(Received for publication, October 1, 1996, and in revised form, April 16, 1997)
From The School of Food and Nutritional Sciences, University of Shizuoka, 52-1 Yada, Shizuoka-shi, Shizuoka 422, Japan
To identify the essential motifs of the promoter
of the human gene for thymidylate synthase (TS), we constructed a set
of deletion mutants from the 5-terminal region of the human TS gene. From the results of assays of the expression of chloramphenicol acetyltransferase (CAT), we identified two functional elements with
positive effects on the promoter activity: a CACCC box (CCACACCC) and
an Sp1-binding motif (GAGGCGGA) that was homologous to the Sp1-binding
site in the mouse TS gene. In addition, negative regulatory sequences
were identified between the two positive elements and in the region
upstream of the CACCC box. The results of gel mobility shift analyses
suggested that Sp1 binds to the Sp1-binding motif of the human TS gene
promoter and that multiple nuclear factors that are related to Sp1 bind
to the CACCC box. Furthermore, the binding of Sp1 to mutated
Sp1-binding motifs in the promoter region of the human TS gene was
correlated with the promoter activity, as measured by the CAT assay.
Therefore, the Sp1 motif seems to be a major contributor to the basic
promoter activity of the human TS gene, although multiple positive and
negative regulatory elements are involved in the regulated expression
of this gene.
Thymidylate synthase (TS; N5,N10-methylenetetrahydrofolate:dUMP C-methyltransferase; EC 2.1.1.45)1 catalyzes the conversion of deoxyuridylate to thymidylate, and the enzyme is known as the key enzyme in nucleotide metabolism. Tight regulation of TS activity is essential for the normal replication of DNA, and impairment of this enzyme causes various biological and genetic abnormalities, such as thymine-less death (1), chromosome breakage and exchange (2), the expression of heritable fragile sites (3), and genetic recombination (4-6). The regulation of the expression of the TS gene depends on the proliferative state of cells (7-10). For example, the activity of the enzyme and the level of mRNA for human TS (hTS) are very low in quiescent human fibroblast cells, but both increase dramatically when the cells are stimulated by serum to enter the G1/S phase (11).
To study the regulated expression of the hTS gene, we cloned and
characterized the hTS gene (12-15). The hTS gene is 18 kilobase pairs
in length and consists of seven exons. Regions essential for the
regulation that is dependent on the stage of the cell cycle have been
identified both in the first intron (15, 16) and in the 5-flanking
region (16) of the hTS gene. Although the expression of the hTS gene is
mainly regulated at the post-transcriptional level (11), these findings
suggest the involvement of the promoter region in the cell
cycle-dependent regulation of the gene. However, functional
motifs in the promoter region and mechanism for regulation of the
promoter activity of the hTS gene remain to be elucidated. The
5
-flanking region of the hTS gene has neither a TATA box nor a CAAT
box (13). The mouse TS gene also has neither a TATA box nor a CAAT box
in its promoter region and an Sp1-binding site in the promoter region
has been reported to be essential for the promoter activity (17). The
Sp1- binding site in the mouse TS gene exhibits 90% homology with the
corresponding region of the hTS gene, suggesting that the human and
mouse genes for TS might share common motifs that are essential for
promoter activity. However, the regulatory elements identified in the
region upstream from the Sp1-binding motif of the mouse TS gene (17)
are not conserved in the hTS gene. Furthermore, the region upstream of the cap site of the hTS gene exhibits only 36% homology over 100 bp
with the corresponding region of the mouse TS gene. These data suggest
that most of the regulatory elements are not conserved between the
human and mouse genes for TS in the upstream region, implying the
possibility that some elements might contribute specifically to the
regulation of the promoter of the hTS gene. In previous studies, we
showed that the essential promoter sequence is in the region from
242
to
148 (nucleotide positions in the hTS gene in this report are
numbered from the first nucleotide of the codon for initiation of the
translation of the hTS gene) and that, in addition to sequences
upstream of the cap site, the tandemly repeated sequences downstream
from the cap site are necessary for the sufficient expression of the
hTS gene (18, 19). In this study, we examined sequences that were
essential for promoter activity using a transient expression system and
we demonstrated that a CACCC box (CCACACCC) and the region that
includes the sequence homologous to the Sp1-binding site of the mouse
TS gene are necessary for efficient promoter activity. We also
identified two negative regulatory sequences in the promoter region of
the hTS gene and characterized the nuclear factors that interact with
the positive elements in the promoter region of the hTS gene.
