Analysis of histone variants and epigenetic marks is dominated by genome-wide approaches in the form of chromatin immunoprecipitation-sequencing (ChIP-seq) and related methods. Although uncontested in their value for single-copy genes, mapping the chromatin of DNA repeats is problematic for biochemical techniques that involve averaging of cell populations or analysis of clusters of tandem repeats in a single-cell analysis. Extending chromatin and DNA fibers allows us to study the epigenetics of individual repeats in their specific chromosomal context, and thus constitutes an important tool for gaining a complete understanding of the epigenetic organization of genomes. We report that using an optimized fiber extension protocol is essential in order to obtain more reproducible data and to minimize the clustering of fibers. We also demonstrate that the use of super-resolution microscopy is important for reliable evaluation of the distribution of histone modifications on individual fibers. Furthermore, we introduce a custom script for the analysis of methylation levels on DNA fibers and apply it to map the methylation of telomeres, ribosomal genes and centromeres.

DNA and chromatin fiber extension techniques are single-molecule methods based on the stretching of DNA or chromatin on microscopic slides. Stretching of the template is followed by immunofluorescence (IF) and fluorescence in-situ hybridization (FISH) detection to investigate replication dynamics, physical genome mapping or epigenetic profiles of repeats using microscopy. Fiber techniques are great tools to study the chromatin of repetitive regions that are hard to map using genome-wide approaches. Preparing DNA fibers to study them under the microscope has been pioneered by Bensimon et al. (1994), who demonstrated that isolated DNA can be extended on a silanized glass surface. High molecular weight DNA can be prepared in agarose plugs, and then purified and extended on a poly-L-lysine-coated or silanized coverslip, a process termed molecular combing (Kaykov et al., 2016; Schurra and Bensimon, 2008). DNA fibers can be also extended directly by immersing cells or isolated nuclei on microscope slides into a lysis buffer (Fransz et al., 1996; Jackson and Pombo, 1998). From a technical standpoint, the coupling of replication labeling or FISH with DNA fiber preparation is relatively straightforward and has been reviewed in depth previously (Nieminuszczy et al., 2016; Quinet et al., 2017). In terms of applications, DNA fibers are mainly used to investigate the dynamics of replication in the genome and to physically map repeats and examine their methylation. It has been shown that a subset of ribosomal RNA genes (rDNA) in human cells displays palindromic organization (Caburet, 2005) and that non-canonical rDNA repeat arrangement is linked to DNA hypermethylation (Zillner et al., 2015).

Preparing chromatin fibers (Garcia-Blanco et al., 1995) relies on using a milder lysis of nuclei, during which some proteins are retained on the chromatin template. This method has since been leveraged to study centromeric regions in yeast, Drosophila (Blower et al., 2002), plants (Jin et al., 2004; Zhang et al., 2008) and mammals (Cohen et al., 2009; Sullivan and Karpen, 2004). Fiber techniques can be relevant when analysing more complex DNA structures through the combination of fiber extension and super-resolution microscopy. It has been shown that telomeric T-loops can be visualized using a modified protocol for fiber extension, and these loops can be detected by super-resolution microscopy (Benarroch-Popivker et al., 2016; Doksani et al., 2013). The advantage of this approach is the possibility to quantify the frequency of T-loops and analyze their length in a given mutant background. Several protocols and methodical resources are now available, describing chromatin fiber preparation from yeast and mammalian cells in detail (Sabatinos and Green, 2018; Sullivan, 2010; Wooten et al., 2020). These can be a good starting point for working with fibers, but extensive optimization is usually required, especially if the IF labeling of proteins is coupled with FISH detection afterwards. Recently, two modifications of the chromatin fiber technique have been introduced, termed APEX-chromatin fibers (Kyriacou and Heun, 2018) and SRCF (Wooten et al., 2019). APEX-chromatin fibers is an extension of the chromatin fiber technique (APEX stands for ascorbate peroxidase 2) and provides the possibility of testing protein interactions between specific centromeric proteins. SRCF simply refers to the use of super-resolution microscopy for chromatin fiber imaging, which is essential for proper data interpretation, as we argue in this paper.

As mentioned above, one of the key applications of fiber technology is the study of chromatin in repetitive regions. These constitute a major portion of most eukaryotic genomes. The biological roles of repetitive regions are evident only in some cases (e.g. ribosomal genes, telomeres and centromeres) and remain enigmatic for most of the other repeats, with some repeats even posing a threat to genome integrity (for example, transposons). Ribosomal genes are essential housekeeping genes, coding structural RNA elements of the ribosome. Usually, hundreds of copies are present in the genome; some of these are actively transcribed, while the rest are silenced (McStay and Grummt, 2008). Telomeres are nucleoprotein structures that protect chromosome termini and are composed of tandem microsatellite repeats in most taxonomic groups of organisms (e.g. TTAGGG in mammals). Centromeres are composed of longer tandem satellite sequences (e.g. the 180 bp-long α-satellite in primates) that create a distinct genomic environment that favors CENH3 (also known as CID in Drosophila) incorporation and kinetochore–microtubule interaction during cell division (Blower and Karpen, 2001). Deeper characterization of these repeats is outside of the scope of this article (for reviews see Achrem et al., 2020; Hartley and O'Neill, 2019; Procházková Schrumpfová et al., 2019), but we highlight some of their common features. All of these repeats have been shown to be transcribed to some extent (Luke and Lingner, 2009; Mayer et al., 2008; Wong et al., 2007), suggesting that viewing these repeats only as silent heterochromatin is simplistic. Moreover, epigenetic regulation of these repeats is important for genome stability (Muchová et al., 2015). It is obvious that the epigenetic regulation of these repeats is more complex than assumed previously, and study of this regulation is valuable, but not straightforward.

Owing to the difficulty of assembling larger repeats from sequencing data, repetitive regions are often excluded from genome-wide analyses. Although it is reasonable to assume that non-functional repeats might be quite homogeneous in their epigenetic signature, functional repeats will most likely contain a mixture of alternative epigenetic states (e.g. active and inactive clusters of ribosomal genes). Failing to capture the heterogeneity of clusters of repeats can lead to somewhat controversial data. For instance, plant telomeres are considered to be bivalent chromatin regions, meaning that they exhibit both euchromatin and heterochromatin-related features (Dvořáčková et al., 2015; Fojtová and Fajkus, 2014). Chromatin immunoprecipitation-sequencing (ChIP-seq) data indicates that telomeric regions are enriched in transcription-associated histone variant H3.3, as well as some euchromatin marks such as histone H3 dimethylation at lysine 4 (H3K4me2) and acetylation at lysine 9 (H3K9Ac), when compared with centromeric regions (Vaquero-Sedas and Vega-Palas, 2013; Vaquero-Sedas et al., 2012). On the other hand, results from a dot-immunoblot analysis of telomeres in Nicotiana tabacum have revealed the presence of both histone H3 lysine 9 dimethylation (H3K9me2; constitutive heterochromatin) and lysine 27 trimethylation (H3K27me3; repressed ‘euchromatin’), as well as histone H3 lysine 4 trimethlyation (H3K4me3), an activation mark (Majerová et al., 2014). Recent results (Adamusová et al., 2020) have shown the existence of two combinatorial patterns of histone modifications in plants, with telomeric histones decorated predominantly by H3K9me2 or H3K27me3 marks. Interestingly, neither of the two patterns are related to the sequence of the telomere motif, the lengths of the telomeres, or the phylogenetic position of the plant species. Besides direct detection, functional studies in plants have shown that the histone deacetylases HDT4 and HDA6, as well as the histone methyltransferase Polycomb Repressive Complex 2 (PRC2), play a role in the epigenetic status of telomeres, underlining the partially repressive character (Lee and Cho, 2016; Zhou et al., 2018). Hence, fiber techniques could prove essential to decipher the actual epigenetic profile of these repeats.

In this sense, the main incentive to use extended fibers (a term used here to refer to both DNA fiber and chromatin fiber preparations) in epigenetics is to study DNA methylation, histone modifications and histone variant distribution in repetitive regions. Although ChIP-seq is essential for our understanding of the epigenetic make-up of individual genes, it is not clear how to interpret the data gathered for repeats. At best, ChIP-seq data can show what an average repeat in a population of cells looks like, disregarding the possibility that there can be alternative epigenetic states associated with a subset of these repeats. Recently, single-cell biochemical approaches that aim to capture cell-to-cell heterogeneity in complex samples have been introduced. DNA methylation profiles (Clark et al., 2017), chromatin accessibility (Cusanovich et al., 2018) and histone mark profiles (Grosselin et al., 2019; Rotem et al., 2015) have been successfully probed in individual cells using DNA barcoding and microfluidics technologies, upgrading commonly used genome-wide analyses. Distinct cellular subpopulations have been identified in this manner, which is relevant, for example, in evaluating epigenetic heterogeneity in tumors and therapy-resistant subclones. Single-cell biochemical characterization does not discriminate between possible alternative epigenetic states within a cluster of repeats, but this information can be obtained using fiber techniques.

Currently, the main limitations impeding the wider adoption of fiber techniques are experimental reproducibility, the lack of appropriate tools for data analysis and data interpretation. One of the major issues for chromatin fiber preparations is the variable degree of stretching and protein retention, which causes low experimental reproducibility. Routinely, it is necessary to prepare several samples for each experimental condition to obtain at least one or two samples with sufficiently stretched fibers that still retain histones. Moreover, there is a considerable variation in stretching even on a single coverslip. Although DNA fiber imaging does not require super-resolution microscopy, chromatin fibers (which are stretched to a lesser extent than DNA fibers) should ideally be imaged using either stimulated emission depletion microscopy (STED; Wooten et al., 2019) or stochastic optical reconstruction microscopy (STORM; (Ribeiro et al., 2010). Imaging of DNA methylation in repetitive regions is usually performed on extended DNA fibers, producing discrete, dotted signals for both the FISH probe and the anti-5-methylcytosine antibody. This can be problematic for commonly used colocalization-based approaches used to determine the extent of DNA methylation on a particular repeat.

In this work, we address both experimental reproducibility and data analysis. First, we show that both chromatin and DNA fibers from various models (cultured mammalian cell lines and isolated plant nuclei) can be prepared using an alternative, more reproducible approach based on lysis buffer aspiration with a pipette tip attached to a vacuum pump. Second, we demonstrate that using super-resolution microscopy when imaging chromatin fibers helps to reliably evaluate the distribution of histone modifications and is useful to parse single fibers from fiber bundles, which might otherwise obscure later analysis. Last, but not least, we illustrate the versatility of using DNA fibers in tracking methylation profiles of telomeres, ribosomal genes and centromeres, and develop an image analysis tool that quantifies the methylation levels along selected repeats. Overall, this work displays the value of using super-resolution microscopy in the evaluation of chromatin fibers, offers guidelines that help to avoid errors in fiber evaluation and provides a new tool for the analysis of single-molecule DNA methylation on DNA fibers.

