The Fanconi anemia core complex associates with chromatin during S phase

Fanconi anemia (FA) is a genetic disease of cancer susceptibility marked by congenital defects, bone marrow failure, and myeloid leukemia.^1-4  To date at least 11 complementation groups have been defined^5-7  and 8 genes have been cloned.^8-17  However, the gene products resemble no known proteins and have few identifiable functional protein motifs.

Cells derived from patients with the disease exhibit characteristic hypersensitivity to DNA cross-linking agents and generalized decreased survival.^18-22  In addition, a well-described G₂ phase cell cycle delay has been described that is thought to be secondary to a defective S or G₂ checkpoint.^23-25  Other studies have implicated cytokine dysregulation, sensitivity to oxidative damage, and defects in DNA repair.^23,26-29  However, no defined biochemical mechanism for this hypersensitivity has been elucidated. Patient and cellular phenotypes across all the complementation groups are similar, suggesting an inter-relatedness or cooperativity between the FA proteins.

This cooperativity has been borne out by work we have done in showing binding of FANCA and FANCC in a core protein complex in both nucleus and cytoplasm.^30-32  Recent work has found the FANCG, FANCE, FANCF, and FANCL proteins in the complex as well.^17,33-36  One group has shown that BRCA2/FANCD1 binds to FANCG.³⁷  A large complex is suggested by our recent work,³⁸  and formation of the core complex does not occur in any of the complementation groups except the FA-D1, FA-D2, FA-I, and FA-J groups.⁷ 

Our work has established that FA proteins bind to chromatin.³⁹  Immunoblotting experiments revealed that the increased FA proteins were bound to chromatin after DNA damage and that the complex underwent egress from the nucleus at mitosis. This movement appears to be regulated, as the FANCG protein becomes doubly phosphorylated at mitosis yet remains bound to the core complex.³⁹  Recent work has implicated FANCD2 in chromatin, reinforcing the importance of the FA core complex, since it is required for FANCD2 monoubiquitination.^40,41 

In this paper we take those observations further in an attempt to better define at the level of chromatin (1) the dynamics of movement of the complex by microscopy during the cell cycle, (2) how the complex responds to DNA damage, and (3) the manner in which the complex localizes to chromatin fibers. In order to look exclusively at chromatin, we strip away soluble nuclear components and cytoplasmic proteins by in situ chromatin preparations. In addition, we demonstrate true localization to chromatin fibers by vertical extraction in a urea buffer, which causes release and elongation of fibers. Our data suggest that the FA protein complex achieves a focal presence on chromatin by G₁-S, which is accentuated during S phase, and that the increase in FA complex caused by DNA damage is likely a manifestation of the S-phase concentration of FA complex.

Cells were grown at 37°C in a 5% CO₂ incubator. FA fibroblast cell lines were grown in Dulbecco modified Eagle medium (DMEM) supplemented with 15% (vol/vol) fetal bovine serum (FBS). HeLa cells were grown in DMEM with 10% FBS. FA-A primary mutant cells (GANO; gift of Hans Joenje, Free University, The Netherlands) were grown in F12 nutrition medium and 15% FBS. HSC536N (FA-C) and EUFA143 (FA-G) mutant lymphoblasts were grown in RPMI 1640 and 15% FBS.

The plasmid pECFP (cyano fluorescent protein)–FANCG was constructed with the FANCG gene fragment from pMMP-FANCG through the BglII and EcoRI site generated by polymerase chain reaction (PCR). The plasmid pEGFP (green fluorescent protein)–FANCA was constructed with the FANCA gene fragment from pcDNA3-Flag-FANCA through the HindIII and EcoRI site generated by PCR. The plasmid pEYFP (yellow fluorescent protein)–FANCC was constructed with the FANCC gene fragment from pLXSN-FANCC through KpnI and BamHI site generated by PCR.

To generate plasmid single-point mutations, PCR was performed with site-directed mutagenesis kit (Quickchange, Stratagene, La Jolla, CA). pEGFP-FANCA (H1110P), pECFP-FANCG (G546R), and pEYFP-FANCC (L554P) were constructed, respectively, by using following pairs of primers: 5′-GCAGTTCTTCCCCTTGGTCA-3′ and 5′-GTTGACCAAGGGGAAGAACT-3′, 5′-GTGCCCACGTAATCGAGACA-3′ and 5′-CTCGATTACGTGGGCACATC-3′, 5′-TTAAAGAGCCGCGAACTCAAG-3′ and 5′-TTGAGTTCGCGGCTCTTTAAG-3′. Resulting constructs were verified by DNA sequencing.