Enzymes were purchased from Takara Shuzo (Kyoto,
Japan) and Toyobo (Osaka, Japan). ES medium (20) and fetal calf serum
for cell culture were products of Nissui Seiyaku (Tokyo, Japan) and HyClone Laboratories (Logan, UT), respectively. DNA polymerase from
Thermus aquaticus (AmpliTaq) was obtained from Perkin-Elmer (Foster City, CA), agarose S for electrophoresis from Nippon Gene (Toyama, Japan), [-32P]dCTP (3,000 Ci/mmol) and
D-threo-[1,2-14C]chloramphenicol (40-60
mCi/mmol) from ICN Biomedicals Inc. (Costa Mesa, CA). A mouse
monoclonal antibody against Sp1 was purchased from Santa Cruz
Biotechnology, Inc. (Santa Cruz, CA). A nonspecific mouse IgG was
purchased from Bio-Rad.
The 587-bp ScaI-BglI
fragment, including the cap sites and the site of initiation of
translation of the hTS gene, was prepared from pHRR68, a subclone of
the hTS gene (12), and inserted into the HindIII site of the
pIGCAT vector (18) via a 10-bp HindIII linker, such that the
initiation codon of the hTS gene was in frame with a CAT gene on the
pIGCAT vector. From the resultant plasmid (pScBg6CAT), we prepared a
set of deletion plasmids using exonuclease III. Each plasmid was
sequenced to determine the deletion site. A cassette mutation or a
point mutation was introduced into the promoter region of the hTS gene
by PCR (21). To introduce a BglII linker sequence at a
specific site in the promoter region of the hTS gene, we synthesized
oligonucleotide primers (LS2a, etc. in Table I) that
contained a BglII site at the 5-end and used them with a
5
-primer and a 3
-primer (Primer 1 and Primer 2, respectively, in Fig.
2) for amplification of the promoter region of the hTS gene. A
HindIII site was added to the 5
-end of Primer 2. The
sequences of the primers used for PCR are listed in Table I. For the
PCR, pScBg6CAT was used as template. The amplified fragment included
the region from
283 to +22 and had a HindIII site at its
3
-end. The products of PCR were cloned into pUC19, and the sequences
of inserted fragments were determined. The resultant plasmids were
digested with PvuII and HindIII, and each
PvuII-HindIII fragment containing the promoter
region of the hTS gene was purified by electrophoresis on a 4% agarose
gel. The fragments with mutations were subcloned between a blunt-ended SalI site and a HindIII site of the pIGCAT vector
and used for assays of CAT activity. A point mutation was introduced
into the Sp1-binding site of the promoter of the hTS gene by PCR (21). The plasmid carrying the promoter of the hTS gene with a point mutation
in the Sp1-binding motif was digested with XbaI and
HindIII, and the XbaI-HindIII fragment
containing the promoter region with the point mutation was used to
replace the XbaI-HindIII fragment of the promoter
of the hTS gene in the deletion plasmid, d229 (Fig. 1). Structures of
the cassette and point mutants are summarized in Fig. 2.
|
Cell Culture and Transfection
HeLa S3 (SC) cells were
obtained from the Japanese Cancer Research Resources Bank (Tokyo,
Japan) and cultured in ES medium (20) that contained 10% fetal calf
serum. The calcium phosphate coprecipitation method was used for
transfection of cells with plasmid DNA. Each transfection reaction was
performed with 2.5 µg of CAT plasmid and 5.0 µg of
pSV--galactosidase plasmid (Promega, Madison, WI) per 60-mm Petri
dish. The conditions for transfection and quantitative analysis of CAT
activity have been described previously (18).
Nuclear extracts were
prepared from HeLa cells as described by Schreiber et al.
