ABSTRACT
We have reconstituted salt-treated SV40 minichromosomes with differentially phosphorylated forms of
histone H1 extracted from either G0-, S- or M-phase cells. Sedimentation studies revealed a clear difference
between minichromosomes reconstituted with S-phase histone H1 compared with histone H1 from G0- or M-phase cells, indicating that the phosphorylation state of
histone H1 has a direct effect on chromatin structure. Using reconstituted
minichromosomes as substrate in the SV40
in vitro
replication system, we measured a higher replication efficiency for SV40
minichromosomes reconstituted with S-phase histone H1 compared with G0- or M-phase histone H1. These data indicate that the chromatin
structure induced by the phosphorylation of histone H1 influences the
replication efficiency of SV40 minichromosomes
in vitro
.
A major element in the control of chromatin organization is histone H1, which
plays a role both in the stabilization of the nucleosomal structure and in the
higher order coiling of the chromatin fiber (reviewed in
1
).
Histone H1 has been shown to be differentially phosphorylated during the cell
cycle (reviewed in
2
-
4
). This has been examined in a number of different organisms and cell types,
most notably with the slime mold
Physarum polycephalum
(
5
,
6
) and with synchronized Chinese hamster ovary cells (
7
,
8
). Levels of H1 phosphorylation are usually lowest in the G1 phase of the cell
cycle and continuously rise during S-phase and mitosis. During mitosis, phosphorylation becomes maximal just
before or at metaphase and sharply decreases thereafter (
5
,
8
,
9
). Serine and threonine residues are the acceptors of phosphate groups in H1
histones. The short N-
and long C-terminal basic regions of the molecule appear to contain all the
phosphorylation sites in conserved (K[S/T]PXK) or (K[S/T]PK) sequence motifs.
These motifs are consensus sequences for the cdc kinase family and Langan
et al
. (
10
) could show that the growth-associated histone H1 kinase is identical with a mammalian homolog of the
cdc2/CDC28 yeast kinase (reviewed in
11
,
12
).
Conflicting data exist about the influence of this cell cycle dependent
modification on chromatin structure (reviewed in
13
). A correlation between histone H1 phosphorylation and chromatin condensation
was demonstrated in numerous
in vitro
and
in vivo
experiments. The decreased phosphorylation of histone H1 in a temperature
sensitive growth mutant of FM3A cells results in an abnormal chromosome
condensation at the non-permissive temperature (
14
,
15
), and hyperphosphorylation of histone H1 in tsBN2 cells leads to premature
chromatin condensation (PCC) at the non-permissive temperature (
16
). Treatment of cells with the topoisomerase II inhibitor VM26 (
17
) or the protein kinase inhibitor staurosporine (
18
, both induce a dephosphorylation of histone H1), leads to a concomitant
decondensation of the chromatin, indicating that the phosphorylation of histone
H1 is required to maintain the highly condensed state.
However, H1 phosphorylation is uncoupled from mitosis and chromosome
condensation in the amitotic macronucleus of
Tetrahymena
, where H1 is hyperphosphorylated in exponentially growing cells and completely
dephosphorylated in the highly condensed state (
19
). During sea urchin spermatogenesis histone H1 is also found to be
dephosphorylated in the highly condensed chromatin of mature sperm
(
20
)
.
Similarily, histone H5, a homologue of H1, is dephosphorylated in avian
erythrocytes at a time when compaction of chromatin occurs (
21
)
.
Furthermore, treatment of a temperature-sensitive p34
cdc2
mutant cell line with phosphatase 1 and 2A inhibitors resulted in full
chromosome condensation in the absence of histone H1 phosphorylation (
22
). Thus, the role of histone H1 phosphorylation during chromatin condensation is
still a matter of controversy.
Little is known about whether the phosphorylation of histone H1 has an effect on
chromatin replication. Studies on the phosphorylation sites of human histones
H1A and H1B during the HeLa cell cycle have shown that phosphorylation of one
site in the C-terminal domain of histone H1 precedes, while phosphorylation of a second
site follows the onset of DNA replication (
23
). This led to the proposal that H1 phosphorylation preceding the onset of DNA
replication might produce a chromatin conformation that permits DNA
replication. By using antibodies generated against hyperphosphorylated histone
H1, Lu
et al
. (
24
) observed a replication-dependent phosphorylation of histone H1 during S-phase, which was accompanied by a diffuse nuclear staining. They
propose that H1 phosphorylation acts as a first step mechanism to promote
transient chromatin decondensation, which allows access of specific DNA binding
factors during cell cycle regulated processes such as gene activation, DNA
replication, as well as chromosome condensation. Furthermore
in vivo
experiments with ts mutants, defect in histone H1 phosphorylation, have shown
that the decrease in H1 phosphorylation resulted in an incomplete DNA
replication (
15
).
