Chromatin Higher Order Structure: Chasing a Mirage?

The question as to how the strands of chromatin are packed into the eukaryotic nucleus has intrigued biologists and biochemists for many years. Although there is ample space in the nucleus for the eukaryotic genome, the enormous length of the DNA molecules requires that they be extensively folded. In addition, the folding must be such as to allow regulated access to certain regions of the genome. The discovery of the nucleosome as a basic organizing unit cast the problem in more defined focus (see Refs. 1van Holde K.E. Chromatin. Springer Verlag, New York1988: 289-343Google Scholar and 2Tsanev R. Russev G. Pashev I. Zlatanova J. Replication and Transcription of Chromatin. CRC Press, Inc., Boca Raton, FL1992: 14-24Google Scholar). Even so, it was realized that the condensation afforded to the DNA by nucleosomes (about a 5:1 compaction ratio) could not account for the many thousand-fold compaction that must exist in vivo. Some kind of higher order folding of the nucleosomal fiber must therefore exist. Many models have been proposed for this folding, most of which predicate a regular helical folding of the fiber. It is the purpose of this review to critically re-examine the evidence concerning this condensed structure of chromatin. We must begin, however, with two caveats. First, most of our evidence concerning chromatin fiber structure has been based on in vitro studies of isolated fragments of chromatin. There is some reason for concern that these may not closely resemble the fibers as they exist in the nucleus (see, for example, Ref. 3Giannasca P.J. Horowitz R.A. Woodcock C.L. J. Cell Sci. 1993; 105: 551-561PubMed Google Scholar). Only a few experiments have been conducted in which chromatin fibers have been examined in situ, with minimum perturbation of the nuclear structure. Second, past experiments have all too often utilized cells that are relatively inactive, for example chicken erythrocytes and invertebrate sperm. The folding of chromatin in such cells may not be typical of eukaryotic cells in general. Such concerns must be kept in mind in evaluating existing data. The first electron microscope observations of chromatin fibers to reveal their nucleosomal structure utilized the low salt spreading technique developed by Miller and Bakken (4Miller Jr., O.L. Bakken A.H. Acta Endocrinol. Suppl. 1972; 168: 155-177Crossref PubMed Google Scholar). Both Woodcock (5Woodcock C.L.F. J. Cell Biol. 1973; 59 (abstr.): 368Google Scholar) and Olins and Olins (6Olins A.L. Olins D.E. J. Cell Biol. 1973; 59 (abstr.): 252Crossref Scopus (48) Google Scholar, 7Olins A.L. Olins D.E. Science. 1974; 183: 330-332Crossref PubMed Scopus (936) Google Scholar) reported regularly spaced beadlike objects along extended DNA filaments. It has since been realized that this method, while ideal for visualizing nucleosomes, seriously distorts the chromatin fiber itself. The first evidence that the fiber might be more compact under some circumstances appears to be found in a study of SV40 minichromosomes by Griffith (8Griffith J.D. Science. 1975; 187: 1202-1203Crossref PubMed Scopus (277) Google Scholar). At low salt Griffith observed a zigzag circular chain of nucleosomes, but at 150 m M NaCl the structure had condensed to yield a smooth fiber of close packed nucleosomes about 10 nm in thickness. Similar structures were reported under a variety of circumstances (see Refs. 9Finch J.T. Noll M. Kornberg R.D. Proc. Natl. Acad. Sci. U. S. A. 1975; 72: 3320-3322Crossref PubMed Scopus (135) Google Scholar and 10Finch J.T. Klug A. Proc. Natl. Acad. Sci. U. S. A. 1976; 73: 1897-1901Crossref PubMed Scopus (952) Google Scholar, for example). However, conditions reported by various laboratories for production of the “10-nm filament” were contradictory, and it is now believed that such a structure is an artifact produced by EM 1The abbreviations used are:EMelectron microscopySFMscanning force microscopybpbase pairs. staining procedures (see Ref. 11Widom J. Annu. Rev. Biophys. Biophys. Chem. 1989; 18: 365-395Crossref PubMed Scopus (146) Google Scholar for a recent evaluation). Nevertheless, the term 10-nm filament still persists in the literature and should be discarded. electron microscopy scanning force microscopy base pairs. A major advance was made when Finch and Klug (10Finch J.T. Klug A. Proc. Natl. Acad. Sci. U. S. A. 1976; 73: 