histone h2b (and h2a) ubiquitination allows normal histone octamer and core particle reconstitution

ELSEVIER Biochimica et Biophysica Acta 1218 (1994) 187-193

Biochi~ic~a et Biophysica AFta

Histone H2B (and H2A) ubiquitination allows normal histone octamer and core particle reconstitution

Neil Davies, George G. Lindsey *

Department of Biochemistry, University of Cape Town, Private Bag, Rondebosch 7700, South Africa

(Received 20 July 1993)

Abstract

146 bp core particles were assembled from reconstituted hybrid histone octamers where either histone H2A or H2B were replaced by their ubiquitinated counterparts uH2A and uH2B. No difference in the structure of the core particles was evident upon DNase 1 digestion suggesting that ubiquitination of these histones was not a barrier to normal core particle formation.

Key words: Ubiquitin; Histone; Core particle; Nucleosome

I. Introduction

Histones are subjected to several post-translational modifications including phosphorylation, acetylation, methylation, ADP-ribosylation, glycosylation and ubiquitination (for a review see Ref. 1). Ubiquitination involves the coupling of the C-terminus of the 76 amino acid polypeptide ubiquitin to an E-amino group of the histone via an isopeptide bond. Of the histones, only histones H2A and H2B have thus far been found to be ubiquitinated at lysine-ll9 and lysine-120, respectively [2,3] to yield the ubiquitinated derivatives uH2A and uH2B. These conjugates are present at low concentrations, less than 10% of the parent histone, in a wide variety of species [4].

The role of the ubiquitinated histones in chromatin is still unclear but it has been shown that these histones are absent from condensed chromatin [5-7]. A number of workers [8-11] have therefore investigated whether there is correlation between the presence of these histones and transcribed chromatin, a state in which the chromatin must be maximally accessible to RNA polymerase. The transcriptionally active Tetrahy- mena macronucleus has been shown to have markedly elevated levels of uH2B relative to the transcriptionally inactive micronucleus [11]. Moreover, inhibition of

* Corresponding author. Fax: + 27 21 6855931.

0167-4781/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0167-4781(93)E0289-Z

transcription in breast carcinoma cells has been shown to result in a virtual disappearance of uH2B [12]. A correlation between transcriptional activity and uH2A content is less clear. Although some workers have shown a relative enrichment of uH2A in preparations enriched in 'active' chromatin [8,11,13], this finding has not been confirmed by others in the field [9,14,15]. Thus, inhibition of transcription has not been found to effect uH2A levels in either breast carcinoma cells [12] or HeLa cells [16] although uH2A levels are higher in the Tetrahymena macronucleus compared with the micronucleus.

This absence of correlation between transcribed chromatin (with its putative altered nucleosomal structure [17]) and uH2A content agrees with the finding that nucleosomes containing uH2A in place of histone H2A have an unaltered structure as determined by DNase 1 digestion [18]. Whereas the ubiquitination site in histone H2A is trypsin sensitive (and therefore ex- posed) in nuclei and chromatin, that in histone H2B is not (for a review see Ref. 19). It has therefore been proposed [12] that ubiquitination of histone H2B oc- curs during transcription through alteration of the normal nucleosome structure. This modification then im- pedes nucleosome refolding thereby maintaining the chromatin in an active conformation. In this report we have used reconstitution methodology [20,21] to re- place both molecules of histones H2A and H2B one at a time in the chicken erythrocyte histone octamer with

188 N. Davies, G.G. Lindsey /Biochimica et Biophysica Acta 1218 (1994) 187-193

Table 1 Amino acid analysis of uH2A

Amino acid Moles found Calculated

Asp 14.6 (15) 16 Thr 10.3 (10) 11 Set 7.5 (7) 8 Glu 25.1 (25) 23 Pro 10.7 (11) 8 Gly 19.9 (20) 19 Ala 21.3 (21) 20 Cys 0.0 0 Val 12.0 (12) 12 Met 0.2 1 lie 12.1 (12) 13 Leu 24.6 (25) 25 Tyr 2.7 (3) 4 Phe 3.5 (3) 3 Trp 0.0 0 His 6.4 (6) 4 Lys 20.9 (21) 21 Arg 17.4 (17) 16 Total 209.1 (209) 204

the ubiquitinated derivatives of these histones from calf thymus. The reconstituted octamers could be assembled into core particles which were found to have an unaltered structure as determined by DNase 1 digestion.

