![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Molecular and Cellular Biology, June 2008, p. 3672-3685, Vol. 28, No. 11
0270-7306/08/$08.00+0 doi:10.1128/MCB.00510-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.
,
,
Régis Courbeyrette,
Morgane Agez,
Françoise Ochsenbein,
Carl Mann,* and
Jean-Yves Thuret*
CEA, iBiTec-S, Gif-sur-Yvette F-91191, France
Received 23 March 2007/ Returned for modification 24 May 2007/ Accepted 20 March 2008
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Nucleosome assembly and disassembly occur during DNA replication, repair, transcription, and other events that necessitate access to DNA. In mammals, the major isoforms of the histone H3 family are H3.1, H3.2, and H3.3 (16). H3.1 and H3.2 are expressed and massively incorporated into chromatin during DNA replication in proliferating cells. H3.3 is expressed in all cell cycle phases in proliferating cells, but it is also expressed in quiescent cells (59, 60). Its posttranslational modifications correspond to active gene loci (6, 29), and Ahmad and Henikoff have demonstrated that H3.3, but not H3.1, is deposited in a transcription-dependent manner in Drosophila melanogaster cells (3). Moreover, distinct chaperone complexes copurified with H3.3 or H3.1 histones: in vitro, the complexes containing H3.1 are solely competent for DNA synthesis-coupled (SC) assembly, and the complexes containing H3.3 are solely competent for DNA synthesis-independent (SI) assembly (54). The chaperone complexes isolated with H3.1 include the evolutionarily conserved chromatin assembly factor 1 (CAF-1) (54). CAF-1 consists of three subunits (p150, p60, and p48 in human cells) (21, 57) and had previously been isolated as a chaperone for the assembly of newly synthesized H3 and H4 on replicating DNA (52). Through interaction with proliferating cell nuclear antigen (PCNA), CAF-1 localizes to replication foci in S-phase cells (23, 49) and at sites of damaged DNA in UV-treated cells (12). CAF-1 also promotes H3.1 deposition on UV-damaged DNA in vitro (11, 54) and in vivo (42). CAF-1 is essential for viability in early mouse embryos and in murine embryonic stem cells (18). CAF-1 is also essential in humans, as depletion of p60CAF-1 triggers apoptosis in proliferating, but not quiescent, cells (35). Interestingly, in vitro assays showed that extracts from cells depleted of p150CAF-1 or p60CAF-1 are defective in the assembly of nucleosomes on a replicating template (17, 35). Moreover, CAF-1 is directly linked to S-phase progression, as depletion of p150CAF-1 by RNA interference triggers an accumulation of cells in early and mid-S phase (17). CAF-1 is thus a major player in chromatin assembly dependent on DNA synthesis.
SI nucleosome assembly of H3.3 has been characterized during transcription in Drosophila cells in vivo, but the histone chaperones involved in this deposition were not identified (3). Chaperone complexes isolated with H3.3 from HeLa cells include the evolutionarily conserved histone repression A factor (HIRA) (54). HIRA is a 110-kDa protein distantly related to p60CAF-1. HIRA belongs to the family of Hir (histone repression) factors implicated in the coupling of histone synthesis and DNA replication. Indeed, overexpression of HIRA in human cells inhibited histone expression and led to an S-phase arrest (39). Importantly, HIRA is an essential gene for development, as HIRA–/– mice die in utero by day 11 (46). HIRA also promotes the formation of senescence-associated heterochromatic foci (SAHF) in primary human cells (61) and replication-independent assembly of H3.3 to replace protamines by histones throughout the paternal genome of the one-cell embryo (26).
Histone H3.1 and H3.3 complexes purified from HeLa cells contain the human ASF1 (hASF1) proteins (ASF1A and ASF1B) (54). ASF1A and ASF1B belong to the ASF1 family of histone H3/H4 chaperones conserved from yeast to humans. Its founding member was first isolated in Saccharomyces cerevisiae as a gene whose overexpression or deletion disrupts silenced chromatin (anti-silencing function 1 [25, 51]). Drosophila melanogaster ASF1 was subsequently isolated as a factor that assembles chromatin onto newly replicated DNA in vitro in synergy with CAF-1 (56). D. melanogaster ASF1 also colocalizes with active replication forks in Drosophila tissue culture cells (48), and in human cells, ASF1A/B is bound through H3/H4 histones to the putative replicative minichromosome maintenance (MCM) helicase complex in chromatin (14). In vertebrate and Drosophila cells, depletion of ASF1 leads to an accumulation of cells in S phase (14, 15, 47, 48). In human cells, this inhibition of S-phase progression has been attributed at least in part to a defect in progression of the MCM helicase in the absence of ASF1A/B (14).
ASF1A and ASF1B are 84% identical in their first 156 amino acids but highly divergent in their C termini. In SC assembly, ASF1A and ASF1B appear to have common functions. Indeed, both proteins are present in the complexes copurified with H3.1 that have in vitro SC assembly activity (54). In contrast, ASF1A and ASF1B may function differently in SI assembly, although both proteins copurify with H3.3. Indeed, ASF1A is dramatically enriched in the complexes containing HIRA that are active for SI assembly in vitro (54). Moreover, ASF1A, but not ASF1B, interacts and synergizes with HIRA in vivo for the formation of SAHF (61). Additionally, quiescent human cells contain ASF1A but no detectable ASF1B (43). Recently, Tang et al. studied the ASF1A-HIRA interaction (55). They defined the B domain of HIRA that contains 37 evolutionarily conserved residues that are necessary and sufficient for interaction with the core domain of ASF1A. Interestingly, ASF1A residues involved in binding to the B domain are strictly conserved in ASF1B. Consequently, both ASF1A and ASF1B are able to interact with the B domain. Tang et al. showed that the preferential interaction of HIRA with ASF1A depends on both the N-terminal 30 residues and the C-terminal 50 residues of hASF1 as well as uncharacterized regions of HIRA outside the B domain.
During replicative stress, the amount of newly synthesized histone H3 interacting with hASF1 increases (15). Groth et al. (15) proposed that hASF1 functions as a histone buffer that donates H3/H4 histones to CAF-1 and HIRA, which would be more directly involved in depositing them onto DNA in SC and SI nucleosome assembly pathways, respectively. Indeed, in Xenopus egg extracts that contain an abundance of stored histones, ASF1 is dispensable for direct de novo histone deposition in either pathway (45).
