ABSTRACT
Isolation of cDNA clones for the mouse CCAAT binding factor (mCBF) has revealed
the expression of two distinct forms of mCBF that are generated by alternative
splicing of a single primary transcript from a gene that maps to chromosome 17.
The mCBF1 mRNA encodes a protein of 997 amino acids, whereas the mCBF2 protein
is predicted to be only 461 amino acids in length; mCBF1 and human CBF (hCBF)
share >80% amino acid sequence identity. Analysis of adult mouse tissue RNAs
has revealed that the mCBF1 and mCBF2 mRNAs are ubiquitously expressed, but
that mCBF1 mRNA is 5- to 10-fold more abundant than mCBF2 mRNA. Similarly, mCBF mRNA was
detected throughout the placenta and in all tissues of the developing embryo
from day 8 to day 18 of gestation. Overexpression of the two forms of mCBF in
mammalian cells has demonstrated that the mCBF1 and mCBF2 proteins localize to
different cellular compartments, with mCBF1 found predominantly in the nucleus
and mCBF2 restricted to the cytoplasm. Co-expression of these two forms influences their localization, however,
indicating that CBF activity can be regulated by the relative amounts of the
two forms expressed in a cell.
Previous investigations of human heat shock protein 70 gene (
hsp70
) expression led to the identification of a CCAAT box at -70 as a promoter element critical for serum-inducible transcription (
1
-
3
) and the protein CBF as the transcription factor that acts through this element
(
4
). Although several CCAAT factors have been identified, some of which have been
shown to be able to bind to the CCAAT element of the
hsp70
gene promoter (
4
-
8
), CBF appears to be unique in its ability to activate transcription from this
promoter (
4
). In addition, CBF mediates activation of the
hsp70
gene promoter by the adenovirus E1a oncoprotein (
9
) and repression of this promoter by the p53 tumor suppressor protein (
10
). The effects of E1a and p53 on this promoter are apparently due to the ability
of these regulatory factors to form protein-protein complexes with CBF (
9
,
10
). Thus CBF appears to represent a critical node in mammalian cells for both
growth promoting and growth repressing signaling pathways.
To date the analysis of CBF has been restricted to the human factor. To expand
these studies we sought to isolate the mouse homolog of hCBF and to use this
cDNA clone to characterize the forms of CBF synthesized in the mouse, to
compare the sequences of the mouse and human proteins to reveal conserved
domains, to map the chromosomal location of the
Cbf
gene in the mouse and to analyze the developmental expression and the tissue
distribution of CBF mRNA. The results reported here identify a previously
undetected form of CBF and demonstrate that the ability of this transcription
factor to move into the nucleus depends on the relative amounts of the
different CBF forms present in the cell.
BALB/c 3T3 and COS cells were maintained in Dulbecco's modified Eagle's medium
(DMEM; Gibco-BRL) supplemented with 10% fetal calf serum (Gibco-BRL), glutamine, penicillin and streptomycin. Cell cultures were
starved by incubation in medium containing 0.5% serum for 48 h and then
stimulated by addition of fresh medium containing 15% serum for varying lengths
of time. DNA transfections of COS cells were performed using DEAE-dextran (
11
). For these transfections 5 * 10
5
COS cells were transferred into each 10 cm dish 24 h before addition of 5-20 [mu]g plasmid DNA. Cells were harvested 48 h post-transfection.
Total RNA was purified from BALB/c 3T3 cells that had been stimulated with serum
for 2, 4, 6 and 8 h by centrifugation of guanidinium thiocyanate lysates
through CsCl cushions (
12
). Equal amounts of RNA from these four time points were combined, selected by
oligo(dT)-cellulose chromatography for poly(A)
+
RNA and reverse transcribed into single-stranded cDNA. RNA-DNA duplexes were converted into double-stranded DNA using RNase H and DNA polymerase I (
12
) and inserted into the [lambda]ZAP vector (Stratagene), resulting in a library of 1.3 * 10
6
independent clones. A random primed probe was prepared (
12
) from the hCBF cDNA (
4
) to screen the library. The mCBF1 and mCBF2 cDNAs were recovered from the [lambda]ZAP clones by phagemid excision and transferred into the pSP72 bacterial
plasmid (Promega) and into the pMT2 mammalian expression vector (
13
). Dideoxy sequencing of the cDNA clones was performed using Sequenase 2.0
(United States Biochemicals); the sequences have been submitted to GenBank
(U19891 and U19892).
