J. Mol. Rid. (1991) 219, 391-392
Can One Measure the Free Energy of Binding of the Histone Octamer to Different DNA Sequences by Salt-dependent Reconstitution? Horace
R. Drew
CSIRO Division of Biomolecular Engineering P.O. Box 184, North Ryde. LVSW %I 13, Australia (Received 18 December 1990; accepted 13 February
1991)
1 explain why many recently reported measurements for t.he “free energy” of positioning of the histone octamer on different DNA sequencesare likely to be in error: i.e. because histone octamers do not equilibrate between different DlVA molecules at’ low salt, but only at high salt. Thus, the report,ed “free energies” refer to an equilibrium at high salt, under nearly dissociating conditions between protein and DNA, and they are likely to be much too small on an absolute scale. There are many other lines of evidence to suggest that the preferences of the histone octamer for different DNA sequencesare rather strong and of importance in biological systems. Keywords: histone octamer; nucleosome: sequence-dependent positioning: chromatin reconstitution; exchange of histone octamers
Recently. several workers (Jayasena & Behe. 1989: Shrader 8 (‘rothers. 1989, 1990) have used for another purpose a method that 1 developed for the r~onst itution of small amounts of radioactive DNA wifh the histone octatner: i.e. to measure “free ~~nrrgirs” for binding of t’he hist,one octamer to different I)!iA sequences. Tn brief: a, concentrated mixture of (I) histone octamer, (2) mixed-sequence chic*k~~n(*ore I)NA in excess, and (3) a small amount of radioactive> DNA of defined sequence. is adjusted to 0.i to 1.0 wK\‘;~(I: then the mixture is diluted slowl,~~ to 04k5 to WlO M-XaCl (Drew ~5 Travers. l!M). .\ portion of the mixture may then be applied to ;II~ ;~,~~rosr ,g,lcbI. in which the radioactivfl l)NA movc5 more slo\vl\~ during elec:trophor,esis if it is I~or~r~tlto 1hr histo&! oc-tamer than if it remains free. SOM. it’ this radioactive DXA binds to the histone oc~tatnf’r more‘ tightly t.han does misetl-sequence (‘orf’ t,S.-I. thrbn more of the radioacti1.r DSA will run t hrongh t ht. gel as a histone-l)NL4 c~)rnplex than iis ;t t’rr,clrnc,tec*u It>: and /-in CWSU.From doing t,his c~s[)~~r~n~ricsof protein-to-1)SA ratios. a “f’&(L pr,cqYr,r?-” of binding of the histoma octamer to rzltli(,;~(~ti\-fs I)S:I of’ any lcbngth or sc~ql~erlc*e can be tl~ic~rrr~inf~l.()1’(~011rs~~. stIc*han experiment does not >,icsltlil /l/ttr/ f’rl;tti\.cxto t hr. mixture of chickcsncore 11X.4 rnolfY~ult~x. il
caomplexesin the living cell? No, they are not; for thr simple and obvious reason that the exchange of histones among different DNA molecules ceases brlo\~ about 0.5 to O-6 M-NaCt (which is why the high salt is necessary in the first place. to achieve r.t~c,on~titution). Thus. thu che~mid equ.ilibrium, used to mcn.Wrr O-.i .)I-,VaC’l
thr ’ yrfree or higher,
energy”
refcvs
to
n
state
qf
which is not relevant to DEA
in a living cell. LZ’r ran expect two &ds of error to result from tnaking such measurements in W.5IV-XaCl. rather t ban in 0.1 M-KaCl. First, differences of free energy among different DNA sequences in their affinities toward the histone octamer might be much less at high salt t ban at low salt. becausr the l)NA is not \vr~pped so firmly about the protein at high salt as coomparedt,o tow salt. Indeed, Shrader &r (‘r-others (1990) see variations in free energy among different DNA seyuencars of only about) 1 kcal/mot (1 cal = 4.1% J). when studying the chemical equilibrium at 0.~ hf-?Ja(‘l. Such small differencaeshave encouraged biochemists to think that such preferences are not significant. Yet when the reconstituted structures are diluted to 0.1 n-NaCl, and probed wit’h nucleases (as in Fig. 2(a) of Hhrader & (‘rothers, 1990), one observes very sharp patterns of nuclrase digestion that require the DK.4 to be bound firmly, as a function of its base sequence, about t,he histone octamrr. So it seemsvery likely that t.hradifferences in free energy of binding to t’he oc+arner among
