This article is contained in the following reference:
Neutrons in Biology, edited by Schoenborn and Knott
Plenum Press, New York, 1996
Vito Graziano, Sue Ellen Gerchman, Dieter K. Schneider, and Venki Ramakrishnan1
Biology Department Brookhaven National Laboratory Upton, New York 11973 and
1Department of Biochemistry, University of Utah School of Medicine, Salt Lake City, UT 84132
We have been engaged in studies of the structure and condensation of chromatin into the 30nm filament using small-angle neutron scattering. We have also used deuterated histone H1 to determine its location in the chromatin 30nm filament. Our studies indicate that chromatin condenses with increasing ionic strength to a limiting structure that has a mass per unit length of 6-7 nucleosomes/11nm. They also show that the linker histone H1/H5 is located in the interior of the chromatin filament, in a position compatible with its binding to the inner face of the nucleosome. Analysis of the mass per unit length as a function of H5 stoichiometry suggests that 5-7 contiguous nucleosomes need to have H5 bound before a stable higher order structure can exist.
The genome of higher organisms is organized as chromatin, a complex of DNA with proteins called histones. The fundamental unit of chromatin is the nucleosome, which consists of two copies each of the core histones H2a, H2b, H3 and H4, around which are wrapped approximately two turns of DNA (van Holde, 1989).
Ever since small-angle neutron scattering began to be used in biology - a state brought about by the advent of high-flux sources and area detectors - it has made important contributions to chromatin structure. In fact, contrast variation using neutron scattering on nucleosomes in solution (Hjelm et al., 1977; Pardon et al., 1975) provided the earliest physical evidence that DNA was on the outside of the histone octamer in the nucleosome.
The linker histone H1, or its counterpart H5 in avian erythrocyte chromatin, binds to the outside of the nucleosome, and is required for the organization of nucleosomes into a higher-order structure called the 30nm filament. The filament becomes highly compact at ionic strengths that approach the physiological range. A definite model for how nucleosomes are organized into the 30nm filament was first proposed by Finch & Klug (1976), based on electron microscopy. In this model, called the 'solenoidal model', nucleosomes are arranged in a helical array, with 6-8 nucleosomes per turn of the helix. Each nucleosome is connected by linker DNA to its neighbor in the helix. In this model, linker DNA and histone H1 are proposed to be inside the filament. Since then, other models for the structure of the 30nm filament have been proposed, based on various experimental data. The continuously bent linker model of McGhee et al. (1983) differs from the solenoidal model in that the linker DNA, and histone H1/H5 can alternate between the inside and outside. In the helical ribbon model of Woodcock et al. (1984), neighboring nucleosomes make a zigzag pattern to form a ribbon. The entire zigzag ribbon then winds into a helix to form the 30nm filament. The crossed-linker model of Staynov (1983) proposes that a single helical array of nucleosomes has straight linker DNA that threads back and forth through the interior of the fiber. In the crossed-linker model of Williams et al. (1986) two arrays of nucleosomes are connected by a zigzag pattern of linker DNA; the double array twists about its axis to form the fiber, so that the linker DNA goes back and forth through the interior.
The models differ in some important measurable properties. For example, the solenoidal class of models propose a mass per unit length that is almost half that of the helical ribbon and crossed-linker models. Another important difference is the predicted location of the linker histone H1 in the filament.
Neutron scattering on long filaments can reveal directly the mass per unit length and cross-sectional radius of gyration of the filaments. Moreover, if it is possible to remove the linker histones H1/H5 and replace them by a deuterated counterpart, then by contrast variation one could determine the location of the deuterated component relative to the rest of the structure. In this article, we shall review some of our results in this area. All our neutron scattering measurements were done using the small-angle spectrometer on beamline H9B of the High Flux Beam Reactor at Brookhaven National Laboratory (Schneider & Schoenborn, 1984).