A modified protocol to prepare extended chromatin and DNA fibers

We developed a modified method for the preparation of extended fibers, given that previously described techniques (coverslip dragging, tilting the slide or drying) show low reproducibility. In our setup, we cytocentrifuge or spot isolated nuclei onto microscope slides or coverslips and immerse these into a Coplin jar or coverslip holder containing the lysis buffer (SDS-based buffer for DNA fibers; Triton X-100-based buffer for chromatin fibers). Instead of removing the slides or coverslips from the Coplin jar one by one, we simply aspirate the lysis buffer with a pipette tip attached to a vacuum pump, similar to removing cell culture medium when working with adherent cell lines. We immerse the entire tip into the Coplin jar to ensure the steady removal of the lysis buffer at a constant rate. In our experience, using a 1–10 µl pipette tip and relatively slow aspiration of the lysis buffer works best. The key advantages of this approach are that both DNA and chromatin fibers can be prepared in this way and several slides (4–8) can be prepared at the same time without introducing errors caused by manually dragging the coverslips or tilting the slides too much or too little. Although the incubation time in the lysis buffer and the buffer composition (in the case of chromatin fibers) still needs to be optimized for different models, this protocol greatly reduces the variability among samples prepared in the same run.

While inter-sample variability in the same experiment is minimal, we reasoned that inter-experimental variability in fiber preparation might be more substantial. In order to demonstrate the reproducibility of this method, we performed extension of both chromatin and DNA fibers using the aspiration technique and compared the stretching of fibers between independent experiments. For chromatin fibers, we evaluated the stretching of clusters of compact heterochromatin in mouse embryonic fibroblasts (MEFs), termed chromocenters (Fig. 1A–C) while for DNA fibers we tested the length of telomeres in fiber preparations of N. tabacum (Fig. 1D–F). For chromatin fibers, we observed the stretching of chromocenters in two replicate experiments to be on average 13.3 µm (n=64, s.d.=6.9 µm) and 20.8 µm (n=66, s.d.=8.4 µm), showing a moderate degree of variability. For DNA fibers, the results of the two replicate experiments were almost identical, with the telomeric tracts stretching for 30.4 µm (n=50, s.d.=20.6 µm) and 30.9 µm (n=52, s.d.=19.6 µm) (Fig. 1D–F). Importantly, we noticed a considerable difference in the extension of heterochromatin compared to the rest of the genome. Fibers extruding from the partially lysed nuclei of MEFs that were labeled with histone H3 often extended over several fields of view (Fig. S1), whereas regions that were labeled with histone H3 lysine 9 trimethylation (H3K9me3) remained in close proximity to the lysed nucleus (Fig. 1B). For histone H3 and H3K9me3 labeling, we used slides from the same technical replicate, indicating that differential stretching is not related to inter-experimental variability but is due to an actual difference in stretching between heterochromatic and euchromatic regions. We compared the aspiration technique with a previously used drying method to prepare chromatin fibers (Cohen et al., 2009). Although the drying technique can produce acceptable fibers, we noted that the stretching was not homogenous across the slide (Fig. S2).

Fig. 1.

Extension of chromatin and DNA fibers using buffer aspiration. (A–C) Chromatin fibers produced in two independent experiments by buffer aspiration from a Coplin jar, following partial lysis of MEFs. Fibers are stained for DNA using DAPI (A) and for H3K9me3 (B), with merged images shown beneath (C). Heterochromatin, decorated with the H3K9me3 histone mark, shows a variable degree of extension in chromatin fibers, averaging 17.1 µm (C1,C2). (D–F) DNA fibers produced in two independent experiments by the lysis of isolated nuclei of N. tabacum using aspiration. Fibers are stained for methylation (5-methylcytosine, 5mC; D), and temores are labelled using FISH (E), with merged images shown beneath (F). The lengths of chromatin fibers and telomeres (indicated by white lines) were measured for the purposes of reproducibility assessment using the segmented line tool in ImageJ. Scale bars: 10 µm.

Fig. 1.

Extension of chromatin and DNA fibers using buffer aspiration. (A–C) Chromatin fibers produced in two independent experiments by buffer aspiration from a Coplin jar, following partial lysis of MEFs. Fibers are stained for DNA using DAPI (A) and for H3K9me3 (B), with merged images shown beneath (C). Heterochromatin, decorated with the H3K9me3 histone mark, shows a variable degree of extension in chromatin fibers, averaging 17.1 µm (C1,C2). (D–F) DNA fibers produced in two independent experiments by the lysis of isolated nuclei of N. tabacum using aspiration. Fibers are stained for methylation (5-methylcytosine, 5mC; D), and temores are labelled using FISH (E), with merged images shown beneath (F). The lengths of chromatin fibers and telomeres (indicated by white lines) were measured for the purposes of reproducibility assessment using the segmented line tool in ImageJ. Scale bars: 10 µm.

Super-resolution microscopy of chromatin fibers

We tested and evaluated three main super-resolution microscopy approaches by studying the mutual positioning of H3K4me3 (active) and H3K27me3 (repressive) histone marks in chromatin fibers, depicted in Fig. 2A. These included structured-illumination microscopy (SIM), STED and direct STORM (dSTORM) (Fig. 2B). A succinct report on the advantages and disadvantages of the different methods when imaging chromatin fibers is presented in Table 1, and examples from each of the techniques are shown in Fig. 2, alongside widefield or confocal images of the same fiber.

Fig. 2.

Analysis of histone modifications on chromatin fibers using super-resolution microscopy in HEK293T cells. (A) Schematic representation of chromatin fiber histone modifications on the N-terminal histone tails. (B) Distribution of H3K4me3 (red) and H3K27me3 (green) histone marks along HEK293T chromatin fibers probed using conventional microscopy (panels 1,3 and 5) and SIM (panel 2), STED (panel 4) and dSTORM (panel 6) microscopy. Boxes indicate regions shown in C. (C) Detailed views of the numbered regions highlighted in B. (D,E) Two-color dSTORM microscopy of H3K4me3 (red) and H3K27me3 (green) histone marks along HEK293T chromatin fibers. Boxed regions (i–v) in D are shown in detail in E. Note the co-occurrence of H3K4me3 and H3K27me3 in some parts of the fiber (i,iii), whereas other parts show predominant H3K4me3 (ii) or H3K27me3 signals (iv). Images are representative of two independent experiments per imaging condition. Scale bars: 2 µm (B), 5 µm (D).

Fig. 2.

Analysis of histone modifications on chromatin fibers using super-resolution microscopy in HEK293T cells. (A) Schematic representation of chromatin fiber histone modifications on the N-terminal histone tails. (B) Distribution of H3K4me3 (red) and H3K27me3 (green) histone marks along HEK293T chromatin fibers probed using conventional microscopy (panels 1,3 and 5) and SIM (panel 2), STED (panel 4) and dSTORM (panel 6) microscopy. Boxes indicate regions shown in C. (C) Detailed views of the numbered regions highlighted in B. (D,E) Two-color dSTORM microscopy of H3K4me3 (red) and H3K27me3 (green) histone marks along HEK293T chromatin fibers. Boxed regions (i–v) in D are shown in detail in E. Note the co-occurrence of H3K4me3 and H3K27me3 in some parts of the fiber (i,iii), whereas other parts show predominant H3K4me3 (ii) or H3K27me3 signals (iv). Images are representative of two independent experiments per imaging condition. Scale bars: 2 µm (B), 5 µm (D).

Table 1.

Comparison of SIM, STED and STORM in chromatin fiber imaging

Comparison of SIM, STED and STORM in chromatin fiber imaging
Comparison of SIM, STED and STORM in chromatin fiber imaging

We report that using super-resolution microscopy approaches to study chromatin fibers can yield new and essential information. As expected, many of the foci that appeared to indicate colocalization of the two examined histone marks in images taken using widefield or confocal microscopy actually represented multiple discrete positions along the fiber (Fig. 2C). This degree of resolution is crucial for the proper analysis of any region of interest. During the stretching procedure, fibers tend to align with one another, creating fiber bundles that are difficult to analyze. In such cases, we found that STED and STORM microscopy could discriminate between single fibers and smaller bundled fibers (Fig. 3). As a general rule, individual fibers showed very weak DAPI staining and were less densely covered by histones (Fig. 3A–C, insets 1,2,5 and 6). Fibers that in some cases could be evaluated as single fibers using conventional microscopy could be discriminated as bundles using a super-resolution modality (compare Fig. 3A,C,E with B,D,F). For the evaluation of fibers we recommend considering the following criteria: (1) parallel orientation of histone marks along the fibers (Figs 3, 4); (2) relative intensity of DAPI fluorescence; and (3) branching of chromatin fibers (both outlined in Fig. 3H–K). These criteria are detailed in greater depth in the Discussion.

Fig. 3.

Discrimination of single fibers and fiber bundles in chromatin fiber preparations using STED and dSTORM microscopy. (A,B) Comparison of confocal (A) and STED (B) analysis of chromatin fibers extended from HEK293 cells and stained for H3K4me3, H3K27me3 and DNA (DAPI, blue). Single fibers (boxed regions 1 and 2) show lower levels of histone modifications as opposed to fiber bundles (boxed regions 3 and 4) and tend to align, showing very weak DAPI staining. (C–F) Widefield (C,E) and dSTORM (D,F) imaging of H3K4me3, showing a single fiber (C,D; boxed regions 5 and 6) and a bundle of several fibers (E,F; boxed regions 7 and 8). Regions highlighted by boxes in A–F are aligned and shown in detail on the right. (G) Illustration of a single chromatin fiber (left) and bundled fibers (right). (H–K) Examples of fibers not matching the criteria for single fiber analysis (widefield microscopy). Vastly different fluorescence intensities of a fiber bundle (indicated by arrowheads) and a presumed single fiber (indicated by a dotted line) are shown. Branching of fibers is indicated by arrows. Images are representative of three experiments. Scale bars: 2 µm (A–F), 1 µm (A–F, magnifications), 5 µm (H–K).

Fig. 3.

Discrimination of single fibers and fiber bundles in chromatin fiber preparations using STED and dSTORM microscopy. (A,B) Comparison of confocal (A) and STED (B) analysis of chromatin fibers extended from HEK293 cells and stained for H3K4me3, H3K27me3 and DNA (DAPI, blue). Single fibers (boxed regions 1 and 2) show lower levels of histone modifications as opposed to fiber bundles (boxed regions 3 and 4) and tend to align, showing very weak DAPI staining. (C–F) Widefield (C,E) and dSTORM (D,F) imaging of H3K4me3, showing a single fiber (C,D; boxed regions 5 and 6) and a bundle of several fibers (E,F; boxed regions 7 and 8). Regions highlighted by boxes in A–F are aligned and shown in detail on the right. (G) Illustration of a single chromatin fiber (left) and bundled fibers (right). (H–K) Examples of fibers not matching the criteria for single fiber analysis (widefield microscopy). Vastly different fluorescence intensities of a fiber bundle (indicated by arrowheads) and a presumed single fiber (indicated by a dotted line) are shown. Branching of fibers is indicated by arrows. Images are representative of three experiments. Scale bars: 2 µm (A–F), 1 µm (A–F, magnifications), 5 µm (H–K).

Fig. 4.

