Tài liệu Báo cáo khoa học: Glucose sensing in the intestinal epithelium pdf

Glucose sensing in the intestinal epithelium
Jane Dyer
1,
*, Steven Vayro
1,
*, Timothy P. King
2
and Soraya P. Shirazi-Beechey
1
1
Epithelial Function and Development Group, Department of Veterinary Preclinical Sciences, University of Liverpool, England, UK;
2
Rowett Research Institute, Aberdeen, Scotland, UK
Dietary sugars regulate expression of the intestinal Na
+
/
glucose cotransporter, SGLT1, in many species. Using sheep
intestine as a model, we showed that lumenal monosaccha-
rides, both metabolisable and nonmetabolisable, regulate
SGLT1 expression. This regulation occurs not only at the
level of transcription, but also at the post-transcriptional
level. Introduction of
D
-glucose and some
D
-glucose ana-
logues into ruminant sheep intestine resulted in > 50-fold
enhancement of SGLT1 expression. We aimed to determine
if transport of sugar into the enterocytes is required for
SGLT1 induction, and delineate the signal-transduction
pathways involved.

A membrane impermeable
D
-glucose analogue, di(glucos-
6-yl)poly(ethylene glycol) 600, was synthesized and infused
into the intestines of ruminant sheep. SGLT1 expression was
determined using transport studies, Northern and Western
blotting, and immunohistochemistry. An intestinal cell line,
STC-1, was used to investigate the signalling pathways.
Intestinal infusion with di(glucos-6-yl)poly(ethylene gly-
col) 600 led to induction of functional SGLT1, but the
compound did not inhibit Na
+
/glucose transport into
intestinal brush-border membrane vesicles. Studies using
cells showed that increased medium glucose up-regulated
SGLT1 abundance and SGLT1 promoter activity, and
increased intracellular cAMP levels. Glucose-induced acti-
vation of the SGLT1 promoter was mimicked by the protein
kinase A (PKA) agonist, 8Br-cAMP, and was inhibited
by H-89, a PKA inhibitor. Pertussis toxin, a G-protein
(G
i
)-specific inhibitor, enhanced SGLT1 protein abundance
to levels observed in response to glucose or 8Br-cAMP.
We conclude that lumenal glucose is sensed by a glucose
sensor, distinct from SGLT1, residing on the external face of
the lumenal membrane. The glucose sensor initiates a sig-
nalling pathway, involving a G-protein-coupled receptor
linked to a cAMP–PKA pathway resulting in enhancement
of SGLT1 expression.

Keywords: intestine; Na
+
/glucose cotransport; nutrient
transport; sugar sensing.
The dietary monosaccharides,
D
-glucose and
D
-galactose,
are transported across the brush-border membrane of
intestinal absorptive cells (enterocytes) by the Na
+
/glucose
cotransporter, SGLT1. It has been demonstrated that
lumenal glucose enhances the number of functional SGLT1
molecules in the intestinal brush-border membrane, and
that the metabolism of glucose is not required for the
induction [1–5].
We have used sheep intestine, which is an excellent model
system, for the study of monosaccharide regulation of
intestinal sugar transport [3,6]. We have shown that dietary
monosaccharides regulate the expression of intestinal brush-
border membrane Na
+
/glucose cotransporter at both the
transcriptional and post-transcriptional levels [3,7,8].
In preruminant lambs (birth to 3 weeks), milk sugar lactose
is hydrolysed by the intestinal lactase into
D
-glucose and