2. Materials and methods

Ubiquitinated histones H2A and H2B were pre- pared from calf thymus glands essentially as described previously [3,22] except that high molecular weight contaminants were removed by ion-exchange chromatography after separation of total histones H2A and H2B on Biogel P60. This ion-exchange chromato- graphic step was carried out on a CM 52 column in 50 mM sodium acetate 6 M urea 50 mM NaC1 (pH 5.4); uH2A and uH2B were eluted with a linear gradient of 125 mM to 175 mM NaC1 in this same buffer. The eluted ubiquitinated histones were enriched by this procedure as they eluted at a lower ionic strength than histones H2A and H2B presumably due to the presence of the ubiquitin moiety. Final purification was by Biogel P60 chromatography in 2 mM Tris-HCl 7 M urea (pH 2.2).

H1

H3

H 2 B

H 2 A

H4

5 U U H H 2 2 A (5

Fig. i. SDS-PAGE of purified ubiquitinated histones. The standard (S) is a total acid extract of calf thymus nuclei.

Reconstitution of histone octamers containing either uH2A or uH2B replacing either histone H2A or H2B respectively in the chicken erythrocyte histone octamer was carried out as described previously [23,24]. In brief, to equimolar quantities of histones uH2A (or uH2B) in 2 mM Tris-HC1 7 M urea (pH 2.2) and histones H3, H2B (or H2A) and H4 in 50 mM NaC1 (pH 1.7), urea and 2-mercaptoethanol were added to 7 M and 0.1 mM respectively before the mixture was dialysed against 2 M NaC1 10 mM Tris-HC1 (pH 7.4). Octamers were separated from contaminating H3-H4 aggregates and H2A-H2B dimers by Sephacryl $300 chromatography in this same buffer.

The preparation of hybrid core particles was also carried out as described previously [20]. Hybrid histone octamers in 2 M NaC1 10 mM Tris-HC1 (pH 7.4), polyglutamic acid (PG 340) in 10 mM Tris-HCl (pH 7.4) and long (> 1000 bp) DNA in 0.1 mM EDTA were mixed in the ratio 1:3:1.7 by weight before dialysis against 20 mM NaCI 10 mM Tris-HC1 (pH 7.4) so as to assemble octamers onto long DNA (polycores). The polycore preparation was then digested with Micrococ- cal nuclease (MNase) in this same buffer after the

Fig. 2. (a) Sephacryl $300 chromatography of the products of octamer reconstitution of chicken erythrocyte histones H3, H2A and H4 together with calf thymus uH2B. The front half of the first eluting peak (fractions 30-33 inclusive) was pooled as was the second peak (fractions 37-43 inclusive). On the basis of the elution volumes and the histones present in these pooled fractions, the peaks were assessed to represent histone octamers and histone H2A-H2B dimers, respectively. Essentially identical elution profiles were observed for control and uH2A octamers (not shown). (b-d) SDS-PAGE of the proteins present in the first eluting peak, lanes 1, and the second eluting peak, lanes 2, from Sephacryl $300 chromatography. The standard (S) is a total acid extract of chicken erythrocyte nuclei. Octamers were reconstituted from chicken erythrocyte core histones (b), with uH2A replacing H2A (c) and with uH2B replacing H2B (d). (e) 6% non-denaturing PAGE of the products of MNase digestion of uH2B hybrid octamers (Fig. 2d, lane 1) assembled onto long DNA. The digestion was stopped at 0, 0.5, l, 2, 5, 10, 20 and 40 min (lanes 1-8). The standard (S) is a HpalI digest of pBR 322.

