higher-order structure of saccharomyces cerevisiae chromatin

Higher-Order Structure of Saccharomyces cerevisiae ChromatinAuthor(s): P. T. Lowary and J. WidomSource: Proceedings of the National Academy of Sciences of the United States of America,Vol. 86, No. 21 (Nov. 1, 1989), pp. 8266-8270Published by: National Academy of SciencesStable URL: http://www.jstor.org/stable/34850 .

Accessed: 01/05/2014 15:40

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

National Academy of Sciences is collaborating with JSTOR to digitize, preserve and extend access toProceedings of the National Academy of Sciences of the United States of America.

http://www.jstor.org

This content downloaded from 130.132.123.28 on Thu, 1 May 2014 15:40:49 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=nas

http://www.jstor.org/stable/34850?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Proc. Naitl. Acad. Sci. USA Vol. 86, pp. 8266-8270, November 1989 Biochemistry

Higher-order structure of Saccharomyces cerevisiae chromatin P. T. LOWARY AND J. WIDOM Departments of Chemistry and Biochemistry, University of Illinois, Urbana, IL 61801

Communicated by G. Felsenfeld, July 27, 1989 (received for review Februlary 10, 1989)

ABSTRACT We have developed a method for partially purifying chromatin from Saccharomyces cerevisiae (baker's yeast) to a level suitable for studies of its higher-order folding. This has required the use of yeast strains that are free of the ubiquitous yeast "killer" virus. Results from dynamic light scattering, electron microscopy, and x-ray diffraction show that the yeast chromatin undergoes a cation-dependent folding into 30-nm filaments that resemble those characteristic of higher-cell chromatin; moreover, the packing of nucleosomes within the yeast 30-nm fllaments is similar to that of higher cells. These results imply that yeast has a protein or protein domain that serves the role of the histone Hi found in higher cells; physical and genetic studies of the yeast activity could help elucidate the structure and function of HI. Images of the yeast 30-nm filaments can be used to test crossed-linker models for 30-nm filament structure.

We are interested in using the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe as model organisms in studies of eukaryotic chromatin structure, principally for two reasons. (i) We would like to use yeast genetics to identify and study the proteins necessary for the highest levels of chromosome folding. (ii) Although the manner in which nucleosomes are packaged together in 30-nm chromatin filaments is known approximately, the connectivity of the structure is not known (1). S. cerevisiae and neuronal cells have approximately no linker DNA (2-4), so studies of the higher-order folding of such chromatin could place strong constraints on possible connectivities. Previous studies of neuronal chromatin have shown it to fold into a higher-order structure that appears similar to the 30-nm filaments of higher cells (3, 4). One would like to extend these results by using methods such as electron microscopy of negatively stained samples or x-ray diffraction to study the packing of nucleosomes within the neuronal chromatin 30-nm filaments. How- ever, because it has not been possible to isolate long oligomers of this neuronal chromatin, no such studies have been carried out. One might hope that yeast chromatin would present no such difficulty.

One must first address the question: are the folded states of yeast chromatin similar to those of higher cells? The sequences of the four core histones (particularly H3 and H4) of both yeasts are quite homologous to those of higher eukaryotes (5), and nuclease digestion experiments reveal a similar pattern of DNA protection except for subtle differ- ences near the ends of the nucleosomal DNA (6). In the absence of further information it is reasonable to suppose that yeast nucleosomes will be similar in size and shape to those of higher cells. Surprisingly, it is not known whether this similarity extends to higher levels of chromosome structure. One study reported electron microscopic evidence for the existence of a 30-nm filament-like state of S. cerevisiae chromatin (7), but this has remained a matter of debate (8). A protein homologous to higher-cell histone Hi has yet to be

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. ?1734 solely to indicate this fact.

definitively identified in either yeast (8); but, in all higher cells, histone Hi must be present for the chromatin to fold into 30-nm filaments (1, 9).