D
-galactose, and these sugars are transported by SGLT1.
Lambs are normally weaned at 3–10 weeks of age and, as
the diet changes from milk to grass, the rumen develops.
Dietary carbohydrates are fermented by rumen microflora
to short chain fatty acids, and under these conditions
negligible levels of monosaccharides reach the small intes-
tine [9,10]. Associated with the decline in lumenal sugars,
there is a decrease of over 50-fold in the levels of SGLT1
protein and mRNA [8]. Introduction of either
D
-glucose
or nonmetabolisable analogues of
D
-glucose, via duodenal
cannulae, into the intestinal lumenal contents of ruminant
sheep enhances the levels of functional SGLT1 protein and
mRNA to those detected in the preruminant state [4,8,11].
Intestinal infusions of
D
-glucose induced SGLT1 expression
in the brush-border membrane of enterocytes just below the
crypt–villus junction, with SGLT1 expression spreading to
the villus tip, with cell migration along the crypt–villus axis
[4,12]. We cloned and characterized the ovine SGLT1
promoter [13], and using intestinal STC-1 cells as a suitable
in vitro model [8], we identified (a) the basal SGLT1
promoter, (b) a glucose-responsive element within the
promoter, and (c) a sugar-induced transcription factor
involved in the transcriptional regulation of SGLT1 [8].

In this study, we set out to assess if the transport of sugar
across the brush-border membrane into the enterocyte is
required for enhancement in the expression of intestinal
Correspondence to S. P. Shirazi-Beechey, Epithelial Function and
Development Group, Department of Veterinary Preclinical Sciences,
University of Liverpool, Brownlow Hill, Liverpool L69 7ZJ, UK.
Fax: + 44 (0) 151 794 4244, Tel.: + 44 (0) 151 794 4255,
E-mail:
Abbreviations: GPCR, G-protein coupled receptor; H-89, N-[2-
(p-bromocinnamylamino)-ethyl]-5-isoquinolinesulfonamide; PKA,
protein kinase A; SGLT1, Na
+
/glucose cotransporter; BBMV,
brush-border membrane vesicle; PEG, poly(ethylene glycol).
*Note: These two authors contributed equally to this work.
Note: A web site is available at />(Received 11 April 2003, accepted 16 June 2003)
Eur. J. Biochem. 270, 3377–3388 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03721.x
SGLT1. To this end, we synthesized a membrane imper-
meable glucose analogue, di(glucos-6-yl)poly(ethylene gly-
col) 600 [di(glucos-6-yl)PEG
600
]. Introduction of this
compound into the lumenal content of the ruminant sheep
led to an increase in the expression of intestinal
SGLT1. This glucose analogue did not, however, inhibit
Na
+
-dependent glucose transport activity into ovine brush-
border membrane vesicles. We conclude that the monosac-
charide in the lumen of the intestine is sensed by a sugar

sensor, which is located on the lumenal surface of the
intestinal epithelial cell membrane, and is distinct from
SGLT1. To delineate the signal-transduction pathway by
which the glucose sensor might operate, we investigated the
role of some modulators of cAMP levels in induction of
SGLT1 protein expression and SGLT1 promoter activity.
Using STC-1 cells, we report that 8-bromo-cAMP
(8Br-cAMP), a protein kinase A (PKA) agonist, mimi-
cked the glucose-induced activation of the SGLT1
promoter. 8Br-cAMP also increased the levels of SGLT1
expressed endogenously in the cell line. The glucose-
induced SGLT1 promoter activity was inhibited, in a
dose-dependent manner, by the PKA antagonist H-89.
There was a 47% increase in the level of intracellular
cAMP when cells were exposed to increased medium
D
-glucose concentration; this paralleled the enhancement
in SGLT1 abundance. The potential role of G-proteins in
the pathway was investigated. Addition of pertussis toxin,
a G-protein (G
i
)-specific inhibitor, to the intestinal cell
line grown in low-glucose conditions enhanced the
SGLT1 abundance to that observed with high-glucose
or with 8Br-cAMP.
We propose that the intestinal epithelial cells have a
glucose sensor that resides on the external face of the
lumenal membrane. Glucose binds to the sensor and
generates an intracellular signal leading to enhancement of
the expression of SGLT1. It is evident that the generated

