INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Initiation of DNA replication in Escherichia coli, mammalian viruses, and the budding yeast Saccharomyces cerevisiae (23, 67) is controlled primarily by trans-acting initiator proteins that interact with cis-acting DNA sequence elements (the replicators) (46). In these simple replicons, the cis-acting element consists of an essential core sequence, containing initiator protein binding sites and easily unwound sequences (DNA unwinding elements [DUEs]), and auxiliary sequences which enhance the efficiency of replication initiation (23, 67). In S. cerevisiae, the initiation of replication requires an autonomously replicating sequence (ARS) element in cis and occurs within a region flanked by a DUE and the binding sites for the origin recognition complex (ORC) and other initiator proteins (10). The initiation site for leading-strand synthesis appears to be restricted to a single nucleotide in a chromosomal ARS element (11). Replication initiation in fission yeast is also controlled by the sequence-specific recognition of a cis-acting replicator by ORC and associated initiator proteins, but the fission yeast replicators are larger than those of its budding yeast counterpart (18, 19, 50, 66, 68, 69). Initiation activity of yeast ARS elements in chromosomal DNA depends not only on specific DNA sequences in the ARS but also on chromatin structure and chromosomal position (13, 34, 36, 51, 63, 67, 70, 79, 88).

Extensive mapping of replication start sites in mammalian chromosomes has revealed that replication at most but not all loci begins at a few high-frequency start sites contained within a broad zone of initiation (4, 37, 59, 62, 82, 83; see also studies reviewed in references 12, 24, and 40). One of the most thoroughly mapped high-frequency initiation regions (IRs) in mammalian chromosomes is the region downstream from the dihydrofolate reductase (DHFR) gene in Chinese hamster ovary (CHO) cells (Fig. 1A). This region contains a 55-kb zone of delocalized origin activity containing three preferred start sites: origin beta (ori-), centered approximately 17 kb downstream from the DHFR gene; ori-' just downstream from ori-; and ori-, located 23 kb further downstream (5, 14, 42, 43, 44, 52, 55, 57, 71, 85, 86, 89).

View larger version (38K): [in this window] [in a new window]   FIG. 1.   (A) Features of the endogenous DHFR ori- IR. Preferred start sites of DNA replicationthe ori-, ori-', and ori- sites; a 55-kb initiation zone (shaded area) between the DHFR- and 2BE2121-coding sequences; and matrix-attached regions (MARs, stippled boxes)are indicated (5, 14, 42, 43, 44, 52, 55, 57, 71, 85, 86, 89). The 5.8-kb fragment of the DHFR ori- region (pMCD), extending from the BamHI to the KpnI recognition sequences, is indicated. Arrows denote the positions of the primers used in the competitive PCR assay for origin function and correspond to primer pairs used in previous DHFR ori- mapping studies (71). (B) Strategy for quantitating initiation occurring in exogenous DHFR ori- fragments in ectopic chromosomal locations. A 5.8-kb wild-type or mutated DHFR ori- fragment was coelectroporated with a neomycin resistance gene into the DR12 cell line, a CHO derivative containing a 150-kb deletion encompassing the entire DHFR locus (47). After selection with G418, total DNA was isolated, heat denatured, and size fractionated on a 5 to 30% linear neutral sucrose gradient. The fraction containing single-stranded DNA with a length of 1 to 2 kb, representing nascent DNA, was isolated, and the abundance of ori- target sequences contained in the fraction was quantitated by competitive PCR.

Despite the progress in mapping initiation sites, the specific genetic elements necessary to direct initiation of mammalian DNA replication remain poorly understood. The existence of preferred IRs raises the question of whether specific DNA sequences within or neighboring an IR may direct initiation of replication at the IR or even in a broad zone more distant from the preferred IR. Several lines of evidence are consistent with the notion that sequences neighboring a preferred IR may constitute a mammalian replicator element. In the human lamin B2 IR, DNase I footprinting has revealed that a specific sequence is protected in a cell cycle-dependent manner, suggesting that it may be a binding site for an initiator protein complex (1, 29, 39). Moreover, replication at the lamin B2 origin was shown to initiate within a 3-bp sequence that overlaps the footprint region (2). A site hypersensitive to micrococcal nuclease has been mapped in the DHFR ori- IR locus (72), and there is preliminary evidence that hamster ORC2, a subunit of the ORC complex, binds within the DHFR ori- IR (cited in reference 12). Recent genetic analysis of the human -globin (4) and the human c-myc (62) IRs at an ectopic chromosomal locus demonstrated that a defined sequence of 2 to 8 kb can be sufficient to direct initiation in the ectopic locus. Furthermore, the putative c-myc replicator even induced new start sites in the flanking chromosomal DNA (62). Taken together, these studies suggest that initiation of chromosomal DNA replication in mammalian cells may be directed by specific cis-acting DNA sequence elements that constitute a replicator similar to those characterized in yeasts.

