretic mobility shift assays and mutational analysis in regions ؊65/؊54 and ؊185/؊174. Cotransfection of Zic1 and Zic2 expression vector and different APOE promoter-luciferase reporter constructs in U87 glioblastoma cell line showed that the three binding sites partially contributed to the trans-stimulation of the luciferase reporter. Ectopic expression of Zic1 and Zic2 in U87 cells also trans-stimulated the expression of the endogenous gene, increasing the amount of apolipoprotein E produced by glial cells. These data indicate that Zic proteins might contribute to the transcriptional activity of the apolipoprotein E gene and suggest that apolipoprotein E could mediate some of the developmental processes in which Zic proteins are involved.
Apolipoprotein E (apoE) 1 is a major component of various classes of plasma lipoproteins. It is a single-chain polypeptide of 299 amino acids, which plays a prominent role in transport and metabolism of plasma cholesterol and triglycerides, resulting from its ability to interact with lipoprotein receptors (1). The amino acid sequence of the human protein presents two polymorphic sites that generate three isoforms (apoE2, apoE3, and apoE4). The major site of synthesis is the liver, but the protein is also produced in extrahepatic tissues such as adrenals and nervous system. In the rat brain, the synthesis takes place in astrocytic cells (2,3), whereas in peripheral nervous system apoE is synthesized by nonmyelinating glial cells and resident macrophages (2). In the human brain, the synthesis occurs both in glial cells and in subpopulations of neurons in the cortex and hippocampus (4). A number of previous in vitro and in vivo studies as well as recent experiments with apoEdeficient mice and human APOE transgenic mice reveal that apoE plays an important role in neuronal maintenance and repair (5)(6)(7)(8)(9)(10)(11)(12). Genetic studies have identified the apoE4 allele as a major risk factor for developing Alzheimer's disease (AD), both in sporadic and in familial late onset forms of the disease (13,14). This allele is also responsible for poor outcome after acute brain injury (15,16), stroke (17), or neurotoxic damage (12).
ApoE synthesis is regulated in hepatic and steroidogenic cells by complex interaction of developmental, hormonal, and dietary factors (18 -22). The regulatory complexity emerges from interactions of a number of proteins which bind to proximal regions of the APOE gene promoter, as well as to far downstream elements involved in its tissue-specific expression (23)(24)(25)(26)(27)(28)(29)(30). In brain, the regulation of this gene remains largely unexplored, despite its importance in processes of degeneration and regeneration of the nervous system. The importance of the regulatory region of the APOE gene in the determination of the apoE levels in the brain is emphasized by the recent identification of a number of polymorphisms within the promoter (31,32). Some of these polymorphic sites are associated with functional changes in the activity of the promoter and with increased risk of AD (33)(34)(35). We have initiated a search of possible regulatory proteins that bind to proximal regions of the APOE gene promoter. Previously, we identified the transcription factor AP-2 as a mediator of the cAMP stimulation of apoE synthesis in glial cells (36). In the present report, we analyze the regulatory region that lies between Ϫ163 and Ϫ124, a region that contains positive regulatory elements in HELA, HepG2, COS, and U87 cells (23,24,36,37). We identify transcription factors Zic1 and Zic2 as positive transcriptional regulatory elements for the APOE gene.
The reporter constructs were subsequently linearized and sequentially integrated into the yeast strain YM4271. Yeast were transformed first with APOE1-pLacZi, followed by APOE1-pHIS-1. Transformants selected on leu Ϫ , ura Ϫ minimal medium were used as a dual reporter host yeast strain for the library screen.
Screening of the cDNA Library-The host yeast strain was transformed with a MATCHMAKER human brain cDNA library constructed in the pGAD10 vector (CLONTECH) by the LiAc/polyethylene glycol method. Approximately 5 ϫ 10 4 transformants were plated per 150-mm dish containing his Ϫ leu Ϫ minimal selective medium supplemented with 15 mM 3-aminotriazole. Approximately 2.5 ϫ 10 6 cDNA plasmids were screened. Based on large colony size and rapid growth, a total of 19 positive clones were selected. These clones were tested for ␤-galactosidase activities using the filter replica method. Ten clones showed strong blue color compared with the host strain. Plasmids were recovered from His ϩ /LacZ ϩ colonies by transformation into DH5␣ cells. The plasmids were partially sequenced and the nucleotide sequences were compared with sequences in the GenBank/EMBL data bases using the Fasta program.
