The One-Kilobase DNA Fragment Upstream of the ardC Actin Gene of Physarum polycephalum Is Both a Replicator and a Promoter

Abstract

The 1-kb DNA fragment upstream of the ardC actin gene of Physarum polycephalum promotes the transcription of a reporter gene either in a transient-plasmid assay or as an integrated copy in an ectopic position, defining this region as the transcriptional promoter of the ardC gene (PardC). Since we mapped an origin of replication activated at the onset of S phase within this same fragment, we examined the pattern of replication of a cassette containing the PardC promoter and the hygromycin phosphotransferase gene, hph, integrated into two different chromosomal sites. In both cases, we show by two-dimensional agarose gel electrophoresis that an efficient, early activated origin coincides with the ectopic PardC fragment. One of the integration sites was a normally late-replicating region. The presence of the ectopic origin converted this late-replicating domain into an early-replicating domain in which replication forks propagate with kinetics indistinguishable from those of the native PardC replicon. This is the first demonstration that initiation sites for DNA replication in Physarum correspond to cis-acting replicator sequences. This work also confirms the close proximity of a replication origin and a promoter, with both functions being located within the 1-kb proximal region of the ardC actin gene. A more precise location of the replication origin with respect to the transcriptional promoter must await the development of a functional autonomously replicating sequence assay in Physarum.

Replication origins can be detected either as genetic elements that confer autonomous replication to plasmids or as physically mapped sites of initiation (40). However, in recent years, it has been shown that both these complementary approaches are needed to define eukaryotic replication origins. It is now well established that some of the autonomously replicating sequences isolated from the yeast Saccharomyces cerevisiae are not active as origins of replication in their native chromosomal context (27, 30, 46). On the other hand, mapping an initiation site does not give information on the spatial distribution of the genetic determinants encoding or controlling origin activity. It has been shown that the amplification control element of the chorion genes in Drosophila does not coincide with the main sites of initiation of DNA replication (21, 38), whereas the locus control region has been implicated in the remote control of an origin located in the promoter region of the β-globin gene in human cells (2, 39). Nevertheless, in Escherichia coli, simian virus 40, or S. cerevisiae, there is a coincidence between the replicator, defined by a specific cis-acting sequence, and the origin, which is the initiation site of DNA replication (40, 55). In each case, the replicator is recognized by a trans-acting protein complex, the initiator. It is believed that the binding of the initiator induces torsional stress that unwinds flanking sequences, leading to local melting of the double helix and to sequential recruitment of the various components of the replication forks (45, 55). This provides a mechanistic justification of the coincidence between the sequence recognized by the initiator and the actual site of initiation. Under these conditions, the priming of the leading strands is confined to the boundary of the initiator binding site, as recently demonstrated for the origin ARS1 of S. cerevisiae (10).

Thus far, eukaryotic chromosomal origins have been characterized in detail only in S. cerevisiae (44, 54, 56). They are characterized by a modular organization with a short 11-bp consensus sequence that is the binding site for the initiator proteins called the origin recognition complex (5, 23). In turn, the origin recognition complex nucleates the cell cycle-dependent assembly of a large prereplicative protein complex (pre-RC) (24, 45). Although, it is clear that the components of the pre-RC are conserved and play a role in DNA replication of multicellular organisms like Drosophila, Xenopus, and humans (34, 45), the organization of the replication origins in these organisms is not known and is a subject of controversy. Conflicting results have been generated by studies aimed at defining either the genetic elements (18, 61) or the specificity of the initiation sites (14, 48, 58, 60) at various loci in mammalian cells. These uncertainties have precluded a detailed description of the relative distribution of genes and replication origins within eukaryotic genomes. This aspect of the functional organization of the genome is central to our work on the slime mold Physarum polycephalum.

Chromosomal DNA replication is highly regulated, both temporally and spatially, in the multinucleated, naturally synchronous plasmodium of Physarum (32, 49). Specific genes replicate in a strictly defined temporal order, and two-dimensional (2D) agarose gel electrophoresis studies (13) have pinpointed replication initiation sites within defined restriction fragments (6–8, 50, 51). This control extends to the simultaneous activation of allelic origins (6). Thus far, the five replication origins that we mapped were located within the promoter regions of abundantly transcribed genes (6, 8, 9). This is in agreement with previous electron microscopic observations on chromatin spreads that show transcribed genes located at the central part of nascent replicons (52).

One of these five origins is situated within the promoter region of the ardC actin gene. A 2D gel electrophoresis analysis showed that the replication initiation site is located about 500 nucleotides (nt) upstream of the transcription initiation site (6). The transcriptional promoter, PardC, was defined in standard chloramphenicol acetyltransferase and luciferase assays as a 1.1-kb piece of DNA upstream of the gene (4, 15). It was further shown that the ectopic integration of one copy of the hph gene (encoding hygromycin phosphotransferase), under the control of this promoter, is sufficient to confer hygromycin resistance to uninucleated Physarum amoebae (16, 17). However, in these experiments, circular plasmids containing the PardC-hph cassette were not maintained as stable episomes, suggesting that the transcription promoter is not efficient in sustaining DNA replication. This could indicate that the genetic determinants specifying the origin activity are located away from the actual site of initiation. To address this question, we studied the pattern of replication of two different ectopic copies of PardC, stably integrated at heterologous sites and promoting the transcription of the hygromycin resistance gene (16, 17). In both cases, we found 2D gel electrophoresis patterns consistent with the firing, at the proper time in S phase, of a replication origin associated with the displaced copy of PardC. These results demonstrate that the 1.1 kb upstream of the actin C gene controls both the transcription of the gene and its timely replication.

MATERIALS AND METHODS

Strains and determination of S-phase time points.