different DNA sequences are much larger than 1 kral/mol, but that t’hese true energies have not ?;t% been measured accurately. A second kind of error is perhaps more subtle: cnan all of the interactions between the protein and the DNA at 0.1 rv-Ka’aCl be maintained on going to 0.5 M-NaCI? It seems likely that only the strong int’eractions will persist to 0.5 M-NaCI. while t.he weak ones will be lost. Indeed. Shrader & C’rothers (1990) cannot detect, by their study of chemical equilibria in 0.5 M-KaCI. one of the most striking details of the nucleosome core structure: the change in 1)NA sequences preferences (Satchwell rt al., 1986) and helix twist (Hayes et al., 1990) around t)he centre or dyad of the histone octamrr. I believe that such preferences arise from the binding of t.he amino-termini of histones H3 and H4 to A+T-rich minor grooves on the outer parts of the I)KA in these locations. It is no wonder that such weak int,eractions are lost at high salt. In summary. although the method I described in 1985 for efficient reconstitution of small amount,s of I)NA with the histone octamer is st,ill useful for making such material, it should not be used t.o measure “free energies” of binding of the o&amer smong different’ DNA sequences. Such “free energies” refer to an equilibrium at high salt. under near-dissociating conditions. They may give a rough indication of sequence-dependent binding of I)NA toward the histone octamer. hut they arc unsuitablr for precise work on an absolute scale. In an early study. Ramsay r! nl. (1984) used 0.X *~-KaCl to trans:fer the h&one oct,amer from nucleosome cores to small amounts of radioactive IINA. Their work was based on the very early work
M&d
of (Germond rt nl. (lYi6). octamers would rxchangr mol~c~ules at 0.8 >vXa(‘l.
\vhc) shonrd t,hat histonta between diffrrrnt 1)N.A
References Drew. H. It. & Travrrs. A. A (IO%). I)SA bending and its relation to nuf+fxom~ positioning. J. Mol. Wiol. 186. 773-780. (iermond, .I. E.. Bellard, 11.. Oudet. I’. Hr (‘hamboll. I’. (1976). Stability of nucleosomrs in natural and rt*f.onstituted c-hromatins. Sucl. Acids Krs. 3, 3173 Z192. Hayes. ,I.. Tullius, T. I). &, \Volffe. ;\. I’. (I!j!N). The struf-ture of I)XA in a nuclrosome. I’,oc. Snf. .Icctd. Sri.. /‘.S.A 87. 7406 -7409. Jayasena. S. I). & Rrhe. M. -1, (1989). (‘ompetitivc~ nllvleoof f)(~l~drox~~~u~It~otidrs some rt~fwnstitution containing oligoguanosine tracts .I. Mol. Hid. 208. 297 ~306. I’iiia. I).. Harrttino. I).. Truss. $1. & Beato. JI. (lY90). Structural features of’ a regulatory nu~~l~osom~~. J. Mol.
f?iot.
216.
975-990.
Ramsay, Ii.. Frlsenfeld, G., Rusht,on, B, & Mc(:hre. .J. I). (1984). A 145 bp DK;A sequence that, positions itself precisely and asymmetrically on the nuc~leosomr (WY. EMNO J. 3, 2605-2611. Satchwrll. S. C’.. Drew. H. R. R: Travrrs. A. A. (l!M). Sequence periodicitirs in cahickrn nuc*leosome (‘ore I)X;A. .I. Mol. Hiol. 191. (i59-675. Schrader. T. E. N: (“rothers, 1). M. (19X9). Artitic-ial nu(.leosome positioning sequences. I’roc. .\‘nt. .-Icud. Sci.. 1:.s.rl. 86. 741%7422. Schrader. T. E. bt (‘rot,hrrs. I). 51. (1990). Effects of I)?iA seyuenc*r and histone-histone interac.tion on nuc~lrosome placement. ./. Mol. Hiol. 216. (iR- X4. Straka. (‘. B Horz. 11’. (1991). A functional role fol nucleosomes in the repression of a yeast promoter. EMRO J. 10. 361-368.
by A. Klug
Note added in proof. Recent experiments by I’ifia et al. (1990) and Straka & Horz (1991) show that the sequence-dependent binding of DISU’Ato the histone octamer is comparable in energy to its binding of transcription factors. Thus, hi&one-DNA int’eractions within a promoter would seem to btx of sufficient energy to regulate genes.
