THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 262, No. 25, Issue of September 5, pp. 1221&12222,1987 FFinted in U.S.A.

Chromatin Structure FURTHER EVIDENCE AGAINST THE EXISTENCE OF A BEADED SUBUNIT FOR THE 30-nm FIBER*

(Received for publication, April 6, 1987)

P. Roy Walker and Marianna Sikorska From the Cellular Oncology Group, Division of Biological Sciences, National Research Council of Canada, Ottawa, Ontario K1A ORs, Canada

The size distribution of chromatin fragments re- leased by micrococcal nuclease digestion of liver chro- matin at various ionic strengths was examined. Below 20 mM ionic strength, gradient profiles with a peak centered at 6 nucleosomes are generated, whereas be- tween 20 and 50 mM the peak is always centered on 12 nucleosomes, and above 50 mM ionic strength the 30-nm fiber becomes less accessible to the nuclease and there is a corresponding increase in the size distribu- tion of fragments in the gradients. However, extensive digestions always give profiles with a peak of 12 nu- cleosomes as nuclease-resistant dodecamers accumu- late. All of these observations are consistent with the winding of the 10-nm polynucleosome chain into a helical coil commencing at about 20 mM ionic strength. The helical turns are stabilized by histone H1 interac- tions between 20 and 50 mM ionic strength producing stable dodecamers. Above 50 mM ionic strength the coil condenses longitudinally and the profiles are consist- ent with a random attack of this fiber by the nuclease. Consequently it is not necessary to invoke the existence of a subunit bead to explain the profiles. We further define the conditions at which specific structural tran- sitions take place and provide methodology for the preparation of chromatin at various levels of conden- sation.

Several models have been proposed for the structure of the 30-nm chromatin fiber (1-10). These models can be divided into two fundamentally different categories, those based on some kind of helical coil arrangement of the nucleosomes (1- 8) and those based upon a beaded arrangement of the nucleo- some subunits (9, 10). The latter group consisting of the superbead model proposed by Renz (9) and the nucleomer model proposed by Kiryanov et al. (10) is the most controver- sial. Two major criticisms, a lack of consensus on the actual size of the subunit bead (ranging from 8 nucleosomes/bead (11) to 48 (12)) and a lack of information on the specific structural organization of the particle still remain. However, Zentgraf and Franke (12) have recently extended the beaded fiber concept by proposing that the size of the particle varies from species to species and is a function of the transcriptional activity of the chromatin. For example, in a transcriptionally active tissue such as liver the bead contains 8 nucleosomes, whereas in a transcriptionally inactive tissue such as sea urchin spermatozoa the particles contain 48 nucleosomes.

* This is National Research Council Publication No. 27792. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The two lines of evidence that have been cited in support of the bead models are electron microscopic images of irreg- ular-shaped or “knobby” 30-nm fibers (10, 12-14) and the presence of peaks in sucrose density gradient profiles of nuclease digests of chromatin (9-12). However, both the ir- regular shapes and the gradient peaks are only generally seen at certain intermediate ionic strengths when the 30-nm fiber is partially unfolded, suggesting that these observations relate to some aspect of the 10-30-nm folding process rather than the existence of a specific beaded subunit structure (15).

In this report we have extended our initial observations on chromatin structure (15) into a detailed study of the size distribution of chromatin fragments in sucrose gradients. The data shows that the size of the peak is a function of ionic strength and of the extent of digestion, varying from 6 to 250 nucleosomes for chromatin from a single tissue. The changes in the position of the peak as a function of ionic strength are entirely consistent with the concept that the 30-nm fiber is a helical coil with 12 nucleosomes/turn (8) rather than being only interpretable in terms of a beaded subunit structure. Moreover, we have further defined the conditions at which critical structural transitions occur and provide methodology for the isolation of chromatin fragments at virtually any level of fiber folding.

