Orc5 induces large-scale chromatin decondensation in a GCN5-dependent manner

In eukaryotes, the initiation of DNA replication requires the coordinated action of a multiprotein pre-replication complex (preRC) at the origins (Bell and Dutta, 2002). The origin recognition complex (ORC), a six-subunit complex, binds to replication origins during the G1 phase of the cell cycle, and this is followed by a sequential assembly of other preRC components (Bell and Stillman, 1992). In addition to their role in replication initiation, ORC subunits contribute to other cellular processes including transcriptional silencing, heterochromatin organization, sister chromatid cohesion, centrosome duplication, telomere maintenance and cytokinesis (Sasaki and Gilbert, 2007).

Open chromatin structures are known to regulate the efficiency of preRC formation, thereby facilitating replication initiation (Papior et al., 2012). However, the molecular mechanisms that affect chromatin structure and how the preRC components establish themselves on the chromatin remain to be understood. The accessibility of the replication factors is influenced by the chromatin structure, and the chromatin architecture dictates the efficiency of origin usage and firing (Brown et al., 1991; Ferguson and Fangman, 1992; Simpson, 1990; Stevenson and Gottschling, 1999). Histone acetylation is known to play a key role in the regulation of origins of DNA replication in yeast and Drosophila, and there is accumulating evidence that the deacetylation of histones negatively affects origin activity (Aggarwal and Calvi, 2004; Groth et al., 2007a; Knott et al., 2009b; Unnikrishnan et al., 2010; Vogelauer et al., 2002). Furthermore, the replication timing of the β-globin gene domain in human cells is also modulated by histone modifications at the origin (Goren et al., 2008). There is accumulating evidence that histone acetyl transferases act as positive regulators of replication origins in yeast and Drosophila as well as human cells (Groth et al., 2007b; Knott et al., 2009a). In yeast, GCN5p (also known as KAT2A in humans), a histone acetyl transferase (HAT), has been found to positively stimulate DNA replication by negating the inhibitory effect of the histone deacetylases (Espinosa et al., 2010; Vogelauer et al., 2002). Further, Hat1p and its partner Hat2p interact with the ORC (Suter et al., 2007). In Drosophila, the HATs Chameau (Chm) and CBP (Nejire) stimulate origin activity (Aggarwal and Calvi, 2004; McConnell et al., 2012). In human cells, HBO1 (also known as KAT7), another HAT, associates with ORC and is required for the loading of the minichromosome maintenance complex (MCM) onto chromatin and for replication fork progression (Iizuka et al., 2006; Iizuka and Stillman, 1999; Miotto and Struhl, 2008, 2010). A recent study has pointed out that the acetylation of some histone lysine residues depends on the binding of ORC to the origin and that the acetylation is at its maximum on the nucleosomes adjacent to one side of the major initiation site (Liu et al., 2012). How the ORC regulates such chromatin modifications and how the chromatin structure at origins is organized remain to be defined.

The ORC comprises six subunits, and in human cells they are highly dynamic. The largest subunit, Orc1 is degraded at the end of G1, and rebinding of the protein to chromatin is an obligatory step for the establishment of the preRC in G1 (Méndez et al., 2002). The smallest subunit of ORC, Orc6, binds to the ORC in a transient manner and also has independent roles in cytokinesis (Bernal and Venkitaraman, 2011; Prasanth et al., 2002). Orc2, Orc3, Orc4 and Orc5 in human cells constitute the core ORC and are associated with each other throughout the cell cycle (Dhar et al., 2001; Vashee et al., 2001). Orc1, Orc4 and Orc5 are members of the AAA+ family of ATPases and contain consensus motifs. Mutations in the ATP-binding sites on Orc4 and Orc5 impair complex assembly, whereas the ATP binding of Orc1 is dispensable (Ranjan and Gossen, 2006; Siddiqui and Stillman, 2007). Orc2 and Orc3 also have a structure that is common to AAA+ proteins, but they do not possess a consensus ATP-binding motif.

Multiple subunits of human ORC – including Orc1, Orc2, Orc3 and Orc5 – and the protein ORC-associated (ORCA) have roles in heterochromatin organization (Giri et al., 2015; Prasanth et al., 2010; Shen et al., 2010). In yeast, Orc5 has distinguishable functions in replication initiation and silencing (Dillin and Rine, 1997). Further, in Drosophila and humans, the loss of multiple ORC subunits leads to chromosome segregation defects (Pflumm and Botchan, 2001; Prasanth et al., 2004b). In this manuscript, we report that Orc5 has a distinct function in chromatin unfolding. Ectopic tethering of Orc5 to a chromatin locus leads to dramatic chromatin decondensation, predominantly during G1 phase of the cell cycle. This chromatin-opening role of Orc5 requires the activity of the HAT GCN5. We propose that the Orc5 subunit of the ORC plays a key role in mediating large-scale chromatin-opening during G1 that, in turn, facilitates the loading of other preRC components onto the origins.