We have used salt-treated SV40 minichromosomes reconstituted with differentially
phosphorylated forms of histone H1 to investigate the influence of H1
phosphorylation on chromatin structure and replication. We found that at a
physiological ratio of one molecule histone H1 per nucleosome different
chromatin structures are induced depending on the phosphorylation state of
histone H1. These minichromosomes replicate with different efficiencies
indicating a direct effect of H1 phosphorylation on chromatin replication.
African green monkey kidney (CV-1) cells were grown in Dulbecco's modified Eagle medium supplemented with
5% calf serum (DMEM). Cells were synchronized for G0-arrest by incubation in isoleucine-deficient DMEM medium supplemented with 5% dialysed serum for 72 h (
25
). For S-phase synchronization cells were first incubated in isoleucine-deficient DMEM medium for 48 h, followed by incubation in DMEM
medium containing 7.5 mM thymidine for 18 h to arrest in early S-phase (
26
)
.
Release from this block was for 5.5 h by incubation in DMEM medium. As
monitored by
3
H-thymidine pulse labelling, cells reached mid S-phase after this time. For mitotic arrest, cells were grown in DMEM
medium containing 1 [mu]g/ml nocodazole for 20 h (
27
). Cell cycle analysis was done by flow cytometry.
Histone H1 was isolated from synchronized CV-1 cells by extraction with 5% perchloric acid (PCA) (
28
,
29
). All buffers contained the following phosphatase inhibitors: 50 mM NaF, 80 mM [beta]-glycerophosphate and 0.1 mM sodium orthovanadate. Histone H1 was
analysed by 15% SDS-PAGE (
30
) and silver staining (
31
). H1 concentrations were determined by comparison with commercially available
calf thymus H1 (Calbiochem) by densitometric scanning of the gel (Biometra,
Scanpack).
For
in vivo
labelling of histone H1 cells were synchronized as described but for the last 4
h of synchronization the medium was changed to the corresponding phosphate-free medium containing 40 [mu]Ci/ml [
32
P]orthophosphoric acid. Histone H1 was then extracted as described before and
analysed on 15% SDS-PAGE.
HEPES-histidine cationic disc electrophoresis at neutral pH was used to control
the phosphorylation of unlabelled histone H1 (
32
). Treatment of histones H1 with alkaline phosphatase was performed as described
(
38
).
Preparation of SV40 DNA from infected CV-1 cells was performed according to the Hirt procedure (
33
)
.
The SV40 T-Ag was purified from infected insect cells (Sf9) by immunoaffinity
chromatography (
34
). Cytosolic S100 extracts were prepared from HeLa cells exactly as described (
35
). The SV40 minichromosomes were isolated 38 h after infection of CV1 cells (
36
,
37
). To remove RNA particles comigrating on sucrose gradients, the minichromosome
eluate was incubated with 100 [mu]g/ml RNase A for 10 min at room temperature before loading on 5-30% sucrose gradients (
29
).
Salt-treated SV40 minichromosomes (50 [mu]g/ml) were incubated with increasing amounts of differentially
phosphorylated forms of histone H1 (10, 15, 17.5 and 20 [mu]g/ml). Histone H1 and minichromosomes were mixed on ice, dialysed against a
linear salt gradient and centrifuged through 5-30% sucrose gradients (SW40, 39 000 r.p.m., 3 h, 4oC) (
29
). The DNA concentration was determined by measuring the absorbance at 260 nm
and by comparison with SV40 DNA standards on agarose gels. The protein
composition of the reconstituted minichromosomes was analysed by
electrophoresis on 15% denaturing polyacrylamide gels (
30
), followed by silver staining (
31
) and densitometric scanning of the gel. The amount of bound H1 per nucleosome
was determined by comparison with commercially available calf thymus H1. Based
on the fact that the histone to DNA ratio (w/w) appears to be ~1 (
1
), 1.5 [mu]g chromatin contain 258 ng histone H1 at a H1:nucleosome ratio of 1.
Minichromosomes (in 10 mM HEPES, pH 7.8; 30 mM NaCl) were fixed with 0.1%
glutaraldehyde and spread with 2 * 10
-4
% BAC exactly as described (
38
). For contrast enhancement the grids were rotary-shadowed with tungsten at an angle of 8o.
Phosphorylation of H1 during replication.
