Modulation of the Nucleosome Structure by Histone Acetylation

Eur. J. Biochem. 110,143-152 (1980) 0 by FEBS 1980

Modulation of the Nucleosome Structure by Histone Acetylation

Jurgen BODE, Karsten HENCO, and Edgar WINGENDER

Abteilung Molekularbiologie, Geselischaft fur Biotechnologische Forschung, Braunschweig-Stockhein1

(Received February 27/June 2, 1980)

A rapid procedure for the isolation of core particles from Chinese hamster ovary cells is described which permits measurements, usually at the day of their preparation. Particles of 145 f 5 base pairs, derived from interphase cells, will be compared with the analogue specimens from butyrate-treated cells, metaphase cells and a standard preparation from chicken erythrocytes. Butyrate cause an increase in the acetylation of histones H3 and H4, which induces alterations of the interhistone and histone-DNA interactions. Changes in the interhistone contacts, correlated to an extension of a-helical segments, lead to an altered accessibility of the H3 cysteine side-chains and to a different histone displacement by protamines. On the other hand, histone-DNA contacts are loosened in parts and this is particularly evident from the changes in the premelting region of a thermal- denaturation profile.

In the structural element of chromatin two each of histones H2a, H2b, H3 and H4 form an octameric protein core which is surrounded by a 140-base-pair segment of DNA. These ‘core particles’ are linked by a variable segment of DNA to reveal a repeating subunit structure (nucleosome) [l]. Specific lysine side-chains within the highly basic aminoterminal regions of histone H3 (residues 9, 14, 18 and 23) and histone H4 (residues 5, 8, 12 and 16) are subject to rapid acetylation and deacetylation reactions [2]. H3 and H4 have been highly conserved in evolution and play the predominant role in organizing the DNA [3]. It is therefore likely that these modifications cause distinct conformational changes, which are related to processes such as their deposition on DNA [4], their removal during spermiogenesis [5] and the transcriptional activity of chromatin districts [2]. Since the steady-state level of monoacetylation (30 - 40% of H3 and H4) is much larger than the amount of histones estimated to be involved in active genes, it has been suggested that only a ‘hyperacetylation’ might bring about the required structural changes [6]. Until recently, chemical acetylation provided the only access to substantial fractions of acetylated chromatin [5 ] . The discovery by Riggs et al. [7] that sodium butyrate, a known inducer of cellular responses [8], causes an accumulation of highly acetylated histones has initiated an increasing number of investigations.

Abbreviations. PhMeSOlF, phenylmethylsulfonyl fluoride ; CD, circular dichroism.

Butyrate has now been found to act via a non-com- petitive inhibition of deacetylase to stabilize histones which carry several acetyl groups at the correct sites [9]. We have taken advantage of this effect to isolate core particles from butyrate-treated Chinese hamster ovary cells. These specimens will be compared with control preparations from the interphase and metaphase of these cells and from chicken erythrocyte nuclei. Particular emphasis will be placed on the rapid succession of preparation steps and physico- chemical measurements, because of the known dete- rioration of core particles [lo], especially the increased sensitivity of acetylated chromatin towards intrinsic proteases [l l] . The results, elaborated by several approaches, will add evidence to the conclusions by Vidali et al. [12], Nelson et al. [13] and Simpson [14], who demonstrated an altered internal structure of acetylated nucleosomes, mainly by an increased sensitivity towards DNase I.

MATERIALS AND METHODS Cell Growth and Chromatin Isolation

Chinese hamster ovary cells were grown as a monolayer in 1200-nil stoppered Roux bottles containing 100 ml M 199 medium with 15 % newborn calf serum. The M199 formula contained Hank’s salts and 20 mM N-[2-hydroxy-l, 1-bis(hydroxymeth- yl)ethyl]glycine (Tricine) pH 7.6; it had been enriched with minimum essential vitamins. At a cell number

144 Nucleosome Acetylation

of about 107/bottle, loosely attached cells were washed off with medium and the monolayer was treated by one of the following procedures. For the preparation of interphase chromatin, the medium was replaced by 10 ml cold 10 mM NaCI, 1.5 mM MgClZ, 10 mM Tris . HCl, pH 7.4, which had been adjusted to 0.2 mM PhMeSOzF immediately prior to use (buffer A). After a 30-min incubation at 472, the rounded cells were scraped off and centrifuged (1000 xg, 10 min) The pellets collected from three bottles were homo- genized (Virtis homogenizer, setting 50, 2 s) in 10 ml 0.25 M sucrose, 3 mM CaC12, 10 mM Tris . HCl, which had also been supplied with PhMeSOzF (buffer B), followed by centrifugation at 1000 x g. After two additional washings in buffer B, the nuclear pellet was dispersed in 10 ml digestion buffer C (0.25 M sucrose, 0.3 mM CaClZ, 0.2 mM PhMeSOzF, 5 mM Tris . HC1, pH 8.0) by magnetically stirring for 30 min at 4 "C. The chromatin, which was pelleted after this treat- ment, was redissolved in 500 p1 C and degraded as described below.

For the preparation of hyperacetylated chromatin, Chinese hamster ovary cells were exposed at the above density for 20 h to M 199 medium containing 7 mM sodium butyrate. After this addition, the cell number remained essentially unchanged. All further steps were performed as above but with 7 mM sodium butyrate additions to buffers A, B and C.

