Vol.
179,
September
No.
2, 1991
16,
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
1991
Pages
Neutron
810-816
Scattering Studies of Chromatosomes
Stan Lambert*, SergeMuyldermans, John Baldwin*, JosephineKilner*, Konrad Ibel”, and Lode Wijns *Liverpool Polytechnic, Byrom Street, Liverpool L3 3AF, U.K. “Institut Laue Langevin, F-38042 Grenoble, France Vrije Universiteit Brussel,Instituut voor Molekulaire Biologie, Paardenstraat65, 1640 Sint GenesiusRode, Belgium Received
June
27,
1991
Neutron scattering data establish that the radius of gyration of the DNA in chicken erythrocyte chromatosomeparticles is significantly higher, by about 0.3 nm, than the radius of gyration of the DNA in the core particle. Correspondinginformation of the radius of gyration of the protein component in the chromatosomes(3.75 nm) indicated an enlargement, comparedto the radius of gyration of the octamer of histone proteins both in core particles and in the histone octamer stabilisedin 2 M NaCl (3.25 nm). From the latter data, we could calculate the distancein the chromatosomebetween the centre of massof the linker histone and the histone octamer as 5.5 nm. These results impose severe limitations for the organisation of the 22 bp extra DNA and the possible location of Hl/H5 in the chromatosome,implying that the Hl/H5 is close to the centre turn of the core particle DNA. 0 1991
Academic
Press,
Inc.
The DNA in eukaryotic chromatin winds around an octamer core composedof two copieseach of histonesH2A, H2B, H3 and H4. The octamer is able to protect 146basepairs of the DNA againstattack by micrococcal nucleaseto form the nucleosomecore particle (1). Micrococcal nucleasedigestion of chromatin containing a fifth histone, Hl (H5 in avian erythrocytes), hasa pausein digestiondue to protection of 168basepairs (bp) of DNA, and the protected particle is called the chromatosome. This particle consists of the nucleosomecore particle plus 22 bp of DNA and the histone Hl (1,2). Chromatosomesare strung together in the chromatin of cell nuclei by between 0 and 70 bp of linker DNA, the size of linker being speciesand cell-type specific. Two very important questionsregarding the structure of the chromatosomeare: 1) What is the position of histoneHl (or H5) in the chromatosomeand 2) What is the location and the conformation of the extra 22 bp of DNA? In this paper we useneutron scatteringto attempt to answerthesequestions. MATERIALS
AND METHODS
Prenarationand characterisationof chromatosomes. The preparation of chicken erytbrocyte chromatosomeswas essentially as described by Simpson (2). The amounts of Hl/H5 relative to the core histonesin our chromatosome 0006-291x/91 $1.50 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
810
Vol.
179,
No.
2, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
preparations were essentially identical to the amounts in the nuclei, as deduced from Coommassie staining of the proteins separated by SDS gel electrophoresis. The size heterogeneity of the DNA as analysed under denaturing conditions, indicated a very small quantity of 146 nucleotides, and no material in the size range of 178, most of the material migrating at a length of 168 nucleotides. DNase I digestions on end-labelled chromatosomes indicated a symmetrical location of the DNA on the histone octamer (2),(see below). Small-angle neutron scattering. Neutron scattering from particles, containing regions of different neutron scatteringlength densities (e.g. DNA and protein in this case), can provide valuable information on structure. Solutions of the particles containing different ratios of H20 to D20 are irradiated by a cold neutron beam. The scattered neutrons are detected as a function of scattering angle 2 8 to provide a neutron scattering intensity spectrum, I(Q), where Q = 4 II sin 0/h. and h is the wavelength of the neutron. Spectra at different percentages of D20 in the solvent are taken because the coherent elastic neutron scattering length density from D20 is strongly positive whereas that from H20 is slightly negative. The mean scattering density for DNA is matched by a scatteringlength density of 63% D20, whereas the scattering density for the histone protein is matched Thus in 63% D20 the neutron scatter spectrum contains structural by 40% D20. information on the protein only, whereas in 40% D20 the spectrum contains information on the DNA only, since one component is contrast matched when its scattering density equals that of the solvent (internal heterogeneities can be neglected in our case). The analysis of the neutron spectra can be done, in terms of Guinier plots of the LnI(Q) versus Q2. The slope of the plot at small Q gives Rg2/3 where Rg is the radius of gyration. The intercept of the extrapollated intensities with the ordinate axis gives I(Q=O) values which, in absolute units, allows the determination of the molecular weight of the particle, and each of its components. The m) varies linearly with solvent D20 percenta e and the D20 percentage for zero contrast (i.e. contrast matching) occurs when Jr--+ I(Q=O) is zero. The slope of the w) versus scattering