By lightly digesting nuclei with micrococcal nuclease, it is possible to obtain chromatin fragments that have in excess of 100 nucleosomes. These filaments are highly extended. The scattering from dilute solutions of such filaments may be analyzed using the rod approximation (Porod, 1982). In this approximation, one considers chromatin filaments to be highly elongated rods, and at low angles, the scattered intensity obeys the relation:
The scattering vector q = 4
sin
/
,
where 2
is the scattering angle and
is the wavelength. The scattering is related to the mass per unit length µ
and the cross-sectional radius of gyration Rx. Thus cross-sectional
Guinier plots, which are plots of ln[qI(q)] vs. q2
should be linear at low angles, and the cross-sectional radius of gyration and
mass per unit length of the filaments can be estimated respectively from the
slope and intercept of a linear fit to the data. Such an analysis was applied
to chromatin filaments using both X-rays (Sperling & Tardieu,1976) and
neutrons (Baudy & Brain, 1978; Suau et al., 1979). In our work,
we have relied almost exclusively on this type of analysis. The advantage of
neutrons in this context is the ability to deuterate specific components and
vary the contrast, or the difference between the scattering length densities
of macromolecule and solvent. To analyze the data at various contrasts, we have
used the formalism developed by Stuhrmann and his coworkers (Ibel &
Stuhrmann, 1975) for the three-dimensional case and applied it to the
two-dimensional case of the rod approximation.
In our first study, we asked how the mass per unit length and cross-sectional radius of gyration of chromatin filaments changed with the ionic strength of the solvent. To some extent, this question had been addressed using neutron scattering before by Suau et al. (1979), who analyzed chromatin at both high and low ionic strength. However, we wanted to resolve the controversy regarding the mass per unit length of chromatin filaments by doing scattering and scanning transmission electron microscopy (STEM) experiments on the same sample. Further, electron microscopy (Thoma et al., 1979) and later scattering experiments (Bordas et al., 1986) suggested that at low ionic strength, chromatin already had an organization into a higher-order structure. This meant that it was important to study chromatin at several ionic strengths to determine its course of condensation.
We compared the results from scattering and STEM studies on native chicken erythrocyte chromatin as a function of ionic strength (Gerchman & Ramakrishnan, 1987). In a second study (Graziano et al., 1988), we asked whether it was possible to reconstitute depleted chromatin with pure histone H5 and obtain scattering curves that behaved like native chromatin. The results from these two studies are summarized in Figures 1 and 2. Figure 1a shows scattering curves from native chicken erythrocyte chromatin in NaCl concentrations ranging from 10 to 60mM. As can be seen, both the slope and the intercept of the linear region at low angles increase, suggesting that the fibers are becoming thicker and more compact. The steep inner slopes yield radii of gyration that range from 80-125 Å, which is characteristic of a higher-order structure. Figure lb shows the scattering curves for chromatin in 80, 100 and l20mM NaCl. There is little difference between these curves, showing that chromatin has reached a limit of compaction. We then examined depleted chromatin, which was made by selectively removing linker histones H1 and H5 from chromatin under conditions in which nucleosome sliding does not take place. In Figure 1e, which shows the scattering curves for depleted chromatin, we see that in the absence of H1/H5, there is no steep inner slope, and the outer shallow slope is characteristic of a '10nm' filament, which represents nucleosomes without any higher level of organization.
In Figures 1c and 1d, we address the question of whether it is possible to add back linker histone under appropriate conditions and get back the structure of native chromatin. As can be seen in Figure 1c, the scattering curves for depleted chromatin reconstituted with H5 are very similar to the corresponding curves for native chromatin shown in Figure la. Similarly, reconstituted chromatin reaches a limiting structure (Figure 1d) just as native chromatin does.
The mass per unit length can be calculated from the intercepts made by the linear fits in Figure 1, using a calibration for the absolute intensity (Jacrot & Zaccai, 1981). The values are expressed in units of 'nucleosomes per 11nm'. This is because the nucleosome is approximately 11nm on edge, so the repeating unit along a filament in which nucleosomes are in contact is of this order. Thus the 11nm corresponds to a 'turn' of the helix in several models for chromatin.
