Spatial Preservation of Nuclear Chromatin Architecture during Three-Dimensional Fluorescence in Situ Hybridization (3D-FISH)

Experimental Cell Research 276, 10 –23 (2002) doi:10.1006/excr.2002.5513, available online at http://www.idealibrary.com on

Spatial Preservation of Nuclear Chromatin Architecture during Three-Dimensional Fluorescence in Situ Hybridization (3D-FISH) Irina Solovei,* ,1 Antonio Cavallo,† Lothar Schermelleh,* Franc¸oise Jaunin,‡ Catia Scasselati,‡ Dusan Cmarko,‡ Christoph Cremer,† Stanislav Fakan,‡ and Thomas Cremer* *Department of Biology II, Ludwig-Maximilians University of Munich, Germany; †Kirchhoff Institute for Applied Physics, University of Heidelberg, Heidelberg, Germany; and ‡Centre of Electron Microscopy, University of Lausanne, Lausanne, Switzerland

jor tool for studies of the higher order chromatin architecture in the interphase nucleus [1– 6] including the topology of individual genes with regard to their chromosome territory (CT) [7–9]. Replication labeling of DNA during S phase has become another highly useful approach for the investigation of chromatin arrangements in the interphase nucleus. Halogenated thymidine analogues incorporated into replicating chromatin (BrdU, CldU, IdU) can be immunodetected after fixation and chromatin denaturation [10, 11]. More recently replication labeling using nucleotides directly conjugated to fluorochromes (such as or FITC-, Cy3-, or Cy5-dUTP) was employed. The latter method has allowed for the first time the visualization of replication-labeled chromatin in living cells [12–15]. Both approaches have provided evidence that during S phase replicating chromatin forms replication foci with a DNA content in the order of 1 Mb. Notably, these ⬃1-Mb chromatin domains are apparently maintained at all stages of the cell cycle and presumably from one cell cycle to another (for review see [1]). Furthermore, a procedure was recently described that allows even the visualization of single genes in living cells [16]. In vivo approaches, however, have been carried out only on cultured cells and do not yet allow studies of cells in tissue sections. Therefore, 3DFISH will continue to be an indispensable tool in pinpointing similarities and differences of the higher order chromatin architecture of different cell types in tissues and organs of a given species. It is also a promising tool for evolutionary comparisons of nuclear architecture in the same cell types of different species. Such comparisons are essential in defining evolutionary conserved motifs of higher order chromatin architecture [4]. 3D-FISH requires harsh pretreatments for the optimal accessibility of probes to nuclear target DNA. This raises the question to what extent the spatial relations of CTs and chromatin domains detected in 3D-FISH experiments reflect the 3D higher order chromatin topology of living cells. Despite a few studies that have approached this problem to a limited extent [17–20],

3D-FISH has become a major tool for studying the higher order chromatin organization in the cell nucleus. It is not clear, however, to what extent chromatin arrangement in the nucleus after fixation and 3D-FISH still reflects the order in living cells. To study this question, we compared higher order chromatin arrangements in living cells with those found after the 3D-FISH procedure. For in vivo studies we employed replication labeling of DNA with Cy3-conjugated nucleotides and/or chromatin labeling by GFP-tagged histone 2B. At the light microscope level, we compared the intranuclear distribution of H2B-GFP-tagged chromatin and the positions of replication-labeled chromatin domains in the same individual cells in vivo, after fixation with 4% paraformaldehyde, and after 3D-FISH. Light microscope data demonstrate a high degree of preservation of the spatial arrangement of ⬃1-Mb chromatin domains. Subsequent electron microscope investigations of chromatin structure showed strong alterations in the ultrastructure of the nucleus caused mainly by the heat denaturation step. Through this step chromatin acquires the appearance of a net with mesh size of 50 –200 nm roughly corresponding to the average displacement of the chromatin domains observed at light microscope level. We conclude that 3D-FISH is a useful tool to study chromosome territory structure and arrangements down to the level of ⬃1-Mb chromatin domain positions. However, important ultrastructural details of the chromatin architecture are destroyed by the heat denaturation step, thus putting a limit to the usefulness of 3D-FISH analyses at nanometer scales. © 2002 Elsevier Science (USA)

Key Words: 3D-FISH; replication labeling in vivo; H2B-GFP; nucleus ultrastructure.

INTRODUCTION

Fluorescence in situ hybridization on three-dimensionally preserved nuclei (3D-FISH) has become a ma1 To whom correspondence and reprint requests should be addressed. Fax: (⫹49) 89 21806719. E-mail: [email protected].

0014-4827/02 $35.00 © 2002 Elsevier Science (USA) All rights reserved.

10

USE OF 3D-FISH TO STUDY CHROMATIN ARRANGEMENT

the question of artifacts has not been answered convincingly. In this study we employed replication labeling of cultured human neuroblastoma cells and HeLa cells with Cy3-conjugated nucleotides. In some experiments we used HeLa cells expressing green fluorescent protein-tagged histone 2B (H2B-GFP) resulting in green fluorescence of chromatin. Light and electron microscope (LM and EM, respectively) approaches were applied to investigate nuclear morphology following fixation of cells and 3D-FISH. At the LM resolution level, we compared the intranuclear distribution of chromatin visualized by H2B-GFP and the positions of replication-labeled chromatin domains in the same individual cells at three different stages: in living cells, after fixation, and after 3D-FISH. These data show a high degree of preservation of the spatial arrangement of labeled ⬃1-Mb chromatin domains. Distance measurements between individual replication foci in the living cell and after 3D-FISH demonstrated that average displacement of the chromatin domains due to both fixation and 3D-FISH was about 300 nm. In addition, transmission electron microscopy and DNA immunolabeling were applied to investigate the chromatin structure following fixation of cells and 3D-FISH. These studies showed strong alterations in the ultrastructure of the nucleus caused mainly by the heat denaturation step routinely employed in 3D-FISH experiments. Through this step chromatin acquires the appearance of a net with mesh size of ca. 50 –200 nm which roughly corresponds to the shift value observed at LM level. Furthermore, obvious DNA redistribution takes place as indicated by a reduction of condensed chromatin areas and the leakage of some nuclear DNA into the cytoplasm. MATERIALS AND METHODS Cells. The human neuroblastoma cell line SH-EP N14 was kindly donated by Prof. W. W. Franke (DKFZ, Heidelberg), and HeLa cells expressing H2B-GFP by Dr. Kevin F. Sullivan (The Scripps Research Institue, La Jolla). The two cell lines were cultured in RPMI supplemented with 10% fetal calf serum and gentamycin. Human primary skin fibroblasts were grown in DMEM supplemented with 10% fetal calf serum and gentamycin. All cells were maintained at 37°C in an atmosphere of 5% CO 2. For standard 3D-FISH experiments cells were subcultured on 26 ⫻ 76-mm coverslips (thickness 0.17 ⫾ 0.01 mm, Assistent, Germany) placed in Quadriperms (In Vitro Systems). For in vivo observations cells were grown on round coverslips (40 mm in diameter) fitting to a FCS2 living cell chamber (Biotechs). To enable the individual relocation of cells, a grid with a mesh of approximately 100 ␮m was manually scratched in the center of each coverslip using a tungsten marker. For the continued study of individual living cells and also after fixation and FISH, the central area of round coverslips containing these cells was cut out with the help of a diamond pen to fit a microscopic slide. For experiments that involved observations of the same nuclei with confocal microscopy and with electron microscopy, cells were grown on round gridded Cellocate coverslips (Eppendorf ).

