The c-myc DNA-unwinding Element-binding Protein Modulates the Assembly of DNA Replication Complexes in Vitro*

The presence of DNA-unwinding elements (DUEs) at eukaryotic replicators has raised the question of whether these elements contribute to origin activity by their intrinsic helical instability, as protein-binding sites, or both. We used the human c-myc DUE as bait in a yeast one-hybrid screen and identified a DUE-binding protein, designated DUE-B, with a predicted mass of 23.4 kDa. Based on homology to yeast proteins, DUE-B was previously classified as an aminoacyl-tRNA synthetase; however, the human protein is ∼60 amino acids longer than its orthologs in yeast and worms and is primarily nuclear. In vivo, chromatin-bound DUE-B localized to the c-myc DUE region. DUE-B levels were constant during the cell cycle, although the protein was preferentially phosphorylated in cells arrested early in S phase. Inhibition of DUE-B protein expression slowed HeLa cell cycle progression from G1 to S phase and induced cell death. DUE-B extracted from HeLa cells or expressed from baculovirus migrated as a dimer during gel filtration and co-purified with ATPase activity. In contrast to endogenous DUE-B, baculovirus-expressed DUE-B efficiently formed high molecular mass complexes in Xenopus egg and HeLa extracts. In Xenopus extracts, baculovirus-expressed DUE-B inhibited chromatin replication and replication protein A loading in the presence of endogenous DUE-B, suggesting that differential covalent modification of these proteins can alter their effect on replication. Recombinant DUE-B expressed in HeLa cells restored replication activity to egg extracts immunodepleted with anti-DUE-B antibody, suggesting that DUE-B plays an important role in replication in vivo.

The initiation of DNA replication in eukaryotes relies on the sequential assembly of protein complexes at replicator sequences, controlled by the activities of kinases and phospha-tases (1). Genetic and biochemical studies in Saccharomyces cerevisiae, Drosophila melanogaster, and Xenopus laevis suggest that the origin recognition complex (ORC) 1 is a component of the replication initiator that recruits Cdc6, Cdt1, and the minichromosome maintenance (MCM) proteins to origins late in mitosis to form the pre-replication complex (pre-RC). Activation of the pre-RC for replication requires the activity of S phase cyclin-dependent kinases plus the Cdc7/Dbf4 kinase and involves binding of MCM10, Cdc45, and replication protein A (RPA) to unwind DNA and to load DNA polymerases. A complex containing MCM proteins and Cdc45 may function as a replicative helicase to extend the unwound origin DNA (2,3).
In S. cerevisiae, chromosomal replication origins cloned in plasmids display autonomously replicating sequence (ARS) activity and characteristically comprise a set of modular elements, including an ARS consensus sequence (ACS) (4)-binding site for ORC (5), a region of helical instability termed a DNAunwinding element (DUE) that contributes to origin activity through template unwinding or binding of pre-RC proteins (6 -10), and transcription factor-binding sites that can promote the assembly of replication complexes through protein-protein interactions and modification of chromatin structure. In mammalian chromosomes, no consensus DNA sequence analogous to the yeast initiator-binding site has been identified. Instead, the feature most common to mammalian origins is a region of helical instability (11). Whereas defined sequences derived from the ␤-globin, lamin B 2 , and c-myc loci display replicator activity at ectopic loci (12)(13)(14)(15)(16)(17)(18), deletion of the 40-kb region encompassing the dihydrofolate reductase ori-␤ does not eliminate replication initiation in that endogenous location (19), suggesting that ectopic assays reveal the minimal elements essential for replication.
The 2.4-kb upstream region of the human c-myc gene contains multiple transcription factor target-binding sites (20). Our laboratory initially reported that replication begins in this region (21)(22)(23), and Vassilev and Johnson (24) were the first to use PCR quantitation of nascent DNA to define the replication initiation zone. Mapping of DNA nascent strands by our laboratory (15,18,(21)(22)(23)(25)(26)(27)(28) and confirmed by others (24, 29 -34) showed that replication initiates at multiple sites within this core domain and at flanking sites at the endogenous c-myc chromosomal location (27,28,32). Site-specific chromo-somal integration at an ectopic site mediated by the S. cerevisiae Flp recombinase showed that the c-myc origin core satisfies the genetic criteria of a chromosomal replicator and that a short segment of the c-myc replicator containing the DUE (26,35,36) and three matches to the yeast ACS is essential for replicator activity (15,18).
In eukaryotes, chromosomal replication contrasts with the replication of certain viral genomes (e.g. SV40) in terms of the number of proteins that intervene between the replication initiator and the cellular DNA polymerases. In an attempt to identify additional proteins that might modulate c-myc origin activity, we used a yeast one-hybrid assay to isolate proteins that bind to the c-myc DUE/ACS region. We report one such protein, designated DUE-B (for DUE-binding protein), that has a predicted molecular mass of 23.4 kDa and that shows strong evolutionary conservation in yeast, mice, frogs, and flies. In HeLa cells, DUE-B is a constitutively expressed protein found attached to chromatin and in the soluble fraction of lysed cells. During gel exclusion chromatography of HeLa cell nuclear extracts, DUE-B migrated at a size of ϳ46 kDa. Similarly, DUE-B expressed from a baculovirus vector chromatographed with an apparent molecular size of 46 kDa and co-purified with ATPase activity. In contrast, a significant portion of baculovirus-expressed DUE-B mixed with Xenopus egg or HeLa cell nuclear extracts eluted as a high molecular mass species (Ͼ250 kDa). Chromatin immunoprecipitation assays also show that DUE-B bound at or near the c-myc replicator DUE in a cell cycle-dependent manner in vivo.