To investigate the chromatin changes that occur when preRC proteins, including ORC proteins, bind to origins, we tethered individual subunits of the ORC to a heterochromatic locus using an in vivo human U2OS osteosarcoma cell system (CLTon) (Fig. 1A). This reporter carries a stably integrated 200-copy transgene array with lac operator repeats, and this heterochromatic locus is visualized through the stable expression of Cherry–lac-repressor (Cherry–LacI). Upon transcriptional activation of this reporter locus, through the addition of doxycycline, this locus shows chromatin decondensation (Janicki et al., 2004; Shen et al., 2010). We generated triple-fusion proteins of YFP–LacI–ORCs, and these were tethered to the CLTon locus. Targeting YFP–LacI to this locus showed association of LacI with the heterochromatic CLTon locus (Fig. 1Ba). Surprisingly, tethering of YFP–LacI–Orc5 caused dramatic decondensation at the CLTon locus, whereas none of the other ORC subunits, including Orc1, Orc2, Orc3, Orc4 and Orc6, caused any changes to the chromatin architecture at the locus (Fig. 1Ba). 81% of YFP–LacI–Orc5-tethered cells showed decondensation of the heterochromatic locus (Fig. 1Bb). Furthermore, the extent of decondensation upon tethering Orc5 to the locus was determined by calculating the area of the decondensed chromatin. Measurement of the area of decondensation upon tethering Orc5 revealed a range of chromatin decondensation, ranging 2–35 µm² (Fig. 1Bc), whereas the control YFP–LacI cells showed condensed loci with sizes in the range 0.2–1.3 µm² (Fig. 1Bc). The average area of the U2OS nuclei was found to be 360±101 µm² (n=52 cells). Based on the area of decondensation, we categorized the Orc5-mediated decondensation phenotype into three categories: medium (2–6 µm²), large (6–10 µm²) and very large (10–35 µm²) (Fig. 1C). The tethering of Orc5 to the locus resulted in 37%, 34% and 29% of cells showing medium, large and very large ranges of decondensation, respectively.

We investigated the role of Orc5 in chromatin decondensation by utilizing another system, in this case a CHO-derived A03 cell line that contains 90 Mb of a homogenously staining region generated through stable integration and amplification of the LacO-DHFR vector (Li et al., 1998). Tethering Orc5 to the A03 locus also resulted in dramatic decondensation of this locus (Fig. 1D). The decondensation upon tethering of Orc5 was in the range 4.5–27 µm², whereas tethering of YFP–LacI resulted in decondensation in the range 0.6–1.2 µm².

We next determined the minimum domain of Orc5 that is required for its ability to mediate chromatin decondensation. Triple-fusion Orc5 truncation mutants (including fusions comprising amino acid residues 1–100, 101–200, 201–300, 301–400 and 301–435 of Orc5) were generated (Fig. 2A), and their ability to cause chromatin unfolding was examined (Fig. 2B). As described earlier, full-length YFP–LacI–Orc5 caused decondensation of 81% of the CLTon locus (Fig. 2C). The Orc5 truncation mutants 1–100, 101–200, 201–300 and 301–400 failed to mediate chromatin unfolding, but the N-terminal truncation 301–435 mutant resulted in chromatin unfolding in ∼40% of cells (Fig. 2B,C). The decondensation upon tethering YFP–LacI–Orc5 (301–400) was in the range 0.5–1.0 µm², whereas tethering of YFP–LacI–Orc5 (301–435aa) resulted in decondensation in the range 2.5–19.0 µm² (Fig. 2D). Our results indicate that the last 35 amino acid residues at the C-terminus of Orc5 are crucial for its chromatin decondensation function.

Upon examination of the C-terminal 400–435 residues, we found that it was enriched with acidic residues (Fig. S1A). It has previously been reported that targeting of acidic activators to heterochromatic chromatin domains can cause large-scale chromatin decondensation (Carpenter et al., 2005). We generated a mutant of Orc5 where multiple aspartic acid residues were replaced by alanine residues (Fig. S1A). However, this mutant, when tethered to the locus, showed similar levels of chromatin decondensation, suggesting that the ‘acidic domain’ within Orc5 is not required for decondensation (Fig. S1B). We next determined whether the ATP-binding ability of Orc5 is required for its chromatin-unfolding function. We generated Walker A mutant (K43A) and an arginine finger mutant (R166A) (Fig. S1A), and tethered these to the CLTon locus (Fig. S1C). The extent of chromatin decondensation upon tethering these mutants was comparable to that of the wild-type Orc5, suggesting that the ATP-binding ability of Orc5 is also dispensable for its chromatin decondensation function.

