Portal de Periódicos da CAPES

Structural diversity and dynamics of genomic replication origins in Schizosaccharomyces pombe

2010; Springer Nature; Volume: 29; Issue: 5 Linguagem: Inglês

10.1038/emboj.2009.411

1460-2075

Cristina Cotobal, Mónica Segurado, Francisco Antequera,

Bacterial Genetics and Biotechnology

Article21 January 2010free access Structural diversity and dynamics of genomic replication origins in Schizosaccharomyces pombe Cristina Cotobal Cristina Cotobal Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Salamanca, Spain Search for more papers by this author Mónica Segurado Mónica Segurado Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Salamanca, Spain Search for more papers by this author Francisco Antequera Corresponding Author Francisco Antequera Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Salamanca, Spain Search for more papers by this author Cristina Cotobal Cristina Cotobal Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Salamanca, Spain Search for more papers by this author Mónica Segurado Mónica Segurado Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Salamanca, Spain Search for more papers by this author Francisco Antequera Corresponding Author Francisco Antequera Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Salamanca, Spain Search for more papers by this author Author Information Cristina Cotobal1, Mónica Segurado1 and Francisco Antequera 1 1Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Salamanca, Spain *Corresponding author. Instituto de Microbiología Bioquímica, CSIC/Universidad de Salamanca, Edificio Departamental, Campus Miguel de Unamuno, Salamanca 37007, Spain. Tel.: +34 923 121778; Fax: +34 923 224876; E-mail: [email protected] The EMBO Journal (2010)29:934-942https://doi.org/10.1038/emboj.2009.411 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info DNA replication origins (ORI) in Schizosaccharomyces pombe colocalize with adenine and thymine (A+T)-rich regions, and earlier analyses have established a size from 0.5 to over 3 kb for a DNA fragment to drive replication in plasmid assays. We have asked what are the requirements for ORI function in the chromosomal context. By designing artificial ORIs, we have found that A+T-rich fragments as short as 100 bp without homology to S. pombe DNA are able to initiate replication in the genome. On the other hand, functional dissection of endogenous ORIs has revealed that some of them span a few kilobases and include several modules that may be as short as 25–30 contiguous A+Ts capable of initiating replication from ectopic chromosome positions. The search for elements with these characteristics across the genome has uncovered an earlier unnoticed class of low-efficiency ORIs that fire late during S phase. These results indicate that ORI specification and dynamics varies widely in S. pombe, ranging from very short elements to large regions reminiscent of replication initiation zones in mammals. Introduction Eukaryotic replication origins (ORIs) are chromosomal regions in which the origin recognition complex (ORC) and other initiator proteins assemble to initiate DNA replication (DePamphilis, 2005). Although endowed with the same function, there are significant structural differences and sequence degeneracy between individual ORIs even within the same genome, making it difficult to predict their localization on the basis of their sequence. Among eukaryotes, ORIs have been studied in great detail in the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe because of their compact genomes and the feasibility of genetic approaches. In S. cerevisiae, a consensus sequence 11–17 bp long (Theis and Newlon, 1997) is required for origin activity, although only a small proportion of all of these elements colocalizes with active ORIs in the genome (Breier et al, 2004). Even in this yeast, ORIs are degenerated in the sense that auxiliary elements, that are also required for ORI activity, vary between different ORIs (Theis and Newlon, 2001; Dershowitz et al, 2007). In S. pombe, it has long been known that autonomus replication sequence (ARS) elements have a high content in adenine and thymine (A+T) (Maundrell et al, 1988) and deletion analyses have failed to identify any essential consensus element (Clyne and Kelly, 1995; Dubey et al, 1996). A+T richness has also been found to be a property of genomic ORIs (Dubey et al, 1994; Okuno et al, 1997; Gómez and Antequera, 1999) and this feature led to the identification of A+T-rich islands, which colocalize with active genomic ORIs in 90% of the cases tested. A+T-rich islands were initially defined as regions 0.5–1 kb long with an A+T content ranging between 72 and 75% (Segurado et al, 2003). Recent studies have shown that the majority of these regions bind ORC in vivo (Hayashi et al, 2007) and colocalize with distinct peaks of early replication during the S phase (Feng et