Strain 41T1 is haploid and corresponds to transformant 1 generated by electroporation of amoebae with recombinant plasmid pTB41 (16). Strain 44T28 corresponds to hygromycin-resistant transformant 28 generated by electroporation of amoebae with plasmid pTB44 (17). In both cases, the amoebal strain LU352, which carries a gad mutation (for “greater asexual differentiation”) that permits asexual differentiation (selfing) into haploid plasmodia (20), was used to obtain the transformants. Frozen amoebal strains 41T1 and 44T28 were thawed and grown on agar plates with live bacteria at 26°C until haploid plasmodia formed. These plasmodia were then transferred to shaken axenic liquid cultures and grown by standard procedures in the absence of hygromycin selection (20).

Synchronous plasmodia were cultured on Whatman paper filters. Mitosis was observed under a phase-contrast microscope. Stages of the cell cycle are identified with respect to telophase, e.g., a +5-min DNA sample corresponds to a preparation extracted from a plasmodium harvested 5 min after the synchronous division of the nuclei. Wild-type (WT) strains were M3CIV for gene dosage experiments or TU291 for bromodeoxyuridine (BUdR) incorporation; both are diploid strains usually used for DNA replication studies (32, 50–52).

Probes.

Vector-free inserts were used as probes. The hph DNA probe was the 1.3-kb BamHI fragment from pLG83 (36). The DNA probe for the promoter of the actin C gene (PardC) was the BglII-HindIII 1.1-kb genomic fragment (16). The recipient site of the ectopic copy of the PardC-hph cassette in strain 41T1 was detected by using a 501-bp KpnI-NheI DNA fragment spanning the junction between the integration site and the 5′ end of the ectopic copy of PardC. It was obtained from a 700-bp PCR product (see Fig. 2B). The probes were labelled by random priming (NEN kit) with [³²P]dCTP as a tracer.

FIG. 2.

Analysis of the transformation event in strain 41T1. (A) Structure of the HindIII-digested 8.4-kb pTB41 plasmid (pTB41). The plasmid DNA (pGEM7Zf; Promega) is represented by a dashed line. For a description of the Physarum elements, see the text. Below, a partial restriction map of the ectopic copy of PardC in plasmodial DNA of strain 41T1 is shown (41T1). The first pair of arrows indicates the inserted fragment. Restriction sites relevant for sequence analysis of the 5′ and 3′ junctions of the insertion are shown. The second pair of arrows indicates the genomic fragment isolated for sequencing the 5′ junction between the PardC-hph cassette and the recipient site. Genomic 1.5-kb KpnI-BamHI fragments were selected on an agarose gel and ligated into a KpnI-BamHI-digested pBluescript plasmid, represented by a dashed line (5′ Junction). The convergent arrows represent primers used to amplify a 700-bp fragment spanning the 5′ junction of the integrated DNA. A 501-bp KpnI-NheI fragment was purified, subcloned, and sequenced to delineate the border of the ectopic copy of PardC. (B) This 501-bp fragment (5′-JCT probe) was used to measure the size of the inserted DNA. Southern blotting with HindIII-digested samples shows that this composite probe is specific for the endogeneous PardC copy (6.6-kb fragment) and for a fragment present only in the WT strain (at 3.5 kb) whereas, as expected, it also recognizes the integrated copy of PardC in strain 41T1 (5.6-kb fragment). This demonstrates that the 3.5-kb HindIII fragment of the recipient site was converted into a 5.6-kb fragment, suggesting that only 2.1 kb of the transforming pTB41 plasmid has been integrated. This was confirmed by sequencing the 3′ junction of the insertion (see Materials and Methods). (C) Sequence of the 5′ junction in strain 41T1. (Top) Sequence of the 5′ end of the pTB41 plasmid linearized by HindIII. The first nucleotide from PardC is the G · C base pair following the dash. Nucleotides shown in bold originated from plasmid engineering. (Bottom) Sequence of the 5′ junction. The recombination has taken place precisely at the HindIII site, eliminating the protruding 5′ single-stranded DNA.

PCR amplification of the 5′ and 3′ junctions of the inserted DNA in strain 41T1.

To characterize the junctions between the recipient site and the exogenous DNA in the 41T1 transformant, relevant restriction fragments were size selected on agarose gel. For the 5′ junction, 150 μg of KpnI- and BamHI-digested DNA was electrophoresed on a 0.5% agarose gel. Restriction fragments of about 1.5 kb were eluted from the gel by centrifugation, purified by phenol-chloroform extraction, and ligated for 2 h at 15°C to a KpnI-BamHI pBluescript II KS+ plasmid. A 10-fold excess of the inserts with respect to the plasmid was used. The 5′ junction was then amplified with primers complementary to M13 (5′-CAGGAAACAGCTATGACCAT-3′) and to PardC (5′-AGCCACATACATCCCTAACC-3′, nt 333 to 314 of accession no. M73459) (see Fig. 2, 5′ junction). As expected, a 700-bp DNA fragment was the main product of the amplification reaction (results not shown). This product was reamplified with a second, partly overlapping set of primers (5′-TGACCATGATTACGCCAAGC-3′ and 5′-TAACCACGTTTCCCATTGCC-3′). The 700-bp fragment spanning the 5′ junction was the only amplified product (results not shown). It was further digested with KpnI and NheI (see Fig. 2), and the 501-bp resulting fragment was ligated into a KpnI-XbaI-double-digested pBluescript II KS+ plasmid. A recombinant plasmid was selected following transformation of competent “sure” bacteria (Stratagene), and the sequence of the insert was determined. It is composed of 198 nt from PardC, with only one change from the published sequence, and 288 nt from the chromosomal insertion site, separated by 15 nt of a polylinker originating from plasmid pTB41 (see Fig. 2B and C).