Histone acetylation, catalyzed by various HATs, is linked with the open chromatin state and is known to facilitate transcription (Narlikar et al., 2002). Because Orc5 was found to induce chromatin decondensation, we asked whether it achieves this through its association with known HATs. We transfected T7–Orc5 into U2OS cells and performed immunoprecipitation with an antibody against T7 (Fig. 3A). We found that T7–Orc5 interacted strongly with endogenous GCN5 (Fig. 3A). Next, we co-transfected T7–Orc5 and Flag–GCN5, and performed immunoprecipitation with an antibody against Flag. Orc5 and GCN5 were found to interact with one another (Fig. 3B). A reverse immunoprecipitation experiment recapitulated the interaction (Fig. 3E, full-length T7–Orc5 lane). Further, tethering of GCN5 to the CLTon locus also showed strong recruitment of Orc5 to the site, corroborating the immunoblot results (Fig. 3C).

We next mapped the region within Orc5 that associates with GCN5. We generated several truncation mutants of Orc5 (Fig. 3D) and co-transfected each of these with Flag–GCN5. Immunoprecipitation with the antibody against T7 revealed that full-length Orc5 and the 100–435 fragment efficiently associated with GCN5 (Fig. 3E), suggesting that the first 100 amino acids within the N-terminus of Orc5 are dispensable for GCN5 binding. Orc2 was found to bind to full-length Orc5 and to fragments 100–435 and 300–435 (Fig. 3D,E). These results suggest that Orc2 and GCN5 could associate with Orc5 simultaneously and that the binding might not be mutually exclusive. Because the T7–Orc5 100–435 mutant could bind to GCN5, we tested the ability of this mutant to cause chromatin decondensation. We tethered it to the CLTon locus (Fig. 3Fa) and found that it could cause chromatin decondensation comparable to that of the full-length protein (Fig. 3Fb). It is interesting to note that the fragment of Orc5 comprising amino acids 300–435 did not interact with Flag–GCN5 (Fig. 3E), but did cause partial chromatin decondensation (Fig. 2B–D).

We next examined the status of various chromatin marks at the Orc5-tethered locus. Immunofluorescence analysis using an antibody against trimethylated histone 3 (H3) at residue K9 (H3K9me3) showed robust accumulation of this mark at the LacI-containing heterochromatic locus (Fig. S3A). However, upon tethering Orc5 to the locus, H3K9me3 was distinctly devoid at these sites (Fig. S3A). Because Orc5 associates with GCN5, we examined whether H3 acetylation accumulated in Orc5-tethered cells. We did not observe robust accumulation of acetylated H3 marks in the highly decondensed Orc5-tethered cells; however, the control cells were clearly devoid of this mark at the condensed CLTon locus (Fig. S3Ba–Bc).

We then evaluated whether GCN5 could itself cause chromatin decondensation in our CLTon assay. YFP–LacI–GCN5 was tethered to the locus, and the status of chromatin architecture at the CLTon locus was evaluated (Fig. 4Aa). GCN5 could also mediate chromatin decondensation; however, this decondensation looked visually different from that observed for Orc5. Tethering YFP–LacI–Orc5 caused considerable large-scale decondensation, whereas tethering YFP–LacI–GCN5 caused a smaller scale ‘puffy’ chromatin appearance, but still significant decondensation when compared to that of YFP–LacI control (Fig. 4Aa). Although the range of chromatin decondensation upon tethering YFP–LacI–Orc5 varied from 2 to 35 µm², the extent was much smaller for YFP–LacI–GCN5, which showed decondensation in the range 3–9.5 µm² (Fig. 4Ab).

To gain an insight into the functional relevance of the Orc5 and GCN5 interaction, we evaluated whether GCN5 is required for the Orc5-mediated chromatin decondensation. We depleted GCN5 using small interfering (si)RNA in CLTon cells (Fig. 4B) and examined whether tethering of Orc5 could still induce chromatin decondensation. Remarkably, in GCN5-depleted cells, we observed a significant decrease (32% decrease, **P<0.01, Student's t-test) in the extent of Orc5-mediated chromatin decondensation (Fig. 4C–E). In control cells, 84.3% of the cells showed chromatin decondensation upon tethering Orc5. By contrast, this number reduced to 52% upon loss of GCN5 (Fig. 4C). To better understand this reduction in Orc5-mediated chromatin decondensation, we scored CLTon cells based on medium or large decondensation of the heterochromatic locus. Upon loss of GCN5, there was a striking reduction in the percentage of cells showing large decondensation (from 44% in control cells to 16% in GCN5-knockdown cells) (Fig. 4D). This result was corroborated by examining the area of the loci in control and GCN5-depleted cells. The area of the CLTon locus varied from 2 to 35 µm² in control cells and 1 to 7 µm² in GCN5-depleted cells (Fig. 4E).