For the preparation of metaphase chromatin, cells were grown in M 199 media including 480 mg EGTA/l to complex most of the CaZ+ ions. At the above cell number, a fresh medium containing 0.06 pg/ml colce- mide was provided. This was repeated after 1 h. Mitotic cells were then harvested by shaking at 3-h intervals up to 9 h. Mitotic indices were determined by the aceto-orcein method [15] and were found to be in the order of 80-90 %. Cells were transferred to buffer A containing 50 mM NaHSO3 and then processed in buffers B and C, to which the same addition had been made.

Preparation of Chinese Humster Ovary Core Particles

Each of the three above chromatin preparations in 500 p1 of the respective digest buffer C was pre- warmed to 37°C for 10 min prior to the addition of 100 units of micrococcus nuclease. After 2 min (hyperacetylated chromatin) or 5 min (interphase and metaphase chromatin) the reaction was stopped on ice by adjusting the media to 1 mM NaEDTA, 0.2 M NaCI. The entire suspension was layered on the top of a freshly packed column with 5rnl Bio-Gel A5m and eluted at the same ionic strength. The emerging monomeric fractions were free from the linker histone (Hl) as it binds preferentially and cooperatively to the higher-molecular-weight species at quasiphysio- logical ionic strength [10,16]. In cases where some

turbidity emerged near the void volum it was easily removed by centrifugation. Fractions from the monomer range were subjected to a rechromatography in 1 mM Na-EDTA, 0.4 M NaCl on a similar column of Bio-Gel A5m to remove the remainder of accom- panying non-histone proteins. When feasible, experiments were performed directly on the 0.4 M NaCl eluates at the day of their preparation; in one example the effect of non-histones was examined by comparing the properties of the eluates at 0.4 M and 0.2 M NaCl (Fig. 5). Thermal denaturation experiments required 2-3 days because of the dialysis steps required. Freezing was avoided as it was found to cause irreversible changes and some precipitation. The preparation of core particles from chicken erythrocytes and the methods of their characterization have been published [lo].

Analyses on Hyperacetyluted Chromatin and Core Particles

A DNA gel [lo] from the unfractionated digest of hyperacetylated chromatin reflects a major com- ponent with an average of about 145 base pairs, but there are also other monomeric species up to a size of 180 base pairs, and in addition some contribution of dimers (340 base pairs), trimers (550 base pairs) and higher oligomers. After the gel chromatography at 0.2 M NaCl, the only remaining specimen contained 145 & 5 base pairs and has been subjected to further characterization.

The protein composition of core particles has been analyzed on 14 - 17 acrylamide/sodium dodecylsulfate gradient gels, essentially as described by Laemmli [17]. If proteins were to be recovered, a 17-20% disulfide-containing gel was applied instead, which had been prepared with the cleavable crosslinking reagents bisacryloylcystamine [18,19]. For a colori- metric quantification, Coomassie-blue-stained protein bands were cut from the gel and eluted with a 75:25 mixture of 2-mercaptoethanol and 8 M urea, 0.4 M Tris . HCl, pH 8.4 [19]. Total protein contents of core particles were compared by the fluorescamine test [20], referencing the fluorescence readings at 490 nm to the absorbances at 260 nm as described previously [21]. A combination of these methods clearly showed that after gel chromatography at 0.4 M NaCl, all core particle species were indistinguish- able with respect to their protein content and histone composition. Between 0.4 M and 0.65 M NaCl, only the hyperacetylated sample suffered some further loss of protein due to a partial desorption of each of the four core histones.

For the determination of acetyl contents, the individual histone bands from the disulfide-containing dodecylsulfate gel were processed as recently described [19] and re-electrophoresed on acetic acid/6 M urea gels; the results have been summarized

J. Bode, K. Henco, and E. Wingender 145

H4 H4 H2a 2 10 H2a 2 10 J==Yrf+ _f_rft_

H4

I

A 0

r c .- c

H3 H3 c 10 H2b 10 H2b

5 m v

Q

H4 H4 H2a 4 3 2 1 0 H2a 4 3 2 1 0 * -

H3 H3 3 2 1 0 H2b 3 2 1 0 H2b

-hTl-- -

Fig. 1 . Analysis of histone H3 and H4 subfractions on acetic acidlurea gels. Histones H3 and H4 were enriched by electrophoresis on a cleavable dodecylsulfate gel and recovered as described [19]. A subsequent electrophoresis on an acetic acidlurea gel [19] led to a subfractionation according to the acetyl content as shown above. The correlation between the mobility and the degree of modification has been established by the incorporation of radioactivity [22]. (A,A’) Histones from chicken erythrocyte core particles; (B,B’) histones of Chinese hamster ovary core particles from interphase; (C,C’) histones of core particles from butyrate-treated Chinese hamster ovary cells; (D, D’) the same histones but isolated from whole chromatin. Subscripts mark the number of acetyl residues

Table 1 . The percentage of modified hisfones H3 and H4 in preparations of whole chromatin or the corresponding core particles (numbers in parentheses) Subscripts indicate the extent of a modification, i.e. H44 is a histone with four acetyl residues attached to it. All data have been derived spectrophotometrically by cutting stained protein bands from acid/urea gels and eluting them as described in Materials and Methods