11

Replication labeling. SH-EP N14 cells were grown to 50% confluency on marked round coverslips as described above. Nuclei of about 30 cells with location near the grid lines were microinjected with Cy3-AP3-dUTP (Amersham–Pharmacia) in PBS (50 mM) using a Zeiss AIS2 (automated injection system) in combination with a 5246 transjector (Eppendorf ). After microinjection, cells were grown for 42 h (some two cell cycles) to obtain nuclei with segregated CTs [15]. HeLa cells were labeled by scratch labeling [14] and allowed to grow for 24 h for interphase chromosome segregation. 3D-FISH. Methods of cell fixation, pretreatment, and FISH to three-dimensionally preserved nuclei employed in this study have been described previously [21] and are similar to those described in [22]. Coverslips with growing cells were briefly washed with PBS, fixed in freshly made 4% paraformaldehyde in PBS for 10 min, washed in PBS (3 ⫻ 5 min), incubated in 0.5% Triton X-100 in PBS for 10 min, and equilibrated in 20% glycerol in PBS for 30 min. Cells were then frozen in liquid nitrogen and thawed at room temperature; this step was repeated four times. After washing cells again in PBS, they were incubated for 5 min in 0.1 N HCl, briefly rinsed in 2⫻ SSC, and stored in 50% formamid/2⫻ SSC for 1 h at minimum but mostly for several days. Labeled probes were dissolved in a hybridization mixture (50% formamide, 10% dextransulfate, 1⫻ SSC), applied to cells, and sealed under coverslips with rubber cement. DNA–probe and cellular DNA were denatured simultaneously on a hot block at 75°C for 2 min. Hybridization was carried out in a humid atmosphere at 37°C for 2 days. Whole chromosome paint probes for human chromosomes 1, 2, 3, 4, 5, 15, 19, and X were a kind gift from Dr. Ferguson-Smith (University of Cambridge). These probes were depleted from repetitive sequences using human Cot1 [23, 24] and labeled by DOP-PCR with biotin– dUTP or digoxigenin– dUTP. Posthybridization washings included 2⫻ SSC (1⫻ SSC is 0.15 M NaCl and 0.015 M Na citrate) at 37°C and 0.1⫻ SSC at 60°C. Standard procedures were employed for probe detection [21]. Briefly, cells were preincubated in 4⫻ SSC with 4% BSA and then incubated in the same solution with appropriate antibodies or avidin conjugated to different fluorochromes. Nuclei were counterstained with DAPI or propidium iodide and mounted in Vectashield antifade. In some experiments a mock-hybridization procedure (mock 3D-FISH) was performed including the same steps as a full 3D-FISH except that no probe DNA was added to the hybridization mixture and probe detection was omitted. Sequential observations of the same cells in living state, after paraformaldehyde fixation, and after hybridization. For in vivo observations coverslips with replication-labeled SH-EP N14 or HeLa cells were mounted in a FCS2 living cell chamber (Biotechs) with 5% CO 2 adapted RPMI medium containing Hepes buffer and maintained at 37°C by heating the chamber as well as the objective lens of the microscope. Light optical serial sections of living cells with Cy3labeled replication chromatin domains and/or H2B-GFP were collected. Thereafter cells were fixed immediately. The coverslip with cells was removed from the living cell chamber, briefly rinsed in prewarmed PBS, fixed in freshly made 4% paraformaldehyde for 10 min, and incubated in 0.5% Triton X-100 in PBS for 10 min. Since living cells and their nuclei move relatively quickly, only one nucleus was scanned from each coverslip so that the time between scanning of a living cell and the beginning of fixation did not exceed 1 min. The area of the coverslip with microinjected cells was cut out, mounted with Vectashield on a microscope slide and optical confocal sections were collected again. In one experiment DNA counterstaining with TO-PRO-3 (Molecular Probe) was employed prior to mounting in antifade. Then the coverslips with cells were unmounted, antifade medium was washed away in PBS, and cells were subjected to 3D-FISH or mock 3D-FISH. The third series of optical confocal sections through the same nuclei was then collected to map Cy3labeled replication chromatin domains, H2B-GFP, and hybridized CTs.

12

SOLOVEI ET AL.

