Low angle x-ray diffraction studies of HeLa metaphase chromosomes: effects of histone phosphorylation and chromosome isolation procedure.

To test whether gross changes in chromatin structure occur during the cell cycle, we compared HeLa mitotic metaphase chromosomes and interphase nuclei by low angle x-ray diffraction. Interphase nuclei and metaphase chromosomes differ only in the 30-40-nm packing reflection, but not in the higher angle part of the x-ray diffraction pattern. Our interpretation of these results is that the transition to metaphase affects only the packing of chromatin fibers and not, to the resolution of our method, the internal structure of nucleosomes or the pattern of nucleosome packing within chromatin fibers. In particular, phosphorylation of histones H1 and H3 at mitosis does not affect chromatin fiber structure, since the same x-ray results are obtained whether or not histone dephosphorylation is prevented by isolating metaphase chromosomes in the presence of 5,5'-dithiobis(2-nitrobenzoate) or low concentrations of p-chloromercuriphenylsulfonate (ClHgPhSO3). We also compared metaphase chromosomes isolated by several different published procedures, and found that the isolation procedure can significantly affect the x-ray diffraction pattern. High concentrations of ClHgPhSO3 can also profoundly affect the pattern.

Low angle x-ray diffraction studies of chromatin structure in vivo and in isolated nuclei and metaphase chromosomes.

Diffraction of x-rays from living cells, isolated nuclei, and metaphase chromosomes gives rise to several major low angle reflections characteristic of a highly conserved pattern of nucleosome packing within the chromatin fibers. We answer three questions about the x-ray data: Which reflections are characteristic of chromosomes in vivo? How can these reflections be preserved in vitro? What chromosome structures give rise to the reflections? Our consistent observation of diffraction peaks at 11.0, 6.0, 3.8, 2.7 and 2.1 nm from a variety of living cells, isolated nuclei, and metaphase chromosomes establishes these periodicities as characteristic of eukaryotic chromosomes in vivo. In addition, a 30-40- nm peak is observed from all somatic cells that have substantial amounts of condensed chromatin, and a weak 18-nm reflection is observed from nucleated erythrocytes. These observations provide a standard for judging the structural integrity of isolated nuclei, chromosomes, and chromatin, and thus resolve long standing controversy about the "tru" nature of chromosome diffraction. All of the reflection seen in vivo can be preserved in vitro provided that the proper ionic conditions are maintained. Our results show clearly that the 30-40-nm maximum is a packing reflection. The packing we observe in vivo is directly correlated to the side-by-side arrangement of 20- 30-nm fibers observed in thin sections of fixed and dehydrated cells and isolated chromosomes. This confirms that such packing is present in living cells and is not merely an artifact of electron microscopy. As expected, the packing reflection is shifted to longer spacings when the fibers are spread apart by reducing the concentration of divalent cations in vitro. Because the 18-, 11.0-, 6.0-, 3.8-, 2.7-, and 2.1-nm reflections are not affected by the decondensation caused by removal of divalent cations, these periodicities must reflect the internal structure of the chromaticn fibers.

Radial density distribution of chromatin: evidence that chromatin fibers have solid centers.

