of the L-pyruvate kinase gene in a hepatoma cell line is independent of insulin and cyclic AMP. FASEBJ. 8: 89-96;. 1994. Key Words: glucose signaling pathway.
RESEARCH COMMUNICATIONS
Glucose-dependent in a hepatoma cyclic
cell line
of the L-pyruvate is independent
of insulin
KAHN,’
LEFRANc0Is-MAR’FINEz, AND BENDICTE ii’roi
MARIA-JOSE
M. DIAZ-GU.ERRA,
Laboratoire de Recherches en G#{233}n#{233}tique et Pathologic Mol#{233}culaires,Institut National de Ia Sante et dc Ia Recherche MCdicale, 75014 Paris, France
Hepatocyte-like mhAT3F cells have been derived from the hepatoma of a transgenic mouse expressing the SV4O large T antigen under the control of the antithrombin III gene regulatory region (Antoine, B., Levrat, F., Vallet, V., Berbar, T., Cartier, N., Dubois, N., Briand, P., and Kahn, A. (1992) Gene expression in hepatocyte-like lines established by targeted carcinogenesis in transgenic mice. Exp. Cell. Res. 200, 175-185; F. Levrat et al., unpublished results). In these cells, the LPK gene is transcriptionally activated by glucose, as it is in vitro and in cultured hepatocytes. However, in contrast to the L-PK gene regulation in the liver and isolated hepatocytes, the glucose responsiveness does not require insulin and is not blocked by cyclic AMP. In mhAT3F cells, the insensitivity to insulin might be due to the replacement of insulin-dependent glucokinase by insulinindependent hexokinases able to phosphorylate glucose in the absence of the hormone. The glucose-dependent activation of the L-PK gene is delayed, requires ongoing protein synthesis, and is mediated by the same glucose response element as in vivo and in isolated hepatocytes. These results suggest that the glucose-dependent signaling pathway responsible for the transcriptional activation of glycolytic and lipogenic genes requires glucose phosphorylation, a phenomenon that is insulin-dependent in the liver but insulin-independent in cultured hepatoma cells. Nevertheless, the action of glucose 6-phosphate is most likely indirect.Lefrancois-Martinez, A. -M., Diaz-Guerra, M.-J. M., Vallet, V., Kahn, A., Antoine, B. Glucose-dependent regulation of the L-pyruvate kinase gene in a hepatoma cell line is independent of insulin and cyclic AMP. FASEBJ. 8: 89-96; 1994. ABSTRACT
Key Words: glucose signaling pathway . transcription diet . hepatocyte-like cell lines glycolytic enzyme
.
vate
gene
and
ISOENZYME
OF PYRuvATE
02-phosphotransferase,
0892-6638/94/0008-0089/$01
.50. © FASEB
Determination
Cochin
VERONIQUE
VALLET,
de G#{233}n#{233}tique Mol#{233}culaire, Institut
for the and hormones was achieved using transgenic mice or hepatocytes in primary culture; L-PK constructs driven by 183 bp of 5’ flanking sequence were responsive to glucose and hormones in a tissue-specificmanner (4-6). Different binding sites (boxes Li to L4) for transcription factors were identified; they are able to interact with, from 3’ to 5’: hepatocyte nuclear factor 1 (HNF 1),nuclear factor 1 (NF 1),hepatocyte nuclear factor4 (HNF 4), and major latetranscriptionalfactor (MLTF) (7). The glucose/insulin response element of the L-pyruvate kinase gene was determined to be a perfect palindrome located from nucleotide -168 to -144 with L-PK
respect
gene
of the DNA
expression
to the
cap
elements
in response
site.
This
element
responsible
to carbohydrates
is partially
homologous
to the MLTF binding site and its full efficiency requires cooperation with the contiguous binding site for HNF 4 located from nucleotide -145 to -125 (6, 8, 9). To identify the factors involved in the response of the LPK gene to carbohydrates and hormones, hepatocytes in primary culture provide a powerful tool, their responsiveness to hormones and nutrients being well conserved (3). However, the inabilityof these cellsto divide constitutesa serious limitation for further functional investigations. Several lines of well-differentiatedhepatoma cells had been described that,unfortunately,had lostthe normal control of the L-PK and other glycolytic genes by nutrients and hormones (10). Recently we established by targeted carcinogenesis in transgenic mice novel lines of differentiated and immortalized hepatocyte-like cells that exhibited a good liver-specific pattern of gene expression, and whose endogenous L-PK gene was regulated by glucose but not by insulin (11, 12).
high glucose
KINASE (L-PK)(ATP: pyruEC 2.7.1.40) is a tissue-specific unidirectional glycolytic enzyme that plays a key role in the regulation of glycolysis and gluconeogenesis in the liver. In rat liver, the L-PK expression is under dietary and hormonal control; it is induced by a high glucose diet and suppressed by starvation or in diabetes. In vivo studies and ex vivo experiments in cultured hepatocytes have shown that the L-PK gene expression is transcriptionally regulated positively by the couple glucose and insulin, and negatively by glucagon through its second messenger, cyclic AMP (cAMP) (1-3).