-~ ~~ . .. - ~ _ _ % . _~

Chicken erythrocytes (52) (35) (1 3) (55) (35) (10)

In terphase ( > 60) (50) (41) ( 9) Chinese hamster ovary cells

Metaphase 15(45) 55(37) 30(18) Butyrate-treated 23(18) 28(26) 31(34) 18(22) 18(17) 30(27) 21(25) 17(19) 14(12)

in Fig. 1 and Table 1 together with the respective data from the other preparations to be investigated. With the exception of metaphase specimens, the degree of modification in the original chromatin is essentially maintained at the core particle stage. It has been established that, at metaphase, H3 modification is mainly due to a phosphate residue [23,24], which is easily lost by the action of phosphatases [24].

Fluorescence Measurements

N- Ip-(2-Benzimidazolyl)phenyl]maleimide was synthesized by the procedure of Kanaoka et al. [25]. All reactions were performed in 40mM sodium phosphate buffers of pH 5.5, where the reagent is specific for thiols [26]. The addition of thiols to the maleimide moiety triggers the fluorescence of the benzimidazolyl-


phenyl chromophore [25] at 374nm, which was observed with an excitation wavelength of 308 nm. Measurements were performed on a Schoffel RRS 1000 fluorimeter as described [26].

Circular Dicliroism

CD spectra were recorded on a Jouan Mark I1 Dichrograph using cuvettes of 1-cm path length; measurements were taken directly from the eluates of the second Bio-Gel A5m column at 0.4 M NaCI.

Thei-mal Denaturation

Melting curves were registered using a thermo- stated microquartz cell mounted in a Zeiss PMQ I11 spectrophotometer equipped with a micro-optic as described [lo, 271. Temperatures were raised linearly by 0.3"C/min. For some of the experiments, absorbances were stored on a punched tape to enable the generation of a continous differentiated melting profile on a computerized plotter. Prior to the measurements all samples had been dialyzed overnight against 1 mM sodium cacodylate, 0.25 mM NaEDTA, pH 8.0 and saturated with helium.

RESULTS AND DISCUSSION

The Thiol Content of Core Particle Species

With the possible exception of yeast [28], an H3 cysteine residue 110 has been preserved in evolution. In mammals more advanced than the rodents, an H3 serine in position 96 has generally been substituted by another cysteine. The presence of only a single cysteine in rodent cells has been concluded from the inability of this residue to form an intramolecular crosslink with another cysteine [29,30]. The presence of an equal number of free H3 cysteines in Chinese hamster ovary core particles and in those obtained from chicken erythrocytes is supported by Fig. 2 with the aid of a fluorogenic, thiol-specific reagent, N-[p-(2- benzimidazol yl)phenyl]maleimide.

Little is yet known about the biological role of the H3 cysteines and whether they become oxidized during any state of chromosomal activity. That a S-S linkage between the H3 molecules of a histone core is compatible with a correct nucleosome structure has been shown by reconstitution [31]. Older findings suggested a minimal content of free thiols in condensed chromosomes [32] and recently it has been found that the H3 phosphorylation observed at metaphase [33] may enhance the tendency of isolated H3 histones to dimerize [34]. These results opened the possibility that H3-dimers could add stability to metaphase chromosomes.

What would be the consequences of H3 dimers upon the results shown in Fig.2? In our procedure,

0 5 10 0 0 Time (rnin)

Fig. 2. Fluorimetric quantifkation ofthe thiols in various core particle species. 2 ml 1.4 M NaCI, 40 mM sodium phosphate (pH 5.5) were adjusted to a concentration of 1 pM N-[p-(2-benzimidazolyl)phenyl]maleimide. Reactions were started by the addition of 100 pl of the respective core particle stock solution. Stock solutions had an absorbance ( A w , ~ ) of 0.8. (0) Peracetylated Chinese hamster ovary core particles after gel chromatography at 0.2 M or 0.4 M NaCI; (0) interphase core particles after gel chromatography at 0.2 M and 0.4 M NaCl (cf. Materials and Methods). (0) Metaphase Chinese hamster ovary core particles and core particles from chicken erythrocytes

metaphase cells have been collected in 1.5 mM MgC12, 10 mM NaCl 10 mM Tris-HC1 (pH 7.4) containing 50 mM NaHS03 as a phosphatase inhibitor 1331. All subsequent media contained the same amount of the sulfite up to the chromatographic purification, where it was removed. By the action of this reagent, accessible S - S bonds would undergo sulfitolysis to yield R-S- and R-S-SO; [35], i.e. at most 50% of the total thiols should be titrated. As this is not the case, there is no indication as of an H3 dimeriza- tion at metaphase.

Fig. 2 also comprises a comparison of core particles at different stages of purification. As the presence of non-histones in the 0.2 M NaCl eluates has no influence upon the final fluorescence generated by N-[p-(2-benzimidazolyl)phenyl]maleimide, they appear to contain negligible amounts of cysteine. This observation opened the possibility of studying their influence upon the reactivity of the H3 thiols, which is a sensitive function of the nucleosomal conformation [26,36].