Methylation analysis pipeline for DNA fibers. Depiction of key steps in the workflow for telomere measurement and methylation analysis. Individual channels are exported from two-color images after thresholding (panels 1 and 2). The channel with the telomere FISH probe signal is then segmented (panel 3) in order to join the interrupted signals of the probe along the fiber and the longest object is selected (panel 4). The skeleton of the longest object (panel 5) is then used to calculate the approximate length of the object and to generate a mask (panel 6) that is then used to evaluate methylation levels (panel 7). Image noise is then calculated in a mask surrounding the fiber (panel 8), which works as an indicator to alter the mask width in the previous step in cases where the noise is high. Dashed boxes indicate regions shown in detail on the right. For clarity, the skeleton is highlighted with green lines (panel 5a), the width of the adapted mask is indicated by magenta arrows (panel 6a), and the signals from the anti-5-methylcytosin antibody identified in the masks are highlighted in magenta (panels 7a and 8a). 5mC, 5-methylcytosine; kbps, kilobase pairs; px, pixels; Tel, telomere.

Fig. 4.

Methylation analysis pipeline for DNA fibers. Depiction of key steps in the workflow for telomere measurement and methylation analysis. Individual channels are exported from two-color images after thresholding (panels 1 and 2). The channel with the telomere FISH probe signal is then segmented (panel 3) in order to join the interrupted signals of the probe along the fiber and the longest object is selected (panel 4). The skeleton of the longest object (panel 5) is then used to calculate the approximate length of the object and to generate a mask (panel 6) that is then used to evaluate methylation levels (panel 7). Image noise is then calculated in a mask surrounding the fiber (panel 8), which works as an indicator to alter the mask width in the previous step in cases where the noise is high. Dashed boxes indicate regions shown in detail on the right. For clarity, the skeleton is highlighted with green lines (panel 5a), the width of the adapted mask is indicated by magenta arrows (panel 6a), and the signals from the anti-5-methylcytosin antibody identified in the masks are highlighted in magenta (panels 7a and 8a). 5mC, 5-methylcytosine; kbps, kilobase pairs; px, pixels; Tel, telomere.

Methylation analysis pipeline

We aimed to develop an image analysis tool that would be suitable for a quantitative evaluation of methylation along DNA fibers. The basic operations performed during the analysis are outlined in Fig. 4. Overall, individual fibers containing the FISH probe signal are cropped from raw data, and both the methylation and FISH probe channels are exported. Then, the channel with the FISH signal is used for image segmentation in the MATLAB Image Segmenter app to join the discrete signals along the fiber. Afterwards, the longest object is selected in order to eliminate background signals. A skeleton is then constructed for the FISH signal, which can be used to estimate the length of the repeat and is used afterwards to generate the adapted mask. This mask is then applied to the methylation channel, where the methylation levels are calculated as the ratio of pixels in the mask with intensity higher than 20% of the maximum pixel intensity in the mask area. Below, we use our tool for methylation mapping of functional repeats, namely telomeric repeats in N. tabacum, ribosomal genes in Arabidopsis thaliana, and centromeres in human embryonic stem cells before and after induced differentiation.

Telomeres

We analyzed the length of telomeres and methylation patterns in telomeric arrays of N. tabacum and MEFs, both models with relatively long telomeres. In contrast to plants, mammalian cells lack the system for cytosine methylation in the asymmetric (CpHpH) sequence context, and telomeric DNA is thus not methylated (Blasco, 2007; Vera et al., 2008) (Fig. 5A). Therefore, we used MEFs to calibrate and optimize individual steps in the image analysis. We observed ∼3–5% methylation content in the telomeres of MEFs, which is caused by the non-specific background antibody signal in the image. In agreement with previous findings (Majerová et al., 2014; Cokus et al., 2008), we detected methylated regions in the telomeres of N. tabacum (Fig. 5B). The mean methylation content was 35.41%, with high heterogeneity between individual fibers (n=66; s.d.=18.73%, Fig. 5C). Importantly, we did not see a correlation between telomere length and methylation levels (Fig. 5D; Spearman's ρ=0.32; Kendall's τ=0.21). Interestingly, the density of 5-methylcytosine signals was observed to be similar over the entire telomere fiber length, which clearly confirms our previous result obtained using a biochemical approach (Majerová et al., 2014) with greater credibility. Representative images of fibers from MEFs and N. tabacum are presented in Fig. S3 and Fig. S4, respectively.

Fig. 5.

Analysis of telomere length and methylation in DNA fibers extended from N. tabacum. (A) A telomeric fiber from a MEF used as a negative control shows no methylation. (B) Example of a methylated telomere in N. tabacum. (C) Average percentage methylation in the telomeres of N. tabacum (n=66). The line marks the mean, and the mean and s.d. values are shown. (D) Analysis of the correlation between the length and the methylation content of telomeres. n=66 pooled from two independent experiments. Linear regression (line) and confidence interval (shaded region) are indicated. Spearman's and Kendall's correlation coefficients are shown. Scale bars: 10 µm. 5mC, 5-methylcytosine; Tel, telomere.

Fig. 5.

Analysis of telomere length and methylation in DNA fibers extended from N. tabacum. (A) A telomeric fiber from a MEF used as a negative control shows no methylation. (B) Example of a methylated telomere in N. tabacum. (C) Average percentage methylation in the telomeres of N. tabacum (n=66). The line marks the mean, and the mean and s.d. values are shown. (D) Analysis of the correlation between the length and the methylation content of telomeres. n=66 pooled from two independent experiments. Linear regression (line) and confidence interval (shaded region) are indicated. Spearman's and Kendall's correlation coefficients are shown. Scale bars: 10 µm. 5mC, 5-methylcytosine; Tel, telomere.

Given that the degree of stretching in telomeric repeats might influence the evaluation of methylation density, we assessed whether the measured lengths of telomeres in our protocol match previously published data. For MEFs, we measured the average telomere length to be 34.7 kbp (s.d.=6.3 kbp), and previously reported lengths range from 30 to 60 kbp (Blasco, 2007; Varela et al., 2011). For N. tabacum, the average length we measured was 75 kbps (s.d.=42.2 kbp), and lengths cited in the literature range from 60 to 160 kbp (Fajkus et al., 1995; Majerová et al., 2011). Notably, telomere restriction fragment (TRF) analysis in both models often shows a large smear on the electrophoretogram, indicating telomeres with highly variable lengths. This is in line with our findings, which show quite significant standard deviations in the statistical analysis of telomere lengths.

Ribosomal genes

Next, we mapped the methylation of ribosomal genes in the plant model species A. thaliana. We identified three main types of rDNA fibers (n=84) with respect to the methylation profile. One type appeared to be largely unmethylated (Fig. 6A), while the other two showed different extents of DNA methylation (Fig. 6B,C). The unmethylated subset (37%), which is probably derived from the transcriptionally active intranucleolar rDNA, appeared unmethylated all along the fiber (Fig. 6A). This suggests that the promoter, spacer and coding sequences are all unmethylated, in line with results from a previous work showing that intranucleolar rDNA is unmethylated all along the rDNA repeat (Pontvianne et al., 2013). Some fibers with high levels of methylation had relatively uniform methylation (44%; Fig. 6B). Curiously, we sometimes observed clustering of methylation along fibers (19%; Fig. 6C); this might reflect the distribution of rDNA sequence variants, which are known to be differentially expressed in A. thaliana (Pavlištová et al., 2016). The average extent of methylation in the ribosomal genes of A. thaliana was measured to be ∼25.6% (Fig. 6D). In Arabidopsis, there are two clusters of rDNA repeats, which are located on chromosomes 2 and 4. One of these is inactivated during development and is not associated with the nucleolus. In some cases, we detected methylated rDNA fibers extending across several fields of view during microscopy, covering ∼250 µm (500 kbp, given a commonly used value of 2 kbp per 1 µm stretching; Fig. S5). We presume these fibers represent the rDNA cluster inactivated during development.

Fig. 6.

Methylation of ribosomal genes in Arabidopsis thaliana DNA fibers. (A–C) Examples of unmethylated rDNA fibers (A) and methylated rDNA fibers with diffuse or clustered methylation patterns (B and C, respectively). Two examples are shown for each group. Arrows in C indicate high levels of DNA methylation in rDNA clusters. (D) Evaluation of rDNA methylation (5-methylcytosine, 5mC) in fiber preparations. The line marks the mean, and the mean and s.d. values are shown. Data pooled from two independent experiments (n=84 in total). Scale bars: 10 µm.

Fig. 6.

Methylation of ribosomal genes in Arabidopsis thaliana DNA fibers. (A–C) Examples of unmethylated rDNA fibers (A) and methylated rDNA fibers with diffuse or clustered methylation patterns (B and C, respectively). Two examples are shown for each group. Arrows in C indicate high levels of DNA methylation in rDNA clusters. (D) Evaluation of rDNA methylation (5-methylcytosine, 5mC) in fiber preparations. The line marks the mean, and the mean and s.d. values are shown. Data pooled from two independent experiments (n=84 in total). Scale bars: 10 µm.

Centromeres

It has been reported that centromeric methylation is required for successful differentiation in mouse embryonic stem cells (Kanellopoulou, 2005) and that α-satellites in mouse embryonic stem cells are hypomethylated in comparison to differentiated MEFs (Tosolini et al., 2018). We quantified DNA methylation levels in the centromeres of pluripotent (n=42; Fig. 7A) and differentiated human embryonic stem cells (n=50; Fig. 7B). As described in the introduction, centromeres are essential chromosomal assembly sites for the kinetochores, involved in the attachment of spindle microtubules during cell division. We detected centromeres 21 and 13 using a FISH probe derived from the BAC clone pZ21A (Archidiacono et al., 1995; Politi et al., 2002), which is specific for the α-satellite subtypes in the centromeres of these chromosomes. We compared human stem cells pre- and post-differentiation to see whether differentiation reinforces the silencing of centromeres. We did not find statistically significant differences between pluripotent and differentiated cells, and both cell types showed considerable methylation in the centromeric region of chromosomes 13 and 21 (mean=32.48% and 39.45% for differentiated and human embryonic stem cells, respectively; Fig. 7C).

Fig. 7.

Comparing pericentromeric methylation patterns in DNA fibers prepared from human embryonic stem cells pre- and post-differentiation. (A,B) Examples of DNA methylation of human embryonic stem cells pre-differentiation (A) and post-differentiation (B) in the centromeric regions of chromosomes 13 and 21 (n=42 and n=50, respectively). (C) Statistical evaluation (Mann–Whitney U-test) of methylation levels between the two cell types. Points in the dot-plot indicate individual fibers, and the line marks the mean. Values of the mean, s.d. and statistical significance are shown. Data pooled from two independent experiments. DIF, differentiated cells; hESCs, human embryonic stem cells; pZ21A, centromeric FISH probe for chromosomes 13 and 21; 5mC, 5-methylcytosine. Scale bars: 10 µm.

Fig. 7.

Comparing pericentromeric methylation patterns in DNA fibers prepared from human embryonic stem cells pre- and post-differentiation. (A,B) Examples of DNA methylation of human embryonic stem cells pre-differentiation (A) and post-differentiation (B) in the centromeric regions of chromosomes 13 and 21 (n=42 and n=50, respectively). (C) Statistical evaluation (Mann–Whitney U-test) of methylation levels between the two cell types. Points in the dot-plot indicate individual fibers, and the line marks the mean. Values of the mean, s.d. and statistical significance are shown. Data pooled from two independent experiments. DIF, differentiated cells; hESCs, human embryonic stem cells; pZ21A, centromeric FISH probe for chromosomes 13 and 21; 5mC, 5-methylcytosine. Scale bars: 10 µm.

