Vol. 168, No. 2, 1990 April 30, 1990

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 637-643

USE OF AN IMMOBILIZED ENZYMEAND SPECIFIC ANTIBODIES TO ANALYSETHE ACCESSIBILITY AND ROLEOF HISTONE TAILS IN CHRUNATIN SIWJCTURE Marie-Frangoise HACQUES’ , Sylviane HULLER*, Gilbert De HURCIA3, Marc H.V. Van REGENHORTEL * and Christian MARION ‘* ‘Laboratoire de Physico-Chimie Biologique, LBTN-CNRS UH 24, Universite Lyon-l, 43 Bd. du 11 Novembre1918, 69622Villeurbanne Cedex, France * Laboratoire d’bmnunochimie and 3 Laboratoire de Biochimie, IBHC, 15 rue ReneDescartes, 67084Strasbourg Cedex, France Received

March

5, 1990

Using limited proteolysis with subtilisin bound to collagen membranes,the degradation of the histone proteins revealed by specific antibodies wascorrelated to changes in chromatin conformation and condensation monitored by circular dichroism and electric birefringence. This new approach allows us to detect for the first time a hierarchy of histone tails cleavages. The terminal domainsof Hi, the NH=-terminal tail of H3 and the carboxy-terminal ends of histones H2Aand H2Bwere found to be cleaved already at the early stages of proteolysis and this led to a decondensationof polynucleosomal chains. Thereafter the C-terminal part of H3 and both NH=-terminal regions of H2Aand H2Bbecamerapidly cleaved, resulting in relative reorientation of swinging nucleosomesor partially unfolded segments.Unexpectly, this removal of tails of Hi, H2B, H2A and H3 is not accompagnied by significant changesin DNA-protein interactions resulting in free-oriented DNA.This might suggest that hi&one-hi&one interactions play a central role in stabilizing the solenoid. 01990 Academic Press, Inc.

The extremely polymorphic and highly modified histones Hi, which are essential for the formation and/or stabilization of the higher-order chrwatin structure, have been proposed as repressors of gene activity (1). However the transition from condensed to open chromatin is thought to be also modulatedby differential binding of various proteins (non-histone proteins, histone variants or modified histones), as well as by DNA modifications. Knowledge of the accessible histone domainswithin polynucleosomal chains is therefore of particular interest since these are potential sites of both post-translational modifications and interactions with non-histone proteins. Relatively little is knownabout the exact role in higher-order structure of not only the different danains of Hl but also core histone tails, which appear not to be involved in the stabilization of the nucleosomes.Hence, the primary function of core histone tails, as well as Hl, may be the stabilization of the solenoid (2,3). Proteolytic digestion of chromatin, which has proven useful to define accessible histone danains or to unfold chranatin structure by removing histone tails, has so implicated the tails of H4 or H3 in chromatin expansion (for a review, see reference 4). The involvement of H3, especially of its NHz-domain,was also suggestedby *To whanall correspondenceshould be addressed.

637

0006-291X/90 $1.50 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

Vol.

168,

No.