(22). The conditions for the electrophoretic mobility shift assay were
the same as those described previously (23), with the exception that
the mixture after the binding reaction was analyzed by electrophoresis
on a 4% polyacrylamide gel. A PvuII-XbaI
fragment and an XbaI-BssHII fragment from the promoter region of the hTS gene were used for gel mobility shift assays
as probes (see Fig. 2). To identify binding sites of nuclear factors,
DNA fragments with cassette or point mutations were used. These
fragments were labeled with [-32P]dCTP using the
Klenow fragment of DNA polymerase I and used for the assay as probes.
Competitor DNA fragments, each including a binding site for a nuclear
factor, for the gel mobility shift assays were purchased from
Stratagene and Promega. The sequences of the respective fragments were
as follows: AP1, 5
-CTAGTGATGAGTCAGCCGGATC-3
; AP2,
5
-GATCGAACTGACCGCCCGCGGCCCGT-3
; AP3,
5
-CTAGTGGGACTTTCCACAGATC-3
; Sp1, 5
-GATCGATCGGGGCGGGGCGATC-3
;
NFI/CTF, 5
-ATTTTGGCTTGAAGCCAATATG-3
; NF-
B,
5
-AGTTGAGGGGACTTTCCCAGGC-3
; TFIID,
5
-GCAGAGCATATAAGGTGAGGTAGGA-3
; GRE,
5
-TCGACTGTACAGGATGTTCTAGCTACT-3
; CREB,
5
-AGAGATTGCCTGACGTCAGAGAGCTAG-3
; OCT, 5
-TGTCGAATGCAAATCACTAGAA-3
.
For competition studies, various amounts of an unlabeled competitor
fragment were mixed with binding buffer and a nuclear extract, as
indicated in legends to the figures, for 5 min prior to the addition of
the labeled probe. To identify nuclear factors that bind to DNA probes,
0.25 or 0.5 µg of mouse monoclonal antibody against Sp1 was added to
the reaction mixture containing 2.5 µg of protein of nuclear
extracts, prior to the addition of the labeled probe.
In a previous study, the
essential region of the promoter of the hTS gene was identified as the
region from 242 to
148 (18). To identify the essential motifs in
this promoter region, we constructed a set of deletion mutants that
covered the region from
441 to +28 of the hTS gene. The structures of
the deletion mutants were indicated on the left in Fig.
1. Among the fragments shown in Fig. 1, the fragments
including the region upstream of position
270 had no significant
promoter activity. Promoter activity was detected with the fragments
that included the regions from
229 to
223 and from
187 to
147.
These results suggest that the negative regulatory element is in the
region from
342 to
269. Although deletion mutant plasmids d223 and
d212 included the region from
187 to
147, these two plasmids had
lower promoter activity than d201. This observation suggests that the
region from
212 to
201 decreased the promoter activity of the
region from
187 to
147. Thus, the results shown in Fig. 1 suggest
that the promoter region of the hTS gene contains two negative
regulatory sequences and two positive elements that influence the
promoter activity.
We found several DNA
motifs in the two positively acting regions that we had identified from
the results of the CAT assay with the deletion mutants. The upstream
positively acting region, namely the region from 229 to
223,
overlapped with the CACCC box (Ref. 24; CCACACCC from
228 to
221).
The downstream positive region, from
187 to
147, contained a
sequence homologous to the Sp1-binding site (AAGAGGCGGA, from
152 to
143 of the hTS gene) of the mouse TS gene (17). This region contains
the consensus sequence of an Sp1-binding site (27). To examine the
function of these DNA motifs, we introduced cassette mutations and a
point mutation into the promoter region of the hTS gene. The structures of these mutants are shown in Fig. 2 and the results of
CAT assays with these constructs are shown in Fig. 3.