Transfections were performed by lipofection per manufacturer's protocol (Lipofectamine; Gibco, Carlsbad, CA). For all transient and stable transfection, plasmid DNA was prepared by column purification (Qiagen, Valencia, CA). 5 × 10⁵ cells were placed on 60-mm petri dishes one day prior to transfection and grown until the cells were 60% to 80% confluent. Plasmid DNA was diluted in 100-μL Opti-MEM medium (Gibco) and mixed with plus reagent (Gibco); LipofectAMINE was diluted in 100 μL OPTI-MEM medium (Gibco). These were gently mixed together 15 minutes after incubation at room temperature, and incubated another 15 minutes at room temperature. The final complex was placed on prewashed cells with phosphate-buffered saline (PBS) and incubated for 5 hours.

HeLa cells were treated overnight with 2 mM thymidine, washed, released into regular media, and treated again overnight with thymidine, as previously described.³⁹  Cells were then released for variable times or incubated overnight in the presence of 1 μM nocodazole (Sigma, St Louis, MO). Mitotic cells were collected the next morning by shaking the plate and collecting the cells in suspension. Verification of cell cycle state was achieved using FACScan (Becton Dickinson, San Jose, CA) analysis after resuspension of 100 000 cells in 0.3% citrate, 0.01% NP40, 100 μg/mL propidium iodide, and 10 μg/mL RNAse.³⁹ 

Asynchronous cells or cells synchronized at G₁-S border or S phase were incubated with mitomycin (MMC) (0.1 μM) for 4 hours. For the time course, asynchronous cells were incubated with MMC for 0 hours, 6 hours, 12 hours, and 24 hours. For transiently transfected HeLa cells, 48 hours after transfection, cells were treated for 12 hours with MMC.

For sorting, 10⁶ asynchronously growing HeLa cells transfected with pEGFP-FANCA and pECFP-FANCG were collected and pelleted. The cells were resuspended sequentially in 500 μL PBS and 500 μL 2% paraformaldehyde and incubated on ice for 1 hour. After pelleting, the cells were washed and then incubated overnight at 4°C in 70% ethanol. The cells were pelleted again and resuspended in 1 mL of 40 μg/mL propidium iodide and 100 μg/mL RNAse in PBS. This mix was incubated at 37°C for 30 minutes. The cells were then sorted into G₁, S, and G₂-M groups on the FACS instrument (Becton Dickinson, San Jose, CA).

Cells grown on chamber slides were first washed in cold PBS. Soluble proteins were removed by extraction in cytoskeletal buffer [CSK: 10 mM PIPES (piperazine diethanesulfonic acid), pH 6.8/300 mM sucrose/100 mM NaCl/3 mM MgCl₂/1 mM EGTA (ethylene glycol tetraacetic acid)/RNAse inhibitor (Gibco)/1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride] containing 0.5% Triton X-100 for 2 minutes at 4°C. The structure remaining was extensively cross-linked by treatment with 4% formaldehyde in CSK for 40 minutes at 4°C. The slides were covered with mounting solution containing 2 ug/mL DAPI (4′,6-diamidino-2-phenylindole; Vector Laboratories, Burlingame, CA) after washing with cold PBS.^39,42 

Chromatin fibers were prepared according to the method of Blower et al.⁴³  Cells were trypsinized 48 hours after transfection, washed twice with cold PBS, then suspended in hypotonic solution (75 mM KCl in PBS) for 10 minutes. Cells adjusted to a density of 1 to 2 × 10⁵ cells/mL were cytospun onto charged slides (double funnels) at 800 rpm (150g) with high acceleration for 4 minutes, 150 μL cells per slide. For attached/monolayer cells, 1 × 10⁵ cells/mL were used, and for suspension cultures, 2 × 10⁵ cells/mL. After cytospin, slides were immersed in a Coplin jar filled with lysis buffer [25 mM Tris (tris(hydroxymethyl)aminomethane) Cl, pH 7.5/0.5 M NaCl/1% Triton X-100/0.2 M urea for 3 to 10 minutes. The slide was slowly lifted vertically from the lysis buffer, and the chromatin was allowed to move down the slide. After fixation in 4% paraformaldehyde in PBS-Tween20 (0.05%) for 20 minutes, the slides were covered with mounting solution containing 2 ug/mL DAPI (4′6-diamidino-2-phenylindole 2HCl) (Vector Laboratories) after wash in PBS. For fluorescent microscopy and analysis, inverted fluorescent microscope (Nikon, Melville, NY) was used, with images captured by digital camera (Hamamatsu, Hamamatsu City, Japan).

