Site-specific ORC binding, pre-replication complex assembly and DNA synthesis at Schizosaccharomyces pombe replication origins

Abstract

Previous studies have shown that the Schizo saccharomyces pombe Orc4 subunit is solely responsible for in vitro binding of origin recognition complex (ORC) to specific AT-rich sites within S.pombe replication origins. Using ARS3001, a S.pombe replication origin consisting of four genetically required sites, we show that, in situ as well as in vitro, Orc4 binds strongly to the Delta3 site, weakly to the Delta6 site and not at all to the remaining sequences. In situ, the footprint over Delta3 is extended during G(1) phase, but only when Cdc18 is present and Mcm proteins are bound to chromatin. Moreover, this footprint extends into the adjacent Delta2 site, where leading strand DNA synthesis begins. Therefore, we conclude that ARS3001 consists of a single primary ORC binding site that assembles a pre-replication complex and initiates DNA synthesis, plus an additional novel origin element (Delta9) that neither binds ORC nor functions as a centromere, but does bind an as yet unidentified protein throughout the cell cycle. Schizosaccharomyces pombe may be an appropriate paradigm for the complex origins found in the metazoa.

Fig. 1. Orc4 bound to two sites, Δ3 and Δ6, in ARS3001. (A) ARS3001 is located in the non-transcribed spacer of the rDNA repeats and contains two strongly required DNA regions (Δ3, Δ9) and two moderately required DNA regions (Δ2, Δ6) (Kim and Huberman, 1998). (B) DNA band shift assays were carried out with a 452 bp [³²P]DNA fragment (5 ng) (panel A) radiolabeled at both 5′-ends, incubated with the indicated amount of Orc4, and then fractionated by gel electrophoresis. This sequence exhibited full ARS3001 activity (Table I). Control DNA consisted of 688 bp of average sequence taken from pBluescript KSII. (C) Competitive DNA band shift analysis was carried out with 5 ng of ARS3001 [5′-³²P]DNA plus a 52–68 bp DNA fragment containing either Δ2, Δ3, Δ6, Δ9 or control DNA, at the indicated molar ratio. This DNA was then incubated with 8 ng Orc4. Competitor DNA sequences were: (control DNA) 5′-TAAATTTTTCAGGG TCGGTAGAGTCAGAGATGGGTGTGGGAAGGGGTAGTTGTAGG TAGG-3′; (Δ2 DNA) 5′-TTATGGGAAGGTGGAGAGAAAAAATG AAAAAAACAAGGTAATTTGTAGGATT-3′; (Δ3 DNA) 5′-AAT TTGTAGGATTTTTACAAAATAAATAAATACATTTTATATAATT TAACCAAAAGTAATGT-3′; (Δ6 DNA) 5′-AACAAAAAAAGTG CAAACAAATAAAAGAAAAAATAAGAAAACAAAAAAACAACT ACAAAGGTA-3′; and (Δ9 DNA) 5′-ATGAAAAAATAAAGAA AAATTTAATTTATAATTTAACAAAACAATATTTATTGAAAAGCCAATTTTAA-3′.

Fig. 2. Orc4 produced a footprint on the Watson strand of the Δ3 region in vitro that matched a genomic footprint detected in situ. In vitro footprinting was carried out with a 665 bp ARS3001 DNA fragment incubated with either 0, 10, 40 or 80 ng of Orc4 (lanes indicated by right triangle). The 5′-end of the Watson strand (top of gel) was labeled with ³²P. The sequence of the Watson strand (lanes G, A, T and C), beginning at the same 5′-nucleotide, was run in parallel using an appropriate DNA primer. To display the footprint (site A) in the Δ3 region more clearly, the same experiment was carried out with a 452 bp ARS3001 DNA fragment. Both fragments exhibited full ARS activity. Sequences of Δ3 and site A are given in Figure 8. In situ genomic footprinting was carried out on nuclei isolated from cells arrested in G₂ phase. Nuclei (G2 lane) and genomic DNA (DNA) were isolated and subjected to DNase I digestion in parallel. Samples in which the extent of digestion was similar were subjected to primer extension using a primer annealed to the Crick strand. The same primer used to locate the 3′-ends of the Watson strand that was cut by DNase I was also used to display the sequence. Therefore, the sequence shown is the Crick strand, while the DNase I digestion pattern is from the Watson strand. Arrows indicate critical nucleotides that appear in both in vitro and in situ footprints. Sequences and footprinting data for Δ2 and Δ3 are given in Figure 8.