L-TYPE
kinase
AMP
ANNE-MARIE AXEL
regulation
ITo whom correspondence should be addressed, at: Laboratoire de Recherches en G#{233}n#{233}tique et Pathologie Mol#{233}culaires,Institut Cochin de G#{233}n#{233}tique Mol#{233}culaire, Institut National de Ia Sante et de Ia Recherche M#{233}dicale, Unite 129, 24 rue du Faubourg SaintJacques, 75014 Paris,France. 2Abbreviations: AT III, antithrombin III; CAT, chloramphenicol acetyl transferase; 2 dG, 2 deoxyglucose; GK, glucokinase; GIRE, glucose response element; HNF, hepatocyte nuclear factor; MLTF, major late transcriptional factor; L-PK, L-type pyruvate kinase; PABP, poly A binding protein; PEPCK, phosphoenolpyruvate carboxykinase; SV 40, simian virus 40; NF-1, nuclear factor 1; FCS, fetal calf serum; 8-Br-cAMP, 8-bromo-cyclic AMP; 8-4-CPTcAMP, 8-4-chlorophenylthio-cAMP; DO’FAP, N-(1-(2, 3-dioleoyl-
oxyl)
propyl)-N,N,N-trimethyl-ammonium
cyclic
AMP.
methylsulfate;
cAMP,
89
RESEARCH COMMUNICATIONS In the present report, we have characterized in detail the mechanisms of the glucose-dependent induction of the L-PK gene in one of these lines, mhAT3F. Both endogenous L-PK gene and chioramphenicol acetyl transferase (CAT) constructs driven by the 183-bp L-PK gene promoter are induced by glucose, regardless of the presence of insulin or cAMP. However, the glucose-dependent transcriptional activation is mediated by the glucose response element (GIRE) previously characterized in hepatocytes, where its function requires insulin and is blocked by cAMP (9). The glucose effect is partially mimicked by fructose, is slow, and requires ongoing protein synthesis. These results confirm that glucose derivatives are the actual activators of the L-PK gene, insulin playing a permissive role in hepatocytes but being dispensable in this hepatoma cell line. MATERIALS Cell lines
AND
METHODS
and culture
and mhPKT hepatocyte-like cell lines were derived from the tumoral liver of transgenic mice expressing the SV4O early genes under the direction of the liver-specific antithrombin III promoter (ATIII-SV4O) or L-PK promoter (PK-SV4O). Cells were cultured in HamFl2-DMEM (vol/vol) (Gibco. Chagrin Falls, Ohio) medium supplemented with penicillin, streptomycin, 0.1 sM insulin, 1 sM dexamethasone, I ILM triiodothyronin, and 5% (vol/vol) fetal calf serum (FCS) (11). mhAT
Hormonal
regulation
Protein
(I or 20 mM).
synthesis
Cycloheximide
inhibitors
(10 tM)
or 25 tM
anisomycin
was added
in the glucose-
containing culture medium and the results were analyzed by Northern after different times of culture.
Northern Total
blot
blot analysis
RNAs
were isolated
from
cell lines by lysis in guanidinium,
followed
by phenol extraction (13). The RNA concentration was determined spectrophotometrically. RNAs were denaturated with methylmercury (II) hydroxide, electrophoretically separated on formaldehyde agarose gels, and finally transferred and U. V. cross-linked to nylon filters (Hybond-N, Amersham, Arlington Heights, Ill.). Prehybridization (I h with 100 gg/ml salmon sperm DNA) and hybridization were carried Out as previously described (14) except that the temperature was 55#{176}C when probes of rat or human origin were used.
Figure
ribosomal probe as standard (R45 eDNA probe, a gift from J. P. Hugnot, ICGM, Paris). The intensity of the autoradiographic bands was measured using a Shimadzu CS930 densitometer. The results of mRNA abundance were expressed in arbitrary units after standardization by the PABP or R45 signals, as previously reported. Transfection
experiments
The L-PK-CAT constructs used have been described previously (6, 9). Transfection was performed by lipofection using N-(1-(2, 3-dioleoyloxy) propyl)-N,N,N-trimethyl-ammonium methylsulfate (DUTAP, Boehringer Mannheim, Mannheim, Germany) according to the manufacturer’s instruc-
tions, in cells cultured under the lactate conditions. Five micrograms of CAT constructs and 2 sg of the pRSV-luciferase plasmid were cotransfected when cells were 60-80% confluent in 60-mm plastic dishes (Falcon, Oxnard, Calif.). After 12-14 h, the media containing the liposome-DNA complex were removed and replaced with medium as described above for hormonal regulation studies. Cells were harvested 36 h after transfection end. CAT and luciferase activity assays were performed as described (9). CAT activity was calculated as the percentage of the diacetylated form (upper form) of chloramphenicol vs. the nonmetabolized substrate. Results were expressed as a ration of CAT activity vs. luciferase activity so correct for the transfectability
variability.
