Alteration of domain architecture of chicken erythrocyte chromatin

Alteration of domain architecture of chicken erythrocyte chromatin

Biochimica et Biophysica Acta, 782 (1984) 415-421 415 Elsevier BBA 91377 ALTERATION OF D O M A I N ARCHITECTURE OF CHICKEN ERYTHROCYTE CHROMATIN AR...

495KB Sizes 0 Downloads 57 Views

Biochimica et Biophysica Acta, 782 (1984) 415-421

415

Elsevier BBA 91377

ALTERATION OF D O M A I N ARCHITECTURE OF CHICKEN ERYTHROCYTE CHROMATIN ARUPA GANGULY and BUDDHADEB BAGCHI

Department of Physics, University College of Science, 92, Acharya Prafulla Chandra Roa~ Calcutta 700009 (India) (Received February 27th, 1984)

Key words: Chromatin structure; Domain architecture; Ionic strength; Nuclear protein; (Chicken)

The higher-order organisation of chromatin in chicken erythrocyte nuclei as a function of the ionic strength of the nuclear suspension buffer and also of the time of incubation in this buffer prior to nuclease digestion has been investigated. This organisation is described in terms of a physical parameter called the domain length. The 45-kbp-long domains of control nuclei were unravelled to give rise to domains of length 150 kbp on overnight equilibration at 0oC of the nuclei in standard isolation buffer containing 0.135 M NaC! prior to nuclease digestion. However, transition to the equilibrium state was preceded by a metastable and irregular domain architecture when the nuclei were incubated for only 1 h. In contrast, the domain length remained unchanged when nuclei were incubated in the isolation buffer alone for identical periods of time. The proteins dissociated at the higher ionic strength were characterised and their role in stabilising the domain structure is discussed.

Introduction

The organisation of DNA in eukaryotic nuclei follows a hierarchy of compaction. This compaction is induced by its association with a large number of proteins, both chromosomal and nonchromosomal. At one level of organisation, the 250 A, nucleofilaments generated from 100 A thick beaded chains [1-3] are folded into loops or domains [4,5]. In most systems a supporting structure for the attachment of the ends of the domains can be defined [6]; however, for chicken erythrocyte the topological constraints are probably relatively free [7]. As yet, the nature of the constraints is not well defined. Also, how the modification in the strength of the protein-protein and protein-DNA interaction, responsible for the maintenance of the loops, changes the domain architecture is a subject of investigation. In the present paper, the effects of predigestion incubation of the nuclei in the standard isolation buffer of Hewish and Burgoyne [8], having a total 0167-4781/84/$03.00 © 1984 Elsevier Science Publishers B.V.

ionic strength of 0.08 M, and in a buffer of increased ionic strength (0.22 M) on the higherorder organisation of chromatin were compared. In the case of standard buffer, the length distribution of soluble chromatin obtained after various times of nuclease digestion could be fitted with the domain model of chromatin structure [7]. The results obtained after incubation in higher ionic strength buffer suggested an alteration in chromatin organisation which could be partially interpreted in the light of the domain model. The proteins dissociated in the higher ionic buffer were analysed on sodium dodecyl sulphate polyacrylamide gels. Materials and Methods

Isolation of chicken erythrocyte nuclei Chicken erythrocyte nuclei were isolated in buffer A (0.34 M sucrose/0.01 M NaC1/0.06 M KC1/0.15 mM spermine/0.5 mM spermidine/15 mM fl-mercaptoethanol/O.1 mM PMSF/15 mM

416

Tris-HC1 (pH 7.4) plus 2 m M E D T A and 0.5 m M E G T A following the method of Wilhelm et al. [9].

