- Email: [email protected]
Disulfide bonds and the structure of the chromatin complex in relation to aging
Mechanisms of Ageing and Development, 12 (1980) 65-80
65
©Elsevier Sequoia S.A., Lausanne - Printed in the Netherlands
DISULFIDE BONDS AND THE STRUCTURE C O M P L E X IN R E L A T I O N T O A G I N G
OF THE CHROMATIN
SINAN TAS, CHICK F. TAM and ROY L. WALFORD Department of Pathology, Center for Health Sciences, University of California, Los Angeles, CA 90024 (U.S.A.)
(Received May 18, 1979; in revised form August 6, 1979) SUMMARY Triton X-100 washed nuclei from livers of young-adult and old mice were digested with micrococcal nuclease and pelleted. Supernatants (1SF) were saved and the pellets lysed in a hypotonic EDTA buffer. A second supernatant (2SF) and a final pellet (P) were obtained by recentrifugation (7000 g, 7 minutes). The 1SF/2SF ratio, which has been shown to be an index of the transcriptionally active to inactive chromatin ratio, was lower in older mice. The fraction relatively resistant to solubilization by the nuclease (P) was found by isopycnic sucrose gradient centrifugation to be in a more compact, condensed state when prepared from older mice. Higher amounts of heavy density chromatin were obtained from nuclei of old than young mice by hypotonic lysis plus minimal mechanical shearing. 2-Mercaptoethanol (2ME) treatment brought the density of the material of P from old mice back to the levels of young mice. In both age groups 2ME decreased the densities of mechanically sheared chromatin as well as of the whole Triton X-IO0 washed nuclei. In nuclease digestion experiments treatment of the nuclei from both age groups with S-S reducing agents increased the release of DNA from P into the supernatants. The results are consistent with S-S bonds being involved in the condensed structure of chromatin in young and old mice and in the shift of the chromatin complex towards a more compact, condensed state in old age.
INTRODUCTION While an initial level of organization ofchromatin involves association of histones and DNA in nucleosomes [ 1], the nature of higher levels of organization is poorly understood. It is known, however, that accessibility of DNA in the chromatin complex is highly restricted. This restriction, which in part may be due to supranucleosomal organizational features, seems to become more pronounced, at least in some cases, with progressive cellular differentiation and is accompanied by a more compact, condensed chromatin structure [2, 3]. The condensed structure, detected mainly by cytological procedures, may be observed in identical regions of chromatin in all cells of an organism (constitutive heterochromatin) as well as in different regions in different cell types (facultative hetero-
66 chromatin) [4, 5]. Facultative heterochromatinization is considered to be a mechanism for gene inactivation operating during cellular differentiation, and previous evidence suggested a role for disulfide bonds in maintaining the condensed structure of facultative heterochromatin in interphase cells [6]. Several studies have suggested that chromatin of both postmitotic and proliferative cell populations of old animals may exist in a more compact, condensed state than chromatin from similar cells of young animals, and that the amount of freely accessible DNA in chromatin decreases with age [7-14]. Evidence presented elsewhere was not inconsistent with the idea that disulfide bonds may be involved in this age-related condensation, and that heterochromatinization of the genome may be an important feature of cellular aging [6, 11 ]. Recent investigations have provided evidence that genetically active chromatin regions in nuclei can be preferentially attacked by micrococcal nuclease under mild digestion conditions, and that the sequences for active genes in a digestion product are enriched in mononucleosomal-length DNA while depleted in oligonucleosomal-length DNA [15, 16]. Based in part on this preferential digestibility of active genes by micrococcal nuclease, Bloom and Anderson [ 15] recently described a procedure for separating nuclear DNA into three fractions: one enriched in active genes, one enriched in inactive genes, and a fraction relatively resistant to solubilization by the enzyme [15]. In the present study we applied this fractionation procedure to Triton X-100 washed nuclei of young-adult and old mice in the presence or absence of disulfide reducing agents. The chromatin species fractionated by micrococcal nuclease were then subjected to isopycnic and isokinetic sucrose-gradient centrifugations, with and without prior treatment with disulfide reducing agents. Results of these experiments showed that in both young-adult and old mice treatment of hepatic nuclei with S-S reducing agents significantly increased the release of DNA from the fraction relatively resistant to solubilization by micrococcal nuclease. We noted that this fraction assumed a more compact, higher density state in old age, and that treatment with 2-mercaptoethanol (2ME) prior to centrifugation brought the density of the fraction back to that characteristic of young animals. The densities of other chromatin fractions did not appear to be influenced by age or S-S reducing agents. The decondensing effect of S-S reducing agents on the chromatin complex was also demonstrable with the Triton X-IO0 washed nuclei of old and to a lesser degree of young mice, and with the denser fraction of mechanically sheared chromatin in both age groups. Finally, the active/inactive chromatin ratio measured by the micrococcal nuclease digestion procedure [15] appeared to decrease significantly with increasing age. MATERIAL AND METHODS Animals
Male mice of the long-lived (C57BL]6 X Balb/c)Fl hybrid were used. The youngadults were 7-11 months old and the aged animals 22-27 months old at the time of sacrifice. They were fed ad libitum with standard laboratory chow.
67
Preparation of nuclei Nuclei were prepared according to the method of Widnell and Tata [17], adjusted to a smaller amount of tissue. Animals were killed by decapitation between 8.30 and 10.30 a.m., their livers immediately taken into ice-cold homogenization medium, weighed, and equal amounts from two or (in the micrococcal nuclease digestion experiments) four animals of each age group pooled for use as starting material. Animals with enlarged or abnormally appearing livers or with gross evidence of disease elsewhere were excluded from the study and a healthy animal substituted. The protease inhibitor phenylmethylsulfonylfluoride (PMSF) in a concentration of 1 mM was included in the initial homogenization medium and at 0.2 mM in later media. Under the conditions of our experiments, PMSF was shown by means of the 5-5'-dithiobis-2-nitrobenzoic acid reaction [18] not to alter the free -SH content of chromatin.
Digestion of nuclei with micrococcal nuclease Freshly purified nuclei [17] were washed with 15 ml of 1% Triton X-100, 0.5 M sucrose, 3 mM MgC12, 0.2 mM PMSF, 10 mM Tris-HC1 (pH 7.4) for 10 minutes at ice-cold temperature and centrifuged at 150 g for 7 minutes in a round-bottom 25 mm diameter tube. The resulting loose nuclear pellets were resuspended in 0.35 M sucrose, 25 mM KCI, 5 mM MgC12, 0.5 mM CaCI2, 0.1 mM PMSF, 10 mM Tris (pH 7.4) and adjusted for both age groups to a concentration of A260 = 4.5--6.0. Two-ml aliquots were then incubated at 4 °C for 40 minutes in the presence or absence of 20 mM of either 2ME or dithiothreitol (DTT), warmed to 37 °C for 2 minutes, and incubated at 37 °C with micrococcal nuclease at concentrations and time periods indicated below (see Results). Following digestion, the aliquots were chilled in ice and centrifuged at 7000 g for 7 minutes at 4 °C. In every experiment separate 2.0 ml aliquots from both age groups and from 2ME treated and untreated groups were centrifuged as above but without prior incubation at 37 °C. These aliquots did not contain micrococcal nuclease and were used to determine the zero-minute values. The resulting first supernatant fractions (1SF)were set aside and the pellets lysed by suspension in 2.0 ml of 0.1 mM PMSF, 5 mM EDTA (pH 7.0) for 10 minutes. The lysed suspensions were centrifuged at 7000 g for 7 minutes at 4 °C to yield a second supernatant (2SF) and a final pellet (P) fraction. The pellet fractions were resuspended by a pasteur pipette in 2.0 ml of 0.1 mM PMSF, 5 mM EDTA (pH 7.0). In every experiment the buffers described above (with and without 2ME or nuclease) were run in parallel with the nuclear suspensions to provide controls for the chemical tests and blanks for the absorbancy measurements. Preparations of chromatin samples for the isopycnic sucrose gradient centrifugations were made by using 5.5 ml nuclear suspensions of A~o = 15.0 for each age group. Enzyme concentrations and digestion times were as described (see Results).
Preparation of mechanically sheared chromatin A mo.dification of the method described by Monahan and Hall [19] was used. Freshly purified nuclei [17] were washed twice for 15 minutes each with 15 ml of 1% Triton X-IO0, 0.25 M sucrose, 3 mM MgC12, 0.2 mM PMSF and 10 mM Tris-HCI (pH 7.9)
68 and centrifuged at 1000 g. The nuclear pellet was resuspended by vortexing in 15 ml of 0.14 M NaC1, 1 mM MgC12, 0.2 mM PMSF and 5 mM Tris-HC1 (pH 7.9). The suspension was allowed to stand in ice for I0 minutes, then centrifuged at 630 g for 5 minutes. This step was repeated and fluid was carefully drained from the pellet. Next, 10 ml of a solution consisting of 2 mM EDTA, 0.1 mM PMSF and 5 mM Tris (pH 7.9)were added to the nuclear pellet. The pellet was disrupted by three strokes in a tightly fitting glass-to-glass homogenizer. With maximum hand pressure and as quickly as possible the disrupted material was passed three times through 23, 27 and 30 gauge needles. The final solution was centrifuged at 2500 g for 10 minutes to pellet unbroken nuclei, nucleoli and any aggregates of chromatin which may have formed [ 19]. The supernatants (mechanically sheared chromatin) from both age groups were removed with a pasteur pipette and diluted with the 0.1 mM PMSF, 2 mM EDTA, 5 mM Tris (pH 7.9) lysis solution to equal concentrations of about A 260 = 1.5.
Isopycnic and isokinetic sucrose-gradient centrifugation Ice-cold sucrose solutions were prepared in the buffers used for the solubilization or suspension of the chromatin fractions or nuclei to be applied to the gradients. Linear sucrose gradients as indicated in the figures were generated by a Buchler density gradient engine (Buchler Instruments, New Jersey). For sedimentation rate experiments linear gradients from 23 to 53% sucrose were prepared with a 2.0 ml cushion of 63% sucrose at the bottom of the tubes. Two.ml samples of chromatin solutions or of nuclear suspensions at equal A2~ values (1.0-1.8, but the same in each experiment) for both age groups were layered on top of the gradients (31 ml) in 25 mm diameter nitrocellulose centrifuge tubes. In two additional tubes the layered samples from the two age groups were pretreated with 20 mM 2ME at 4 °C for 30--90 minutes in sealed tubes prior to layering. Two-ml samples of the buffer, or buffer containing 20 mM 2ME, were also layered on top of the gradients in separate tubes. The tubes were centrifuged at 18000 rpm for 48 hours at 4 °C (average g = 42 820) in a Beckman-Spinco SW 27 rotor. Equilibrium was reached by 36 hours and the position of peaks did not change between the 36th and 48th hours. Isokinetic sucrose-gradient centrifugation of the 2ME treated and untreated nuclei was done at 5000 rpm for a total of 11 minutes with an acceleration time of 2 minutes. The rotor (SW 27) made a total of about 53 000 revolutions. The tubes were punctured and 1.8-ml fractions collected (1.65-ml fractions for the sedimentation rate measurements). Absorbance of the fractions at 260,280 and 320 nm was determined in 1 cm path-length quartz spectrophotometry cells. Refractive indices in the fractions at 20 °C were measured with a refractometer (Abbe-3L, Bausch and Lomb, New York) and the corresponding weight percentages of sucrose determined. Absorption curves for the buffer and 20 mM 2ME in the buffer were drawn in order to subtract the corresponding absorbance values from those of the investigated samples. (Because 2ME absorbs at the above wavelengths, this subtraction step is essential in order not to overestimate the absorbance on top of the gradients.) Recoveries of the total loaded A~o in the gradients varied from 92 to 103%. In all experiments freshly prepared chromatin samples or nuclear suspensions were used.