Unique insights

Fiber preparations can offer unique insights into the arrangement of chromatin states in situ. The key application in epigenetics is the mapping of various tandem repeats, such as telomeres, rDNA genes and centromeres. Although the emerging single-cell ChIP-seq approach can remove the problem of epigenetic heterogeneity within a cell population, it is still unable to distinguish identical repeats on the same or different chromosomes. Fiber extension techniques bypass this problem by allowing for a simultaneous detection of a chromosome-specific marker and a particular repeat, albeit with a lower resolution. We reason that fiber techniques could be used together with single-cell biochemical techniques to resolve both variability within cellular populations and along individual tandem repeat clusters. Notably, by analyzing individual fibers, these techniques make it possible to distinguish between a truly uniform pattern and the averaged result of diverse patterns in a specific chromosomal region. In terms of basic research, chromatin fibers could be used to map the distribution of proteins such as those of the structural maintenance of chromosomes (SMC) family and CCCTC-binding factor (CTCF), which are major chromosomal architectural proteins in mammalian cells (Banigan et al., 2020; Williams and Flavell, 2008). Fiber extension techniques can be also coupled to pulse replication labeling with 5-bromo-2′-deoxyuridine (BrdU) and 5-ethynyl-2′-deoxyuridine (EdU) to study the distribution of histone marks and replication factors with respect to a particular origin of replication, as recently reported by Wooten et al. (2020).

DNA fiber analysis of telomeres, as a single-molecule approach in terms of DNA, can yield important information on possible correlations between parameters like telomere length (Kahl et al., 2020) and methylation status (Wang et al., 2013). This is hardly achievable using classical biochemical approaches, which provide just averaged results. Outliers in telomere length can be identified (e.g. extremely long or short telomeres), and the density of DNA methylation can be mapped along the entire telomere length. It is also possible to evaluate dozens of telomeres on a single coverslip. We have further demonstrated that it is possible to perform a quantitative analysis of DNA methylation for other key repeats, using a custom MATLAB script.

Besides their value in basic research, DNA fibers have already been used in clinical settings to detect deletions in relevant targets, for example in the gene coding tuberous sclerosis 2 (TSC2) protein (Michalet, 1997) or in mapping deletions in the subtelomeric 4q35 region, which is related to facioscapulohumeral dystrophy (Nguyen et al., 2017). Integrating DNA methylation detection into DNA fiber analysis could also reveal methylation and demethylation patterns on target proto-oncogenes or tumor suppressor genes. Hypermethylation of CpG islands in the promoters of tumor suppressor genes causes their inactivation (Esteller, 2002), whereas demethylation of proto-oncogenes can lead to their aberrant expression (Agrawal et al., 2007). We envision that with the introduction of a well-designed battery of labeled oligonucleotide FISH probes (Yamada et al., 2011), it would be possible to determine the methylation status of several important genes in a single analysis, with multiplexing made possible by the use of different fluorophore tags for each target. In summary, it can be stated that DNA and chromatin fiber approaches clearly have strong potential for use in gaining a mechanistic understanding of epigenetic processes and genome architecture, as well as for applications in clinical settings.

In this paper, we show that mapping the methylation of repeats can produce novel findings. First, we demonstrate that although long stretches of ribosomal genes in A. thaliana can be demethylated all along the fiber, some display clusters of high DNA methylation, which could reflect the organization of different rDNA variants that are known to be present in the A. thaliana genome (Chandrasekhara et al., 2016; Havlová et al., 2016). A second insight is that telomeres of N. tabacum display very even methylation patterns, where the subtelomere-adjacent or terminal part of the telomere cannot be differentiated based on the levels of DNA methylation. Finally, our data also validates previous work showing that centromeric repeats in both stem cells and their differentiated counterparts are heavily methylated (Bar and Benvenisty, 2019; Chen et al., 2003; Tosolini et al., 2018).

Technical challenges of fiber preparation

General

In this paper, we propose three objective criteria that should be considered when evaluating extended fibers. First, super-resolution microscopy can be used to discern the parallel positioning of histone marks on the fiber, which indicates a bundle. Second, branching of fibers (Fig. 3H–K; Fig. S6E–G) is another indication of clustering, and fibers that branch into smaller fibers should not be evaluated. Third, the relative intensity of DNA counterstains works as a good guideline for discarding fiber bundles. In brief, only the fibers with the lowest intensity of DNA counterstain should be evaluated in any field of view (Fig. 3H–K; Fig. S6). We submit that following these recommendations (for both DNA and chromatin fibers) will improve the quality of fiber analysis. Moreover, we stress that the aspiration technique introduced in this paper for fiber preparation is essential for processing several slides the same way in one experiment and for low inter-experiment variability. Nevertheless, high-quality fibers can be prepared using other methods, as demonstrated in the literature (Sullivan, 2010; Wooten et al., 2020).

DNA fibers

In general, DNA fibers can be prepared with relatively good reproducibility, even without molecular combing, and are much easier to optimize than chromatin fibers. This is because the cells used are fully lysed and proteins are removed from the DNA template. It is important to consider that the determination of telomere length will be much more precise in models that have relatively long (>10 kbp) telomeres and that generate signals spanning at least a few micrometers. When it comes to resolution, imaging of DNA fibers does not require super-resolution microscopy, since the physical stretching of the DNA generally leads to discrete signals from FISH probes or antibodies. While the improved fiber-stretching technique presented in this paper remedies the issue of slide-to-slide variability, technical replicates can still show some variability in the DNA methylation levels, especially when using different anti-5-methylcytosine antibodies. For this reason, it is crucial to optimize the immunodetection protocol for each antibody so that the non-specific background is minimized. Importantly, we tested whether the stretching of telomeres can be influenced by using different types of glass coating. When we extended DNA fibers from MEFs, we observed a notable difference in telomere lengths when the fibers were produced on poly-L-lysine-coated coverslips or SuperFrost Plus coverglasses. On poly-L-lysine-coated coverslips, the average length of MEF telomeres was 28.9 kbp (n=73, s.d.=12.6), while on SuperFrost Plus coverglass the stretching was lower (20.4 kbp, n=77, s.d.=7.9).

Chromatin fibers

One of the major problems with chromatin fiber preparations is the low reproducibility in terms of the degree of stretching and protein retention. It has been shown that stretching can vary from ∼29 kbp/µm to ∼72 kbp/µm in the same experiment (Sullivan et al., 2011). Routinely, it is necessary to prepare several samples for each experimental condition to obtain at least one or two samples with properly stretched fibers that still retain histones. We propose that stretching using steady buffer removal (as described in this paper and similar to Wooten et al., 2020) yields the most consistent results because the alternatives (dragging with parafilm or dipping slides into the lysis buffer; Blower et al., 2002; Jin et al., 2004) require manual extension, which can be difficult to reproduce. Overall, chromatin fiber extension presents a compromise between trying to obtain maximum resolution and not losing most of the associated proteins at the same time. The homogeneity of the starting material is important to obtain relatively even stretching of fibers and produce reproducible results. As an example, producing fibers from cultured mammalian cells is much easier to optimize than preparing fibers from nuclei isolated from plant seedlings, where nuclei with different ploidy are present.

From our experience, chromatin fibers can be used to assess the presence and the distribution of a particular protein or histone mark in a genomic region. This analysis, however, struggles to produce quantitative results due to issues with protein retention, variable stretching and partial exclusivity of immunodetection and FISH.

Data interpretation and experimental design

Several pitfalls need to be avoided when evaluating the results from chromatin and DNA fiber analysis. When investigating the distribution of a particular protein of interest, gaps in the signal of this protein along the fiber can mean that the protein is absent (either because it is not bound at that site or because it has dissociated during extension), or that it is present and not detected by antibodies. Bundles of chromatin or DNA fibers, as described here and previously (Sullivan, 2010), should be evaluated with caution. In our analysis of methylation profiles, we avoided quantifying methylation on fibers that overlapped with other fibers in the field of view to avoid inaccuracies. Our motivation to develop a script to analyze methylation levels stemmed from the fact that the signal from FISH probes and anti-5-methylcytosine antibodies can produce discrete signals along individual fibers that do not necessarily colocalize, even though they are clearly on the same fiber. This means that using colocalization tools (for a review see Lagache et al., 2015) to measure methylation levels on fibers can be inaccurate for quantitative analysis.

It is important to contemplate all the challenges referenced above during experimental design. We recommend choosing the model system for a particular problem with consideration of the inherent limitations of the techniques. For instance, organisms with short telomeric repeats (e.g. Physcomitrella patens or A. thaliana) are not suitable for the study of methylation patterns in telomeric regions. In such cases, the region of interest is relatively short (a few kbp, which translates into ∼1 µm stretch on a fiber), and it is difficult to interpret the results. As a rule of thumb, the more homogeneous the starting cell population, the better for producing extended fibers. In this sense mammalian cell lines are a ‘go to’ choice. For plant models, it might be advantageous to use suspension cell lines, but this would require protocol optimization to include a cell wall digestion step. Finally, validation of both antibodies and FISH probes on isolated nuclei from plant cells or mammalian cell lines is important, since strong background or low signal-to-noise ratio can lead to false positive results (for the validation of probes used in this paper, see Fig. S7).

In conclusion, our paper introduces an optimized, reproducible fiber preparation protocol that facilitates parallel preparation of several samples at the same time. Importantly, we show how different super-resolution microscopy techniques can be successfully applied to study the distribution of histone marks in chromatin fibers and how DNA fibers can be a versatile tool to study overall methylation patterns in telomeric, pericentromeric and ribosomal repeats. Even though widefield imaging is sufficient for some types of analysis (e.g. analysis of DNA fibers), we note that using super-resolution microscopy can be useful when assessing samples with clustered chromatin fibers and overlapping histone mark signals. First, colocalization analysis can yield false positive results due to limited resolution of widefield imaging. Second, the fluorescence signals of particular histone marks (especially those marking actively transcribed regions, such as H3K4me3 or H3 pan-acetylation) might appear continuous in widefield microscopy, but they are clearly discrete when examined using super-resolution microscopy. Last, but not least, STED and STORM microscopy allow clear discrimination between fiber bundles and single fibers. This distinction is vital for data analysis and might not be straightforward for inexperienced users of conventional microscopy. We have demonstrated that DNA fiber extension is useful for the study of individual clusters of repeats (e.g. one telomere or one nucleolus organizer region) and their methylation profiles using our newly developed image analysis tool. Besides the experimental advances presented above, we have also discussed technical challenges and limitations concerning these techniques and offered suggestions for experimental design considerations when planning to use extended fiber methods.

Plant growth

Arabidopsis thaliana seeds were sterilized (90% ethanol, 5 min) and plated on half-strength agar Murashige and Skoog medium (½ MS medium; Duchefa Biochemie, Haarlem, The Netherlands) with 1% sucrose. Seeds of Nicotiana tabacum were first drenched in distilled water, then sterilized by two washes in 90% bleach and 10% Tween-20 (2 min each). Seeds were then thoroughly washed in distilled water (3×10 min) and transferred onto ½ MS medium with 1% sucrose. After 1 day stratification (4°C in the dark), plates were transferred to a growth chamber and grown for up to 2 weeks under long-day conditions (16 h light at 21°C, 8 h dark at 19°C; 50–60% relative humidity). Seedlings were then collected directly from plates and used for isolation of nuclei.

