Isolation and characterization of a novel zinc-finger protein with transcription repressor activity.

To identify genes that can repress the expression of growth regulatory molecules, a human fetal cDNA library was screened with a degenerate oligonucleotide that corresponds to the conserved stretch of 6 amino acids connecting successive zinc-finger regions in the Wilms' tumor suppressor/Egr-1 family of DNA-binding proteins. One clone, designated zinc-finger protein 174 (ZNF174), corresponds to a putative transcription factor with three zinc fingers and a novel finger-associated domain, designated the SCAN box. The three Cys-His-type zinc fingers are positioned at the carboxyl terminus, while the 65-amino acid finger-associated SCAN box is located near the amino terminus. Chromosomal localization using somatic cell hybrid analysis and fluorescent in situ hybridization mapped the gene for ZNF174 to human chromosome 16p13.3. The 2.5-kilobase transcript from this gene is expressed in a variety of human organs, but most strongly in adult testis and ovary. Fusion of the upstream regulatory region of ZNF174 to the DNA-binding domain of GAL4 revealed that the gene could confer a repression function on the heterologous DNA-binding domain. ZNF174 selectively repressed reporter activity driven by the platelet-derived growth factor-B chain and transforming growth factor-β promoters and bound to DNA in a specific manner. This member of the CH-type zinc-finger family is a novel transcriptional repressor.

To identify genes that can repress the expression of growth regulatory molecules, a human fetal cDNA library was screened with a degenerate oligonucleotide that corresponds to the conserved stretch of 6 amino acids connecting successive zinc-finger regions in the Wilms' tumor suppressor/Egr-1 family of DNA-binding proteins. One clone, designated zinc-finger protein 174 (ZNF174), corresponds to a putative transcription factor with three zinc fingers and a novel finger-associated domain, designated the SCAN box. The three Cys 2 -His 2type zinc fingers are positioned at the carboxyl terminus, while the 65-amino acid finger-associated SCAN box is located near the amino terminus. Chromosomal localization using somatic cell hybrid analysis and fluorescent in situ hybridization mapped the gene for ZNF174 to human chromosome 16p13.3. The 2.5-kilobase transcript from this gene is expressed in a variety of human organs, but most strongly in adult testis and ovary. Fusion of the upstream regulatory region of ZNF174 to the DNA-binding domain of GAL4 revealed that the gene could confer a repression function on the heterologous DNA-binding domain. ZNF174 selectively repressed reporter activity driven by the platelet-derived growth factor-B chain and transforming growth factor-␤ 1 promoters and bound to DNA in a specific manner. This member of the C 2 H 2 -type zinc-finger family is a novel transcriptional repressor.
Cellular responsiveness to environmental signals is mediated, at least in part, by the induction or modulation of gene transcription. Relatively detailed models of basal and activated transcription have been generated in eukaryotes (reviewed by Tjian and Maniatis (1994)). Repression, an equally important aspect of transcriptional control, is not nearly as well characterized. Transcriptional repressors can inhibit gene expression by at least two general mechanisms. Passive repressors downregulate the activity of transcription factors by competing for their DNA binding sites; alternatively, active repressors have an intrinsic repressing activity and directly inhibit transcription (reviewed by Cowell (1994) and Johnson (1995)). The regulation of growth factor gene expression can involve both types of repressive mechanisms. There is increasing evidence that members of the Cys 2 -His 2 zinc-finger class of transcription factors are involved in the transcriptional repression of growth factor gene expression. The Wilms' tumor suppressor gene (wt-1) encodes a zinc-finger DNA-binding protein, which can interact with the insulin-like growth factor type II promoter to repress expression of the gene during nephrogenesis (Drummond et al., 1992;Pritchard-Jones et al., 1990). Similarly, WT-1 can also repress expression driven by the transforming growth factor TGF-␤ 1 1 (Dey et al., 1994), PDGF-A chain (Gashler et al., 1992), colony-stimulating factor-1 (Harrington et al., 1993) and the insulin-like growth factor-I (IGF-I) receptor (Werner et al., 1994) promoters. Recent studies have revealed that WT-1 expression can be autoregulated through multiple potential WT-1 binding sites located in its promoter (Fraizer et al., 1994;Malik et al., 1994). The loss of function of the wt-1 gene product is thought to contribute to neoplastic transformation. Structural alteration of the wt-1 gene or its abnormal expression have been implicated in the continued proliferation of embryonic kidney blastemal cells seen in Wilms' tumors. Mutations in the wt-1 gene have been detected in 5-10% of Wilms' tumors Gessler et al., 1990;Call et al., 1990;Huff et al., 1991).