MATERIALS AND METHODS

Isolation and Digestion of Nuclei-Nuclei were isolated from the livers of male rats as described previously (15, 16) using a buffer consisting of 0.25 M sucrose, 50 mM Tris-HC1, pH 7.5, 150 mM KCl, 5 mM MgCl,, and 0.5 mM phenylmethylsulfonyl fluoride. Following the final wash the nuclei were resuspended in 2.0 ml of digestion buffer a t a concentration of 0.8-1.0 mg/ml of DNA. The digestion buffer consisted of 10 mM Tris-HC1, pH 8.2 (at 30 “c) , 0.5 mM phenylmethylsulfonyl fluoride, and the concentration of KC1 indi- cated in the figure legends. Micrococcal nuclease (Sigma) was added at a concentration of 50 units/ml (15) and digestions were carried out for the times indicated at 30 “C. The reaction was then terminated by the addition of EDTA to a final concentration of 1 mM followed by rapid cooling on ice. The suspension was centrifuged at 25,000 X g for 15 min in the 50Ti rotor of a Beckman L8-70 ultracentrifuge at 5 “C to generate a supernatant containing released, soluble chromatin.

Gradient Centrifugation-The size distribution of a released chro- matin fragments was determined by layering 0.5 or 1.0 ml of super- natant onto a 10-35% (w/w) sucrose gradient prepared in 10 mM Tris-HC1, pH 8.2 (at 5 “C) containing 1 mM EDTA and the concen- tration of salt indicated in the figure legends. Centrifugation was carried out a t 5 “C in a SW40 rotor at 40,000 rpm to preset w2t values of 2 X 10” radiansz/s or in some cases 1 X 10” radians*/s (195 or 98 min, respectively) as described previously (8). The gradients were fractionated using an ISCO model 185 gradient fractionator con- nected to an ISCO P absorbance monitor, set to 260 nm, and fraction collector. 0.7-ml fractions were collected.

Size Analysis of DNA Fragments-DNA was extracted from gra- dient fractions as described previously (8, 15). The size of the DNA fragments was determined by electrophoresis in 0.8% agarose gels using suitable markers as previously described (15).

12218

30-nm Fiber Structure

Chemical Measurement of Total DNA-The DNA content of sam- ples of supernatant and pellet generated after nuclease digestion was measured by the diphenylamine reagent following acid extraction as described previously (15).

RESULTS

Effect of Ionic Strength on the Position of the Peak in Sucrose Gradients-In all these experiments the nuclei were isolated in a buffer of physiological ionic strength to maintain chromatin in its native configuration (15, 17). For the exper- iments described in Fig. 1, digestions were then carried out at ionic strengths ranging from 10 to 130 mM for 5-15 min in order to digest 25-35% of the chromatin (longer digestion times are required at higher ionic strengths (17)). Samples of each supernatant were loaded onto gradients prepared at the same ionic strength at which the digestions were carried out (buffer and EDTA were assumed to contribute 5 mM to the total ionic strength) and centrifuged to a a2t value of 2 x 10”. Under these conditions chromatin fragments containing from 1 to 80 nucleosomes are resolved (8). The profiles, shown in Fig. 1, indicate that the position of the peak moved further down the gradient tube as the ionic strength was increased. Although increases in ionic strength cause chromatin of a given size to compact and sediment more quickly (a), the shifts observed in Fig. 1 cannot be explained by this alone and the profiles likely reflect an increase in the size distribu- tion of the chromatin fragments digested at higher ionic strengths.

