J. Mol. Biol. (1990) 212, 495-511

Structure of Nucleosomes and Organization Internucleosomal DNA in Chromatin

of

S. G. Bavykin’, S. I. Usachenko’, A. 0. Zalensky3 and A. D. Mirzabekov’f‘Institute of Molecular Biology USSR Academy of Sciences Vavilov Street, Moscow 117984, USSR Uzbek

21nstitute of Bio-organic Chemistry of Sciences, Tashkent 700143, USSR

Academy

‘Institute

Vaschnil,

of Agricultural Microbiology Leningrad 188620, USSR

(Received 28 June 1988; accepted 27 November

1989)

We have compared the mononucleosomal pattern produced by micrococcal nuclease digestion of condensed and unfolded chromatin and chromatin in nuclei from various sources with the repeat length varying from 165 to 240 base-pairs (bp). Upon digestion of isolated HI-containing chromatin of every tested type in a low ionic strength solution (unfolded chromatin), a standard series of mononucleosomes (MN) was formed: the core particle, MNib5, and Hl-containing, MNleS, MNlT5, MN1s5, MNlg5, MNZo5 and MNzls (the indexes give an approximate length of the nucleosomal DNA that differs in these particles by an integral number of 10 bp). In addition to the pattern of unfolded chromatin, digestion of whole nuclei or condensed chromatin (high ionic strength of Ca2+) gave rise to nucleispecific, Hl-lacking MNiS5. Digestion of Hl-lacking chromatin produced only MN,,,, MN‘,55 and MNic5 particles, indicating that the histone octamer can organize up to 165 bp of nucleosomal DNA. Although digestion of isolated sea urchin sperm chromatin (repeat length of about 240 bp) at a low ionic strength gave a typical “unfolded chromatin pattern”, digests of spermal nuclei contained primarily MNlb5; MNlS5, MNzj5 and MNzd5 particles. A linear arrangement of histones along DNA (primary organization) of the core particle was found to be preserved in the mononucleosomes, with the spacer DNA length from 10 to 90 bp on one (in MNlS5) or both sides of core DNA being a multiple of about 10 bp. In MN2a5, the core particle occupies preferentially a central position with the length of the spacer DNA on both sides of the core DNA being usually about 30+60 or 40+50 bp. Histone Hl is localized at the ends of these particles, i.e. close to the centre of the spacer DNA. The finding that globular part of histones H3 and sea urchin sperm H2B can covalently bind to spacer DNA suggests their involvement in the organization of chromatin superstructure. Our data, indicate that decondensation of chromatin is accompanied by rearrangement of histone Hl on the spacer DNA sites adjacent to the core particle and thus support a solenoid model for the chromatin superstructure in nuclei in which the core DNA together with the spacer DNA form a continuous superhelix.

of nucleosomal core crystals (Richmond et al., 1984; Uberbacher & Bunick, 1985; Burlingame et al., 1985), DNA-histone (Mirzabekov et al., 1978; Bavykin et al., 1985a; Mirzabekov, 1986) and histone-histone (McGhee & Felsenfeld, 1980; Mirzabekov, 1980) crosslinking studies, etc. (for reviews, see Lilley & Pardon, 1979; McGhee & Felsenfeld, 1980; Mirzabekov, 1980; Pederson et aZ., 1986).

1. Introduction The structure of the core particle, which represents the first level in chromatin organization, is at present fairly well known, owing to X-ray analyses t Author addressed.

to whom

0022-2836/90/070495-17

all correspondence

$03.00/O

should

be

495

0

1990 Academic

Press Limited

S. G. Bavykin

At the next level of chromatin organization, core particles toget,her with internucleosomal (spacer) DNA are folded with the help of histone HI into a 30-nm fibre, the arrangement of which is still unclear in many essential details (Butler, 1983; Felsenfeld & McGhee, 1986; Crane-Robinson et al., 1984). This paper is concerned with the organization of nucleosomes and internucleosomal DNA in chromatin with a different repeat length originating from different sources, particularly sea urchin sperm. Sea urchin sperm nuclei were chosen as a convenient model for chromatin structure studies. This completely repressed and highly condensed chromatin seems t,o be packed in a most regular way. The length of its nucleosomal repeat is about 240 bpt (Spadafora et al., 1976), which is longer than in any other chromatin studied so far. Histone H2B is present here as a spermal variant having a rather long positively charged insert at the N-terminal moiety (Strickland et al., 1978). In t,his work, we have found, using sea urchin sperm mat,erial, a striking difference in micrococcal nuclease digestion patterns of unfolded chromatin and nuclei. The digestion pattern of unfolded Hlcontaining chromatin is the same in all the material that was studied. Following condensation of chromat,in, the digest,ion pattern becomes similar to that of nuclei. The digestsion patterns of whole nuclei depended on the source and included a nucleispecific Hl-lacking particle MNIs5 in addition to the set of particles in the “unfolded chromatin pattern”. The primary organization (sequential arrangement of histones along DNA) of the spermal nucleosomes MN 155 and MN,,, was analysed by covalent crosslinking of histones to DNA within nuclei

and

compared

with

that

of

those

studied

previously, MN,,,, MN,,,, MN1,5 and MN1,, et aZ.: 1980; Karpov et al., 1982; (Belyavsky Bavykin et aZ., I985a). On the basis of these data we shall discuss structural transitions in chromatin induced by its decondensation or isolation from nuclei and solenoid models of chromatin organization. 2. Experimental (a) Isolation

of nuclei

Procedures and chromatin

The sperm of sea urchin Strongylocentrotus intenned&s was stored at - 70°C. Nuclei were extracted at 4°C as described by Zaienskaya et al. (1981). Sperm chromatin was obtained as described by Spadafora & Geraci (1975); with some modifications. In order to obtain 200 A,,, units of chromatin, I g of sperm was suspended in a Dounce homogenizer (plunger L) in 50 ml of 0.57 &I-KC], filtered and centrifuged at 2000 g for 10 min. The pellet was washed twice in the same buffer, resuspended in 40 ml of 0.05 M-Tris.HCl (pH 7.5); 0.25 M-SUCrOSe, 0.005 x-MgCl,, 1% Triton X-100 in a Dounce homogenizer (plunger S) and freshly prepared @03 M-diisopropylfluorophosphate was added to a. final 7 Abbrevia.tions used: bp, base-pair(s); MN, mononucleosomes; b, bases.

et al. concentration of 30 ELM and incubated for 30 min with steady mixing at 4°C. The material obtained was centrifuged, separated and washed twice wit,h the same buffer without diisopropylfluorophosphate and once without Triton X-100. The nuclei were suspended in 20 ml of 0.005 M-MgCl,, &05 Af-Tris. HCl (pH 75), 2 M-SUCI’OSe, layered on an equal volume of the same buffer and centrifuged at 14,000 g for 10 min. then suspended in 50 ml of 0.1 M-Tris . HCl (pH 7.5) in a Dounce homogenizer (plunger L) and centrifuged at 2000 g for 10 min, stirred by hand in 0.01 M-Tris” KC1 (pH 7.5) and centrifrnged again under t’he same conditions. In order to obtain chromatin, the nuclei were lysed in a Dounce homogenizer in 30 ml of bidistilled water. The nuclei of Drosophila embryos, rat liver; mouse ascites carcinoma, yeast and chicken erythrocytes were isolated as described @hick et al., 1950, 1985: Bavykin et al., 1985a), and lysed in 20 to 40 vol. of bidistilled water in a Dounce homogenizer t,o isolate chromatin. The HI and H5 histones were removed from chicken erythrocytes chromatin as described (Shick et al., 1985). (b) Crosslinking