Consistent with a possible role in replication initiation, DUE-B was preferentially phosphorylated in HeLa cells arrested early in S phase relative to cells blocked at G 1 or G 2 /M phase. Baculovirus-expressed DUE-B bound saturably to Xenopus sperm chromatin and inhibited its replication in Xenopus egg extracts. In this system, DUE-B did not inhibit pre-RC formation, but reduced replication in proportion to its inhibition of RPA binding. In HeLa cells, down-regulation of DUE-B protein expression by small interfering RNA (siRNA) was associated with a prolonged G 1 phase and the induction of cell death. These data suggest that the c-myc DNA-unwinding element-binding protein DUE-B plays a role in regulating replication initiation in HeLa cells.

EXPERIMENTAL PROCEDURES
One-hybrid Screen-The wild-type DUE/ACS region of the c-myc origin (nucleotides 735-832; GenBank TM /EBI accession number X00364) was cloned into vector pHisi-1 and transformed into S. cerevisiae strain YM4271 (MATa,112, trp1-901, tyr1-501, gal4-⌬512, gal80-⌬538, ade5::hisG). The wildtype and mutant DUE/ACS bait sequences are given in Table I. The bait sequences were amplified by PCR using primers to allow cloning between the SstI and EcoRI sites of pHisi-1. Yeast transformants were selected for growth on histidine-deficient synthetic dextrose medium (Clontech). The DWAW reporter strain was transformed with a HeLa cDNA library cloned into pGAD-GH (Clontech) to produce Gal4 activation domain (Gal4AD) fusion proteins, and colonies were selected for growth at 30°C on His/Leu-deficient medium containing 15 mM 3-aminotriazole. Plasmids were isolated from crude yeast lysates and cloned into Escherichia coli according to standard procedures. The identity of yeast strains containing mutant bait sequences was con-firmed by PCR amplification and restriction enzyme digestion of the bait sequences, and these strains were transformed with a plasmid expressing the Gal4AD-DUE-B fusion protein. Yeast doubling times were calculated by overnight growth in selective medium and dilution to A 600 ϭ 0.1 in selective medium containing 2 mM 3-aminotriazole. Doubling times were calculated from the linear portion of growth curves by monitoring the growth at A 600 from 0.3 to 0.9. Doubling times in the absence of 3-aminotriazole were the same for all four reporter strains, whereas yeast transformed with a plasmid containing only Gal4AD did not grow in medium containing 2 mM 3-aminotriazole. A plasmid expressing the Gal4AD-DUE-B fusion protein also resulted in elevated expression of a pLacZi-DUE/ACS reporter integrated at the URA3 locus in strain YM4271 (data not shown).
RNA and Protein Analyses-RNA was isolated using TRIzol, and DUE-B RNA was visualized by Northern blotting of total RNA electrophoresed on denaturing formaldehyde-agarose gels using a DUE-B cDNA probe labeled with [␣-32 P]dCTP by random primer extension. A cDNA encoding the DUE-B protein with a C-terminal His 6 tag was cloned by PCR, inserted into the bacterial expression vector pTRCHis2B, and expressed in E. coli induced by 1.0 mM isopropyl ␤-D-thiogalactopyranoside. The protein was isolated using nickel-nitrilotriacetic acid (Ni-NTA)-agarose (Qiagen Inc.) under nondenaturing conditions. Polyclonal antibody to DUE-B was produced commercially in rabbits (Cocalico Biologicals Inc.) by injection of DUE-B expressed in E. coli. HeLa cells were synchronized by overnight incubation with 1 g/ml aphidicolin, 1 nM okadaic acid, 200 M mimosine, 2 mM hydroxyurea, 10 M purvalanol A, or 100 ng/ml or 400 ng/ml nocodazole. HeLa cells were lysed using NE-PER buffers (Pierce) to yield nuclear and soluble fractions. Western blotting was performed on proteins resolved on 13% SDS-polyacrylamide gels transferred to Immobilon membranes by standard procedures. For phosphate labeling of DUE-B, cells were grown overnight in phosphate-free medium and labeled for 4 h with 30 Ci/ml [␥-32 P]ATP. For expression in insect cells using the MaxBac kit (Invitrogen) DUE-B cDNA was cloned into the pBlueBac4.5 vector and cotransformed with Bac-N-Blue Autographa californica multinucleocapsid nuclear polyhedrosis virus DNA into Sf9 cells according to the manufacturer's directions. Baculovirus-expressed recombinant DUE-B or control Sf9 cell lysates were chromatographed on Ni-NTA resin (Qiagen Inc.) under nondenaturing conditions. Chromatography of recombinant DUE-B (200 -1000 ng in 20 mM Tris-Cl (pH 7.5), 150 mM NaCl, and 1 mM MgCl 2 ) or cell extracts from HeLa cells or Xenopus eggs (2 mg in 50 mM HEPES (pH 7.5), 5 mM MgCl 2 , 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 10% glycerol, 0.01% Tween 20, and 400 mM NaCl) was performed on a 1-m Sephacryl S-200 column. HeLa extracts were prepared according to Vashee et al. (37), except that nuclei were extracted with 500 mM NaCl and clarified by microcentrifugation and 0.22-m filtration. Protein elution was monitored by immunoblotting or enzyme-linked immunosorbent assay (ELISA) using antibodies to DUE-B or the His 6 tag. ATPase assays (25 l) contained 200 mM HEPES (pH 7.5), 0.01% Nonidet P-40, 500 mM NaCl, 10 mM MgCl 2 , 10 mM dithiothreitol, 0.5 mg/ml bovine serum albumin, 200 M ATP, and 2.7 M [␥-32 P]ATP. ATPase activity was monitored by thin layer chromatography on polyethyleneimine-cellulose (38).