density of solvent is a check on the molecular volume. The R data as a function of solvent scattering density contrast (relative to the contrast matche fi density) allows a refinement processs (3) from which Rg values may be calculated under 1) contrast matched conditions for the protein, 2) contrast matched conditions for the DNA, and 3) under infinite contrast where the Rg value corresponds to the shape of the particle. RESULTS l.Zero-angle Scattering. The measured neutron scattered intensities I(Q) for chromatosomes after correction for the scatter contribution of the sample container, were divided by the scattered intensities from a standard (in our case, incoherent scattered intensities from H20). LnI(Q) versus 42 (the Guinier plots) were straight lines at low values of Q. Extrapolated to I(0) values, when put on an absolute scale are given by -2 p .V 2c where p is the contrast scattering density of the particle. Vc is the effective volume of the particle and is given by (VF - VE) where VF is the geometrical volume within the Van der Waals envelopes of the particle and VE is a correction term due to D H exchange. A graph of m) versus solvent scatteringlength density, psol is a straight line of slope proportional to the sqaure root of the particle molecular weight, and intercept Pcm on the psol axis. Figure 1 is such a plot in this study of chromatosomes. From figure 1 the molecular weight for the chromatosome corresponded to 227,000, if a VE value of 0.15 VF is assumed (in comparison with the value used in earlier nucleosome core particle calculations (4)), together with a partial specific volume of 0.65 ml/g and a ratio of particle molecular weight to DNA molecular weight of 2.15. This ties in quite well with the value expected for a chromatosome of 242,250 in view of the systematic errors and statistical deviations of the data and its reduction to an absolute scale. 811
Vol.
179, No. 2, 1991
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Chromotosomes
Figure 1. The square root of the zero angle neutron scattered intensities versus solvent
scattering-length densityfor chromatosomes. The molecularweight of the particleis given by theproductof the squareof the gradientof the graphand
whereTw (= 0.4576)is the neutrontransmission of 1 mmtickuessof H20 at the wavelength of the experiment,NA is Avogadro’snumber(6.022x 1026perKgm mole),d (= 0.001m) is the samplecell ticlmess,VR is the fraction of exchangeable protonsin the particle,(0.15 assumed), 5 is thepartialspecificvolumeof thechromatosome (6.5x 104 m3/Kg),andRAT (= 2.18assumed) is theratio of total particlemolecularweightto theDNA molecularweight.
2.Guinier Dlots. At low Q, graphs of Ln(I(Q,@) versus Q2 for spectra at a given contrast fi (called Guinier plots) had a gradient R&5)/3, where Rg@) is the radius of gyration of the particle about a point and is given by Rg2 = (Zbiri2)/(Xbi) where bi values at radius ri are the excessneutron scatter lengths of the nuclei of the atoms which constitute the particle. The Guinier plots for different contrasts are given for the chromatosomeparticles in figure 2. A full analysisof Rg(p) data involves a ‘Stuhrmann’ plot of (Rg(p^))2versus (l/p) (3) and this graph for chromatosomeparticles is given in figure 3. If we take the data from figure 3 and compareit with the nucleosomecore particle data of Suau et a1.(4),we conclude: First, that the radius of gyration of the nrotein in the chromatosomeparticle is 3.75 nm, significantly larger (as expected) than the correspondingvalue for the core particle of 3.25 nm (4,5,6). Second, the radius of gyration of the shwe of the chromatosomeparticle is 4.44 nm, significantly larger than the correspondingcore particle value of 4.05 nm. Third, the radius of gyration of the DNA in the chromatosomeparticle is 5.09 +/- 0.2 nm, significantly larger than the correspondingvalue for the core particle of approximately 4.7 +I- 0.2 nm. 3.Distance betweenthe centre of Hl and of the histoneoctamer. Wood et al. (6) showed that the isolated histone octamer from chick erythrocytes stabilised in 2 M NaCl has a radius of gyration of 3.25 nm and that, at 63% D20 the nucleosomecore particle in a 10 mM ionic strength buffer had the sameradius of gyration. Thus the DNA is very effectively contrast matched at 63% D20 so that the true protein radii 812
Vol.
179,
No.
BIOCHEMICAL
2, 1991
Chmotosomes
AND
BIOPHYSICAL
(H.0)
RESEARCH
Ckornatosmes
COMMUNICATIONS
(637,
0.0)
-10
1
chromotosoms
(30%
4
= 3.725
Chromotosomes
D,O)
(80X
+/-
0.836
m-n
D,O)
-0.5 Rg = 4.831
1
Chromotoscmes
(40%
+/-
0.41
rm
D,O)
Chromotosomes
(D,O)
-10,
--
I 0.0
d.1
0;
oh
d.4
0’ Ini’
Figure 2. Guinier plots for chromatosomes over a range of contrasts in different percentage D20 buffers.
of gyration can be measured at that D201H20 ratio in the buffer. We measured a radius of gyration of the protein in the chromatosome about its centre of gravity of 3.75 +/- 0.8 nm. The globular part of histone Hl has a molecular weight of ca 8000 and, if it formed a tight packed sphere of partial specific volume of 0.73 ml/g, it would have a radius of 1.27 nm 813
Vol.