We next determined whether the Orc5-mediated chromatin decondensation occurred throughout the cell cycle or was restricted to a specific stage of the cell cycle. To investigate this, we tethered YFP–LacI–Orc5 to the CLTon locus and performed double fluorescence labeling of MCM (through labeling of MCM3) and PCNA. MCM and PCNA show distinct temporal patterns of chromatin loading (Prasanth et al., 2004a). MCM, the replication helicase, loads onto chromatin in G1, shows spatio-temporal patterns during S phase and is gradually lost by the end of S phase, whereas PCNA, the clamp, is associated with chromatin only during S phase. Based on the patterns of MCM and PCNA loading, it is possible to discern whether the cell is in G1 or S phase of the cell cycle (G1 cells are MCM+ and PCNA−, whereas S-phase cells are PCNA+). We observed that the YFP–LacI–Orc5-expressing cells with the most dramatic chromatin decondensation at the CLTon locus were positive for MCM staining (Fig. 5Aa). At the same time, YFP–LacI–Orc5-tethered cells in S phase (PCNA+ cells) showed a lower degree of chromatin decondensation (Fig. 5Ab, inset 3). The most dramatic decondensation of the CLTon loci happened in MCM+ PCNA− cells (Fig. 5Ab, inset1 and 2, and Fig. 5Ac). These results suggest that the Orc5-mediated decondensation occurs predominantly during G1 phase of the cell cycle.

Because we observed a decrease in the extent of Orc5-mediated chromatin decondensation in cells that lacked GCN5, we examined if this was due to an effect of cell cycle perturbation. We evaluated the cell cycle profile of control and GCN5-depleted cells with flow cytometry (Fig. S2Aa,Ab) and observed a marginal increase in the G1 population of cells upon loss of GCN5. Because the Orc5-mediated chromatin decondensation is most dramatic in G1, the loss of the decondensation in the absence of GCN5 is not due to perturbation of cell cycle progression. We propose that GCN5 and Orc5 cooperate in mediating this chromatin decondensation.

Chromatin changes are related to the efficiency of DNA replication. We therefore profiled the replication timing of the locus upon tethering YFP–LacI–Orc5. To determine the stage of S phase during which the locus replicates, we performed immunofluorescence analysis of PCNA. PCNA shows distinct localization patterns during S phase, and based on the PCNA patterns, S phase can be categorized into early, mid or late stages (O'Keefe et al., 1992). We tethered YFP–LacI or YFP–LacI–Orc5 to the locus and examined the accumulation of PCNA at the locus (Fig. 5Ba,Bb, Fig. S4). The replication timing of the CLTon locus was discerned by examining the PCNA pattern of the cell nucleus. Upon tethering YFP–LacI, the locus did not show any accumulation of PCNA in early S-phase (Fig. 5Ba). In late S-phase, we could observe robust accumulation of PCNA at the locus, indicating that the YFP–LacI-labeled locus replicated in late S-phase (Fig. 5Ba,Bb). This wasn't surprising owing to the inherent heterochromatic nature of the locus. Next, we tethered YFP–LacI–Orc5 to the locus and examined the replication timing of the locus (Fig. 5Bb; Fig. S4). We could observe robust accumulation of PCNA at the locus preferentially in early S-phase and not in late S-phase, indicating that the replication timing of the locus shifted to early S-phase upon tethering of Orc5. Note that even the few dramatically decondensed Orc5-tethered loci that we observed during S phase replicated early (Fig. S4).

Histone acetylation is required for origin activation during S phase (Unnikrishnan et al., 2010), and it has also been shown to regulate the timing of replication origin firing (Vogelauer et al., 2002). We asked whether Orc5-mediated chromatin-opening facilitates histone acetylation in cooperation with GCN5 at the origins of replication. We conducted acetylated H3 chromatin immunoprecipitation (ChIP) in control and Orc5-depleted cells (Fig. 6A), and found a significant reduction in the acetylation of H3 at select origins – including those of lamin B1, MCM5 and lamin B2 – but not at distal sites (either upstream or downstream) of the specific origins (Fig. 6Ba–Bd) or other origins – including those of AKNA, TRPM1, MCM6 and MCM3 (data not shown). We next wanted to determine whether the reduced acetylation of origins causes inefficient origin firing and replication defects in S phase. Therefore, we depleted Orc5 and performed double immunofluorescence staining of MCM and PCNA to examine the cell cycle profile upon loss of Orc5 (Fig. 6C). We observed a 21% reduction in S-phase cells (PCNA+ cells) from 58 to 37%, and a concomitant increase in G1 (MCM+ PCNA− cells) from 28% to 46% upon loss of Orc5, indicating that cells accumulated in G1 in the absence of Orc5.

