Mammalian interphase and mitotic cells were analyzed for their cation composition using a three-dimensional high resolution scanning ion microprobe. This instrument maps the distribution of bound and unbound cations by secondary ion mass spectrometry (SIMS). SIMS analysis of cryofractured interphase and mitotic cells revealed a cell cycle dynamics of Ca2+, Mg2+, Na+, and K+. Direct analytical images showed that all four, but no other cations, were detected on mitotic chromosomes. SIMS measurements of the total cation content for diploid chromosomes imply that one Ca2+ binds to every 12.5–20 nucleotides and one Mg2+ to every 20–30 nucleotides. Only Ca2+ was enriched at the chromosomal DNA axis and colocalized with topoisomerase IIα (Topo II) and scaffold protein II (ScII). Cells depleted of Ca2+ and Mg2+ showed partially decondensed chromosomes and a loss of Topo II and ScII, but not hCAP-C and histones. The Ca2+-induced inhibition of Topo II catalytic activity and direct binding of Ca2+ to Topo II by a fluorescent filter-binding assay supports a regulatory role of Ca2+ during mitosis in promoting solely the structural function of Topo II. Our study directly implicates Ca2+, Mg2+, Na+, and K+ in higher order chromosome structure through electrostatic neutralization and a functional interaction with nonhistone proteins.
Introduction
Although multiple indirect experiments have proven the over 70-yr-old hypothesis that DNA in nuclei is complexed with monovalent and divalent cations (Hammarsten, 1923), there have only been a few attempts to obtain direct three-dimensional (3D)* high resolution images of cations in cells and on chromosomes without the use of secondary fluorescent indicators (Kearns and Sigee, 1980; Chandra et al. 1984; Horoyan et al., 1992). To date, all cellular cation distributions have only been shown at a spatial resolution ≥500 nm and, in addition, even less is known regarding specific interactions between cations and chromatin-binding proteins and their influence on chromosome structure. Therefore, revealing cation distributions and cation–protein interactions in cells is important to understand the roles of cations during the cell cycle and in the maintenance of higher order chromosome structure.
The most abundant cations in the eukaryotic cell are Ca2+, Mg2+, Na+, and K+. These cations are fundamental for multiple cellular processes in every phase of life including cell growth and differentiation, development, cell–cell interactions, morphology, motility, and apoptosis leading to cell death (for review see Boynton et al., 1982). The major Ca2+ storage sites in the cell are the ER, Golgi complex, mitochondria, secretory granules, and nuclear envelope (for review see Rottingen and Iversen, 2000). The K+ storage sites are mainly the cytosol, Golgi complex, and the nucleus (Schapiro and Grinstein, 2000). Multiple ion transmembrane pumps (ATPases) and exchangers are responsible for Ca2+, Mg2+, Na+, and K+ cellular influx and efflux to regulate the cellular cation concentrations from internal storage sites and to maintain osmolarity (for review see Scheiner-Bobis, 1998).
Cations have been implicated in the regulation of the cell cycle (for reviews see Boynton et al., 1982; Hepler, 1994). For example, increasing concentrations of Ca2+, Mg2+, and Na+ in the media of growing cells stimulated the mitotic rate almost twofold, whereas in contrast, K+ had no effect (Atkinson et al., 1983). In other studies, cellular Na+ concentration levels appeared to fluctuate throughout the cell cycle peaking in M and S phase, whereas the K+ nuclear and cytoplasmatic concentration levels remained unchanged (Cameron et al., 1979; Warley et al., 1983). Several studies have implicated a major role of Ca2+ in mitosis, correlating with nuclear envelope breakdown and entry into mitosis, microtubular breakdown at the meta- to anaphase transition, and a brief Ca2+ increase at the anaphase onset (Poenie et al., 1986), which led to activated chromosome motion (Groigno and Whitaker, 1998). The meta- to anaphase transition could be prevented with EGTA or the more specific Ca2+-chelator 1,2-bis[o-aminophenoxy]ethane-N,N,N′,N′-tetraacetic acid (BAPTA) in Ca2+-free medium (for review see Hepler, 1994).
The main cations interacting with DNA are Ca2+, Mg2, Na+, and K+. The divalent cations bind to the negatively charged phosphate residues of DNA in a stoichiometry of 1 mol Ca2+ or Mg2+ to 2 mol phosphate (e.g., Mathieson and Olayemi, 1975). Recent crystallization studies of B-DNA decamers or dodecamers in the presence of Mg2+ or Ca2+ confirmed a direct cation interaction with the major and minor grooves as well as phosphate oxygen atoms contributing to DNA stabilization and conformation (Minasov et al., 1999; Chiu and Dickerson, 2000). These crystallization studies resolved that Ca2+ has a higher affinity to DNA, inducing a greater DNA bending and thermal stabilization than Mg2+.