qf histones to DNA

Crosslinking of DNA to histones was performed nuclei as described by Bavykin et al. (1985a). (c) NydroEysis of chromatdn and nuclei microcouzl nucleate, analysis of DNA isolation of nucleosomes

w:ithin

with and

Crosslinked and uncrosslinked nuclei or chromatin (50 units/ml) were hydrolysed in different soiuto 150 A,,, tions (see Result,s) with micrococcal nuclease (Wort,hington, 85 units/ml) at 37°C as described by Shick et al. (1980). Before digestion, chromatin was precipitated (excluding Fig. 2, lane s) in 50 mw-sodium cacodylate (pH 7.0), 2.5 rnM-EDTA. Following hydrolysis of proteins with Pronass (Calbiochem, 0.5 mg/mI), the length of single-stranded nucleosomal DNA was measured by comparing t,he electrophoretic mobility of each sample in denaturing polyacrylamide gel in the presence of 7 M-urea ($Ianiatis et al., 1975), with the mobility of DNA fragments of known size obtained by digestion of rat liver nuclei with DNase I (Noll, 1974; Lohr & Van Holde, 1979; Prune11 et al.; 1979). One volume of buffer (7 M-urea in formamide) was added to each sample before electrophoresis. units of nuclei or chroma,tin Nucleosomes (4000 A,,, hydrolysate) were separated by preparative electrophoresis in a 796 polyaerylamide gel (40 cm x 20 cm x et al., 1976; 0.06 cm) at pH 8.3 as described (Varshavsky Shick et al., 1980). Positions of nucleosomes were visualized in ultraviolet light (A = 254 nm), the gel being applied to a thin-layer chromatography plate “Silufol UV-254” (Czechoslovakia). The gel was pestled in a Dounce homogenizer and t(he nucleosome fractions were eluted overnight with 4 vol. electrode buffer. (d) Two-dimensional electro?horesis of crosslinked DNA-protein complexes from nucleosomes

Removal of free histones and DNA and 2-dimensional elect’rophoresis of crosslinked DNA-histone complexes were carried out as described @hick et al., 1980: Bavykin et al., 1985a). In order to improve the resolution of the separation of DNA diagonals crosslinked to different histones, eysteine residues in histones H3 and H4 from sea urchin sperm

Organization

Chromatin:

of Nucleosomes

(Strickland et al., 1974; Sures et al., 1978). were, in some experiments, oxidized to cystines in the presence of l,lOphenanthroline with the formation of dimers (Kuehl, 1979). To do so, the crosslinked nucleosomes (0.8 A,,,), after removal of the main part of free histones and DNA, were dissolved in 50 ~1 of 10 mM-Tris.HCl (pH &O), 6 M-urea, incubated in boiling water for 1 min to denature DNA and thereby prevent its splitting by 1 ,lO-phenanthroline through intercalating double-stranded DNA, then 10 ,LL~of calf thymus histone H3 (to 6 mg/ml) and 2 ~1 of a freshly prepared solution of @32 mMl,lO-phenathroline plus @6 mM-CuSO, were added. Incubation was for 2 h at 37°C. The reaction was terminated by addition of 32 ,LJ of 2 mm-EDTA (pH 8.0). Under these conditions only 10% of the cysteinecontaining histones were transformed to dimers. Half the volume of the first-direction electrophoresis buffer (Shick et al., 1980; Bavykin et al., 1985a) without dithiothreitol was added to each sample. The solution was then incubated at 55°C for 1 min and applied to t;he gel.

3. Results (a) Differences

in micrococcal n&ease of chromatin and nuclei

hydrolysis

Electrophoresis of DNA under denaturing conditions (Maniatis et al., 1975) provides a better separation of fragments than is possible by electrophoresis of native DP\rA. Accordingly, we have found new

peculiarities in the microeoccal nuclease digestion of chromatin (Finch et al., 1975). At the initial stage of sea urchin sperm chromatin hydrolysis under low ionic strength conditions (10 mw-Tris . HCl, pH 8.0, 0.4 mM-CaCl,), there appear‘s a wide spectrum of nucleosomal DNA fragments with the DNA length varying from about 145 to 225 nucleotides (Fig. l(b), core particles and fractions 1 to 8). The difference in the DNA length remained as multiples of about 10 bp, and the predominating fragments were “core particle” MNlb5, MN,,,, MN,,, and MN,s, particles containing DNA fragments of 145, 165, 175 and 185 nucleotides long, respectively (Fig. l(b), fractions: core, 2, 3 and 4). At the final stages of hydrolysis, the nucleosomes MNlG5 and MN145 become the basic particles occurring in hydrolysates (not shown). It should be noted here that particle MN,55 is present at all stages of hydrolysis in quite small amounts (Fig. l(b), fraction 1). This pattern of unfolded chromatin digestion by micrococcal

nuclease

seems to be of general

significance as the same general picture was observed earlier for unfolded chromatin from various sources, namely Drosophila embryos (Fig. 2; lane a), rat liver (Fig. 2, lane c) (Belyavsky et al., 1980; Karpov et al., 1982), mouse ascites carcinoma (Fig. 2, lane m) and chicken erythrocytes (Fig. 2, lane e), with the length of the nucleosomal repeat being 185, 195, 180 and 212 bp, respectively (Kornberg, 1977; Bakayev et al., 1977). In contrast intermediate sperm nuclei

to unfolded

chromatin,

stages of hydrolysis yield

only MN,,,,

MNZd5 (Fig. l(a)), corresponding

the initia’l

and

of sea urchin

MN1,,,

MN235 and

to fractions

1 to 4

and #pacer

DNA

497

in Figure 4(a), below, whereas particles with DNA of 165, 175, 185 bp, etc., are present in negligible amounts only at early stages of digestion. Upon deep hydrolysis, MNZs5 and MNZb5 disabppear (Fig. l(a), lane d), transforming probably to MN1,, being present in and MN155r the latter particles about equal amounts at all stages of hydrolysis. The MN,55 particles are formed in hydrolysis of all types of nuclei studied by us: Drosophila embryos (Fig. 2, lane g), mouse ascites carcinoma (Fig. 2, lane i), chicken erythroeytes (Fig. 2, lane k), lily sepals (not shown), yeast (Fig. 3(a)). The MN 155 particles are present also in digests of condensed chromatin in the same amounts as in nuclei. Thus, an increase in the Ca*+ concenl;ration up to 1 mM (Fig. 2, lane o) or in the ionic strength up to 130 mM (Fig. 2, lane q) during chromatin digestion leads to the appearance of MN,,5, which is absent after digestion of chromatin in a low ionic strength buffer, containing 0.25 to @4 ml&a’+ (Fig. l(b) and Fig. 2, lanes a, c, e and m). Our data correlate with the data of Bordas et al. (1986a) showing that chromatin condensation starts at a high chromatin concentration (3 to 6 mg DNA per ml) only when the Mg*+ concentration is over 1 mM. It is necesary to mention here that the disappearance of MN,,, does not depend on the mel;hod of chromatin isolation. The MN,,, particles were observed after digestion of chromatin obtained by lysis of nuclei in bidistilled water (Figs l(b) and 2(a)) as well as chromatin isolated after mild nuclease digestion of nuclei (Worthington: 1 unit per 5 mg DNA, 37°C for 30 min, the reaction being terminated by addition of EDTA to 2i.5 mM), followed by their lysis by dialysis against 0.25 mM-EDTA, pH 8.0 (Fig. 2, lane m). The essential moment is the content of divalent cations in chromatin. Thus, digests of chromatin that have not been washed with EDTA before digestion by nuclease (see Experimental Procedures) contain MN,,, particles even if digestion was carried out at low ionic strength buffer (Fig. 2, lane s). In contrast to sea urchin sperm nuclei, digests of Drosophila embryos, mouse ascites carcinoma and chicken erythroeyte nuclei contain a continuous spectrum of nucleosomes: MNXd5, MNl,,, MN1,,, MNas etc. (Fig. 2(b), fractions c and 1 to 7). At ihe same time, digests of yeast nuclei (Fig. 3(a)) from which histone HI is known to be absent (Wintersberger et al., 1973; Davie, 1982; Certa et al., 1984) or chicken erythrocyte chromatin (Fig. 3(b)) that has been depleted of Hl and H5 (Shiuk et al., 1985) prior to nuclease treatment contain only MNlb5, MN155 and MN,,,. So, the histone octamer per se organizes 145, 155 or 165 nucleotide pairs and protects those against digestion by nucleases. The presence of Hl in a nucleosome favours a discrete shielding of longer DNA stretches, up to 215 bp in the nuclear chromatin of chicken erythrocytes (Fig. 2, lane k). The presence of MN1,5 in the set of nucleosomes obtained from nuclei or from condensed chromatin, but not unfolded chromatin, digested by micro-