Purification of Recombinant His-Zic1 and His-Zic2-To produce histidine-tagged Zic1 and Zic2 proteins (His-Zic1 and His-Zic2), we prepared bacterial expression constructs for both proteins. First, the Zicencoding cDNA inserts from the pGAD10 cloning vector, isolated after the one-hybrid procedure, were transferred into the BamHI/BglII sites of the prokaryotic expression vector pTrcHisA (Invitrogen Inc, San Diego, CA). Then, Escherichia coli BL21 strain was transformed with the expression constructs and grown in LB culture medium containing 50 g/ml ampicillin and 70 g/ml chloramphenicol. Expression was induced by addition of 1 mM isopropyl-␤-thiogalactoside for 5 h. Bacteria were collected and lysed by sonication in 50 mM NaH 2 PO 4 , 10 mM imidazole, 1 mM phenylmethylsulfonyl fluoride, 0.5 mg/ml lysozime. Fusion proteins were purified with Ni-NTA resins (Qiagen Inc, Valencia, CA) using a batch protocol as recommended by the manufacturer.
Plasmid Constructions-The luciferase reporter plasmid pXP2 (39) was used to harbor different fragments of the APOE promoter. The fragments were generated by PCR using oligonucleotides from the desired regions as primers and the APOE-pCRII construct (36) as a template. Amplified fragments were ligated to the pCRII vector (Invitrogen, San Diego, CA), and the identity was confirmed by sequencing. Fragments were subcloned in the MCS of pXP2, in front of the luciferase reporter gene. Mutation were introduced by PCR by using mutant oligonucleotides as described (40). Mouse Zic1 and Zic2 expression vectors were prepared by cloning the full-length genes into the pEBOS vector as described. 2 Cell Culture and Transfections-U87 cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. The day before transfection, confluent cells were subcultured by trypsinization and 1-3 ϫ 10 4 cells/well were plated in 24-well tissue culture plates. For transient transfections, cells were transfected with 0.5 g of DNA/well by the calcium phosphate method using the CalPhos mammalian transfection kit (CLONTECH) according to the instructions of the manufacturer.
For stable transfection U87 cells were transfected by electroporation with Zic cDNAs cloned into the expression vector pcDNA3 that carries the neomycin resistance gene for selection. After selection with 0.2 g/liter G418 for 12-15 days, single colonies were isolated with cloning cylinders. Clones stably transfected were maintained in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum, 2 mM L-glutamine, and 0.2 g/liter G418.
Luciferase and ␤-Galactosidase-Cells were harvested on day 2 following transfection with 150 l of a lysis buffer containing 25 mM Tris-phosphate, 2 mM dithiothreitol, 2 mM EDTA, 10% glycerol, 1% Triton X-100. Unsolubilized material was removed by 2 min of centrifugation, and the luciferase and ␤-galactosidase activities of the extracts were determined. Luciferase was measured using the Luciferase Assay System (Promega, Madison, WI) in a Monolight 2010 luminometer (Analytical Luminescence Laboratory) by incubation of 40 l of cell extract with 90 l of luciferase assay reagent as recommended by the manufacturer. ␤-Galactosidase was determined in a 96-well microtiter plate by incubating 20 l of cellular extract with 20 l of a solution containing 3 mg/ml o-nitrophenyl-␤-D-galactopyranoside as described (41). Absorbance at 405 nm was determined in a microplate reader.
Analysis of Protein Expression-Cells grown to confluence were lysed in 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride. The protein concentrations were determined with the Bio-Rad protein assay reagent, and SDS-PAGE was performed in the presence of 2-mercaptoethanol. Samples were transferred by electroblotting onto a nitrocellulose membrane in a semidry electroblotting system (Life Technologies, Inc.). Nonspecific protein binding to the blot was blocked by incubation of the filter with 3% nonfat milk in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 ⅐7H 2 O, 1.4 mM KH 2 PO 4 , pH 7.3) containing 0.1% Triton X-100. The blot was probed with anti-apoE antibody (Calbiochem, Bad Soden, Germany) diluted 1:5000 for 1 h at room temperature. After washing, the blot was probed with a peroxidase-linked anti-goat IgG. Bands were visualized with the ECL detection method (Amersham Pharmacia Biotech).