Interspecific backcross progeny were generated by mating (C57BL/6J *
Mus spretus
) F
1
females and C57BL/6J males as described (
14
). A total of 205 N
2
mice were used to map the
Cbf
locus (see text for details). DNA isolation, restriction enzyme digestion,
agarose gel electrophoresis, transfer to Zetabind nylon membranes (AMF-Cuno) and filter hybridization were performed essentially as described (
15
). The probe, a 2.2 kb
Xho
I fragment of the mCBF1 cDNA, was labeled with [[alpha]-
32
P]dCTP using a nick-translation labeling kit (Boehringer-Mannheim) and following hybridization the filters were washed at a
final stringency of 0.5* SSCP (75 mM NaCl, 7.5 mM sodium citrate, 2 mM sodium phosphate), 0.1%
SDS at 65oC. Fragments of 19.5 and 7.8 kb were detected in
Eco
RV-digested C57BL/6J DNA and fragments of 12.5 and 5.7 kb were detected in
Eco
RV-digested
M.spretus
DNA. The presence or absence of the 12.5 and 5.7 kb
M.spretus
-specific
Eco
RV fragments, which co-segregated, was followed in backcross mice.
The probes and restriction fragment length polymorphisms (RFLPs) for the loci
linked to
Cbf
, including laminin A subunit (
Lama
), mouse homolog-1 of Sos (
Msos1
) and antiphosphotyrosine immunoreactive kinase (
Tik
), have been described previously (
16
,
17
). Recombination distances were calculated as described (
18
) using the computer program SPRETUS MADNESS. Gene order was determined by
minimizing the number of recombination events required to explain the allele
distribution patterns.
Reverse transcription polymerase chain reaction (RT/PCR) assays were performed
as described previously (
19
). In brief, 5 [mu]g total RNA were reverse transcribed using random hexamer primers (Pharmacia
Biotech) and the resulting cDNA was subjected to PCR in the presence of [[alpha]-
32
P]dATP with the mCBF-specific oligonucleotide primers 5'-TAAGCTGGGAGATCCTCAGAACAG-3' and 5'-GGCGGCATCTGTGTGCAGGTGACC-3'; ribosomal L19
oligonucleotide primers were included as an internal control (
19
). Products were extracted with phenol/chloroform, precipitated with ethanol,
resolved by polyacrylamide gel electrophoresis and visualized by
autoradiography.
For
in situ
hybridization mouse fetuses were collected from pregnant Swiss-Webster mice (Harlan Breeding Laboratory) at days 8, 10, 12, 14, 16 and 18
of gestation and frozen at -80oC. Hybridizations were performed as described previously (
19
) with antisense and sense riboprobes generated by
in vitro
transcription of the linearized pSP72-mCBF1 construct in the presence of [[alpha]-
33
P]UTP (DuPont-New England Nuclear).
For immunofluorescence detection of CBF COS cells were grown on glass coverslips
and transfected with CBF expression constructs or vector alone. Cells were
fixed with 2% paraformaldehyde at room temperature for 10 min, followed by
permeabilization with ice-cold methanol for 5 min. Cells were then treated with 3% bovine serum
albumin (BSA) in phosphate-buffered saline (PBS) for 1 h at room temperature. Purified IgG from a
rabbit antiserum raised against recombinant hCBF was diluted to a final
concentration of 50 [mu]g/ml (the polyclonal antiserum recognizes both the human and the mouse
proteins). The secondary antiserum, Texas red-conjugated goat-anti-rabbit IgG, was purchased from Vector Laboratories and was
visualized with a Zeiss Axiophot microscope.