studies
For the hormonal regulation studies, cells were first grown in the presence of FCS. Twenty four hours before the experiment, cells were cultured in a serum-free, glucose-free medium supplemented with 1 tM triiodothyronine, 1 gM dexamethasone, 10 sg/ml transferrin (Sigma, St. Louis, Mo.), 100 sg/ml albumin (Sigma), and 10 mM lactate. The various induction studies were then performed under the following conditions: 17 mM gluLose; 17 mM glucose, and 0.1 sM insulin; 17 mM glucose, 0.1 sm insulin, and 1 mM 8-bromo-cyclic-AMP (8-Br-cAMP) and 0.1 mi 8-4-chlorophenylthio-cAMP (8-4-CP’PcAMP); 17 mM glucose, 0.1 sM insulin, and 5 gM glucagon. Induction of the L-PK gene expression was also analyzed in the presence of various concentrations of glucose, fructose (2 mM), or 2-deoxyglucose
The probes for rat L-PK, rat phosphoenolpyruvate carboxykinase (PEPCK), and human poly (A) binding protein (PABP) sequences have been described previously (11). The rat glucokinase (GK) probe was a 1.8-kb eDNA fragment provided by P. lynedjian (15). The PABP probe was used as a standard for mRNA quantifications because its level is relatively unvariant in different cultured cell lines (unpublished data). However, the in vivo abundance of the PABP mRNA is much lower, for instance in the liver, than cx vivo in cultured cells. Therefore, comparisons between the abundance of mRNAs in vivo and cx vivo were performed through the use of a 18S
1. L-PK promoter
activity
in hepatocyte-like
cell lines according
RESULTS
AND
DISCUSSION
Expression according
of the L-PK to carbohydrate
gene in hepatocyte-like culture conditions
cell lines
In hepatocyte-like cell lines derived by targeted carcinogenesis, expression of the L-PK gene was studied by Northern blot analysis. The 32P-labeled rat L-PK cDNA probe hybridized, at the low stringency used (Fig. 1A), with the L-PK and M2-PK mRNAs. In cells cultured in the glucose/insulin-containing regular medium, the 3.2-kb L-PK mRNA, the only species in mice (16), accounted for 5 to 50% of the signal in the carbohydrate-induced mouse liver (lane h). Among the different hepatocyte-like cell lines (11, 12), we selected the line mhAT3F because its expression of the L-PK gene was strongly regulated by glucose. When these cells were cultured for 24 h with lactate in the absence of glucose and insulin (Fig. 1B, lane 1), no significant accumulation of the L-PK mRNA occurred, whereas the M2-PK mRNA, whose expression is not sensitive to diet in vivo, was still present. The glucose-dependent accumulation of the L-PK mRNA in mhAT3F cells was independent of insulin (compare lanes 2 and 3 Fig. IB) in contrast to the need for insulin in hepatocytes in primary culture (9). In some other
to carbohydrate
and hormonal
culture
conditions.
A, B)
Northern
cell lines. Fifteen micrograms of total RNAs were loaded in each slot. The with a probe recognizing the unvariant poly (A) binding protein mRNA. The cDNA probe used to hybridize the mouse L-PK mRNA was of rat origin and allowed to detect mRNA encoded by the M2-PK isoforin gene. A) Cells were cultured in the presence of 17 mM glucose: lane a, mhAT3F cells; lane b, mhPKT cells; lane c, mhAT1F8 cells; lane blot analysis of the L-type PK transcripts amount of loaded RNA was standardized
in hepatocyte-like by hybridization
d, mhPKTF3 cells; lane e, mhATIA3 cells; lane f, mhAT1A6 cells; lane g, mhATlG8 cells; lane h, carbohydrate-induced mouse liver. B) Northern blot analysis of the L-type transcripts as a function of the culture condition in mhAT3F cells (lanes 1 to 4) and in mhPKT cells (lanes 6 to 9). Cells were first cultured in a lactate culture medium for 24 h before glucose and hormone supplementation, as described in Materials and Methods; then for 36 h culture conditions were: 1 and 6, 10 mM lactate; 2 and 7, 17 mM glucose; 3 and 8, 17 mM glucose and 0.1 sM insulin; 4 and 9, 17 mM glucose, 0.1 sM insulin, 1 mM 8-Br-cAMP, and 0.1 mM 8-4 CVIcAMP. C) Representative CAT experiments performed on extracts of mhAT3F cells and mhPKT cells transfected with 5 ig of L4L3-1l9 PKJCAT (lanes 1 to 4 and 6 to 9) and 2.5 tg of kSV2/CAT plasmids (lanes 5 and 10). Culture conditions are the same as in B. (See next page.)
90
Vol. 8
January
1994
The FASEB Journal
LEFRANcOIS-MARTINEZ
ET AL.
RESEARCH COMMUNICATIONS A.
isoform M2 isoform
L
a
bcdef
gh
Poly A Binding Protein control
B. mhPKT
mhAT3F
1234
+ -
6789
+ -
+ + -
+ + +
Lactate Glucose Insulin cAMP
+ -
+ -
+ + -
+ + +
C. mhPKT
mhAT3F
4
.5
JLL3J
+ -
+ -
+ +
+ +
-
-
-
+
GLUCOSE-DEPENDENT,
INSULIN-INDEPENDENT
+
-
REGULATION
Lactate Glucose Insulin cAMP
OF THE
L-PK GENE
+ + -
+ +
+ + +
+
91
RESEARCH COMMUNICATIONS an immediate adjustment of the glucose phosphorylation rate to the intracellular glucose concentration over a physiological range, in contrast to the other hexokinases whose Km for glucose is at least 10- to 100-fold lower (18). In addition, in contrast to the other hexokinase isoenzymes, GK is accurately regulated at the pretranslational level by insulin (19) and at the post-transcriptional level by the ratio fructose 1-P to fructose 6-P (20). In liver and in cultured hepatocytes, transcription of the GK gene is quickly activated by insulin regardless of the glucose concentration (19), and inhibited by glucagon and cAMP (19, 21). In mhAT3F cells, however, the GK gene is extinguished and cannot be reactivated by insulin (Fig. 2B, lanes c-c), whereas insulin is perfectly efficient in inhibiting transcription of the PEPCK gene (Fig. 2B, lane f). This demonstrates that neither the insulin-insensitive extinguishment of the GK gene in these cells nor the insensitivity to,insulin of the glucose effect on L-PK gene expression results from the absence of insulin receptor or their uncoupling to the downstream signaling pathway. In the absence of GK, glucose phosphorylation in mhAT3F cells is assured by other insulin-insensitive hexokinase isozymes known