al, 2006; Heichinger et al, 2006; Mickle et al, 2007). See Mickle et al (2007) for a detailed comparison between predictions of S. pombe genomic ORIs and for the different names given to them using different approaches. In contrast with the diversity of ORI sequences, the six subunits of ORC are conserved among all eukaryotes studied (Bell and Dutta, 2002). An important exception arose with the identification of the gene encoding the Orc4 subunit of S. pombe (Chuang and Kelly, 1999). This protein contains a carboxy domain of about 450 amino-acids homologous to the Orc4 protein in other eukaryotes in addition to a—so far exclusive—amino-domain of a similar size that includes nine AT-hook motifs capable of binding A+T-rich stretches 4–8 bp long in vitro regardless of their primary sequence (Reeves and Beckerbauer, 2001). This domain is not required for the assembly of ORC in vitro, but mutant complexes lacking it are unable to bind to ORIs (Lee et al, 2001) and this is probably the reason why the expression of the carboxy domain alone cannot rescue the lethal phenotype caused by the absence of the wild-type (wt) protein (Chuang et al, 2002). This two-component system made up of A+T-rich ORIs and Orc4p provides a reasonable explanation of how ORC is targeted to specific sites in the genome to initiate replication. Nevertheless, this general model still leaves unanswered the question of how ORIs are specified in the S. pombe genome; that is which are the minimal requirements in terms of size and base composition for a genomic region to recruit ORC and initiate replication. This question is relevant because only a small proportion of intergenic regions displays detectable ORI activity in the genome (Gómez and Antequera, 1999; Segurado et al, 2003) and bind ORC in vivo (Hayashi et al, 2007) despite having an average composition as high as 70% A+T. In this study, we have found that ORIs in the S. pombe genome range from short A+T-rich elements to initiation zones made up of several degenerate elements each of which has the potential to drive replication individually. These results dramatically reduce earlier estimates of ORI size in S. pombe based on plasmid assays and have led to the finding of a new class of low-efficiency and late-firing genomic ORIs. Results Artificial A+T-rich DNA fragments specify active replication origins in S. pombe Earlier work from several laboratories has proposed that ARS elements in S. pombe are large and lack consensus sequences (Maundrell et al, 1988; Clyne and Kelly, 1995; Dubey et al, 1996; Okuno et al, 1997; Dai et al, 2005) and deletion experiments have identified important elements for replication initiation in plasmids (Clyne and Kelly, 1995; Dubey et al, 1996; Okuno et al, 1999) and in the chromosome (Takahashi et al, 2003). These studies have invariably used S. pombe DNA either as a donor for ARS assays or have manipulated endogenous genomic sequences to identify relevant functional elements. We have taken a more radical approach to address the question of what are the minimum requirements to specify active ORIs by asking whether custom-made synthetic DNA fragments covering a range of sizes and A+T content could mimic the activity of endogenous ORIs. As a first step, we synthesized three sets of DNA fragments 100 bp long with an A+T content of 75, 83 and 90%, respectively. Every 100 bp monomer within each class had a different sequence and none of them derived from S. pombe or showed significant similarity to any sequence in its genome (see Materials and methods for construction and cloning details and Supplementary Figures 1 and 2 for sequences). The three sets of fragments were self-ligated independently and cloned into the 2.9 kb pBSK+ plasmid lacking an S. pombe replication origin, but harbouring the ura4 gene as a selectable marker. Ligation mixtures were used directly to transform an S. pombe ura-d18 strain auxotroph for uracil, and cells that incorporated plasmids capable of autonomous replication were selected on minimal medium plates. Sequence analysis of cloned inserts from a random selection of the transformant colonies showed that the minimum size of fragments displaying ARS activity was approximately 100, 200 and 500 bp in the 90, 83 and 75% A+T sets, respectively (Figure 1A). Figure 1.Plasmid replication by artificial ARS elements. (A) Size of ARS fragments of 90% (red), 83% (yellow) and 75% (green) A+T content capable of replicating the pBSK+ plasmid harbouring the ura4 gene as selectable marker. Sequences of the 27 fragments are shown in Supplementary Figure 1. (B) ARS activity of 15 fragments from 102 to 472 bp long with an 83% A+T content. Efficiencies are relative to the positive control ars1 (1.2 kb) (black bar). Plasmid pBSK+ with the ura4 gene without any ARS element was used as a negative control (C−). The results are average of two independent experiments. The