The M-CA mutation in Fig. 2 had three substitutions of nucleotides in
the CACCC box in the promoter of the hTS gene. The promoter activity of
the mutant plasmid (M-CA) was 65% of the activity of the native
promoter of the hTS gene ("No mutation" in Fig. 3). The
fragment with the mutation indicated by M-Sp had five substitutions of
nucleotides in the Sp1-binding motif of the promoter of the hTS gene
and had no significant promoter activity in the CAT assay
(M-Sp in Fig. 3). A single nucleotide substitution in the
Sp1-binding motif reduced the promoter activity to half of that of the
parental fragment (see d229 and d229(146A) in
Fig. 3). These results suggest that, in the hTS gene, the CACCC box and
the sequence homologous to the Sp1-binding site of the mouse TS gene
function as positive elements for the promoter activity and that the
Sp1-binding site is essential for the expression of the gene.
Furthermore, we introduced a cassette mutation into the negative
regulatory region that we had identified in the CAT assays of the
deletion mutants. The M-NRS mutant had four substitutions in the
negative regulatory sequence that was located between two positive
elements in the promoter region. Its promoter activity was 2.5-fold
higher than that of the native promoter of the hTS gene. This result
suggests that this region represses the promoter activity of the hTS
gene.
Sp1 Binds to the CACCC Box and the Sp1-binding Motif of the hTS Gene Promoter
To identify nuclear factors that regulate the
promoter activity of the hTS gene, we examined the specific binding of
nuclear factors to the functional motifs in the promoter of this gene. To examine the individual nuclear factors that bind to the CACCC box or
the Sp1-binding motif in the promoter, we used a
PvuII-XbaI fragment and an
XbaI-BssHII fragment (Fig. 2) from the promoter region of the hTS gene as probes in the gel mobility shift assay. Fig.
4 shows the formation of nucleoprotein complexes that
contained the XbaI-BssHII fragment. The mutated
fragments used in the CAT assays were also used for the preparation of
the probes. When the fragment with the mutation indicated by M-Sp (Fig.
2) was used as the probe, the rate of formation of the major
nucleoprotein complex was markedly reduced (compare the band
a in lanes 1 and 4 in Fig. 4A).
B in Fig. 4 indicates that the point mutation in the
Sp1-binding motif also inhibited the formation of the major complex
with the XbaI-BssHII fragment. The results
obtained in the competition assays with DNA fragments that contain
known binding motifs for specific transcription factors are shown in
C in Fig. 4, and the results indicate competition for the
formation of the major complex upon the addition of a DNA fragment with
the binding motif for Sp1 (lanes 9 and 10 in Fig.
4C). These results suggest that the major complex involved
the Sp1-binding motif in the XbaI-BssHII fragment
of the hTS gene promoter together with Sp1. Taking into consideration
the fact that the M-Sp mutant had no significant promoter activity in
the CAT assay, as compared with other mutants or the parental fragment
of the promoter (Fig. 3), and that the point mutation in the
Sp1-binding motif reduced the promoter activity of the fragment in the
CAT assay, we concluded that the formation of the complex that involved
the Sp1-binding motif was correlated with the promoter activity of the
mutants that had been measured by the CAT assay.
When the PvuII-XbaI fragment (Fig. 2) was used as
a probe, several shifted bands were observed in the gel mobility shift
assay (lane 1 in Fig. 5A). Among
these bands, two bands (bands b and e in Fig.
5A) were not detected when the fragment with the mutation in
the CACCC box (M-CA) was used as the probe. To examine the sequence
specificity of the nuclear factors that bind to the fragment, competitor DNA fragments that contain known binding motifs for specific
transcription factors were added to the binding reaction mixture. The
results of the competition assay showed that intensities of almost all
the bands, including the bands whose intensities diminished when the
fragment with the mutation indicated by M-CA was used as the probe,
were reduced upon the addition of the fragment with the Sp1-binding
motif (bands b, d, and e in lanes 9 and 10 in Fig. 5B). The remaining band was
subject to competition by the DNA fragment with the AP2-binding motif
(band c in lanes 5 and 6 in Fig.
5B). These results suggest that multiple nuclear factors
that recognize the Sp1 binding motif bind to the
PvuII-XbaI fragment of the promoter of the hTS
gene, and at least two of them interact with the CACCC box in the
fragment, while a nuclear factor that binds to the consensus sequence
for binding of AP2 also binds to the PvuII-XbaI
fragment. In fact, computer search did reveal that an AP2 consensus
motif (YCSCCMNSSS; Ref. 25) was present in the region from 206 to
197 in the XbaI-BssHII fragment (see Fig. 2).