179, No. 2, 1991
BIOCHEMICAL
AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Chromotosomes
Figure 3. Squareof radiusof gyration versus the reciprocalof contrast(the so-called Stuhrmannplot (3)). R DNA at 40% D20 = 5.09 nm, Rg,proteinat 63% D20 = ) = 3.76 nm andRg,shape (l/P-3= ) = 4.44nm.
and a radius of gyration of 0.98 nm. An extreme situation of an elongatedmolecule (of the samemolecular weight) would be e.g. an ellipsoid with a correspondingradius of gyration 1.3 nm. The correspondingcalculation for the complete histone Hl including the basic tail region would give a radius of 1.4 nm, if it were sphericaland around 1.8 nm if it were quite elongated. These radii of gyration are about the gravity centres of scattering density of the particles. Simple calculations of moments of scattering-length density, (knowing that the molecular weight of the octamer is about four timesthe molecular weight of Hl) show that, if y is the distancebetween the centre of the octamer to the centre of the chromatosome,then 4y is the distancebetween the chromatosomecentre and the centre of Hl. The calulation neededthe following formulae: Rg2(H1 + octamer) =
and (xb),mer
= 4(xb)Hl, where Rg values are about the centre of scattering length
density of the chromatosome. Therefore, using the parallel axis theorem Rgg,chromatosome= 4/5(Rgg20ctamer+ Y2) + 1/5(Rgg2H1 + 16~2)~ where Rgg values are about the centre of scattering density of the particles themselves. Using 3.75 nm for Rgg,chromatosome,3.25 nm for Rgg,octamerand 1.4 nm for Rgg,Hl (Hl spherical), the value of the distance between the centre of the Hl and the centre of the octamer in the chromatosomeis 5.7 nm. If we took an elongatedHl (Rgg,Hl = 1.8 nm), the distancebecomes5.5 nm (Figure 4a). Taking errors in the measurementof the radius of gyration of the chromatosomeprotein we may say that the distanceis between5.1 and 5.9 nm. An exactly similar calculation for the DNA in the chromatosomeand the DNA in the core particle componentof the chromatosome,knowing the approximate shapeand molecular weight of the 22 bp extra in the chromatosome,showsthat the observed 0.3 nm increaseof the Rg of the chromatosomeDNA compared with the Rg for core particle DNA could not arise with the 22 bp as 11 bp extensions either side of the core particle DNA bend as a 814
Vol.
179,
No.
2, 7991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Figure 4.Models of chromatosome structure. (a) The inner and outer radius of the DNA are given as well as the radius of the superhelical axis of the DNA. The distance between the centre of the octamer (little open square) and the centre of Hl (little black square) was calculated to be 5.5 nm. In model (b) and (c) we show the possible locations of the 2 x 11 bp extra DNA to account for the measured increase in Rg,~~~ between the core particle and the chromatosome. Only model c is in acordance simultaneously with neutron scatter and DNA digestion data.
smooth continuation of the core particle DNA curvature (figure 4a). However, the calculations show that the observed increase in DNA radius of gyration could arise from the models in figure 4b, where the 22 bp is one side of the chromatosome with Hl presumably located on one side only of the centre turn of the core particle DNA, or figure 4c, where the DNA straighten out, away from the core particle component. DISCUSSION To summarize, we have established that the radius of gyration of DNA on the chromatosome, to within quite broad limits of error, is about 5.0 nm. It is significantly higher, by about 0.3 nm, than the radius of gyration of the DNA in the core particle. Corresponding information from the X-ray results show that the chromatosome has a radius of gyration of 4.8 nm compared with the core particle radius of gyration of 4.6 nm and the radius of gyration of the Hl/HS-depleted chromatosome of 4.6nm (J. Kilner, S. Muyldermans, J. Baldwin & L. Wijns unpublished results). Although the X-ray results measure scattering from protein and DNA, the contrast of the DNA is much higher than that of the protein for solutions in buffer containing no small molecules to alter the solvent X-ray scattering-length density. Here again the X-ray results would indicate a significant increase in radius of gyration of the DNA in the chromatosome. From our measurements of the radius of gyration of the protein in the chromatosomes and the knowledge of the radius of gyration of the protein both in the core particle (4) and the histone octamer (6), we could calculate, to a first approximation, the distance between the centre of mass of Hl and the octamer as 5.5 nm (Figure 4a). Clearly if the centre of histone Hl is 5.5 nm from the centre of the octamer it cannot lie outside the centre turn of the core particle DNA as it was proposed (7). Based on our data of the radius of gyration of the protein, and DNA, and the distance of HI to the octamer, there are two possibilities to arrange the linker histone and the chromatosomal DNA: 1) the Hl lies on one side of the chromatosome centre turn of DNA only, i.e. the extra 22 base pairs are on one side (see figure 4b). Such a length of DNA could extend outwards enough to explain the observed increase in the radius of gyration of the DNA. However, this asymmetric model for the DNA should be excluded as it is in clear contradiction with the DNase I digestion mapping on end-labelled chromatosomes (2,12), proving a symmetrical location of the octamer on the DNA. 815
Vol.
