Biochimica et Biophysica Acta, 1130 (1992) 307-313 © 1992 Elsevier Science Publishers B.V. All rights reserved 0167-4781/92/$05.00
307
BBAEXP 92361
Nuclear distribution of histone deacetylase: a marker enzyme for the internal nuclear matrix M i c h a e l J. H e n d z e l a n d J a m e s R. D a v i e Department of Biochemistry and Molecldar Biology, FacuJJyof Medicble. Unirersityof Manitoba, Winnipeg, Manitoba (Canada) (Received 25 September 1991)
Key words: Hislone deacetylase; Internal nuclear matrix; Histone acetylation Nuclear matrins are proteins that localize to the internal nuclear matrix, in a previous study, we reported that histone deacetylase is a component of the internal matrix, suggesting that histone dcaeetylase is a nuclear matrin. Here, we demonstrate that the majority of the histone deacetylase activity is associated with the internal nuclear matrices of chicken and trout liver. Thus, the association of the histone deacetylase with the internal nuclear matrix is neither tissue- nor species-s,Jecific. Using historic deacetylase as a marker enzyme for the partitioning of the internal nuclear matrix during nuclear fractionations, we show that in contrast to the internal nuclear matrices of trout liver, trout hepatocellular carcinoma and chicken liver, the stability of the chicken erythrocyte internal nuclear matrix is temperature-dependent. Our results support a model that has the histone deaeetylase mediating transient interactions between the internal nuclear matrix and chromatin regions undergoing dynamic acetylation, for example transcriptionally active chromatin regions.
Introduction The nuclear matrix is an operationally defined structure typically prepared by high salt extraction of isolated nuclei following extensive mtclease digestion. The insoluble material is composed of residual nucleoli, the nuclear pore-lamina complex, and the internal nuclear matrix [1]. Many nuclear metabolic processes are associated with nuclear matrix preparations, including replication, transcription, RNA splicing and topoisomerase activity [1-5]. The nuclear matrix is thought to represent a structural framework for nuclear/chromatin organization [6]. This model of the nucleus views topologically independent chromatin domains as the units of regulation [4,6-.8]. These domains may contain one to several genes and are fixed to the nuclear matrix at both ends through DNA sequences termed nuclear matrix attachments (MARs) (reviewed in Refs. 4,6). Specific nuclear matrins, which are proteins of the internal nuclear matrix, are thought to interact with MARs (e.g., attachment region binding protein) [9,10]. Current evidence suggests that MARs are stable attachment sites [6]. In addition, there exist dynamic
Correspondence: J.R. Davie, Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Manitoba, 770 Bannatyne Ave., Winnipeg, Manitoba R3E 0W3, Canada.
interactions between transcriptionally active gene chromatin regions and the nuclear matrix [11]. Recently, we demonstrated that approx. 50% of the total nuclear histone deacetylase activity is associated with the internal nuclear matrix of chicken erythrocyte [12]. EM studies, however, have questioned the presence of an internal nuclear matrix in chicken erythrocyte. Lafond and Woodcock [13] used 'mild' isolation procedures in an attempt to visualize the nuclear matrix of the chicken mature erythrocyte. These conditions involved DNase I digestion at 4°C and a single extraction of the nuclease-digested nuclei with a 2 M NaCl-containing buffer. Upon visual inspection of these preparations by electron microscopy, they found only the nuclear pore-lamina complex and no evidence for an internal nuclear matri~ [13]. Biochemical studies, in contrast, demonstrate the presence of internal nuclear matrix proteins in chicken e rythrocytes [9,10,12]. In this report, we determined the tissue- and species-specificity of the association of histone deacetylase with the internal nuclear matrix. We show that the majority of the histone deacetylase activity of chicken and trout liver is found with the internal nuclear matrix, demonstrating that histone deacetylase is a nuclear matrin and a marker enzyme for the internal nuclear matrix. By analyzing the activity of the histone deacetylase in various nuclear matrix preparations, we demonstrate that first, the stability of the internal
3~ nuclear matrix of chicken erythrocytes is temperaturedependent and second, that neoplastic transformation of trout hepatocytes does not alter the stability of the internal nuclear matrix. Materials and Methods
Isolation of nuclei Chicken immature erythrocyte nuclei were isolated as described by Delcuve and Davie [14]. Trout liver, trout hepatocellular carcinoma and chicken liver nuclei were prepared by homogenization in 10 mM Pipes/1.0 M hexylene glycol/l% v / v thiodiglycol/2 mM MgCi,/30 mM sodium butyrate (pH 7.0) (buffer A) with 0,25% NP40 present during the first two of three washes. After the first homogenization, the he. mogenate was filtered through cheese-cloth to remove connective tissue, The nuclei were collected by centrifugation at 3500 rpm in an $$34 rotor for 10 rain. After the third wash, nuclei were resuspended in a small volume of buffer A and layered onto 50 mM Tris-HCI (pH 7,5)/150 mM KCI/5 mM MgCI2/0.7 M sucrose. The nuclei were collected by centrifugation at 4000 rpm in an SS34 rotor for 10 rain. DNase I digestions were carried out in 10 mM Tris-HC! (pH 7.5)/10 mM NaCl/4 mM MgCi2/10 mM sodium butyrate/0.25 M sucrose (RSB-0.25 M sucrose) essentially as described [12] at a concentration of approx. 1 m g / m l of DNA. Nuclei were digested for I to 2 h at 4°C with 250 ~ g / m l DNase l and at room temperature (23°C) with 100 ~ g / m l DNase I.
Preparation of the intermediate.salt~high.salt KCI matrix and nuclear pore-lamina complexes Nuclear matrices were prepared by sequential extractions with first 10 mM Tris-HCI (pH 7.4)/0.4 M KCI/0.2 mM MgC12/0.25 M sucrose and then 10 mM Tris-HCI (pH 7.4)/2.0 M KCI/0.2 mM MgC12/0.25 M sucrose extractions as described previously [12]. Nuclear pore-lamina complexes were prepared by digestion with DNase ! and RNase A and extraction with 1.6 M N a C I / l % v / v 2-mercaptoethanol as described [15], Ammonium sulfate matrices were prepared by DNase ! digestion and extraction with 0.2 M (NH4)2SO 4 as described [3] except that DNase I digestions were performed at 250/~g/ml DNase l in RSB0.25 M sucrose.