For the 3′ junction, a 2.8-kb fraction of BamHI-KpnI-double-digested total DNA of strain 41T1 was size selected and ligated into the corresponding pBluescript II KS+ vector. A 1-μl volume of the ligation reaction mixture was used as a substrate for PCR with a forward primer from the hygromycin gene (5′-GGCTCTCGATGAGCTGATGC-3′, nt 757 to 776 of V01499) and a reverse primer from within pBluescript (M13 reverse primer). The expected 2.2-kb fragment was the main amplification product. It was reamplified with an internal hph-specific primer (5′-AGCAGACGCGCTACTTGGAG-3′, nt 960 to 979 of V01499). The resulting 2.0-kb fragment was SacII-HindIII digested (see Fig. 2A) and subcloned into pBluescript IIKS+. The junction between the hph gene and the recipient site was determined by sequencing: 156 nt identical to the hph gene and 100 nt of the insertion site were read. The junction occurs at the second nucleotide of the unique ScaI site of the hph gene, truncating the gene such that the last 18 amino acids of the protein are missing (data not shown).

DNA isolation.

For Southern blotting, including gene dosage experiments, soluble DNA preparations were obtained following phenol-chloroform extractions of isolated nuclei (7, 8). For 2D gel electrophoresis analysis, isolated nuclei were embedded into agarose plugs as described previously (6). Density shift experiments with BUdR-substituted DNA were carried out as described previously (50, 51).

RT-PCR.

Total RNA was extracted following solubilization of a plasmodium in guanidium hydrochloride and overnight centrifugation onto a CsCl cushion (50). The pellet was resuspended in diethylpyrocarbonate-treated water, ethanol precipitated, and frozen. Reverse transcription-PCR (RT-PCR) was carried out with a commercial kit from Stratagene. A 10-μg sample of total RNA was subjected to RT with random primers. A 1-μl volume of the cDNA was used as a substrate for RT-PCRs primed with two sets of specific primers (coamplification) and catalyzed by a Taq⁺ long polymerase (Stratagene). The hph-specific primers, 5′-GGCTCTCGATGAGCTGATGC and 5′-TCTACACAGCCATCGGTCCA (nt 757 to 776 and 1197 to 1177, respectively, of V01489), direct the amplification of a 440-bp product, whereas, as a control, the profilin P primers, 5′-ACCCCCGCCAATGTCTTTGC and 5′-CTCAATGAGGTAGTCGGCTAG (nt 703 to 722 and 987 to 967, respectively, of M38038), specify a 211-bp intron-minus band in RT-PCRs and a 284-bp intron-plus product in PCRs (11). The reaction mixture was heated at 91°C for 5 min and cooled to 55°C for 5 min, and the Taq polymerase was added. The reaction mixtures were covered with mineral oil, and 35 cycles of amplification were performed with the schedule 91°C for 1 min, 54°C for 1 min, and 72°C for 1 min. Finally, the reaction mixtures were incubated for another 10 min at 72°C. A 100-ng portion of genomic DNA was used as the substrate in PCR amplification of DNA.

2D gel electrophoresis and hybridization.

2D gel electrophoresis and hybridization were performed as previously described (6–8). Fuji X-ray films were scanned with an Agfa Studioscan IIsi.

RESULTS

Production of transformed synchronous plasmodia of Physarum.

We chose two transformed strains of Physarum amoeba, 41T1 and 44T28, generated by electroporation with plasmids that had in common the PardC-hph selectable cassette (pTB41 [see Fig. 1] and pTB44 [see Fig. 6]). It was shown that only one copy of the transforming DNA was integrated into an ectopic position, leaving the endogeneous PardC intact in both strains (16, 17). To compare the replication pattern of the endogenous and ectopic copies of PardC within the same synchronous cell, transformed amoebae were differentiated into haploid plasmodia (Fig. 1A). Southern blotting with genomic DNA extracted from synchronous plasmodia (Fig. 1B) reproduced the restriction pattern previously obtained with DNA from the transformed amoebae (16, 17). The 1:1 ratio of the hybridization signals from the ectopic (5.6- and 2.1-kb HindIII fragments in strains 41T1 and 44T28, respectively) and endogenous (6.6-kb fragment in WT and transformed strains) PardC copies indicates that the many nuclei of the plasmodium have inherited one copy of the integrated DNA and that this DNA is replicated once per cell cycle (Fig. 1B). Finally, such a transformed plasmodium shows resistance to 200 μg of hygromycin per ml (data not shown), suggesting that the hph expression is maintained during the amoebal-plasmodial transition, even in the absence of hygromycin selection for many cell generations. To ascertain the expression of the hph gene, RT-PCR experiments were conducted with RNA extracted from transformed plasmodia, and in both strains amplification of an hph cDNA was obtained (Fig. 1C). We then studied the two transformed strains individually, determining first which part of the electroporated plasmid was integrated.

FIG. 1.

Production of a transformed haploid plasmodium containing only one expressed copy of the PardC-hph construct. (A) A linear plasmid containing the PardC-hph cassette is shown. Following electroporation into haploid amoebae (zigzag arrow), cells were grown on hygromycin to select for resistant colonies (16, 17). Transformed amoebae, because they carry a gad mutation (for “greater asexual differentiation”), can differentiate spontaneously (selfing) into haploid plasmodia (20). The origin previously mapped by 2D gel electrophoresis (6) within the genomic copy of PardC is schematized. (B) HindIII-digested plasmodial genomic DNA preparations were probed with PardC, revealing the endogeneous copy of the promoter in a 6.6-kb fragment in WT and transformed strains and the ectopic copy at 5.6 kb in strain 41T1 and at 2.1 kb in strain 44T28. (C) Expression of the hph gene in the plasmodia of strains 41T1 and 44T28, as shown by RT-PCR. Total RNAs subjected to RT and extracted from plasmodia of strains 41T1 and 44T28 were used as substrates for RT-PCR amplifications (lanes 2 and 3). Since the hph gene has no intron, as an internal control the profilin proP cDNA was coamplified with a set of primers spanning intron 2 of the gene. In both strains, the expected 440-bp product of the hph cDNA was coamplified with a 211-nt intronless proP cDNA (proP i−). No bands corresponding to the intron-plus proP gene at 284 nt were seen by RT-PCR, ruling out the possibility of an amplification resulting from contaminating genomic DNA. In contrast, only the 284-nt band (proP i+) was amplified with the proP primers in a PCR with genomic DNA from strain 41T1 and 44T28 (lanes 1 and 4, respectively). This confirms that the expression of the hph gene under the control of PardC is maintained during the amoebal-plasmodial transition in strains 41T1 and 44T28.

