SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Proliferating cell nuclear antigen (PCNA) is loaded on chromatin upon initiation of the S phase and acts as a platform for a large number of proteins involved in chromosome duplication at the replication fork. As duplication is completed, PCNA dissociates from chromatin, and thus, chromatin-bound PCNA levels are regulated during the cell cycle. Although the mechanism of PCNA loading has been extensively investigated, the unloading mechanism has remained unclear. Here, we show that Elg1, an alternative replication factor C protein, is required for the regulation of chromatin-bound PCNA levels. When Elg1 was depleted by small interfering RNA, chromatin-bound PCNA levels were extremely increased during the S phase. The number of PCNA foci, regions in the nucleus normally representing DNA replication sites, was increased and PCNA remained on chromatin after DNA replication. Various chromatin-associated protein levels on chromatin were affected, and chromatin loop size was increased. During mitosis, cells with aberrant chromosomes and lagging chromosomes were frequently detected. Our findings suggest that Elg1 has an important role in maintaining chromosome integrity by regulating PCNA levels on chromatin, thereby acting as a PCNA unloading factor.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Genomic information is maintained by faithful replication during the S phase and segregation in mitosis (Nurse 1994). For chromosomal DNA replication to occur, replication origins are licensed for replication at the end of mitosis or early G1 phase by forming a prereplicative complex (pre-RC). This is accomplished by loading an MCM2-7 complex on the origin-bound recognition complex-bound origin with the aid of Cdc6 and Cdt1 (Bell & Dutta 2002; Nishitani & Lygerou 2002; Blow & Dutta 2005). At the onset of the S phase, replication origins are activated by CDK and DDK, which lead to the formation of an active DNA helicase comprising MCM2-7, GINS and Cdc45, called the CMG (Cdc45-MCM2-7-GINS) complex (Ilves et al. 2010; Leman & Noguchi 2013; Li & Araki 2013). After origin unwinding by the CMG complex, polymerase (Pol) α/primase primes DNA synthesis on both strands, and then, Pol ε and Pol δ elongate the leading and lagging strands, respectively. For these Pols to synthesize new strands processively and stably, they are attached and tethered to the sliding clamp, proliferating cell nuclear antigen (PCNA). Pol α/primase synthesizes the primer, and then, PCNA is loaded to the end of the primer by a protein complex known as the clamp loader, the replication factor C (RFC) complex. RFC is a heteropentameric complex composed of a largest subunit RFC1 and four small subunits, RFC2 to RFC5, called RFC1-RFC (Waga & Stillman 1998). RFC1-RFC-dependent loading of PCNA onto the primer terminus is required at every Okazaki fragment on the lagging strand. Synthesis of the Okazaki fragments is terminated when Pol δ meets the 5′ end of the RNA portion of the previously synthesized fragment, and two PCNA-binding proteins, Fen1 and LigI, remove the RNA part, fill the gap and ligate the two adjacent fragments (Stillman 2008; Burgers 2009; Leman & Noguchi 2013). PCNA is required not only for DNA replication itself, but also for replication-linked processes, chromatin assembly and remodeling, epigenetic inheritance, sister chromatid cohesion, etc., by recruiting and orchestrating the crucial players (Moldovan et al. 2007; Alabert & Groth 2012).

Recent findings indicate that PCNA has another role in preventing re-replication of chromosomal DNA. Cdt1 is an important licensing factor required for MCM2-7 loading onto replication origins. Once DNA replication has started, Cdt1 is inactivated by proteolysis and/or geminin binding, which are important for preventing the re-licensing of replicated chromosomal DNA (Machida et al. 2005; Arias & Walter 2007; Abbas & Dutta 2011; Havens & Walter 2011). Cdt1 has a PCNA interaction protein motif (PIP-box) at its amino terminal end. Upon the initiation of DNA replication, PCNA is loaded onto chromatin, Cdt1 associates with PCNA through the PIP-box, and the surface created by this protein–protein interaction is recognized by CRL4Cdt2 ubiquitin ligase. CRL4Cdt2, a Cullin-ring ligase 4-Cdt2, is composed of Cul4, DDB1 and Cdt2 and plays a central role in Cdt1 degradation from yeast to mammals. The WD40 repeat protein Cdt2 is the substrate-recognizing subunit of CRL4Cdt2. Both Cdt1 and Cdt2 were originally isolated as Cdc10-dependent transcripts 1 and 2 in fission yeast (Hofmann & Beach 1994; Arias & Walter 2006; Higa et al. 2006; Jin et al. 2006; Nishitani et al. 2006; Sansam et al. 2006; Senga et al. 2006; Guarino et al. 2011). Cdt1 is also degraded when cells are exposed to DNA-damaging agents, such as UV irradiation, by the same PCNA-dependent CRL4Cdt2 pathway.