Our results suggest that Orc5 might make the chromatin more accessible for the establishment of preRCs during G1 at specific origins (Fig. 6D), and this in turn is important for efficient origin firing and DNA replication in S phase. Our results provide new insights into the specific role of one of the ORC subunits in facilitating chromatin opening for the establishment of preRCs.

Replication of DNA occurs once and only once per cell division cycle. The licensing of replication origins requires the sequential binding of the ORC, Cdc6 and Cdt1 that is then needed for the loading of the Mcm2-7 complex onto the chromatin (Bell and Dutta, 2002). ORC, comprising six subunits, serves as the landing pad for the assembly of this multiprotein complex at the origins of replication. The contribution of individual subunits in this process remains to be understood. We demonstrate that the Orc5 subunit is unique, and when tethered ectopically to a transgene array, induces large-scale chromatin decondensation. It associates with the HAT GCN5, which acetylates H3. The ability of Orc5 to cause chromatin decondensation requires GCN5 (Fig. 6D), and the loss of Orc5 causes decreased acetylation at specific origins. Our data suggest that there could be multiple mechanisms for the chromatin decondensation that is mediated by Orc5 during G1, a GCN5-dependent mechanism involving residues 100–200 of Orc5 and a GCN5-independent mechanism involving the last 35 amino acids of Orc5. Once the chromatin is decondensed during G1, it facilitates origin licensing, and it is likely that Orc5 and GCN5 cooperate at select origins and that Orc5 cooperates with other ‘unknown factors’ at other origins.

GCN5 is a global regulator of gene expression (Baker and Grant, 2007; Robert et al., 2004). More recently, GCN5 has been found to associate with yeast origins, albeit weakly (Espinosa et al., 2010), and positively regulates DNA replication by counteracting the inhibitory effects of HDACs. In addition, GCN5-mediated acetylation of H3 lysine residues has also been proposed to function in replication-coupled nucleosome assembly (Burgess et al., 2010). We propose that Orc5 at origins helps the recruitment of GCN5 to these sites and that this in turn facilitates the opening of chromatin, thus enabling the loading of other preRC components.

The acetylation of histones H3 and H4 is known to be dynamically regulated around the origins of replication that facilitate origin firing (Unnikrishnan et al., 2010). Furthermore, acetylation of H4 at residue K79 is enriched at origins; acetylation of H4 at residue K16 is enriched at early-firing origins and also limits the spread of heterochromatin at origins. Acetylation of H3 at residues K9 or K14, and of H4 at residues K5, K8 or K12 is enriched at active origins, and these modifications are believed to promote firing (Dorn and Cook, 2011).

Hbo1 is an H4-specific HAT that interacts with the human ORC and MCM proteins (Iizuka and Stillman, 1999). It is required for replication licensing (Iizuka et al., 2006) and is known to associate with replication origins (Miotto and Struhl, 2008). It has recently been demonstrated that Hbo1-mediated hyperacetylation of H4 increases Mcm2-7 loading onto chromatin (Miotto and Struhl, 2010). It is generally believed that histone acetylation loosens up the compacted chromatin, thereby increasing the accessibility for loading of the MCM helicase complex, or acts as a molecular tag to which the helicase is tethered (Chadha and Blow, 2010). Hbo1 association to origins is dependent on Cdt1, and it has been shown that Cdt1 can modulate chromatin accessibility through temporal recruitment of Hbo1 to origins (Miotto and Struhl, 2008). Interestingly, ectopic tethering of Cdt1 is also known to induce large-scale chromatin unfolding at a transgene array (Wong et al., 2010). It is well established that the origins comprise open chromatin during G1 and then become less accessible as cells exit out of G1 phase (Djeliova et al., 2002; Pemov et al., 1998). The replication origins are known to exhibit temporal dynamics in chromatin structure, with a highly open structure during G1 and more closed architecture during S phase. Elegant work, using a quantitative-PCR-based approach on DNase-1-treated chromatin samples, has shown that endogenous replication origins, including those of MCM4 and lamin, display a more open chromatin structure during G1 than in S phase (Wong et al., 2010).