MATERIALS AND METHODS Materials, Strains, and Buffers. Zymolyase-lOOT was pur-

chased from ICN. S. cerevisiae strain SKQ2N (10) was the gift of JoAnn Wise (University of Illinois at Urbana). Strain 1160 (11) (a killer-virus-free strain) was the gift of R. Sclafani (University of Colorado, Denver) and R. Wickner (National Institutes of Health). The following buffers were used (all buffers contained 0.5 mM phenylmethylsulfonyl fluoride, added immediately before use from a stock solution in isopropanol): preincubation buffer (1.0 M sorbitol/50 mM potassium phosphate, pH 7.5/10 mM MgCl2/30 mM 2- mercaptoethanol); spheroplasting buffer (1.0 M sorbitol/25 mM potassium phosphate/25 mM sodium succinate, pH 5.5/10 mM MgC92/10 mM 2-mercaptoethanol); lysis buffer [10 mM sodium cacodylate, pH 6.0/18% (wt/vol) Ficoll 400]; buffer A (10 mM sodium cacodylate, pH 6.0/140 mM NaCI/ 1.0 mM MgC92); and buffer B (10 mM Tris HCI, pH 7.5/140 mM NaCl/1.0 mM MgCl2).

Preparation of S. cerevisiae Chromatin. Cultures (500 ml) were grown to early stationary phase [in 1% yeast extract/2% (wt/vol) peptone/2% (wt/vol) dextrose]. The cells were harvested by low-speed centrifugation, resuspended in 0.1 vol preincubation buffer, incubated at 30?C for 15 min with gentle agitation, and collected again by centrifugation. Zy- molyase 100-T (5 mg) was dissolved in 50 ml of spheroplasting buffer and clarified by centrifugation at 1500 x g for 10 min. The cells were resuspended in this clarified zymolyase solution and incubated at 30?C with gentle shaking. Spheroplas- ting was considered complete when 90% of the cells lysed when diluted 1:50 into deionized water, typically after 45-60 min of incubation.

Spheroplasts were harvested by low-speed centrifugation and washed once with spheroplasting buffer plus 2 mM iodoacetate. They were resuspended in 10 ml of lysis buffer. vortex mixed briefly, diluted 1:5 with buffer A plus 2 mM iodoacetate, and centrifuged at 1500 x g for 15 min. The pellet was resuspended in 50 ml of buffer A, layered onto 60% (wt/vol) sucrose in buffer A, and centrifuged 8 min at 7000 rpm in a Sorvall SS-34 rotor. The cloudy top layer was aspirated away, and the middle layer (containing primarily nuclei) was collected. All further centrifugations were for 8 min at 7000 rpm in the SS-34 rotor.

The nuclei were then washed once with buffer A containing 0.20% Nonidet P-40, once with buffer A, and once with buffer B. They were then suspended to 50 ml in buffer B and prewarmed to 37?C for 5 min. The solution was brought to 2.5 mM CaC12, and micrococcal nuclease (Sigma) was added to a final concentration of 0.035 unit/ml (to generate chromatin having average DNA lengths of 4-6 kilobases; less nuclease was added when longer chromatin was desired). Digestion

Abbreviation: DT, translational diffusion coefficient.

8266



Biochemistry: Lowary and Widom Proc. Natl. Acad. Sci. USA 86 (1989) 8267

was for 3 min at 37?C. The solution was then brought to 6 mM EDTA, cooled on ice, and centrifuged.

Chromatin was solubilized by resuspending nuclei in 10 mM Tris HCI, pH 7.5/1 mM EDTA and incubating on ice for 16 hr. Nuclear debris was pelleted by centrifugation, leaving soluble chromatin in the supernatent.

Column Chromatography. Sephacryl S-1000 chromatography was performed on a column 5 cm x 70 cm, in 10 mM Tris HCI, pH 7.5/50 mM NaCl/0.1 mM EDTA/0.5 mM phenylmethylsulfonyl fluoride. Chromatin was dialyzed against the column buffer prior to loading and eluted at a flow rate of 1 ml/min.

