Architecture and biological significance of the nucleus: an overview

The nucleus is the biggest organelle in eukaryotic cells (10–20 μm in a diameter in mammalian cells), and it packages the entire genome. It has long been thought that the nucleus is a container that holds nuclear events including gene duplication, transcription, and damage repair. Recent reports, however, have shown that the architecture of the nucleus itself plays important roles in the regulation of genome functions and structures. Namely, the nuclear architecture is closely related to the nuclear functions [1, 2].

In eukaryotes, the genomic DNA is separated from the cytoplasm by the nuclear envelope and forms chromatin inside the nucleus [3–6]. Chromatin formation is achieved through several DNA-folding steps. First, 146-bp DNA segments wrap around a histone octamer (two each of H2A, H2B, H3, and H4) in 1.65 turns to form a nucleosome [7]. The linker histones, which include H1, H5, and other subtypes, are major components of the chromatin and play a significant role in the higher-order packaging of the array of nucleosomes called the “beads-on-a-string” fiber to form the 30-nm fiber [8–11]. In addition to the linker histones, non-histone proteins, such as high mobility group proteins, condensins, and topoisomerases, are also involved in the higher-order structure of the chromatin [12]. In particular, the condensin complex and topoisomerase II (Topo II) are thought to hold two DNA strands in remote regions to create tension in the trapped DNA region [13, 14]. The higher-order chromatin forms transcriptional active and inactive regions named euchromatin and heterochromatin, respectively [15].

A well-known function of the nuclear envelope is protection of the genome from environmental damage, such as ultraviolet irradiation. Moreover, the nuclear envelope also plays other important roles including nuclear transport and regulation of chromatin functions. The nuclear envelope is composed of double lipid bilayers [16] (an outer and an inner nuclear membrane; ONM and INM, respectively), the nuclear pore complex (NPC), and the nuclear lamina. Heterochromatin attaches to both the INM and the nuclear lamina (for a detailed description of the nuclear envelope, see reviews [17–20]). It is thought that the ONM and the endoplasmic reticulum (ER) are the same compartment, as the ONM continues from the ER and ribosomes exist on the ONM. On the other hand, the INM contains specific INM proteins, such as the lamin B receptor (LBR) [21, 22], lamina-associated polypeptides [23], emerin [24], MAN1 [25], and the SUN proteins [26, 27]. Recent proteomic analysis has revealed that about 80 proteins exist in the INM [28]. Some of the INM proteins interact with chromatin and/or the nuclear lamina [26, 29] (see “Interaction between the nuclear envelope and chromatin”).

The inside of the nucleus can be roughly divided into two regions: chromosome territory and the interchromatin compartment. The chromosome territory is the region that is occupied by chromatin. Heterochromatin is located under the nuclear envelope, around the nucleolus, and in the inner region of the chromatin territory, whereas euchromatin with transcriptional activities is situated at the boundary between the chromatin territory and the interchromatin compartment [30, 31]. The interchromatin compartment seems to exclude chromatin and contains sub-compartments including the nuclear matrix, nuclear speckle, Cajal body, promyelocytic leukemia (PML) body, and nucleolus (see “Inner-nuclear structures that interact with chromatin”). These sub-compartments function as the scaffolds for ribosome biogenesis [32], RNA maturation, and other processes. In the interchromatin region, proteins and RNA cooperate to promote biological activities such as transcription and RNA splicing under a constant cycle of assembly and disassembly.

The nucleus possesses a highly complex but well-organized architecture, and the dynamic functions of the nucleus depend on this architecture. Thus, it is important to elucidate the molecular mechanisms of individual nuclear events in relation to specific nuclear structures, hopefully at the nanometer, millisecond, and piconewton levels. Due to the complexity of such nuclear events, the elucidation of the mechanisms will require a coordinated combination of techniques in nanotechnology, biochemistry, and cell and molecular biology.

New approaches to studies of nuclear architecture

The recent development of atomic force microscopy (AFM) [33] has been a breakthrough in the study of nuclear architecture [16, 34, 35]. The instrumentation and application of AFM have been enormously fruitful due to the development of sharp cantilevers for molecular imaging [36–38], cantilever modification techniques for single-molecule force measurement [39–43], and fast-scanning devices for real-time imaging [44–48].