In addition to RFC1-RFC, eukaryotic cells have three additional RFC-like complexes, Ctf18-RFC, Elg1-RFC and Rad17-RFC, involved in sister chromatid cohesion, genome stability and checkpoint activation with another clamp, the 9-1-1 complex, respectively (Kim & MacNeill 2003; Majka & Burgers 2004). Therefore, we carried out a knockdown analysis of large RFC complex subunits to determine which RFC complex is required for the PCNA-dependent degradation of Cdt1 and recently reported that Ctf18 and RFC1 are required separately for CRL4Cdt2-mediated Cdt1 proteolysis in the S phase and after UV irradiation, respectively (Shiomi et al. 2012). During this analysis, we unexpectedly observed that PCNA levels on chromatin were extremely increased in Elg1-depleted cells, and thus, we further characterized the role of Elg1. Elg1 was originally isolated as a gene whose defect results in an enhanced level of genomic instability in budding yeast (Ben-Aroya et al. 2003; Kanellis et al. 2003; Smith et al. 2004), and its role in chromosome stability in mammalian cells has been showed (Ishii et al. 2005; Sikdar et al. 2009; Bell et al. 2011). Elg1 is involved in DNA replication, recombination, and the regulation of telomere length (Banerjee & Myung 2004; Smolikov et al. 2004) and the mono-ubiquitination levels of PCNA (Lee et al. 2010). It is not clear, however, how the Elg1 defect is linked to chromosome instability. In this report, we show that Elg1 depletion leads to the accumulation of chromatin-bound PCNA during the S phase and that PCNA is retained on chromatin even in G2 phase cells. The Elg1-depleted cells showed an extended chromatin loop structure in interphase and chromosome aberrations, such as breaks and lagging chromosomes, during mitosis. Our results suggest that, in contrast to other RFC complexes, the Elg1-containing complex plays an important role in maintaining genomic stability by regulating PCNA unloading.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Chromatin-bound PCNA levels are increased in Elg1-depleted cells

In our analyses to determine which PCNA loaders are required for chromatin binding of Cdt2 and for PCNA-dependent Cdt1 degradation by depleting the large RFC complex subunits (Shiomi et al. 2012), we observed that PCNA levels on chromatin were extremely increased in Elg1-depleted cells compared with cells depleted of other large subunits, such as RFC1, Ctf18 and Rad17 in HeLa cells (Fig. 1A). Knockdown analysis of Cdt2, PCNA and DDB1 was included as control. Such an increase in PCNA on chromatin was observed with three small interfering RNAs (siRNAs) targeted for different positions of Elg1 mRNA and using a different human cell line, U2OS cells (Fig. 1B). Reverse transcription–polymerase chain reaction (RT-PCR) analysis confirmed that Elg1 mRNA was knocked down after transfection with each siRNA both in U2OS and in HeLa cells (Fig. 1C, Fig. S1 in Supporting Information). Among the cells, we detected high amounts of chromatin-bound PCNA in Elg1 siRNA #2 or #3 transfected cells. In addition, the mono-ubiquitinated form of PCNA was detected in Elg1 siRNA #2-transfected cells, as reported previously (Lee et al. 2010) (Fig. 1B). Therefore, we used Elg1 siRNA #2 for all the experiments described below. This siRNA efficiently knocked down the expression of exogenously transfected Halo-tagged Elg1 (Fig. S2 in Supporting Information). The PCNA detected in the chromatin-containing fraction was actually associated with chromatin, functioning as a sliding clamp, based on the recovery of PCNA in the supernatant after micrococcal nuclease treatment (Fig. 1D).