sequences of all the fragments are shown in Supplementary Figure 2. (C) Inactive monomers (m) (a, b and c) from panel (B) display increasing ARS activity when self-ligated as dimers (d) or trimers (t). Ars1 and C-, positive and negative controls as in (B). Download figure Download PowerPoint These results indicated that A+T-rich DNA sequences alien to the S. pombe genome could drive efficient plasmid replication. It remained possible, however, that cells might have selected fragments containing some cryptic sequence pattern meaningful for them, but unidentifiable to us. To test this possibility, we cloned 15 individual fragments of between 102 and 472 bp and 83% A+T into the 2.9 kb pBSK+ plasmid and recovered them in E. coli. ARS analysis in S. pombe showed that all fragments longer than 200 bp were active, ruling out the possibility that specific sequence elements were required (Figure 1B). In addition, and consistent with the results in Figure 1A, none of the six fragments approximately 100 bp long was active, probably because of their small size. To test whether these fragments could be active in the context of a larger plasmid, we repeated the ARS assay after cloning them into the 5.3 kb pJK148 plasmid. None of the six fragments generated any colony except the 111 bp long that generated a reduced number of very small and heterogeneous colonies (data not shown). To test whether the lack of ARS activity of these fragments was indeed because of their small size, we self-ligated three different inactive 100 bp monomers (a, b and c in Figure 1B) into dimers and trimers and cloned them in the pBSK+ plasmid. Figure 1C shows that the three sets of oligomers made up of individually inactive monomers reached an ARS activity comparable with that of the positive control ars1 depending on their size. Could these ARS elements of arbitrary sequence also initiate replication in the genome? To answer this question, we selected 11 fragments of between 100 and 700 bp and 75, 83 or 90% A+T identified in Figure 1A and targeted them by homologous recombination to position 108845 in chromosome III, between the SPCC330.19c and SPCC330.03c genes (Figure 2A). The nearest predicted ORIs were 32 kb upstream and 10 kb downstream from this site (Segurado et al, 2003; Hayashi et al, 2007) and we have earlier confirmed by two-dimensional electrophoresis that this region does not contain a replication origin (Gómez and Antequera, 1999; Figure 2B, wt). Integration of the ura4 gene by itself did not change the passive replication pattern of this locus (Figure 2B, ura4). To test whether this region was permissive for replication initiation, we showed that ars1 was equally efficient as an ORI when integrated into this region (Figure 2B, ars1) as in its endogenous location (Gómez and Antequera, 1999). Analysis of the integrated exogenous fragments showed that 9 out of 11 were capable of initiating replication. Only one out of the three fragments of 75% A+T showed weak ORI activity (Figure 2C, 75.5) despite their relatively large size, suggesting that fragments of this base composition probably represent the lowest boundary required for genomic ORIs to be specified. This notion was supported by the fact that fragments with 75% A+T yielded a significantly lower number of transformants than those of 83 and 90% in plasmid ARS assays (data not shown). In contrast, all eight 83 and 90% A+T fragments tested were capable of initiating replication efficiently (Figure 2D and E). The presence of the ura4 marker linked to the integrated fragments was not required for ORI activity, as shown by the unchanged activity of the 83.1 fragment on removal of ura4 from this region in the genome (Figure 2D, 83.1—ura4). Figure 2.Artificial ORIs initiate replication in the S. pombe genome. (A) Diagram of the insertion site (black arrow) of artificial ARS elements in chromosome III. White arrowheads indicate nearest predicted ORIs. rDNA represents ribosomal RNA locus at the left end of the chromosome. The expanded region below shows genes surrounding the insertion site. Bracket indicates the genomic region centred on the insertion site tested by two-dimensional electrophoresis. (B) Replication of the 3.8 kb Xba I restriction fragment in wild-type cells (wt), and after integration of the ura4 marker or ars1. The presence of a bubble replication arc in ars1, which indicates the presence of an active ORI, is indicated by a black arrow. Bubble arcs in all the other 2D gels in this work are also indicated by black arrows. (C–E) Replication pattern on insertion of the indicated fragments from Figure 1A with a 75, 83 and 90% A+T content, respectively. Download figure Download PowerPoint Taken together, these results indicated that A+T-rich fragments of arbitrary sequence could drive replication initiation as efficiently as endogenous ORIs, and confirmed the lack of the requirement of a