In addition, tethering of replication protein Cdc45 leads to chromatin decondensation in a Cdk2-dependent manner (Alexandrow and Hamlin, 2005). Such decondensation is mediated by phosphorylation of H1. The phosphorylation of H1, in turn, is Cdk2-dependent and leads to chromatin decondensation in S phase. These observations point towards a role for Cdc45 in replication fork progression by regulating large-scale chromatin changes (Alexandrow and Hamlin, 2005). Our observations of the CLTon locus – which reveal that Orc5 can efficiently cause chromatin decondensation, predominantly during the G1 phase of the cell cycle – provide a strong indication that similar events occur at endogenous origins, whereby Orc5 might cause local chromatin changes in collaboration with GCN5. Another piece of evidence to support this idea comes from a recent study that shows preferential association of Orc5 with acetylated H3 at K27 (Ji et al., 2015) as compared to several other markers of activation, such as trimethylated H3 at K4, dimethylated H3 at K79 and trimethylated H3 at K36. We propose that at an endogenous locus at which ORC binds, the amount of Orc5 present would be sufficient for a smaller scale of decondensation, which might not be observable at the resolution of light microscopy. But that decondensation would be sufficient for the cognate cellular function, such as replication initiation.

Orc5 is one of the ORC subunits, and ATP binding to Orc5 is involved in efficient ORC formation (Siddiqui and Stillman, 2007; Takahashi et al., 2004). Orc5 has also been implicated in silencing at the HML and MHR loci in yeast and in heterochromatin organization in human cells (Dillin and Rine, 1997; Prasanth et al., 2010). The gene encoding Orc5 maps to chromosome 7q22 and is frequently deleted in adult acute myeloid leukemia, myelodysplastic syndrome, uterine leiomyomas and malignant myeloid diseases (Fröhling et al., 2001; Quintana et al., 1998). We have observed dose-dependent chromatin decondensation at the CLTon locus, and this was directly correlated with the expression level of Orc5. This finding is supported by the fact that tethering of Orc5 to the CLTon locus can cause dramatic chromatin decondensation. However, tethering of other ORC subunits does not, despite the fact that Orc5 can be recruited to the locus by other ORC subunits. Our data imply that excessive levels of Orc5 in the cell could result in aberrant chromatin decondensation and cause genomic instability.

pEGFP-LacI vector was a kind gift from Dr Miroslav Dundr (Rosalind Franklin University of Medicine and Science, North Chicago, IL) (Kaiser et al., 2008) and used to generate the pEYFP-LacI vector. YFP-LacI-Orc1, -Orc2, -Orc3, -Orc4, -Orc5 and -Orc6 were cloned by amplifying and inserting the ORC proteins into the pEYFP-LacI vector. YFP–LacI–Orc5 D414A,D426A,D433A; YFP–LacI–Orc5 K43A; YFP–LacI–Orc5 R166A were generated using site-directed mutagenesis (Stratagene) with YFP–LacI–Orc5 wild type as template. T7–Orc5 full length and truncations were cloned by amplifying and inserted into pCGT vector. Flag–GCN5 was a kind gift from Brian Freeman (University of Illinois at Urbana–Champaign, Champaign, IL).

U2OS osteosarcoma cells were grown in high glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Hyclone). U2OS-2-6-3 CLTon cells were cultured in DMEM supplemented with 10% Tet-system-approved FBS (Clonetech).

For visualizing YFP–LacI-tagged proteins, cells were fixed with 2% formaldehyde in PBS (pH 7.4) for 15 min at room temperature, followed by permeabilization with 0.5% Triton X-100 in PBS for 7 min on ice and by blocking. For immunofluorescent labeling of H3K9me3, acetylated H3, MCM3 and PCNA, cells were pre-extracted before fixing with 0.5% Triton X-100 in cytoskeletal buffer (CSK; 100 mM NaCl, 300 mM sucrose, 3 mM MgCl₂, 10 mM PIPES at pH 6.8) for 5 min on ice followed by fixing with 1% formaldehyde in PBS (pH 7.4) for 5 min at room temperature and then blocking. This was followed by blocking for 30 min with 1% normal goat serum in PBS. After that, cells were incubated for 1 h with primary antibody in a humidified chamber, followed by incubation with secondary antibody for 25 min. Nuclei were then stained with DAPI and mounted using Vectashield (Vector Laboratories). The following antibodies were used for immunofluorescence: anti-H3K9me3 (1:200, Millipore 07-523), anti-acetylated-H3 (1:500, Millipore 06-599), anti-MCM3 (1:300, 738 rabbit polyclonal antibody) and anti-PCNA mAbPC10 (1:150, PC10 monoclonal antibody).

