From the Department of Cell Biology and Cancer
Center, University of Massachusetts Medical Center, Worcester,
Massachusetts 01655, the § Division of Molecular Oncology,
Departments of Medicine and Pathology, Washington University School of
Medicine, St. Louis, Missouri 63110, and the ¶ Department of
Immunology, Faculty of Medicine, University of Tokyo, Hongo 7-3-1, Bunkyo-ku 113, Tokyo
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Histone genes display a peak in transcription in
early S phase and are ideal models for cell cycle-regulated gene
expression. We have previously shown that the transcription factor
interferon regulatory factor 2 (IRF-2) can activate histone H4 gene
expression. In this report we establish that a mouse histone H4 gene
and its human homolog lose stringent cell cycle control in synchronized embryonic fibroblasts in which IRF-2 has been ablated. We also show
that there are reduced mRNA levels of this endogenous mouse histone
H4 gene in the IRF-2/
cells. Strikingly, the
overall mRNA level and cell cycle regulation of histone H4
transcription are restored when IRF-2 is reintroduced to these cells.
IRF-2 is a negative regulator of the interferon response and has
oncogenic potential, but little is known of the mechanism of these
activities. Our results suggest that IRF-2 is an active player in
E2F-independent cell cycle-regulated gene expression at the
G1/S phase transition. IRF-2 was previously considered a
passive antagonist to the tumor suppressor IRF-1 but can now join other
oncogenic factors such as c-Myb and E2F1 that are predicted to mediate
their transforming capabilities by actively regulating genes necessary
for cell cycle progression.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Interferon regulatory factor 2 (IRF-2)1 and the related factor IRF-1 are negative and positive regulators, respectively, of the interferon response and are implicated in growth control with IRF-2 having oncogenic potential and IRF-1 demonstrating tumor suppressor activity (1-3). One putative mechanism of the oncogenic activity of IRF-2 is that IRF-2 antagonizes the antiproliferative action of IRF-1 by competing for binding sites in the promoters of several growth-suppressing genes (4-6). An alternative hypothesis is that IRF-2, also known as histone nuclear factor M (HiNF-M), plays a positive role in the cell cycle regulation of the human histone H4 gene FO108 (7-10). Unlike many other genes that are regulated at the G1/S transition point such as dihydrofolate reductase and thymidine kinase, FO108 is regulated by an E2F-independent mechanism (11-13). The proximal promoter element of FO108 that interacts with IRF-2 is essential for the peak in transcription of this gene in early S phase and was designated a cell cycle element (CCE) (8). However, other protein factors can also interact with the CCE, namely HiNF-P (H4TF2) and HiNF-D, a CDP/cut, Cdc2, cyclin A, pRB complex (9, 10, 14, 15).
Point mutations in the FO108 promoter that disrupt in
vitro binding of any of these factors individually have limited
effects on the cell cycle regulation of this gene
(8).2 This result could be
due to partial functional redundancy among these factors, the use of
alternative binding sites, or differential binding properties of these
factors in vivo. Weak interactions of IRF-2 and
CDP/cut with alternative binding sites in the
FO108 promoter have been observed in electrophoretic
mobility shift assays
(EMSAs).3 Therefore, to
definitively establish the role of IRF-2 in histone H4 transcription
regulation, we have performed cell cycle analyses using embryonic
fibroblasts from IRF-2 knockout (IRF-2/
) mice (16).
![]() |
EXPERIMENTAL PROCEDURES |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Stable Cell Lines--
NIH 3T3 or embryonic fibroblasts derived
from IRF knockout mice, maintained as described previously (7), were
transfected with a wild type H4 promoter chloramphenicol
acetyltransferase (CAT) construct (7) along with pSV-2neo (17) by the
calcium phosphate precipitation method (18). 48 h after
transfection, cells were maintained in media containing 400 µg/ml
Genticin (Life Technologies, Inc.). Resistant colonies were pooled and
assayed for CAT enzymatic activity as described (7). To generate the IRF-2 "add-back" cell lines, H4-CAT IRF-2/
cells
were transfected with pcDNA1-IRF-2 (7, 19) along with the
hygromycin-resistant plasmid pCEP4 (Invitrogen). 48 h after transfection, cells were maintained in 400 µg/ml Geneticin and 60 units/ml hygromycin B (Calbiochem). Individual resistant colonies that
expressed both CAT mRNA and IRF-2 protein were isolated and expanded.
