Hepatocyte-Nuclear
Factor 3p
Gene Transcripts
Generate Protein
Isoforms with Different
Transactivation
Properties
on the
Glucagon
Gene
Jacques
Philippe
Department of Genetics and Microbiology
Medicine Centre M6dical Universitaire
1211 Geneva 4, Switzerland
and
Hepatocyte-nuclear
factor 36 (HNF96),
a member
of the HNF-3 gene family, is expressed
in glucagon-producing
islet cells and represses
glucagon
gene expression.
We show here that at least three
different HNF-36 transcripts
that encode HNF-36
protein variants are present in glucagon-producing
cells, HNF-BP,, HNF-3&., and HNF-3&.
Compared
with the HNF-36, cDNA, HNF-36, cDNA lacks sequences of exon 1 while exons 1 and 4 are absent
from the HNF-36, cDNA. The deduced
amino-acid
(aa) sequence
of HNF-36,
and HNF-36,
proteins
differs from HNF-36, by a 6-aa amino-terminal
extension and by the absence
of the first 30 aa, respectively.
HNF-BP,, HNF-3&,
and HNF-3&
bind
to the major enhancer
of the rat glucagon
gene G2
with similar affinity. By contrast to HNF-36,, which
represses
glucagon
gene expression
when overexpressed
in the glucagon-producing
cell line
InRlG9, HNF-36, and HNF-36, do not affect transcriptional
activity. Furthermore,
cotransfection
of
HNF-3P,
or HNF-36,
along with HNF-3&
decreases
the negative
effects
of HNF-36,.
We
conclude
that glucagon
gene expression
may be
regulated
by the relative abundance
of the three
different
HNF-3P variants
in cu-cells. (Molecular
Endocrinology
9: 366-374, 1995)
INTRODUCTION
ulation of the glucagon gene may thus implicate
The glucagon
gene is expressed
in the pancreatic
islets, the intestine, and the brain (1). Pancreatic-specific expression is conferred by 300 base pairs (bp) of
the 5’-flanking region of the rat glucagon gene, which
contains three DNA elements, the upstream promoter
element G, and two enhancers, G, and G,. G, acts as
a potent activator of glucagon gene expression and is
only active in islet cells (2, 3).
0888-8809/95/$3.00/O
Molecular
Endocrmology
Copyright
0 1995 by The Endocme
We recently identified two major protein complexes
binding to G,; whereas one complex is islet-specific
and appears to act positively on glucagon gene transcription, the second complex contains hepatocytenuclear factor-36
(HNF-3P), a DNA-binding
protein
found in abundance in the liver (4). HNF-36 belongs to
a gene family characterized
by a winged-helix
(WH)
DNA-binding
domain and is found predominantly
in
tissues derived from the endoderm (5, 6). Overexpression of HNF-30 in glucagon-producing
cells results in
a repression of glucagon transcription
(4).
Since different mRNAs were identified by Northern
analysis using an HNF-3 P-specific probe in glucagonproducing
cells, we searched for HNF36 cDNA variants. We report here the cloning of two additional
HNF36
cDNA clones, HNF-36, and HNF-36,. The
putative proteins they encode differ from HNF-36 (referred to now as HNF-36,) at the amino-terminal
end.
HNF-36,,
the most abundant form has a six-amino
acid (aa) amino-terminal
extension whereas HNF-36,
lacks the first 30 aa. Both HNF-36, and HNF-36, bind
G, with a similar affinity compared
with HNF-36,.
However, by contrast to HNF-Sp,, they do not modify
glucagon gene transcriptional
activity. We hypothesize
that the lack of effect of HNF-3/3, and HNF-36, on
transcription
results from differences in their aminoterminal ends compared
with HNF-3&
The aminoterminal end of HNF-36, is a transactivation
domain
(7) and its modification
may impair its function. Reg-
Society
multiple
element.
transcription
factors
RESULTS
AND DISCUSSION
binding
the
G,
We previously reported that two different mRNAs of 2
and 2.2 kilobases (kb) were detected in islet glucagonproducing cells with an HNF-S/3-specific
probe (4). To
further investigate whether multiple mRNAs encoding
different HNF-36 protein isoforms were present in
Regulation
of Glucagon
Gene
Expression
369
glucagon-producing
cells, we screened
a mouse
aTC1 cDNA library with an HNF-36 cDNA probe. We
obtained
14 HNF-38 clones; IO of these contained
sequences
limited to exon 6 (Fig. 1). One cDNA
corresponded
to the originally
described
HNF-36
sequence,
HNF-38,
(5, 8, 9), and three additional
cDNAs contained
sequences
at the 5’-end potentially encoding two different HNF-38 isoforms, HNF3& and HNF-36,.
