THE JOURNAL OF BIOLOGICAL CHEMWIXY Vol. 265, No. 32, Issue of November 15, pp. 19639-19847,199O Printed in U.S.A.

Isolation and Characterization of Acetylated Histones H3 and H4 and Their Assembly into Nucleosomes*

(Received for publication, April 26,199O)

Keith W. Marvin+, Peter YauS, and E. Morton Bradbury+ From the $Department of Biological Chemistry, School of Medicine, University of California, Davis, California 95616 and the ILife Sciences Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545

Nucleosome and chromatin structure/function rela- tionships of histone acetylations are not understood. To address these questions we have developed chro- matographic procedures that separate subtypes of H3 and the acetylated states of histone H3 and H4 in exceptionally pure forms. The sites of acetylation of the intermediately acetylated states of H3 have been determined and show a specific pattern of acetylation. An unexpected finding was the identification of a fifth site of acetylation in H3 at lysine 27. Nucleosome par- ticles with fully acetylated H3 and H4 have been as- sembled on the Lytechinus uariegutus 6 S rRNA DNA phasing sequence and characterized. These defined acetylated H3 and H4 particles migrate more slowly in polyacrylamide nucleoprotein particle gels than the control particles indicating a subtle effect of acetyla- tion in nucleosome structure. However, DNA foot- printing of these particles using DNase I show only small changes when compared to control particles over the core particle DNA length. It is shown further that H3 cysteines in the particle containing fully acetylated H3 and H4 were not accessible to iodoacetamide indi- cating that protein factors additional to H3 and H4 acetylation are required to make H3 cysteines acces- sible to the label. These findings are consistent with the proposal that histones H3, H4 acetylations exert their major effects outside of the core particle 146- base pair DNA, either on the DNA segment entering and leaving the nucleosome or possibly on the inter- nucleosome interactions that involve the amino-ter- minal domains of the core histones in organization and stability of higher order chromatin structures.

The four core histones of the nucleosome have well defined structural domains. Each has a flexible, basic amino-terminal domain extending from a structured globular domain. H2A and H3 also have short COOH-terminal tails extending from the globular central domain (1). The amino-terminal domains of the four core histones contain all the sites of reversible acetylation (2). The known sites are: in H2A, lysine 5; in H2B, lysines 5, 12, 15, 20; in H3, lysines 9, 14, 18, and 23; in H4, lysines 5, 8, 12, and 16. In this study, a fifth site of acetylation of histone H3 (Lys-27) has been identified. So far the modes of interaction and functions of the NHz-terminal domains in nucleosomes and chromatin are not well under- stood. However, it has been recently shown that high levels

* This work was supported by National Institutes of Health Grant GM 26901 and Department of Energy Grant DE-FGO348ER60673. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 USC. Section 1734 solely to indicate this fact.

of acetylation of all four core histones cause a reduction in the linking number change per nucleosome in assembled circular minichromosomes from -1.01 to -0.82 (3). Thus histone acetylation releases DNA supercoils previously con- strained on nucleosomes, which has implications for changing the topological state of chromatin domains.

There is a long established association of reversible histone acetylation of the core histones with gene activity (see Allfrey (2) for a review of the earlier work) based on a variety of correlative data. Studies of the cell cycle in Physarum poly- cephalum show elevated incorporation of labeled acetate into modified H4 during periods of gene activity (4). DNase I has been shown to digest preferentially the active gene sequences in chromatin (5-8). Although this nuclease sensitivity may be conferred by the non-histone proteins, e.g. HMG 14 and 17 (g-12), the histones released by such digestion contain ele- vated levels of acetylation (13, 14). Treatment of tissue cul- tured cells with the histone de-acetylase inhibitor, n-sodium butyrate (15), enhances both the DNase I digestion of chro- matin and the release of more highly acetylated histones (16, 17). An increased DNase I susceptibility has been reported for positions 60 bp’ from each end of the DNA in hyperace- tylated core particles (18, 19). Nuclease digestions using mi- crococcal nuclease (20, 21) or DNase II (22) release a 2 mM Mg2’-soluble fraction of chromatin which is enriched in both active gene sequences and in acetylated histones (23, 24) and the solubility of this active chromatin is selectively enhanced by butyrate treatment (25-27). Nucleosome core particles containing very high levels of acetylation have been obtained by their differential solubility (18) or fractionation on HAP columns (28). Strong correlations of histone acetylation with gene expression come from studies of accessible sulfhydryls found in active chromatins. In P. polycephalum rDNA chro- matin, the ordinarily protected H3 sulfhydryls in inactive nucleosomes are accessible to thiol-specific alkylating agents in active chromatin (30, 34) especially in the AcB- and Ac4H3 (34). Mercurial agarose retains a chromatin fraction which is enriched in acetylated H3 and H4 (29,35). Retention is tightly correlated with gene activity; histone gene sequences from S- phase but not Gn-phase chromatin are retained (35), and c- myc and c-fos are retained only and precisely during their periods of induced activity (36). Different states of histone acetylation have been associated with different chromatin functions: acetylation of all four core histones with DNA replication and hyperacetylation of H3 and H4 with transcrip- tional activity (38).