Predigestion incubation of the nuclei in two different buffers Portions of the nuclear suspension at a concentration of 100 .4260 units (measured in 0.1 N N a O H ) were incubated at 0°C for different times either in buffer A / 2 m M E D T A / 0 . 5 m M E G T A (henceforth referred to as 'control nuclei') or in this buffer containing additionally 0.135 M NaCI (' treated nuclei'). The final concentrations of NaCI in the two buffers were 0.015 M and 0.15 M, respectively. The total ionic strengths were thus around 0.08 M and 0.22 M. After the incubation, nuclei were collected by centrifugation at 4°C. The pellet was washed with isolation buffer (buffer A / 2 m M E D T A / 0 . 5 m M EGTA) and dispersed in buffer A / 2 m M CaC12 at the same nuclear concentration. These were next used for digestion with micrococcal endonuclease and chromatin isolation. Digestion of the nuclei and preparation of soluble chromatin Following incubation in the respective buffers, control and treated nuclei were digested with micrococcal endonuclease (EC 3.1.31.1, Worthington, Calbiochem) at a concentration of 1.5 u n i t s / m l for various times at 37°C. The method of extraction of soluble chromatin was the same as that reported earlier [7]. The fraction of chromatin solubilised was obtained as the ratio of A260 of soluble chromatin to the A260 of the initial nuclear suspension after correction for the hyperchromicity of nucleic acid in 0.1 M N a O H . DNA gel electrophoresis Estimation of the molecular weight of soluble chromatin by D N A gel electrophoresis was performed as in Ref. 7. Protein gel electrophoresis Proteins of the whole nuclei, those extracted in buffer A, with or without 0.135 M NaC1 and those associated with the released chromatin were analysed on sodium dodecyl sulphate polyacrylamide gels following the method of Laemmli [10] as modified by Thomas and Kornberg [11]. All the

samples were adjusted to a buffer consisting of 0.025 M Tris-HC1 (pH 6.8)/2% S D S / 1 5 % g l y c e r o l / 2.5% f l - m e r c a p t o e t h a n o l / 0 . 0 0 1 % Bromophenol blue and heated for 3 min at 95°C before loading on the gels. The gels were stained with 0.05% Coomassie blue in m e t h a n o l / w a t e r / acetic acid mixture (40 : 50 : 10, v / v ) overnight and destained in the same buffer without dye. Gels were stored in 7.5% acetic acid. Molecular weight markers included in each run were albumin (66000), pepsin (34700), trypsinogen (24000), lactoglobulin (18400) and lysozyme (14300). The distances of migration of the different protein components were converted into a molecular weight scale with the help of these markers. The gels were scanned at 570 nm using a gel-scanner attached to a Hitachi double-beam spectrophotometer (model 200-20). Results

Time-course of chromatin solubilisation Fig. 1 shows the time-course of chromatin solubilisation from chicken erythrocyte nuclei following micrococcal endonuclease (1.5 U / m l ) dio i.d o9 ._J nn :D

-J 0

i

z

I

80

i

z

0 IX: 1-

(j,

40

U-

o Iz w (..) Ix

,,,

Q_

I

0 5

lO TIME

15

20

(rain.)

Fig. 1. Time-course of chromatin solubilisation from control and treated nuclei following micrococcal endonuclease digestion (1.5 U / 1 0 0 A260 units of lysed nuclei) at 37°C. Curve (a) is for control nuclei not preincubated (e) or preincubated at 0°C in buffer A / 2 m M E D T A / 0 . 5 m M EGTA for 1 h (A) or overnight (11). Curves (b) and (c) are for treated nuclei preincubated in the above buffer plus 0.135 M NaCI for 1 h and overnight, respectively. Percent of chromatin solubilised was calculated as described in Materials and Methods.

417

gestion at 37°C. Curve (a) represents chromatin released from control nuclei, while (b) and (c) represent solubilisation from nuclei preincubated in the higher ionic strength buffer for 1 h and overnight, respectively. The rates of solubilisation are different for the three samples. For control nuclei, the solubilisation steadily increases with time of digestion. In case of treated nuclei, maximum solubilisation is only 55% (1 h) or 35% (overnight). In contrast, predigestion incubation of nuclei in standard isolation buffer for different time intervals had no effect on the rate of solu-

18.

1 I 150 65 40 20 10 2.9 MOLECULAR WEIGHT (kbp)

130 65 40 20 10 2~9 MOLECULAR WEIGHT (kbp)

Fig. 2. Microdensitometric traces of the gel negatives for the D N A of the soluble chromatin obtained by micrococcal nuclease digestion of chicken erythrocyte nuclei, exposed to buffer A / 2 m M E D T A / 0 . 5 m M E G T A plus 0.135 M NaCI for 1 h at 0°C. The numbers indicate the degree of solubilisation of the sample electrophoresed. The number-average molecular weights of each individual population (see text) are indicated by dots on the curves. The lowest trace shows D N A gel scans of the marker substances h D N A (49 kbp) and BamHI-cut fragments of the same (33, 26.5, 15.2, 13.5, 13.3, 13, 12 kbp). Fig. 3. Microdensitometric scans of the gel negatives for D N A obtained from the soluble chromatin released by micrococcal nuclease digestion of chicken erythrocyte nuclei exposed to buffer A / 2 m M E D T A / 0 . 5 m M E G T A plus 0.135 M NaCI overnight at 0°C. The numbers and dots have the same meaning as in Fig. 2. Molecular weight markers were the same as in Fig. 2.

bilisation or the maximum attainable solubilisation at this nuclease concentration. The three symbols on curve (a) represent different times of preincubation in the standard isolation buffer [9]. There is a decreasing trend in the solubilisation curves (b) and (c) after the maximum value has been obtained. Such a change could be observed for control nuclei only after prolonged digestion.

