Cell Type-Specific
Regulation
of
Transcription
by Cyclic Adenosine
3,‘5’-Monophosphate-Responsive
Elements within the Calcitonin
Promoter
Yomna
T. Monia,
Sara Peleg,
and Robert
Section of Endocrinology
Department of Medical Specialties
The Universitv of Texas M. D. Anderson
Houston, Texas 77030
F. Gage1
Cancer
Center
A CAMP-induced
enhancer
was
previously
mapped
to nucleotides
-255 to -85 of the calcitonin (CT) gene Y-flanking
DNA. To determine
the functional
&-acting
elements
within this region, we transfected
medullary
thyroid
carcinoma (MTC) cells with CT 5’-flanking
DNA/GH
fusion genes containing
potential
CAMP-responsive elements
and assessed
their transcriptional
activities
with and without
CAMP. In CT-expressing MTC cells (the TT line), we identified
by deletions and point mutations
three transcriptionally active motifs: a CAMP-responsive
element
(CRE), TGACGTCA,
at -253 to -248, and a hybrid
site containing
a CRE-like element (CREL; TGACCTCA, - 189 to - 182) adjacent
to an equally transcriptionally
active octamer
(0) sequence
(ATGCAAAT, -181 to -154). These three motifs acted
synergistically
and their transcriptiohal
activity
was completely
dependent
on CAMP. In HeLa
cells their synergistic
activity was more constitutive than CAMP induced,
whereas
in CT-negative MTC cells (the RO-D81-1
line) these motifs
were inactive.
Gel mobility
shift assays with antibodies
against CRE-binding
protein (CREB) and
activating
transcription
factor 1 (ATF-1) showed
that both CREB and ATF-1 interacted
with the
CRE in MTC cells, whereas
in HeLa cells only
ATF-1 bound to the CRE. Specific
binding
to the
CREUO
motif was detectable
in extracts
from
tumors induced by injection
of TT cells but not in
extracts from any of the three cultured
cell lines.
We conclude
that CAMP-induced
transcription
of
the CT gene is modulated
in a cell-specific
manner by the CRE and the CREUO elements.
(Molecular Endocrinology
9: 784-793,
1995)
osss-9909/95/$3.00/0
Molecular
Endocrinology
Copyright 0 1995 by The Endocrine
INTRODUCTION
The gene encoding
calcitonin
(Cl) and CT-generelated-peptide
is expressed in neuronal and thyroid C
cells (1). It is also expressed
in neoplasms
derived
from neuroendocrine
cells, such as medullary thyroid
carcinoma (MTC) (2).
We and others previously demonstrated
that transcription of the CT gene in transformed
C cells (IT
cells) is modulated by a distal neuroendocrine-specific
enhancer and a proximal CAMP-induced
enhancer.
The distal enhancer is responsible
for cell-restricted
basal expression
of the gene, whereas the proximal
enhancer is responsible for CAMP-induced
transcription (3, 4). We showed previously that the two enhancers can function
independently
but together
act in synergism. Vitamin D down-regulates
the CT
gene, probably by interfering with this synergistic interaction (5). The distal elements involved in this interaction are two CANNTG motifs, which are potential
binding sites for helix-loop-helix
proteins (6).
To date, the functional &-acting
elements within
the proximal CAMP-induced
enhancer have not been
clearly identified. To understand better the mechanism
of CAMP-induced
transcription
and the synergistic interaction between the two enhancers, we had to identify the functional
&-acting
elements
within the
CAMP-induced
enhancer
and characterize
the proteins that bind to these proximal elements. A computer
analysis of the CAMP-induced
enhancer region (-255
to -85) revealed multiple sequences that may function
as binding sites for CAMP-regulated
transcription
factors (7, 8). These include a CAMP responsive element
(CRE) that may be a binding site for members of the
CREB/ATF family (9-l 1); a CRE-like (CREL) motif, one
base different from CRE, flanked by two octamer sequences (0) that are potential binding sites for home01
Pit-l, OTF 1 and 2, and UNC 86 (POU) domain proteins (12-16); and two potential
activator
protein
(AP)-2 binding sites (17) (Fig. 1).
Society
784
induction of Transcription
by CAMP
785
I
Al?
II
-189/-180
-89/-98
TCCCC(A/C)N(G/C)
TCCCC C
A
T'XCC C
T
C
G
(G/C)
C
G
(G/C)
C
G
I
Fig. 1. Structure of the CT CAMP-Induced Enhancer Region
Shown are the locations of the sequences that are potential binding sites for CAMP-regulated transcription factors. The
consensus sequences (upper sequences in each box) and their homologs (lower sequences) within the CAMP-induced enhancer
are given. Pit-l, Pituitary-specific transcription factor from the homeo/POU domain family (12-14); Ott 1-6, octamer-binding
proteins that are members of the homeo/POU domain family (15, 16).
In this study, we further analyzed CT CAMP-induced
enhancer function in CT-expressing
and CT-negative
tumor lines and mapped the functional &-acting
elements within this enhancer region using DNA transfer
techniques (4, 5). Furthermore, we began characterizing the CAMP enhancer-binding
proteins using electrophoretic mobility shift assays and specific antibodies against members
of the CREB/ATF
family of
transcription
factors.
RESULTS
Mapping
the Functional
within the CAMP-Induced
c&-Acting
Elements
CT Enhancer
In previous studies, we showed that the CT 5’-flanking
DNA contained
a proximal CAMP-induced
enhancer
that functioned independently
of and synergized with
a distal constitutive
enhancer (5). To map functional
cis-acting
elements
within the CAMP-induced
enhancer, it was necessary to test their activities independently
of the constitutive
enhancer
elements.
Therefore, we created the construct pCAMP-1, which
lacked the basal enhancer but contained the potential
CAMP-induced
enhancer elements shown in Fig. 1.
Then, we prepared a series of mutant constructs
by
sequential 5’-deletion
of the pCAMP-1
and by introducing point mutations in the potential cis-acting elements. These constructs were transfected into lT cells
(18) and their activities were tested before and after
addition of the CAMP analog. The CAMP-induced
activity of these mutants was expressed as a percentage
of the CAMP-induced
activity of pCAMP-1 and not as
fold induction of basal promoter activity because in TT
cells this activity was undetectable.