to replace GK in hepatoma cells (22). Accordingly, the glucose dose-response curve (Fig. 2C) shows that the L-PK gene was activated by glucose concentration in mhAT3F cells lower than that in hepatocytes in Effects of hormones on glucose responsiveness of the primary culture (3), as expected because GK has a lower L-PK gene in mhAT3F cells affinity for glucose than the other hexokinase isoforms. However, this experiment does not exactly reflect this hexInsulin was inactive by itself (Fig. 2A) and was unable to okinase affinity because of the continuous glucose consumpmodify glucose responsiveness of both the endogenous L-PK cells, progressively lowering glucose concengene and the L4L3-119 PK/CAT construct in mhAT3F cells tion by cultured tration during the 36-h induction. It is tempting to correlate (Fig. IB, Fig. 1C and Fig. 2A, lane 4). the insulin-independent glucose phosphorylation in Glucagon and/or cAMP addition were also inactive on mhAT3F cells and the loss of the insulin action on glucoseboth glucose-dependent L-PK mRNA accumulation (15.5dependent activation of the L-PK gene. Indeed, it has reand 14.6-fold stimulation when cAMP or cAMP and glucacently been suggested that the fatty acid synthase and acetylgon were added instead of an 18.1-foldin their absence) and coA carboxylase lipogenic genes whose activation in hepatoCAT activityof the L4L3-119 PK/CAT construct (Fig. 2A). cytes requires the presence of both glucose and insulin beForskolin,an activatorof the intracellularcAMP generating come sensitive to glucose alone in adipocytes, which are also system, gave the same negative results as cAMP (Fig. 2A). devoid of GK. Moreover, in this case a correlation was estabIn liver, the first step of glycolysis is catalyzed by between mRNA accumulation and intracellular conglucokinase (GK). GK (hexokinase type IV) is a distinctive lished of glucose 6-P or 2 deoxyglucose 6-P synthetized member of the mammalian hexokinase gene family (17).It centration from the nonmetabolized 2 deoxyglucose analog (23). The differs from the other three hexokinases (type I-Ill) by its question of whether a similar phenomenon could be obsmaller size, its greater specificity for glucose, and above all, served on the L-PK gene expression in mhAT3F cells could its lack of end product inhibition by glucose 6-P and its high therefore be raised. Michaelis-Menten constant (Km) for glucose, which permits hepatocyte-like cell lines that exhibited good levels of L-PK gene expression, e.g.,mhPKT cells(Fig. 1B, lanes 6 to 9), glucose-dependent regulation of the L-PK gene expression was attenuated and did not seem to be specific to the L isoform as compared with the M2 mRNA. To determine whether the glucose-dependent regulation of the L-PK gene was transcriptional,we investigatedthe ability of the mhAT3F cellsto reproduce the glucose-regulated expression of a L-PK CAT construct (9). Transfectability of mhAT3F cellsby lipofection,as estimated by the expression of the luciferasegene, was about 50-fold higher than that of hepatocytes in primary culture. In addition, the L4L3-119 PK/CAT construct, which contains all cis-actingDNA/elements necessary for the glucose responsiveness (9), was strongly stimulated by glucose in these cells (Fig. IC, lanes 1 and 2) in contrast to the results with other lines, e.g., mhPKT cells (Fig. 1C, lanes 6 and 7). Figure 2A shows that when mhAT3F cellswere firstcultured for 24 h in a medium containing 10 mM lactatewithout glucose and insulin,and then for 36 h in a glucose (17 mM) culture medium, accumulation of L-PK mRNA (stimulated 18.4-fold by glucose) and CAT activity of the L4L3-119 PK/CAT construct (stimulated 15.5-fold) were similarly increased.
2. Relationships between glucose-regulated L-PK promoter activity and hormonal culture conditions in mhAT3F cells. A) Variation of L-PK mRNA levels under various glucose and hormonal culture conditions (empty box) were quantified by scanning the same Northern blot autoradiograms as in Fig. lB. The values are the means of three distinct experiments and are expressed relative to the value obtained under the lactate culture condition. CAT activity (black box) of the L4L3-119 PK/CAT construct transfected in mhAT3F cells under various glucose and hormonal culture conditions was determined by the percentage of chloramphenicol conversion to its acetylated forms. Fold stimulation represents % conversion under each culture condition/% conversion under lactate culture condition. The values are the mean of five to six experiments for each culture condition. It was also verified that the kSV2CAT plasmid was similarly active under lactate and glucose culture conditions. B) Glucokinase and phosphoenolpyruvate carboxykinase expression in mhAT3F cells analyzed by Northern blotting. Fifteen micrograms of total RNAs were loaded in each slot. The cDNA probes used to hybridize mouse glucokinase (GK) (lanes a to e) and phosphoenolpyruvate carboxykinase analyzed by Northern blotting (PEPCK) (lanes f and g) were of rat origin. Culture conditions were: lane a, carbohydrate-induced mouse liver; lane b, rat hepatocytes in primary culture with 17 mM glucose, 0.1 iM insulin for 8 h; lanes c to e, mhAT3F with 17 mM glucose, 0.1 tM insulin for 8 h (c), 12 h (d), 16 h (e); lane f, mhAT3F cells with 17 mM glucose, 0.1 sM insulin; lane g, mhAT3F cells with 10 mM lactate, 1 mM 8-Br-cAMP, 0.1 mM 8-4 CP1CAMP for 36 h. C) Glucose dose-response curve of L-PK gene induction in mhAT3F cells compared with primary hepatocytes. Results are expressed as the percentage of the value obtained with 40 mM glucose of either L-PK promoter activity in mhAT3F cells (black boxes) or endogenous L-PK mRNA abundance in hepatocytes in primary culture (empty boxes). The results in hepatocytes in primary culture are reproduced from ref 3. (See next page.) Figure
92
Vol. 8
January
1994
The FASEB
Journal
LEFRAN#{231}0IS-MARTINEz El AL.