8.0)/150 mM NaCI/1.0 mM EDTA and incubated at 37°C for 60 min. Deacetylase activity was quantitated as described [12]. Background activity was determined from the amount of dpm released into the ethyl acetate phase using a boiled matrix fraction as an enzyme source and subtracted from the measured activity of each assayed sample. Substrate for histone deacetylase activity was prepared by labelling chicken immature erythrocytes with [all]acetic acid as described [12] except that labelling proceeded for 1 h and 10 mM sodium butyrate was present during the entire incubation period. Note that the same substrate preparation and quantity were used for all experiments. Thus, where given, measures of enzymatic hydrolysis can be directly compared. Results
Histone deacetylase activity is located with the htternal nuclear matrix of chicken liver We determined the nuclear distribution of histone deacetylase in chicken liver, a tissue in which the transcriptional activity is considerably higher than that of erythrocytes. The fractionation protocol shown in Fig. I was used throughout this study. Nuclear matrices isolated by this procedure were less susceptible to spurious association with proteins than other nuclear matrix preparations [12,16]. Chicken liver nuclei were digested for 1 h with DNase ! at room temperature (23°C) and then fractionated. Fig. 2 shows that for all the fractions the histone deacetylase activity was linear with increasing INTERMEDIATE - / H I G H - SA LT MATRIX NUCLEI JDNase I S
P 04M
0,4 M KCI
KCI
0.4 M KCI
S 2 0 M KCI
A n y for histone deacetylase activity Isolated soluble fractions were prepared for histone deacetylase assay by dialysis against deionized distilled H20. When a precipitate formed during dialysis, these samples were assayed as suspensions. The insoluble pellet fractions were resuspcnded in deionized distilled H20. A sample of 0.26 ml of each fraction was assayed in a final volume of 0.30 ml in 10 mM Tris-HCI (pH
,~.0 M KCl S
2 0 M KCI P (MATRIX)
Fig. 1. Isolationprocedure for intermediate-/high-saltnuclear matrix. Nuclei were isolated, digested with DNase I, and extracted sequentiallywith bufferscontaining0.4 M KCI and 2.0 M KCI. Four fractions were generated and assayed for deacetylase activity as describedin Materialsand Methods.
309 stabilize the nuclear matrix [3]. Fig. 3 shows that 70.7 + 4.5% (n = 3 separate experi~,~ents) of the total nuclear histone deacetylase acti,.,ty was found in the nuclear matrices (fraction 2.0 M KCI P, 23°t?) prepared from chicken liver nuclei digest~:d at room tcmperaturc with DNase !. There was a modest reduction in the amount of histone deacetylase ~ctivity (57.6 + 10.2%, n = 3) partitioning with nuclem matrices (fraction 2.0 M KC! P, 4°C) prepared from nuclei digested at 4°C consistent with the observation that increased temperature stabilizes the internal nuclear matrix [3]. We have previously shown that histone deacetylase is a component of the internal nuclear matrix but not of the nuclear pore-lamina complex of chicken erythrocyte [12]. Nuclear pore-lamina complexes were isolated from chicken liver nuclei according to the procedure of Kaufmann et al. [15]. Fig. 3 shows that 9l).8_+ 2.8~ (n = 2) of the total nuclear histone deacetylase activity was solubilized in the 1% 2-mercaptoethanol/l.6 M NaCI fraction. Less than 10% of the histonc deacety lase activity was retained by the nuclear pore-lamina complex (fraction 2-ME/NaCI P in Fig. 3). A gentle procedure for the isolation of nuclear matrices has recently been described using ammonium
protein concentrations up to between 1 to 2 mg protein per reaction. At higher concentrations, inhibition of deacetylase activity was seen. Additionally, Fig. 2 shows that the specific activity of histone deacetylase differed between the four fractions with the 0.4 M KCI/2.0 M KC! insoluble nuclear material (nuclear matrix) having the greatest specific activity (approx. 22500 dpm released/mg protein), followed by the 0.4 M KCI soluble fraction (approx. 4400 dpm released/rag protein), the digestion supernatant (approx. 4000 dpm released/mg protein) and finally the 2.0 M KCI soluble fraction contained the lowest specific activity (approx. 2300 dpm released/rag). Once the linear range of the enzyme activity of each fraction was established, we routinely assayed samples containing between 2(10 and 500 /zg of protein. This ensured an accurate assessment of both the enzyme activity in each fraction and the distribution of enzyme activity amongst the fractions. We prepared nuclear matrices following incubation of chicken liver nuclei with DNase ! at 40C and room temperature. Nuclease digestions at different temperatures were used because it has been suggested that elevated temperatures may alter the partitioning of material with the nuclear matrix [2] or alternatively
E ~ "t~
t
5
"
~'
f
I oOO[:O:4M'KCl s U,i,i -
Dig.sup.
4 ooo;
•
/
3000-
\]
¢
lOOOO-
""---]
g
' :::i2
£J)
in,|
-r
0.0
#3
.
•........
-L
0.5 1.0 1.5 Protein (rag)
J
2.0
::I:
P3
¢-L
8000U
_m O; i.
! 2.0 M KCI Soluble
~
.
oooo/ 4 000- j~ /
"0
0.0
2.5 5.0 7.5 10.0 20.C Protein (mg)
3.5000 . . . . w ' 0.4M/20MKCIInsoluble 1 30000 ~ i 25000
~o t~
2
?
i
t.
2000C
o
I 5000
t:n
10000
/ 1 I
< %)
I ZE O3
0.0
2.5
5.0 7.5 10.C 20.0 P r o t e i n (rag)
tn
5000
do
i
2
6
8
Protein (rag)
Fig. 2. q'he histone deacetylase assay a s a quantitative measure of histone deacclylase activity in nuclear subfractions. Isolated nuclear subfractions (chicken liver, 4°C digest) were characterized over a range of protein concentrations. Protein content in mg represents the amount o1" protein present in the 300 #l assay volume. [3H]acetyl groups released represents the dpm contained in one-half of the volume of ethyl acetate used to extract the free acetate enzymatically released from a labelled histone substrate. The fraction headings for each of the graphs correspond to the fractions outlined in the flow chart in Fig. I.
310
r-o
"6 u B.
00~
• '~. o o
a~'c
r~
.x" .& 0 O
~g~:
4 "c
~e~
0
mm
~g
Eg
4"c
23"c
4"c
I i
0
i~i~
0
b~
0
Fig. 3. Histoue deacetylase activity partitions with the internal nuclear matrix of chicken liver. Isolated chicken liver nuclei prepared by digestion with DNase I at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. ! or alternatively according to the procedures described by Kaufmann et al. [15] and Bclgrader et al. [3]. 2-ME/NaCI S and P represent material soluhilized by extraction with 1% v/v 2-mercaptoethanol/I.6 M NaCI (which includes the internal nuclear matrix) and the insoluble material (the nuclear Ix~re-lamina complex), respectively. AS S represents the material solubilized b} extraction with 0.2 M Jmmonium sulfate, and AS P represents the residual material insoluble in 0.2 M ammonium sulfate (the ammonium sulfate nuclear matrix). Between 200 and 500 #g protein per reaction was assayed and ;~he relative distribution of histone deacetylase activity was determined by calculating the total activity in each subnuclear fraction. Each value represents the mean + S.E. from the mean from two or more experiments.