1897-1901Crossref PubMed Scopus (952) Google Scholar) demonstrated that even low concentrations of Mg2+ (0.2 m M) produced condensation of the filaments of nucleosomes into an irregular fiber about 30 nm in diameter. The term “30-nm fiber” has since become associated with the salt-compacted structures. However, we believe that there are now good reasons to substitute for this designation the term “condensed fiber” (see below). The fibers observed by Finch and Klug (10Finch J.T. Klug A. Proc. Natl. Acad. Sci. U. S. A. 1976; 73: 1897-1901Crossref PubMed Scopus (952) Google Scholar) showed occasional diagonal striations, interpreted as evidence for helical structure. Earlier x-ray diffraction studies on chromatin fibers had indicated maxima corresponding to spacings of about 11, 5.5, 3.7, and 2.7 nm. Accordingly, Finch and Klug (10Finch J.T. Klug A. Proc. Natl. Acad. Sci. U. S. A. 1976; 73: 1897-1901Crossref PubMed Scopus (952) Google Scholar) postulated a helical “solenoid” model with a pitch of about 11 nm and suggested that the series of maxima represented successive orders of diffraction from such a helix. The model was simple and seemed to accommodate much of the data available at the time. It took root in the textbooks and continues to be regarded by many as the appropriate description of the condensed fiber. Several major electron microscopic studies appeared in the succeeding years; especially important are the careful analyses by Rattner and Hamkalo (12Rattner J.B. Hamkalo B.A. J. Cell Biol. 1979; 81: 453-457Crossref PubMed Scopus (44) Google Scholar) and Thoma et al. (13Thoma F. Koller T. Klug A. J. Cell Biol. 1979; 83: 403-427Crossref PubMed Scopus (1179) Google Scholar). Thoma et al. described a progressive folding from a zigzag structure at very low ionic strength to the solenoid at NaCl concentration above about 60 m M or at Mg2+ concentration above 0.3 m M. The beads-on-a-string form was considered by these authors to be an artifact of fiber stretching and/or H1 depletion, whereas the 10-nm nucleofilament was viewed as an artifactual consequence of staining conditions. The model of Thoma et al. (13Thoma F. Koller T. Klug A. J. Cell Biol. 1979; 83: 403-427Crossref PubMed Scopus (1179) Google Scholar) for chromatin condensation (folding with increasing ion concentrations from an open zigzag through a closed zigzag to a helical condensed fiber) has dominated thinking in this field for many years. Indeed, it is this picture that helped to generate the rash of alternative models for the condensed fiber that appeared in the 1980s. The solenoid model had fixed the idea that some kind of regular structure must describe the condensed fiber, but there was widespread controversy concerning the details of that structure. During the years immediately following, a considerable number of very specific models for the condensed fiber were put forward. Some of these, such as the models proposed by McGhee et al. (14McGhee J.D. Rau D.C. Charney E. Felsenfeld G. Cell. 1980; 22: 87-96Abstract Full Text PDF PubMed Scopus (182) Google Scholar) and Butler (15Butler P.J.G. EMBO J. 1984; 3: 2599-2604Crossref PubMed Scopus (60) Google Scholar), were essentially refinements of the Finch and Klug (10Finch J.T. Klug A. Proc. Natl. Acad. Sci. U. S. A. 1976; 73: 1897-1901Crossref PubMed Scopus (952) Google Scholar) solenoid, making specific disposition of the linker DNA, a question that Finch and Klug had left open. Others, like those of Worcel et al. (16Worcel A. Strogatz S. Riley D. Proc. Natl. Acad. Sci. U. S. A. 1981; 78: 1461-1465Crossref PubMed Scopus (163) Google Scholar), Staynov (17Staynov D.Z. Int. J. Biol. Macromol. 1981; 5: 310-311Google Scholar), Woodcock et al. (18Woodcock C.L.F. Frado L.-L.Y. Rattner J.B. J. Cell Biol. 1984; 99: 42-52Crossref PubMed Scopus (263) Google Scholar), and Williams et al. (19Williams S.P. Athey B.D. Muglia L.J. Schappe R.S. Gough A.H. Langmore J.P. Biophys. J. 1986; 49: 233-248Abstract Full Text PDF PubMed Scopus (169) Google Scholar) seem to have been inspired by the zigzag chain presumed to exist at lower ionic strength. The first of these represents a twisting of the zigzag chain, Staynov's model has a crossed linker pattern, and the latter two are two-start helices built on the zigzag motif. Illustrations of most of these models are to be found in Refs. 1van Holde K.E. Chromatin. Springer Verlag, New York1988: 