69
Protein, R N A and D N A measurements Protein was determined according to the method of Lowry et al. [20] with bovine serum albumin as standard; RNA was measured as described by Ceriotti [ 2 1 ] . DNA, extracted by a final concentration of 1.5 N of HC104 at 70 °C for 20 minutes, was measured according to the description of Richards [22]. The buffers used for solubilization or suspension of the chromatin fractions or nuclei to be tested by the above procedures were run in parallel in each test as controls. DNA in the nuclei was measured in 0.1 mM PMSF, 5 mM EDTA (pH 7.0). (Tests indicated that prompt centrifugation o f the nuclear suspension in the digestion mixture prevented loss of any DNA into the supernatant.), In micrococcal nuclease digestion experiments estimation o f DNA in ISF was satisfactory either (1) when in a direct estimate the error introduced by interference of 2ME and the I)uffer was corrected by appropriate controls, or (2) when the DNA values were calculated indirectly by subtracting the DNA of 2SF + P from that of the nuclei suspension used for digestion and closely correlated with the A260 values.
o(ISF) 25
f ~,,...~
/~;....;;;'~ "
c{P)
20.
50-
~ 15 ¸
45-
~ ' 10
40-
354
5' Q
O 2
I
65-
~
5 t0 Digestion time (minutes) D(2SF)
'
zs-
I
g.
f
E "6
~ 55I g_
'it
5O
45
o 2
30-"
20
~ tb Digestion time (minutes)
20
o z
3 tb Digestion time (minutes)
20
Fig. la-c. Time course of fractionation by micrococcal nuclease digestion of the Triton X-T00 washed and 2-mercaptoethanol treated or untreated nuclei from livers of young-adult and old mice. Micrococcal nuclease concentrations were 8 units to 2 ml of nuclear suspension. Each point represents the mean of duplicate samples. • o, old; O O, young-adult; A...A, old treated with 2ME; zx...~ young-adult'treated with 2ME. It is seen that 2ME releases an increased amount of DNA from P into supernatants in both age gxoups, flaat this release occurs mainly with the zero-minute nuclei and mainly into 2SF, and that for ISF, young > old, and for 2SF, young < old,
70
RESULTS
Micrococcal nuclease digestion of the Triton X-I O0 washed nuclei Results of these experiments can be outlined as follows: (1) The amount of nuclear DNA recovered in 1SF showed a steep increase within the first few minutes and kept increasing, albeit at a slower rate, with increasing digestion time, while that recovered in 2SF declined with increasing time of digestion after an initial steep increase during the first minute (Figs. 1 a, b and 2a, b). The level of 2SF was influenced both by its rate of release from P and also by its rate of conversion to 1SF. Figures lb, c and 2b, c show that the main event in the 0-1-minute interval was the release of the material of 2SF from P. A nearly linear decrease was observed in the level of 2SF during the 2-10-minute digestion period, apparently due mainly to conversion to 1SF material (Figs. la, b, c and 2a, b, c). We noted that the amount of nuclear DNA released into 1SF was higher in young-adult than in old mice at all times of digestion and the reverse was true for 2SF (Figs. 1 a, b and 2a, b). This relationship could be expressed by
o(lSF) 30
I ~ / ~ c(P)
I 55~ I
50-
~ t53 c ~ tO-
45-
5-
35-
30-
o z
~,
~b
Digestion time (minutes)
20 25-
20-
15.
5.'J I =
10-
"6 5O 5-
45 2
5 tO Digestion time (minutes)
20
0
2
5 tO Digestion time (minutes)
20
Fig. 2a--¢. Same experiment as in Fig. la-c but with 50 units of microcoecal nuclease to 2 ml of nuclear suspension, o ~ e , old; O O, young-adult; A...A, old treated with 2ME;/x.../x, youngadult treated with 2ME.
71 Ratio A = (amoun_____._~tof DNA recovered in 1 S F ) o l d ] / l ( a m o u n t of DNA recovered in 1SF)young]
l
/
- - I / /
(amount of DNA recovered in 2SF) o l d ] / [(amount o f DNA recovered in 2SF) youngj
Since the value o f this ratio appeared to be independent of the digestion periods within the 1-20-minute range (see Figs. la, b and 2a, b) the 1st, 2nd, 5th, 10th and 20th minute values were assumed to form a single population and their mean as well as standard error o f the mean were calculated for each experiment (Table I). From Table I the overall mean of ratio A in the seven separate experiments without 2ME and involving 28 young-adult and 28 old mice was 0.906. This was significantly less than 1 (p < 0.005). (2) The amount of DNA recovered in P showed a steep decline within the first minute and a slow decline thereafter (Figs. lc and 2c), With higher concentrations o f enzyme (Fig. 2c versus Fig. 1c) the amounts of nuclear DNA recovered in P were lower at the first minute, but the differences related to enzyme concentration tended to become smaller through the end of the 20-minute digestion period. Amounts of DNA recovered in P did not significantly differ between the two age groups within the 8 - 1 3 0 units of micrococcal nuclease to 2.0 ml nuclear suspension range (not shown). (3) With low concentrations of enzyme (Fig. la) the amount of DNA recovered in 1SF increased with 2ME over the values for 1SF in the absence of 2ME. With higher enzyme concentrations (Fig, 2a) this effect became less pronounced, and could not be detected at all by an enzyme concentration of 130 units to 2.0 ml nuclei suspension of TABLE I RATIO A, COMPARING THE AMOUNTS OF NUCLEAR DNA RECOVERED IN THE FIRST AND SECOND SUPERNATANTS BY MICROCOCCAL NUCLEASE DIGESTION OF LIVER NUCLEI IN YOUNG-ADULT AND OLD MICE, WITHOUT AND WITH ADDED 2ME Units o f micrococcal nuclease added to 2.0 ml o f nuclei suspension
8 8 50 50 50** 85** 130"**
A26o o f the nuclei suspension
4.5 4.5 4.5 4.5 6.0 6.0 5.5
Mean o f ratio A * for 1, 2, 5, 10 and 20 minute digestions within each experiment ± S.E.M. without 2ME
with 2ME
0.883 0.904 0.892 0.910 0.905 0.912 0.933
1.007 1.017 0.854 0.903 0.879 0.928 0.924
± 0.022t ± 0.029* ± 0.026t ± 0.028* ± 0.019'
-+0.039 ± 0.025 ± 0.014t ± 0.022t ± 0.047
*Ratio A, see text for definition. Duplicate samples were used for each digestion period in all experiments. **Means of the 2rid, and 20th minute digestion values. Other digestion periods were not studied. ***Means of 1', 2, 5 and 10 minute digestion values. tRatio of significantly less than 1 (p < 0.005). *Ratio is significantly less than 1 (p < 0.025).
72 A26o = 5.5 (not shown). 2ME treatment of the nuclei also enhanced the release of the material of 2SF from P and this effect was seen most dramatically with the zero-time nuclei (i.e. before incubation at 37 °C and without nuclease) (Figs. lb, c and 2b, c). The ratio (in terms of DNA) Ratio B =
[(ISF + 2SF)/P] with 2ME [(1SF + 2SF)/P] without 2ME
accordingly was significantly larger than unity for both age groups and approached unity with increasing digestion time (Fig. 3). The variability of the value of ratio B in seven separate experiments using 8-130 units of nuclease to 2.0 ml nuclei suspension is indicated in Fig. 3.
1+i+
0 2 5 10 20 Digestion time(minutes)
Fig. 3. Value of ratio B as a function of the digestion period. Ratio B, an index of the 2ME-induced DNA release from P, is defined as (1SF + 2SF)/P with 2ME/[(ISF + 2SF)/P],without 2ME. It is seen that the amount of DNA released by 2ME into the supernatants 1SF + 2SF decreases with increasing digestion time. Data were derived from the same set of experiments as in Table I. Bars represent the standard errors of the means of experiments using 8--130 units of nuelease. O - - O , old; O - - O , young-adult. (4) When digestions were carried out in the presence of 2ME, ratio A remained significantly smaller than unity with enzyme concentrations of 50 units or more to 2.0 ml nuclei suspensions (Table I, p < 0.005). With 8 units of nuclease, however, ratio A was increased by treatment with 2ME in two separate experiments (Table 1). When the Triton X-IO0 washed nuclei that had been incubated in the absence of 2ME for 40 minutes at 4 °C were centrifuged at 7000 g for 7 minutes, only a small amount of nuclear DNA (not exceeding 2-3% and usually much less) was released into the supernatant (1SF). Reliable comparisons of ISF between young and old mice in these experiments were therefore not possible by the chemical test used. (In a limited number of experiments we incubated nuclei in the digestion buffer without micrococcal nuelease at 37 °C. Such experiments suggested that as much as 13% of nuclear DNA could be released into the 1SF within 20 minutes, probably as a result of the endogenous Mg2÷, Ca2+-activated nucleolytic activity.) In earlier experiments the nuclei for micrococcal nuclease digestion were prepared as described by Bloom and Anderson [15]. Although such nuclear preparations gave
73 similar 2ME effects and age-related decline of the 1SF/2SF ratio to those described above, microscopic examination revealed clumps which were difficult to separate into individual nuclei. Because of possible ambiguities these clumps might cause in comparing the amounts of DNA recovered in P, we employed the low-speed centrifugation procedure described in Methods. This gave pure, light-microscopically intact nuclear suspensions virtually free of clumps. Also, we used I% instead of 0.2% Triton X-IO0 in order to remove nuclear membranes completely. Although these modifications of the original procedure [15 ] enhanced the releasability of DNA into 1SF by the micrococcal nuclease, the cause of the nearly complete release of the material of 2SF within the first minute of digestion in our experiments, as opposed to increasing release of 2SF with increasing time of digestion of the nuclei from chick cells in the original study [15], is not known. It was not possible to carry out the micrococcal nuclease digestion experiments for prolonged periods because the material of 2SF started to aggregate after 40 minutes under the described conditions and with or without 2ME in the digestion buffer (not shown). All of the 2ME effects noted above were reproducible with 20 mM DTT.
lsopycnic and isokinetic sucrose-gradient centrifugation Isopycnic sucrose gradient centrifugation of the second supernatant fractions obtained by the micrococcal nuclease digestion of the Triton X-100 washed nuclei revealed a single band at 26.2% sucrose density in both age groups. Samples reduced with 2ME prior to loading onto the gradients peaked also at the same density (Fig. 4). The positioning of the peak of 2SF at 26.2% sucrose density was strictly reproducible in three different experiments involving digestions with 50, 100 and 200 units of nuclease (added to 5.5 ml nuclear suspensions of A260 = 15.0) for 15, 20 and 20 minutes, respectively, of digestion. It is therefore possible to conclude that the density of 2SF obtained as above was not measurably altered by age or S-S reduction. Pellet fractions resuspended in the lysis solution showed considerable light scattering. In four experiments the means of the relative absorbancies at 260[280[320[ 550 nm were 5.6/4.3/2.1/1 for both age groups. A brief low-speed centrifugation (3 minutes, 350 rpm) caused considerable decrease in the light scattering (the above values became 7.7/5.5/1.8/1) while pelleting approximately 11% of DNA and 43% of RNA in both ages (means of three determinations). Therefore, P fractions were routinely subjected to this centrifugation before loading onto the gradients. Figure 5 shows that the P fractions prepared in this way yielded a main peak and a small shoulder on the denser side of the peak in isopycnic sucrose-density gradient centrifugation profiles. The main peak seemed to consist mainly of a DNA--protein complex judging from its A26o/A28o ratios (around 1.7, not shown) and from the chemical composition of the unfractionated P determined in aliquots taken before loading onto the gradients (Fig. 5). The A26o/A 28o ratio of the shoulder was also around 1.7 but its nature was not otherwise determined. Density of the main peak was higher in old age than in young adulthood (Fig. 5). Reduc. tion of S-S bonds in the P fractions with 20-50 mM 2ME for 30--90 minutes at 4 °C prior to loading onto the gradients dramatically shifted the peak positions to a lower density in
74
!251 .15
~ 20 -6
" ~ o 15 E I0-
-6
5 ~
J
20
25
50 55 40 45 Percenl Sucrose
(u ~
#
:.