specific sequence pattern. A second conclusion was that fragments as short as 100–200 bp could drive robust ORI activity, indicating that genomic ORIs could be specified by regions much shorter than those found in plasmid ARS assays (Maundrell et al, 1988; Dubey et al, 1996; Okuno et al, 1999; Dai et al, 2005). Replication initiation zones in the S. pombe genome are made up of individually active redundant elements Many S. pombe chromosomal ORIs colocalize with regions of 0.5–1 kb with a high A+T content that we have called A+T-rich islands (Segurado et al, 2003). The results depicted in Figures 1 and 2 showed, however, that efficient replication could be driven from shorter regions and this raised the question of whether A+T-rich islands were an absolute requirement for replication initiation. To address this question, we chose the intergenic region encompassing A+T island 1090 (Figure 3A). This region is located between the vma5 and SPAPB2B4.04c genes in chromosome I and colocalizes with a strong ORI (Figure 3C, wt). The NcoI restriction fragment of 5.6 kb in wt cells tested by two-dimensional electrophoresis encompassed the A+T island and the entire intergenic region, making it impossible to distinguish whether independent elements could contribute to ORI activity. As a first step in the functional dissection of this region, we assayed the ARS activity of five non-overlapping fragments across the regions 1, 5, 6, 7 and 8 in Figure 3A. We found that fragment 5 (321 bp and 79% A+T), which spanned the A+T island, was active, consistent with the results in Figure 1 (Figure 3B). In addition, fragment 1 (1.1 kb) displayed an even higher activity despite the fact that its global A+T content was 68%, which is very similar to the 70% intergenic average. Cleavage of this fragment into two halves generated fragment 2, which was inactive, and fragment 4, which retained over 50% of the activity. Close sequence examination revealed the presence of an uninterrupted stretch of 29 A+Ts (29W) at the 5′ end of this fragment. Deletion of this element generated fragment 3 and resulted in the complete loss of ARS activity (Figure 3B). The 29W stretch also lacked ARS activity by itself when cloned into the plasmid. Fragments 6, 7 and 8 were also inactive despite their relatively large size. Figure 3.Functional dissection of the intergenic region encompassing A+T-rich island 1090. (A) A+T content in 300 bp windows (step 50 bp) across the region between the divergently transcribed genes SPAPB2B4.04c (red) and vma5 (blue) genes. A+T island 1090 is indicated by an arrow. Vertical lines represent restriction sites for NcoI flanking the 5.6 kb region tested by two-dimensional electrophoresis. The genomic intergenic average is 70% A+T (red line). Fragments tested for ARS activity in panel (B) are indicated by numbered colour bars and the 29 bp stretch of A+Ts by a yellow triangle (29W). (B) ARS activity of fragments in panel (A) relative to ars1 (black bar). (C) Two-dimensional electrophoresis of the region indicated in panel (A) in exponential wild-type cells (wt) and in cells deleted for region 5 (Δ5), 29W (Δ29W), both simultaneously (Δ5+Δ29W), and regions 5 and 1 combined (Δ5+Δ1). Download figure Download PowerPoint To evaluate the relative contribution of fragments 5 and 29W to ORI activity, we deleted them independently or in combination in their chromosomal context. Two-dimensional electrophoresis revealed that in both cases replication initiation was diminished, but not abolished, indicating the presence of other elements capable of sustaining residual activity (Figure 3C). Deletion of half of the intergenic region encompassing fragments 5 and 1 (which included 29W) (Δ5+ Δ1) resulted in a further reduction in activity, which probably derived from cooperation between fragments 6, 7 and 8, each of which lacked ARS activity when assayed individually. Indeed, when these three fragments were ligated into a plasmid to generate a unique fragment in the genomic configuration (6–8), their ARS activity was comparable with that of ars1, although the colonies grew at a much smaller rate comparable with that shown in Figure 4C (see later). Figure 4.ORI specification by individual modules at an ectopic location. (A) Two-dimensional gel analysis of fragments 5 and 29W integrated at the insertion site (IS, black arrow) in chromosome III shown in panel (B). (B) A 1 kb region encompassing the insertion site (ISR 1 kb), the 29W element alone (29W) and the 29W element integrated into the ISR 1 kb fragment (ISR+29W) were cloned into the pBSK+/ura4 vector and tested for ARS activity. The histogram shows that neither of the two individual fragments showed any ARS activity, but the composite fragment yielded the same number of colonies as the positive control ars1 (black bar). (C) Colony growth in plates of minimal medium after 4 days at 32°C. Cells transformed with