Nanometer scale imaging by AFM

AFM scans the sample surface with a very sharp probe and reveals the surface topography [34, 49, 50]. Thus, AFM routinely identifies the binding site of a DNA-binding protein with a resolution of several tens of base pairs and determines the degree of protein polymerization on a liner DNA of several kilobase pairs [51]. AFM has also revealed the distinct higher-order structures of DNA, including supercoiling [52], stem-loop structures [53], and enhancer-promoted DNA loops [54] as well as nucleosomes and chromatins [55].

The sequential removal of the cell membranes, cytoplasm, and nucleoplasm from cultured cells on a cover slip (Fig. 1a and b) leads to the exposure of interphase chromatin [56]. By using this method and nanoscale imaging, the inner chromatin structures inside the interphase nucleus have been revealed (Fig. 1c and d) [57]. The depletion of cytoplasmic and nucleoplasmic materials from HeLa cells exposes the interphase chromosome that is composed of an ∼80-nm diameter granular unit that forms an ∼80-nm-wide fiber (Fig. 1e and f). In contrast, cytoskeletal fibers are always thinner, with a width of 10–60 nm (Fig. 1g and h).

Fig. 1
figure 1

Micro-dissection of cell nucleus. HeLa cells cultured on a cover glass can be chemically dissected by successive treatment with 0.1% Triton X-100 and 250 mM (NH4)2SO4 and 10 U DNase I (a–c), followed by 100 U DNase I (d). The specimens were then fixed, subjected to a critical-point drying and observed by AFM. Partial treatment with 10 U DNase I exposed chromatin fibers in the nucleus (c), whereas the following harsh treatment with 100 U DNase I removed all of these fibers and revealed the nuclear matrix (d). e Section analysis of the chromatin fiber observed in (c) shows granular units (arrowheads). f The width of the chromatin fiber was measured and summarized after subtracting the effect of the AFM tip curvature. The average value was 78.1 nm. g and h The widths of the cytoskeletal fibers observed in (a) were measured and summarized after subtracting the AFM tip curvature. Most of the fiber widths were distributed between 10 and 70 nm with four major peaks with the center values of 11.6, 22.6, 39.1, and 66.4 nm

Full size image

In human mitotic chromosomes, 70- to 80-nm granular structures become evident under varying concentrations of Mg2+ in electron microscopy (EM) [58] and after treatment with RNase and pepsin in AFM [59]. In the mitotic chromosome of Drosophila melanogaster, the fibrous structures with ∼40 and 80 nm widths can be observed by EM [60]. On the surface of condensed chromosomes purified from sperm cells, ∼70-nm particles have also been identified [61]. When the chromosome is dissected with the AFM tip and its internal structures are visualized, ∼70-nm particles that form a fibrous structure are revealed [61]. These lines of evidence suggest that the 30- to 40-nm thin fiber and the 70- to 80-nm granular fiber are the fundamental structural units of eukaryotic chromosomes; these fibers are relatively stable in a wide range of experimental conditions. Recent reports have demonstrated that these hierarchical structures exist in both eukaryotes and prokaryotes [35] and that nascent single-stranded RNAs are part of the 30–80 nm fiber structures [62, 63].

Piconewton scale force measurement

Other useful biological applications of AFM are to measure the physical properties of a sample surface and to measure interactive forces between biological macromolecules and even between domains within a molecule ([64–68]; also see “Inner-nuclear structures that interact with chromatin”). When the cantilever pushes the sample surface or pulls up one of two interacting molecules, it deflects upward or downward depending on the force applied to it. The force is proportional to the amount of the deflection as described in the following equation: F = kΔx, where k and Δx stand for the spring constant of the cantilever and the deflection, respectively. Thus, by measuring the cantilever deflection, interactions between biological molecules can be characterized. The application of single-molecule force measurement is described in “Nuclear envelope in interphase.”

The force measurement mode can be used to measure the elasticity of living cells. When an AFM cantilever approaches and pushes against the cell surface, a large indentation in the cell and the cell surface is usually observed after the probe first contacts the cell surface. This indentation can be plotted against the force and fitted to the Hertz model equation [69] to estimate the Young’s modulus that describes the elasticity of the sample. The Young’s moduli of the areas over the nucleus (1.8 ± 1.5 kPa) are smaller than those of the cell peripheries (5.1 ± 2.9 kPa). The actin network is responsible for the elasticity of the cell [70]; i.e., the degradation of cellular actin by cytochalasin D leads to a decrease in the measured Young’s modulus [71, 72]. Immunostaining of actin filaments in HeLa cells indicated that these filaments mainly exist in the peripheral areas of the cell and that only a few filaments can be seen over the nucleus [73]. Thus, the actin distribution might be one reason why the areas over the nucleus appear to be softer than the peripheral areas. The elasticity measurements have shown that both the plasma membrane and the nuclear envelope are soft enough to absorb a large deformation formed by the AFM probe. Penetration of the plasma membrane and the nuclear envelope are possible when the probe indents the cell membrane far down close to a hard glass surface. These types of experiments will provide useful information for the development of single-cell manipulation techniques.