Sucrose Gradients. Linear 33-ml 5-30% (wt/vol) sucrose gradients in 5 mM Tris'HCI, pH 7.5/0.5 mM EDTA/0.5 mM phenylmethylsulfonyl fluoride and the desired concentration of NaCI, were loaded with 5 ml of soluble chromatin and centrifuged at 12,000 rpm for 16 hr at 4?C in a Beckman SW 28 rotor. Fractions were analyzed by agarose gel electropho- resis with ethidium bromide staining. For analytical pur- poses, 5 ml of an isokinetic 10-25% sucrose gradient in the above buffer was loaded with 200 Al of chromatin. Centrif- ugation was for 45 min at 50,000 rpm and 4?C in a Beckman SW 55 rotor.

Other Methods. The purification, characterization, and size-fractionation of chicken erythrocyte chromatin with and without Hi and H5 was carried out as described (12-14). Samples were prepared for electron microscopy as described (12) or using the alcian blue method of Sogo and Thoma (15) (see Fig. 2g). The dynamic light scattering measurements were obtained with a commercial instrument, at 23?C. Typ- ical chromatin concentrations were =20 ,tg/ml (A260, 0.2) or less. X-ray diffraction patterns were obtained as described (12), except that the source was an Elliot GX-20 rotating anode.

RESULTS Biochemical Characterization. Our initial attempts to iso-

late S. cerevisiae chromatin from strain SKQ2N included a step of Percoll gradient centrifugation for the purification of nuclei (16). Electron micrographs of this chromatin showed it to be badly contaminated by large numbers of regular and irregular particles. We believed that these contaminants might be residual Percoll particles; we, therefore, eliminated

FIG. 1. Electron micrograph of crude soluble yeast chromatin prior to column chromatography or sucrose gradient centrifugation. The chromatin was adsorbed in 10 mM Tris HCI, pH 7.5/1 mm Na2 EDTA and negatively stained. The chromatin is contaminated by -40-nm-diameter round particles (small arrow) and by larger more varied "husks" (large arrows). (Bar = 100 nm.) (b) Yeast chromatin electrophoresed in 18% polyacrylamide gels purified by size- exclusion column chromatography (lane 1) and the proteins that are soluble in 0.5 M HCI (lane 4). Chicken erythrocyte histones are used as standards (lanes 2 and 3); bands are (from the top) Hi (doublet), H5, H3, H2B, H2A, and H4.

Percoll from our procedure and developed another method for purifying the nuclei. Fig. la shows that chromatin prepared without Percoll was still badly contaminated by such particles, at levels of contamination comparable to those obtained when Percoll was used. These contaminants were divided into two classes. The small arrow indicates one of many small round regularly shaped particles. Larger arrows indicate contaminants that are larger and more varied in size and shape. We refer to this latter class of contaminants as husks. We have determined that the smaller discrete size class of contaminants was the yeast "killer" virus (17). This identification was based on physical and biochemical analysis of the crude yeast chromatin and of purified virus particles (unpublished results). Having identified these contaminants, we avoided them by using yeast strain 1160, which does not carry killer virus, but chromatin prepared from strain 1160 was still contaminated by husks. We have utilized two methods for further purifying this chromatin: size-exclusion column chromatography and sucrose gradient centrifugation.

Fig. lb (lane 1) shows that chromatin purified by column chromatography contains proteins that comigrate with core histones from chicken erythrocytes. Also shown (lane 4) are the proteins that can be extracted from yeast chromatin by 0.5 M HCI, a property common to histones and other basic proteins. The four putative yeast core histones were present, as were several other proteins. Redigestion of the isolated yeast chromatin with micrococcal nuclease revealed a nucleosome repeat length of approximately 165 base pairs (data not shown), in accord with previous observations (2).