Analysis of Cysteine-110 Accessihitities us a Function of Histone Modijkations or the Presence of Non-histone Proteins

At a closer inspection the modification time courses of core particles from Chinese hamster ovary cells and chicken erythrocytes (Fig. 2) exhibit distinct differences despite of a common final fluorescence (Fm). These differences become amenable to an interpretation in

J. Bode, K . Henco, and E. Wingender 147

15

I

i 10 I

t f 0

5

0 10 20 30 40 50 60 Tirne,t(min)

Fig. 3. Labelling kinetics obiuined by iwo d$ftJreni modes ofrni,uing. (A) Core particles (40 nM H3 thiols) from chicken erythrocytes (@) or Chinese hamster ovary rnetaphase chromosomes (0) were incubated for 3 h in 1 M NaC1, 40 mM sodium phosphate, pH 5.5. The reaction was started by adding N-[p-(2-benzimidazolyl)phenyl]- maleimide to a final concentration of 1.5 pM. (B) Same concentrations. Core particles were pipetted to the dye at f = 0

a replot of kt versus t , according to the equations of a second-order reaction :

. with c i , c i , initial concentrations of the fluorogenic dye and core particles respectively, cxr concentration of reacted thiols at time t , and F I , Fm, fluorescence readings during or at the end of the reaction. Ob- viously, in this plot the slope equals the second-order rate constant (Fig. 3).

At the concentration of core particles chosen for Fig.3 (chicken erythrocyte specimens at 40nM of H3 thiols) the cysteine side-chains are completely inaccessible at NaCl concentrations below 0.8 M NaCl (see [26]). If a sample of core particles in 0.4 M NaCl is pipetted into a solution containing 1 M NaCl and dye, the resulting unfolding process is evident from an initial lag phase of a few minutes, during which there is barely an increase of fluorescence. As unfolding proceeds, the thiols become successively available for a reaction with the dye, up to a maximal rate constant. If the sequence of additions is reversed, i.e. the dye is pipetted to a pre-equilibrated solution of core particles in 1 M NaCl, there is an initial burst with H3 which has already dissociated from DNA and a second, slower phase, which is apparently governed by the dissociation rate constant of the H3-DNA complex (see [26]). As the second part of

the reaction obeys the same rate constant in both cases, this parameter was usually derived from the first approach.

Fig. 3 includes the corresponding data from metaphase core particles, which are seen to be identical with those from chicken erythrocytes. The massive loss of histone H3 phosphates during the preparation of chromatin [24] (Table 1) may contribute to the lack of any effect which could be ascribed to this particular modification.

A comparison of peracetylated core particles and such from interphase with the corresponding specimens from chicken erythrocytes is seen in Fig.4. The most obvious influence of peracetylation concerns the characteristic positive curvature found in the rate profiles at intermediate ionic strength (Fig. 4). Apparently, a significant portion of H3 thiols is accessible at the instance of mixing already, whereas in the chicken erythrocyte particles they have first to be liberated by unfolding (cf. Fig.4B and 3). At all ionic strengths, the traces for interphase exhibit intermediate characteristics, although the degree of H3/ H4 acetylation is rather comparable to the histones from chicken erythrocytes (Table 1). It is, therefore, tempting to speculate that conformational changes at the nucleosome level occur during the life-cycle of Chinese hamster ovary cells. The identical properties of nucleohistones from chicken erythrocytes and metaphase chromosomes (Fig. 3) would then reflect the inactive nature of the parent chromatins. In both cases, inactivation involves rather serious rearrangements leading to highly condensed states.

These conclusions may be extended by an inspection of the thiol reactivities (in terms of the second

order rate constants k ) as a function of the ionic strength (Fig. 5). A typical reactivity profile (trace A) consists of an increasing part (0.8 - 1.4 M NaCl), which signifies the ionic-strength-dependent opening and dissociation of core particles [26]. A decreasing branch beyond 1.4 M NaCl is superimposible with the corresponding function from the pure core histones in the absence of DNA and it is controlled by the tightening of the core histone complex at higher ionic strength [26]. Trace A corresponds to core particles from chicken erythrocytes and from Chinese hamster ovary metaphase chromatin. The profile derived from interphase particles (B) exhibits related though less pronounced features. Peracetylated particles, however, behave significantly differently (C). Firstly, 70 % of the total thiols are already accessible to the dye at low ionic strength (0-0.8 M NaCl), only the remaining 30 % require a dissociation for their reaction (this distribution has been calculated from P, values). Secondly, in the acetylated histone complex, desorbed from DNA above 1.3 M NaCl, cysteine accessibilities are already close to the minimum which is found for a strongly tightened octamer of unmodified histones


0.8 MNaCl

I

Time,t (min) t (min) l o o lo

t (min) t (rnin)

I Fig. 4. The salt-induced exposure of liistone H3 tliiols in different core particle preparations. Second-order plos of labelling kinetics were derived from the core particles from (-0) chicken erythrocytes, (-0) Chinese hamster ovary interphase and (0 -0) butyrate- treated cells. Core particles were pipetted to a solution of N-[p-(2-benzimidazolyl)phenyl]maleimide containing the indicated amount of NaCI. The other conditions correspond to Fig. 3