Since the expression pattern of WT-1 is restricted to specific cell types (reviewed by Rauscher (1993)), we reasoned that there may be additional Cys 2 -His 2 factors, which could play an important role in negative growth factor gene regulation. In this report, we have cloned and characterized a novel zincfinger-containing gene, ZNF174, from a human fetal library. ZNF174 corresponds to a putative transcription factor with three zinc fingers and a novel finger-associated structural element. When the amino-terminal region of ZNF174 was fused to the DNA-binding domain of GAL4, the chimera repressed transcription of a reporter construct containing multiple GAL4 binding sites. Moreover, overexpression of full-length ZNF174 cDNA with a series of promoter-reporter constructs revealed that ZNF174 could repress expression driven by the human PDGF-B chain and TGF-␤ 1 promoters, but not the PDGF-A chain, Egr-1, c-Fos, or c-Jun promoters. ZNF174 repression of PDGF-B promoter activity was dose-dependent, and 5Ј-deletion analysis mapped the putative negative element to the proximal promoter. DNase I footprint and gel shift analysis revealed that recombinant ZNF174 protein binds to an element near the transcriptional start site, indicating that ZNF174 repression may be mediated by direct interaction with DNA. These results suggest that ZNF174 may contribute to the negative regulation of certain growth factor genes.

RNA Isolation and Construction of a Human Fetal cDNA Library-
Total RNA was extracted with guanidine isothiocyanate from each organ and purified by spinning through a cesium-chloride cushion (Sambrook et al., 1989). Poly(A ϩ ) RNA was prepared from 10-and 12-week human embryos according to standard procedures. Tissue was obtained under an approved Brigham and Women's Hospital Human Subjects Protocol for discarded human tissue. The techniques of Gubler and Hoffman (1983) were used for cDNA synthesis. Briefly, oligo(dT) was used as a primer for first strand synthesis. Following second strand synthesis, NotI/EcoRI linkers were added. The library was size-selected for inserts greater than 1.0 kb, ligated to calf intestinal phosphatasetreated gt10 arms, and recombinant bacteriophage generated with high efficiency packaging extract. The library had a primary size of 2,750,000 recombinants with a vector background of less than 10%. Randomly chosen cDNA inserts ranged in size from 300 to 7000 bp. The partial cDNA clones were obtained from the amplified portion of the library.
cDNA Cloning-A region from the conserved "knuckle" section of the WT-1/Egr zinc fingers (TGEKPFQC) was used to generate a degenerate 24-bp oligonucleotide (5Ј-ACIGG(C/A)GAGAAGCC(T/A)TTCCA(A/ G)TGT-3Ј). The oligonucleotide was used as a forward primer with a 3Ј reverse primer (5Ј-CTTATGAGTATTTCTTCCAGGGTA-3Ј) from the gt10 phage arms. The human fetal library was used as a DNA source with the following polymerase chain reaction (PCR) protocol. Thirty PCR cycles were performed: denaturation at 92°C for 30 s, annealing at 50°C for 30 s, and extension for 1 min at 72°C. The prominent 590-bp reaction product from a parallel set of six PCR reaction tubes was pooled and purified. The ends of the PCR product were repaired with the Klenow fragment of DNA polymerase I and phosphorylated with T4 kinase, and the fragment was subcloned into the HincII site of the plasmid vector pBS (Stratagene, La Jolla, CA), according to standard procedures (Sambrook et al., 1989). A reverse primer from the 5Ј end of the initial partial cDNA clone (5Ј-CTAAAGCATGGAGTGCTCCTTT-TAGTC-3Ј) was used with a 5Ј forward gt 10 primer (5Ј-AGCAAGT-TCAGCCTGGTTAAG-3Ј) and the human fetal library to clone the 5Ј portion of the cDNA. PCR conditions were: denaturation at 92°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min. A 1170-bp amplification product was subcloned into pBS, as described above, and designated ZNF174.
DNA Sequencing and Sequence Analysis-Nucleotide sequence was determined by the dideoxynucleotide chain termination procedure with modified T7 DNA polymerase (U. S. Biochemical Corp.) and 35 S-dATP. Oligonucleotide primers were synthesized using an oligonucleotide synthesizer (Applied Biosystems) and were used without purification. Protein secondary structure predictions were made using the "HYBRID" system (Zhang et al., 1992).
Chromosomal Mapping and in Situ Hybridization-Southern blots were prepared from BamHI-digested DNA from 34 human-mouse somatic hybrid cell lines involving 19 unrelated human cell lines and 4 mouse cell lines (Shows et al., 1982). The hybrids were characterized by karyotypic analysis and by mapped enzyme markers (Shows, 1983). Blots were hybridized with an EcoRI-HindIII restriction fragment of the ZNF174 cDNA labeled by random priming, washed at high stringency, and autoradiographed.
Fluorescent in situ hybridization to metaphase chromosomes was used to localize an ZNF174 genomic clone to the p13.3 region of chromosome 16 (Cherif et al., 1989;Fan et al., 1990). Metaphase chromosomes were prepared from 5-bromodeoxyuridine-synchronized lymphocyte cultures. The probe was biotinylated, hybridized to the chromosome spreads, and detected by fluorescein-conjugated avidin (Vector Laboratories). Slides were evaluated with a Nikon florescence microscope. Forty-one metaphases were examined. Q (DAPI counterstaining) and R banding (propidium iodide counterstaining) were used to confirm the identity of the chromosome. The idiogram demonstrates the localization of an ZNF174 genomic clone to 16p13.3 by double fluorescent signals, one on each chromatid.