The association ~f the histone deacetylase activity with the chicken erythrocyte internal nuclear matrix is dependent upon temperature When chicken erythrocyte nuclear matrices were prepared by digestion with nucleases at room temperature or 37°C using a variety of protocols, histone deacetylase was associated with t h e internal nuclear matrix [12]. However, chicken erythrocyte nuclear matrices prepared from nuclei incubated with DNase I at 4°C lack an internal nuclear matrix that can be visualized by electron microscopy [13]. We addressed this discrepancy by using histone deacetylase as a marker enzyme for the partitioning of the internal nuclear matrix during nuclear fractionation. Nuclei were digested with DNase 1 for 1 h at 4°C or 23°C. Fig. 4 shows the results obtained from three separate experiments done at each temperature. When digestions were performed at 23°C with DNase I, approx. 50.5 + 4.4% of the total nuclear histone deacetylase activity in chicken erythrocyte partitioned with the 0.4 M KCi/2.0 M KC! insoluble material (nuclear matrix). In contrast, when digestions were performed with DNase I at 4°C, over 79.5 + 3.3% of the total nuclear deacetylase activity was solubilized by extraction of nuclei with 0.4 M 100 F Chicken
Immature
Erythrocyte
90"
~o~
J-
70, ~3
sulfate to extract nuclease digested nuclei [3]. This procedure results in less aggregation and an apparently more preserved nuclear morphology relative to the high salt extraction procedures typ,.',:ally used while extracting 98% of the total DNA, 78% of the total protein and 73% of the total RNA [3]. Additionally, pzei~cubation of nuclei at 37°C results in a stabilization of the internal nuclear matrix prepared under these conditions [3]. This pretreatment results in a quantitatively improved recovery of internal nuclear matrix proteins with the insoluble material but no major changes in the spectrum of proteins associated with this structure. Chicken liver nuclei preincubated in buffer A for I h at 37°C or 4°C were digested with DNase 1 for one-half hour at 4°C in RSB-0.25 M sucrose and then extracted with 0.2 M ammonium sulfate. Fig. 3 shows that more than g0% of the total nuclear histone deacetylase activity was associated with this nuclear matrix preparation (fraction 0.2 M AS P). These observations suggest that the majority, if not all, of the histone deacetylase activity is localized with the imerna! nuclear matrix of chicken liver. Thus, histone deaeetylase is a nuclear martin.
C
0
60~50 40
0
30 20 i_
~0 O 0
0
tO
t~
0
0
~
r~
c I-
2 3 °C
4 °C
Fig. 4. The nuclear matrix of chicken immature erythrocyte is destabilized by digestion at 4°C. Isolated chicken immature erythrocyte nuclei prepared by digestion with DNase ! at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. 1. Between 200 and 500 #g per reaction was assayed and the relative distribution of histone deacetylase activity was calculated frorl the total activity in each subnuclear fraction. ND represents no,~-detectable activity for the assayed fraction. Each value represen,ts the mean _+S.E. from the mean from three experiments.
311 KCi-containing buffer. A further 14.5 + 0.7% of the total activity was extracted in 2.0 M KCl-containing buffer. Less than 6% of the histone deacetylase activity remained associated with the 0.4 M KCI/2.0 M KCI insoluble material. These observations suggest that the internal nuclear matrix of chicken erythrocyte is a fragile structure and that incubations at low temperature are sufficient to prepare nuclear matrices devoid of the internal nuclear matrix component. This result is consistent with the electron microscopy studies [13]. Histone deacetylase is a nuclear matrin in trout fleer We next determined whether the association of the histone deacetylase activity with the internal nuclear matrix was species-specific. Nuclear matrices were prepared from trout liver nuclei that were incubated with DNase I at either 4°C or room temperature. Fig. 5 shows that, regardless of the digestion conditions used, the majority (84.8 5: 2.5% at room tetnperature and 71.2 :I: 5.0% at 4*C) of the histone deacetylase activity partitioned with the nuclear matrix (fraction 2.0 M KCI P). Further, trout liver nuclear pore-lamina complexes retained a low amount of the total histone deacetylase activity, with approx. 98% of the enzyme activity being present in the I% 2-mercaptoethanol/l.6 M NaCI 100 Trout
~
9O
~g
80
$o
7o t
Liver
90
;
1~J
I
~c•o
O.~:-
Q
/:
~5
3o '
L,j z.
20-
m
~L t~l
Ix
0
0
,n
.
23"C
,t~
.
CD
0
~)
4 *C
Fig. 6. The nuclear matrix in irout hepatocellular carcinonla is resistant to low temperature destabilization. Trout aflatoxin Bt-induced hepatocellular carcinoma nuclei that were incubaled with DNase 1 at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. 1, Between 200 and 501) #g per reaction was assayed and the relative distribution of histone deacetylase activity was determined by extrapolating the total activity in each subnuclear fraction. Each value represents the mean _+SE. from tbt. mean from two fractionations wilh individual tumors.
extract (not shown). Thus, we conclude that the association of histone deacetylase with the internal nuclear matrix is not a species-specific phenomenon.
4o) 30"
~
.":
--~
~4epatocel [U',QF CQr'ctnori~a
T
1
~-
....
! 00 F :~
20~
Ill
o L
~ ,7,
.~
o
o
,~
,,.
0
o
c
2 3 *C
4°C
Fig. 5. Histone deacetylase is associated with the nuclear matrix of trout liver. Isolated trout liver nuclei prepared by digestion with DNase I at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. I. Between 200 and 500 tzg per reaction was ~ssayed and the relative distribution of histone deacetylose activity was calculated from the total activity in each subnuclear fraction. Each value represents the mean + S.E. from the mean from two experiments.
Neoplastic transformation of trout lirer does not alter the association of histone deacetylase with the nuclear matrix Neoplastic transformation of a cell is a consequence of the abnormal expression of certain genes. The aberrant expression of these genes may be a result of modification at the gene level a n d / o r changes in chromatin and nuclear structure. Transformed cells have been shown to have altered nuclear matrix composition [17-19]. We investigated whether neoplastic transformation of trout liver altered the association of the histone deacetylase with the nuclear matrix. Aflatoxin Brinduced trout hepatocellular carcinomas differ from the parent trout liver tissue in that they have an altered H1 subtype composition, more nuclease sensitive chromatin, and moderately or greatly enlarged hyperchromatic nuclei [20,21]. Nuclei isolated from individual hepatocellular carcinomas were incubated with DNase 1 at 23°C and 4°C. Fig. 6 shows that the distribution of histone deacetylase activity amongst the various nuclear subfractions from hepatocellu!ar carcinoma and