Laser scanning microscopy. Series of equidistant 250-nm 8-bit grayscale images, 512 ⫻ 512 pixels in size, from nuclei (in living cells, after fixation, and with painted CTs) were collected using laser scanning microscopy (Leica TCS SP confocal microscope or ZEISS LSM410). Galleries of confocal images were assembled using NIH program; RGB images were generated in Adobe Photoshop 5.5. Computer evaluations. To estimate the degree of geometrical displacements of labeled chromatin domains, an interactive computer program was developed. The program allows to identify labeled replication foci in different confocal image stacks and to perform distance measurements between chosen foci. The analysis of each data stack was carried out in the following way. First, the frames, stored in separate TIFF files, were put together and converted into KDF file format (as used in Cantata environment, a commercial environment for multidimensional image processing). The Khoros program (in the Cantata environment) was used to process the data in a way suitable for feeding of our self-written routines for image analysis. These evaluation routines were able to correct for both shifts and rotations of the optical sections of a 3D data stack, produced by mechanical drifts of the object stage of the Zeiss CLSM used for registration. The procedures are based on a special cross-correlation algorithm developed by Dr. R. Heintzmann in C. Cremer’s laboratory. At the end of the “preprocessing” phase we obtained three 3D files for each experiment, containing the 3D state of the cell analyzed in the living, fixed, and hybridized status. The next step was done using a self-made program to apply a global thresholding algorithm (developed by Dr. R. Heintzmann) with user-driven interface to select “spots” (i.e., chromatin domains). The global threshoding was done in such a way that all the voxels above a given value were counted to belong to a spot. The program allows the user to eliminate interactively any spot which was considered by the investigator to be a mistake of the global thresholding algorithm, and to process only the “true” spots. Note that the program was written in such a way that it easily allows also the use of a different algorithm for segmentation. The user can choose the threshold level interactively, the segmentation result can be visualized on the screen, and the segmented spots are shown as small ellipsoids in 3D. For better localization, a spot position was determined as the barycenter (gravity center of fluorescence intensity) computed from the algorithm. For visualization of the results, the program assigns three different colors for the segmented chromatin domains, corresponding to their positions, shapes, and recorded volumes, in living, fixed, and hybridized cells (Fig. 3F). By clicking the mouse over chosen chromatin domains in each of the three colored images, the barycenter of an identified individual site was automatically determined, and its coordinates were added to a table supported by the program (using a simple ascii file for further analysis). Measurements of distances were done between individual chromatin domains clearly identified to be the same in all three stacks (living, fixed, hybridized). The results were scaled for real distances in nanometers, using the known magnification factors of the system. For this purpose, a set of programs was developed in Python (a scripting language) with numerical extensions. Preparation procedure for electron microscopy. Primary human fibroblasts after standard fixation with 4% paraformaldehyde in PBS and after different steps of 3D-FISH were prepared for electron microscopy. Coverslips with cells were washed overnight in PBS and then either postfixed with 2% glutaraldehyde or with 2% paraformaldehyde/0.2% glutaraldehyde in 0.1 M So¨ rensen phosphate buffer (pH 7.4) for 30 min at 4°C. Some samples were dehydrated directly, without postfixation. In all cases dehydration was carried out in increasing concentrations of ethanol and the cells were then infiltrated with LR White resin. Coverslips were finally attached to LR White resin-filled capsules and allowed to polymerize for 48 h at 60°C. The embedded cells were separated from the coverslip by a short treatment with liquid nitrogen and cut parallel to the coverslip plane using a Leica Ultracut UCT ultramicrotome. In order to iden-

tify the same interphase cells previously analyzed by confocal microscopy after 3D-FISH for a subsequent EM study, the cells were relocated using the alphanumeric imprints of Cellocate coverslips on the surface of the resin blocks. Ultrathin sections on Formvar– carbon-coated copper or nickel grids were stained with uranyl acetate and lead citrate, and examined in a Philips CM10 or CM12 electron microscope, at 80 kV, using a 30- to 40-␮m objective aperture. Immunogold labeling. Ultrathin sections on Formvar– carboncoated nickel grids were incubated on a drop of normal goat serum (NGS; Nordic Immunology Laboratories, Tilburg, The Netherlands) diluted 1:100 in PBS for 3 min. Anti-DNA antibody (Progen, Heidelberg, Germany; clone AC-30-10 specific for both single- and doublestranded DNA) was diluted 1:10 in PBS containing 0.1% bovine serum albumin and 0.05% Tween 20; the incubation was carried out at 4°C for 17 h. For control, sections were incubated in the same solution without the primary antibody. After a rinse with PBS/ Tween 20 and PBS, grids were incubated in NGS as above. The secondary antibody coupled with colloidal gold particles was diluted 1:10 (GAM IgG, IgM; 12 nm; Jackson, Immunoresearch) in PBS and incubation was carried out for 20 min at 20°C. After washing with PBS and distilled water, grids were air-dried and stained with uranyl acetate and lead citrate before examination.

RESULTS

3D-FISH requires a series of fixation, pretreatment, hybridization, and detection steps, which are listed in Table 1. We wished to elucidate whether some of these steps cause artifacts that seriously affect conclusions on 3D chromatin architecture and arrangement in the cell nucleus. In addition to light microscope observations and quantitative measurements, electron microscopy was performed at appropriate stages of the procedure. Light Microscopic Observations The experiments, listed in Table 2, included observations of the same individual cells at three successive steps: in vivo, after paraformaldehyde fixation, and after 3D-FISH or mock 3D-FISH (Fig. 1). SH-EP N14 cells and HeLa cells with segregated Cy3-labeled CTs were chosen so that the arrangement of individual Cy3-labeled chromatin domains in these Cy3-labeled CTs could be compared at each successive step. 1. Spatial distribution of chromatin domains identified by replication labeling. In the first experiment neuroblastoma SH-EP N14 cells were replication-labeled using Cy3-dUTP. Cells were allowed to proliferate for two or three cell cycles in order to obtain nuclei with segregated Cy3-labeled CTs. Only cells with a sufficiently low number of labeled chromatin foci (mid– late replication pattern) were chosen for scanning to facilitate the identification of individual foci (Figs. 1B, 1D, and 1F). We traced the 3D positions of labeled individually identified chromatin domains (Fig. 1) before and after mock 3D-FISH (Table 2, Experiment 1) or after true 3D-FISH with whole chromosome paint probes (Table 2, Experiment 2). The results of both

13

USE OF 3D-FISH TO STUDY CHROMATIN ARRANGEMENT

TABLE 1 Key Steps of 3D-FISH and Survey of Confocal Laser Scanning Microscopy (CLSM) and Electron Microscopy (EM) Studies Steps 1. 2. 3. 4. 5. 6. 7. 8. 9.

Cells grow on coverslips Fixation Permeabilization: membrane destruction Permeabilization: cytoplasm breakage Removal of proteins DNA denaturation Hybridization Removal of DNA probes not hybridized to target DNA Visualization of hybrid DNA

Treatment

CLSM

EM

Transfer in living cell chamber 4% paraformaldehyde in PBS Incubation in 0.5% Triton X-100 in PBS Incubation in 20% glycerol in PBS and freezing/thawing Incubation in 0.1 N HCl Heating at 75°C in 50% formamide Incubation at 37°C over 1–3 days Washings in 0.1⫻ SSC at 60°C Detection of haptens (biotin, digoxigenin), mounting in antifade, microscopy

⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹

⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺ ⫺ ⫹

Note. ⫹, cells were used for LM or EM observations; ⫺, observations were not performed.

