Analyzing changes of chromatin-bound replication proteins occurringin response to and after release from a hypoxic block of replicon initiation in T24 cells Maria van Betteraey-Nikoleit,
Trang 1Analyzing changes of chromatin-bound replication proteins occurring
in response to and after release from a hypoxic block of replicon initiation in T24 cells
Maria van Betteraey-Nikoleit, Karl-Heinz Eisele, Dirk Stabenow and Hans Probst
Physiologisch-Chemisches Institut der Universita¨t Tu¨bingen, Germany
It was shown previously [Riedinger, H J., van
Betteraey-Nikoleit, M & Probst, H (2002) Eur J Biochem 269,
2383–2393] that initiation of in vivo SV40 DNA replication
is reversibly suppressed by hypoxia in a state where viral
minichromosomes exhibit a nearly complete set of
repli-cation proteins Reoxygenation triggers fast completion
and post-translational modifications Trying to reveal such
fast changes of chromatin-bound replication proteins in the
much more complex replication of the cellular genome
itself, we developed a protocol to extend these studies using
the human bladder carcinoma cell line T24, which was
presynchronized in G1by starvation Concomitantly with
stimulation of the cells by medium renewal, hypoxia was
established This treatment induced T24 cells to contain a
large amount of replicons arrested in the ‘hypoxic
preini-tiation state’, ready to initiate replication as soon as normal
pO2was restored Replicons in other stages of replicative
activity were not detectable Consequently the arrested replicons were rapidly released into synchronous initiation and succeeding elongation Extraction of T24 nuclei with a Triton X-100 buffer yielded a fraction containing the cellular chromatin, including DNA-bound replication proteins, while unbound proteins were removed The use-fulness of this protocol was tested by the proliferation marker PCNA We demonstrate here that this protein switches from the remainder cellular protein pool into the Triton-extracted nuclear fraction upon reoxygenation Employing this protocol, analyses of chromatin-bound MCM2, MCM3, Cdc6 and cdk2 suggests that the ‘classi-cal’ prereplication complex is already formed during hypoxia
Keywords: chromatin; DNA replication; hypoxia; nuclei; synchronization
Apart from control by cell cycle signals, DNA replication in
mammalian cells is subject to a regulation which depends on
the O2tension in the cellular environment Presumably, this
regulatory phenomenon, adapting the intensity of DNA
replication of growing cells to the supply of O2, is important
during embryonic growth and wound healing, and
influen-ces the propagation of malignant tumors The O2
-depend-ent regulation concerns cells which are in S-phase or are
definitively committed to enter S-phase When the
concen-tration of O2drops to about 0.2–0.02%, scheduled replicon
initiations are suppressed and already-active replication
forks are slowed down Of the cell lines examined so far,
only Ehrlich ascites cells exhibit suppression of replicon
initiation without a significant slowing down of fork
progression [1–3] During hypoxia cells accumulate
repli-cons in a state ready to initiate (almost instantaneously)
within a few minutes after oxygen recovery Thus, sudden
reoxygenation after several hours of hypoxia triggers a
synchronous burst of initiations of the accumulated repli-cons followed by normal replication So far, this regulatory phenomenon has been published for Ehrlich ascites, HeLa and CCRF cells [2,4,5] Further cell lines examined so far, e.g T24, A549, PC3, TC7, BHK, SW2, HL60 and HUVEC are also sub ject to the fast O2-dependent control of replication (G Probst, H Probst & M van Betteraey-Nikoleit, unpublished results) We therefore suggest that it represents a general phenomenon in mammalian cells, although the molecular mechanisms involved are still largely obscure The remarkably fast resumption of initiations after reoxygenation suggests that the O2-dependent replication control acts very directly on the replication apparatus itself
As published earlier, replication of the SV40 genome in virus infected cells also obeys the oxygen-dependent regu-lation [6,7] Reoxygenation of virus infected cells after several hours of hypoxia triggers a burst of hypoxically accumulated viral replicon initiations followed by a syn-chronous round of completely regular replication of viral genomes
Studying several replication proteins bound to SV40 minichromosomes before and after reoxygenation, i.e before and after triggering initiation, we found that a large number of polypeptides taking part in viral replication were bound to the SV40 minichromosome already under hypo-xia However, the multiprotein complexes necessary for unwinding, primer synthesis and elongation lacked essential components and remained incomplete as long as hypoxia lasted [8] Reoxygenation triggered fast completion to a
Correspondence to M van Betteraey-Nikoleit,
Physiologisch-Chem-isches Institut der Universita¨t Tu¨bingen, Hoppe-Seyler-Straße 4,
D-72076 Tu¨bingen, Germany.
Fax: + 49 7071293339, Tel.: + 49 70712973329,
E-mail: maria.van-betteraey@uni-tuebingen.de
Abbreviations: PCNA, proliferating cell nuclear antigen; BrdU,
5¢-bromodeoxyuridine; FITC, fluorescein isothiocyanate.