Isolation of cDNA Clones Encoding APOE Promoter-binding
Proteins-We are interested in identifying transcription factors that could regulate the expression of the APOE gene in the nervous system. The yeast one-hybrid system was used to screen for human brain cDNAs encoding for proteins that bind to sequences of the APOE promoter located upstream of the TATA box, a region of the promoter that has been previously shown to contain a number of functional promoter elements (23,24,36,37). Five tandem copies of the sequence spanning Ϫ163 to Ϫ124 were cloned upstream of a minimal GAL4 promoter either in pHIS1 or in pLacZi reporter plasmids and integrated into the yeast genome of YM4271 (his3, leu2). A hybrid expression library (MATCHMAKER) consisting of human brain cDNAs fused to the GAL4 activation domain was then screened to identify proteins that bound to the APOE promoter and activated HIS3 transcription from the reporter construct. Transformants growing on selective medium were assayed for ␤-galactosidase expression. After screening 2.5 million independent colonies, 10 transformants produced strong blue color on filter assay after 1 h of incubation in the presence of 5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside. Plasmids of these transformants were then recovered and sequenced. Sequence analysis indicated that five clones corresponded to different portions of the transcription factor Zic1 and five to the closely related Zic2. The longest Zic1 clone lacked regions encoding the 72 amino acids of the amino terminus as well as 10 amino acids in the carboxyl end of the open reading frame of the human gene. The longest Zic2 clone spans from amino acid 182 to 530. Both clones contained the respective zinc finger domains and were in frame with the activation domain of GAL4.
The Bait Oligonucleotide Forms Specific Complexes with Recombinant Zic Proteins-To confirm the presence of a Zicbinding site in the APOE gene promoter, we first produced His-tagged Zic1 and Zic2 proteins by transferring the Zic-encoding inserts from the yeast vector pGAD10 to the prokaryotic expression vector pTrcHisA. The recombinant proteins were then purified with nickel-containing resins. The binding activity of the recombinant proteins was assayed by EMSA with the 32 P-labeled, double-stranded bait oligonucleotide (bp Ϫ163 to Ϫ124). We detected the formation of stable complexes of this DNA fragment with both His-Zic1 and His-Zic2, but not with control bacterial extracts (Fig. 1). The binding of His-Zic1 was specifically competed by increasing amounts of the unlabeled bait oligonucleotide (Ϫ163/Ϫ124) (Fig. 2, lanes 5-7). Interestingly, the binding was also competed by double-stranded oligonucleotides derived from adjacent regions of the APOE promoter spanning bp Ϫ70 to Ϫ40 and bp Ϫ188 to Ϫ169 (Fig. 2,  lanes 2-4 and 8 -10), suggesting the existence of additional Zic1 binding elements in the promoter. The binding was not displaced by an excess of unlabeled oligonucleotides Ϫ39/ϩ1 and Ϫ113/Ϫ80 (data not shown) or by an unrelated oligonucleotide (Fig. 2, lane 11). Identical results were obtained by using His-  5, and 8), 100-fold (lanes 3, 6, and 9), or 1000-fold (lanes 4, 7, 10, and 11) excess in the EMSA. B, diagram of the proximal APOE promoter showing the location of some putative regulatory elements. TATA, the TATA box element; SP1, the GC box element; AP-2, AP2 binding sites; BEF1, BK virus enhancer factor-1 (27); URE1 and URE3, upstream response elements 1 and 3 (24,28); ERE, estrogen response element. The regions covered by DNA fragments used in the displacement analysis are also shown. Fragments labeled with an asterisk formed stable complexes with His-Zic1 and His-Zic2 in EMSA.
Zic2 (data not shown). We did not detect displacement by a number of oligonucleotides derived from other sequences of the proximal region of the APOE gene promoter (data not shown).