FIG. 6.

Simultaneous replication of the endogenous and ectopic copies of PardC in strain 44T28. (A) The structure of the 8.7-kb transforming plasmid pTB44, linearized by NotI, is shown above a restriction map of the inserted DNA in strain 44T28. Sequential probing with ardD, PardC, and hph probes showed that these three elements are adjacent within a single 5.6-kb KpnI fragment. The upstream KpnI site is provided by the inserted DNA, whereas the downstream site belongs to the recipient locus. A HindIII chromosomal site found 1.1 kb downstream of the initiation site of the hph sequence marks the truncated 3′ end of the plasmid DNA. (B) 2D gel electrophoresis analysis was performed on DNA preparations extracted from synchronous plasmodia of strain 44T28 either shortly after the onset of S phase (+5 min) or after about one-third of genome replication (+40 min). From each DNA preparation, a KpnI digest and a HindIII digest were probed with the hph and the PardC probe, respectively. The KpnI digest (left) allows the detection of the 5.6-kb fragment that contains the ectopic copy of PardC. A bubble–to–Y-arc transition indicates that the fragment is actively replicated from a bidirectional origin located within the central one-third of the fragment (white portion of the bar below the restriction map). This pattern is consistent with activation of the origin contained within the ectopic copy of PardC. The HindIII digest (right) allows the detection of the endogenous promoter in a 6.6-kb fragment already analyzed in Fig. 4 (see the restriction map). A complete Y-arc is found, indicating a complete dispersion of the leftward fork of the replicon on the fragment at +5 min (arrow). This demonstrates a simultaneous activation of the endogenous and ectopic copies of the PardC origin. Later in S phase (+40 min), only hybridization spots corresponding to linear restriction fragments are seen, demonstrating a temporally controlled replication of the two loci.

Molecular analysis of the integration event in strain 41T1.

In the 8.4-kb plasmid pTB41 (Fig. 2A), the hygromycin gene was bracketed by the PardC and TardC elements (16). The 1.1-kb PardC fragment was previously shown to be efficient as a transcriptional promoter in standard CAT and luciferase assays (4, 15). The 0.6-kb TardC fragment contains the 3′ untranslated region of the actin C gene and provides a polyadenylayion site (37). A 2.1-kb XbaI fragment of Physarum that contains a repetitive element was added downstream of TardC in an attempt to stimulate integration of the electroporated DNA (16). To determine which of these genetic elements are inserted at the ectopic site in strain 41T1, we established a restriction map of the integrated DNA (Fig. 2A) by standard Southern blotting. Using a PCR approach (see Materials and Methods), we then sequenced the junctions between the exogenous DNA and the recipient site. Sequence analysis of the 5′ junction (Fig. 2C) reveals that PardC was integrated in its entirety into the 41T1 transformant. We used the probe generated in this experiment that spans the junction between the recipient site and the integrated DNA (5′-JCT, Fig. 2B) to measure the length of the inserted DNA. As expected, this composite probe recognizes both the ectopic (5.6-kb HindIII fragment) and the endogenous (6.6-kb HindIII fragment) copies of PardC in strain 41T1. By contrast, in a WT strain, the probe recognizes, in addition to the endogenous promoter, a 3.5-kb HindIII fragment that served as a recipient site for the integration (Fig. 2B). This Southern analysis demonstrates that only 2.1 kb of the 8.4-kb electroporated plasmid was inserted by recombination, converting the WT 3.5-kb HindIII fragment into a 5.6-kb fragment. Since the PardC element is 1.1 kb long, our results suggest that the DNA located downstream of the hph gene in the pTB41 plasmid was not integrated. This was verified by sequencing the 3′ junction. A similar strategy to the one used for the 5′ junction was used (see Materials and Methods). This experiment shows that the hph gene is truncated and is missing the coding sequence for the last 18 amino acids (data not shown). The deletion does not inactivate the enzyme since strain 41T1 is hygromycin resistant. These results indicate that we can determine whether the genetic determinants of the replication origin are confined to the 1.1-kb promoter-containing PardC fragment, since no other piece of Physarum DNA has been integrated into strain 41T1 (Fig. 2A).

Pattern of replication of the PardC-hph unit in strain 41T1.