RESEARCH COMMUNICATIONS A.
stimulation
Fold
DU
-U .
,.9
. .
10
1
Lactate
+
+
+
Glucose +
Insulin cAMP
+
+
+
+
+
+
+
+
-
+
+ +
Glucagon
+
-
Forskolm
B. OK mRNA
PEPCK
PABP mRNA
C.
mRNA
PABP mRNA
0 -.
SI)
q
100
100
50
50
5 0 (lactate)
GLUCOSE-DEPENDENT,
INSULIN-INDEPENDENT
REGULATION
10
20
OF THE i-PK GENE
U-
40 mM glucose
93
RESEARCH COMMUNICATIONS Influence of 2 deoxyglucose of the L-PK gene promoter
and fructose in mhAT3F
The 2 deoxyglucose (2 dG) is an effective competitor cose transport and subsequent phosphorylation. The yglucose
not
6-phosphate
further
is trapped
for glu2 deox-
within
cells
endogenous
L-PK
mRNA
or
CAT
activity
of
the
L4L3-119 PK/CAT construct was observed after 16 h of culture (Table 1). These results seem contradictory with those obtained in the rat insulinoma cell line (INS-I cells) in which the 2 dG was partially effective in mimicking the effectof glucose on the L-PK gene expression (25). However, the important loss of 2 dG 6-P from liver cells, either through the action of glucose 6-phosphatases or by metabolism of 2 dG 6-P to other labile products, could introduce a strong artifact in interpreting the results in hepatocytes and hepatocyte-like cell lines, such that our data are not necessarily inconsistent with the hypothesis that glucose 6-P and derivatives are, directlyor indirectly, the transcriptionalstimulatorsof glycolytic
genes
(e.g.,
the
L-PK
gene)
or lipogenic
genes
(e.g.,
the
fatty acid synthase and acetyl coA carboxylase gene (23)). In normal rat liver, dietary fructose markedly stimulates transcription of the L-PK gene (14). Previous studies have showed
that
this
fructose-dependent
mechanism
tion ture,
to test
similar
the
efficiency
to that of glucose.
a similar
effect
of fructose
of this
sugar
at a concentra-
In hepatocytes in primary on L-PK gene expression
culhas
been reported (26). In this case, however, the fructose effect was insulin-dependent. Fructose in mhAT3F cellscan probably be phosphorylated in fructose 6-P by hexokinases and fructose 1-P by fructokinase, both metabolites being able to be metabolized through the glycolytic pathway or to generate glucose 6-P.
TABLE
1. Effect
aciivity in
mhA
of fructose
and 2-deoxyglucose
on the L-PK
promoter
T3F ce11s Fructose, 2mMfor36h
Endogenous L-PK mRNA levels (fold stimulation) L4L3-119 PKICAT
2-deoxyglucose, lmMforl6h
2-deoxyglucose, 20mMforl6h
6
1.5
2.4
3.4
1.1
2.2
activity (fold
stimulation)
Wariation of L-PK mRNA levels were quantified by scanning the Northern blot autoradiograms. Results represent the means of three distinct experiments and are expressed relative to the value under the lactate culture condition. CAT activity of the L4-L3-119 PK/CAT construct transfected in mhAT3F cells was determined as described in Fig. 2. The values are the means of four to six experiments for each culture condition and are
expressed in fold stimulation lactate culture condition.
94
Vol. 8
January
with respect to the value obtained
1994
15
under the
The
.51 a
10 EQ #{149}*
5
+ anisomyCin
+ cycloheximid -
4
mhAT3F
cells
8
6
3. Time
Figure
10
12
14
18
16
course of L-PK promoter and
inhibition
by
protein
24
20
activation synthesis
hours
by glucose in inhibitors.
After
decreasing the level of the L-PK mRNA by culture in 10 mM lactate for 24 h, mhAT3F cells were cultured with 17 mM glucose alone (black square or stars) or with 17 mM glucose and 25 tM anisomycin (shaded square) or with 17 mM glucose and 10 LM cycloheximide (white square) for the indicated times. The amount of L-PK
mRNA
zation culture
Each relative
point to the
conditions.
transfected mM glucose
was
(squares)
analysis.
is expressed
is powerful
enough, in vivo, to overcome the inhibitory effect of cAMP in the liver of fasted animals without inducing insulin secretion (14).In mhAT3F cells,2 mM fructose was substituted for glucose in the insulin-freeculture media (Table 1).The abundance of the L-PK mRNA was stimulated sixfold by 2 mM fructose (instead of 18.4-fold by glucose) and the CAT activity of the L4L3-119 PK/CAT construct was also subjected to the same variation (Table 1). Due to the known toxic effect of fructose at higher concentrations (26), it was impossible
by glucose
but
metabolized through the Embden-Meyeroff 24). In the presence of 1 mM or 20 mM 2 dG, for glucose in the culture medium, no increase
pathway (23, as a substitute of
so formed
stimulation
on the activity cells
CAT
in mhAT3F incubations
determined is the
value
activity
at zero of the
as well
time,
blot
hybridi-
determinations under
L4L3-1l9
and
10 mM
PK/CAT
then
as in hepatocytes
lactate
construct
induced
as described
for CAT assay 36 h after transfection of three
by 17 in Fig.
end.
experiments.