(Received 30 April 2003, revised 24 July 2003,
accepted 28 July 2003)
Trang 2functional complex, indicating a specific influence of
O2-dependent cellular changes on critical steps of the
assembly of a functional viral replication machinery
Consequently, we wanted to extend these studies from
the SV40 model to the far more complex replication of
the cellular genome of mammalian (human) systems For
this purpose we had to meet two demands Firstly to
find a cell line that can be induced to bear a maximal
number of replicons arrested in the ‘hypoxic preinitiation
state’ and as few as possible in other states of replication,
and secondly the elaboration of a protocol for preparing
a cell fraction that contains the cellular chromatin and
specifically retains the functionally bound (replication)
proteins
In this communication we demonstrate that by a simple
starvation procedure followed by stimulation with fresh
medium and concomitant establishment of hypoxia, the
human bladder carcinoma cell line T24 can be induced to
accumulate replicons scheduled to initiate in the early
S-phase in most cells, while other stages of replicon
activity are virtually absent Reoxygenation triggers these
replicons to initiate replication at a high degree of
synchrony, followed by subsequent normal elongation
The immediate answer to sudden reoxygenation resembles
in principle that of SV40 replicons in virus infected cells
However, replicons started in the noninfected T24 cells are
much longer and not identical Extraction of T24 nuclei
with a Triton X-100 buffer yields a fast sedimenting
nuclear fraction containing the cellular DNA and proteins
associated with replicating chromatin We take this
material as a functional equivalent to SV40
minichromo-somes elutable from the nuclei of virus infected cells and
operationally define replication proteins remaining after
the Triton X-100 extraction (in accordance with [9]) as
functionally chromatin-bound To justify this view we
examined whether reoxygenation triggers the association
of PCNA, the processivity clamp of polymerase delta in
mammalian replication, with the fast sedimenting
Triton-extracted fraction
In a first experiment we employed this protocol thereby
demonstrating that important components of the ‘classical’
prereplication complex [10] become bound to chromatin
already during the hypoxic incubation This suggests that
the ‘hypoxic preinitiation state’ is similar to the known
‘classical’ prereplication complex, and that hypoxia directly
influences mechanisms activating this complex
Materials and methods
Cell culture, transient hypoxia, reoxygenation and
radioactive labeling
T24 cells (gift from Altana Pharma, Konstanz, Germany)
were grown in plastic flasks in DMEM supplemented with
10% fetal bovine serum and penicillin/streptomycin (100 U/
100 lgÆmL)1) The cells were subcultured when they
reached confluence Under these conditions the cells
exhib-ited a partially tetraploid caryotype
For synchronization, the desired number of glass Petri
dishes was seeded from an almost confluent large culture
with 150 000 cellsÆmL)1 (35 mm, 1.5 mL; 145 mm,
25 mL) 44 h before the start of an experiment Thereby,
most cells became arrested in G1due to starvation When
a 14C prelabel was desired, the seeding medium was supplemented with 2.5 nCiÆmL)1 [14C]Thd Experiments started with stimulation of the cells by a complete exchange of the culture medium with prewarmed fresh medium supplemented with 10% (v/v) fetal bovine serum Subsequent gassing of the cell cultures was performed with a continuous flow of humidified artificial air containing 5% (v/v) CO2 for normoxic incubations, and with 0.02% O2, 5% CO2, and Ar to 100% for hypoxic gassing For gassing, the equipment and the procedures described by [7] were used For reoxygenation 0.25 volumes of medium equilibrated with 95% O2/5% CO2 (v/v) were added to hypoxic cell cultures, and gassing was continued with artificial air
[Methyl-3H]deoxythymidine was added either directly to the cells, or under hypoxic culture conditions by plunging a spatula carrying the appropriate quantity in dried form into the culture medium To stop incubations medium was removed by aspiration and the cells were washed once with ice-cold phosphate-buffered saline (NaCl/Pi: 150 mMNaCl,
10 mMNaHPO4, pH 7) and either processed for determin-ation of acid insoluble radioactivity as described [11] or otherwise for analyses as described below
Alkaline sedimentation analyses of cellular DNA For analyzing the length distribution of growing daughter strands of T24 DNA, cultures on 35 mm glass Petri dishes were pulse-labeled for 8 min with 7 lCi [methyl-3 H]deoxy-thymidineÆmL)1 Labeling was stopped by washing the cells with ice cold phosphate-buffered saline (NaCl/Pi: 150 mM
NaCl, 10 mMNaHPO4, pH 7) The cells were trypsinized for 5 min at 4C and layered onto the top of 10–30% alkaline sucrose gradients [12] After denaturation of the DNA for 6 h, centrifugation was performed at
20 000 r.p.m., 23C for 10 h in a Beckman SW28 rotor 1.2 mL fractions were collected from the top of the gradient and processed to analyze acid insoluble radioactivity DNA cytofluorometry
For cytofluorometry of cellular DNA cells were trypsinized, washed with NaCl/Pi and fixed with 90% methanol Histograms of DNA contents were recorded with a FACSCalibur (Becton-Dickinson) after staining the cells with propidium iodide (0.05 mgÆmL)1 in 0.1% sodium citrate) and simultaneous RNase digestion (1 mgÆmL)1) for
30 min at 37C
Cell fractionation Cells were washed once with NaCl/Pi and twice with hypotonic buffer (20 mM Hepes, pH 7.5, 20 mM NaCl,
5 mMMgCl2) and suspended in 10 mL of hypotonic buffer After 10 min on ice, cells were disrupted to free nuclei by
25 strokes with the tight fitting pestle of a dounce homogenisator and were then centrifuged for 5 min and