Can One Measure the Free Energy of Binding of the Histone Octamer to Different DNA Sequences by Salt-dependent Reconstitution? Horace
R. Drew
CSIRO Division of Biomolecular Engineering P.O. Box 184, North Ryde. LVSW %I 13, Australia (Received 18 December 1990; accepted 13 February
1991)
1 explain why many recently reported measurements for t.he “free energy” of positioning of the histone octamer on different DNA sequencesare likely to be in error: i.e. because histone octamers do not equilibrate between different DlVA molecules at’ low salt, but only at high salt. Thus, the report,ed “free energies” refer to an equilibrium at high salt, under nearly dissociating conditions between protein and DNA, and they are likely to be much too small on an absolute scale. There are many other lines of evidence to suggest that the preferences of the histone octamer for different DNA sequencesare rather strong and of importance in biological systems. Keywords: histone octamer; nucleosome: sequence-dependent positioning: chromatin reconstitution; exchange of histone octamers
Recently. several workers (Jayasena & Behe. 1989: Shrader 8 (‘rothers. 1989, 1990) have used for another purpose a method that 1 developed for the r~onst itution of small amounts of radioactive DNA wifh the histone octatner: i.e. to measure “free ~~nrrgirs” for binding of t’he hist,one octamer to different I)!iA sequences. Tn brief: a, concentrated mixture of (I) histone octamer, (2) mixed-sequence chic*k~~n(*ore I)NA in excess, and (3) a small amount of radioactive> DNA of defined sequence. is adjusted to 0.i to 1.0 wK\‘;~(I: then the mixture is diluted slowl,~~ to 04k5 to WlO M-XaCl (Drew ~5 Travers. l!M). .\ portion of the mixture may then be applied to ;II~ ;~,~~rosr ,g,lcbI. in which the radioactivfl l)NA movc5 more slo\vl\~ during elec:trophor,esis if it is I~or~r~tlto 1hr histo&! oc-tamer than if it remains free. SOM. it’ this radioactive DXA binds to the histone oc~tatnf’r more‘ tightly t.han does misetl-sequence (‘orf’ t,S.-I. thrbn more of the radioacti1.r DSA will run t hrongh t ht. gel as a histone-l)NL4 c~)rnplex than iis ;t t’rr,clrnc,tec*u It>: and /-in CWSU.From doing t,his c~s[)~~r~n~ricsof protein-to-1)SA ratios. a “f’&(L pr,cqYr,r?-” of binding of the histoma octamer to rzltli(,;~(~ti\-fs I)S:I of’ any lcbngth or sc~ql~erlc*e can be tl~ic~rrr~inf~l.()1’(~011rs~~. stIc*han experiment does not >,icsltlil /l/ttr/ f’rl;tti\.cxto t hr. mixture of chickcsncore 11X.4 rnolfY~ult~x. il
caomplexesin the living cell? No, they are not; for thr simple and obvious reason that the exchange of histones among different DNA molecules ceases brlo\~ about 0.5 to O-6 M-NaCt (which is why the high salt is necessary in the first place. to achieve r.t~c,on~titution). Thus. thu che~mid equ.ilibrium, used to mcn.Wrr O-.i .)I-,VaC’l
thr ’ yrfree or higher,
energy”
refcvs
to
n
state
qf
which is not relevant to DEA
in a living cell. LZ’r ran expect two &ds of error to result from tnaking such measurements in W.5IV-XaCl. rather t ban in 0.1 M-KaCl. First, differences of free energy among different DNA sequences in their affinities toward the histone octamer might be much less at high salt t ban at low salt. becausr the l)NA is not \vr~pped so firmly about the protein at high salt as coomparedt,o tow salt. Indeed, Shrader &r (‘r-others (1990) see variations in free energy among different DNA seyuencars of only about) 1 kcal/mot (1 cal = 4.1% J). when studying the chemical equilibrium at 0.~ hf-?Ja(‘l. Such small differencaeshave encouraged biochemists to think that such preferences are not significant. Yet when the reconstituted structures are diluted to 0.1 n-NaCl, and probed wit’h nucleases (as in Fig. 2(a) of Hhrader & (‘rothers, 1990), one observes very sharp patterns of nuclrase digestion that require the DK.4 to be bound firmly, as a function of its base sequence, about t,he histone octamrr. So it seemsvery likely that t.hradifferences in free energy of binding to t’he oc+arner among