Thus, the AP2-binding motif is a candidate for the binding site of an
AP2-like factor.
Then, we examined the binding of Sp1 to the
XbaI-BssHII fragment of the hTS gene using
monoclonal antibody against Sp1. Fig. 6 shows the effect
of the antibody on the formation of nucleoprotein complexes when the
Sp1 monoclonal antibody was added to the binding mixture of the gel
mobility shift assay. The result indicated that the major retarded band
disappeared by the addition of the antibody (band a in
lane 6 in Fig. 6). The band that disappeared corresponds to
the nucleoprotein complex formed with the Sp1-binding motif of the hTS
gene (band a in Fig. 4). These results suggested that the
nucleoprotein complex formed with the Sp1-binding motif in the
XbaI-BssHII fragment includes Sp1.
The promoter
region of the hTS gene does not contain any DNA motifs that are typical
of eukaryotic promoters, such as a TATA box, a CAAT box, or a typical
GC box. Therefore, we attempted to identify the motifs that are
essential for the promoter activity. In studies of the mouse TS gene,
multiple nuclear factors, including Sp1 and Ets-like factors, were
reported to bind to the promoter region, and the Ets/Sp1-binding motifs
were shown to play an important role in the expression of the mouse TS
gene (17, 26). In a previous study, we identified the region that is
essential for the promoter activity of the hTS gene, as well as the
negative regulatory region that is present upstream of the essential
promoter region (18). In this study, we characterized the functional DNA motifs in the essential region of the promoter of the hTS gene. Two
positively acting motifs were found in the promoter region of the hTS
gene: a CACCC box in the region from 228 to
221 and an Sp1-binding
motif in the region from
150 to
142.
The Sp1-binding motif found in the promoter region of the hTS gene is GAGGCGGAG and it matches the consensus sequence, KRGGCGKRRY (27). However, it does not include a typical GC box (GGGCGG). The motif is highly homologous to the Sp1-binding site of the mouse TS gene that has been reported to be essential for the expression of the mouse TS gene (17). Indeed, in the case of the human TS gene, introduction of four substitutions around the Sp1-binding motif reduced the promoter activity to the background level in the transient expression assay, indicating that the motif is essential for the promoter activity of the hTS gene. The Sp1-binding motif is conserved in the human, mouse (17), rat (28), and monkey (29) genes for TS. These findings suggest that the Sp1-binding motif is the core sequence of the promoter that maintains the basic promoter activity of genes for TS.
The major cap sites of the mouse TS gene have been identified at
positions downstream from the Sp1-binding site (30). By contrast, the
major cap site of the hTS gene was located at position 179 from the
results of a primer extension experiment (14). This position is located
upstream of the Sp1-binding motif. In a previous study, Takeishi
et al. (13) suggested the presence of cap sites downstream
from the Sp1-binding site from the results of S1 nuclease mapping.
However, the results could not be confirmed by primer extension
experiments, because the unique, tandemly repeated structure of the hTS
gene interfered with the identification of the cap sites around the
Sp1-binding site (14). The finding that the common Sp1-binding motif is
essential for the promoter activity of the human and mouse genes for TS
suggests that the major cap site of the hTS gene is also present in the
region downstream of the Sp1-binding motif.
The CACCC box in the hTS gene is located at 71 bp upstream of the
Sp1-binding motif, and it appears to have positive effect on the
promoter activity of the hTS gene. This motif is not conserved in the
corresponding region of the mouse and rat genes for TS. However, in the
case of the mouse and rat genes, a CACCC box is found about 160 bp
upstream of the Sp1-binding motif. The function or biological
significance of the CACCC box in the mouse and rat genes is unknown.