Fig. 2 shows the results of a more detailed study of this phenomenon in which time courses of digestion were carried out at several ionic strengths and gradients run at each time point as described above. After fractionation the size of the DNA in the peak fraction was determined and converted to the number of nucleosome equivalents, assuming 200 base pairs/liver nucleosome. The data is expressed in terms of the number of nucleosomes in the peak fraction as a function of the percentage of chromatin digested and is split into two graphs for the sake of clarity. When digestions and runs were carried out at very low ionic strengths of 5 and 10 mM, the peak was always centered on oligomers containing 6 nucleo- somes even at the lowest levels of digestion. In contrast, digestions carried out at ionic strengths in the range of 20-50

0 -1

FRKTION NUMBER

FIG. 1. The effect of ionic strength on the sucrose density gradient profiles of chromatin fragments. Nuclei, isolated in the presence of 150 mM KC1 were digested with micrococcal nuclease at the ionic strengths indicated by the number associated with each profile in the figure. Digestion was timed to solubilize 25-35% of the chromatin. Gradients were run at the same ionic strength as the digestions and fractionated as described under “Materials and Meth- ods.” The horizontal arrow indicates the direction of sedimentation.

12219

PERCENT Of DNA RELEASED INTO S1

FIG. 2. Effect of ionic strength on the changes in the size distribution of chromatin fragments as a function of digestion. Nuclei were digested at the ionic strengths indicated for 2-60 min. The number of nucleosomes in the fragments isolated from the peak fraction of each gradient profile was determined as described under “Materials and Methods” and is plotted against the percentage of DNA released into the supernatant. For the digestions carried out at 90 and 130 mM ionic strength, 0.1 mM CaC12 was added to the digestion buffer. At the lowest levels of digestion at 130 mM ionic strength the gradients were run to a w2t value of 1 X 10”.

mM produced profiles in which the position of the peak was centered a t 10-12 nucleosomes regardless of the extent of digestion (except at 20 mM where the peak shifted slightly at high levels of digestion). Profiles for digestions carried out a t higher ionic strengths of 60 and 80 mM are shown in the left panel and for 70, 90, and 130 mM in the right panel of Fig. 2. At all of these ionic strengths the size of the chromatin fragments at the peak of the profile was a function of both the ionic strength and the extent of digestion. At low levels of digestion the size of the fragments in the peak varied from 40 at 60 mM ionic strength to 280 at 130 mM ionic strength. As digestion progressed, the number of nucleosomes in the peak fractions decreased at all ionic strengths until the size was reduced to 12 nucleosomes. At 60 mM ionic strength this was reached by the time only 15% of chromatin had been released, whereas at 130 mM much more extensive digestion was required indicating a marked decrease in the overall sensitivity of the 30-nm fiber to the nuclease a t higher ionic strengths (at 130 mM ionic strength digestions were carried out in the presence of 0.1 mM calcium chloride to stimulate the rate of digestion (17)).

Effect of Extent of Digestion on the Position of the Peak- Fig. 3 shows, more clearly, the relationship between the size of the peak and the extent of digestion over the entire range of ionic strengths that effect the 10-30-nm fiber transition. The extent of digestion is a good measure of the nucleolytic activity of the enzyme at each ionic strength. The data shows that, depending on the digestion conditions, a gradient profile peak corresponding to anything from 6 to 270 nucleosomes can be generated for chromatin fragments from the same tissue source. The greatest range of peak sizes is observed during short incubations when only approximately 10% of the chromatin has been solubilized by the nuclease. Typical pro- files are shown in Fig. 4 for digestions carried out a t 5, 30, 70, and 120 mM ionic strength. In these short digestions virtually none of the chromatin has been cleaved to monomers or short oligomers, particularly at higher ionic strengths.

The data shown in Fig. 3 can be conveniently discussed in three groups. The first group consists of samples digested a t ionic strengths below 20 mM. At these ionic strengths, the

12220 30-nm Fiber Structure

I I

I O N I C STRENGTH (mM)

FIG. 3. Effect of the extent of digestion on the changes in the size distribution of chromatin fragments. The data is derived from Fig. 2 at selected levels of digestion to illustrate digestion- dependent and digestion-independent changes in the size distribution as a function of ionic strength.