1500 g at 4C to separate the cytosolic supernatant from the nuclear pellet Nuclei were resuspended in extraction buffer (50 mMHepes, pH 7.5, 100 mMKCl, 0.25% Triton X-100, 2.5 mM MgCl, 1 mM dithiothreitol) containing
Trang 3aprotinine (1 lM), leupeptine (50 lM),
4-(2-aminoethyl)-bezenesulfonylfluoride/HCl (1 mM) and NaF (10 mM) and
centrifuged for 3 min and 600 g at 4C Nuclei were
resuspended in extraction buffer three more times to fully
lyse the nuclear envelope and complete extraction
Super-natants were combined and yielded nucleosolic proteins
The remaining pellet contains all DNA and structure bound
proteins and is further referred to as chromatin-fraction
Electrophoresis of proteins and Western blotting
Cytosolic and nucleosolic proteins were precipitated from
the respective supernatants by adding five volumes of
ice-cold acetone Proteins remaining in Triton-extracted nuclei
were recovered after nuclease digestion with DNase and
RNase Nuclei prepared as above were suspended in
extraction buffer containing DNase (0.1 mgÆmL)1), RNase
(0.025 mgÆmL)1) and MgCl2(5 mM) Digestion was for 1 h
on ice Proteins for Western blot analyses were then either
isolated by phenol extraction and subsequent acetone
precipitation as described [13] or directly denatured and
solubilized with SDS electrophoresis sample buffer, as it
turned out that remaining DNA fragments did not
interfere with the running properties of the proteins in
SDS-gels
Proteins were separated on an 8% SDS/polyacrylamide
gel [14], blotted to Nylon-P membrane (Amersham) and
subsequently immunodetected using the ECL Western
blotting procedure (Amersham) according to the
manufac-turer’s instructions Dilution of antibodies used were as
follows: PCNA (mouse monoclonal antibody, Santa Cruz)
1 : 3000, Cdc6 (mouse monoclonal antibody, Santa Cruz
Biotechnologies) 1 : 500, MCM2 (rabbit polyclonal
anti-body, Transduction Laboratories, Heidelberg, Germany)
1 : 10,000, MCM3 (rabbit polyclonal antibody,
Transduc-tion Laboratories, Heidelberg, Germany) 1: 3000, Cdk2
(rabbit polyclonal antibody, Santa Cruz Biotechnologies)
1 : 500
Immunofluorescence staining of total PCNA
and chromatin-bound PCNA
Cells grown on coverslips were washed once with ice-cold
NaCl/Pi For subsequent staining of total PCNA, cells were
directly fixed with ice-cold acetone/methanol (1 : 1, v/v) for
10 min at 4C When only chromatin-bound PCNA had to
be stained, soluble proteins were extracted by washing the
cells three times with extraction buffer (see Cell fractionation)
and afterwards fixed with acetone/methanol (1 : 1, v/v)
for 10 min at 4C Subsequently all coverslips were
processed for detection of PCNA after air drying Cells
were blocked with 1% (w/v) BSA in NaCl/Pifor 20 min and
incubated with anti-PCNA Ig (Boehringer Mannheim,
dilution 1 : 100) in NaCl/Pi/BSA for 1 h at room
tempera-ture After washing three times with NaCl/Pifor 5 min they
were further incubated for 30 min with anti-mouse IgG
labeled with Alexa Fluor 586 (Molecular Probes, dilution
1 : 200) in NaCl/Pi/BSA Cells were again washed three
times for 5 min with NaCl/Pi During the last wash total
DNA was stained with bisbenzimide (2 lgÆmL)1in NaCl/
Pi) Finally PCNA (Alexa Fluor 568 stain) and total
DNA (bisbenzimide stain) were visualized with a Zeiss
fluorescence microscope (Axioskop) using the appropriate filter combinations
Immunofluorescence staining of chromatin-bound PCNA and of replicating DNA
Cells grown on coverslips were labeled by adding 15 lM
5¢-bromodeoxyuridine (BrdU) 15 min before the end of the respective incubation conditions, in case of hypoxic labeling by plunging a spatula carrying the appropriate quantity in dried form into the cell culture medium To stop incubations cells were washed once with NaCl/Pi For extraction of soluble proteins cells were washed three times with extraction buffer (see Cell fractionation) and subsequently fixed with methanol for 10 min at 4C Cells were then sequentially stained and fixed as reported previously in [15] Briefly, cells were blocked with 1% (w/v) BSA in NaCl/Pi for 20 min, incubated with anti-PCNA Ig (Boehringer Mannheim, dilution 1 : 100) in NaCl/Pi/BSA for 1 h at room temperature and for 30 min with Alexa Fluor 568 antibody (red fluorescence, dilu-tion 1 : 200) in NaCl/Pi/BSA The primary and secondary antibodies were fixed in place with 4% (v/v) formaldehyde for 20 min at room temperature Subsequently cells were washed twice with NaCl/Pi For DNA denaturation cells were treated with 2M HCl at 37C for 1 h After neutralization with NaCl/Pi they were finally incubated for one h with a fluorescein isothiocyanate (FITC)-labeled anti-BrdU Ig (green fluorescence, Boehringer Mannheim, dilution 1 : 50) Between the antibody incubation steps cells were washed three times for 5 min with NaCl/Pi During the last wash total DNA was stained with bisbenzimide (2 lgÆmL)1 in NaCl/Pi) Finally PCNA (Alexa Fluor 568 stain), replicating DNA (FITC stain) and total DNA (bisbenzimide stain) were visualized with a Zeiss fluorescence microscope (Axioskop, Zeiss, Go¨ttin-gen, Germany) using the appropriate filter combinations
Results
Inducing the ‘hypoxic preinitiation state’ in cellular replicons
About 35 h after infection with SV40 virus, CV1 cells replicate almost exclusively SV40 minichromosomes at high intensity These represent a highly homogenous population
of conveniently small subcellular entities During a hypoxic period of 6–7 h, a large amount of them is arrested in the
‘hypoxic preinitiation state’ [6,7] and can be released within 2–3 min into effective initiation and a succeeding synchron-ous replication round by reoxygenation Thereby, among total viral genomes exhibiting replicative activity (‘viral replicons’), the fraction of hypoxically synchronized repli-cons reaches > 90% Consequently, the equipment of the minichromosomes with replication proteins reflects with sufficient reliability the state of the replication machinery before and after oxygen recovery, respectively