In Figure 2, we show the mass per unit length of native and reconstituted chromatin as a function of the NaCl concentration in the solvent. Clearly, the mass per unit length levels off at a value of 6-7 nucleosomes/11nm, and the values for reconstituted chromatin are in good agreement with those of native chromatin. This value for the plateau is inconsistent with models for the 30nm filament that require much higher mass per unit length, and is in agreement with the prediction of the solenoidal class of models. Similar results were obtained when chromatin was reconstituted with pure chicken H1 rather than H5 (Graziano & Ramakrishnan, 1990). These experiments establish that we are able to obtain reasonable scattering curves for chromatin and that we can reconstitute depleted chromatin with linker histone to get scattering behavior indistinguishable from that of native chromatin.
Deuteration of eukaryotic proteins has been greatly facilitated by the development of systems for overexpression of recombinant genes in E. coli. One of the most successful systems for expressing proteins in E. coli is the T7 expression system (Studier et al., 1990). Unfortunately, neither this system nor various other systems are capable of expressing full-length chicken H5 in E. coli (Gerchman et al., 1994). However, the gene for histone H1a, which is one of the six chicken histone genes, can be expressed well in E. coli using the T7 system. We were able to fully deuterate H1 by growth of E. coli in D2O in a fully deuterated minimal medium with deuterated succinate as the carbon source. The resulting H1 was judged to be greater than 99% deuterated from electro-spray ionization mass spectrometry. We were also able to produce 65% deuterated H1 by growth of E. coli in 80% D2O using protonated glucose as the carbon source.
We now describe the results of our experiments to locate H1 in the 30nm filament by using deuterated H1 and contrast variation (Graziano et al., 1994). We studied native chromatin (N), chromatin reconstituted with protonated recombinant H1 (H), chromatin reconstituted with partially deuterated H1 (DH) and chromatin reconstituted with fully deuterated H1 (D). These samples were all studied at 85mM NaCl, because our previous studies had shown that chromatin is almost fully compact at this ionic strength, but is still reasonably soluble. Each sample was dialyzed into H2O and D2O buffers and the appropriate H2O/D2O mixtures made by mixing.
Figure 3 shows cross-sectional Guinier plots on N- and H-chromatin (Figure 3a) and D-chromatin (Figure 3b). The curves for N- and H-chromatin are nearly identical in both 30% and 80% D2O. This was true for the other contrasts measured as well, showing that the scattering curves for native chromatin and chromatin reconstituted with recombinant H1 are indistinguishable even as a function of contrast. The slopes for 30% and 80% D2O are very similar. On the other hand, the slopes for D-chromatin in 30% and 80% D2O are quite different.
Figure 4 shows the variation of the normalized intercept [qI(q)]0 ½ as a function of the percent D2O in the solvent. Again, the curves for N- and H-chromatin are indistinguishable, and the match-point of 49.8% D2O is close to that calculated for chicken erythrocyte chromatin. The match-point for D-chromatin is shifted to a value of 63.1% D2O, again in excellent agreement if one assumes that the H1 is fully deuterated. The match point for DH-chromatin has an intermediate value of 58.6% D2O which is exactly what one would expect for H1 that is 65% deuterated. In Figure 5, the variation of Rx2 as a function of inverse contrast (Stuhrmann plot) is shown. The steep negative slope for D-chromatin and the insignificant slope for N- and H-chromatin shows clearly that the deuterated material is on the inside of the filament. Again, the values for DH-chromatin lie intermediate between those for H- and D-chromatin. The lack of a significant curvature suggests that there is little difference (in projection down the fiber axis) between the centers of mass of the H1 and chromatin components.
An estimate of the separate cross-sectional radii of gyration of H1 and the rest of chromatin can be obtained by plotting Rx2 as a function of f, the fractional excess scatter of H1, which will depend on the contrast (Serdyuk, 1975). The contrast in turn, will depend on both the percent D2O in the solvent, and the level of deuteration. In Figure 6, we see that Rx2 decreases linearly with f. As expected, the points for both partially deuterated chromatin and fully deuterated chromatin fall on the same line. From the extrapolated values at f=1 and f=0, we can estimate the values for the cross-sectional radius of gyration of H1 and the rest of chromatin respectively. If one uses the parallel axis theorem and the measured values of 30Å and 45Å for the radii of gyration of H1 and the nucleosome respectively, then one estimates that the chromatin component (largely nucleosomal) lies about 115Å from the fiber axis, whereas the H1 component lies about 60-65Å from the fiber axis. If one draws a box of roughly the dimensions of a nucleosome, then it can be seen from Figure 7 that the center of mass of H1 roughly coincides with the inner face of the nucleosome.