The effects of histone hyperacetylation at the core particle level have been subtle. Imai et al. (28) confirmed the report of Bode et al. (33) that hyperacetylated core particles have

’ The abbreviations used are: bp, base pair(s); HPLC, high perform- ance liquid chromatography.

19839

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

19840 Nucleosomes Containing Fully Acetylated H3 and H4

retarded mobility on nucleoprotein gels relative to control core particles. However, they did not observe significant struc- tural differences between hyperacetylated and control core particles at low resolution by neutron scatter and suggested that the major structural effects of histone acetylation may act above the level of the core particle, e.g. on DNA entering and exiting the nucleosome (28). This is supported by the small changes reported for the effects of acetylation on the sedimentation coefficient and circular dichroism of core par- ticles (18) and in the thermal denaturation of DNA in hyper- acetylated mono-, di-, and trinucleosomes (37).

TO fully understand histone acetylation structure/function relationships it is necessary to assemble nucleosomes and chromatin models from fully defined acetylated histones and DNA sequences. This requires the development of procedures to fractionate and characterize the different acetylated forms of the core histones. Although two methods using ion ex- change celluloses have been available for some time for the fractionation of H4 into acetylated forms, one using ethanolic formic acid (39,40) and the other using acetate-buffered urea (41), neither has been found effective for the fractionation of acetylated forms of H3 (39).’ H3 may be separated into two classes of variants differing in hydrophobicity by reversed phase HPLC (42,43). The procedure described here involves rapid isolation of H3 as two classes of variants by reversed phase HPLC followed by fractionation on a weak cation exchange HPLC column with low nonspecific binding prop- erties. The cation exchange step is generally applicable to all the purified core histones, although only the fractionation of H3 and H4 will be described here.

The locations and occupancy of the sites of acetylation in each acetylated form of the less hydrophobic H3 have been determined; in particular a fifth site of acetylation in histone H3 (Lys-27) has been identified. Nucleosome particles which are fully acetylated with respect to H3 and H4 have been assembled and characterized. It is shown that the H3 cysteines in the acetylated nucleosome particles are not accessible to iodoacetamide, suggesting that protein factors additional to histone acetylation are required to make their cysteines ac- cessible. DNA footprints of nucleosome particles assembled using fully acetylated H3 and H4 show only small changes when compared to control particles over the core particle DNA length.

EXPERIMENTALPROCEDURES

RESULTS

Occupation of Sites of Acetylation in H3 and Identification of a Fifth Site-The phenylthiohydantoins of both the acet- ylated and the monomethylated lysines were readily distin- guishable from that of lysine. For each of the acetylated forms of H3.2/.3 isolated from butyrate-treated cells, the levels of acetylation at each of the five sites has been established (Fig. 4). In the first three acetylated forms, recovery of total lysine derivatives at lysine 9 was low. This may be attributed to dimethylation, which has been reported at this site, but which was not analyzed for in the present experiments. The per- centages of acetylated and monomethylated lysine at this residue were therefore calculated against an interpolated value for total lysine. From these data it is clear that the order of

* K. Marvin, unpublished results. 3 Portions of this paper (including “Experimental Procedures,” part

of “Results,” and Figs. 1-3) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

100 AC,

4 9 14 10 23 27

Position in The Sequence of H3

FIG. 4. Occupation of the sites of acetylation in H3. For each of the five acetylated forms of H3.2/.3 (Acl through AC& the per- centages of acetyllysine (Ad) to total lysine at the lysines in the amino-terminal region are plotted against the number of cycles Ed- mon degradation required to reach the residue. Because dimethylly- sine was not determined in these experiments, the total lysine at cycle 9 in the first three acetylated forms was estimated by multiply- ing the total lysine at cycle 4 by [yield15, where yield is the average of [recovery at Ala-15/recovery at Ala-1]“24. Yield = 0.945 f 0.006. For Acr through Acz sequencing was stopped before lysine 27.

acetylation of lysines in histone H3 is not random. In the monoacetylated H3, lysines 14 and 23 are each acetylated almost to the 50% level with a trace of acetylation on lysine 18. The levels of acetylation of lysines 14 and 23 increase to 90% in the diacetylated H3 with low levels of acetylation of lysines 18 and 9. In the triacetylated form the major increase in acetylation is of 18, whereas in the tetraacetylated form the major increase in acetylation is of lysine 9. In the tetra- acetylated H3 lysines 9, 14, 18, and 23 are fully acetylated and this situation appears to be a prerequisite for the acety- lation of lysine 27, which is found to be fully acetylated in the pentaacetylated state. Monomethylation was detected at both lysines 4 and 9 (Fig. 5) of the first three acetylated forms and at lysine 4 only in the pentaacetylated form. The lysines 27 and 36 of A~H3.21.3 and lysine 36 of the AcbH3.21.3 were all reported in only trace amounts.

Natural Levels of Acetylation-To determine whether the fifth site of acetylation of histone H3 resulted from the butyrate treatment of the cells or whether it occurs naturally, the state of acetylation of control H3 was analyzed. H3.1 and H3.2/.3 from control HeLa cells were chromatographed by cation exchange HPLC. All of the forms detected in butyrate- treated cells are still present (Fig. 6, A and B) but the nonacetylated form is now the most prominent and the higher acetylated forms are much less so. AcbH3.2/.3 from butyrate- treated cells used as a standard eluted at the same retention time as the penta-modified form in H3.2/.3 from control cells (Fig. 6, C and D). The tetra- and pentaacetylated forms combined represent about 1% of the H3 of cycling HeLa cells.