312 liver was similar. Of the total histone deacetylase activity, 79.8 + 7.9% (n = 2) and 69.5 + 8.5% (n = 2) was associated with the 0.4 M KCI/2.0 M KCI insoluble material (tumor nuclear matrix) at room temperature and 4°C, respectively, The partitionings of the histone deacetylase activity among the nuclear fractions of the individual tumors and livers were similar. Discussion The internal matrix consists of a class of proteins called nuclear matrins. Recently, Nakayasu and Berezney [22] used indirect immunofluoresccncc microscopy to show that nuclear matrins are organized as a tightly packed network of fibrogranular structures that arc localized in the nuclear interior of intact cells, in this study, we demonstrate that greater than 85% of the total nuclear histone deacetylase activity of chicken and trout liver is localized with the internal nuclear matrix, Thus, the a~ociation of the histoue deacetylase with the internal nuclear matrix is neither tissue-specific nor species.specific. It is concluded from these studies that histone deacetylas¢ is a nuclear matrin. Furthermore, this study substantiates the use of histone deacetylase as a marker enzyme for the internal nuclear matrix, Because of the ease of performing this enzyme agsay, histone deacetylase activity may be a useful and valuable marker when it is of interest to know the partitioning of the internal nuclear matrix during nuclear fractionation procedures. it has been reported that chicken erythrocyte nuclei do not have an internal nuclear matrix [13]. This study provides biochemical evidence that there is an internal nuclear matrix in chicken ¢rythrocytc nuclei. However, the presence of this structure in nuclear matrix preparations is dependent upon temperature. When chicken erythrocyte nuclei are digested with DNase 1 at 4°C, histone deacetylase activity is almost quantitatively solubilized by salt extraction. In contrast, when chicken erythrocyte nuclei arc digested with DNasc i at 23°C, approx, 50% of the total nuclear histone deacctylase activity partitions with the nuclear matrix. Temperature is an important parameter when preparing nuclear matrices. There is considerable evidence that the internal nuclear matrix is stabilized by treating nuclei at 37°C before nuclcase treatment (reviewed in Refs, 2 and 3). Importantly, this temperature-induced stabilization causes quantitative but not qualitative changes in the compositioa of the internal nuclear matrix [3]. Our results indicate that exposure of nuclei to room temperature conditions also results in the stabilization of the internal nuclear matrix. There was a modest, but consistent, increase in the histone deacetylase activity partitioning with liver nuclear matrices isolated from nuclei incubated at 23°C versus
4°C. Based upon the partitioning of the histone deacetylase activity, the effect of temperature-induced stabilization of the internal nuclear matrix was most marked for chicken erythrocytes, Our data suggest that internal nuclear matrices cannot be prepared in crythrocytes without such temperature-induced stabilization. it should be noted, however, that temperature-induced alterations in nuclear matrix integrity may be restricted to nuclei isolated under hypotonic conditions
[23]. Using histone deacetylase as a marker enzyme for the internal nuclear matrix, we determined whether neoplastic transformation of trout hepatocytes was accompanied by an alteration in the structural integrity of the internal nuclear matrix, Previous studies have noted changes in the nuclear matrix of transformed cells relative to their normal counterparts [17-19]. Our results provide evidence that the structure of the hepatocellular internal nuclear matrix was not altered in trout hepatocellular carcinomas. • Our observations that the majority of the total nuclear histone deacetylase activity of chicken and trout liver partitions with the internal nuclear matrix suggest that histone deacetylation is restricted to the liver internal nuclear matrix. Preliminary results suggest that histone acetyltransferase is also associated with the internal nuclear matrix of chicken erythrocytes and liver (M.J. Hendzel, P.N. Lewis and J.R. Davie, unpublished results). We and others have shown that dynamically acetylated histones, which attain high acetylation levels and are rapidly deacetylated, are complexed principally to transcriptionally active DNA [12,14,2429]. Furthermore, we observed that (!) the chromatin distribution of dynamically acetylated histones did not parallel that of the competent DNA sequences (i.e., DNase I-sensitive, non-transcribed chromatin) and (2) the majority of dynamically acetylated histones were found associated with the chromatin fragments of the low salt insoluble nuclear material [12]. Transcriptionally active gene chromatin also has an insoluble nature [14,30,31]. For example, Stratling et al. [31] demonstrated a transcription-dependent partitioning of lysozyme coding sequences with the low salt insoluble nuclear material. We propose that regions of chromatin with dynamically acetylated histones have an insoluble nature due to the dynamic interaction of the histones with the internal nuclear matrix-bound histone deacetylase and acetyltransferase, in this model, the preferential association of dynamically acetylated histones with transcriptionally active, but not competent or repressed, DNA results in the selective partitioning of transcriptionally active gene chromatin fragments with the low salt insoluble nuclear material. The juxtapositioning of the transcriptionally active gene chromatin domains may facilitate nuclear processes such as transcription and RNA processing [4,5].
313
Acknowledgements This project was supported by grants from the Medical Research Council of Canada and the National Cancer Institute of Canada, and by the awards of a Medical Research Council Studentship to M.J. Hendzel and a Medical Research Council Scientist to J.R. Davie. References I Nelson, W.G., Pienta, K.J., Barrack, E.R. and Coffey, D.S. 119801 Annu. Rev. Biophys. Biophys. Chem, 15, 457-475. 2 Kaufmann, S.H. and Shaper, J.H. (19911 Exp. Cell Res, 192, 51 ! -523. 3 Belgrader, P., Siegel. A.J. and Berezney, R. (1991)J. Cell Sci. 98, 281-291. 4 Jackson, D.A. (1991) Bioessays 13, I - IfI. 5 Vcrheijen R., Van Venrooij, W. and Ram:tekers, F. (198X)J. Cell Sci. 90, I 1-36. 6 Gasser. S.M. and Lacmmli, tJ.K. 119871 Trends Genct. 3. 16-22. 7 Goldman, M.A. (1988) Bioessays 9, 50-55. 8 Cook, P.R. 119891 Eur, J, Biochem. 185, 487-501. 9 Phi-Van, L., Von Kries, J.P., Ostertag, W. and Stratling., W.lt. (19901 Mol. Ceil. Biol. 10, 23112-23117. 10 Von Kries, J.P., Buhrmester, H. and Stmtling, W.H. (19911 Cell 64. 123-135. II Bodnar, J.W. 119881 J. Theor. Biol. 132, 479-507. 12 Hendzel, M.J., Delcuve, G.P. and Davie, J.R. 119911 J. Biol. Chem. 266, 21936-21942. 13 Lafond, R.E. and Woodcock, C.L.F. 0983) Exp. C¢1~ Res. 147, 31-39.
14 Delcuve, G.P. and Davie, J.R. 11989) Biochem. J. 263, 179-186. 15 Kaufmann, S.H., Gibson. W. and Shaper. J.t]. (i9831 J. Biol. Chem. 258, 2710-2719. 16 Roberge, M., Dahmus, M.E. and Bradhury, E.M. (1988) J. Mol. Biol. 2{ll. 545-555. 17 Fey, E.G. and Penman, S. (1988) Proc. Natl. Acad. Sci. USA 85. 121-125. 18 Brancolini, C, and Schneider, C. 119911 Proc. Natl. Acad, Sci. USA 88, 6936-69411. 19 Kuzmina, S.N., Buldyaeva, T.V., Akopov, S.B. and Zharsky. I.B. 119841 Mol. Cell. Biochem. 58. 183-180. 211 Davie, J.R. and Delcuve. G.P. 11991) Bioehem. J. 280, 491-497. 21 Sinnhuber, R.O., Hendricks, J.D., Wales, J.H. and Putnam, G.B. (19771 Ann. NY Acad. Sci. 298, 389-4118. 22 Nakayasu, Ft. and Berezney, R. (1991) Proc. Natl. Acad. Sci. USA 88. 1(1312-10316. 23 Boyle. W.J. and Baluda, M.A. 0987) Mul. ('ell. Biol. 7. 3345-3348. 24 Zhang, D.-E. and Nelson. D.A, (19881 Biochem. J. 250, 241-247. 25 Zhang, D.-E. and Nelson. D.A. 11988) Bi~chem. J. 250, 233-2411, 26 Ridsdale, J.A., Hendzel, M.J., Delcuve. G.P. and Da~,'ie, J.R. (1990) J. Biol. Chem. 265. 51511-5151~. 27 ip. Y.T., Jackson. V., Meier, J. and Chalkley, R. 11988) J. Biol. Chem. 263, 14044-141152. 2~ lteld~cs. T.W., Thornc, A.W. and Crane-Robinson. C. (198~1 EMBO ,I. 7, 1395-14112. 29 Turner, B.M. 119911J. ('ell Sci. 99, 13-211. 3(I Gro~is, D.S. and Garrard, W.T. (1987) Trends Biochem. Sci. 12, 293-297 31 Stratling, W.li., Dolle, A. and Sippel. A.E. 1198111 BiochemistD 25, 495-51~2.