Cellular replicons on the other hand, are highly hetero-geneous in their sizes and replication states within an asynchronous cell cycle Therefore in addition to ‘hypoxic pre-initiation states’, 7 h of hypoxia accumulated significant amounts of replicons hit by hypoxia in other states of
Trang 4activity [4] Thus, accumulation of cellular ‘hypoxic
prein-itiation states’ cannot be achieved as easily as that of SV40
Therefore we tried to subject cell populations enriched with
G1 cells to hypoxia, as successfully performed previously
with Ehrlich ascites cells, by selecting G1 cells by zonal
zentrifugation [3,16] In the course of investigating several
cell lines (see Discussion), we came across the human
bladder carcinoma cell line T24, which is easily arrested in
G1by starvation [17] Using this cell line, we developed an
appropriate protocol Briefly, cells were grown for 44 h after
seeding which caused shortage of nutrients and growth
factors in the medium Starved cells were stimulated by
exchanging the medium with prewarmed fresh medium,
followed by hypoxic or normoxic gassing of the cells The
experiments described below demonstrate that replicative
activity released immediately after O2 admission to
pre-treated hypoxic T24 cells represents almost exclusively
synchronous replicon initiation followed by normal
elon-gation
DNA synthesis rate
The course of the [methyl-3H]deoxythymidine
incorpor-ation rate into DNA of starved T24 cells was monitored
after stimulation by medium renewal under normoxic,
hypoxic and reoxygenated incubation conditions Figure 1
shows that, in normoxically incubated cells, the
incorpor-ation rate remained relatively low up to 4 h after medium
exchange and then gradually increased up to 10 h, when
maximal incorporation was attained This was followed by
a decrease Under hypoxia, in contrast, incorporation
decreased to a background level during the first 2 h and
remained at this level until reoxygenation Immediately after
reoxygenation, incorporation of radioactivity increased strongly within a very short interval and decreased 6–8 h later The profile of [3H]Thd incorporation after reoxygen-ation appears double-peaked The first peak is possibly caused by cells that proceded to the end of G1phase during the 7 h hypoxic gassing, accumulating replicons ready to initiate immediately after reoxygenation Cells causing the second peak possibly had not yet reached this border during the 7 h hypoxic period
Alkaline sedimentation analyses of growing daughter strands
A fast increase of the DNA synthesis rate either reflects release of replicon initiations or stimulation of elongation,
or both To determine the cause of the increase in Fig 1, we analyzed the chain length distribution of pulse-labeled nascent daughter strands by means of alkaline sedimenta-tion Synchronous replicon initiations first produce homo-geneously sized small daughter strands, which subsequently grow homogeneously to longer sizes, thus causing a synchronous shift of growing DNA chains to higher S-values
Figure 2A shows a survey of alkaline sedimentation profiles of acid-insoluble radioactivity from pulse labels applied to normoxic, hypoxic and reoxygenated T24 cells
The cells were prelabeled with [14C]Thd when seeded,
44 h before the start of the experiment The resulting [14C]Thd profile (Fig 2B, crosses) typically exhibits a peak
in the last third of the gradient representing matured bulk DNA The [14C]Thd gradients were omitted from Fig 2A for clarity After medium exchange the normoxically incubated cultures exhibited a sedimentation profile (Fig 2A, first profile) attributable to asynchronously acting replicons, because of a typical label distribution across the gradient, resulting from the normal steady-state of asyn-chronous initiation, elongation and termination The gradient of hypoxically treated T24 cells contains almost
no [3H]Thd, as expected according to the incorporation curve (Fig 1) As soon as 15 min after reoxygenation, a strong incorporation of [3H]Thd into growing daughter strands occurs, preferentially sedimenting in the first third
of the gradient and attributable to short chains originating from newly initiated replicons In the course of further
25 min of reoxygenated growth, the incorporation of [3H]Thd still increased, while the peak shifted to higher S-values To visualize the chain growth between 15 and
40 min better, the last two profiles of Fig 2A are depicted
as the percentage of total c.p.m in Fig 2B From 15 to
40 min after reoxygenation the peak distinctly shifted to higher S-values The extent of the shift reflects the chain elongation during 25 min and can be calculated as about 0.5 lmÆmin)1at either end of growing daughter strands This is a very common elongation rate for mammalian cells Note that up to 15 min after reoxygenation there is hardly any incorporation into fast sedimenting ‘old’ daughter strands Thus, almost no active replicons occur that have been initiated before reoxygenation The shapes
of the two gradient profiles are narrow and very similar, suggesting that the cellular replicons grow synchronously
at relatively homogenous elongation rates Thus, alkaline
Fig 1 Rate of [3H]Thd incorporation into DNA of starved T24 cells
under normoxic (s) and hypoxic/reoxygenated (d) incubation
condi-tions after medium renewal T24 cells were prelabeled with [14C]Thd
and grown for 44 h Subsequently the medium was renewed and cells
were either incubated normoxically for 7 h, or hypoxically for 7 h and
then reoxygenated At the times indicated cells were pulse-labeled for
8 min with 7 lCiÆmL)1 [ 3 H]Thd while maintaining the respective
incubation conditions during labeling and processed for measuring the
ratio between acid-insoluble3H and14C radioactivity.