Chromatin:

Organization

of Nucleosomes

coccal nuclease indicates that some rearrangements seem to follow decondensation of chromatin. (b) Xtructure

and #pacer

499

DNA

Table 1 Localization of histone crosslinking DNA in MN,,,, MN,,, and MN,,,

sites on particles

of nucleosomes

The primary structure of core particles obtained from sperm nuclei and MN1,5 and MN1s5 nucleosomes isolated from sea urchin sperm chromatin has been established (Karpov et al., 1982; Bavykin et al., 1985a). It was of interest to us to elucidate also the structure of MNIS, which is specific for all micrococcal nuclease digests of nuclei, and MN235, which is found only in hydrolysates of sea urchin sperm nuclei. These particles were isolated from the nuclear digests in a homogeneous state by electrophoresis in polyacrylamide gels (Fig. 4(a)). Analysis of the protein composition in fractions 1 to 4 (Fig. 4(a)) has shown the presence of histones H2A, H2B, H3 and H4 in about equimolar amounts. Fractions 3 to 4 contained histone Hl also (not shown). The length of nucleosomal DNA was determined by comparison with single-stranded DNA fragments of known size obtained from DNase I digests of rat liver nuclei and was found to be approximately 145, 155, 235 and 245 bp for fractions 1 to 4, respectively (Fig. 4(b)). As regards the DNA length, fractions 1 to 2 were fairly homogeneous, fraction 3 was more heterogeneous and contained DNA of 235( + 10) bp and fraction 4 was very heterogeneous; therefore, we made no attempt to analyse the primary organization of MNZd5. To find out the localization of histones on the nucleosomal DNA, histones were covalently crosslinked within nuclei (Bavykin et aZ., 1985a) to partially depurinized DNA under mild conditions (Levina et al., 1981). At the site of crosslinking, the DNA was split in such a manner that the histone molecule remained attached only to the 5’-end DNA fragment. Then the nuclei were hydrolysed with micrococcal nuclease, and the nucleosomes were fractionated by preparative gel electrophoresis (Varshavsky et al., 1976; Shick et al., 1980). The location of histone crosslinks on the nucleosomal DNA was found relative to the 32P-labelled 5’-end by measuring the length of the 5’.terminal nucleosomal DNA fragments bound to each histone in a system of two-dimensional gel electrophoresis (Shick et al., 1980; Bavykin et al., 1985a). In such a system, the crosslinked protein-DNA complexes were separated in the first dimension by electrophoresis under denaturing conditions. Upon digestion of histones by Pronase in the gel, the

H2A

35 112-122 145

112-132 145, 155

65-95@’ 140-205”’

ml3

25-35 48-58 95, 105, 115, 125

35-68

55_gp’,‘b’.‘d’ 88-118’“’ 125-195”’

95; 105, 115, 125, 135

200-235’“’ H3

H4

Hl

5 (68), 75. 85, 95 135-145

68, 75, 85, 95, 105 135-155

55, 65

55, 65, 75

88

88, 98

-

45-65’d’ 89-155’“’ 165-195”’ 200-235”’ (65), (75), 85, 95, 105, ‘115, 125’*‘,(b) 118, 128, 138. 148, (158_168)@‘.@‘.‘d’ (loo-195)‘“‘, 195-235’“’

Values represent the size in nucleotides of DNA fragments covalently bound to histones, which also corresponds to the distance of a particular histone crosslinking site from the g-ends of nucleosomal DNA. Values in parentheses represent weak crosslinking sites. 7 See Fig. 5(c) and Bavykin et al. (1985a). $ See Fig. 5(a) and (b). 4 (a) to (d) correspond to (a) to (d) in Fig. 6.

32P-labelled DNA fragments were fractionfated in the second dimension according to their lengths and thus they fell on different diagonals, each diagonal corresponding to a particular histone bound to DNA in the first-dimension electrophoresis. Figures 5 and 6 show the two-dimensional gels for the core particle MNlh5, MNls5 and MN235 nucleosomes. The data on the location of histones on DNA derived from these gels are summarized in Table 1. (c) Primary

organization

of MNI,,

Comparing the length of DNA fragments bound to each histone fraction in the core particle with those in MN,,, (Fig. 5; Table l), one can :see that two sets of DNA fragments exist in the latter particle: the first set is the same as for the core nucleosome, while the other is elongated with re:spect to

Figure 1. Electrophoretic separation under denaturing conditions of DNA fragments from micrococcal nuclease digests of (a) sea urchin sperm nuclei and (b) isolated chromatin. Digestion was done under low ionic strength conditions (10 mrvr-Tris. HCl, pH 8.0, 0.4 mM-CaCl,). The digests were centrifuged at 1000 g, the pellet was discarded, and the supernatant was treated with Pronase. Then sample buffer was added, the mixture was boiled and applied on the gel. The denatured DNA hydrolysates were separated in 9% polyacrylamide gel in the presence of 7 M-urea (Maniatis et aZ., 1975). Denatured DNA fragments of known size (markers) were obtained by digestion of rat liver nuclei with DNase I and run in lanes e and j; their lengths in nucleotides are given on the right-hand side of the Figure (Noll, 1974; Lohr & Van Holde, 1979; Prune11 et al., 1979). Fractions 1 to 8 in (b) correspond to DNA of MN,,,, MN1,,, MN,,,, MN1s5. MN,,,, MNzo5, MN,,, and MN,,,, respectively. The digestion time (min) is indicated above the lanes.

dr

ef

432-

-176

c-

-

21

-

2-

-176 -

c-

-14.2 -

- 142

-

-

-142

-

-

-

-

-

-

CI

b

d

C

e

7

dr

--

-I76 -I42

-

432/-

176

-

142

-176

c-

-I42

-

4

h

I

j ( b)

Fig. 2.

k

i

Chromatin: asc

Organization

of Nucleosomes

asc

501

and Spacer DNA

asc

asc

-

4 3 2

-I76

76

C

-I42

42

76

I-176 -

42

- 142

(c) Figure 2. Comparison of micrococcal nuclease digests of isolated (a) and (c) chromatin and (b) nuclei from Drosophila embryos (dr, lanes a and g); rat liver (rl, lane c), chicken erythrocytes (er, lanes e and k) and mouse ascites carcinoma (asc, lanes i, m, o, q and s). Chromatin and nuclei were digested in low ionic strength buffer ((a) and (b) as for Fig. 1; lane m, 10 mM-Tris.HCl (pH 80), 0.25 mM-CaCl,, 0.25 miv-EDTA; lane s, the same buffer without EDTA or in the conditions ensuring the folded chromatin state (lane o, 10 mnil-Tris.HCl (pH 8.0), 1 mM-CaCl,, 0.25 mM-EDTA; lane q, 10 mM-Tris-sodium phosphate (pH 8.0), 80 mM-Nacl, 025 mM-CaCl,, 0.25 mM-EDTA)., Chromatin was obtained by lysis of nuclei in bid&tilled water (lanes a, c, e and s) or after mild digestion with micrococcal nuclease (lanes m, o and q; (see the text)). Electrophoretic separation of DNA fragments was carried out under denaturing conditions, as for Fig. 1. The treatment by nucleases was for 20 min. Lanes b, d, f, h, j, 1, n, p, r and t contained DNA markers (see Fig. 1); band c, nucleosomal core particle; bands 1 to 7, mononucleosomes, as for Fig.l(b).