Several studies have shown that mono- and divalent cations are essential in maintaining higher order chromatin structure. For example, chromatin at low ionic strength and in the absence of divalent cations has an extended structure representing the 10-nm coil. The transformation to a more compact or 30-nm structure, that is described as a solenoid or arrangement of superbeads, could be induced by an increase of Na+ or Mg2+ and Ca2+ (for review see Felsenfeld and McGhee, 1986). Models of mitotic chromosome structures have been proposed for the folding of the chromatin fiber >30 nm, each based on the hierarchical organization of eukaryotic chromatin into loops and coils (Ohnuki, 1968; Rattner and Lin, 1985; Filipski et al., 1990). In the loop-scaffold model, highly ademine-thymine (AT)-rich DNA elements, named scaffold-associated regions (SARs) interact dynamically with nonhistone proteins to form loop anchorage sites (Paulson and Laemmli, 1977). SARs play a key role as cis elements of chromosome dynamics and as initiation elements for chromosome condensation (Strick and Laemmli, 1995). Nonhistone binding proteins like topoisomerase IIα (Topo II) and scaffold protein II (ScII) (homologue to hCAP-E, an SMC protein), have been implicated as partners in a nuclear complex (Ma et al., 1993) and also colocalize at the chromosomal axis (Lewis and Laemmli, 1982; Saitoh et al., 1994). Topo II and protein complexes, called condensins, including hCAP-C and -E and other SMC proteins, are essential for chromosome condensation, structure maintenance, and sister chromatid separation (Adachi et al., 1991; Schmiesing et al., 1998; Hirano, 1999).
There are several methodologies available to measure both free and bound cations in cells. Fluorescent indicators, like fura-2 were developed to detect free cations, especially Ca2+ (Poenie et al., 1986). However, these indicators encounter several technical problems; for example, they can bind nonspecifically to cell constituents or other cations (McCormack and Cobbold, 1991). Although X-ray crystallography can detect Ca2+ and Mg2+ on DNA oligonucleotides (Minasov et al., 1999), Na+ and K+ cannot be easily distinguished because of the interference with H2O. In addition, cryogenic temperatures and crystal packing effects occurring during X-ray crystallography may shift the ion distribution of the sample. X-ray microanalysis has been used to measure cations in the nucleus and cytoplasm during the cell cycle (Cameron et al., 1979; Warley et al., 1983). Although this technique involves lengthy exposures and needs substantial spectral corrections due to the presence of a high continuous background, this method is a valuable tool for chemical quantitation.
In contrast to X-ray microanalysis, which determines the atomic composition only at specific points within a biological sample and with no direct microscopic representation and depth information, secondary ion mass spectrometry (SIMS) (Benninghoven, 1987) measures the isotopic composition either in the form of stable or tracer isotopes with high sensitivity, high spatial resolution, and essentially no background. Previous SIMS investigations using BrdU-labeled human and polytene chromosomes demonstrated the use of tracer isotopes (Levi-Setti et al., 1997). SIMS signals provide a rapid visualization of the isotopic distribution within a sample, and SIMS sequential mapping renders in-depth analytical information for constructing 3D compositional images (SIMS tomography).
In this investigation, we show for the first time high-resolution analytical images of the cation composition of mammalian interphase and mitotic cells as well as of isolated metaphase chromosomes using the University of Chicago (Chicago, IL) scanning ion microprobe (Levi-Setti, 1988; Chabala et al. 1995). To preserve the ionic integrity of the analyzed cells and prevent the well-known occurrence of analytical artifacts due to the high diffusivity of cations in biological samples (Morgan et al., 1975), we used fast cryopreservation methods (freeze drying and freeze fracture) (Echlin, 1984), without any prefixations or washes. This study presents SIMS imaging evidence of cation redistribution between cytosol and chromosomes during the cell cycle and proves that Mg2+, Ca2+, K+, and Na+ are an integral part of mitotic chromatin. Our results indicate that Ca2+ directly binds to Topo II in the absence of DNA, supporting a regulatory role of Ca2+ in the reversible transition of a catalytically active Topo II to a structural DNA binding protein during mitosis.
Results
Cryopreserved and fractured interphase and mitotic cells show cell cycle dynamics of cations
3D-SIMS analysis of cryopreserved and cryofractured Indian muntjac (IM) deer fibroblasts showed a specific cation distribution inside the cell and throughout the cell cycle (Fig. 1, A–E). In interphase cells 40Ca2+ distributed throughout the cytosol with specific accumulations, possibly representing the ER and Golgi complex, whereas the nucleus was reduced of Ca2+ (Fig. 1 B). In contrast to interphase, mitotic cells showed high concentrations of Ca2+ on chromatin (Fig. 1, C–E). A similar distribution was detected for 24Mg2+ in interphase and mitotic cells (Fig. 1, B–E). A SIMS analysis of 10 nocodazole-arrested IM cells demonstrated Ca2+ and Mg2+ enrichment on mitotic chromatin. A colocalization of these cations with mitotic chromatin could be proven using 81Br−dU as a tracer for DNA (Fig. 2). In contrast to Ca2+ and Mg2+, the monovalent cations 23Na+ and 39K+ remained unchanged in their cellular distribution throughout the cell cycle. Na+ in interphase and mitotic cells distributed throughout the entire cytosol and nucleus and associated with mitotic chromatin. K+ was enriched in the nucleus and especially in the nucleolus during interphase and also on chromatin during mitosis (Fig. 1, A–E). Fig. 1 A shows three successive K+ layers of an entire interphase cell. The first layer shows in addition to the nucleus positive K+ signals at the rim of the cell, which could represent Na+/K+ ATPase pumps and K+ channels. Using SIMS, we observed that omitting buffer washes of cells before the cryopreservation maintained morphology and consistent cation distributions among multiple samples tested (unpublished data).