A number of caveats must be applied to the kind of analysis used in this work. The first is that chromatin filaments are not infinitely long, rigid and uniform rods. Thus the Porod approximation used in the analysis is not rigorously true. Hjelm (1985) has studied the effect of finite length and non-uniform cross-section on the rod approximation. In general, the results of this study should be close to the 'real' values, but it is not clear how much the deviation is at low ionic strengths when the filaments are highly extended. Another point is that chromatin filaments are intrinsically heterogeneous. It is not merely that the cross-section is not uniform, but it changes from one section of the filament to the next. There may also be whole classes of filaments whose structure deviates from the mean significantly. This and other issues have led to some question of whether in fact there is a specific structure for the 30nm filament (Giannasca & Horowitz, 1993; Zlatanova et al., 1994). The values that we measure are root-mean-squared values, and there may be considerable variation between individual filaments. However, there is no question that regardless of individual variation in the filaments, on average, the location of H1 is interior relative to the nucleosome.
Although we have determined that the center of mass of the linker histone H1 is inside relative to the nucleosome, this does not by itself place the individual parts of H1 with great accuracy. The reason is that all linker histones have a tripartite structure. They consist of a central globular domain that binds to the nucleosome, and are flanked by extended, highly basic arms. In particular, the C-terminal arm makes up about half the protein, and almost half the residues in the arm consist of lysine or arginine. Thus our study cannot say anything about the relative disposition of the C-terminal and globular domains. However, if the arms could be preferentially deuterated relative to the globular domain, then it might be possible to determine their location using the methods described here.
While small-angle scattering, especially with neutrons, is useful to answer global questions, it cannot, unfortunately, provide definitive answers to the problem of chromatin higher-order structure. The reason is that the models differ mainly in the connectivity of DNA: What is the path of the linker DNA as it goes from one nucleosome to the next? Is this path topologically the same in all filaments or does it vary? What is the chemical nature of the interaction between linker histones and chromatin? To answer these questions, it is likely that one will have to construct uniform filaments that consist of tandem repeats of a deigned sequence of DNA that positions histone octamers precisely, and obtain fiber diffraction pattems or even single crystals from these filaments. Even if it is possible to construct such filaments and obtain their structure, it will not necessarily settle the question of the structure of chromatin in the cell nucleus, as there is some evidence that the structure formed by refolding in vitro is not the same as that in the cell nucleus (Giann.sca & Horowitz, 1993). Despite these reservations, it is our opinion that neutron scattering experiments have provided some definite answers in a field that is riddled with controversy.
This work was supported by the Office of Health and Environmental Research of the U.S. Department of Energy, and grant GM 42796 from the NIH.
Baudy, P., & Bram, S., (1978). Chromatin fiber dimensions and nucleosome orientation: a neutron scattering investigation. Nucl. Acids Res., 5:3697-3714.
Bordas, J., Perez-Grau, L., Koch, M.H.J-., Vega, M.C., & Nave, C., (1986). The superstructure of chromatin and its condensation mechanism. Eur Biophys. J., 13:157-173.
Finch, J.T., & Klug, A., (1976). Solenoidal model for superstructure in chromatin. Proc. Natl. Acad. Sci. USA,73:1897-1901.
Gerchman, S.E., Graziano, V., & Ramakrishnan, V., (1994). Expression of chicken linker histones in E.coli: Sources of problems and methods for overcoming some of the difficulties. Protein Expr Purif., 5:242-251.
Gerchman, S.E., & Ramakrishnan, V., (1987). Chromatin higher-order structure studied by neutron scattering and scanning transmission electron microscopy. Proc. Natl. Acad. Sci. USA, 84:7802-7806.