Fiber diameter, radial distribution of density, and radius of gyration were determined from scanning transmission electron microscopy (STEM) of unstained, frozen-dried chromatin fibers. Chromatin fibers isolated under physiological conditions (ionic strength, 124 mM) from Thyone briareus sperm (DNA linker length, n = 87 bp) and Necturus maculosus erythrocytes (n = 48 bp) were analyzed by objective image-processing techniques. The mean outer diameters were determined to be 38.0 nm (SD = 3.7 nm; SEM = 0.36 nm) and 31.2 nm (SD = 3.6 nm; SEM = 0.32 nm) for Thyone and Necturus, respectively. These data are inconsistent with the twisted-ribbon and solenoid models, which predict constant diameters of approximately 30 nm, independent of DNA linker length. Calculated radial density distributions of chromatin exhibited relatively uniform density with no central hole, although the 4-nm hole in tobacco mosaic virus (TMV) from the same micrographs was visualized clearly. The existence of density at the center of chromatin fibers is in strong disagreement with the hollow-solenoid and hollow-twisted-ribbon models, which predict central holes of 16 and 9 nm for chromatin of 38 and 31 nm diameter, respectively. The cross-sectional radii of gyration were calculated from the radial density distributions and found to be 13.6 nm for Thyone and 11.1 nm for Necturus, in good agreement with x-ray and neutron scattering. The STEM data do not support the solenoid or twisted-ribbon models for chromatin fiber structure. They do, however, support the double-helical crossed-linker models, which exhibit a strong dependence of fiber diameter upon DNA linker length and have linker DNA at the center.

The diameters of frozen-hydrated chromatin fibers increase with DNA linker length: evidence in support of variable diameter models for chromatin.

The diameters of chromatin fibers from Thyone briareus (sea cucumber) sperm (DNA linker length, n = 87 bp) and Necturus maculosus (mudpuppy) erythrocytes (n = 48 bp) were investigated. Soluble fibers were frozen into vitrified aqueous solutions of physiological ionic strength (124 mM), imaged by cryo-EM, and measured interactively using quantitative computer image-processing techniques. Frozen-hydrated Thyone and Necturus fibers had significantly different mean diameters of 43.5 nm (SD = 4.2 nm; SEM = 0.61 nm) and 32.0 nm (SD = 3.0 nm; SEM = 0.36 nm), respectively. Evaluation of previously published EM data shows that the diameters of chromatin from a large number of sources are proportional to linker length. In addition, the inherent variability in fiber diameter suggests a relationship between fiber structure and the heterogeneity of linker length. The cryo-EM data were in quantitative agreement with space-filling double-helical crossed-linker models of Thyone and Necturus chromatin. The data, however, do not support solenoid or twisted-ribbon models for chromatin that specify a constant 30 nm diameter. To reconcile the concept of solenoidal packing with the data, we propose a variable-diameter solid-solenoid model with a fiber diameter that increases with linker length. In principle, each of the variable diameter models for chromatin can be reconciled with local variations in linker length.

Crosslinking of proteins to DNA in human nuclei using a 60 femtosecond 266 nm laser.

We developed appropriate conditions to use a laser with 60 femtosecond pulses, a frequency of 1 KHz and a wavelength of 266 nm to efficiently crosslink proteins to DNA in human nuclei for the purpose of using immunoprecipitation to study the binding of specific proteins to specific sequences of DNA under native conditions. Irradiation of nuclei for 30 min with 1-3 GW/cm(2)pulses crosslinked 10-12% of total protein to DNA. The efficiency of crosslinking was dose and protein specific. Histones H1 and H3 were crosslinked by 15 min of irradiation with 20-25% efficiency, at least 10 times more strongly than the other histones, consistent with experiments using conventional UV light. Irradiation for 15 min did not damage proteins, as assayed by SDS-PAGE of Ku-70 and histones. Although the same level of irradiation did not cause double-strand breaks, it did make the DNA partially insensitive to Eco RI restriction enzyme, probably through formation of thymidine dimers. Immuno-analysis of crosslinked nucleoprotein showed that Ku crosslinking to nuclear DNA is detectable only in the presence of breaks in the DNA, and that nucleosomes are bound to a significant fraction of the telomeric repeat (TTAGGG) (n).

Psoralen crosslinking of active and inactive sea urchin histone and rRNA genes.