Trang 5gradient centrifugation confirmed that replicon initiation is
inhibited under hypoxia Upon reoxygenation, suppressed
initiations are released very fast in a highly synchronous
fashion
DNA cytofluorometry
A large portion of the partially tetraploid T24 cells exhibited
G DNA content at 44 h growth after seeding (Fig 3A)
Subjecting such cells after medium renewal to a 7-h hypoxic period markedly increased the cell fraction with G1DNA content (Fig 3B) Three hours after reoxygenation, the majority had entered the S-phase while a minor part still exhibited G1 DNA content (Fig 3C) This result also supports the assumption that after medium renewal most of the cells proceed during hypoxia through G1up to the point
at which the first replicons of a (scheduled) S-phase would normally be activated However, for a minor part of the cells
a 7-h hypoxic incubation following medium exchange does not seem to be sufficient to accumulate replicons ready to initiate immediately after O2 recovery Perhaps these cells already were in G0at the time of medium exchange Yan
et al [18] presented flow cytometric analyses of T24 cells
4 days after seeding in high density and following release from contact inhibition In the ATCC catalogue T24 cells are described as hypertriploid with 8% polyploidy In contrast to the diploid T24 cells used by Yan et al [18], the T24 cells we used were tetraploid for unknown reasons Nevertheless their flow cytometric analyses also show that the cells are arrested with a G1DNA content As they enter S-phase about 20 h after replating, the cells must have been
in a G0state before this We intended to arrest the cells in
G1, from where they can proceed to DNA synthesis within about 6 h As shown in Fig 3C, the majority of the cells exhibiting S-phase DNA content 3 h after reoxygenation probably only experienced a G1 arrest These cells are obviously identical to those initiating immediately upon reoxygenation, and may be the cause of the first peak in Fig 1 and the sedimentation profiles shown in Fig 2B Mitotic index
To demonstrate that after release of the hypoxic block T24 cells further proceed through the cell cycle normally and at high synchrony, we determined the percentage of mitotic cells Figure 4 shows that after medium exchange and further normoxic gassing first mitotic cells appear after about 13 h, their number increases within the next 5 h and decreases again at longer incubation A similar increase of DNA synthesis occurs in the same cells 8–10 h before (Fig 1), compatible with an elapse of a S- and G2-phase Cells exposed to hypoxia directly after medium renewal and reoxygenated 7 h later exhibited sharp rise of mitotic cells
10 h after reoxygenation, which resembles the sharp rise in the DNA synthesis rate directly after reoxygenation The
Fig 3 Histograms of cellular DNA content recorded by flow cyto-fluorometry T24 cells were grown for 44 h Subsequently the medium was renewed and the cells were incubated hypoxically for 7 h or reoxygenated for 3 h thereafter After stopping the respective incu-bation conditions, cells were trypsinized, fixed and stained as described
in the Materials and methods (A) Cells after medium renewal; (B) cells after medium renewal and 7 h of hypoxia (200 p.p.m.); (C) the same cells after 7 h of hypoxia (200 p.p.m) and 3 h aerated incubation.
Fig 2 Alkaline sedimentation patterns of pulse-labeled T24 DNA after
lysis on top of the gradients T24 cells were grown for 44 h, after which
the medium was renewed and cells were either incubated normoxically
or hypoxically for 7 h, or reoxygenated after 7 h of hypoxia Nascent
daughter DNA chains were pulse-labeled with 10 lCi [3H]ThdÆmL)1
8 min before the end of the respective incubation conditions (A)
Comparison of the gradient profiles of normoxic, hypoxic, 15 min and
40 min reoxygenated T24 cells Profiles are depicted consecutively in
total c.p.m Normoxia, 97215 c.p.m.; hypoxia, 365 c.p.m.;
reoxygen-ated 15 min, 57735 c.p.m.; reoxygenreoxygen-ated 40 min, 176754 c.p.m Each
profile consists of 31 fractions (B) Comparison of the profiles of
15 min (m) or 40 min (d) reoxygenated cells (same as in Fig 2A) in
percentage of total c.p.m ·, Matured 14 C-labeled bulk DNA of T24
cells Sedimentation was from left to right.
Trang 6mitotic index also exhibits a double-peaked profile similar to
the profile of the [3H]Thd incorporation The double peak
of the3H incorporation curve is therefore possibly caused
by cells entering S-phase in succession
Separating a cell fraction containing DNA
bound proteins
Entire replicative SV40 minichromosomes bearing
func-tionally bound replication proteins can be eluted from
nuclei of virus infected cells by hypotonic buffer [19] The
DNA of mammalian chromatin, however, is organized into
loops of about 5–150 kb firmly attached to the nuclear
matrix [20] Thus, intact cellular chromatin cannot be eluted
from isolated nuclei Interrupting the continuity of the
DNA (e.g by suitable endonucleases) yields elutable
chromatin fragments preferably originating from regions
far from matrix attachment points As DNA replication foci
are probably located near the nuclear matrix, preferably
nonreplicative chromatin fragments might be eluted while
replicative chromatin regions remain attached Therefore,
preserving the natural chromatin/matrix relations and
extracting unbound replication proteins from the nuclei
seemed to be more appropriate for studying the influence of
oxygen recovery after a hypoxic period on DNA-bound
proteins For this purpose, we adopted a protocol described
in [9] with some modifications The modified protocol yields
three fractions which are denoted according to the proteins
they contain Fraction 1 includes all ‘non-nuclear proteins’,
i.e cytosolic proteins separated during hypotonic
prepar-ation of nuclei, fraction 2 contains ‘soluble nuclear proteins’
which are extractable from nuclei by Triton X-100
contain-ing buffer, and in fraction 3, the chromatin fraction, all
proteins remain that resist Triton extraction We supposed that the latter fraction included, besides the common chromatin proteins, functionally DNA-bound replication proteins Because the PCNA protein is loaded by a well-defined actively controlled process onto replicative DNA structures [21], it can be taken as an example of replication proteins recruited to DNA according to the demands of replication
Western blot analyses T24 cells synchronized by starvation/hypoxia were fract-ionated as described Equal amounts of protein from each fraction were separated by SDS gel electrophoresis, blotted onto a Nylon-P membrane and PCNA was immunodetec-ted Figure 5 shows the results obtained from cells incuba-ted hypoxically for 7 h and then stopped or reoxygenaincuba-ted for 5 min, 30 min or 1 h
In cytosolic and soluble nuclear proteins, the amounts of PCNA did not vary under any incubation conditions By contrast, in hypoxic chromatin only very little PCNA was detected However PCNA increased strongly as soon as
5 min after reoxygenation and continued to increase after
30 min and 1 h The pattern of chromatin-bound PCNA suggests that the protein is recruited to DNA as soon as its function in replication is required, after replicon initiation had taken place
Immunofluorescence staining of total cellular and chromatin-bound PCNA
T24 were grown and incubated on coverslips Two sets of hypoxic cells and of cells reoxygenated for 10 min and
30 min were prepared One set of cells was directly fixed after the incubation From the second set, soluble proteins were extracted by washing with buffer containing Triton X-100 prior to fixation As shown in Fig 6A, directly fixed cells show a very similar PCNA content after any incubation condition No visible differences exist between hypoxically incubated and reoxygenated cells The mainly nuclear localization of PCNA is due to the fixation procedure Acetone fixation leads to cell shrinkage and loss of membranes Therefore cytosolic PCNA is not as prominent as in the Western blot (Fig 5) In contrast, when the cells were extracted prior to fixation by Triton
Fig 4 Mitotic index of starved T24 cells under normoxic (s) and
hypoxic/reoxygenated (d) incubation conditions after medium renewal,
respectively T24 cells were grown on coverslips for 44 h The medium
was then renewed and cells were either incubated normoxically for 7 h,
or hypoxically for 7 h and then reoxygenated At the times indicated
incubations were stopped, cells were fixed with acetone/methanol and
total DNA was stained with bisbenzimide Subsequently cells were
photographed and counted The percentage of mitotic cells was
calculated as indicated.