In this work, we analyzed the functions of hASF1 in SC and SI assembly of newly synthesized histones H3.1 and H3.3 in live HeLa cells.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
GCT, Gly68 GGAGGT, Arg69 AGGCGA, His70 CATCAC, Phe72 TTTTTC, and Val73 GTAGTC silent mutations. Mutants were obtained by PCR-mediated mutagenesis (Roche's Expand High Fidelity). All constructs were sequenced. Cell culture, transfections, and siRNA. HeLa cells were grown in Dulbecco modified Eagle medium (4.5 g of glucose/liter) with 10% fetal bovine serum (Gibco). For synchronization, cells were blocked in mitosis with nocodazole (10 ng/ml, 5 to 6 h). Mitotic cells were harvested by mechanical shake-off and then transfected. For S-phase synchronization, thymidine was added 5 h later (400 µg/ml). The cells were incubated for 13 more hours and then released in S phase by washing the cells with warm medium. siRNAs were designed and synthesized by Invitrogen (Stealth siRNA), and used at final concentrations of 20 nM or 50 nM (Asf1B-II). The sequences of the siRNAs are as follows: ASF1A-I, CCGCAGGAAGGCAUAUGUUUGUAUU; ASF1A-II, GGCAAAGGUUCAGGUGAACAAUGUA; ASF1B-I, CAGCAGGGAGACACAUGUUUGUCUU; and ASF1B-II, CGACCUGGAGUGGAAGAUCAUUUAU. All transfections were done with Lipofectamine 2000 (Invitrogen).
Immunoprecipitations, SDS-PAGE, and Western blotting.
HeLa cells (
10e7) were harvested 24 h after transfection, washed in PB buffer (100 mM potassium acetate, 30 mM KCl, 10 mM Na2HPO4, 1 mM MgCl2, pH 7.4) and lysed on ice in 1 ml of PB buffer containing 0.2% Triton X-100, 200 mM NaCl, and protease inhibitor (Complete; Roche). The lysate was centrifuged for 1 h at 20,000 x g. Anti-GFP microbeads (60 µl; Miltenyi Biotec) were added to the supernatant for 30 min at 4°C. The suspension was loaded on magnetic columns, which were then extensively washed with lysis buffer. The microbeads were harvested, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) 5x loading buffer was added, and the samples were incubated at 95°C for 5 min. After SDS-PAGE, proteins were transferred to nitrocellulose membranes (Schleicher & Schuell) and detected with anti-GFP (Ab290; Abcam), anti-p60CAF-1 (NB500-212A1; Novus Biologicals), anti-hASF1, anti-H3 (Ab1791; Abcam), anti-glyceraldehyde-3-phosphate dehydrogenase (Ab9485; Abcam), or anti-HIRA antibody and horseradish peroxidase-linked secondary antibodies.
Fixed and live cell microscopy. For fixed cell microscopy, cells were grown on three-well glass slides (Cel-line). All incubations were carried out at room temperature. The cells were washed with phosphate-buffered saline (PBS) and fixed in PBS containing 2% formaldehyde. The cells were washed with PBS, incubated in PBS containing 0.2% Triton X-100, and washed (W buffer [PBS containing 0.05% Triton X-100 and 0.1% Tween 20]). The cells were then incubated in filtered PBS containing 0.1% Tween 20 and 5% bovine serum albumin (Sigma) and then incubated with rabbit anti-GFP (Ab290; Abcam), mouse anti-Flag (M2; Sigma), or mouse anti-p150CAF-1 (NB500-207A1; Novus Biologicals) antibodies, washed (buffer W), and finally incubated with Alexa Fluor 488- or Alexa Fluor 594-conjugated secondary antibodies (Molecular Probes). For dual anti-GFP/antibromodeoxyuridine (anti-BrdU) immunolabeling, the cells were first treated with anti-GFP antibodies and anti-rabbit secondary antibodies, then postfixed in PBS containing 2% formaldehyde, and treated with anti-BrdU (clone BU1/75; Harlan Sera Lab) antibodies after 10 min of DNA denaturation in 4 N HCl. The slides were mounted in Prolong Gold (Invitrogen). Image acquisition was done on a Zeiss LSM510 Meta confocal microscope (Plan-Apochromat 63x 1.4-numerical-aperture [NA1.4] objective). For Fig. 2B and C and 3B and C, a representative plane was selected and imaged. For Fig. 2A and 3A, images were taken along the z axis for both anti-GFP and anti-p150CAF-1 probes, and a representative plane was selected. For the profile plots, intensity values were determined with ImageJ (http://rsb.info.nih.gov/ij/) and expressed as a percentage of the mean fluorescence value along the line. Colocalization was assessed in at least 500 nuclei from five different experiments; cells where clearly defined foci were unambiguously visible, even on a diffuse background, were classified as "cells with H3.1-FP foci." Images of mitotic cells stained with anti-GFP were taken along the z axis. The whole stack was projected on one plane using ImageJ and the "standard deviation" z-projection method.
|
|
|
T open glass dishes after shake-off of nocodazole-treated cells; incubated in CO2-independent medium [Gibco] and kept at 37°C with a Bioptechs TC3 Bioptechs controller; evaporation of the medium was prevented by a layer of mineral oil [Sigma]). Photobleaching and image acquisition were carried out on a Zeiss LSM510 Meta confocal microscope with a Plan Neofluar 100x NA1.3 oil immersion objective. The pinhole aperture was set at the maximum value. EGFP was bleached with 100% intensity of lines of 477 nm and 488 nm, and Venus was bleached with lines of 488 nm and 514 nm of a 40-mW laser (argon) set at 85% output power. For each experiment, 200 images were acquired in 88 s. Images were 256 by 200 pixels, 12 bits, acquired in 192 ms, separated by 250 ms. Fifty images were acquired before photobleaching. A 40- by 10-pixels rectangular area was photobleached during 116 ms. Acquisition of the next image started 2 ms later. For all frames, the mean intensity in a background region over time [Ibkg(t)], the mean intensity in the photobleached area over time [IFRAP(t)], and the mean intensity for the whole nucleus over time [Itot(t)] were determined. We carried out double normalization (41), which corrects for prebleach intensity, background noise, and photobleaching as follows: I(t) = [Itot(pre) – Ibkg(t)]/[Itot(t) – Ibkg(t)][(IFRAP(t) – Ibkg(t)]/[IFRAP(pre) – Ibkg(t)], where Itot(pre) and IFRAP(pre) were calculated as the mean of 50 values of Itot(t) and IFRAP(t) before bleaching. These calculations were done in Microsoft Excel. For each condition, at least 20 acquisitions were done in two different experiments. The curves corresponding to the mean I(t) value, as well as the 99% confidence interval, are shown on the graphs. The estimates of the mobile fractions were also calculated according to an alternative normalization as described in the figure legend to Fig. S3 in the supplemental material, which allows a direct reading of the mobile fraction and the standard errors. In order to get a better average value, the estimates correspond to the mean of the last 15 values of all experiments (t = 82.01 s to 88.22 s) and are within a 99% confidence interval (these values were not significantly different from those observed between t = 75.35 s and t = 81.56 s [data not shown]). For FRAP experiments on cells in S phase, a control was carried out to check that cells were in S phase. During the FRAP experiment, the coordinates of the treated cells were noted. After the completion of the FRAP experiments, the cells were subjected to anti-CAF-1p150 immunofluorescence staining, the treated cells were observed, and only the FRAP data for cells with replication foci were analyzed.
