SEARCH

SEARCH BY CITATION

Keywords:

  • chromatin structure;
  • histone exchange;
  • DNA intercalator;
  • minor groove binder;
  • DRAQ5;
  • Syto17;
  • fluorescence confocal microscopy

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED
  8. Supporting Information

DNA-binding dyes are useful in contrasting cell nuclei and chromatin for live cell fluorescence microscopy; however, they may interfere with nuclear and DNA structure and function. We investigated the influence exerted by two DNA dyes on chromatin and nuclear structure, as well as histone–DNA interactions in live HeLa cells. A membrane-permeant fluorescent DNA intercalator DRAQ5 (anthracycline derivative), at a concentration of 1 μM, caused microscopically detectable changes of nuclear architecture. Following DRAQ5 intercalation into DNA, chromatin aggregated into distinct areas and foci. The loss of 3D chromatin distribution was exerted via interference with a dynamic exchange of a linker histone (H1), which is a known chromatin stabilizing factor. At higher concentrations (3 and 7.5 μM), DRAQ5 interfered with binding of H2B core histones to DNA. Similar effects resulted from intercalation of chemotherapeutic drugs, adriamycin and daunomycin, but were not observed after binding to DNA of a minor groove binder, Syto17. © 2008 International Society for Advancement of Cytometry

Chromatin—a complex of DNA, histones, and nonhistone proteins—yields poor contrast in optical microscopy. In standard transmitted light mode, or even in DIC observation, chromatin is barely detectable. Throughout the cell division cycle, chromatin becomes clearly distinguishable only for a relatively short time, when chromosomes condense for mitosis. Thus, chromatin of live cells is commonly visualized using specific fluorescent probes that exhibit affinity to DNA. Most of these probes are intercalators or minor groove binders. A fluorescent probe bound to chromatin reveals the presence of spatial distribution and a local concentration of DNA in nuclei of live cells. However, binding of a probe to DNA may cause changes in chromatin or nuclear structure, which are dependent on the chemical nature and concentration of the probe. In the absence of an independent way of visualizing chromatin, such structural changes remain unknown. Availability of GFP-tagged histones for noninvasive visualization of chromatin in vivo opens a new window of opportunity for investigation of the influence of various DNA-binding dyes on nuclear and chromatin structure.

We used HeLa cells expressing eGFP-tagged linker (H1) and core (H2B) histones to investigate the influence of a membrane-permeant DNA intercalator, DRAQ5 (1), and a minor groove binder Syto17, on nuclear and chromatin structure. We describe concentration-dependent aggregation of chromatin and altered interactions between histones and DNA, induced by the bound DRAQ5. We also demonstrate that a minor groove binder Syto17 causes only barely detectable structural changes in the nuclear architecture and histone–DNA interactions.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED
  8. Supporting Information

Reagents

DRAQ5 was obtained from Biostatus (Shepshed, UK); Syto17 from Molecular Probes (Eugene, OR); fetal calf serum from Gibco (Gibco Products, Invitrogen Corp., Grand Island, NY) other cell culture reagents from Sigma.

Cells

HeLa cells expressing histones tagged with eGFP were kindly provided by Dr. T. Kanda (2), Dr. H. Kimura, and Prof. P. R. Cook (3). Cell cultures were maintained according to standard procedures.

Staining Procedure

DRAQ5 was added directly from the original stock solution to culture medium in a microincubator on a microscope stage. Syto17 solution was prepared from original stock by dilution in PBS, and subsequently added to medium.

Confocal Microscopy

Bio-Rad MRC1024 confocal system (Bio-Rad Microscience, Hemel Hempstead, UK) was interfaced with a Nikon Diaphot microscope (Nikon, Amsterdam, The Netherlands), 100-mW Ar laser (ILT, Salt Lake City, UT, USA) and KrAr laser (ALC, Salt Lake City, UT, USA). A 60× NA1.4 lens and a microincubator (Life Science Resources, Cambridge, UK) were used. Details of live cell imaging are given in Ref.4. eGFP and DRAQ5 or Syto17 were excited simultaneously by 488-nm light; filters used were primary dichroic 510DCLP, secondary 565DRLP, emission 540/30 and 630OG.