Construction of GAL4-ZNF174 Chimeras-The starting plasmid pSG-424 (Sadowski et al., 1988), contains the DNA-binding domain of GAL4 driven by the SV40 early promoter/enhancer followed by a polylinker. The region of ZNF174 upstream of the zinc-finger elements was fused in frame COOH-terminal to GAL4 amino acids 1-147. The regions of ZNF174 that were used in these constructs (amino acids 3-128 and 3-320) were amplified by PCR, digested with BamHI and XbaI, and cloned into the corresponding sites of pSG424. The amplified region and the junction of the construct were verified by dideoxy sequencing. To measure repression, a reporter with basal activity and five GAL4 binding sites, pGAL4 5 tkCAT (Shi et al., 1991), was selected. To measure activation, a reporter with minimal basal activity and five GAL4 binding sites, pGAL4 5 E1bCAT (Lillie et al., 1989), was chosen.
Transfection and Analysis of Chimeric Constructs-COS cell cultures (10 6 cells/100-mm culture dish) were transfected by a modified calcium phosphate precipitation technique (Sambrook et al., 1989) using 3 g of reporter and 20 g of expression constructs to test for activation. BALB/c 3T3 cells were transfected by the same method using 10 g of reporter and 2 g of expression constructs to test for repression. COS cells could not be used for repression studies in light of apparent activation of the TK promoter-reporter constructs by the pSG-424 expression vector in this cell type. Relative transfection efficiency was determined by cotransfection with CMV-␤-galactosidase (2 g). After transfection, cell extracts were prepared as described previously (Neish et al., 1992). Transfection of each promoter-reporter construct was performed in duplicate for each experiment and the experiments were repeated at least three times using at least two independent plasmid preparations.
Conversion of radiolabeled acetyl-coenzyme A to acetylated chloramphenicol was assayed by the two-phase fluor diffusion technique (Sambrook et al., 1989). Assays for ␤-galactosidase activity were performed in a 100-l volume containing 10 l of extract, 0.1 M sodium phosphate, 0.01 MgCl 2 , 0.1 M 2-␤-mercaptoethanol, and 3 mM O-nitrophenyl-␤-D-galactopyranoside (Sigma). Reactions were incubated at 37°C for 1 h and stopped by the addition of sodium bicarbonate to 0.6 M. Assays were compared to a standard curve on a Dynatech Lab MR700 microtiter plate reader at 410 nm.
Bacterial Expression and Protein Purification-A plasmid containing a portion of the ZNF174 cDNA was digested with PstI to excise a 520 bp fragment between a PstI site upstream from the zinc-finger region and a PstI site in the polylinker just downstream from the translational stop site. This fragment was treated with T4 DNA polymerase to create blunt ends, and then ligated into pET vector 15b (Novagen) that had previously been digested with XhoI and filled in with DNA Polymerase I (Klenow fragment). pET expression vectors contain the T7 promoter, as well as transcription and translation signals. pET-15b has a His-tag leader sequence that codes for 6 consecutive histidine residues, which serve as an affinity tag for protein purification using a nickel chromatography column. Inserted genes are expressed by transforming the expression vector into a bacterial host (BL21(DE3)) that contains the T7 RNA polymerase gene under lacUV5 control. Transformed bacteria are grown to an A 600 of 0.8 then induced with 0.4 mM isopropyl-1-thio-␤-Dgalactopyranoside for 3 h. Cell extracts were prepared according to the pET system manual and the recombinant protein purified by column chromatography using Novagen His-Bind resin under denaturing conditions of 6 M urea. The protein was renatured slowly by dialysis against buffer containing 25 mM HEPES, pH 7.6, 100 mM KCl, 0.1 mM EDTA, 12.5 mM MgCl 2 , 10 M ZnSO 4 , 10% glycerol, 0.1% Nonidet P-40, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, with 5, 4, 3, 2, 1.5, 1, 0.75, 0.5, 0.25, and 0 M urea, respectively.
DNase I Footprint Analysis-A 530-bp EcoRI restriction fragment from the human PDGF-B promoter construct d36 (Resnick et al., 1993) was dephosphorylated with calf intestinal alkaline phosphatase (New England Biolabs) and isolated by 1% agarose gel electrophoresis and glass bead purification. The fragment was end-labeled with [␥-32 P]dATP using T4 polynucleotide kinase (New England Biolabs) and purified using Chromaspin-10 spin columns (Clontech). The fragment was cut once with HindIII to yield two labeled fragments of 250 and 280 bp in length. The smaller 32 P-labeled fragment contained 202 bp of the PDGF-B promoter in the sense orientation and extended 77 bp into the 5Ј-untranslated region. This fragment was isolated by 3% agarose gel electrophoresis and glass bead purification. G ladders were generated by methylating the probe with dimethyl sulfate (Aldrich) and piperidine (Sigma) cleavage.
Electrophoretic Mobility Shift Assay-An oligonucleotide bearing the PDGF-B chain transcriptional start site sequence and its flank was 32 P-labeled and incubated with 1 g of ZNF174 protein in 10 mM Tris, pH 7.5, 50 mM NaCl, 1 mM EDTA, 5% glycerol, 1 mg/ml bovine serum albumin, 0.5% Nonidet P-40, and 1 mM dithiothreitol for 30 min at 22°C. In competition studies, unlabeled oligonucleotide was added 10 min prior to the addition of probe. Non-denaturing sample buffer was added to the binding reaction and loaded onto a 5% polyacrylamide/Tris borate/EDTA gel and run at 200 V for 2 h at 22°C.