Nucleotide Sequences Required for Zic Binding-To determine more precisely the nucleotide sequence required for binding of Zic proteins, a number of double-stranded 20-mers included within the displacing oligonucleotides were synthesized and assayed in EMSA experiments. DNA-protein complexes were only observed when we used the DNA fragments labeled with an asterisk in Fig. 2, and these were selected for further mutational analyses. Various substitution mutations were introduced within these oligonucleotides, and then used in EMSA experiments. Fig. 3 shows results of mutagenesis in DNA fragment Ϫ143/Ϫ124. Mutation in the 5Ј end of this fragment produced a marked decrease in His-Zic1 binding with respect to control, although the retarded DNA-protein complex was clearly detectable (Fig. 3, lane 2). However, when bases from Ϫ136 to Ϫ125 were exchanged in groups of three for adenine, binding of His-Zic1 was severely impaired and the retarded complexes were undetectable (lanes 3-6). Similar results were obtained when His-Zic2 was used in the assays (data not shown). These experiments define the binding site for Zic proteins within the initial bait. Binding sequence was also investigated within the DNA fragment spanning Ϫ69 to Ϫ50. As shown in Fig. 4, binding of both His-Zic1 and His-Zic2 was disrupted by mutations in bases located between Ϫ65 to Ϫ54 (lanes 3-6 and 10 -13). Mutations in bases Ϫ69 to Ϫ66 and Ϫ50 to Ϫ53 had a weaker effect on binding of both proteins (lanes 2, 7, 9, and 14) as compared with the respective controls (lanes 1  and 8). Finally, another site was defined for both recombinant proteins within the DNA fragment Ϫ188/Ϫ169. Mutations in positions between Ϫ185 and Ϫ174 completely abolished the formation of retarded complexes (Fig. 5, lanes 3-6 and 10 -13), whereas mutations in bases Ϫ173 to Ϫ171 produced a weaker reduction of binding (Fig. 5, lanes 7 and 14). In summary, the three binding sites described above were termed Zic-BS1 (Ϫ65/ Ϫ54), Zic-BS2 (Ϫ136/Ϫ125), and Zic-BS3 (Ϫ185/Ϫ174), according to their positions with respect to the transcriptional starting point.
Transcriptional Activation of the APOE Promoter by Zic Proteins-To analyze the ability of Zic proteins to stimulate transcription from the APOE promoter, and to assess the relative contribution of each binding site to the measured activities, several APOE promoter-luciferase constructs were prepared and cotransfected with an expression vector containing the full-length coding region of the mouse Zic1 or Zic2 genes into the glioblastoma cell line U87. Along with the test constructions, each plate was cotransfected with a ␤-galactosidase expression vector that served as an internal reference for transfection efficiency. Luciferase and ␤ galactosidase activities were determined 48 h later. The luciferase activity was strongly stimulated by both Zic1 and Zic2 in a construct containing the Zic-BS1 (Ϫ70/ϩ1) as compared with controls (6 and 27.9 times, respectively) (Fig. 6). The stimulatory effect was greatly reduced after mutation of this site (construct Ϫ70/ϩ1 m(Ϫ62/Ϫ57)), although not completely abolished (the activity for Zic1 and Zic2 was 4 and 8 times over controls, respectively) (Fig. 6). However, the promoter activity was more severely impaired in the deletion mutant Ϫ60/؉1 (Ϫ60/ϩ1 in Fig. 6) (1.4 and 2.7 over controls, respectively), suggesting that the mutant Ϫ70/ϩ1 m(Ϫ62/Ϫ57) still conserved some binding capability. Both transcription factors also stimulated the luciferase activity in a longer construct including also Zic-BS2 (bp Ϫ143 to Ϫ1) (Fig. 6, construct Ϫ143/ϩ1), although to a lower extent