0 0.5 1.0 1.5 2.0 NaCl (M)

Fig. 5 . Reactivity profile of the histone H3 thiols as a function of ionic strength. Reactivities were derived from the terminal slopes of second-order kinetic plots such as Fig.3. (-) Core particles which were free from non-histone proteins. (----) Core particles prepared at 0.2 M NaCl and containing some non-histone proteins (see Materials and Methods). Particles are from (0) chicken erythrocytes, (0) Chinese hamster ovary interphase cells and (0) butyrate-treated hamster cells

at much higher ionic strength. In conjunction with the results to follow, these findings may be explained by the concept that a hyperacetylation strengthens histone-histone contacts at the expense of histone-DNA interactions. Both interactions have been found to shield the H3 cysteines, but only a strong binding to DNA could render them fully inaccessible [26,36].

Histone modifications are not the only factors controlling the conformation of nucleosomes. With the hyperacetylated species, the influence of foreign

proteins upon the cysteine accessibility is minor, since the cysteine environments are already relatively open. The group of non-histone proteins, however, which stays bound to interphase particles at 0.2 M NaCI, obviously influences the thiol reactivities (dashed lines in Fig.5), probably by competing with the segment of DNA which usually controls the access to that part of the H3 molecule.

Circular Dichroism

The position of the acetylation sites within the highly basic aminoterminal regions of histones H3 and H4 was expected to affect their interaction with DNA [2]. DNA spectra on artificial binary complexes from H4 histones and DNA have indicated that H41 is far less effective than the non-acetylated parent protein (H40) to induce a B + A type of conformational shift into the nucleic acid [37]. The relevance of this finding, however, is not clear; since native nucleohistones 1381 and especially core particles [39] display a reduction of the positive CD band of DNA at 275 nm rather than the enhancement caused by H4 alone. It has further been speculated that a multiple histone acetylation would allow those regions to assume an a-helical conformation by reducing electrostatic repulsions [2], but this hypothesis has not yet gained experimental support.

Roughly, a nucleohistone CD spectrum is divided into two regions. The spectral range above 240 nm is almost exclusively determined by the DNA conformation, while the ellipticity at 223 nm is conditioned by the a-helical content of the histone. It is evident from Fig.6 that the conservative B-structure of DNA is


20

10

- 0 '5 c

i -0

6 -20

-

-30

-40

I

Fig.6. CD spectra of various core particle preparations. (---, -----) 0.4 M NaC1, 1 mM NaEDTA (pH 8) eluates consisting of hyperacetylated and interphase specimens from Chinese hamster ovary cells respectively (for preparation see Materials and Methods). (- ~ - - ) Chicken erythrocyte core particles under comparable conditions. (.......) Conservative spectrum of DNA in its B conformation

significantly distorted in core particles but that this effect is most pronounced for the particles from chicken erythrocytes. In the specimens from Chinese hamster ovary cells there is some more resemblence with the DNA spectrum, both with regard to the ellipticity and to the crossover point (257 nm). Qualitatively similar data and conclusions have been derived comparing the particles from chicken erythrocytes and butyrate-treated HeLa cells [14].

The inspection of 8223 values indicated the lowest a-helix content for chicken erythrocytes and increasing contributions for interphase and peracetylated particles. An increase in the helix content has frequently been correlated with the induction of specific cross- interactions between the histones (cf. [40]). Our results would hence support a model in which the interhistone contacts are strengthened due to an extension of a-helical sections. These changes would simultaneous- ly release some of the constraints imposed onto the DNA structure by a reduction of electrostatic interactions. It should be mentioned, however, that Simpson [14] has recently observed the opposite trend of helici- ties, mainly by comparing particles which had under- gone a salt wash at 0.6 M NaCl. While such a treat- ment is apt to remove non-histones completely, it was not applicable to our systems as it tended to release some portion of all histones from the acetylated Chinese hamster ovary core particles. The conditions, which in our hands rendered fully comparable preparations, have been detailed in the experimental section.

Thermal Denaturation

Heating causes two prominent transitions in core particles (cf. [41] and [lo]). A so-called premelting

0.02

- u 0-

.+ . w 2 0.01

0 30 40 50 60 70 80

Ternperature,t ("C)

Fig.1. Differential melting curves of core particles; the eflert of a histone hyperacetylation. Measurements were performed on solutions with an AZ~O,ZO"C value of 0.8 in 1 mM sodium cacodylate, 0.25 mM NaEDTA, pH 8.0. The profiles have been normalized to A 2 6 0 . 2 0 0 ~ = 1 .O. Influences of non-histones upon thermal melting were negligible, i.e. identical traces were derived for the specimens purified by gel chromatography at 0.2 M and 0.4 M NaC1. ( x ~ x ) 140-base-pair particles from chicken erythrocytes; (0) corresponding particles from the interphase of Chinese hamster ovary cells. The continous profile refers to the hyperacetylated species and is based on a computerized replot of hyperchromicities

transition, which in the medium used by Weischet et al. [41] is generally centered at 6 2 T , is though to reflect a denaturation of some loosely bound parts of the 145-base-pair piece of DNA. With the particles from chicken erythrocytes these parts are most likely the ends [41] and the transition is almost fully reversible [10,41]. The final and irreversible thermal break- down is centered at 75 "C; it is highly cooperative [lo]. In the light of the conclusions derived from the accessibility data and CD spectra, one would expect an increased premelting contribution for the core particles from active chromatin, especially if they contain a peracetylated histone kernel. Fig. 7 and Table 2 show that this prediction is borne out.