RESULTS
Isolation of cDNA Clones-A predicted peptide from the conserved "knuckle" region of the WT-1/Egr zinc fingers (TGEK-PFQC) was used to synthesize a degenerate 24-mer oligonucleotide. This probe was used in a PCR strategy to screen a human fetal cDNA library generated from 10 -12-week embryos. The degenerate oligonucleotide was used as the forward primer with a 3Ј-reverse primer derived from the bacteriophage arms. DNA sequence analysis of one of the amplification products revealed a novel zinc-finger-containing structure. The open reading frame was completed by generating a series of overlapping partial cDNA clones using standard anchored PCR, as well as by rescreening additional cDNA libraries. DNA sequence from at least two independent clones for each amplification product was obtained to eliminate PCR-generated sequence errors.
Sequence analysis of the cDNA clones revealed a single open reading frame of 1221 nucleotides (Fig. 1A). The open reading frame is closed by a termination codon located 93 bp upstream of the translational start site. The polypeptide predicted from the open reading frame has a calculated relative molecular mass of 46,414 Da and an estimated isoelectric point of 10.1. A long 5Ј-untranslated region containing multiple translational start and stop codons flanks the open reading frame. The cDNA has a relatively short 3Ј-untranslated region with a consensus polyadenylation signal (AAUAAA) and poly(A) tail. A sche-matic representation of ZNF174 cDNA appears in Fig. 1B.
Structural Properties of the Putative ZNF174 Polypeptide-The predicted protein contains three tandemly repeated sequence motifs related to the zinc fingers of the WT-1/Egr-1/Sp1 or Drosophila Kruppel gene family of transcription factors ( Fig.  2A). All three zinc fingers encoded by the cDNA of ZNF174 fit the consensus sequence for this type of zinc finger. The conserved amino acid "knuckle" (or H/C link) between zinc fingers, typified by the amino acid sequence TGE(R/K)P(F/Y)X, is also conserved in the deduced amino acid sequence. From these features it is reasonable to predict that ZNF174 could encode a DNA-binding protein with transcriptional regulatory properties.
Outside the putative zinc-finger DNA-binding domain, the predicted protein contains proportionately high numbers of serine (9.3%), glycine (8.9%), proline (8.6%), and glutamine (8.2%) residues (Fig. 1A). Outside the strikingly high content and stretches of these residues, characteristics of the other transcription factors in the WT-1/Egr/Sp1 family were not seen. Although amino acid sequence homology within these enriched regions is not generally observed among transcription factors, the high content of these amino acids in the non-zinc-finger segment of the predicted protein fits the motif of a number of known and putative transcription factors (Mitchell and Tjian, 1989). A single consensus phosphorylation site (Ser*/Thr*-Pro-X-Lys/Arg) for Cdc2 kinase was identified in the open reading frame (SPLK, Fig. 1A). An adjacent region rich in basic amino acid residues (LKKSKGGKR) (Fig. 1A) suggests the existence of a nuclear localization signal within the protein.
Upstream Finger-associated SCAN Box-The deduced amino acid sequence contains a novel element upstream of the zincfinger domain that consists of approximately 65 amino acids. This region, which we have designated the SCAN box, was identified by its homology with similar elements in several other zinc-finger transcription factors such as SRE-ZBP, CTfin-51, #18 cDNA, ZNF165, and 3c3 (Fig. 2B). The term "SCAN" box was derived from the first letters of the names of the four proteins initially found to contain this domain (SRE-ZBP, CTfin-51, AW-1, 2 Number 18 cDNA). The SCAN box is separated from the zinc-finger domains by amino acid regions that vary in length and do not share extensive sequence homology. The element is enriched in hydrophobic and negatively charged residues with the L(X 6 )L motif at its core. This core is flanked by certain residues (e.g. A, E, L, M, H, C) that are frequently found in ␣-helices. This element also has a high proportion of glutamic acid residues, suggesting that it constitutes a negatively charged acidic domain, commonly found in transcription factors. When the SCAN box amino acid sequence is plotted as a helical wheel representation (Fig. 2C), there appears to be a higher proportion of hydrophobic residues on one side of the helix as compared to the opposite side, especially in the first five turns of the helix. This feature suggests that ZNF174 may have the ability to multimerize with either itself or other molecules with a similar domain.
Tissue Distribution of the ZNF174 Transcripts-ZNF174 cDNA was used as a probe to determine the expression pattern of the zinc-finger protein in a variety of human organs. Northern blot analysis defined a major transcript of approximately 2.5 kb, which is consistent with the size of the overlapping cDNA clones. The gene is widely expressed in many adult organs, with the highest levels found in ovary and testis. Interestingly, two minor transcripts of about 1.5 and 3 kb in size were also detected and found to be expressed at different levels in various tissues (Fig. 3). In contrast to low transcript levels detected in adult kidney, the ZNF174 gene is highly expressed FIG. 2. Amino acid alignment of zinc fingers and finger-associated domains of ZNF174 with those of other zinc-finger proteins. A, zinc-finger sequence alignment. B, alignment of zinc-finger-associated domains. Regions of identity with ZNF174 are highlighted. The putative ␣-helical region is indicated (H) below the sequence. C, helical wheel representation of the SCAN box amino acid sequence from ZNF174. Each residue is offset from the preceding one by 100°, the typical angle of rotation for an ␣-helix. Analysis was done with the assistance of the GCG Helical Wheel graphics program. Boxed residues indicate hydrophobic amino acids.
