The forkhead transcription factor FoxI1 remains bound to condensed mitotic chromosomes and stably remodels chromatin structure - PubMed (original) (raw)

The forkhead transcription factor FoxI1 remains bound to condensed mitotic chromosomes and stably remodels chromatin structure

Jizhou Yan et al. Mol Cell Biol. 2006 Jan.

Abstract

All forkhead (Fox) proteins contain a highly conserved DNA binding domain whose structure is remarkably similar to the winged-helix structures of histones H1 and H5. Little is known about Fox protein binding in the context of higher-order chromatin structure in living cells. We created a stable cell line expressing FoxI1-green fluorescent protein (GFP) or FoxI1-V5 fusion proteins under control of the reverse tetracycline-controlled transactivator doxycycline inducible system and found that unlike most transcription factors, FoxI1 remains bound to the condensed chromosomes during mitosis. To isolate DNA fragments directly bound by the FoxI1 protein within living cells, we performed chromatin immunoprecipitation assays (ChIPs) with antibodies to either enhanced GFP or the V5 epitope and subcloned the FoxI1-enriched DNA fragments. Sequence analyses indicated that 88% (106/121) of ChIP sequences contain the consensus binding sites for all Fox proteins. Testing ChIP sequences with a quantitative DNase I hypersensitivity assay showed that FoxI1 created stable DNase I sensitivity changes in condensed chromosomes. The majority of ChIP targets and random targets increased in resistance to DNase I in FoxI1-expressing cells, but a small number of targets became more accessible to DNase I. Consistently, the accessibility of micrococcal nuclease to chromatin was generally inhibited. Micrococcal nuclease partial digestion generated a ladder in which all oligonucleosomes were slightly longer than those observed with the controls. On the basis of these findings, we propose that FoxI1 is capable of remodeling chromatin higher-order structure and can stably create site-specific changes in chromatin to either stably create or remove DNase I hypersensitive sites.

PubMed Disclaimer

Figures

FIG. 1.

FIG. 1.

Generation of doxycycline (Dox)-inducible FoxI1-GFP and FoxI1-V5 stable cell lines. (A) Constructs for Dox-inducible FoxI1-GFP or FoxI1-V5 used for generation of FoxI1-GFP and FoxI1-V5 PAC2 cell lines. (B) FoxI1-GFP Pac2 cell line in the absence of Dox (−Dox). (C) FoxI1-GFP Pac2 cell line in the presence of Dox (+Dox) visualized by GFP fluorescence. (D) FoxI1-V5 cell line in the absence of Dox. (E) FoxI1-V5 cell line in the presence of Dox, visualized by anti-V5 antibodies. (F) Confocal images of FoxI1-V5 staining, results were similar for the GFP fusion (not shown). The lower right corner has a higher-magnification inset showing the punctate staining in more detail. (G) DAPI staining of the same cells. (H) Differential interference contrast (DIC) image of the same cells. (I) Merged image from panels F, G, and H.

FIG. 2.

FIG. 2.

FoxI1-GFP associates with mitotic chromatin at metaphase (left) and telophase (right). FoxI1-GFP cells were counterstained with DAPI and imaged with a confocal microscope. (A) FoxI1-GFP signal at metaphase. (B) DAPI staining of the same cell as shown in panel A. (C) DIC image of the cell in panel A. (D) Merged image of panels A, B, and C. Note that the GFP signal colocalizes with the DAPI signal in the condensed chromosomes. (E) FoxI1-GFP signal in a dividing cell in telophase. (F) DAPI staining of the telophase cell in panel E. (G) DIC image of the cell in panel E. (H) Merged image of panels E, F, and G. Similar results were seen with the FoxI1-V5 fusion (not shown).

FIG. 3.

FIG. 3.