The complex organization of the DHFR initiation zone raises the question of whether it differs structurally from the putative replicators containing the c-myc, lamin B2, and -globin IRs. For example, the cluster of preferred IRs may represent a novel replicator organization with multiple elements dispersed throughout the 55-kb initiation zone that are interdependent and cannot function independently as replicators. Alternatively, the three preferred start sites, ori-, ori-', and ori-, within the delocalized initiation zone could represent separable but redundant genetic elements that ensure the faithful replication of this locus (72). In this latter case, the regions surrounding each of the three start sites might serve as individual replicators able to direct initiation when placed at an ectopic chromosomal location.

To address the question of whether a defined DNA sequence encompassing DHFR ori- can direct initiation of DNA replication in chromosomal DNA, we have generated stably transfected cells containing the exogenous ori- IR in random ectopic chromosomal locations and measured ori- initiation activity by using a competitive PCR-based nascent strand abundance assay. The results presented here demonstrate that a 5.8-kb sequence containing the ori- IR is sufficient to direct initiation in multiple ectopic chromosomal sites and that specific DNA sequences within and flanking the IR are essential for efficient initiation activity at the DHFR ori- region.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell culture. Diploid Chinese hamster ovary K1 (CHOK1) cells and DR12 cellsCHOK1 derivatives containing 150-kb deletions of the entire DHFR locus (47)were grown in Ham's F12 medium supplemented with 10% fetal calf serum (Life Technologies, Rockville, Md.) at 37°C and 4% CO₂.

Plasmid constructs. The plasmid pMCD, containing the 5.8-kb BamHI-KpnI fragment, nucleotides (nts) 1 through 5793 (GenBank accession no. Y09885) of the DHFR ori- locus in the vector pUC19 (90), was the kind gift of N. H. Heintz (16). Deletion mutant pAKO was generated by digestion of pMCD with ApaI (nt position 1127), removal of overhanging ends using T4 DNA polymerase, and religation. Mutant pMCDAT was created by partial digestion of pMCD with SphI (nt position 2163), followed by complete digestion with EcoRV (nt position 2507), releasing a 344-bp fragment; the other fragment was then blunt ended and religated. The mutant pMCDDNR was created by partial digestion of pMCD with XbaI (nt position 4454) followed by complete digestion with NheI (nt position 4219), releasing a 235-bp fragment; the remaining fragment was blunt ended and religated. Mutant pMCDTR was created by using PCR and mutagenic primers (5'GACAAAAACAATCGATAAATAAG and 5'CTTATTTATCGATTGTTTTTGTC) to insert a ClaI restriction site at nt position 686, relative to the BamHI site. The resulting plasmid was partially digested with PvuII (nt position 949), followed by complete digestion with ClaI, releasing a 263-bp fragment; the remaining fragment was blunt ended and religated. The pSV2neo plasmid contains a full-length neomycin resistance gene (Clontech Laboratories, Inc., Palo Alto, Calif.). Plasmid K126 contains the 3-kb BamHI-XbaI fragment of the DHFR ori- region in the vector pBluescript. Plasmid Mut8-2 contains the SacI/AccI fragment of the DHFR ori- region in the vector pBluescript.