Specific cations are a fundamental part of mammalian metaphase chromosomes
The specific association of 40Ca2+, 24Mg2+, 39K+, and 23Na+ with mitotic chromatin in cryopreserved cells led to a further analysis of fractionated metaphase chromosomes for cation binding using SIMS. To rule out methodological artifacts, we used two different chromosome harvesting and fixation procedures with IM and human BV173 cells. Both methods resulted in similar SIMS cation maps of chromosomes, where Ca2+, Mg2+, K+, and Na+ signals followed the chromatids (Fig. 3, A–D). However, we observed that fractionated p-formaldehyde–fixed chromosomes were more compact, rigid, and impervious to erosion with the ion probe than fractionated methanol-acid fixated chromosomes (Fig. 3 C). We also tested simultaneously the same chromosomal preparations for six additional specific divalent cations, like 55Mn2+, 56Fe2+, 58Ni2+, 59Co2+, 63Cu2+, and 65Zn2+. SIMS analysis demonstrated that these tested divalent cations were not associated with IM and BV173 chromosomes (unpublished data).
The specific binding of Ca2+ and Mg2+ was investigated further by SIMS to search for chromosome substructures, such as the chromosome axes, for example. To visualize the chromosome axes, which contain AT-rich DNA (Saitoh and Laemmli, 1994), IM cells were grown in the presence of the thymidine analogue BrdU (Levi-Setti et al., 1997). BrdU-labeled metaphase chromosomes at the second cell division (where the label content of sister chromatids is in the ratio 2:1) were then analyzed for 81Br−, 40Ca2+, and 24Mg2+. As shown in Fig. 4, A and B, the 81Br− (as seen for the doubly Br−-labeled chromatid), and Ca2+ signals were both localized at the AT-rich axes, whereas Mg2+ was more equally distributed over the entire chromatid. The 26CN− map, (Fig. 4 C) representing the overall protein/DNA chromosomal profile, is broader than those of Br− and Ca2+. The Br−/Ca2+ colocalization was further examined by obtaining average signal intensity profiles of Br−, Ca2+, and CN− (counts/pixel), measured across the chromatids and plotted as a function of distance (μm) from the chromosomal symmetry axis (Fig. 4, D–F). We found that the Br− signals peaked at ∼0.3 μm (setting the chromosomal symmetry axis as 0 μm), and the maximal Ca2+ signals occurred at 0.3–0.5 μm, at the chromatid axis. In contrast, the CN− map, signifying the chromatid borders (0–1.0 μm) showed a valley at the Ca2+ peak. At the chromosomal symmetry axis (0 μm), we detected a Ca2+ and Br− valley, but a sharp CN− peak, indicating a protein enriched region (Fig. 4 F). The structure of metaphase chromosomes is described in many different models, which also can coexist, like radial loops and helical coils (Rattner and Lin, 1985). This study shows a tripartite chromosomal structure, a center, axes, and an outer chromatid region, but does not reveal if loops or coils or both are present.
Interestingly, 10% of chromosomes showed a more refined coiling pattern for Ca2+ along the AT-rich axes, which appears to mimic the distribution of the main scaffold proteins Topo II and ScII (Fig. 5). We also observed that, in some of these cases, each Ca2+/sister chromatid coiling was in opposite helicity, as described by Ohnuki (1968) and Boy de la Tour and Laemmli (1988) (unpublished data). Our results imply that Ca2+ colocalizes with Topo II and ScII at the chromatid axes.
Cation quantitation on chromosomes
The Ca2+ and Mg2+ cell and chromosome concentration measurements shown in Table I were obtained from the SIMS images by averaging the signal intensities over the entire cell, cellular compartments or chromosomes and using the calibration plots of Fig. 4, G and H. For example, we determined the total Ca2+ concentration of a cryopreserved interphase IM cell as 7.0–9.0 mM; and of a mitotic IM, cell 4.0–8.0 mM (n = 6). In contrast, the total Mg2+ concentration of an interphase and mitotic IM cell was 1.0–3.0 mM, respectively. In addition, we determined that the averaged divalent cation concentrations for a diploid set of chromosomes are for Ca2+ in the range of 20–32 mM; and for Mg2+, 12–22 mM (n = 15) (Table I; Fig. 4, G and H). Assuming there are 6 × 109 DNA base pairs in a diploid eukaryotic cell, we calculated ratios of one Ca2+ for every 12.5–20 nucleotides and one Mg2+ for every 20–30 nucleotides, which is equivalent to 10–16 Ca2+ and 6.7–10 Mg2+ per nucleosome (200 bp). In addition, after scanning across chromosomes, we also determined a 3:1 concentration ratio of Ca2+ to Mg2+ on the chromatid axis, supporting that Ca2+ is enriched at the AT-rich axes.