Chromatin structure is highly correlated with the transcriptional activity of specific genes. For example, it has been found that the regularity of nucleosome spacing is compromised when genes are transcribed. The rRNA genes from fungi, plants, and animals give distinctly bimodal distributions of psoralen crosslinking, which has led to the suggestion that these genes might be largely devoid of nucleosomes when transcriptionally active. We investigated the chromatin structure of the multicopy rRNA and histone genes during sea urchin early embryogenesis. The rRNA genes, which are weakly expressed, give a unimodal distribution of weak psoralen crosslinking, in contrast to the situation in all other organisms studied. The early histone genes were more accessible to psoralen crosslinking when active than inactive. The pattern of crosslinking suggests that these polII genes have a homogeneous structure and are still highly protected by nucleosomes when in the active conformation, unlike the situation in polI genes.

The major component of a large, intracellular proteinase accumulated by inhibitors is a complex of alpha 2-macroglobulin and thrombin.

A large, intracellular proteinase accumulated by inhibitors (PABI) was found in cultured mammalian cells as a large, multicatalytic proteinase with a greatly elevated concentration in the presence of small peptide proteinase inhibitors (Tsuji and Kurachi (1989) J. Biol. Chem. 264, 16093). Electron microscopic analysis showed that the tertiary structure of PABI highly resembled that of alpha 2-macroglobulin complexed with a proteinase(s). Isolation of the anti-PABI cross-reacting material from calf serum added to the culture media of baby hamster kidney cells further supported that the primary component of PABI was alpha 2-macroglobulin. Immunoblot analyses and the substrate specificity of PABI indicated that the major proteinase component contained in PABI was thrombin. When alpha 2-macroglobulin was added to the PABI-depleted serum, a significant accumulation or a degradation of the intracellular alpha 2-macroglobulin was observed in the presence or absence of leupeptin, respectively. Similarly, when thrombin was added to the PABI-depleted fetal calf serum supplemented with fresh alpha 2-macroglobulin, a significant amount of intracellular thrombin was found only in the presence of leupeptin. These results indicate that the major component of the intracellular PABI molecules is a complex of alpha 2-macroglobulin with thrombin which is internalized from the culture media. Intracellular accumulation of PABI, therefore, is a phenomenon primarily relevant to the culture cells. Whether or not PABI is also generated in certain physiological or pathological conditions requires further study.

Quantitation of molecular densities by cryo-electron microscopy. Determination of the radial density distribution of tobacco mosaic virus.

We have determined the absolute mass and radial scattering density distribution of tobacco mosaic virus in the frozen-hydrated state by energy-filtered low-dose bright-field transmission electron microscopy. The absolute magnitude of electron scattering from tobacco mosaic virus in 150 nm of ice was within 3.0% of that predicted, with inelastic scattering accounting for approximately 80% of the scattering contrast. In order to test the accuracy of the radial reconstruction, a computer model of tobacco mosaic virus was built from the atomic co-ordinates assuming uniform solvent density. The validity of the model was confirmed by comparison of X-ray scattering and predictions of the model (R factor = 0.05). First-order corrections for the microscope contrast transfer function were necessary and sufficient for conversion of the cryo-electron microscopy images into accurate representations of the mass density. At 1.9 nm resolution the compensated reconstruction and model had density peaks of similar magnitude at 2.4, 4.2, 6.0 and 7.8 nm radius and a central hole of 2 nm radius. Equatorial Fourier transforms of the corrected electron images were in excellent agreement with predictions of the model (R factor = 0.12). Thus, the uniform solvent approximation was adequate at 1.9 nm resolution to describe quantitatively X-ray scattering in liquid water and electron imaging in vitreous ice. This is the first demonstration that cryo-electron microscopy images can be used to quantitate the absolute mass, mass per unit length and internal density distributions of proteins and nucleic acids.

A polarized photobleaching study of chromatin reorientation in intact nuclei.