For observing cells and obtaining statistics (area of decondensation, and PCNA and MCM3 staining), we used a Zeiss Axioimager z1 fluorescence microscope (Carl Zeiss Inc.) equipped with chroma filters (Chroma Technology). The images were acquired using a 60×, 1.4NA oil immersion objective. Digital imaging was performed using an Hamamatsu ORCA cooled CCD camera. Axiovision software (Zeiss) was used for processing the images.

A delta vision optical sectioning deconvolution instrument (Applied precision) on an Olympus microscope was also used for observing cells and for obtaining the images used in the manuscript. The images were acquired using a 60×, 1.4NA oil immersion objective. softWoRx imaging workstation was used to calculate the area of the decondensed locus. The ‘Edit Polygon’ tool was used as it is suitable for calculating irregular-area statistics. The extent of overlap of H3K9me3 with YFP–LacI or YFP–LacI–Orc5 was measured by using the ‘Colocalization’ tool and defining a region of interest (ROI); the Pearson coefficient of colocalization was then determined. Line profiles for H3K9me3 and acetylated H3 immunofluorescence were generated by using the ‘Line Profile’ tool with the band value set to 2.

U2OS cells were grown to 30% confluence, followed by two rounds of knockdown with a control siRNA against the luciferase gene (Shen et al., 2010) or with an siRNA against Orc5 (5ʹ-UGACUUUGUUCGCUUGUUUUU-3ʹ) (Prasanth et al., 2010), 24 h apart at a final concentration of 100 nM using Lipofectamine RNAiMax (Invitrogen). This was followed by collection of cells for ChIP analysis 24 h after the second knockdown.

For GCN5 knockdown, CLTon cells were grown to 30% confluence on coverslips followed by two rounds of knockdown with a control siRNA against the luciferase gene (Shen et al., 2010) or with an siRNA against GCN5 (Palhan et al., 2005), 24 h apart at a final concentration of 40 nM. YFP–LacI–Orc5 was transfected into cells while carrying out the second round of knockdown, and the cells were fixed for microscopy analysis 24 h later. The first round of knockdown was performed using Lipofectamine RNAimax (Invitrogen), and the second round of knockdown with transfection of the YFP-tagged constructs was performed using Lipofectamine 2000 (Invitrogen).

For co-immunoprecipitation experiments, co-transfections were performed in U2OS cells. Flag–HATs and T7–Orc5 were transiently transfected, and cells were lysed, 24 h post-transfection, in immunoprecipitation buffer (50 mM HEPES pH 7.9, 10% Glycerol, 200 mM NaCl, 0.1% Triton X-100, 1 mM CaCl₂) supplemented with the protease and phosphatase inhibitors. 4U of MNase (Sigma) was then added per 10-cm plate of sample, followed by rocking at room temperature for 20 min. EDTA (final concentration 5 mM) was added to stop the reaction, and samples were centrifuged at 12,500 rpm (Eppendorf centrifuge 5424), for 5 min at 4°C. The supernatant was pre-cleared. Pre-clearing was performed using Gammabind Sepharose beads for 1 h, and the lysates were incubated with appropriate antibody overnight. For pulldown of the antibody-bound complexes, agarose beads were washed in the same immunoprecipitation buffer and were incubated with lysate containing antibodies for 2 h. This was followed by three washes of the pulled-down complexes, and finally denaturation of the pulled-down proteins by the addition of Laemmli buffer and incubation in a heat block for 10 min. The complexes were then analyzed by western blotting.

For semi-endogenous immunoprecipitation, the cells were initially lysed in IP-100 buffer (50 mM Tris pH 7.5, 100 mM NaCl, 1 mM MgCl₂, 0.2% NP-40, 1% glycerol supplemented with protease and phosphatase inhibitors). The lysate was then sonicated for 2 min with Bioruptor Power-up (Diagenode) using 30-s-on and 30-s-off cycles. Benzonase (250 U/10⁷ cells) was then added to the lysate, and the sample was incubated at room temperature for 30 min with rotation. The reaction was stopped by adding EDTA (final concentration 5 mM), and the salt concentration was raised to 200 mM with NaCl. The sample was then pre-cleared, and subsequent steps of immunoprecipitation were performed as explained in the previous paragraph. All subsequent washes were performed with IP-200 buffer (50 mM Tris pH 7.5, 200 mM NaCl, 1 mM MgCl₂, 0.2% NP-40, 1% glycerol supplemented with protease and phosphatase inhibitors).

For immunoprecipitations and immunoblotting, the following antibodies were used: anti-Flag M2 (1:500, Sigma), anti-T7 (1:5000, Novagen), anti-ORC2 polyclonal antibody 205-6 (1:1000) and anti-GCN5 (1:200, Cell Signaling).