0.6

0 0

0 .5 c-

O t n e4 0.4

u 0.3 t-

O .I0 £.

o 0.2 ol

.<

0.1

5 10 15 20 25 30 35 4 0 4 5 5 0

N. Davies, G.G. Lindsey /Biochimica et Biophysica Acta 1218 (1994) 187-193 189

F r a c t i o n

H I / H 5

H3 H2B H2A H4

u H 2 A uH2A

s 1 s 1 2 s 1

c d

s I 2 3 4 5 6 7 8 5

bp

e

190 N. Davies, G. G. Lindsey / Biochimica et Biophysica Acta 1218 (1994) 187-193

addition of CaC12 to 1 mM. Core particles were then 0.5 purified from the digest by 5-20% sucrose gradient centrifugation in this same buffer.

DNase 1 digestion of the assembled core particles, 0.4 labelled with [32p]ATP using T4 kinase, was carried "E

¢..

out as described previously [29]. Autoradiographs of ,, Ln 0.3

the DNase 1 gels were scanned with a laboratory built ¢~ densitometer. This instrument uses a microscope lens to collect light transmitted from a halogen source ~ 0.2

0 through the autoradiograph to a photocell with a spec- tral response between 400 and 600 nm. Accordingly the ,,°

.ID

instrument has a resolution < 0.1 mm. < 0.1

3. Results and discussion

The identity of the purified ubiquitinated histones (Fig. 1) was confirmed by gas phase protein sequencing a [26,27]. and amino acid analysis [28]. UH2B yielded equivalent quantities of two amino acid residues at each cycle of the Edman degradation; these were Met and Pro, Gin and Glu, lie and Pro. These amino acids H I / H 5 correspond with the first three residues of ubiquitin (M-Q-I) and calf thymus H2B (P-E-P). UH2A yielded only the sequence M-Q-I; no sequence from the histone component of uH2A was detected as the N- H3 H2B terminus of H2A is acetylated and thereby blocked to H2A the Edman chemistry. Amino acid analysis (Table 1) of H4

uH2A showed the combined amino acid composition of ubiquitin and histone H2A.

Hybrid histone octamers were therefore reconstituted using either uH2A in place of histone H2A or uH2B in place of histone H2B; control octamers containing non-ubiquitinated histones were also reconsti- b tuted. Essentially identical results were observed for both octamer reconstitution and core particle assembly whether non-ubiquitinated chicken erythrocuye core histones were used or whether uH2A or uH2B replaced their non-ubiquitinated counterparts. The re- suits shown are for uH2B hybrid octamers. The products of octamer reconstitution were individually ap- plied to a Sephacryi $300 column (Fig. 2a); in all cases two peaks eluted. Analysis by SDS-PAGE (Fig. 2b) showed that the first peak contained all the histone species in stoichiometric quantities with the second peak having an excess of histones H2A and H2B (or ubiquitinated derivatives thereof) over histones H3 and

Fig. 3. (a) Isokinetic sucrose gradient ultracentrifugation profile of uH2B hybrid core particles. (b -d) SDS-PAGE of histones present in control (b) and hybrid (uH2A: c, uH2B: d) core particles purified by sucrose gradient centrifugation. The standard (S) is a total acid extract of chicken erythrocyte nuclei. (e) 6% non-denaturing PAGE of DNA present in control, uH2A and uH2B hybrid core particles (lanes 1-3, respectively). The standard (S) is a HpalI digest of pBR 322.

0 0

/

I I I I I I I 5 10 15 20 25 30 35

F r a c t i o n

uH2A uH2B

s 1 s 1 S

S 1 2 3

1 8 0 - 1 6 0 - - 1 4 7 - -

1 2 2 - -

bp

I 40 4 5

N. Davies, G.G. Lindsey / Biochimica et Biophysica Acta 1218 (1994) 187-193 191

H4. The first peak had the identical elution volume as control reconstituted octamers (not shown) and this, together with the stoichiometric equivalence of the histones in this peak, implied that hybrid histone octamers had indeed been reconstituted. Unfortunately, insufficient material was reconstituted to allow cross linking, the normal criterion for checking octamer au- thenticity, to be carried out. In our experience, however, elution volume and a stoichiometric histone equivalence are a reliable way of assessing the oc- tameric nature of the reconstitute. The efficiency of reconstitution for all octamers was in the range 25- 40%. The hybrid octamer containing fractions were pooled, assembled onto long DNA using polyglutamic acid and digested with MNase. Digestion to the 146 bp core particle proceeded through 167 bp and 157 bp intermediates (Fig. 2e) as reported previously for both natural [29] and reconstituted [25] core particles. No protection of linker DNA was observed as was the case with the S(T)PXX containing variants of both histone H2A [20] and histone H2B [21].