Despite the apparent integrity of the core histones, we were, nevertheless, concerned about the possibility of proteolysis affecting our yeast chromatin preparations; we, therefore, carried out several experiments to test for proteolytic activity. We could detect no change in protein gel patterns after prolonged storage of the yeast chromatin at 4?C; we also could detect no changes when a large number of additional protease inhibitors were included in the preparation procedure (data not shown). Finally, since Hi is known to be extremely sensitive to proteolysis, we directly tested for the presence of an Hi-proteolytic activity in S. cerevisiae extracts by adding chicken H5 (an Hi varient) to a stoichiometry of one H5 molecule per yeast nucleosome and then following the fate of the H5 over time with a specific antiserum. No degradation was detected even over several hours at room temperature (J. Godde and J.W., unpublished data). We conclude that our preparations of S. cerevisiae chromatin are probably not affected by proteolysis.

Cation-Dependent Folding. Higher-cell chromatin in vitro is induced to fold progressively from an extended (or "unfold- ed") nucleosome filament state into a more compact higher- order structure, the 30-nm filament, by the addition of a wide variety of cations (1). In Fig. 2, the Na+ concentration- dependent appearance of S. cerevisiae chromatin was compared by electron microscopy with that of chicken erythrocyte chromatin and with that of Hi- and H5-depleted chicken erythrocyte chromatin. When the Na+ concentration was increased from 10 mM to 75 mM, native chicken chromatin folded from -10-nm-wide filaments (Fig. 2e) into -30- nm-wide filaments (Fig. 2b). When Hi and H5 were removed using either of two methods, there was no apparent tendency for the depleted chromatin to fold into 30-nm filaments (1, 9) (Fig. 2 c and f). S. cerevisiae chromatin in the same conditions is shown in Fig. 2 a and d. It is apparently folded into =30-nm-wide filaments; in both ionic conditions, it resembles the native chicken erythrocyte chromatin and not the Hi- and H5-depleted chromatin.

Not all chromatin types are capable of such compaction. For example, our current preparations of chromatin from the distantly related yeast Schizosaccharomyces pombe (Fig. 2h) mimicked the behavior of Hi- and H5-depleted chicken



8268 Biochemistry: Lowary and Widom Proc. Natl. Acad. Sci. USA 86 (1989)

FIG.~~~~~ ;* 2. N dpedet foling ofyat choai compAred to, chicken erythroce cma

(IG) S2.cerevisiepnetfligoyas chromatin in5mcapeae ormicrosctopya dcribked brythSogot anchromati (15). (arnd 100ou nm. (h) SH5oSac charomycsipome chromatin in 100 mM Na' a n 0m preared) forhicros- prthotassiu phrospaten (pH 7.5mMNb) and 0.0m NaMEDa. e;Hl n cHromapltin. Thiske corytrelates wih,omatndi7 may be due to, the presNce of aCproteoltic wactfivity pointhvel sthinoed,candromay-

prearatowe.(ans)fthatidegrads texogmeineouslyr i added0 ciknmH5 Ths teabltyo S. cerevisiae chromatin to5m a'peae fo ldirocp int

Alsorbdb Shown ind Figh2gisa micrograph of0 Sm.)h Screizsiae chrmcspmechromatin in 1 00 mM Na ' nwihtesml a preparedfomirs forasimicrospyae(H75 asnecrbd by1 Sog NanEDThoA(5. Wt thisomethod, Thes crexatendedth formmof higer-cel chomatin aperedec as a filamenlti ofditincty inuclosoeS,hizratherothnya peaat1-nilamnt; thet yegadst chromatneul beavded shimilarly. weti havueaso carie Goute hydrody nambicstudiresuofthe.

Nhs,t-depedn foilding of y.ceevsast chromatin. We measuredo tra0nsflational iffsincefint(T a quatittive measure ofs compcnes) uinFg g dyna micrlightsattering. Threisiaexpr

imn eurdchromatinin1mMN'nwhc thataml was welsz-racioated. thimerefore, the column-prfed S.r cerevisiael chromatinwa fuperter s frcionatedt by sursrdietntntuclment riuatiorthn; fac- tio0ns contaieninghyas chromatin wihbehanedN smlengthyo 3W54e kiaobaseso wererued foruth hydrodynamic studies.o h

sam-epaeraedength aslin the yeast chromatin. The reasultred sown iopcneFig. 3. n yai lgtsatrng hsepr