179, No. 2, 1991
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
2) the centre turn of DNA is slightly inflected in towards the core histonesat the position of interaction with Hl, and part of Hl penetrates into the grooves of the DNA (Figure 4~). Indeed in Richmond’s model from the X-ray crystallography of the core particle there is an inflection of the DNA at the central region (8), both sidewaysand inwards; this may provide the space for the globular part of Hl. Moreover, from theoretical work (10,ll) it was inferred that the globular part of Hl might additionally penetratethe DNA grooves. Hence, this model is compatible with the 5.5 nm radius. The original experimentsof Allan et al. (9) showedprotection of chromatosomeDNA to micrococcal nucleaseattack by the globular part of histone H5 complexed onto chick chromatin depleted of histoneH5 and Hl . In trying to build a model of a chromatosomewith 11 base pairs either side of the core particle, and touching the globular part of Hl, it is difficult to account for the 0.3 nm increasein radius of gyration of the chromatosomeDNA comparedwith that of the core particle. All modelsin which the 2 x 11 bp extra are coiled as a smooth continuation of the core particle DNA (figure 4a) are certainly incompatible. The model of Richmond et al. (8), with 1.8 turns of DNA on the core particle, can only fit if we assumethat the DNA ends of the core particle component of DNA also rearrange and straightenout in the chromatosome(Figure 4c). There is evidence (13,14) that the ends of the core particle DNA are bound to the octamer in a different way in the core particle comparedto the chromatosome. If the core particle model of Uberbacherand Bunich (15) is taken with 1.9 superhelical turns of DNA to the core particle and if the DNA is assumedto bend outwards from within the core particle component when H5 is present, then the model can explain the measured DNA radius of gyration and a closing of the two turns to allow H5 to interact acrossthe ovaloid centre turn of DNA. ACKNOWLEDGMENTS This work had financial support of the N.F.W.O. and the Nationale Loterij, Belgium, from the Science and Engineering Research Council, U.K., and from the Daresbury Laboratory Synchrotron Radiation Source, Warington, U.K. REFERENCES 1. i4: 5. t* 8: 9. 10. 11. 12. 13. 14. 15.
Igo-KemenesT., W. H&z, & H. Zachau (1982) Ann.Rev. B&hem. 5l, 89-121. SimpsonR.T. (1978) Biochemistry fl,5524-5531 Ibel K., 8z H.B. Stuhrmann (1975) J.Mol.Biol. 93,255-265 Suau P., G.G. Kneale, G.W. Braddock, J.P. Baldwin, E.M. Bradbury (1977) Nucl.Acids Res.&3769-3786 Hjelm R.P., G.G. Kneale, P. Suau, J.P. Baldwin, E.M. Bradbury, K. Ibel (1977) Cell 1Q, 139-151 Wood et al. J.Biol.Chem. (1991) =,5696-5702. Crane-RobinsonC. & O.B.Ptitsyn (1988) Prot.Eng. 2,577-582 Richmond T.J., J.T. Finch, B. Rushton, D. Rhodes, A. Klug, 5-(1984) Nature 311, 532-537 Allan J.D., P.G. Hartman, C. Crane-Robinson,F.X. Aviles (1980) Nature 288, 675679 Tumell W.G., S.C. Satchwell, A.A.Travers (1988), FebsLetters =,263-268 SegersA., L. Wijns, I. Lasters(1991) Biochem.Biophys.Res.Com.174,898-902 Staynov D.Z. & C. Crane-Robinson(1988) EMBO J. 2,3685-3691 Weishet W.O., K.Tatchell, K.E. Van Holde, H. Klump (1978) Nucl.Acids Res. 5, 139-160 Belyavsky A.V., S.G.Bavykin, E.G. Goguadze, A.D. Mirzabekov (1980) J.Mol.Biol. m,519-536 UberbacherE.C. & G.J. Bunich (1985) J.Biomol.Snucture & Dyn. 2,1033-1055. 816
179,
September
No.