Time course of glucose-dependent in mhAT3F cells vivo
Northern
of two
cells in 10 mM lactate, (stars) was determined
2. All cells were harvested Each value is the mean
In
by
mean
L-PK in primary
gene
induction
culture,
it was
shown that insulin and carbohydrates induce transcription of the L-PK gene expression with a maximum between 12 and 16 h of induction (14, 26), and that this induction can be blocked by the protein synthesis inhibitor cycloheximide (2, 26); however, itisdifficult to determine in these resultswhat is due to the action of insulin, probably through the necessary induction of GK synthesis, and to that of glucose itself. In contrast, the insulin independence of the L-PK gene regulation by glucose in mhAT3F cells allowed us to address this question directly. When mhAT3F cells are cultured and transfected without glucose (lactate condition) and then incubated for various periods with glucose, no glucose-induced enhancement of CAT activity of the L4L3-119 PK/CAT construct was observed before the 12th h. CAT activity under glucose stimulation was maximal between 12 h and 14 h, reaching an approximate 12-fold stimulation and then reaching a plateau until at least the 36th h (Fig. 3). Similarly, glucose-induced mRNA accumulation was not observed before 8 and 12 h of culture and reached a maximum about 24 h after the addition of glucose (about ninefold stimulation) (Fig. 3). In the insulinoma-derived INS-i cells,induction of the LPK gene transcription by glucose was shown to be more rapid than in mhAT3F cells,but still delayed because itwas detected only 2 h after glucose addition (25). Such a delay was not in favorof a directpost-translationaleffectof quickly accumulated
glucose
whose steady-state be reached within
FASEB Journal
metabolite
such
as
the
glucose
6-P
intracellular concentration is expected to a few minutes. This inference was also
LEFRANCOIS-MARTINEZ
Er AL.
RESEARCH supported cells.
Ten
which
are
expression
micrograms sufficient
concentrations
sis (26), totally abrogated endogenous L-PK mRNA
and
24 h of culture
cycloheximide-sensitive
to inhibit
protein
the inductive
effect
accumulation
observed
(Fig.
3). Therefore,
transcriptional
the
of glucose
on
between
8
delayed
induction
Cis acting
glucose
Identification ment (G1RE) or
response
element
of
in mhAT3F
in
primary
culture
the
responsiveness.
As opposed
cells
response eletransgenic mice (9, 27). This DNA nucleotides -168 and
to the
results
in hepatocytes
where a functional GIRE also mediated the cAMPdependent inhibition of the L-PK gene promoter (9), CAMP was inefficient in inhibiting the glucose-dependent transcriptional activation (data not shown). The parallel loss in mhAT3F cells of the permissive effectof insulin and of the inhibitory effect of cAMP on this glucose-dependent activation of the L-PK gene might suggest that a regulatory component common to the insulin and cAMP signaling pathways is lost. However, the CAMP-dependent transcription blockage of the L-PK gene isan early phenomenon independent of protein synthesis (1),which consequently cannot be ascribed only to the negative effectof CAMP on expression of the GK gene. Further investigationsare clearly required to solve this paradox. Sequences upstream of the G1RE have been demonstrated to play complex roles on the expression of the L-PK gene, which differ,in vivo, in differenttissuesof transgenic mice (5) and, ex vivo, in transiently transfected hepatocytes (9). In these latter cells,the element L5 borne in the -283 L-
GLUCOSE-DEPENDENT,
INSULIN-INDEPENDENT
.%
.50
REGULATION
n
I
ri,j
CAT -.
,
l
CAT
.me
L-
L4) is located between -144 with respect to the cap site and contains two putative MLTF-binding sites. It is located in the immediate vicinity of a HNF4 binding site. In hepatocytes, any mutation of the box L4 that impairs MLTF binding, or of the box L3 that impairs HNF4 binding, abolished the glucose/insulin responsiveness of the L-PK gene. However, oligomerized L4 boxes were able to reconstitutea glucose/insulin responsive promoter in the absence of the L3-HNF4 binding site (9, 28). Figure 4 shows that the L-PK G1RE also mediates the glucose responsiveness of the L-PK gene in mhAT3F cells: those mutations deleterious for the G1RE activity in hepatocytes are deleterious in mhAT3F cells as well. Whereas the L4L3-119 PK/CAT construct was stimulated about 15-fold in these experiments, the two L4 mutations (L4 me and L4 mi) that suppressed binding activity of the box L4 (9) also resulted in a very low glucose responsiveness (in the range of a twofold stimulation). The same results were observed for the mutant devoid of the box L4 and for the L3 mutant impairing HNF4 binding (L3 mi). In contrast, the L3 me mutant neither altered HNF4 binding (9) nor blocked the glucose
(box
.fl9
.150
_LCAT
and
of the L-PK gene glucose/insulin was previously described using
hepatocytes
#{149}3
CAT activity (% kSV2CAT) itt lactate
#{149} 15.4
0.7
x3
1.5
.2.6
1.6
a 7.25
0.85
a1.2
0.62
xS.2
0.8
a7.6
0.6
a1.4
0.6
synthe-
PK gene by glucose and insulin in livercellsdoes not only rely on the need for a prior induction of GK synthesisby insulin, which is a rapid phenomenon, but also seems to result from the characteristics of the glucose signaling pathway itself. This suggests that glucose 6-P, a probable intermediate in the glucose-dependent transcriptional activation, is not likely to act on its own, for instance, by the post-translational activation of a pre-existing transcription factor active on the L-PK gene, but rather could either stimulate synthesis of such a factor or give rise to active metabolites whose progressive accumulation would require the de novo synthesis of some enzymatic proteins.