~.~
°~ 5
',
5
50
'l
20
25
30 35 40 Percenl sucrose
I
45
50
Fig. 4. Second supernatant fractions prepared from nuclei from livers of young-adult and old mice and subjected to isopycnic sucrose-density gradient centrifugation with or without prior treatment with 2ME. 5.5 ml of nuclei suspensions of A26o = 15.0 were digested with 50 units of nuclease for 15 minutes at 37 °C. • ~ •, old; O ~ O, young-adult; A...A, old treated with 2ME; Z~...A, youngadult treated with 2ME. The density of 2SF was the same in both age groups and was not influenced by treatment with 2ME. Fig. 5. Influence of age and treatment with 2ME on the densities of the material of P prepared by micrococcal nuclease digestion of Triton X-100 washed nuclei and subjected to isopycnic sucrosedensity gradient centrifugation. 5.5 ml of nuclei of A26o = 15.0 were digested with 50 units of nuclease for 20 minutes at 37 °C and the P fractions prepared as described in the text. In aliquots of the P fractions taken before layering onto the gradients, the relative amounts of DNA, RNA, and protein were DNA 1, RNA 0.20, protein 1.97 in old mice, and DNA 1, RNA 0.31, protein 2.03 in young-adult mice. • ~ • , old; O O, young; A . . . i , old reduced with 2ME; ~---A, young reduced with 2ME. It is seen that for P before 2ME the density of old > young, and that this difference was nearly abolished by 2ME treatment.
material from old animals. This 2ME-iliduced decrease in density was much less pronounced in material from young animals (Fig. 5). The above-mentioned aging and 2ME influences on the densities of P fractions were reproducible with four separate preparations involving 50--100 units of nuclease digestion for 20 minutes, and did not appear to be influenced much by the enzyme concentration within this range (for example, 20 minutes digestion with 100 units of nuclease yielded P fractions peaking at the densities of 27.8% and 25.6% sucrose for old and young adult mice, respectively, and these values became 25.4% and 25.2% after reduction with 2ME). It can be seen from Fig. 5 that, even if one assumes all of the 1 I% loss of DNA from the P o f young adult mice during the brief low-speed centrifugation step (3 minutes, 350 rpm) as a loss of high density material, the difference between the densities of the material of P of young-adult and old mice cannot be accounted for, and must therefore be regarded as a true age-related alteration. Figure 6 shows the fractionation patterns of the mechanically sheared chromatin preparations u p o n isopycnic sucrose-gradient centrifugation. Particles of the mechanically sheared chromatin preparations were detectable basically in two regions, and the low-
75
E
18-
~!l
~ 44 ~o ~ o g~
i' AFTERREDUCTO IN
BEFORE REDUCTION
1 6-
e~2 20
3o
4b
so
20
Percen! Sucrose
30 40 Percent Sucrose
50
Fig. 6. Fractionation of mechanically sheared chromatin from young-adult and old mice by isopycnic sucrose-density gradient eentrifugation before and after reduction with 2ME. e ~ e, old; O ~ O, young. TABLE II COMPOSITION OF CHROMATIN SOLUTIONS PREPARED BY HYPOTONIC LYSIS PLUS MINIMAL MECHANICALSHEARING OF THE NUCLEI FROM LIVERS OF YOUNG-ADULTAND OLD MICE Age
Pro tein/DNA *
DNA/RNA **
7-11 months 22-27 months
1.91 1.82
10.5 ± 0.7 13.4 ± 1.0
*Means of the measurements with three separate preparations for each age group. **Mean ± S.E.M. (p < 0.05). Data derived from five separate preparations for each age group.
density region peak tended to have a subpeak at around 28% sucrose density. The percentage of nuclear DNA recovered in the mechanically sheared chromatin solutions ranged from 72 to 90% in different preparations with no statistical difference between the two age groups. Their chemical compositions are presented in Table II. Although the overall pattern of fractionation (two main peaks and a subpeak at around 28% sucrose density) was reproducible, position of the high-density region peak of the mechanically sheared chromatin ranged from 37% to 40% sucrose (in young) and from 40% to 47% sucrose (in old) in different experiments. The percentage of the total loaded A26o recovered in the high-density region, however, was reproducibly higher in old age in all of the 8 separate experiments we carried out, and reduction of the samples with 20 mM 2ME prior to centrifugation always caused decreases in the percentages of A26o recovered in the high-density region, as exemplified in Fig. 6. The A ~o/A2so ratios of both the lowdensity and the high-density region peaks of the mechanically sheared chromatin in both ages were above 1 and usually about that of the unfractionated chromatin (~1.6). This and the absolute values of the densities would indicate that the materials of the peaks were neither pure DNA nor pure protein. Finally, Fig. 7 suggests that treatment with 20 mM 2ME caused decreases in the sedimentation rates of the Triton X-100 washed nuclei (intact by light microscopy) of old mice and possibly of young-adult mice (reproducible with three separate preparations).
76 E 1480
't2-
/;
~ a 'to
e~ e~ 20 18 TOP
t6
t4 12 10 8 Fraction Number
6
4 2 80TTOM
20 `18 '16 '14 '12 'tO {3 TOP Fraclion Number
6
4 2 BOTTOM
Fig. 7. Influence of 2ME treatment on the sedimentation rates of Triton X-IO0 washed nuclei from livers of young-adult and old mice. t ~ e, old; © ~ ©, young-adult;~ ..A, old with 2ME;~- - -~, young-adult with 2ME. Resolution of the isokinetic sucrose-gradient fractionation of 2SF was not high enough under the conditions studied to decide unambiguously whether the average DNA size in 2SF was larger in material from older compared to younger animals.
DISCUSSION Electrophoretic analysis of the DNA of micrococcal nuclease digestion fractions has shown that 1SF consists of mononucleosomes, 2SF of small amounts of mononucleosomes but mainly oligonucleosomes, and P mononudeosomes and still longer fragments of chromatin [15]. Chromatin regions containing transcriptionally active genes have been reported to be preferentially digested by micrococcal nuclease [15, 16] and also by deoxyribonucleases I [23, 24] and II [25, 26], suggesting that these regions might have a more open conformation in the nucleus. Such would make their DNA relatively more accessible to certain exogenously added enzymes under defined conditions of digestion. Enrichment of active genes .in ISF might therefore be a product of their preferential digestibility by the enzyme and possibly of the preferential solubility of RNA-associated mononucleosomes in Mg2÷.containing buffers [25]. Electron-microscopic and biochemical studies provide evidence that a 1% Triton X-100 wash of nuclei removes both the outer and the inner nuclear membranes [27, 28]. Nuclear morphology, however, remains otherwise largely intact and the possibility cannot be excluded that smaller particles digested by the nuclease in the inner aspects of the nucleus might be released more easily than larger particles into the digestion buffer. The influence of each of these factors on the amount of DNA recovered in the fractionation products remains to be elucidated. The amount of nuclear DNA recovered in 1SF was lower and that in 2SF higher in old mice than in young mice. Because of this dual effect of aging, occasional fluctuations in the comparative values of ISF or 2SF were overcome and ratio A stayed significantly below unity in all experiments (Table I). Although we did not carry out mRNA-derived eDNA hybridization expefimenls,~the results of Bloom and Anderson [15] suggest that the decreased 1SF/2SF ratio in old age may indicate a decrease in the total tramcriptive
77 activity with aging. The age-related decrease in the amount of RNA associated with chromatin (Table II) might in part be responsible for the lower DNA recoveries in 1SF in old mice. It is possible that an age-associated decrease in the accessibility of DNA in the chromatin complex might also play a role in the decreased 1SF/2SF ratio in old mice. It is of interest that 2ME did not increase ratio A in experiments with 50 units or higher concentrations of nuclease (Table I). This may suggest the presence of age-related alterations in the nucleosomal or supranucleosomal organizational features that cannot be corrected by 2ME, or alternatively might be due to a decrease in the percentage of nucleosomes associated with RNA in old age. In contrast to its inability to increase ratio A in five separate experiments with 50 or more units of nuclease, 2ME caused considerable increases in ratio A with 8 units of nuclease in two separate experiments (Table I), suggesting the possibility that age-related alterations involving S-S bonds and influencing micrococcal nuclease digestibility of the nuclei might be present but could be detected only by limiting levels of nucleolytic activity. (The decondensing influence of 2ME on the comparatively more condensed, compact structure of the chromatin complex in older mice (Figs. 5-7) would not be expected to enhance the digestibility of various chromatila regions in every situation if the restrictions presumed to be imposed were not absolute and could have already been overcome by relatively higher concentrations of added enzyme.) Further studies are needed, however, to confirm and explain the influence of 2ME on the 1SF/2SF ratio in aging mice with the relatively small amounts of nuclease. The increase in the amount of nuclear DNA recovered in 1SF + 2SF with 2ME and the decrease in P (Fig. 3) would suggest that either (1) certain chromatin organizational units are bound to a nuclease-resistant skeleton (perhaps similar to the "nuclear matrix" [27-29]) by proteins with S-S bonds and are released into the supernatants upon reduction of these bonds, or (2) reduction of S-S bonds causes a more open conformation in the condensed, compact regions in the chromatin complex, rendering them more accessible to the micrococcal and/or the endogenous nucleases. The decrease in value of ratio B with increasing digestion time, and the fact that the highest 2SF release from P by 2ME was seen in the absence of micrococcal nuclease and without incubation at 37 °C (Figs. lb, 2b and 3), might favor possibility (1), but the two possibilities need not be mutually exclusive. Both would be compatible with the 2ME-induced decondensation of the chromatin complex revealed by the isopycnic and isokinetic sucrose-gradient centrifugation experiments (Figs. 5-7). Whereas we noted that 2ME would nullify the density differences between the relatively nuclease-resistant fractions (P) from old and young chromatin on isopycnic gradient centrifugation (Fig. 5), 2ME did not cause release of more DNA from old than from young Triton-X-100 washed nuclei (Fig. 3). Further studies where the endogenous nucleolytic activity is kept to a minimum are being conducted to clarify these relationships. Figure 6 shows that the decondensing effect of S-S reducing agents can be demonstrated with mechanically sheared chromatin. Particles of these minimally sheared chromatin preparations appeared to be rather large as judged from their s~o.w values measured by isokinetic sucrose-gradient centrifugation (from s2o,w 0.47 X 103 to 1.75 × lOa). It has been noted that shearing of chromatin by homogenization induces S-S bonds