image

Figure 1. Elg1 depletion leads to increased proliferating cell nuclear antigen (PCNA) levels on chromatin. (A) Analysis of chromatin-associated PCNA levels. HeLa cells were transfected with small interfering RNAs (siRNAs) for CRL4Cdt2 subunits, Cdt2 and DDB1, for clamp loader subunits RFC1, Ctf18, Rad17 and Elg1(#1), for PCNA and for luciferase as a control (Cont), collected and lysed. Whole-cell extract (W) was separated into supernatant and chromatin-containing pellet fractions (C). ORC2 represents the loading control. (B) Levels of chromatin-bound PCNA in U2OS cells transfected with three different siRNAs for Elg1. WCE, whole-cell extract. The relative amounts of chromatin-bound PCNA, normalized to PCNA in WCE, are standardized to the sample of control siRNA-transfected cell, set as 1.0. Arrow indicates the position of mono-ubiquitinated PCNA. (C) RT-PCR analysis of Elg1 mRNA knockdown. U2OS cells were transfected with the indicated siRNAs, and mRNA levels were analyzed. The mRNA levels of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were used as a control. The graph shows mRNA levels of Elg1 normalized to GAPDH. Error bars indicate SD from two independent results. (D) Chromatin association of PCNA. Chromatin-containing fraction (C) was prepared from whole-cell extracts (W), treated with micrococcal nuclease (+MN) and separated into the soluble fraction (S) and insoluble pellet fraction (P). RCC1 was used as a loading control.

Download figure to PowerPoint

S-phase cells contain a high amount of chromatin-bound PCNA

To examine whether the increased PCNA levels on chromatin in Elg1-depleted cells was due to an increase in the population of S-phase cells, we carried out a flow cytometry analysis. The results of the analysis indicated that the number of S-phase cells increased approximately 1.3 times in Elg1-depleted cells (Fig. 2A). We then examined S-phase progression using a synchronized cell culture. The siRNA-transfected cells were synchronized during early S phase with thymidine and released. The S-phase progression of Elg1 siRNA-transfected cells was slightly delayed from mid- to late-S-phase compared with control siRNA-transfected cells (Fig. 2B, 9 h to 15 h). The delay was not due to activation of the DNA replication checkpoint nor the DNA damage checkpoint, because activation of Chk1 and γH2AX was not detected (Fig. 2D,E). Although these results indicated that the S-phase progression was delayed and the population of S-phase cells was actually increased, we found, using a synchronized cell culture of Elg1-depleted cells, that the chromatin-bound PCNA levels were very high from the early S phase (thymidine-arrested) and throughout the S phase (Fig. 2C). In contrast, PCNA levels in control cells were low in early S-phase cells, and the PCNA levels increased as the cells progressed into S phase. These findings suggest that the increased levels of chromatin-bound PCNA observed in Elg1-depleted cells were not merely a result of the increased number of S-phase cells, but that the S-phase cells contained high amounts of PCNA on chromatin.

image

Figure 2. Elg1-depleted S-phase cells have high levels of proliferating cell nuclear antigen (PCNA) on chromatin and increased numbers of PCNA foci. (A) Analysis of S-phase population. U2OS cells were transfected with the indicated siRNAs and collected for flow cytometry analysis. The population of S-phase cells is shown (%). (B) Cell cycle progression after release from thymidine arrest. The cells were transfected with each siRNA and cultured (asy) or synchronized with thymidine and released. ‘Nco’ represents nocodazole-arrested cells. (C) Immunoblot analysis of PCNA on chromatin. Cell extracts were prepared with the cell sample used in B. PCNA in the whole-cell extract (WCE) and chromatin-bound fraction (Chr) are shown. Histone H3 was used as a chromatin-bound control. The relative amounts of chromatin-bound PCNA, normalized to PCNA in WCE, are standardized to the sample of control siRNA-transfected cells (asy), set as 1.0. (D) Elg1 depletion did not activate the replication checkpoint, as examined by the anti-CHK1(S296) antibody. HU-treated cells were used as a positive control. (E) The DNA damage checkpoint was not activated by Elg1 depletion based on examination with anti-γH2AX antibody. Zeocin, an inducer of DNA breaks, was used as a positive control. Bar, 50 μm (F) Immunofluorescent analysis with anti-PCNA antibody of U2OS cells transfected with indicated siRNAs. Arrowheads, both open and closed, indicate PCNA foci-positive cells. Cells indicated by open arrowheads were enlarged and are shown below in each panel. The fraction of PCNA foci-positive cells among the indicated siRNA-transfected cells is shown (%). At least 300 nuclei were counted for each type of siRNA-transfected cell. Bar, 10 μm.