The major differences between HNF-Sp,, HNF-3&,
and HNF-36, cDNAs are the absence of exon 1 in
p2 and of exons 1 and 4 in &; in addition, p2 and p3
cDNAs contain exons 3 and 5, respectively
(Fig. 1,
A and B). These transcripts
result from alternative
splicing mechanisms of the HNF-36 gene, which contains 6 exons (10, 11). Interestingly,
an in-frame stop
codon is present in the V-untranslated
sequences of
both HNF-36, and p2 cDNAs (Fig. 1A). The HNF-38,
cDNA contains a single large open reading frame that
would encode a protein of 465 aa. The initiator methi-
Fig. 1. Structure
and Sequence
of the
HNF-36
cDNA Variants
A, Y-Ends
of the HNF-38,,
&, and
&, cDNAs and the predicted
aa sequence of the encoded
proteins.
The
initiator ATGs are in bold letters; it is
preceded
by an in-frame
termination
codon
in the HNF-36,
and p2 cDNA.
The sequence
underlined
represents
the Y-primers
(a, b, and c) used to
amplify the respective
cDNAs. The arrow indicates
the boundary
of exon 6.
B, Structure
of the HNF-3 gene (16-l 8)
and of the HNF-36
variant cDNAs.
Exons (E) of the HNF-3
gene are represented by boxes and introns by lines.
Each cDNA is schematized
by the juxtaposition
of the exons it contains.
The
expected
size of each HNF-36
protein
isoform is indicated
on the right.
onine conforms well to the consensus for eukaryotic
initiation (12). Comparison
of the aa sequence with
that of HNF-38, indicates that both proteins only differ
by the presence of six additional
aa at the aminoterminal end of HNF-36,. The HNF-38, cDNA contains
a shorter open reading frame of 429 aa which lacks the
30 first amino-terminal
aa of HNF-36,; except for the
absence of these 30 aa, the coding sequences
of
HNF-36, are identical to that of HNF-38,. An additional potential ATG is present in the rat HNF-36,
cDNA 234 bp upstream of the initiator ATG (11). It is,
however, not conserved in the mouse cDNA.
To examine whether the two different mRNAs observed on Northern analysis correspond
to the cloned
cDNAs (Fig. 2, lane l), we used DNA fragments specific for each cDNA for hybridization.
Whereas probes
specific for HNF-36, and & hybridized to the 2.2 kb
mRNA (Fig. 2, lane 2 and 3, respectively),
no signal
was seen with the HNF-38,
probe, indicating
that
mRNA encoding HNF-38, is less abundant than those
fiNF30
B
El
E3 Et
E2
AT6
GENE
E5
cDNAs
E6
protefns
HNF 381
459 m
h
El
E4
E6
E3 E4
E6
AT6
i
HNF3E2
465 m
HW363
429 m
MOL
370
END0
. 1995
Fig. 2. Northern
Blot Analysis
of HNF-36
mRNAs
Total RNA extracted
from the glucagon-producing
cell line
RIN 56 A was hybridized
with 32P-labeled
rat HNF-36,
cDNA
probe (lane l), exon 1 of the rat HNF-36
gene (16) (lane 2), or
the ?-untranslated
sequence
of the mouse HNF36,
cDNA
along with the rat glucagon
cDNA (lane 3). -I
indicates
the
2.0 kb and 2.2 bands and -1-indicates
glucagon
mRNA.