To detect CBF by immunoblotting, extracts were prepared from transfected COS
cells as described (
20
) and 100 [mu]g protein were fractionated by SDS-PAGE and transferred to nitrocellulose (Biotrace). Filters were incubated with 5% non-fat milk in low salt buffer (20 mM Tris-HCl, pH 7.6, 150 mM NaCl, 0.05% Triton X-100) before addition of 2 [mu]g/ml rabbit polyclonal anti-CBF IgG. Filters were washed in low salt
buffer and then in high salt buffer (20 mM Tris-HCl, pH 7.6, 1 M NaCl, 0.4%
N
-lauryl sarcosine). Goat-anti-rabbit IgG coupled to alkaline phosphatase (Cappel) was added
and detected with an alkaline phosphatase conjugate kit (BioRad).
Epitope-tagged forms of CBF were generated by inserting double-stranded oligonucleotides encoding either a single copy of the
influenza hemagglutinin (HA) tag (YPYDVPDYA) or the FLAG tag (DYKDDDDK)
immediately after the translation initiation ATG codon in the mCBF1 and mCBF2
cDNAs. The tagged proteins were detected with anti-HA monoclonal antibody 12CA5 (Berkeley Antibody) or anti-FLAG monoclonal antibody M2 (IBI-Eastman Kodak). In transfections with one tagged construct
binding of the primary antibody was detected with FITC-conjugated horse anti-mouse IgG (Vector Laboratories). In co-transfection experiments with both tagged constructs FITC-conjugated rat anti-mouse IgG1 and biotinylated rat anti-mouse IgG2b (both from Zymed Laboratories)
were used to detect anti-FLAG M2 antibody and anti-HA 12CA5 antibody respectively. Texas red avidin D (Vector
Laboratories) was used to detect the presence of biotinylated secondary
antibody. The secondary antibodies were found to be specific, such that rat
anti-mouse IgG1 did not recognize the 12CA5 antibody and the rat anti-mouse IgG2b failed to interact with the M2 antibody. Immunoblots of
HA- and FLAG-tagged proteins were incubated with the 12CA5 and M2 antibodies and
then developed with alkaline phosphatase-conjugated goat anti-mouse IgG.
The mouse chromosomal location of
Cbf
was determined by interspecific backcross analysis using progeny derived from
matings of (C57BL/6J *
M.spretus
) F
1
* C57BL/6J mice. This interspecific backcross mapping panel has been typed
for over 1800 loci that are well distributed among all the autosomes as well as
the X chromosome (
14
). C57BL/6J and
M.spretus
DNAs were digested with several enzymes and analyzed by filter hybridization
for informative RFLPs using the mCBF1 cDNA (see Materials and Methods). The
mapping results indicated that
Cbf
is located in the distal region of mouse chromosome 17, linked to
Lama
,
Tik
and
Msos1
. Although 136 mice were analyzed for every marker and are shown in the
segregation analysis (Fig.
3
), up to 180 mice were typed for some pairs of markers. Each locus was analyzed
in pairwise combinations for recombination frequencies using the additional
data. The ratios of the total number of mice exhibiting recombinant chromosomes
to the total number of mice analyzed for each pair of loci and the most likely
gene order are: -
Lama
-9/176-
Tik
-0/180-
Cbf
-1/149-
Msos1
. The recombination frequencies (expressed as genetic distances in cM +- SE) are: -
Lama
-5.1 +- 1.7-[
Tik
,
Cbf
]-0.7 +- 0.7-
Msos1
. No recombinants were detected between
Tik
and
Cbf
in 180 animals typed in common, suggesting that the two loci are within 1.7 cM
of each other (upper 95% confidence limit).