I I I I 1 0 3 6 9 12 15 18

FFWTlON NUMBER FIG. 4. Sucrose density gradient profiles of chromatin frag-

ments released during the early stages of digestion at the indicated ionic strengths. The number of nucleosomes in the peak fractions are also indicated. The percentage digestion at each ionic strength was 7.6% (5 mM), 10.5% (30 mM), 8.2% (70 mM), and 12.1% (120 mM). The horizontal arrow indicates the direction of sedimen- tation.

chromatin fibers are thought to be completely unfolded into 10-nm polynucleosomes chains (8). Since the 30-nm fiber does not exist, the peak at 6 nucleosomes cannot represent a feature of this fiber’s structure. Clearly, therefore, the 10-nm fiber has some sort of structural periodicity that generates a nuclease-sensitive site every 6 nucleosomes (as digestion pro- ceeded these hexamers were rapidly reduced to monomers, data not shown). Moreover, this site is masked in the 30-nm fiber since a peak of 6 nucleosomes is never seen at ionic strengths greater than 20 mM (Fig. 3). The second group of data consists of digestions carried out at ionic strengths between 20 and 50 mM where it is apparent that the 10-nm fiber has folded up into a loose helical coil with 12 nucleo- somes/turn (8). At these ionic strengths there appeared to be a barrier to degradation of particles containing 12 nucleo- somes. A typical profile carried out at 30 mM ionic strength is shown in Fig. 4. There was a relatively broad distribution of sizes with a definite peak at 12 nucleosomes. There was

also some indication of 12-mers being degraded to monomers and smaller fragments suggesting that the 12-mers were only partially stabilized at this ionic strength. For the third group of data consisting of all digestions carried out at ionic strengths above 50 mM, the size of the peak was strictly a function of the extent of digestion. The size distribution was gradually reduced as nuclease activity proceeded, being faster at lower ionic strengths. At low levels of digestion the position of the peak was, therefore, much further down the tube as shown in Fig. 4 for digestions carried out to less than 10% at 70 and 120 mM ionic strength.

Fragments Containing 12 Nucleosomes Are More Resistant to Degradation-It is noteworthy that all the profiles shown in Figs. 1 and 2 that were carried out at ionic strengths above 50-60 mM showed very little degradation to monomers or short oligomers. We studied this in more detail in the exper- iment described in Fig. 5 with a digestion carried out at 80 mM ionic strength. As much as 46% of the chromatin could be solubilized with little or no degradation to small fragments leading to the production of a profile with a substantial peak of 12-mers. Analysis of the optical density profile revealed that only 2.5% of the total optical density on the gradient was present in material containing from 1 to 6-mers. More recent experiments have shown that under similar conditions as much as 85% of chromatin can be solubilized without signif- icant accumulation of monomers and small oligomers (data not shown).

Stratling and Klingholz (18) observed a similar peak at 12 nucleosomes in gradients of digests of liver chromatin. How- ever, they concluded that it was due to the elevated stability of a particle containing 12 nucleosomes which, although the linker DNA was cut by the nuclease, was prevented from dissociating by histone H1 interactions. If this was the case we would expect to see considerable degradation of the DNA in the peak fractions of the profiles shown in Fig. 5A. How- ever, when examined by agarose gel electrophoresis and ana- lyzed by densitometry we found (Fig. 5B) that each of the peak fractions contained only DNA of a discrete size with no indication of a “ladder” which would suggest cutting by the nuclease within the dodecamer. Quite clearly, therefore, the oligomers containing 12 nucleosomes were resistant to the nuclease and not just stabilized against dissociation.

Effect of Ionic Strength on the Stability of 12-mers-The digestion and centrifugation conditi,ms used by Stratling and Klingholz (18) were generally of lower ionic strength than those used here raising the possibility that the 12-mer is more sensitive to the nuclease at lower ionic strengths. To investi- gate this, digestions and gradients were run at several differ- ent ionic strengths and the size distribution of the DNA in the 12-mer peak was examined. The results shown in Fig. 6 indicate that at the intermediate ionic strengths of 20-50 mM there was substantial cleavage of the linkers generating a ladder of oligomers containing from 1 to 12 nucleosomes, particularly at the lowest ionic strength of 20 mM. However, above 60 mM ionic strength the 12-mers were essentially resistant to the nuclease. Despite the cleavage of some of the linker DNAs at the lower ionic strengths the oligomers were resistant to dissociation as observed by Stratling and Klingh- olz (18).