In TT cells, we detected an approximately
S-55%
loss in CAMP-induced
transcription when the upstream
CRE (TGACGTCA) motif was deleted or mutated (Fig. 2).
Deletion of CT nucleotides -247 to -175, which includes the potential upstream AP-2 site and part of the
upstream octamer sequence, had no further effect on
cAMP?nduced
transcription.
There was an additional
2530%
loss of CAMP-induced
activity when the CREL
(TGACCTCA) site was mutated or when CT nucleotides
-175 ,to -129 were deleted. Similar results were obtained”by mutating or deleting the adjacent downstream
octamer ATGCAAAT sequence. ‘Hence the hybrid element TGACCTCAATGCAAAT
is now identified as
CREUO. Interestingly, a mutation of the CREUO motif in
a construct containing the CRE (pCAMP-7) reduced
CAMP-induced
transcription by 80%. These results suggest that the CREUO element was essential for transcriptional activity of the CRE even though the CREUO
element could also function independently of the CRE in
lT cells.
Figure 2C demonstrates
the inverse relationship between passage number and the CAMP-induced
activity of CREUO. At early passages (30-40), this motif
contributed
80% of CAMP-induced
transcription,
whereas at later passages (170-180)
it contributed
only 20% of this activity. Taken together, these results
suggest that in lT cells CAMP-induced
transcription of
the CT gene is due to a CRE and to another, equally
important, hybrid element (CREUO) containing both a
CREL motif and an octamer motif.
MOL
786
END0
.1995
Vol 9 No. 7
100
50
Qrouth
flormona
(% &UP-traatmd
&CAMP-l)
0
B.
=-p
QRSV-CAT
,
0
c.
1.0
p 3040
q pCAnP-1 (CRB+CREL/O)
n pcAmP-4
(CRBL/O)
0
0.5
CAT activity
(unit&
pQ protein)
,
PCTQS
(promoter
p 110-130
only)
-0
(I
Qrcmth
Bornone
CAMP-treated
pC?XP-1)
Fig. 2. Mapping
the Functional
&-Acting
Elements
of the CT CAMP-Induced
Enhancer
Region in lT Cells
A, Left pane/, CT/GH expression
plasmids;
right pane/, GH production.
Each bar represents
the mean 2 SEM of two to eight
experiments
of three transfections
each. To control for transfection
efficiency,
the results were expressed
as percentage
of the
CAMP-induced
activity of pCAMP-1
(which ranged between
1 .Q and 8.6 r-g/ml). None of the constructs
had constitutive
activity.
6, Expression
of CAT by the RSV late terminal
repeat/CAT
fusion gene (RSV-CAT).
CAT activity
was assessed
by measuring
acetylated
‘%-labeled
chloramphenicol
after incubation
with enzyme-containing
cell extracts.
Each bar shows the mean of three
plates. C, Differential
activation
of CRE+CREL/O
(in pCTGH-1)
and CREUO
alone (in pCTGH-4)
at increasing
passage numbers
of lT cells. CAMP-induced.GH
production
was expressed
as in panel A. *, statistically
different
from pCAMP-4
values (P < 0.05).
The proximal sequences of the CT promoter region
contributed
only 20% of the total CAMP-induced
transcriptional
activity. However,
replacing
these
sequences
with the heterologous
thymidine
kinase
promoter completely diminished
CAMP-induced
transcription of CRE and CREL (data not shown). Therefore, we concluded that CAMP-induced
transcriptional
activity of CRE and CREUO is dependent
on other
proximal promoter elements of the CT gene. Further
mapping of the proximal promoter region is required to
identify these elements.
Because the CT CAMP-induced
&-acting
elements
contained octamer sequences that could potentially
function as binding sites for homeo/POU
domain proteins or other octamer-binding
transcription
factors,
we thought that these elements might play a role in
cell-specific
CT transcription.
We therefore examined
the use of the CT CAMP-induced
c&-acting elements
Induction
of Transcription
by CAMP
787
in cell lines that do not transcribe the CT gene. Plasmids pCTGH and pCAMP-1
and other mutant constructs were transfected
into HeLa and CT-negative
undifferentiated
MTC @O-D81 -1) cells (19) (Fig. 3). In
HeLa cells, transcriptional
activity of pCAMP-1
was
primarily constitutive and was not augmented
by addition of the CAMP analog. Mutating CRE in the presence of CREUO (construct pCAMP-2)
or mutating
CREUO in the presence of CRE (construct pCAMP-7)
diminished transcription
by 90%. Similar results were
obtained by deleting the CRE. We concluded
that in
HeLa cells as in TT cells, the transcriptional
activity of
the CRE was dependent
on CREUO. On the other
hand, while CREUO was somewhat active in lT cells
in the absence of CRE, it did not act independently
in
HeLa.
In RO-D81-1
cells, transcription
of the constructs
tested was mainly constitutive activity of the proximal
promoter, whereas the CAMP-induced
elements were
not functional. In summary, the mapping experiments
clearly established
that the CRE, the downstream
CREUO element, and the proximal promoter of the CT
gene were used differently in CT-positive
and CTnegative cells.
A.CONSTRUCTS
Identification
and Characterization
of tramActing Factors that Bound to the CT CRE
To characterize
the factors binding to the CT CRE, we
performed
mobility-shift
assays. Incubating
the CT
CRE probe with lT nuclear extracts resulted in the
formation of two complexes
(Fig. 4). The principal
complex was completely competed by the CRE oligonucleotide,
which contains the wild type consensus
CRE motif (TGACGTCA),
but not by the CREM oligonucleotide,
which contains a mutated version of the
CRE motif (TMCGGCA).
There was no clear competition with oligonucleotides
containing the consensus
sequence for either activator protein 1 (AP-1, 20) or
AP-2 binding sites. On the autoradiogram
of the gel
shown in Fig. 4, there seems to be partial competition of the CRE complex with the AP-1 oligonucleotide, but this competition
was not reproducible.