the first one by 10 bp. This implies that MN1,, is a precursor of the core nucleosome whose DNA is extended by IO bp at one of its termini (Fig. 7). The MN 155 particles do not contain histone HI, Upon micrococcal nuclease digestion of nuclei from various sources, they are formed in about the same amounts as the core particles (Figs l(a), 2(b) and 3(a)). By contrast, upon hydrolysis of unfolded chromatin isolated from these nuclei (Figs l(b) and 2, lanes a, c, e and m) mainly the core particles and were produced whereas MN1 55 particles were MNI,, scarce. Thus one can conclude that, along with the core particles and chromatosomes (MNiG5), other basic particles of digestion of nuclei and unfolded chromatin are, respectively, MN,,, and MNlT5. (d) Primary

organization

of MN,,,

It did not seem feasible to separate completely the DNA fragments crosslinked to different histones within MN,,,, because of the inherent heterogeneity of this particle. In order to obtain a clearcut picture, two-dimensional electrophoresis was performed under four different conditions (Fig. 6(a) to (d)). After removal of uncrosslinked histones and DNA,

the crosslinked DNA-histone complex was pretreated with l,lO-phenanthroline (Fig. 6’(a) and (c)), which under the conditions used oxidized a certain proportion of cysteine residues in spermal histones H3 and H4 to cystines and this led to a dimerization of these histones and some narrowing of histone diagonals for monomers and dimers (see Experimental Procedures). The nature of the latter phenomenon is unclear. The appearance of the diagonal corresponding to the dimer of histone Hl (Fig. 6(a)) suggests that the spermal variant of HI in Xtrongylocentrotus intermedius perhaps contains cysteine, just as the embryonic variant of SCrongylocentrotus purpuratus (Levy et al., 1982), or the spermal variant of the mollusc Glycymeris yessoensis (Odintsova et al., 1989), or histone Hl from chicken, Drosophila and newt (Wells, 1986). Comparison of the primary organization of the core nucleosomes and the MNIS, particles (Fig. 5) obtained from nuclei with the core particles, MNlG5, derived from chromatin and MN185, MN175 et al., 1982; (Belyavsky et al., 1980; Karpov Bavykin et al., 1985a), has revealed a close similarity in the arrangement of histones in all of them. These nucleosomes are formed by extension of the

502

8. G. Bavykin

et ai.

min)

DN

-

-

-

176

-

-

-I

-

-

0

c

b

d

f

Figure 3. Electrophoretic separation nuclease digests of (a) yeast nuclei and conditions were as for Fig. 1. Fractions free of histone Hl. DK, dinucleosomes. min.

9

h

(b:

lo)

under denaturing conditions (see Fig. I) of’ DNA fragments from micrococea; (b) chicken erythrocyte chromatin depleted of’ histones Hl and H5. Digestion 1 to 3 correspond, respectively, to mononucleosomes MN,,,, MN,,S and Mh’,,, Lanes d a.nd h contained DNA markers (see Fig. 1). Digestion time is given in

by 10 bp on one I)NA side for DNA in MN,,, MN15, and by 10, 20 or 30 bp on both DNA sides for other nucleosomes (Fig. 7). Analysis of the data in Figure 6 and Table 1 shows that MNZS5 is actually a mixture of particles arranged in a similar manner to the core particle alt#hough the DNA in each particle is longer than the core particle DNA on two ends by 10 and 80, 20 and 70, 30 and 60 or 40 and 50 bp (see Fig. 7). Indeed, for different histone fractions the size of histone-bound DNA

fragments differs discretelv by ILO, 20, 30 , or 80 nucleotides. For example, in Figure 6(a) and (b) one can see on the diagonals of histone H4 and its dimer (H4), more or less distinct spots of discrete DNA fragments: H4 (85, 95, 105, 115, 125) and H4 (118, 128, 138, 148). These fragments are derived from H4 (55, 65) and H4 (88) fragment’s of the core nucleosome elonga,ted by 30, 40, 50 and 60 nucleotides. So, MN235 corresponds to nucleosomes in which 40 and 50 or 30 and 60 bp are linked to the core particle

Figure 4. Kucleosomes in sea urchin sperm nuclear digests. (a) Separation in 7 o/b polyacrylamide gel (Varshavsky- rb al., 1976). Fractions 1 to 4 correspond to MN,,5, MN1,,, Mi’J235 and MN,,,. Dr\‘, dinucleosomes. (b) Determination of the UK-A length in these nucleosomes obtained from micrococcal nuclease digests of nuclei. The nuoleosomes were extracted from preparative gels (Varshavsky et al., 1976), treated with Pronase; the DNA was denatured and separated by electrophoresis in t,he presence of 7 M-urea (NJaniatis et al., 1975). Lanes c, b, g and e contained DNA from fractions 1 to 4; respectively. J,anes a, d, f and h contained DNA ma,rkers (see Fig. 1).

ii

rr, SUO!+SDJJ

504

S. G. Bavykin et al.

Figure 5. Two-dimensional electrophoretic separation of “P-1abelled DNA fragments of nucleosomes MX145 arltl MNISS crosslinked to histones. Electrophoresis of crosslinked DNA-histone complexes in the first dimension (from left to right) was conducted in 17% polyacrylamide slab gel (200 mm x 200 mm x 0.6 mm); electrophoresis of DNA in the second dimension (from top to bottom) in 15% polyacrylamide slab gel (300 mm x 400 mm x 1 mm) after crosslinked histones had been digested by Pronase directly in the gel. Broken lines indicate t,he position in the gel and numbers show approximate sizes of markers (see Fig. 1). 32P-labelled DNA fragments crosslinked to histones H2A, H2B, H3 and H4 fall on the separate diagonals shown in the upper part of the Figure. (a) MNIS,; (b) a short exposure of the gel in (a); (c) the core nucleosome MN,,, (Bavykin et al.; 1985a). Positioning of DNA fragments crosslinked to histones is indicated by continuous lines.

DNA on the two sides (Fig. 7). In Figure 6(a) and (b) one can also see weak spots of H4 (65, 75) and H4 (158, 168), which demonstrate the existence of a small amount of such MN,,, p articles in which the nucleosomal core DNA is elongated on both sides by

20 and 70 or 10 and 80 bp, respectively. This follows also from the distribution of intensities of t,he DNA fragments in the diagonals of histones H2A, H2B and H3 (see Fig. 6(c) and Table 1). Besides, a very weak spot H4 (55) in Figure 6(b) and (d) suggests

l~h~llrvl h:ntenn hn,,mJ nl\T4, lllYY”lll, U”UllU YI,A I f11ranments -b-~--‘-~derive? from MN235 particles. (a) and (b) Electrophoresis in the first and Figure 6. ‘Two-dimensionai eieetrophoreiic separation 011‘32n I -IaJuFjllDu second dimensions in 15% polyacrylamide gel (200 mm x 400 mm x O-6 mm and 300 mm x 400 mm x 1.0 mm). Histone dimers are shown in parenthesis. (c) and (d) Electrophoresis in the first dimension in 17 y0 polyacrylamide gel and in the second dimension in 15% polyacrylamide gel. The gel sizes were as in (a) and (b). Before electrophoresis in (a) and (c), the samples were treated with I,lO-phenanthroline (see Experimental Procedures}. For other designations see Fig. 5.