Chromatin extract was prepared and run on sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), as previously described, essentially a modification of the above-described in situ preparation.^39,42  Immunoblotting was conducted with anti-FANCG(N terminal), anti-FANCA(N terminal), and Ku86 (Santa Cruz Biotechnology, Santa Cruz, CA), respectively.³⁹ 

For the purpose of visualizing FANCA, FANCG, and FANCC by direct fluorescent microscopy, we constructed FA cDNAs with fluorescent tags: EGFP-FANCG, EYFP-FANCC, and ECFP-FANCG. In order to determine if the fluorescent versions of the FANCA, FANCC, and FANCG were functional, we transfected the fluorescent and nonfluorescent tagged constructs into FA-A (HSC72), FA-C (HSC536N), and FA-G (EUFA143) mutant cells, respectively. After transfection the cells were analyzed for cytotoxicity against MMC. In Figure 1, each version of the FA genes was able to restore normal MMC resistance in the corresponding mutant line, indicating that the fluorescent tag did not interfere with normal FA function. Thus, use of these constructs is valid in experiments to explore the normal function of the FA proteins.

In previous work we showed by immunoblotting that the endogenous FA core complex localized to chromatin. To determine if tagged versions of the FA proteins localized to chromatin, we doubly transfected HeLa cells with either EGFP-FANCA and ECFP-FANCG or ECFP-FANCG and EYFP-FANCC. After 48 hours, we extracted the cells in situ to leave behind what we term a “chromatin prep.” Direct microscopy of either whole cells or chromatin prepared in situ revealed that FANCA, FANCC, and FANCG all localized to the nucleus and chromatin and totally colocalized (Figure 2A). Most expressing cells contained fluorescent protein in both cytoplasm and nucleus. No chromatin binding was observed in control experiments for cells expressing only EGFP, ECFP, or EYFP (Figure 2B). Transfection efficiency was estimated to be at least 50% in all cases, as determined by counting fluorescent cells prior to extraction. In order to show that mutation of the FA proteins could disrupt proper chromatin localization, we prepared similarly tagged versions of FANCA (H1110P), FANCC (L554P), and FANCG (G546R) by PCR-mediated mutagenesis. Each of these mutants exist in nature in FA patients, and the derivative proteins are expressed at near wild-type levels.^8,44,45  The FANCA (H1110P), FANCC (L554P), and FANCG (G546R) mutants all were cytoplasmic and were not visible upon chromatin preparation (Figure 2C). Thus, patient-derived point mutations resulted in abrogation of nuclear or chromatin localization, further validating the use of our tagged constructs.

Our previous work did not show any marked change of FA protein amount bound to chromatin during S phase, as compared to the G₁-S border.³⁹  In order to show qualitative changes in the FA core complex, we synchronized HeLa cells by double thymidine block, then released into regular media. Chromatin preparations at the G₁-S border and during S phase revealed that the FA core complex formed foci, which were not readily apparent in whole cells (Figure 3A). In all cases, FANCA and FANCG as well as FANCG and FANCC colocalized. Interestingly, cells from G₁-S to mitosis contained predominately fluorescent FA proteins in the nucleus. The whole cells in the micrographs were underexposed in order to diminish cytoplasmic staining and increase visibility of the nucleus. As the cells continued into G₂, the foci appeared to diffuse and migrate to the nuclear periphery by late G₂ (Figures 3B-D). This also was apparent in whole cells. Cells that became detached from the plate after nocodazole arrest, assumed to be in mitosis, were collected and cytospun onto slides. Examination of these cells revealed complete exclusion from nuclei (prior to chromosome condensation) in the whole cell as well in chromatin preps (Figure 3E). Once cells are released from G₁-S block, approximately 10 to 12 hours later the cells are “shakable” from the surface of the plate and represent mitotic cells. These cells were collected and allowed to reattach to the slide surface for 4 hours. These cells, seen in Figure 3F, had divided, but the FA proteins were still cytoplasmic. These data are consistent with our previous work in which we showed the FA complex is excluded from condensed chromosomes at mitosis. In early G₁ it is apparent that the complex is still cytoplasmic. Representative FACS analyses from HeLa cells and EGFP-FANCA/ECFP-FANCG are depicted in Figure 3G; no differences in DNA histograms were seen in any of the transfectants and HeLa cells alone. Cell number at this stage was no different in nontransfected and transfected cells.