Thus, we were unable to demonstrate
any definite
binding to the CT CRE of members of the Fos/Jun
protein family, although the latter can bind to a CRE
after cross-heterodimerization
with members of the
CREB/ATF family (21).
Because the CRE sequence was identical to the
CREL motif except for one base mismatch,
we
repeated
the competition
studies with the CREL
B.
HeLa
C. RO-Dal-1
pCAHR-5
pCmH
pRSV-CAT
CAT activity
(unite./
pg protein)
Fig. 3. Analysis
of CAMP-Induced
Enhancer
Activity
in CT-Negative
Cells
CT-GH expression
plasmids
(A) were transfected
individually
into HeLa cells (B) and RO-D81-1
cells (C). Each bar represents
the mean + SEM of a representative
experiment
of three transfections.
The results were expressed
as percentages
of the
CAMP-induced
expression
of pCAMP-1,
which ranged between
7 and 10 rig/ml in HeLa cells and between
4.5 and 7 rig/ml in
RO-D81-1
cells. Underneath
panels B and C are the results of CAT assays after transfection
of RSV-CAT
into HeLa (B) and
RO-D81-1
(C) cells.
MOL END0 . 1995
788
A.Binding
Vol 9 No. 7
Specificity
B.Imrnunoreactivity
Fig. 4. Characterization
of CRE-Binding Proteins in CT-Producing TT Cells
A, An autoradiogram of a gel showing the specific binding pattern of nuclear proteins from TT cells to the CT CRE. The type
and amount (in nanograms) of oligonucleotide competitor are indicated below each lane. The CRE probe used contained
nucleotides -280 to -202. The arrow indicates the position of the specific protein-WE complexes. The partial competition seen
with AP-1 was not reproducible in other experiments. B, An autoradiogram of a gel showing that ATF-1 and CREB were
components of the CRE-specific complexes. The arrows indicate the position of the CRE-specific complexes and the partial
super-shift after incubation with antibodies against CREB and ATF-1 for 2 h at 4 C. PI, Preimmune serum; ATF, antibodies against
ATF-1; CREB, antibodies against CREB; CRE, CRE oligonucleotide (50 ng).
oligonucleotide,
which contained
the CREL motif
(TGACCTCA),
and with the CRELM
oligonucleotide, which contained
a mutated
version of the
CRELmotif(TAACCGCA).
There was no cross-competition with either oligonucleotide.
We concluded
that
the proteins that bound to the CRE element specifically require the TGACGTCA
motif for high-affinity
binding. Moreover, these binding proteins are probably different from those that bound to the CREL
motif, because
there was no cross-competition
between the two.
To test whether CAMP-mediated
transcription
involved up-regulation
of CRE-binding
proteins, we repeated the binding assays with the CRE probe and
extracts from untreated TT cells and TT cells treated
with a CAMP analog for 24 h. Treating the cells with
CAMP had no effect on either the mobility or the abundance of the CRE-binding
proteins (data not shown).
These findings suggest that these factors respond to
CAMP by modifying their transcriptional
activity rather
than their abundance or DNA binding (10, 11).
To identify the factors within the specific CRE complexes, we used antibodies against two members of
the CREB/ATF family, CREB and ATF-1. The antibodies against CREB recognize
homodimers
and heterodimers of CREB, whereas the antibodies
against
ATF-1 recognize only ATF-1 homodimers
(22). We repeated the band-shift assays, incubating lT extracts
with the antibodies
and the CRE probe (Fig. 4). The
ATF-1 antibodies
induced a partial supershift of the
CRE-protein complex, whereas preimmune serum did
not. The super-shift with the ATF-1 antibodies
indicated that ATF-1 was a component of the CRE-protein
complex. The CREB antibodies had two effects on the
CRE-specific
complexes:
a partial supershift and an
inhibition of complex formation. We think that disrup-
tion of the CRVprotein
complexes by these antibodies
occurred because they were raised against a peptide
in the C terminus of CREB, a region that is important
for the DNA-binding
activity of this protein.
These results show that both CREB and ATF-1
bound to CT CRE element in lT cells. However, we
cannot exclude the possibility
that there are other,
unidentified components
of the CRE-protein
complex,
because both antibodies reacted only partially with the
CRE-protein
complexes.
Characterization of the WE-Binding Proteins in
CT-Negative Cells
We next examined whether protein binding to the CRE
explained the differences in CAMP-induced
transcriptional activity in lT cells and the CT-negative cell lines
RO-D81-1 and HeLa. We repeated the electrophoretic
mobility shift assays using the CRE probe and nuclear
extracts from RO-D81-1 and HeLa cells (Fig. 5A). A
DNA-binding
pattern similar to the one seen with TT
cells was observed. There were only minor differences
in the intensities of the complexes
in the extracts,
suggesting
that binding-protein
abundance
did not
affect transcriptional
activity of the CT CRE element in
the different cells.
Further characterization
with the antibodies against
CREB and ATF-1 (Fig. 5B) revealed a qualitative difference between the CRE-binding proteins from HeLa and
TT cells. There was a complete supershift of the CREbinding proteins from HeLa by the ATF-1 antibodies, but
no supershift or disruption of complex formation was
noted with the CREB antibodies. Because the ATF-1
antibodies recognize only the ATF-1 homodimer, we
concluded that only ATF-1 homodimers bound to the CT
CRE in HeLa cells. This difference in CRE-binding pro-
induction
of Transcription
A,
by CAMP
TT
nnn
789
RO-Del
HaLa
123456
769
B.
l-r
1234
no-D41
5678
R.la
9101112
Fig. 5. Characterization
of the CRE-Binding
Proteins
in
CT-Negative
RO-Dal-1
and HeLa Cells
A, Autoradiogram
of a gel comparing
the binding specificity
of nuclear extracts from lT, RO-Dal-i,
and HeLa cells to the
CT CRE. The type of unlabeled
oligonucleotide
competitor
present (+) or absent (-) is indicated
below each lane. The
arrow indicates the position of the specific protein-CRE
complexes. CRE, CRE oligonucleotide
(50 ng); CREM, CREM oligonucleotide (50 ng). 8, Characterization
of the components
of the
CRE complex in CT-positive
and CT-negative
cells. The anows
indicate the position of the CRE-specific
complexes
and the
partial supershii
after incubation
with antibodies
against CREB
and ATF-1.