EMSA-- Whole cell extracts from the various cell lines were prepared as described (20). EMSAs were carried out with 10 µg of each extract and 10 fmol of radioactive CCE probe as described previously (7). Where indicated, 1 pmol of unlabeled oligonucleotide competitor DNA (either wild type or mutant CCE as described previously (7)) was included.
Cell Synchronization and DNA Synthesis Assays-- Cells maintained as monolayer cultures were synchronized by the double thymidine block method (21). After the final release from block, cells were collected at hourly time points for RNA isolation and DNA synthesis assays by [3H]thymidine uptake (21).
Northern Analysis-- Total RNA was isolated using Trizol reagent according to the manufacturer (Life Technologies, Inc.). 10 µg of total RNA from each time point was separated in an agarose/formaldehyde gel as described (18), blotted on Zeta-Probe (Bio-Rad), and probed with a rat histone H4 fragment (22).
RNase Protection and S1 Nuclease Assays--
Antisense RNA
probes for CAT and cyclophilin (internal control) were prepared from
pTriCAT and pTriCyclophilin plasmids using a MaxiScript kit according
to the manufacturer (Ambion) except that the cyclophilin probe was made
with 200 µM UTP and 20 µCi of
[-32P]UTP. RNase protection assays were performed with
the RPA II kit according to the manufacturer (Ambion) using 10 µg of
total RNA, 8 × 104 cpm of CAT probe, and 3.5 × 104 cpm of cyclophilin probe and digesting with 0.5 units
of RNase A and 5 units of RNase T1.
![]() |
RESULTS AND DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
To compare the cell cycle regulation of a histone H4 gene in
synchronized IRF-2+/+ and IRF-2/
cells, a
construct containing the wild type FO108 promoter sequence fused to CAT (7) was stably transfected into NIH 3T3 or
IRF-2
/
embryonic fibroblasts. Integration of the H4-CAT
in each cell type was assayed by CAT enzymatic activity (data not
shown).
The H4-CAT stably transfected lines were synchronized by the double
thymidine block method, which arrests cells at the G1/S phase boundary (21). The synchronization was monitored by DNA synthesis
assays and Northern analysis of cellular histone H4 mRNA levels
performed at hourly time points after release from block (Fig.
1). Northern analysis revealed an
8-10-fold peak in total cellular histone H4 mRNA levels at the
onset of S phase (Fig. 1A), which is consistent with
previous studies (8, 23, 24). This peak in total histone mRNA
levels is due to both transcriptional and post-transcriptional
mechanisms, with the latter involving sequences in the histone
3-untranslated region (reviewed in Ref. 25). Therefore, to monitor
only the transcriptional contribution to the regulation of histone H4,
we measured CAT mRNA levels by RNase protection, which in this case
is a direct measure of the stably integrated human histone H4 promoter
activity.
|
A single protected fragment of CAT mRNA was detected for each time point as seen at the top of Fig. 1B. In the IRF-2+/+ NIH 3T3 cells, the levels of CAT mRNA peaked concomitantly with the peak of DNA synthesis and showed a 2.5-3-fold increase over base-line levels (Fig. 1B). This magnitude of increase in transcriptional activity agrees with many previous studies from several laboratories in which a 2-5-fold increase in transcription of histone genes occurred at the onset of S phase both in vivo and in vitro (8, 23, 24, 26-28). In contrast, cyclophilin mRNA, which was used as an internal control, showed no fluctuations in mRNA levels over the synchronized time course (Fig. 1C). These results indicate that human H4 transcription is cell cycle-regulated in mouse cells in a manner similar to the regulation of this H4 promoter in human cells (8).