Transfection. The BamHI-KpnI ori- fragment cloned in pUC19 was linearized with AatII to generate the 5.8-kb DHFR insert flanked by 1,271 and 1,401 bp of vector DNA. Four micrograms of the digest mixed in a 3:1 molar ratio with PvuI-linearized pSV2neo DNA was electroporated into 5 × 10⁶ DR12 cells using a BioRad Gene Pulser at 360 V and 600 µF. Cells were plated in 75-cm² tissue culture flasks. After 24 h, the medium was replaced with Ham's F12 supplemented with 10% fetal calf serum and 0.5 mg of G418 (Life Technologies) per ml. After 4 weeks of growth under selection, exponentially proliferating cells from four 150-cm² flasks, grown to 70% confluency, were harvested, and total DNA was isolated and analyzed as described below. Under these transfection conditions, the average number of copies of exogenous DHFR ori- integrated per cell in pools of cells was highly reproducible (see below). Single cell clones were isolated through serial dilution into a 96-well plate and expanded under drug selection for 14 weeks to four 150-cm² flasks. Isolation of total genomic DNA and its analysis were carried out as described below for pools of transfected cells.

Quantitation of stably transfected DNA. The integration of ectopic DHFR ori- fragments into DR12 genomic DNA was monitored by PCR analysis of DNA from transfected cells. PCR amplification was carried out with primer set 2 (Table 1) and with 5 ng of total genomic DNA, isolated from transfected DR12 cells or from CHOK1 cells, in a Perkin-Elmer 4800 thermal cycler (35 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s). Amplification products were resolved by 7% polyacrylamide gel electrophoresis (PAGE), stained with ethidium bromide, and quantified by densitometry (IPLab Gel software; Signal Analytics Corp., Vienna, Va.). The ratio of the exogenous DHFR ori- products to the endogenous CHOK1 ori- products, termed the integration ratio, was consistently close to 1 after 4 weeks of drug selection. Thus, although the integration ratio is not used to calculate the initiation activity, the copy number of integrated pMCD fragments in DR12 cells, on average, mimics the copy number of the endogenous DHFR ori- in CHOK1 cells. In some experiments, the PCR analysis was carried out in parallel with primer pairs 2 and 3; the resulting ratios did not differ significantly.

DNA isolation and gradient centrifugation. Total genomic DNA was isolated from harvested cells according to the manufacturer's instructions (Nucleon II; Scotlabs). Five-hundred micrograms of resuspended total genomic DNA in 1 ml of TE (10 mM Tris-HCl [pH 8.0]-1 mM EDTA) was heat denatured at 100°C for 10 min, followed by a 10-min incubation in ice-water. DNA was loaded onto a 5 to 30% linear neutral sucrose gradient and centrifuged at 26,000 rpm (Sorvall AH-629 rotor) for 17 h at 20°C. Fractions of 1 ml each were collected, and the fraction enriched in nascent single-stranded DNA (ca. 1 to 2 kb in length) was dialyzed against TE and used as a template for PCR amplification. As size markers, 50 µg of pMUT8-2 digested with KpnI and NdeI to produce fragment sizes of 921 and 2,972 bp and 50 µg of pK126 digested with StuI and BgIII to produce fragment sizes of 1,636 and 4,338 bp were loaded on a sucrose gradient and run concurrently with each set of experiments. A 50-µl aliquot of each marker fraction was subjected to electrophoresis on a 1% agarose gel at 200 mA for 2 h, and DNA was visualized by ethidium bromide staining. For control PCR, total genomic DNA was sheared to fragments 1 to 2 kb in size by sonication for 15 s at 25% power (250/450 Sonifier; Branson Ultrasonics Corp., Danbury, Conn.), heat denatured, and fractionated on a 5 to 30% linear neutral sucrose gradient as described above.

Construction of PCR competitive templates. Competitors were constructed essentially as previously described (71). Briefly, PCR amplification was performed with pMCD template using the standard primer sets 8, 2, 6, and 3 and their corresponding competitor primer sets, as denoted by the suffix COM, which contain 20-bp insertions (Table 1). These PCR products were used as templates for another round of PCR amplification with the standard primer sets, resulting in PCR amplification products which were identical to their pMCD target sequence except for the 20-bp insertion. The mutated PCR amplification products were cloned into pMCD to create competitive templates for the PCR-based nascent strand abundance assay.