experiments were identical and we describe them together. To assess the effect of 3D-FISH on the shape and volume of the nucleus, we counted the number of optical sections necessary to record a given nucleus in vivo, after fixation with 4% paraformaldehyde, and after 3D-FISH. In eight cells studied in this way, we noted that the nuclei swelled slightly (approximately 7%) as a result of fixation, but shrank again during the 3DFISH procedure, so that the height of the nuclei finally decreased, on average, to about 80% of that in the living cell. The X,Y-diameters of the nuclei measured in mid plane optical sections changed similarly. A more detailed account of the changes of the size of the nuclei was obtained using computer evaluations (see below). A visual comparison of matching optical sections through individual nuclei recorded in living, fixed, and hybridized states suggested that individual chromatin domains after fixation and 3D-FISH retained apparently the same size and shape as in living cells (compare Fig. 1B with 1D and 1F; Fig. 1L). Small-scale swelling and shrinking of the nucleus during fixation (Fig. 1G) and 3D-FISH (Fig. 1H) described above did not notably affect the relative positions of labeled chromatin domains. The topology of individual chromatin domains within chromatin domain clusters were well preserved after fixation and 3D-FISH, and could be

easily aligned with the topology observed in the living cell nucleus, while distant clusters of chromatin domains sometimes showed more pronounced shifts in relation to each other (Figs. 1H and 1L). Occasionally, we noted that some clusters of chromatin domains were shifted from a given confocal section to a neighboring one (compare Figs. 1D and 1F). Such shifts likely represent treatment artifacts, although we cannot exclude in vivo movements in the short period (about 1 min) between CLSM scanning of the living cell and fixation. Nevertheless, within such shifted clusters the small-scale topology of labeled chromatin domains remained nearly identical (compare Fig. 1D and insert in Fig. 1F). To determine the shifts of labeled chromatin domains in a quantitative way, a computer program was written that aligns corresponding optical sections through a nucleus in vivo, after fixation, and after FISH. The program determined the 3D coordinates of the intensity gravity centers of individual chromatin domains, visually identified in all three image stacks, and calculated the distances between each pair of chromatin domains in vivo (L), after fixation (F), and after hybridization (H) (Fig. 2A). The resulting differences between the distances obtained for the in vivo state and after fixation (⌬LF), between fixation and after 3D-FISH (⌬FH), and between the in vivo state and

TABLE 2 Experiments for Light Optical Studies Cells SH-EP N14 HeLa

Experiment number 1 2 3 4

In vivo labeling

Treatments

Cy3-replication labeling Cy3-replication labeling H2B-GFP H2B-GFP ⫹ Cy3-replication labeling

Mock-FISH 3D FISH Mock-FISH Mock-FISH, whole DNA staining

Fluorochromes scanned 1: 2: 1: 3:

Cy3 Cy3 ⫹ FITC or Cy5 GFP GFP ⫹ Cy3 ⫹ TO-PRO-3

Number of observed cells 6 4 2 3

14

SOLOVEI ET AL.

FIG. 1. Neuroblastoma cells were replication labeled with Cy3-dUTP and cultivated for two to three additional cell cycles to allow segregation of labeled and unlabeled chromatids. Confocal serial sections of nuclei were taken from live cells (A, B), after fixation with 4% paraformaldehyde (C, D), and after 3D-FISH (E, F). (A–H) and (I–L) Results from two typical nuclei. (A) Transmission light view of the cell nucleus of a living cell, (C) the same nucleus after fixation, and (E) after mock 3D-FISH. (B, D, F) Corresponding optical sections from this nucleus show Cy3-labeled ⬃1-Mb chromatin domains. Note the preservation and similar distribution of labeled chromatin domains seen in the live cell nucleus (B), after fixation (D), and 3D-FISH (F). One cluster of chromosome domains (B, D, arrowhead) was no longer observed after 3D-FISH in the corresponding (best match) section (F), but its characteristic pattern was easily identified in the neighboring section

USE OF 3D-FISH TO STUDY CHROMATIN ARRANGEMENT

15

FIG. 2. 3D distances between clearly identified pairs of Cy3-labeled chromatin domains measured in nuclei of live neuroblastoma cells, after fixation, and after 3D-FISH. (A) Schematic example of ⌬LF, ⌬FH, and ⌬LH determination; L, distance in the live cell nucleus; F, after fixation; and H, after 3D-FISH. (B–D) Scatter diagrams of ⌬LF, ⌬FH, and ⌬LH measurements (in nm) obtained in one nucleus shown in Fig. 1I–1L. Compared to the distances in live cells, distances between chromatin domains were slightly reduced after fixation (B) and slightly increased after hybridization (C). Accordingly, distances between two chromatin domains observed after 3D-FISH corresponded very well to those in living cells (D). Lines across the scatter diagrams show linear regression, ⌬ ⫽ kL.

(F, insert). (G) Overlay of corresponding sections from the living (green) and fixed (red) cell nucleus demonstrates that fixation caused a slight shift of chromatin domains. (H) Overlay of corresponding optical sections from fixed nucleus (green) and nucleus after hybridisation (red) shows a more pronounced shift of labeled chromatin domains although their size, shape, and topography were still well preserved. Note that in any nuclear region it is possible to get an almost perfect overlap of a local cluster of chromatin domains (yellow); however, this might lead to more pronounced shifts of more distant clusters. (I) Transmission light view of the nucleus of another living neuroblastoma cell and overlay of a mid optical section shows the distribution of Cy3-labeled chromatin domains (red) within this nucleus. (K) Maximum intensity projection from the stack of confocal serial sections through this nucleus after 3D-FISH demonstrates two painted chromosome 15 territories (green) together with Cy3-labeled chromatin domains (red). (L) Overlay of best matching optical nuclear sections from the live cell (green chromatin domains), after fixation (red domains), and after 3D-FISH (blue domains) emphasizes the preservation of chromatin domains with regard to their size, shape, and topography. Bars, 5 ␮m.

16

SOLOVEI ET AL.