Polarized fluorescence recovery after photobleaching (pFRAP) was used to monitor the effects that condensation, i.e. compaction and aggregation, have on the (microseconds and ms) internal dynamics of chromatin in intact nuclei. When divalent cations were present with physiological (approximately 90 mM) monovalent salt the chromatin was found to exist in a compact and aggregated state which was characterized by rotational immobilization over timescales that range from 10 microseconds to 40 milliseconds. This immobilization is attributed to suppression of internal dynamics by intermolecular interactions. When the divalent cations were removed, the compact fibers no longer aggregated and were free to reorient with a characteristic decay time of about 1.2 milliseconds. It is shown that this millisecond relaxation could represent rigid rotation of topologically independent structural domains. Dilution of the monovalent salt induced a gradual change in the structural state of the chromatin that was manifest as a dramatic increase in internal flexibility. At the lowest salt concentration studied (11 mM-monovalent salt) the chromatin reorients in fewer than ten microseconds. These changes in flexibility are continuous with salt concentration, indicating that there are no well-defined endpoints to structural transitions and that the microsecond-millisecond internal dynamics of chromatin are a sensitive measure of structure. Measurements made on nuclei from cells that are either transcriptionally quiescent or active indicate that the dynamics mirrors biological activity.

Quantitative energy-filtered electron microscopy of biological molecules in ice.

The theoretical and experimental bases for quantitative electron microscopy of frozen-hydrated specimens are described, with special considerations of energy filtration to improve the images. The elastic and inelastic scattering from molecules in vacuum and in ice are calculated, and simple methods to approximate scattering are introduced. Multiple scattering calculations are used to describe the scattering from vitreous ice and to predict the characteristics of images of frozen-hydrated molecules as a function of ice thickness and accelerating voltage. Energy filtration is predicted to improve image contrast and signal-to-noise ratio. Experimental values for the inelastic scattering of ice, the energy spectrum of thick ice, and the contrast of biological specimens are determined. The principles of compensation for the contrast transfer function are presented. Tobacco mosaic virus is used to quantify the accuracy of interpreting image intensities to derive the absolute mass, mass per unit length, and internal mass densities of biological molecules. It is shown that compensation for the contrast transfer function is necessary and sufficient to convert the images into accurate representations of molecular density. At a resolution of 2 nm, the radial density reconstructions of tobacco mosaic virus are in quantitative agreement with the atomic model derived from X-ray results.

Implementation of an NSOM system for fluorescence microscopy.

We describe our progress toward an NSOM system intended for fluorescence imaging of biological samples. This process included integration of shear-force feedback into an existing NSOM system. Topographic images acquired using uncoated tips are presented. We also present our initial effort at simultaneous acquisition of topographic and fluorescence data using an aluminum coated tip.

Small angle x-ray scattering of chromatin. Radius and mass per unit length depend on linker length.

Analyses of low angle x-ray scattering from chromatin, isolated by identical procedures but from different species, indicate that fiber diameter and number of nucleosomes per unit length increase with the amount of nucleosome linker DNA. Experiments were conducted at physiological ionic strength to obtain parameters reflecting the structure most likely present in living cells. Guinier analyses were performed on scattering from solutions of soluble chromatin from Necturus maculosus erythrocytes (linker length 48 bp), chicken erythrocytes (linker length 64 bp), and Thyone briareus sperm (linker length 87 bp). The results were extrapolated to infinite dilution to eliminate interparticle contributions to the scattering. Cross-sectional radii of gyration were found to be 10.9 +/- 0.5, 12.1 +/- 0.4, and 15.9 +/- 0.5 nm for Necturus, chicken, and Thyone chromatin, respectively, which are consistent with fiber diameters of 30.8, 34.2, and 45.0 nm. Mass per unit lengths were found to be 6.9 +/- 0.5, 8.3 +/- 0.6, and 11.8 +/- 1.4 nucleosomes per 10 nm for Necturus, chicken, and Thyone chromatin, respectively. The geometrical consequences of the experimental mass per unit lengths and radii of gyration are consistent with a conserved interaction among nucleosomes. Cross-linking agents were found to have little effect on fiber external geometry, but significant effect on internal structure. The absolute values of fiber diameter and mass per unit length, and their dependencies upon linker length agree with the predictions of the double-helical crossed-linker model. A compilation of all published x-ray scattering data from the last decade indicates that the relationship between chromatin structure and linker length is consistent with data obtained by other investigators.