The levels of expression of mCBF1 and mCBF2 mRNAs in adult tissue and in the
developing conceptus at different stages of gestation were determined by RT/PCR
analysis. To detect both forms of mCBF mRNA primers were utilized that flank
the splice site, so that amplification of mCBF1 and mCBF2 would yield fragments
of 634 and 347 bp respectively. As an internal control primers were included to
amplify the mouse ribosomal L19 mRNA (
19
). Both the mCBF1 and mCBF2 mRNAs were present in all tissues examined (Fig.
4
). The identities of the PCR fragments were confirmed by cloning and sequencing,
thereby demonstrating that the mCBF2 cDNA clone represents a naturally
occurring mRNA. Although the level of each of these mRNAs was found to be
approximately constant among all of these tissues, the concentration of the
mCBF2 mRNA was significantly lower than that of the mCBF1 mRNA in each sample
(Fig.
4
), consistent with their representation in the cDNA library. The intensities of
the mCBF1 and mCBF2 RT/PCR products were determined by phosphorimager analysis
and normalized to the amount of the L19 product. In all tissues the amount of
the mCBF1 mRNA was ~5- to 10-fold greater than the mCBF2 mRNA; for the BALB/c 3T3 cell line
the ratio of mCBF1 to mCBF2 mRNA was ~20:1 (Fig.
4
).
Both mCBF1 and hCBF contain a potential nuclear localization signal near their C-termini (residues 942-946, Leu-Arg-Lys-Ala-Arg in mCBF1 and 943-947, Thr-Lys-Lys-Ser-Lys
in hCBF) which is not present in mCBF2. To determine if the mCBF1 and mCBF2
proteins localize to distinct cellular compartments expression constructs
containing the mCBF1 and mCBF2 cDNAs were transfected into COS cells. By
immunofluorescence staining mCBF1 was found to translocate efficiently to the
nucleus, whereas mCBF2 remained in the cytoplasm and accumulated in the
perinuclear region (Fig.
6
). A low level of fluorescence in both the nucleus and in perinuclear structures
was detected in untransfected cells (Fig.
6
), probably from endogenous CBF1 and CBF2 proteins.
In a screen for the mouse homolog of hCBF cDNA we have identified two mRNA
forms, mCBF1 and mCBF2, that arise by alternative splicing of a single
transcript. The N-terminal 454 amino acids of mCBF1 and mCBF2 are identical. This region
includes residues 1-192, which in hCBF are sufficient to mediate specific interactions with
both the adenovirus E1a oncoprotein and the p53 tumor supressor protein (
9
,
10
). Thus mCBF1 and mCBF2 may overlap in the set of proteins with which they
interact, including mouse p53.
The mCBF2 protein lacks the C-terminal half of mCBF1, which includes the putative nuclear localization
signal. Consistent with this difference, mCBF1 was detected predominantly in
the nucleus, whereas mCBF2 was found to localize primarily in the cytoplasm.
The detection of some mCBF1 in the cytoplasm (and some cells with mainly
cytoplasmic mCBF1) and some mCBF2 in the nucleus (with some cells containing
mostly nuclear mCB2) indicates that these two forms of CBF can move between
these two major cellular compartments. Since the transfected cell cultures were
not synchronized, one possible explanation for these results is that mCBF1 and
mCBF2 localization is cell cycle regulated.