78 between histone H 3 molecules, whereas preparation of chromatin by hypotonic lysis plus minimal mechanical shearing of nuclei does not [30]. Since the mechanically sheared chromatin solutions used in our experiments were prepared by hypotonic lysis plus minimal mechanical shearing in the presence of a protease inhibitor, the S-S bonds we observed in such preparations might be expected to have existed in the original nuclei. It may be of interest that the percentages of the total loaded A26 o recovered in the highdensity region of the gradients in our experiments were persistently higher in old age. If minimal mechanical shearing of nuclei does not preferentially release the condensed chromatin fraction from the nuclei in old compared to young mice, then the age-related increase in the percentages of the total loaded A26o (mechanically sheared chromatin) recovered in the denser regions of the gradients (Fig. 6) may in part reflect a phenomenon similar to the age-associated increase in the density of P obtained by micrococcal nuclease digestion (Fig. 5). Since the experimental results shown in Fig. 6 were obtained with chromatin material lysed in EDTA, the possibility that complexing away of metals from chromatin by the added -SH compounds, rather than the reduction of S-S bonds, could have been responsible for the observed decondensing effects of 2ME and DTT, does not appear tenable. Decreases in the sedimentation rates of freshly prepared, Triton X-IO0 washed "whole nuclei" following 2ME treatment (Fig. 7) would suggest that the S-S bonds shown in the present study to be involved in the condensed structure of the chromatin complex are not an artifact of the experimental conditions. The possibility of an artifact is made further unlikely by prior evidence for the decondensing effects of 2ME and DTT on heterochromatin in situ in intact ceils [6]. A recent study involving the lysis of exponentially growing mouse carcinoma cells by sodium dodecyl sulfate and analysis of the released "DNA-protein complexes" by various procedures suggested that DNA in these cells may be folded by certain tightly binding nonhistone proteins carrying S-S bonds, and that intactness of S-S bonds may be essential for such a function [31 ]. These results [31 ], accord in general with our f'mdings with mechanically or enzymatically sheared or intact chromatin complexes prepared from principally nondividing ceils. It might be important, however, in evaluating the results with dividing cell populations to consider the possibility of alterations in the -SH/S-S status of certain chromatin proteins with the cell cycle [32 ]. Results of the present study are not inconsistent with the idea that S-S bonds may play an important role in the organization and condensed structure of the chromatin complex throughout life and also in its shift to a more condensed, compact state in older age [6, 11 ]. It seems that the portions of the complex relatively resistant to solubflization by micrococcal nuclease may constitute important sites of aging changes involving S-S bonds. Such age-associated alterations of the chromatin complex might have consequences on the state and regulation of genetic activity as well as on the replication and repair of DNA in old age [6, 11, 33]. Studies directed to the further characterization of the protein species involved and their interactions with other components of the cell nucleus may improve our understanding of the regulation of genetic activity in youth and old age.
79 A portion o f the present study was presented in the 31st Annual Scientific Meeting o f the Gerontological Society held in Dallas, U.S.A., on November 16-20, 1978.
ACKNOWLEDGEMENTS The present study was supported by USPHS Research Grant Ag-00424 from the NIA. S. Tag acknowledges the fellowship support o f the Scientific and Technical Research Council o f Turkey, and o f the Glenn Foundation for Medical Research. C.F. Tam is supported by USPHS postdoctoral fellowship AG-05063.
REFERENCES 1 R.D. Kornberg, Structure of chromatin. Ann. Rev. Biochem., 46 (1977) 931-954. 2 J. Brachet and N. Hulin, Binding of tritiated actinomycin and cell differentiation. Nature, 222 (1969) 481-482. 3 J. H. Frenster, Ultrastructural probes of chromatin during cell differentiation and cell division within living human bone marrow cells. In E. J. Hidv~gi, J. Stimegi and P. Venetianer (eds.), Biochemistry of the Cell Nucleus, Mechanism and Regulation of Gene Expression, American Elsevier, New York, 1974, pp. 419--428. 4 M. F. Lyon, Gene and chromosome inactivation. Genetics, 78 (1974) 305-308. 5 N. Maclean and V. A. Hilder, Mechanisms of chromatin activation and repression, Int. Rev. Cytol., 48 (1977) 1-54. 6 S. Ta~, Involvement of disulfide bonds in the condensed structure of facultative heterochromatin and implications on cellular differentiation and aging. Gerontology, 24 (1978) 358-364. 7 G. Beneke, K. Ervig and W. Schmitt, Age related polyploidisation and heterochromatinization of tendon cell nuclei. Beitr. Pathol., 141 (1970) 19-32. 8 H. Krug and H. vonWenk, Ver~inderungen der Feulgen-Hydrolysekurve w~ihrend des postnatalen wachstums in hippocumpus der ratte. Acta Histochem. {Jena), 45 (1974) 305-321. 9 I. Zs-Nagy and V. Zs-Nagy, Age dependence of heat induced strand separation of DNA in situ in postmitotic cells of rat brain as revealed by acridine orange microfluorimetry. Mech. Ageing Dev., 4 (1975) 349-360. 10 C. Pieri, I. Zs°Nagy, C. Giuli, V. Zs-Nagy and C. Bertoni-Freddari, Age dependent increase of thermal stability of in situ chromatin of rat liver and its reversal after hepatectomy. Experientia, 32 (1976) 891-893. 11 S. Ta$, Disulfide bonding in chromatin proteins with age and a suggested mechanism for aging and neoplasia. Exp. Gerontol., 11 (1976) 17-24. 12 A. M~sliwski, J. Mysliwska and S. Kimie6, Age related changes in chromatin of liver cell nuclei of different ploidity. Histochemistry, 52 (1977) 91-96. 13 A. M. Preumont, P. U. Gansen and J. Brachet, Cytochemical study of human lymphocytes stimulated by PHA in function of donor age. Mech. Ageing Dev., 7 (1978) 23-32. 14 S. P. Modak, C. Gonet, C. Unger-Ullmann and M. Chappuis, Chromatin structure in ageing mouse liver. Experientia, 34 (1978) 949. 15 K. S. Bloom and J. N. Anderson, Fractionation of hen oviduct chromatin into transcriptionally active and in.active regions after selective micrococcal nuclease digestion. Cell, 15 (1978) 141-150. 16 W. B. Levy and G. H. Dixon, Partial purification of transcriptionally active nucleosomes from trout testis cells. Nucleic Acids Res., 5 (1978) 4155-4163. 17 C. C. Widnell and J. R. Tara, A procedure for the isolation of enzymically active rat liver nuclei. Biochem. J., 92 (1964) 313-317. 18 G. L. Ellman, Tissue sulfhydryl groups. Arch. Biochem. Biophys., 82 (1959) 70-77,
80
19 J. H. Monahan and R. H. Hall, Preparation of chromatin from tissue culture cells - a convenient method. Anal Biochem., 65 (1975) 187-203. 20 O. H. Lowry, N. J. Roscbrough, A. L. Farr and R. J. Randall, Protein measurement with the Folin phenol rcagent. J. Biol. Chem., 193 (1951) 265-275. 21 G. Ceriotti, Determination of nucleic acids in animal tissues. J. Biol. Chem., 214 (1955) 59-70. 22 G. M. Richards, Modifications of the diphenylaminc reaction giving increased sensitivity and simplicity in the estimation of DNA. Anal, Biochem., 57 (1974) 369-376. 23 H. Weintraub and M. Groudine, Chromosomal subunits in active genes have an altered conformation. Science, 193 (1976) 848-856. 24 A. Garel and R. Axel, Selective digestion of transcriptionally active ovalhumine genes from oviduct nuclei. Proc. Natl. Acad. Sci. U.S.A., 73 (1976) 3966-3970. 25 J. M. Gottesfeld and P. J. G. Butler, Structure of transcriptionally active chromatin subunits. Nucleic Acids Res., 4 (1977) 3155-3173. 26 R. J. Billing and J. Bonner, The structure of chromatin as revealed by deoxyribonuclease digestion studies. Biochim. Biophys. Acta, 281 (1972) 453-462. 27 R. P. Aaronson and G. Blobel, On the attachment of the nuclear pore complex. J. Cell Biol., 62 (1974) 746-754. 28 D. E. Comings and T. A. Okada, Nuclear proteins. Exp. Cell Res., 103 (1976) 341-360. 29 L. D. Hodge, P. Mancini, F. M. Davis and P. Heywood, Nuclear matrix of HeLa $3 cells, J. Cell BioL, 72 (1977) 194-208. 30 W. T. Garrard, P. Nobis and R. Hancock, Histone H a disulfide reactions in interphase mitotic and native chromatin. J. Biol. Chem., 252 (1977) 4962-4967. 31 M. Nakane, T. Ide, K. Anzai, S. Ohara and T. Andoh, Supercoiled DNA folded by nonhistone proteins in cultured mouse carcinoma ceils. J. Biochem., 84 (1978) 145-157. 32 A. Sadgopal and J. Bonner, Proteins of interphase and metaphase chromosomes compared. Biochim. Biophys. Acta, 207 (1970) 227-239. 33 S. Ta~ and R. L. Walford, Possible role of disulfide bonds in chromatin in differentiation, meiotic rejuvenation and aging. Gerontologist, (Part II) 18 (1978) 131.
Note Added in Proof We have recently fractionated the Triton X-100 washed liver nuclei of young adult and old mice by parallel digestions with DNAse I and endogenous nucleases. The 1SF/2SF ratio, similar to the results with micrococcal nuclease, was significantly lower in older mice with DNAse I. Treatment of the nuclei with S-S reducing agents increased the release of DNA in chromatin into 2SF in both age groups but did not correct the decreased 1SF/2SF ratio of old mice. The increase into 2SF was more pronounced in old than in young mice at tl~e start of digestions and this was not attributable to activation of endogenous nucleases but appeared to be due to a more open conformation of the ehromatin complex after S-S reduction. Other experiments involving lysis of live lymphocytes of young adult and old mice on top of neutral sucrose gradients with Triton X-100 plus varying concentrations of NaC1, and in the presence or absence of S-S reducing agents, followed by centrifugation, suggested that the reducing agents exert a more pronounced decondensing effect on the chromatin complex of old than young mice, and that the protein species carrying the S--S bonds involved in the more condensed structure of the chromatin complex of older mice are, at least in part, one or more of the tightly DNA binding nonhistones which cannot be dissociated from DNA by 1.6 M NaCI.