NTiveor,te chicken-erytrocytScrviie chromatin exiitd asNa ionscntration-dpndenromtinces winh DT thatN corlaednwtho

thme ompactio oberved in the mistchrograpi.hs ofeFi.l2s and

qhiwn i

9.50 -

8.50

4 7.50 E 0

o 6.50

5.50

4.50f 0 40 80 120 160

[Na+] mM

FIG. 3. Na+ concentration dependence of the DT for yeast chromatin and for chicken erythrocyte chromatin with and without Hi and H5. All samples have mean DNA lengths of 3.5-4 kilobases. (o) Chicken erythrocyte chromatin. (A) Hi- and H5-depleted chicken erythrocyte chromatin. (A) Yeast chromatin purified by sizing column chromatography followed by sucrose gradient centrifugation. (A) Yeast chromatin purified by two stages of sucrose gradient centrifugation in two Na+ concentrations.

e. In contrast, Hi- and H5-depleted chromatin exhibited little change in DT over this range of Na+ concentrations (ref. 18; see Discussion); this correlated with the micrographs of Fig. 2 c and f that show little compaction for the depleted chromatin. We obtained quantitatively similar results using two procedures for removing the Hi and H5 (12-14).

S. cerevisiae chromatin showed a significant Na+ concentration-dependent increase in DT, similar to that for native chicken erythrocyte chromatin of the same length and much larger than that for Hi- and H5-depleted chicken chromatin. This correlated with the Na+-dependent folding of yeast chromatin observed in the electron micrographs of Fig. 2 a and d.

We have also tested for a Na+ concentration dependence to the sedimentation coefficient of yeast chromatin, using isokinetic sucrose gradients. We found that an increase in Na+ from 5 mM to 60 mM produced an increase in the sedimentation coefficient equal to that observed for native (Hi and H5 containing) chicken erythrocyte chromatin. This suggests an alternative method for purifying yeast chromatin, in which the material is centrifuged through two successive sucrose gradients, one in low salt and one in higher salt; this should remove contaminants that are unresponsive to the change in ionic conditions. Also shown in Fig. 3 are diffusion coefficient measurements of yeast chromatin that was purified by this method; there is good quantitative agreement with the data for chromatin purified over a sizing column plus one sucrose gradient.

Structure of Yeast 30-nm Filaments. To learn more about the structure of folded yeast chromatin filaments, we have examined them by x-ray diffraction and by electron microscopy using both negative stain and linear shadowing. Fig. 4 shows a gallery of images from a variety of conditions in which higher-cell chromatin folds into 30-nm filaments. Just as when prepared for electron microscopy by other methods, the yeast chromatin folded into filaments with an approxi- mate diameter of 30 nm. In many images, lateral striations with a spacing of 10-15 nm were visible (arrows). The lengths of the shadows in images of linearly shadowed molecules