DNA gel electrophoresis The pattern of molecular weight distribution of soluble chromatin was analysed by D N A gel electrophoresis on 0.7% agarose gels. Figs. 2 and 3 are the densitometric traces of the D N A gel negatives of chromatin released from nuclei preincubated in higher ionic strength buffer either for 1 h or overnight. The chromatin samples were taken from the linear part of curves (b) and (c) in Fig. 1. The number against each trace marks the degree of solubilisation given as a percentage of the total D N A in the starting nuclear suspension. D N A gel scans for chromatin released from control nuclei have been published earlier [7]. It can be seen that there is a broad distribution of molecular weights at each time point, the broadness increasing with progress of digestion. 1 h high ionic strength treatment gives rise to a size distribution that often implied the presence of two different overlapping populations; this population heterogeneity is less discernible at longer times of digestion. The number-average molecular weights of the different samples were determined from the microdensitometric traces using the relation:

M,

Y~gi/Mi

where gi is the fluorescent intensity corresponding to a molecular weight M i. In case of Fig. 2, where the presence of two different populations was distinct, a vertical line was drawn up to the abscissa, and the separated regions were used for the computation of the molecular weight of each presumably independent population. This approach is subject to error due to spillover of one distribution into the other, but the magnitude of the error will not be large enough to overrule the conclusions to be drawn from the results of such calculations.

418

The number-average molecular weights are indicated by dots on each curve. The value decreases as the degree of solubilisation increases. The spectrophotometric data for the percent of solubilisation (see Materials and Methods) and the number average molecular weight calculated for different points of solubilisation can now be extrapolated to fit with the theoretical model proposed by Igo-Kemenes and Zachau [4] and the domain length of chromatin under the different conditions (control or treated) can be determined. The domain length for control nuclei was estimated earlier [7]. Repeated experimentation corrects this value to 45 + 5 kbp. This length remains unaltered even on overnight incubation in standard isolation buffer. Calculation of domain length demands the knowledge of the percent solubilisation corresponding to each number average molecular weight. For 1 h incubation (Fig. 2), this parameter for each subpopulation was obtained by multiplying the experimental value of the percent solubilisation by the proportion of fluorescent intensity contributed by each segment of the curve to the total intensity.

Fig. 4 gives the plot of number-average molecular weight versus the percent of solubilisation. The solid lines are theoretical curves computer-generated for domain lengths of 150 (a), 100 (b) and 34 kbp (c). These lengths correspond respectively to domains in nuclei treated with NaC1 overnight, and two subdomains in nuclei treated for 1 h.

49

65

72

(c)

46

13C

14-,5

jO[

E

391157~j 5A ~3g

¢_

76105

l,m

JD

2

t..

13 v

E

r~ lOO " bd (#) .J

E

I

I

I

I

._1 0 Z

~ so 0 T 0 h 0 pZ t~ 0IT O.

I0 0

i

20 40 60 80 NUMBER AVERAGE MOLECULAR WT. OF SOLUBLE CHROMATIN( l ~ t) kbp

Fig. 4. Dependence of the number-average molecular weight of soluble chromatin at any degree of solubilisation on the domain length. The solid lines are theoretical curves, computer-generated for domain lengths of 150 kbp (a), 100 kbp (b), and 34 kbp (c). The d o s e d circles, open circles and triangles indicate the experimentally obtained number-average molecular weights. These refer to overnight-treated samples and the two populations (see text) obtained after 1-h treatment of the nuclei with high ionic strength buffer.

14 21 26",534 44 60 80 MOLECULAR

I00

W E I G H T ( x I 0 -3 )

Fig. 5. Scans of the extracted proteins on a SDS-polyacrylamide gel. Nuclear suspension at a concentration of 100 A26o u n i t s / m l was incubated at 0°C either in buffer A / 2 m M E D T A / 0 . 5 m M E G T A overnight (scan a) or in the same buffer plus 0.135 M NaC1 for 1 h (scan b) or overnight (scan c). Proteins contained in 0.25 ml of the extract was precipitated by acetone and collected by centrifugation. The precipitates were dissolved in 0.025 M Tris-HCl (pH 6.8)/2% SDS/5% glycerol/ 2.5% fl-mercaptoethanol (sample buffer) and approx. 10-15/~g protein loaded on the gels, after heating for 3 min at 95°C. These gels were run with an applied voltage of 180 V and a constant current of 32 m A for 5 h.