PCR-coupled ChIP analyses of mitotic chromatin. Doxycycline (Dox)-induced (+Dox) and uninduced (−Dox) FoxI1-V5 cells were treated with nocodazole as described in Materials and Methods. Equal amounts of input DNA were subject to ChIP, and the enriched DNA was amplified using primers designed to amplify sequences identified in the subcloned ChIP fragments (T1 to T15, Ts). Data were compared to amplifications by primers designed to random genomic sequences (R1 to R16). The Ts primers were selected from type I satellite DNA, which flank the 186-bp repeat unit which creates a ladder when amplified. Nine of the 16 ChIP fragments amplified in the presence of FoxI1, while only 1 of 16 amplified in the random control, and this fragment only amplified in the presence of FoxI1, suggesting it may be a bona fide target of FoxI1 binding. Two of the ChIP-identified sequences amplified in the uninduced state, one of which was the satellite DNA (T9, Ts). These may represent high-affinity targets that are bound by a low level of FoxI1 that is detectable even in the uninduced state.

FIG. 4.

FIG. 4.

Confirmation of FoxI1-DNA binding by southwestern analysis of FoxI1 protein. _E. coli_-expressed FoxI1-V5, FoxI1+, and a FoxI1-null vector, FoxI1- (FoxI1 coding region was inserted into the opposite orientation), were induced with 0.02% arabinose. Protein extract (100 μg) was subject to SDS-PAGE and transferred to a nitrocellulose membrane. Probes were DIG labeled by PCR, amplifying the inserts from ChIP-cloned targets (T1, T2, T3, T4, and Ts) and random targets (R4, R10, and R11). DIG-labeled probes (10 μg) were hybridized with the bound proteins in the presence of 100 μg of sonicated herring sperm DNA and a 10-fold excess of competitor DNA (unlabeled probe DNA, 100 μg). The ChIP probes specifically recognized the thioredoxin-FoxI1-V5 fusion protein (predicted size of 60 kDa) and could be competed with unlabeled probe, while random probes (R4, R10, and R11) did not bind to FoxI1.

FIG. 5.

FIG. 5.

Strategy for SELEX immunoprecipitation coupled with a yeast inverse one-hybrid test system. (A) Enrichment of genomic DNA by FoxI1-V5 using SELEX immunoprecipitation. Lane 1, DNA marker; lane 2, sonicated genomic DNA; lane 3, PCR amplification of the first round of precipitated DNA; lane 4, PCR amplification of the third round of enriched DNA; lane 5, PCR amplification of the final selected DNA fragments. (B) Immunoblot detection of Gal4AD-HA-FoxI1 fusion proteins in yeast (+), with an empty vector used as a control (−). FoxI1 expression was detected by Western blotting using anti-HA-tagged antibody. (C) Construction of the URA3 reporter plasmid and the FoxI1 fusion protein. The EcoRI site was used for shotgunned insertion of test DNA fragments. The FoxI1 protein was fused to the transcription-activating domain of the yeast GAL4 protein to ensure transcriptional activation in yeast. The protein was constitutively expressed in the test strain. Three unique sequences were identified using this technique including type 1 satellite DNA.

FIG. 6.

FIG. 6.

Quantitative analyses of DNase I sensitivity by real-time PCR. (A and B) FoxI1-V5 cells were induced for 24 h or grown without induction. The intact nuclei were digested with increasing concentrations of DNase I (0.25 to 4 U). Primer sets were either designed using ChIP-isolated DNA sequences (ChIP primers) or primers randomly selected from zebra fish STS sequences (random primers). ΔCt indicates the number of additional cycles necessary to amplify the target to the CT compared to a 0.25-U DNase I control. The number at the top of each target indicates the Δ_CT_ derived by subtracting the CT at 0.25 U DNase I from the CT at 4 U of DNase I. A Δ_CT_ of 1 is equivalent to a twofold change in sensitivity, for example, T1 changes from 3.4 to 6.9 in the presence of FoxI1, this is equal to an ∼11-fold increase in sensitivity. (A) Expression of FoxI1 alters DNase I hypersensitivity in condensed chromatin treated with nocodazole. Mitotically arrested cells were separated by shake off of the culture after 16 h in nocodazole (400 ng/ml). The x axis indicates different specific genomic targets in the presence (+) or absence (−) of FoxI1 and ranges of DNase I concentration from 1 to 4 U. (B) Alterations in DNase I hypersensitivity in unsynchronized cells. Δ_CT_s are not as large under these conditions compared to condensed chromatin. (C) Δ_CT_s for all ChIP targets compared to the random targets in the absence and presence of FoxI1 expression. Note the shift to the left for both ChIP and random targets and the bimodal distribution of ChIP targets in the presence of FoxI1 (red line). Dox, doxycycline.