Soluble histone H1 as well as chromatin bound H1 were incubated in cytosolic
S100 extracts under replication conditions in the presence or absence of SV40 T-Ag. To detect phosphorylation of H1 under these conditions the reaction
was done in the presence of 40 [mu]Ci [[gamma]-
32
P]ATP and phosphatase inhibitors. After 120 min incubation at 37oC proteins were separated on 12% SDS-PAGE, labelled proteins were visualized by autoradiography.
Dephosphorylation of histone H1 reconstituted minichromosomes.
Histone H1 was phosphorylated
in vitro
with purified cdc2/cyclin B kinase (
39
). Salt-treated minichromosomes were reconstituted with
in vitro
labelled H1, purified as described above and incubated for increasing times (0-120 min) in cytosolic S100 extract. One half of the sample was taken for
protein gel analysis (12% SDS-PAGE), the other for precipitation with 25% trichloracetic acid (TCA).
Stability of H1 phosphorylation in the presence of phosphatase inhibitors. In
vivo
labelled histone H1 from M-phase cells (see above) was incubated in cytosolic S100 extract at 37oC for 120 min in the presence of increasing amounts of one of the
phosphatase inhibitors: 2-20 mM NaF (Merck), 5-30 mM [beta]-glycerophosphate (Merck) or 0.01-5 [mu]M okadaic acid (Calbiochem).
The S100 extracts were pretreated with immobilized RNase A (
29
) and 2.5 [mu]M OA for 30 min at 37oC before adding equal amounts (500 ng) of the reconstituted
minichromosomes. Replication was done for 120 min at 37oC exactly as described (
29
).
The effect of histone H1 phosphorylation on chromatin structure and chromatin
replication was studied with SV40 minichromosomes. Minichromosomes were
prepared from infected African monkey kidney (CV-1) cells at 500 mM potassium-acetate to obtain salt-treated minichromosomes lacking histone H1 and most non-histone chromatin proteins (
36
,
37
,
40
). These minichromosomes were reconstituted with exogenously added histone H1,
prepared from CV-1 cells, arrested at different stages of the cell cycle.
Cells were arrested in G0 by isoleucine deprivation (
25
)
,
S-phase cells were obtained by thymidine block (
26
) with subsequent release for 5 h, and mitotic cells by treatment with
nocodazole (
27
)
.
The cell cycle state was monitored by flow cytometry, showing that we obtained
80-90% synchronization for the individual cell cycle phases (data not
shown). We have determined the phosphorylation state of histone H1 either by
labelling with radioactive [
32
P] phosphate
in vivo
or by analysis of unlabelled H1 using cationic disc electrophoresis (
32
)
.
Histone H1 was prepared from isolated nuclei of
32
P-labelled cells and investigated by PAGE in the presence of SDS and silver
staining (Fig.
1
A). The gel was further analysed by autoradiography showing that histone H1 is
not phosphorylated in G0-phase, moderately phosphorylated in S-phase and hyperphosphorylated in M-phase (Fig.
1
B).
We used salt-gradient dialysis to reconstitute salt-treated SV40 minichromosomes with increasing amounts of histone H1
from either G0-, S- or M-phase cells (
29
). We use the terms G0-H1 chromatin, S-H1 chromatin or M-H1 chomatin for these reconstituted minichromosomes.
Reconstituted minichromosomes were purified on sucrose gradients to remove
excess histone H1. Minichromosome containing fractions were pooled and
concentrated on 30% sucrose cushions. To investigate the protein composition of
the reconstituted minichromosomes equal amounts of chromatin were analysed on a
protein gel. The amount of chromatin bound H1 was estimated by comparison with
known amounts of calf thymus H1 (Fig.
2
). We found that the differentially phosphorylated forms of histone H1 bind with
similar affinities to chromatin.
Differences in chromatin structure induced by the differentially phosphorylated
forms of histone H1 might be below the resolution of the electron microscope.
For this reason we have investigated the hydrodynamic properties of
reconstituted minichromosomes using sucrose gradient centrifugation (Fig.
4
). At a ratio of 0.5 molecules histone H1 bound per nucleosome we found no
difference in the sedimentation rate of G0-H1, S-H1 and M-H1 chromatin. In all cases the S-value increased from 50S for H1-depleted minichromosomes to 66S for H1-carrying chromatin. At ratios of 0.75
molecules H1 per nucleosome we detected a small reduction of 4S in the
sedimentation rate of S-H1 chromatin compared with G0-H1 and M-H1 chromatin. This difference became more pronounced at the
physiological ratio of one molecule H1 per nucleosome (
42
), resulting in an S value of 71S for S-H1 chromatin compared with 80S for G0-H1 and M-H1 chromatin. At a histone H1:nucleosome ratio of >1,
minichromosomes tend to aggregate and differences in the hydrodynamic
properties of the differentially reconstituted minichromosomes disappear. The
sedimentation profiles indicate a distributive binding of histone H1 to the
chromatin for all ratios and all phosphorylation states used, resulting in a
homogenous chromatin population.