than construct Ϫ70/ϩ1 (4.9 and 19.9 times over controls, respectively). Mutation in the Zic-BS2 decreased the simulation mediated by Zic1 and Zic2 by more than one half (2.3 and 6.5 times stimulation over controls respectively) (Fig. 6, construct Ϫ143/ϩ1 m(Ϫ133/Ϫ129)). The contribution of the Zic-BS3 was analyzed in construct Ϫ189/ϩ1. Both factors stimulated the luciferase activity by 5.2 and 12.6 times, respectively, in the wild form of this construct (Fig. 7, construct Ϫ189/ϩ1). Mutations in Zic-BS3 decreased drastically the activity of the promoter (to 2.3 and 2.5 times over controls, respectively) ( Fig.  7, construct Ϫ189/ϩ1 m(Ϫ182/Ϫ177)), indicating that in the context of this construct the Zic-BS3 seems to be preponderant over Zic-BS2 and Zic-BS1. Nevertheless, these sites also contributed to the observed activity as the stimulation of the promoter activity was between 60 and 75% lower in mutants of Zic-BS2 or Zic-BS1 in the context of construct Ϫ189/ϩ1 (Fig. 7, constructs Ϫ189/ϩ1 m(Ϫ133/Ϫ130) and Ϫ189/ϩ1 m(Ϫ55/ Ϫ50)), when compared with the stimulation observed in the wild type version of this construct (construct Ϫ189/ϩ1). Moreover, the Zic1-and Zic2-mediated stimulation of the promoter activity was completely abolished in a mutant in which the three binding sites were eliminated (Fig. 7, construct Ϫ189/ϩ1  3 m).
Zic Proteins Also Stimulate the Endogenous APOE Gene Expression-We also investigated the effect of Zic proteins on the expression of the endogenous APOE gene in the glioblastoma cell line U87. First we generated stably transfected U87 cell lines that constitutively expressed either Zic1 or Zic2. The expression of mRNA for Zic1 or Zic2 was analyzed by RT-PCR (Fig. 8A, upper panel). Only stably transfected U87 cells (Zic-1U87 and Zic2-U87 cell lines) expressed the corresponding Zic mRNA. The expression of apoE mRNA was also assayed by RT-PCR in mock-, Zic1-, and Zic2-transfected U87 cells and compared with that of two other cell lines, the glioblastoma SW1088 and the hepatoma HepG2. ApoE mRNA was detected in all cell lines except in the parental mock-transfected U87 ( Fig. 8A, middle panel). These observations were paralleled by protein expression profile obtained by Western blot analysis with a specific anti-apoE antibody (Fig. 8B). ApoE appeared in blots as a band with the same electrophoretic mobility as the recombinant apoE (34 kDa) (Fig. 8B, rApoE). Whereas the expression in mock-transfected cells was undetectable, the expression in the Zic1-U87 and Zic2-U87 cell lines was greater (Fig. 8, Zic1-U87) or similar (Fig. 8, Zic2-U87) to the endogenous expression levels observed in the glioblastoma SW1088, and slightly lower than that measured in HepG2 cells.

DISCUSSION
In the present report, we identified transcription factors Zic1 and Zic2 as potent transcriptional activators of the human APOE gene. This regulatory activity is mediated by three Zicbinding sites located in the proximal region of the APOE promoter. These sites were identified by screening a human brain library, by the yeast one-hybrid technique, by band mobility shift experiments, and by functional assay of luciferase reporter gene activity in various constructs of the APOE promoter. The Zic-mediated stimulation not only was observed on artificial APOE promoter constructs, but also on the endogenous gene in a human glioblastoma cell line. The determination of the molecular mechanisms involved in the regulation of apoE synthesis in brain is a matter of importance, as this protein seems to be involved in processes of brain repair after traumatic injury or in the pathogenesis of AD (13,14,42).