Download figure to PowerPoint

To study this phenomenon in more detail, we conducted additional analyses. During S phase, PCNA was detected in immunofluorescent analysis as discrete foci that represent DNA replication sites (Bravo & Macdonald-Bravo 1987). Therefore, we examined the PCNA foci in Elg1 siRNA-transfected cells. Consistent with the increased number of S-phase cells among Elg1 siRNA-transfected cells, the frequency of cells with PCNA foci was more than two times higher in the Elg1-depleted cell population compared with control or RFC1- or Ctf18-depleted cells (Fig. 2F). Importantly, among cells containing PCNA foci, the intensity of the PCNA foci was stronger in Elg1-depleted cells. In addition, Elg1-depleted cells had a higher number of PCNA foci than control cells. When either RFC1 or Ctf18 was depleted, we observed no such change in PCNA staining, probably because RFC1 and Ctf18 redundantly function to load PCNA (Shiomi et al. 2004, 2012). These results together suggest that Elg1 depletion leads to an increase in chromatin-bound PCNA levels in S-phase cells.

PCNA remains on chromatin after DNA replication

We then examined how the increase in PCNA levels was linked to DNA replication in Elg1-depleted cells by analyzing the incorporation of thymidine analogues, 5-bromo-2′-deoxyuridine (BrdU) or 5-ethynyl-2′-deoxyuridine (EdU). siRNA-transfected cells were labeled with BrdU for 30 min and costained with BrdU and PCNA after washing out the chromatin-unassociated PCNA by pre-treatment with detergent. In control siRNA-transfected cells, PCNA foci-positive cells were also positive for BrdU and vice versa, and these foci signals completely overlapped, confirming that PCNA foci represent DNA replication sites. In Elg1 knockdown cells, however, we detected a number of cells with PCNA foci, but that were negative for BrdU staining (Fig. 3A, arrows). These cells appeared to be in G2 phase. During a cell cycle, cyclin B protein is detected in the cytoplasm in G2 phase cells (Hagting et al. 1999). Thus, when costained with cyclin B and PCNA, cyclin B-positive cells were negative for PCNA foci, as shown in control cells (Fig. 3B). In contrast, many of the cyclin B-stained cells had PCNA foci when Elg1 was depleted, indicating that PCNA remained bound to chromatin in G2 phase. In addition, PCNA was detected at significantly higher levels in the chromatin-containing fraction prepared from nocodazole-arrested cells (Fig. 2C, Nco and Fig. 4, Nco) and even from G1-synchronized cells (Fig. S3 in Supporting Information) than in control cells. Notably, when we examined EdU-PCNA double foci-positive cells, we detected some PCNA foci that were not stained well with EdU in Elg1 knockdown cells (Fig. 3C). Such PCNA foci were rarely observed in control cells. We predict that such PCNA foci represent chromatin sites where PCNA remained on chromatin after completion of DNA replication. The cells in early-mid and late S phases, as categorized by the pattern of PCNA foci, are shown. Taken together, these results suggest that Elg1 has an important role in unloading PCNA after DNA replication.

image

Figure 3. Proliferating cell nuclear antigen (PCNA) remains on chromatin after DNA replication. (A) Immunofluorescent analysis with anti-BrdU and anti-PCNA antibodies after 30-min pulse label with BrdU. Arrows indicate cells with PCNA foci, but negative for BrdU incorporation. Bar, 50 μm. (B) Immunofluorescent analysis with anti-PCNA and anti-cyclin B antibodies. Arrowheads, both open and closed, indicate cells having PCNA foci. The cells indicated by open arrowheads are costained with cyclin B. Bar, 30 μm. (C) Asynchronous growing cells were labeled with EdU for 5 min and costained with EdU and PCNA after washing out the chromatin-unassociated PCNA by detergent pre-treatment. Images of a single confocal section are shown. Intensities of both green and red signals on the white line in merge images are shown in the right graph. Bar, 20 μm.