PI, Preimmune
serum; ATF, antibodies
against
ATF-1; CREB, antibodies
against CREB.
teins from HeLaand TT cells may be one of the reasons
why the combinedtranscriptionalactivity of the CT CRE
and CREUO in HeLa cells was constitutive and not
CAMP induced.We were unableto find any differences
betweenthe CRE-bindingproteinsfrom RO-D81-1 and
lT by using the antibodies against CREB and ATF-1.
Therefore our tests cannot explain the lack of CAMPinduced transcriptionin RO-D81-1 cells.
Characterization
of Protein Binding to the
Downstream
CREUO Element in Tl Cells
To characterize the proteins binding to the CREUO
element, we first considered the complexity of this
element. As shown in Fig. 1, it contains a CREL se-
quence flanked by two octamer motifs. Since these
three sequencesare adjacent, we realized that a mutation in one binding site might affect binding to a
neighboring site. Therefore, binding pattern was first
tested with the OKREUO probe (CT nucleotides
-202 to - 129)which contained all three sites, unmutated. Incubating this probe with nuclearextracts from
cultured lT, RO-D81-1, or HeLa cells did not result in
the formation of a specific complex (data not shown).
This lack of binding strongly suggested that the
TGACCTCA motif within the CREUO region, although
similar to a CRE, was not a CREB/ATF binding site.
There are two possible explanations for the lack of
detection of specific binding with the lT nuclear extracts. First, the proteins binding to CT nucleotides
-202 to -129 may have been present in low abundance in cultured lT cells and therefore could not be
detected in crude extract. Second, the binding affinity
of the transcription factors interacting with this region
may be too low to be detected by the electrophoretic
mobility shift assay. Because there are recorded differences in the pattern of CTKGRP gene expression
in MTC tumors carried in animalmodelsand in culture
(23), we considered the possibility that transcription
factors that regulate expression of the CT gene will be
represented differently or perhaps more abundantly in
tumors generated from lT cells than in culture. We
tested extracts from 5- to 7-week-old tumors induced
in nude mice by SC injection with lT cells. Figure 6
(lanes 1-5) shows that proteins from tumor extract
formed one slowly migratingcomplex (complex 1) and
three fast migrating complexes (complexes 2-4) with
the O/CREL/O probe. Competition with the CREL oligonucleotide (which contains the CREL and octamer
motifs) abolished complex 1 formation but had a
minimaleffect on formation of complexes 2-4. Somewhat surprisingly, a mutated CREL oligonucleotide
(CRELM)alsocompeted with complex 1. There was no
competition of complex 1 with the AP-1 (lane6) or the
AP-2 oligonucleotides(data not shown).
These results suggested that the proteins within
complex 1 were probably not the Vans-activators that
bound to CREL, as the proteins within this complex
did not specifically require the TGACCTCA motif to
bind. Because the oligonucleotide competitors ,CREL
and CRELM both contained octamer motifs, we tested
whether the binding site of proteins in complex 1 overlapped the upstream octamer motif. We repeated the
binding assays using the CREUO probe, which contained a truncated version of the upstream octamer.
Fig 6 (lanes6-8) showsthat this truncation prevented
the formation of complex 1 but did not have any effect
on complexes 2-4. We therefore concluded that the
binding activity of the proteins within complex 1 was
dependent on the upstream octamer sequences. Because the upstream octamer was not transcriptionally
active in the neuroendocrine lT cells (compare
pCAMP-2 to pCMAP-4 in Fig. 2), we do not know the
function of the DNA-binding proteins within complex
1. However, since other sites for expressionof the CT
MOL
790
END0
.I995
Vol 9 No. 7
123456
I
t
I
I
Complex 1
Complexes
2-4
Competitor
(n!7)
Fig. 6. Analysis
of OKREUO-Binding
Proteins
in Nuclear Extracts
of TT Tumors
The probes
used are illustrated
above
the corresponding
lanes. The arrows
indicate
the protein
complexes
detected.
Lanes l-6, Incubation
with the OKREUO
probe; lanes 7-9, incubation
with the CREUO
probe; lanes 1 O-12, incubation
with
the CREUOm
probe;
lanes 13-15,
incubation
with the CRELm/O
probe.
The type and amount
(nanograms/reaction)
of
oligonucleotide
competitor
used are indicated
below each lane.
gene are in the central and peripheral nervous systems
(1) it is possible that the upstream octamer and its
transacting factors are active in these other cell types.
We next investigated
the identity of the proteins
within complexes 2-4. For this, we used the CRELm/O
and the CREUOm probes, which were identical to
CREUO except for two-point
mutations in the CREL
site or three-point
mutations in the downstream
octamer motif, respectively. Either one of these mutations diminished transcriptional
activity of the CREUO
element (Fig. 2). Examination
of binding to these
probes revealed that the formation of complexes 24
was abolished
by the three-point
mutations
in the
octamer sequence (Fig. 6, lanes 1 O-l 2) but not by the
two-point
mutations in the CREL site (Fig. 6, lanes
13-15). We therefore concluded
that complexes 2-4
probably contained octamer-binding
proteins.
By mutating either the CREL or the octamer motif
we created binding sites for unidentified proteins (Fig.
6, lanes 10-15). Especially noteworthy was a “CRELspecific” complex
formed by mutating the downstream octamer motif (Fig. 6, lanes 10-12). Our interpretation
of these results is that mutation of the
octamer motif facilitated
the binding of low-affinity
binding proteins to the CREL and the octamer motifs.
We do not know whether these proteins are relevant to
regulation of CT gene transcription.
From these experiments
we concluded
that both
transcriptional
activity of, and protein binding activity
to, the CREUO motif were complex. Further studies
are necessary to clarify the identity of the CREL- and
octamer-binding
proteins.