the plasmid carrying the ISR+29W fragment grew at a much lower rate than the ars1 control plasmid. Neither fragment generated any transformants when assayed independently (top panels). Download figure Download PowerPoint To test whether fragments 5 and 29W contained enough information to drive replication by themselves in the genome, we integrated them independently at the ectopic location in chromosome III (Figure 4B). Fragment 5 drove replication from this site (Figure 4A, 5/Chr III), consistent with the observation that fragments of a comparable size and A+T content were also active in this locus (Figure 2). Contrary to our expectations, however, we found that the 29W element also generated an active, although less efficient, ORI on integration in the same site (Figure 4A, 29W/Chr III). To confirm the presence of the weak bubble arc relative to the passive replication arc, we repeated this 2D gel in triplicate from cultures grown from independent colonies with the same result in the three cases (data not shown). This unanticipated result raised the question of how such a small fragment could specify a genomic ORI despite the fact of being inactive when assayed in a plasmid. We speculated that 29W might recruit ORC through the AT-hooks of the Orc4p subunit and that this binding could be stabilized by the surrounding context of the intergenic region that had a higher A+T content than when cloned in the plasmid. To test this possibility, we measured the ARS activity of the 29W element cloned in the middle of a 1 kb fragment encompassing the insertion site region (ISR, Figure 4B). This region was unable to initiate replication in the genome (Figure 2B, wt) and did not display ARS activity in a plasmid either (Figure 4C, ISR). However, insertion of 29W into this inactive fragment (ISR+29W) generated a similar number of colonies in the ARS assay as the positive ars1 control, although the colonies grew at a much slower rate (Figure 4C), which is consistent with the different efficiency of ars1 and 29W to initiate replication on integration in chromosome III (compare ars1 in Figure 2B with 29W/Chr III in Figure 4A). The inability of 29W to sustain ARS activity by itself could explain why short elements on this kind had not been identified earlier in ARS assays. These results supported the hypothesis that 29W could recruit ORC to DNA and the complex could be further stabilized by binding to flanking regions and implied that the minimum size of a DNA sequence to specify an active ORI could be even shorter than the 100–200 bp found in Figures 1 and 2. We tested this possibility directly by chromatin immunoprecipitation (ChIP) with an antibody against the Orc4 subunit of ORC. The results of four immunoprecipitation assays from independent cultures of the S. pombe strain carrying the ectopic 29W insertion and another four of the wt strain as a control showed a 23% increase in the amount of immunoprecipitated Orc4 in the strain harbouring the ectopic 29W element relative to the wt control (Supplementary Figure 3). Although small, this increase relative to background is consistent with the ORC binding profile of other poly-W ORIs found in an earlier ChIP-on-chip genome-wide analysis using antibodies against Orc4, Orc1 and Mcm6 (Hayashi et al, 2007). Poly-W stretches colocalize with late-firing genomic ORIs There are 460 non-overlapping stretches of 29 contiguous A+Ts in the S. pombe genome, approximately half of which map to intergenic regions devoid of A+T-rich islands and do not colocalize with the sites of ORC binding (Hayashi et al, 2007). Considering that perhaps even shorter stretches could drive replication, an immediate implication of these results is that the number of active ORIs in the genome could be much higher than current estimates (Segurado et al, 2003; Heichinger et al, 2006; Hayashi et al, 2007; Mickle et al, 2007). To test this possibility, we selected 12 genomic regions that contained uninterrupted stretches of homogenous Ts (or As) or a combination of both bases (W) between 21 and 41 bp long and analysed them by two-dimensional electrophoresis. All regions tested were at least 12 kb distant from the nearest predicted ORI (Segurado et al, 2003; Hayashi et al, 2007). The results depicted in Figure 5 show that 8 out of 12 regions displayed bubble arcs of an intensity comparable with that generated by the 29W element when inserted in chromosome III (Figure 4A, 29W/Chr III). In sharp contrast with ORIs colocalizing with A+T-rich islands such as 1090 (Figure 3C, wt and Segurado et al, 2003), a common feature of poly-W ORIs was their low efficiency. This was indicated by the low ratio of the intensity between bubble and fork arc signals that, in some cases, bordered the limits of detection (Figure 5). Figure 5.Short poly-W tracts colocalize with active ORIs. Two-dimensional gel electrophoresis of 12 genomic regions