Ca2+ and Mg2+ are essential for the integrity of chromosome structure
Previous investigations have demonstrated that incubation of cells with chelators results in partially decondensed mitotic chromosomes (Zelenin et al., 1982; Herzog and Soyer, 1983; Earnshaw and Laemmli, 1983; Staron, 1985). These studies also showed rescue of the condensed chromosome state after addition of Ca2+ or Mg2+. To quantitate the chromosomal loss of cations due to chelators and identify cation–protein interactions, IM cells were released from a G2/M phase block and then depleted of divalent cations using the chelators BAPTA-AM/BAPTA or EGTA as they entered mitosis. Cation-depleted chromosomes were then isolated and analyzed for Ca2+ and Mg2+ by imaging SIMS. The ion-induced secondary ion (ISI) topographical maps, and the images obtained using fluorescent DAPI (Figs. 6 and 7) and YOYO-1 staining, exhibited highly swollen partially decondensed chromosome structures, in contrast to the compacted control chromosomes. Using for example 5 mM EGTA, a 5-fold and 10-fold reduction of Mg2+ and Ca2+ occurred at the chromosomes, respectively (Figs. 6, E and F, and 7 D). We also observed sister chromatid disjunction (except at the centromeres) (Figs. 6 I and 7 A, arrows).
We further examined the Ca2+ and Mg2+ depleted chromosomes for cation–protein interactions by immunofluorescence (IF) and immunoblotting using specific antibodies against histone H1 and the nonhistone scaffold proteins Topo II, ScII, and hCAP-C. A 10-fold reduction of antibody staining for Topo II and ScII proteins was observed after Ca2+/Mg2+ depletion (Fig. 7, A–D), which was confirmed using SDS gel electrophoresis and immunoblotting (Fig. 7, B and C). In contrast, the condensation protein hCAP-C, the linker histone H1 and the core histones, remained bound on Ca2+- and Mg2+-depleted chromosomes using IF and immunoblotting (Figs. 6, H and J, and 7, B and C). In particular, IF staining for histone H1 on isolated Ca2+- and Mg2+-depleted chromosomes showed similar signal intensities to the control chromosomes, but the chromatin was partially decondensed (Fig. 6, H and J).
Ca2+ directly binds and inactivates the enzymatic activity of chromosomal Topo II
It is well documented that Topo II localizes at the mitotic chromosome axes (Earnshaw et al., 1985; Saitoh and Laemmli, 1994). From our SIMS analyses, we derived a ratio of 3:1 Ca2+/Mg2+ on the chromosomal axis (Figs. 4 and 5). Therefore, we determined the Topo II enzymatic activity in vitro in the presence of different cation concentration ratios. This assay determines the Topo II relaxation activity of negatively supercoiled plasmid DNA or catenated kinetoplast DNA in the presence of Mg2+, which is essential for catalytic activity (Osheroff and Zechiedrich, 1987). Our results showed that Topo II activity was inhibited 57–90% when the molar ratio of Ca2+/Mg2+ was in the range 1:1–3:1, respectively (Fig. 8 A). This supports the notion that Topo II may be enzymatically inactive at the metaphase chromosome axes.
To directly detect Ca2+ binding on chromosomal proteins, we applied the fluorescent filter binding assay using the Ca2+ marker quin-2, which has a Ca2+ binding constant of 1.3 × 107 (Tatsumi et al., 1997). Although fluorescent Ca2+ markers, like quin-2, have difficulties detecting Ca2+ inside of cells (McCormack and Cobbold, 1991), these markers are more sensitive than 45Ca for detecting Ca2+-binding proteins directly on polyvinyldifluoride (PVDF) membranes after SDS gel electrophoresis (Tatsumi et al., 1997). In addition to known Ca2+-binding control proteins, like calmodulin and albumin (Tatsumi et al., 1997), we also observed direct Ca2+ binding with both purified human Topo II and chromosomal IM Topo II (Fig. 8 B). In contrast, no Ca2+ binding was detected with other chromosomal binding proteins, including hCAP-C, ScII, and histones.