DISCUSSION
Our previous reports and those of others have described in great detail a constitutive neuroendocrinespecific enhancer of CT gene transcription
in humans
and rats (4-6, 24). In addition, a CAMP-induced
enhancer was identified by other investigators,
but its
role in the regulation of neuroendocrine-specific
CT
expression
was not determined
(3, 8). In the study
reported here, we showed that CAMP induction of CT
gene transcription
was also restricted to a CT-producing neuroendocrine
cell line (‘IT). This transcriptional
activity was attributed to a typical CRE and to a novel
hybrid CREUO element that acted synergistically.
In
the CT gene, neither the CRE nor the CREUO is a true
enhancer element, because their combined
activity
depended
either on the proximal CT promoter sequences or the presence of a distal upstream neuroendocrine-specific
enhancer (5). A synergism between a
CRE and other elements may be a common mechanism for maximizing
CAMP-induced
transcription
of
neuroendocrine
genes, e.g. the interaction
between
CRE and an AP-2 site in the promoter of the proenkephalin gene and the synergism between the two
CREs and a tissue-specific
element in the proximal
promoter region of the gene encoding the a-subunit of
the glycoprotein
hormone (25-27).
Although the CT CRE is a typical binding site for the
ubiquitous
CREB and ATF-related
transcription
factors (1 l), its activity is clearly cell specific. One possible mechanism for this cell-specific
regulation of CRE
activity may be differential interaction of the CREB/
ATF-related
factors with the CRE. For example, in
Induction
of Transcription
by CAMP
HeLa cells, we found ATF-l-binding
activity but not
CREB-binding
activity. The preferential
binding
of
ATF-1 coincided with constitutive transcriptional
activity of the CRE and CREUO motifs in these cells. A
similar constitutive
transcriptional
activity of CREs
from other genes in HeLa cells has been attributed to
the predominant
expression of ATF-1 and to the structure of ATF-1, which is devoid of the glutamine-rich
N-terminal domain and the a-region that are essential
for CAMP-induced
activity of CREB (28, 29).
Differential interaction of CREB/ATF with the CRE
cannot account for the lack of transcriptional
activity of
this element in the undifferentiated
CT-negative MTC
RO-D81-1 cells because we found identical binding patterns for CREB and ATF-1 in extracts of these cells and
of CT-positive lT cells. One explanation is provided by
an analogous cell culture model, the teratocarcinoma
cell
line F9 (30). In the undifferentiated
state, the transcriptional activity of CRE is low in F9 cells despite the presence of CREB and ATF. However, when F9 cells are
induced to differentiate by treatment with retinoic acid,
the CAMP-induced
activity of CRE is restored. The low
activity of the CRE in undifferentiated F9 cells is correlated with low levels of the regulatory subunit of protein
kinase A and high levels of CBPl 00, an inhibitor of CREB
action (30-32). The increase in CRE activity in the differentiated F9 cells was correlated with elevated levels of
the regulatory subunit of protein kinase A and reduced
levels of CBPl 00. Another explanation may be related to
the dependence
of CT CRE on the proximal promoter
and the CREUO motif. If a CREUO-specific
or promoterspecific transacting factor is absent from the CT-negative MTC line, it may also lead to diminished constitutive
and CAMP-induced
transcription of the CRE. These are
two examples of auxiliary mechanisms that regulate the
transcriptional
activity of CREB/ATF family members,
and they may apply to the differential activation of the CT
CRE in the cell lines tested in the study reported here.
The transcriptionally
active component
in CT nucleotides -175 to -129 was the sequence TGACCTCAATGCAAAT.
The activity of this element was also
cell specific: it was partially active in the absence of
CRE in lT cells, was completely dependent on CRE in
HeLa cells, and was inactive in RO-D81-1
cells. It is
possible that this sequence functions as a binding site
for a single cell-specific
transcription
factor that has
not yet been identified. However, because it contains
two potential &-acting
elements, one a CREL motif
(TGACCTCA),
and the other an octamer motif (ATGCAAAT), we favor the hypotliesis
that its transcriptional activity depends on two transcription
factors, a
CREL-binding
protein and an octamer-binding
protein. We base this hypothesis on transcription
assays
showing that mutations in either the CREL sequence
or the octamer
sequence
diminished
the CAMPinduced transcription
attributable
to that region. Further support for this hypothesis was derived from the
DNA-protein interaction assays: although we were unable to clearly demonstrate
binding to the TGACCTCA
sequence, we did detect strong binding to the ATG-
791
CARAT sequence. This latter binding activity was not
affected by mutation in the CREL sequence but was
abolished by mutating the octamer motif.
The strong homology of the octamer in the CREUO
element to binding sites for POU/homeodomain
proteins and the detection of octamer-binding
activity
suggest that CREUO is a &-acting
motif for a related
factor and may explain both its CAMP-induced
activity
and its cell-specific
behavior. Several members of the
POU/homeo
family are regulated by CAMP, and most
are involved in regulation of cell-specific transcription
and differentiation
(12-l 6). Recently, an octamerbinding protein was reported to be involved in the
regulation of the distal neuroendocrine-specific
constitutive enhancer of the rat CT gene (24). It will be
interesting to determine whether the upstream and
downstream
octamers are binding sites for the same
transcription
factors.
We were unable to identify the CREL-binding
protein,
but our competition studies and protein interaction assays with the CRE motif clearly indicated that the CREL
motif was not a binding site for members of the CREB/
ATF family, although it differs from a typical CRE by only
one base. Further analysis of the complement of TGACCTCA revealed that it was homologous to a half-binding
site for proteins of the steroid/thyroid/retinoic
acid receptor family (33). Although response elements for members
of this family usually contain two or more repeats of the
AGGTCA motif, two members of this family act through
a single motif. One is the nerve growth factor-induced
binding protein (NGFI-B) (34, 35). The other is retinoid X
receptor 8, which binds to the TGAGGTCA
sequence
within region II of the major histocompatibility
complex
class I transcriptional
regulatory element (36-38). Future
experiments will determine whether or not the CREL is a
binding site for a protein related to the nuclear receptor
family.
MATERIALS
Plasmid
AND METHODS
Construction
We previously
described
the restriction
endonuclease-mediated deletions
used to create pCTGH
and pCTGH-1,
which
contain CT nucleotides
-129 to +90 and -1460
to +90, respectively,
attached to the GH gene (4, 5). Plasmid pCAMP-1
(CT nucleotides
-829 to +90) was created by double digestion
of pCTGH-1
with HindIll and BspMl and self-ligation.