The higher order structure of chicken erythrocyte chromosomes in vivo.

Recently eukaryotic chromosomes have been shown to consist of a repeating subunit, called the nucleosome. Although electron microscopy, neutron scattering and X-ray diffraction have been used to determine the low resolution structure of the nucleosome, these techniques have yielded little information about the disposition of nucleosomes within chromosomes. Electron microscopy has produced many models for chromosome structure based on uniform fibres of 150-500A diameter or on globular 'superbeads. Unfortunately the models are based on microscope images that fail to reveal the strong structural periodicities shown by X-ray scattering to be characteristic of isolated chromatin in solution. Moreover it has not been demonstrated that the chromosomes of living cells are composed of such fibres. We have used low-angle X-ray scattering to investigate the organization of chromosomes in vivo and to account for the previously observed inconsistencies in many X-ray and electron microscope observations. We report here that chicken erythrocytes have a 400 A periodicity due to a nuclear structure that is directly related to the 300 A side-by-side packing of chromosome fibres revealed by electron microscopy of embedded cells, and that this periodicity can be preserved in isolated nuclei provided that the proper buffers are used.

Chromatin architecture: investigation of a subunit of chromatin by dark field electron microscopy.

Dark field scanning electron microscopy of unstained, unfixed samples of chromatin, histone-1-depleted chromatin, and nucleohistone has been used to identify an apparent subunit of chromatin, namely a disk-shaped structure we term the unit particle, which is probably about 135 A wide and 50 A thick in the hydrated state. The unit particles are found at rather uniform intervals along thin DNA-like fibers. Histone 1 depletion leads to a bimodal distribution of these spacings. Our observations suggest that the unit particle consists of a loop of nucleoprotein, perhaps around a histone core.

Efficient solubilization and partial purification of sea urchin histone genes as chromatin.

Soluble chromatin fragments are rapidly and efficiently produced when nuclei are digested with restriction endonucleases in buffers containing very low concentrations of magnesium. Under these conditions, the sequence specificity of the restriction endonucleases is maintained, resulting in release of specific genes as fragments with discrete molecular weights that can be fractionated by size on glycerol gradients. Gradient fractions can be chosen to be significantly enriched in specific genes and their associated proteins. For instance, we can achieve a 16-fold enrichment of the chromatin containing the early histone genes of sea urchin. The enrichments produced by these methods are useful as a first step in techniques to purify specific genes as chromatin. Glycerol gradient analyses can also be used to test whether putative gene-specific proteins are actually bound to the same sequences in vivo.

Nucleoprotein hybridization: a method for isolating specific genes as high molecular weight chromatin.

We describe a new technique designed to isolate specific eukaryotic genes as native oligonucleosome fragments. The isolation method consists of hybridization of single-stranded termini of chromatin restriction fragments to complementary mercurated DNA probes, followed by isolation of the hybrids by sulfhydryl-Sepharose chromatography. SV40 minichromosomes were used to test the effectiveness of the technique. About 80% of KpnI- or BamHI-restricted and lambda exonuclease treated SV40 minichromosomes hybridized to an appropriate DNA probe after a 12-h hybridization reaction under mild conditions (0.1 M aqueous salt, 37 degrees C, pH 8). When the restricted minichromosomes were mixed with a 15-fold excess of "background" chromatin from sea urchin embryos, nucleoprotein hybridization was able to reisolate the SV40 chromatin to 88% purity with a 63% yield. This represented a 115-fold enrichment of specific genes as chromatin. Results of electron microscopy and polyacrylamide gel electrophoresis indicate that the hybridized SV40 chromatin has not lost the major chromosomal proteins characteristic of SV40 nor acquired significant amounts of protein due to exchange with background chromatin. Our experimental results show that it is currently possible to isolate repeated genes from higher eukaryotes for structural and biochemical study of the proteins involved with gene regulation.