289-343Google Scholar and 2Tsanev R. Russev G. Pashev I. Zlatanova J. Replication and Transcription of Chromatin. CRC Press, Inc., Boca Raton, FL1992: 14-24Google Scholar. We shall not discuss them in greater detail here for a simple reason: in our opinion (see below) it is unlikely that any regular helical model describes significant portions of the chromatin fiber. In certainty, there exists no convincing evidence for any one of the specific models that have been proposed. For many years there have been those who were skeptical of helical models. The skeptics fell into two classes. First were the proponents of the “superbead” hypothesis, which held that the condensed fiber represented a linear array of “beads,” each containing some roughly defined number of nucleosomes (see Refs. 1van Holde K.E. Chromatin. Springer Verlag, New York1988: 289-343Google Scholar and 2Tsanev R. Russev G. Pashev I. Zlatanova J. Replication and Transcription of Chromatin. CRC Press, Inc., Boca Raton, FL1992: 14-24Google Scholar for a description of this model and criticism of it). Second, there were a few who, although not wedded to the superbead hypothesis, remained skeptical of any regular ordering in the chromatin fiber (see, for example, Refs. 20Subirana J.A. Monoz-Guerra S. Ayami J. Radermacher M. Frank J. J. Biomol. Struct. & Dyn. 1983; 1: 705-714Crossref PubMed Scopus (15) Google Scholar-25Woodcock C.L. McEwen B.F. Frank J. J. Cell Sci. 1991; 99: 107-114PubMed Google Scholar). However, such skeptics have been a decided minority, with most scientists who are concerned about chromatin higher order structure arguing the merits of one or another of the helical models. Perhaps it is time to reopen the whole question by asking: “how substantial is the evidence for a regular, helical structure in chromatin fibers?” The transmission electron microscope, using negative staining or metal shadowing, provided much of the earlier data concerning chromatin structure. It was realized by many workers in the field that these treatments, in addition to the chemical fixation and extreme dehydration usually employed, could well produce artifacts. In fact, the possibility of distortion during sample preparation has been put forward as an explanation for the remarkable paucity of EM images showing regular helical structure in condensed fibers (25Woodcock C.L. McEwen B.F. Frank J. J. Cell Sci. 1991; 99: 107-114PubMed Google Scholar). Paper after paper presents multiple (presumably selected) images of fibers in which tiny bits are pointed out by arrows as indicating whatever model the authors champion. The vast majority of the fibers in such images show no evidence at all for regular structure. The argument that it was there in the nucleus but was destroyed in extraction is hardly supported by recent studies of sectioned or frozen nuclei, which also show little evidence for regular structure (26McDowall A.W. Smith J.M. Dubochet J. EMBO J. 1986; 5: 1395-1402Crossref PubMed Scopus (119) Google Scholar, 27Horowitz R.A. Agard D.A. Sedat J.W. Woodcock C.L. J. Cell Biol. 1994; 125: 1-10Crossref PubMed Scopus (188) Google Scholar, 28Woodcock C.L. J. Cell Biol. 1994; 125: 11-19Crossref PubMed Scopus (98) Google Scholar). Scanning transmission electron microscopy has been used by Gerchman and Ramakrishnan (29Gerchman S.E. Ramakrishnan V. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 7802-7806Crossref PubMed Scopus (112) Google Scholar) in a very careful study of chicken erythrocyte chromatin. Again, fibers of very non-uniform diameters were observed, a result that correlates with the large variation in mass per unit length derived from these studies. Both factors argue against the idea of regular helices. There exist two electron microscope studies that do provide some evidence for helical structures. Williams et al. (19Williams S.P. Athey B.D. Muglia L.J. Schappe R.S. Gough A.H. Langmore J.P. Biophys. J. 1986; 49: 233-248Abstract Full Text PDF PubMed Scopus (169) Google Scholar) made optical transforms of selected regions of EM images; these show cross-like patterns that constitute evidence for local helical structure. However, these experiments must be evaluated with the understanding that the regions selected were of very limited extent, and the chromatin utilized had unusually uniform linker lengths. It is quite possible (even likely) that small local regions of regular structure exist, perhaps in regions of unusual regularity in nucleosome spacing. In a more recent study of end-on views of short chromatin fibers, Bartolomé et al. (30Bartolomé S. Bermdez A. Daban J.