As histone H1 is one of the few known substrates of the cdc-kinase family (
10
), it may be possible that phosphorylation of histone H1 has a regulatory effect
on chromatin replication. We investigated this question in the SV40
in vitro
replication system, by using salt-treated SV40 minichromosomes, reconstituted with G0-, S- and M-phase histone H1, as template for replication in
unfractionated extracts.
Since it was not known whether the H1 phosphorylation state remains unchanged
during incubation under replication conditions, we determined whether H1 is
phosphorylated in the cytosolic S100 extract used for replication. For that
purpose, we included [[gamma]-
32
P]ATP in the reaction mixture, which was incubated in the absence or presence of
the SV40 T-Ag. Phosphorylation of histone H1 could not be detected under replication
conditions, showing that a phosphorylation of unphosphorylated G0-H1 or hemiphosphorylated S-H1 does not occur during incubation (data not shown).
We also wanted to demonstrate that H1 is not dephosphorylated when incubated in
the replication extract. We therefore reconstituted salt-treated SV40 minichromosomes with H1, either labelled
in vitro
with purified p34
cdc
-kinase (Fig.
5
A and B) or isolated from M-phase cells and phosphorylated
in vivo
(Fig.
5
C and D). The reconstituted minichromosomes were purified by
sucrose gradients and incubated at 37oC under replication conditions in cytosolic S100 extract. Samples, taken
during incubation, were investigated on protein gels followed by
autoradiography (Fig.
5
A). For quantitation, the phosphate content of histone H1 was determined by TCA
precipitation (Fig.
5
B). We found that already after 10 min incubation 50% of the input histone H1
was dephosphorylated, and after 2 h incubation only 10-20% of histone H1 was still phosphorylated, indicating that active
phosphatases are present in the S100 extract.
With the exception of
Saccharomyces cerevisae
(
43
) all eukaryotic organisms analysed so far contain a linker histone. It is now
widely accepted that members of the H1/H5 family of linker histones are
involved in the condensation of chromatin filaments into both 30 nm fibers and
higher-order chromosomal structures (
44
). Nevertheless, the precise nature of the interactions of these lysine-rich histones with DNA and other chromosomal constituents during chromatin
packaging remains largely elusive. One of the uncertainties surrounding histone
H1 is the role(s) played by the cell cycle dependent phosphorylation of the
protein. Mitotic hyperphosphorylation of histone H1 has been assumed to trigger
or promote chromosome condensation (
5
), it is however still controversial whether H1 phosphorylation plays an active
or passive role during condensation (
13
). So far, no special function has been linked to the moderate phosphorylation
of histone H1 during S-phase. This is of particular interest, because histone H1 is a specific
substrate of the cdc kinases (
10
), which are thought to regulate the activity of proteins involved in DNA
replication by phosphorylation (reviewed in
12
,
45
).
To get further insights into these processes we have reconstituted H1-depleted SV40 minichromosomes (
40
,
41
) with increasing amounts of histone H1 extracted from G0, S or M-phase cells. Phosphorylation and ADP-ribosylation are the only known cell-cycle dependent post-translational modifications of histone H1. Because there
is no evidence at present that poly (ADP-ribose) synthesis on nuclear proteins other than poly (ADP-ribose) polymerase is of any physiological relevance (
46
), we assume that the observed effects are due to the different phosphorylation
state of histone H1.
The phosphorylation state of histones H1, prepared from different phases of the
cell cycle, was determined by
in vivo
labelling and analysis on HEPES-histidine gels (
32
) (Fig.
1
). Proper binding of histone H1 to the nucleosome was confirmed by electron
microscopy (Fig.
3
), where we observed in all cases a condensation of the minichromosomes induced
by the addition of histone H1 as described earlier (
41
)
.
Furthermore micrococcal nuclease digestion of reconstituted minichromosomes in
comparison to salt-treated minichromosomes showed the expected increase in protection of the
monomeric DNA band.
Comparison of the chromatin structure of the reconstituted minichromosomes by
sedimentation studies revealed significant differences only in a narrow range,
corresponding to the physiological ratio of one molecule histone H1 per
nucleosome. Whereas both the completely unphosphorylated form of histone H1
prepared from G0-phase cells and the hyperphosphorylated form of histone H1 prepared from M-phase cells induce the same degree of compactness, resulting in an S-value of 80S, we determined a more retarded sedimentation for
the minichromosomes reconstituted with the same amount of the moderately
phosphorylated form of histone H1 prepared from S-phase cells, resulting in an S-value of only 71S (Fig.