For salt extraction experiments, the cells were grown on Bioptechs T glass dishes and treated with increasing salt concentrations on the stage of a fully automatic Leica DMIRE2 microscope. First, the cells were washed with PB buffer (see "Immunoprecipitations, SDS-PAGE, and Western blotting" above), several representative fields were selected, and their positions were recorded by using Metamorph (Roper Scientific). Images of all the fields were acquired with a 40x, N Plan, NA0.55 CORR objective. PB buffer was replaced by PB buffer containing 0.2% Triton X-100 and 0.75 M NaCl, and images of the same fields were acquired with identical exposure times. The extraction buffer was then replaced by PB buffer containing 0.2% Triton X-100 and 2 M NaCl, and images of the same fields were acquired with identical exposure times. NaCl at a concentration of 0.75 M extracted H3-FP that was not deposited into chromatin, whereas 2 M NaCl extracted chromatin-bound H3-FP but did not extract H3-FP found in bright fluorescent nucleoli for the minority of transiently transfected cells that expressed aberrantly large amounts of H3-FP (see Fig. S1A in the supplemental material). This observation supports the idea that this population represents a sequestered population of hydrophobically aggregated H3-FP molecules. For all image acquisitions, the focal plane with maximal intensity was selected to correct for the possible focus drift due to changes in the media. The total fluorescence of single nuclei before and after extraction was determined after subtraction of local background noise, and fluorescence at 0.75 and 2 M was expressed as a percentage of the initial fluorescence. The values are within a 99% confidence interval. For Fig. 4E, the minimum and maximum gray values displayed were set identically for all images. As some cells never enter S phase after synchronization, the extent of H3.1WT remaining after 0.75 M salt extraction was assessed for H3.1WT-transfected cells that did not have S-phase foci and found to be lower than 25%. We kept only the cells where more than 25% of the fluorescence remained after extraction in 750 mM NaCl in order to ensure that we were analyzing cells in S phase.
|
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Endogenous ASF1A and ASF1B coimmunoprecipitated with transiently expressed H3.1-FP and H3.3-FP, as illustrated in Fig. 1A, lane 2, and Fig. 1B, lanes 2 and 3. In order to further analyze these interactions in vivo, we turned to the BiFC assay, where the two putative partners are fused to the first 172 amino acids or to the last 83 amino acids of CFP. If the two parts of CFP stay in close proximity, they form a functional fluorescent protein. Fluorescence thus indicates that the two putative partners interact (19). In our assay, where the transfection efficiency was about 70%, we detected the interaction of wild-type ASF1A with wild-type H3.1 in 60 to 70% of the cells (Fig. 1C and D). In contrast, using a V94R mutant of ASF1A and wild-type H3.1, fluorescence was seen in less than 7% of the cells. We showed previously that the ASF1-V94R mutation abolishes the ASF1-H3 interaction (32), and recent structural studies explain the disruptive effect of this mutation (2, 4, 10, 38). The BiFC assay thus faithfully reflects the in vivo interaction of ASF1A and H3.1.
|
We mutated R128, R129, and R134 and assessed the ability of the resulting H3 mutants to interact with ASF1A and ASF1B. All tagged mutants were expressed at levels similar to those of the tagged wild-type protein in transient-transfection experiments (Fig. 1A and B) (see Fig. S1C and D in the supplemental material). As illustrated in Fig. 1A, lanes 3 and 4, and Fig. 1B, lanes 4 and 5, the triple mutation of R128, R129, and R134 to alanine (3RA), as well as the single mutation of R129 to glutamate (R129E) completely abolished interaction of either H3.1 or H3.3 with ASF1A and ASF1B in our coimmunoprecipitation assay. The 3RA mutation also abolished interaction of H3.1 with ASF1A in vivo in the BiFC assay (Fig. 1C and D). A single mutation of R129 to alanine or a double mutation of R128 to alanine and R134 to glutamate (RA/RE) resulted in a dramatically reduced interaction of H3.1 with ASF1A and ASF1B (Fig. 1A, lanes 5 and 6). We estimate that these mutants coimmunoprecipitated less than 5% of the ASF1A and ASF1B proteins coimmunoprecipitated with the wild type (data not shown). Interestingly, no interaction between ASF1A and these mutants could be detected in vivo in the BiFC assay, suggesting that this assay, albeit providing important information about in vivo interactions, is less sensitive in detecting residual interaction than the coimmunoprecipitation assay. Importantly, the 3RA, R129E, R129A, and RA/RE mutants of H3.1 maintained their interaction with p60CAF-1 (Fig. 1A), and the 3RA and R129E mutations did not affect the ability of H3.3 to interact with HIRA (Fig. 1B). We conclude from these data that the C-terminal arginine 129 and at least one of the two arginines (arginine 128 and 134) of H3 are involved in the interaction of tagged H3 with hASF1 but are not required for the interaction with p60CAF-1 and HIRA. These results prompted us to analyze the consequences of these mutations for the assembly of histones into nucleosomes.
Histone H3.1 mutants defective in ASF1 binding are diffusely distributed in the nucleoplasm during S phase. We first set out to compare the localization of newly synthesized wild-type and mutant histones during the S phase. Kimura and Cook showed that in heterokaryons formed between wild-type HeLa (nonfluorescent nuclei) and HeLa cells expressing H3.1-GFP (fluorescent nuclei), newly synthesized H3.1-GFP imported into nonfluorescent S-phase nuclei colocalized with "replication factories" (22). Replication factories are nuclear structures where DNA replication occurs. They were identified in the nuclei of mammalian cells as foci where analogues of nucleotides (e.g., BrdU) and proteins associated with the replication fork (e.g., PCNA) are detected by immunofluorescence microscopy (5, 31, 37). CAF-1, which interacts with PCNA, is also present in the replication factories (23). The apparent size, number, and localization of replication foci are different in early, mid, and late S phase. To confirm that newly synthesized H3.1-Venus was indeed present in replication foci in transiently transfected HeLa cells, we analyzed its localization 5 h after transfection. A small fraction of the transfected cells were already in or entered S phase during the incubation. As illustrated in Fig. 2A, H3.1-Venus was present in the replication factories identified by anti-p150CAF-1 immunofluorescence microscopy in patterns corresponding to early, mid, and late S phase. The foci seen with the anti-p150CAF-1 and the anti-GFP antibodies colocalized to a large extent as expected.
