Two subunits of human ORC are dispensable for DNA replication and proliferation

The six-subunit Origin Recognition Complex (ORC) is believed to be an essential eukaryotic ATPase that binds to origins of replication as a ring-shaped heterohexamer to load MCM2-7 and initiate DNA replication. We have discovered that human cell lines in culture proliferate with intact chromosomal origins of replication after disruption of both alleles of ORC2 or of the ATPase subunit, ORC1. The ORC1 or ORC2-depleted cells replicate with decreased chromatin loading of MCM2-7 and become critically dependent on another ATPase, CDC6, for survival and DNA replication. Thus, either the ORC ring lacking a subunit, even its ATPase subunit, can load enough MCM2-7 in partnership with CDC6 to initiate DNA replication, or cells have an ORC-independent, CDC6-dependent mechanism to load MCM2-7 on origins of replication DOI: http://dx.doi.org/10.7554/eLife.19084.001


Introduction
The discovery of the six-subunit ORC (Bell and Stillman, 1992) identified the long sought initiator protein that binds to replicator sequences to initiate DNA replication in eukaryotes. ORC is an essential six-subunit, ring-shaped ATPase complex that recruits and co-operates with the CDC6 protein to promote the loading of CDT1 and then the MCM2-7 subunits of the replicative helicase during the 'licensing' of origins of replication (Bleichert et al., 2015;Yeeles et al., 2015;Bell and Stillman, 1992;Blow and Tada, 2000;Masai et al., 2010). Of the six subunits of human ORC, ORC2-5 form a tight core complex (Dhar et al., 2001). ORC1 is the only subunit responsible for the ATPase activity of ORC (Chesnokov et al., 2001;Giordano-Coltart et al., 2005;Klemm et al., 1997) All six ORC subunit genes are essential for the viability of S. cerevisiae and S. pombe (Bell et al., 1993;Foss et al., 1993;Herskowitz, 1993;Loo et al., 1995;Micklem et al., 1993). It is therefore expected that eukaryotic cells will not be viable, and will not replicate, if any of the ORC subunit genes are deleted. However, we have now deleted both alleles of human ORC2 or of ORC1, to discover that cells can still survive and replicate in the complete absence of either of these two critical subunits of Origin Recognition Complex.

Results
Biallelic disruption of the ORC2 gene CRISPR/Cas9 was used to insert a~600 bp blasticidin gene and poly A site into exon 4 (amino acid 40) of ORC2 in HCT116 p53-/- (Bunz et al., 1998) colon cancer cells (WT: HCT116 p53-/-, ORC2+/+) ( Figure 1A and B). All clones were viable and proliferated for months despite the disruption of the ORC2 gene. Immunoblotting and immunoprecipitation-immunoblotting of cell extracts showed that the ORC2-/-B2 and BP8 clones had no detectable ORC2 protein ( Figure 1D-H). An anti-ORC2C antibody recognizing the C terminal half of ORC2 ( Figure 1C) ensured that a C-terminal fragment of ORC2 was not being expressed from an alternatively spliced transcript using an internal methionine ( Figure 1E). Quantitative immunoblotting showed that if any ORC2 was expressed, it must be at a level <1% of wild type levels ( Figure 1D). Quantitative immunoblotting of cell lysates and carefully measured amounts of recombinant bacterially produced ORC2 fragment showed that wild type cells express~153,000 ORC2 molecules per cell ( Figure 1I,J). Thus, if the B2 or BP8 clones contain any ORC2 molecules below the level of detection, they can have no more than 1530 molecules/cell. ORC2 was also deleted in 293T human embryonic kidney cells or HBEC human bronchial epithelial cells immortalized with CDK4 and hTERT, and these cells too continued proliferating in the absence of ORC2 protein ( Figure 1K).

CDC6, CDT1 and MCM2-7 loading on chromatin in the absence of ORC2
The ORC2-/-cells suffer a decrease of ORC3, ORC4 and ORC5, which are destabilized when not complexed with ORC2 ( Figure 1G). ORC1 was also decreased, but ORC6, CDC6 and CDT1 were unchanged. There was no activation of the DNA damage checkpoint, measured by the phosphorylation of Chk1 or H2AX, as would have occurred with impaired DNA replication. Chromatin association of ORC2-5 and of ORC1 was decreased in ORC2-/-cells (as expected from the decrease of these proteins in cell lysates) ( Figure 1H). The chromatin association of MCM3, 5 and 7 of the MCM2-7 was reduced, but not completely eliminated. Surprisingly, chromatin association of ORC6 or CDT1 was relatively unchanged, while CDC6 association was slightly increased.