307
BBAEXP 92361
Nuclear distribution of histone deacetylase: a marker enzyme for the internal nuclear matrix M i c h a e l J. H e n d z e l a n d J a m e s R. D a v i e Department of Biochemistry and Molecldar Biology, FacuJJyof Medicble. Unirersityof Manitoba, Winnipeg, Manitoba (Canada) (Received 25 September 1991)
Key words: Hislone deacetylase; Internal nuclear matrix; Histone acetylation Nuclear matrins are proteins that localize to the internal nuclear matrix, in a previous study, we reported that histone deacetylase is a component of the internal matrix, suggesting that histone dcaeetylase is a nuclear matrin. Here, we demonstrate that the majority of the histone deacetylase activity is associated with the internal nuclear matrices of chicken and trout liver. Thus, the association of the histone deacetylase with the internal nuclear matrix is neither tissue- nor species-s,Jecific. Using historic deacetylase as a marker enzyme for the partitioning of the internal nuclear matrix during nuclear fractionations, we show that in contrast to the internal nuclear matrices of trout liver, trout hepatocellular carcinoma and chicken liver, the stability of the chicken erythrocyte internal nuclear matrix is temperature-dependent. Our results support a model that has the histone deaeetylase mediating transient interactions between the internal nuclear matrix and chromatin regions undergoing dynamic acetylation, for example transcriptionally active chromatin regions.
Introduction The nuclear matrix is an operationally defined structure typically prepared by high salt extraction of isolated nuclei following extensive mtclease digestion. The insoluble material is composed of residual nucleoli, the nuclear pore-lamina complex, and the internal nuclear matrix [1]. Many nuclear metabolic processes are associated with nuclear matrix preparations, including replication, transcription, RNA splicing and topoisomerase activity [1-5]. The nuclear matrix is thought to represent a structural framework for nuclear/chromatin organization [6]. This model of the nucleus views topologically independent chromatin domains as the units of regulation [4,6-.8]. These domains may contain one to several genes and are fixed to the nuclear matrix at both ends through DNA sequences termed nuclear matrix attachments (MARs) (reviewed in Refs. 4,6). Specific nuclear matrins, which are proteins of the internal nuclear matrix, are thought to interact with MARs (e.g., attachment region binding protein) [9,10]. Current evidence suggests that MARs are stable attachment sites [6]. In addition, there exist dynamic
Correspondence: J.R. Davie, Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Manitoba, 770 Bannatyne Ave., Winnipeg, Manitoba R3E 0W3, Canada.
interactions between transcriptionally active gene chromatin regions and the nuclear matrix [11]. Recently, we demonstrated that approx. 50% of the total nuclear histone deacetylase activity is associated with the internal nuclear matrix of chicken erythrocyte [12]. EM studies, however, have questioned the presence of an internal nuclear matrix in chicken erythrocyte. Lafond and Woodcock [13] used 'mild' isolation procedures in an attempt to visualize the nuclear matrix of the chicken mature erythrocyte. These conditions involved DNase I digestion at 4°C and a single extraction of the nuclease-digested nuclei with a 2 M NaCl-containing buffer. Upon visual inspection of these preparations by electron microscopy, they found only the nuclear pore-lamina complex and no evidence for an internal nuclear matri~ [13]. Biochemical studies, in contrast, demonstrate the presence of internal nuclear matrix proteins in chicken e rythrocytes [9,10,12]. In this report, we determined the tissue- and species-specificity of the association of histone deacetylase with the internal nuclear matrix. We show that the majority of the histone deacetylase activity of chicken and trout liver is found with the internal nuclear matrix, demonstrating that histone deacetylase is a nuclear matrin and a marker enzyme for the internal nuclear matrix. By analyzing the activity of the histone deacetylase in various nuclear matrix preparations, we demonstrate that first, the stability of the internal
3~ nuclear matrix of chicken erythrocytes is temperaturedependent and second, that neoplastic transformation of trout hepatocytes does not alter the stability of the internal nuclear matrix. Materials and Methods
Isolation of nuclei Chicken immature erythrocyte nuclei were isolated as described by Delcuve and Davie [14]. Trout liver, trout hepatocellular carcinoma and chicken liver nuclei were prepared by homogenization in 10 mM Pipes/1.0 M hexylene glycol/l% v / v thiodiglycol/2 mM MgCi,/30 mM sodium butyrate (pH 7.0) (buffer A) with 0,25% NP40 present during the first two of three washes. After the first homogenization, the he. mogenate was filtered through cheese-cloth to remove connective tissue, The nuclei were collected by centrifugation at 3500 rpm in an $$34 rotor for 10 rain. After the third wash, nuclei were resuspended in a small volume of buffer A and layered onto 50 mM Tris-HCI (pH 7,5)/150 mM KCI/5 mM MgCI2/0.7 M sucrose. The nuclei were collected by centrifugation at 4000 rpm in an SS34 rotor for 10 rain. DNase I digestions were carried out in 10 mM Tris-HC! (pH 7.5)/10 mM NaCl/4 mM MgCi2/10 mM sodium butyrate/0.25 M sucrose (RSB-0.25 M sucrose) essentially as described [12] at a concentration of approx. 1 m g / m l of DNA. Nuclei were digested for I to 2 h at 4°C with 250 ~ g / m l DNase l and at room temperature (23°C) with 100 ~ g / m l DNase I.
Preparation of the intermediate.salt~high.salt KCI matrix and nuclear pore-lamina complexes Nuclear matrices were prepared by sequential extractions with first 10 mM Tris-HCI (pH 7.4)/0.4 M KCI/0.2 mM MgC12/0.25 M sucrose and then 10 mM Tris-HCI (pH 7.4)/2.0 M KCI/0.2 mM MgC12/0.25 M sucrose extractions as described previously [12]. Nuclear pore-lamina complexes were prepared by digestion with DNase ! and RNase A and extraction with 1.6 M N a C I / l % v / v 2-mercaptoethanol as described [15], Ammonium sulfate matrices were prepared by DNase ! digestion and extraction with 0.2 M (NH4)2SO 4 as described [3] except that DNase I digestions were performed at 250/~g/ml DNase l in RSB0.25 M sucrose.
8.0)/150 mM NaCI/1.0 mM EDTA and incubated at 37°C for 60 min. Deacetylase activity was quantitated as described [12]. Background activity was determined from the amount of dpm released into the ethyl acetate phase using a boiled matrix fraction as an enzyme source and subtracted from the measured activity of each assayed sample. Substrate for histone deacetylase activity was prepared by labelling chicken immature erythrocytes with [all]acetic acid as described [12] except that labelling proceeded for 1 h and 10 mM sodium butyrate was present during the entire incubation period. Note that the same substrate preparation and quantity were used for all experiments. Thus, where given, measures of enzymatic hydrolysis can be directly compared. Results
Histone deacetylase activity is located with the htternal nuclear matrix of chicken liver We determined the nuclear distribution of histone deacetylase in chicken liver, a tissue in which the transcriptional activity is considerably higher than that of erythrocytes. The fractionation protocol shown in Fig. I was used throughout this study. Nuclear matrices isolated by this procedure were less susceptible to spurious association with proteins than other nuclear matrix preparations [12,16]. Chicken liver nuclei were digested for 1 h with DNase ! at room temperature (23°C) and then fractionated. Fig. 2 shows that for all the fractions the histone deacetylase activity was linear with increasing INTERMEDIATE - / H I G H - SA LT MATRIX NUCLEI JDNase I S
P 04M
0,4 M KCI
KCI
0.4 M KCI
S 2 0 M KCI
A n y for histone deacetylase activity Isolated soluble fractions were prepared for histone deacetylase assay by dialysis against deionized distilled H20. When a precipitate formed during dialysis, these samples were assayed as suspensions. The insoluble pellet fractions were resuspcnded in deionized distilled H20. A sample of 0.26 ml of each fraction was assayed in a final volume of 0.30 ml in 10 mM Tris-HCI (pH
,~.0 M KCl S
2 0 M KCI P (MATRIX)
Fig. 1. Isolationprocedure for intermediate-/high-saltnuclear matrix. Nuclei were isolated, digested with DNase I, and extracted sequentiallywith bufferscontaining0.4 M KCI and 2.0 M KCI. Four fractions were generated and assayed for deacetylase activity as describedin Materialsand Methods.