In order to analyze a large number of S-phase cells, we synchronized the cells with nocodazole and thymidine. The cells were transfected with the tagged histone constructs. The localization of the tagged proteins was assessed by immunofluorescence microscopy after the cells were released into S phase. We performed dual immunofluorescence labeling with anti-GFP and anti-BrdU antibodies in cells that had been released for 1 hour (Fig. 2B) or 2 hours (Fig. 2C) into S phase and where BrdU had been added to the culture medium at 37°C for 10 min before fixation. Under these experimental conditions, H3.1-Venus staining corresponds to the localization of tagged H3.1 that was synthesized during the synchronization protocol and after release into S phase, whereas the foci revealed with the anti-BrdU antibodies correspond to sites where DNA replication is taking place at the time of fixation. As illustrated in the blown-up area of the nucleus in Fig. 2B, one hour after the start of S phase, there is either colocalization or a good overlap between H3.1-Venus staining and the replication foci. In contrast, 2 hours after the start of S phase, H3.1-Venus staining showed only partial overlap with the replication foci (Fig. 2C). We interpret this first by the fact that the ratio of tagged H3.1 to endogenous H3.1 will be higher in very early S phase than later, thus making regions where chromatin has been assembled early in S phase potentially containing more label than at later times. Second, most of the assembled chromatin containing tagged H3.1 may simply have moved away from the actual sites of BrdU incorporation.
Since there was some disjunction between the replication foci and the tagged histone localization in synchronized cells, we decided to directly compare the localization of wild-type and mutant histones. We simultaneously transfected cells with one construct expressing wild-type H3.1-Flag and another expressing mutant H3.1-Venus and visualized their localization by dual anti-Flag and anti-GFP immunofluorescence. The localization of the tagged proteins was assessed by immunofluorescence microscopy 90 min after the cells were released into S phase. As illustrated in Fig. 3A, H3.1WT-Flag and H3.1WT-Venus were highly colocalized. In contrast, H3.1-3RA-Venus was more diffusely distributed in the nucleus and therefore did not colocalize well with H3.1WT-Flag. We also performed dual immunofluorescence labeling as described above on synchronized cells transfected with H3.1-3RA-Venus. As shown in Fig. 3C and D, at both 1 and 2 hours into S phase, H3.1-3RA-Venus staining was more diffuse and not clearly colocalizing with or adjacent to the replication foci (see Discussion).
We quantified the number of synchronized S-phase cells where the H3.1-Venus proteins precisely colocalized with H3.1WT-Flag (Fig. 3A, left column) versus cells where the localization of H3.1-Venus was diffuse (Fig. 3A, right column). The H3.1-3RA and H3.1R129E mutants colocalized with H3.1WT in 5.6% ± 2% and 4.4% ± 1% of the cells, respectively, whereas the H3.1R129A and H3.1RA/RE mutants colocalized with H3.1WT in 43% ± 5% and 20% ± 6% of the cells, respectively (Fig. 3B). Interestingly, H3.1R129A and H3.1RA/RE conserved a weak but detectable interaction with ASF1A and ASF1B (Fig. 1A and data not shown). The localization of H3.1 mutants in the foci revealed by wild-type H3.1 was thus in excellent correlation with their ability to interact with ASF1A and ASF1B.
Newly synthesized histone H3.1 and H3.3 deposition in chromatin during S and G1 phases. We further characterized our H3 mutants using FRAP to assess the intranuclear mobile fraction of tagged histones. FRAP has previously been used in several studies of histone dynamics in animal cells. We reasoned that any difference in the assembly of the H3.1 mutants into nucleosomes would have consequences on the fraction of mobile histones. Indeed, Kimura and Cook have shown qualitative differences in the mobile fractions of H4-GFP in S-phase and non-S-phase nuclei (22). Additionally, while this study was being performed, Meshorer et al. described the use of FRAP to assess the mobile fraction of newly assembled histone H3 (30).
First, we confirmed that we could actually measure the difference of mobile fractions between assembled and unassembled proteins. FRAP was carried out on a rectangular region of the nucleus representing less than 5% of the nuclear surface in the focal plane. All the results presented here correspond to the analysis of at least 20 cells from two different experiments. The fluorescence recovery values for all experiments were normalized, which allowed a statistical analysis of the results and a direct estimation of the mobile and immobile fractions. Using this protocol, 100% ± 0.9% of GFP fused to a nuclear localization signal was mobile (n = 29) (Fig. 4A) (see Fig. S3A in the supplemental material). In contrast, only 30% ± 0.5% of H3.1WT-EGFP in HeLa cells stably expressing the protein was mobile (n = 98) (Fig. 4B) (see Fig. S3B in the supplemental material). We synchronized cells as previously described and found that 32% ± 0.9% of newly synthesized H3.1WT-Venus after 1.5 h into S phase was mobile (n = 28) (Fig. 4C) (see Fig. S3C in the supplemental material). These results are consistent with those of Meshorer et al. (30) and Kimura and Cook (22).
We next analyzed the fraction of H3.1 that is mobile under conditions where it is not normally assembled, i.e., outside of S phase. We synchronized HeLa cells by mitotic shake-off as described above and found that 96% ± 1% of H3.1 was mobile after 7 h into G1 phase (n = 59) (Fig. 4D) (see Fig. S3D in the supplemental material). This demonstrated that when it is not incorporated, wild-type H3.1 is almost entirely mobile in the nucleoplasm. We then measured the fraction of H3.1-3RA-Venus that is mobile in S phase. Cells were synchronized and transfected as described above, and 72% ± 2% of H3.1-3RA-Venus was mobile (n = 34), which is much greater than the percentage of H3.1WT-Venus that was mobile (Fig. 4E) (see Fig. S3C in the supplemental material).