Electrophoretic Mobility Shift Assay-A 123-bp probe containing the c-myc DUE/ACS was labeled by PCR in the presence of [␣-32 P]dCTP. The sequence of the probe is GAAGGAATTCATGAGAAGAATGTTTT-TTGTTTTTCATGCCGTGGAATAACACAAAATAAAAAATCCCGAGG-GAATATACATTATATATTAAATATAGATCATTTCAGGGAGCTCGA-GAAACAA. Additional probes prepared with substitution mutations correspond to the sequences described under "One-hybrid Screen." Recombinant DUE-B was purified from baculovirus-infected Sf9 insect cells. Binding reactions (10 mM Tris (pH 7.5), 4% glycerol, 50 mM NaCl, 0.5 mM EDTA, 0.5 mM dithiothreitol, and 1 mM MgCl 2 ) contained 25 fmol (2 ng) of probe, 6 pmol (0.15 g) of DUE-B, and 250 ng of poly- Immunocytochemistry-DUE-B cDNA including C-terminal V5 and His 6 epitope tags was cloned into pcDNA3.1 and transfected into HeLa cells using Lipofectamine 2000 (Invitrogen). 24 h post-transfection, the cells were fixed, permeabilized, and sequentially incubated with monoclonal antibody to either the V5 (Zymed Laboratories Inc.) or His 6 (C terminus-specific; Invitrogen) epitope and then fluorescein isothiocyanate-conjugated goat anti-mouse IgG (Sigma). Cells were counterstained with Hoechst 33258.
ELISA-Polystyrene 96-well plates were coated with Sf9 insect cellexpressed DUE-B (100 ng/well) or 100 l of alternate column chromatography fractions. The primary antibodies were used at dilutions of 1:1000 (anti-DUE-B polyclonal antiserum) and 1:2000 (anti-His 6 antibody), and the horseradish peroxidase-conjugated secondary antibodies were used at 1:10,000 dilutions. Signals were assayed by adding 100 l of o-phenylenediamine/H 2 O 2 substrate (Sigma). Reactions were quenched by the addition of 25 l of 3 N HCl, and absorbance was read at 490 nm. In sperm chromatin binding assays, wells were coated with ϳ23,000 demembranated sperm.
Chromatin Immunoprecipitation (ChIP)-ChIP was carried out as described (39) with the following modifications. Cross-linked chromatin was resuspended in Tris/EDTA buffer and sonicated (Branson Sonifier cell disrupter 200; 50% duty cycle, 10-s pulses, seven pulses with 1-min intervals). The chromatin was digested with 0.1 unit of micrococcal nuclease (Sigma)/100 g of nucleoprotein at 37°C for 5 min to yield fragments Ͻ500 bp. 250 g of nucleoprotein preparation was used for each ChIP assay. The nucleoprotein was diluted with 11ϫ NET (550 mM Tris-Cl (pH 7.4), 1.65 M NaCl, 5.5 mM EDTA, and 5.5% Nonidet P-40) to a final concentration of 1ϫ NET. 15 g of anti-DUE-B polyclonal antiserum or an equivalent amount of normal rabbit antiserum (Upstate Biotechnology) was used for immunoprecipitation. The antibodies were allowed to bind the chromatin complex for 2 h at room temperature. Protein A-agarose beads supplemented with salmon sperm DNA (Upstate Biotechnology) were incubated with the antibody complex for another 2 h at room temperature. Washing of the antibody complex and purification of coprecipitated DNA were carried out as described (40). Real-time PCR was performed as described (15). 1.6% and 0.833% aliquots of immunoprecipitated and input DNAs, respectively, were used for each real-time PCR. Primer sequences used for real-time PCR are shown in Table II. RNA Interference-Cells were plated in 6-well dishes at 3-5 ϫ 10 4 cells/well in MEM ␣ medium and 5 mM glutamine with 10% fetal bovine serum (Invitrogen). siRNA (2 g in 100 l of Opti-MEM) was mixed with 60 l of 1:4 Oligofectamine (Invitrogen)/Opti-MEM. Half of the mixture was added to each of two wells, and cells were incubated for 48 h, re-transfected, and incubated for an additional 48 h before harvesting. The targets for siRNA are the DUE-B sequences AAGCACUG-GUCGAAGAGUGUG (siRNA1) and AAACAGUACGAGAUUCUGUGU (siRNA2). A TT dinucleotide was located at the 3Ј-end of each siRNA.
Replication in Xenopus Egg Extracts-Egg high speed cytosol extracts, membranes, sperm chromatin, and nucleoplasmic extract (NPE) were prepared and used in replication assays as described (41,42). Replication reactions (30 l) contained 150 ng of M13 phage DNA or chromatin from 50,000 Xenopus sperm. Sperm chromatin binding assays were performed with 1 g of recombinant DUE-B or an equal volume (1 l) of control buffer preincubated in 15 l of cytosol for 30 min before the addition of 400,000 demembranated sperm to a final volume of 20 l. The reaction was incubated for 30 min at room temperature and centrifuged twice through a 0.75 M sucrose cushion in a Beckman Microfuge E. The sucrose cushions and wash buffers contained 0.5% Triton X-100. The final pellet was resuspended in Tris/EDTA buffer, and a 20% aliquot was quantitated using SYBR Green fluorometry (Molecular Probes, Inc.) using a sperm chromatin standard curve. The remaining solubilized pellet was assayed by Western blotting using monoclonal antibody to MCM7 (Sigma) or RPA70 (NeoMarkers) and SuperSignal West Pico chemiluminescent substrate (Pierce).
To deplete X. laevis DUE-B from egg extracts, 0.2 volume of preimmune serum-or anti-DUE-B antiserum-coupled protein G-agarose beads was incubated three times with extract for 25 min at 4°C as described (42). HeLa DUE-B was purified by nickel affinity chromatography (Ni-NTA-agarose) from whole cell lysates (prepared with M-PER buffers (Pierce) and protease inhibitor mixture (Sigma)) from a clonal HeLa cell line stably overexpressing the DUE-B cDNA with a His 6 tag.