Initiation of DNA replication in ORC2-/-cells
The proliferation rate of the ORC2-/-cells was >50% that of WT cells ( Figure 2A) and ORC2 did not re-appear even after passage of the cells for over a year. The ORC2-/-cells did not accumulate in S phase ( Figure 2B) and completed DNA replication after release from an early S-phase block in the same time as WT cells ( Figure 2C). The percentage of cells synthesizing DNA during a 30 min pulse was not significantly decreased ( Figure 2D). By molecular combing the median distance between bidirectional origins of replication in ORC2-/-cells was marginally increased to 113-118 kb from 96 kb and the fork progression rate was slightly increased to 1.3-1.5 kb/min from 1.2 kb/min ( Figure 2E). Given the total DNA content of six billion bp in these cells, this measurement suggests that the ORC2-/-cells fire about 52,000 origins of replication. eLife digest Most of the DNA in human cells is packaged into structures called chromosomes.
Before a cell divides, the DNA in each chromosome is carefully copied. This process begins at multiple sites (known as origins) on each chromosome. A group of six proteins collectively known as the Origin Recognition Complex (or ORC for short) binds to an origin and then recruits several additional proteins. When the cell is ready, the assembled proteins are activated and DNA copying begins. It is thought that all of the ORC proteins are essential for cells to survive and copy their DNA.
Here, Shibata et al. reveal that human cells can survive without ORC1 or ORC2, two of the six proteins in the ORC complex. Disrupting the genes that encode the ORC1 and ORC2 proteins in human cancer cell lines had little effect on the ability of the cells to copy their DNA and survive. Furthermore, these cells spend the same amount of time copying their DNA and use a similar set of origins as normal cells.
However, the experiments also reveal that cells without ORC1 or ORC2 are more dependent on the presence of one particular protein recruited to the origin after the ORC assembles. Reducing the availability of this protein, CDC6, decreased the ability of these cells to survive and divide. Future efforts will aim to identify the mechanism by which cells bring together the proteins required to copy DNA in the absence of a complete ORC. We mapped replication initiation sites by a second method: enriching for BrdU-labeled, origincentered nascent strands by immunoprecipitation and sequencing (BrIPseq) (Karnani et al., 2010).~13,000 BrIPseq origins were mapped to unique DNA sequence in the ORC2-/-cells compared to~20,000 in WT cells ( Figure 2F). Taking into consideration repetitive DNA and the diploid nature of the human genome, this method also suggests that about 52,000 origins are fired in the ORC2-/-cells. 40% of the BrIP-seq peaks in ORC2-/-cells overlapped with that in WT cells, comparable to the overlap reported among origins mapped by different groups (Karnani et al., 2010). The <100% of overlap is explained by plasticity of origin usage (Cadoret et al., 2008). The inter-origin distances for the BrIPseq origins in ORC2-/-cells (26 kb) were slightly longer than in the WT (25 kb) cells (P value~4.3e-07, Wilcoxon rank sum test) and the chromosome-by-chromosome distribution of inter-origin distances was similar to that in WT cells ( Figure 2G). The 4-fold shorter inter-origin distance in a population-based approach of origin mapping (BrIPseq) compared to molecular combing is also due to the plasticity of origin usage: only one out of four possible origins in a single DNA segment fire in one S phase, while all four origins are used in the entire population of cells (see discussion in Karnani et al., 2010). Origins are enriched in gene-rich domains and near transcription start sites ( Figure 2H and I) as reported in previous studies (Karnani et al., 2010;Cadoret et al., 2008;Danis et al., 2004;MacAlpine et al., 2004;Mesner et al., 2011;Sequeira-Mendes et al., 2009).

Survival of ORC1-/-cells
We next disrupted the only subunit of ORC demonstrated to have ATPase activity, ORC1 (Chesnokov et al., 2001;Giordano-Coltart et al., 2005;Klemm et al., 1997). A blasticidin or a plasmid-derived DNA fragment was inserted in exon 1 of ORC1 at the initiator methionine ( Figure 3A,B). Western blotting and immunoprecipitation-western blotting of ORC1 protein showed no ORC1 protein in the ORC1-/-clones ( Figure 3C-E). ORC1 is loosely associated with the other subunits of ORC, and unlike ORC2, ORC1 depletion did not decrease ORC2, ORC3, ORC4, ORC5, ORC6, CDT1, and CDC6 ( Figure 3C). Although ORC1 was absent, the remaining five subunits of ORC along with CDC6 and CDT1 (at least in one clone) could associate with chromatin. However, the chromatin loading of MCM3, 5, and 7 subunits of MCM2-7 is significantly decreased ( Figure 3D).