PCR-based nascent strand abundance assay. The competitive PCR nascent strand abundance assay was performed using primer sets 8, 2, 6, and 3 (31, 53, 71; Table 1) essentially as previously described. Each PCR mixture included the corresponding competitor plasmid DNA, which had been precalibrated against 20 ng of total genomic DNA from asynchronously growing CHOK1 cells. Assuming 3 × 10⁹ bp per haploid genome, 20 ng of genomic DNA would correspond to 6,000 molecules of competitor (31). Amplification reactions with each primer set contained increasing amounts of the size-fractionated nascent DNA and the precalibrated amount of the corresponding competitor DNA. PCR was carried out in a Perkin-Elmer 4800 thermal cycler (50 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s). Similar results were obtained using 35 cycles with the same program (data not shown), consistent with the expectation that competitive PCR should be independent of cycle number (31). Amplification products were resolved by 7% PAGE, stained with ethidium bromide, and quantified by densitometry (IPLab Gel software; Signal Analytics Corp.). The intensity of the signal was linearly proportional to the amount of stained DNA under the conditions used.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

A DHFR ori- fragment functions efficiently as an origin in ectopic chromosomal locations. Given that DNA fragments containing the -globin (4) and the c-myc (62) IRs supported efficient DNA replication at ectopic chromosomal locations, we asked whether a small DNA fragment containing DHFR ori- and flanking DNA but lacking the downstream ori-' and ori- regions would direct initiation of DNA replication when located at ectopic chromosomal sites. To address this question, we chose the experimental strategy diagramed in Fig. 1B. A 5.8-kb fragment containing the ori- IR (pMCD [16]) but not the ori-', ori-, or matrix attachment regions or the DHFR coding sequences was cotransfected by electroporation with an unlinked neomycin resistance marker into DR12 cells, a CHOK1-derived line lacking both DHFR loci (47). We chose an unlinked resistance marker to ensure that any initiation of DNA replication occurring within the ori- fragment would not be influenced by transcription of the neomycin gene in the same construct. To facilitate the stable integration of the DNA fragments into the chromosomal DNA, plasmids containing the ori- DNA fragment were first linearized in the vector portion by restriction endonuclease digestion. Under these conditions, integration of exogenous DNA fragments has been reported to be random and without significant loss of DNA sequences at the ends of the fragment (35, 77).

View larger version (56K): [in this window] [in a new window]   FIG. 2.   Initiation of DNA replication at the endogenous DHFR ori- site in asynchronous CHOK1 cells. (A) PCR amplifications were performed with each of the four primer pairs and with size-fractionated nascent DNA template from asynchronous CHOK1 cells in the presence of a precalibrated amount of the corresponding competitor DNA. Amplification products were analyzed by PAGE and ethidium bromide staining. Control lanes: C, competitor template only; G, nascent genomic DNA template only; , no template. Numbers above each lane represent the volume in microliters of nascent DNA added to the PCR mixture. Amplification reactions with pp2 used a 1:10 dilution of the nascent DNA. (B) PCRs were performed and analyzed as in panel A, except that the template was a 1:10 dilution of sheared, denatured total genomic CHOK1 DNA 1 to 2 kb in length. (C) Amplification products generated with either nascent DNA from asynchronous CHOK1 cells (black bars) or sheared total genomic DNA from asynchronous CHOK1 cells (white bars) were quantitatively evaluated for five independent experiments with each type of template. As a measure of initiation activity, the abundance of each target sequence in nascent genomic DNA was normalized to the abundance of pp3 target sequences in the corresponding experiment, which was set equal to 1 (see example in Table 2), and the average of five experiments is shown. Bars indicate the standard error of the mean (SEM).

View this table: [in this window] [in a new window]   TABLE 2.   Abundance of DHFR ori- target sequences in nascent DNAa