Chemical modification of the AFM cantilever and recognition imaging

The most important issue to be considered in single-molecule force measurement is how to attach the molecule of interest to the cantilever. Specific interactions between glutathione and glutathione S-transferase (GST) or between Ni–NTA (nickel–nitrilotriacetic acid) and (His)6 have been used to attach GST-/(His)6-tagged proteins to the AFM cantilever. Methods to covalently bind glutathione or Ni–NTA to an AFM cantilever via a polyethylene glycol (PEG) linker and to attach GST or (His)6-fused protein to this cantilever have been developed [41, 42] (Fig. 2).

Fig. 2
figure 2

Chemical modifications to the AFM cantilever. Schematic illustrations of the covalent coupling of glutathione [42] or antibody [115] to an AFM cantilever. The cantilever made of silicon nitride was amino-functionalized by amino-propyl-tetraethoxy-siliane. To avoid steric hindrance from the cantilever surface and to confer enough flexibility, a PEG linker was inserted between the glutathione and the cantilever surface. A hetero-bifunctional PEG linker with N-hydroxysuccinimidyl (NHS) at one end and the maleimide group at the other end was incubated with the functionalized cantilever. Glutathione was then reacted with the PEG-coupled cantilever to covalently attach to the end of PEG (a). To attach antibody, 5-carboxy 1-pentanethiol was reacted with the PEG-coupled cantilever. The carboxyl group was activated by EDC and NHS, and then attached to an antibody (b)

Full size image

Glutathione contains a cysteine residue in the middle of the tri-peptide (NH2-Gln-Cys-Gly-COOH). The thiol group of the cysteine residue can react to a maleimide group on the PEG-linked AFM cantilever. The rupture force between glutathione and GST was measured to be ∼150 pN [42]. Similarly, Ni–NTA can also be attached to an AFM cantilever via a PEG linker with a rupture force of 150–194 pN against (His)6 [41, 74].

The recent development of recognition imaging under the TREC™ mode has enabled researchers to simultaneously obtain a topographic image and a recognition signal and, thus, to identify a specific molecule within the AFM image. Antibody can also be attached to an AFM cantilever via a PEG linker, and the topographic image and recognition image are obtained simultaneously. With this technique, histone proteins in the reconstituted chromatin have been successfully recognized [75, 76]. The application of TREC™ imaging to the inner nuclear structures is described in “Nuclear matrix.”

From nucleosome to chromatin

Physical properties critical for the nucleosomal formation

The first step of genome folding is nucleosome formation, which involves the wrapping of DNA around the core histone octamer. DNA has an elastic property, and physicochemical studies have estimated the persistence length of DNA as 140–180 bp [77–80]. This property of DNA as a stiff polymer largely contributes to the basis of nucleosome maintenance and dynamics. DNA stiffness influences the efficiency of nucleosome formation [81]. The elastic property of DNA also affects the positioning of the histone octamer. Reconstitution with the histone octamer and a short DNA segment (437 bp) has shown a preferential positioning of the core histone at the end of the DNA segment [82] (Fig. 3a).

Fig. 3
figure 3

From DNA to chromatin in vitro. a Nucleosomes slide along DNA and tend to stack at the end of DNA [82]. b The nucleosome-fiber reconstitution on 100-kbp plasmid in the absence of histone H1. The nucleosome-fiber was reconstituted by a salt dialysis procedure and observed by AFM [8]. c The reconstituted chromatin with linker histone H1. Histone H1 was added to the reconstituted nucleosome-fiber after the salt dialysis was completed, and then observed by AFM [8]. d Topo II was mixed with 14-kbp plasmid DNA in the absence of ATP and observed by AFM [13]. Topo II forms a homodimer with a ring-like shape. e Condensin pentamers were mixed with DNA and subjected to AFM observation [101]. Arrowheads indicate condensin–DNA complexes. f AFM images of the H1 and Topo II-added chromatin fiber. The H1-added chromatin fiber containing 80 ng DNA was treated with 2 ng of Topo II [13]

Full size image

DNA supercoiling results from DNA winding around DNA-binding proteins and/or from the topological constraint imposed on closed circular DNA. Supercoiling plays important roles in genomic events, such as transcription, replication, and recombination, as well as the maintenance of the genome architecture [83–89].