DNA replication in ORC1-/-clones
At early passage (at one month), the ORC1-/-clones proliferated at 10-20% of the rate of WT cells but by 6 months of passage, their proliferation rates at about 50% of WT cells without reappearance of ORC1 protein ( Figures 4A and 3E). The cells did not accumulate in S phase ( Figure 4B) and passage through S phase was not slowed ( Figure 4C). By molecular combing, inter origin distance in ORC1-/-cells was unchanged or slightly decreased ( Figure 4D). Fork progression was decreased to 0.8-1.0 kb/min from 1.2 kb /min. BrIPseq analysis showed that the ORC1-/-cells fire~13,000 unique origins, with 43% of them overlapping with origins in WT cells ( Figure 4E). The distribution of interorigin distances between chromosomes, enrichment of origins in gene-rich segments and near transcription start sites was similar to that of WT cells or ORC2-/-cells ( Figure 4F-H). ORC5 protein remains important for DNA replication in ORC1-/-or ORC2-/-cells To test whether any of the other subunits of the six-subunit ORC remain important for replication in the ORC1-/-or ORC2-/-cells, we depleted ORC5 by siRNA ( Figure 5D). Knockdown of ORC5 did not change the cell cycle profile in ORC-/-mutant cells ( Figure 5E), and in fact more severely repressed the % of actively replicating cells in WT cells than in the ORC1-/-or ORC2-/-cells ( Figure 5F). However knockdown of ORC5 still inhibited BrdU incorporation in the ORC1-/-and ORC2-/-cells. We have not yet succeeded in knocking out the ORC4 and ORC5 genes, though of course such clones may emerge in future attempts. At the present state, our results suggest that the other subunits of ORC may support DNA replication independent of ORC1 or ORC2.

Discussion
The surprising results from this genetic investigation of ORC in human cell lines suggest that two subunits of ORC are dispensable for DNA replication. Although we believe that no ORC2 protein is synthesized in the ORC2-/-cells, we have to entertain the caveat that up to 1530 ORC2 molecules can escape the limits of detection with our current antibodies. This number is still too few to license the 52,000 origins of replication mapped by two independent methods. It is unlikely that a single ORC complex can catalytically load thirty-five MCM2-7 hexamers distributed 100 kb apart over 3.5 Mb of chromosomal DNA.
Since knockdown of ORC5 still affects DNA synthesis in the ORC1-/-or ORC2-/-cells, we cannot yet conclude that replication initiation can occur in the absence of all subunits of ORC. To reach that conclusion we have to successfully delete both alleles of the other four subunits, ORC3-6. However, even if the remaining ORC subunits are functional for loading MCM2-7 our result requires reconsideration of existing models of ORC function in replication initiation. (1) The crystal structure suggests that ORC2-3-5-4-1 are arranged in a gapped ring (with a central channel that is wide enough to surround a DNA double-helix), and that later in licensing, CDC6 slips into the gap between ORC2 and ORC1 to close the gap. The ORC-CDC6 ring is proposed to interact with the MCM2-7 ring end-onend during the loading of MCM2-7 (Bleichert et al., 2015). Loss of one subunit in the five- membered ORC ring makes it difficult for the remaining subunits to form a ring large enough (i) to surround a DNA double-helix in the same manner as wild type ORC or (ii) to interact with the MCM2-7 ring end-on-end. (2) Human ORC1 and ORC4 are the only subunits that have intact Walker A and B motifs. Multiple groups have shown that the ATPase activity of ORC (in S. cerevisiae, in D. melanogaster and in H. sapiens) depends exclusively on the Walker A and B motifs of the ORC1 subunit, and that this ATP binding and hydrolysis activity is essential for ORC function    (Chesnokov et al., 2001;Giordano-Coltart et al., 2005;Klemm et al., 1997). Even if an altered or partial ORC is initiating replication, we have to conclude that any ATPase activity necessary can be provided by ORC4 or CDC6.
There is a report arguing that ORC1 is not essential for endoreplication in Drosophila, because ORC1-/-larvae still allowed endoreplication in salivary gland cells, with only a twofold reduction of ploidy (Park and Asano, 2008). However the paper did not show the sensitivity of the western-blot to measure the level of residual ORC, so that it is theoretically possible that there was enough residual maternal ORC in the endoreplicating cells.
The classic model of replication initiation where ORC first associates with the DNA, helps load CDC6, which then helps load CDT1 and MCM2-7 may still be important for efficient MCM2-7 loading. The surprise is that inefficient MCM2-7 loading, perhaps with the help of CDC6, is sufficient for replication initiation and cell survival in the absence of six-subunit ORC. The other surprise is that the six-subunit ORC does not appear to associate with chromatin as a holocomplex. Clearly ORC6 associates with chromatin normally despite a decrease in ORC1 (in the ORC2-/-or ORC1-/-cells) or ORC2-5 loading (in the ORC2-/-cells). Similarly ORC2-5 loading is independent of ORC1 loading (in the ORC1-/-cells).
The origin plasticity of eukaryotes is attributed to the loading of excess subunits of MCM2-7 on chromatin. It is thus also surprising that the plasticity persists despite a significant reduction in the association of MCM2-7 on chromatin. In summary, we suggest that the absolute requirement for sixsubunit ORC for licensing bi-directional origins of replication can be bypassed in some cell lines.