different DNA sequences are much larger than 1 kral/mol, but that t’hese true energies have not ?;t% been measured accurately. A second kind of error is perhaps more subtle: cnan all of the interactions between the protein and the DNA at 0.1 rv-Ka’aCl be maintained on going to 0.5 M-NaCI? It seems likely that only the strong int’eractions will persist to 0.5 M-NaCI. while t.he weak ones will be lost. Indeed. Shrader & C’rothers (1990) cannot detect, by their study of chemical equilibria in 0.5 M-KaCI. one of the most striking details of the nucleosome core structure: the change in 1)NA sequences preferences (Satchwell rt al., 1986) and helix twist (Hayes et al., 1990) around t)he centre or dyad of the histone octamrr. I believe that such preferences arise from the binding of t.he amino-termini of histones H3 and H4 to A+T-rich minor grooves on the outer parts of the I)KA in these locations. It is no wonder that such weak int,eractions are lost at high salt. In summary. although the method I described in 1985 for efficient reconstitution of small amount,s of I)NA with the histone octamer is st,ill useful for making such material, it should not be used t.o measure “free energies” of binding of the o&amer smong different’ DNA sequences. Such “free energies” refer to an equilibrium at high salt. under near-dissociating conditions. They may give a rough indication of sequence-dependent binding of I)NA toward the histone octamer. hut they arc unsuitablr for precise work on an absolute scale. In an early study. Ramsay r! nl. (1984) used 0.X *~-KaCl to trans:fer the h&one oct,amer from nucleosome cores to small amounts of radioactive IINA. Their work was based on the very early work
M&d
of (Germond rt nl. (lYi6). octamers would rxchangr mol~c~ules at 0.8 >vXa(‘l.
\vhc) shonrd t,hat histonta between diffrrrnt 1)N.A
References Drew. H. It. & Travrrs. A. A (IO%). I)SA bending and its relation to nuf+fxom~ positioning. J. Mol. Wiol. 186. 773-780. (iermond, .I. E.. Bellard, 11.. Oudet. I’. Hr (‘hamboll. I’. (1976). Stability of nucleosomrs in natural and rt*f.onstituted c-hromatins. Sucl. Acids Krs. 3, 3173 Z192. Hayes. ,I.. Tullius, T. I). &, \Volffe. ;\. I’. (I!j!N). The struf-ture of I)XA in a nuclrosome. I’,oc. Snf. .Icctd. Sri.. /‘.S.A 87. 7406 -7409. Jayasena. S. I). & Rrhe. M. -1, (1989). (‘ompetitivc~ nllvleoof f)(~l~drox~~~u~It~otidrs some rt~fwnstitution containing oligoguanosine tracts .I. Mol. Hid. 208. 297 ~306. I’iiia. I).. Harrttino. I).. Truss. $1. & Beato. JI. (lY90). Structural features of’ a regulatory nu~~l~osom~~. J. Mol.
f?iot.
216.
975-990.
Ramsay, Ii.. Frlsenfeld, G., Rusht,on, B, & Mc(:hre. .J. I). (1984). A 145 bp DK;A sequence that, positions itself precisely and asymmetrically on the nuc~leosomr (WY. EMNO J. 3, 2605-2611. Satchwrll. S. C’.. Drew. H. R. R: Travrrs. A. A. (l!M). Sequence periodicitirs in cahickrn nuc*leosome (‘ore I)X;A. .I. Mol. Hiol. 191. (i59-675. Schrader. T. E. N: (“rothers, 1). M. (19X9). Artitic-ial nu(.leosome positioning sequences. I’roc. .\‘nt. .-Icud. Sci.. 1:.s.rl. 86. 741%7422. Schrader. T. E. bt (‘rot,hrrs. I). 51. (1990). Effects of I)?iA seyuenc*r and histone-histone interac.tion on nuc~lrosome placement. ./. Mol. Hiol. 216. (iR- X4. Straka. (‘. B Horz. 11’. (1991). A functional role fol nucleosomes in the repression of a yeast promoter. EMRO J. 10. 361-368.
by A. Klug
Note added in proof. Recent experiments by I’ifia et al. (1990) and Straka & Horz (1991) show that the sequence-dependent binding of DISU’Ato the histone octamer is comparable in energy to its binding of transcription factors. Thus, hi&one-DNA int’eractions within a promoter would seem to btx of sufficient energy to regulate genes.