The superhelical constraint of DNA is a critical determinant of the formation of nucleosomal arrays. When the superhelical constraint is completely removed by topoisomerase I, nucleosomes cannot be formed [90]. The histone-to-DNA-weight ratio during the reconstitution is a critical determinant of the manner of compaction. A large transition between the dense and dispersed states in the reconstituted chromatin depends on the weight ratio. This transition is energetically predicted as a function of the nucleosome–nucleosome distance [91]. These analyses suggest that the interaction between nucleosomes as well as the physical properties of DNA (length and superhelicity) play fundamental roles in chromatin dynamics [16].

Histone H1 converts the “beads-on-a-string” fiber to the 30-nm fiber

Structural and biochemical studies have demonstrated that the linker histone plays a significant role in the higher-order packaging of the “bead-on-a-string” fiber (Fig. 3b). The removal of histone H1 results in the unfolding of the 30-nm fiber to the “beads-on-a-string” fiber [11, 92]. Trypsin digestion of histone H1 and the N terminus of histone H3 leads to a loss of the “zig-zag” arrangement of the nucleosomes and deforms the 30-nm fiber [10]. When purified histone H1 is added to the reconstituted nucleosomes formed on a 100-kbp supercoiled plasmid, the H1-containing 30-nm chromatin fibers are formed (Fig. 3c) [8]. These fibers are assumed to be composed of solenoid-like repetitive turns along the fiber axis; namely, they are formed by a three-dimensional arrangement of a nucleosomal array. Several higher-order folding models of the “beads-on-a-string” nucleosomes have been proposed [8, 11, 93]. Because histone H1 reduces the nucleosome–nucleosome distance when it binds to linker DNA [55] and because nucleosomes can freely slide along DNA strands [82], H1 coordinates the assembly of nucleosomes along the DNA strand and contributes to the formation of a well-organized solenoid structure.

Beyond the 30-nm fiber in the nucleus

Topo II and the condensin complex localize to the nuclear matrix fraction of interphase cells [94] and the scaffold fraction of the mitotic chromosome [95]. Topo II forms a ring-shaped homodimer [96, 97] and catalyzes the decatenation and relaxation of the DNA double strand (Fig. 3a) [98]. In Schizosaccharomyces pombe, chromosomes cannot be condensed without functional Topo II (Fig. 3d) [99]. The condensin structure maintenance of chromatin (SMC) assumes a head–tail structure, where the head region assembles with non-SMC proteins (Fig. 3e) [100, 101].

When Topo II is added to 30-nm chromatin fibers that have been reconstituted with histone H1, a large complex is formed regardless of the presence of adenosine triphosphate (ATP; Fig. 3f) [13]. This is not the case for the “beads-on-a-string” fibers (i.e., H1-free chromatin fibers). A pull-down assay using GST-H1 also supports this result [13], suggesting that the chromatin compaction by Topo II is not due to a direct interaction between Topo II and histone H1 but rather an interaction between Topo II and H1-containing chromatin regardless of the presence of ATP.

Histone H1 has two possible roles in compaction: (1) It increases the contact area, strengthening the attraction between fibers, and (2) it reduces the negative charge of the nucleosomal complex, decreasing repulsion between fibers. An H1-containing nucleosome has a smaller negative charge (−100 e) than an H1-free nucleosome (−150 e), suggesting that the strength of Coulombic repulsion is smaller when H1 is present. In bulk solution, the H1-induced 30-nm fibers do not aggregate, which suggests that the fibers have a charged colloidal nature [13]. On the other hand, Topo II appears to induce a parallel alignment of double-stranded DNA chains and to promote the formation of thick filaments in the absence of ATP, indicating that Topo II mediates assembly of DNA strands independently of its enzymatic activity. It is assumed that this interaction also occurs within the DNA–histone complex (i.e., nucleosome) in a histone H1-dependent manner. In the absence of H1, each Topo II mediates a weak attraction between the nucleosomes, whereas in the presence of H1, as the contact areas between the fibers are increased, the attraction between two clusters also increases due to the added attraction via Topo II. Therefore, the attraction between 30-nm fibers would be much stronger than the attraction between the pair of nucleosomes, even if Topo II can enhance both attractions.