2, 1991
16,
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
1991
Pages
Neutron
810-816
Scattering Studies of Chromatosomes
Stan Lambert*, SergeMuyldermans, John Baldwin*, JosephineKilner*, Konrad Ibel”, and Lode Wijns *Liverpool Polytechnic, Byrom Street, Liverpool L3 3AF, U.K. “Institut Laue Langevin, F-38042 Grenoble, France Vrije Universiteit Brussel,Instituut voor Molekulaire Biologie, Paardenstraat65, 1640 Sint GenesiusRode, Belgium Received
June
27,
1991
Neutron scattering data establish that the radius of gyration of the DNA in chicken erythrocyte chromatosomeparticles is significantly higher, by about 0.3 nm, than the radius of gyration of the DNA in the core particle. Correspondinginformation of the radius of gyration of the protein component in the chromatosomes(3.75 nm) indicated an enlargement, comparedto the radius of gyration of the octamer of histone proteins both in core particles and in the histone octamer stabilisedin 2 M NaCl (3.25 nm). From the latter data, we could calculate the distancein the chromatosomebetween the centre of massof the linker histone and the histone octamer as 5.5 nm. These results impose severe limitations for the organisation of the 22 bp extra DNA and the possible location of Hl/H5 in the chromatosome,implying that the Hl/H5 is close to the centre turn of the core particle DNA. 0 1991
Academic
Press,
Inc.
The DNA in eukaryotic chromatin winds around an octamer core composedof two copieseach of histonesH2A, H2B, H3 and H4. The octamer is able to protect 146basepairs of the DNA againstattack by micrococcal nucleaseto form the nucleosomecore particle (1). Micrococcal nucleasedigestion of chromatin containing a fifth histone, Hl (H5 in avian erythrocytes), hasa pausein digestiondue to protection of 168basepairs (bp) of DNA, and the protected particle is called the chromatosome. This particle consists of the nucleosomecore particle plus 22 bp of DNA and the histone Hl (1,2). Chromatosomesare strung together in the chromatin of cell nuclei by between 0 and 70 bp of linker DNA, the size of linker being speciesand cell-type specific. Two very important questionsregarding the structure of the chromatosomeare: 1) What is the position of histoneHl (or H5) in the chromatosomeand 2) What is the location and the conformation of the extra 22 bp of DNA? In this paper we useneutron scatteringto attempt to answerthesequestions. MATERIALS
AND METHODS
Prenarationand characterisationof chromatosomes. The preparation of chicken erytbrocyte chromatosomeswas essentially as described by Simpson (2). The amounts of Hl/H5 relative to the core histonesin our chromatosome 0006-291x/91 $1.50 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
810
Vol.
179,
No.
2, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
preparations were essentially identical to the amounts in the nuclei, as deduced from Coommassie staining of the proteins separated by SDS gel electrophoresis. The size heterogeneity of the DNA as analysed under denaturing conditions, indicated a very small quantity of 146 nucleotides, and no material in the size range of 178, most of the material migrating at a length of 168 nucleotides. DNase I digestions on end-labelled chromatosomes indicated a symmetrical location of the DNA on the histone octamer (2),(see below). Small-angle neutron scattering. Neutron scattering from particles, containing regions of different neutron scatteringlength densities (e.g. DNA and protein in this case), can provide valuable information on structure. Solutions of the particles containing different ratios of H20 to D20 are irradiated by a cold neutron beam. The scattered neutrons are detected as a function of scattering angle 2 8 to provide a neutron scattering intensity spectrum, I(Q), where Q = 4 II sin 0/h. and h is the wavelength of the neutron. Spectra at different percentages of D20 in the solvent are taken because the coherent elastic neutron scattering length density from D20 is strongly positive whereas that from H20 is slightly negative. The mean scattering density for DNA is matched by a scatteringlength density of 63% D20, whereas the scattering density for the histone protein is matched Thus in 63% D20 the neutron scatter spectrum contains structural by 40% D20. information on the protein only, whereas in 40% D20 the spectrum contains information on the DNA only, since one component is contrast matched when its scattering density equals that of the solvent (internal heterogeneities can be neglected in our case). The analysis of the neutron spectra can be done, in terms of Guinier plots of the LnI(Q) versus Q2. The slope of the plot at small Q gives Rg2/3 where Rg is the radius of gyration. The intercept of the extrapollated intensities with the ordinate axis gives I(Q=O) values which, in absolute units, allows the determination of the molecular weight of the particle, and each of its components. The m) varies linearly with solvent D20 percenta e and the D20 percentage for zero contrast (i.e. contrast matching) occurs when Jr--+ I(Q=O) is zero. The slope of the w) versus scattering density of