309 stabilize the nuclear matrix [3]. Fig. 3 shows that 70.7 + 4.5% (n = 3 separate experi~,~ents) of the total nuclear histone deacetylase acti,.,ty was found in the nuclear matrices (fraction 2.0 M KCI P, 23°t?) prepared from chicken liver nuclei digest~:d at room tcmperaturc with DNase !. There was a modest reduction in the amount of histone deacetylase ~ctivity (57.6 + 10.2%, n = 3) partitioning with nuclem matrices (fraction 2.0 M KC! P, 4°C) prepared from nuclei digested at 4°C consistent with the observation that increased temperature stabilizes the internal nuclear matrix [3]. We have previously shown that histone deacetylase is a component of the internal nuclear matrix but not of the nuclear pore-lamina complex of chicken erythrocyte [12]. Nuclear pore-lamina complexes were isolated from chicken liver nuclei according to the procedure of Kaufmann et al. [15]. Fig. 3 shows that 9l).8_+ 2.8~ (n = 2) of the total nuclear histone deacetylase activity was solubilized in the 1% 2-mercaptoethanol/l.6 M NaCI fraction. Less than 10% of the histonc deacety lase activity was retained by the nuclear pore-lamina complex (fraction 2-ME/NaCI P in Fig. 3). A gentle procedure for the isolation of nuclear matrices has recently been described using ammonium
protein concentrations up to between 1 to 2 mg protein per reaction. At higher concentrations, inhibition of deacetylase activity was seen. Additionally, Fig. 2 shows that the specific activity of histone deacetylase differed between the four fractions with the 0.4 M KCI/2.0 M KC! insoluble nuclear material (nuclear matrix) having the greatest specific activity (approx. 22500 dpm released/mg protein), followed by the 0.4 M KCI soluble fraction (approx. 4400 dpm released/rag protein), the digestion supernatant (approx. 4000 dpm released/mg protein) and finally the 2.0 M KCI soluble fraction contained the lowest specific activity (approx. 2300 dpm released/rag). Once the linear range of the enzyme activity of each fraction was established, we routinely assayed samples containing between 2(10 and 500 /zg of protein. This ensured an accurate assessment of both the enzyme activity in each fraction and the distribution of enzyme activity amongst the fractions. We prepared nuclear matrices following incubation of chicken liver nuclei with DNase ! at 40C and room temperature. Nuclease digestions at different temperatures were used because it has been suggested that elevated temperatures may alter the partitioning of material with the nuclear matrix [2] or alternatively
E ~ "t~
t
5
"
~'
f
I oOO[:O:4M'KCl s U,i,i -
Dig.sup.
4 ooo;
•
/
3000-
\]
¢
lOOOO-
""---]
g
' :::i2
£J)
in,|
-r
0.0
#3
.
•........
-L
0.5 1.0 1.5 Protein (rag)
J
2.0
::I:
P3
¢-L
8000U
_m O; i.
! 2.0 M KCI Soluble
~
.
oooo/ 4 000- j~ /
"0
0.0
2.5 5.0 7.5 10.0 20.C Protein (mg)
3.5000 . . . . w ' 0.4M/20MKCIInsoluble 1 30000 ~ i 25000
~o t~
2
?
i
t.
2000C
o
I 5000
t:n
10000
/ 1 I
< %)
I ZE O3
0.0
2.5
5.0 7.5 10.C 20.0 P r o t e i n (rag)
tn
5000
do
i
2
6
8
Protein (rag)
Fig. 2. q'he histone deacetylase assay a s a quantitative measure of histone deacclylase activity in nuclear subfractions. Isolated nuclear subfractions (chicken liver, 4°C digest) were characterized over a range of protein concentrations. Protein content in mg represents the amount o1" protein present in the 300 #l assay volume. [3H]acetyl groups released represents the dpm contained in one-half of the volume of ethyl acetate used to extract the free acetate enzymatically released from a labelled histone substrate. The fraction headings for each of the graphs correspond to the fractions outlined in the flow chart in Fig. I.
310
r-o
"6 u B.
00~
• '~. o o
a~'c
r~
.x" .& 0 O
~g~:
4 "c
~e~
0
mm
~g
Eg
4"c
23"c
4"c
I i
0
i~i~
0
b~
0
Fig. 3. Histoue deacetylase activity partitions with the internal nuclear matrix of chicken liver. Isolated chicken liver nuclei prepared by digestion with DNase I at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. ! or alternatively according to the procedures described by Kaufmann et al. [15] and Bclgrader et al. [3]. 2-ME/NaCI S and P represent material soluhilized by extraction with 1% v/v 2-mercaptoethanol/I.6 M NaCI (which includes the internal nuclear matrix) and the insoluble material (the nuclear Ix~re-lamina complex), respectively. AS S represents the material solubilized b} extraction with 0.2 M Jmmonium sulfate, and AS P represents the residual material insoluble in 0.2 M ammonium sulfate (the ammonium sulfate nuclear matrix). Between 200 and 500 #g protein per reaction was assayed and ;~he relative distribution of histone deacetylase activity was determined by calculating the total activity in each subnuclear fraction. Each value represents the mean + S.E. from the mean from two or more experiments.