in normal 18 -20-week human fetal kidney specimens. 3 The 2.5-kb transcript was detected in total RNA from a variety of individual cell types, including human umbilical vein endothelial cells, human foreskin fibroblasts, and HepG2 cells. 3 Chromosomal Localization of the ZNF174 Gene-The gene for the zinc-finger protein maps to a locus at 16p13.3. DNA isolated from 34 human-mouse somatic hybrid cell lines and parental controls was examined for the presence or absence of the human gene by Southern blotting techniques. On BamHIdigested DNA, the ZNF174 cDNA produced readily distinguishable patterns of restriction fragments (data not shown). The presence or absence of the human gene was tabulated against the presence or absence of human chromosomes in each hybrid (ZNF174/chromosome) for concordance (ϩ/ϩ, Ϫ/Ϫ) and discordancy (Ϫ/ϩ, ϩ/Ϫ). A 0% discordance indicated a matched segregation of the zinc-finger gene with a chromosome and was the basis for the initial chromosomal assignment (Table I). Thus, the ZNF174 gene mapped to human chromosome 16 by analysis of the somatic cell hybrids.
To map the ZNF174 gene more precisely, we performed fluorescent in situ hybridization to metaphase chromosomes. An idiogram indicates the distribution of fluorescent signal at 16p13.3 on both chromatids of a single chromosome 16 in 21 out of the 41 cells (51%) (Fig. 4A). Double signals also appeared once at 7q22 and twice at 16q12.1 3 ; however, the frequency of the appearance of these signals suggests false hybridization sites.
The ZNF174 gene was further localized to the polycystic kidney disease (PKD1) region of chromosome 16 by analysis of selected somatic cell hybrids. The breakpoints of two der(16) chromosomes isolated in rodent-human somatic cell hybrids (N-OH1, 23HA) approximate the telomeric and centromeric boundaries of the PKD1 region (Fig. 4C) (Germino et al., 1992). Southern blot analysis revealed that the gene for ZNF174 was contained within the breakpoint interval defined by these hybrids (Fig. 4B).
ZNF174 Encodes a Portable Transcriptional Repression Element-To assess whether ZNF174 can affect the transcriptional activity of other genes, the putative regulatory region of ZNF174 was fused to a heterologous DNA-binding domain of the yeast transcriptional activator GAL4. The amino-terminal region of ZNF174 was amplified by PCR and cloned into the expression vector pSG-424, which encodes GAL4 amino acids 1-147. The resulting chimeras were tested for their ability to transactivate a reporter construct containing five GAL4 binding sites in front of either the adenovirus E1b or thymidine kinase (TK) minimal promoters (Fig. 5A). To determine whether ZNF174 encoded a repressor function, the GAL4-ZNF174 construct was cotransfected with a GAL4 reporter with high basal expression suitable for examining transcriptional repression (pGAL4 5 TK-CAT). The GAL4-ZNF174 chimera containing ZNF174 amino acids 3-320 suppressed transcription about 5-fold in this assay (Fig. 5B). However, the chimera containing amino acids 3-128, which is the portion of the gene that corresponds to the SCAN box, was not able to suppress transcription. Repression was dependent on the presence of the GAL4 binding sites (Fig. 5, A and B). These observations demonstrate that ZNF174 contains a transcriptional repression element that is portable in the context of a heterologous DNA-binding domain.
Certain zinc-finger transcription factors, such as Egr-1, have the capacity to regulate transcription in a bimodal manner (Gashler et al., 1993). To determine whether ZNF174 also encodes an activator function, the GAL4-ZNF174 chimera was 3 A. Williams and T. Collins, unpublished observations.

FIG. 3. Expression pattern of the ZNF174 gene.
Northern blots containing 2 g of poly(A ϩ ) RNA/lane from 16 different human adult tissues (Clontech) were hybridized with an EcoRI fragment of ZNF174 that did not contain the zinc-finger region. Lane 1 contains RNA size markers (not shown). Lanes 2-17 contain mRNA from human heart, brain, placenta, lung, liver, skeletal muscle, kidney, pancreas, spleen, thymus, prostate, testis, ovary, small intestine, colon, and peripheral blood leukocytes, respectively. The Northern blots were also hybridized with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to normalize for loading differences.  (Shows et al., 1982) digested with BamHI. The hybrids were characterized by karyotypic analysis and by mapped enzyme markers (Shows, 1983). A t in the table indicates a translocated chromosome with no intact chromosome present and is not tabulated for percent discordance. Scoring was determined by the presence (ϩ) or absence (Ϫ) of human bands in the hybrids on the blots and was compared to the presence or absence of human chromosomes in each hybrid. A 0% discordance indicates a matched segregation of the DNA probe with a chromosome. The gene maps to human chromosome 16. cotransfected with a reporter containing five GAL4 binding sites in front of the E1b minimal promoter (pGAL4 5 E1bCAT) (Fig. 5A). This construct, however, failed to stimulate the expression of the reporter gene (Fig. 5B). In contrast, a similarly constructed Egr-1 (amino acids 3-281) chimera (Gashler et al., 1993) stimulated transcription by approximately 100-fold. Thus, ZNF174 (amino acids 3-320) does not confer positive transcriptional activity on a heterologous DNA-binding domain.