Plasmid
pCAMP-2
was made by insertion of a SalI linker (5’-GGTCGACC-3’)
at the Aatll site in pCTGH-1,
deletion of nucleotides
- 1460 to -829 by double digestion with BspMl and HindIll, and
self-ligation.
Plasmid pCAMP-3
was also made by inserting a
Sal linker at the A&II site in pCTGH-1,
but this time the .%/I/
HindIll fragment
was deleted and the plasmid self-ligated.
The
polymerase
chain reaction
technique
was used to create
pCAMP-4.
A fragment
spanning
CT nucleotides
- 175 to +90
was amplified by using pCTGH-1
as template and primers from
nucleotides
- 175 to - 155 of the CT gene and from nucleotides
+50 to +30 of the GH gene. The polymerase
chain reaction
product was digested with BarnHI, and the resulting DNA fragment (-175
to -129)
was inserted
into compatible
(blunt/
BarnHI) sites in pCTGH.
MOL
792
END0
. 1995
Vol 9 No. 7
Plasmids
pCAMP-5
and pCAMP-6
were produced
like
pCAMP-4,
except
that the CT primer had either two point
mutations
in the CREL site (T&XCGCA
instead of TGACCTCA) to create pCAMP-5
or three point mutations
in the
downstream
octamer
sequence
(ATGCAB
instead
of
ATGCAAAT)
to create pCAMP-6.
The in-context
mutation
in pCAMP-7
was produced
with a
site-directed
mutagenesis
system (Promega,
Madison,
WI),
after CT nucleotides
-829 to -129 were subcloned
into the
plasmid pALTER. The mutation
was inserted by following
the
manufacturer
instructions
by using a 20 mer primer containing the mutated
CREL motif (TAACCGCA).
The sequence
of each clone was confirmed
by restriction
mapping
and DNA sequencing.
Cell Culture
and Transfection
The human MTC cell lines, the well differentiated
line TT (18) and
undifferentiated
line RO-D81-1
(19) were maintained
in RPM1
1640 medium
supplemented
with 10% fetal bovine
serum.
HeLa cells were grown in Dulbecco’s
modified Eagle’s medium
supplemented
with 10% fetal bovine serum. Twenty four hours
before transfection,
the cells were plated in 35mm
dishes at a
density of 4 x lo5 cells per dish for TT, 1 x l@ cells per dish
for RO-D81-1,
and 5 X 104 cells per dish for HeLa. The cells
were then transfected
with 8 w plasmid DNA per dish by using
the diethylaminoethyl
(DEAE)-dextran
method (4, 5) and incubating at 37 C for 30 min for TT and RO-D81-1
cells and 4 h for
HeLa cells. A CAMP analog, 8(4-chlorophenylthio)-CAMP
(1 mM),
was added 48 h (to Hela cells) or 72 h (to lT and RO-D81-1
cells) after transfection.
Medium
samples were collected
24 h
after addition of the CAMP analog. GH production
by the reporter gene was measured
by a two-site
IRMA assay as described by the manufacturer
(Nichols Institute, San Juan Capistrano, CA) The limit of detection
of GH by this assay was
0.3-0.5
rig/ml. No GH was detectable
in medium alone or in
medium from mock-transfected
cells.
Chloramphenicol
acetyltransferase
(CAT)
activity
was
measured
in cell extracts
3 days after transfection
with the
Rous sarcoma
virus (RSV)-CAT
plasmid.
Monolayers
of lT,
HeLa, and RO-D81-1
cells were washed in PBS, scraped into
cold 0.25 M Tris-HCI,
pH 7.5, lysed by three freeze-thaw
cycles, and centrifuged.
Aliquots
of supernatant
containing
50 mg (Tl) or 10 mg (HeLa and RO-D81-1)
protein
were
incubated
with 1 mM acetylcoenzyme
A and [14C]chloramphenicol
for 3 h at 37 C. Substrate
and products
were resolved by TLC as described
by Gorman
et al. (39) and autoradiographed
for 2448
h. The levels of enzyme
activity were
determined
by measuring
the amount of [‘4C]chloramphenicol products
after excision
from the TLC plates. The excised
substrate
and products
were counted in 5 ml scintillation
fluid
in a Beckman
scintillation
counter.
One arbitrary
unit of CAT
activity was 1% conversion
of substrate
to product.
Electrophoretic
Mobility
the
upstream
CRE
motif (TGACGTCA);
the downstream
which contained
CT nucleotides
-202 to
-129
and included
the CREL site and its flanking
octamer
sequences:
the CREUO
probe, which contained
CT nucleotides -175 to -129 and included
the CREL site, the downstream
octamer,
and a truncated
version
of the upstream
octamer;
the CRELm/O
and the CREUOm
probes,
which
were ho.mologous
to the CREUO
probe except
for either
two-point
mutations
in the CREL site or three-point
mutations
in the downstream
octamer
sequence,
respectively.
The oligonucleotides
used for competition
were the
CRE,
5’-GAAATTCACCATGACGTCAAACTGCCCTCA-3’;
the mutated
CRE (CREM),
5’-GAAATTCACCATMCGGCAAACTGCCCTCA-3’;
CREL, 5’-TTCCATCAATGACCTCAATGCAAATAC-3’;
the mutated
CREL
(CRELM),
5’TTCCATCAAT&ICCGCAATGCAAATAC-3’;
the
AP-1
consensus
oligonucleotide
5’-CTAGTGATGACTCAGCCGGATC-3’;
and the AP-2 consensus
oligonucleotide,
5’GATCGAACTGACCGCCCGCGGCCCGT-3’.
For the immunoreactivity
studies with antibodies
against
CREB and ATF-1, electrophoretic
mobility
shift assays were
performed
as described
above, except that the reaction
mixture was preincubated
with 3 ~1 crude antiserum
for 2 h at 4
C before the probe was added. The preimmune
sera and the
antibodies
against
CREB and ATF-1 were a generous
gift
from Drs. Helen C. Hurst and Kevin A. W. Lee (22).