coding for HNF-36, and p2 (data not shown). No hybridization to the relatively abundant 2.0-kb message
was detected. The same results were obtained from
Northern analyses with total RNA extracted from the
three glucagon producing
cell lines RIN 56A (Fig. 2),
OirCl , and InRl G9 (data not shown). To investigate
whether an additional variant HNFS cDNA that was
not represented
in our 14 original clones was present
in (uTC1, RIN 56 A, or lnRlG9 cells, we performed
reverse transcription
(RT) from cytoplasmic
RNA extracted from these cell lines and amplified by polymerase chain reaction (PCR) most of exon 6 se-
Fig. 3. RNase Protection
Analysis
of
HNF-36
mRNAs
A, A 590-bp fragment
containing
the
Y-end
(534 bp) of the rat HNF-36,
cDNA
and 56 bp of unrelated
sequences
was transcribed
by the T7
polymerase
into a ‘“P-labeled
riboprobe that was used for RNase protection analysis with 30 pg RIN 56 A cytoplasmic
RNA. M indicates
markers
whose sizes are indicated
on the left:
arrows point to the expected
protected
fragments
of 534 bases (HNF-Sp,),
243
bases (HNF-36,)
and 173 bases (HNF363). B, Schematic
representation
of
the RNA probe used in the RNase protection
assays
and of the respective
expected
fragments
from HNF-36,,
P2,
and p3 mRNAs.
Vol 9 No. 3
quences. A single 1 .l kb band was obtained from each
RT-PCR, indicating that no additional
cDNA results
from exon 6 splice variants (data not shown). In addition, when we looked for the presence of the HNF-36,,
&, and p3 cDNAs in different cells by RT-PCR, we only
obtained a single amplification
product with each set
of oligonucleotides
(see below).
To assess the relative abundance
of the HNF-36,,
HNF-3&, and HNF-3& mRNAs, we performed RNase
protection
assays with 30 pg cytoplasmic
RNA from
RIN 56 A and a riboprobe of 590 bases corresponding
to 534 bases of the 5’-end of the rat HNF-36, cDNA
and 56 bases of unrelated sequences.
As shown in
Fig. 3, three different bands were detected at 534,243,
and 173 bases, corresponding
to the expected protected fragments from HNF-36,, P2, and &, respectively. The most abundant mRNA corresponded
to that
encoding HNF-36,. The same relative distribution was
observed in the insulin-producing
cell line RIN 38 (data
not shown). We also investigated
whether HNF-36,,
p2, and p3 were present in the glucagon-producing
cell lines aTC1 and InRl G9 and in primary rat islets by
RT-PCR using 5’-primers specific for each cDNA (Fig.
1A) and a common 3’-primer.
Expected amplification
products for HNF-36,, &, and p3 were of 232, 218,
and 184 bp, respectively. As shown in Fig. 4, all three
products were obtained using primary rat islet cDNA;
the same profile was also observed when using glucagon- ((uTC~ and InRl G9) and insulin-producing
cell
line (RIN 38 and HIT-15) cDNAs (data not shown).
We recently showed that HNF-36, binds to the major enhancer
G, of the rat glucagon
gene, and its
overexpression
by transient transfection
assays leads
to a decrease in G,-mediated
chloramphenicol
acetyltransferase (CAT) activity. We thus assessed whether
HNF-36, and p3 could bind G, and affect its activating
properties. We inserted HNF-36,, &, and p3 cDNAs
Regulation
of Glucagon
Gene
Expression
371
into an expression vector driven by the cytomegalovirus (CMV) promoter (5) and transfected
these constructs into BHK-21 cells, a fibroblast cell line that
does not normally synthesize HNFB proteins. We
prepared nuclear extracts from DNA- and sham-transfected cells 48 h later. No G, binding activity was detected in nuclear
extracts from sham-transfected
BHK-21 cells (Fig. 5 A); by contrast, a complex comigrating with HNF38 from InRl G9 cells (the same pattern
was obtained with 56 A and crTC1 nuclear extracts) was
observed in nuclear extracts from cell lines transfected
with HNF36, (Fig. 5A, lane 2), p2 (lane 3), and p3 (lane 4).
HNF-36, was consistently more abundant than HNF-38,
and p3 despite the fact that similar amounts of total
proteins were assayed (all nuclear extracts were verified
for their ability to bind the octamer sequence). Formal
proof that HNF-38, with its amino-terminal
extension is,
in fact, synthesized awaits, however, the availability of
specific antibodies. Of note, HNF-3&, the smallest protein in size, migrated slightly faster than HNF-38, and &
(Fig. 5). To investigate the relative affinity of HNF-36,, &,,
and p3 for G,, we added increasing amounts of unlabeled G, oligonucleotide
competitor. Competition efficiency by G, was found to be similar for all three proteins
123M
.267
0184
0124
Fig. 4. PC!? Amplification
of the HNF-36
cDNA Variants
from Rat Islets
Total RNA was isolated
from rat islets and reverse
transcribed
with random
hexamers.