TABLE 3 Mean Shift Values of Chromatin Foci Resulting from Fixation and 3D-FISH Absolute shift

Nucleus 1

Nucleus 2

Nucleus 3

Nucleus 4

Average and standard deviation for four nuclei

兩⌬LF兩 兩⌬FH兩 兩⌬LH兩

488 ⫾ 273 221 ⫾ 201 368 ⫾ 274

308 ⫾ 214 160 ⫾ 129 336 ⫾ 273

428 ⫾ 321 382 ⫾ 270 258 ⫾ 179

296 ⫾ 235 294 ⫾ 178 235 ⫾ 181

383 ⫾ 92.9 265 ⫾ 95.7 303 ⫾ 62.4

Note. For details see Fig.2 and text.

after FISH (⌬LH) were plotted against L (Figs. 2B– 2D). Although the relative contributions of shrinkage and swelling varied somewhat in individual nuclei, the plots confirmed that nuclei swelled slightly after fixation and then shrank slightly after 3D-FISH. For example, in the nucleus shown in Figs. 1I–1L, the shrinkage and swelling effects compensated each other (Fig. 2D). The absolute shift values ⌬LF, ⌬FH, ⌬LH for 313 chromatin domains determined in four nuclei were on the order of 300 nm (Table 3). 2. Spatial distribution of GFP-tagged histone 2B. The use of HeLa cells stably expressing H2B-GFP (Table 2, Experiment 3) allowed study of the effect of

3D-FISH on histone distribution. Comparison of matching optical sections through the nuclei of living and fixed cells showed that fixation and permeabilization steps did not cause major changes in the GFP pattern (Fig. 3). After HCl treatment and heat denaturation (either alone or combined) the fluorescence intensity decreased greatly. While the “borders” between chromatin and lacunas of the interchromatin compartment (which contain speckles and nuclear bodies) appeared rather distinct in living cells and after fixation, they became more blurred after these treatments (compare Figs. 3B and 3F). This effect in particular was noted for condensed chromatin around the

FIG. 3. Transmission light views (A, C, E) and corresponding mid confocal sections of H2B-GFP tagged chromatin (B, D, F) of the nucleus in a living HeLa cell (A, B), after fixation with 4% paraformaldehyde in PBS (C, D), and after mock-3D FISH (E, F). At LM resolution the higher order chromatin structure is well preserved after fixation in spite of slight nuclear swelling. After additional pretreatment steps required for 3D-FISH, the structure of the chromatin is markedly less distinct. Bar, 5 ␮m.

USE OF 3D-FISH TO STUDY CHROMATIN ARRANGEMENT

17

FIG. 4. Corresponding optical sections (best matches from confocal serial image stacks) of a HeLa cell nucleus after fixation (A, C, E) and after mock 3D-FISH (B, D, F). Pattern of Cy3-labeled chromatin domains (A, B) and DNA component of the chromatin (C, D) in general are well preserved after mock 3D-FISH with regard to domain size, shape, and topography, while pattern of GFP-H2B component of the chromatin is less obvious or even lost (E, F). Bar, 5 ␮m.

nucleoli and chromatin at the nuclear periphery. We also traced the distribution of H2B-GFP in several cells fixed 2 days after replication labeling of DNA with Cy3-dUTP (Table 2, Experiment 4). After fixation and mock 3D-FISH, the DNA in these cells was visualized by staining with TO-PRO-3. Again we observed a more blurred distribution of the GFP signal (compare Fig. 4C with 4D). However, neither a pronounced shift of Cy3-labeled chromatin domains (compare Fig. 4A with 4B) nor any obvious redistribution of TO-PRO-3 stained DNA (compare Fig. 4E with 4F) was noted at the light microscope level. Electron Microscopic Observations After each of the key steps of 3D-FISH, human primary fibroblasts were fixed and prepared for electron microscope analysis (Table 1). Cells fixed with paraformaldehyde and postfixed with glutaraldehyde demonstrated a good morphological preservation of the nuclei (Fig. 5A). The nuclei contained small compact aggregates of condensed chromatin throughout the nucleoplasm and larger accumulations of condensed chroma-

tin adjacent to the nuclear envelope. Perichromatin granules [25] located near condensed chromatin areas and interchromatin granules in the interchromatin space were well conserved. The nucleoli exhibited wellrecognizable perinucleolar chromatin, the dense fibrillar component, the granular component, and small fibrillar centers. Permeabilization treatments (incubations in Triton X-100, incubation in glycerol, freezing/ thawing, and brief incubation in HCl—see Table 1) did not greatly modify the fine structure of nuclei (Fig. 5B). Condensed chromatin regions still remained well recognizable, although the overall nuclear structure already became somewhat reticulate. However, drastic changes took place after denaturation of chromatin by heating in 50% formamid/10% dextran sulfate at 75°C for 2 min. A thin filamentous network now appeared nearly throughout the entire nuclear space (Fig. 5C). The size of meshes in this network varied between 50 and 200 nm. Surprisingly, aggregates of condensed chromatin were no longer seen in the proximity of the nuclear envelope and inside the nucleus. A chromatinfree, alveolar zone was observed along the nuclear en-

18

SOLOVEI ET AL.

FIG. 5. Ultrathin sections of human primary skin fibroblasts at different stages of 3D-FISH. (A) Well-preserved ultrastructure after paraformaldehyde/glutaraldehyde fixation. (B) Reasonably preserved ultrastructure after permeabilization and HCl treatment, condensed chromatin, and interchromatin granules are still recognizable. (C) Heat denaturation in formamide results in drastical changes of nuclear ultrastructure, appearance of a homogeneous filamentous network. (D) After completion of 3D-FISH. Sections were stained with uranyle acetate and lead citrate. c, condensed chromatin; ig, interchromatin granules; nu, nucleolus; arrows indicate filamentous network; arrowheads indicate nuclear periphery. Bars, 0.5 ␮m.

USE OF 3D-FISH TO STUDY CHROMATIN ARRANGEMENT

19

FIG. 6. Comparison of chromatin morphology in a fibroblast nucleus after 3D-FISH at light (A–E) and electron microscopic (F, G) levels. Maximum intensity projection of three channels confocal stack (A) and a single optical section from the upper two-thirds of the nucleus (B) with DNA stained with propidium iodide (blue), painted chromosome 2 territories (green) and chromosome 3 territories (red). (C–E) Separate recordings of the three channels: channel with propidium iodide fluorescence (C) exhibits brightly stained nucleoli (n) and unstained lacunas of the interchromatin compartment; painted chromosome 2 territories (D) and chromosome 3 territories (E) reveal chromatin density variations. (F, G) Electron microscopic section through the same nucleus at low and higher magnification. Note homogeneous filamentous network structure of the chromatin. Bars (A–F), 5 ␮m; (G), 0.5 ␮m.

velope. While interchromatin granules were still recognizable (not shown), nucleoli had a rather homogenous appearance (Fig. 5C). Small perforations were observed in the nuclear envelope, some of them probably corresponding to nuclear pores. These can also result from the freezing–thawing steps performed in order to increase the accessibility of the nuclear interior to chromosome paint probes. Further steps of 3DFISH (Table 1, steps 7–9) did not induce additional major changes in the structure of the nucleus, but resulted in an even more pronounced filamentous network (Fig. 5D).