We next examined the regulation of histone H4 in synchronized H4-CAT
IRF-2/
cells. Strikingly, cells that lack IRF-2 show no
transient increase in the levels of CAT mRNA in a synchronized time
course (Fig. 2). This result was
reproducibly seen in several experiments (Fig. 2, inset).
Therefore, IRF-2 is critical for the cell cycle regulation of this
promoter.
|
Interestingly, the presence of IRF-1, which is able to bind and
activate transcription of the H4 promoter in vitro (7), is
unable to promote the proper cell cycle regulation of the H4-CAT gene
in IRF-2/
cells (Fig. 2). This supports our previous
study in which IRF-1 was unable to compensate for IRF-2-mediated
activation of the H4 promoter in vivo (7). These results
demonstrate the specificity of IRF-2 for this cell cycle-regulated
gene.
We directly assessed whether IRF-2 is a key mediator in H4 gene
transcription during the cell cycle by reintroducing IRF-2 into the
H4-CAT IRF-2/
cells. These cells were stably
transfected with a human IRF-2 expression vector, and colonies were
assayed for the presence of IRF-2 by EMSA. Four positive isolates are
shown in Fig. 3 along with the other cell
types used in this study. As expected, no IRF-2-specific complexes are
seen in extracts from IRF-2
/
cells (Fig. 3, lane
4), while at least two protein DNA complexes are seen in extracts
from NIH 3T3, IRF-1
/
, HeLa, and the IRF-2 add-back cell
lines (Fig. 3, lanes 1-3 and 5-8). Assignment
of the complexes from NIH 3T3, IRF-1
/
, and HeLa as
specific for IRF-2 are based on oligonucleotide competition and
antibody identification (Refs. 7 and 10 and data not shown). The exact
migration of the protein-DNA complexes from the cells in which human
IRF-2 was introduced into mouse cells varies slightly from both the
human HeLa cells and the mouse NIH 3T3 and IRF-1
/
cells, but these complexes are IRF-2-specific as shown by
oligonucleotide competition (Fig. 3, lanes 9-11).
Therefore, we have successfully reintroduced IRF-2 protein in the
IRF-2
/
cells.
|
In a cell cycle analysis of one of the IRF-2 add-back cell lines, a transient peak of H4-CAT transcription is again concomitant with the peak in DNA synthesis (Fig. 4). Similar results were obtained with all four of the IRF-2 add-back cell lines analyzed (Fig. 4, inset). Therefore, the reintroduction of IRF-2 is sufficient to restore the cell cycle-regulated transcription of the human histone H4 gene in the IRF-2 knockout cells.
|
Having established that IRF-2 is necessary for cell cycle-regulated
transcription of the introduced H4-CAT gene, one would expect an effect
on the endogenous histone H4 genes. To test this prediction, we
examined the regulation of mouse histone H4 genes in
IRF-2/
and IRF-2+/+ cells. Eukaryotic cells
have multiple H4 genes that display nearly identical coding sequences
but divergent promoter regions. It is thought that these multiple genes
are required to ensure proper chromatin structure in various cell types
or stages of cell development. We selected two mouse H4 genes, one of
which is the apparent homolog of FO108 (designated A (29))
and another that is quite divergent (designated B (30)). We established
the binding properties of the CCE interacting factors HiNF-M/IRF-2,
HiNF-P, and HiNF-D to the two mouse promoter sequences by EMSA (Table
I). Mouse H4-A has the same binding
properties as the human sequence, while H4-B does not interact with
IRF-2 and has only weak interaction with HiNF-D. We therefore
anticipated that H4-A would be regulated in a manner analogous to the
human H4 gene FO108 and that H4-B would not be regulated by
IRF-2.
|
We determined the effect of IRF-2 on endogenous histone mRNA by
comparing the expression levels of H4-A and H4-B in proliferating IRF-2/
, IRF-2 add-back, or IRF-1
/
control cells by S1 nuclease protection analysis using specific oligonucleotide probes complementary to a unique segment of the 5
-untranslated regions of H4-A and H4-B. Fig.