PCR amplification
of the
Y-end
of the HNF-36,
3p2, and p3 was performed
with an
oligonucleotide
specific
for each cDNA (Y-primer,
see Materials and Methods)
and a common
3’-oligonucleotide
annealing to the exon 3 sequences.
PCR reactions
were run on
a 2.5% agarose gel. The identity of the reaction products
was
verified
by Southern
analysis
using a 32P-labeled
HNF-38,
cDNA as a probe. Lane 1 corresponds
to a 184-bp
amplification product
of the HNF-38,
cDNA, whereas
lanes 2 and 3
contain the products
of the HNF-38,
(218 bp) and 8, (232 bp)
cDNAs, respectively.
%2345
6
Fig. 5. Binding
1
2
3
4
5
6
7
8
9
10
11
12
13
C
1
2
3
4
5
6
7
a
9
19
II
12 13
of the HNF-38
Protein lsoforms
on G,
Gel retardation
assays were performed
with a 3”P-labeled
oligonucleotide
containing
the G2 binding site (4) and 2 pg InRlG9
or 6 pg BHK-21 nuclear extracts.
BHK-21
cells were transfected
by the calcium phosphate
precipitation
method with expression
vector alone or containing
the respective
HNF-38
cDNAs. BHK-21
nuclear extracts
were then obtained.
A, Nuclear extracts
from
InRl G9 cells (lane l), BHK-21
cells transfected
with expression
vector containing,
respectively,
the HNF-38,
(lane 2), HNF-38,
(lane 3) HNF-38,
(lane 4) cDNAs or expression
vector alone (lane 5) were incubated
with 32p-labeled
G,. B, Nuclear extracts
from
lnRlG9
cells (lane 1) or cells expressing
either HNF-38,
(lanes 2-5), HNF-36,
(lanes 6 to 9), or HNF-36,
(lanes 10 to 13). Comp
indicates
the addition
of G, oligonucleotides
as competitors
in a molar excess of 10 (lanes 3, 7, and 1 l), 25 (lanes 4, 8, and 14),
and 50 (lanes 5, 9, and 13). C, Same as for panel A except that the competitor
oligonucleotide
contains
the HNF-3 binding site
of the lTR gene (7). Arrow
indicates
the HNF-36
complex
and dots indicate
nonspecific
complexes.
MOL
372
END0
1995
(Fig. 5B); similar results were obtained with the HNF3
binding site of the lTR gene (Fig. 5C). These results
indicate that the three HNF3P isoforms can bind G, with
similar affinities.
To examine whether HNF-S/3,, P2, and p3 affect
glucagon
gene expression
differently, we cotransfected their respective cDNAs inserted into a CMVdriven expression vector with a reporter plasmid containing G, linked to the glucagon gene promoter and
the CAT gene. HNF-3/3,, &, and p3 appeared to be
overexpressed
in these experiments
as nuclear extracts prepared from transfected cells showed a relative increase in the HNF-36 complex (data not shown).
Whereas we observed a dose-dependent
negative effect with HNF-36,) no change in transcriptional
activity
was seen with HNF-3P, and p3, indicating that the
latter isoforms, although capable of binding G,, do not
affect transcription
(Fig. 6). Titration studies were performed with different ratios of reporter to expression
vector (3 Kg/O.1 pg, 3 pgIO.5 Fg, 3 pgll pg, 1 pg/l
pg, 1 pg/2.5 pg, and 1 pg/5 Fg). Although these ratios
of reporter/HNF-30,
expression
vector resulted in
quantitative
differences
on transcriptional
activity,
they consistently led to a decreased activity. The negative effects of HNF-36, could be reduced by cotransfection of both HNF-3& or p3 cDNAs (Fig. 6). HNF-3
Fig. 6. Effects of HNF-36
lsoforms
on G,-Mediated
Transcriptional
Activity
lnRlG9
cells were cotransfected
with 1 wg of a reporter
plasmid
containing
wild type or mutated
G, or G, and the
glucagon
gene promoter
linked to the CAT gene (G,-136
CAT, G,M,-136
CAT, and G,-136
CAT, respectively),
1.25 or
5 pg HNF-36
cDNAs subcloned
into a CMV-driven
expression vector, and 1 pg of a control plasmid,
pSV,Apap
(28) to
assess transfection
efficiency.
The increasing
amounts
of
expression
vector or HNF-36
cDNAs cotransfected
with the
reporter
plasmids
are indicated
as molar ratios (l/l, 2.5/l, 5/l
or 2.5 + 2.5/l).