4
). As we have found that the same amount of the differentially phosphorylated
forms of histone H1 is bound to the reconstituted minichromosomes, the slower
sedimentation of the S-H1 chromatin is indicative for a more open chromatin structure.
The results concerning the variations in chromatin structure induced by the
differentially phosphorylated forms of histone H1 do not follow a simple
pattern because both unphosphorylated G0-H1 chromatin and hyperphosphorylated M-H1 chromatin have identical hydrodynamic properties, whereas only S-H1 chromatin shows a reduced sedimentation rate.
The basic N- and C-terminal tails of histone H1 in the unphosphorylated G0-H1 can neutralize negative charges of the phosphodiester
backbone of linker DNA (
47
), which may result in a coiling or bending of the linker DNA (
48
,
49
) and in a moderate condensation of the chromatin. This situation is comparable
with the condensation observed in the presence of dephosphorylated H1 (H5) in
Tetrahymena
macronuclei (
19
), avian erythrocytes (
21
) and mature sperm (
20
).
However, hyperphosphorylation of histone H1 during mitosis may weaken the
interaction between histone tails and linker DNA (
50
), and this could cause a repulsion of the tails from the linker DNA and an
opening of the chromatin structure. The induction of a similar chromatin
structure as observed with G0-H1 reconstituted minichromosomes could be generated by a modulation of the
interaction between H1 histones themselves, or between H1 and H2A (
51
) or H1 and H3 (
52
,
53
) bringing nucleosomes into closer contact. The degree of condensation we
observe after reconstitution with mitotic histone H1 is certainly only one of
the reactions required for chromatin condensation. During mitosis
hyperphosphorylation of histone H1 may be required to recruit `compaction
factors' as topoisomerase II (
54
) or the SMC (structural maintenance of chromosomes) proteins (
55
) to get access to the DNA and to further condense the chromatin by an as yet
unknown mechanism.
An interesting result of our studies is a clear difference in hydrodynamic
properties after reconstitution of SV40 minichromosomes with histone H1
prepared from S-phase cells. As known from studies in the CHO cell cycle (
8
), histone H1 phosphorylation occurs during S-phase only on serine residues in the C-terminal domain of histone H1, resulting in an asymmetric charge
distribution. It has been suggested that the SPKK motif, which is
phosphorylated in the C-terminal domain of H1, forms [beta]-turn secondary structures, which bind to the minor groove of
DNA (
56
). Phosphorylation in this domain is likely to disrupt this conformation and to
weaken the interaction of these sequences with DNA (
57
). Thus, the moderate and asymmetric phosphorylation of histone H1 may lead to a
dissociation of the C-terminal histone tails from the linker DNA. The linker DNA may be more
extended compared with G0-chromatin resulting in a larger distance between individual nucleosomes
and in a more open chromatin structure.
Interestingly, the differences observed in chromatin structure are reflected in
the replication efficiencies of the reconstituted minichromosomes. Both G0-H1 and M-H1 chromatin show a reduced replication efficiency compared with S-H1 chromatin (Fig.
6
). The relaxed structure of the S-H1 chromatin could either increase the accessibility for replication
factors to the DNA, or this type of H1 phosphorylation may directly affect the
interaction with replication proteins. It has been shown for example that the
SV40 T-Ag hexamer directly interacts with histone H1 in an ATP-dependent reaction (
58
). It is however not known whether this interaction is influenced by the
phosphorylation of H1.
It is widely assumed that the chromatin structure is transiently disrupted
during passage of the replication fork, whereby histone H1 is removed from the
replication fork and reassociates with the daughter strands during maturation
of the chromatin (reviewed in
59
). Weakening of H1-DNA interactions could facilitate the detaching of histone H1 and by this
way stimulate the movement of the replication machinery.
Thus it seems that the process of replication is not only controlled on the
level of the activity of replication proteins but also on the level of the
chromatin structure. These data are in agreement with
in vivo
experiments using cell mutants, which are defect in histone H1 phosphorylation (
15
). The decrease in H1 phosphorylation resulted in an incomplete DNA replication
and a defect in chromosome condensation, indicating that histone H1
phosphorylation also plays a role during S-phase of the cell cycle.
We thank Rolf Knippers for helpful discussions and critical reading of the
manuscript and Rosmarie Mettke for technical assistance. This work was
supported by Sonderforschungsbereich 156 (C4).
REFERENCES
Return