2, 1990

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

However proteolysis leads to ‘limit” polypeptides and phosphorylation studies f5,6). no study has been reported describing the precise order of degradation of histone domains nor the exact correlation between their removal and the salt-induced filament = solenoid transition of chranatin. Indeed all histones have several regions that are particularly accessible to soluble proteases and the multiple cleavages lead to ambiguous evaluation of the hierarchy of domain importance for stabilizing the higher-order structure. In this paper we used a new approach allowing the investigation of both the topography and the structural role of histones in chrcxnatln. The accessible protein regions at the chranatin surface were selectively proteolysed with inrnobillzed subtilisin, which cleaves a wide range of residues, and the precise identification of these regions was made by antibodies specific for histones and histone peptides. These cleveages were correlated to structural changes revealed by sensitive biophysical techniques, i.e. circular dichroism and electric birefringence. MATERIALSAND METHODS The preparation and characterization of rat liver chromatin were identical to those described previously (7). The intnobilization of subtilisin (3.4.21.14; from Bacillus subtilis, Boehringer) on 15 x 15 nm collagen films was as described W, and the specific activities of enzymatic membranes, determined at 22’C by the rate of hydrolysis of casein, varied from 0.11 to 0.12 AzF3 unit/min/fmn* of film. Polynucleoscmal chains (45 + 10 nucleosomes) were adjusted to a concentration of 1.25 AzaD units/ml in TE buffer (10 midTris-HCI, pH 7.4, 0.2 d EDTA). Samples consisting of 10 ml of the chromatin suspension were digested at 20’C for times varying fran 0 to 60 min, prior to being put on ice and divided into aliquots, in order to do the analysis by biophysical methods and to characterize the proteolytic products using antibodies. Native chromatin or digested samples were analysed by SDS-16% polyacrylamide gel electrophoresis (9), revealed with Coanassie blue (9 ug protein per well), or transferred to nitrocellulose paper Gchleicher and Schiill; 0.2 Pm; 7 pg protein per weI1) according to Towbin et al. (10). After blocking remaining free sites with 1% BSA, the filters were incubated with various specific antisera, revealed by 12sI-labelled protein A (Amersham IM 144; 0.7 mCi/ml) and exposed for autoradiography at -70’C (11). The synthetic peptides corresponding to sequences in calf thymus histones were prepared by the solid-phase method of Barany and Merrifield (12) and their purity and amino acids composition were controlled as described (13). Antibodies to synthetic peptides were obtained as previously described (14.15). Antisera to each individual histone were obtained by immunizing rabbits with each histone canplexed to yeast RNA (16). The analysis of chranatin samples by electron microscopy and circular dichroism have been described elsewhere (7). The apparatus, procedures and calculations for electro-optical measurements were as described (5,7). RESULTSAND DISCUSSION Significant electrophoresis and bmnunoblottlng experiments on native and various digested samples are described in Figure 1. The specificity of anti-histone and anti-histone peptide antisera was checked in immunoblotting and in ELISA using pure isolated histones. The imnrnochemical study of native chromatin shows no significant cross-reactivity between antibodies and non-related histones (Fig. 1A). Antibodies against the N-terminus of a histone reveal polypeptides lacking its C-terminal end since histone basic termini are the most exposed to proteolytic cleavage, contrary to their globular domain (4). Antibodies specific for the C-terminal hexapeptide 130-135 of H3 (Fig. lA,d) reveal a natural fragment fram H3 lacking its N-terminal end, which was called P’l by B&m et al. (17). As expected, native chromatin exhibits a 638

Vol. 166, No. 2, 1990

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Ii l-l lil.zt Hlor

"k HZBH2Ay tl4r

a

bcdef

q

hi

j

a

bcdef

!

0’ A

I

9h1

1

----t=z==

___-_

I

I

;

I

I

Pigrre 1 - Study in immrnoblottlngof chrcemtin digestedby immobilizedeubtilisin using antibodies to hlstonesand hlstone peptides. Native chranatln (A) or chranatln samples digested for 5 mln (8). 10 mln (0. 33 mln (D) or 60 mln (E) were analysed by SDS-1611; polyacrylamide gel electrcphoresisand revealed with Cocmrassie blue (a) or transferred to nitrocellulose paper, incubated with varlous speclflc antlsera, and

revealed by *2sI-labelled protein A (b-j). Antlsera were as follows: anti-globular drzmain of Hl’ fb), anti-peptide l-21 of H3 fc), anti-peptide MI-135 of H3 (d), antipeptide l-25 of H2B fe), anti-peptlde 110-125 of H2B ff), anti-H2A fg), antl-peptlde

12-26of H2A(h), anti-peptlde 116-129of H2A(I) and antl-H4 (j). diluted 1:lOO(c.f,h,l), 1:500 (b,e), 1:lOOO (d,j) or 1:1200(g).