In the nucleus, chromatin fibers are held by scaffold/matrix structures that occur every several tens or hundreds of kilobases [102], and the matrix-associated region/scaffold associate region (MAR/SAR) forms a complex with many proteins including Topo II [103]. On the other hand, Topo II by itself does not possess a specific affinity for the MAR/SAR sequences and binds directly to any DNA sequence with low affinity [13]. Therefore, in vivo, mechanisms for loading Topo II on the MAR/SAR must exist, and Topo II condenses 30-nm chromatin fibers in concert with other scaffold proteins such as the condensin complex and SAF-A/SP120/hnRNP U (hereafter called as SP120) [104]. A comparison of in vitro and in vivo chromatin structures is required for determining the significance of the MAR/SAR in Topo II-dependent chromatin condensation.

Inner-nuclear structures that interact with chromatin

In addition to the chromosomal compartment, the nucleus contains nonchromosomal compartments such as nucleolus, PML body, nuclear speckle [105], Cajal body [106], and nuclear matrix. Recent studies identified protein components localized in each compartment [107]. Especially, the studies utilizing fluorescence microscopy revealed the localization and dynamics of these proteins in the nucleus [108, 109]. The application of various nano-techniques to the structural and functional analyses of nuclear matrix, nucleolus, and PML body has just become promising in the last 5 years.

Nuclear matrix

In the early 1960s, a non-chromatin internuclear network of ribonucleoproteins with 10-nm branched filaments was identified as the nuclear matrix/scaffold by EM [56, 110, 111]. On the other hand, the nuclear matrix/scaffold was defined biochemically as a remnant structure after a eukaryotic cell is treated with detergent, high salt buffer, and DNase I [112]. It is important to realize that there are no such “insoluble” materials inside the cell nucleus and that the nuclear matrix/scaffold visible under the microscope is the solidified complex formed during chemical treatments [113]. Nevertheless, the nuclear matrix can be observed by AFM. When HeLa cells are gently treated with Triton X-100, ammonium sulfate, and DNase I, fibrous and dotted structures inside the nucleus are observed (Fig. 5a). Several nuclear proteins have been biochemically identified as the component of the nuclear matrix/scaffold [104]. Among these proteins, the MAR/SAR-binding proteins bind to the MAR/SAR of chromosomal DNA and hold them inside the nucleus.

One of the major MAR-binding proteins, SP120, was identified as one of the heteroribonucleoprotein complexes. This protein has a MAR-binding domain in the N terminus and an arginine–glycine-rich RNA-binding domain in the C terminus [104, 114]. In the nucleus, SP120 interacts and co-localizes with transcriptional factors such as the glucocorticoid receptor and p300 via the RNA-binding domain. In addition, SP120 might interact with several partners (e.g., Topo IIβ) and lead them to different nuclear sub-domains to assure their region-specific localization and function.

Pico-TREC™ can recognize a specific protein by using an antibody-coupled cantilever while obtaining a topographic image with the same probe [75]. The anti-SP120 antibody can be attached to an AFM cantilever with the consumption of less than 10-μg antibody [115], and the localization of SP120 can be elucidated by using this cantilever; individual SP120 molecules can be identified within nuclear components (Fig. 4a–c). This technique will be useful for further structural investigation of the inner-nuclear organization.

Fig. 4
figure 4

Nuclear scaffold/matrix protein visualized by recognition imaging with Pico TREC™ (a–c) and hybrid imaging by AFM and fluorescence microscopy (d–f). The recognition signal and topographical signal were simultaneously recorded. Disrupted nuclei purified from HeLa cells were covalently attached on the amino-modified glass surface and imaged by TREC™-mode AFM with the anti-SP120 antibody-coupled cantilever. The topographical images (a) and corresponding recognition images (b) are shown. Scale bars = 1,000 nm. The merged image is shown in (c) where the recognition signals are in blue. The dotted SP120 molecules were recognized within the nuclear components (c). d–f Enhanced green fluorescent protein-fused PML was transiently expressed in HeLa cells on a cover glass. The cells were successively treated with detergent, high-salt solution and DNase I to expose nuclear matrix. The specimen was then observed by inverted fluorescence microscopy equipped with an AFM scanner on the stage. The AFM image (d) and the fluorescence image (e) are merged (f). The location of PML bodies can be mapped on the AFM image [purple dots in (f)]. Scale bars in (a–c) and (d–f) are 1 and 2 μm, respectively