In order to verify the phenomenon of protein shuttling in asynchronously growing cells, we sorted by flow HeLa cells cotransfected with GFP-FANCA and CFP-FANCG into G₁, S, and G₂-M groups. Microscopy revealed that FANCA and FANCG were cytoplasmic and colocalized in G₁ and G₂-M cells, while in S phase cells they were predominately nuclear (Figure 3H). This is consistent with the idea that FA proteins shuttle between nucleus and cytoplasm during the cell cycle and that synchronization does not affect this event.

To demonstrate that synchronization does not prevent cells from proceeding to the next cell cycle, we collected HeLa cells transfected with EGFP-FANCA and ECFP-FANCG and arrested at mitosis. These cells were washed and then placed in fresh regular media and observed under time-lapse photography. The film shows that initially FANCA and FANCG are cytoplasmic. However, after 6 hours, FANCA and FANCG can be observed in the nucleus, mostly at the periphery (Supplementary Video 1 [see the Video link at the top of the online article on the Blood website]). FACS (fluorescence-activated cell scans) analysis demonstrates that the cells are at the G₁-S border (Figure 3G). Time-lapse photography taken of asynchronous cells demonstrates a similar phenomenon: initially cytoplasmic proteins become nuclear after 2 hours of observation (Supplementary Video 2, also available on the Blood website).

Chromatin preparations afford the ability to see structures beyond soluble nuclear material, yet chromatin still represents a densely packed version of the genome. In order to visualize distinct chromatin fibers, we extracted nuclei by vertical immersion in a urea-based extraction buffer after cotransfection with EGFP-FANCA and ECFP-FANCG. In the asynchronous cells, colocalizing FANCA and FANCG were seen in irregularly spaced foci along the length of the fiber (Figure 4). In contrast, upon G₁-S and S-phase entry, the proteins could be seen slightly more prominently but more homogeneously spread along the fiber. At G₂, the proteins appeared to be more clumped. Strikingly and consistent with our other data, a chromatin fiber prep performed on mitotic cells revealed the complete absence of FA proteins, while the DAPI-stained image showed predictably a heterochromatically staining fiber.

Our previous work showed that increased FA proteins localized to chromatin after MMC treatment. To demonstrate this phenomenon in situ, we treated an asynchronous cell population with 0.1 μM MMC and examined chromatin and whole cells at various time points of MMC treatment. No obvious differences can be seen in whole cells (Figure 5A). However, by 12 hours, foci seen in untreated chromatin had become diffusely stained in MMC-treated chromatin (Figure 5B). We analyzed a slide containing cells similarly treated with MMC and performed chromatin fiber preparation. Increased and more-spread FA proteins stained the fibers (Figure 5C), consistent with the in situ appearance in Figure 5B and the immunoblotting data of our prior work.³⁹ 

Because HeLa cells are easily synchronized, we have used them for the bulk of the work presented. In order to show that these data are applicable to primary cells as well, we transfected primary FA-A mutant cells with EGFP-FANCA and ECFP-FANCG. In the untreated cells, analysis of both whole cells and chromatin revealed that FANCA and FANCG colocalized to foci in the nucleus of whole cells (Figure 6A) and chromatin (Figure 6B). Cells treated with MMC displayed increased chromatin staining, consistent with the data presented in the HeLa cells. Even in the whole cell, the nucleus can be seen homogeneously stained after MMC treatment. Thus, for FA core complex function, behavior of the complex in immortalized and primary cells appears similar.

Chromatin fiber experiments and synchronization experiments suggest similarities in appearance of FA proteins during S phase and after MMC treatment. This suggests that the DNA damage inducibility of the FA protein localization on chromatin may be a function of engagement of an S-phase checkpoint. This has been suggested by work done by Akkari et al,²⁹  who showed that cytotoxicity induced by MMC was dependent on traverse through S phase.