-R. J. Cell Sci. 1994; 107: 2983-2992PubMed Google Scholar) show images depicting what appears to be a helical coiling of the fiber periphery. It is interesting to note that in this as well as in many other studies (see, for example, Ref. 13Thoma F. Koller T. Klug A. J. Cell Biol. 1979; 83: 403-427Crossref PubMed Scopus (1179) Google Scholar) both left and right helices have been reported, sometimes in the same chromatin preparation. Interpretation of such observations, in terms of models, is complicated by the fact that some models demand left helices, some right, and some are ambiguous. Advances in electron microscopic techniques over the past two decades have not changed the picture materially. Careful attempts, using such modern techniques as EM tomography, to map the location of individual nucleosomes in a fiber produced no evidence for regular helices (20Subirana J.A. Monoz-Guerra S. Ayami J. Radermacher M. Frank J. J. Biomol. Struct. & Dyn. 1983; 1: 705-714Crossref PubMed Scopus (15) Google Scholar, 27Horowitz R.A. Agard D.A. Sedat J.W. Woodcock C.L. J. Cell Biol. 1994; 125: 1-10Crossref PubMed Scopus (188) Google Scholar). In summary, despite two decades of careful observation, it cannot be said that electron microscopy has provided convincing evidence for more than fragments of regular helical structure in the chromatin fiber. A wide variety of scattering experiments, using x-rays and neutrons on chromatin fibers in solution, in swollen gels, and in partially oriented quasi-crystalline arrays, have been carried out in attempts to deduce structural parameters. Because it is difficult to obtain highly oriented samples, the scattering patterns observed in most of these studies are almost or completely radially symmetric. This led to misinterpretation of early experiments; the scattering maxima at 5.5, 3.7, and 2.7 nm were thought to represent successive higher orders of the 11-nm reflection originating from a solenoid of 11-nm pitch. This is now believed not to be the case, since the latter three reflections are observed from chromatin solutions at low ionic strength in which the condensed fiber no longer exists; indeed all maxima other than that at 11 nm can be generated by scattering from individual nucleosomes (22Koch M.H.J. Saenger W. Heineman V. Protein-Nucleic Acid Interactions. MacMillan, London1989: 163-204Google Scholar). Furthermore, the preferential orientation of the latter reflections is at right angles to that of the 11-nm reflection (see, for example, Ref. 31Widom J. Klug A. Cell. 1985; 43: 207-213Abstract Full Text PDF PubMed Scopus (203) Google Scholar). Nevertheless, the fact that the 11-nm maximum in the scattering curves is observed only under conditions in which the condensed fiber is stable is widely held to argue for a helical structure with 11-nm pitch. There are, we believe, good reasons to be skeptical even of this inference. Theoretical calculations by Koch (22Koch M.H.J. Saenger W. Heineman V. Protein-Nucleic Acid Interactions. MacMillan, London1989: 163-204Google Scholar) demonstrate that nucleosome-size objects, packed tightly but randomly to give a cylinder with 30-nm outer diameter, produce a scattering curve remarkably similar to that observed experimentally, including a maximum near 11 nm. Thus, a regular helical structure is not required to produce the observed pattern of scattering maxima. A better case for a pervasive helical structure could be made if there were strong evidence in oriented samples for splitting of the 11-nm reflection about the axis, providing at least a hint of a “cross” pattern in the reflections. However, the results of the most careful x-ray diffraction studies of oriented fibers (31Widom J. Klug A. Cell. 1985; 43: 207-213Abstract Full Text PDF PubMed Scopus (203) Google Scholar) show no such effect. Very weak splitting of the 11-nm reflection has been observed in neutron scattering from fibers (32Baldwin J.P. Carpenter B.G. Crespi H. Hancock R. Stephens R.M. Simpson J.K. J. Google Scholar). However, is complicated by the fact that similar results were observed with both chromatin and samples from which H1 had been B.G. J.P. 1976; 3: PubMed Scopus Google the latter should not exist as condensed