RESULTS
One-hybrid Screen-Deletion or substitution of the c-myc DUE/ACS region strongly suppresses chromosomal replicator activity (15). To identify proteins that bind to the DUE/ACS, this region was cloned upstream of a HIS3 reporter gene in the S. cerevisiae His Ϫ strain YM4271 (Fig. 1A). This strain was transfected with a Gal4AD-HeLa cDNA fusion protein expression library, and selection was applied for elevated expression of HIS3. One cDNA promoted rapid colony growth under selective conditions in strains containing the wild-type DUE/ACS sequence, but not in otherwise isogenic strains with mutations in the DUE (Fig. 1, A and B). These results imply that the expressed protein interacted specifically with the wild-type DUE reporter. The expressed protein was named DUE-B to designate it as a DNA-unwinding element-binding protein. A reporter strain containing a DUE/ACS bait sequence with mutations in the ACS that prevent S. cerevisiae ORC binding (4) produced a slight further enhancement of cell growth (  (Table III) that had been provisionally annotated as a human histidyl-aminoacyl-tRNA synthetase (HARS2; GenBank TM /EBI accession number NM080820) (43). However, this designation has not been functionally validated. Human DUE-B has notable similarity to proteins of comparable size in Mus musculus (98% similarity) and X. laevis (89% similarity). Proteins of ϳ60 fewer C-terminal amino acids are found in S. cerevisiae (46% similarity) and other organisms (Caenorhabditis elegans, 72% similarity; Schizosaccharomyces pombe, 45% similarity; and E. coli, 41% similarity) (data not shown). The 148-amino acid protein in S. cerevisiae with homology to human DUE-B has been identified as a D-tyrosyl-tRNA Tyr deacylase and is nonessential (43).
DUE-B amino acids 29 -147 are Ͼ90% similar to a domain of unknown function (DUF154) evolutionarily conserved in bacteria, yeast, and mammals (Fig. 1C). The C-terminal 70 amino acids of DUE-B show ϳ50% similarity to nuclear proteins found in humans, flies, and mice. The C-terminal 60-amino acid extension of the human protein not found in the worm and yeast enzymes is conserved in the frog (70% identical) and mouse (93% identical) proteins. DUE-B cDNA spans seven 5Ј-TGGAGTCTTGCGAGGCG exons on human chromosome 20, with the proposed initiator methionine located in exon 2. Northern blot analysis revealed a single species of ϳ1.4-kb DUE-B mRNA ( Fig. 2A), which is sufficient to encode a protein of 209 amino acids.

DUE-B Expression in HeLa Cells and Xenopus
Eggs-Rabbit polyclonal antiserum was raised against the protein expressed in E. coli from the open reading frame of DUE-B fused to a His 6 tag (Fig. 2B, lane 1). This antiserum reacted with a major band migrating at the predicted size of the DUE-B protein (23.4 kDa) (Fig. 2B, lane 2). The 23.4-kDa band was immunoprecipitated with anti-DUE-B antiserum, but not with preimmune serum (data not shown). Cross-reacting proteins of like size have been detected in human WS1 and HCT116 cells and in D. melanogaster SN2 cells. Additionally, X. laevis cell-expressed se-quences predict a 23.2-kDa DUE-B homolog of 207 amino acids that is 81% similar to human DUE-B. The antiserum produced against human DUE-B was able to detect a band at this molecular mass in Xenopus egg cytosol (see below).
DUE-B Is Bound to HeLa Chromatin-To determine the intracellular location of DUE-B, nuclear and soluble lysate fractions were prepared using NE-PER buffers from HeLa cells synchronized by cell cycle inhibitors. DUE-B was detected in both the tightly bound nuclear fraction and in the lysate fraction representing soluble cytoplasmic proteins or proteins loosely associated with nuclei (Fig. 2C). Under these fractionation conditions ϳ 70% of DUE-B was found in the solubilized fraction, whereas the histones were quantitatively retained in the nuclei (data not shown). However, when cells were lysed in  RSB containing 0.5% Nonidet P-40 ϳ70% of DUE-B was in the nuclear pellet (Fig. 2D, lanes 1 and 2). In cells lysed in NE-PER buffers, the proportion of DUE-B in the nuclear and lysate pools did not vary appreciably between cells arrested in S phase by aphidicolin or hydroxyurea or in G 2 /M phase by nocodazole (Fig. 2C). To ascertain whether DUE-B was bound to chromatin, nuclei were isolated from an asynchronous population of HeLa cells and resuspended in RSB (Fig. 2D, lanes 1 and  2) or RSB containing micrococcal nuclease (lanes 3 and 4) or ribonuclease A (lanes 5 and 6). DUE-B was quantitatively released from nuclei upon treatment with micrococcal nuclease. The partial release of DUE-B during incubation in RSB with or without RNase was presumably due to the action of endogenous DNases.
To assess the distribution of DUE-B in intact cells, the protein tagged with V5 and His 6 epitopes was expressed in an asynchronous population of HeLa cells. Immunocytochemical analysis using antibody against either the V5 or His 6 epitope ( Fig. 2E) revealed that DUE-B was localized primarily to the nucleus, whereas controls using an empty expression vector or omitting the primary antibody against the V5 or His 6 epitope revealed no nuclear staining (data not shown). Taken together with the cell lysis results, these data suggest that, in vivo, DUE-B is present primarily in the nucleus in tightly and loosely bound pools, with the loosely bound portion able to leak into the cytoplasmic lysate during fractionation.