FIG. 7.

FIG. 7.

Nucleosome structural analysis in FoxI1-expressing cells. Nuclei were isolated from induced non-FoxI1 integration Pac2-tet-on cells (Pac2/+Dox), uninduced FoxI1-V5 cells (FoxI1/-Dox), and induced FoxI1-V5 cells (FoxI1/+Dox). (A) Isolated nuclei were incubated with the indicated concentrations (0.25, 1, or 2 units/ml) of MNase; 0 indicates incubation without nuclease. Purified DNA fragments were electrophoresed on a 1.4% agarose gel. Marker is the 1-kb Plus DNA ladder (Invitrogen). (B) MNase digests (2 μg each) were digested with NdeI and separated on a 1.4% agarose gel at 16°C, 70 V for 6 h, stained with ethidium bromide (EtBr), and after photography (shown in panel EtBr, on the left), blotted to nylon membranes and probed with 32P-labeled ChIP target T1 (right panel). The T1 target is 279 bp in length and contains 72-bp SINE repeat elements. The line passes through the center of the trioligonucleosome at 0.25 U/ml MNase in the absence of FoxI1 to facilitate visualization of the nucleosome size change.

FIG. 8.

FIG. 8.

Chromatin fractionation analysis. Isolated nuclei were digested with 2 units/ml of MNase for 10 min. Three fractions were obtained: S1, soluble first fraction; S2, soluble second fraction; P, the insoluble pellet fraction. Aliquots from each cell line were separated on a NuPAGE 12% Bis-Tris gel. M, SeeBlue Plus2 prestained standard (Invitrogen). (A) Gels were stained with SimplyBlue Safe Satin (Invitrogen), and the duplicate gel was immunoblotted with anti-V5 monoclonal antibody. FoxI1-labeled lanes have the induced FoxI1 protein. Linker H1 protein is clearly seen as a 40-kDa protein enriched in the S2 fraction in all three cell lines. FoxI1 was detectable in the S1 fraction (marked with black arrow), but the majority was present in the pellet (P). (B) An ethidium bromide-stained agarose gel showed that isolated DNA in the S2 fraction was low, relative to the S1 and P fractions. (C) The initial pellet fraction (P) was subjected to DNase I digestion and high-salt extractions, as described in Materials and Methods. The soluble extraction (P1) and insoluble fraction (P2) were examined using the ECL Western blot analysis system (Amersham Bioscience). A band detected by anti-V5 antibody in insoluble chromatin (P2) (marked with black arrow) was 62 kDa in size, and the untreated pellet fraction was 48 to 62 kDa (P). The mobility anomaly may be associated with the high-salt treatment of the chromatin.

Similar articles

Cited by

References

    1. Al-Awqati, Q., and G. J. Schwartz. 2004. A fork in the road of cell differentiation in the kidney tubule. J. Clin. Investig. 113:1528-1530. - PMC - PubMed
    1. Allan, J., P. G. Hartman, C. Crane-Robinson, and F. X. Aviles. 1980. The structure of histone H1 and its location in chromatin. Nature 288:675-679. - PubMed
    1. Azumi, K., M. Fujie, T. Usami, Y. Miki, and N. Satoh. 2004. A cDNA microarray technique applied for analysis of global gene expression profiles in tributyltin-exposed ascidians. Mar. Environ. Res. 58:543-546. - PubMed
    1. Bell, P. J., I. W. Davies, and P. V. Attfield. 1999. Facilitating functional analysis of the Saccharomyces cerevisiae genome using an EGFP-based promoter library and flow cytometry. Yeast 15:1747-1759. - PubMed
    1. Berezney, R. 1980. Fractionation of the nuclear matrix. I. Partial separation into matrix protein fibrils and a residual ribonucleoprotein fraction. J. Cell Biol. 85:641-650. - PMC - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources