Proc. Natl. Acad. Sci. USA Vol. 93, pp. 10548-10555, October 1996 Review What determines the folding of the chromatin fiber? Kensal van Holde*t and Jordanka Zlatanova4 *Department of Biochemistry and Biophysics, Oregon State University, Corvallis, OR 97331-7305; and PInstitute of Genetics, Bulgarian Academy of Sciences, Sofia 1113, Bulgaria ABSTRACT In this review, we attempt to summarize, in a DOES THE LINKER DNA BEND? critical manner, what is currently known about the processes of condensation and decondensation of chromatin fibers. We The solenoid model appeared to gain substantial support from begin with a critical analysis of the possible mechanisms for the studies of Yao et al. (6), which provided evidence that the condensation, considering both old and new evidence as to linker DNA in dinucleosomes did in fact contract or fold in whether the linker DNA between nucleosomes bends or re- some fashion as the salt concentration was raised from 0 to 20 mains straight in the condensed structure. Concluding that mM. The picture (and in particular, the putative role of linker the preponderance of evidence is for straight linkers, we ask histones) was complicated by a subsequent study (7), which what other fundamental process might allow condensation, showed that the same changes could be observed with dinu- and argue that there is evidence for linker histone-induced cleosomes from which linker histones had been removed. Yao contraction of the internucleosome angle, as salt concentra- et al. (6, 7) used two techniques to provide evidence for linker tion is raised toward physiological levels. We also ask how contraction: (i) direct visualization of fixed dinucleosomes by certain specific regions of chromatin can become decon- transmission electron microscopy (EM) and (ii) measurement densed, even at physiological salt concentration, to allow of the translational diffusion coefficient of dinucleosomes by transcription. We consider linker histone depletion and acet- dynamic light scattering. The latter measurements showed an ylation of the core histone tails, as possible mechanisms. On increase in the translational diffusion coefficient (D) with the basis ofrecent evidence, we suggest a unified model linking increasing salt concentration (Fig. IA, dashed line). This is targeted acetylation of specific genomic regions to linker most easily explained by a compaction of the particle, for such histone depletion, with unfolding of the condensed fiber as a compaction should produce a decrease in the frictional coef- ficient, f, and hence an increase in D, since D is inversely consequence. proportional tof. The results from hydrodynamic experiments were supported by EM studies, in which dinucleosomes in 20 The belated discovery by molecular biologists that chromatin mM Na+ or 2 mM Mg2+ were observed to be more compact structure might be of major importance in regulating DNA than those in 2 mM Na+. However, in a recent publication, transcription and replication has sparked a renewed interest in Bednar et al. (9) have repeated both experiments with quite some old questions: How does a chain of nucleosomes fold to different results. As Fig. 1A shows, they observe the diffusion produce the condensed fibers observed in the eukaryotic coefficient to be independent of salt concentration over the nucleus? What makes it unfold to allow transcription or same range. In addition, Bednar et al. (9) find by cryo-EM that replication? The latter question has been especially perplexing. the center-to-center distance between nucleosomes in dimers The earliest electron microscopy studies of isolated chromatin does not decrease as salt concentration is raised. Interestingly, fibers showed that at low salt concentrations an extended string they do, like Yao et al. (6, 7), find a contraction at higher salt of nucleosomes could be observed, whereas raising the ionic when the dinucleosome is studied by conventional transmis- strength to levels close to physiological led to the formation of sion EM. This indicates that either transmission EM or an irregular, highly condensed fiber about 30 nm in diameter. cryo-EM is giving an artifactual result. Speculations concerning the mechanisms of such folding and Thus, we have, at the focus of an important issue, a clear the structure of the "30 nm fiber" were rife in the early contradiction: Does the linker DNA contract with increasing postnucleosome years (for review, see refs. 1 and 2, and salt or does it not? On the resolution of this question depends references therein). Although many models for the condensed how we may visualize the condensed chromatin fiber, and what fiber structure were proposed, and hotly debated, the "sole- role the linker histones may play in that condensation. To noid" model of Finch and Klug (3) or variants thereof (4, 5) approach this issue, we have first asked: Are there other data gained acceptance by most researchers. In such structures, in the literature that might support one view or the other? nucleosomes adjacent on the DNA strand are packed cheek- In fact, there exists a wealth of relevant evidence, both in the by-jowl into a regular helix. earlier literature and from more contemporary work. Con- A corollary of the solenoid model is that the linker DNA sider, for example, the measurement of diffusion coefficients between adjacent nucleosomes must, at least at physiological of dinucleosomes by dynamic light scattering. In an earlier salt concentration, be bent or curled in some fashion to allow study, Marion et al. (8) find no change in D in up to 80 mM salt; adjacent nucleosomes to contact one another. This salt- their data are in almost exact quantitative agreement with dependent bending or coiling was postulated to be facilitated those of Bednar et al. (9) (Fig. 1A) and are inconsistent with by the interaction of linker DNA with "linker histones" (Hi, those of Yao et al. (6). H5, and the like), for these proteins have been demonstrated There is, in addition, extensive evidence from sedimentation to be essential for proper chromatin fiber condensation (see studies of dinucleosomes. If a dinucleosome contracts with below). In this review, we will try to critically evaluate the data increasing salt concentration, the sedimentation coefficient s on (i) linker DNA bending, (ii) nucleosome-nucleosome in- should increase, for like D, s is inversely proportional to the teractions, (iii) the role of linker histones, and (iv) the role of the core histone tails and their acetylation in the folding of the Abbreviations: EM, electron microscope/microscopy; SFM, scanning chromatin fiber. In addition, we present some speculations force microscope. concerning the unfolding of the fiber in transcription-related tTo whom reprint requests should be addressed. e-mail: vanholdk@ processes. bcc.orst.edu. 10548 Downloaded by guest on October 1, 2021 Review: van Holde and Zlatanova Proc. Natl. Acad. Sci. USA 93 (1996) 10549 A. A. I*- 21.1 nm- 2.4 center-to-conte (.t..::: distance ( \ --------- 2.3 i : t j < t -- J ^ 5.5 nm _1- linker DNA 2.2 - .1.. , U-' 2.1 [ U1 center-to-center o- X distance linker blNAj. .. 2.0 [ : -- :. 5.5 nm 23.6 nm \ J}. 1.9 F 1.8I'L.j B. 0 10 20 80 NaCl concentration (mM) B. 25 Predicted for compact dimers 20 _ _ _ _ _ __ _ 0 ...: . 4.4 15 omMom ,,.9j 11 1*Predicted for extended dimers .. 10 V)d) FIG. 2. Models of dinucleosomes and trinucleosomes. (A) Dinu- cleosome. The relationship between center-to-center distance and 5 linker length depends slightly on relative orientation of nucleosomes. Two extremes are shown for chicken erythrocyte dinucleosomes. (B) . Trinucleosome. The distance R13 depends on the angle between DNA duplexes entering and exiting nucleosome 2. 0 20 40 60 80 100 NaCl concentration (mM) each other) leads to a prediction of s2 = 19S (Fig. 1B, broken line). The difference is clearly far greater than can be ac- FIG. 1. Hydrodynamic studies of dinucleosomes as a function of counted for by experimental error. salt concentration. (A) Diffusion coefficient measurements from As a test for the credibility of the experimentally measured quasi-elastic light scattering: l, data of Marion et al. (8); *, data of values for D and s, we may ask if the combination of the best Yao et al. *, data of Bednar et al. Sedimentation (6); (9). (B) estimates forD (2.25 x 10-7 cm2/sec) and s (15.4 x 10-13 sec) coefficient measurements: *, data of Wittig and Wittig (10); *, data a molecular for an average dinucleo- of Stratling (11); O, data of Butler and Thomas (12). gives reasonable weight some. It is realized of course, that individual dinucleosomes frictional coefficient. Published results from a number of may vary in linker length, degree of external DNA trimming, laboratories are shown in Fig. 1B; all lead to the same and content of linker histones. Calculation of the molecular conclusion: there is no significant change in the sedimentation weight of a typical dinucleosome, based on reasonable struc- coefficient of dinucleosomes over the entire salt concentration tural parameters (total DNA length plus two histone octamers range from 0 to 100 mM. All data points lie close to an average plus two linker histones) gives a value ofM = 5.13 x 105 g/mol. value of 15.4S. We can make a rough estimate of the change This is in almost exact agreement with that calculated from s D in s to be expected from contraction of the linker by using the and from the Svedberg equation Kirkwood (13) formalism as adopted by Bloomfield et al. (14). RTS This expresses s, the sedimentation coefficient for an n-mer, M = 5.10 g/mol [2] in terms of si, the sedimentation coefficient of the monomer, =D(l1-p) the Stokes' radius of the monomer (r), and the set of distances when we have used the value given by Wittig and Wittig (10) (Rij) between units i and j in the n-mer, for the partial specific volume, P.
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages8 Page
-
File Size-