signal is independent of glucose metabolism and appears to
work via a G-protein-coupled receptor and cAMP/PKA
signalling cascade.
Materials and methods
Synthesis and characterization of di(glucos-6-yl)PEG
600
Di(glucos-6-yl)PEG
600
was synthesized by the route shown
in Fig. 1.
Synthesis of dibromoPEG. Triphenylphosphine, final
concentration 2
M
, was added to a magnetically stirred
solution of 0.67
M
PEG
600
and 1.67
M
tetrabromometh-
ane in 15 mL dry dichloromethane at 40 °C. The
reaction mixture was refluxed, in the dark, for 4 days,
by which time the reaction was complete, as indicated by
TLC using fluorescent aluminium-backed silica plates
(Merck type 5556) using methyl ethyl ketone/methanol/
water/ 27% (w/w) concentrated ammonia (65 : 20 : 5 :
10, v/v/v/v) as irrigant. TLC plates were developed
initially using iodine vapour and, after evaporation of the
iodine, visualization was with methyl red spray, which

gave a bright red colour with the product [14]. The
mixture was filtered and the filtrate washed three times
with deionized water (15 mL) to remove triphenylphos-
phine oxide. The organic layer was concentrated and the
residue swirled with 20 mL deionized water for 6 h. The
mixture was filtered, the residue washed with deionized
water (10 mL), and the aqueous portions freeze-dried,
redissolved in 5 mL water, and separated by gel filtration
on a Sephadex G15 gel column (75 · 2cm) and eluted
with deionized water. An initial 25 mL was collected, and
then aliquots of 5 mL were taken. The halogenated PEG
(Fig. 1) was isolated from fractions 1–5 as a clear,
viscous, oil, which, when freeze-dried, gave an azure blue
followed by a green colour in a flame test [15].
Reaction of methyl-a,
D
-glucopyranoside with halogenated
PEG in aqueous KOH. Methyl-a,
D
-glucopyranoside was
added to a mixture of 2
M
KOH in 2 mL dimethyl sulfoxide
to a final concentration of 1
M
, followed immediately by
an equimolar amount of the halogenated PEG derivative
(Fig. 1). The reaction mixture was stirred in the dark for
 4 days until there was no remaining starting material, as
assessed by TLC using a p-anisaldehyde spray. The sugar

derivatives gave blue spots on a pink background. The
reaction mixture was then purified by gel filtration, as
described above, and the pure product was isolated from
fractions 1–3.
Oxidation of the product with periodate indicated the
presence of two 6-O-glucosyl units per unit of PEG. Infrared
spectroscopy and
1
H NMR confirmed the presence of a
methylglucose unit on either end of the PEG
600
backbone.
Hydrolysis of di(methylglucos-6-yl)PEG
600
using H
2
SO
4
.
To remove the methyl groups di(methylglucos-6-yl) PEG
600
was dissolved in 10 mL 0.5
M
H
2
SO
4
to a concentration of
0.5 m
M

and heated under reflux to produce the target
compound. At all stages of the synthesis, reaction products
were analysed by TLC, infrared spectroscopy, and
1
H
NMR. The di(glucos-6-yl)PEG
600
was tested for any
potential free glucose using a commercial glucose testing
kit (Boehringer-Mannheim).
Fig. 1. Synthesis of di(glucos-6-yl)PEG
600
.
D
-Glucose was linked by
ether bonds to PEG
600
by the synthesis pathway outlined.
3378 J. Dyer et al.(Eur. J. Biochem. 270) Ó FEBS 2003
Biological stability of di(glucos-6-yl)PEG
600
. To deter-
mine if di(glucos-6-yl)PEG
600
was stable, and resistant to
hydrolysis when introduced into the intestinal contents,
the following experiments were undertaken. Di(glucos-6-
yl)PEG
600
was mixed with 10 mL ovine intestinal digesta to