The synchrony of S phase within the plasmodium permits 2D gel replication analysis of single-copy genes by using total nuclear DNA (6, 7), despite the complexity of the Physarum genome (3 × 10⁸ bp). On the other hand, it requires that the replication timing of the DNA fragment of interest be known precisely. So far, the timing of replication of specific genes of Physarum has always been determined in diploid strains (49). First, we verified that the early replication of genes like the endogenous actin C gene and the profilin P gene is maintained in the haploid strain 41T1 (data not shown). In doing these experiments, we observed that the ectopic copy of PardC is replicated concomitantly with the endogenous copies of actin C (see below) and profilin P (results not shown) at the onset of S phase. We therefore studied the pattern of replication of the translocated copy of PardC in a synchronous DNA preparation obtained 5 min after the onset of S phase. As revealed by 2D gel electrophoresis analysis, the 5.6-kb HindIII fragment containing the integrated copy of PardC in its center is replicated actively (Fig. 3A). The clear transition from a bubble to a Y-arc signal is indicative of a bidirectional, efficient, site-specific origin located close to the center of the restriction fragment. These conclusions are reinforced by the shape of the replication intermediates within partially overlapping restriction fragments. Accordingly, a 4.4-kb KpnI restriction fragment in which the hph gene is centrally located is replicated mainly by one fork, giving rise to a Y-arc signal (Fig. 3B), ruling out the possibility of an initiation taking place within the hph sequence. Similarly, the 4.0-kb BamHI fragment spanning the upstream junction is replicated as a Y-arc when hybridized with the PardC probe (Fig. 3C). This suggests very strongly that the origin detected within the HindIII fragment is located between the KpnI and BamHI restriction sites (Fig. 3). These results are compatible with active initiation from within PardC at the appropriate time in S phase. To estimate the size of the nascent, ectopic replicon at this early stage in S phase, the BamHI blot of Fig. 3C was rehybridized with the hph probe. This generated an incomplete Y-arc that illustrates the partial replication of this 8-kb downstream fragment (Fig. 3D). In the nuclei that were the most advanced into the cell cycle, the forks have not yet reached the center of the fragment as deduced from the absence of the inflexion point in the Y-arc signal. This indicates a degree of dispersion of the rightward forks of at most 3 or 4 kb within this fragment (Fig. 3D). If we consider that the origin is most probably located 500 nt upstream of this BamHI fragment, we can estimate that the largest nascent replicons were about (3.5 + 0.5) × 2 = 8 kb. We then compared these results with the size of the endogenous replicon of the native PardC in the same DNA preparation.

FIG. 3.

Replication of the translocated PardC-hph cassette in strain 41T1: a 2D gel electrophoresis analysis. Genomic DNA was extracted from a synchronous plasmodium of strain 41T1 at the onset of S phase (+5 min), restricted with the appropriate enzyme, and electrophoresed on agarose gels. A map of the relevant restriction sites is shown below the 2D gel patterns. The sizes of the fragments analyzed are indicated next to interpretative drawings of the results. In one fragment, the portion that possibly contains a replication origin is in white whereas portions in which the location of an origin is excluded are in black. The distribution of the forks on only half of fragment D is depicted by arrows. For an analysis of the 2D gel patterns generated by replicating molecules, see the original paper of Brewer and Fangman (13). (A) A 5.6-kb HindIII fragment was probed with the hph probe. The clear bubble–to–Y-arc transition demonstrates the presence of a bidirectional, site-specific origin within the central one-third of that fragment (white portion in bar A). (B) A Y-arc was obtained by probing with hph the 4.4-kb KpnI fragment spanning the 3′ end of the hph gene. (C) Similarly, the 4.0-kb BamHI upstream fragment probed with PardC generated a Y-arc. These results restrict the position of the origin to the upstream and downstream extremities of the KpnI and BamHI fragments, respectively (white portions), confirming that it is confined to the central one-third of the HindIII fragment. (D) The BamHI blot shown in panel C was rehybridized with the hph probe. The 8.0-kb downstream fragment is only partially replicated in this synchronous DNA sample, as revealed by the presence of a partial Y-arc. This illustrates the limited dispersion of the replication forks, estimated at about 0 to 3.5 kb in the BamHI fragment (arrowheads), since no forks have reached the inflexion point of the Y-arc at this time point. Taking the position of the origin about 500 nt upstream of the 8.0-kb BamHI fragment, this would correspond to a replicon size distribution from about 0 to 8 kb in that particular DNA preparation.

Simultaneous activation in strain 41T1 of the origins contained within the endogenous and ectopic copies of PardC.

The HindIII blot of Fig. 3A was reprobed with the PardC probe (Fig. 4). Under these conditions, branched DNA molecules generated by the activation of the endogenous and ectopic copies of the PardC origin are revealed on the same blot, demonstrating the simultaneous firing of the two replicons. However, different patterns of hybridization signals are observed because the origins occupy different positions in their respective restriction fragments (see the restriction maps and schematic representation of the fork movements in Fig. 4). The bubble–to–Y-arc transition seen with the hph probe (Fig. 3A) is again revealed by the PardC probe in the recombinant 5.6-kb fragment. This is consistent with the central location of the ectopic copy of PardC in this fragment. In contrast, in its 6.6-kb native restriction fragment, the PardC origin is located at one extremity, about 0.5 kb from the downstream site, such that this fragment is replicated mainly by one fork. At this early stage in S phase (+5 min), the incomplete Y-arc signals the ongoing replication of the fragment by the leftward fork of the replicon. From the presence of the inflexion point in the Y-arc signal, it can be deduced that some forks progressed beyond the middle of the fragment, at about 4 kb away from the origin. Since we know that the rightward fork of that particular replicon diverges from the origin with a similar rate of elongation (6), we concluded that the size distribution of this replicon extends up to 8 kb in this DNA preparation. This size is similar to the estimated size of the ectopic replicon (Fig. 3) and further demonstrates a simultaneous firing, at the onset of S-phase and with a high level of temporal resolution, of the two, nonallelic copies of the PardC origin in the transformed strain 41T1. To find whether the early replication of the integrated copy of PardC is brought about by the displaced origin or is an intrinsic property of the recipient locus, we defined the timing of replication of the recipient site in a WT strain.

FIG. 4.

Simultaneous activation of the endogenous and ectopic copies of the PardC replication origin. The 2D gel in Fig. 3A was reprobed with the PardC probe. This reveals simultaneously the endogenous copy of PardC in its native 6.6-kb fragment and the ectopic copy in its 5.6-kb recombinant fragment (left). The detection of replication intermediates within both fragments at this cell cycle time point (+5 min after the onset of S-phase) demonstrates their simultaneous replication. The schematic representation illustrates the pattern of replication of each fragment as deduced from the 2D gel electrophoresis result. The 5.6-kb fragment (right) is replicated from within the ectopic copy of PardC, generating the bubble–to–Y-arc transition seen in Fig. 3. Arrows below the restriction map illustrate bidirectional replication of the fragment. The 6.6-kb fragment (middle) is replicated from within the endogenous copy of PardC. It is therefore replicated mainly by the leftward fork (arrow). The replication of this fragment is slower, since it is being carried out mainly by only one fork. It is incomplete at this time point, generating a partial Y arc, from which we can deduce the size distribution of the endogenous PardC replicon at about 0 to 8 kb in the sample (see the text). This value is similar to the size distribution measured for the ectopic replicon as seen in Fig. 3D.