Download figure to PowerPoint

image

Figure 4. Elg1 depletion leads to a decrease in chromatin remodeling factors in chromatin during S phase. Immunoblotting of various chromatin-associated proteins in U2OS cells asynchronously growing (asy), released from thymidine block for 0 h (thym), 3 h and 6 h, and in nocodazole (Nco)-arrested M phase cells. Proteins in the whole-cell extract (WCE) and chromatin-bound fraction (Chr) are shown. The (s) and (l) for proliferating cell nuclear antigen immunoblotting represents short and long exposures, respectively. The ratio of chromatin-bound proteins at 3 h, standardized by control siRNA, set as 1.0, is shown in the graph. ND indicates ‘not determined’ due to higher background.

Download figure to PowerPoint

Elg1 depletion causes change in levels of chromatin-associated proteins bound to chromatin

Cdt2 levels on chromatin were increased in Elg1-depleted cells (Fig. 1A). This finding was consistent with the increased levels of chromatin-bound PCNA, because Cdt2 is a subunit of CRL4Cdt2, which associates with chromatin-bound PCNA. We further examined the chromatin levels of Cdt2, other PCNA-interacting proteins, and chromatin-associated proteins using a synchronized culture. As with PCNA levels, Cdt2 levels in the chromatin fraction were high even in thymidine-arrested cells when Elg1 was depleted. The levels of replication factory proteins that interact with PCNA directly, Pol δ, DNA ligase I and MSH2, and those that do not interact directly with PCNA, such as MCM6, did not change upon Elg1 depletion. The behavior of other PCNA noninteracting chromatin proteins, histone H3 and CENP-A, was similar between control and Elg1-depleted cells. In contrast, we observed decreased levels of the chromatin-associated proteins, RanGEF, RCC1; cohesin subunit, SMC3; histone acetyl transferase, HBO1; chromatin remodeling complex, SNF2H; heterochromatin protein, HP1α; and Rap1-interacting factor-1, Rif1, on the chromatin in Elg1-depleted cells (Fig. 4). These findings suggested that Elg1 defects induce changes in the chromatin architecture.

Chromatin loop size increases after Elg1-depletion

To examine the above-mentioned possibility, we studied chromatin loop structure. Chromatin DNA is organized into loop structures in the interphase nucleus (Gerdes et al. 1994), and depletion of Rif1 leads to an increase in chromatin loop size (Yamazaki et al. 2012). Therefore, we carried out a halo assay to examine whether Elg1 depletion affects chromatin loop size. Cells were arrested at early S phase with aphidicolin to neglect the variability in loop size in the cell cycle, permeabilized with detergent and extracted with high salt buffer. Compared with control siRNA-treated cells, chromatin loop sizes were dramatically increased in Elg1-depleted cells (Fig. 5). A similar increase in loop size was also observed with asynchronous growing cells (Fig. S4 in Supporting Information). These results suggest that Elg1 is required to maintain chromatin loop size in interphase cells.

image

Figure 5. Elg1 depletion leads to an increase in chromatin loop sizes. The control and Elg1 siRNA-transfected U2OS cells were synchronized at G1/S-phase with aphidicolin for 24 h. Cells were harvested, and DNA halo assays were conducted. (A) Representative images of control and Elg1 siRNA-transfected cells. Bar, 20 μm. (B) Chromatin loop sizes (halo radius) were calculated. At least 160 halo nuclei were counted for each siRNA-transfected cell.

Download figure to PowerPoint

Elg1 depletion induces chromosomal abnormalities and segregation defects

Based on the finding that Elg1 depletion affected the binding of the chromatin remodeling factor to chromatin during S phase and increased chromatin loop size and that PCNA foci were abnormally detected in G2 phase, we speculated that Elg1-depleted cells would have chromosomal aberrations. To investigate this possibility, we examined the chromosome structure in mitotic cells using a chromatin spread assay. Although chromosome condensation appeared normal in Elg1-depleted cells, we observed that many of the cells had aberrant broken and gapped chromosomes (Fig. 6A). In association with this observation, we detected lagging chromosomes during anaphase to cytokinesis in Elg1-depleted cells, whereas CENP-A staining showed that sister kinetochores were correctly split into two daughter cells (Fig. 6B). A cell survival assay showed that colony formation was impaired in Elg1-depleted cells (Fig. S5 in Supporting Information). Taken together, these results show that Elg1 depletion leads to a chromosome instability.