8. G. Bavylcin

506

Spacer

Core

i bp)

et ai.

(146

bp)

Spacer

MN,,, #a 3.

50

1MN 185

40

60

70----------80------..-------5

-w-m-_------

-~-~~~-~-~-~~3'

HZB’ H42H2!3* H2AZ H4’ H4’ H3’ H3* H3’ H2BZH2BZ I I I I I I / I 50 60 70 80 90 100 110 120

HZB’ H2B’ I I I 20 30 40

I IO

0 146 140 130 120 110 100 SO 80 70 60 50 1 I / I I I H3’ H3’ H2E’H2B’ H3’ H3’ H3’ H4” H4’ H2A’ HZA’ HZB’ H4’ H2BZ

30

H2A’ , H3’ H3 / I ,3’-,O 130 140146

---20

MkcJ

,f, , , , , , ( , 40

30

HZB’

20

IO

0

H2B2

166

MN,,,

“‘-:-i

I

I

I

I

I

I

/

I

-15 -14 -13 -12 -II -10 -9

-8

I

!

I

-7 -6

-5

I

I

I

I

I8

-4 -3 -2

-I

0

I

I

/

L

i

8

I

I

s

2

3

4

5

6

7

8

9

I

/

IO II

I

I

I

/

12 13 14 !5

Figure 7. The primary organization of the symmetric core nucleosome 3I\iIN145 (Bavykin et al., 1985rc) aml mononucleosomes MN,,S (see Fig. 5) MNlG5, MN,,,, MN,,, (Karpov et aZ., 1982) and MKZSS (see Fig. 6). In the cent,ral part of Figure, there can be seen the main histone-binding sites on the core nucleosome double-stranded DNA (see Bavykin et al., 1985a; Table 1). The scheme explains the multiplicity of variants for every nucleosome. For example. lMNl5, is derived from the core particle by a lo-bp extension of the core DNA at one end; the series of MN,,, nucleosomes by adding spacers of 10 and 80, or 20 and 70, or 30 and 60, or 40 and 50 bp on 2 sides of the core DNA. Frequent particles are marked by continuous lines, rare particles by broken lines. Each DNA strand is measured in nucleotides from the 5’-end for the core or in superhelical turns (about 10.4 bp per turn) for ail other particles (bottom line), taking the dyad axis as the starting pointy the presence of trace amounts of MNz3, particles in which the 90-bp spacer DNA is located on one side of the core nucleosome. It should be stressed here that a discrete length of histone-crosslinked DNA fragments and their periodicity of about 10 bp in MN,,,, MN,,,, (Belyavsky et aE., 1980; Karpov et MNl,s> MN,,5 al., 1982; Bavykin et al., 1985a), MN,,, (Fig. 5) and (Fig. 6) have been shown (Bavykin et al., MR’,,, 1985b) not to be a consequence of any selectivity of crosslinking or micrococcal nuclease action on specific nucleotide sequences. The existence of the same discrete periodicity in the size of DNA fragments crosslinked to histones in nucleosomal cores and HI -containing nucleosomes actually indicates a similarity in the organizat’ion of bhe core particle DNA and spacer DNA and in their interaction with histones. In Figure 6(a) and (b) the diagonals of H3 and H2B are merged into one diagonal containing, among other regions, an area of 200 to 235 bp DNA fragments crosslinked to these histones. In Figure 6(c) the two diagonals are resolved. Owing to this, we were able to determine separately the crosslinking sites for histone H3 and those for histone H2B at the ends of the DNA spacer in MNzs5. Interaction between histone H3 and spacer DNA has been shown for LMN,6s, MNiY5 and MN1s5 (Belyavsky et al., 1980; Karpov et al., 1982). As for the spermal variant of HZB, its binding to spacer DNA may be connected with the presence of a positively charged insertion within the N-terminal region (Strickland et al., 1978), which seems to affect the chromatin organization in sea urchin sperm nuclei and MNZ3, nucleosomes. On the histone H3 diagonal (Fig. 6(d)), there are 45 to 65 some weakly crosslinked DNA fragments, nucleotides long. This weak interaction of H3 wit,h the first five nucleotides in the 5’-terminal DNA of the core particle was not observed for MN,,,, since

our electrophoretic fragments shorter

system could not separate than 20 nucleotides.

DSA

. Discussion (a) Regular

organization

of spacer DNA

Upon hydrolysis of chromatin and nuclei wit’b micrococcal nuclease a series of nucleosomes was found with t’he same linear arrangement of the core histones along the core particle DNA and with the length of spacer DNA varying from 10 bp in MNlj5 by a discrete step of about to 90 bp in MN,,, 10 xn bp, where n = 0, 1, 2, 3, . . . 9. The BO-bp periodicity seems to refer to the whole of spacer DNA as observed for internucleosomal DNA in dinucleosomes (Karpov et al., 1982; Strauss & Prunell, 1983) and in chromatin from various carcinoma (Fig. 2(c)), mouse ascites sources: Drosophila embryos (Karpov et al., 1982): rat liver (Belyavsky et al., 1980), chicken erythrocytes and sea urchin sperm (Figs l(b) and 2(a)), with the lengths of internucleosomal DNA being about 30, 30, 50, 70 and 90 bp, respectively (Bakayev et al.? 1977; Kornberg, 1977). The same periodicity, if somewhat limited, is also t,ypical of nuclei and chromatin depleted of histone HI (Fig. 3). The lo-bp regularity in t,he cleavage of the core particle DKA by different nucleases, including micrococcal nuclease (Kornberg, 1977; Felsenfeld, 1978; McGhee & Felsenfeld, 1980; Prunell, 1983) is a consequence of the accessibility to nuclease action of the outer surface of the nucleosomal DNA that is wrapped around the histone octamer, while the internal surface of the DNA is laterally shielded by histones (Mirzabekov & Rich, 1979; Richmond et al., 1984; Lambert & Thomas, 1986). The same periodicity in the splitting of DNA by nucleases was found when the DNA was laterally shielded by a Aat

Chromatin:

Organization

of

surface, for example by crystals of mica or calcium phosphate (Rhodes & Klug, 1980). At present, three basic models have been put forward to account for the organization of spacer DNA in chromatin (Felsenfeld & McGhee, 1986): (1) DNA is coiled in a superhelix similarly to the DNA of the core particle (Finch & Klug, 1976; Worcel & Benyajati, 1977; McGhee et al., 1980; Lee et al., 1981; Karpov et al., 1982; McGhee et al., 1983); (2) the DNA is extended and links the adjacent nucleosomes arranged as a zigzag (Worcel et al., 1981; Staynov, 1983; Lesters & Staynov, 1983; Woodcock et al., 1984; Makarov et al., 1985; Williams et al.; 1986; Bordas et al., 1986b); (3) the DNA is irregularly folded inside a solenoid (Thoma et al., 1979; Butler, 1984; Subirana et al., 1985; Widom & Klug, 1985; Sen et al., 1986). The observed lo-bp periodicity in splitting of the core particle DNA and spacer DNA by various nucleases is a strong argument in favour of the concept that chromatin DNA forms a continuous superhelix. At the same time, these data do not fit well with two other models of spacer organization. Neither the model of zigzag location of adjacent nueleosomes linked by extended DNA fragments nor the model of irregularly arranged spacer DNA are capable of explaining the lo-bp periodicity in digestion of nucleosomal and internucleosomal DNA by DNase I (Nell, 1974; Lohr & Van Holde, 1979; Prune11 et al., 1979), or the formation of a series of mononucleosomes from MNlb5 to MN2s5 and dinucleosomes with lo-bp periodicity (Karpov et al., 1982; Strauss & Prunell, 1983; see also Figs 1 and 2). (b) Interaction with