Image Analysis

Image analysis was performed using LaserSharp (Carl Zeiss MicroImaging GmbH, Göttingen, Germany) and ImagePro+ (Media Cybernetics, Bethesda, MD, USA).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED
  8. Supporting Information

To investigate the influence of DNA-binding drugs on nuclear and chromatin structure, we exposed live HeLa cells expressing GFP-tagged histones to different concentrations of DRAQ5 or Syto17. Figures 1–5 demonstrate five processes that occur when live cells are exposed to increasing concentrations of DRAQ5: (i) binding of DRAQ5 to nuclear DNA, (ii) subsequent dissociation of H1 and H2B histones from DNA, (iii) aggregation of chromatin, (iv) accumulation of H1 histones in nucleoli (at 3 μM DRAQ5 and higher), and (v) quenching of eGFP-H1 and eGFP-H2B fluorescence by chromatin-bound DRAQ5.

thumbnail image

Figure 1. Changes of nuclear structure and spatial distribution of chromatin in live HeLa cells following intercalation of DRAQ5 into DNA. Green – fluorescence of histone H1 or H2B; red – DNA-bound DRAQ5; n – nucleus; nl – nucleolus. (A, K, D, N) Nuclear architecture prior to incubation with DRAQ5; (B, L) after 15-min incubation with DRAQ5 at 1 μM; (C, M) after additional 15-min incubation with DRAQ5 at 3 μM; (E, O) after 15-min incubation with DRAQ5 at 7.5 μM. Note: cells shown in D, E are different than in panels A–C; the same stands for K–M vs. N, O. Bars: 1 μm. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

thumbnail image

Figure 2. Aggregation of chromatin into heterochromatic regions in live cells. (A) Line profiles showing histone H2B fluorescence intensity prior (dotted) and after DRAQ5 intercalation (solid), measured within a region marked in images B (before) and C (after adding DRAQ5). Local fluorescence intensity increases in DNA-rich region, and diminishes in a region with a lower DNA content, suggesting large scale chromatin contraction toward heterochromatic foci.

Download figure to PowerPoint

thumbnail image

Figure 3. Accumulation of GFP-tagged histones in nucleoli following intercalation of DRAQ5. A ratio of histone fluorescence in nucleolus vs. areas occupied by DNA (fl-NL/fl-DNA) is plotted against DRAQ5 concentration. Intensity of fluorescence of H1-GFP detected in nucleoli of live cells increases with concentration of DRAQ5; accumulation of H2B-GFP is very weak.

Download figure to PowerPoint

thumbnail image

Figure 4. Dynamics of histones accumulated in nucleoli in the presence of DRAQ5 (7.5 μM). (A, B, C) FRAP of H1-GFP histones accumulated in nucleoli; images were collected before, immediately after, and 25 min after bleaching one nucleolus in the center (arrow). (D, E, F) FRAP of H2B histones; images were collected before, immediately after, and 25 min after bleaching a rectangular area embracing one nucleolus. Histones H1 and H2B, which are accumulated in nucleoli, are exchanging with their mobile pools, as demonstrated by FRAP. Bars: 1 μm. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

thumbnail image

Figure 5. Quenching of GFP histone tags by intercalated DRAQ5, in live and fixed cells. (A) Histograms of intensities of DNA-bound H2B-GFP in areas of the nucleus rich in chromatin (fixed cells), demonstrating a decrease of GFP fluorescence intensity as a function of a concentration of DRAQ5 added to culture medium. Inset: DRAQ5-dependent decrease of intensity of fluorescence of DNA-bound H2B-GFP in the brightest pixels in chromatin. (B, C) Integrated fluorescence of H1-GFP and H2B-GFP histones in chromatin-rich areas of the nucleus, in fixed (B) and live (C) cells, plotted against a concentration of DRAQ5. Fixation abolishes histone dynamic exchange and chromatin spatial rearrangements, and thus in fixed cells fluorescence loss is entirely due to quenching, but in chromatin of live cells signal intensity depends on histone detachment, quenching, and contribution from the unquenched mobile histone pool.