CAT activities
were measured
48 h after
transfection
and expressed
as a percentage
of the activity
obtained
with the reporter
plasmids
and expression
vector
alone. Results are corrected
for the amount
of proteins
and
the alkaline phosphatase
activities
and represent
the mean
t- ssM of six experiments.
Vol 9 No. 3
p2 and p3 may thus compete with HNF-3&
for the
G,-binding
site and their respective abundance
in glucagon-producing
cells along with that of the uncharacterized activator A, (4) may determine the activation
potential of G,. Overexpression
of HNF-36, did not
modify the transcriptional
activity obtained with G, or
a mutant of G, (G,M,) which binds HNF-36, poorly (4).
HNF-3 proteins were first identified in the liver and
shown to be important for liver-specific gene expression (5, 13, 14); they belong to a family of proteins
characterized
by a highly conserved DNA-binding
domain of 100 aa referred to as the WH motif (6). It is now
becoming apparent that distinct WH proteins are distributed in most tissues with patterns of expression
that are developmentally
and spatially restricted.
We recently reported the presence of HNF36 in all
islet cell phenotypes and its effect on glucagon gene
expression (4). We show here that HNF-36 consists of
at least three isoforms, p,, &, and &. Compared with
p,, p2 is characterized
by an amino-terminal
extension
of six aa, a feature also found for another WH member,
XFKHl (15, 16). &, by contrast, lacks the first 30 aa.
Whether additional HNF-36 forms are present in islet
cells is still unclear; this is suggested by the presence
of a 2.0 kb mRNA that hybridizes with an HNF-36specific probe, but neither our cloning nor our PCR
attempts were successful in identification
of cDNAs
corresponding
to the 2.0-kb mRNA.
The HNF-3 p2 and p3 cDNAs have recently been
reported to be expressed in the adult liver and the
notochord and midline neural plate cells of mammalian
embryos (10, 11). HNF-36 proteins may thus serve as
critical factors of endoderm
differentiation
and floor
plate development
(11, 17); their respective role and
importance
in these processes
are undetermined,
however.
Although
HNF-3 p2 and p3 both bind G, with a
similar affinity compared with p,, they do not affect,in
contrast to p,, the transcriptional
activity conferred by
G,. One potential explanation
lies in the difference
between the amino terminus of the three isoforms.
HNF-36, has indeed two transcriptional
activation domains, one of which is located within the 50 aminoterminal aa, is rich in serine, and contains two putative
casein kinase I phosphorylation
sites (7). Deletion of
the amino-terminal
activation domain of HNF-36, results in a 50% loss in the ability to activate transcription from the lTR promoter. The first 30 aa in HNF-36,
could thus be critical in the negative transcriptional
effects on glucagon gene expression. How the aminoterminal extension of HNF-36, abolishes the effects
observed with 6, is less clearly apparent; however,
three of the six aa are serines and potential sites of
phosphorylation.
Changes
in the phosphorylation
state of HNF-3&
may result in different negative
activation potential.
Alternative splicing of transcripts from a single gene
is often used to generate protein isoforms with different functions.
Many genes coding for transcription
factors are subject to alternative splicing, a mecha-
Regulation of Glucagon Gene Expression
nism directly responsible
for the generation
of transcription factors with distinct or opposite activities
(18-25). Our results suggest that the three HNF-3P
proteins
play different
roles on glucagon
gene
expression
and may compete for the G, binding site.
Several questions,
however, are raised by our data.
How is the synthesis of the different HNF-3/3 isoforms regulated?
Are different promoters
under the
control of different transcription
factors responsible
for the generation
of all three transcripts?
What are
the physiological
roles of the HNF-3 isoforms in the
developmental
regulation
of the glucagon
gene?
Their role may not be limited to glucagon
gene expression inasmuch as these proteins are present in
the different islet cell phenotypes;
identification
of
genes regulated by HNF-3fl may help to understand
islet development.
Further work will be needed to
better define the structure of the HNF-3P gene and
the function and regulation
of its protein products.