Discussion
This study has directly analyzed the cation composition of cryopreserved mammalian interphase and mitotic cells, as well as fractionated untreated and cation-depleted chromosomes at a resolution of 50 nm using SIMS. Quantitative direct SIMS imaging of cryopreserved cells demonstrated that Na+ and K+ were associated with chromatin throughout the cell cycle, whereas in contrast Ca2+ and Mg2+ exhibited a localization change during interphase (mainly cytoplasm) and mitosis (chromatin). Our study provides the first high resolution images of cations inside interphase and mitotic cells without using fluorescent dyes and confirms the SIMS analysis of Chandra et al. (1984) using rat interphase cells at a resolution of ∼500 nm. The direct association of Ca2+, Mg2+, Na+, and K+ with cryopreserved mitotic chromatin and fractionated mitotic chromosomes finally confirms earlier experiments of cation binding to DNA and chromatin in vitro. Using SIMS and Ca2+ calibration standards in agarose, we quantified 7–9 mM in interphase and 4–8 mM in mitotic cells, and 12–24 mM peak intensities on mitotic chromatin (Table I). Comparable Ca2+ concentrations were detected in different cells, but they were detected using X-ray microanalysis and different ion calibration matrices. For example, with similar cell preparation techniques, a cellular Ca2+ concentration of ∼8.8–10.2 mM was detected in rabbit cells (Wroblewski et al., 1983) and ∼4.7–7.3 mM in T cells (Kendall et al., 1985), but using gelatin as a calibration matrix. A Ca2+ concentration of ∼2.4–3.4 mM was found in mouse interphase cells, but a concentration ∼9.5–11.0 mM was found at metaphase chromatin using X-ray microanalysis and BSA as a calibration matrix (Cameron et al., 1979). The increase of Ca2+ localized to mitotic chromatin compared with interphase nuclei was approximately eightfold in Cameron et al. (1979) and 3.6-fold using SIMS (Table I). However, we determined that the total cellular Ca2+ and Mg2+ concentrations between interphase and mitosis changed only minimally. We propose that during the cell cycle an intracellular redistribution of Ca2+ and Mg2+ occurs, but with no major cellular influx/efflux of these cations.
Our main research interests have focused on cation–DNA and cation–protein interactions in terms of their roles in higher order structure. Therefore, the chromatin association of Na+ and K+ supports important roles of these cations in both interphase and mitotic chromatin compaction, whereas Ca2+ and Mg2+ binding points to an essential function in mitotic chromatin compaction. Consequently, the depletion of Ca2+ and Mg2+ mitotic chromosomes, which resulted in partly decondensed structures, a process which is reversible (Zelenin et al., 1982; Earnshaw and Laemmli, 1983; Staron, 1985), points to key roles for Ca2+ and Mg2+ in specific cation–chromatin binding and in the dynamics of chromatin condensation and decondensation. DNA condensation is a multimolecular, highly cooperative and delicately balanced process, which occurs in a rapid time span during each cell cycle. In the presence of cations, DNA condensation is determined by charge neutralization and not by binding to DNA per se as determined by calculations of electrostatic forces (for review see Bloomfield, 1998). The binding of cations specifically to the DNA phosphates results in decreasing the overall electrostatic Coulomb repulsion between free phosphates and adjacent DNA structures. For example, the DNA neutralization fraction of core histones was calculated at 57% by circular dichroism of nucleosome cores (Morgan et al., 1987, and references therein). Therefore, the remaining 43% of the DNA net negative charge must be neutralized by histone H1, the nonhistone proteins, polyamines, like spermine4+, spermidine3+, and putrescine2+, and especially Ca2+, Mg2+, Na+, and K+. As shown for DNA oligomer crystals (Minasov et al., 1999), the DNA charge neutralization of Ca2+, Mg2+, Na+, and K+ together resulted in a greater DNA radius reduction. The effect of Ca2+ and Mg2+ on DNA compaction and overwinding has been shown with supercoiled DNA (Adrian et al., 1990) as well as with helical DNA (Xu and Bremer, 1997). The potential of cation binding for chromosome condensation and maintenance can be seen with histone-free dinoflagellate chromosomes, which are exclusively compacted and stabilized by Ca2+ and Mg2+, at two different binding sites (Herzog and Soyer, 1983). For naked DNA, the maximal binding and compaction for Ca2+ and Mg2+ was found to be 0.63 cations/bp (or one every 3.17 nucleotides) (Koltover et al., 2000). From our total Ca2+ and Mg2+ chromosome concentration values, we determined that one Ca2+ binds to every 12.5–20 nucleotides (1–2 helical turns) and one Mg2+ to every 20–33 nucleotides (2–3 helical turns). The discrepancy between naked DNA and chromatin can be explained due to the occupation of cation binding sites with Na+, K+, and chromatin binding proteins. In addition, even after incubating IM cells with 10 μM of the Ca2+ specific ionophore A23187 for 6 h, we detected no further increase in concentration levels of Ca2+ or other cations tested for binding on mitotic chromosomes (unpublished data). This finding supports the notion that metaphase chromosomes have no additional binding sites for Ca2+ or other cations. Therefore, we conclude that the detected Ca2+, Mg2+, Na+, and K+ cations together with polyamines, histones, and nonhistone proteins result in charge neutral mitotic chromosomes and represent the highest compacted state. During cation neutralization, this process may lead to modifications in local DNA structures, like bending toward the neutralized region, and thus facilitate nucleosome folding. Subsequently, the charge neutralization of chromosomes may be necessary to facilitate free chromosomal movement throughout mitosis.