Zic1 and Zic2 belong to the Cys 2 -His 2 family of zinc finger transcription factors and are the orthologues of the Drosophila gene odd-paired (44). They seem to be important genes in the embryonic development. The expression of both proteins is observed in the gastrula stage. Later, the expression becomes restricted to the neural tube, where Zic2 seems to play an essential role in neurulation (45). Indeed, mutant mice lacking this gene show incomplete closure of the neural tube, producing holoprosencephaly and spina bifida as well as alterations in digits and vertebrae (46). In the human, mutations in ZIC2 gene are associated with holoprosencephaly (47). Later, the expression in adults is mainly observed in the cerebellum, although Zic1 is observed in other areas of the brain, and Zic2 has also been reported in testis (48). Zic proteins are also detected in a number of tumor cells (48,49). Previous studies have shown that Zic1 is able to bind to the consensus sequence of the closely related Gli proteins (50). The present report indicates that the sequence for Zic binding is rather ambiguous, although there are changes of a few nucleotides that are deleterious for the binding. The triplet CTG, followed by a GC-rich region occurs in the three Zic-binding sites of the APOE promoter. Both proteins bind similarly to these sites, although, in functional studies, Zic2 behaved as a more potent activator of the promoter than Zic1. Despite the similarity for the target sequence for both proteins that would suggest a functional redundancy, the phenotype of the Zic1 mutant mice differs markedly from that of Zic2 mutant (45,51). Further clarification of the target genes would be necessary.
The region of the promoter used in this study had been previously identified as the binding site for a number of proteins in several cell types, although never in nervous cells. For instance, the region from Ϫ161 to Ϫ141, included in our original bait and upstream to the Zic-BS2 (bp Ϫ136 to Ϫ125), was protected in DNase I footprinting experiments by nuclear extracts of HepG-2, HeLa, and Chinese hamster ovary cells, defining the regulatory element URE1 (24). This sequence was able to bind two transcription factors in a competitive fashion, one of which was identified as Sp1, the other remaining unidentified (37). Additionally, the Zic-BS1 is flanked by a Sp1binding site and the Zic-BS3 overlaps with a third Sp1-binding site (37). In extracts of HeLa and HepG2 cells, the protein BEF-1 has been identified as a negative regulator that binds to the Ϫ94/Ϫ84 sequence (27), and an unidentified protein was reported to bind to URE3 between Ϫ89 and Ϫ101 (28). Finally, a binding site for AP2 overlaps with Zic-BS1 and is responsible for the cAMP-mediated induction of apoE in glial cells (36). These results indicate that this region proximal to the APOE gene promoter must play a critical role in controlling the expression of apoE. Whether Zic proteins interact with some of these factors remains to be determined.
We do not know whether Zic proteins will play a functional role in the expression of apoE in the mature brain. In the adult, both proteins do not share the expression pattern. However, the situation could be different in the developing embryo. ApoE is highly expressed in several developing organs, including brain, eye, or bone (52). ApoE is expressed during osteogenic and chondrogenic differentiation of murine mesenchymal progenitor cells, where it seems to be induced by bone morphogenetic protein-2 (BMP2) (52). Interestingly, a BMP2/4-mediated regulation has been suggested for Zic2 in the differentiating bone of the limb buds (45), and thus, the inductive action of BMP2 on the APOE gene could be mediated by Zic2.
Another relevant relationship between Zic2 and APOE gene could be found in holoprosencephaly. It is known that some alterations in the metabolism of cholesterol such as inhibition of the 7-dehydrocholesterol reductase (53,54), or alterations in the apoE receptor megalin (55) result in holoprosencephaly. Although it is thought that some of these alteration in the metabolism of cholesterol result in alterations in the postranscriptional processing of sonic hedgehog and consequently in holoprosencephaly (56), we suggest that alterations in apoE could also play a role in those cases of holoprosencephaly that are mediated by mutations in the Zic2 gene.
ApoE plays a crucial role in a number of physiological processes, including cholesterol transport in peripheral circulation (57) and central nervous system (43). ApoE is also involved in the response to neural injury (43,58,59), maintenance of dendritic arborizations (60), and neuronal remodeling in vitro (61,62) and in AD (63). The recent association of different polymorphisms in the promoter region with AD (32-35) strongly suggests that transcriptional regulation of APOE gene may play an important role in the development of this deleterious disease. Thus, identification and characterization of the transcriptional machinery involved in the regulation of ApoE expression may be relevant to devise a possible pharmacological modulation of apoE levels. For instance, recent observations suggest that the protective effect of estrogens, which transcriptionally regulate the expression of apoE (64,65), on AD (66) could be mediated in part by apoE (67). Here we contribute to the characterization of the APOE transcriptional machinery by identifying two members of the Zic family as strong transcriptional activators of this gene.