Intrinsic timing of replication of the recipient site of integration in strain 41T1.

The 5′-JCT probe (Fig. 2A), which encompasses the 5′ junction of the inserted DNA in strain 41T1, detects in WT HindIII-digested DNA the PardC-containing 6.6-kb fragment, known to replicate at the onset of S phase, and the 3.5-kb HindIII fragment of the recipient site, whose timing of replication is unknown. We took advantage of this property to compare the replication timing of these two loci. First, we studied the distribution of these two fragments in HindIII-digested light-light (LL) and heavy-light (HL) DNA fractions obtained from a WT plasmodium treated with BUdR for the first 40 min of S phase. As expected, the early-replicating 6.6-kb fragment was found almost exclusively in the HL, density-shifted fraction (Fig. 5A). In contrast, the 3.5-kb fragment was found in the LL fraction, suggesting that it had not been replicated by 40 min in S phase. To confirm this result, we further compared the timing of replication of the two loci by gene dosage analysis, a procedure that does not require any treatment of the syncytial plasmodium (7, 50, 51). In this case, a Southern blot made with WT HindIII-digested DNA preparations obtained at specific time points of the cell cycle was probed with the 5′-JCT probe. As seen in Fig. 5B, in the three samples from the G₂-phase DNA preparation, a cell cycle stage at which the two loci are replicated, the two restriction fragments produce hybridization signals of similar intensities. This defines the relative ratio of hybridization of the probe with its two targets. However, in early-S-phase samples, the relative copy number has changed such that the 6.6-kb signal is twice as intense as the 3.5-kb signal, indicating that the 6.6-kb fragment replicates first, in agreement with the BUdR density shift experiment. This unbalanced ratio persists in S phase as long as the second fragment has not replicated. As shown in Fig. 5B, a comparable intensity for the two bands is attained at +120 min in S phase, a cell cycle stage at which about 80% of the genome is duplicated. This gene dosage analysis demonstrates unambiguously that the PardC-hph construct has inserted into a very late-replicating compartment of the genome. Therefore, as a result of this integration, the timing of replication of the recipient site has been advanced by about 2 h, becoming replicated at the onset of S phase, like the endogenous ardC actin gene replicon, as previously shown in Fig. 4.

FIG. 5.

Late replication of the recipient locus in which DNA was inserted in strain 41T1. The 5′-JCT probe derived from strain 41T1 (Fig. 2) was used to determine the timing of replication of the recipient site of the PardC-hph cassette. As shown in Fig. 2B, this probe recognizes, in DNA extracted from a WT strain, both the 6.6-kb HindIII fragment of the endogeneous PardC and the recipient 3.5-kb HindIII fragment. (A) This probe was first hybridized to LL and HL fractions separated on a CsCl gradient after BUdR incorporation for the first 40 min of S phase of a synchronous plasmodium (about one-third of genome replication). As expected, the early-replicating 6.6-kb fragment was density shifted and enriched in the HL DNA fraction. In contrast, the 3.5-kb fragment is enriched in the LL fraction, showing that it is not replicated during the first 40 min of S phase. (B) To expand these results, a gene dosage experiment was carried out. HindIII-digested genomic DNA samples extracted from M3CIV plasmodia at distinct stages of the cell cycle were electrophoresed on a 0.7% agarose gel, blotted, and hybridized to the 5′-JCT probe. Hybridization signals were recorded on a PhosphorImager (Molecular Dynamics) and quantified with ImageQuant. The ratio of the intensities of the 6.6- and 3.5-kb bands at one time point is indicated below each lane. In G₂-phase samples, in which the two fragments are expected to be replicated, the hybridization signals are of similar intensities. However, in S-phase DNA samples, the 6.6-kb band becomes roughly twice as intense as the 3.5-kb band. This corresponds to early replication of the endogeneous PardC and confirms the late replication of the 3.5-kb fragment. This unbalanced ratio persists up to the cell cycle stage where the 3.5-kb band is duplicated, between 90 and 120 min in S phase (about 80% genome replication). This establishes the intrinsic late replication of the locus in which the PardC-hph cassette was integrated in strain 41T1.

These data demonstrate that the PardC-hph construct acts as a dominant replicator with respect to the chromosomal DNA context in transformant 41T1. To further establish this property, we analyzed another independent heterologous integration of PardC resulting from electroporation of amoebae with a plasmid, pTB44, that had a different genetic makeup (17).

Replication pattern of the inserted PardC origin in strain 44T28.

As shown in Fig. 6, the 8.7-kb plasmid pTB44 contained, upstream of the selectable PardC-hph cassette, a 2.3-kb genomic fragment from a mutated allele of the actin-like ardD gene. This gene, which is weakly expressed and replicated early in the plasmodium, encodes a protein that is 84% homologous to the actin encoded by the ardC gene (1). The fragment inserted into pTB44 begins at codon 83 of the gene, contains a deletion spanning most of intron 5 and the first 128 bp of the last exon, and ends with a 1.2-kb 3′ noncoding region (17). The recombinant plasmid was electroporated into haploid amoebae as a linear NotI-digested DNA fragment (Fig. 6A). We chose a transformant, 44T28, in which the unique integration event took place neither at the ardD locus nor at the endogenous PardC site (17). Amoebae of the transformed strain were allowed to differentiate spontaneously into haploid plasmodia that were further grown in the absence of hygromycin selection. This provided us with another possibility to study the pattern of replication of a copy of PardC in an ectopic position.