We also studied the chromatin ultrastructure in nuclei of primary human fibroblasts, in which CTs 2 and 3 were first painted by two-color 3D-FISH and recorded by laser scanning microscopy (Fig. 6). Light optical serial sections were collected from four nuclei and the positions of the cells on gridded coverslips were recorded. The identified cells were embedded into resin and sectioned. Fibroblast nuclei were rather flat (only about 3 ␮m in height) and portions of the painted CTs were present in almost all optical sections. Figure 6A shows the maximum intensity projection of a nucleus showing CTs 2 and 3 in contact with the nuclear pe-

20

SOLOVEI ET AL.

FIG. 7. Ultrathin sections of nucleus from human primary skin fibroblasts after fixation (A) and after mock 3D-FISH (B). Sections were immunoreacted with mouse anti-DNA (12-nm gold grains) and stained with uranyl acetate and lead citrate. cy, cytoplasm; ig, interchromatin granules; large arrows, some DNA labeling; arrowheads, nuclear periphery; small arrow, perichromatin granule. Bars, 0.5 ␮m. After fixation, DNA labeling was observed over condensed chromatin and in the perichromatin region, whereas after subsequent 3D-FISH, labeling was found over the filamentous network and in the cytoplasm.

riphery. In this nucleus one CT 2 (green) was found in immediate neighborhood to both CTs 3 (red). Figure 6F shows an EM section at low magnification approximately corresponding to the single confocal section shown in Figs. 6B– 6E. At higher magnification this EM section clearly demonstrates the altered fine morphology characteristic for the nuclei after 3D-FISH (Fig. 6G, see also the previous paragraph). It is noteworthy that LM observations indicated focal aggregations of DNA within the painted CTs (Figs. 6C– 6E), while such aggregations were not apparent at the EM level (Fig. 6F). In particular, condensed chromatin is missing at the nuclear periphery, where “empty alveoli” strongly prevail over electron-dense material (Fig. 6G). DNA was also detected by EM immunolabeling on sections through cells after mock 3D-FISH. Labeling in the nucleus occurred within the filamentous network (Fig. 7B). In addition to nuclear sites, EM immunolabeling revealed DNA in the cytoplasm around the nuclei (Fig. 7B, arrow). This finding is in clear contrast to DNA immunostaining in nuclei after fixation with paraformaldehyde: in these cells labeled DNA was restricted entirely to the nucleus (Fig. 7A).

DISCUSSION

Light microscope observations of individual ⬃1-Mb chromatin domains in the same cells in vivo, after fixation, and after 3D-FISH demonstrated that number, size, and spatial distribution of ⬃1-Mb chromatin domains were generally well preserved throughout the 3D-FISH procedure. This finding confirms other studies which indicate that both the distribution of DNA in general and the localization of specific nuclear structures, e.g., PML bodies and centromere regions of interphase chromosomes, do not change profoundly at the LM level after 3D-FISH [17, 18, 20]. In contrast, electron microscopy showed that 3D-FISH caused drastic changes in the fine structure of the nucleus. Most importantly, nucleoplasm acquired the appearance of a thin filamentous network with a mesh size of 50 –200 nm. While aggregates of condensed chromatin were still visible in the EM after formaldehyde fixation, they disappeared almost fully after heat denaturation of chromatin. Mongelard et al. [19] showed that the volume of a hybridization signal increased when the heat denaturation time was prolonged. In agreement with our observations these authors concluded

USE OF 3D-FISH TO STUDY CHROMATIN ARRANGEMENT

that denaturation is the most damaging step with regard to nuclear morphology. Slight shifts in the positioning of individual domains were identified by light microscopy in experiments, where the positions of domains were compared in nuclei of living cells and after fixation and 3D-FISH. The average shift was about 300 nm and was more pronounced for chromatin domains located far apart from each other, i.e., in different CTs, than for domains that were close neighbors, i.e., part of an individual CT. The chromatin domain shifts observed at the LM level may reflect the formation of the filamentous network seen at the EM level. Larger shift values, detected for more remote chromatin domains, may be caused by some deformations in nuclear shapes that induce shifts of whole CTs in relation to each other. We do not know to what extent the damage observed in nuclei after identical pretreatment steps may differ from cell type to cell type. Accordingly, protocols need to be adapted to the special requirements of the studied cell type in order to obtain an optimal signal-to-background ratio together with a reasonable preservation of the 3D structure [21]. In case of overtreatment with HCI and/or overdenaturation by heating in 50% formamid one can note protuberances of DNA from the nuclear periphery into the cytoplasm, or even a halo of DNA around the nuclear boundary at the LM level using DNA-specific dyes, e.g., DAPI or TO-PRO-3 (data not shown). Such preparations should not be used for studies of higher order chromatin arrangements, although the position of entire CT may still not be largely affected. Immunoelectron microscopic assays showed some leakage of DNA from the nucleus into the cytoplasm. It is obvious that detergent and formamide treatments give rise to solubilization of lipid constituents as well as to partial extraction of nuclear envelope proteins. This may result in the enlargement of nuclear pore remnants, even the formation of holes in the nuclear envelope, thus opening ways for DNA leakage. A nonhomogeneous pattern of chromatin and DNA was noted in LM studies of (i) chromatin with H2BGFP in living and fixed cells, (ii) fixed nuclei counterstained with various DNA specific dyes, and (iii) painted CTs [1, 3, present article]. In formaldehydefixed and DNA-stained nuclei the regions of more condensed chromatin seen at the LM level corresponded to electron-dense material seen at the EM level in cells prior to denaturing steps. In 3D-preserved nuclei subjected to chromosome painting, this heterogeneous density was still observed at the LM level but was no longer apparent at the EM level. When comparing data obtained at the LM and EM levels, one should keep in mind the differences in resolution of LM and EM, and in the thickness of sections used for EM (80 nm) and LM studies (nominal thickness of 200 nm with the axial resolution of the confocal microscope about 800