Nucleosome Assembly-Nucleosomes assembled using fully acetylated H3 and H4 (i.e. using a H2A/H2B mixture from control cells) consistently exhibit a lower mobility in nucleo- protein polyacrylamide gels than do either those reconstituted with control or with nonacetylated inner core histones (Fig.

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Nucleosomes Containing FJI& Acetylnted H3 and H4

1 2 3 4 5

Level of Acetylation

FIG. 5. Monomethylation of lysine residues in acetylated H3. The percentage of monomethyllysine (MML) at lysines 4 and 9 of the acetylated forms of H3.2/.3 were plotted against the number of acetyl residues per molecule of H3. Reliable data were not obtained for cycle 4 of the tetraacetylated H3 rendering the percentage at that cycle highly uncertain.

time @in)

C

SO 40

time (min)

FIG. 6. Fractionation of the acetylated forms of H3 from control cells. A, cation exchange HPLC chromatogram H3.2/.3 from control cells plotted as A230 nm against time; B, a similar plot but for H3.1 from control cells: C. nurified Ac5 from butvrate-treated HeLa cells H3.2/.3; D, the b& i&cated in Ehromatogiam A expanded to show detail; E, the box indicated in chromatogram B expanded to show detail.

7). This effect is independent of the DNA which is used for the reconstitution as it is seen both with the core particle length DNA (Fig. 7A) and with the monomeric 5s-207 phasing sequence (Fig. 7B). The latter produces a more complex pattern of reconstituted bands but all of the nucleosomal bands shift as a result of acetylation. The shift is also inde- pendent of which non-allelic variant of H3 was used, as can be seen in Figs. 7A and 1OA.

Three closely spaced bands were observed in the nucleo- somes reconstituted on the 5s-207 DNA. However, when the bands were excised from a nucleoprotein particle gel and loaded onto a second dimension sodium dodecyl sulfate gel, all four core histones were present in each at similar ratios (data not shown). The origin of the multiple reconstituted bands is not fully understood. However, it has been shown that nucleosomes assembled onto tandem repeats of the 5s- 207 phasing sequences are located on several positions mul- tiples of 10 bp away from a dominant position (59),5 and this

’ G. Meersseman, S. Pennings, and E. M. Bradbury (1990) J. Mol. BioL, submitted for publication.

constituted nucleosomal particles. A, lane I, 147 + 3-bp HeLa core narticle DNA: lanes 2 and 9. MSDI digest of uBR322 (New England Biolabs); ianes 3 and 8, nati;e chsken erhhrocyte core particles; lane 4, reconstituted core particles containing unacetylated H3.2/.3 and H4 and bulk H2A and H2B assembled at a molar ratio of histone to 147 + 3-bp DNA of 1.2:1; lane 5, reconstituted core particles containing tetraacetylated H3.2/.3 and H4 and bulk H2A and H2B assembled at a molar ratio of histone to 147 & 3-bp DNA of 1.2:1; lanes 6 and 7 same as in lanes 4 and 5, respectively, but with H3.1 instead of H3.2/.3. B, lanes 1 and 5, PM2 HaeIII digest + 5s- 207 DNA; lane 2, unacetylated H3.2/.3 and H4 with control H2A/ H2B reconstituted on 5s-207 DNA; lane 3, tetraacetylated H3.2/.3 and H4 with control H2A/H2B reconstituted on 5s-207 DNA; lane 4, control histones (native levels of acetylation) reconstituted on 5s-207 DNA. (The reconstitutions were at a 1:l molar ratio of histone to DNA.) The standard sizes were estimated against a partial AuaI digest of ~5~207-18 (not shown).

could be an explanation for the multiple bands. Thiol Reactivity of Hktone H3 Cysteines-It has been shown

that the cysteines of H3 in the very transcriptionally active P. polycephulum rDNA chromatin are fully accessible tc ;o- doacetate or iodoacetamide derivatives (30, 34). In contrast, transcriptionally inactive nucleosomes are not labelled by these reagents (29-33). The active rDNA nucleosomes, called lexosomes, have a more open configuration than inactive nucleosomes and not only have hyperacetylated H3 (34) but also contain stoichiometric amounts of the non-histone chro- mosomal proteins LP30 and LP32 (30). The ability to assem- ble nucleosome core particles with fully acetylated histone H3 and H4 allows us to determine whether these acetylated states, which are associated with active genes, are sufficient to allow access of H3 cysteines to iodoacetamide. Whereas the free H3 in urea-denatured core histone mixtures can be labeled with the reagent, H3 in either nonacetylated assembled nucleo- somes or in fully H3 and H4 acetylated nucleosomes could not be labeled (Fig. 8). This is consistent with electron micro- graphs of cc5S-207-M minichromosomes assembled with fully acetylated histones which show normally compact nucleosome particles, not extended ones (74).