MATERIALS
Plasmids
AND METHODS
and Oligonucleotides
Oligonucleotides containing G, [nucleotides (nt) -201 to
-1651. G,M, (4). or G, (nt -274 to -234) with BamHl compatible ends’were ins&t& into a BamHl site 5’ the glucagon
373
plasmid;
1, 2.5, or 5 Fg effector
plasmid
(expression
vector
containing
the HNF-3/3,,
-p2 or -p3 cDNAs),
and 1 pg of the
plasmid
pSV,Apap
to monitor
transfection
efficiency.
pSV,Apap
is a plasmid
containing
the human alkaline phosphatase
gene driven
by the simian virus 40 long terminal
repeat (30). Cell extracts
were prepared
48 h after transfection and analyzed
for CAT and alkaline phosphatase
activities
as described
previously
(31). BHK-21
cells were grown in the
same medium
as the islet cell lines and transfected
by the
calcium
phosphate
precipitation
technique.
Cell Extracts
Nuclear
Schreiber
performed
and Gel Retardation
extracts
were
et al. (32).
as previously
Assays
prepared
by
Gel
retardation
described
(31).
the
method
of
assays
were
RT and PCR
Total RNA was extracted
from whole rat islets and cytoplasmic RNA was obtained
from mouse, rat, and hamster glucagon-producing
cell lines by standard
procedures
(33). RNA
was reverse-transcribed
into cDNA with random
hexamers
by Moloney
murine leukemia
virus reverse transcriptase
(Superscript,
Bethesda
Research
Laboratories,
Rockville,
MD)
and 100 ng of cDNA were used for 30 cycles of PCR amplification (1 min at 94 C, 1 min at 58 C, and 1 min at 72 C). The
different
HNF-38
cDNA variants
were amplified
with oligonucleotides
a and d (232 bp of HNF-36,)
b and d (218 bp of
HNF-36,)
and c and d (184 bp of HNF-38,).
PCR amplification products
were run on a 2.5% low melt agarose
gel. The
nature of the amplification
products was confirmed
by Southern
analysis.
promoter
(nt -136 to +51) (4). HNF-38,,
&. and p3 cDNAs
were inserted
into a CMV-driven
expression
vector (5). The
HNF-3 binding site from the transthyretin
gene promoter
(7)
was generously
given by M. Raymondjean
(Hopital
Cochin,
Paris). Sequences
of oligonucleotides
(5’-3’) used in the PCR
were as follows (see Fig. 1A): a) TCTGGAGCAGCGGCCAGCGAG [nt 162 to 182 from 8, (26)]; b) GCACTCGGCTTCCAGTATGCT (representing
the sequence
encoding
the six aminoterminal aa of ._.
8,); c) GAGCTCAGCCTAGGTGCTAACCTG;
d)
ACATAGGATGACATGTTCATGGAG
[nt 372 to 395 from 8,
(26)l: e) TATTGGCTGCAGCTAAGCGG
fnt 10 to 29 from
&, ‘[26)]; 9 GTTACAGTTAAGTCCCAGGGA
[nt 130 to 150
from 6, (26)]. The rat HNF-38,
cDNA was generously
provided by Dr. E. Lai (Memorial
Sloan-Kettering
Cancer Center,
New York).
Northern
blots were performed
with total RNA as previously
described
(27). HNF-B&-specific
probe was represented
by
exon 1, which was amplified
from rat genomic
DNA by oligonucleotides
e and f; HNF-36,
and &-specific
probes were
the 5’-untranslated
sequences
of the respective
cDNA
clones.
HNF-38,,
p2, and p3 transcripts
were assayed
in
cytoplasmic
RNA (30 kg) from the RIN 56 A cell line by RNase
protection
analysis (33). A rat HNF-381
probe was used and
obtained
by cloning
a 534 bp EcoRI-Aval
fragment
of the
HNF-38,
cDNA (26) along with 56 bp of unrelated
sequence
into Bluescript.
HNF-3p
Acknowledgments
cDNA Cloning
Polyadenylated
mRNA was isolated by the Poly A tract mRNA
isolation kit (Promega,
Madison,
WI). A random and oligo (dT)
primed cDNA library was made from 5 pg poly A+-selected
RNA from the mouse glucagon-producing
cell line, LY TCl
which was generously
provided
by Dr. D. Hanahan (University
of California,
San Francisco,
CA). After addition
of EcoRl
(Notl) adaptors,
cDNAs were ligated into the Agt 11 vector.
The cDNA library was screened
with a random-primed
32Plabeled HNF-3 p-specific
DNA fragment
corresponding
to the
first 340 bp of the HNF-38,
cDNA (4) according
to standard
procedures.