Fig. 3. Orc4 produced a footprint on the Crick strand of the Δ3 region in vitro that matched a genomic footprint detected in situ. The same footprinting analyses described in Figure 2 were carried out on the Crick strand of the Δ2–Δ3 region using a 320 bp ARS3001 DNA fragment that lacks the Δ9 region. Only the pyrimidine sequencing lanes are shown for simplicity. Sequences and footprinting data for Δ2 and Δ3 are given in Figure 8.

Fig. 4. Orc4 produced a footprint on the Crick strand of the Δ6 region in vitro that matched a genomic footprint detected in situ. The same footprinting analyses described in Figure 3 were carried out on the Crick strand of regions Δ3 and Δ6 using ARS3001-320. Genomic footprinting of the Δ6 region was analyzed using ARS3001-2 (Table II). A footprint (site B) was observed both in vitro (A) and in situ (B). No footprint was detected in the Δ9 region in vitro (C), although two DNase I-hypersensitive sites (*) were detected in situ (D). Region Δ9 was analyzed using the Crick strand of a 665 bp ARS3001 DNA fragment. In situ genomic footprinting of the Δ9 region was analyzed using ARS3001-3 (Table II). Sequences and footprinting data for Δ6 and Δ9 are given in Figure 8.

Fig. 5. Orc4 preferentially bound the T-rich strand at ORC DNA binding sites. (A) DNA band shift assays containing 1 ng [5′-³²P]oligonucleotide plus the indicated amounts of Orc4 and ORC-5 complex were incubated at room temperature for 10 min before fractionating the material by gel electrophoresis. The A-rich strand of Orc4 binding site A in ARS3002 DNA (Kong and DePamphilis, 2001) was represented by 5′-TAATACTATTTTTTATATTAATTAAAAA AAAAAAAAAAAAAAACCT-3′. The T-rich strand was represented by 5′-AGGTTTTTTTTTTTTTTTTTTTTAATTAATATAAAAAATA GTATTA-3′. (B) A non-ARS ssDNA sequence from the S.pombe Orc3 gene was used as a control DNA [5′-CCGGCCTCGAGATGC ATCACCATCACCATCACTCAGCAATACTACAATATGATTC-3′]. Poly(dA) and poly(dT) sequences were each 46 residues long.

Fig. 6. The genomic footprint at Δ3 was extended to Δ2 in the presence of pre-RCs. (A) Schizosaccharomyces pombe cells were arrested in G₁ phase either by addition of P-factor, or by inactivating Cdc10 or Orc1, or by depletion of Cdc18. Chromatin was isolated and its proteins were fractionated by SDS–PAGE and then analyzed with antibodies against Mcm2, Mcm6 and Orc4, as described previously (Kong and DePamphilis, 2001). (B) In situ genomic footprinting was performed on the Crick strand as in Figures 3 and 4. The footprint (A) in Δ3 was extended (A*) to Δ2 in P-factor arrested cells (a), but it was not extended in Cdc10^–, Cdc18^– or Orc1-4 arrested cells (a and b). Similar changes were not detected in the Δ6 (c) or Δ9 regions (d).

Fig. 7. Leading strand DNA synthesis began in Δ2. The nucleotide locations of nascent strand start sites were mapped using replicating intermediates (RI) enriched for RNA-primed nascent DNA. Primers 1 and 6 (Table II) were annealed to the Crick and Watson strands, respectively, to identify the ends of nascent DNA strands in the Δ2–Δ3 region, primer set 2 and 5 were used for the Δ6 region, and primer set 3 and 4 examined the Δ9 region (data not shown). The same primers were used to display the sequence of these regions. Arrowheads indicate transition points from discontinuous to continuous DNA synthesis on each strand of Δ2 region (sequence given in Figure 8).

Fig. 8. Schizosaccharomyces pombe ARS3001 contains two strongly required regions (Δ3, Δ9) and two moderately required regions (Δ2, Δ6). ORC binds (through the AT-hook domain of its Orc4 subunit) strongly to Δ3 and moderately to Δ6 (capital letters), but not to either Δ2 or Δ9. Arrowheads in Figures 2–4 are reproduced here. Leading strand DNA synthesis begins at Δ2, marking this as the OBR. Δ9 is required for origin function, not for plasmid segregation, and binds a protein(s) of unknown function throughout the cell cycle.