Several proteins associated with yeast replicators undergo cyclin-dependent kinase phosphorylation prior to origin firing, although the specific function of these reversible phosphorylations is not known. When HeLa cells were treated with cell cycle inhibitors (Fig. 3A) and labeled with [␥-32 P]ATP, DUE-B was seen to be phosphorylated in cells arrested in early S phase by aphidicolin, but showed lower levels of phosphorylation in cells arrested in G 1 phase by mimosine or in cells arrested in G 2 /M phase by the trisubstituted purine purvalanol under conditions that selectively inhibit Cdk1/2 (44) and an increased level of phosphorylation in cells treated with the protein phosphatase 2A inhibitor okadaic acid (Fig. 3, B and C). These results are consistent with the direct or indirect cyclin-dependent kinase-catalyzed phosphorylation of DUE-B upon entry into S phase and dephosphorylation by protein phosphatase 2A, both of which have been implicated in origin firing (45).
Gel Filtration of Recombinant and Endogenous DUE-B Proteins-The DUE-B protein expressed in bacteria eluted from a Sephacryl S-200 column at a position consistent with its predicted monomeric molecular mass of 23.4 kDa (Fig. 4A). A His 6 -tagged version of DUE-B was also expressed from a baculovirus vector in Sf9 cells and purified to apparent homogeneity using Ni-NTA resin (Fig. 2B, lane 3). Whereas bacterially expressed DUE-B eluted as a monomer, DUE-B purified from Sf9 cells eluted as an apparent dimer of ϳ46 kDa (Fig. 4A). Thus, expression in eukaryotic cells may enhance DUE-B dimerization by post-translational modification; interestingly, the related yeast tRNA deacylases are homodimeric enzymes. To determine whether DUE-B exists in a monomeric or multimeric state in vivo, HeLa nuclear extracts were chromatographed. Endogenous DUE-B eluted at a position corresponding to a size of ϳ46 kDa, with a small percentage of the protein eluting at a higher apparent molecular mass (Ͼ250 kDa) (Fig.  4B). However, when DUE-B expressed in Sf9 cells was mixed with HeLa nuclear extract, roughly three-fourths of the recombinant DUE-B eluted as a dimer, whereas about one-fourth of the exogenous protein eluted in a high molecular mass complex near the column void volume (Fig. 4C). The complex has a size of ϳ250 -350 kDa on Sepharose 4B chromatography and was not eliminated by nuclease treatment or by passing the mixture through a 0.22-m filter. The difference in behavior between the endogenous and exogenous proteins suggests that DUE-B is subject to post-translational modification that affects its ability to interact with other proteins.
In contrast to extracts from an asynchronous population of cultured cells, Xenopus egg extracts offer a highly concentrated pool of proteins poised for rapid and efficient DNA replication. When baculovirus-expressed DUE-B was added to an egg high speed cytosol extract (41), virtually all of the recombinant DUE-B migrated as a Ͼ250-kDa complex (Fig. 4D). The antibody raised against human DUE-B also detected a band in the Xenopus cytosol preparation that chromatographed at ϳ46 kDa, but electrophoresed at ϳ23 kDa. The 23-kDa band was not visible when Xenopus cytosol immunoblots were probed with preimmune serum from the same rabbit (data not shown). Based on the similarity of its chromatographic, electrophoretic, and immunoreactive properties to those of human DUE-B, we refer to the ϳ23-kDa Xenopus protein band as xlDUE-B.
Baculovirus-expressed DUE-B Co-purifies with ATPase Activity-Several potential purine nucleotide-binding sites (motif A and P-loop) were detected in the DUE-B primary sequence. Gel filtration chromatography of purified baculovirus-expressed DUE-B confirmed that an ATPase activity co-eluted with the affinity-purified immunoreactive protein, supporting the view that the ATPase activity is intrinsic or tightly bound to DUE-B (Fig. 5A). This pattern of ATP hydrolysis was not seen with control Sf9 cell lysates (data not shown). Fractions 29 -34 of baculovirus-expressed DUE-B were mixed with a 1000-fold molar excess of ATP, and the hydrolyzed product was quantitated by thin layer chromatography (Fig. 5B). At the lowest DUE-B concentration, the initial kinetics were linear, with a cleavage rate of ϳ0.35 pmol of ATP hydrolyzed per min/pmol of DUE-B. In comparison, under similar conditions of substrate excess in the absence of DNA, purified yeast ORC is reported to hydrolyze ϳ0.04 pmol of ATP/min/pmol of ORC (46).

Selectivity of DUE-B Binding in the Presence of HeLa
Nuclear Extract-An electrophoretic mobility shift assay was used to test the specificity of DUE-B binding to the c-myc DUE/ACS origin fragment in vitro. When baculovirus-expressed DUE-B was added to a radiolabeled c-myc DUE/ACS probe, DUE-B produced a strongly retarded protein-DNA complex (Fig. 6A,  lanes 2-4) that was effectively competed by poly(dI-dC)⅐poly(dI-dC), suggesting that the isolated protein can bind polynucleotides nonspecifically. In the presence of poly(dI-dC)⅐poly(dI-dC), baculovirus-expressed DUE-B (Fig. 6B, lane 1) did not produce a retarded band, whereas the HeLa extract produced a retarded band of the labeled wild-type c-myc DUE/ACS probe, band a (Fig. 6B, lane 3). When recombinant DUE-B was added in the presence of a HeLa nuclear extract, band a was decreased in favor of a new band (band b) that was resistant to the nonspecific competitor (Fig. 6B, lane 2). This result suggests that DUE-B is modified by or forms a complex with proteins in the HeLa nuclear extract to overcome the nonspecific competition for binding to the DUE/ACS probe. The retarded band b was reactive with anti-DUE-B antiserum after Western transfer and was supershifted by anti-His 6 antibody (data not shown).