Cell cycle alterations in protein localization might simply reflect the period
just after the completion of M phase, when some mCBF1 is in the cytoplasm and
has not yet translocated back into the reformed nucleus and some mCBF2 has been
captured in the newly formed nucleus but not yet been transported back into the
cytoplasm. However, co-expression of elevated levels of mCBF1 and mCBF2 resulted in nuclear mCBF2
and cytoplasmic mCBF1 in a high percentage of cells. Thus the mechanism of
mCBF1 and mCBF2 redistribution in the cell cannot be explained solely by
nuclear envelope breakdown and reassembly. Instead, mCBF1 and mCBF2 probably
interact, either directly or through accessory factors, to regulate
localization. Therefore, variations in the relative amounts of these two forms
of CBF, or variations in mCBF post-translational modifications that may influence protein-protein interactions, may occur at specific stages of the cell
cycle to regulate localization and function. Transient alterations in the
cellular distributions of these proteins may have signficant effects,
especially considering that the one known target gene for CBF, the
hsp70
gene, is serum inducible (
1
) and cell cycle regulated (
21
) and that the CBF binding protein p53 also undergoes a conformational change in
response to serum (
22
). Furthermore, a truncated form of hCBF similar in length to mCBF2 is able to
stimulate transcription from the
hsp70
gene promoter and to target E1a to this promoter (
9
), indicating that this form is capable of binding DNA.
Since mCBF1 and hCBF both contain the p53 and E1a interaction domain it is
possible that mCBF2 also acts as a dominant interfering form of CBF,
sequestering proteins in the cytoplasm that are needed for mCBF1-mediated transcriptional activation in the nucleus. Alternatively, mCBF1
and mCBF2 might share co-factors but act on different processes. For example, mCBF1 may interact
with p53 in the nucleus to regulate transcription, whereas mCBF2 might
cooperate with p53, which has been shown to bind to the 5.8S rRNA (
23
) to regulate protein synthesis in the cytoplasm. The accumulation of mCBF2
protein in the perinuclear region of the cytoplasm suggests that it may
function in association with specific structures or compartments.
One possible means of identifying potential physiological effects of mCBF1 and
mCBF2 is to correlate the chromosomal location of the
Cbf
gene with mapped mutations that cause abnormalities in the mouse. We have
compared the interspecific map of chromosome 17 with a composite mouse linkage
map that reports the location of many uncloned mouse mutations (compiled by
M.T.Davisson, T.H.Roderick, A.L.Hillyard and D.P.Doolittle and provided from
GBASE, a computerized database maintained at The Jackson Laboratory, Bar
Harbor, ME). However,
Cbf
was found to map in a region of the composite map that lacks known mouse
mutations (data not shown). The distal region of mouse chromosome 17 shares a
region of homology with human chromosomes 18p and 2p (summarized in Fig.
3
). In particular,
Tik
has been placed on human 2p22-p21. The tight linkage between
Tik
and
Cbf
in the mouse suggests that
Cbf
will also reside on human 2p.
Both mCBF1 and mCBF2 mRNAs were found in all the tissues of the adult mouse that
were examined, in the placenta, throughout the developing embryo from day 8 to
day 18 of gestation and in actively growing 3T3 cells. Furthermore, no tissues
could be detected in the embryo or adult that lacked mCBF mRNA. Thus mCBF may
be active in most, if not all, cell types during fetal and placental
development and in the adult. One difference that was consistently observed
between mCBF1 and mCBF2 was a 5- to 10-fold greater level of mCBF1 compared with mCBF2 mRNA in each tissue.
Since mCBF2 can apparently regulate mCBF1 translocation into the nucleus, it
will be of interest to determine if the relative levels of mCBF2 and mCBF1
change under various physiological conditions.
The ability of hCBF to mediate transcriptional induction and repression of the
growth-regulated
hsp70
gene by regulatory factors such as the adenovirus E1a and cellular p53 proteins
suggests that mCBF1 and mCBF2 may be important components of cell growth
regulation in the mouse. The finding that mCBF1 and mCBF2 are ubiquitously
expressed during fetal development and in the adult further suggests that their
actions will be found to be of general importance in the regulation of cell
function.
We thank Ellise Estes and B.Cho for excellent technical assistance and Doug
Engel for comments on the manuscript. This work was supported by NIH grants to
BW (GM42465) and DL (HD29962), by the P30 Center in Reproductive Biology at
Northwestern University (HD28048) and by the National Cancer Institute, DHHS,
under contract N01-CO-46000 with ABL.
REFERENCES
Return