Mammalian cell culture

The human embryonic stem cell line CCTL14 has been characterized previously (The International Stem Cell Initiative, 2007) and was propagated on mitotically inactivated mouse embryonic fibroblasts (MEFs) from the CD1 mouse strain (Dvorak et al., 2005), and experiments were performed in feeder-free culture (Krutá et al., 2014; Kunova et al., 2013). For differentiation, CCTL14 human embryonic stem cells were propagated in human embryonic stem cell medium (hES; Fojtík et al., 2021) without FGF2 and supplemented with 10% fetal bovine serum (10270106; Thermo Fisher Scientific, Waltham, MA, USA) for three successive subcultures (12 days) until they started showing morphological changes related to differentiation.

MEFs from the BL6 mouse strain (derived from C57BL/6 mice at the Department of Biology, Faculty of Medicine, Masaryk University) used for experiments were cultivated in KnockOut Dulbecco's modified Eagle's medium (10829018; Thermo Fisher Scientific) supplemented with 10% fetal bovine serum, 0.1 mM MEM non-essential amino acids (11140-035; Thermo Fisher Scientific), 1% L-glutamine (XC-T1715; Biosera, Boussens, France), 1% penicillin-streptomycin (Gibco, Waltham, MA, USA) and 100 µmol/l β-mercaptoethanol (Sigma-Aldrich, St Louis, MO, USA).

HEK293T cells (from the DSMZ cell culture collection; ACC635) were maintained in Dulbecco's modified Eagle's Medium supplemented with 10% fetal bovine serum, 0.1 mM non-essential amino acids, 1% penicillin-streptomycin and 1% L-glutamine.

After reaching 70–80% confluency, cells were trypsinized, washed three times with ice-cold 1× phosphate-buffered saline (PBS), counted and stored in 50% glycerol in 1×PBS at −20°C.

Isolation of unfixed plant nuclei

Arabidopsis thaliana

A total of 0.3 g of 10–14-day-old seedlings of A. thaliana was chopped with a razor blade in 500 µl of nuclear isolation buffer (NIB; 0.5 M sucrose, 10 mM EDTA, 2.5 mM dithiothreitol, 100 mM KCl, 1 mM spermine and 4 mM spermidine in 10 mM Tris-HCl, pH 9.5). After chopping, the solution was filtered through a 50 μm and 30 μm pore size disposable filters (CellTrics, Sysmex, Germany). Afterwards, the filtrate was supplemented with 1/10 volume of 10% Triton X-100 in NIB and centrifuged at 2000 g for 10 min at 4°C. Supernatant was removed after centrifugation, and the resulting pellet was resuspended in 1× NIB, then supplemented with one volume of 100% glycerol. These aliquots were stored at −20°C.

Nicotiana tabacum

To extract N. tabacum nuclei, 15-day-old seedlings were flash frozen in liquid nitrogen, then ground to a powder in liquid nitrogen using mortar and pestle. Powder was transferred into Galbraith buffer (45 mM MgCl2, 20 mM MOPS, 0.1% Triton X-100 and 30 mM sodium citrate; Galbraith et al., 1983), incubated on ice for 10 min, then filtered through a 50 µm pore size nylon filter (CellTrics, Sysmex, Germany). Filtrate was then centrifuged at 2000 g for 10 min at 4°C. Supernatant was removed and the remaining pellet was resuspended in Galbraith buffer, aliquoted and supplemented with an equal volume of 100% glycerol. Aliquots were stored at −20°C.

DNA fiber preparation

To produce DNA fibers, aliquots of unfixed isolated nuclei or trypsinized cells were centrifuged at 2000 g for 10 min at 4°C. The pellet containing nuclei was resuspended in 75 mM KCl and left at room temperature for 20 min. Nuclei were then spotted onto slides by drawing a line vertically on the glass slide (3 µl per line). Several lines can be spotted in tandem, spaced ∼1 cm apart. Slides were left to dry at room temperature until the point where the edges of the spotted lines appeared viscous and sticky. Then, slides were immersed in the lysis buffer (0.5% SDS, 5 mM EDTA and 100 mM Tris-HCl) in a Coplin jar and incubated for 10 min. Fibers were extended by the aspiration of the lysis buffer using a vacuum pump equipped with a 10 µl pipette tip. After aspiration, slides were briefly air-dried, fixed in cold methanol:acetic acid (3:1) fixative for 10 min, air-dried and baked at 60°C for 1 h in an oven. Slides were then rinsed in 2× SSC, fixed in 4% paraformaldehyde, washed again twice in 2× SSC and dehydrated in ethanol series (70%, 80%, 96%; 2 min each). Slides were finally air-dried again and aged overnight at 4°C. Denaturation of the probe and the slides was performed.

FISH probe design and labeling

To visualize human and mouse telomeric repeats, we used a custom biotin-labeled Affinity Plus probe (Integrated DNA Technologies, Coralville, IA, USA) with the sequence 5′-biotin-(CCC*TAA)4-3′, where the asterisk indicates the Affinity plus modified nucleotide. Canonical telomeric repeats in plants (N. tabacum) were visualized using a custom biotin-labeled LNA probe (Qiagen, Germantown, MD, USA) with the sequence: 5′-biotin-CCTAAA(CCCTAA)3-biotin-3′. The plant rDNA probe was prepared by nick translation, as described previously (Mandáková and Lysak, 2008). BAC clone T15P10 (containing 45S rDNA) was isolated using Macherey Nagel NucleoBond Xtra Midi, and 1 µg of plasmid was used as a template for nick-translation labeling with biotin-dUTP (NU-803-BIO16-S; Jena Bioscience, Jena Germany). The centromeric probe for chromosomes 21 and 13 was prepared by nick-translation labeling of 1 µg of pZ21A BAC (Archidiacono et al., 1995) using the Atto488 NT Labeling Kit (PP-305S-488; Jena Bioscience).

FISH and immunodetection of 5-methylcytosine on DNA fibers

Slides were denatured on a hotplate (80°C, 2 min), independently from the hybridization mix. The hybridization mix consisted of the FISH probe diluted in 50% hybridization buffer (2× SSC, 50% deionized formamide, 10% dextran sulfate) in a 1:40 ratio. The hybridization mix was denatured in a heating block (80°C, 4 min) and was then added to the slides immediately after denaturation of the slides, covered with a clean 22×40 mm coverslip and allowed to hybridize for 3 h or overnight at 37°C without sealing off the coverslip with rubber cement. Samples were then washed at 42°C, twice in 50% formamide and once in 2× SSC, prior to blocking. Blocking was performed at room temperature using 5% fetal calf serum (FCS) in 2× SSC for 30 min. Incubation with anti-5-methylcytosine antibody (C15200006-100; Diagenode, Seraing, Belgium) diluted 1:200 in 2× SSC containing 5% FCS was conducted for 30 min at room temperature. The washing steps included incubation in 2× SSCT (2× SSC with 0.05% Tween 20) for 3×5 min. Secondary antibody (Alexa Fluor 488-conjugated goat anti-mouse IgG; A11001; Invitrogen, Waltham, MA, USA) and Alexa Fluor 594-conjugated streptavidin (S11227; Thermo Fisher Scientific) were diluted 1:200 in 2× SSC containing 5% FCS and incubated for 30 min at room temperature, followed by washing in 2× SSCT for 3×5 min. Samples were post-fixed in 4% paraformaldehyde for 5 min, then rinsed in 1× PBS and mounted in DAPI (4′,6-diamidino-2-phenylindole, dihydrochloride; D1306; Invitrogen) and Vectashield (H-1000-10; Vector laboratories, Burlingame, CA, USA).

Chromatin fiber preparation

Chromatin fiber preparation after the extension step followed the protocol outlined by Cohen et al. (2009), with minor modifications. Glycerol aliquots of human or plant nuclei were centrifuged at 2000 g for 10 min at 4°C. Afterwards, nuclei were resuspended 75 mM KCl and incubated for 20 min at room temperature. Then, nuclei were cytocentrifuged (8 µl of nuclei suspension per cytocentrifugation chamber) onto poly-L-lysine-coated coverslips or glass slides (400 g; 4 min; 4°C). After cytocentrifugation, excess liquid was aspirated with a pipette and the slides were immersed in the nuclei lysis buffer (NLB; 350 mM NaCl, 1% Triton X-100, 500 mM urea and 25 mM Tris-HCl). In this case, the coverslips were mounted in a coverslip holder and introduced into a 100 ml beaker. After 10–14 min of incubation (time of incubation needs to be adjusted empirically), chromatin fibers were extended by the aspiration of the lysis buffer using a vacuum pump equipped with a 10 µl pipette tip, which ensures steady removal of the buffer.

To produce chromatin fibers by drying, we added 25 µl of NLB directly onto the spot of cytocentrifuged cells and covered the drop with a 1×1 cm parafilm coverslip. After 60 min, we removed the parafilm coverslip and proceeded with the protocol as described below. Note, the composition of the NLB for MEFs was adjusted to 500 mM NaCl, 1% Triton X-100, 500 mM urea and 25 mM Tris-HCl. Coating of coverslips and coverglasses with poly-L-lysine (P8920; Sigma-Aldrich, St Louis, MO, USA) was performed according to the manufacturer's instructions.

Immunofluorescence on chromatin fibers

After fiber extension, coverslips were immersed in KCM buffer (120 mM KCl, 20 mM NaCl, 10 mM Tris-HCl, 0.5 mM EDTA and 0.1% Triton X-100) for 30 min. Coverslips were then blocked in 5% BSA in 1× KCM for 30 min. After a quick rinse in KCM, primary antibodies (anti-H3K4me3, ab8580, Abcam, Cambridge, UK; anti-H3K27me3, ab6002, Abcam; anti-H3, ab1791, Abcam; anti-H3K9me3, 07-442, Sigma-Aldrich) were added in a 1:100 dilution to KCM buffer containing 5% BSA. Incubation took place at room temperature for 1 h in a humid chamber. Samples were then rinsed 3× in KCM buffer and incubated with following secondary antibodies: Alexa Fluor 647-conjugated anti-rabbit IgG (ab150075; Abcam) and CF 680-labeled anti-mouse IgG (20065-1; Biotium, Fremont, CA, USA) for dSTORM; Alexa Fluor 594-conjugated anti-rabbit IgG (A21207; Thermo Fisher Scientific) and Abberior Star RED-labeled anti-mouse IgG (STRED-1001; Abberior, Gottingen, Germany) for STED. Antibodies were diluted 1:200 in KCM containing 5% BSA, the incubation time was 45 min in a humid chamber at room temperature. Coverslips were washed 3× in KCM buffer and then fixed in fresh 4% paraformaldehyde for 10 min at room temperature. After two washes in 1× PBS, the samples were either prepared for FISH or mounted in DAPI and Vectashield.

Widefield microscopy

DNA fibers were imaged on an upright epifluorescence microscope Zeiss AxioImager Z2 using a 63× (1.40 NA) Plan Apochromat or a 100× (1.40 NA) Plan Apochromat objective, using appropriate filters. Images were captured on a Hamamatsu ORCA Flash camera with 2048×2048 pixel resolution.