element
Fold stimulation by glucose
by the necessity
glucose-activating
of ongoing protein synthesis for of the L-PK gene in mhAT3F cycloheximide or 25 LM anisomycin,
COMMUNICATIONS
lEnIri3lMtiL,
rCAT
--i--
253
rEn rrn
EE
I
-ISO
#{149} 1111.. 4. Cis-acting
Figure cells.
Several
DNA
V1
glucose segments
.
I
-
Al’
response
element
spanning
from
(G1RE) nt
-172
containing point mutation of the L4 or L3 binding serted in front of the I19PK/CAT construct. Two ments upstream of the box L4, up to nt -283 or added to the wild-type sequence. Finally, the box Broken lines between the boxes represent the 8 bp ing the preparation of the constructs. CAT activities culture conditions for each construct are expressed of kSV2CAT values (100%). The extent of induction the ratio between CAT activities with and without value represents the mean of six experiments.
PK/CAT modify result activity
in mhAT3F to
-122,
each
sites, were inadditional fragnt -3193, were L4 was deleted. introduced durunder lactate relative to that by glucose is glucose. Each
construct behaved as an inhibitor that did not the glucose responsiveness. Exactly the same type of is found in mhAT3F cells, as evidenced by the lower of the
compared
SUMMARY
with
-3193
the
AND
and
-183
the
-283
(or L4L3-119)
L-PK/CAT
PK/CAT
constructs
constructs.
CONCLUSIONS
Our results demonstrate that the L4 element of the L-PK gene promoter is actually a glucose response element that is able to respond independently of insulin in some cells, e.g., hepatocyte-like mhAT3F cells, in which the insulindependent GK (hexokinase IV) is replaced by hexokinase isozymes able to phosphorylate glucose in the absence of insulin. The insulin-independent transcriptional effect of glucose on the L-PK gene is relatively delayed and requires ongoing protein synthesis such that, although its accumulation is most likely necessary, glucose 6-P is not likely to act as the direct activator of a pre-existing transcription factor. The cAMP-dependent inhibition of the L-PK gene, also mediated through the G1RE in hepatocytes, and in vivo in the liver of transgenic mice, is similarly lost in mhAT3F cells as is the insulin dependence. This cAMP negative effect is known to be a very rapid phenomenon, independent of protein synthesis (25), which cannot be explained simply by the inhibition of GK synthesis. Consequently, further investigations are necessary to understand the mechanisms of the negative cAMP effect on the transcription of genes activated by glucose and/or insulin, and to explain the parallel loss of the insulin and cAMP responsiveness in cells such as mhAT3F cells. The answer to this problem will probably require the molecular dissection of the glucose-dependent signaling pathway, from glucose 6-P to the transcriptional machinery. After a long period of time, during which the
OF THE L-PK GENE
95
RESEARCH
COMMUNICATIONS
mechanisms of the glucose action on gene transcription not understood at all, the progress in our knowledge field should be rapid.
were in this
B. (1993) Influence of the content in transcription factors on the phenotype of mouse hepatocyte like cell lines (mhAT) Exp. Cell. Ret. In press 13. Chomzynski, P., and Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction.
Anal. Bioc/zern. 162, 156-159
This work was supported by grants from L Association pour Ia Recherche sur le Cancer (ARC), from La Fondation Francaise pour Ia Nutrition (FFN), from Ia Ligue Nationale Francaise Contre Ic Cancer (LNFCC), and from l’Institut National de Ia Sante et de Ia Recherche M#{233}dicale (INSERM). We thank Alexandra Henrion from our laboratory for careful revision of the text. REFERENCES 1. Vaulont, S., Munnich, A., Decaux, J. F., and Kahn, A. (1986) Transcriptional and post-transcriptional regulation of L-type pyruvate kinase gene expression in rat liver. J. Biol. Che,n. 261, 7621-7625 2. Noguchi, T., Inoue, M., and Tanaka, T. (1985) Transcriptional and post-transcriptional regulation of L-type pyruvate kinase in diabetic rat liver by insulin and dietary fructose. j Biol. C/tern. 260, 14393-14397 3. Decaux, J. F., Antoine, B., and Kahn, A. (1989) Regulation of the expression of the L-type pyruvase kinase gene in adult rat hepatocytes in primary culture. J. Biol. C/tern. 264, 11584-11590 4. Thompson, K. S., and Towle, H. C. (1991) Localization of the carbohydrate response element of the rat L-type pyruvate kinase gene. j Biol. C/tern. 266, 8679-8682 5. Cuif, M. H., Cognet, M., Boquet, D., Trenip, G., Kahn, A., and Vaulont, S. (1992) Elements responsible for hormonal control and tissue specificity of L-type pyruvate kinase gene expression in transgenic mice. Mo!. Cell. Biol. 12, 4852-4861 6. Cognet, M., Bergot, M. 0., and Kahn, A. (1991) Cis-acting DNA ele-
ments regulating
expression of the liver pyruvate kinase gene in hepato-
cytes and hepatoma cells. j Biol. C/tern. 266, 7368-7375 7. Vaulont, S., Puzenat, N., Levrat, F., Cognes, M., Kahn, A., and Raymondjean, M. (1989) Analysis by cell.free transcription of the liver specific pyruvate kinase gene promoter. J. Mol. Biol. 209, 205-219
8. Yamada, K., Noguchi, T., Matsuda, T., Takenaka, M., Monaci, P., Nicosia, A., and Tanaka, T. (1990) Identification and characterization of hepatocyte-specific regulatory regions of the rat pyruvate kinase L gene. J. Biol. Chern. 265, 19885-19891 9. Bergot, M. 0., Diaz.Guerra M., M. J., Puzenat, N., Raymondjean, M., and Kahn, A. (1992) Cis- regulation of the L-type pyruvate kinase gene promoter by glucose, insulin and cyclic AMP. Nucleic Acids Ret. 20, 1871-1878
10. Meienhofer, M. C., de Medicis, E., Cognet, M., and Kahn, A. (1987) Regulation of genes for glycolytic enzymes in cultured rat hepatoma cell lines.