In addition to the overall chromosome binding of Ca2+, Mg2+, Na+, and K+, we also detected that Ca2+ specifically binds to the chromosome axis, the location of the scaffold proteins and the highly AT-rich SARs, in a 3:1 Ca2+/Mg2+ ratio. Different DNA binding properties of Ca2+ and Mg2+ have been shown in DNA oligomer crystals where both divalent cations bound to the DNA phosphate groups over oxygen atoms; but in contrast to Mg2+, Ca2+ was specifically found in the minor groove of AT-rich DNA (Minasov et al., 1999). Recent NMR studies of DNA oligomeres in the presence of K+ demonstrated a stabilization of quadruplex DNA structures as implicated at the telomeres and centromeres (Marathias and Bolton, 2000). The K+ nucleolar enrichment we observed may reflect the centromeric regions mapping adjacent to the nucleolar organizing regions. Lewis and Laemmli (1982) proposed that Cu2+ or Ca2+ is needed for stabilization of chromosome scaffolding proteins, including Topo II (ScI) and ScII. It has previously been shown that 10% of the total nuclear Ca2+ is bound to the nonhistone–scaffold protein fraction (Schibeci and Martonosi, 1980). Our direct binding experiments of Ca2+ with Topo II supports that Topo II is the main Ca2+-binding protein in the scaffold fraction (Fig. 8 B). Since we demonstrated that Ca2+ and not Cu2+ is enriched on the chromosomal axis using SIMS, we propose that Ca2+ is the most likely candidate for stabilization of the chromosomal scaffolding proteins, particularly Topo II. The finding of Lewis and Laemmli (1982) concerning the role of Cu2+ can be explained in that, in contrast to Ca2+, Cu2+ can bind multiple oxygen atoms of proteins with the result of oxidizing the peptide backbone and being reduced to Cu+ (Legler et al., 1985). The fact that Ca2+ also widened the minor groove, whereas Mg2+ contracted it (Minasov et al., 1999), may also have important implications for the structure and function of SARs at the chromosome axes as well as for axis-binding proteins like Topo II and ScII.
Topo II and ScII proteins comprise ∼40% of the overall proteins in the chromosome scaffold (Earnshaw and Laemmli, 1983), and are involved in chromosome structure maintenance at the chromosome axes (Boy de la Tour and Laemmli, 1988; Saitoh et al., 1994). Andreassen et al. (1997) colocalized Topo II and ScII only after prophase on the axes, supporting that the mitotic ScII may be present in two complexes, with the condensins and with Topo II. We showed that Topo II directly binds Ca2+ without binding DNA, whereas chromosomal ScII did not (Fig. 8 B), although Ca2+ chelation experiments depleted Topo II and ScII from mitotic chromosomes (Fig. 7). This result could be explained that Topo II binds to both Ca2+ and ScII and after chelation of Ca2+ both proteins are depleted due to possible strong protein–protein binding (Lewis and Laemmli, 1982; Ma et al., 1993). We observed that over 80–90% chelation of chromosome bound Ca2+ and Mg2+ resulted in partially decondensed chromosomes and, more importantly, in a 90–95% loss of Topo II and ScII, but in no loss of hCAP-C and histones. These partially decondensed structures are most likely due to both the loss of compaction and neutralization by Ca2+ and Mg2+, as well as by the simultaneous depletion of Ca2+/Mg2+-bound Topo II. A complete chromosome collapse was not detected because of the remaining Na+ and K+, histones, and other nonhistone proteins, like hCAP-C as well as DNA–DNA interactions.
Finally, we found that the 3:1 chromosome axis ratio of Ca2+/Mg2+ fully inhibited Topo II catalytic activity in vitro (Fig. 8 A). In Ca2+ and Mg2+ competition studies by us and Osheroff and Zechiedrich (1987), results showed that an interaction between Ca2+ and Topo II led to inactivation of the catalytic activity by trapping Topo II onto DNA in a stabilized cleavage complex. It is reasonable that the recent findings of catalytically inactive Topo II on chromosomes at metaphase and especially anaphase (Shamu and Murray, 1992; Meyer et al., 1997; Bojanowski et al., 1998) may be explained by Ca2+ binding, and that the actual mechanism involving Ca2+-induced Topo II enzymatic inactivation could be due to the larger ionic radii of Ca2+ (0.99A) as compared with Mg2+ (0.66 A), thus altering the tertiary structure of Topo II and converting it to a solely structural DNA binding protein. If Topo II has two (or more) different binding sites for Ca2+ and Mg2+ like DNase A (Poulos and Price, 1972) or only one site is still unresolved. Interestingly, in the presence of Ca2+, DNase A had different DNA cleavage specificity and changes of the protein structure protecting the enzyme against proteolysis (Poulos and Price, 1972). It will be important to determine if the enzymatic activity of Topo II is inhibited after binding of Ca2+ due to the changing of protein structure.