Structured-illumination microscopy

Structured-illumination microscopy was performed on Nikon 3D N-SIM microscope (inverted Nikon Eclipse Ti-E; Nikon, Tokyo, Japan) equipped with a Nikon CFI SR Apo TIRF objective (100× oil, NA 1.49). Structured illumination pattern projected into the sample plane was created on a diffraction grating block (100 EX V-R 3D-SIM) for laser wavelengths 488, 561 and 647 nm. Excitation and emission light were separated using filter cubes with appropriate filter sets SIM488 (excitation, 470–490 nm; emission, 500–545 nm), SIM561 (excitation, 556–566 nm; emission, 570–640 nm) and SIM647 (excitation, 590–650 nm; emission, 663–738 nm). Emission light was projected through a 2.5× relay lens onto the chip of the EM CCD camera (Andor iXon Ultra DU897, 10 MHz at 14-bit, 512×512 pixels). Laser intensity, EM gain and camera exposure time were set independently for each excitation wavelength. Intensity of fluorescence signal was held within the linear range of the camera. Fifteen images (three rotations and five phase shifts) were recorded for every plane and color. SIM data were processed in NIS-Elements AR (Nikon). Before sample measurement, the symmetry of the point spread function was checked with 100 nm red fluorescent beads (580/605, carboxylate-modified microspheres, Life Technologies, Waltham, MA, USA) mounted in Abberior Liquid Mount medium (Abberior, Gottingen, Germany), and optimized by adjusting the objective correction collar.

STED microscopy

Two-color STED microscopy was performed on an Abberior Instruments Expert Line STED system equipped with a Nikon Eclipse Ti-E microscope body and a Nikon CFI Plan Apo Lambda 60× Oil, NA 1.40 objective. The sample was illuminated with pulsed 561 nm and 640 nm lasers and depleted by a pulsed 775 nm STED laser with a 2D donut shape formed by a spatial light modulator. The fluorescence signal passed the pinhole set to 1 AU, was filtered by emission filters (580–630 nm and 650–720 nm) and was detected by single-photon counting modules (Excelitas Technologies, Waltham, MA, USA). STED images were scanned with a pixel size of 20 nm×20 nm, 10 μs dwell time, and in-line interleaved acquisition with time-gated detection using the Inspector software (Abberior, Gottingen, Germany). DAPI images were acquired using a 405 nm laser in conventional confocal mode.

Single-molecule localization microscopy

Single-molecule localization microscopy experiments (using STORM) were acquired on a N-STORM system (Nikon) equipped with a Nikon Eclipse Ti body, a Nikon CFI HP Apo TIRF 100× Oil/NA 1.49 objective, a perfect focus system (hardware autofocus for the stabilization of z-position) and an EM CCD Andor iXon Ultra DU897 camera. For STORM, a 647 nm laser beam (fiber output power 125 mW) was focused by 2× magnifying lenses in the TIRF Illuminator unit to reach higher laser intensity in the sample. The sample was illuminated by highly inclined laser beam to improve the signal-to-noise ratio. Excitation and emission light were separated by a 405/488/561/647 nm Laser Quad Band filter cube (TRF89902, Chroma), specifically, the far-red emission was collected in the emission range 674–785 nm. For spectral de-mixing, the microscope was adapted by insertion of W-VIEW GEMINI Image splitting optics (A12801-01, Hamamatsu) to the right port of microscope in front of the EM CCD camera. The fluorescence emission was split to shorter and longer wavelengths by a 700 nm dichroic beam splitter (FF700-Di01, Semrock). All emission wavelengths were detected by the EM CCD camera. Sequences of frames (typically 20,000–30,000 frames) were acquired in NIS-Elements software (version 5.11.02) in Fast Time-Lapse mode to capture the images with a frame rate of 31.3 frames per second. The time series were measured in a selected region of interest with an image size of 512×128 pixels and a pixel size of 107 nm (the result of combining the 100×/1.49 NA objective, a 1.5× zoom adapter and the camera pixel size of 16×16 μm2). Before the acquisition, the sample was illuminated by 647 nm laser to switch off most of the molecules to the dark state. The acquisition was started after detecting single molecules only. The imaging buffer [50 mM Tris-HCl, 10 mM NaCl and 10% (w/v) glucose (pH 8) with 50 mM β-mercaptoethylamine (MEA, 30070; Sigma-Aldrich), 1.1 mg/ml glucose oxidase from Aspergillus niger (G2133; Sigma-Aldrich) and 100 µg/ml catalase from bovine liver (C40; Sigma-Aldrich)] was freshly prepared on ice before imaging. Image reconstruction was performed in rapidSTORM software (Wolter et al., 2012) followed by drift correction based on cross-correlation analysis in ThunderSTORM (Ovesný et al., 2014). Localizations were filtered by localization intensity and precision. For two-color images, spectral de-mixing of Alexa Fluor 647 and CF 680 signals was performed in SD-mixer software (Tadeus et al., 2015). Transformations of localization tables between the software used was performed using custom-written Python scripts.

Methylation profile and telomere length analysis, statistical evaluation and image presentation

For methylation profile analysis, we thresholded raw images and exported the channels with the FISH signal (rDNA, centromeres or telomeres) and the 5-methylcytosine signal separately. Images were then converted from RGB to grayscale. The image with the FISH signal was segmented in MATLAB Image Segmenter using mathematical morphological operations in order to join interrupted FISH signals. We developed a custom script to perform the following operations. We filtered out the longest detected object in the FISH image. Then we obtained the skeleton of the longest object and calculated the length of the repeat (relevant mostly for telomeres) by summing the pixels of the skeleton. In order to evaluate the methylation profile, we generated a mask based on the skeleton of the FISH signal, then overlapped this mask with the methylation channel. Last, we calculated the coverage of the methylation along the fiber, while excluding pixels in the methylation channel that were below 20% intensity of the highest intensity pixel in the mask of the methylation channel. Methylation levels and telomere lengths were evaluated using non-parametric statistical tests, given that data did not show normal distribution. Correlation between methylation levels and telomere length was evaluated using Spearman's rank correlation coefficient and Kendall's rank correlation coefficient. Methylation levels in human embryonic stem cells and differentiated cells were compared using a Mann–Whitney U-test. The code for our tool is available in a GitHub repository (https://github.com/amkilar/MELANIE). The images showing insets and detailed views of chromatin and DNA fibers in this article have been smoothened using bicubic interpolation.

We are grateful to T. Janovič from the Laboratory of Interactions and Functions of Essential Biomolecules (CEITEC, Brno) for providing HEK293T cells. We thank V. Peška for kindly providing a custom telomeric FISH probe for the mammalian-type telomeres. We would like to acknowledge these core facilities (CF) for their support with obtaining data presented in this paper: CELLIM of CEITEC MU and Imaging Methods CF at BIOCEV, both supported by MEYS CR (LM2018129 Czech-BioImaging).

Author contributions

Conceptualization: M.F., V.R., M.D., J.F.; Methodology: M.F., A.K., P.F., M.O.; Software: A.K.; Validation: A.K.; Investigation: M.F., A.K., P.F., M.O.; Resources: P.F., A.B., V.R., M.D., J.F.; Data curation: M.F.; Writing - original draft: M.F.; Writing - review & editing: M.F., M.D., J.F.; Visualization: M.F., M.O.; Supervision: A.B., M.D., J.F.; Project administration: M.F., M.D.; Funding acquisition: A.B., V.R., M.D., J.F.

Funding

This research was supported by the Grantová Agentura České Republiky (project 19-11880Y to M.F and M.D.); by the Ministerstvo Školství, Mládeže a Tělovýchovy (projects INTER-COST LTC18048 and LTC20003 to J.F., M.F. and M.D.); and by the European Regional Development Fund (project SINGING PLANT CZ.02.1.01/0.0/0.0/16_026/0008446 to A.K., J.F. and M.F.). M.O. acknowledges financial support from the European Regional Development Fund (project CZ.02.1.01/0.0/0.0/16_013/0001775). P.F. and V.R. were supported by the Ministerstvo Školství, Mládeže a Tělovýchovy and the European Regional Development Fund (project ENOCH CZ.02.1.01/0.0/0.0/16_019/0000868) and by project LQ1605 from the National Program of Sustainability II (Ministerstvo Školství, Mládeže a Tělovýchovy). V.R. was further supported by the grant agency of the Ministerstvo Zdravotnictví Ceské Republiky (NU20-06-00156).