Biochemistry: Lowary and Widom Proc. Natl. Acad. Sci. USA 86 (1989) 8269

~~~~~~~.~~i ' , ^ t . 3

FIG. 4. Yeast chromatin 30-nm filaments. (a-c) Negatively stained samples. (d-f) Positively stained and linearly shadowed samples at an angle of 260 (d and e) or 18?(f). Arrows indicate lateral striations having a spacing of 10-15 nm. (a) Fixed, in 50 mM Na+. (b-f) Unfixed samples in 3 mM Mg2+. (Bar = 100 nm.)

(Fig. 4 d-f) implied that the yeast 30-nm filaments had a height or thickness of 30-45 nm.

X-ray diffraction patterns obtained from the yeast chromatin in three solution conditions are illustrated in Fig. 5. In low Na+ (Fig. 5, curve a), the principal feature observed was a broad band at =6 nm (s = 0.017 A-1), previously attributed to the diffraction of a nucleosomal disk (19). Compared to chicken erythrocyte chromatin (12, 19), the yeast chromatin in these conditions had much less contrast in the bands at 3.7 and 2.7 nm (s = 0.027 and 0.037 A-1, respectively) (previously attributed to intranucleosomal diffraction (19), and, as expected, no band at =20 nm (s = 0.005 A-') (previously

a

$/) b \.002 .005

0.00 0.02 0.04

S (A-1)

FIG. 5. X-ray diffraction patterns from yeast chromatin purified by column chromatography. Data are plotted as In S2I(s) versus S, where I is the measured scattering intensity, S = (2/A)sin 6, A = 1.54 A, and 20 is the scattering angle. The yeast chromatin is in 5 mM Tris-HCI, pH 7.5/0.5 mM Na2EDTA (curve a), 5 mM TrisAHCI, pH 7.5/0.5 mM Na2EDTA/150 mM NaCI (curve b), or 10 mM Tris HCI, pH 7.5/5 mM MgCI2 (curve c). (Inset) Low-angle data from curve b are plotted on expanded axes.

attributed to the internucleosome spacing) (1). In 150 mM Na+ (Fig. 5, curve b), where the yeast chromatin folded into 30-nm filaments, the full set of diffraction bands (at 30-40, 11, 16, 3.7, and 2.7 nm) characteristic of higher-cell 30-nm filaments (19) were seen. The low-angle band occurred at 33 nm (s = 0.003 A-1) (Fig. 5 Inset). Compared to chicken erythrocyte 30-nm filaments in these conditions, all of the diffraction bands had less contrast. In 10 mM Tris/5 mM Mg2+ (Fig. 5, curve c), the yeast 30-nm filaments again behaved like those of higher cells (12); the relative strength of the 11-nm band increased, and the 33-nm band disap- peared, presumably behind the beamstop cutoff, which was at 1/s = 44 nm for this data set (see Discussion).

DISCUSSION We consider two questions: (i) Does S. cerevisiae chromatin undergo cation-dependent folding into some higher-order structure? (ii) If so, is this structure similar to that of higher cells? The electron micrographs of Fig. 2 show that the yeast chromatin behaves similarly to native chicken erythrocyte chromatin in that it appears to fold from =10-nm filaments in 10 mM Na+ into =30-nm filaments in 75 mM Na+. By contrast, chicken erythrocyte chromatin that has been depleted of histones Hi and H5 shows little folding, with no apparent tendency to form 30-nm filaments. By this criterion, the yeast chromatin does indeed form a higher-order structure that resembles the structure of higher cell chromatin.

A compaction detected by electron microscopy should also be detected in hydrodynamic studies of the molecules in solution. We find (Fig. 3) that yeast chromatin and native chicken erythrocyte chromatin size-fractionated to the same length show a similar increase in DT during a titration with Na+. This correlates with the micrographs and provides support for the conclusion that the yeast and chicken erythrocyte chromatin filaments form higher-order structures that are similar in shape. It should be noted that the measurements of DT do not prove identity of the yeast and chicken 30-nm filaments. Since yeast chromatin has a shorter repeat length than chicken, the samples that were size-fractionated to the same average DNA length in fact contain different numbers of nucleosomes.

Our measurements of DT also correlate with the micrographs in showing much less compaction for chicken chromatin depleted of histones HI and H5. We obtain similar values of DT using two methods for removing the Hi and H5, and our data are in agreement with a previous study (18). These results are in contrast with reports of a small but significant increase in the sedimentation coefficient of depleted chromatin over this range of Na+ concentrations (14). The results could be reconciled if part of the observed increase in the sedimentation coefficient was due to aggre- gation, since this has opposite effects on s and DT.