View larger version (60K): [in this window] [in a new window]   FIG. 3.   Initiation of DNA replication in the exogenous 5.8-kb ori- fragment in DR12 cells. (A) The integrated exogenous DHFR ori- fragment in 5 ng of DNA from uncloned pools of DR12 cells and the endogenous ori- in 5 ng of CHOK1 DNA were amplified by PCR. The ratio of the amplification products in DR12 relative to those of endogenous ori- in CHOK1 is indicated, and this ratio suggests that the copy number of exogenous ori- fragments present in the pool of transfectants mimics that of the endogenous locus in CHOK1. (B) The structure of the integrated DHFR ori- fragments in DR12 pools was determined by PCR amplification of either pMCD plasmid DNA (lanes P) or large single-stranded genomic DNA isolated from uncloned drug-resistant pools of pMCD-transfected DR12 cells (lanes G). Amplification reactions were performed with a panel of six primer sets spanning the 5.8-kb DHFR ori- fragment and the flanking vector sequences (bottom). The resulting overlapping PCR amplification products, labeled 1 through 6, were analyzed by PAGE and ethidium bromide staining. M indicates a DNA size marker. The sizes of the resulting PCR amplification products are indicated. (C) PCR amplifications were performed with each of the four primer pairs and size-fractionated nascent DNA from asynchronous pMCD-transfected DR12 cells in the presence of a precalibrated amount of the corresponding competitor DNA. Amplification products were analyzed by PAGE and ethidium bromide staining. Control lanes: C, competitor template only; G, nascent genomic DNA template only; , no template. Numbers above each lane represent the volume in microliters of nascent DNA added to the PCR mixture. Amplification reactions for pp2 and pp6 used a 1:10 dilution of the nascent DNA. (D) Amplification products generated with either nascent DNA from a pool of asynchronous pMCD-transfected DR12 cells (black bars) or sheared total genomic DNA from asynchronous pMCD-transfected DR12 cells (white bars) were quantitatively evaluated for seven independent transfection experiments. As a measure of initiation activity, the abundance of each target sequence in nascent genomic DNA was normalized to the abundance of pp3 target sequences in the corresponding experiment, which was set equal to 1 (see example in Table 2), and the average of the seven experiments is shown. Bars indicate the SEM.

Ori- functions in multiple ectopic locations, but exhibits some position effects. Since chromosomal context is an important determinant of origin function in mammalian cells (3, 17, 27, 30, 33, 54, 56), the function of the exogenous ori- IR may be sensitive to position effects that were not detected in the uncloned cell pools. If initiation activity of the 5.8-kb fragment is affected by chromosomal context, one might expect to find that initiation activity of the ectopic ori- region would vary among clonal cell lines with different ori- integration sites. To test for variability in initiation activity, six individual cell lines were cloned from a resistant-cell pool. To estimate the amount of integrated exogenous ori- fragment in each clone, total genomic DNA isolated from each cell line was used as a template for PCR amplification with pp2. For comparison with the endogenous ori- region, an equal amount of total genomic DNA isolated from CHOK1 cells was used as a template in a parallel amplification. The amplification of the exogenous ori- DNA fragments differed slightly from clone to clone. Clones 1, 3, and 5 contained about the same copy number found for the endogenous ori- region in CHOK1 DNA (Table 3). Clones 2, 4, and 6 had 20 to 30% less ori- fragment, possibly suggesting some loss of ori- sequences during expansion of the cloned cell lines.

View this table: [in this window] [in a new window]   TABLE 3.   Abundance of DHFR ori- target DNA sequences in nascent DNAa

Specific cis-acting sequences are required for efficient DHFR ori- initiation activity. To assess whether specific DNA sequences within the 5.8-kb fragment of the ori- region were required for efficient initiation at the preferred IR, several ori- deletion mutants were constructed (Fig. 4A). The regions chosen for deletion contain DNA sequences that could be functionally important in mammalian origins based on their resemblance to DNA elements identified in other origins as discussed below. To permit comparison of the initiation activity of mutated fragments with that of the parental pMCD ori- fragment, mutations that would delete primer sites for PCR amplification were avoided.