of histones HI, H3 and HZB DNA

internucleosomal

Histones of the core nucleosomes are capable of organizing 10 bp of the spacer DNA on both sides of the nucleosome. This is shown by the formation of MN155 and MNle5 particles upon micrococcal nuclease digestion of yeast nuclei free of histone Hl or chicken erythrocyte chromatin depleted of Hl and H5 (Fig. 3). The interaction between core histones and spacer DNA was demonstrated in numerous experiments (Thoma et al., 1979; Ruiz-Carillo et al., 1979; Weischet et al., 1979; Marion & Roux, 1980; Allan et al., 1982; Morse & Cantor, 1985; Smirnov et al., unpublished results). To organize longer stretches of internucleosomal DNA, histone Hl is needed. The fact that histone Hl interacts with spacer DNA has been shown by many investigators (Varshavsky et al., 1976; No11 & Kornberg, 1977; Simpson, 1978; Thoma et al., 1979). The data on DNA-histone crosslinking suggest that histone Hl binds not only to the terminal sites in a chromatosome but also, to a lesser degree, to the core DNA (Sperling & Sperling, 1978; Belyavsky et al., 1980). In has been shown recently (Pruss et al., 1988) that the globular part of the histone H5 molecule is the main region of crosslinking to DNA in nucleosomes, isolated chromatin and chromatin within nuclei. It

Nucleosomes and Spacer DNA

507

transpires, from our data on the primary organization of MN2s5, that Hl in sea urchin sperm n.uclei is strongly bound to the central region in the spacer DNA and not so strongly to some other sites on the spacer and nucleosomal core DNA (Figs 6 and 8). If one supposes that histone Hl interacts with the sites close to the histone octamer (for example, within about 20 bp on both its sides), then in the case of MNzs5, whose terminal DNA regions are mostly from 30 to 60 bp long (Fig. 7), one would see the strongest Hl-binding sites in the area,3 10 to 60 b and 175 to 215 b from the DNA g-ends with maxima within 25 to 45 b and 190 to 210 b, respectively (Fig. 8(b) and (c)). However, this disagrees with our experimental data. Actually, we observed preferential binding of Hl in the region of bases 195 to 235 (Figs 6(a) , 8(d) and (e), and Table 1) with a maximum close to the 235th base (Figs 6(b) and 8(a’); see also the position of the spot (Hl), in Fig. 6(a)) and no binding in the region from the 30th to the 60th base from the 5’.end (Fig. 8(a)). It was further shown by DNA-histone crosslinking that histone H3 interacts also with spacer DNA (Belyavsky et al., 1980; Karpov et al., 1982). In the tetramer (H3, H4),, histone H3 is localized in the central part of the folded nucleosomaJ DNA, from -2.5 to + 2.5 sites (see Figs 7 and 9), which region is close to the -5.5 to - 10.5 spacer DNA sites on the adjacent turn of the DNA superhelix in the lower part of the nucleosome (Fig. 9). The spatial proximity makes possible simultaneous binding of the same H3 molecule to both DNA fragments. This conclusion was supported by X-ray analysis of nucleosomes (Bentley et al., 1984). Similarly, the binding site of histone HZB on the core particle DNA from - 1 to -5,5 is brought close to the - 9 to - 13,5 region of the spacer DNA. So, if the spacer is in a superhelical state, histone H3 can be involved in the spacer organization up to the 35th nucleotide from the end of the core particle DNA, and histone H2B, in its organization from the 20th to the 65th nucleotide. The sea urchin sperm histone H2B contains a positively charged insertion of about 20 amino acid residues in its N-terminal part (Strickland et al., 1978). This feature of the sperm H2B may be responsible for the extremely long spacer of this chromatin (Zalenskaya et al., 1981), although other factors, such as structural peculiarities of sperm H2A and Hl (Sures et al., 1978; Wouters et al., 1978; Allan et a:Z., 1980; Strickland et al., 1980), could also play a role. The observed covalent binding of H3 and H2B to the spacer DNA in MN,,, from sea urchin sperm is in line with the above conclusion, and the model in Figure 9 can provide a structural explanation for the role of H2B in the organization of this unusually long spacer. It should be noted here that the interaction of H2B and H3 with the spacer would be difficult to explain within a model in vvhich the spacer is extended beyond the nucleosome and links the neighbouring nucleosomes in a zigzag construction. If this were the case, the H2B and H3 molecules belonging to the histone octamer would

.-- ,-w-p .---I

r---* I --c ---

I I I J cm*m-r -r -

-Y* c--c x---c ---L -

Chromatin:

Organization

of Nucleosomes

Figure 9. A model for the arrangement of histones Hl, H2B and H3 on super-helical nucleosomal and spacer DNA in sea urchin sperm nuclei. The core DNA together with the spacer (for the sake of simplicity, the latter is shown on only one side of the core DNA) form a lefthanded superhelix containing about 80 bp per turn (Finch et al.; 1977; Klug et al., 1980). Long thin arrows mark the sites of additional interactions of the core histones with the spacer DNA on the adjacent superhelical turn. A short filled arrow indicates the centre of the spacer DNA. Histones interacting with the core DNA and spacer DKA are designated by c and s, respectively. Distances along the nucleosomal DNA are given by numbers of superhelical turns from the dyad axis [see Fig. 7).

509

and MN,,S particles (Figs l(a) and 4), the MN,,, latter two being free of histone Hl. This may have something to do with the 240 bp-long nucleosomal repeat in this chromatin. If the nucleosomal core DNA together with the spacer DNA forms a continuous superhelix (Finch & Klug, 1976; Worcel & Benyajati, 1977; McGhee et al., 1980; Lee et al., 1981; Karpov et al., 1982; McGhee et al., .1983) in which 80 bp make one superhelical turn as in the core particle (Finch et al., 1981; Bentley et al., 1981), then the 240-bp nucleosomal repeat must contain three superhelical turns. In this arrangement, splitting by nucleases at 80 bp periodicity of the sites that are outwardly exposed and accessible to nuclease attack on the spacer DNA in the solenoidal structure of chromatin (Finch & Klug, 1976) with 80-bp periodicity would result in a preferential cleavage of MN23s-MN245 (Figs 7 and 9). Within the nucleosomal repeat of the spermal chromatin, the histone octamer can occupy a number of positions on DNA, each one shifted by 10 bp from the next one, although the central positions are much preferred. The arrangement of the core particles in three positions 10 bp apart has been shown for the nucleosomal repeat of Drosophila embryos chromat#in (Karpov et al., 1982); this correlates with the variable spacer length in this chromatin always remaining a multiple of IO bp (Karpov et aZ., 1982) as was shown allso to be the case for rat liver chromatin (Strauss & PruneIl, 1983). In’ contrast to this, digestion by microccocal nuclease of various other kinds of chromatin -in nuclei, for example from Drosophila embryos, mouse ascites carcinoma and chicken erythrocytes, produces a whole spectrum of discrete nucleosomes, from to MN2i, (Fig. 2(b)). The corresponding MN,e repeat lengths in chromatin from the above sources are 185; 180 and 215 bp, i.e. they are not divisible by 80 bp and thus different outwardly exposed sites on the spacer DNAs must be accessible to nuclease attack

out along the linearized spacer for as much as 30 A (1 A = 0.1 nm) or even more. This is difficult to imagine, especially because 90% of DNA-histone crosslinking under our experimental conditions was provided by histidines (Ebralidse et al., 1988; Nacheva et al., 1989), which are absent in the N and C fragments of the core histones (Wells, 1986). At the same time, any interaction that would hinder superspiralization of the spacer DNA (such as expulsion of histone Hl , absence of Mg2+, modification of other histones, etc.) would lead to linearization of the spacer DNA with the consequent appearance of a zigzag arrangement of adjacent nucleosomes in chromatin.