Full size image

PML body

The PML bodies are nuclear matrix-attached particle-like structures that are larger than the nuclear speckles (0.2–1 μm in diameter) and were originally found in PML cells. In the AFM image (Fig. 4d–f), the PML bodies appear as several speckles within the nuclear matrix. One mammalian nucleus has 10–30 PML bodies. The core component of the PML body is a protein called PML, which is a tumor suppressor gene product. Three lysine residues in the PML protein are modified by a ubiquitin-like protein, SUMO-1, and this modification is needed for the formation of PML bodies. The SUMOylation is also essential for PML protein to form a complex with other proteins such as transcription factors (Sp100, CBP), heterochromatin protein HP1, tumor suppressor proteins (Rb, p53), SUMOylation enzymes, and kinases [116–120]. Because of these components, PML bodies are thought to be involved in biological events such as DNA repair, apoptosis, and transcription. In fact, γ-ray irradiation results in the recruitment of p53 to PML, and cells that lack PML fail to induce p53-dependent gene expression [118, 121]. PML bodies could also stabilize transcription co-activator-transcription factor complexes because of their ability to interact with several transcriptional co-activators such as CBP and p300 [122].

Nucleolus

The nucleolus is the most prominent substructure of the nucleus. It is a large assembly of rRNA genes (rDNA), rRNAs, and more than 700 proteins containing RNA transcription factors, processing enzymes, and ribosomal protein subunits [123, 124]. The nucleolus is structurally distinctive among the subcellular organelles because it has no membrane. Nevertheless, there are three different sub-domains in the interphase nucleolus, and each compartment is involved in different steps of ribosomal biogenesis [125, 126]. The fibrillar center (FC) is the center of rRNA transcription and contains the rRNA transcription machinery such as RNA polymerase I and upstream binding factor. The dense fibrillar component (DFC) contains the rRNA early processing factors such as fibrillarin, and the granular component (GC) contains the late processing factors and ribosomal maturation enzymes such as nucleolin and B23 (nucleophosmin). This hierarchical organization of the interphase nucleolus has been shown by EM and fluorescence microscopy; these imaging techniques have revealed that the FC consists of several foci (100–500 nm in diameter) and is surrounded by the DFC [127]. The GC is distributed throughout the whole nucleolus and forms intricate structures with heterochromatin in the periphery [128]. AFM and immunofluorescent microscopy revealed that the nucleolus in HeLa cells is a rigid structure and that it remains on the glass substrate after sequential treatment by detergent, high salt buffer, and DNase I (Fig. 5a, b) [57].

Fig. 5
figure 5

Nucleolar structure and dynamics. a An AFM image of the nucleolus on the nuclear matrix. HeLa cells were sequentially treated with detergent, high-salt buffer, and DNase I. Arrows indicate nucleoli. b The subcellular localizations of the nucleolar proteins in interphase. Left double staining of the FC protein (our unpublished observation, green) and fibrillarin (DFC protein, red). Right localization of the nucleolin (GC protein). c The subcellular localizations of the nucleolar proteins during mitosis. Upper panels double staining of FC-localizing antigen (green) and DNA (red). Lower panels double staining of GFP-fibrillarin (green) and nucleolin (red). The PNBs are observed at telophase (arrows) and gradually targeted to the NORs (arrowheads). Scale bars = 10 μm