O/CREL/O probe,
Statistical
Analysis
GH production
in untreated
compared
with the analysis
comparison
tests
wherever
considered
significant.
and CAMP-treated
cells was
of variance
and Duncan multiple
appropriate.
P < 0.05 was
Acknowledgments
We wish to thank Drs. Helen C. Hurst and Kevin A. W. Lee of
the Imperial Cancer Research
Fund, London,
England, for the
CREB and ATF-1 antibodies,
and Dr. Guy Juillard,
of the
University
of California,
Los Angeles, for the medullary
thyroid
carcinoma
cells, RO-D81-1
(passage
22).
Received
June 10, 1994. Revision
received
March
29,
1995. Accepted
April 12, 1995.
Address
requests
for reprints to: Sara Peleg, Ph.D., Section of Endocrinology
(Box 15), Department
of Medical Specialties,
The University
of Texas M. D. Anderson
Cancer
Center,
1515 Holcombe
Boulevard,
Houston,
Texas 77030.
This work
was supported
by NIH Grant
DK-38146
(to R.F.G.).
Shift Assays
Nuclear extracts
were prepared
as described
by Dignam et
al. (40) from five sources:
untreated
lT cells, lT cells treated
with CAMP, RO-D81-1
cells, HeLa cells, and tumors induced
in nude mice by SC injection
of lT cells.
Each binding reaction
was performed
by incubating
5-10
@ protein extract
with 0.5 ng 32P-labeled
DNA probe, 2 pg
poly [d(l-C)] as nonspecific
competitor
DNA, 12% glycerol,
12
mM HEPES-NaOH,
pH 7.9,4 mM Tris-HCI,
pH 7.9.60 mM KCI,
1 mM EDTA, and 1 mM dithiothreitol
at room temperature
for
30 min. The complexes
were resolved
by electrophoresis
through
4% polyacrylamide
gels in Tris-glycine
buffer at 4 C.
For competition
experiments
the same conditions
were used,
except that specific
oligonucleotides
and nonspecific
DNA
were added before the probe to the binding reactions.
The probes
used in these assays were the CRE probe,
which contained
CT nucleotides
-280 to -202 and included
Rosenfeld
MG, Amara SG, Evans RM 1984 Alternative
RNA
processing:
determining
neuronal
phenotype.
Science 225:1315-l
320
Edbrooke
MR, Parker D, McVey JH, Riley JH, Sorenson
GD, Pettenaill
OS. Craig RK 1986 Expression
of the
human calczonin/CGRP
gene in lung and thyroid carcinoma. EMBO J 4:715-724
de Bustros A, Lee RY, Compton
D, Tsong TY, Baylin SB,
Nelkin BD 1990 Differential
utilization
of calcitonin
gene
regulatory
DNA sequences
in cultured
lines of medullar-y
thyroid
carcinoma
and small-cell
lung carcinoma.
Mol
Cell Biol 10:1773-l
778
Peleg S, Abruzzese
RV, Cote GJ, Gage1 RF 1990 Transcription
of the human calcitonin
gene is mediated
by a
Induction
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
of Transcription
by CAMP
C ceil-specific
enhancer
containing
E-box-like
elements.
Mol Endocrinol
11 :1750-l
757
Peleg S, Abruzzese
RV, Cooper
CW, Gage1 RF 1993
Down-regulation
of calcitonin
gene transcription
by
vitamin
D requires
two widely
separated
enhancer
sequences.
Mol Endocrinol
7:999-l
008
Ball DW, Compton
D, Nelkin BD, Baylin SB, de Bustros A
1992 Human calcitonin gene regulation
by helix-loop-helix
recognition
sequences.
Nucleic Acids Res 20:117-l
23
Cote GJ, Abruzzese
RV, Lips CJM, Gage1 RF 1990 Transfection of calcitonin
gene regulatory
elements
into a cell
culture model of the C cell. J Bone Miner Res 5:165-l
71
de Bustros A, Ball DW, Peters R, Compton
D, Nelkin BD
1992 Regulation
of human calcitonin
gene transcription
by cyclic AMP.
Biochem
Biophys
Res Commun
189:
1157-1164
Montminy
MR, Sevarino
KA, Wagner
JA, Mandel
G,
Goodman
RH 1986 Identification
of a cvclic-AMP-responsive
element within the rat somatostatin
gene. Proc
Nat1 Acad Sci USA 83:6682-6686
Dean DC, Blakeley
MS, Newby
RF, Ghazal
P, Hennighausen
L, Bourgeois
S 1989 Forskolin
inducibility
and
tissue-specific
expression
of the fibronectin
promoter.
Mol Cell Biol 9:1498-1506
Habener
JF 1990 Cyclic AMP response
element binding
proteins:
A Cornucopia
of transcription
factors.
Mol Endocrinol
4:1087-l
094
Bodner
M, Castrillo
J-L, Theill LE, Deerinck
T, Ellisman
M, Karin M 1988 The pituitary-specific
transcription
factor GHF-1
is a homeobox-containina
- protein. Cell 55:
505-518
lngraham
HA, Chen R, Mangalam
HJ, Elsholtz HP, Flynn
SE, Lin CR, Simmons DM, Swanson L, Rosenfeld MG 1988
A tissue-specific
transcription
factor. containing
a homeodomain specifies a pituitary phenotype.
Cell 55:519-529
Kapiloff MS, Farkash Y, Wegner
M, Rosenfeld
MG 1991
Variable
effects of phosphorylation
of Pit-l dictated
by
the DNA response
elements.
Science 253:786-789
Faisst S, Meyer
S 1991 Compilation
of vettebrateencoded
transcription
factors.
Nucleic
Acids Res. 20:
3-26
Segil N, Roberts
SB, Heintz N 1991 Mitotic phosphorylation of the Ott-1
homeodomain
and regulation
of Ott-1
DNA binding activity.
Science 254:1814-l
816
lmagawa
M, Chiu R, Karin M 1987 Transcription
factor
AP-2 mediates
induction
by two different
signal-transduction pathways:
protein C and CAMP. Cell 51:251-260
Leong SS, Horoszewicz
JS, Shimaoka
K, Friedman
M,
Kawinski
E, Song MJ, Zeigel R, Chu TM, Baylin SB,
Mirand
EA 1981 A new cell line for study of human
medullaty
thyroid
carcinoma.