The association ~f the histone deacetylase activity with the chicken erythrocyte internal nuclear matrix is dependent upon temperature When chicken erythrocyte nuclear matrices were prepared by digestion with nucleases at room temperature or 37°C using a variety of protocols, histone deacetylase was associated with t h e internal nuclear matrix [12]. However, chicken erythrocyte nuclear matrices prepared from nuclei incubated with DNase I at 4°C lack an internal nuclear matrix that can be visualized by electron microscopy [13]. We addressed this discrepancy by using histone deacetylase as a marker enzyme for the partitioning of the internal nuclear matrix during nuclear fractionation. Nuclei were digested with DNase 1 for 1 h at 4°C or 23°C. Fig. 4 shows the results obtained from three separate experiments done at each temperature. When digestions were performed at 23°C with DNase I, approx. 50.5 + 4.4% of the total nuclear histone deacetylase activity in chicken erythrocyte partitioned with the 0.4 M KCi/2.0 M KC! insoluble material (nuclear matrix). In contrast, when digestions were performed with DNase I at 4°C, over 79.5 + 3.3% of the total nuclear deacetylase activity was solubilized by extraction of nuclei with 0.4 M 100 F Chicken
Immature
Erythrocyte
90"
~o~
J-
70, ~3
sulfate to extract nuclease digested nuclei [3]. This procedure results in less aggregation and an apparently more preserved nuclear morphology relative to the high salt extraction procedures typ,.',:ally used while extracting 98% of the total DNA, 78% of the total protein and 73% of the total RNA [3]. Additionally, pzei~cubation of nuclei at 37°C results in a stabilization of the internal nuclear matrix prepared under these conditions [3]. This pretreatment results in a quantitatively improved recovery of internal nuclear matrix proteins with the insoluble material but no major changes in the spectrum of proteins associated with this structure. Chicken liver nuclei preincubated in buffer A for I h at 37°C or 4°C were digested with DNase 1 for one-half hour at 4°C in RSB-0.25 M sucrose and then extracted with 0.2 M ammonium sulfate. Fig. 3 shows that more than g0% of the total nuclear histone deacetylase activity was associated with this nuclear matrix preparation (fraction 0.2 M AS P). These observations suggest that the majority, if not all, of the histone deacetylase activity is localized with the imerna! nuclear matrix of chicken liver. Thus, histone deaeetylase is a nuclear martin.
C
0
60~50 40
0
30 20 i_
~0 O 0
0
tO
t~
0
0
~
r~
c I-
2 3 °C
4 °C
Fig. 4. The nuclear matrix of chicken immature erythrocyte is destabilized by digestion at 4°C. Isolated chicken immature erythrocyte nuclei prepared by digestion with DNase ! at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. 1. Between 200 and 500 #g per reaction was assayed and the relative distribution of histone deacetylase activity was calculated frorl the total activity in each subnuclear fraction. ND represents no,~-detectable activity for the assayed fraction. Each value represen,ts the mean _+S.E. from the mean from three experiments.
311 KCi-containing buffer. A further 14.5 + 0.7% of the total activity was extracted in 2.0 M KCl-containing buffer. Less than 6% of the histone deacetylase activity remained associated with the 0.4 M KCI/2.0 M KCI insoluble material. These observations suggest that the internal nuclear matrix of chicken erythrocyte is a fragile structure and that incubations at low temperature are sufficient to prepare nuclear matrices devoid of the internal nuclear matrix component. This result is consistent with the electron microscopy studies [13]. Histone deacetylase is a nuclear matrin in trout fleer We next determined whether the association of the histone deacetylase activity with the internal nuclear matrix was species-specific. Nuclear matrices were prepared from trout liver nuclei that were incubated with DNase I at either 4°C or room temperature. Fig. 5 shows that, regardless of the digestion conditions used, the majority (84.8 5: 2.5% at room tetnperature and 71.2 :I: 5.0% at 4*C) of the histone deacetylase activity partitioned with the nuclear matrix (fraction 2.0 M KCI P). Further, trout liver nuclear pore-lamina complexes retained a low amount of the total histone deacetylase activity, with approx. 98% of the enzyme activity being present in the I% 2-mercaptoethanol/l.6 M NaCI 100 Trout
~
9O
~g
80
$o
7o t
Liver
90
;
1~J
I
~c•o
O.~:-
Q
/:
~5
3o '
L,j z.
20-
m
~L t~l
Ix
0
0
,n
.
23"C
,t~
.
CD
0
~)
4 *C
Fig. 6. The nuclear matrix in irout hepatocellular carcinonla is resistant to low temperature destabilization. Trout aflatoxin Bt-induced hepatocellular carcinoma nuclei that were incubaled with DNase 1 at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. 1, Between 200 and 501) #g per reaction was assayed and the relative distribution of histone deacetylase activity was determined by extrapolating the total activity in each subnuclear fraction. Each value represents the mean _+SE. from tbt. mean from two fractionations wilh individual tumors.
extract (not shown). Thus, we conclude that the association of histone deacetylase with the internal nuclear matrix is not a species-specific phenomenon.
4o) 30"
~
.":
--~
~4epatocel [U',QF CQr'ctnori~a
T
1
~-
....
! 00 F :~
20~
Ill
o L
~ ,7,
.~
o
o
,~
,,.
0
o
c
2 3 *C
4°C
Fig. 5. Histone deacetylase is associated with the nuclear matrix of trout liver. Isolated trout liver nuclei prepared by digestion with DNase I at 4°C or at room temperature were fractionated according to the procedure outlined in Fig. I. Between 200 and 500 tzg per reaction was ~ssayed and the relative distribution of histone deacetylose activity was calculated from the total activity in each subnuclear fraction. Each value represents the mean + S.E. from the mean from two experiments.
Neoplastic transformation of trout lirer does not alter the association of histone deacetylase with the nuclear matrix Neoplastic transformation of a cell is a consequence of the abnormal expression of certain genes. The aberrant expression of these genes may be a result of modification at the gene level a n d / o r changes in chromatin and nuclear structure. Transformed cells have been shown to have altered nuclear matrix composition [17-19]. We investigated whether neoplastic transformation of trout liver altered the association of the histone deacetylase with the nuclear matrix. Aflatoxin Brinduced trout hepatocellular carcinomas differ from the parent trout liver tissue in that they have an altered H1 subtype composition, more nuclease sensitive chromatin, and moderately or greatly enlarged hyperchromatic nuclei [20,21]. Nuclei isolated from individual hepatocellular carcinomas were incubated with DNase 1 at 23°C and 4°C. Fig. 6 shows that the distribution of histone deacetylase activity amongst the various nuclear subfractions from hepatocellu!ar carcinoma and