419

SDS-polyacrylamide gel electrophoresis of proteins Fig. 5 shows the gel scans of proteins extracted from the nuclei during incubation at 0°C in the control and high ionic strength buffer either overnight (scans a and c) or for 1 h (scan b). It is seen that inctibation in control buffer A alone extracts many proteins (scan a), but the overall quantity is much less than that extracted in the above buffer containing additional 0.135 M NaC1 (scan b and c). The extraction of proteins in latter buffer is dependent on the time of incubation both in quantity and in quality. Some of the peaks (14.5, 49, 58, 76 kDa) remaining unresolved after I h incubation (scan b) become very pronounced on incubation overnight (scan c). The peak height of some proteins (e.g., 27, 39, 130 kDa) is not altered in these two curves. These possibly represent proteins that

Cb)

Y . ~

are loosely bound, and a brief exposure to added NaC1 is sufficient for their release. The gel patterns of the whole nuclear proteins of the various control and treated samples were not much different (data not shown). In Fig. 6, comparison was made between the proteins associated with chromatins released after micrococcal nuclease digestion of the control and treated nuclei. The patterns represent samples after overnight treatment only. It has to be mentioned here that the gel picture of the proteins associated with chromatin released from control nuclei is independent of the time of keeping the nuclei in the isolation buffer prior to nuclease digestion. This implies absence of proteolysis. This was also expected, as the proteolysis inhibitor PMSF (phenylmethylsulphonylfluoride) was present in all the buffers. For treated nuclei, there is an incubationtime-dependent change in this pattern. The overnight pattern represents the equilibrium situation. From the scans it can be seen that the content of nonhistone proteins (41-72 kDa) is remarkably reduced in the treated nuclei (scan b) compared to that in control (scan a). Also the proportion of H 1 and H s relative to the core histone (marked H , ) was much reduced in the treated sample. Discussion

I

"

f I 14.0

21

31 3 6 4 4

60

80

I00

MOLECULAR WEIGHT (xlO -3) Fig. 6. Scans of the chromatin-associated proteins on SDSpolyacrylamidegels. Curves(a) and (b) refer to the chromatins obtained from control and treated nuclei, respectively. The time of incubationwas overnightfor both the samples. Samples containing about 0.5 A2~o units were loaded directly on the gels after mixing with twice-concentrated sample buffer as described in the legend to Fig. 5.

The results of the present experiments show that (i) the time-course of solubilisation of chromatin from the nuclei and (ii) the molecular weight distribution of the D N A is soluble chromatin, both change on predigestion incubation of the nuclei in higher ionic strength buffer. These features suggest an alteration of the domain architecture of chromatin present in the control nuclei. The 45 kbp length of the domain observed for control nuclei [7] changes to 150 kbp length on overnight incubation of the nuclei in the higher ionic environment. I h incubation produces a metastable state consisting of two populations of chromatin fragments that correspond to domain lengths of 100 and 34 kbp. Also, chromatin isolated from treated nuclei contains fewer nonhistone chromosomal bands and the quantity of H 5 and H 1 associated with it is much reduced compared to those in the control at identical level of solubilisation (Fig. 6, scans a and b). The above

420 results provide a picture of the higher-order structure of chromatin in the light of the domain model. Since a 150 kbp length corresponds to a domain that is uniform (see Results section) and was exposed after overnight treatment with NaCI, this can be looked upon as a basic structural unit under the conditions of the present experiments. In control nuclei - that is, in a lower ionic environment - this long domain is constrained by several proteins, extractable under the higher ionic strength condition (Fig. 5, scans b and c), into three subdomains of approximate lengths 45 kbp each. The transition from 45 to 150 kbp can result from the gradual elimination of the constraint between two neighbouring subdomains, which at some point during reorganisation may yield a distribution of domains of length 90 kbp (complete removal of the constraint between the two neighbours) and 45 kbp (incomplete removal of the constraint). These values are not very far from experimental values of 100 kbp and 34 kbp (Fig. 4), considering the crudeness of the domain model. From the protein gel study we have three related pieces of information: (a) the content of H 1 / H 5 bound to chromatin released from overnight treated nuclei after micrococcal nuclease digestion was less than that in corresponding control; (b) none of the H ~ / H 5 was extracted from the whole nuclei during incubation in either buffer; and (c) it was found that the proportion of H 1 / H 5 in the insoluble nuclear pellet obtained after chromatin solubilisation was higher in treated nuclei compared to that in control (scan not given). Taken together, these data suggest a change in the binding locus of the histones H J H 5 during the higher ionic treatment. It has been observed that the histones H ~ / H 5 preferentially bind to and stabilise compact superhelical structures [12-14]. Thus, relaxation of the chromatin structure during NaC1 treatment may be associated with sliding of the H I / H 5 to particular regions of the long chromatin. Due to the presence of additional H ~ / H 5, these regions remain superhelical and condensed. This condensation might be the reason why not more than 55% or 35% of chromatin could be solubilised when the nuclei were treated with high ionic buffer. All the above discussions are based on the assumption that the domain model is applicable in