The most significant deviations from a standard denaturation profile concern the premelting step of the hyperacetylated particles. With a half-width of almost 20 "C (Table 2), this transition is considerably broadened as a probable consequence of the different levels of acetylation it comprises. Furthermore, the transition midpoint has shifted by almost 5°C to lower temperatures, proving that the histone-DNA interaction has actually been weakened. Finally, cooling down a solution of peracetylated core particles from 65°C shows that premelting is mostly (70%) irreversible. In these specimens the recognition sites between DNA and the histone core, which have been conserved in evolution, have obviously been disturbed to an extent which would make a deacetylation one prerequisite for a proper reassiociation.


Table 2. Numerical evulutation of Fig. 7 concerning the melting properties of core particle preparations

Property Butymte- Interphase Chicken treated erythrocyte

Premelting transition t m ("C) 56.8f 1 62 + 1 61.4+ 0.3 Hyperchromicity ( x Of total) 46 + 2 35 k 2 23 + 1 Base pairs 64 k 3 49 + 3 32 +2 Half-width ('C) 19.5+1 14 & 1 12 +0.2

I m ("C) 74 f 1 14.5+ 1 75 +0.2

( x O f tOtd1) 54 + 2 65 + 2 77 + 1

Main transition

Hyperchromicity

Base pairs 76 + 3 91 + 3 108 + 3 Half-width ('C) 10.8+0.5 7.6k0.5 8 +0.2

Acetylation does not prevent the remainder (54 %) of DNA from melting at the usual temperature of the main transition (74- 75 "C) and a similar observation has been taken by others to indicate but minor alterations of the core particle structure [14]. It must be considered, however, that this transition should be dominated by those particles containing H3 and H4 molecules which are either not or only monoacetylated (both types are present in chicken erythrocyte particles, cf. Table 1). Nevertheless, a noticeable broaden- ing from a half-width of 8 "C to almost 11 "C in- dicates a slight destabilization of this part of DNA in hyperacetylated species.

The Interaction of Core PurticIes and Protarnines

During spermiogenesis, the somatic histones are replaced by a group of newly synthesized, highly basic proteins (protamines). This process is assisted by a number of enzymatic modifications, among which a protamine phosphorylation/dephosphorylation sequence and histone peracetylations are the most prominent [5]. Various aspects of the protamine- histone competition have been studied by models in vitro [5,11,42]. Owing to the late discovery of the butyrate effect [7] the role of histone acetylation has as yet only been evaluated by a chromatin from calf thymus, which had been treated with acetic anhydride up to the approximate acetyl level in vivo.

Some major points which have emerged from the simulations of spermiogenesis in vitro are as follows. (a) H1 (H5) is the only histone which can directly be displaced by protamine from calf thymus [42] and chicken erythrocyte nuclei [43] or from the corresponding chromatins [5,42]. (b) The nucleosomal entity provides a perfect protection of the core histones from protamic in nuclei, chromatin and core particles [5,16,42]. (c) A chemical acetylation of calf thymus chromatin does not support the replacement of the

I T P1 P2

T P1 P2 Migration

Fig. 8. The stability of Chinese lzarnster ovury core pariicles towards protamines. At a concentration of 0.235 mM nucleotides, core particles were treated with increasing amounts of protarnine (clupeine YI) in 0.2 M NaCI, 1 mM NaEDTA, pH 8.0. Protamine- arginine/nucleotide ratios of 0.4 were sufficient to render all core particles pelletable at 3000xg (15 min), i.e. the absorbance at 260 nm in the supernates was close to zero after this step. The histone Composition of the pellets was registered at two protarnine loads (Pl, Arg/nucleotides = 0.4; P2, Arg/nucleotides = 4) and compared with the initial composition, T. T had been analyzed on a pellet of core particles in 18 "/, trichloroacetic acid. Similar conditions were used to confirm that histones missing in the pellets were part of the supernates. Top, particles from interphase cells; bottom, , particles from butyrate-treated cells

core histones, unless it is assisted by their proteolytic degradation [5,11]. (d) A disruption of the native interhistone contacts has been achieved by shearing forces or by the incorrect reconstitution of chromatin under direct mixing conditions ; from these complexes variable amounts of the core histones are lost by the action of protamines [42]. It is our hypothesis that a distortion of interhistone and histone-DNA recognition sites could also be a consequence of an enzymatic modification or an incorrect refolding of histones upon the reversal of such a step (see [42] and above).

Fig.8 shows the effect of two different concentrations (Pl, P2) of protamine upon Chinese hamster ovary core particles from interphase and from butyrate-treated cells; it should be kept in mind that particles from chicken erythrocytes would be un- affected by these conditions [21,43].