To test the specificity of this interaction, the wild-type c-myc DUE/ACS probe (DWAW), recombinant DUE-B, HeLa nuclear extract, and poly(dI-dC)⅐poly(dI-dC) were incubated with a 32or 64-fold molar excess of unlabeled wild-type or mutant versions of the DUE/ACS probe. The unlabeled wild-type competitor effectively displaced the DWAW probe (Fig. 6B, lanes 4 and  5), whereas the mutant competitors (DMAW (lanes 6 and 7), DWAM (lanes 8 and 9), and DMAM (lanes 10 and 11)) were less effective at competing for the wild-type DUE/ACS probe. Because DUE-B appeared to bind to the DWAM sequence in the one-hybrid assay, it was expected that the DWAM sequence would compete efficiently in the gel retardation assays. That it did not may reflect differences in the proteins bound in the in vivo and in vitro assays. Nonetheless, the electrophoretic results suggest that recombinant DUE-B can bind the DUE/ACS region of the c-myc origin with apparent selectivity in the presence of other HeLa nuclear protein(s). cross-linked chromatin was isolated from HeLa cells growing exponentially or arrested in the G 1 phase of the cell cycle and immunoprecipitated with anti-DUE-B antibody. The abundance of c-myc sequences in the immunoprecipitated chromatin was quantitated by real-time PCR (Fig. 7). In asynchronously growing cells and in cells arrested in G 1 phase by mimosine, the DUE-B immunoprecipitate showed a significant enrichment of sequences near the c-myc replicator DUE relative to the input chromatin (Fig. 7, A and B) or sequences immunoprecipitated with preimmune serum. The enrichment of DUE sequences in the DUE-B immunoprecipitates was quantitatively comparable with that seen for MCM4/PRKDC and TOP1 origin sequences in ORC and MCM chromatin immunoprecipitates (47)(48)(49). Similar enrichments were observed for ORC1, ORC2, MCM3, MCM7, and Cdc6 proteins at the c-myc replicator. 2 The temporal and spatial similarities in the binding of DUE-B and the MCM complex, as exemplified by MCM4, are especially striking (Fig. 7, C and D), possibly reflecting a functional rela-tionship between the DUE-B and the helicase.

DUE-B Binds Near the c-myc DUE in Vivo-Formaldehyde
DUE-B Down-regulation Inhibits Cell Cycle Progress-In S. cerevisiae, disruption of the DUE-B ortholog is not lethal (43). To test whether DUE-B function is essential for cell cycle progression, HeLa cells were transfected twice with siRNA directed against DUE-B mRNA (siRNA1), once at t ϭ 0 h and again at t ϭ 48 h, and examined at t ϭ 96 h without intervening trypsinization (see "Experimental Procedures"). DUE-B siRNA1 reduced DUE-B levels by Ͼ95%, but not the levels of the unrelated protein Ku80 (Fig. 8A). A scrambled siRNA of similar composition did not reduce DUE-B or Ku80 levels. In four independent experiments, compared with cultures transfected with scrambled siRNA, cultures transfected with DUE-B siRNA1 showed a 3.5-8.5-fold increase in cells with sub-G 1 DNA content (Fig. 8B), a decrease in population doubling (2.5-4-fold), and an increase in cell death (31.5% versus 9.6%) as measured by trypan blue exclusion. Transfection of another RNA, DUE-B siRNA2, decreased DUE-B levels more slowly and less completely. DUE-B siRNA2 induced a smaller increase in cell death (12.5%) and a smaller increase in the population of sub-G 1 cells (3-5-fold) (data not shown).
Cells that had been treated with siRNA1 were trypsinized, synchronized with nocodazole (Fig. 8C), and released from the G 2 /M phase arrest for 12 h. Relative to cultures treated with scrambled siRNA, cultures treated with DUE-B siRNA1 showed an increase in the population of G 1 phase cells (25% versus 18%, control) and a decrease in the fraction of S phase cells (38% versus 47%, control) (Fig. 8D). This effect was even more apparent when cells were synchronized in G 1 phase with mimosine (Fig. 8E) and released (Fig. 8F). Here, DUE-B siRNA caused a 50% decrease in the S phase cell population (31% versus 64%). Taken together, these results suggest that inhibition of DUE-B expression delays entry into S phase and promotes cell death.

DUE-B Controls Replication in Vitro-
To determine whether DUE-B has a direct effect on replication, Xenopus egg extracts were used. These extracts assemble sperm chromatin into nuclei that undergo a complete round of semiconservative replication, dependent on the ordered assembly of pre-RC components, Cdk2, Cdc7, Cdc45, RPA, and DNA polymerase (41,42,50). Xenopus egg high speed cytosol mimics the in vivo conditions for pre-RC formation, and the subsequent addition of an egg membrane fraction or NPE concentrates factors that promote activation of the pre-RC and replication (41). As shown by gel filtration chromatography (Fig. 4), exogenous DUE-B (but not endogenous xlDUE-B) was modified in egg high speed cytosol to form a high molecular mass complex. In an ELISA, baculovirus-expressed DUE-B bound saturably to demembranated sperm chromatin, and the presence of egg extract increased the amount of DUE-B required for 50% sperm chromatin binding by ϳ10-fold, from Ͻ0.02 to 0.2 g (Fig. 9A). In the absence of exogenous DUE-B, binding of endogenous xl-DUE-B was not detected when Xenopus cytosol was incubated with sperm chromatin (data not shown).
When DUE-B expressed in bacteria or baculovirus-infected Sf9 cells was added to the cytosol at roughly the same concentration as endogenous DUE-B (10 ng/l) simultaneously with sperm chromatin and membranes, the addition of recombinant DUE-B had no apparent effect on DNA replication (Fig. 9, B and C). However, when recombinant DUE-B was allowed to incubate for 30 min with cytosol and sperm chromatin before the addition of membranes, there was a dramatic inhibition of chromatin replication by Sf9 cell-expressed DUE-B and a lesser inhibition of replication by bacterially expressed DUE-B (Fig.  9, B and D). In immunoblot assays, both the recombinant and endogenous forms of DUE-B were found to be stable in Xenopus cytosol and after the addition of membranes (data not shown). Baculovirus-expressed DUE-B preincubated with cytosol and sperm chromatin was able to inhibit DNA replication in a dose-dependent manner (Fig. 9E). By contrast, preincubation of baculovirus-expressed DUE-B with chromatin did not inhibit the formation of sperm pseudo-nuclei (data not shown), nor did it inhibit the replication of single-stranded M13 DNA (Fig. 9F), which occurs independently of pre-RC formation (51). These observations imply that the inhibitory effect of recombinant DUE-B on replication is not due to a gross perturbation of the egg extract system.