In summary, the electron micrographs and the hydrodynamic data suggest that yeast chromatin behaves similarly to native chicken erythrocyte chromatin and quite differently from chicken chromatin that has been depleted of Hi and H5.

The second question to be considered is whether the yeast chromatin higher-order structure is similar in internal ar- rangement to that of higher cells. Images from negatively stained and from linearly shadowed yeast chromatin preparations (Fig. 4) bear a striking resemblance to images obtained from higher-cell 30-nm chromatin filaments; indeed, images like these, obtained from rat liver chromatin (9), originally led to the proposal of the solenoid model for the structure of 30-nm chromatin filaments.

Since the micrographs show the yeast 30-nm filaments to be approximately equal in width and in thickness, we pre- sume that they are round in cross-section. We, therefore, interpret the 10- to 15-nm spaced lateral striations as arising



8270 Biochemistry: Lowary and Widom Proc. Natl. Acad. Sci. USA 86 (1989)

from helical grooves in the surface of the filaments. Within the context of the solenoid model, the interpretation of these images is that one is resolving the groove between adjacent helical ramps of nucleosomes that are oriented radially, with their flat faces parallel to the 30-nm filament axis (19). The spacing of the grooves is, therefore, equal to the diameter of the disk-shaped nucleosomes, which, for all higher-cell types, is known or believed to be 11 nm (20). We assume that the dimensions of yeast nucleosomes are also conserved. In that case, the 10- to 15-nm spacing of the lateral striations visible in the yeast 30-nm filaments implies that the yeast nucleosomes, too, are packed radially and oriented parallel to the direction of their filament axis.

The x-ray diffraction patterns in Fig. 5 b and c provide further evidence that the packing of nucleosomes in the yeast 30-nm filaments is similar to that in higher cells. It is difficult to say much about the intensities of the various chromatin x-ray diffraction features, because no attempt has yet been made to analyze them even for higher-cell chromatin; we, therefore, consider only the positions of the bands. The yeast 30-nm filaments show the full set of diffraction bands characteristic of higher-cell 30-nm filaments in vivo and in vitro (1). The band at 33 nm provides evidence for the existence of 30-nm filaments packed loosely together (19). As with higher- cell chromatin (12), this band disappears in 5 mM Mg> because, under those conditions, 30-nm filaments aggregate tightly together and electron density contrast between them is lost. The band at 11 nm is a direct manifestation of the =11-nm spaced grooves visible in the electron microscopic images (19). The appearance of this band in the x-ray patterns of yeast 30-nm filaments confirms that the striations visible in the micrographs of Fig. 4 are indeed significant features of the structure.

Structure of the 30-nm Chromatin Filament. With the results of this study, 30-nm filaments from chromatin types having the full range of allowed DNA linker lengths of -0-80 base pairs (0-270 A) are found to give an 11-nm diffraction band (1, 19, 21). The packing of nucleosomes that leads to the 11-nm lateral striations and to this diffraction band must be an important determinant of 30-nm filament structure.

While the packing of nucleosomes within 30-nm filaments is now understood, the connectivity of the structure is not known (1). In the solenoid model (9, 19), nucleosomes that neighbor laterally in three dimensions are also neighbors in one dimension-i.e., along the nucleosome filament. Other models have been proposed that have a similar packing of nucleosomes but that differ in their linker DNA paths from the solenoid in that laterally neighboring nucleosomes come from non-consecutive locations along the nucleosome filament. One such class of models are referred to as crossed- linker models (1); in these models, linker DNA connects between nucleosomes on opposite sides of the 30-nm filament. It has been shown by model building (22) that for such a higher-order structure to be assembled from chromatin having approximately no linker DNA, the filaments will be thin (-25 nm in diameter), and the pitch angle (the angle off horizontal of the visible striations) will be steep (-35?).