The isolated
cDNA clones
were inserted
into
Bluescript
(Stratagene,
La Jolla, CA) for sequence
analysis.
Cell Culture
and Transfection
Northern
We thank
de Peyer
and RNase
Protection
Isabel Pacheco for expert
for typing the manuscript.
Analyses
technical
help and Marie
Received September
20, 1994. Revision received November
7, 1994. Accepted
December
6, 1994.
Address
requests
for reprints to: Jacques
Philippe,
M.D.,
Department
of Genetics
and Microbiology,
Centre
Medical
Universitaire,
9, avenue
de Champel,
1211 Geneva
4,
Switzerland.
This work was supported
by the Swiss National
Science
Foundation.
Studies
The glucagon-producing
cell lines RIN 56A (rat) (27) (rTC 1
(mouse) (28), and InRl G9 (hamster)
(29) were grown in RPM1
1640 medium
containing
5% fetal calf serum and 5% newborn calf serum. InRl G9 cells were transfected
in suspension
by the diethylaminoethyl-dextran
method
with 1 kg reporter
REFERENCES
1. Philippe
J 1991 Structure
the insulin and glucagon
and pancreatic
expression
of
genes. Endocr Rev 12:252-271
MOL
374
END0
1995
2. Drucker
DJ, Philippe J, Jepeal L, Habener
JF 1987 Glucagon gene 5’-flanking
sequences
promote
islet cellspecific glucagon
gene transcription.
J Biol Chem 262:
15659-l
5665
3. Philippe J, Drucker
DJ, Knepel W, Jepeal L, Misulovin
Z,
Habener
JF 1988 Alpha cell-specific
expression
of the
glucagon
gene is conferred
to the glucagon
promoter
element by the interactions
of DNA-binding
proteins.
Mol
Cell Biol 8:4877-4888
4. Philippe
J, Morel C, Prezioso
V 1994 Glucagon
gene
expression
is negatively
regulated
by hepatocyte
nuclear
factor 3p. Mol Cell Biol 14:3514-3523
5. Costa R, Grayson
DR, Daniel1 JE 1989 Multiple
hepatocyte-enriched
nuclear factors function
in the regulation
of
transthyretin
and a,-antittypsin
genes.
Mol Cell Biol
9:415-1425
6. Clark KL, Halay ED, Lai E, Burley SK 1993 Co-crystal
structure
of the HNF-S/fork
head DNA-recognition
motif
resembles
histone H5. Nature 364:412-420
7. Pani L, Overdier
DG, Porcella A, Qian X, Lai E, Costa RH
1992 Hepatocyte
nuclear
factor 3p contains
two transcriptional
activation
domains,
one of which is novel and
conserved
with the Drosophila
fork head protein. Mol Cell
Biol 1213723-3732
8. Kaestner
KH, Hiemisch
H, Luckow
B, Schijtz G 1994 The
HNF-3 gene family of transcription
factors in mice: gene
structure,
cDNA
sequence
and mRNA
distribution.
Genomics
20:377-385
9. Pani L, Qian X, Clevidence
D, Costa RH 1992 The restricted promoter
activity
of the liver transcription
factor
hepatocyte
nuclear factor 3p involves a cell-specific
factor and positive autoactivation.
Mol Cell Biol 12:552-562
10. Ruiz i Atalba A, Prezioso
VR, Darnell JE, Jesell TM 1993
Sequential
expression
of HNF-3P
and HNF-3a
by embryonic
organizing
centers:
the dorsal lips/node,
notochord and floor plate. Mech Dev 44:91-l
08
11. Sasaki H, Hogan BLM 1994 HNF-3P
as a regulator
of
floor plate development.
Cell 76:103-l
15
12. Kozak M 1986 Point mutations
define a sequence
flanking the AUG initiator codon that modulates
translation
by
eukaryotic
ribosomes.
Cell 44:283-292
13. Lai E, Prezioso
VR, Smith E, Litvin 0, Costa RH, Darnell
JEJr 1990 HNF-3A,
a hepatocyte-enriched
transcription
factor of a novel structure
is regulated
transcriptionally.
Genes Dev 4:1427-l
436
14. Lai E, Darnell JE 1991 Transcriptional
control in hepatocytes: a window
on development.
Trends Biochem
Sci
16r427-430
15. Dirksen
ML, Jamrich
M 1992 A novel, activin-inducible,
blastopore
lip-specific
gene of Xenopus laevis contains
a
fork head DNA-binding
domain.