That single-stranded DNA replicated normally in the presence of Sf9 cell-expressed recombinant DUE-B suggested that DUE-B is not a general DNA polymerase inhibitor. To determine whether the addition of recombinant DUE-B inhibits replication before or after pre-RC formation, the soluble egg cytosol/NPE system was used in a sperm chromatin binding assay. MCM7 and RPA binding was assessed as markers of pre-RC and initiation complex formation, respectively (42,50). DUE-B was preincubated with egg cytosol and sperm chromatin prior to the addition of NPE. Sperm chromatin was separated from the cytosol by sucrose gradient centrifugation in the presence of Triton X-100. The results show a dramatic reduction of RPA binding, but no significant decrease in MCM7 loading on sperm chromatin (Fig. 9G). To quantitate this effect, reactions were performed in triplicate, and the results were normalized against the amount of sperm chromatin pelleted (Fig. 9H). DUE-B did not inhibit MCM7 loading, but decreased RPA loading and replication by ϳ80%. The stability of MCM7 binding despite the inhibition of RPA binding and replication argues for the selectivity of DUE-B action on replication initiation in the Xenopus egg extract system.
The inhibitory effect of baculovirus-expressed recombinant DUE-B on replication was counterintuitive; however, because the recombinant and endogenous DUE-B proteins behaved differently upon gel filtration, we decided also to test for a role of endogenous DUE-B in replication. Egg extracts were immunodepleted of xlDUE-B by Ͼ95% (Fig. 10A), and sperm chromatin replication was measured. In these immunodepleted extracts, replication was inhibited by Ͼ85% (Fig. 10, B and C).
To determine whether replication could be rescued by DUE-B that had not been expressed in insect cells, HeLa cells were constructed to express His 6 -tagged recombinant DUE-B. The co-isolation of the recombinant and endogenous DUE-B pro- teins from HeLa cells by nickel affinity chromatography indicates that these proteins fold normally to heterodimerize in vivo. The addition of the HeLa cell-expressed recombinant DUE-B fraction quantitatively restored replication when added at a 4-fold molar excess over the level of endogenous xlDUE-B in undepleted extracts (Fig. 10, B and C), whereas HeLa cell lysate-or Sf9 cell-expressed recombinant DUE-B similarly eluted from Ni-NTA did not (data not shown). The observations that removal of DUE-B by immunoprecipitation inhibited replication and that replacement of DUE-B purified by another method, i.e. affinity chromatography, restored replication strongly indicate that DUE-B is involved in the initiation of DNA replication. DISCUSSION A previously uncharacterized protein, which we have termed DUE-B, has been identified based on its selective affinity for the DUE of the human c-myc replicator in a yeast one-hybrid screen. In addition to its interaction with an essential element of the c-myc replicator, DUE-B coordinately inhibited sperm chromatin replication and RPA binding in the Xenopus egg extract system, and down-regulation of DUE-B slowed entry into S phase and decreased the proliferation of HeLa cells in culture. These observations suggest that DUE-B may play a role in modulating DNA replication in vivo.
Sequencing and translation of the DUE-B cDNA predicted a 23.4-kDa protein of 209 amino acids. Northern blot analysis revealed a 1.4-kb mRNA, sufficient to encode a protein of this size, and Western analysis using antibody raised against recombinant DUE-B detected a protein of the predicted size in HeLa cells. Over its N-terminal 148 amino acids, DUE-B shows strong similarity (Ͼ45% identical and Ͼ65% homologous) to yeast and bacterial homodimeric D-tyrosyl-tRNA Tyr deacylases. Although we have not tested for deacylase activity in DUE-B, the S. cerevisiae enzyme is nonessential and cytoplasmic, whereas a significant portion of HeLa DUE-B is bound to chromatin, and its depletion is associated with decreased cell proliferation. Thus, the yeast and human enzymes appear to have some dissimilar properties. However, the ATPase activity that co-purifies with baculovirus-expressed human DUE-B may be related to the hydrolase activity of the yeast enzymes based on the observation that the seven-motif arrangement of invariant amino acids essential for catalysis in the PPM phos- Immunocytochemical analysis showed that epitope-tagged DUE-B expressed in HeLa cells localized to the nucleus in the absence of an added nuclear localization sequence, consistent with the observation that a fraction of endogenous DUE-B was recovered from isolated HeLa nuclei and released from the nuclei by high salt extraction (data not shown) or nuclease digestion, similar to the reported distribution of other pre-RC proteins (55). DUE-B appears to be present at roughly constant levels throughout the cell cycle, and agents that inhibit replication (e.g. aphidicolin and hydroxyurea) did not affect the levels or cellular distribution of DUE-B. Hence, DUE-B does not appear to redistribute between cellular compartments in response to DNA damage. The observation that DUE-B showed preferential binding to the wild-type DUE/ACS sequence in vivo and in vitro in the presence of HeLa nuclear proteins suggests that DUE-B binding to the DUE/ACS may be indirect or involve other proteins that influence the structure of DNA. Recently, Kinoshita and Johnson (56) reported that MCM4 binds near the DUE region of the c-myc replicator preferentially during G 1 phase. Our data on the binding of MCM4 are consistent with this observation and allow speculation that DUE-B and the MCM helicase recognize or modulate the structure of the DUE. The idea that the binding of DUE-B to the DUE is affected by the proximal binding of other proteins is consistent with the enhanced expression of the one-hybrid reporter in yeast when the ARS consensus elements flanking the c-myc DUE were mutated. Because the Gal4AD-DUE-B protein activated the HIS3 reporter as well or better when bound to the DWAM bait compared with the wild-type sequence in the one-hybrid assay, it was expected that the DWAM sequence would compete effi-ciently in the gel retardation assays. However, comparison of these two assays is difficult because the reporter system is complex and more directly measures transcription than binding. On the other hand, the HeLa proteins that interact with DUE-B do appear to impart sequence-selective binding in the context of a non-chromatinized template. The identity of the DUE-B-interacting protein(s) is currently under investigation.