Antisera were

positive birefringence and the longest relaxation time ~~~ which represents the rotation of the whole chranatin particles (5,?f, is of the order of 170-175ps (Table 1). Increase of the NaCl concentration causes a reduction of TV, while its contribution to the decay kinetics a2 increases. This is consistent with the formation of condensedcylindrical particles (5). After 5 min digestion (Fig. lB), the fragnents from Hl lack residues in their Cand/or N-terminal tails (b), whereasthe N-terminal part of H3 (d) and the C-terminal parts of H2B (e) and H2A(g,h) are rapidly attacked. The C-terminal part of H3 cc> and the N-terminal part of H2B(f) also begin to be cleaved but these molecules are still intact after 2 min digestion (data not shown). The degradation of these tails of Hl, H3, H2Rand H2B, induces a decondensationof chranatin fibres revealed by the increase of both the positive An and TV. The latter indicates that the linear dimensions of the fibres lengthen by about 15%(Fig. 2). The increase of the Kerr constant is about two-fold (Table 1). A negative contribution with relaxation time of the order of 2-3 ps appears in the birefringence signal (Fig. 2). As the magnitudeof this negative An is very low compared to free DNAand the c.d. is surprinsingly constant, this might suggest that it arises frcnn hinge points betweennucleosomesor orientation of partially unfolded segnents rather than from the presence of extended linker DNA. After 10 min digestion (Pig. IC), the N- and C-terminal tails of H3 (c,d) and H2B (e,f) along with the C-terminal part of H2A continue to be degraded. Digestion products of H4 Cj) and polypeptides from H2A, lacking its N-terminal part (i), appear. 639

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Vol. 168, No. 2, 1990

TABLE t Analysis

of electric

blrefringence

and circular

dichroisn

of control

and selectively

Im imic drvagth mclrlc BlrvfdWJMFc

Kerr cmetant 8.W (em cge)

proteolyzed

chrmatin

hi@ ionic slremlh

Rclualim ttr W

?kdbilltg cavfflcient

hiar ellipticitv kkg.&/hl)

KlecWc llrefrlvLEl?ce

Kerr metant B.W (em cm)

Relarrtim tiac (ye)

?lexlbliity cafflcicat

Mar clliptlcitl (deg.mvaol

0

Pmltlve

t 0.132

172

0.63

ml

PrnlUW

+ 0.033

42

1.00

1965

2

emItIre

4 0.24a

190

0.58

208

emitin hegatirr part)

4 0.055

42

1.00

1930

5

Pmitire

I 0.260

298

0.52

i?m

tkgatire ~pmltlve part1

- 0.011

2-3

1.00

Ipso

10

Pvsitise

4

a.245

214

0.68

2175

lk~livc tpvsftlve part)

- 0.014

4-5

1.00

mo

30

Pmitire

4

0.240

222

0.55

2315

Wtmtlvc

- 0.023

26

0.52

2170

En

Povitivv (negatlvv part)

4

0.143

ala

0.66

2v2s

lkg3llw

- a.032

4a

0.54

ma

j

Blrefringcnce meacurcacnts are performed In 10 mH Trls-flCI (pII 7.4) 0.2 d KDTA, cootainlng 60 d-Hal3 or not, at an applied electric field of 1600 V/cm. The values of specific Kerr constants, B, and of relaxation times are given as standard deviations of 6 to @& and of g to 10%. reqwctively. e.s.u.. electrostatic units. ‘Ike flexibility coefficient is the cmtributim of the longest relaxation Urn, I +, to the birefringence &cay secbanim. llolar elllpticlty values are f ZOO.

During this step, the A n amplitude decreases at low ionic strength and the signal becomes predominantly negative at 60 mH-NaCl, the positive contribution arising from the orlentatlon of canpact structures which stll l exist, as revealed by electron microscopy (data not shown). The slight increase of c.d. is consistent with a modest unfolding of chrcnnatin (Table 1). After 30 min digestion (Fig. lD), no intact Hl (b), H3 (c,d) and H2B (e,f) are revealed. Kost of Hi-digestion products have disappeared and only two Hi peptides are detected (b): the first one, which presents the higher electrophoretic mobility, is the globular domain of Hi and the second one probably corresponds to Hl devoid of its C-terminal region (18). An has still decreased at low ionic strength and is became entirely negative at 60 mM-NaC1 (Table 11, as a result of the very limited number of nucleosrmes engaged in canpact structures. Asswning that DNA has either its proper native chranatin folding or the B-form, the c.d. change can be explained by the contribution of 6-8% of protein-free DNA. When no more intact histone is detectable (Fig. lE), the molar ellipticity increases about 1.5 fold compared to native chranatin while the An signal become composite at low ionic strength, as it was observed when very lysine-rich histones were removed by salt treatment (19). However, these drastic changes resulting fran the unfolding of the chranatin fibres together with an overall rotation of nucleosarnes, were observed as the globular part of Hi still persists (b). Besides, the globular parts of H3 (c,d), H2B (e,f) and H4 C.i) are also resistant, contrary to histone KU, whose numerous fragnents still appear (g,h,i). 640

Vol. 168, No. 2, 1990

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

1

200

400

0

I,

I

40

1

I

I

80

1

120

Time(ps)

Time (ps)

Figure 2 - Analysis of blrefringence decays and determination of relaxation times of control chranatin (0) and chraaatin proteolyzed for 5 min with inmcbiiized eubtilisin (0). Heasurements were made In 10 sin Tris-HCI (pH ?.4), 0.2M EDTA containing indicated NaCl concentrations. The field strength used was 1500 V/cm. In a2 Is calculated by extrapolating the linear part of the plot.