65
©Elsevier Sequoia S.A., Lausanne - Printed in the Netherlands
DISULFIDE BONDS AND THE STRUCTURE C O M P L E X IN R E L A T I O N T O A G I N G
OF THE CHROMATIN
SINAN TAS, CHICK F. TAM and ROY L. WALFORD Department of Pathology, Center for Health Sciences, University of California, Los Angeles, CA 90024 (U.S.A.)
(Received May 18, 1979; in revised form August 6, 1979) SUMMARY Triton X-100 washed nuclei from livers of young-adult and old mice were digested with micrococcal nuclease and pelleted. Supernatants (1SF) were saved and the pellets lysed in a hypotonic EDTA buffer. A second supernatant (2SF) and a final pellet (P) were obtained by recentrifugation (7000 g, 7 minutes). The 1SF/2SF ratio, which has been shown to be an index of the transcriptionally active to inactive chromatin ratio, was lower in older mice. The fraction relatively resistant to solubilization by the nuclease (P) was found by isopycnic sucrose gradient centrifugation to be in a more compact, condensed state when prepared from older mice. Higher amounts of heavy density chromatin were obtained from nuclei of old than young mice by hypotonic lysis plus minimal mechanical shearing. 2-Mercaptoethanol (2ME) treatment brought the density of the material of P from old mice back to the levels of young mice. In both age groups 2ME decreased the densities of mechanically sheared chromatin as well as of the whole Triton X-IO0 washed nuclei. In nuclease digestion experiments treatment of the nuclei from both age groups with S-S reducing agents increased the release of DNA from P into the supernatants. The results are consistent with S-S bonds being involved in the condensed structure of chromatin in young and old mice and in the shift of the chromatin complex towards a more compact, condensed state in old age.
INTRODUCTION While an initial level of organization ofchromatin involves association of histones and DNA in nucleosomes [ 1], the nature of higher levels of organization is poorly understood. It is known, however, that accessibility of DNA in the chromatin complex is highly restricted. This restriction, which in part may be due to supranucleosomal organizational features, seems to become more pronounced, at least in some cases, with progressive cellular differentiation and is accompanied by a more compact, condensed chromatin structure [2, 3]. The condensed structure, detected mainly by cytological procedures, may be observed in identical regions of chromatin in all cells of an organism (constitutive heterochromatin) as well as in different regions in different cell types (facultative hetero-
66 chromatin) [4, 5]. Facultative heterochromatinization is considered to be a mechanism for gene inactivation operating during cellular differentiation, and previous evidence suggested a role for disulfide bonds in maintaining the condensed structure of facultative heterochromatin in interphase cells [6]. Several studies have suggested that chromatin of both postmitotic and proliferative cell populations of old animals may exist in a more compact, condensed state than chromatin from similar cells of young animals, and that the amount of freely accessible DNA in chromatin decreases with age [7-14]. Evidence presented elsewhere was not inconsistent with the idea that disulfide bonds may be involved in this age-related condensation, and that heterochromatinization of the genome may be an important feature of cellular aging [6, 11 ]. Recent investigations have provided evidence that genetically active chromatin regions in nuclei can be preferentially attacked by micrococcal nuclease under mild digestion conditions, and that the sequences for active genes in a digestion product are enriched in mononucleosomal-length DNA while depleted in oligonucleosomal-length DNA [15, 16]. Based in part on this preferential digestibility of active genes by micrococcal nuclease, Bloom and Anderson [ 15] recently described a procedure for separating nuclear DNA into three fractions: one enriched in active genes, one enriched in inactive genes, and a fraction relatively resistant to solubilization by the enzyme [15]. In the present study we applied this fractionation procedure to Triton X-100 washed nuclei of young-adult and old mice in the presence or absence of disulfide reducing agents. The chromatin species fractionated by micrococcal nuclease were then subjected to isopycnic and isokinetic sucrose-gradient centrifugations, with and without prior treatment with disulfide reducing agents. Results of these experiments showed that in both young-adult and old mice treatment of hepatic nuclei with S-S reducing agents significantly increased the release of DNA from the fraction relatively resistant to solubilization by micrococcal nuclease. We noted that this fraction assumed a more compact, higher density state in old age, and that treatment with 2-mercaptoethanol (2ME) prior to centrifugation brought the density of the fraction back to that characteristic of young animals. The densities of other chromatin fractions did not appear to be influenced by age or S-S reducing agents. The decondensing effect of S-S reducing agents on the chromatin complex was also demonstrable with the Triton X-IO0 washed nuclei of old and to a lesser degree of young mice, and with the denser fraction of mechanically sheared chromatin in both age groups. Finally, the active/inactive chromatin ratio measured by the micrococcal nuclease digestion procedure [15] appeared to decrease significantly with increasing age. MATERIAL AND METHODS Animals
Male mice of the long-lived (C57BL]6 X Balb/c)Fl hybrid were used. The youngadults were 7-11 months old and the aged animals 22-27 months old at the time of sacrifice. They were fed ad libitum with standard laboratory chow.
67
Preparation of nuclei Nuclei were prepared according to the method of Widnell and Tata [17], adjusted to a smaller amount of tissue. Animals were killed by decapitation between 8.30 and 10.30 a.m., their livers immediately taken into ice-cold homogenization medium, weighed, and equal amounts from two or (in the micrococcal nuclease digestion experiments) four animals of each age group pooled for use as starting material. Animals with enlarged or abnormally appearing livers or with gross evidence of disease elsewhere were excluded from the study and a healthy animal substituted. The protease inhibitor phenylmethylsulfonylfluoride (PMSF) in a concentration of 1 mM was included in the initial homogenization medium and at 0.2 mM in later media. Under the conditions of our experiments, PMSF was shown by means of the 5-5'-dithiobis-2-nitrobenzoic acid reaction [18] not to alter the free -SH content of chromatin.
Digestion of nuclei with micrococcal nuclease Freshly purified nuclei [17] were washed with 15 ml of 1% Triton X-100, 0.5 M sucrose, 3 mM MgC12, 0.2 mM PMSF, 10 mM Tris-HC1 (pH 7.4) for 10 minutes at ice-cold temperature and centrifuged at 150 g for 7 minutes in a round-bottom 25 mm diameter tube. The resulting loose nuclear pellets were resuspended in 0.35 M sucrose, 25 mM KCI, 5 mM MgC12, 0.5 mM CaCI2, 0.1 mM PMSF, 10 mM Tris (pH 7.4) and adjusted for both age groups to a concentration of A260 = 4.5--6.0. Two-ml aliquots were then incubated at 4 °C for 40 minutes in the presence or absence of 20 mM of either 2ME or dithiothreitol (DTT), warmed to 37 °C for 2 minutes, and incubated at 37 °C with micrococcal nuclease at concentrations and time periods indicated below (see Results). Following digestion, the aliquots were chilled in ice and centrifuged at 7000 g for 7 minutes at 4 °C. In every experiment separate 2.0 ml aliquots from both age groups and from 2ME treated and untreated groups were centrifuged as above but without prior incubation at 37 °C. These aliquots did not contain micrococcal nuclease and were used to determine the zero-minute values. The resulting first supernatant fractions (1SF)were set aside and the pellets lysed by suspension in 2.0 ml of 0.1 mM PMSF, 5 mM EDTA (pH 7.0) for 10 minutes. The lysed suspensions were centrifuged at 7000 g for 7 minutes at 4 °C to yield a second supernatant (2SF) and a final pellet (P) fraction. The pellet fractions were resuspended by a pasteur pipette in 2.0 ml of 0.1 mM PMSF, 5 mM EDTA (pH 7.0). In every experiment the buffers described above (with and without 2ME or nuclease) were run in parallel with the nuclear suspensions to provide controls for the chemical tests and blanks for the absorbancy measurements. Preparations of chromatin samples for the isopycnic sucrose gradient centrifugations were made by using 5.5 ml nuclear suspensions of A~o = 15.0 for each age group. Enzyme concentrations and digestion times were as described (see Results).
Preparation of mechanically sheared chromatin A mo.dification of the method described by Monahan and Hall [19] was used. Freshly purified nuclei [17] were washed twice for 15 minutes each with 15 ml of 1% Triton X-IO0, 0.25 M sucrose, 3 mM MgC12, 0.2 mM PMSF and 10 mM Tris-HCI (pH 7.9)
68 and centrifuged at 1000 g. The nuclear pellet was resuspended by vortexing in 15 ml of 0.14 M NaC1, 1 mM MgC12, 0.2 mM PMSF and 5 mM Tris-HC1 (pH 7.9). The suspension was allowed to stand in ice for I0 minutes, then centrifuged at 630 g for 5 minutes. This step was repeated and fluid was carefully drained from the pellet. Next, 10 ml of a solution consisting of 2 mM EDTA, 0.1 mM PMSF and 5 mM Tris (pH 7.9)were added to the nuclear pellet. The pellet was disrupted by three strokes in a tightly fitting glass-to-glass homogenizer. With maximum hand pressure and as quickly as possible the disrupted material was passed three times through 23, 27 and 30 gauge needles. The final solution was centrifuged at 2500 g for 10 minutes to pellet unbroken nuclei, nucleoli and any aggregates of chromatin which may have formed [ 19]. The supernatants (mechanically sheared chromatin) from both age groups were removed with a pasteur pipette and diluted with the 0.1 mM PMSF, 2 mM EDTA, 5 mM Tris (pH 7.9) lysis solution to equal concentrations of about A 260 = 1.5.
Isopycnic and isokinetic sucrose-gradient centrifugation Ice-cold sucrose solutions were prepared in the buffers used for the solubilization or suspension of the chromatin fractions or nuclei to be applied to the gradients. Linear sucrose gradients as indicated in the figures were generated by a Buchler density gradient engine (Buchler Instruments, New Jersey). For sedimentation rate experiments linear gradients from 23 to 53% sucrose were prepared with a 2.0 ml cushion of 63% sucrose at the bottom of the tubes. Two.ml samples of chromatin solutions or of nuclear suspensions at equal A2~ values (1.0-1.8, but the same in each experiment) for both age groups were layered on top of the gradients (31 ml) in 25 mm diameter nitrocellulose centrifuge tubes. In two additional tubes the layered samples from the two age groups were pretreated with 20 mM 2ME at 4 °C for 30--90 minutes in sealed tubes prior to layering. Two-ml samples of the buffer, or buffer containing 20 mM 2ME, were also layered on top of the gradients in separate tubes. The tubes were centrifuged at 18000 rpm for 48 hours at 4 °C (average g = 42 820) in a Beckman-Spinco SW 27 rotor. Equilibrium was reached by 36 hours and the position of peaks did not change between the 36th and 48th hours. Isokinetic sucrose-gradient centrifugation of the 2ME treated and untreated nuclei was done at 5000 rpm for a total of 11 minutes with an acceleration time of 2 minutes. The rotor (SW 27) made a total of about 53 000 revolutions. The tubes were punctured and 1.8-ml fractions collected (1.65-ml fractions for the sedimentation rate measurements). Absorbance of the fractions at 260,280 and 320 nm was determined in 1 cm path-length quartz spectrophotometry cells. Refractive indices in the fractions at 20 °C were measured with a refractometer (Abbe-3L, Bausch and Lomb, New York) and the corresponding weight percentages of sucrose determined. Absorption curves for the buffer and 20 mM 2ME in the buffer were drawn in order to subtract the corresponding absorbance values from those of the investigated samples. (Because 2ME absorbs at the above wavelengths, this subtraction step is essential in order not to overestimate the absorbance on top of the gradients.) Recoveries of the total loaded A~o in the gradients varied from 92 to 103%. In all experiments freshly prepared chromatin samples or nuclear suspensions were used.