DNA Footprints of Nucleosomes Containing Fully Acetylated H3 and H4-There are several explanations for the remark- able result that hyperacetylation of core histones results in a decrease in the linking number difference per nucleosome from -1.04 to -0.82 (3). First, histone hyperacetylation could release 9.3 bp of DNA from the ends of the DNA entering and exiting the core particle (3). Second, a change in the shape of the histone octamer (60), acting as a spool for the coiling of DNA in the nucleosome could result in a change of the DNA linking number difference of a core particle without release of DNA from the ends of the core particle. Third, histone acetylation might affect internucleosomal interac- tions and higher order nucleosome arrangements. To test these possibilities we have compared the DNase I footprints of fully acetylated (with respect to H3 and H4) nucleosomes

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Nucleosomes Containing Fully Acetylated H3 and H4

,H3 ‘-H2B ?$A

B

- $j;B ‘H2A ‘H4

FIG. 8. Sodium dodecyl sulfate-polyacrylamide gel electro- phoresis of iodo[ 1-l ‘Clacetamide-treated reconstituted nu- cleosomes. A, Coomassie-stntncd and dried gel; H, autoradiogram of the above gel. Lnne,~ I-3, reconstituted 5s.207 nucleosomes treated with iodoacetamide as deacrihed; lnnc 1, unacetylated; lane 2, fully acetylated m H:l and H4, lane .I. control; lane 4, HeLa htstone marker (not treated), lane .5, iodoacetamide-treated, urea-denatured control histones. A field inhomogeneity caused the greater migratton m lane 2; the histones were intact Acetylation of H3 typtcally causes poor resolution between H:3 and HZH (74).

assembled on the 5s207 DNA with control nucleosomes sim- ilarly assembled.

Acetylation does appear to confer a modest increased sen- sitivity to DNase I at several sites well inside the nucleosome. The clearest of these are the triplets around 102 and 82 nucleotides from the labeled end, and the doublets at 89, 91, and 69-70 on the coding strand (Fig. 9). However, no major change in the cutting pattern has occurred, and the same periodicity has been retained.

The reported position of the phased nucleosome obtained by DNase I footprinting of both strands would correspond to between 168 nucleotides on the coding strand and 185 on the noncoding strand or 23-168 and 185-40, respectively (61). The footprints in Fig. 9 closely resemble those obtained by Simpson and Stafford (61) on the same strand except that no strong cleavage near 168-171 bases was observed for these reconstituted nucleosome particles. Digestion in 10 mM

MgClr, 3 mM CaCl, under the same conditions used previously (61) did not restore this cleavage site (data not shown). The region between 143 and 171 nucleotides from the 3’ end of the coding strand falls within the transcription factor IIIA binding site. The transcription factor IIIA binding site is about 50 bp long and spans the region 138-185 bp from the end of the 207.nucleosome phasing sequence.

The formation of the three complexes with the phasing sequence DNA could have obscured some differences between acetylated and unacetylated nucleosomes. Therefore, the com- plexes were separated on the nucleoprotein gel in Fig. 1OA after the DNase I footprinting reaction and the DNA run on denaturing gels. The noncoding strand was labeled in these experiments because of the greater distance from the reported edge of the phased nucleosome (40 nucleotides from the labeled end).

The footprints in Fig. 10B show that the three complexes do not mask differences between acetylated and unacetylated nucleosomes. The three particles footprint differently, partic- ularly in the 128. and 150.nucleotide region. In contrast, each acetylated particle footprints only slightly differently from its unacetylated control (compare lanes j to h, 1 to m, and n to 0). The footprints extend from approximately 55-186 nucleotides for the upper two of the three particles from the nucleohistone gel, but somewhat less at both ends for the third.

111 - 106-107-~

lo1 35 94-95c

; ii- 72 77 .-

i

72

u 60-62:

; 54-56 51 46 43

37 36

C A F -r- G 2 3 4 5 6 7 8 9 10111213

-

-

- .- -._

FIG. 9. DNase I footprint of control and acetylated nucleo- some reconstituted on 3’-coding strand-labeled 5s-207 DNA. The digestions were for 2 min at room temperature m a final volume of 40 ~1 with the amount of DNase I Indicated in the final reconsti- tutton buffer with 100 mM NaCl and 3 mM added M&l,. Lanes 2-r?. control histone reconstitutes digested wtth; lnrze 2, 2 ~1: lane .3, 4 ~1. and lane 4, 16 ~1 of a s-fold dilution of the stock enzyme t-80 ‘C stored stock); and lane Fi, 12.8 ~1 of the undiluted stock. Lanes c-9, acetylated reconstitutes dtgested with the same conditions as for lane.5 2-ii, respectively. Lanen 10-13, the free labeled 5s207 DNA digested with; lane IO, 2 ~1; lane 1 I, 4 ~1; lane 12. 16 ~1 of a 250-fold dilution of the stock enzyme; and lane 19, 1.2X ~1 01 a 3-told dilution. Lane G. (CH ,)SO,/pipertdine-cleaved. labeled 5s.‘07.