The immobilization technique of enzymes on collagen membranes, which has been developed in our laboratory (8,20), Presents numerous advantages, as an enhanced stability, an easy control of Proteolysis, no contamination of the digests and a wide range of enzymatic activities. Furthermore, it avoids filtration and inhibitor addition necessary in using camnercial try&n-glass (21). The use of imnobiIized subtilisin demonstrates that it is possible to study the effect of limited digestion of histone domains, the removal of which can be correlated with chranatin structural changes. Such a study has previously been described neither for Hi daaains, nor for core histone domains at the level of polynucleoscmal chains. This new approach shows for the first time both the accessibility of the terminal regions of H3, H2A and H28, as well as Hl, and the consequences of their removal for the ability of chromatin to adopt a condensed structure. Up to now, the involvement of core histone tails had essentially been postulated from their “non-role’ in stabilizing core particles and as being the locations of post-synthetic modification sites, except for the N-terminal part of H3 for which the data appeared to be most conclusive from protease digestion and phosphorylatlon studies. Our results lead us to conclude that the unstructured N- and C-terminal tails of HI are accessible at the surface of fibres, contrary to the globular danaln which seems very resistant to immobilized enzymes and therefore internally located. This suggests that the in vivg poly(ADP-ribosyljation of histone Hl, whose sites are located in both terminal regions, can contribute to the decondensation of chrazatin, as that has been demonstrated JJ vitrp (22). However, other regions of histones also take part in this process. The N-terminal region of H3 has a very external I ocat on in chraaatin, which is an additional argument to involve it in the Moreover its accessibility may stabi ization of higher-order structure (6,20,23). expla ,i n that during mitosis, all the H3 molecules can be phosphorylated on serine 10 (24). The C-terminal regions of H2A and H2B appear also to be very accessible in chromatin fibres and their cuts probably play a role in the decondensation of chrcmatin. This exposed location has already been suggested for H2A but never for H2B, which has no trypsin-susceptible C-terminal tail (4). This finding is particularly important since nuclear conjugation of ubiquitin to H2A and H2B, which could be 641

Vol. 168, No. 2. 1990

BIOCHEMICAL AND BiOPHYSlCAL RESEARCH COMMUNICATIONS

associated with transcribed regions of the genane (251, or with DNA repair mechanisms (261, is confined to these carboxy-terminal ends. The C-terminal hexapeptide of H3 (IRGBRA) and the N-terminal regions of H2A and H2B seem also accessible at the chromatin surface and might be involved in chromatin expansion. On the contrary, H4 appears to be the most protected histone and is not implicated in the structural changes observed during the beginning of digestion. This could suggest that the major part of histone H4 might be involved in nucieosome stabilization rather than in chromatln condensation. Such a very specific role could explain the discrepancies between hyperacetyiation data. During the course of this study, an unexpected observation was made. The decondensation of chranatin at the beginning of digestion as well as further changes in structural properties are not accompanied by the appearance of naked DNA. This could mean that the cleaved domains were involved in protein-protein interactions rather than in DNA-protein binding. Intimate histone-histone contacts might then play a crucial role in the stabilization of chranatin, as previously_suggested (6,27). Starting from this working hypothesis, studies are now in progress to obtain information on the precise role of each histone domain. Indeed, at the earliest time point for which digestion data are presented, the cleavage of Hl is almost complete and this alone may cause the chromatln not to fold up. It Is therefore not possible from these data to assess additional effects of the cleavage of H3, H2A and H2B tails. Besides, it will be interesting to study the Na+ dependence on the protease cleavage kinetics, and to determine whether the changes observed here are due to cleavage or to the possible dissociation of cleavage products. ACKNOWLEDGMENTS We are grateful to Drs E.M. Bradbury and K. Logan (Davis) for a glft of anti G-Hl’ antiserum, Dr J.P. Briand and S. Plaue for the preparation and purification of synthetic peptides, and C. Van Herrewege for the photographs. M.F.H. has a feliowshlp with the Association pour ia Recherche sur le Cancer. REFERENCES

a:

3. i:

6. ii: 1:: :a: 13. :54: 16.