image

Figure 6. Elg1 depletion induces chromosomal aberration in mitotic cells. (A) Chromatin spread assay. U2OS cells were transfected with control siRNA (a) or Elg1 siRNA (b) and cultured in the presence of nocodazole for 24 h before making the chromosome spread preparations. Mitotic chromosomes were processed for Giemsa staining. Examples of normal (c), gapped (d and e) and broken (f and g) chromosomes are shown. The number of chromosome aberrations was counted for each cell and frequency was plotted (h). At least 40 mitotic nuclei were counted for each experiment. Error bars indicate SD from two independent results. Arrows and arrowheads in (a) to (g) represent chromosome gap and break, respectively. Bars, 5 μm. (B) Immunofluorescent analysis of mitotic cells stained with Hoechst (DNA), anti-tubulin and anti-CENP-A antibodies. Arrows indicate lagging chromosomes. Bars, 10 μm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Eukaryotic cells have three different PCNA-associated RFC complexes, but their specific roles have not been well elucidated. The findings of the present study showed that Elg1 plays an important role in regulating chromatin-bound PCNA levels. First, Elg1 depletion leads to the accumulation of PCNA on chromatin. Synchronized cell cultures showed that the early-S-phase cells had high levels of chromatin-bound PCNA. Second, Elg1-depleted cells were strongly stained and contained a high number of PCNA foci. Third, PCNA remained on chromatin after DNA replication, as showed by the presence of G2 cells that were negative for BrdU staining, but positively stained for PCNA. In addition, some PCNA foci were not stained with EdU (Fig. 3C). We speculate that such PCNA foci represent the chromatin-bound PCNA that should have been removed from the replicated DNA. Taken together, our findings suggest that the Elg1-containing RFC complex is involved in coupling the unloading of PCNA with the completion of chromosome duplication. Inactivation of ctf18+ or over-expression of elg1+ is synthetically lethal in the yeast rfc1 mutant, and the synthetic lethality of rfc1 and ctf18 double mutants is rescued by elg1+ deletion (Kim et al. 2005). These findings are consistent with the notion that both RFC1 and Ctf18 function in loading PCNA on chromatin, whereas Elg1 functions in unloading. Although this paper was in preparation, similar findings were reported in human and yeast cells, demonstrating that an Elg1-containing RFC complex acts as a PCNA unloader (Kubota et al. 2013; Lee et al. 2013). Interestingly, however, chromatin-associated PCNA in nocodazole-arrested cells was dramatically reduced even in Elg1-depleted cells, although only a small amount of PCNA remained on chromatin in comparison with control cells (Figs 2C,4). PCNA might be unloaded by RFC1-RFC or Ctf18-RFC or by another unknown mechanism before entering into or during mitosis. The unloading activity of Ctf18-RFC was reported previously (Bylund & Burgers 2005). Thus, although the Elg1-knocked down cells showed reduced cell survival compared with control cells (Fig. S5 in Supporting Information), such a back-up system might have suppressed Elg1 depletion to some extent.

PCNA was loaded onto double-stranded DNA at a nick or a primer template junction, and such loading processes have been extensively investigated. In contrast, the mechanism of PCNA unloading is not well understood. During a lagging strand replication, loading of PCNA is required for every Okazaki fragment. After maturation of the Okazaki fragment and/or after chromatin reformation, PCNA is unloaded. In these situations, PCNA is present on chromatin, encircling the double-stranded DNA. Thus, Elg1-RFC may recognize this form of PCNA and open the ring for unloading. Detailed biochemical studies are required to clarify how Elg1-RFC as well as other RFC complexes functions in PCNA unloading.