The interphase particles are seen to tolerate binding of 5 - 6 protamine molecule (corresponding to a


protamine-arginine/nucleotide ratio of 0.4) but they suffer some loss of the slightly lysine-rich histones H2a and H2b at more than saturating concentrations. With the hyperacetylated specimens, on the other hand, there is a simultaneous loss of about 30 % of all core histones at low levels of protamine (Pl). While composite particles comprising the complete histone core and the protamines have been traced in preparations from chicken erythrocytes (containing some H42, H32, H41 and H31, cf. Table 1 and [16,43]), the present results are taken to mean that in hyperacetylated fractions the histone complement is totally replaced. At higher loads (P2) a slight preference is seen of the H2a/H2b group as in the control from interphase.

A safe conclusion, which can be drawn from this experiments, is that the histone-binding areas are severely impaired by a hyperacetylation, and to a certain extent also by modifications which have oc- curred in interphase chromatin. Although there is an apparent contradiction between our findings and those which have been obtained by Wong et al. on chemically acetylated chromatin [ 5 ] , it should be considered that our study was performed at the monomer structural level, i.e. with core particles lacking a superstructure. Some comparative experiments in our laboratoiy did, in fact, indicate that protamine-mediated effects are reduced considerably if intact Chinese hamster ovary nuclei are submitted to this test.

CONCLUSIONS

Core particles undergo a variety of structural alterations upon the association of foreign proteins like non-histones (Fig. 5 ) , the linker-specific histones H1 or H5 [16,43] and protamines (Fig.8 and [21]). They are also modulated by enzymatic modifications, among which the acetylation has become amenable to experimentation as it is conveniently stabilized by the addition of butyrate to the medium of a cell culture. For Chinese hamster ovary cells, the average number of acetyllysine side-chains per core particle could be raised from about 2 to 7; this estimate does not include the possible modification of histones H2a and H2b [9].

We have traced a number of changes which could be ascribed to the peracetylated state of nucleosomes. A segment of DNA is transferred into a loosely bound state, as seen by a portion of 64 base pairs melting in the range of 40 "C to 70'C (Fig. 7). The binding of this DNA piece is tight enough to resist a spontaneous degradation by micrococcus nuclease, yet it renders parts of the protein core accessible, as seen from the reaction of the H3 cysteines with a thiol-specific dye at 0-0.8 M NaCl (Fig.5). A general weakening of histone-DNA interactions is also evident from the

displacement of the histones by protamine (Fig. 8). While the histone loss from interphase particles is moderate and only concerns the H2a/H2b group (Fig. 8, top), a nearly stoichiometric complex of all core histones is liberated by the first additions of protamine to the hyperacetylated variant (pellet P1 in Fig.8, bottom). In the core complex, interhistone contacts are tightened to an extent which is otherwise only attained by a high concentration of counterions (Fig. 5) and this is probably facilitated by the expan- sion of a-helical regions, which has been derived from the CD spectrum (Fig. 6).

Besides the direct effects of a histone acetylation, there are other influences which appear to be correlated with the activity of the parent chromatins. Core particles from the interphase of actively growing Chinese hamster ovary cells yield a protein pattern comparable to inactive chromatin, but the structural analyses show some resemblence with the acetylated particles as well. It is proposed that a modification, like the acetylation of lysine residues, may lead to conformational changes which persist in part after the reaction has been reversed, e.g. by deacetylation. Reconstitu- tion experiments are in progress to clarify this point.

The authors wish to thank Prof. K . G. Wagner for his interest and for critically reading the manuscript. The kind gift of a starter culture of Chinese hamster ovary cells by Dr Tiffe and Prof. Thies- sen from the Medizinisclze Hoclzsclzule Hannover is gratefully acknowledged, as is the valuable advice by Mrs Baranowski on the cultivation conditions. We are particularly grateful to Mrs MaaR for her skilled help during all the preparative procedures, to Mrs Pollrnann for some of the gels and to Mrs Jahne and Mr Kiihne for their assistance in preparing this manuscript. This work was supported by the Deutsclze Forsclzungsgemeinscli(~t (Wa 19) and the Fonr1.Y der Clzemiscken Industrie.

REFERENCES 1. 2.

3. 4.

5.

6. 7.

8. 9.

10. 11.

12.

13.

14. 15.

Felsenfeld, G. (1978) Nuture (Lond.) 271, 115-122. Ruiz-Carillo, A., Wangh, L. J. & Allfrey, V. G. (1975) Science

Bina-Stein, M. & Simpson, R. T. (1977) Cell, I I , 609-618. Louie, A. J. and Dixon, G. H. (1972) Proc. Nutl Acud Sci.

Wong, T. K . & Marushige, K. (1976) Bioclzemisiry, 15, 2041 -

Sealy, L. & Chalkley, R. (1978) cell, 14, 115-121. Riggs, M. G., Whittaker, R. G., Neumann, J . R. & Ingram,

Prasad, K. N. & Sinha, P. K . (1976) In vitro, 12, 125-132. Cousens, L. S., Gallwitz, D. & Alberts, 9. M. (1979) J . Biol.