DISCUSSION

A two-column HPLC procedure has been developed for the fractionation and purification of all of the acetylated states of histones H3 and H4 essentially to homogeneity as shown in Figs. 2 and 3 in the “Miniprint.” It is well known that there are four sites of acetylation in histone H4 and four acetylated states of H4 have been obtained. However, for histone H3 we have established that there are five sites of acetylation in both butyrate-treated and control cells. Lysine 27 is the fifth site. HeLa cell acetyltransferase clearly uses this and the other four sites of acetylation in H3 in a nonrandom order (Fig. 4). Lysines 14 and 23 are acetylated first, followed by lysines 18 and 9, and, finally, lysine 27. With increased acetylation of H3, monomethylation of lysine 9 decreases while that of lysine 4 increases (Fig. 5), so that in the pentaacetylated state the only monomethylated site is lysine 4, which is largely occu- pied. In the first three acetylated forms of H3 the dimethy- lation of lysine 9, estimated from the under recovery of total lysine at that residue, was quite substantial, about 75, 45, and 55% for AC,-, AC,-, and Ac,,H3, respectively. A possible role for methylation is to limit acetylation to the correct sites in the amino-terminal domain of H3.

The apparent order of usage of the acetylation sites differs from that found from Tetrahymena thermophilia macronu- clear H3 (microsequenced from bands on gels) where the

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Nucleosomes Containing Fully Acetylated H3 and H4 19843

12345678 91oTt1213141516

A abcdefghi jk lmnop

---a--_-L - . - .^. B-

.U ]m

.Aa .:=52

FIG. 10. DNase I footprinting of the acetylated and unacet- ylated 5s-20’7 reconstituted nucleosome particles from three particle bands separated on a 5% nucleoprotein particle gel. A, autoradiogram of the nucleoprotein gel. Reconstitutes on 3’. noncoding strand-labeled 5s-207 DNA were digested at room temper- ature in final reconstitution buffer containing 50 mM NaCl and 2 mM added MgCl, with 1.2 ~1 of a X-fold dilution of stock enzyme (-20 “C stored) for the times specified. All reconstitutions contained control H2A and H2B; the acetylated contained Ac,H3 and Ac,H4, and the unacetylated, AcoH3 and Ac0H4. Lanes I and 2, 5-min digests of unacetylated and acetylated reconstituted nucleosomal particles, re- spectively, in which the H3 is H3.2/.3; lanes 3 and 4, same as lanes 1 and 2 but with H3.1; lanes 5-8, 9-12 and 13-16, same as lanes l-4 but digested for 10, 20, and 40 min, respectively. B, footprints of the samples excised and extracted from bands in lanes 3, 4, 7, and 8 above. (These contained H3.1 but similar results were obtained with H3.2/.3.) Lanes a and p. (CH&S04/piperidine-cleaved noncoding strand 5s-207 markers. Lanes b-h were 5-min digests and lanes i-o, lo-min digests of the same samples. Reading from left to right in each set of digests are: free DNA (lanes b and i); lowest mobility 5s-207 mononucleosomal particle from the unacetylated sample (lanes c and

j) and from the acetylated sample (lanes cl and k); intermediate mobility particles, unacetylated (lanes e and 1) and acetylated (lanes f and m); highest mobility particles, unacetylated (lanes g and n) and acetylated (lanes h and 0).

order in the presence or absence of butyrate was 9 and 14 followed by 18 (62). Either the mono- and diacetylated forms of H3 in the butyrate-treated HeLa cells represent largely the replication related forms or alternatively the utilization pat- terns of the sites of acetylation in the two organisms are different. The methylation of lysine 9 suggests the latter interpretation because the cells appear to maintain the inter- mediate acetylated states in a class of chromatin. Two studies show different patterns of acetylation of H4 for transcription compared to replication (62, 63).

Hyperacetylation of histone H3 and H4 has been associated with transcriptionally active genes. We have assembled 147

+ 3-bp core particles and 5s-207 nucleosomal particles with fully acetylated H3 and H4 and bulk H2A and H2B and compared their properties with control particles containing the same H2A and H2B and nonacetylated H3 and H4. As can be seen in Fig. 10 the fully acetylated (H3, H4) particles show gel retardation compared to the control particles. This behavior is identical to core particles in which all four core histones are hyperacetylated. This effect on gel migration therefore, can be attributed to the acetylations of H3 and H4. Whereas the acetylated core particles (Fig. 7A) migrate as a single band the acetylated 5s-207 particle (Fig. 7B) migrates as three closely spaced bands. This complexity may be related to the lo-bp separated phasing positions relative to the mon- omer sequence seen in tandom repeats of 5s-207 (59).” Each of the three components of the 5s-207 particle is equally affected by acetylation of histones H3 and H4 in their gel migration behavior. The retardation of the acetylated nucleo- some particles indicates an altered conformation, not a simple charge effect. These nucleosome particles have a net negative charge under the conditions of electrophoresis and would be expected to have a higher mobility upon the charge neutrali- zation of the lysines by acetylation if that were the only alteration in the particle.

In transcriptionally active rDNA chromatin in P. polyce- phalum it has been shown that the histone H3 sulfhydryls are fully accessible to thiol reagents (30, 34). These active rDNA particles contain both hyperacetylated histone H3 and stoi- chiometric amounts of protein LP30 and LP32. They have a more extended conformation than inactive nucleosomes. We have shown (Fig. 8) that full acetylation of H3 and H4 is not a sufficient condition to allow thiol reagent labeling of the H3 sulfhydryls. This is consistent, also, with electron micro- graphs of hyperacetylated nucleosome particles on closed circular 5s-207-18 DNA that look identical to control particles (74). Thus, histone hyperacetylation probably provides the conditions for factors (e.g. LP30 and LP32 in the case of physarum rDNA chromatin) to bind to transcriptionally ac- tive nucleosomes in control of their expression.