DISCUSSION

The superbead or nucleomer models proposed for the struc- ture of the 30-nm fiber rely heavily on the interpretation of the significance of the peak in sucrose density gradient pro- files of chromatin digests (9-12). These models proposed that peaks in gradient profiles arise from the release of beads as

FIG. 5. Stability of dodecamers isolated from gradient profiles of chromatin fragments digested at 80 mM ionic strength. A, optical density profiles at different levels of digestion. The horizontal arrow indicates the direc- tion of sedimentation. B, densitometry profiles of the DNA fragment found in Fraction 10 of each profile shown in A. DNA was isolated and electrophoresed as described under "Materials and Meth- ods." The inset shows the gel of the DNA fragment found in Fraction 10 a t 46.6% digestion.

FIG. 6. Densitometry of the DNA fragments found in the dodecamer peak at various ionic strengths. Digestions were carried out at each ionic strength to approximately 30% digestion (45.6% at 130 mM ionic strength) and the peak fraction from each gradient pro- file was extracted and the DNA analyzed as described under "Materials and Meth- ods." Each inset shows the ethidium bro- mide-stained gel.

30-nm Fiber Structure

n

3 6 9 12 15 FRACTION NUMBER

t W V z a m Lz

v) 0 m a

the nuclease cuts DNA a t sites on the 30-nm fiber that are between the beads and more exposed to the enzyme. The beads themselves are either more resistant to the nuclease or do not depend on DNA continuity for their stability. At high levels of digestion this would be expected to generate a mon- odisperse size distribution of chromatin fragments centered around the mean number of nucleosomes/bead. Gradient pro- files considered to reflect this are sometimes seen (9-12). However, at low levels of digestion when some of the sites between beads remain uncut, multimers of the beaded sub- units should exist in solution and would be easily resolvable on sucrose gradients in the same way that multimers of nucleosomes (see, for example, Fig. 1) or polyribosomes (19) are. Such profiles are never seen, even at the lowest levels of digestion, as shown in Fig. 4. Instead there is either a broad- ening of the size distribution with the peak remaining a t 12 (at lower ionic strengths) or an actual increase in the size

DISTANCE MIGRATED (cm)

12221 -

- C

DISTANCE MIGRATED (cm)

distribution with a concomitant increase in the position of the peak. Even at higher levels of digestion the position of the peak is not constant (Fig. 1) as would be required by the superbead model. Moreover, this latter observation invali- dates the conclusion that there is a different supranucleoso- mal organization of chromatin in different tissues or species based solely on the position of the peak (12).

The data summarized in Figs. 2 and 3 demonstrate that two variables determine the position of the peak, or more correctly the size distribution, of chromatin fragments in the gradients. The most significant variable is the ionic strength and since changes in ionic strength affect the structure of the 30-nm fiber (8, 15), the data must be discussed in this context. The overall shape of the curves in Fig. 3 are remarkably similar to the changes in the power law dependence of sedimentation coefficient upon mass as a function of ionic strength (Fig. 5 of Ref. 8). This latter graph, based upon a sedimentation

12222 30-nm Fiber Structure

analysis of chromatin structure, gives an indication of the changes in shape of the 30-nm fiber as it folds from a 10-nm polynucleosome chain at low ionic strength into a fully com- pacted 30-nm fiber at physiological ionic strength. Two sharp transitions are observed, the first at approximately 20 mM ionic strength which we consider to reflect the folding of the 10-nm polynucleosome chain into a loose helical coil, and the second at approximately 60 mM ionic strength which we believe marks the completion of the stabilization of the loops of the helical coil and marks the onset of the longitudinal compaction of the coil towards the smooth fiber seen in electron micrographs taken at high ionic strengths (3, 14).