Full-length ZNF174 Inhibits Reporter Gene Expression-To determine whether ZNF174 could repress the expression of authentic promoter-reporter gene constructs, full-length ZNF174 cDNA was inserted into an expression vector and introduced with a series of chimeric genes in a cotransfection protocol. The reporter genes included a series of growth factor gene promoters (TGF-␤ 1 , IGF-II, PDGF-A, and PDGF-B), transcription factors (egr-1, c-myc, c-fos, and c-jun), and endothelial leukocyte adhesion molecules (E-selectin and vascular cell adhesion molecule-1). Expression driven by the human TGF-␤ 1 (ph7-1-1) and PDGF-B (d77) promoters was significantly inhibited when ZNF174 was coexpressed (Fig. 6). In contrast, expression driven by the IGF-II, PDGF-A, egr-1, c-fos, and c-jun promoters was unaffected by ZNF174 (Fig. 6). In addition, E-selectin and vascular cell adhesion molecule-1 reporter gene expression was unaffected by the gene (data not shown). These findings demonstrate that ZNF174 is able to suppress authentic promoter-reporter gene constructs in a selective manner.

ZNF174 Interacts with the PDGF-B Promoter and Represses
Expression Driven by the PDGF-B Promoter-The preceding data indicate that ZNF174 has the capacity to modulate the expression of pathophysiologically relevant genes in the context of promoter-reporter constructs in transfected cells. Bovine aortic endothelial cells (BAEC) were used in cotransfection studies to determine whether overexpression of ZNF174 can repress PDGF-B promoter-reporter gene expression in cells in which the PDGF-B gene is expressed constitutively at high levels (Khachigian et al., 1994). To determine whether ZNF174 could affect expression driven by the PDGF-B promoter, a reporter construct bearing 1.3 kb of promoter sequence (6a-CAT) was used in a transient cotransfection setting with the ZNF174 expression construct. ZNF174 inhibited expression of 6a-CAT in a dose-dependent manner (Fig. 7). In contrast, the pcDNA backbone alone had no effect (Fig. 7). Additionally, a construct directing expression of a region of ZNF174 without the zinc-finger domain fused to an irrelevant DNA binding element (pGAL4 ZNF174 aa3-320), had no effect on expression of the PDGF-B chain promoter. This result indicates that the DNA-binding domain of ZNF174 is required for repression. Deletion derivatives of the B chain promoter (Khachigian et al., 1994) were used to map the ZNF174 binding site. Expression driven by the PDGF-B promoter deletion construct (d77-CAT), extending 82 bp upstream of the TATA box and defined previously as the minimal region required for basal expression in BAEC (Khachigian et al., 1994), was inhibited by ZNF174 in a concentration-dependent manner (Fig. 7). The ability of ZNF174 to repress expression driven by several reporter constructs including the minimal PDGF-B promoter and its structural similarity with a number of zinc-finger transcription factors suggests that ZNF174 repression of gene expression may be mediated by direct interaction with DNA elements within the core promoter.
To examine whether ZNF174 protein interacts directly with the PDGF-B promoter, DNase I footprint and gel shift assays were performed with a recombinant form of ZNF174. Since ZNF174 could suppress expression of d77-CAT as effectively as 6a-CAT (Fig. 7), a footprint probe encompassing the minimal promoter was used to allow visualization of bases protected from DNase I digestion. Bacterially expressed ZNF174 bound to a defined region in the PDGF-B promoter in a dose-dependent manner (Fig. 8A). To demonstrate the specificity of the interaction, another zinc-finger protein, Egr-2, failed to interact with this site (Fig. 8A).
In support of these observations, a single nucleoprotein complex was observed when a 32 P-labeled oligonucleotide (B-prom) spanning the region bound by ZNF174 was used in electrophoretic mobility shift assay with the recombinant protein (Fig.  8B). The shift was abolished by the presence of a 50-fold molar excess of the unlabeled cognate (Fig. 8B), whereas 100-fold excess of an unrelated oligonucleotide, PEA-3, failed to compete (Fig. 8B). The region bound by ZNF174 spans the major transcriptional start site in the human PDGF-B gene (Rao et al., 1986). These findings suggest that ZNF174 may serve as a competition-type transcriptional repressor (reviewed by Levine and Manley (1989)). Thus, ZNF174 can interact with the PDGF-B chain promoter in a specific manner and suppress the activity of B chain reporter constructs.