View larger version (62K): [in this window] [in a new window]   FIG. 4.   Initiation of DNA replication in wild-type and mutant exogenous ori- DNA fragments. (A) Unusual DNA sequences in the 5.8-kb ori- fragment and the location of the deletion mutants are indicated on the restriction map of the region. Positions of AluI repetitive elements (white boxes) and AT-rich sequences homologous to the S. pombe origin consensus motifs D, C, Z, and M (50) (small black boxes) are marked. AT-rich sequences homologous to a cell cycle-dependent DNase I genomic footprint in the human lamin B2 IR (1, 29, 39) (hatched box) and to the ORC-binding region in the Drosophila chorion ACE3 (6, 80) (checkered boxes) are noted. AT-rich sequences containing stably bent DNA (B) and binding sites for a zinc finger protein of unknown function, RIP60 (black arrowheads), are indicated (15, 16, 21, 45, 64). The position of a cell cycle-dependent nuclease-hypersensitive site (72) is indicated (striped box). (B) The integrated mutant DHFR ori- fragments in 5 ng of DNA from uncloned pools of DR12 cells and the endogenous ori- region in 5 ng of CHOK1 DNA were amplified by PCR and visualized by gel electrophoresis and ethidium bromide staining. The ratio of the mutant amplification products in DR12 relative to those of the endogenous ori- region in CHOK1 is indicated and suggests that the copy number of the exogenous ori- region present in the pool of transfectants mimics that of the endogenous locus in CHOK1. (C) PCR amplifications were performed with each of the four primer pairs and size-fractionated nascent DNA from asynchronous mutant-transfected DR12 cells in the presence of a precalibrated amount of the corresponding competitor DNA. Amplification products were analyzed by PAGE and ethidium bromide staining. Control lanes: C, competitor template only; G, nascent genomic DNA template only; , no template. Numbers above each lane represent the volume in microliters of nascent DNA added to the PCR mixture; note that for pMCDTR, the nascent genomic DNA was used at a 1:10 dilution with pp2 and pp6. (D) The abundance of nascent DNA from pools of mutant-transfected DR12 cells from three independent transfection experiments was quantitatively evaluated. As a measure of initiation activity, the abundance of each target sequence in nascent genomic DNA was normalized to the abundance of pp3 target sequences in the corresponding experiment, which was set equal to 1 (see Table 4 for a typical experiment), and the average of three experiments with each mutant is shown. For comparison, the average initiation activity measured with nascent DNA from a pool of asynchronous wild-type pMCD-transfected DR12 cells (Fig. 3D) is also shown. Bars indicate the SEM.

View this table: [in this window] [in a new window]   TABLE 4.   Abundance of DHFR ori- target DNA sequences in nascent DNAa

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A 5.8-kb region surrounding the DHFR ori- IR functions as an independent chromosomal replicator. Evidence presented in this study indicates that a 5.8-kb fragment of the DHFR ori- region placed in random ectopic chromosomal locations was sufficient to direct initiation of DNA replication from the ori- start site (Fig. 3C). These data were obtained by using a competitive PCR-based nascent strand abundance assay with nascent template DNA enriched only by heat denaturation and size selection. Thus, a possible concern is that the template fraction contained not only nascent DNA but probably also small DNA fragments generated by shearing in vitro and single-strand breaks in vivo. To validate the preparation methods and assays used to assess the initiation activity of the 5.8-kb ectopic ori- fragment, we initially tested these procedures with the endogenous DHFR locus (Fig. 2C). Direct comparison of our data on nascent strand abundance at the ori- site in CHOK1 cells with data obtained by using lambda exonuclease digestion to further enrich for RNA-primed nascent DNA after the size fractionation (38) reveals remarkable congruence (52). The ratio of nascent DNA at the center of the ori- IR to flanking nonreplicating DNA in our study averaged about 15-fold in multiple experiments (Fig. 2C), quantitatively comparable to the 12-fold enrichment reported by Kobayashi et al. for an experiment with CHOK1 cells (compare Fig. 2C of this report with Fig. 6B in reference 52). The reproducibility of the results obtained with the two different methods of enrichment for nascent DNA in the endogenous ori- locus in CHOK1 cells argues strongly that the application of our methods to analysis of the ectopic ori- fragments in DR12 hamster cells accurately reflects the initiation activity emanating from the ectopic fragments. However, it should be noted that the variability in our experiments was somewhat greater than was observed with the additional enrichment for nascent DNA using lambda exonuclease (52).

Is the ori- region composed of essential modular elements? Studies of model systems have shown that replicators usually have a modular organization (reviewed in reference 22). The origin core consists of discrete elements: a DUE, an origin recognition element that contains binding sites for initiator proteins, and sometimes an AT-rich region. The core is often flanked by auxiliary elements that bind transcription factors and enhance the replication activity of the core element up to 1,000-fold.