312 liver was similar. Of the total histone deacetylase activity, 79.8 + 7.9% (n = 2) and 69.5 + 8.5% (n = 2) was associated with the 0.4 M KCI/2.0 M KCI insoluble material (tumor nuclear matrix) at room temperature and 4°C, respectively, The partitionings of the histone deacetylase activity among the nuclear fractions of the individual tumors and livers were similar. Discussion The internal matrix consists of a class of proteins called nuclear matrins. Recently, Nakayasu and Berezney [22] used indirect immunofluoresccncc microscopy to show that nuclear matrins are organized as a tightly packed network of fibrogranular structures that arc localized in the nuclear interior of intact cells, in this study, we demonstrate that greater than 85% of the total nuclear histone deacetylase activity of chicken and trout liver is localized with the internal nuclear matrix, Thus, the a~ociation of the histoue deacetylase with the internal nuclear matrix is neither tissue-specific nor species.specific. It is concluded from these studies that histone deacetylas¢ is a nuclear matrin. Furthermore, this study substantiates the use of histone deacetylase as a marker enzyme for the internal nuclear matrix, Because of the ease of performing this enzyme agsay, histone deacetylase activity may be a useful and valuable marker when it is of interest to know the partitioning of the internal nuclear matrix during nuclear fractionation procedures. it has been reported that chicken erythrocyte nuclei do not have an internal nuclear matrix [13]. This study provides biochemical evidence that there is an internal nuclear matrix in chicken ¢rythrocytc nuclei. However, the presence of this structure in nuclear matrix preparations is dependent upon temperature. When chicken erythrocyte nuclei are digested with DNase 1 at 4°C, histone deacetylase activity is almost quantitatively solubilized by salt extraction. In contrast, when chicken erythrocyte nuclei arc digested with DNasc i at 23°C, approx, 50% of the total nuclear histone deacctylase activity partitions with the nuclear matrix. Temperature is an important parameter when preparing nuclear matrices. There is considerable evidence that the internal nuclear matrix is stabilized by treating nuclei at 37°C before nuclcase treatment (reviewed in Refs, 2 and 3). Importantly, this temperature-induced stabilization causes quantitative but not qualitative changes in the compositioa of the internal nuclear matrix [3]. Our results indicate that exposure of nuclei to room temperature conditions also results in the stabilization of the internal nuclear matrix. There was a modest, but consistent, increase in the histone deacetylase activity partitioning with liver nuclear matrices isolated from nuclei incubated at 23°C versus
4°C. Based upon the partitioning of the histone deacetylase activity, the effect of temperature-induced stabilization of the internal nuclear matrix was most marked for chicken erythrocytes, Our data suggest that internal nuclear matrices cannot be prepared in crythrocytes without such temperature-induced stabilization. it should be noted, however, that temperature-induced alterations in nuclear matrix integrity may be restricted to nuclei isolated under hypotonic conditions
[23]. Using histone deacetylase as a marker enzyme for the internal nuclear matrix, we determined whether neoplastic transformation of trout hepatocytes was accompanied by an alteration in the structural integrity of the internal nuclear matrix, Previous studies have noted changes in the nuclear matrix of transformed cells relative to their normal counterparts [17-19]. Our results provide evidence that the structure of the hepatocellular internal nuclear matrix was not altered in trout hepatocellular carcinomas. • Our observations that the majority of the total nuclear histone deacetylase activity of chicken and trout liver partitions with the internal nuclear matrix suggest that histone deacetylation is restricted to the liver internal nuclear matrix. Preliminary results suggest that histone acetyltransferase is also associated with the internal nuclear matrix of chicken erythrocytes and liver (M.J. Hendzel, P.N. Lewis and J.R. Davie, unpublished results). We and others have shown that dynamically acetylated histones, which attain high acetylation levels and are rapidly deacetylated, are complexed principally to transcriptionally active DNA [12,14,2429]. Furthermore, we observed that (!) the chromatin distribution of dynamically acetylated histones did not parallel that of the competent DNA sequences (i.e., DNase I-sensitive, non-transcribed chromatin) and (2) the majority of dynamically acetylated histones were found associated with the chromatin fragments of the low salt insoluble nuclear material [12]. Transcriptionally active gene chromatin also has an insoluble nature [14,30,31]. For example, Stratling et al. [31] demonstrated a transcription-dependent partitioning of lysozyme coding sequences with the low salt insoluble nuclear material. We propose that regions of chromatin with dynamically acetylated histones have an insoluble nature due to the dynamic interaction of the histones with the internal nuclear matrix-bound histone deacetylase and acetyltransferase, in this model, the preferential association of dynamically acetylated histones with transcriptionally active, but not competent or repressed, DNA results in the selective partitioning of transcriptionally active gene chromatin fragments with the low salt insoluble nuclear material. The juxtapositioning of the transcriptionally active gene chromatin domains may facilitate nuclear processes such as transcription and RNA processing [4,5].
313
Acknowledgements This project was supported by grants from the Medical Research Council of Canada and the National Cancer Institute of Canada, and by the awards of a Medical Research Council Studentship to M.J. Hendzel and a Medical Research Council Scientist to J.R. Davie. References I Nelson, W.G., Pienta, K.J., Barrack, E.R. and Coffey, D.S. 119801 Annu. Rev. Biophys. Biophys. Chem, 15, 457-475. 2 Kaufmann, S.H. and Shaper, J.H. (19911 Exp. Cell Res, 192, 51 ! -523. 3 Belgrader, P., Siegel. A.J. and Berezney, R. (1991)J. Cell Sci. 98, 281-291. 4 Jackson, D.A. (1991) Bioessays 13, I - IfI. 5 Vcrheijen R., Van Venrooij, W. and Ram:tekers, F. (198X)J. Cell Sci. 90, I 1-36. 6 Gasser. S.M. and Lacmmli, tJ.K. 119871 Trends Genct. 3. 16-22. 7 Goldman, M.A. (1988) Bioessays 9, 50-55. 8 Cook, P.R. 119891 Eur, J, Biochem. 185, 487-501. 9 Phi-Van, L., Von Kries, J.P., Ostertag, W. and Stratling., W.lt. (19901 Mol. Ceil. Biol. 10, 23112-23117. 10 Von Kries, J.P., Buhrmester, H. and Stmtling, W.H. (19911 Cell 64. 123-135. II Bodnar, J.W. 119881 J. Theor. Biol. 132, 479-507. 12 Hendzel, M.J., Delcuve, G.P. and Davie, J.R. 119911 J. Biol. Chem. 266, 21936-21942. 13 Lafond, R.E. and Woodcock, C.L.F. 0983) Exp. C¢1~ Res. 147, 31-39.
14 Delcuve, G.P. and Davie, J.R. 11989) Biochem. J. 263, 179-186. 15 Kaufmann, S.H., Gibson. W. and Shaper. J.t]. (i9831 J. Biol. Chem. 258, 2710-2719. 16 Roberge, M., Dahmus, M.E. and Bradhury, E.M. (1988) J. Mol. Biol. 2{ll. 545-555. 17 Fey, E.G. and Penman, S. (1988) Proc. Natl. Acad. Sci. USA 85. 121-125. 18 Brancolini, C, and Schneider, C. 119911 Proc. Natl. Acad, Sci. USA 88, 6936-69411. 19 Kuzmina, S.N., Buldyaeva, T.V., Akopov, S.B. and Zharsky. I.B. 119841 Mol. Cell. Biochem. 58. 183-180. 211 Davie, J.R. and Delcuve. G.P. 11991) Bioehem. J. 280, 491-497. 21 Sinnhuber, R.O., Hendricks, J.D., Wales, J.H. and Putnam, G.B. (19771 Ann. NY Acad. Sci. 298, 389-4118. 22 Nakayasu, Ft. and Berezney, R. (1991) Proc. Natl. Acad. Sci. USA 88. 1(1312-10316. 23 Boyle. W.J. and Baluda, M.A. 0987) Mul. ('ell. Biol. 7. 3345-3348. 24 Zhang, D.-E. and Nelson. D.A, (19881 Biochem. J. 250, 241-247. 25 Zhang, D.-E. and Nelson. D.A. 11988) Bi~chem. J. 250, 233-2411, 26 Ridsdale, J.A., Hendzel, M.J., Delcuve. G.P. and Da~,'ie, J.R. (1990) J. Biol. Chem. 265. 51511-5151~. 27 ip. Y.T., Jackson. V., Meier, J. and Chalkley, R. 11988) J. Biol. Chem. 263, 14044-141152. 2~ lteld~cs. T.W., Thornc, A.W. and Crane-Robinson. C. (198~1 EMBO ,I. 7, 1395-14112. 29 Turner, B.M. 119911J. ('ell Sci. 99, 13-211. 3(I Gro~is, D.S. and Garrard, W.T. (1987) Trends Biochem. Sci. 12, 293-297 31 Stratling, W.li., Dolle, A. and Sippel. A.E. 1198111 BiochemistD 25, 495-51~2.