21

nm). Redistribution and leakage of DNA, taking place during the 3D-FISH procedure, becomes more obvious at the EM level. The precise proportions of proteins, DNA, and RNA that are lost or redistributed during the course of a 3D-FISH experiment remain unknown. In the only semiquantitative study concerning DNA preservation during FISH, Mongelard et al. [19] noted that heat denaturation of cell nuclei increased their affinity for DAPI staining and explained this phenomenon as the release of proteins associated with DNA. Raap et al. [26] describe a 40% DNA loss after heat denaturation of nuclei fixed with methanol/acetic acid. However, cells for 3D-FISH were fixed with 4% paraformaldehyde, a fixative known to preserve nucleic acids relatively well. The network of electron-dense material observed in EM after HCl and denaturation steps apparently comprises proteins, DNA, and RNA, making it difficult to determine the extent to which the 3D distribution of DNA is altered at the nanometer scale. Such knowledge is important in the assessment of the usefulness of 3D-FISH in studying the 3D positioning of individual genes. Our LM experiments suggest both loss and redistribution of H2B-GFP in HeLa cell nuclei after the HCl and denaturation steps, while no pronounced redistribution of DNA was observed at this resolution level. However, EM immunolabeling studies suggest a redistribution of DNA in nuclei subjected to 3D-FISH. Further studies are necessary to explore the structural and topological relationships between ⬃1-Mb chromatin domains seen at the LM level and chromatin aggregations seen at the EM level. The limited preservation of nuclear structure after 3D-FISH is underscored by our EM studies and should be taken into account when interpreting data on higher order 3D chromatin organization obtained by this approach. A careful comparison of LM and EM data obtained sequentially in the same nuclei, however, revealed that the dispersion of DNA noted at the EM level is locally constrained most likely by crosslinking effects of paraformaldehyde fixation. While glutaraldehyde fixation should preserve the in vivo structure even better, more extensive crosslinks produced by glutaraldehyde interfere with hybridization efficiency. Multicolor 3D painting of CTs at the LM resolution level [1, 4] indicates that painted CTs fill most of the nucleus, although it was shown [27] that chromatin occupies only about 50% of nuclear volume. Recent observations clearly reveal that in vivo labeled interphase CTs are often surrounded by large lacunes of interchromatin space and that this space also extends into the CT interior [3, 28, 29]. These findings are compatible with model proposing that a part of the interchromatin space is located within CTs [1]. In conclusion, our study demonstrates that reliable conclusions about the relative positions of CTs, arm

22

SOLOVEI ET AL.

domains, and band domains down to the level of ⬃1-Mb chromatin domains can be reliably obtained by 3DFISH. Shifts of genes located on individual chromatin loops can also be detected in the lower micrometer range [9, 30 –32]. Light optical procedures to map 3D positions of hybridized DNA segments on the nanometer scale have recently become available [33]. These upcoming technical developments make it possible in principle to perform a light optical analysis of chromatin loop domains in the order of 100 kb, which in turn may contribute to the formation of ⬃1-Mb chromatin domains [1, 34]. Attempts to study gene positions visualized by 3D-FISH at nanometer levels of resolution run great risk for artifacts of altered chromatin configurations to be measured, and provide little, if any, information on chromatin architecture at the nanometer scale in the nucleus of a living cell. It is probable that milder FISH procedures avoiding the heat denaturation step will provide better preservation of the nuclear morphology. Recent developments have made it possible to visualize CTs, chromatin domains, and transgenes in the nuclei of living cells [14 –16] and should provide new opportunities to overcome these limitations. We thank Prof. W. W. Franke (DKFZ, Heidelberg) and Dr. K. F. Sullivan (The Scripps Research Institute, La Jolla) for supplying the cell lines, and Mrs J. Fakan, F. Voinesco and V. Mamin for excellent technical assistance with electron microscopy. This work was supported by grants from the Deutsche Forschungsgemeinschaft to T.C. (Cr 59/20-1) and to C.C. (FOR 240/2-3), and from the Swiss National Science Foundation to S.F. (31-53944.98). C.S. benefited from a fellowship of the Fondation du 450e Anniversaire de I’Universite´ de Lausanne and Fonds Alain Gautier.

7.

Dietzel, S., Schiebel, K., Little, G., Edelmann, P., Rappold, G. A., Eils, R., Cremer, C., and Cremer, T. (1999). The 3D positioning of ANT2 and ANT3 genes within female X chromosome territories correlates with gene activity. Exp. Cell Res. 252, 363–375.

8.

Kurz, A., Lampel, S., Nickolenko, J. E., Bradl, J., Benner, A., Zirbel, R. M., Cremer, T., and Lichter, P. (1996). Active and inactive genes localize preferentially in the periphery of chromosome territories. J. Cell Biol. 135, 1195–1205.

9.

Volpi, E. V., Chevret, E., Jones, T., Vatcheva, R., Williamson, J., Beck, S., Campbell, R. D., Goldsworthy, M., Powis, S. H., Ragoussis, J., Trowsdale, J., and Sheer, D. (2000). Large-scale chromatin organization of the major histocompatibility complex and other regions of human chromosome 6 and its response to interferon in interphase nuclei. J. Cell Sci. 113, 1565–1576.

10.

Visser, A. E., Eils, R., Jauch, A., Little, G., Bakker, P. J. M., Cremer, T., and Aten, J. A. (1998). Spatial distributions of early and late replicating chromatin in interphase chromosome territories. Exp. Cell Res. 243, 398 – 407.

11.

Zink, D., Bornfleth, H., Visser, A., Cremer, C., and Cremer, T. (1999). Organization of early and late replicating DNA in human chromosome territories. Exp. Cell Res. 247, 176 –188.

12.

Manders, E. M., Kimura, H., and Cook, P. R. (1999). Direct imaging of DNA in living cells reveals the dynamics of chromosome formation. J. Cell Biol. 144, 813– 821.

13.

Pepperkok, R., and Ansorge W. (1995). Direct visualization of DNA replication sites in living cells by microinjection of fluorescein-conjugated dUTPs. Methods Mol. Cell. Biol. 5, 112–117.

14.

Schermelleh, L., Solovei, I., Zink, D., and Cremer, T. (2001). Two-color fluorescence labeling of early and mid-to-late replicating chromatin in living cells. Chromosome Res. 9, 77– 80.

15.

Zink, D., Cremer, T., Saffrich, R., Fischer, R., Trendelenburg, M. F., Ansorge, W., and Stelzer, E. H. (1998). Structure and dynamics of human interphase chromosome territories in vivo. Hum. Genet. 102, 241–251.

16.

Tsukamoto, T., Hashiguchi, N., Janicki, S. M., Tumbar, T., Belmont, A. S., and Spector, D. L. (2000). Visualization of gene activity in living cells. Nat. Cell Biol. 2, 871– 878.

17.