Below 20 mM ionic strength, the peak in the gradient profiles in the present study is centered on 6 nucleosomes even at the lowest levels of digestion (Fig. 4) and as digestion proceeds this is rapidly broken down to produce essentially equal amounts of 1-6-mers and eventually completely to monomers indicating that the 10-nm fiber is extremely sen- sitive to the nuclease and has little structural organization. At the intermediate ionic strengths (between 20 and 50 mM) the optical density profiles show a peak at 12 nucleosomes which is consistent with our earlier conclusion that the poly- nucleosome chain is now wound into a loose helical coil with 12 nucleosomes/turn. As the ionic strength increases, the nucleosomes within each turn become stabilized and less sensitive to nuclease attack as indicated in Fig. 6. This in- crease in stability leads to the accumulation of 12-mers in the gradients, producing a "superbead-like'' profile. Above 50-60 mM ionic strength there is a marked decrease in the overall sensitivity of the fiber to the nuclease (15) which parallels its increasing compaction (8). These changes are reflected in the gradient profiles of the size distribution of digests carried out at higher ionic strengths (Fig. 4). At each ionic strength the size distribution is initially large and decreases as digestion proceeds, eventually reaching a peak of 12 as nuclease-resist- ant 12-mers accumulate (Figs. 2 and 5B). The rate at which this occurs decreases as the ionic strength increases.

To counter the argument that superbead profiles are only seen in digests carried out at intermediate ionic strengths (typically 20-60 mM) Zentgraf and Franke (12) carried out digestions and gradient analyses at a "near physiological ionic strength" of 100 mM monovalent cation. Under these condi- tions, nuclei isolated from tissues with vastly different tran- scriptional activities produced gradient profiles with peak sizes that were inversely proportional to the level of gene expression. Thus liver nuclei showed a peak at 8 nucleosomes while erythrocytes had a peak at 20 nucleosomes when di- gested under identical conditions. The authors concluded that chromatin has a different higher order structure in different tissues, with the number of nucleosomes packaged into each superbead being related to the transcriptional activity of the tissue. As indicated quite clearly in Figs. 2 and 3 there is a continuum of peak positions at the higher ionic strengths and indeed under certain conditions profiles can be generated from transcriptionally active liver chromatin with peaks that exceed those of the inactive sea urchin spermatozoa chromatin reported by Zentgraf and Franke (12). In other words, conclu- sions about the supranucleosomal organization of chromatin cannot be drawn from the size of gradient peaks carried out under only a single set of conditions.

Furthermore, for a given set of conditions, a larger size distribution may be expected for chromatin fragments derived from cells such as erythrocytes which appear to have chro- matin that is more resistant to unfolding as the ionic strength is decreased (20). Therefore, the presence of a different sized

peak may reflect differences in chromatin stability, due to the presence of histone H5 for example, rather than a fundamen- tal difference in higher order structure. Moreover, 100 mM monovalent cation ionic strength cannot be considered phys- iological since the 30-nm chromatin fiber is capable of undergoing further compaction as the salt is increased to more physiological levels of 150 mM (3, 8).