In: Andreoli
M, Monaco
F,
Robbins
F (eds) Advances
in Thyroid
Neoplasia.
Field
Educational
Italia, Rome, pp 95-l 08
Abemayor
E, Sidell N, Juillard G 1989 Human medullary
thyroid
carcinoma.
Initial characterization
and in vitro
differentiation
of two new cell lines. Arch Otolaryngol
Head Neck Sura 115:478-483
Lee W, Mitchei
P, Tjian R 1987 Purified transcription
factor
AP-1
interacts
with
TPA-inducible
enhancer
elements.
Cell 49:741-752
Hai T, Curran
T 1991 Cross-family
dimerization
of
transcription
factors
Fos/Jun
and ATF/CREB
alters
DNA binding
specificity.
Proc Natl Acad Sci USA 88:
3720-3724
Hurst HC, Masson
N, Jones NC, Lee KAW 1990 The
cellular
transcription
factor
CREB
corresponds
to
activating
transcription
factor
47 (ATF-47)
and forms
complexes
with a group
of polypeptides
related
to
ATF-43.
Mol Cell Biol 10:6192-6203
Rosenfeld
MG, Lin RC, Amara SG, Stolorsky
L, Roos BA,
Ong ES, Evans EM 1982 Calcitonin
mRNA polymorphism:
peptide switching
associated
with alternative
RNA splicing
events. Proc Natl Acad Sci USA 79:1717-l
721
793
24. Tverberg
LA, Russo AF 1993 Regulation
of the calcitonin/
calcitonin
gene-related
peptide
gene by cell-specific
synergy
between
helix-loop-helix
and octamer-binding
transcription
factors.
J Biol Chem 268:15965-l
5973
25. Deutsch PI, Jameson
JL, Habener
JF 1987 Cyclic AMP
responsiveness
of human
gonadotropin-alpha
gene
transcription
is directed
by a repeated
18-base
pair
enhancer:
alpha-promoter
receptivity
to the enhancer
confers
cell-preferential
expression.
J Biol Chem 262:
12169-12174
26. Andersen
B, Kennedy
GC, Hamerlink
DL, Bokar JA,
Bohinski
R, Nilson JH 1990 Amolification
of the transcriptional
signal
mediated
by the tandem
CAMP response
elements
of the glycoprotein
hormone
alphasubunit
gene
occurs
through
several
distinct
mechanisms.
Mol Endocrinol
4:573-582
27. Comb M, Mermod
N, Hyman SL, Pearlberg
J, Ross ME,
Goodman
HM 1988 Proteins bound at adjacent DNA elements act synergistically
to regulate human proenkephalin
CAMP inducible transcription.
EMBO J 7:3793-3805
28. Hurst HC, Totty NF, Jones NC 1991 Identification
and
functional
characterization
of the cellular activating
transcription
factor 43 (ATF-43)
protein.
Nucleic Acids Res
19:4601-4609
29. Liu F, Thompson
MA, Wagner
S, Greenberg
ME, Green
MR 1993 Activating
transcription
factor-l
can mediate
Ca2’and CAMP-inducible
transcriptional
activation.
J Biol Chem 268:6714-6720
30. Masson
N, Ellis M, Goodbourn
S, Lee KAW 1992 Cyclic
AMP response
element-binding
protein and the catalytic
subunit of protein kinase A are present
in F9 embryonal
carcinoma
cells but are unable to activate
the somatostatin promoter.
Mol Cell Biol 12:1096-l
106
31. Plet A, Evain D 1982 Effect of retinoic acid treatment
of
F9 embryonal
carcinoma
cells on the activity and distribution of cyclic AMP-dependent
protein
kinase. J Biol
Chem 257:889-893
32. Masson
N, Hurst HC, Lee KAW 1993 identification
of
proteins
that interact with CREB during differentiation
of
F9 embryonal
carcinoma
cells. Nucleic
Acids Res 21:
1163-1169
33. Evans RM 1988 The steroid and thyroid hormone
receptor superfamily.
Science
240:889-895
34. Wilson TE, Fahrner
TJ, Johnston
M, Milbrandt
J 1991
Identification
of the DNA binding
site for NGFI-B
by
genetic selection
in yeast. Science 252:1296-l
300
35. Wilson TE, Mouw AR, Weaver CA, Milbrandt
J, Parker KL
1993 The orphan nuclear receptor
NGFI-B regulates
expression
of the gene encoding
steroid 21-hydroxylase.
Mol Cell Biol 13:861-868
36. Hamada
K, Gleason SL, Levi B-Z, Hirschfeld
S, Appella
E 1989 H-2RIIBP.
a member
of the nuclear
hormone
receptor
superfamily
that binds to both the regulatory
element of major histocompatibility
class I genes and the
estrogen
response
element.
Proc Natl Acad Sci USA
86:8289-8293
37. Marks MS, Hallenbeck
PL, Nagata T, Segars JH, Appella
E, Nikodem
VM. Ozato
K 1992 H-PRIIBP
(RXR 8) heterodimerization
provides
a mechanism
for combinatorial
diversity
in the regulation
of retinoic
acid and thyroid
hormone
responsive
genes. EMBO J 11:1419-l
435
38. Nagata T, Segars JH, Levi B-Z, Ozato K 1992 Retinoic
acid-dependent
transactivation
of major histocompatibiliky complex
class I promoters
by the nuclear hormone
receptor
H-2RIIBP
in undifferentiated
embryonal
carcinoma cells. Proc Natl Acad Sci USA 89:937-941
39. Gorman
CM, Moffat
LF, Howard
BH 1982 Recombinant
genomes
which express
chloramphenicol
acetyltransferase in mammalian
cells. Mol Cell Biol 2:1044-1051
40. Dignam
DJ, Lebowitz
RM, Roeder
RG 1983 Accurate
transcription
initiation
by RNA polymerase
II in soluble
extract
from isolated
mammalian
nuclei. Nucleic Acids
Res 11 :1475-l 489