solvent is a check on the molecular volume. The R data as a function of solvent scattering density contrast (relative to the contrast matche fi density) allows a refinement processs (3) from which Rg values may be calculated under 1) contrast matched conditions for the protein, 2) contrast matched conditions for the DNA, and 3) under infinite contrast where the Rg value corresponds to the shape of the particle. RESULTS l.Zero-angle Scattering. The measured neutron scattered intensities I(Q) for chromatosomes after correction for the scatter contribution of the sample container, were divided by the scattered intensities from a standard (in our case, incoherent scattered intensities from H20). LnI(Q) versus 42 (the Guinier plots) were straight lines at low values of Q. Extrapolated to I(0) values, when put on an absolute scale are given by -2 p .V 2c where p is the contrast scattering density of the particle. Vc is the effective volume of the particle and is given by (VF - VE) where VF is the geometrical volume within the Van der Waals envelopes of the particle and VE is a correction term due to D H exchange. A graph of m) versus solvent scatteringlength density, psol is a straight line of slope proportional to the sqaure root of the particle molecular weight, and intercept Pcm on the psol axis. Figure 1 is such a plot in this study of chromatosomes. From figure 1 the molecular weight for the chromatosome corresponded to 227,000, if a VE value of 0.15 VF is assumed (in comparison with the value used in earlier nucleosome core particle calculations (4)), together with a partial specific volume of 0.65 ml/g and a ratio of particle molecular weight to DNA molecular weight of 2.15. This ties in quite well with the value expected for a chromatosome of 242,250 in view of the systematic errors and statistical deviations of the data and its reduction to an absolute scale. 811
Vol.
179, No. 2, 1991
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
Chromotosomes
Figure 1. The square root of the zero angle neutron scattered intensities versus solvent
scattering-length densityfor chromatosomes. The molecularweight of the particleis given by theproductof the squareof the gradientof the graphand
whereTw (= 0.4576)is the neutrontransmission of 1 mmtickuessof H20 at the wavelength of the experiment,NA is Avogadro’snumber(6.022x 1026perKgm mole),d (= 0.001m) is the samplecell ticlmess,VR is the fraction of exchangeable protonsin the particle,(0.15 assumed), 5 is thepartialspecificvolumeof thechromatosome (6.5x 104 m3/Kg),andRAT (= 2.18assumed) is theratio of total particlemolecularweightto theDNA molecularweight.
2.Guinier Dlots. At low Q, graphs of Ln(I(Q,@) versus Q2 for spectra at a given contrast fi (called Guinier plots) had a gradient R&5)/3, where Rg@) is the radius of gyration of the particle about a point and is given by Rg2 = (Zbiri2)/(Xbi) where bi values at radius ri are the excessneutron scatter lengths of the nuclei of the atoms which constitute the particle. The Guinier plots for different contrasts are given for the chromatosomeparticles in figure 2. A full analysisof Rg(p) data involves a ‘Stuhrmann’ plot of (Rg(p^))2versus (l/p) (3) and this graph for chromatosomeparticles is given in figure 3. If we take the data from figure 3 and compareit with the nucleosomecore particle data of Suau et a1.(4),we conclude: First, that the radius of gyration of the nrotein in the chromatosomeparticle is 3.75 nm, significantly larger (as expected) than the correspondingvalue for the core particle of 3.25 nm (4,5,6). Second, the radius of gyration of the shwe of the chromatosomeparticle is 4.44 nm, significantly larger than the correspondingcore particle value of 4.05 nm. Third, the radius of gyration of the DNA in the chromatosomeparticle is 5.09 +/- 0.2 nm, significantly larger than the correspondingvalue for the core particle of approximately 4.7 +I- 0.2 nm. 3.Distance betweenthe centre of Hl and of the histoneoctamer. Wood et al. (6) showed that the isolated histone octamer from chick erythrocytes stabilised in 2 M NaCl has a radius of gyration of 3.25 nm and that, at 63% D20 the nucleosomecore particle in a 10 mM ionic strength buffer had the sameradius of gyration. Thus the DNA is very effectively contrast matched at 63% D20 so that the true protein radii 812
Vol.
179,
No.
BIOCHEMICAL
2, 1991
Chmotosomes
AND
BIOPHYSICAL
(H.0)
RESEARCH
Ckornatosmes
COMMUNICATIONS
(637,
0.0)
-10
1
chromotosoms
(30%
4
= 3.725
Chromotosomes
D,O)
(80X
+/-
0.836
m-n
D,O)
-0.5 Rg = 4.831
1
Chromotoscmes
(40%
+/-
0.41
rm
D,O)
Chromotosomes
(D,O)
-10,
--
I 0.0
d.1
0;
oh
d.4
0’ Ini’
Figure 2. Guinier plots for chromatosomes over a range of contrasts in different percentage D20 buffers.
of gyration can be measured at that D201H20 ratio in the buffer. We measured a radius of gyration of the protein in the chromatosome about its centre of gravity of 3.75 +/- 0.8 nm. The globular part of histone Hl has a molecular weight of ca 8000 and, if it formed a tight packed sphere of partial specific volume of 0.73 ml/g, it would have a radius of 1.27 nm 813
Vol.