The changes in mobile fractions could result from differences in chromatin assembly but also from differences in interactions with macromolecular complexes in the nucleus. To check whether macromolecular interactions distinct from assembly in nucleosomes could play a significant part in the extent of changes between the wild-type histones and the mutant histones, we performed salt extraction on cells synchronized and transfected as for the FRAP experiments. Whereas the soluble nuclear proteins and some of the nucleosomal H2A and H2B are extracted from chromatin with 750 mM NaCl, nucleosomal histones H3 and H4 are resistant to extraction at this salt concentration, but they are extracted at 2 M NaCl (22, 53). Indeed, for wild-type H3.1-EGFP stably expressed in HeLa cells, about 20% of the fluorescence was extracted with 750 mM NaCl in asynchronous cells (Fig. 4F), and the remainder was subsequently extracted with 2 M NaCl (not shown). We also measured about 20% soluble fluorescent H3.1 by Western blotting with anti-GFP antibodies (see Fig. S2C in the supplemental material). As shown in Fig. 4F, about 30% of wild-type fluorescent H3.1 was extracted from early S-phase nuclei with 750 mM NaCl (all H3.1WT was subsequently extracted with 2 M NaCl [not shown]). This showed that 70% of newly assembled wild-type H3.1 behaved as nucleosomal histone H3. Interestingly, as shown in Fig. 4F, more than 50% of the H3.1-3RA-Venus fluorescence was extracted from early S-phase nuclei with 750 mM NaCl (all H3.1-3RA was subsequently extracted with 2 M NaCl [not shown]). We conclude from this experiment that about 46% of H3.1-3RA is assembled in nucleosomes under these experimental conditions, though this mutant does not interact with hASF1 and does not colocaliz well with replication foci.
Importantly, the salt-resistant H3 fraction was comparable to the immobile fraction observed in the FRAP experiments, which confirmed that the immobile fraction indeed corresponds to the fraction of H3-FP incorporated into chromatin. Since the cells are kept alive and under culture conditions during the FRAP experiments, they are more likely to reveal physiological differences in the assembly of histones in chromatin. All further experiments were carried out using exactly the same FRAP protocol.
We then analyzed the mobile fractions of WT and mutant histone H3.3-Venus. Wild-type H3.3 is assembled into nucleosomes by SI assembly pathways that are active both inside and outside of the S phase. We synchronized HeLa cells in G1 as described above and found that 66% ± 1% of H3.3WT-Venus was mobile after 7 h in G1 (n = 43) (Fig. 4G) (see Fig. S3E in the supplemental material). This confirmed, in comparison with the mobile fraction of H3.1, that H3.3 is incorporated into chromatin outside of S phase. Interestingly, 91% ± 2% of H3.3-3RA-Venus was mobile after 7 h in G1 (n = 33) (Fig. 4H) (see Fig. S3E in the supplemental material), which is thus significantly more mobile than H3.3WT-Venus, showing that the H3.3-3RA mutant is defective for assembly during G1. We also found that 30% ± 0.8% of newly synthesized H3.3WT was mobile in early S phase in cells synchronized as described above for H3.1 (n = 48) (Fig. 4I) (see Fig. S3F in the supplemental material). The mobile fractions of wild-type H3.1 and H3.3 in early S phase are thus similar. Note that, under our experimental conditions, H3.3 was expressed in synchronized cells for 17 h before release into S phase and was therefore also assembled into chromatin in G1 by replication-independent mechanisms. In comparison, 60% ± 2% of H3.3-3RA-Venus was mobile in S phase (n = 45) (Fig. 4J) (see Fig. S3F in the supplemental material), showing that the H3.3-3RA mutant is also defective for assembly during S phase.
In addition to the results of the salt extraction experiment, we also obtained independent confirmation that the H3.1-3RA mutant was assembled into chromatin. First, when we analyzed the localization of the H3.1 mutants at 3 days posttransfection, we observed that the H3.1 mutants were all present on mitotic chromosomes (Fig. 5A). Second, we performed FRAP analysis on asynchronous interphase cells 3 days after transfection and compared the results to those of the mobile fractions of H3.1WT-Venus and H3.3WT-Venus under the same conditions. We obtained values of 25% ± 1% (n = 20), 24% ± 1% (n = 21), 23% ± 1% (n = 20), and 21% ± 0.9% (n = 22) for H3.1WT-Venus, H3.1-3RA-Venus, H3.3WT-Venus, and H3.3-3RA-Venus, respectively (Fig. 5B and C and data not shown). These values confirm the stable deposition of the mutant H3 histones into chromatin.
|
Both ASF1A and ASF1B participate in H3.3 chromatin deposition in G1 phase. The assembly defects of H3.1 and H3.3 mutants unable to interact with ASF1A/B suggested that their interaction with these chaperones was crucial for efficient assembly into chromatin. We further tested this idea by examining the effect of depleting ASF1A and/or ASF1B on the nucleosomal assembly of newly synthesized wild-type H3 histone. Simultaneous depletion of hASF1A and hASF1B causes cells to accumulate in S phase (14, 15, 47, 48). The dramatic S-phase block would make it difficult to distinguish direct from indirect effects of ASF1A/B depletion on the deposition of H3.1 in S phase. Therefore, we decided to concentrate on the effect of hASF1 depletion on the deposition of H3.3 in G1-phase cells. We first determined the effect of hASF1 depletion on G1-phase progression. We used two unrelated siRNAs for ASF1A (A-I and A-II) and ASF1B (B-I and B-II) for hASF1 knockdown. Cells were treated with siRNA and synchronized in mitosis as shown in Fig. 6A. The extent of hASF1 knockdown is shown in Fig. 6B. To analyze the consequences of hASF1 depletion on the duration of the G1 phase, we added BrdU immediately after mitotic release to allow the detection of S-phase cells at the end of G1. As shown in Fig. 6C, between 9 and 13 h after the release from the nocodazole block, we observed a linear accumulation of S-phase cells for depleted and undepleted cells. However, at all time points, there were more cells in S phase in the undepleted population than in the population depleted for hASF1. From these data, it can be deduced that about 14% of the undepleted cells enter S phase every hour compared with 10% for the cells depleted of hASF1. Interpolation to the origin suggested that the first undepleted cells started S phase about 8 h after the release from the nocodazole block compared with 8.5 h for the cells depleted of hASF1. It can be concluded that depletion of hASF1 results in a modest lengthening of G1 phase.