However, it is possible that the motif might play a role in the
regulation of expression of the mouse or rat gene. The CACCC motif is
found in the promoter region of the gene for -globin, in the
enhancer region of the early gene of SV40, and in many other genes
(31). Multiple transcription factors, including Sp1 and EKLF (erythroid
Krüppel-like factor), have been reported to bind to a CACCC box
(31, 32). This observation suggests that the motif can function as
either a promoter element or an enhancer element. In the case of the
hTS gene, inactivation of the Sp1-binding motif downstream of the CACCC
box reduced the promoter activity to the background level, whereas
three substitution of nucleotides in the CACCC box reduced the promoter
activity to 65% of that of the native hTS gene promoter. Therefore,
the CACCC box seems not to be a promoter element but, rather, to be a
regulatory element essential for the appropriate promoter activity of
the hTS gene.
We also examined the specific interaction of nuclear factors with the functional DNA motifs found in the promoter region of the hTS gene. The results of gel mobility shift analysis revealed that Sp1 bound to the Sp1-binding motif of the hTS gene (Figs. 4 and 6). Sp1 has been reported to bind also to the corresponding region of the mouse TS gene, and the binding is important for the expression of the gene (17). In the case of the mouse TS gene, Johnson and his colleagues (33, 34) reported that an E2F-binding motif and Ets-binding motifs (GGAAG) play a role in the growth-regulated expression of the gene. These functional motifs in the mouse TS gene are not conserved in the corresponding regions of the human TS gene, whereas an Ets-binding motif, GGAAG, adjacent to the Sp1-binding motif is conserved in the human TS gene. When the DNA fragment with an Sp1-binding motif was used as a competitor, the band due to formation of the nucleoprotein complex that included the Sp1-binding site was completely lost (Fig. 4C). Furthermore, the formation of the complex that included the Sp1-binding site was inhibited by the addition of the monoclonal antibody against Sp1 (Fig. 6). This result suggests that, at least in HeLa cells, the binding of Sp1 is essential for the formation of the complex at the Sp1-binding site.
The results of gel mobility shift assays suggest that nuclear factors
bound to the CACCC box (bands b and e in Fig.
5A). The binding of the factors was affected by the
introduction of mutations into the CACCC box (Fig. 5A) and
subject to competition by the DNA fragment with the Sp1 binding motif
(Fig. 5B). Based on these results, we suggest that the major
factor that bound to the CACCC box of the hTS gene is Sp1 or an
Sp1-related factor. This finding is consistent with the previous
finding that Sp1 binds weakly to the CACCC box (31). The CACCC box of
the gene for -globin is reported also to be a target of EKLF, and
the binding of EKLF is known to regulate tissue-specific expression of
the gene (32). It will be of interest to examine whether the CACCC box
of the hTS gene is a target for tissue-specific factors, including
EKLF, because some growth-related genes, such as the gene for DNA
polymerase
(35), have been reported to be expressed in a
tissue-specific manner.
Two negative regulatory sequences were identified by CAT assays with
the deletion mutants of the promoter of the hTS gene (Fig. 1), and the
function of one of these sequences was confirmed by CAT assays with the
cassette mutant of the promoter (M-NRS in Fig. 3). Negative
regulation of the expression of the TS gene should be important,
because it is necessary that the level of thymidylate in cells should
remain constant throughout the cell cycle and the overexpression of the
TS gene is toxic to cells, as is a defect in the expression of the
gene. Comparison of the sequences of these two negative regulatory
regions revealed the presence of a consensus sequence, TTCCC (Fig.
7). The TTCCC sequence is included in the consensus
motif of the E2F-binding sequence (36), which is important in the
regulation of transcription that is dependent on the stage of the cell
cycle. Many genes expressed in the G1-S phase, such as
genes for dihydrofolate reductase and thymidine kinase and the mouse TS
gene, have E2F-binding motifs in the promoter region (36). The hTS gene
also has a potential E2F-binding site in the region from 128 to
121. We previously examined the function of the motif and found that
the motif decreased the expression of the hTS gene in a transient
expression assay in HeLa cells (19). These findings suggest that the
TTCCC motif or a related motif functions as a negative motif in the
regulation of the hTS gene.
In conclusion, we have identified the basic functional motifs in the promoter region of the hTS gene. Among these motifs, the Sp1-binding motif was identified as a core promoter motif that is conserved in the human, mouse, rat, and monkey genes for TS. Furthermore, other positive and negative regulatory motifs that are specific to the hTS gene are present in the promoter region of the gene.