Hyperacetylation of histones H3 and H4 is correlated strictly with the requirements of active genes (29, 35, 73), though how this is involved in nucleosome and chromatin structure/function relationships is poorly understood. At the level of the core particle, neutron scatter studies (28) have shown that at low resolution there is little or no effect of hyperacetylation on the solution structure or shape of the core particle. Thus the nucleosome core particle shape and conformational changes that are responsible for the gel retar- dation (Fig. 10) of the acetylated particles must be very subtle. This is further supported by the DNA footprints of fully acetylated nucleosome particles compared with the controls (Figs. 9 and 10). These DNA footprints indicate that in the fully acetylated H3 and H4 nucleosome particles the strong histone DNA interactions are maintained with unchanged periodicity compared to the control. Some DNA sites were more exposed to DNase I digestion in the acetylated particle but the overall pattern of DNA cutting was not markedly affected. The simplest explanation for the increased suscep- tibility of DNA sites flanking the dyad axis of the core particle is that the amino-terminal domains of H3 and H4 interact with the DNA in these regions and that acetylation abolishes those interactions without causing a major structural change in the core particle. It has been shown by nuclear magnetic resonance spectroscopy (64) that acetylation of the four ly- sines in the H4 amino-terminal peptide (l-23) abolishes the binding of this peptide to DNA. The fact that histidine 18 of H4 cross-links at 1.5 helical turns from the dyad (65) supports

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

19844 Nucleosomes Containing Fully Acetylated H3 and H4

the idea that the histone tails bind in this region. Core particle DNA does not coil in a regular manner around

the histone octamer but follows a path of gentle and tight bends (69,73). In data from 177 core particle DNA sequences a lo-bp periodicity of AT and AA/TT sequences on the inside of the DNA coil and of CG and GG/CC sequences on the outside was disrupted at f10.5 bp from the dyad axis (66, 67). Pyrimidine dimerization follows a similar pattern in UV- irradiated core particles (68). The x-ray crystal structure of the core particle also shows discontinuities at these positions (69) which provides the potential for the DNA sequences on the dyad axis to be subjected to strains that are different from those acting on the DNA in the flanking coils. Acetylation of H3 and H4 may be a mechanism to release strain that has been localized in the region of the dyad axis and introduce structural lability into the acetylated nucleosome. Such labil- ity, however, may need other factors associated with the active state of chromatin to induce a structural transition in the nucleosome. Physical stress may also induce structural changes in acetylated nucleosomes more easily than unmodi- fied nucleosomes as suggested by Oliva et al. (72) from electron microscopy studies.

The reported position of the the phased nucleosome on the Lytechinus uariegutus 5s gene (61) overlaps the the transcrip- tion factor IIIA binding site (the internal control region) by about 30 bp (70). It is over much of this region that the nucleosomal footprints become less distinct. (The internal control region spans 138-185 nucleotides from the 3’-labeled end in Fig. 9 and 72-25 nucleotides from the 3’-labeled end in Fig. 10B). The situation resembles somewhat that of the positioned nucleosome in the promoter of PA globin gene of chickens in which 32 bp in the downstream side of the nucleosome footprints poorly because of the presence of sites for the binding of the G-string binding protein and transcrip- tion factor Spl (71).

The major effect of hyperacetylation of the core histones is a reduction in the linking number difference of the nucleo- some particle (3). It is possible that this is caused by flanking DNA segments external to core particle DNA unwinding from the ends of the DNA coiled on the nucleosome particle. The possibility that histone acetylation induces a polymorphic change in the shape of the histone octamer such as promoting a more facile cylinder to ellipsoid transition of the type proposed by White et al. (60) cannot be excluded. This may explain the small differences observed in the DNA footprints, but, overall, the evidence in support of this model is not strong. The possibility remains that the amino-terminal do- mains of the core histones are involved in inter-nucleosomal interactions in chromatin and that histone acetylation modi- fies this interaction and changes the writhe of the DNA linker regions.

Acknowledgments-We acknowledge Dr. R. T. Simpson for the kind gift of p&-207-18; Dr. Brian Imai, Dr. Gary Cook, and Gary Schroth for helpful advice and discussion; Dr. Liane Mende-Mueller, Protein/Nucleic Acid Shared Facility of the Cancer Center of the Medical College of Wisconsin and the Protein Structure Laboratory of the University of California at Davis for protein sequencing; and also for the quantitative amino acid analyses; and Alan J. Smith, director, for advice on the quantitation.

REFERENCES

1. Bradburv. E. M. (1983) CIBA Found. S.ymp. 93,246-265 2. Allfrey, v. (1977).in Chromatin and Chio&tin Structure (Li, H.