Bode, J. & Henco, K. (1980) Int. J . Biol. Mucromol. 2, 122- 128. Marushige, K., Marushige, Y. & Wong, T. K . (1976) Biocliem-

istry, 15, 2047 - 2053. Vidali, G., Boffa, C., Bradbury, E. M. & Allfrey, V. G. (1978)

Proc. Nutl Acud. Sci. USA, 75, 2239-2243. Nelson, D., Perry, M., Sealy, L. & Chalkley, R. (1978) Bio-

chem. Bioplzys. Res. Commun. 82, 1346- 1353. Simpson, R. T. (1978) Cell, 13, 691 -699. Hsu, T. C. (1958) J . Nut1 Cancer Ins!. 21, 437-473.

( W U S h . DC) 190, 117-190.

USA, 69, 1975- 1979.

2046.

V. M. (1977) Nuture (Lond.) 268, 462-464.

Cliem. 254, 1716-1723.

152

16. Hofmann, K. W., Arellano, A,, Arfmann, H. A. & Bode, J. (1980) Int. J . Biol. Macromol. 2, 25-32.

17. Laemmli, U. K. (1970) Nature (Lond.) 227, 680-685. 18. Hansen, J. N. (1976) Anal. Biochem. 76, 37-44. 19. Bode, J., Schroter, H. & MaaD, K. (1980) J. Chromafogr. 190,

20. Bode, J. (1979) Anal. Biochem. YY, 274-280. 21. Wingender, E., MaaD, K. und Bode, J. (1980) Int. J . Biol.

22. Wangh, L. J., Ruiz-Carillo, A. & Allfrey, V. G. (1972) Arch.

23. DAnna, J. A,, Tobey, S., Barham, S. S. & Gurley, L. R. (1977)

24. D’Anna, J. A,, Gurley, L. R. & Deaven, L. L. (1978) Nucleic

25. Kanaoka, Y., Machida, M., Ando, K. & Sekine, T. (1970)

26. Bode, J. & Wagner, K . G. (1980) Int. J . Biol. Macromol. 2,

27. Romer, R. Riesner, D., Coutts, S. M. & Maass, G. (1970) Eur.

28. Brandt, W. F. &van Holt, C. (1976) FEBS Lett. 65, 386-390. 29. Panyim, S., Sommer, K. R. & Chalkley, R. (1971) Biochemistry,

30. Marzluff, W. F., Sanders, L. A., Miller, D. M. & McCarty,

437-444.

Macromol. in the press.

Biochem. Biophys. 150,44- 56.

Biochem. Biophys. Res. Commun. 77, 187- 194.

Acids Res. 5, 3195-3207.

Biochim. Biophys. Acta, 207, 269- 277.

129 - 136.

J . Biochem. 15, 77-84.

10,3911 -3917.

K. S. (1972) J . Biol. Chem. 247, 2026-2033.

J. Bode, K. Henco, and E. Wingender: Nucleosome Acetylation

31. Camerini-Otero, R. D. and Felsenfeld, G. (1977) Proc. Nut1

32. Sadgopal, A. & Bonner, J. (1970) Biochim. Biophys. Acta, 207,

33. Gurley, L. R., DAnna, J. A., Earham, S. S., Deaven, L. L.

34. Garrard, W. T., Nobis, P. & Hancock, R. (1977) J. Biol. Chem.

35. Ando, T. & Steiner, M. (1973) Biochim. Biophys. Acta, 311,

36. Bode, J . & Standt, U.-D. (1978) 2. Naturforsch. 33c, 884-890. 37. Adler, A. J. & Fasman, G. D. (1974) J . Biol. Chem. 249,291 1 -

38. Wagner, T. & Spelsberg, T. C . (1971) Biochemistry, 10, 2599-

39. Zama, M., O h , D. E., Prescott, B. & Thomas, G. J., Jr (1978)

40. D’Anna, J. A. & Isenberg, D. (1974) Biochemistry, 13, 2098 -

41. Weischet, W. O., Tatchell, K., van Holde, K. E. & Klump, H.

42. Bode, J., Willmitzer, L. & Opatz, K . (1977) Eur. J. Biochem.

43. Bode, J., Willmitzer, L., Arfmann, H.-A. & Wehling, K . (1979)

Acad. Sci. USA, 74, 5519-5523.

227-239.

& Tobey, R. A . (1978) Eur. J . Biochern. 84, 1 - 15.

252, 4962 - 4967.

38-44.

2914.

2605.

Nucleic Acids Res. 10, 3881 -3897.

2104.

(1978) Nucleic Acids Res. 5, 139- 160.

72, 393-403.

Int. J. Biol. Macromol. I , 73 - 78.

J. Bode and E. Wingender, Gesellschaft fur Biotechnologische Forschung mbH, Mascheroder Weg 1, D-3300 Braunschweig-Stockheim, Federal Republic of Germany

K. Henco, Institut fur Organische Chemie der Universitat Darmstadt, D-6100 Darmstadt, Federal Republic of Germany

https://www.researchgate.net/publication/18234840_Aspects_of_chromosomal_structure_I_Circular_dichroism_studies?el=1_x_8&enrichId=rgreq-37cecb236228b09f8b362de3dea9f53a-XXX&enrichSource=Y292ZXJQYWdlOzE1NzYwMjU2O0FTOjE0OTY3Mjc0ODM5MjQ0OEAxNDEyNjk2MTcyMTEy

Documents

Modulation of the Nucleosome Structure by Histone Acetylation