the present case, even when there is rearrangement of the structure due to the dissociation of the proteins. However, the domain model is a qualitative model and helps to depict a particular level of organisation on the basis of a parameter which is a length. Taking 60% as the maximum nucleaseaccessible zone under the present experimental conditions, the lengths of the chromatin domains in the 1-h or overnight-treated nuclei could be modified, but that would not alter the general picture proposed. Moreover, the domain lengths as determined in the present paper are not absolute. In repeated experiments, it was found that the absolute values of these lengths varied widely. But what is important is that the general scheme was valid everywhere - the length of domains in overnight-treated nuclei was always around 2- to 3times the control value. Cook and Brazell [17] observed that the sedimentation coefficient of D N A isolated from chicken erythrocyte nuclei after 1 M NaC1 treatment was not that characteristic of supercoiled DNA, fixed at the ends. However, we observed that under our experimental conditoins the micrococcal nuclease digestion pattern can be fitted with a domain model [7]. These two data can be simultaneously satisfied if the constraints are assumed to be relatively free, that is, not attached to any extrachromosomal supporting structure. The present work extends the idea and shows that some of the constraints (i.e., between two subdomains) can be removed or made to slide along the length of the chromatin with 0.135 M NaC1 treatment.

Acknowledgements One of the authors (A.G.) is indebted to Indian Council of Medical Research, Government of India, for a research fellowship. Particular thanks are due to Dr. Bharati Ghosh of Bose Institute, Calcutta for constant cooperation and allowing us to use the gel scanner attached to Hitachi doublebeam spectrophotometer. Thanks are due to Mr. D. Chatterjee of the same Institute for giving us B a m H I - c u t fragments of lambda DNA.

References 1 Worcel,A. and Benyajati, C. (1971) Cell 12, 83-100 2 Mcghee, J.D. and Felsenfeld, G. (1980) Annu. Rev. Biochem. 49, 1115-56

421 3 Igo-Kemenes, T., Horz, W. and Zachau, H.G. (1982) Annu. Rev. Biochem. 51, 89-121 4 Igo-Kemenes, T. and Zachau, H.G. (1977) Cold Spring Harbour Symp. Quant. Biol. 42, 109-118 5 Benyajati, C. and Worcel, A. (1976) Cell 9, 393-407 6 Comings, D.E. (1978) in The Cell Nucleus, Vol. IV (Busch, H., ed.), pp. 345-368, Academic Press, New York 7 Ganguly, A., Bagchi, B., Bera, M., Ghosh, A.N. and Sen, A. (1983) Biochim. Biophys. Acta 739, 286-290 8 Hewish, D.R. and Burgoyne, L.A. (1973) Biochem. Biophys. Res. Commun. 52, 504-510 9 Wilhelm, F.X., Wilhelm, M.L., Erard, M.P. and Daune, M.P. (1978) Nucleic Acids Res. 5, 505-521

10 Laemmli, U.K. (1970) Nature 227, 680-685 11 Thomas, J. and Kornberg, r.D. (1975) Proc. Natl. Acad. Sci. USA 72, 2626-2630 12 Vogel, T. and Singer, M.F. (1976) J. Biol. Chem. 251, 2334 13 Singer, D.S. and Singer, M.F. (1976) Nucleic Acids Res. 3, 2531 14 Bauer, W.R. (1978) Annu. Rev. Biophys. Bioeng. 7, 287-313 15 Ruiz-Carrillo, A., Puigdomenech, P., Eder, G. and Lurz, R. (1980) Biochemistry 19, 2544-2554 16 Shaw, B.R., Herman, T.M., Kovacic, R.T., Beaudreau, G.S. and Van Holde, K.E. (1976) Proc. Natl. Acad. Sci. USA 73, 505-509 17 Cook, P.R. and Brazell, I.A. (1976) J. Cell Sci. 22, 287-306