Download figure to PowerPoint

Binding of DRAQ5 and Aggregation of Chromatin

DRAQ5 added to culture medium readily crossed plasma and nuclear membranes and stained nuclear DNA in a concentration-dependent manner. The increasing intensity of red fluorescence in nuclei reflected the growing amount of the DNA-bound dye (Figs. 1F–1J and 1P–1U). However, even in the presence of 1 μM DRAQ5 (when DNA staining is still poorly detectable), chromatin began to aggregate, as demonstrated by comparing the spatial distribution of DNA-bound histones prior and following intercalation of the dye (Figs. 1A vs. 1B and 1K vs. 1L, see also Supplementary Movies 1–3). Fluorescence of DNA-bound H2B-GFP increased in heterochromatic regions, and decreased in regions of low DNA concentration (Fig. 2), indicating that chromatin had a tendency to contract toward heterochromatic foci. This process was most evident in perinucleolar heterochromatin. Further increase of DRAQ5 concentration to 3 μM caused a more pronounced aggregation and a gradual destruction of chromatin organization (Figs. 1C and 1M). At 7.5 μM, nucleus adopted an appearance, which suggested a random homogenous distribution of chromatin, where hetero- or euchromatin could no longer be discerned (Figs. 1 E and 1O). As expected, drug-induced aggregation did not occur in formaldehyde-fixed cells (data not shown), where protein crosslinking prevented such structural changes.

Detachment of Linker Histones (H1) from DNA

Analysis of images in Figure 1 demonstrates that intercalation of DRAQ5 into DNA leads to detachment of H1 histones. Initially, fluorescence of H1 in DNA-rich regions decreased due to histone loss as well as quenching of GFP fluorescence by the intercalated DRAQ5 (see below). Subsequently histone H1 began to accumulate in the nucleoli (Figs. 1C, 1E, and 3). The process of H1 accumulation was quite rapid—it occurred within minutes (at 3 μM DRAQ5) or even seconds (at 7.5 μM DRAQ5) of adding the drug. Thus, it is highly probable that accumulated H1 was derived from a pool displaced from DNA. The nucleoli-accumulated H1 molecules were not attached to their target permanently, but were still exchanging with the mobile and the DNA-bound pools, as demonstrated by FRAP experiments (Figs. 4A–4C).

Detachment of Core Histones (H2B) from DNA

Association of the H2B core histones with DNA was influenced by DRAQ5 to a lesser degree. The H2B histones are firmly bound within a nucleosomal structure and undergo only very limited dynamic exchange [t1/2 = 2 h (5)]. At 1 μM DRAQ5, distribution of H2B (Fig. 1L) matched that of DNA (Fig. 1R) and reflected chromatin aggregation. At 3 μM DRAQ5, the distribution of H2B became diffused (Fig. 1M), which is consistent with the postulated loss of chromatin higher order structures and suggests formation of H2B mobile pool. Indeed, formation of a slowly exchanging pool (t1/2 = ∼5 min; data not shown) of H2B at 7.5 μM DRAQ5 was confirmed by FRAP (Figs. 4D–4F, see also Supplementary Movie 4). Under these conditions the distribution of H2B (which was now present in a form of a bound as well as a mobile pool) in nuclei became almost uniform. These observations are consistent with a notion that when DNA is modified by the intercalator, a subpopulation of H2B binds less tightly. The weaker binding promotes formation of a soluble pool. We hypothesize that at DRAQ5 concentrations exceeding 3 μM some nucleosomes are dismantled and H2B is liberated, and thus in many areas not only higher but also lower orders of chromatin structure are lost.

In contrast to H1, accumulation of H2B in nucleoli was quite weak (Fig. 3, also compare Fig. 1E with 1O). This difference may be explained by lower efficiency of H2B removal from DNA and a smaller fraction of mobile H2B. It might alsoindicate that H2B has a lower affinity to its target in nucleolus.

Quenching

Intercalated DRAQ5 interacts with GFP fused with histones and causes quenching of its fluorescence, as demonstrated by histograms and graphs in Figure 5. In formaldehyde-fixed cells, quenching of H2B-GFP by DRAQ5 appears to be less pronounced than H1-GFP (for instance, 69% quenching of eGFP-H1 but only 35% of eGFP-H2B at 3 μM DRAQ5; Figs. 5B and 5C). Higher quenching of GFP tag attached to H1 than H2B may result from a shorter distance separating H1 from an intercalated DRAQ5.

At first glance the decrease of histone-GFP fluorescence caused by DRAQ5 appears almost the same in live and fixed cells (Figs. 5B and 5C). This might lead to a conclusion that quenching efficiency of each of the histones by DRAQ5 is the same in fixed and live cells. However, this assessment may not be fully accurate. Live cells contain various proportions of detached histones, which are not quenched by the DNA-bound DRAQ5. Thus, histone detachment and quenching have opposing influences on the net measured H1-GFP or H2B-GFP fluorescence in live cells. Moreover, intermolecular distances separating DRAQ5 and GFP and frequency of intermolecular collisions are likely to be different in fixed and live cells. Thus, a direct comparison between graphs B and C does not appear to be possible, nor is it possible to accurately dissect contributions of histone detachment and quenching to the overall loss of green fluorescence in live cells exposed to DRAQ5.