We analyzed the replication pattern of a 5.6-kb KpnI fragment that, except for the last 1 kb close to the downstream site, is composed of DNA derived from the transforming plasmid (Fig. 6A). A 2D gel electrophoresis analysis demonstrates that this fragment is replicated actively, from a bidirectional origin, located within the central one-third of the fragment i.e., a position that is compatible with the firing of the PardC origin (Fig. 6B, top left). Furthermore, this ectopic origin is activated at the onset of S phase. This is demonstrated by the simultaneous replication of the ardC actin gene (Fig. 6B, top right). The firing of the ectopic origin at the very beginning of S phase suggests that the inserted DNA replicates earlier than its flanking sequences. This assumption is also supported by the absence, in the bubble–to–Y-arc transition shown in Fig. 6B, of a complete Y-arc or of a termination signal that would have resulted from forks entering the region from neighboring origins. Finally, it is shown that at +40 min in S phase, neither the endogenous nor the ectopic copy of PardC is replicating (Fig. 6B, lower panels), providing evidence that both are under the control of the genome temporal order of replication. Together, these results suggest strongly that the ectopic PardC element is also acting as a dominant replicator in strain 44T28.

DISCUSSION

Ectopic activity of the PardC origin.

In this study, we established that the PardC-hph cassette is sufficient, as a translocated piece of DNA, to promote early replication of the recipient locus, even when integrated into a late-replicating chromosomal fragment of the Physarum plasmodium. Our study exploits several advantages. First, only one copy of the exogenous DNA was integrated in each of the two transformants and was found in a different location (Fig. 1B). Second, the 1.1-kb PardC fragment is the only piece of exogenous Physarum DNA in common between the two inserted fragments. Third, a gad mutation allowed us to differentiate the transformed amoebae into plasmodia and hence to compare the replication of the endogenous and ectopic copies of PardC within the same synchronous, untreated cell (Fig. 1A).

Our results confirm the presence of a replication origin within the 1.1-kb promoter-containing PardC fragment. This origin was mapped within the promoter region of the gene by 2D gel electrophoresis, at about 500 nt upstream of the transcription initiation site (6). In the two independent transformants, prominent bubble structures were observed when the ectopic PardC element was in the central one-third of the restriction fragment analyzed (Fig. 3 and 6). This identifies the active replication of the inserted DNA. The absence of a complete Y-arc below the bubble arc is typical of a localized and efficient initiation and mimics the results obtained for the PardC origin in its native genomic site. In the same vein, partial simple Y-arcs, reflecting the distribution of forks on only part of a restriction fragment, were found on regions flanking the ectopic and the endogenous origin as well (Fig. 3D and 4, respectively). Finally, cutting on either side of the ectopic copy of PardC in strain 41T1 converts the bubble-shaped molecules into simple Y-arcs, as predicted if the nascent bubble structures and the restriction sites were coincident (Fig. 3). We conclude that DNA replication initiates within the transferred PardC fragment. Therefore, within the limit of resolution of this study, the initiation site of DNA replication and the replicator coincide within the 1.1-kb PardC element. In turn, this demonstrates that sequences upstream and downstream of PardC, including the introns and the 3′ end of the gene, are dispensable for the origin activity of the endogenous ardC gene.

However, in previous work, it was shown that circular plasmids, containing the selectable PardC-hph cassette, were not maintained as stable episomes when electroporated into haploid amoebae of Physarum. Stable transformants were obtained only following linearization of the vector and resulted, with a low frequency, from integration of the plasmid DNA into chromosomal DNA (16, 17). In other words, in the amoebae, these PardC-hph containing plasmids behave as genetic elements lacking a replication origin. The PardC element is fully functional as a transcription promoter in amoebae (37), but it is not known whether it acts as a chromosomal replication origin in this cell type. Alternatively, the failure to maintain plasmids in amoebae could be explained by the inability of these cells to replicate circular DNA or by the instability of acentric extrachromosomal DNA. As an example, in the yeast Yarrowia lypolitica, a chromosomal origin of replication displays ARS activity only when linked to a centromeric element which, in this organism, is as essential as the origin for the establishment of a replicative plasmid (59).

Firing at the onset of S phase of the displaced PardC origins.

Our data also show that the ectopic copies of PardC are replicated at the onset of S phase. This timing of replication was confirmed by comparison to the endogenous ardC actin gene, known to be duplicated in the first minutes of the 3-h S phase (6). The detection of replicating structures on 2D gels unambiguously demonstrates that in both strains, the two unlinked replicons containing a copy of PardC are firing simultaneously (Fig. 4 and 6). At this cell cycle stage, initiation events are spaced, on the average, every 150 kb in the plasmodium genome (49). The very early replication of the inserted DNA suggests that it is replicated earlier than its flanking sequences. This assumption is supported by the absence of a complete simple Y-arc beneath the bubble arc in Fig. 3 and 6, ruling out the possibility of a passive replication by invading forks originating from nearby origins. Similarly, the absence of termination signals in Fig. 3 and 6 further argues against forks approaching in a direction opposite the direction used by the ones derived from the ectopic PardC origin. Finally, in one case, strain 41T1, our determination of the intrinsically very late timing of replication of the acceptor site (Fig. 5) verifies this point. These findings imply that not only the origin activity but also the timing of activation of the ectopic replicons is controlled by the exogenous PardC-hph DNA.

PardC is a discrete genetic element that dictates the temporal order of replication of flanking sequences.

In a WT diploid plasmodium, the two allelic origins of PardC are firing simultaneously at the onset of S phase (6). Here we show that in a haploid plasmodium, the unique copy of the endogenous origin is still activated at the same cell cycle stage. Furthermore, when an ectopic copy of PardC is introduced into the haploid genome, the two nonallelic copies are again firing simultaneously (Fig. 4 and 6). We conclude that the simultaneous activation of the allelic origins is the result of independent events driven by the presence, in cis, of the PardC replicator.