Weintraub, H. (1985) Cell, 42, 705-711. McGhee, J.D., Rau, D.C., Charney, E. and Pelsenfeld, G. (1980) Cell, 22, 87-96. Allan, J., Harborne, N., Rau, D.C. and Gould, H. (1982) J. Cell. Biol. 93, 285-297. Btihm, L. and Crane-Robinson, C. (1984) Blosci. Rep. 4, 365-386. Marion, C., Martinage, A., Tirard, A., Roux, B., Daune, M. and Mazen, A. ( 1985) J. Mol. Biol. 186, 367-379. Hazen, A., Hacgues, M.F. and Marion, C. (1987) J. Mol. Biol. 194, 741-745. Marion, C., Roche, J., Roux, B. and Gorka, C. (1985) Biochemistry, 24, 6328-6335. Marion, C., Roux, B., Pallotta, L. and Coulet, P.R. (1983) Biochem. Biophys. Res. Conmiun. 114, 1169-1175. Laemnli, U.K. (1970) Nature, 215, 360-363. Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl. Acad. Sci., U.S.A. 76, 4350-4354. Huller, S. and Van Regenmortel, N.H.V. (1989) Methods in Enzymoi. 170, 251-263. Barany, G. and Merrlfleld, R.B. (1980) in The Peptldes (Cross, E. and Nelenhofer, J. eds.) Vol. 2, pp. l-284, Academic Press, New York. Nuller, S., Couppez, H., Briand, J.P., Gordon, J., Sautiere, P. and Van Regenmortei, M.H.V. (1985) Blochlm. Blophys. Acta, 747, 100-106. Huller, S.. Hiamelspach, K. and Van Regenmortel, N.H.V. (1982) HNBD J. 1, 421-425. Huller, S., Piaue, S., Couppez, M. and Van Regenmortel, N.H.V. (1986) Holec. Iimnunol. 23, 593-601. Stoilar, B.D. and Ward, M. (1970) J. Bioi. &em. 245, 1261-1266. 642

Vol. 168, No. 2, 1990

BIOCf-fEMiCAL AND BfOPHYSICAL RESEARCH COMMUNlCATfONS

17. Bf%un,L., Briand, G., Sautiere, P. and Crane-Rhinaon, C. (1981) Eur. J. Biochem. 119, 67-74. 18. Thoma, F., Losa, R. and Keller, T. (1983) J. Mol. Biol. 167, 619-640. 19. Roche, J., Marion, C., Gorka, C., Roux, B. and Lawrence, J.J. (1984) Biochem. Biophys. Res. Camnun. 121, 538-537. Marlon, C., Roux, B. and Coulet, P.R. (1983) FEBSLett. 157, 317-321. ff : Ausio, J., Bong, F. and van Holde, K.E. (1989) J. Mol. Blol. 206, 451-463. 22. Poirier, G.G., de Hurcia, G., Jongstra-Bifen, J., Niedergang, C. and Handel, P. (1982) Proc. Nat]. Acad. Sci. USA, 20, 3423-3427. Saccone, G.T.P., Skinner, J.D. and Burgoyne, L.A. (1983) FEBSLett. 157, 111-114. E Gurley, L.R., Tobey, R.A., Walters, R.A., Hildebrand, C.E., Hohmann, P.G., D/Anna, J.A., Barham, S.S. and Deaven, L.L. (1978) in Cell Cycle Regulation (Jeter, J.T., Cameron, I.L., Padilla, G.M. and Zimmerman, A.M., eds), pp. 37-60, Academic Press, NewYork. 25. Reeves, R. (1984) Biochim. Biophys. Acta, 782, 343-393. Jentsch, S., HcGrath, J.P. and Varshavsky, A. (1987) Nature, 329,131-134. 5 Allan, J., Mitchell, T., Harborne, N., B6hm, L. and Crane-Roblnson, C. (1986) J. Mol. Biol. 187, 591-601.

643