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.258374

Achrem
,
M.
,
Szućko
,
I.
and
Kalinka
,
A.
(
2020
).
The epigenetic regulation of centromeres and telomeres in plants and animals
.
Comp. Cytogenet.
14
,
265
-
311
.
Adamusová
,
K.
,
Khosravi
,
S.
,
Fujimoto
,
S.
,
Houben
,
A.
,
Matsunaga
,
S.
,
Fajkus
,
J.
and
Fojtová
,
M.
(
2020
).
Two combinatorial patterns of telomere histone marks in plants with canonical and non-canonical telomere repeats
.
Plant J.
102
,
678
-
687
.
Agrawal
,
A.
,
Murphy
,
R. F.
and
Agrawal
,
D. K.
(
2007
).
DNA methylation in breast and colorectal cancers
.
Mod. Pathol.
20
,
711
-
721
.
Archidiacono
,
N.
,
Antonacci
,
R.
,
Marzella
,
R.
,
Finelli
,
P.
,
Lonoce
,
A.
and
Rocchi
,
M.
(
1995
).
Comparative mapping of human alphoid sequences in great apes using fluorescence in situ hybridization
.
Genomics
25
,
477
-
484
.
Banigan
,
E. J.
,
van den Berg
,
A. A.
,
Brandão
,
H. B.
,
Marko
,
J. F.
and
Mirny
,
L. A.
(
2020
).
Chromosome organization by one-sided and two-sided loop extrusion
.
eLife
9
,
e53558
.
Bar
,
S.
and
Benvenisty
,
N.
(
2019
).
Epigenetic aberrations in human pluripotent stem cells
.
EMBO J.
38
,
e101033
.
Benarroch-Popivker
,
D.
,
Pisano
,
S.
,
Mendez-Bermudez
,
A.
,
Lototska
,
L.
,
Kaur
,
P.
,
Bauwens
,
S.
,
Djerbi
,
N.
,
Latrick
,
C. M.
,
Fraisier
,
V.
,
Pei
,
B.
et al. 
(
2016
).
TRF2-mediated control of telomere DNA topology as a mechanism for chromosome-end protection
.
Mol. Cell
61
,
274
-
286
.
Bensimon
,
A.
,
Simon
,
A.
,
Chiffaudel
,
A.
,
Croquette
,
V.
,
Heslot
,
F.
and
Bensimon
,
D.
(
1994
).
Alignment and sensitive detection of DNA by a moving interface
.
Science
265
,
2096
-
2098
.
Blasco
,
M. A.
(
2007
).
The epigenetic regulation of mammalian telomeres
.
Nat. Rev. Genet.
8
,
299
-
309
.
Blower
,
M. D.
and
Karpen
,
G. H.
(
2001
).
The role of Drosophila CID in kinetochore formation, cell-cycle progression and heterochromatin interactions
.
Nat. Cell Biol.
3
,
730
-
739
.
Blower
,
M. D.
,
Sullivan
,
B. A.
and
Karpen
,
G. H.
(
2002
).
Conserved organization of centromeric chromatin in flies and humans
.
Dev. Cell
2
,
319
-
330
.
Caburet
,
S.
(
2005
).
Human ribosomal RNA gene arrays display a broad range of palindromic structures
.
Genome Res.
15
,
1079
-
1085
.
Chandrasekhara
,
C.
,
Mohannath
,
G.
,
Blevins
,
T.
,
Pontvianne
,
F.
and
Pikaard
,
C. S.
(
2016
).
Chromosome-specific NOR inactivation explains selective rRNA gene silencing and dosage control in Arabidopsis
.
Genes Dev.
30
,
177
-
190
.
Chen
,
T.
,
Ueda
,
Y.
,
Dodge
,
J. E.
,
Wang
,
Z.
and
Li
,
E.
(
2003
).
Establishment and maintenance of genomic methylation patterns in mouse embryonic stem cells by Dnmt3a and Dnmt3b
.
Mol. Cell. Biol.
23
,
5594
-
5605
.
Clark
,
S. J.
,
Smallwood
,
S. A.
,
Lee
,
H. J.
,
Krueger
,
F.
,
Reik
,
W.
and
Kelsey
,
G.
(
2017
).
Genome-wide base-resolution mapping of DNA methylation in single cells using single-cell bisulfite sequencing (scBS-seq)
.
Nat. Protoc.
12
,
534
-
547
.
Cohen
,
S. M.
,
Chastain
,
P. D.
,
Cordeiro-Stone
,
M.
and
Kaufman
,
D. G.
(
2009
).
DNA replication and the GINS complex: localization on extended chromatin fibers
.
Epigenetics Chromatin
2
,
6
.
Cokus
,
S. J.
,
Feng
,
S.
,
Zhang
,
X.
,
Chen
,
Z.
,
Merriman
,
B.
,
Haudenschild
,
C. D.
,
Pradhan
,
S.
,
Nelson
,
S. F.
,
Pellegrini
,
M.
and
Jacobsen
,
S. E.
(
2008
).
Shotgun bisulphite sequencing of the Arabidopsis genome reveals DNA methylation patterning
.
Nature
452
,
215
-
219
.
Cusanovich
,
D. A.
,
Hill
,
A. J.
,
Aghamirzaie
,
D.
,
Daza
,
R. M.
,
Pliner
,
H. A.
,
Berletch
,
J. B.
,
Filippova
,
G. N.
,
Huang
,
X.
,
Christiansen
,
L.
,
DeWitt
,
W. S.
et al. 
(
2018
).
A single-cell atlas of in vivo mammalian chromatin accessibility
.
Cell
174
,
1309
-
1324.e18
.
Doksani
,
Y.
,
Wu
,
J. Y.
,
de Lange
,
T.
and
Zhuang
,
X.
(
2013
).
Super-resolution fluorescence imaging of telomeres reveals TRF2-Dependent T-loop formation
.
Cell
155
,
345
-
356
.
Dvořáčková
,
M.
,
Fojtová
,
M.
and
Fajkus
,
J.
(
2015
).
Chromatin dynamics of plant telomeres and ribosomal genes
.
Plant J.
83
,
18
-
37
.
Dvorak
,
P.
,
Dvorakova
,
D.
,
Koskova
,
S.
,
Vodinska
,
M.
,
Najvirtova
,
M.
,
Krekac
,
D.
and
Hampl
,
A.
(
2005
).
Expression and potential role of fibroblast growth factor 2 and its receptors in human embryonic stem cells: expression and potential role of fibroblast growth factor 2 and its receptors in human embryonic stem cells
.
Stem Cells
23
,
1200
-
1211
.
Esteller
,
M.
(
2002
).
CpG island hypermethylation and tumor suppressor genes: a booming present, a brighter future
.
Oncogene
21
,
5427
-
5440
.
Fajkus
,
J.
,
Kovařík
,
A.
,
mKrálovics
,
R.
and
Bezděk
,
M.
(
1995
).
Organization of telomeric and subtelomeric chromatin in the higher plant Nicotiana tabacum
.
Molec. Gen. Genet.
247
,
633
-
638
.
Fojtík
,
P.
,
Beckerová
,
D.
,
Holomková
,
K.
,
Šenfluk
,
M.
and
Rotrekl
,
V.
(
2021
).
Both hypoxia-inducible factor 1 and MAPK signaling pathway attenuate PI3K/AKT via suppression of reactive oxygen species in human pluripotent stem cells
.
Front. Cell Dev. Biol.
8
,
607444
.
Fojtová
,
M.
and
Fajkus
,
J.
(
2014
).
Epigenetic regulation of telomere maintenance
.
Cytogenet Genome Res.
143
,
125
-
135
.
Fransz
,
P. F.
,
Alonso-Blanco
,
C.
,
Liharska
,
T. B.
,
Peeters
,
A. J. M.
,
Zabel
,
P.
and
Jong
,
J. H.
(
1996
).
High-resolution physical mapping in Arabidopsis thaliana and tomato by fluorescence in situ hybridization to extended DNA fibres
.
Plant J.
9
,
421
-
430
.
Galbraith
,
D. W.
,
Harkins
,
K. R.
,
Maddox
,
J. M.
,
Ayres
,
N. M.
,
Sharma
,
D. P.
and
Firoozabady
,
E.
(
1983
).
Rapid flow cytometric analysis of the cell cycle in intact plant tissues
.
Science
220
,
1049
-
1051
.
Garcia-Blanco
,
M. A.
,
Miller
,
D. D.
and
Sheetz
,
M. P.
(
1995
).
Nuclear spreads: I. Visualization of bipartite ribosomal RNA domains
.
J. Cell Biol.
128
,
15
-
27
.
Grosselin
,
K.
,
Durand
,
A.
,
Marsolier
,
J.
,
Poitou
,
A.
,
Marangoni
,
E.
,
Nemati
,
F.
,
Dahmani
,
A.
,
Lameiras
,
S.
,
Reyal
,
F.
,
Frenoy
,
O.
et al. 
(
2019
).
High-throughput single-cell ChIP-seq identifies heterogeneity of chromatin states in breast cancer
.
Nat. Genet.
51
,
1060
-
1066
.
Hartley
,
G.
and
O'Neill
,
R.
(
2019
).
Centromere repeats: hidden gems of the genome
.
Genes
10
,
223
.
Havlová
,
K.
,
Dvořáčková
,
M.
,
Peiro
,
R.
,
Abia
,
D.
,
Mozgová
,
I.
,
Vansáčová
,
L.
,
Gutierrez
,
C.
and
Fajkus
,
J.
(
2016
).
Variation of 45S rDNA intergenic spacers in Arabidopsis thaliana
.
Plant Mol. Biol.
92
,
457
-
471
.
Jackson
,
D. A.
and
Pombo
,
A.
(
1998
).
Replicon clusters are stable units of chromosome structure: evidence that nuclear organization contributes to the efficient activation and propagation of S phase in human cells
.
J. Cell Biol.
140
,
1285
-
1295
.
Jin
,
W.
,
Melo
,
J. R.
,
Nagaki
,
K.
,
Talbert
,
P. B.
,
Henikoff
,
S.
,
Dawe
,
R. K.
and
Jiang
,
J.
(
2004
).
Maize centromeres: organization and functional adaptation in the genetic background of oat
.
Plant Cell
16
,
571
-
581
.
Kahl
,
V. F. S.
,
Allen
,
J. A. M.
,
Nelson
,
C. B.
,
Sobinoff
,
A. P.
,
Lee
,
M.
,
Kilo
,
T.
,
Vasireddy
,
R. S.
and
Pickett
,
H. A.
(
2020
).
Telomere length measurement by molecular combing
.
Front. Cell Dev. Biol.
8
,
493
.
Kanellopoulou
,
C.
(
2005
).
Dicer-deficient mouse embryonic stem cells are defective in differentiation and centromeric silencing
.
Genes Dev.
19
,
489
-
501
.
Kaykov
,
A.
,
Taillefumier
,
T.
,
Bensimon
,
A.
and
Nurse
,
P.
(
2016
).
Molecular combing of single DNA molecules on the 10 megabase scale
.
Sci. Rep.
6
,
19636
.
Krutá
,
M.
,
Šeneklová
,
M.
,
Raška
,
J.
,
Salykin
,
A.
,
Zerzánková
,
L.
,
Pešl
,
M.
,
Bártová
,
E.
,
Franek
,
M.
,
Baumeisterová
,
A.
,
Košková
,
S.
et al. 
(
2014
).
Mutation frequency dynamics in HPRT locus in culture-adapted human embryonic stem cells and induced pluripotent stem cells correspond to their differentiated counterparts
.
Stem Cells Dev.
23
,
2443
-
2454
.
Kunova
,
M.
,
Matulka
,
K.
,
Eiselleova
,
L.
,
Salykin
,
A.
,
Kubikova
,
I.
,
Kyrylenko
,
S.
,
Hampl
,
A.
and
Dvorak
,
P.
(
2013
).
Adaptation to robust monolayer expansion produces human pluripotent stem cells with improved viability
.
Stem Cells Transl. Med.
2
,
246
-
254
.
Kyriacou
,
E.
and
Heun
,
P.
(
2018
).
High-resolution mapping of centromeric protein association using APEX-chromatin fibers
.
Epigenetics Chromatin
11
,
68
.
Lagache
,
T.
,
Sauvonnet
,
N.
,
Danglot
,
L.
and
Olivo-Marin
,
J.-C.
(
2015
).
Statistical analysis of molecule colocalization in bioimaging: statistical analysis of molecule colocalization in bioimaging
.
Cytometry A
87
,
568
-
579
.
Lee
,
W. K.
and
Cho
,
M. H.
(
2016
).
Telomere-binding protein regulates the chromosome ends through the interaction with histone deacetylases in Arabidopsis thaliana
.
Nucleic Acids Res.
44
,
4610
-
4624
.
Luke