An important consequence of Elg1 depletion in our study was an increase in chromosome aberrations (Fig. 6). Actually, Elg1 was originally isolated as a gene whose defect resulted in an enhanced level of genomic instability in budding yeast, and its role in preventing chromosome instability upon exposure to DNA-damaging reagents and tumorigenesis was also reported in mammalian cells (Sikdar et al. 2009; Bell et al. 2011). As showed here, chromatin-bound levels of various chromatin-associated factors were affected, and Elg1-depleted cells showed abnormally large chromatin loop sizes. Because PCNA acts as a platform for many proteins involved not only in DNA replication but also in chromatin formation, directly and indirectly, it is likely that increased and sustained PCNA levels on chromatin during DNA replication and during cell cycle progression lead to the de-regulation of chromatin-associated proteins and thus a change in chromatin structures, resulting in chromosomal aberrations and tumorigenesis. Retention of PCNA on chromatin also altered the recruitment of PCNA-interacting factor in the yeast elg1Δ strain (Kubota et al. 2011). Elg1 might also be required for the recruitment of chromatin-associated proteins to the replication factory for the construction of the chromatin architecture as chromosome duplication proceeds. A previous report showed that Elg1 physically and genetically interacts with Rad27 (FEN1) (Kanellis et al. 2003), and RFC1 and Ctf18 also interact with various proteins other than PCNA (Levin et al. 2004; Franco et al. 2005; Shiomi et al. 2007; Farina et al. 2008; Redondo-Munoz et al. 2013).

The larger chromatin loop size suggests a larger replicon size (Buongiorno-Nardelli et al. 1982; Courbet et al. 2008). This change in the replicon size might affect S-phase progression. Consistent with this, the population of S-phase cells increased and S-phase progression was slightly delayed when Elg1 was depleted. The delay of S-phase progression is also observed in ELG1-deleted budding yeast (Kanellis et al. 2003). Elg1-depleted cells released from thymidine block appeared to progress normally during the early S phase, although progression was slightly delayed from mid- to late S phase (Fig. 2C). In control cells, PCNA levels on chromatin were low in thymidine-blocked (early S phase) cells, and PCNA levels increased with the progression of S phase. However, when Elg1 was depleted, PCNA levels were already increased in thymidine-blocked cells, and its levels remained at the same high level after release. As S phase progresses, PCNA is controlled so that it is loaded onto replicating DNA and unloaded from replicated DNA. If the unloading of PCNA from the regions replicated early in S phase was delayed in Elg1-depleted cells, the amount of PCNA available for the mid- to late-S-phase replication would be limited. However, we found that Rif1 was drastically reduced from chromatin in Elg1-depleted cells (Fig. 4). Interestingly, a recent report indicated that depletion of Rif1 also leads to an increase in chromatin loop size and a specific loss of replication timing domain structures (Yamazaki et al. 2012). Based on these findings, the progression of mid-S phase is also affected by the loss of Rif1 function.

Although DNA replication was delayed and aberrant chromosomes with breaks and gaps were frequently detected in Elg1-depleted cells, activation of the checkpoint was not observed in our analysis. This finding suggests that chromosome breaks were not induced during S phase. Because PCNA foci were still detected in G2 cells, however, late replicated regions might not be correctly organized before the cells enter mitosis. Such regions would be susceptible to chromosome breaks and thus result in aberrant chromosome structures and lagging chromosome formation. In contrast, the late S-phase-replicated centromere regions appeared normal, as sister kinetochores were correctly separated.

In conclusion, our findings showed that, in contrast to other RFC complexes, the Elg1-RFC complex is an important regulator of chromosome stability by controlling PCNA unloading during chromosome duplication.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Cell culture

Cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum in a 5% CO2 atmosphere. HeLa cells, a cell line derived from cervical cancer, and osteosarcoma U2OS cells were used in the present study. To synchronize cells in early S phase and M phase, siRNA-transfected cells were treated with 0.5 mm thymidine or 5 μg/mL aphidicolin and 0.1 μg/mL nocodazole for 24 h, respectively. To analyze the DNA content, flow cytometry was carried out as described previously (Shiomi et al. 2012).

RNA-interference knockdown experiments

Double-stranded RNAs were transfected at 100 μm using HiPerFect (Qiagen). Twenty-four hours after the first transfection, the second transfection was carried out, and the cells were cultured for two more days. The details of PCNA, Cdt2, DDB1, RFC1, Ctf18 and Rad17 siRNAs were described previously (Shiomi et al. 2012). siRNA for siLuc, known as GL2, was used as a control siRNA. The three Elg1 siRNAs (#1, HSS129124; #2, HSS129125; and #3, HSS129126) were purchased from Invitrogen.