DISCUSSION
This paper describes the isolation and characterization of a novel zinc-finger protein, ZNF174, located on human chromosome 16p13.3 with the apparent ability to repress expression driven by the promoters of a number of pathophysiologically relevant genes. ZNF174 was also found to have a number of interesting structural features. The primary sequence of the putative ZNF174 polypeptide revealed a potential single phos- FIG. 4. Chromosomal location of the ZNF174 gene. A, idiogram of the distribution of ZNF174 signals on both chromatids of chromsome 16. Each dot represents two fluorescent signals, one on each chromatid of a chromosome 16. 51% of the cells examined (21/41) had signals on both chromatids of a single chromosome 16 at p13.3. B, mapping of the ZNF174 gene within the PKD1 region of chromosome 16. ZNF174 does not hybridize to a somatic cell hybrid with a breakpoint at 23HA, but is able to hybridize to a different somatic cell hybrid whose breakpoint extends to N-OH1. C, schematic representation of chromosome 16p13.3 indicating the location of the somatic cell hybrid breakpoints. Also shown are the approximate locations of the ␣ globin gene, MS205.2 (a microsatellite marker), TSC-2 (tuberous sclerosis gene), and PKD-1 (polycystic kidney disease gene). ZNF174 maps to the region between the N-OH1 and 23HA breakpoints. phorylation site and nuclear translocation sequence. The SPLK sequence (Fig. 1A) fits the consensus phosphorylation sequence recognized by Cdc2 kinase (Ser*/Thr*-Pro-X-Lys/Arg). Cdc2 is a highly conserved cell cycle-regulatory protein serine kinase that has been reported to phosphorylate nuclear transcription factors (reviewed by Hunter and Karin (1992)). Located near FIG. 5. Repression by GAL4-ZNF174 chimeras. A, schematic representation of various plasmid constructs used to assess the function of upstream domains of ZNF174. B, CAT activities from cotransfections with GAL4-ZNF174 fusion constructs and promoter-CAT reporter plasmids. The left portion of the figure shows transfections with the GAL4x5 E1B CAT reporter, which contains five copies of the GAL4 binding site placed in front of the minimally active E1B promoter and CAT. The plasmid pE1B CAT does not contain GAL4 binding sites and is used as a control. As a positive control, a cotransfection was done with a known activator GAL4-Egr-1 (amino acids 3-281) (Gashler et al., 1993) The right portion of the figure depicts transfections with the reporter GAL4x5 TK-CAT, which is a pBL-CAT2-based plasmid with 5 copies of the GAL4 binding site placed in front of the basally active TK promoter. The histograms represent the mean of four replicative assays. The error bars indicate standard deviation from the mean. Where the error bars are not visible, this indicates that the standard deviation was smaller than the bar data point. the phosphorylation site, and just upstream of the zinc-finger domain, is a region that contains a high proportion of basic amino acid residues (LKKSKGGKR) (Fig. 1A). Nuclear localization sequences often share a sequence of four predominantly basic amino acids (Lys-Arg/Lys-X-Arg/Lys) flanked by acidic and proline residues (reviewed by Powers and Forbes (1994)). The core of this short stretch of basic residues in ZNF174 could constitute a nuclear localization signal or may be part of more complex bipartite signal involving another region of the protein, as observed with Egr-1 (Gashler et al., 1993). Phosphorylation at the SPLK site may affect the function of the adjacent nuclear localization signal and regulate import of the protein into the nucleus (reviewed by Silver (1991)).
The ZNF174 gene contains a novel finger-associated element located upstream of the zinc-finger domain. This domain, or SCAN box, is also found in a number of other zinc-finger proteins including SRE-ZBP, CTfin-51, #18 cDNA, ZNF165, and 3c3 (Fig. 2B). This domain may contain amphipathic ␣-helices (Fig. 2C), which would permit interactions with other proteins containing similar domains. CTfin-51 is a mouse protein with seven tandemly repeated carboxyl-terminal zinc fingers, which has been functionally associated with meiosis in both male and female gametogenesis (Noce et al., 1992) and has also been shown to be a strong transcriptional activator (Chowdhury et al., 1992). SRE-ZBP is a human transcription factor that binds to the c-fos serum response element. This protein also contains seven tandemly repeated carboxyl-terminal zinc fingers and is thought to function as a repressor of c-fos transcription (Attar and Gilman, 1992). ZNF165 4 is a zincfinger gene located on chromosome 6p21, which is expressed specifically in the testis. The element was also identified in two novel partial cDNAs termed #18 cDNA (Pengue et al., 1993) and 3c3 (Calabro et al., 1995), both of which correspond to transcribed sequences from chromosome 3p21. The 3p region is of particular interest because of its frequent involvement in rearrangements and/or deletions associated with various hu-man tumors, including lung and renal carcinomas (Naylor et al., 1987;Bergenhein et al., 1989). The conserved SCAN box does not correspond with previously reported modules linked to zinc-finger domains such as the 75-amino acid Kruppel-associated box (KRAB), which occurs in one-third of the Kruppel-type finger genes (Bellefroid et al., 1991;Rosati et al., 1991) and mediates transcriptional repression (Witzgall et al., 1994;Margolin et al., 1994), or the finger-associated box (FAX), which has been found in Xenopus finger proteins (Knochel et al., 1989). Although the function of the SCAN box has not yet been elucidated, the conservation of this module and its ␣-helical structure suggests that it may serve ZNF174 as a dimerization domain or a site that interacts with components of the transcriptional machinery resulting in the repression of gene expression.