The G2 checkpoint of mammalian cells requires Topo II–dependent decatenation of DNA duplexes before entry in mitosis (Downes et al., 1994). Topo II specific drugs, like VP16 given in G1 or S phase inhibit Topo II, resulting in undecatenated DNA and G2 arrest (Tobey et al., 1990). Ca2+ binds Topo II and DNA in a covalent complex, but leaves Topo II kinetically active (Osheroff and Zechiedrich, 1987). In contrast to VP16, a depletion of Ca2+ at the chromatin due to the redistribution during the cell cycle could result in active Topo II, religating DNA cleavages, and thus, explain the lack of DNA breaks during/after mitosis. After prophase, the cells are insensitive to Topo II inhibitors (Rowley and Kort, 1989), presumably because chromosomal Topo II is complexed with Ca2+ and DNA as stable cleavage complexes.
In conclusion, this investigation identifies the cations Ca2+, Mg2+, Na+, and K+ as essential participants in the maintenance of higher order structure in mammalian chromosomes particularly at mitosis due to their functions in (a) DNA electrostatic neutralization and chromosome condensation, (b) a direct interaction of Ca2+ with Topo II, and (c) regulation of Topo II as a structural chromosomal binding protein through cation–protein interactions with Ca2+ at the chromosomal axis. The distribution of these chromatin-binding cations and the scaffold proteins are represented in our chromosome model (Fig. 9). Thus, the cations Ca2+, Mg2+, Na+, and K+ in addition to polyamines, histones, and nonhistone proteins are pivotal to complete and maintain “maximal chromosome condensation” during mitosis.
Materials And Methods
Scanning ion microprobe
In the University of Chicago scanning ion microprobe (Chabala et al., 1995), a 30-pA beam of gallium ions extracted from a liquid metal ion source, typically accelerated to 45 keV and focused to a spot ∼50 nm in diameter, is rastered over a specimen to erode the outer surface layers. The sputtered ionized atoms or molecular clusters are discriminated on the basis of their mass/charge ratio with a high performance magnetic sector mass spectrometer (Finnigan MAT 90). Both positive and negative ion species can be mass analyzed. By recording the secondary ion signal counts, detected by an active film electron multiplier (ETP AF820), as a function of the position of the scanning beam, two-dimensional compositional distribution (SIMS) maps are obtained. The sputter erosion depth during the acquisition of one analytical image (map), inverse function of the magnification, can be controlled over a wide range from a few atomic monolayers to tens of nanometers. A detector overlooking the sample collects secondary ions yielding topographic (ISI) images similar to those generated by a scanning electron microscope. The detection sensitivity in analytical images can reach the ppm range, due to the high transmission efficiency (∼20%) of the SIMS system. The digital images, containing 512 × 512 picture elements from single square raster scans are analyzed with a KONTRON IMCO image processing system.
Cell lines and culture
Male IM deer primary fibroblast cells (American Type Culture Collection) and the BV173 human leukemia T/B-progenitor cell line (gift from Dr. J.D. Rowley, University of Chicago, Chicago, IL) were grown in log phase from previously seeded cell cultures at 0.07 × 106 cells/ml in F10 media (Life Technologies), 10% FCS, or 0.5 × 106 cells/ml in RPMI (Life Technologies) with 10% FCS, respectively.
Cryopreservation and fracturing of cells
The cryopreservation was performed according to Chandra et al. (1986) with modifications. IM cells were grown on 99.99% pure silicon substrate discs, or fractionated mitotic cells were layered onto silicon discs and immediately dipped into a liquid N2 plus 1,1,1-trichloroethane slush bath (−150°C) at a speed of 2 m/s to a depth of 5 cm for 10 s. Cryofracturing of cells was obtained by splitting open a sandwich of two discs. The frozen discs were then transferred under N2/1,1,1-trichloroethane into a –65°C vacuum chamber for 16 h at a pressure of 0.5 millitorr and then brought to room temperature in 2 h under vacuum.
Cell synchronization
IM and BV173 cells were synchronized at the G1/S phase border using aphidicolin (20 μM) for 16 h. The block was then released for 3 h, and 0.3μM nocodazole was added for an additional 18 h. Mitotic cells were collected in the supernatant fraction, and DAPI staining showed >98% pro- or metaphase cells with only 1–2% interphase cells. For experiments involving chelators, we first synchronized IM cells to the G1/S border and then added 0.2 μM nocodazole for 18 h. Using this approach, >80% of the total cells were synchronized to the G2/M border, as determined by optical microscopy and DAPI staining. After releasing the G2/M block, either EGTA (5 mM) or EDTA (5 mM) or BAPTA (0.3 mM) and BAPTA-AM (50 μM) was added for 6 h, and mitotic chromosomes were harvested using the chromosome isolation methods described below for comparison.