Numerous studies in the Brewer and Fangman laboratory have shown the importance of the chromosomal context on the timing of activation of a replication origin in S. cerevisiae (28, 29, 31). It is apparent from these studies that in yeast, early activation is the default state of an origin. Late activation is not an intrinsic property of the corresponding origins and is controlled by separable cis-acting flanking sequences (31). Although the mechanisms by which the chromosomal context influences the origin activation time are not understood, it has been shown that the late activation pattern imposed at ARS501 by proximity to the telomere is a cell cycle event taking place in each G₁ phase (53). Moreover, recent reports indicate that the temporal order of replication and the cell cycle progression are coordinated through the sequential action of cyclin-dependent protein kinases and of the related Cdc7/Dbf4 kinase on the early and late origins (25, 26). The data presented in this paper suggest either that PardC behaves as an element insensitive to the chromosomal context or that the integration sites are neutral with respect to origin activity in the two transformants analyzed. Such potentially neutral chromosomal domains have not been demonstrated in yeast, where targeted, homologous integration is always used to exchange early and late origins. At any rate, our data suggest that the late-replicating chromatin of Physarum, in which the 41T1 integration has taken place, has no structural feature that prevents its early replication. It is simply the lack of a nearby early origin that explains the delayed replication. The insertion of the PardC-hph cassette converts a late-replicating domain (Fig. 5) into an early one in which the forks propagate with kinetics that are indistinguishable from the ones of the native early-replicating domain (Fig. 3 and 4).

PardC controls both the replication and the transcription of the ardC actin gene.

The two transformants analyzed were selected on the basis of their hygromycin resistance, suggesting that hph expression is under the control of PardC. For strain 41T1, it was further shown that the hygromycin resistance phenotype was maintained in plasmodia even in the absence of selection (16). Our results extend these conclusions to the transformant 44T28 and demonstrate the presence of an hph transcript in plasmodia of both strains by RT-PCR. Therefore, the displaced copy of PardC acts not only as an origin of DNA replication and a timer for the newly created replicons but also as a transcriptional promoter. In this context, it is noteworthy that our results do not exclude the possibility that the downstream transcription of either the reporter gene or the endogenous ardC actin gene plays a role in the replicator activity located within the promoter-containing DNA fragment. However, it is unlikely that the early replication of the reporter gene is simply the consequence of its transcriptional activity, since we previously identified two genes that, although actively transcribed within the plasmodium, are replicated late in S phase (50, 52).

A physical linkage between active genes and origins of DNA replication has been found in various eukaryotic loci (3, 21, 33, 35, 38, 39, 43, 57), raising a number of questions concerning the functional relationship between the two processes. In several instances, transcription is known to induce DNA replication by providing a RNA primer in the form of a truncated transcript (40). Primer extension of the ardC actin gene mRNA has revealed only one transcription start site, located 33 nt upstream of the ATG (37). This short mRNA leader is present at the downstream extremity of the translocated PardC element and is probably used for the hph transcription. No other transcriptional event in the origin region of PardC is currently known.

Preferential DNA unwinding upstream of transcribed genes, due to negative supercoiling generated by RNA polymerase elongation, has also been postulated to play a role in DNA replication (47). Since there is no G₁ phase in the Physarum plasmodium, it is tempting to speculate that the resumption of transcription that follows mitosis could play a role in the activation of the origins found in the promoter regions of abundantly transcribed genes. However, on early-S-phase chromatin spreads, nascent bubbles were always seen on chromatin regions devoid of transcriptional activity whereas transcripts appeared on both sides of some of the replicons once they have reached a few kilobases (52). These direct observations suggest that initiation of DNA replication precedes transcription at these loci.

It is also known that transcription factors, rather than transcription per se, can facilitate DNA replication. Studies on viral origins have shown the importance of auxiliary sequences that are found adjacent to the core origin with which the initiator protein interacts. These auxiliary sequences contain binding sites for transcription factors and act synergistically to increase initiation frequency (22). A similar situation has been found at the ARS1 origin of S. cerevisiae (42). It has been postulated that resident transcription factors prevent nucleosome repression of an origin in the same way that they prevent nucleosome repression of a promoter (19, 22, 41, 42). In that respect, replication origins and transcriptional promoters are sharing chromatin structure requirements that might provide a selective advantage to a colocalization of these two controlling elements. Such a genomic organization would also ensure a codirectional replication and transcription of the genes (12). This property is seen in other Physarum genes. The actin ardB gene and the profilin proP gene are actively transcribed and duplicated at the onset of S phase from an origin located in their promoter regions (6, 8). Likewise, the two unliked histone H4 genes of Physarum are simultaneously replicated at the onset of S phase from a bidirectional origin located in the immediate 5′ region of these genes (9). Clearly, a mutational analysis is required to determine whether the genetic elements controlling the replication and the transcription of these genes are coincident or whether they can function independently, as it is the case for the origin found in the promoter of the rRNA genes of Tetrahymena (33).

So far, we have mapped five different origins of replication that are activated at the onset of S phase in the plasmodium and are associated with abundantly transcribed genes (6, 8, 9). We have been unable, however, to identify an origin consensus sequence. It should be pointed out that origins mapped by 2D gel electrophoresis are generally localized with a resolution of about 1 kb. As pointed out by Bielinsky and Gerbi (10), in simian virus 40 and S. cerevisiae, the transition between the leading and lagging strands is closely associated with the cis-acting binding site of the trans-acting initiator protein. In Physarum, too, the site of initiation of the different origins, if mapped with a high resolution, might lead to the identification of a consensus cis-acting sequence that has so far escaped our scrutinity.