We next examined the effect of hASF1 depletion on the mobile fraction of H3.3WT-Venus in G1 phase. As shown in Fig. 6D and E, the fraction of H3.3WT-Venus that remains mobile in G1-phase cells was not significantly different in cells depleted either for ASF1A or for ASF1B and in untreated cells (compare Fig. 4G, [66% ± 1%], Fig. 6D [A-I, 64% ± 1%; n = 43] [A-II, 67% ± 1%; n = 45], and Fig. 6E [B-I, 65% ± 1%]; n = 49] [B-II, 68% ± 1%; n = 48]) (also see Fig. S3G and H in the supplemental material). In contrast, in G1-phase cells depleted for both ASF1A and ASF1B, the mobile fraction of H3.3WT-Venus was significantly higher (Fig. 6F) (see Fig. S3I in the supplemental material): 85% ± 2% (n = 49) for A-I and B-I and 82% ± 1% (n = 42) for A-II and B-II. The mobile fraction of H3.3WT-Venus in G1 cells depleted for both ASF1A and ASF1B resembled the mobile fraction of H3.3-3RA-Venus in untreated G1 cells (91% ± 2%) (Fig. 4H). These results suggested that G1-phase cells depleted for both ASF1A and ASF1B have a defect in the assembly of newly synthesized H3.3 into chromatin.
Although the phenotypes observed after depletion of proteins with two different siRNAs are usually attributed to impaired function of those proteins, it remains formally possible that they are due to off-target effects of the siRNAs. We therefore set out to confirm that the increased mobile fraction of H3.3WT-Venus in G1 cells depleted for hASF1 was indeed due to impairment of ASF1A and ASF1B function by restoring the expression of ASF1A. To this end, we first constructed an ASF1A-mCherry expression plasmid in which the ORF of ASF1A had been rendered insensitive to siRNA A-I by silent point mutations. Following the same experimental protocol, cells were depleted of hASF1 by siRNAs A-I and B-I, synchronized, and cotransfected with H3.3WT-Venus and ASF1A-mCherry constructs in a 1:2 ratio. This allowed for the identification of cells coexpressing both fusion proteins. All cells expressing H3.3WT-Venus were found to express ASF1A-mCherry to various extents (not shown). Surprisingly, when we performed FRAP experiments on those cells, we obtained two strikingly distinct populations: in the cells with a low level of ASF1A-mCherry fluorescence, where fluorescence was exclusively nuclear, the H3.3WT-Venus mobile fraction was 66% ± 3% (n = 14) (graph not shown), similar to what we observed for undepleted cells, and indicative of complementation of function by the ASF1A-mCherry construct. In contrast, in cells with a higher level of ASF1A-mCherry fluorescence, where fluorescence was both nuclear and cytoplasmic, the H3.3WT-Venus mobile fraction was 91% ± 2% (n = 13) (graph not shown), similar to what was observed for H3.3-3RA-Venus in undepleted cells. These results suggested that the overexpression of the ASF1A-mCherry fusion protein hindered ASF1A function. We were worried that the presence of mCherry itself could cause or aggravate the lack of complementation in cells expressing high levels of the fusion protein. Since we had observed that all cells expressing H3.3WT-Venus also expressed ASF1A-mCherry, we decided to introduce a stop codon between ASF1A and mCherry, so that the protein expressed after transfection of the construct was wild-type ASF1A only. As shown in Fig. 6G and in Fig. S3J in the supplemental material, using this construct and the same experimental conditions, a single-cell population was observed in the FRAP experiments, and the H3.3WT-Venus mobile fraction was 65% ± 1% (n = 55). Figure 6H shows the extent of hASF1 depletion and of wild-type ASF1A reexpression. This result strongly suggests that wild-type ASF1A is able to complement the depletion of both ASF1A and ASF1B for the incorporation of H3.3WT-Venus in chromatin in G1 cells. To further confirm that this was due to a function of ASF1A linked to its interaction with histone H3, we performed the same experiment with the ASF1A-V94R mutant. Figure 6H shows the extent of ASF1A-V94R expression in hASF1-depleted cells. As shown in Fig. 6G and in Fig. S3K in the supplemental material, the mobile fraction in this assay was 82% ± 1% (n = 50), indicating that histone binding was required for ASF1A activity in G1. The H3.3WT-Venus mobile fraction was 66% ± 2% (n = 25) (graph not shown) in undepleted G1 HeLa cells where ASF1A-V94R was expressed, showing that ASF1A-V94R had no dominant-negative activity.
The ASF1A-HIRA interaction has been characterized, and several mutations that disrupt the ASF1A-HIRA interaction have been described. In particular, Daganzo et al. have shown that the ASF1A-H36R+D37R does not bind HIRA in vitro (8). Recently, Tang et al. carried out structural characterization of the HIRA-ASF1A interface and showed that the single D37A mutation was sufficient to disrupt interaction of ASF1A with a minimal interaction domain of HIRA in vitro and also showed that E39 was involved in the interaction (55). We have previously reported that the S. cerevisiae Asf1-D37R+E39R mutant has in vivo silencing defects but is still able to bind histone H3 (32). These phenotypes are most likely attributable to lack of Hir binding. Altogether, these results strongly suggest that the ASF1A-D37R+E39R mutant is unable to bind HIRA. We therefore decided to assay the complementation of hASF1 depletion by the ASF1A-D37R/E39R mutant. As shown in Fig. 6G and in Fig. S3K in the supplemental material, the mobile fraction of H3.3WT-Venus in cells reexpressing ASF1A-D37R+E39R (Fig. 6H) was 70% ± 1% (n = 43), showing that this mutant is able to complement hASF1 depletion, albeit slightly less efficiently than wild-type ASF1A (this may be due in part to lower expression of this mutant compared to the wild type [Fig. 6H]). This result strongly suggests that HIRA binding is not required for ASF1A function in H3.3 deposition in G1 cells.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The hASF1-H3 interaction involves C-terminal arginines of H3. The C-terminal arginine 129 together with at least one of the two arginines (arginine 134 and 128) of H3 are involved in the interaction of tagged H3 with hASF1 (Fig. 1). These residues are located near the interface between the two H3 polypeptides in the nucleosome. However, they do not directly participate in the interface or in H4 binding, which are therefore unlikely to be affected. Indeed, even though H3-R129E and H3-3RA are not efficiently integrated into chromatin (Fig. 3 and 4), they are slowly deposited by less efficient deposition pathways in HeLa cells (Fig. 5). Had the mutations severely impaired the formation of the H3/H3 interface or of the H3/H4 dimer, we would have expected a measurable difference with the dynamic behavior of wild-type H3. Instead, we found that these mutants behaved dynamically like wild-type H3 once incorporated into chromatin (Fig. 4C and 5B and C), supporting the lack of an important role for these arginines in the stability of the assembled nucleosome.