and #pacer DNA

in the solenoid

structure.

be extended

(c) Arrangement of nucleosomes in chromatin superstructure A feature of hydrolysis of sea urchin sperm nuclei by micrococcal nuclease is the formation of a limited

set of nucleosomes

containing

MN2s5,

MN,,,,

(d) Differences in structure between chrclmatin in nuclei and isolated chromatin

Interestingly, Hl-free particle MN155 is formed in appreciable amounts only upon hydrolysis of nuclei or condensed chromatin and not unfolded chromatin from any source that we have studied. A special role in this process is played by divalent cations (Fig. 2, lanes rn; o and s). As has been shown earlier, irreversible alterations in the regular chromatin structure have been demonstrated in a low ionic strength buffer, namely at 62 mM-EDTA, by Brust & Harbers (1981), Williams et al. (1986) and Dimitrov et al. (1987). Reversibility of these distortions depends on the length of the spacer DNA (Dimitrov et al., 1987). We think that isolation of involve certain structural chromatin must rearrangements that seem to be connected with a partial rearrangement of histone Hl, when EDTA is used in this procedure.

8. G. Bavykin et al.

510

One should also note that the series of mononucleosomes that was obtained from micrococcal nuclease digests of unfolded chromatin isolated from all studied sources (see Belyavsky et al., 1980; Karpov et al.: 1982; and Figs l(b) and 2(a)): namely, rat liver (Fig. 2, lane c), Drosophila embryos (Fig. 2, lane a), sea urchin sperm (Fig. l(b)), mouse ascites carcinoma (Fig. 2, lane m) and chicken erythrocytes (Fig. 2, la,ne e), were actually indistinguishable. Unlike unfolded chromatin, the hydrolysis patterns of nuclei clearly depended on the source: the longer a nucleosomal repeat the great,er the number of mononucleosome fractions (except for urchin sperm) (Figs l(a) and 2(b)). In isolated nucleosomes; as well as in unfolded chromatin, the hydrophobic central part of the same HI molecule can bind with 10 bp at both ends of the DNA in the chromatosome MN1,,, thereby shielding them against nuclease attack (Varshavsky et al., 1976; No11 & Kornberg, 1977; Simpson, 1978; Thoma et al., 1979; Alan et al., 1980). The appearance of an Hl-free particle, MN1,,, in nuclear chromatin or condensed chromatin digest may correspond to depletion of histone Hl from the 10 bp DNA on one side of the chromatosome. We suppose that within nuclei histone Hl interacts with spacer DNA at just one side of the nucleosome so that its globular part interacts with the central part of the spacer, with additional contacts being localized on spacer DNA of neighbouring nucleosomes or on adjacent superhelical turns of the solenoid. Thus, by fixing the neighbouring solenoid turns histone Hl can stabilize the solenoid structure as a whole. Upon decondensation of chromatin, Hl is redistributed in such a manner that its globular part becomes bound to the spacers on both sides of the same nucleosome, which leads to the appearance of chromatosomes and destruction of the lo-nm fibrils with a zigzag arrangement of chromatin. Additionally, this leads to partial decondensation of the solenoid structure and thus the spacer DNA becomes available for splitting by nuclease along its entire length. As a consequence, a limited set of particles in sea urchin sperm nuclei digests is replaced by a much wider spectrum of nucleosomes in digests of unfolded chromatin. Alternatively, the presence of asymmetric MN, 55 in nuclear digests could indicate some structural asymmetry in the nucleosomal core pa,rticle that can be induced by factors other than histone Hl; for example, by the polarity of the solenoid in replication and transcription. Moreover, it has been shown recently that in nuclear chromatin of erythrocytes interaction of histone H5 with DNA occurs in a different manner than in unfolded chromatin and nucleosomes (Pruss et al., 1988). as evidenced by difference in crosslinking to DNA of Thrl and His62. References Allan, J., Hartman, P. G., Crane-Robinson, C. & Aviles, F. X. (1980). Wature (London), 288, 675-679.

Allan, J., Harborne, N., Rau, D. C. & Gouid, H. (1982). J. Cell Biol. 93, 285-297. Bakayev, 17. V.. Bakayeva, T. G. & Varshavsky, A. J. (1977). Cell, 11, 619-629. Bavykin, S. G., Usachenko, S. I., Lishanskaya, A. h.. Shick, V. V., Belyavsky, A. V., Undrit,sov, I. 31.. Strokov, A. A.: Zalenskaya, I. A. & Mirzabekov. A. D. (1985a). N&. Acids Res. 13, 3439-3459. Bavykin, S. G., Usachenko, S. I.; Shick, V. V., Belyavsky, A. V.: Lishanskaya. A. I., Undritsov, I. M., Zalenskaya, I. A. & Mirzabekov, A. D. (198%). Mol. Biol. (USSR), 19, 144-161. Belyavsky, A. V., Bavykin, S. G.. Goguadse, E. G. & Mirzabekov, A. D. (1980). J. Mol. Biol. 139, 519-536. Bentley, G. A., Finch, J. T. & Lewit-Bentley, A. (1981). J. Mol. Biol. 145. 771-784. Bentley, G. A., Lewit-Bentley, A., Finch, J. T., Podjamy, A. D. & Roth, M. (1984). J. Mol. Biol. 176, 55-75. Bordas, J., Perez-Grau, I,., Koch, M. H. J., Vega, M. C. 85 Nave, C. (1986a). Eur. Biophys. J. 13, 157-173. Bordas, J., Perez-Grau, I,., Koch, M. H. J., Vega, M. C. & Nave, C. (1986b). Eur. Biophys. J. 13, 175-185. Brust, R. & Harbers, E. (1981). Eur. J. Biochem. ll7! 609-615. Burlingame, R. W., Love, W. E.; Wang, B.-C., Hamlin. R.; Xuong, N.-H. & Moudrianakis, E. N. (1985). Science; 228, 546-553. Butler, P. J. G. (1983). CRC Crit. Rev. Biochem. 15, 57-91. Butler, P. J. G. (1984). EHBO J. 3, 2599-2604. Cert,a, U., Colavito-Shepanski, M. & Grunstein, M. (1984). Nucl. Acids Res. 12: 7975-7985. Crane-Robinson, C., Staynov, D. Z. & Baldwin, J. P. (1984). Comm. Mol. Cell Biophys. 2, 219-265. Davie; J. R. (1982). An&. Biochem. 120, 276-281. Dimitrov, S. I., Smirnov, I. V. & Makarov, V. L. (1987). Mol. Biol. (USSR), 21. 1400-1410. Ebralidse, K. Ii. (1988). Nature (London), 331, 365-367. Felsenfeld, G. (1978). Nature (London), 271; 115-122. Felsenfeld, G. & McGhee, J. D. (1986). Cell, 44. 375-377. Finch, J. T. & Klug; A. (1976). Proc. Nat. Acad. Sk., U.S.A. 73, 1897-1901. Finch, J. T., Nell: M. & Kornberg, R. D. (1975). PTOG. Nat. Acad. Sk., U.S.A. 72. 3320-3322. Finch, J. T., Lutter, L. C., Rhodes, D., Brown, R. S., Rushton, B.; Levitt, M. 8i Klug, A. (1977). il’ature (London), 269. 29-36. Finch, J. T., Brown, R. S., Rhodes, D., Richmond, T., Rushton, B.; Lutter, L. C. & Klug, A. (1981). j. &lob. Biol. 145, 757-770. Karpov, V. L., Bavykin, S. G.. Preobrazhenskaya, 0. V., Belyavsky, A. V. & Mirzabekov, A. D. (1982). Nucl. Acids Res. 10, 4321-4337. Klug, A., Rhodes, D.; Smith, J.. Finch, J. T. & Thomas, J. 0. (1980). Nature (London), 287, 509-516. Kornberg, R,. D. (1977). Annu. Rev. Biochem. 46. 931-954. Kuehl, L. (1979). J. Biol. Chem. 254, 7876--7881. Lambert, S. F. & Thomas, J. 0. (1986). Eur. J. Biochem. 160; 191-201. Lee, K. S., Mandelkern, $. & Crothers, 1). M. (1981). Biochemistry, 20. 1438-1445. Lest’ers, I. 8 Staynov, D. Z. (1983). Int. J. Biol. Macromol. 5, 310-311. Levina, E. S., Bavykin, S. G., Shick, V. V. & Mirzabekov, A. D. (1981). Anal. &o&em. 110, 93-101. Levy. S., Sures, I. & Kedes, L. (1982). J. Biol. Chem. 257> 9438-9443.