Interaction Between a Minor Groove Binder, Syto17, and Chromatin

To establish if the drugs that bind DNA by different modes can differ in their influence on histone–DNA interaction, we also investigated the influence of Syto17, a membrane permeant minor groove binder, on chromatin structure. Incubation of live cells with Syto17 at a low concentration of 0.5 μM (i.e. when the DNA-bound drug becomes detectable) demonstrated only slight chromatin aggregation. The loss of histone fluorescence was only ∼10% (H1) and 3% (H2B) (Figs. 6A vs. 6B; 6K vs. 6L). No histone accumulation in nucleoli was detected, which hints at no detachment of histones from chromatin. In higher (1–7.5 μM) concentrations (Figs. 6C–6E and 6M–6O) we observed only minor aggregation of chromatin and fluorescence quenching (this was confirmed on fixed cells; data not shown). Still no evidence of histone detachment from DNA was observed. In general, Syto17 caused only negligible changes of chromatin and nuclear structure, which is consistent with a notion that histone–DNA interactions were not significantly affected.

thumbnail image

Figure 6. Exposure to Syto17 causes barely detectable chromatin aggregation and no histone accumulation in nucleoli. Images of histones H1 and H2B (green, left columns) in live HeLa cells incubated with Syto17 (red, right columns). Bars: 1 μm. (A, K) Chromatin distribution prior to incubation with Syto17. (B, L; C, M; D, N; E, O) Histone-GFP after 15-min incubation periods with Syto17 at 0.5, 1, 3, or 7.5 μM, respectively. GFP fluorescence decrease is due to quenching by Syto17. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED
  8. Supporting Information

Changes of Chromatin Structure Caused by DRAQ5

The data presented above demonstrate that DRAQ5, an intercalator, which binds to nuclear DNA in a live cell, interferes with chromatin and nuclear structure. Aggregation of chromatin becomes detectable at the drug concentration of 1 μM, when chromatin-bound DRAQ5 becomes detectable by a typical confocal instrument. This implies that even at this low dye concentration, images of fluorescently labeled chromatin in nuclei may represent a spatial distribution, which is already modified by the fluorescent probe, and differs from the native state. It seems possible that less prominent changes of chromatin structure occur even at lower drug concentrations, i.e. below 1 μM, but are not readily detectable by optical microscopy, due to limited spatial resolution and sensitivity of this technique. Higher concentrations of the intercalator (3–7.5 μM) cause significant aggregation and almost complete loss of chromatin and nuclear spatial organization. We observed similar effects caused by several other DNA intercalators, including adriamycin and daunomycin (manuscript in preparation). In contrast, no such influence was exerted by a minor groove binder Syto17. This suggests that the ability to strongly interfere with histone–DNA binding and disrupt nuclear and chromatin structure may be a common feature of DNA intercalators, but not minor groove binders.

Mechanism of Chromatin Structural Changes

Nucleoplasm is characterized by high molecular crowding; however, in a HeLa cell a complex of DNA and histones occupies only 3–6% of the volume of the nucleus. The intricate, highly complicated spatial distribution of this complex within a nucleus is maintained and controlled by the machinery, whose major component is histone H1 (5). The synthesis of H1 (as well as H2B) is tightly controlled and the levels of histones, which are not bound to DNA, are very low. H1 plays its chromatin-stabilizing role being, at the same time, highly dynamic. Molecules of H1 detach from DNA and bind to a different stretch every 3–4 min (6). The data presented in this report demonstrate that DRAQ5 interferes with dynamic exchange of H1 histones in a concentration-dependent manner. Based on the existing knowledge and our data, the following hypothesis explaining the phenomenon of intercalator-induced chromatin aggregation can be put forward: DRAQ5 (at 1 μM) enters live cells and nuclei and intercalates into DNA, pushes bases apart, and unwinds DNA (7). The affinity of histone H1 to such altered DNA is weaker. Since H1 is very dynamic (8), drug intercalation into DNA shifts the balance toward the unbound pool of H1. The effective detachment of histone H1 from DNA brings about loss of higher order chromatin structures. The postulated loss of chromatin structure is depicted schematically in Figure 7. At higher concentrations (3–7.5 μM) of the intercalator, histone H2B also begins to detach from DNA. This leads to dismantling of nucleosomes and a loss of lower order chromatin structures. The proposed sequence of events is consistent with the phenomena we observed and the reports of others, indicating that linker histone H1 is a principal factor stabilizing chromatin structure in vivo (9). As intercalation leads to detachment of histone H1 and exposing new stretches of DNA, new binding sites open for intercalation. Thus, the process becomes self-promoting. Apparently, the detached histone H1 is harbored in nucleoli. This accumulation may result from binding of the H1 histones, which were detached from DNA, to the next best negatively charged polymer, which is not affected by DRAQ5—the RNA.