In our previous work (8, 15, 17) we have stressed the importance of isolating nuclei in buffers containing monova- lent cations of physiological ionic strength. This approach maintains chromatin in its condensed state providing a better starting material for structural studies. The fiber can then be unfolded to the desired degree by carrying out digestions and subsequent analyses at lower ionic strengths. The data pre- sented in this paper show that as the ionic strength is de- creased from 150 to 60 mM the fiber becomes more accessible to the nuclease as the reduced charge neutralization causes the coil to expand (8). Since the majority of the linker regions between the nucleosomes within the loops remain inaccessible to the nuclease this can be exploited to solubilize large amounts of chromatin with little or no degradation to mono- mers and short oligomers, providing an excellent source of material for further studies. In addition, we have defined more precisely the ionic strengths at which specific structural tran- sitions take place. For example, 20 mM ionic strength marks the level at which a fiber with a 12-mer as an essential element first appears in solution. Studies at this ionic strength should provide insight into the rearrangements of the nucleosomes that must occur as the 10-nm fiber becomes organized into a loosely-coiled 30-nm fiber. The role played by histone H1 in stabilizing the 12-mer by binding to the linker DNA should be more amenable to study in the 20-50 mM ionic strength range. Under appropriate conditions large amounts of 12- mers with little or no degradation of the linkers can be isolated. This stabilization is essentially complete at 60 mM ionic strength, but the fiber continues to compact until pre- cipitation occurs at approximately 170 mM ionic strength (8). Although histone H1 is apparently necessary for the stabili- zation of the 12-mer, possibly by a polymerization process (21), the nature of the forces promoting the further compac- tion of the fiber above 60 mM ionic strength remain unknown.

Acknowledgments-We wish to thank J. L. Sherwood and J. Le- Blanc for their excellent technical assistance and D. Gillan for preparation of the photographs and drawings.

REFERENCES 1. Finch, J. T., and Klug, A. (1976) Proc. Natl. Acad. Sci. U. S. A. 7 3 , 1897-

1 om 2. Azorin, F., Martinez, A. B., and Subirana, J. A. (1980) Int. J. Biol. Ma-

3. Woodcock, C. L. F., Frado, L.-L. Y., and Rattner, J. B. (1984) J. Cell Biol.

l""*

cromol. 2,81-92

4. Worcel, A,, and Benyajati, C . (1977) Cell 12,83-100 5. Staynov, D. Z. (1985) Int. J. Biol. Macromol. 5,3-9 6. Makarov, V., Dimitrov, S., Smirnov, V., and Pashev, I. (1985) FEBS Lett.

99.42-52

181,357-361

8. Walker, P. R., and Sikorska, M. (1987) J. Biol. Chem. 2 6 2 , 12223-12227 7. Fulmer, A. W., and Bloomfield, V. A. (1982) Biochemistry 21,985-992

9. Renz, H. (1979) Nucleic Acids Res. 6,2761-2767 10. Kiryanov, G., Smirnova, T. A., and Polyakov, V. Yu. (1982) Eur. J .

11. Stratling, W. H., Muller, U., and Zentgraf, H. (1978) Exp. Cell Res. 117 , Biochem. 124,331-338

mn-m 1

12. Zentgraf, H., and Franke, W. W. (1984) J. Cell Bwl. 99,272-286 13. Azonn. F., Perez-Grau, L., and Subirana, J. A. (1982) Chromosoma (Berl)

"" ""

14. Thoma, F., Koller, Th., and Klug, A. (1979) J. Cell Biol. 8 3 , 403-427 15. Walker, P. R., Sikorska, M., and Whitfield, J. F. (1986) J. Biol. Chem. 2 6 1 ,

8 5 , 251-260

7nAA-7n.51 16.

17. 18. 19. 20. 21.

Sikorska, M., MacManus, J. P., Walker, P. R., and Whitfield, J. F. (1980)

Walker, P. R., an Slkorska, M. (1986) Biochemistry 25,3839-3845 Stratling, W. H., and Klingholz, R. (1981) Biochemistry 20, 1386-1392 Walker, P. R., and Whitfield, J. F. (1985) J. Bioi. Chem. 260,765-770 Olins, D. E., and Olins, A. L. (1972) J. Cell Btol. 53,715-736 Glotov, B. O., and Nikolaev, L. G. (1984) Mol. Biol. (Mosc.) 17,717-739

."" .""* Biochem. Biophys.,Res. Commun. 9 3 , 1196-1203