including runs of contiguous thymines (T) or combined thymines and adenines (W) in asynchronous, exponentially growing wild-type cells. Figures indicate the number of nucleotides in each track. The genomic localization of poly-W stretches and of the restriction fragments analysed are indicated in Supplementary Figure 6. Download figure Download PowerPoint A general correlation between ORI efficiency and time of activation during S phase has been described in S. pombe (Heichinger et al, 2006). To address whether poly-W ORI firing was biased towards any particular period during S phase, we synchronized S. pombe cdc10-129 temperature-sensitive mutant cells by arresting them in G1 at the restrictive temperature. On release at the permissive temperature, we monitored the firing of the strong origin AT1090 and five of the poly-W ORIs identified in Figure 5 at the indicated times during the ensuing S phase (see Supplementary Figure 4 for the FACS analysis). As shown in Figure 6, poly-W ORIs 31W, 39T and 41W fired at 60 min, and 33T and 28W at 90 min into S phase, whereas bubble intermediates were already detectable in the 30 min sample in origin AT1090. As discussed below, the low efficiency and late firing of poly-W ORIs is probably because of a low affinity for ORC, as indicated by the fact that only three out of the eight positives in Figure 5 (21T, 33T and 34T) were identified as sites of ORC binding by ChIP analysis (Hayashi et al, 2007). Figure 6.Poly-W ORIs fire late during S phase. S. pombe cdc10-129 temperature-sensitive mutant cells were arrested in G1 at 36°C for 4 h. Samples were taken at the indicated times during S phase and the activation time of ORIs 1090, 31W, 39T, 41W, 33T and 28W was monitored by two-dimensional electrophoresis. Samples from different time points contained the same amount of DNA and autoradiographs were exposed for the same length of time. Differences in the intensity of passive replication arcs are due to the variation in the amount of replication intermediates at different times during S phase, consistent with the FACS analysis of the progress across S phase shown in Supplementary Figure 4. The same membrane was hybridized successively with probes against 1090, 31W and 28W ORIs. Independent membranes were used for the remaining hybridizations. A sample from exponential asynchronous cdc10-129 cells (Exp) growing at 25°C was also included in the analysis. Download figure Download PowerPoint Discussion ORI specification in S. pombe An essential and as yet unanswered question in the eukaryotic DNA replication field is the identification of the elements that specify active ORIs in the genome. In S. pombe, plasmid replication assays have established that ARS fragments range between 0.5 and over 3 kb (Maundrell et al, 1988; Clyne and Kelly, 1995; Dubey et al, 1996; Okuno et al, 1997; Dai et al, 2005), and it is assumed that similar requirements would apply to genomic ORIs. In fact, size is often cited as one of the major differences between ORIs in S. pombe and S. cerevisiae, in which ARS elements can be shorter than 200 bp (Van Houten and Newlon, 1990; Marahrens and Stillman, 1992). Contrary to this assumption, our results show that poly-W tracks as short as 30 bp long are sufficient to specify replication initiation in the context of intergenic regions (Figure 4) and to reliably predict the localization of active ORIs in the S. pombe genome (Figure 5). An immediate consequence of these results is that current estimates of ORIs in the genome should be expanded, and at the same time raises the question of why this class of ORIs has not been detected in earlier studies. Segurado et al (2003) found that 90% of A+T-rich islands colocalized with relatively strong ORIs (e.g. see A+T island 1090 in Figure 3). However, because that study used a lower size limit of 500 bp, based on earlier ARS data in the literature, many of the ORIs specified by short poly-W tracks were probably excluded from that prediction. Hayashi et al (2007) identified 460 sites of ORC binding by ChIP followed by microarray hybridization, many of which colocalized with A+T-rich islands. However, five out of the eight weak ORIs in Figure 5 (31W, 34W, 28W, 39T and 41W) were not detected in that study, indicating that a large proportion of this class of ORIs either bind ORC very weakly or do so only in a small proportion of the cells, such that their identification falls below the threshold of reliable detection by ChIP. A third study based on microarray hybridization using DNA replicated in the presence of hydroxyurea predicted the existence of 401 high-efficiency and 503 low-efficiency ORIs in S. pombe (Heichinger et al, 2006). It is difficult to ascertain how many of these low-efficiency ORIs colocalize with poly-W tracks because the average density of the microarray used (about 1 probe/kb) often left gaps spanning several intergenic