69
Protein, R N A and D N A measurements Protein was determined according to the method of Lowry et al. [20] with bovine serum albumin as standard; RNA was measured as described by Ceriotti [ 2 1 ] . DNA, extracted by a final concentration of 1.5 N of HC104 at 70 °C for 20 minutes, was measured according to the description of Richards [22]. The buffers used for solubilization or suspension of the chromatin fractions or nuclei to be tested by the above procedures were run in parallel in each test as controls. DNA in the nuclei was measured in 0.1 mM PMSF, 5 mM EDTA (pH 7.0). (Tests indicated that prompt centrifugation o f the nuclear suspension in the digestion mixture prevented loss of any DNA into the supernatant.), In micrococcal nuclease digestion experiments estimation o f DNA in ISF was satisfactory either (1) when in a direct estimate the error introduced by interference of 2ME and the I)uffer was corrected by appropriate controls, or (2) when the DNA values were calculated indirectly by subtracting the DNA of 2SF + P from that of the nuclei suspension used for digestion and closely correlated with the A260 values.
o(ISF) 25
f ~,,...~
/~;....;;;'~ "
c{P)
20.
50-
~ 15 ¸
45-
~ ' 10
40-
354
5' Q
O 2
I
65-
~
5 t0 Digestion time (minutes) D(2SF)
'
zs-
I
g.
f
E "6
~ 55I g_
'it
5O
45
o 2
30-"
20
~ tb Digestion time (minutes)
20
o z
3 tb Digestion time (minutes)
20
Fig. la-c. Time course of fractionation by micrococcal nuclease digestion of the Triton X-T00 washed and 2-mercaptoethanol treated or untreated nuclei from livers of young-adult and old mice. Micrococcal nuclease concentrations were 8 units to 2 ml of nuclear suspension. Each point represents the mean of duplicate samples. • o, old; O O, young-adult; A...A, old treated with 2ME; zx...~ young-adult'treated with 2ME. It is seen that 2ME releases an increased amount of DNA from P into supernatants in both age gxoups, flaat this release occurs mainly with the zero-minute nuclei and mainly into 2SF, and that for ISF, young > old, and for 2SF, young < old,
70
RESULTS
Micrococcal nuclease digestion of the Triton X-I O0 washed nuclei Results of these experiments can be outlined as follows: (1) The amount of nuclear DNA recovered in 1SF showed a steep increase within the first few minutes and kept increasing, albeit at a slower rate, with increasing digestion time, while that recovered in 2SF declined with increasing time of digestion after an initial steep increase during the first minute (Figs. 1 a, b and 2a, b). The level of 2SF was influenced both by its rate of release from P and also by its rate of conversion to 1SF. Figures lb, c and 2b, c show that the main event in the 0-1-minute interval was the release of the material of 2SF from P. A nearly linear decrease was observed in the level of 2SF during the 2-10-minute digestion period, apparently due mainly to conversion to 1SF material (Figs. la, b, c and 2a, b, c). We noted that the amount of nuclear DNA released into 1SF was higher in young-adult than in old mice at all times of digestion and the reverse was true for 2SF (Figs. 1 a, b and 2a, b). This relationship could be expressed by
o(lSF) 30
I ~ / ~ c(P)
I 55~ I
50-
~ t53 c ~ tO-
45-
5-
35-
30-
o z
~,
~b
Digestion time (minutes)
20 25-
20-
15.
5.'J I =
10-
"6 5O 5-
45 2
5 tO Digestion time (minutes)
20
0
2
5 tO Digestion time (minutes)
20
Fig. 2a--¢. Same experiment as in Fig. la-c but with 50 units of microcoecal nuclease to 2 ml of nuclear suspension, o ~ e , old; O O, young-adult; A...A, old treated with 2ME;/x.../x, youngadult treated with 2ME.
71 Ratio A = (amoun_____._~tof DNA recovered in 1 S F ) o l d ] / l ( a m o u n t of DNA recovered in 1SF)young]
l
/
- - I / /
(amount of DNA recovered in 2SF) o l d ] / [(amount o f DNA recovered in 2SF) youngj
Since the value o f this ratio appeared to be independent of the digestion periods within the 1-20-minute range (see Figs. la, b and 2a, b) the 1st, 2nd, 5th, 10th and 20th minute values were assumed to form a single population and their mean as well as standard error o f the mean were calculated for each experiment (Table I). From Table I the overall mean of ratio A in the seven separate experiments without 2ME and involving 28 young-adult and 28 old mice was 0.906. This was significantly less than 1 (p < 0.005). (2) The amount of DNA recovered in P showed a steep decline within the first minute and a slow decline thereafter (Figs. lc and 2c), With higher concentrations o f enzyme (Fig. 2c versus Fig. 1c) the amounts of nuclear DNA recovered in P were lower at the first minute, but the differences related to enzyme concentration tended to become smaller through the end of the 20-minute digestion period. Amounts of DNA recovered in P did not significantly differ between the two age groups within the 8 - 1 3 0 units of micrococcal nuclease to 2.0 ml nuclear suspension range (not shown). (3) With low concentrations of enzyme (Fig. la) the amount of DNA recovered in 1SF increased with 2ME over the values for 1SF in the absence of 2ME. With higher enzyme concentrations (Fig, 2a) this effect became less pronounced, and could not be detected at all by an enzyme concentration of 130 units to 2.0 ml nuclei suspension of TABLE I RATIO A, COMPARING THE AMOUNTS OF NUCLEAR DNA RECOVERED IN THE FIRST AND SECOND SUPERNATANTS BY MICROCOCCAL NUCLEASE DIGESTION OF LIVER NUCLEI IN YOUNG-ADULT AND OLD MICE, WITHOUT AND WITH ADDED 2ME Units o f micrococcal nuclease added to 2.0 ml o f nuclei suspension
8 8 50 50 50** 85** 130"**
A26o o f the nuclei suspension
4.5 4.5 4.5 4.5 6.0 6.0 5.5
Mean o f ratio A * for 1, 2, 5, 10 and 20 minute digestions within each experiment ± S.E.M. without 2ME
with 2ME
0.883 0.904 0.892 0.910 0.905 0.912 0.933
1.007 1.017 0.854 0.903 0.879 0.928 0.924
± 0.022t ± 0.029* ± 0.026t ± 0.028* ± 0.019'
-+0.039 ± 0.025 ± 0.014t ± 0.022t ± 0.047
*Ratio A, see text for definition. Duplicate samples were used for each digestion period in all experiments. **Means of the 2rid, and 20th minute digestion values. Other digestion periods were not studied. ***Means of 1', 2, 5 and 10 minute digestion values. tRatio of significantly less than 1 (p < 0.005). *Ratio is significantly less than 1 (p < 0.025).
72 A26o = 5.5 (not shown). 2ME treatment of the nuclei also enhanced the release of the material of 2SF from P and this effect was seen most dramatically with the zero-time nuclei (i.e. before incubation at 37 °C and without nuclease) (Figs. lb, c and 2b, c). The ratio (in terms of DNA) Ratio B =
[(ISF + 2SF)/P] with 2ME [(1SF + 2SF)/P] without 2ME
accordingly was significantly larger than unity for both age groups and approached unity with increasing digestion time (Fig. 3). The variability of the value of ratio B in seven separate experiments using 8-130 units of nuclease to 2.0 ml nuclei suspension is indicated in Fig. 3.
1+i+
0 2 5 10 20 Digestion time(minutes)
Fig. 3. Value of ratio B as a function of the digestion period. Ratio B, an index of the 2ME-induced DNA release from P, is defined as (1SF + 2SF)/P with 2ME/[(ISF + 2SF)/P],without 2ME. It is seen that the amount of DNA released by 2ME into the supernatants 1SF + 2SF decreases with increasing digestion time. Data were derived from the same set of experiments as in Table I. Bars represent the standard errors of the means of experiments using 8--130 units of nuelease. O - - O , old; O - - O , young-adult. (4) When digestions were carried out in the presence of 2ME, ratio A remained significantly smaller than unity with enzyme concentrations of 50 units or more to 2.0 ml nuclei suspensions (Table I, p < 0.005). With 8 units of nuclease, however, ratio A was increased by treatment with 2ME in two separate experiments (Table 1). When the Triton X-IO0 washed nuclei that had been incubated in the absence of 2ME for 40 minutes at 4 °C were centrifuged at 7000 g for 7 minutes, only a small amount of nuclear DNA (not exceeding 2-3% and usually much less) was released into the supernatant (1SF). Reliable comparisons of ISF between young and old mice in these experiments were therefore not possible by the chemical test used. (In a limited number of experiments we incubated nuclei in the digestion buffer without micrococcal nuelease at 37 °C. Such experiments suggested that as much as 13% of nuclear DNA could be released into the 1SF within 20 minutes, probably as a result of the endogenous Mg2÷, Ca2+-activated nucleolytic activity.) In earlier experiments the nuclei for micrococcal nuclease digestion were prepared as described by Bloom and Anderson [15]. Although such nuclear preparations gave
73 similar 2ME effects and age-related decline of the 1SF/2SF ratio to those described above, microscopic examination revealed clumps which were difficult to separate into individual nuclei. Because of possible ambiguities these clumps might cause in comparing the amounts of DNA recovered in P, we employed the low-speed centrifugation procedure described in Methods. This gave pure, light-microscopically intact nuclear suspensions virtually free of clumps. Also, we used I% instead of 0.2% Triton X-IO0 in order to remove nuclear membranes completely. Although these modifications of the original procedure [15 ] enhanced the releasability of DNA into 1SF by the micrococcal nuclease, the cause of the nearly complete release of the material of 2SF within the first minute of digestion in our experiments, as opposed to increasing release of 2SF with increasing time of digestion of the nuclei from chick cells in the original study [15], is not known. It was not possible to carry out the micrococcal nuclease digestion experiments for prolonged periods because the material of 2SF started to aggregate after 40 minutes under the described conditions and with or without 2ME in the digestion buffer (not shown). All of the 2ME effects noted above were reproducible with 20 mM DTT.
lsopycnic and isokinetic sucrose-gradient centrifugation Isopycnic sucrose gradient centrifugation of the second supernatant fractions obtained by the micrococcal nuclease digestion of the Triton X-100 washed nuclei revealed a single band at 26.2% sucrose density in both age groups. Samples reduced with 2ME prior to loading onto the gradients peaked also at the same density (Fig. 4). The positioning of the peak of 2SF at 26.2% sucrose density was strictly reproducible in three different experiments involving digestions with 50, 100 and 200 units of nuclease (added to 5.5 ml nuclear suspensions of A260 = 15.0) for 15, 20 and 20 minutes, respectively, of digestion. It is therefore possible to conclude that the density of 2SF obtained as above was not measurably altered by age or S-S reduction. Pellet fractions resuspended in the lysis solution showed considerable light scattering. In four experiments the means of the relative absorbancies at 260[280[320[ 550 nm were 5.6/4.3/2.1/1 for both age groups. A brief low-speed centrifugation (3 minutes, 350 rpm) caused considerable decrease in the light scattering (the above values became 7.7/5.5/1.8/1) while pelleting approximately 11% of DNA and 43% of RNA in both ages (means of three determinations). Therefore, P fractions were routinely subjected to this centrifugation before loading onto the gradients. Figure 5 shows that the P fractions prepared in this way yielded a main peak and a small shoulder on the denser side of the peak in isopycnic sucrose-density gradient centrifugation profiles. The main peak seemed to consist mainly of a DNA--protein complex judging from its A26o/A28o ratios (around 1.7, not shown) and from the chemical composition of the unfractionated P determined in aliquots taken before loading onto the gradients (Fig. 5). The A26o/A 28o ratio of the shoulder was also around 1.7 but its nature was not otherwise determined. Density of the main peak was higher in old age than in young adulthood (Fig. 5). Reduc. tion of S-S bonds in the P fractions with 20-50 mM 2ME for 30--90 minutes at 4 °C prior to loading onto the gradients dramatically shifted the peak positions to a lower density in
74
!251 .15
~ 20 -6
" ~ o 15 E I0-
-6
5 ~
J
20
25
50 55 40 45 Percenl Sucrose
(u ~
#
:.