Upon gel exclusion chromatography, DUE-B expressed in bacteria eluted at a position consistent with its predicted monomeric molecular mass, whereas DUE-B expressed from a baculovirus vector eluted as a dimer, suggesting that expression in the eukaryotic cells allows post-translational modifications that affect DUE-B function. Chromatography of HeLa extracts revealed the ϳ46-kDa dimeric form of endogenous DUE-B, along with a minor amount of DUE-B protein eluting at a higher molecular mass (Ͼ250 kDa). Roughly one-fourth of the recombinant DUE-B added to nuclear extract from an asynchronous HeLa culture was also converted to a high molecular mass form. By contrast, virtually all of the recombinant DUE-B added to an equivalent amount of Xenopus egg cytosol was found to elute as a high molecular mass complex, whereas the putative Xenopus DUE-B homolog eluted at the ϳ46-kDa dimer position. Correlating with the change in association state, the affinity of exogenous DUE-B for sperm chromatin binding decreased by 10-fold in the presence of Xenopus egg cytosol. Thus, the high molecular mass complexes containing DUE-B may bind more weakly to chromatin than the dimeric form. The absence of heteromeric complexes in the recombinant and endogenous DUE-B proteins strongly suggests that one or both of these molecules is subject to post-translational modification that induces or precludes high molecular mass complex formation.
In Xenopus egg extracts, incubation with baculovirus-expressed DUE-B before the addition of membranes inhibited sperm chromatin replication in a dose-dependent manner. At the approximate concentration of endogenous xlDUE-B or about one-fourth the molar concentration of MCM proteins in these extracts (41), exogenous DUE-B inhibited sperm chromatin replication by Ͼ50%. This inhibition was selective in that DUE-B did not inhibit the formation of pseudo-nuclei, the replication of single-stranded DNA, or the loading of MCM7. In the soluble cytosol/NPE system, the inhibition of sperm chromatin replication was quantitatively correlated with the decreased loading of RPA. This does not appear to be the result of a direct interaction between DUE-B and RPA inasmuch as DUE-B and RPA did not co-immunoprecipitate or interact in pull-down assays, 3 and recombinant DUE-B did not inhibit the replication of single-stranded DNA. Therefore, to the extent that MCM7 loading reflects pre-RC formation, DUE-B inhibits sperm chromatin replication at a step following pre-RC formation. Endogenous DUE-B does not form high molecular mass complexes and is permissive or stimulatory for DNA replication, whereas exogenous DUE-B expressed from baculovirus forms high molecular mass complexes and inhibits replication. It is plausible therefore that variations in covalent modification are responsible for both of these differences. Preincubation with the extract may allow baculovirus-expressed DUE-B to act as a dominant-negative mutant by competing for or sequestering essential replication factors yet to be identified.
We suggest that one or more components of the high molecular mass complex formed in Xenopus extracts may be homologous to proteins present in limiting amounts in the asynchronous HeLa nuclear extract and that the endogenous and exogenous DUE-B proteins differ in post-translational modification states such that only exogenous DUE-B is able to form the high molecular mass complex, suppress RPA loading, and inhibit replication. A two-state model has also been proposed to explain the difference in stability of endogenous and exogenous geminin in Xenopus egg extracts (57). Together with the observation that ablation of DUE-B expression is associated with reduced cell proliferation, DUE-B may play both positive and negative roles in replication initiation. Several origin-binding proteins are substrates for cyclin-dependent kinase-dependent phosphorylation, including Cdc6, MCM4, ORC2, and ORC6 (58 -61). The increased phosphorylation of DUE-B in early S phase may reflect the transition between these states.
We do not know whether the high molecular mass complexes formed by endogenous DUE-B in asynchronous HeLa extracts are cell cycle-regulated or are the same as those formed by recombinant DUE-B. However, the quantitative difference in the amount of high molecular mass complexes in Xenopus and HeLa extracts points to the cell cycle-dependent regulation of one or more proteins that stoichiometrically interact with DUE-B. It is proposed that DUE-B is present in vivo in a modified dimeric form that distinguishes it from exogenously expressed DUE-B and that, prior to the initiation of replication, DUE-B is further modified to supply a factor necessary for RPA loading at the pre-RC. In this model, baculovirus-expressed DUE-B may sequester this factor in a high molecular mass complex, inhibiting RPA deposition.
Further evidence of physiological differences between Sf9 cell-expressed and endogenous or HeLa cell-expressed DUE-B come from in vitro replication experiments in which depletion of endogenous xlDUE-B or the addition of baculovirus-expressed recombinant DUE-B inhibited replication, whereas HeLa cell-expressed recombinant DUE-B restored replication to immunodepleted extracts when added in 4-fold excess over in vivo levels. These results indicate that endogenous DUE-B is important for DNA replication. Among other possibilities, the need for a molar excess of this fraction could mean that only a portion of DUE-B isolated from asynchronous cells is correctly modified (e.g. phosphorylated) to stimulate replication or that only endogenous HeLa DUE-B that heterodimerizes with His 6  replication and cell cycle progression, we note preliminary results indicating that DUE-B mRNA levels were elevated by 40 -300% in 15 of 20 ovarian and colon tumors tested relative to neighboring normal tissue. 3 Work is under way to address several aspects of a model for DUE-B in mammalian replication.