Cremer, T., Kurz, A., Zirbel, R., Dietzel, S., Rinke, B., Schrock, E., Speicher, M. R., Mathieu, U., Jauch, A., Emmerich, P., et al. (1993). Role of chromosome territories in the functional compartmentalization of the cell nucleus. Cold Spring Harbor Symp. Quant. Biol. 58, 777–792.

18.

Zirbel, R. M., Mathieu, U. R., Kurz, A., Cremer, T., and Lichter, P. (1993). Evidence for a nuclear compartment of transcription and splicing located at chromosome domain boundaries. Chromosome Res. 1, 93–106.

19.

Mongelard, F., Vourc’h, C., Robert-Nicoud, M., and Usson, Y. (1999). Quantitative assessment of the alteration of chromatin during the course of FISH procedures. Cytometry 36, 96 –101.

20.

Verschure, P. J., van der Kraan, I., Manders, E. M. M., and van Driel, R. (1999). Spatial relationship between transcription sites and chromosome territories. J. Cell Biol. 147, 13–24.

21.

Solovei, I., Walter, J., Cremer, M., Habermann, F., Schermelleh, L., and Cremer, T. (2002). FISH on three-dimensionally preserved nuclei. In “FISH: A Practical Approach” (J. Squire, B. Beatty, and S. Mai, Eds.), pp. 119 –157. Oxford Univ. Press, Oxford.

22.

Bridger, J. M. L. (1999). Analysis of mammalian interphase chromosomes by FISH and immunofluorescence. In “Chromosome Structural Analysis” (W. A. Bickmore, Ed.), pp. 103–123. Oxford Univ. Press, Oxford.

REFERENCES 1.

2.

3.

4.

5.

6.

Cremer, T., and Cremer, C. (2001). Chromosome territories, nuclear architecture and gene regulation in mammalian cells. Nat. Rev. Genet. 2, 292–301. Cremer, M., v. Hase, J., Volm, T., Brero, A., Kreth, G., Walter, J., Fischer, C., Solovei, I., Cremer, C., and Cremer, T. (2001). Non-random radial higher-order chromatin arrangements in nuclei of diploid human cells. Chromosome Res. 9, 541–567. Cremer, T., Kreth, G., Koester, H., Fink, R. H. A., Heintzmann, R., Cremer, M., Solovei, I. V., Zink, D., and Cremer, C. (2000). Chromosome territories, interchromatin domain compartment, and nuclear matrix: An integrated view of the functional nuclear architecture. C. R. Eukaryot. Gene Exp. 12, 179 –212. Habermann, F., Cremer, M., Walter, J., v.Hase, J., Bauer, K., Wienberg, J., Cremer, C., Cremer, T., and Solovei, I. (2001). Arrangements of macro- and microchromosomes in chicken cells. Chromosome Res. 9, 569 –584. Boyle, S., Gilchrist, S., Bridger, J. M., Mahy, N. L., Ellis, J. A., and Bickmore, W. A. (2001). The spatial organization of human chromosomes within the nuclei of normal and emerin-mutant cells. Hum. Mol. Genet. 10, 211–219. Croft, J. A., Bridger, J. M., Boyle, S., Perry, P., Teague, P., and Bickmore, W. A. (1999). Differences in the localization and morphology of chromosomes in the human nucleus. J. Cell Biol. 145, 1119 –1131.

USE OF 3D-FISH TO STUDY CHROMATIN ARRANGEMENT 23.

Craig, J. M., Kraus, J., and Cremer, T. (1997). Removal of repetitive sequences from FISH probes using PCR-assisted affinity chromatography. Hum. Genet. 100, 472– 476.

24.

Bolzer, A., Craig, J. M., Cremer, T., and Speicher, M. R. (1999). A complete set of repeat-depleted, PCR-amplifiable, human chromosome-specific painting probes. Cytogenet. Cell Genet. 84, 233–240.

25.

Fakan, S. (1994). Perichromatin fibrils are in situ forms of nascent trancripts. Trends Cell Biol. 4, 86 –90.

26.

Raap, A. K., Marijnen, J. G., Vrolijk, J., and van der Ploeg, M. (1986). Denaturation, renaturation, and loss of DNA during in situ hybridization procedures. Cytometry 7, 235–242.

27.

Lopez, G., Marquez J., Ubaldo E., Corkidi G., Echeverria O., and Vasquez Nin G. H. (1996). Three-dimensional analysis of the arrangement of compact chromatin in the nucleus of G0 rat lymphocytes. Histochem. Cell Biol. 105, 153–161.

28.

Visser, A. E., Jaunin, F., Fakan, S., and Aten, J. A. (2000). High resolution analysis of interphase chromosome domains. J. Cell Sci. 113, 2585–2593.

29.

Solovei, I., Kienle, D., Little, G., Eils, R., Savelyeva, L., Schwab, M., Jager, W., Cremer, C., and Cremer, T. (2000). Topology of double minutes (dmins) and homogeneously staining regions (HSRs) in nuclei of human neuroblastoma cell lines. Genes Chromosomes Cancer 29, 297–308.

Received December 18, 2001 Revised version received February 19, 2002

23

30.

Brown, K. E., Guest, S. S., Smale, S. T., Hahm, K., Merkenschlager, M., and Fisher, A.G. (1997). Association of transcriptionally silent genes with lkaros complexes at centromeric heterochromatin. Cell 91, 845– 854.

31.

Brown, K. E., Baxter, J., Graf, D., Merkenschlager, M., and Fisher, A. G. (1999). Dynamic repositioning of genes in the nucleus of lymphocytes preparing for cell division. Mol. Cell 3, 207–217.

32.

Brown, K. E., Amoils, S., Horn, J. M., Buckle, V. J., Higgs, D. R., Merkenschlager, M., and Fisher, A. G. (2001). Expression of alpha- and beta-globin genes occurs within different nuclear domains in haemopoietic cells. Nat. Cell Biol. 3, 602– 606.

33.

Cremer, C., Edelmann, P., Bornfleth, H., Kreth, G., Muench, H., Luz, H., and Hausmann, M. (1999). Principles of spectral precision distance confocal microscopy for the analysis of molecular nuclear structure. In “Handbook of Computer Vision and Applications,” Vol. 3, “Systems and Applications” (B. Ja¨ hne, H. Haussecker, and P. Geissler, Eds.), pp. 839 – 857. Academic Press, San Diego.

34.

Mu¨ nkel, C., Eils, R., Dietzel, S., Zink, D., Mehring, C., Wedemann, G., Cremer, T., and Langowski, J. (1999). Compartmentalization of interphase chromosomes observed in simulation and experiment. J. Mol. Biol. 285, 1053–1065.