Full size image

The nucleolus undergoes drastic structural changes during mitosis. The hierarchical structures of the interphase nucleolus collapse in the beginning of mitosis and then gradually reassemble during telophase to early G1 phase. In this mitotic step, nucleolar organizing proteins exhibit distinctive dynamics (Fig. 5c). The FC proteins form foci on chromosome throughout mitosis, and the spots correspond to the nucleolar organizing regions (NORs) on the acrocentric chromosomes as shown by immunostaining on the mitotic chromosome spreads [129]. This mechanism enabled the equal distribution of the proteins into two daughter cells and the resumption of rRNA transcription at telophase. Interestingly, not all NORs are actively transcribed. The active NORs are associated with RNA polymerase I, whereas the inactive NORs contain no transcription machinery and are highly methylated [130]. However, structural differences between these active and inactive chromosomes remain unclear. The recently developed AFM technique of chromosome manipulation and imaging in liquid is expected to soon reveal the nanoscale structure of NORs. The DFC and GC proteins disperse to the cytoplasm in the beginning of mitosis and localize to the surface of condensed chromosomes. During anaphase to early telophase, these proteins formed intermediate structures called the pre-nucleolar bodies (PNBs) [131], and then, they are sequentially transported to the NORs, reflecting the roles of the proteins in rRNA processing and the ribosomal maturation process [132–134]. Inhibition of the cyclin-dependent kinase 1 (CDK1)–cyclin B activity during mitosis does not interfere with the reassembly of fibrillarin but does block the relocalization of one of the GC proteins, Nop52. Therefore, CDK1–cyclin B is partially responsible for the nucleolar reassembly, but the detailed mechanisms of the regulation of the entire event remain unknown [135].

Interaction between the nuclear envelope and chromatin

The nuclear envelope is a nanostructure that consists of the ONM, INM, NPC, and nuclear lamina (Fig. 6a). The structure of the nuclear envelope is critical for the control of the chromatin territory. The proteins in the INM, such as LBR, are especially responsible for anchoring specific chromatin domains near the INM [25]. The nuclear lamina lies under the INM and guarantees the strength of the nucleus (Fig. 6a and b). EM observations have shown that the nuclear lamina is a mesh-like structure [136]. The major components of this structure are A- and B-type lamins. Both A- and B-type lamins form a dimer via their rod domains [137, 138]. The dimers further polymerize via the head–tail domains and form a para-crystal structure around 100 nm in length in vitro.

Fig. 6
figure 6

The nuclear envelope during interphase (a–d) and mitosis (e). a Schematic illustration of the nuclear envelope. b AFM images of the nuclear envelope [164]. A germinal vesicle was isolated from a Xenopus oocyte. Its cytoplasmic (left) and nucleoplasmic face (right) were mechanically spread onto the mica and then imaged by AFM. Enlarged images are NPCs. Scale bars = 200 nm. c The rupture force measurement between LBR and chromatin. The GST-fused N-terminal region of LBR (amino acid region corresponding to 1–211) was expressed in Escherichia coli and purified by glutathione Sepharose beads. The purified GST-LBR was bound to an AFM cantilever via glutathione. Reconstituted chromatin was prepared from 26-kbp plasmid DNA and purified human histone octamer by the salt dialysis method and then bound to mica. The rupture force between LBR and reconstituted chromatin was measured, and a histogram of the rupture force is shown. d FRAP analysis of LBR mobility. GFP-tagged wild-type LBR (black) and a deletion mutant of the chromatin-binding region of LBR (Δ1–53, red) were transiently expressed in HeLa cells. The mobility of these proteins in the nuclear envelope was observed under LSM 510 META. e Nuclear membrane reassembly [165]. Xenopus egg crude extract was incubated with Xenopus sperm chromatin at 23°C. The chromatin-targeted membrane was stained with dihexyloxacarbocyanine iodide, and the central (center) and peripheral (surface) regions were observed by confocal microscopy. Arrowheads targeted nuclear membrane. Scale bars = 10 μm

Full size image

The structure of nuclear envelope is important, as the loss of it causes the disease termed nuclear envelopathy or laminopathy [138–140]. Then, the structural analysis of the nuclear envelope at nanoscale is needed.

Nuclear envelope in interphase

For decades, the nuclear envelope was recognized as a static structure in interphase. However, recent investigations with fluorescence recovery after photobleaching (FRAP) have revealed that most nuclear envelope proteins are mobile in the nuclear envelope except for the nuclear lamina and some of the nucleoporins [141–146]. The NPC functions as a gate to import and export substances into and out of the nucleus [147, 148]. It is a huge protein complex (Fig. 6b) with a molecular mass of 125 MDa, and it penetrates both the ONM and the INM [149, 150]. The NPC is composed of about 30 proteins named nucleoporins and forms an eightfold symmetrical structure [151–155]. The NPCs are likely to be newly synthesized and inserted into the nuclear envelope in the interphase cells. Interestingly, a huge NPC-free island can be observed during the early G1 phase in HeLa S3 cells, and this pore-free region disappears at the late G1 phase [156]. These results suggest that the NPCs are also under dynamic mobility control within the nuclear envelope.