J.. and Eckhart. R. A., eds) D. 167, Academic Press, New York 3. Norton, V. G., I&ai, B. S., ?au, P.,‘and Bradbury, E. M. (1989)

Cell67,449-457 4. Chahal, S. S., Mathews, H. R., and Bradbury, E. M. (1983) Nature

287,76-79

5. Weintraub, H., and Groudine, M. (1976) Science 193,848-856 6. Garel, A., and Axel, R. (1976) Proc. Natl. Acad. Sci. U. S. A. 73,

3966-3970 7. Garel, A., Zolan, M., and Axel, R. (1977) Proc. N&l. Acad. Sci.

U. S. A. 74,4867-4871 8. Flint, S. J., and Weintraub, H. M. (1977) Cell 12, 783-794 9. Vidali, G., Boffa, L. C., and Allfrey, V. G. (1977) Cell 12, 409-

415 10. Levy, W. R., Wong, N. C. W., and Dixon, G. H. (1977) Proc. Nutl.

Acad. Sci. U. S. A. 74, 2810-2814 11. Rill, R. L., and Nelson, D. A. (1978) Cold Spring Harbor Symp.

Qua&. Biol. 42,475-482 12. Mayfield, J. E., Serunian, L. A., Silver, L. M., and Elgin, S. C. R.

(1978) Cell 14,539-544 13. Nelson, D., Perry, M. E., and Chalkley, R. (1979) Nucleic Acids

Res. 6,561-574 14. Levy-Wilson, B., Watson, D. C., and Dixon, G. H. (1979) Nucleic

Acids Res. 6,259-274 15. Candido, E. P., M., Reeves, R., and Davie, J. R. (1978) Cell 14,

105-113 16. Vidali, G., Bradbury, E. M., Boffa, L. C., and Allfrey, V. G. (1978)

Proc. Natl. Acad. Sci. U. S. A. 75, 2239-2243 17. Sealy, L., and Chalkley, R. (1979) Nucleic Acids Res. 5, 1863-

1876 18. Ausio, J., and van Holde, K. E. (1986) Biochemistry 25, 1421-

1428 19. Simpson, R. T. (1978) Cell 13,691-699 20. Perry, M., and Chalkley, R. (1981) J. Biol. Chem. 256, 3313-

3318 21. Ridsdale, J. A., and Davie, J. R. (1987) Nucleic Acids Res. 15,

1081-1096 22. Gottesfeld, J. M., Garrard, W. T., Bagi, G., Wilson, R. F., and

Bonner, W. (1974) Proc. Nutl. Acad. Sci. U. S. A. 71, 2193-

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39. 40.

41.

42.

43.

44.

2197 Levy-Wilson, B., Gjerset, R. A., and McCarthy, B. J. (1977)

Biochim. BioDhvs. Acta 475, 168-175 Davie, J. R., andtandido, E. P. M. (1978) Proc. Natl. Acud. Sci.

U. S. A. 75,3574-3577 Hebbes. T. R.. Thorne, A. W., and Crane-Robinson, C. (1988)

EMBb J. 7;1395-1401 Ferenz, C. R., and Nelson, D. A. (1985) Nucleic Acids Res. 13,

1977-1995 Nelson, D. A., Ferris, R. C., Zhang, D., and Ferenz, C. R. (1986)

Nucleic Acids Res. 14. 1667-1682 Imai, B. S., Yau, P., Bgldwin, J. P., Ibel, K., May, R. P., and

Bradbury, E. M. (1986) J. Biol. Chem. 261, 8784-8792 Allegra, P., Sterner, R., Clayton, D. F., and Allfrey, V. G. (1987)

J. Mol. Biol. 196, 379-388 Prior, C. P., Cantor, C. R., Johnson, E. M., Littau, V. C., and

Allfrev. V. G. (1983) Cell 34, 1033-1042 Zama, G:, Olins; D. i., Wilkinson-Singley, E., and Olins, A. L.

(1978) Biochem. Biophys. Res. Commun. 85, 1446-1452 Deterich, A. E., Axel, R., and Cantor, C. R. (1978) Cold Spring

Harbor Symp. Quant. Biol. 42,199-206 Bode, J., Hence, K., and Wingender, E. (1980) Eur. J. Biochem.

110,143-152 Johnson, E. M., Sterner, R., and Allfrey, V. G. (1987) J. Biol.

Chem. 262,6943-6946 Sterner, R., Boffa, L. C., Chen, T. A., and Allfrey, V. G. (1987)

Nucleic Acids Res. 11, 4375-4391 Chen, T. A., and Allfrey, V. G. (1987) Proc. Natl. Acud. Sci. U. S.

A. 84,5252-5256 Yau, P., Thorne, A. W., Imai, B. S., Matthews, H. R., and

Bradbury, E. M. (1982) Eur. J. Biochem. 129,281-288 Waterborg, J., and Matthews, H. R. (1984) Eur. J. Biochem. 142,

329-335 Senshu, T., and Iwai, K. (1970) J. Biochem. (Tokyo) 67,473-485 Wangh. L.. Ruiz-Carrillo. A.. and Allfrey, V. G. (1972) Arch.