Very recently, we and others published structures of the ASF1-H3 or ASF1-H3/H4 interfaces (2, 4, 10, 38). These structures showed an extensive interface between ASF1 and H3, and they revealed the structural basis for the binding of ASF1 to a heterodimeric complex of H3/H4. Interestingly, the structures clearly show an important electrostatic interaction between ASF1-D54 and H3-R129, whereas H3-R134 makes less important contributions to the interaction and H3-R128 is not involved. This is in complete accord with the results of our mutational analyses. Remarkably, the R129E and 3RA mutants of H3.1 and H3.3 still bind CAF-1 and HIRA, respectively. Thus, these chaperones must interact with histone H3 through a distinct binding mode that is presumably involved in the specific association of CAF-I with H3.1 and HirA with H3.3 (54).
Implication of hASF1 in nucleosome assembly in vivo.
ASF1 has been implicated in both nucleosome assembly and disassembly (for a review, see reference 33). In vitro experiments have shown that ASF1 can synergize with CAF-1 and Hir proteins to facilitate SC and SI nucleosome assembly, respectively (13, 54). However, a recent study has shown that Xenopus ASF1 is dispensable for direct de novo histone deposition in either pathway in egg extracts (45). Furthermore, in vivo data supporting a nucleosome assembly role for ASF1 have been less clear. In budding yeast, the asf1 mutant is viable and does not appear to show major defects in nucleosome density (1, 44). In Drosophila and vertebrate cells, depletion of ASF1 slows progression through S phase, but the molecular basis of this inhibition is unknown (14, 15, 47, 48). Indeed, it was recently suggested that inhibition of replication fork progression in the absence of hASF1 could be due to inefficient dissociation of nucleosomes in front of the MCM helicase (14). Nevertheless, depletion of ASF1 in chicken DT-40 cells led to increased nuclease sensitivity of newly synthesized DNA, which is consistent with a role for ASF1 in DNA synthesis-coupled nucleosome assembly (47). In this work, we used fluorescence microscopy to study the roles of hASF1 isoforms in H3 deposition in HeLa cells. First, we showed that histone mutants unable to interact with hASF1 are not efficiently assembled into chromatin in vivo, even though they are still able to bind CAF-1 or HIRA. Second, we showed that simultaneous depletion of ASF1A and ASF1B prevented efficient DNA synthesis-independent assembly of wild-type H3.3-FP into chromatin. We did not determine the effect of ASF1A and ASF1B depletion on DNA synthesis-coupled deposition of H3.1 in S phase, because the known severe inhibition of S-phase progression in this condition would make it impossible to determine whether a defect of H3.1 assembly was a direct or indirect consequence of hASF1 depletion. We have also observed that depletion of hASF1 slightly slows progression through G1 phase in synchronized cells. It thus remains formally possible that the lower chromatin deposition of H3.3 in these cells is an indirect result of the slower progression through G1 phase. However, since the H3.3-3RA mutant is also deposited less efficiently and since the function of ASF1A in G1 requires histone binding, the simplest hypothesis is that hASF1 depletion impairs histone deposition, which in itself might be responsible for a modest lengthening of the G1 phase. Understanding the mechanisms of G1-phase lengthening and the consequences of hASF1 depletion in S phase on histone deposition will be the subject of future enquiries. Nonetheless, our combined results provide the most direct evidence to date for an in vivo role of ASF1A and ASF1B in nucleosome assembly in human cells.
Nucleosome assembly pathways in human cells. We identified H3.1 and H3.3 mutants that are defective in hASF1 interaction but that retain binding to CAF-1 and HIRA, respectively. Interestingly, although these mutants are not efficiently assembled into chromatin by either the SC or SI assembly pathway, they are deposited slowly into chromatin and end up associated with mitotic chromosomes. These results suggest that hASF1 is required for efficient nucleosome assembly in human cells but that inefficient assembly pathways that function with little or no hASF1 may exist. This residual nucleosome assembly may be due to an inefficient activity of CAF-1 and HIRA in the absence of hASF1 in HeLa cells. In this model, hASF1 could be thought of as a processivity factor for efficient nucleosome assembly. Alternatively, CAF-1 and HIRA may be totally inactive for chromatin assembly of newly synthesized H3 in the absence of hASF1, in which case the residual deposition of H3 would be due to distinct histone chaperones that are only able to inefficiently assemble newly synthesized H3 into chromatin. Finally, we cannot exclude the possibility that some assembly of our ectopically expressed H3-fluorescent proteins is due to their constitutive expression. Further work is needed to distinguish between these possibilities.
Both ASF1A and ASF1B participate in DNA synthesis-independent assembly of histone H3.3 into chromatin. Current models of SI assembly implicate a HIRA/ASF1A pathway that would not involve ASF1B. However, our data clearly show that depletion of both ASF1A and ASF1B is required to inhibit H3.3 deposition into G1-phase chromatin. This indicates that ASF1A and ASF1B have redundant functions in SI assembly. This result contrasts with the unique role of ASF1A in SAHF formation in primary human fibroblasts, where depletion of ASF1A alone is sufficient to impair the formation of the SAHF (61). SAHF formation would in this respect be mechanistically different from nucleosome assembly in cycling G1-phase cells. An interesting possibility is that both ASF1A and ASF1B are involved in SI nucleosome assembly in vivo in proliferating cells, whereas only ASF1A would be involved in nucleosome assembly/disassembly in quiescent or senescent cells. The evidence so far shows a strong preference of HIRA for ASF1A. Tang et al. have provided biochemical data showing that this specificity depends on the N-terminal and C-terminal regions of ASF1A that are less conserved in ASF1B (55). Our data suggest that HIRA binding is not required for the G1 deposition of H3.3 that depends on ASF1A, which could explain why either ASF1A or ASF1B is sufficient for G1 deposition of H3.3. Thus, even though hASF1 seems to be important in the efficiency of histone deposition in S phase as a processivity factor and/or a histone donor for CAF-1, it seems that its function in G1 deposition of H3.3 is largely independent of HIRA. Whether hASF1 functions on its own as a bona fide histone deposition factor in G1 cells or whether it acts through an as yet unidentified factor will be the subject for future investigations.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by grants 4470 (C.M.) and 3828 (F.O.) from the Association de Recherche contre le Cancer and by the CEA/Institut Curie Programme Incitatif et Coopératif on Epigenetic Parameters in DNA Damage Response and the Cell Cycle.
|
FOOTNOTES |
|---|
Published ahead of print on 31 March 2008.
Supplemental material for this article may be found at http://mcb.asm.org/.