The ability of ZNF174 to function as a transcriptional repressor is similar to the activities of other zinc-finger transcription factors such as Egr-1 and WT-1 (Madden et al., 1991;Gashler et al., 1993). Repression domains, like activation domains, have been demonstrated to function as independent modular elements in transcription factors (reviewed by Levine and Manley (1989) and Johnson (1995)). Only a small number of repression domains have been well characterized. These include the Drosophila proteins Kruppel (Licht et al., 1990;Zuo et al., 1991), Engrailed (Han et al., 1989;Jaynes and O'Farrell, 1991), and Even-Skipped (Han and Manley, 1993) and the mammalian FIG. 6. Effect of ZNF174 on the expression of various promoter-CAT constructs in COS cells. Fifteen g of the various promoter-CAT constructs were transiently cotransfected into COS cells with 2 g of ZNF174 expression construct (pcZNF174) or 2 g of pcDNA using calcium phosphate technique. CAT activity was assessed as described under "Experimental Procedures." Transcriptional repression is expressed as a percentage of the reporter activity obtained using the empty expression vector (pcDNA) alone. The values (% repression) represent the mean of between two and five separate experiments.

FIG. 7. Effect of ZNF174 on expression driven by PDGF-B chain promoter fragments in BAEC.
Fifteen g of PDGF-B 6a-CAT or d77-CAT constructs were transiently cotransfected into COS cells with 0, 0.5, 2, and 8 g of ZNF174 expression construct (pcZNF174), the backbone alone (pcDNA), or a construct containing the amino-terminal region of ZNF174 without the zinc fingers fused to an irrelevant DNA binding site (GAL4), using a calcium phosphate technique. CAT activity was assessed as described under "Experimental Procedures." The data are representative of two independent experiments. DNA-binding proteins WT-1 (Madden et al., 1991), YY1/NF-E1 (Shi et al., 1991), and Kid-1 (Witzgall et al., 1994). Like ZNF174, these proteins can confer their repression function onto heterologous DNA-binding domains. The specific region(s) within ZNF174 that repress transcription of other genes is not yet clear, although the GAL4 chimera studies were able to show that the SCAN box alone can not confer repression. This implies that either the repression domain of ZNF174 is contained within another portion of the gene, or that for conformational reasons a more extensive region of the gene is required in order to confer repression. The suppression domain in Kruppel was mapped to an alanine-rich domain (Zuo et al., 1991), and that of Egr-1 was mapped to a 34-amino acid element (Gashler et al., 1993). ZNF174, however, does not contain an alanine-enriched element or a region that resembles the repression element of Egr-1 (Fig. 1A). A small number of factors contain modular domains capable of regulating transcription both positively and negatively. These include Egr-1 (Gashler et al., 1993), WT-1 (Wang et al., 1993), Kruppel (Zuo et al., 1991), YY1/NF-E1 (Park et al., 1991), and the immediateearly proteins c-Fos and c-Jun (Abate et al., 1991). The inability of ZNF174 to activate transcription in the context of promoterreporter constructs examined (Figs. 5B, 6, and 7) does not exclude the possibility that the zinc-finger protein has a positive regulatory function with other genes.
Gel retardation and DNase I footprint studies established the ability of the purified zinc-finger domain of ZNF174 to interact with promoter elements that appear in certain reporter constructs it could repress. These assays shed light on a possible mechanism with which ZNF174 suppressed expression driven by the PDGF-B promoter. That ZNF174 bound at the transcriptional start site of the PDGF-B gene (Rao et al., 1986) suggests that it may impair the assembly of various components of the general transcriptional machinery into a functional complex at the initiation site. Thus, ZNF174 could function as a competition-type transcriptional repressor (Levine and Manley, 1989) as recently observed with another zinc-finger protein (Werner et al., 1994). Transcriptional repression of the IGF-I receptor gene by WT-1 is thought to be mediated, at least in part, by direct interaction with the initiator site (Werner et al., 1994).
The gene encoding ZNF174 was mapped to human chromosome 16p13.3. Because two disease loci have been identified in this region near the distal end of the short arm of chromosome 16, tuberous sclerosis and polycystic kidney disease, it has been intensely studied and is known to be rich in transcribed sequences. Autosomal dominant polycystic kidney disease is a common genetic defect in humans that frequently results in renal failure (Gabow, 1993). The candidate gene for this disease occurs within a duplicated region on the chromosome and it encodes a large novel protein of unknown function (European Polycystic Kidney Disease Consortium, 1994). Tuberous sclerosis complex (TSC) is a dominantly inherited disorder that causes mental retardation, seizures and tumors in multiple organs (Gomez, 1988). Linkage studies have demonstrated that TSC has at least two possible genetic loci within the human genome with disease loci on chromosomes 9 (Fryer et al., 1987) and 16 (Kandt et al., 1992). A recently described locus, designated TSC-2, generates a transcript that is widely expressed, and its protein product, tuberin, has a region of homology to the GTPase-activating protein GAP3. Loss of heterozygosity for alleles at 16p has been observed in hamartomatous lesions seen in TSC patients (Green et al., 1994). Often, an alteration in one allele, such as a point mutation or small deletion, is followed by a more extensive loss of sequence from the homologous chromosome. If ZNF174 lies in close proximity to the TSC-2 gene, it may be lost in the second mutational event.
While not the causative gene for TSC, the inclusion or exclusion of the ZNF174 gene in subsequent chromosomal deletions may account for some of the variation of symptoms and complications seen in the disorder.