~.~
°~ 5
',
5
50
'l
20
25
30 35 40 Percenl sucrose
I
45
50
Fig. 4. Second supernatant fractions prepared from nuclei from livers of young-adult and old mice and subjected to isopycnic sucrose-density gradient centrifugation with or without prior treatment with 2ME. 5.5 ml of nuclei suspensions of A26o = 15.0 were digested with 50 units of nuclease for 15 minutes at 37 °C. • ~ •, old; O ~ O, young-adult; A...A, old treated with 2ME; Z~...A, youngadult treated with 2ME. The density of 2SF was the same in both age groups and was not influenced by treatment with 2ME. Fig. 5. Influence of age and treatment with 2ME on the densities of the material of P prepared by micrococcal nuclease digestion of Triton X-100 washed nuclei and subjected to isopycnic sucrosedensity gradient centrifugation. 5.5 ml of nuclei of A26o = 15.0 were digested with 50 units of nuclease for 20 minutes at 37 °C and the P fractions prepared as described in the text. In aliquots of the P fractions taken before layering onto the gradients, the relative amounts of DNA, RNA, and protein were DNA 1, RNA 0.20, protein 1.97 in old mice, and DNA 1, RNA 0.31, protein 2.03 in young-adult mice. • ~ • , old; O O, young; A . . . i , old reduced with 2ME; ~---A, young reduced with 2ME. It is seen that for P before 2ME the density of old > young, and that this difference was nearly abolished by 2ME treatment.
material from old animals. This 2ME-iliduced decrease in density was much less pronounced in material from young animals (Fig. 5). The above-mentioned aging and 2ME influences on the densities of P fractions were reproducible with four separate preparations involving 50--100 units of nuclease digestion for 20 minutes, and did not appear to be influenced much by the enzyme concentration within this range (for example, 20 minutes digestion with 100 units of nuclease yielded P fractions peaking at the densities of 27.8% and 25.6% sucrose for old and young adult mice, respectively, and these values became 25.4% and 25.2% after reduction with 2ME). It can be seen from Fig. 5 that, even if one assumes all of the 1 I% loss of DNA from the P o f young adult mice during the brief low-speed centrifugation step (3 minutes, 350 rpm) as a loss of high density material, the difference between the densities of the material of P of young-adult and old mice cannot be accounted for, and must therefore be regarded as a true age-related alteration. Figure 6 shows the fractionation patterns of the mechanically sheared chromatin preparations u p o n isopycnic sucrose-gradient centrifugation. Particles of the mechanically sheared chromatin preparations were detectable basically in two regions, and the low-
75
E
18-
~!l
~ 44 ~o ~ o g~
i' AFTERREDUCTO IN
BEFORE REDUCTION
1 6-
e~2 20
3o
4b
so
20
Percen! Sucrose
30 40 Percent Sucrose
50
Fig. 6. Fractionation of mechanically sheared chromatin from young-adult and old mice by isopycnic sucrose-density gradient eentrifugation before and after reduction with 2ME. e ~ e, old; O ~ O, young. TABLE II COMPOSITION OF CHROMATIN SOLUTIONS PREPARED BY HYPOTONIC LYSIS PLUS MINIMAL MECHANICALSHEARING OF THE NUCLEI FROM LIVERS OF YOUNG-ADULTAND OLD MICE Age
Pro tein/DNA *
DNA/RNA **
7-11 months 22-27 months
1.91 1.82
10.5 ± 0.7 13.4 ± 1.0
*Means of the measurements with three separate preparations for each age group. **Mean ± S.E.M. (p < 0.05). Data derived from five separate preparations for each age group.
density region peak tended to have a subpeak at around 28% sucrose density. The percentage of nuclear DNA recovered in the mechanically sheared chromatin solutions ranged from 72 to 90% in different preparations with no statistical difference between the two age groups. Their chemical compositions are presented in Table II. Although the overall pattern of fractionation (two main peaks and a subpeak at around 28% sucrose density) was reproducible, position of the high-density region peak of the mechanically sheared chromatin ranged from 37% to 40% sucrose (in young) and from 40% to 47% sucrose (in old) in different experiments. The percentage of the total loaded A26o recovered in the high-density region, however, was reproducibly higher in old age in all of the 8 separate experiments we carried out, and reduction of the samples with 20 mM 2ME prior to centrifugation always caused decreases in the percentages of A26o recovered in the high-density region, as exemplified in Fig. 6. The A ~o/A2so ratios of both the lowdensity and the high-density region peaks of the mechanically sheared chromatin in both ages were above 1 and usually about that of the unfractionated chromatin (~1.6). This and the absolute values of the densities would indicate that the materials of the peaks were neither pure DNA nor pure protein. Finally, Fig. 7 suggests that treatment with 20 mM 2ME caused decreases in the sedimentation rates of the Triton X-100 washed nuclei (intact by light microscopy) of old mice and possibly of young-adult mice (reproducible with three separate preparations).
76 E 1480
't2-
/;
~ a 'to
e~ e~ 20 18 TOP
t6
t4 12 10 8 Fraction Number
6
4 2 80TTOM
20 `18 '16 '14 '12 'tO {3 TOP Fraclion Number
6
4 2 BOTTOM
Fig. 7. Influence of 2ME treatment on the sedimentation rates of Triton X-IO0 washed nuclei from livers of young-adult and old mice. t ~ e, old; © ~ ©, young-adult;~ ..A, old with 2ME;~- - -~, young-adult with 2ME. Resolution of the isokinetic sucrose-gradient fractionation of 2SF was not high enough under the conditions studied to decide unambiguously whether the average DNA size in 2SF was larger in material from older compared to younger animals.
DISCUSSION Electrophoretic analysis of the DNA of micrococcal nuclease digestion fractions has shown that 1SF consists of mononucleosomes, 2SF of small amounts of mononucleosomes but mainly oligonucleosomes, and P mononudeosomes and still longer fragments of chromatin [15]. Chromatin regions containing transcriptionally active genes have been reported to be preferentially digested by micrococcal nuclease [15, 16] and also by deoxyribonucleases I [23, 24] and II [25, 26], suggesting that these regions might have a more open conformation in the nucleus. Such would make their DNA relatively more accessible to certain exogenously added enzymes under defined conditions of digestion. Enrichment of active genes .in ISF might therefore be a product of their preferential digestibility by the enzyme and possibly of the preferential solubility of RNA-associated mononucleosomes in Mg2÷.containing buffers [25]. Electron-microscopic and biochemical studies provide evidence that a 1% Triton X-100 wash of nuclei removes both the outer and the inner nuclear membranes [27, 28]. Nuclear morphology, however, remains otherwise largely intact and the possibility cannot be excluded that smaller particles digested by the nuclease in the inner aspects of the nucleus might be released more easily than larger particles into the digestion buffer. The influence of each of these factors on the amount of DNA recovered in the fractionation products remains to be elucidated. The amount of nuclear DNA recovered in 1SF was lower and that in 2SF higher in old mice than in young mice. Because of this dual effect of aging, occasional fluctuations in the comparative values of ISF or 2SF were overcome and ratio A stayed significantly below unity in all experiments (Table I). Although we did not carry out mRNA-derived eDNA hybridization expefimenls,~the results of Bloom and Anderson [15] suggest that the decreased 1SF/2SF ratio in old age may indicate a decrease in the total tramcriptive
77 activity with aging. The age-related decrease in the amount of RNA associated with chromatin (Table II) might in part be responsible for the lower DNA recoveries in 1SF in old mice. It is possible that an age-associated decrease in the accessibility of DNA in the chromatin complex might also play a role in the decreased 1SF/2SF ratio in old mice. It is of interest that 2ME did not increase ratio A in experiments with 50 units or higher concentrations of nuclease (Table I). This may suggest the presence of age-related alterations in the nucleosomal or supranucleosomal organizational features that cannot be corrected by 2ME, or alternatively might be due to a decrease in the percentage of nucleosomes associated with RNA in old age. In contrast to its inability to increase ratio A in five separate experiments with 50 or more units of nuclease, 2ME caused considerable increases in ratio A with 8 units of nuclease in two separate experiments (Table I), suggesting the possibility that age-related alterations involving S-S bonds and influencing micrococcal nuclease digestibility of the nuclei might be present but could be detected only by limiting levels of nucleolytic activity. (The decondensing influence of 2ME on the comparatively more condensed, compact structure of the chromatin complex in older mice (Figs. 5-7) would not be expected to enhance the digestibility of various chromatila regions in every situation if the restrictions presumed to be imposed were not absolute and could have already been overcome by relatively higher concentrations of added enzyme.) Further studies are needed, however, to confirm and explain the influence of 2ME on the 1SF/2SF ratio in aging mice with the relatively small amounts of nuclease. The increase in the amount of nuclear DNA recovered in 1SF + 2SF with 2ME and the decrease in P (Fig. 3) would suggest that either (1) certain chromatin organizational units are bound to a nuclease-resistant skeleton (perhaps similar to the "nuclear matrix" [27-29]) by proteins with S-S bonds and are released into the supernatants upon reduction of these bonds, or (2) reduction of S-S bonds causes a more open conformation in the condensed, compact regions in the chromatin complex, rendering them more accessible to the micrococcal and/or the endogenous nucleases. The decrease in value of ratio B with increasing digestion time, and the fact that the highest 2SF release from P by 2ME was seen in the absence of micrococcal nuclease and without incubation at 37 °C (Figs. lb, 2b and 3), might favor possibility (1), but the two possibilities need not be mutually exclusive. Both would be compatible with the 2ME-induced decondensation of the chromatin complex revealed by the isopycnic and isokinetic sucrose-gradient centrifugation experiments (Figs. 5-7). Whereas we noted that 2ME would nullify the density differences between the relatively nuclease-resistant fractions (P) from old and young chromatin on isopycnic gradient centrifugation (Fig. 5), 2ME did not cause release of more DNA from old than from young Triton-X-100 washed nuclei (Fig. 3). Further studies where the endogenous nucleolytic activity is kept to a minimum are being conducted to clarify these relationships. Figure 6 shows that the decondensing effect of S-S reducing agents can be demonstrated with mechanically sheared chromatin. Particles of these minimally sheared chromatin preparations appeared to be rather large as judged from their s~o.w values measured by isokinetic sucrose-gradient centrifugation (from s2o,w 0.47 X 103 to 1.75 × lOa). It has been noted that shearing of chromatin by homogenization induces S-S bonds