Present address: INSERM and Université Denis Diderot-Paris7, UMR S 741, Paris F-75005, France.
These two authors contributed equally to this work.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2. Agez, M., J. Chen, R. Guerois, C. van Heijenoort, J. Y. Thuret, C. Mann, and F. Ochsenbein. 2007. Structure of the histone chaperone ASF1 bound to the histone H3 C-terminal helix and functional insights. Structure 15:191-199.[Medline]
3. Ahmad, K., and S. Henikoff. 2002. The histone variant H3.3 marks active chromatin by replication-independent nucleosome assembly. Mol. Cell 9:1191-1200.[CrossRef][Medline]
4. Antczak, A. J., T. Tsubota, P. D. Kaufman, and J. M. Berger. 2006. Structure of the yeast histone H3-ASF1 interaction: implications for chaperone mechanism, species-specific interactions, and epigenetics. BMC Struct. Biol. 6:26.[CrossRef][Medline]
5. Bravo, R., and H. Macdonald-Bravo. 1987. Existence of two populations of cyclin/proliferating cell nuclear antigen during the cell cycle: association with DNA replication sites. J. Cell Biol. 105:1549-1554.
6. Chow, C. M., A. Georgiou, H. Szutorisz, A. Maia e Silva, A. Pombo, I. Barahona, E. Dargelos, C. Canzonetta, and N. Dillon. 2005. Variant histone H3.3 marks promoters of transcriptionally active genes during mammalian cell division. EMBO Rep. 6:354-360.[CrossRef][Medline]
7. Cremisi, C., and M. Yaniv. 1980. Sequential assembly of newly synthesized histones on replicating SV40 DNA. Biochem. Biophys. Res. Commun. 92:1117-1123.[CrossRef][Medline]
8. Daganzo, S. M., J. P. Erzberger, W. M. Lam, E. Skordalakes, R. Zhang, A. A. Franco, S. J. Brill, P. D. Adams, J. M. Berger, and P. D. Kaufman. 2003. Structure and function of the conserved core of histone deposition protein Asf1. Curr. Biol. 13:2148-2158.[CrossRef][Medline]
9. Earnshaw, W. C., B. M. Honda, R. A. Laskey, and J. O. Thomas. 1980. Assembly of nucleosomes: the reaction involving X. laevis nucleoplasmin. Cell 21:373-383.[CrossRef][Medline]
10. English, C. M., M. W. Adkins, J. J. Carson, M. E. Churchill, and J. K. Tyler. 2006. Structural basis for the histone chaperone activity of Asf1. Cell 127:495-508.[CrossRef][Medline]
11. Gaillard, P. H., E. M. Martini, P. D. Kaufman, B. Stillman, E. Moustacchi, and G. Almouzni. 1996. Chromatin assembly coupled to DNA repair: a new role for chromatin assembly factor I. Cell 86:887-896.[CrossRef][Medline]
12. Green, C. M., and G. Almouzni. 2003. Local action of the chromatin assembly factor CAF-1 at sites of nucleotide excision repair in vivo. EMBO J. 22:5163-5174.[CrossRef][Medline]
13. Green, E. M., A. J. Antczak, A. O. Bailey, A. A. Franco, K. J. Wu, J. R. Yates III, and P. D. Kaufman. 2005. Replication-independent histone deposition by the HIR complex and Asf1. Curr. Biol. 15:2044-2049.[CrossRef][Medline]
14. Groth, A., A. Corpet, A. J. Cook, D. Roche, J. Bartek, J. Lukas, and G. Almouzni. 2007. Regulation of replication fork progression through histone supply and demand. Science 318:1928-1931.
15. Groth, A., D. Ray-Gallet, J. P. Quivy, J. Lukas, J. Bartek, and G. Almouzni. 2005. Human Asf1 regulates the flow of S phase histones during replicational stress. Mol. Cell 17:301-311.[CrossRef][Medline]
16. Hake, S. B., B. A. Garcia, E. M. Duncan, M. Kauer, G. Dellaire, J. Shabanowitz, D. P. Bazett-Jones, C. D. Allis, and D. F. Hunt. 2006. Expression patterns and post-translational modifications associated with mammalian histone H3 variants. J. Biol. Chem. 281:559-568.
17. Hoek, M., and B. Stillman. 2003. Chromatin assembly factor 1 is essential and couples chromatin assembly to DNA replication in vivo. Proc. Natl. Acad. Sci. USA 100:12183-12188.
18. Houlard, M., S. Berlivet, A. V. Probst, J. P. Quivy, P. Hery, G. Almouzni, and M. Gerard. 2006. CAF-1 is essential for heterochromatin organization in pluripotent embryonic cells. PLoS Genet. 2:e181.[CrossRef][Medline]
19. Hu, C. D., and T. K. Kerppola. 2003. Simultaneous visualization of multiple protein interactions in living cells using multicolor fluorescence complementation analysis. Nat. Biotechnol. 21:539-545.[CrossRef][Medline]
20. Ishimi, Y., J. Hirosumi, W. Sato, K. Sugasawa, S. Yokota, F. Hanaoka, and M. Yamada. 1984. Purification and initial characterization of a protein which facilitates assembly of nucleosome-like structure from mammalian cells. Eur. J. Biochem. 142:431-439.[Medline]
21. Kaufman, P. D., R. Kobayashi, N. Kessler, and B. Stillman. 1995. The p150 and p60 subunits of chromatin assembly factor I: a molecular link between newly synthesized histones and DNA replication. Cell 81:1105-1114.[CrossRef][Medline]
22. Kimura, H., and P. R. Cook. 2001. Kinetics of core histones in living human cells: little exchange of H3 and H4 and some rapid exchange of H2B. J. Cell Biol. 153:1341-1353.
23. Krude, T. 1995. Chromatin assembly factor 1 (CAF-1) colocalizes with replication foci in HeLa cell nuclei. Exp. Cell Res. 220:304-311.[CrossRef][Medline]
24. Laskey, R. A., and W. C. Earnshaw. 1980. Nucleosome assembly. Nature 286:763-767.[CrossRef][Medline]
25. Le, S., C. Davis, J. B. Konopka, and R. Sternglanz. 1997. Two new S-phase-specific genes from Saccharomyces cerevisiae. Yeast 13:1029-1042.
p150CAF-1) (red) in asynchronous HeLa cells. Transiently transfected H3.1-Venus was revealed with anti-GFP antibodies (
) are indicated on the gels. An anti-glyceraldehyde-3-phosphate dehydrogenase (anti-GAPDH) Western blot on the same membrane is shown as a loading control.