Control and Orc5-knockdown cells were crosslinked for 10 min at room temperature through addition of formaldehyde (Sigma) to culture medium to a final concentration of 1%. The reaction was stopped by addition of glycine at a final concentration of 0.125 M. Fixed cells were then washed twice (quickly) in chilled PBS and harvested with PBS. Chromatin was prepared using two subsequent extraction steps (10 min at 4°C) with Buffer 1 (50 mM HEPES with KOH pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 0.25% Triton X-100) and Buffer 2 (200 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 10 mM Tris-HCl pH 8). Nuclei were then pelleted by centrifugation, resuspended in SDS lysis buffer (50 mM Tris-HCl pH 8, 0.1% SDS, 1% NP-40, 0.1% sodium deoxycholate, 10 mM EDTA, 150 mM NaCl) and subjected to sonication with a Bioruptor Power-up instrument (Diagenode) for 45 min (three times of 15 min of sonication, each cycle comprising 30-s on and 30-s off). This yielded genomic DNA fragments that were 300 bp in length. The obtained chromatin was then precleared with Protein A/G ultralink beads (53133, Pierce) for 1 h at 4°C, and immunoprecipitation with the anti-acetylated-H3 antibody and rabbit IgG was performed overnight at 4°C. Immune complexes were pulled down by adding pre-blocked protein A/G ultralink beads, and the mixture was incubated for 2 h at room temperature. Beads with immune complexes bound to them were washed once with low-salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8, 150 mM NaCl), once with high-salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8, 500 mM NaCl), once with LiCl wash buffer (10 mM Tris-HCl pH 8.0, 1% sodium deoxycholate, 1% NP-40, 250 mM LiCl, 1 mM EDTA) and twice with Tris-EDTA. Beads were then eluted twice (10 min each) in Tris-EDTA+1% SDS+0.1% NaHCO₃ at 65°C, and the cross-links were then reversed overnight at 65°C. DNA from the eluted material was then treated with RNase (10 µg/ml) for 1 h at 37°C followed by 2 h of treatment with proteinase K (4 μl 0.5 M EDTA, 8 μl 1 M Tris-HCl pH 6.9, 1 μl proteinase K 20 mg/ml) at 42°C. DNA was isolated by using the QIAquick PCR purification kit (Qiagen) resuspended in elution buffer, and quantitative PCR was performed using PowerSYBR Green PCR Master mix (Applied Biosystems) and analyzed on a 7300 PCR System (Applied Biosystems). ChIP–quantitative-PCR results were analyzed and plotted as percentages (%) of immunoprecipitation/input signal (% input).

The primer sequences of the regions analyzed were as follows: lamin B1 origin +4.2 Kb, forward 5ʹ-CCTCCCAAAGTGCTAGGATTAC-3ʹ, reverse 5ʹ-CCTAAGGCCCAACCAAGAAT-3ʹ; lamin B1 origin forward 5ʹ-AAGATTCTACGCCTGCTTTAGG-3ʹ, reverse 5ʹ-AGCTAACCAGCTAGCAAACC-3ʹ; MCM5 origin −4 Kb forward 5ʹ-TAACCTCTCTGAGGCTCCAT-3ʹ, reverse 5ʹ-GTGTATGGCGTTGGCATTTC-3ʹ; MCM5 origin forward 5ʹ-CTAGGTGATTGTACGACCTGTG-3ʹ, reverse 5ʹ-CCTGCAACCTCATGTTCTCT-3ʹ; lamin B2 origin −2.2 Kb forward 5ʹ-GAGAATCACTTGAACCCAGGAG-3ʹ, reverse 5ʹ-CCACTAGCTGCCAAGCTTAT-3ʹ; lamin B2 origin forward 5ʹ-TGAGGAGTCCCTCAGATCTTTA-3ʹ, reverse 5ʹ-CAGAATCCGATCATGCACCT-3ʹ; MCM4−5 Kb forward 5ʹ-TTCACATCCACCCAGCTTATC-3ʹ, reverse 5ʹ-AGAGCATTCTTCCCCTGATG-3ʹ; MCM4 origin forward 5ʹ-TTGGGTGGCTACTTGGTGTT-3ʹ, reverse 5ʹ-TAGGCCCCTCGCTTGTTT-3ʹ.

We thank members of the Prasanth laboratory for discussions and suggestions. We thank Drs B. Freeman, C. Mizzen (phosphorylated histone H1 antibodies, University of Illinois at Urbana–Champaign, Champaign, IL), M. Dundr, D. Spector (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY) and B. Stillman (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY) for providing reagents and suggestions. We would like to thank Dr D. Rivier, K. Dovalovsky and A. Matur for critical reading of the paper.