Bi&em.‘Biophys. 150; 44-56 - Kobayashi, Y., and Iwai, K. (1970) J. Biochem. (Tokyo) 67,465-

471 Linder, H., Helliger, W., and Puschendorf, B. (1986) J. Chroma-

togr. 357,301-310 Gurley, L. R., D’anna, J. A., Blumenfeld, M., Valdez, J. G.,

Sebring, R. J., Donahue, P. R., Prentice, D. A., and Spall, W. D. (1984) J. Chromatogr. 297, 147-165

Lawson, G. M., and Cole, R. D. (1979) Biochemistry 18, 2160- 2166

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Nucleosomes Containing Fully Acetylated H3 and H4

45.

46. 47.

48.

49.

50.

51.

52.

53. 54. 55.

56.

51.

58.

Hayashi, T., Ohe, Y., Hayashi, H., and Iwai, K. (1980) J. Biochen. (Tokyo) 88,27-34

Bradford, M. M. (1976) Anal. Biochem. 72,248-254 Wharton, D. C., and McCarty, R. E. (1972) Experimental Methods

in Biochemistry, pp. 213-215, Macmillan Publishing Co. Inc., New York

Simpson, R. T., Thoma, F., and Brubaker, J. M. (1985) Cell 42, 799-808

Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. (1987) Current Protocols in Molecular Biology, pp. 1.7.1-1.7.4, Greene Publishing Associ- ates and Wiley-Interscience, New York

Cook, G. R., Minch, M., Schroth, G. P., and Bradbury, E. M. (1989) J. Biol. &em. 264, 1799-1803

Gaastra, W., and Josephsen, J. (1984) in Methods in Molecular Biology (Walker, J. M., ed) Vol. 2, pp. 273-376, Humana Press Inc., New York

Greyling, H. J., Schwager, S., Sewell, B. T., and von Holt, C. (1983) Eur. J. Biochem. 137, 221-226

Gottesfeld, J. (1987) Mol. Cell. Biol. 7, 1612-1622 Zweidler, A. (1978) Methods Cell Biol. 17, 223 Panyim, S. and Chalkley, R. (1969) Arch. Biochem. Biophys. 130,

337-346 Schreier, M. H., Erni, B., and Staehelin, T. (1977) J. Mol. Biol.

116,727-753 Bode, J., Gomez-Lira, M. M., and Schroter, H. (1983) Eur. J.

Biochem. 130,437-445 Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular

Cloning: A Laboratory Manual, pp. 475-477, Cold Spring Har- bor Laboratory, Cold Spring Harbor, NY

59. Hanson, J. C., Ausio, J., Stanik, V. H., and van Holde, K. E. (1989) Biochemistry 28,9129-9136

60. White, J. H., Gallo, R., and Bauer, W. R. (1989) Nucleic Acids Res. 17,5827-5835

61. Simuson. R. T., and Stafford, D. W. (1983) hoc. N&l. Acad. Sci.

62. Chicoine, L. G., Schulman, I. G., Richman, R., Cook, R. G., and Allis. C. D. (1986) J. Biol. Chem. 261, 1071-1076

Pesis, K. H., and Matthews, H. R. (1986) Arch. Biochem. Biophys. 251,665-673

Cary, P. D., Crane-Robinson, C., Bradbury, E. M., and Dixon, G. H. (1982) Eur. J. Biochem. 127, 137-143

Ebralidse. K. K.. Grachev, S. A., and Mirzabekov, A. D. (1988) Nature ‘33 1, 365-367

Satchwell. S. C.. Drew. H. R.. and Travers. A. A. (1986) J. Mol. Biol. 161,65&675 '

Drew, H. R., and Calladine, C. R. (1987) J. Mol. Biol. 195, 143- 173

63.

64.

65.

66.

67.

68.

69.

70. 71.

72.

73.

74.

Gale, J. M., and Smerdon, M. J. (1988) J. Mol. Biol. 204, 949- 958

Klug, A., Rhodes, D., Smith, J., Finch, J. T., and Thomas, J. 0. (1980) Nature 287, 509-516

Rhodes, D., and Klug, A. (1986) Cell 46,123-132 Kefalas, P., Gray, F. C., and Allan, J. (1988) Nucleic Acids Res.

16,501-517 Oliva, R., Bazette-Jones, D. P., Locklear, L., and Dixon, G. H.

(1990) Nucleic Acids Res. 18, 2739-2747 Uberbacher, E. C., and Bunick, G. J. (1989) J. Biol. Struct. Dyn.

7,1-18 Norton, V. G., Marvin, K. W., Yau, P., and Bradbury, E. M.

(1990) J. Biol. &em. 265, 19848-19852

U: S. i. 80,51-55

4 The presence of two EcoRI sites per monomer phasing sequence noted by Hansen et al. (59) requires an actual length for the monomer unit of 208 bp.

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

19846 Nucleosomes Containing Fully Acetylated H3 and H4

0 .

R a?

0.1

(

A a

A

i0 *

2

k

O\O

5

1-

B abcd~e f

time (min)

L H3.2 H3.3 H2B

H4

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Nucleosomes Containing Fully Acetylated H3 and H4

A 554532SlOS55S

54321OSC

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from

K W Marvin, P Yau and E M Bradburyassembly into nucleosomes.

Isolation and characterization of acetylated histones H3 and H4 and their

1990, 265:19839-19847.J. Biol. Chem. 

  http://www.jbc.org/content/265/32/19839Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/265/32/19839.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on February 1, 2018http://w

ww

.jbc.org/D

ownloaded from