RT-PCR

Total mRNA was collected from U2OS or HeLa cells after siRNA transfection with NucleoSpin RNA XS (Machery-Nagel). The following primer sets were used for RT-PCR: Elg1, ATGAAAGCATTTAGGCAGCC and TCATCAAAACTAGGGACAGG; GAPDH, AATTCCATGGCACCGTCAAGGC and TTACTCCTTGGAGGCCATGTGG.

Antibodies, Western blotting and immunofluorescence

For Western blotting, whole-cell lysates were prepared by lysing cell pellets directly in SDS-PAGE buffer. Chromatin fractionation was carried out as described previously (Shiomi et al. 2012). For immunofluorescence, cells were fixed in ice-cold methanol for 30 min, with or without pre-extraction using 0.1% (v/v) Triton X-100 in phosphate-buffered saline (PBS), and stained with the indicated antibodies, as described previously (Shiomi et al. 2012). The confocal image was processed by LAS AF Lite (Leica). Incorporation of BrdU (11296736001, Roche) or EdU (C10337, Invitrogen) was carried out according to the manufacturer's instructions. The following primary antibodies were used: MCM6 (sc-9843), Pol δ (p125) (sc-17776), DNA ligase I (sc-47703), MSH2 (sc-494), HBO1 (sc-13283, Santa Cruz Biotechnology); histone H3 (#3638), CHK1 (#2345), phospho-CHK1 (#2349), CENP-A (#2186, Cell Signaling); SNF2H (#05-698), γH2AX (#05-698, Upstate Biotechnology); HP1α (BMP001), CAF1 (K0115-3, MBL International); and Rif1 (A300-568A, Bethyl Laboratories). The polyclonal anti-PCNA antibody was a kind gift from Dr. Tsurimoto (Kyushu University). The antibodies for Smc3, Cdt2, PCNA, RCC1, Orc2 and phospho-Histone H3 were described previously (Shiomi et al. 2012). Protein levels were analyzed using Imagej software.

Mitotic chromatin spreads

U2OS cells were synchronized in M phase with nocodazole after siRNA treatment. Cells were harvested, treated with 0.9% sodium citrate for 15 min and then five volumes of Carnoy solution (MeOH : AcOH=3 : 1), and collected immediately by centrifugation. Cells were suspended in Carnoy solution, incubated for 30 min and then dropped onto a glass slide. DNA was stained with 3% Giemsa solution.

Nuclear halo assay

The siRNA-transfected U2OS cells were synchronized at early S phase with aphidicolin. Nuclear halo assays were carried out as previously described (Gerdes et al. 1994). The halo radius of each nucleus was determined by measuring the total area (A) of the nucleus and the central area (B) by Photoshop (Adobe) and using the following formula: halo radius = √(A/π) − √(B/π).

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We would like to thank Mr S. Kanda (University of Hyogo) for his preliminary experiments in this study. We also thank Dr T. Tsurimoto (Kyushu University) for polyclonal anti-PCNA antibody; Drs S. Yamazaki, Y. Kanoh and H. Masai (Tokyo Metropolitan Institute of Medical Science), and Drs K. Nishimura and M. Kanemaki (The National Institute of Genetics) for their practical advice regarding the halo assay and chromatin spread assay. This work was financially supported by Grant-in-Aid for Scientific Research on Innovative Area (23131512) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan and Hyogo Science and Technology Association to HN, and by Grants-in-Aid for Basic Scientific Research (C) (25430171) and Hyogo Science and Technology Association to YS.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
gtc12087-sup-0001-FigureS1.pdfapplication/PDF377K

Figure S1 RT-PCR analysis of Elg1 mRNA knockdown.

gtc12087-sup-0002-FigureS2.pdfapplication/PDF306K

Figure S2 Exogenous expression of Halo-Elg1 was impaired in Elg1 small interfering RNAs (siRNA) transfected cells.

gtc12087-sup-0003-FigureS3.pdfapplication/PDF208K

Figure S3 PCNA remains on chromatin in G1 phase synchronized cells.

gtc12087-sup-0004-FigureS4.pdfapplication/PDF222K

Figure S4 Elg1 depletion leads to an increase in chromatin loop size.

gtc12087-sup-0005-FigureS5.pdfapplication/PDF5620K

Figure S5 Colony formation and growth defects in Elg1-depleted cells.

gtc12087-sup-0006-FigureS1-S5.docxWord document109K 

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.