The INM protein mobility seems to be restricted by the interaction between INM proteins and chromatin and/or nuclear lamina. Emerin and MAN1 are more mobile in the nuclear envelope in mouse embryonic fibroblasts lacking A-type lamin than in wild-type cells [157]. In contrast, a lack of A-type lamin does not affect the LBR mobility. The LBR seems to directly interact with chromatin [158]. Indeed, when the GST-fused N-terminal region of LBR (amino acid region 1–211, which contains the chromatin-binding domain) is attached to a glutathione-modified AFM cantilever, it exhibits a rupture force of ∼50 pN against reconstituted chromatin (Fig. 6c). In live cells, the mobility of LBR in the nuclear envelope is much smaller than that of an N-terminal deletion mutant (Δ1–53) (Fig. 6d), suggesting that the interaction between the LBR and chromatin may be a key factor for chromatin attachment to the nuclear envelope.

Nuclear envelope at the mitotic phase

In many organisms, the nuclear envelope becomes invisible at the beginning of mitosis and reappears at the end of cell division (some organisms, such as most fungi, undergo closed mitosis, in which the nuclear envelope stays visible). At the onset of mitosis in animal cells, all of the nuclear envelope components (the INM, ONM, NPC, and nuclear lamina) are completely or partially broken down into small fragments or vesicles and disperse throughout the entire cell. After chromosome segregation in telophase, the nuclear envelope reassembles around the chromatin.

The molecular mechanism of nuclear envelope reassembly has been well studied in animal cells by using the Xenopus egg cell-free system (for details, see reviews [18, 19, 159]). Observations from fluorescence and EM have revealed that nuclear envelope reassembly occurs in three distinct steps: membrane vesicle targeting, membrane fusion, and growth (Fig. 6e) [160, 161]. At the end of mitosis, Ran, a small G protein, accumulates around the chromosome surface and forms a Ran-GTP gradient that causes the targeting of the nuclear membrane vesicles toward the chromosome surface. The INM proteins are also required in this step. Interestingly, Xenopus egg extracts contain two different vesicles, PV1 and PV2, which can be distinguished by ultracentrifugation [162]. These different vesicles are observed as ribosome-carrying ‘rough’ vesicles and ribosome-free ‘smooth’ vesicles [160]. The targeted membrane-vesicles fuse with each other, develop into a network with further vesicle fusion, and finally form a closed nuclear double-membrane around the entire area of nuclear chromatin. It seems that PV1 is a prerequisite for this step and that PV2 may be required for proper NPC assembly [162]. Finally, nuclear proteins (e.g., lamin) are imported into the growing nucleus through the newly synthesized NPCs, and the nucleus matures into a sphere-like structure.

The NPCs reassemble in parallel with the membrane fusion. However, as the depletion of the Nup 107–160 complex inhibits NPC formation but not nuclear membrane reassembly [163], the nuclear membrane reassembly is independent of the NPC formation. The structural aspects of the underlying mechanisms will hopefully be solved in the near future by using AFM.

Perspectives

The topographical imaging and visualization of biological specimens achieved by AFM have been extremely powerful tools for exploring basic concepts in fundamental biological processes. In this review, we have shown that AFM can be used to visualize the nanoscale structures of the inner nuclear architecture and higher-order folding of chromatin fibers. This approach has increased the understanding of how chromosomes are built up from DNA, histone, and other protein components. The physical properties of DNA were found to be critical for chromatin dynamics [16]. In contrast, the single-force measurement and recognition imaging have not yet been widely used. This is partly because a reliable procedure for attaching the protein of interest to the cantilever had not yet been established. We have developed new, efficient, and reliable methods for the attachment of antibodies and GST-fused proteins to the AFM cantilever [115], and we have used these methods to demonstrate a direct interaction between chromatin and INM proteins.

The field of nano-biology using scanning probe microscopy has been quickly progressing since the invention of AFM [33]. Just as the recent progress in genome research greatly benefited from the information obtained with bioinformatics, nano-biology will proceed to the next level by utilizing information from genome research and newly invented technology including recognition imaging [75] and fast-scanning AFM [47]. The elucidation of nuclear dynamics at the nanometer and piconewton levels will be accelerated. In the near future, it will also be possible to isolate individual processes in nuclear dynamics by fast-scanning AFM [44, 45]. Even beyond our specific objectives, recognition imaging, force measurement with specific probes, fast-scanning AFM, and other nanotechnologies are expected to be powerful tools in the various fields of biology.