Chromatis:

Organization

of Nucleosomes

Lilley, D. M. J. & Pardon, J. F. (1979). Annu. Rev. Tenet. 13, 197-233. Lohr, D. & Van Holde, K. E. (1979). Proc. Nat. Acad. Sci., U.S.A. 76, 6326-6330. Makarov, V., Dimitrov, S., Smirnov, V. & Pashev, I. (1985). FEBS Letters, 181, 357-361. Maniatis, T., Jeffrey, A. & van deSande, H. (1975) Biochemistry, 14, 3787-3794. Marion, C. & Roux, B. (1980). Biochem. Biophys. Res. Commun. 90, 535-541. McGhee, J. D. & Felsenfeld, G. (1980). Annu. Rev. B&hem. 49, 1115-I 156. McGhee, J. D., Rau, D. C., Charney, E. & Felsenfeld, G. (1980). Cell, 22, 87-96. McGhee, J. D., Nickel, J. M., Felsenfeld, G. & R,au; D. C. (1983). Cell, 33, 831-841. Mirzabekov; A. D. (1980). Quart. Rev. Biophys. 13, 255-295. Mirzabekov, A. D. (1986). Sov. Sci. Rev., D. Physicochem. Biol. 6; 33-75. Mirzabekov, A. D. & Rich, A. (1979). Proc. Nat. Acad. Ski., U.S.A. 76, 1118-1121. Mirzabekov, A. D., Shick, V. V., Belyavsky, A. V. & Bavykin, S. G. (1978). Proc. Nat. Acad. Sci., U.S.A. 75,4184-4188. Morse, R. H. & Cantor, C. R. (1985). Proc. Nat. Acad. Sci., U.S.A. 82, 4653-4657. Nacheva, G. A., Guschin, D. Yu., Preobrazhenskaya, 0. v., Karpov, V. L.; Ebralidse, K. K. & Mirzabekov, A. D. (1989) Cell, 58, 27-36. Noll, M. (1974). Nucl. Acids Res. 1, 1573-1578. Noll, M. & Kornberg, R. D. (1977). J. Mol. Biol. 109, 393-404. Odintsova, N. A.: Rozov, S. M. & Zalenskaya, I. A. (1989). Comp. Biochem. Physiol. in the press. Pederson, D. S., Thoma, F. & Simpson, R. T. (1986). Annu. Rev. Cell Biol. 2, 117-147. Prunell, A. (1983). Biochemistry, 22, 4887-4894. Prunell, A.: Kornberg, R. D., Lutter, L., Klug, A., Levitt, M. & Crick, F. H. C. (1979). Science, 204, 855-858. Pruss, D. V., Ebralidse, K. K. & Mirzabekov, A. D. (1988). Mol. Biol. (USSR), 22, 1108-1118. Rhodes, D. & Klug, A. (1980). Nature (London), 286, 573-578. Richmond, T. J., Finch, J. T., Rushton, B., Rhodes, D. & Klug, A. (1984). Nature (London), 311, 532-537. Ruiz-Carrillo, Q., Jorcano, J. L., Eder, G. & Lurz; R. (1979). Proc. Nat. Acad. Sci., U.S.A. 76, 3284-3288. Sen, D.; Mitra, S. & &others, D. M. (1986). Biochemistry, 25, 3441-3447. Shick, V. V., Belyasky, A. V., Bavykin, S. G. & Mirzabekov, A. D. (1980). J. MOE. Biol. 139,491-517. Edited

and Spacer DNA

511

Shick, V. V., Belyavsky, A. V. & Mirzabekov, A. D. (1985). J. Mol. Biol. 185, 329-339. Simpson, R. T. (1978). Biochemistry, 17, 5524-5531. Spadafora, C. & Geraci, G. (1975). FEBS Letters, 57, 79-82. Spadafora, C., Bellard, M., Compton, J. L. & Chambon, P. (1976). FEBS Letters, 69, 281-285. Sperling, J. & Sperling, R. (1978). Nucl. Acids R,es. 5; 2755-2773. Staynov; D. Z. (1983). Int. J. Biol. Macromol. 5, 3-9. Strauss, F. & Prunell, A. (1983). EMBO J. 2, 51-56. Strickland, M., Strickland, W. N., Brandt, W. F. &; Von Holt, C. (1974). FEBS Letters, 40; 346-348. Strickland, M., Strickland, W. N.; Brandt, W. F., Von Holt, C., Wittmann-Liebold, B. & Lehman-n, A. (1978). Eur. J. Biochem. 89, 443-452. Strickland, M., Strickland: W. N., Brandt, W. F., Von Holt, C., Lehmann, & & Wittmann-Lieboled, B. (1980). Eur. J. Biochem. 104, 567-578. Subirana, J. A., Munoz-Guerra, S., Aymami, J., Radermacher, M. & Frank, J. (1985). Chromcsoma, 91, 377-390. Sures, I., Lowry, J. & Kedes, L. H. (1978). Cell, 15; 1033-1044. Thoma, F., Keller, Th. & Klug, A. (1979). J. CeEZBiol. 83, 403-427. Uberbacher, E. C. & Bunick, G. J. (1985). J. BiomoZ. Strut. Dynam. 2, 1033-1055. Varshavsky, A. Ya., Bakayev, V. V. & Georgiev, G. P. (1976). Nucl. Acids Res. 3, 477-492. Weischet, W. O., Allen, J. R., Riedel, G. & Van Holde, K. E. (1979). Nucl. Acids Res. 6, 1843-1862. Wells, D. E. (1986). Nucl. Acids Res. 14 (Suppl.), r119-r149. Widom, J. & Klug, A. (1985). 43, 2077213. Williams, S. P., Athey, B. D., Muglia, L. J., Schappe, R. S.; Gough, A. H. & Langmore, J. P. (1986). Biophys. J. 49, 233-248. Wintersberger, U., Smith, P. & Letnansky, K. (1973). Eur. J. Biochem. 33, 123-130. Woodcock, C. L. F., Frado, L.-L. Y. & Rattner, J. B. (1984). J. Cell Biol. 99, 42-52. Worcel, A. & Benyajati, C. (1977). CeEZ,12, 833100. Worcel, A., Strogatz, S. & Riley, D. (1981). Proc. Nat. Acad. Sci., U.S.A. 78, 1461-1465. Wouters, D.; Sautiere, P. & Biserte, G. (1978). Eur. J. Biochem. 90, 231-239. Zalenskaya, I. A.; Pospelov, V. A., Zalensky, A. 0. & Vorob’ev, V. I. (1981). Nucl. Acids Res. 9, 47:3-487.

by R. Laskey