Fig 5 Western blot analyses of cytosolic, soluble nucleosolic and chromatin-bound PCNA from hypoxic and reoxygenated T24 cells Cytosolic, soluble nucleosolic and chromatin-bound proteins were prepared after the indicated incubation conditions (for details see Materials and methods) and equal amounts were separated on an 8% SDS/polyacrylamide gel After blotting onto Hybond-P membrane (Amersham) PCNA was visualized with an anti-PCNA Ig (Santa Cruz Biotechnologies) using the ECL detection procedure H, hypoxic; 5¢,
5 min reoxygenated; 30¢, 30 min reoxygenated; 1 h, 1 h reoxygenated.
Trang 7Fig 6 Immunofluorescence staining of total cellular PCNA and chromatin-bound PCNA Cells were grown on coverslips for 44 h, after which the medium was renewed and cells were incubated hypoxically for 7 h or subsequently reoxygenated for the indicated periods Cells were then either fixed directly or washed three times with extraction buffer to remove soluble proteins prior to fixation PCNA was visualized using anti-PCNA Ig (Boehringer Mannheim, dilution 1 : 100) as the primary and Alexa Fluor 568 (dilution 1 : 200) as the secondary antibody Total DNA was stained with bisbenzimide (A) PCNA immunfluorescence staining of directly fixed T24 cells (B) PCNA immunfluorescence staining of T24 cells extracted prior to fixation The respective incubation conditions of cell cultures are indicated below the images B 0 corresponds to B 1 and is shown in this special case to visualize all cell nuclei present The other bisbenzimide images are not shown as the red PCNA fluorescence is almost identical to the blue DNA fluorescence.
Trang 8buffer, PCNA was barely detectable in nuclei from hypoxic
cells but became visible in nuclei as soon as 10 min after
reoxygenation The proportion of unextractable PCNA
increased significantly from hypoxic to reoxygenated
incubations These results again confirm that PCNA
becomes chromatin-bound only when required for DNA
synthesis
Simultaneous staining of replicating DNA
and DNA-bound PCNA
To demonstrate the connection between active DNA
replication and the appearance of bound PCNA in nuclei,
simultaneous immunodetection of replicating DNA after
BrdU incorporation and PCNA was performed T24 cells
grown on coverslips were incubated hypoxically and then
stopped or reoxygenated for 30 min Labeling with 15 lM
BrdU was started 15 min before the end of either
incuba-tion The cells were extracted prior to fixation and
processed for BrdU and PCNA immunodetection As
shown in Fig 7, hypoxic cells exhibit neither visible BrdU
incorporation nor bound PCNA However, 30 min after
reoxygenation, BrdU incorporation into replicating DNA
was detectable and the amount of PCNA not extractable by
Triton buffer was high in the same cells These results clearly
show that the PCNA staining is colocated with the BrdU staining and this again signifies that PCNA is only bound to chromatin portions where actively replicating DNA is present
Recruitment of proteins involved in replication
to chromatin during the hypoxic period
In contrast to starved T24 cells, which begin to initiate replication after about 4 h following medium stimulation, T24 cells that were exposed to hypoxia after medium exchange start replicon initiation immediately upon reoxy-genation This suggests that the ‘classical’ prereplication complex was already formed under hypoxia We applied the elaborated protocol to investigate the binding of MCM2, MCM3 and Cdc6, which are known to be important components of the prereplication complex as well as Cdk2, which is considered to be (one of) the activating kinase(s) of the complex, after medium renewal before and at the end of hypoxic gassing as well as under normoxic conditions
As shown in Fig 8 MCM3 and Cdc6 are not, and MCM2 and Cdk2 are barely, detectable on chromatin of starved T24 cells (lane 1) This may be caused partly by different sensitivities of the antibodies used However,
Fig 7 Immunofluorescence staining of replicative T24 DNA and chromatin-bound PCNA under hypoxic and reoxygenated incubation conditions T24 cells were grown on coverslips for 44 h The medium was renewed prior to hypoxic gassing Replicative DNA was labeled by incubating the cells for
15 min with 15 l M BrdU at the end of the respective incubation Cytosolic and soluble nuclear proteins were extracted prior to fixation by washing the cells three times with extraction buffer (see Materials and methods) BrdU incorporated in replicating DNA was visualized after denaturation with anti-BrdU–FITC conjugated Ig PCNA was visualized by using anti-PCNA Ig, followed by anti-mouse IgG labeled with Alexa Fluor 568 Total DNA was stained with bisbenzimide The respective incubation conditions are indicated below the images.