fibers (see Refs. 10Finch J.T. Klug A. Proc. Natl. Acad. Sci. U. S. A. 1976; 73: 1897-1901Crossref PubMed Scopus (952) Google Scholar and 13Thoma F. Koller T. Klug A. J. Cell Biol. 1979; 83: 403-427Crossref PubMed Scopus (1179) Google Scholar). In scattering data provide no evidence for a regular helical structure in the condensed fiber. Indeed, they seem to be more with the picture that has from EM a fiber that is irregular tightly with at a roughly helical structure by occasional short regular It is often that evidence for preferential orientation of nucleosomes, with their short more or to the fiber axis, helical models. Such evidence from two of experiments, and has provided a and neutron scattering studies of oriented chromatin fibers do some It appears that the 11-nm reflection has preferential whereas the reflections are more (see, for example, Refs. 31Widom J. Klug A. Cell. 1985; 43: 207-213Abstract Full Text PDF PubMed Scopus (203) Google B.G. J.P. 1976; 3: PubMed Scopus Google Scholar). However, to our no has been made to the of orientation from such data. attempts have been made to linear and to provide such (see Refs. 1van Holde K.E. Chromatin. Springer Verlag, New York1988: 289-343Google Scholar and M.H.J. Saenger W. Heineman V. Protein-Nucleic Acid Interactions. MacMillan, London1989: 163-204Google Scholar for of The maximum observed are with both and negative reported, but most results A This is a very small when for example, to the maximum of DNA, which This has been interpreted as indicating that the the DNA base and the fiber is close to the of at which the through and This to a of about of the of the nucleosome from the fiber axis, a result with a number of models. However, to the data in this that all of the nucleosomes the same with to the fiber There in fact, an alternative explanation for a very low for the that the nucleosomes are randomly oriented with to the fiber axis, that even of the fibers still results in of DNA This possibility not seem to have been A in linear data from the fact that the orientation of the linker DNA about of the is has been made to for using D. S. 1986; PubMed Scopus Google the results were similar to those described The most from all of the above studies is that there is some evidence for However, the of orientation is and a weak preferential orientation not in argue for a regular structure. Again, the results seem most by the of no more than a very irregular with only limited regions of attempts to details of structure in the condensed fiber have not been very In this results from the of to the problem is to how the more structure at lower ionic strength might be to as salt is The by chromatin fibers in the of and at salt concentrations of only a few has been the of some The structures first reported by Thoma et al. (13Thoma F. Koller T. Klug A. J. Cell Biol. 1979; 83: 403-427Crossref PubMed Scopus (1179) Google Scholar) have been by many as the extended and have in fact the of models for the condensed helix. However, Thoma et al. (13Thoma F. Koller T. Klug A. J. Cell Biol. 1979; 83: 403-427Crossref PubMed Scopus (1179) Google Scholar) pointed out very that the zigzag structure they observed could well be the consequence of of an open by to the EM Indeed, the results of studies of chromatin fibers in salt are in terms of some kind of an open helical J. Koch M.H.J. Biophys. J. 1986; PubMed Scopus Google Scholar, S.E. Ramakrishnan V. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 7802-7806Crossref PubMed Scopus (112) Google Scholar, 5: PubMed Scopus (48) Google Scholar, J. 1981; PubMed Scopus Google Scholar, J. Biomol. Struct. & Dyn. 1984; PubMed Scopus Google Scholar, J. Koch M.H.J. 1984; PubMed Scopus Google Scholar). evidence that the fiber at low salt is not a zigzag from recent scanning force microscopy experiments J. G. Holde Proc. Natl. Acad. Sci. U. S. A. 1994; PubMed Scopus Google Scholar, G. Holde Zlatanova J. Proc. Natl. Acad. Sci. U. S. A. 1994; PubMed Scopus Google Scholar, G. Zlatanova J. Holde Struct. Biol. 1994; 1: PubMed Scopus Google Scholar) do these experiments give evidence for a regular helical an is The of these observations can now be in terms of a simple of chromatin structure G. Holde Zlatanova J. Proc. Natl. Acad. Sci. U. S. A. 1994; PubMed Scopus Google Scholar, G. Zlatanova J. Holde Struct. Biol. 1994; 1: PubMed Scopus Google Scholar, C.L. Horowitz R.A. Proc. Natl. Acad. Sci. U. S. A. 1993; PubMed Scopus Google Scholar). At low ionic we the linker DNA nucleosomes to be an supported by electron microscopic studies on J. J. Proc. Natl. Acad. Sci. U. S. A. PubMed Scopus Google Scholar). the of each nucleosome with to the one be by the number of base in the linker (see, for example, Ref. 20Subirana J.A. Monoz-Guerra S. Ayami J. Radermacher M. Frank J. J. Biomol. Struct. & Dyn. 1983; 1: 705-714Crossref PubMed Scopus (15) Google Scholar). were a regular, open result J. Koch M.H.J. Biophys. J. 1986; PubMed Scopus Google Scholar, M.H.J. Saenger W. Heineman V. Protein-Nucleic Acid Interactions. MacMillan, London1989: 163-204Google Scholar, C.L. Horowitz R.A. Proc. Natl. Acad. Sci. U. S. A. 1993; PubMed Scopus Google Scholar, 1981; Scholar). However, are not of uniform length in chromatin irregular structures like that in should be the and images are very The open are and observed to have a of the order of 30 nm and a of about nm. It is the fact that the low ionic strength structure also has a of about 30 nm that to the term 30-nm fiber as a for the condensed structure. such an irregular structure is to it seem that the might be an not to a of any defined or regularity but to an irregular of such a condensation might on the of linker DNA as the salt concentration is The supported by data of et al. J. J. Proc. Natl. Acad. Sci. U. S. A. PubMed Scopus Google Scholar), is that the linker close of However, data of Butler and P.J.G. J. Biol. 1980; PubMed Scopus Google Scholar) no of over a wide salt that the linker This is in with the study of in and linker DNA in Proc. Natl. Acad. Sci. U. S. A. 1989; PubMed Scopus Google Scholar) and with recent EM observations by Horowitz et al. R.A. Agard D.A. Sedat J.W. Woodcock C.L. J. Cell Biol. 1994; 125: 1-10Crossref PubMed Scopus (188) Google Scholar). This an if might well to the of the fiber reported by (see Refs. 1van Holde K.E. Chromatin. Springer Verlag, New York1988: 289-343Google Scholar and 2Tsanev R. Russev G. Pashev I. Zlatanova J. Replication and Transcription of Chromatin. CRC Press, Inc., Boca Raton, FL1992: 14-24Google Scholar for Furthermore, an condensation of this kind a problem that has been by many of the proponents of helical how can helical chromatin regions or on the on the of the the condensation or the could be by however, the of convincing evidence for a eukaryotic a problem for requires the of negative The only kind of helical chromatin fiber that could a is one in which both of the are in Bartolomé et al. (30Bartolomé S. Bermdez A. Daban J.-R. J. Cell Sci. 1994; 107: 2983-2992PubMed Google Scholar) this from their EM studies of views of chromatin chromatin fibers to structures at ionic conditions those in the for some of helical coiling can be despite years of it to that there is very little reason to believe that the condensed chromatin fiber substantial of any regular helical structure. are there reasons to believe that the fibers observed and in vitro represent of more regular in structures. Furthermore, what we now of the of nucleosome in fibers at low ionic strength for a corresponding under conditions. For these to about the of higher order structures Careful of the chromatin literature that many authors have concerning the of regular helical structure in the fiber (see especially Refs. 20Subirana J.A. Monoz-Guerra S. Ayami J. Radermacher M. Frank J. J. Biomol. Struct. & Dyn. 1983; 1: 705-714Crossref PubMed Scopus (15) Google M.H.J. Saenger W. Heineman V. Protein-Nucleic Acid Interactions. MacMillan, London1989: 163-204Google Scholar). Very evidence from EM R.A. Agard D.A. Sedat J.W. Woodcock C.L. J. Cell Biol. 1994; 125: 1-10Crossref PubMed Scopus (188) Google Scholar) has provided the most to regular models for the condensed fiber. the authors show no of nucleosomes the fibers, but a of structures are now are more by the most of chromatin fibers from a variety of with attempts made to focus on structure in important be the study of chromatin fibers with spaced Such fibers might be by of chromatin or by of both and linker of a very strong and specific nucleosome Such structures be to form regular helices at low ionic which might into regular condensed structures. The of such model structures in and studies provide a against which the data on fibers could be We M. Koch and Woodcock of for of the and and S. H. and G. of for the images in