In order to compare the magnitude of MMC effect upon chromatin localization of the FA proteins, we treated HeLa cells synchronized to the G₁-S border and asynchronous HeLa cells with MMC. S-phase cells were prepared by release into regular media for 4 hours; cells were treated with MMC during this time. After chromatin extract preparation, we performed FANCA immunoblotting. Unexpectedly, peak levels of FANCA in chromatin were seen during S phase, which were greater than the increased levels induced by MMC in asynchronous cells (Figure 7A). Concomitant treatment with MMC during S phase did not stimulate higher FA protein levels in chromatin. A negative control lane with 100 μg of whole cell extract of EUFA143 FA-G mutant cells was included, as these cells do not express FANCG and express little FANCA. FACS data showed that MMC did not alter the DNA histogram in synchronized cells but, as has been well documented, did induce a marked degree of S (12 hours) and G₂-M (24 hours) accumulation (Figure 7B). These data suggest that the MMC inducibility of FA proteins to chromatin is achieved through traverse in S phase and is consistent with the data from Akkari et al.²⁹ 

We previously have implicated the chromatin-nuclear matrix as a functional unit upon which the FA core complex performs its normal function.³⁹  In this paper, we confirm and extend these findings by showing that this is an S-phase–specific process by (1) demonstrating that the FA core complex forms chromatin foci at the G₁-S border and at the beginning of S phase, (2) showing that these foci diffuse in response to DNA damage and during S phase, and (3) showing that the proteins exit the nucleus by the onset of mitosis as a multiprotein complex.

These data suggest a dynamic movement during the cell cycle whereby FA protein foci form on chromatin by G₁-S, the foci diffuse, and proteins increase in amount during S phase. This is recapitulated by the chromatin fiber experiments. The MMC treatment appears to induce a similar change in both chromatin foci as well as in fibers, suggesting that the response to DNA damage and during S phase may be similar. This is supported by literature from the Grompe and Rosselli groups suggesting the importance of the FA pathway in an S-phase process as well as from the D'Andrea group showing that FANCD2 is mono-ubiquitinated during S phase and in response to DNA damage.^29,46-49 

These data have implications for potential function of the FA core complex. First, a specific function for the FAproteins remains obscure in spite of the cloning of 8 genes. Second, the emergence of chromatin regulation as important to genomic stability has become part of the forefront of current research, exemplified by SWI/SNF and by chromatin remodeling activities of cancer related proteins such as core binding protein (CBP) and BRCA1.⁵⁰  Examples of regulation and involvement of histone modification have been shown to affect transcription, replication, and DNA repair. Indeed, although no associated chromatin remodeling activity could be detected, FANCA has been shown to bind to human SWI/SNF.⁵¹  Also, recent work has shown that BRCA1 binds to FANCD2 and that BRCA2 is indeed FANCD1.^48,52  Since FANCD2 ubiquitination is dependent upon wild-type FA core complex, and this ubiquitination occurs in response to DNA damage, it is indeed plausible that the genomic surveillance function of the FA pathway occurs at the level of chromatin. Recently, FANCD2 and its importance in chromatin has been supported by work detailing that mono-ubiqutinated FANCD2 localizes to chromatin and is responsible for BRCA2 loading onto chromatin.^40,41 

Our future studies now will focus on a biochemical function for the FA pathway. These studies suggest a link between replication and repair of DNA damage. Our prior work detailing MMC induction of FA protein to chromatin did not compare directly MMC effect versus S phase. Our direct comparison implies that the 2 are likely inseparable.

Localization to chromatin implies that these biochemical activities may play a role directly in DNA repair, DNA replication, and other genomic surveillance activities. While direct DNA binding of the FA core complex has not been shown, it is plausible that the FA core complex can interact with the functional unit of DNA, namely chromatin. The recent report of Bloom helicase (BLM) binding with the FA core complex suggests work to show how the chromatin interaction occurs.⁵³  We will attempt to show colocalization with BLM and other S-phase proteins in situ, including those involved in recombinatorial repair, replication machinery, and other components of chromatin such as histones.

We thank Han Joenje for the GANO cell line and Maureen Hoatlin and members of the Kupfer laboratory for useful discussions.