Trang 9after medium renewal these proteins become obviously
bound to chromatin under hypoxic and normoxic
condi-tions The signal intensities are slightly stronger under
hypoxic than under normoxic conditions This seems
reasonable, as hypoxic suppression of replicon initiation
accumulates prereplication complexes which disappear
after initiation is completed The latter gradually occur
in cultures not subjected to hypoxia after medium
renewal The lack of Cdc6 may explain the 4 h lag phase
in replication after medium exchange under normoxic
conditions, since prior to replicon initiation prereplication
complexes have to be formed This requires certain
proteins that have to be translated before (especially
proteins with a short half life, such as Cdc6) or whose
mRNA has to be transcribed first
Nevertheless, the known ‘classical’ prereplication
com-plex seems to be formed under hypoxia, rendering the cells
ready to activate scheduled replicon initiations
immedi-ately upon reoxygenation In this context it is noteworthy
that the prereplication complex activating kinase Cdk2
becomes bound to chromatin during hypoxia The
band-ing pattern shows differences compared to Cdk2 of
normoxic cells Under hypoxia the form of the protein
that migrates faster seems to predominate Under
norm-oxic conditions both forms seem to be present at roughly
equal proportions
Discussion
Although common interest focuses on the replication of
cells’ own genome, replication of SV40 DNA frequently
serves as a convenient model of mammalian (human) DNA
replication However, when the cellular replication
equip-ment is abused for viral multiplication, cellular mechanisms
are often falsified or put out of function, in particular the regulatory mechanisms involved Decisive experiments concerning regulatory phenomena have to be performed
in a cellular system in the long term
The aim of the present study was to establish means for extending a recent study [8] on changes of replication proteins bound to SV40 minichromosomes, occurring in the context of the fast O2-dependent regulation of replica-tion [6–8], from the viral system to a (preferably human) cellular system Thus we were confronted with two main problems Firstly, inducing in as many as possible cellular replicons the ‘hypoxic preinitiation state’ and excluding as completely as possible active replicons in other states Secondly, preparing a cell fraction containing only those replication proteins which are functionally associated with cellular chromatin and not those located in cytosolic or nucleosolic fractions
With respect to the first problem, we initially tried to use Ehrlich ascites cells With these cells we first demonstrated the existence of the fast O2-dependent regulation of replication [12,22,23] We had already developed means to select vital G1cells from cell cultures by a zonal centrifu-gation procedure [16] and succeeded to bring them homo-geneously to hypoxic arrest in which they bore exclusively early S-phase replicons in the desired ‘hypoxic preinitiation state’ [3] Although resuming the old experiments principally confirmed the suitability of the Ehrlich ascites cell system for the present purpose, we searched for alternatives because the selection procedure is complicated, time consuming and works only with a mouse cell line (i.e Ehrlich ascites), while most available antibodies are directed against human replication proteins
Consequently we next examined a set of human cell lines, e.g CCRF, HeLa [4], PC3, A549, BHK, TC7, SW2, HL60 and HUVEC with respect to their response to hypoxia and reoxygenation The alkaline sedimentation profiles of HeLa and CCRF cells after hypoxia and reoxygenation already revealed [4] that hypoxic incubation caused significant accumulation of initiation competent replicons, which could
be released into a more or less synchronous round of replication upon reoxygenation However the extent of replicon synchrony attained by the hypoxic incubation alone, i.e absence of active replicons in the state of elongation, turned out to be insufficient for examining the transition reaction between the hypoxic and the reoxygen-ated state in a satisfying specific manner The same problem occurred with the other cell lines examined Inhibitors such
as thymidine or aphidicolin were not used, as they inhibit elongation and not replicon initiation Furthermore, we had shown previously that initiation is not blocked in SV40-infected CV1 cells treated with aphidicolin prior to reoxy-genation [6]
Fortunately, we observed that in the human bladder cancer cell line T24 the effect of hypoxia/reoxygenation could be intensified five- to 10-fold when the medium was renewed prior to hypoxia We suspected that these cells had been (at least partly) arrested in G1 simply by preceding starvation as formerly described by Prescott [17] Our experiments confirmed this suspicion After the optimal starvation conditions were found, starved T24 cells were incubated hypoxically directly after stimulation by medium renewal This treatment accumulated cellular replicons
Fig 8 Western blot analyses of chromatin-bound MCM2, MCM3,
Cdc6 and Cdk2 from normoxic and hypoxic T24 cells
Chromatin-bound proteins were prepared after the indicated incubation
condi-tions (for details see Materials and methods) and equal amounts were
separated on an 8% SDS/polyacrylamide gel After blotting onto
Hybond-P membrane (Amersham) the respective proteins were
immunodetected using the ECL detection procedure Lane 1,
norm-oxia without medium renewal (i.e beginning of the experiment); lane 2,
normoxia 7 h after medium renewal; lane 3, 7 h hypoxia after medium
renewal.