Hybrid and control core particles were purified on isokinetic 5 -20% sucrose gradients. Fractionation of the gradients revealed that all the core particles mi- grated as sharp co-incidental gaussian peaks (Fig. 3a). The purified core particles had stoichiometric amounts of their constituent core histones (Fig. 3b) and discrete DNA lengths of 146 bp (Fig. 3e). Core particle yield was in the range 7-10%. These results demonstrated that both uH2A and uH2B could be incorporated into the core particle with no obvious structural effect. This result for uH2A is in agreement with that reported previously [18].

DNase 1 digestion of core particles results in a 10 bp period due to the DNA being wrapped around the outside of the core histone octamer in a left-handed superhelix [30,31]. The sites of cleavage are not uni- formly susceptible to DNase 1 attack probably due to the sensitivity of the nuclease to the local conformation of the DNA. The positions of maximum protection from cleavage have been correlated with sharply bent regions of the DNA [32] where the DNA has been

1 6 0 - - 147 --

122 -- 110 - -

9 0 - -

7 6 - -

6 7 - -

3 4 - -

2 6 - -

s 1 2 3 4 5 6 7 8 s s 1 2 3 4 5 6 7 8 s s 1 2 3 4 5 6 7 8 s

i i

e

i

m

qlli

Fig. 4. Autoradiographs of denaturing PAGE of the products of DNase 1 digestion of 5'-32p end-labelled control and hybrid core particles. The digestion was stopped at 0, 0.25, 0.5, 1, 1.5, 2, 3 and 5 rain (lanes 1-8). The standard (S) is a HpalI digest of pBR 322 labelled by a 'fill-in' reaction with the Klenow fragment of E. coli DNA Pol 1 using [a-32p]GTP. The results shown are typical of such experiments which were repeated several times.

192 N. Dauies, G.G. Lindsey /Biochimica et Biophysica Acta 1218 (1994) 187-193

postulated to interact with basic regions of the histone octamer [33,34]. DNase 1 digestion of hybrid core particles was therefore carried out in order to investi- gate whether abnormal D N A binding occurred on ubiquitinated core particles. The products of DNase 1 digestion of 5' end-labelled control and ubiquitinated core particles were separated by denaturing P A G E (Fig. 4) and the autoradiographs densitometrically scanned (Fig. 5). The cutting patterns obtained for both control and uH2A core particles showed few differences and are in general agreement with those published previously [18,31,29]. Surprisingly the DNase 1 cutting pat tern for uH2B core particles was very similar to those of both control and uH2A core particles despite the trypsin insensitivity of the ubiquitination site. It would therefore appear that complete H2B ubiquitination had no major effect on the conformation of the core particle as determined by investigation with both MNase and DNase 1 since no dramatic changes in the rate of DNase 1 digestion were found to occur at any one cutting site. Thus, the 1 min (lane 4) digest of control particles, the 3 min (lane 7) digest of uH2A particles and the 2 min (lane 6) digest of uH2B particles show almost identical patterns. Moreover, the products of cleavage of sites known to be cleaved less

. . . . . . , , t , i

. . . . , , , , , i i

b

,,~,, L i J i ~ i

. . . . . . . , i i i

12 8 7 6 5 4 3 2 f

C u t t i n g s i t es ( n u m b e r )

Fig. 5. Densitometric scans of 0.25 min (a-c) and 2 min (d-f) DNase 1 digests of control and hybrid core particles shown in Fig. 4. The scans shown are for uH2A hybrid core particles (a and d), uH2B hybrid core particles (b and e) and control core particles (c and e).