Embryonic regulation of histone ubiquitination in the sea urchin.

We have used quantitative 2-D protein electrophoresis and immunoprecipitation to study the patterns of histone ubiquitination at 10 h and 36 h of embryonic development in Strongylocentrotus purpuratus. Variants csH2A, alpha H2A, beta H2A, gamma H2A, delta HA, H2AF./Z, alpha H2B, beta H2B, and gamma H2B showed up to sevenfold differences in level of monoubiquitination between variants, and individual variants showed up to sixfold changes during development. At 36 h of embryogenesis, the late variants were less ubiquitinated than the early variants, although the overall level of ubiquitination was appreciably greater than at 10 h. Antiubiquitin antibodies were used to precipitate formaldehyde-fixed chromatin fragments in order to estimate the degree of ubiquitination of the early histone genes. The 5' regulatory region of the active H3 gene appeared to be at least twice as ubiquitinated as the adjacent upstream spacer. However, the absolute level of ubiquitination of the early histone gene repeat seemed to be independent of transcriptional activity. These results show that variant-specific ubiquitination of histones is a part of the developmental program in sea urchin embryos, but is not clearly correlated with transcriptional activity of the early histone genes, except perhaps in the regulatory regions.

Conserved nucleoprotein structure at the ends of vertebrate and invertebrate chromosomes.

Eukaryotic chromosomes terminate with telomeres, nucleoprotein structures that are essential for chromosome stability. Vertebrate telomeres consist of terminal DNA tracts of sequence (TTAGGG)n, which in rat are predominantly organized into nucleosomes regularly spaced by 157 bp. To test the hypothesis that telomeres of other animals have nucleosomes, we compared telomeres from eight vertebrate tissues and cell cultures, as well as two tissues from an invertebrate. All telomeres have substantial tracts of (TTAGGG)n comprising 0.01-0.2% of the genome. All telomeres are long (20-100 kb), except for those of sea urchin, human, and some chicken chromosomes, which are 3-10 kb in length. All of the animal telomeres contained nucleosome arrays, consistent with the original hypothesis. The telomere repeat lengths vary from 151 to 205 bp, seemingly uncorrelated with telomere size, regularity of nucleosome spacing, species, or state of differentiation but surprisingly correlated with the repeat of bulk chromatin within the same cells. The telomere nucleosomes were consistently approximately 40 bp smaller than bulk nucleosomes. Thus, animal telomeres have highly conserved sequences and unusually short nucleosomes with cell-specific structure.

DNA methylation pattern changes during development of a sea urchin.

Cytosine methylation of developmentally regulated genes of the sea urchin Strongylocentrotus purpuratus was studied by using restriction-endonuclease digestion and Southern blotting. The single-copy bindin gene, the family of five cytoplasmic actin genes and the 400-fold-repeated set of five early histone genes were mostly unmethylated, but some sites exhibited partial methylation that varied throughout development. This shows that in echinoderms the methylation of DNA is not confined to the non-transcribed portion of the genome, as previously believed [Bird, Tagart & Smith (1979) Cell 17, 889-901], and may play a role in transcriptional regulation.

Long G tails at both ends of human chromosomes suggest a C strand degradation mechanism for telomere shortening.

The chromosomes of lower eukaryotes have short telomeric 3' extensions. Using a primer-extension/nick-translation technique and nondenaturing hybridization, we find long 3' G-rich tails at human chromosome ends in mortal primary fibroblasts, umbilical vein endothelial cells, and leukocytes, as well as in immortalized fibroblasts. For all cells tested, >80% of the telomeres have long G-rich overhangs, averaging 130-210 bases in length, in disagreement with the conventional model for incomplete lagging-strand replication, which predicts overhangs on 50% of the chromosome ends. The observed G tails must exist during most of the cell cycle and probably result from degradation of both chromosome ends. The average lengths of the G tails are quantitatively consistent with the observed rates of human chromosome shortening.