Giannasca, P.J., & Horowitz, R.A., (1993). Transitions between in situ and isolated chromatin. J. Cell Sci., 105:551-561.
Graziano, V., Gerchman, S.E., & Ramakrishnan, V., (1988). Reconstitution of chromatin higher-order structure from histone H5 and depleted chromatin. J Mol. Biol., 203:997-1007.
Graziano,V.,&Ramakrishnan,V.,(1990).Interaction of HMG14 with chromatin. J Mol.Biol., 214:897-910.
Graziano, V., Gerchman, S.E., Schneider, D.K., & Ramakrishnan, V., (1994). Histone H1 is located in the interior of the chromatin 30-nm filament. Nature, 368:351-354.
Hjelm, R.P., (I 985). The small-angle approximation of X-ray and neutron scatter from rigid rods of nonuniform cross-section and finite length. J. Appl Cryst., 18:452-460.
Hjelm, R.P. , Kneale, G.G., Suau, P., Baldwin, J.P., & Bradbury, E.M., (1977). Small angle-neutron scattering studies of chromatin subunits in solution. Cell, 10: 139-15 1.
Ibel, K., & Stuhrmann, H., (1975). Neutron scattering from myoglobin in solution. J. Mol. Biol., 93:255-265.
Jacrot, B., & Zaccai, G., (1981). Determination of molecular weight by neutron scattering. Biopolymers 20:2413-2426.
McGhee, J.D., Nickol, J.D., Felsenfeld, G., & Rau, D.C., (1983). Higher order structure of chromatin: orientation of nucleosomes within the 30 nm chromatin solenoid is independent of species and linker length. Cell, 33:831-841.
Pardon, J. F., Worcester, D.L., Wooley, J.C., Tatchell, K., van Holde, K.E., & Richards, B.M.,(1975). Low-angle neutron scattering from chromatin subunit particles. NucL Acids Res., 2:2163-2176.
Porod, G., (1982). In Small Angle X-ray Scattering (O. Glatter & O. Kratky, editors) pp 1 7-33, Academic Press, London
Schneider, D.K., & Schoenborn, B.P., (1984). In Neutrons in Biology (B.P. Schoenborn, editor) vol. 27, pp 1 19-142, Plenum Press, New York.
Serdyuk, I.N., (1975). Electromagnetic and neutron scattering from the 50S subparticle of E. coli ribosomes. Brookhaven Symp. Biol, 27:IV:49-60.
Sperling, L., & Tardieu, A., (1976). The mass per unit length of chromatin by low-angle X-ray scattering. FEBS Lett., 64:89-91.
Staynov, D., (1983). Possible nucleosome arrangements in the higher order structure of chromatin. Int. J Biol Macromol., 5:3-9.
Studier, F.W., Rosenberg, A.H., Dunn, J.J., & Dubendorff, J.W., (1990). Use of T7 RNA polymerase to direct expression of cloned genes. Meth. Enzymol., 185:61-89.
Suau, P., Bradbury, E. M., & Baldwin, J.P., (1979). Higher-order structure of chromatin in solution. Eur, J Biochem., 97:593-602.
Thoma, F., Koller, T., & Klug, A., (1979). Involvement of histone H1 in the organization of the nucleosome and of the salt dependent superstructures of chromatin. J. Cell Biol., 83:403-427.
van Holde, K.E., (1989). Chomatin, Springer-Veriag, New York.
Williams, S.P., Athey, B.D., Muglia, L.J., Schappe, R.S., Gough, A.H., & Langmore, J.P., (1986). Chromatin fibers are left-handed helices with diameter and mass per unit length that depend on linker length. Biophys. J., 49:233-248.
Woodcock, C.L. F., Frado, L.-L., & Rattner, J.B., (1984). The higher-order structure of chromatin: evidence for a helical ribbon arrangement. J. Cell. Biol, 99:42-52.
Zlatanova, J., Leuba, S.H., Yang, G., Bustamante, C., & van Holde, K., (1994). Linker DNA accessibility in chromatin fibers of different conformations: A reevaluation. Proc. Natl. Acad. Sci. USA, 91:5277-5288.