179, No. 2, 1991
BIOCHEMICAL
AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Chromotosomes
Figure 3. Squareof radiusof gyration versus the reciprocalof contrast(the so-called Stuhrmannplot (3)). R DNA at 40% D20 = 5.09 nm, Rg,proteinat 63% D20 = ) = 3.76 nm andRg,shape (l/P-3= ) = 4.44nm.
and a radius of gyration of 0.98 nm. An extreme situation of an elongatedmolecule (of the samemolecular weight) would be e.g. an ellipsoid with a correspondingradius of gyration 1.3 nm. The correspondingcalculation for the complete histone Hl including the basic tail region would give a radius of 1.4 nm, if it were sphericaland around 1.8 nm if it were quite elongated. These radii of gyration are about the gravity centres of scattering density of the particles. Simple calculations of moments of scattering-length density, (knowing that the molecular weight of the octamer is about four timesthe molecular weight of Hl) show that, if y is the distancebetween the centre of the octamer to the centre of the chromatosome,then 4y is the distancebetween the chromatosomecentre and the centre of Hl. The calulation neededthe following formulae: Rg2(H1 + octamer) =
and (xb),mer
= 4(xb)Hl, where Rg values are about the centre of scattering length
density of the chromatosome. Therefore, using the parallel axis theorem Rgg,chromatosome= 4/5(Rgg20ctamer+ Y2) + 1/5(Rgg2H1 + 16~2)~ where Rgg values are about the centre of scattering density of the particles themselves. Using 3.75 nm for Rgg,chromatosome,3.25 nm for Rgg,octamerand 1.4 nm for Rgg,Hl (Hl spherical), the value of the distance between the centre of the Hl and the centre of the octamer in the chromatosomeis 5.7 nm. If we took an elongatedHl (Rgg,Hl = 1.8 nm), the distancebecomes5.5 nm (Figure 4a). Taking errors in the measurementof the radius of gyration of the chromatosomeprotein we may say that the distanceis between5.1 and 5.9 nm. An exactly similar calculation for the DNA in the chromatosomeand the DNA in the core particle componentof the chromatosome,knowing the approximate shapeand molecular weight of the 22 bp extra in the chromatosome,showsthat the observed 0.3 nm increaseof the Rg of the chromatosomeDNA compared with the Rg for core particle DNA could not arise with the 22 bp as 11 bp extensions either side of the core particle DNA bend as a 814
Vol.
179,
No.
2, 7991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Figure 4.Models of chromatosome structure. (a) The inner and outer radius of the DNA are given as well as the radius of the superhelical axis of the DNA. The distance between the centre of the octamer (little open square) and the centre of Hl (little black square) was calculated to be 5.5 nm. In model (b) and (c) we show the possible locations of the 2 x 11 bp extra DNA to account for the measured increase in Rg,~~~ between the core particle and the chromatosome. Only model c is in acordance simultaneously with neutron scatter and DNA digestion data.
smooth continuation of the core particle DNA curvature (figure 4a). However, the calculations show that the observed increase in DNA radius of gyration could arise from the models in figure 4b, where the 22 bp is one side of the chromatosome with Hl presumably located on one side only of the centre turn of the core particle DNA, or figure 4c, where the DNA straighten out, away from the core particle component. DISCUSSION To summarize, we have established that the radius of gyration of DNA on the chromatosome, to within quite broad limits of error, is about 5.0 nm. It is significantly higher, by about 0.3 nm, than the radius of gyration of the DNA in the core particle. Corresponding information from the X-ray results show that the chromatosome has a radius of gyration of 4.8 nm compared with the core particle radius of gyration of 4.6 nm and the radius of gyration of the Hl/HS-depleted chromatosome of 4.6nm (J. Kilner, S. Muyldermans, J. Baldwin & L. Wijns unpublished results). Although the X-ray results measure scattering from protein and DNA, the contrast of the DNA is much higher than that of the protein for solutions in buffer containing no small molecules to alter the solvent X-ray scattering-length density. Here again the X-ray results would indicate a significant increase in radius of gyration of the DNA in the chromatosome. From our measurements of the radius of gyration of the protein in the chromatosomes and the knowledge of the radius of gyration of the protein both in the core particle (4) and the histone octamer (6), we could calculate, to a first approximation, the distance between the centre of mass of Hl and the octamer as 5.5 nm (Figure 4a). Clearly if the centre of histone Hl is 5.5 nm from the centre of the octamer it cannot lie outside the centre turn of the core particle DNA as it was proposed (7). Based on our data of the radius of gyration of the protein, and DNA, and the distance of HI to the octamer, there are two possibilities to arrange the linker histone and the chromatosomal DNA: 1) the Hl lies on one side of the chromatosome centre turn of DNA only, i.e. the extra 22 base pairs are on one side (see figure 4b). Such a length of DNA could extend outwards enough to explain the observed increase in the radius of gyration of the DNA. However, this asymmetric model for the DNA should be excluded as it is in clear contradiction with the DNase I digestion mapping on end-labelled chromatosomes (2,12), proving a symmetrical location of the octamer on the DNA. 815
Vol.