8

g, 0

~ u b 0 . 2 5

146

bases

Fig. 6. Analysis of the radioactivity associated with DNA of length 146 bases after 0.25 min and after 2 min DNase 1 digestion. The partial scans shown, which are amplifications of the 146 bp region of the gels shown in Fig. 5, are for control (c), hybrid uH2A (ua) and hybrid uH2B (ub) core particles. The abscissa therefore represents distance.

readily, e.g., sites 8 and 11, and more readily, e.g., sites 5 and 9 were under- and over-represented on the gels respectively. Prior to the experiment we had antici- pated finding that ubiquitination of histone H2B in the core particle would result in gross conformational changes causing markedly higher and lower rates of DNase 1 digestion at certain sites. Comparison of the rate of DNase 1 digestion (Fig. 6) of the various core particles, however, showed that the ubiquitinated core particles were digested significantly more slowly than control core particles. The order of increasing digestion was uH2A < uH2B < control core particles. Dif- ferences in the rates of digestion together with other factors such as varying degrees of MNase nicking during particle preparat ion rather than conformational changes brought about by ubiquitination probably ac- count for the slight variations observed in the intensity of individual bands between the various particles (Fig. 6).

X-ray crystallography and two-dimensional N M R spectrosocopy [35,36] have both shown that ubiquitin has a tightly folded globular domain consisting of residues 1-72. The remainder of the molecule, residues 73-76, is in a random coil conformation and acts as a linking arm between the globular domain and the histone. Schematically, ubiquitin might be considered as a balloon, (the globular domain), attached to a larger structure, the core particle, via a string, (the linking arm). Our results suggest that the ubiquitin attached to histones H2A and H2B in the core particle has considerable lateral mobility thereby causing general rather than specific interference with DNase 1

N. Davies, G.G. Lindsey / Biochimica et Biophysica Acta 1218 (1994) 187-193 193

digestion. Additionally, our results would argue against the hypothesis that histone H2B ubiquitination main- tains the structure of transcribed nucleosomes in an open conformation [12]. The situation in vivo, however, would be affected by transient post-synthetic modifications of ubiquitinated histones such as acetylation and phosphorylation [37] which are known to decrease histone-DNA interactions. Alterations in histone structure have been shown [21,38] to prevent histone octamer reconstitution; thus carboxymethylation of Cys- 110 of chicken histone H3 or the replacement of chicken histone H2B by the elongated isohistone from sea urchin sperm would not allow octamer reconstitution although polyglutamic acid mediated assembly of core particle containing these histones could still be carried out.

Acknowledgements

We would like to thank the Foundation for Re- search Development and the University of Cape Town Research Fund for financial support.

References

[1] Van Holde, K.E. (1989) Chromatin, Springer, Berlin. [2] Goldknopf, I.L. and Busch, H. (1977) Proc. Natl. Acad. Sci.

USA 74, 864-868. [3] Thorne, A.W., Sautiere, P., Briand, G. and Crane-Robinson, C.

(1987) EMBO J. 6, 1005-1010. [4] Rechsteiner, M. (1988) Ubiquitin, Plenum Press, New York. [5] Matsui, S., Seon, B.K. and Sandberg, A.A. (1979) Proc. Natl.

Acad. Sci. USA 76, 6386-6390, [6] Wu, R.S., Kohn, K.W. and Bonner, W.M. (1981) J. Biol. Chem.

256, 5916-5920. [7] Mueller, R.D., Yasuda, H., Hatch, C.L., Bonner, W.M and

Bradbury, E.M (1985) J. Biol. Chem. 260, 5147-5153. [8] Levinger, L. and Varshavsky, A. (1982) Cell 28, 375-385. [9] Huang, S., Barnard, M.B., Xu, M., Matsui, S., Rose, S.M. and

Garrard, W.T. (1986) Proc. Natl. Acad. Sci. USA 83, 3738-3742. [10] Ridsdale, J.A. and Davie, J.R. (1987) Nucleic Acids Res. 15,

1081-1096. [11] Nickel, B.E., Allis, C.D. and Davie, J.R. (1989) Biochemistry 28,

958-963. [12] Davie, J.R. and Murphy, L.C. (1990) Biochemistry 29, 4752-

4757.