Genes Dev 6599-608
16. Lai E, Clark KL, Burley SK, Darnell JE 1993 Hepatocyte
nuclear factor B/fork head or “winged
helix” proteins:
a
family of transcription
factors
of diverse
biologic
function. Proc Natl Acad Sci USA 90:10421-10423
17. Ang SL, Wierda
A, Wong D, Stevens
KA, Cascio
S,
Rossant
J, Zaret KS 1993 The formation
and maintenance of the definitive
endoderm
lineage in the mouse:
involvement
of HNF-3/forkhead
proteins.
Development
119:1301-1315
18. Foulques
NS, Sassone-Corsi
P 1992 More is better: ac-
Vol 9 No. 3
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
tivators
and repressors
from the same gene. Cell 68:
411-414
Koenig RJ, Lazar MA, Hodin RA, Brent GA, Larsen PR,
Chin WW, Moore DD 1989 Inhibition
of thyroid
hormone
action by a non-hormone
binding c-erbA
protein generated by alternative
mRNA splicing.
Nature 337:659-661
Roman C, Cohen L, Calame K 1991 A dominant
negative
form of transcription
activator
mTFE3 created
by differential splicing.
Science
254:94-97
Nakabeppu
Y, Nathans
D 1991 A naturally
occurring
truncated
form of FosB that inhibits
Fos/Jun
transcriptional activity.
Cell 64:751-759
Foulkes NS, MellstrZjm
B, Benusiglio
E, Sassone-Corsi
P
1992 Development
switch of CREM function during spermatogenesis:
from antagonist
to activator.
Nature 355:
80-84
Descombes
P, Schibler
U 1991 A liver-enriched
transcriptional
activator
protein,
LAP, and a transcriptional
inhibitory
protein,
LIP, are translated
from the same
mRNA. Cell 67:569-579
Kozmik
Z, Kurzbauer
R, DCirfler P, Busslinger
M 1993
Alternative
splicing of Pax-8 gene transcripts
is developmentally
regulated
and generates
isoforms
with different
transactivation
properties.
Mol Cell Biol 13:6024-6035
Bach I, Yaniv M 1993 More potent transcriptional
activators or a transdominant
inhibitor
of the HNFl homeoprotein family are generated
by alternative
RNA processing. EMBO J 12:4229-4242
Lai E, Prezioso
VR, Tao W, Chen WS, Darnell JE 1991
Hepatocyte
nuclear factor 301 belongs to a gene family in
mammals
that is homologous
to the Drosophila
homeotic
gene fork head. Genes Dev 5:416-427
Philiooe
J. Chick WL. Habener
JF 1987 Multiootential
phe&ypic
expression
of genes encoding
pepiide
hormones in clonal rat insulinoma
cell lines. J Clin Invest
79:351-358
Powers AC, Efrat S, Mojsov
S, Spector
D, Habener
JF,
Hanakan
D 1990 Proglucagon
processing
similar to normal islets in pancreatic
a-like cell line derived from transgenie mouse tumor. Diabetes
39:406-414
Takaki R, Ono J, Nakamura
M, Yokogawa
M, Kumae S,
Hiraoka
T, Yamaguchi
K, Hamaguchi
K, Uchida S 1986
Isolation
of glucagon-secreting
cell lines by cloning
insulinoma
cells. In Vitro Cell Dev Biol 22:120-126
Henthorn
P, Zervos
P, Raducha
M, Harris H, Kadesh T
1988 Expression
of a human
placental
alkaline
phosphatase
gene in transfected
cells: use as a reporter
for
studies
of gene expression.
Proc Natl Acad Sci USA
8516342-6346
Philippe J 1991 Insulin regulation
of the glucagon
gene is
mediated
by an insulin-responsive
DNA element.
Proc
Natl Acad Sci USA 88:7224-7227
Schreiber
E, Mathias
P, Muller MM, Schaffner
W 1988
Identification
of a novel lymphoid
specific octamer
binding protein
(OTF-2B)
by proteolytic
clipping
bandshift
assay. EMBO J 7:4221-4229
Ausubel FM, Brent R, Kingston
RE, Moore DD, Seidman
JG, Smith JA, Struhl K 1994 Current
Protocols
in Molecular Biology.
John Wiley & Sons, New York, vol 1:
4.7.1-4.7.8.