II.
Eur. j
Bloc/tern.
169,
237-243
Antoine, B., Levrat, F., Vallet, V., Berbar, T., Cartier, N., and Kahn, A. (1992) Gene expression in hepatocyte-like cell lines established by targeted carcinogenesis in transgenic mice. Exp. Cell Res. 200, 175-185
12. Levrat,
F., Vallet,
V., Berbar,
T., Miquerol
MODEL
L., Kahn,
14. Munnich, A., Lyonnet, S., Chauvet, D., Van Shaftingen, E., and Kahn, A. (1987) Differential effects of glucose and fructose on liver-L.type pyruvate kinase gene expression in vivo.j Biol. C/tern. 262, 17065-17071 15. Hayzer, D. J., and lynedjian, P. B. (1990) Alternative splicing of glucokinase mRNA in rat liver. Bloc/tern. j 270, 261-263 16. Tremp, G. L., Boquet, D., Ripoche, M. A., Cognet, M., Lone, Y. C., Jami, J., Kahn, A., and Daegelen, D. (1989) Expression of the rat Ltype pyruvate kinase gene from its dual erythroidand liver specific promoter in transgenic mice. J. Biol. C/tern. 264, 19904-19910 17. Weinhouse, 5. (1976) In Current Topics in Cellular Regulation (Horecker, B. L., Stadtman, E. R., eds) pp. 1-49, Academic, New York 18. Magnuson, M. A. (1992) Tissue-specific regulation of glucokinase gene expression. j Cell Biochern. 48, 115-121 19. lynedjian, P. B., Jotterand, D., Nouspikel, T., Asfari, M., and Pilot, P. R. (1989) Transcriptional induction of glucokinase gene by insulin in cultured liver cells and its repression by the glucagon-cAMP system.]. Biol. Che,n. 264, 21824-21829 20. Van Schaftingen, E. (1989) A protein from rat liver confers to glucokinase the property of being antagonistically regulated by fructose 6-phosphate and fructose 1-phosphate. Eur. j Biochern. 179, 178-184 21. Nouspikel, T., and lynedjian, P. B. (1992) Insulin signaling and regulation of the glucokinase gene expression in cultured hepatocytes. Eur. J. Bioche,n. 210, 365-373 22. Schapira, F. (1973) Isozymes and cancer. Advances in Cancer Research, Vol. 18, pp. 77-153, Academic, New York 23. Foufelle, F., Gouhot, B., Pegorier, J. P., Perdereau, D., Girard, J., and Ferre P. (1992) Glucose stimulation of lipogenic enzyme gene expression in cultured white adipose tissue.] Biol. C/tern. 267, 20543-20546 24. Jenkins, A. B., Furler, S. M., and Kraegen, E. W. (1986) 2-Deoxy-6-glucose metabolism in individual tissues of the rat in vivo. mt. J. Biochern. 18, 311-318 25. Marie, S., Diaz-guerra, M. J., Miquerol, L., Kahn, A., and lynedjian, P. B. (1993) The pyruvate kinase gene as a model for studies of glucosedependent regulation of gene expression in the pancreatic a-cell-type. j Biol. C/tern. In press 26. Matsuda, T, Noguchi, T., Yamada, K., Takenaka, M., and Tanaka, T. (1990) Regulation of the gene expression of glucokinase and L-type pyruvate kinase in primary cultures of rat hepatocytes by hormones and carbohydrates. J. Bloc/tern. 108, 778-784 27. Cuif, M. H., Porteu, A., Kahn, A., and Vaulont, S. (1993) Exploration of a liver specific glucose/insulin-responsive promoter in transgenic mice. J. Biol. C/tern. 268, 13 769-13772 28. Liu, Z., Thomson, K. S., and Towle, H. (1993) Carbohydrate regulation of the rat L-pyruvate kinase gene requires two nuclear factors: LF-A1 and a member of the c-myc family. j Biol. C/tern. 268, 12787-12795 Received for publication Accepted for publication
A., and Antoine,
SYSTEMS
September 1, 1993. October 18, 1993.
FOR NEUROGENESIS
An F.! Theme Issue: July 1994 Coordinated
by R. P. Bunge
Research communications on this topic Deadline
96
Vol.
8
January
1994
for
submission
and P. Simpson
will also appear in the July issue.
of manuscripts
The
FASEB Journal
is March
1, 1994.
LEFRAN#{231}OIS-MARTINEZ
ET AL.