[13] Goldknopf, I.L., Wilson, G., Ballal, N.R. and Busch, H. (1980) J. Biol. Chem. 255, 10555-10558.

[14] Parlow, M.H., Haas, A.L. and Lough, J. (1990) J. Biol. Chem. 265, 7507-7512.

[15] Dawson, B.A., Herman, T., Haas, A.L. and Lough, J. (1991) J. Cell. Biochem. 46, 166-173.

[16] Ericsson, C., Goldknopf, I.L. and Daneholt, B. (1986) Exp. Cell Res. 167, 127-134,

[17] Chen, T.A., Smith, M.M., Le, S., Sternglanz, R. and Allfrey, V.G. (1991) J. Biol. Chem. 266, 6489-6498.

[18] Kleinschmidt, A.M. and Martinson, H.G. (1981) Nucleic Acids Res. 9, 2423-2431.

[19] Bohm, L. and Crane-Robinson, C. (1984) Biosci. Rep. 4, 365- 386.

[20] Lindsey, G.G., Orgeig, S., Thompson, P., Davies, N. and Maeder, D.L. (1991) J. Mol. Biol. 218, 805-813.

[21] Lindsey, G.G. and Thompson, P. (1992) J. Biol. Chem. 267, 14622-14628.

[22] Hunter, A.J. and Carey, P.D. (1985) Anal. Biochem. 150, 394- 402.

[23] Lindsey, G.G., Thompson, P., Pretorius, L., Purves, L.R. and Von Holt, C. (1983) FEBS Lett. 155, 301-305.

[24] von Holt, C., Brandt, W.F., Greyling, H.J., Lindsey, G.G., Retief, J.D., Rodrigues, J.A., Schwager, S. and Sewell, B.T. (1989) Methods Enzymol. 170, 428-492.

[25] Lindsey, G.G. and Thompson, P. (1990) Biochim. Biophys. Acta 1049, 9-14.

[26] Hewick, R.M., Hunkapiller, M.W., Hood., L.E. and Dreyer, WJ. (1981) J. Biol. Chem. 256, 7990-7997.

[27] Brandt, W.F., Alk, H., Chauhan, M. and Von Holt, C. (1984) FEBS Lett. 174, 228-232.

[28] Klapper, D.G. (1982) in Methods in Protein Sequence Analysis (Elzinga, M., ed.), pp. 509-515, Humana Press, New Jersey.

[29] Lindsey, G.G. and Thompson, P. (1989) Biochim. Biophys. Acta 1009, 257-263.

[30] Noll, M. (1974) Nucleic Acids Res. 1, 1573-1578. [31] Lutter, L.C. (1978) J. Mol. Biol. 124, 391-420. [32] Richmond, T.J., Finch, J.T., Rushton, B., Rhodes, D. and Klug,

A. (1984) Nature 311,532-537. [33] Manning, G.S., Ebralidse, K.K., Mirzabekov, A.D. and Rich, A.

(1989) J. Biomol. Struc. Dynamics 6, 877-889. [34] Ebralidse, K.K. and Mirzabekov, A.D. (1988) Nature 331, 365-

367. [35] Vijay-Kumar, S., Bugg, C.E. and Cook, W.J. (1987) J. Mol. Biol.

194, 531-544. [36] Weber, P., Brown, S. and Mueller, L. (1987) Biochemistry 26,

7282-7890. [37] Goldknopf, I.L., Rosenbaum, F., Sterner, R., Vidali, G., Allfrey,

V.G. and Busch, H. (1979) Biochem. Biophys. Res. Commun. 90, 269-277.

[38] Lindsey, G.G., Thompson, P. and Von Holt, C. (1985) FEBS Lett. 192, 230-234.

histone h2b (and h2a) ubiquitination allows normal histone octamer and core particle reconstitution

Documents