78 between histone H 3 molecules, whereas preparation of chromatin by hypotonic lysis plus minimal mechanical shearing of nuclei does not [30]. Since the mechanically sheared chromatin solutions used in our experiments were prepared by hypotonic lysis plus minimal mechanical shearing in the presence of a protease inhibitor, the S-S bonds we observed in such preparations might be expected to have existed in the original nuclei. It may be of interest that the percentages of the total loaded A26 o recovered in the highdensity region of the gradients in our experiments were persistently higher in old age. If minimal mechanical shearing of nuclei does not preferentially release the condensed chromatin fraction from the nuclei in old compared to young mice, then the age-related increase in the percentages of the total loaded A26o (mechanically sheared chromatin) recovered in the denser regions of the gradients (Fig. 6) may in part reflect a phenomenon similar to the age-associated increase in the density of P obtained by micrococcal nuclease digestion (Fig. 5). Since the experimental results shown in Fig. 6 were obtained with chromatin material lysed in EDTA, the possibility that complexing away of metals from chromatin by the added -SH compounds, rather than the reduction of S-S bonds, could have been responsible for the observed decondensing effects of 2ME and DTT, does not appear tenable. Decreases in the sedimentation rates of freshly prepared, Triton X-IO0 washed "whole nuclei" following 2ME treatment (Fig. 7) would suggest that the S-S bonds shown in the present study to be involved in the condensed structure of the chromatin complex are not an artifact of the experimental conditions. The possibility of an artifact is made further unlikely by prior evidence for the decondensing effects of 2ME and DTT on heterochromatin in situ in intact ceils [6]. A recent study involving the lysis of exponentially growing mouse carcinoma cells by sodium dodecyl sulfate and analysis of the released "DNA-protein complexes" by various procedures suggested that DNA in these cells may be folded by certain tightly binding nonhistone proteins carrying S-S bonds, and that intactness of S-S bonds may be essential for such a function [31 ]. These results [31 ], accord in general with our f'mdings with mechanically or enzymatically sheared or intact chromatin complexes prepared from principally nondividing ceils. It might be important, however, in evaluating the results with dividing cell populations to consider the possibility of alterations in the -SH/S-S status of certain chromatin proteins with the cell cycle [32 ]. Results of the present study are not inconsistent with the idea that S-S bonds may play an important role in the organization and condensed structure of the chromatin complex throughout life and also in its shift to a more condensed, compact state in older age [6, 11 ]. It seems that the portions of the complex relatively resistant to solubflization by micrococcal nuclease may constitute important sites of aging changes involving S-S bonds. Such age-associated alterations of the chromatin complex might have consequences on the state and regulation of genetic activity as well as on the replication and repair of DNA in old age [6, 11, 33]. Studies directed to the further characterization of the protein species involved and their interactions with other components of the cell nucleus may improve our understanding of the regulation of genetic activity in youth and old age.
79 A portion o f the present study was presented in the 31st Annual Scientific Meeting o f the Gerontological Society held in Dallas, U.S.A., on November 16-20, 1978.
ACKNOWLEDGEMENTS The present study was supported by USPHS Research Grant Ag-00424 from the NIA. S. Tag acknowledges the fellowship support o f the Scientific and Technical Research Council o f Turkey, and o f the Glenn Foundation for Medical Research. C.F. Tam is supported by USPHS postdoctoral fellowship AG-05063.
REFERENCES 1 R.D. Kornberg, Structure of chromatin. Ann. Rev. Biochem., 46 (1977) 931-954. 2 J. Brachet and N. Hulin, Binding of tritiated actinomycin and cell differentiation. Nature, 222 (1969) 481-482. 3 J. H. Frenster, Ultrastructural probes of chromatin during cell differentiation and cell division within living human bone marrow cells. In E. J. Hidv~gi, J. Stimegi and P. Venetianer (eds.), Biochemistry of the Cell Nucleus, Mechanism and Regulation of Gene Expression, American Elsevier, New York, 1974, pp. 419--428. 4 M. F. Lyon, Gene and chromosome inactivation. Genetics, 78 (1974) 305-308. 5 N. Maclean and V. A. Hilder, Mechanisms of chromatin activation and repression, Int. Rev. Cytol., 48 (1977) 1-54. 6 S. Ta~, Involvement of disulfide bonds in the condensed structure of facultative heterochromatin and implications on cellular differentiation and aging. Gerontology, 24 (1978) 358-364. 7 G. Beneke, K. Ervig and W. Schmitt, Age related polyploidisation and heterochromatinization of tendon cell nuclei. Beitr. Pathol., 141 (1970) 19-32. 8 H. Krug and H. vonWenk, Ver~inderungen der Feulgen-Hydrolysekurve w~ihrend des postnatalen wachstums in hippocumpus der ratte. Acta Histochem. {Jena), 45 (1974) 305-321. 9 I. Zs-Nagy and V. Zs-Nagy, Age dependence of heat induced strand separation of DNA in situ in postmitotic cells of rat brain as revealed by acridine orange microfluorimetry. Mech. Ageing Dev., 4 (1975) 349-360. 10 C. Pieri, I. Zs°Nagy, C. Giuli, V. Zs-Nagy and C. Bertoni-Freddari, Age dependent increase of thermal stability of in situ chromatin of rat liver and its reversal after hepatectomy. Experientia, 32 (1976) 891-893. 11 S. Ta$, Disulfide bonding in chromatin proteins with age and a suggested mechanism for aging and neoplasia. Exp. Gerontol., 11 (1976) 17-24. 12 A. M~sliwski, J. Mysliwska and S. Kimie6, Age related changes in chromatin of liver cell nuclei of different ploidity. Histochemistry, 52 (1977) 91-96. 13 A. M. Preumont, P. U. Gansen and J. Brachet, Cytochemical study of human lymphocytes stimulated by PHA in function of donor age. Mech. Ageing Dev., 7 (1978) 23-32. 14 S. P. Modak, C. Gonet, C. Unger-Ullmann and M. Chappuis, Chromatin structure in ageing mouse liver. Experientia, 34 (1978) 949. 15 K. S. Bloom and J. N. Anderson, Fractionation of hen oviduct chromatin into transcriptionally active and in.active regions after selective micrococcal nuclease digestion. Cell, 15 (1978) 141-150. 16 W. B. Levy and G. H. Dixon, Partial purification of transcriptionally active nucleosomes from trout testis cells. Nucleic Acids Res., 5 (1978) 4155-4163. 17 C. C. Widnell and J. R. Tara, A procedure for the isolation of enzymically active rat liver nuclei. Biochem. J., 92 (1964) 313-317. 18 G. L. Ellman, Tissue sulfhydryl groups. Arch. Biochem. Biophys., 82 (1959) 70-77,
80
19 J. H. Monahan and R. H. Hall, Preparation of chromatin from tissue culture cells - a convenient method. Anal Biochem., 65 (1975) 187-203. 20 O. H. Lowry, N. J. Roscbrough, A. L. Farr and R. J. Randall, Protein measurement with the Folin phenol rcagent. J. Biol. Chem., 193 (1951) 265-275. 21 G. Ceriotti, Determination of nucleic acids in animal tissues. J. Biol. Chem., 214 (1955) 59-70. 22 G. M. Richards, Modifications of the diphenylaminc reaction giving increased sensitivity and simplicity in the estimation of DNA. Anal, Biochem., 57 (1974) 369-376. 23 H. Weintraub and M. Groudine, Chromosomal subunits in active genes have an altered conformation. Science, 193 (1976) 848-856. 24 A. Garel and R. Axel, Selective digestion of transcriptionally active ovalhumine genes from oviduct nuclei. Proc. Natl. Acad. Sci. U.S.A., 73 (1976) 3966-3970. 25 J. M. Gottesfeld and P. J. G. Butler, Structure of transcriptionally active chromatin subunits. Nucleic Acids Res., 4 (1977) 3155-3173. 26 R. J. Billing and J. Bonner, The structure of chromatin as revealed by deoxyribonuclease digestion studies. Biochim. Biophys. Acta, 281 (1972) 453-462. 27 R. P. Aaronson and G. Blobel, On the attachment of the nuclear pore complex. J. Cell Biol., 62 (1974) 746-754. 28 D. E. Comings and T. A. Okada, Nuclear proteins. Exp. Cell Res., 103 (1976) 341-360. 29 L. D. Hodge, P. Mancini, F. M. Davis and P. Heywood, Nuclear matrix of HeLa $3 cells, J. Cell BioL, 72 (1977) 194-208. 30 W. T. Garrard, P. Nobis and R. Hancock, Histone H a disulfide reactions in interphase mitotic and native chromatin. J. Biol. Chem., 252 (1977) 4962-4967. 31 M. Nakane, T. Ide, K. Anzai, S. Ohara and T. Andoh, Supercoiled DNA folded by nonhistone proteins in cultured mouse carcinoma ceils. J. Biochem., 84 (1978) 145-157. 32 A. Sadgopal and J. Bonner, Proteins of interphase and metaphase chromosomes compared. Biochim. Biophys. Acta, 207 (1970) 227-239. 33 S. Ta~ and R. L. Walford, Possible role of disulfide bonds in chromatin in differentiation, meiotic rejuvenation and aging. Gerontologist, (Part II) 18 (1978) 131.
Note Added in Proof We have recently fractionated the Triton X-100 washed liver nuclei of young adult and old mice by parallel digestions with DNAse I and endogenous nucleases. The 1SF/2SF ratio, similar to the results with micrococcal nuclease, was significantly lower in older mice with DNAse I. Treatment of the nuclei with S-S reducing agents increased the release of DNA in chromatin into 2SF in both age groups but did not correct the decreased 1SF/2SF ratio of old mice. The increase into 2SF was more pronounced in old than in young mice at tl~e start of digestions and this was not attributable to activation of endogenous nucleases but appeared to be due to a more open conformation of the ehromatin complex after S-S reduction. Other experiments involving lysis of live lymphocytes of young adult and old mice on top of neutral sucrose gradients with Triton X-100 plus varying concentrations of NaC1, and in the presence or absence of S-S reducing agents, followed by centrifugation, suggested that the reducing agents exert a more pronounced decondensing effect on the chromatin complex of old than young mice, and that the protein species carrying the S--S bonds involved in the more condensed structure of the chromatin complex of older mice are, at least in part, one or more of the tightly DNA binding nonhistones which cannot be dissociated from DNA by 1.6 M NaCI.