5 shows that endogenous H4-A mRNA
levels were approximately 3-fold lower in IRF-2
/
cells
compared with the IRF-1
/
cells, which have normal IRF-2
levels (Fig. 5A, compare lanes 1 and
3). Furthermore, the levels of H4-A are restored when IRF-2 is reintroduced to these cells (Fig. 5A, compare lanes
1 and 2). In contrast, the H4-B gene, which does not
interact with IRF-2, generates similar mRNA levels in all three
cell types (Fig. 5A, lanes 5-7).
|
We then monitored the cell cycle regulation of endogenous histone H4
genes in various cell types. Comparison of H4-A and H4-B mRNA
levels during the cell cycle of synchronized IRF-2/
cells shows that there is a transient increase in H4-A mRNA during the cell cycle although the amplitude of the peak is reduced and slightly delayed when compared with the regulation of the H4-B message
(Fig. 6, A and B).
This moderate level of cell cycle regulation of histone H4-A message in
IRF-2
/
cells is not unexpected, because the
post-transcriptional mechanisms of cell cycle regulation discussed
above are presumably operational for this gene (while they are not a
factor in the regulation of CAT mRNA, which was assayed in Figs. 1,
2, and 4). In contrast, maximal cell cycle regulation of histone H4-A
(and H4-B, not shown) is seen in the IRF-2-positive
IRF-1
/
cells (Fig. 6C). Furthermore, the
maximal regulation of mouse histone H4-A (and H4-B, not shown) is
restored when IRF-2 is reintroduced to the IRF-2
/
cells
(Fig. 6D). Therefore, IRF-2 plays a role in the overall expression and cell cycle regulation of a subset of the mouse histone
H4 genes.
|
Along with previous results demonstrating that IRF-2 can activate both a growth control gene (7) and a differentiation-specific gene (gp91phox) (31), the present study further implicates IRF-2 as an active participant in growth control and not only a passive antagonist of the tumor suppressor IRF-1. The mechanism by which IRF-2 contributes to cell cycle regulation is not yet known but may involve the recently described phosphorylation of IRF-2 (32, 33). Although IRF-2 directly interacts with the H4 promoter, it is not yet clear whether the IRF-2 effect on cell cycle regulation is direct. IRF-2 may be involved in the recruiting of other factors to the promoter region of the H4 gene. It is very likely that the cell cycle regulators present in the HiNF-D complex (cyclin A, Cdc2, and pRB), which interact with multiple histone genes (14), are necessary components of histone H4 cell cycle regulation.
In conclusion, the results described in this report clearly establish for the first time that the oncogenic transcription factor IRF-2 is necessary for correct cell cycle regulation of histone H4 gene transcription and that IRF-2 is required for the coupling of histone gene expression with S phase progression in an E2F-independent manner.
![]() |
ACKNOWLEDGEMENTS |
---|
We thank Danielle Lindenmuth, Jack Green, and Elizabeth Buffone for technical assistance and the members of the Stein laboratory for helpful discussions.
![]() |
FOOTNOTES |
---|
* This work was supported by National Institutes of Health Grant GM 32010.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
To whom correspondence should be addressed: Dept. of Cell
Biology and Cancer Center, University of Massachusetts Medical Center, 55 Lake Ave. N., Worcester, MA 01655. Tel.: 508-856-5625; Fax: 508-856-6800; E-mail: gary.stein{at}banyan.ummed.edu.
1 The abbreviations used are: IRF, interferon regulatory factor; CCE, cell cycle element; HiNF, histone nuclear factor; CAT, chloramphenicol acetyltransferase; EMSA, electrophoretic mobility shift assay.
2 F. Aziz, manuscript in preparation.
3 T. J. Last, P. S. Vaughan, J. L. Stein, and G. S. Stein, unpublished observations.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|