Trang 10almost exclusively in the ‘hypoxic preinitiation state’ It
should be mentioned that T24 cells proceed normally
through the cell cycle after hypoxia/reoxygenation for
several days No signs of apototic cell death could be
detected by the CaspaTagTMCaspase (VAD) Activity Kit
(Intergen, Oxford, UK) during and after the hypoxic
treatment (data not shown)
With respect to the second problem, we demonstrated
by means of the PCNA example that T24 nuclei extracted
by Triton X-100 buffer contain functionally
chromatin-bound replication proteins, switching from another cellular
compartment into the chromatin fraction (Figs 5–7) or
undergoing changes of modifications (e.g
phosphoryla-tion) in response to O2 recovery of hypoxic cells PCNA
seemed most suitable because elongation is not affected [2]
or just slowed down under hypoxia [4] and ongoing
elongation is dependent on functional PCNA We suggest
that the absence of chromatin-bound PCNA under
hypoxia is rather a direct consequence of missing
initi-ation, i.e lost activation of the ‘hypoxic preinitiation
complex’, than an impairment of ‘clamp loading’ by
replication factor C
Since T24 cells start to replicate immediately upon
reoxygenation, transcriptional or translational processes
can be excluded as cause of the hypoxic arrest It was
already shown for Ehrlich ascites cells that the expression of
growth related mRNA is not influenced during transient
hypoxia [1]
DNA replication in eukaryotes is initiated by the stepwise
assembly of proteins to the replication origin [10,24,25]
First the hexameric origin recognition complex binds [26],
which then recruits Cdc6 [27,28], cdt1 [29,30] and the
minichromosome maintenance proteins [31] This
prerepli-cation complex is built up during G1of the cell cycle The
complex is presumably activated by cyclin-dependent kinase
Cdk2 [32,33] and the Dbf4/cdc7 [34] kinase, which is
required to load the initiation factor Cdc45 on the
prereplication complex [35–37] To investigate whether this
prereplication complex is built under hypoxia we performed
a first experiment using the above described protocol We
show that MCM2/MCM3 and Cdc6, as well as the
activating kinase Cdk2, present in two modifications with
different electrophoretic mobilities, become bound to
chro-matin already under hypoxia, thus enabeling hypoxic cells
to initiate as soon as the hypoxic suppression of replicon
initiation is released The relative intensities of the two Cdk2
bands differ under hypoxia and normoxia Possibly, this
represents a modification of the kinase influencing its
activity/inactivity Post-translational processes such as
modifications (e.g phosphorylations or
dephosphoryla-tions) of proteins have already been found to be important
regulators in SV40 replication [38,39]
Our ongoing work now focuses on changes arising in the
pattern of further chromatin-bound proteins of hypoxic and
reoxygenated cell T24 cells We use classical Western blot
analyses with immunodetection of replication proteins or
regulators as well as high resolution 2D-gels
With the aid of the T24 system presented here, we hope to
characterize the special state of the protein equipment of the
replication of human cells under the hypoxic block and to
elucidate the fast events occurring as an effect of oxygen
recovery
Acknowledgements
We thank G Probst for critical reading of the manuscript.
References
1 Riedinger, H.J., Gekeler, V & Probst, H (1992) Reversible shutdown of replicon initiation by transient hypoxia in Ehrlich ascites cells Dependence of initiation on short-lived protein Eur.
J Biochem 210, 389–398.
2 Probst, H., Schiffer, H., Gekeler, V., Kienzle-Pfeilsticker, H., Stropp, U., Stotzer, K.E & Frenzel-Stotzer, I (1988) Oxygen dependent regulation of DNA synthesis and growth of Ehrlich ascites tumor cells in vitro and in vivo Cancer Res 48, 2053–2060.
3 Gekeler, V & Probst, H (1988) Synchronization of replicons in Ehrlich ascites cells Exp Cell Res 175, 97–108.
4 Probst, G., Riedinger, H.J., Martin, P., Engelcke, M & Probst, H (1999) Fast control of DNA replication in response to hypoxia and to inhibited protein synthesis in CCRF-CEM and HeLa cells Biol Chem 380, 1371–1382.
5 Brischwein, K., Engelcke, M., Riedinger, H.J & Probst, H (1997) Role of ribonucleotide reductase and deoxynucleotide pools in the oxygen-dependent control of DNA replication in Ehrlich ascites cells Eur J Biochem 244, 286–293.
6 Riedinger, H.J., van Betteraey, M & Probst, H (1999) Hypoxia blocks in vivo initiation of simian virus 40 replication at a stage preceding origin unwinding J Virol 73, 2243–2252.
7 Dreier, T., Scheidtmann, K.H & Probst, H (1993) Synchronous replication of SV40 DNA in virus infected TC7 cells induced by transient hypoxia FEBS Lett 336, 445–451.
8 Riedinger, H.J., Betteraey-Nikoleit, M & Probst, H (2002) Re-oxygenation of hypoxic simian virus 40 (SV40)-infected CV1 cells causes distinct changes of SV40 minichromosome-associated replication proteins Eur J Biochem 269, 2383–2393.
9 Liang, C & Stillman, B (1997) Persistent initiation of DNA replication and chromatin-bound MCM proteins during the cell cycle in cdc6 mutants Genes Dev 11, 3375–3386.
10 Dutta, A & Bell, S.P (1997) Initiation of DNA replication in eukaryotic cells Annu Rev Cell Dev Biol 13, 293–332.
11 Probst, H., Hofstaetter, T., Jenke, H.S., Gentner, P.R & Muller-Scholz, D (1983) Metabolism and non-random occurrence of nonnascent short chains in the DNA of Ehrlich ascites cells Biochim Biophys Acta 740, 200–211.
12 Probst, H & Gekeler, V (1980) Reversible inhibition of replicon initiation in Ehrlich ascites cells by anaerobiosis Biochem Biophys Res Commun 94, 55–60.
13 Englard, S & Seifter, S (1990) Precipitation techniques Methods Enzymol 182, 285–300.
14 Laemmli, U.K (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 227, 680–685.
15 Dimitrova, D.S., Todorov, I.T., Melendy, T & Gilbert, D.M (1999) Mcm2, but not RPA, is a component of the mammalian early G1-phase prereplication complex J Cell Biol 146, 709–722.
16 Probst, H & Maisenbacher, J (1973) Use of zonal centrifugation for preparing synchronous cultures from Ehrlich ascites cells grown in vivo Exp Cell Res 78, 335–344.
17 Prescott, D.M (1976) The cell cycle and the control of cellular reproduction Adv Genet 18, 99–177.
18 Yan, Z., DeGregori, J., Shohet, R., Leone, G., Stillman, B., Nevins, J.R & Williams, R.S (1998) Cdc6 is regulated by E2F and
is essential for DNA replication in mammalian cells Proc Natl Acad Sci USA 95, 3603–3608.
19 Su, R.T & DePamphilis, M.L (1978) Simian virus 40 DNA replication in isolated replicating viral chromosomes J Virol 28, 53–65.