thumbnail image

Figure 7. A schematic representation of the postulated changes of chromatin structure resulting from interaction with an intercalator DRAQ5 (0–3 μM). (A) Nucleosomes and DNA in untreated live cells; (B) DRAQ5 enters cell nucleus; (C) DRAQ5 intercalates into DNA and causes partial unwinding of the DNA double helix; (D) histone H1 is detached from the DNA, higher order chromatin structures are lost, and DNA–histone complex aggregates. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

It seems interesting to ask if the loss of chromatin structure is a consequence of a random detachment of H1, or a dissociation of a specific subpopulation of this histone, located in some specific areas of chromatin. However, we were unable to solve this question, as this requires live cell imaging methods capable of very high sensitivity and spatial resolution (10ths of nanometers).

Importance of Drug-Induced Chromatin Changes in Cytometry

Previous work has shown many advantages of DRAQ5 (10) and other cell-permeant dyes, like DAPI or Hoechst in fluorescent staining of DNA. However, no comparisons were made between chromatin structures existing prior and following staining, and thus structural changes evoked by binding of a dye to DNA were not known. Our study examined the whole range of DRAQ5 concentrations and indicates that, although chromatin structural changes do occur, they can be minimized by optimizing the dye concentration. Obviously no disturbance of chromatin structure, which is detectable by optical microscopy, is expected in cells where proteins were crosslinked by formaldehyde prior to staining with DRAQ5. Thus the applicability of DRAQ5 to DNA detection in fixed cells is not affected by the phenomena described in this report.

The observations described here bear potential relevance for biomedical research as well as live cell imaging. In a wide biological context, they confirm the importance of histone H1 for maintaining nuclear and higher order chromatin structures in vivo. In the context of image cytometry, they show that intercalating dyes, at detectable concentrations, may influence chromatin and nuclear structure and, hence, cause potential disturbance to cell functions, even before imaging of a live cell commences; they also stress the importance of optimizing the dye concentration. The dye–DNA interaction described here is relevant for use of DNA intercalators in flow cytometry, emphasizing that such dyes may be opening new binding sites in DNA of live cells, and possibly even in fixed cells as well. Thus, the binding of a probe may influence quantification of DNA.

Biological Significance of DRAQ5 Interaction with DNA

Dynamic histone exchange regulates higher order chromatin structures (11) and plays an important role in transcription (12), DNA repair (13), and mitosis (14). Previous studies of cytotoxic action of anthracyclines concentrated on the influence on topoisomerase (15, 16), anthracycline free radical generation, and other potential mechanisms (reviewed in17). Intercalator-induced loss of H1 from chromatin and a loss of higher order chromatin structures may be involved in the mechanism of cytotoxic action of intercalating agents like daunorubicin and doxorubicin. The intercalator-driven changes of histone dynamic exchange and chromatin structure may prove relevant to understanding of biological consequences of interaction between antitumor antibiotics and drugs with DNA (Wójcik and Dobrucki, manuscript in preparation).

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED
  8. Supporting Information

The authors are grateful to Dr. Kanda, H. Kimura, and P. R. Cook for making cells expressing eGFP-tagged histones available.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. LITERATURE CITED
  8. Supporting Information

This article includes supplementary video clips, available online at http://www.interscience.wiley.com/jpages/1552-4922/suppmat .

FilenameFormatSizeDescription
cytoa20573-1-H1-DQ5-CONC.gif441KSupporting Information file cytoa20573-1-H1-DQ5-CONC.gif
cytoa20573-2-H2B-DQ5-CONC.gif400KSupporting Information file cytoa20573-2-H2B-DQ5-CONC.gif
cytoa20573-3-H2B-DQ5-TIME.gif608KSupporting Information file cytoa20573-3-H2B-DQ5-TIME.gif
cytoa20573-4-H2B-DQ5-FRAP.gif3387KSupporting Information file cytoa20573-4-H2B-DQ5-FRAP.gif

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.