Chromosome harvests
Two chromosome harvesting procedures were used for SIMS and IF analysis. In one method, chromosomes were extracted from mitotic cells and then fixed in methanol: acetic acid (3:1) according to Ohnuki (1968). This extraction method could lead to artifacts by contaminating chromosomes with cytosolic cations. Therefore, individual IM and BV173 metaphase chromosomes were also fractionated from cells that preserve the morphological integrity of chromosomes as shown by Saitoh and Laemmli (1994). Briefly, after cells were synchronized at mitosis, chromosomes were isolated in Ca2+- and Mg2+-free buffers using a glycerol step-gradient and then fixated with 4% p-formaldehyde. For SIMS analysis, chromosome samples were mounted onto alumina ceramic coverslips, which are devoid of innate K+, Na+, Ca2+, and Mg2+ ions, previously lightly Au coated to ensure substrate conductivity, and then air-dried. The samples were further coated with a thin sputter-deposited layer of gold to prevent electrical charging.
Immunofluorescence
IM and BV173 fractionated chromosomes were fixed with 4% p-formaldehyde and incubated with anti–Topo II monoclonal (Boehringer), anti-ScII polyclonal (Saitoh et al., 1994), anti-H1 monoclonal (Biodesign), and anti–hCAP-C polyclonal (Schmiesing et al., 1998) antibodies, diluted 1:200, 1:50, 1:30, and 1:100, respectively, with 3% BSA in PBS. Rhodamine-conjugated anti–mouse or anti–rabbit secondary antibodies (Boehringer) were used in a dilution of 1:150. The chromosomes were analyzed with a ZEISS Axioplan microscope combined with a digital CCD camera.
Quantitation of Ca2+ and Mg2+ on chromosomes
We established a Ca2+ and Mg2+ reference for SIMS using different concentrations of Ca2+ and Mg2+ mixed with high purity agarose (Seakem Gold, FMC) as a matrix. We also established Cu2+ references for SIMS. Agarose was chosen as a carrier because of the comparable chemical structure to DNA (parallel double helix with left-handed symmetry, tightly bound water, and similar relative density to DNA, which we determined to be equal 1.869 kg/l). Drying of agarose, necessary for SIMS analysis, has only minimal impact on structure (Arndt and Stevens, 1994). The standards were solubilized in bidistilled water, deposited on Au-coated glass coverslips, vacuum dried, Au coated, and mass analyzed for Ca2+ and Mg2+ using SIMS. Control agarose samples were also SIMS analyzed, indicating negligible Ca2+ and Mg2+ background. SIMS sensitivity was greater than 10−8 M for 40Ca2+ and greater than 10−7 M for 24Mg2+ and 63Cu2+. Calibration plots of SIMS signal intensity (cts/pxl) corresponding to different cation concentrations are shown in Fig. 4, G and H. The number of metal atoms/nucleotide was obtained by multiplying the measured local metal atomic concentration (in ppm) by the average number of atoms/nucleotide (taken as 36) divided by 106.
Fluorescent filter-binding assay for the detection of Ca2+-binding proteins
The assay was performed according to Tatsumi et al. (1997), except using MgCl2 in the washing buffer. Protein marker and Calmodulin (Sigma-Aldrich), purified human Topo II (TopoGen), and fractionated IM chromosomes were electrophoresed on a 7.5–15% SDS gradient gel and transferred onto PVDF membrane (Bio-Rad Laboratories). After incubating the membrane with 1 mM CaCl2 and then with quin-2 (Sigma-Aldrich) for 1 h, the fluorescent proteins (Ca2+-binding proteins) were visualized by illumination with UV light at 365 nm, digitally photographed, and analyzed with the Kodak 1D Imaging system.
Topo II relaxation reaction
0.2 μg of supercoiled plasmid pSP72 (Promega) DNA was incubated at 30°C for 5–20 min in the presence of 1 U of purified human Topo II (TopoGen) and Ca2+ and Mg2+ in different ratios in a Topo II relaxation buffer (Osheroff and Zechiedrich, 1987). The reaction was stopped with 1% SDS and 15 mM EDTA, and the DNA phenol/chloroform was extracted, ethanol precipitated, and analyzed on a 1.5% agarose gel. The Topo II relaxation activity was quantified using the ImageQuant analysis program (Molecular Dynamics).
Acknowledgments
We wish to thank Dr. J.D. Rowley for her support and L. Gavrilov for expert technical assistance. We are grateful to Dr. W.C. Earnshaw for providing the ScII antibodies and to Dr. K. Yokomori for the hCAP-C antibodies.
At the Enrico Fermi Institute (Chicago, IL), this work was supported by a grant from the Pritzker Foundation.
R. Strick and P. Strissel contributed equally to this work.
Abbreviations used in this paper: 3D, three-dimensional; AT, ademine-thymine; BAPTA, 1,2-bis[o-aminophenoxy]ethane-N,N,N′,N′-tetraacetic acid; IM, Indian muntjac; ISI, ion-induced secondary ion; PVDF, polyvinyldifluoride; SAR, scaffold-associated region; ScII, scaffold protein II; SIMS, secondary ion mass spectrometry; Topo II, topoisomerase IIα.