Construction of vectors for homologous recombination
Blasticidin or Hygromycin resistance genes terminated by a polyA sequence and flanked by two homology arms (0.9 kb-1.6 kb in length) were cloned into pDONR 221 (Thermo Fisher Scientific) using PCR and In-Fusion cloning (Clontech).

DNA combing assay
Cells were pulse labeled with 100 mM CldU for 30 min, following by 250 mM IdU for 30 min before embedding into agarose plug. DNA was combed on silanized coverslips (Genomic Vision, Bagneux, France), dehydrated at 65˚C for 4 hr and denatured in 0.5 M NaOH and NaCl for 8 min. CldU or IdU were immune-detected with either anti-BrdU antibody that recognizes CldU (MA 182088, Thermo Fisher Scientific, Waltham, MA, RRID:AB_927214) or anti-BrdU antibody that recognizes IdU (347580, BD Biosciences, Franklin Lakes, NJ, RRID:AB_400326). Image acquisition was performed with Zeiss AxioObserver Z1, 63 X objective. DNA lengths were measured using Image J software.

BrdU incorporation
BrdU incorporation was conducted as previously described (Machida et al., 2005) with the following modifications. Cells were labeled with 10 mM BrdU for 30 min and fixed in 70% Ethanol. DNA was denatured in 2 M hydrochloric acid and stained with FITC-conjugated BrdU antibody (556028, BD Biosciences, RRID:AB_396304) and propidium iodide (Sigma-Aldrich, St.Louis) according to the manufacturer's instruction.

Clonogenic assay
To determine the effects of CDC6 knock down, cells were transfected with siRNA twice. 48 hr after the first siRNA transfection, 2000 cells were plated in six well plates. Colonies were fixed, stained with crystal violet, and counted 1 week later. All experiments were conducted in triplicate.

Proliferation assay
1000 cells were plated in 96 well plates. The absorbance of cells were measured every 24 hr using CellTiter 96 Non-Radioactive Cell Proliferation Assay (Promega, Fitchburg, WI) according to the manufacturer's instructions. All experiments were conducted in triplicate and absorbance relative to that on day one was expressed.

Br-IP Seq
The cells were labelled with 100 mM BrdU (Sigma) in exponential phase of their growth (50-60% of confluency) for 1 hr. The cells were lysed and genomic DNA was isolated, denatured and nascent strands were separated on neutral sucrose gradient. The fragments of 0.5 to 3.0 kb were selected. After dialysis against TE, the DNA was sheared, denatured and immune precipitated with anti-BrdU antibody (B8434, Sigma, RRID:AB_476811). The single stranded BrdU immunoprecipitate (10 ng) was then used to prepare next generation sequencing libraries using Takara Chip-Seq kit. The library from the control genomic DNA was prepared the same way as for the BrdU sample but the only difference was that sucrose gradient centrifugation and size selection was not done for the genomic control. Also the BrdU incorporation was for a longer time (36 hr compared to 1 hr for Br-IP sample). Single-end 75 bp reads were obtained for wildtype (WT) and ORC2 or ORC1 knockout cells. BrdU incorporated genomic strands was also sequenced and used as control (CNTL). Perl script was written and used to trim T's present at the 3' end of reads. The trimmed reads were aligned to hg38 using Bowtie2 with the default parameters. Alignment with Bowtie2 resulted in 12597288 (81%) and 10204177 (77%) and 13171609 (93%) mapped reads in WT, KO and CNTL respectively. To define peaks, the genome was divided into 1 kb windows. Any 1 kb window in WT or KO cells with two fold more reads than CNTL were considered to calculate mean and SD. 1 kb windows with reads ! mean + 3 SD number of reads were defined as peaks in WT or KO cells. Dataset for transcription start sites (TSSs) was downloaded from the UCSC genome browser. Circos v0.67 (Krzywinski et al., 2009) was used to construct circular genome visualizations. Peaks coordinates of WT and KO chromosome one was parsed to create files appropriately formatted for input to Circos.

Statistical analysis
We have used the pwr package in R for choosing sample size for all the experiments. Owing to the nature of the performed experiments, no randomization and no blinding were used. All test statistics were calculated with R (http://www.r-project.org/). t-test and Wilcoxon-rank sum test was performed to test the difference in mean for the normal and skewed data respectively.

Data deposition
Source data including BrIP seq data were deposited in Dryad (Shibata et al., 2016).