It should be possible to test this model through quantitative study of the helical parameters of the yeast 30-nm filaments. Yeast offers an important advantage for these experiments because it has not been possible to isolate long oligomers of higher-cell :0-bp-linker chromatin (3, 4); yeast presents no such difficulty. Data presented above already hint that both of the key restrictions on crossed-linker models may not be

met, since the angle of the striations visible in Fig. 4 appears to be less than 350 and the filament diameter implied by the shadow lengths appears to be greater than 25 nm. To settle this question, it will be important to quantitatively study a large number of randomly selected yeast 30-nm filaments.

Histone HI in S. cerevisiae? The data presented above show that S. cerevisiae chromatin folds into 30-nm filaments like those of any higher-cell type. For all higher-cell types, this folding requires that histone Hi (or some variant, such as H5) be present at a stoichiometry of one per nucleosome (23); even the -0-bp-linker neuronal chromatin is reported to contain Hi (3, 4). Although no such protein has yet been identified in either S. cerevisiae or Schizosaccharomyces pombe, evidence for the absence of Hi is entirely negative. It will be important to carry out assays for an Hi-likefunction in yeast.

We are grateful to Drs. J. Jaehning, R. Sclafani, D. Shore, R. Wickner, and J. Wise for providing yeast strains; to Dr. F. Thoma for sending a manuscript in advance of publication; to Drs. M. Grun- stein, J. Jaehning, and J. Wise and Ms. J. L. Allen for helpful advice; and to P. Holland, D. Breaux, and J. Godde for assistance. Electron microscopes were provided by the University of Illinois Center for Microanalysis of Materials. J.W. acknowledges research support from the National Institutes of Health, the Searle Scholars Program of the Chicago Community Trust, and from a National Science Foundation Presidential Young Investigator Award.

1. Widom, J. (1989) Annu. Rev. Biophys. Biophys. Chem. 18, 365-395.

2. Thomas, J. 0. & Furber, V. (1976) FEBS Lett. 66, 274-280. 3. Allan, J., Rau, D. C., Harborne, N. & Gould, H. (1984) J. Cell

Biol. 98, 1320-1327. 4. Pearson, E. C., Butler, P. J. G. & Thomas, J. 0. (1983) EMBO

J. 2, 1367-1372. 5. Wells, D. E. (1989) Nucleic Acids Res. 17, Suppl. r311-r346. 6. Lohr, D. (1986) J. Biol. Chem. 261, 9904-9914. 7. Rattner, J. B., Saunders, C., Davies, J. R. & Hamkalo, B. A.

(1982) J. Cell Biol. 92, 217-222. 8. Certa, U., Colavito-Shepanski, M. & Grunstein, M. (1984)

Nucleic Acids Res. 12, 7975-7985. 9. Finch, J. T. & Klug, A. (1976) Proc. Natl. Acad. Sci. USA 73,

1897-1901. 10. Elliott, S. G. & McLaughlin, C. S. (1979) J. Bacteriol. 137,

1185-1190. 11. Vodkin, M. (1977) J. Bacteriol. 132, 346-348. 12. Widom, J. (1986) J. Mol. Biol. 190, 411-424. 13. Graziano, V., Gerchman, S. E. & Ramakrishnan, V. (1988) J.

Mol. Biol. 203, 997-1008. 14. Butler, P. J. G. & Thomas, J. 0. (1980) J. Mol. Biol. 140,

505-529. 15. Sogo, J. M. & Thoma, F. (1989) Methods Enzymol., in press. 16. Ide, G. J. & Saunders, C. A. (1983) Curr. Genet. 4, 85-90. 17. Wickner, R. B. (1981) in Molecular Biology of the Yeast

Saccharomyces cerevisiae: Life Cycle and Inheritance, eds. Strathern, J. N., Jones, E. W. & Broach, J. R. (Cold Spring Harbor Lab., Cold Spring Harbor, NY).

18. Marion, C., Hesse-Bezot, C., Bezot, P., Marion, M. J., Roux, B. & Bernengo, J. C. (1985) Biophys. Chem. 22, 53-64.

19. Widom, J. & Klug, A. (1985) Cell 43, 207-213. 20. Richmond, T. J., Finch, J. T., Rushton, B., Rhodes, D. &

Klug, A. (1984) Nature (London) 311, 532-537. 21. Widom, J., Finch, J. T. & Thomas, J. 0. (1985) EMBO J. 4,

3189-3194. 22. Williams, S. P., Athey, B. D., Muglia, L. J., Schappe, R. S.,

Gough, A. H. & Langmore, J. P. (1986) Biophys. J. 49, 233- 248.

23. Bates, D. L. & Thomas, J. 0. (1981) Nucleic Acids Res. 9, 5883-5894.



higher-order structure of saccharomyces cerevisiae chromatin

Documents