,
B.
and
Lingner
,
J.
(
2009
).
TERRA: telomeric repeat-containing RNA
.
EMBO J.
28
,
2503
-
2510
.
Majerová
,
E.
,
Fojtová
,
M.
,
Mozgová
,
I.
,
Bittová
,
M.
and
Fajkus
,
J.
(
2011
).
Hypomethylating drugs efficiently decrease cytosine methylation in telomeric DNA and activate telomerase without affecting telomere lengths in tobacco cells
.
Plant Mol. Biol.
77
,
371
-
380
.
Majerová
,
E.
,
Mandáková
,
T.
,
Vu
,
G. T. H.
,
Fajkus
,
J.
,
Lysak
,
M. A.
and
Fojtová
,
M.
(
2014
).
Chromatin features of plant telomeric sequences at terminal vs. internal positions
.
Front. Plant Sci.
5
,
593
.
Mandáková
,
T.
and
Lysak
,
M. A.
(
2008
).
Chromosomal phylogeny and karyotype evolution in x=7 crucifer species (Brassicaceae)
.
Plant Cell
20
,
2559
-
2570
.
Mayer
,
C.
,
Neubert
,
M.
and
Grummt
,
I.
(
2008
).
The structure of NoRC-associated RNA is crucial for targeting the chromatin remodelling complex NoRC to the nucleolus
.
EMBO Rep.
9
,
774
-
780
.
McStay
,
B.
and
Grummt
,
I.
(
2008
).
The epigenetics of rRNA genes: from molecular to chromosome biology
.
Annu. Rev. Cell Dev. Biol.
24
,
131
-
157
.
Michalet
,
X.
(
1997
).
Dynamic molecular combing: stretching the whole human genome for high-resolution studies
.
Science
277
,
1518
-
1523
.
Muchová
,
V.
,
Amiard
,
S.
,
Mozgová
,
I.
,
Dvořáčková
,
M.
,
Gallego
,
M. E.
,
White
,
C.
and
Fajkus
,
J.
(
2015
).
Homology-dependent repair is involved in 45S rDNA loss in plant CAF- 1 mutants
.
Plant J.
81
,
198
-
209
.
Nguyen
,
K.
,
Puppo
,
F.
,
Roche
,
S.
,
Gaillard
,
M.-C.
,
Chaix
,
C.
,
Lagarde
,
A.
,
Pierret
,
M.
,
Vovan
,
C.
,
Olschwang
,
S.
,
Salort-Campana
,
E.
et al. 
(
2017
).
Molecular combing reveals complex 4q35 rearrangements in Facioscapulohumeral dystrophy
.
Hum. Mutat.
38
,
1432
-
1441
.
Nieminuszczy
,
J.
,
Schwab
,
R. A.
and
Niedzwiedz
,
W.
(
2016
).
The DNA fibre technique – tracking helicases at work
.
Methods
108
,
92
-
98
.
Ovesný
,
M.
,
Křížek
,
P.
,
Borkovec
,
J.
,
Švindrych
,
Z.
and
Hagen
,
G. M.
(
2014
).
ThunderSTORM: a comprehensive ImageJ plug-in for PALM and STORM data analysis and super-resolution imaging
.
Bioinformatics
30
,
2389
-
2390
.
Pavlištová
,
V.
,
Dvořáčková
,
M.
,
Jež
,
M.
,
Mozgová
,
I.
,
Mokroš
,
P.
and
Fajkus
,
J.
(
2016
).
Phenotypic reversion in fas mutants of Arabidopsis thaliana by reintroduction of FAS genes: variable recovery of telomeres with major spatial rearrangements and transcriptional reprogramming of 45S rDNA genes
.
Plant J.
88
,
411
-
424
.
Politi
,
V.
,
Perini
,
G.
,
Trazzi
,
S.
,
Pliss
,
A.
,
Raska
,
I.
,
Earnshaw
,
W. C.
and
Della Valle
,
G.
(
2002
).
CENP-C binds the alpha-satellite DNA in vivo at specific centromere domains
.
J. Cell Sci.
115
,
2317
-
2327
.
Pontvianne
,
F.
,
Blevins
,
T.
,
Chandrasekhara
,
C.
,
Mozgova
,
I.
,
Hassel
,
C.
,
Pontes
,
O. M. F.
,
Tucker
,
S.
,
Mokros
,
P.
,
Muchova
,
V.
,
Fajkus
,
J.
et al. 
(
2013
).
Subnuclear partitioning of rRNA genes between the nucleolus and nucleoplasm reflects alternative epiallelic states
.
Genes Dev.
27
,
1545
-
1550
.
Procházková Schrumpfová
,
P.
,
Fojtová
,
M.
and
Fajkus
,
J.
(
2019
).
Telomeres in plants and humans: not so different, not so similar
.
Cells
8
,
58
.
Quinet
,
A.
,
Carvajal-Maldonado
,
D.
,
Lemacon
,
D.
and
Vindigni
,
A.
(
2017
).
DNA fiber analysis: mind the gap!
In
Methods in Enzymology
, Vol.
591
(ed.
B. F.
Eichman
), pp.
55
-
82
.
Elsevier
.
Ribeiro
,
S. A.
,
Vagnarelli
,
P.
,
Dong
,
Y.
,
Hori
,
T.
,
McEwen
,
B. F.
,
Fukagawa
,
T.
,
Flors
,
C.
and
Earnshaw
,
W. C.
(
2010
).
A super-resolution map of the vertebrate kinetochore
.
Proc. Natl. Acad. Sci. USA
107
,
10484
-
10489
.
Rotem
,
A.
,
Ram
,
O.
,
Shoresh
,
N.
,
Sperling
,
R. A.
,
Goren
,
A.
,
Weitz
,
D. A.
and
Bernstein
,
B. E.
(
2015
).
Single-cell ChIP-seq reveals cell subpopulations defined by chromatin state
.
Nat. Biotechnol.
33
,
1165
-
1172
.
Sabatinos
,
S. A.
and
Green
,
M. D.
(
2018
).
A chromatin fiber analysis pipeline to model DNA synthesis and structures in fission yeast
. In
Genome Instability
(ed.
M.
Muzi-Falconi
and
G. W.
Brown
), pp.
509
-
526
.
New York, NY
:
Springer New York
.
Schurra
,
C.
and
Bensimon
,
A.
(
2008
).
Combing genomic DNA for structural and functional studies
. In
The Nucleus
(ed.
R.
Hancock
), pp.
71
-
90
.
Totowa, NJ
:
Humana Press
.
Sullivan
,
B. A.
(
2010
).
Optical mapping of protein–DNA complexes on chromatin fibers
. In
Fluorescence in situ Hybridization (FISH)
(ed.
J. M.
Bridger
and
E. V.
Volpi
), pp.
99
-
115
.
Totowa, NJ
:
Humana Press
.
Sullivan
,
B. A.
and
Karpen
,
G. H.
(
2004
).
Centromeric chromatin exhibits a histone modification pattern that is distinct from both euchromatin and heterochromatin
.
Nat. Struct. Mol. Biol.
11
,
1076
-
1083
.
Sullivan
,
L. L.
,
Boivin
,
C. D.
,
Mravinac
,
B.
,
Song
,
I. Y.
and
Sullivan
,
B. A.
(
2011
).
Genomic size of CENP-A domain is proportional to total alpha satellite array size at human centromeres and expands in cancer cells
.
Chromosome Res.
19
,
457
-
470
.
Tadeus
,
G.
,
Lampe
,
A.
and
Schmoranzer
,
J.
(
2015
).
SDmixer—a versatile software tool for spectral demixing of multicolor single molecule localization data
.
Methods Appl. Fluoresc.
3
,
037001
.
The International Stem Cell Initiative
(
2007
).
Characterization of human embryonic stem cell lines by the International stem cell initiative
.
Nat. Biotechnol.
25
,
803
-
816
.
Tosolini
,
M.
,
Brochard
,
V.
,
Adenot
,
P.
,
Chebrout
,
M.
,
Grillo
,
G.
,
Navia
,
V.
,
Beaujean
,
N.
,
Francastel
,
C.
,
Bonnet-Garnier
,
A.
and
Jouneau
,
A.
(
2018
).
Contrasting epigenetic states of heterochromatin in the different types of mouse pluripotent stem cells
.
Sci. Rep.
8
,
5776
.
Vaquero-Sedas
,
M. I.
and
Vega-Palas
,
M. A.
(
2013
).
Differential association of Arabidopsis telomeres and centromeres with Histone H3 variants
.
Sci. Rep.
3
,
1202
.
Vaquero-Sedas
,
M. I.
,
Luo
,
C.
and
Vega-Palas
,
M. A.
(
2012
).
Analysis of the epigenetic status of telomeres by using ChIP-seq data
.
Nucleic Acids Res.
40
,
e163
.
Varela
,
E.
,
Schneider
,
R. P.
,
Ortega
,
S.
and
Blasco
,
M. A.
(
2011
).
Different telomere-length dynamics at the inner cell mass versus established embryonic stem (ES) cells
.
Proc. Natl. Acad. Sci. USA
108
,
15207
-
15212
.
Vera
,
E.
,
Canela
,
A.
,
Fraga
,
M. F.
,
Esteller
,
M.
and
Blasco
,
M. A.
(
2008
).
Epigenetic regulation of telomeres in human cancer
.
Oncogene
27
,
6817
-
6833
.
Wang
,
K.
,
Zhang
,
W.
,
Jiang
,
Y.
and
Zhang
,
T.
(
2013
).
Systematic application of DNA Fiber-FISH technique in cotton
.
PLoS ONE
8
,
e75674
.
Williams
,
A.
and
Flavell
,
R. A.
(
2008
).
The role of CTCF in regulating nuclear organization
.
J. Exp. Med.
205
,
747
-
750
.
Wolter
,
S.
,
Löschberger
,
A.
,
Holm
,
T.
,
Aufmkolk
,
S.
,
Dabauvalle
,
M.-C.
,
van de Linde
,
S.
and
Sauer
,
M.
(
2012
).
rapidSTORM: accurate, fast open-source software for localization microscopy
.
Nat. Methods
9
,
1040
-
1041
.
Wong
,
L. H.
,
Brettingham-Moore
,
K. H.
,
Chan
,
L.
,
Quach
,
J. M.
,
Anderson
,
M. A.
,
Northrop
,
E. L.
,
Hannan
,
R.
,
Saffery
,
R.
,
Shaw
,
M. L.
,
Williams
,
E.
et al. 
(
2007
).
Centromere RNA is a key component for the assembly of nucleoproteins at the nucleolus and centromere
.
Genome Res.
17
,
1146
-
1160
.
Wooten
,
M.
,
Snedeker
,
J.
,
Nizami
,
Z. F.
,
Yang
,
X.
,
Ranjan
,
R.
,
Urban
,
E.
,
Kim
,
J. M.
,
Gall
,
J.
,
Xiao
,
J.
and
Chen
,
X.
(
2019
).
Asymmetric histone inheritance via strand-specific incorporation and biased replication fork movement
.
Nat. Struct. Mol. Biol.
26
,
732
-
743
.
Wooten
,
M.
,
Li
,
Y.
,
Snedeker
,
J.
,
Nizami
,
Z. F.
,
Gall
,
J. G.
and
Chen
,
X.
(
2020
).
Superresolution imaging of chromatin fibers to visualize epigenetic information on replicative DNA
.
Nat. Protoc.
15
,
1188
-
1208
.
Yamada
,
N. A.
,
Rector
,
L. S.
,
Tsang
,
P.
,
Carr
,
E.
,
Scheffer
,
A.
,
Sederberg
,
M. C.
,
Aston
,
M. E.
,
Ach
,
R. A.
,
Tsalenko
,
A.
,
Sampas
,
N.
et al. 
(
2011
).
Visualization of fine-scale genomic structure by oligonucleotide-based high-resolution FISH
.
Cytogenet Genome Res.
132
,
248
-
254
.
Zhang
,
W.
,
Lee
,
H.-R.
,
Koo
,
D.-H.
and
Jiang
,
J.
(
2008
).
Epigenetic modification of centromeric chromatin: hypomethylation of DNA sequences in the CENH3-associated chromatin in Arabidopsis thaliana and Maize
.
Plant Cell
20
,
25
-
34
.
Zhou
,
Y.
,
Wang
,
Y.
,
Krause
,
K.
,
Yang
,
T.
,
Dongus
,
J. A.
,
Zhang
,
Y.
and
Turck
,
F.
(
2018
).
Telobox motifs recruit CLF/SWN–PRC2 for H3K27me3 deposition via TRB factors in Arabidopsis
.
Nat. Genet.
50
,
638
-
644
.
Zillner
,
K.
,
Komatsu
,
J.
,
Filarsky
,
K.
,
Kalepu
,
R.
,
Bensimon
,
A.
and
Németh
,
A.
(
2015
).
Active human nucleolar organizer regions are interspersed with inactive rDNA repeats in normal and tumor cells
.
Epigenomics
7
,
363
-
378
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information