179, No. 2, 1991
BIOCHEMICAL
AND BIOPHYSICAL
RESEARCH COMMUNICATIONS
2) the centre turn of DNA is slightly inflected in towards the core histonesat the position of interaction with Hl, and part of Hl penetrates into the grooves of the DNA (Figure 4~). Indeed in Richmond’s model from the X-ray crystallography of the core particle there is an inflection of the DNA at the central region (8), both sidewaysand inwards; this may provide the space for the globular part of Hl. Moreover, from theoretical work (10,ll) it was inferred that the globular part of Hl might additionally penetratethe DNA grooves. Hence, this model is compatible with the 5.5 nm radius. The original experimentsof Allan et al. (9) showedprotection of chromatosomeDNA to micrococcal nucleaseattack by the globular part of histone H5 complexed onto chick chromatin depleted of histoneH5 and Hl . In trying to build a model of a chromatosomewith 11 base pairs either side of the core particle, and touching the globular part of Hl, it is difficult to account for the 0.3 nm increasein radius of gyration of the chromatosomeDNA comparedwith that of the core particle. All modelsin which the 2 x 11 bp extra are coiled as a smooth continuation of the core particle DNA (figure 4a) are certainly incompatible. The model of Richmond et al. (8), with 1.8 turns of DNA on the core particle, can only fit if we assumethat the DNA ends of the core particle component of DNA also rearrange and straightenout in the chromatosome(Figure 4c). There is evidence (13,14) that the ends of the core particle DNA are bound to the octamer in a different way in the core particle comparedto the chromatosome. If the core particle model of Uberbacherand Bunich (15) is taken with 1.9 superhelical turns of DNA to the core particle and if the DNA is assumedto bend outwards from within the core particle component when H5 is present, then the model can explain the measured DNA radius of gyration and a closing of the two turns to allow H5 to interact acrossthe ovaloid centre turn of DNA. ACKNOWLEDGMENTS This work had financial support of the N.F.W.O. and the Nationale Loterij, Belgium, from the Science and Engineering Research Council, U.K., and from the Daresbury Laboratory Synchrotron Radiation Source, Warington, U.K. REFERENCES 1. i4: 5. t* 8: 9. 10. 11. 12. 13. 14. 15.
Igo-KemenesT., W. H&z, & H. Zachau (1982) Ann.Rev. B&hem. 5l, 89-121. SimpsonR.T. (1978) Biochemistry fl,5524-5531 Ibel K., 8z H.B. Stuhrmann (1975) J.Mol.Biol. 93,255-265 Suau P., G.G. Kneale, G.W. Braddock, J.P. Baldwin, E.M. Bradbury (1977) Nucl.Acids Res.&3769-3786 Hjelm R.P., G.G. Kneale, P. Suau, J.P. Baldwin, E.M. Bradbury, K. Ibel (1977) Cell 1Q, 139-151 Wood et al. J.Biol.Chem. (1991) =,5696-5702. Crane-RobinsonC. & O.B.Ptitsyn (1988) Prot.Eng. 2,577-582 Richmond T.J., J.T. Finch, B. Rushton, D. Rhodes, A. Klug, 5-(1984) Nature 311, 532-537 Allan J.D., P.G. Hartman, C. Crane-Robinson,F.X. Aviles (1980) Nature 288, 675679 Tumell W.G., S.C. Satchwell, A.A.Travers (1988), FebsLetters =,263-268 SegersA., L. Wijns, I. Lasters(1991) Biochem.Biophys.Res.Com.174,898-902 Staynov D.Z. & C. Crane-Robinson(1988) EMBO J. 2,3685-3691 Weishet W.O., K.Tatchell, K.E. Van Holde, H. Klump (1978) Nucl.Acids Res. 5, 139-160 Belyavsky A.V., S.G.Bavykin, E.G. Goguadze, A.D. Mirzabekov (1980) J.Mol.Biol. m,519-536 UberbacherE.C. & G.J. Bunich (1985) J.Biomol.Snucture & Dyn. 2,1033-1055. 816