a final concentration of 30 m
M
and incubated at 39 °C
(sheep body temperature). Samples were removed at
intervals of time up to 24 h and assayed for free glucose
using a commercial kit (Boehringer-Mannheim), according
to the manufacturer’s instructions. In addition, ovine
intestinal crude cellular homogenate or purified brush-
border membrane vesicles (1 mg protein) were incubated at
39 °C in 0.1 mL of a solution containing 300 m
M
mannitol,
20 m
M
Hepes/Tris, pH 7.4, 0.1 m
M
MgSO
4
and 30 m
M
di(glucos-6-yl)PEG
600
.Samples(10lL) were removed at
1 h intervals for up to 8 h and assayed for glucose using a
glucose assay kit (Boehringer-Mannheim) as above.
Animals and intestinal infusions
Scottish Blackface ewes, all > 1-year-old, were used.
Animals were maintained on a conventional roughage diet,
and fed grass pellets (1 kg a day) throughout the experi-
ment, as described previously [6]. Perspex T-shaped cannu-

lae were fitted into the duodenum 6 cm distal to the pylorus
[6]. Animals were infused, through the duodenal cannulae,
for 3 h with 30 m
M
solutions of
D
-glucose, PEG
600
,or
di(glucos-6-yl)PEG
600
at a rate of 62.5 mLÆh
)1
, as described
[6,12]. They were killed with sodium pentobarbitone
(Euthatal) [4,12], and sections of intestine were removed,
flushed with ice-cold 0.9% (w/v) NaCl, and everted.
Intestinal sections were rinsed clean, blotted with paper
towels to remove mucous, and then wrapped in aluminium
foil before immediate freezing in liquid nitrogen. Additional
samples were frozen in isopentane cooled in liquid nitrogen
for immunohistochemical studies. Tissue was subsequently
stored at )80 °C until use.
All procedures were carried out under an approved UK
Home Office project licence.
Cell culture
Intestinal cells, STC-1 [8,16] (passages 40–90) were grown in
Dulbeccos’ modified Eagle’s medium (Invitrogen) supple-
mented with 10% (v/v) fetal bovine serum or 10% (v/v)
dialysed fetal bovine serum (containing < 200 l

M
D
-glucose), 50 UÆmL
)1
antibiotic solution containing peni-
cillin and streptomycin, and either 25 m
M
(high) or 5 m
M
(low)
D
-glucose, as described [8]. Cells were maintained at all
times at 37 °Cin5%CO
2
. Stock cultures were grown in
75-cm
2
flasks (Corning, High Wycombe, Bucks, UK) and
were fed every 3–4 days. Subsequently cells were washed
twice with 5 mL Hanks’ balanced salt solution and then
trypsinized (1 min at 37 °C, 5% CO
2
)in1mLsolution
containing Versene 1 : 5000 (Invitrogen) and 0.25% (w/v)
trypsin (> 225 UÆmg
)1
; Invitrogen). Culture medium
(10 mL) was added and the cells dispersed using a syringe
fitted with a Venflon 2 (Southern Syringe Services, Man-
chester, UK). The cells were seeded into 12-well plates

(22 mm; Corning) containing 2 mL of the medium at a
density of  0.5 · 10
6
cells, and returned to 37 °C, until
they were 60–70% confluent.
Preparation of brush-border membrane vesicles (BBMVs)
Brush-border membrane vesicles (BBMVs) were prepared
from frozen intestinal sections using a combination of
cation precipitation and differential centrifugation as des-
cribed previously [17]. The final purified BBMVs were
suspended in buffer containing 300 m
M
mannitol, 20 m
M
Hepes/Tris, pH 7.4, and 0.1 m
M
MgSO
4
,andstoredin
liquid nitrogen until use.
The protein concentration in the BBMVs was estimated
by its ability to bind Coomassie blue according to the Bio-
Rad assay technique. Bovine c-globulin was used as the
standard [18]. The plasma membrane origin of the BBMVs
was assessed by determination of the enrichment of the
activity and the abundance of the marker proteins of the
brush-border membrane. BBMV purity was determined by
assessing the levels of marker proteins characteristic of
basolateral and organelle membranes [6,19].
Measurement of monosaccharide transport activity

To assess the activity of SGLT1, the initial rate of 0.1 m
M
D
-glucose transport in BBMVs was measured at 39 °Cin
the presence of NaSCN and KSCN, using the rapid
filtration stop technique, as described before [17,19]. All
initial rate measurements were taken after a 3 s incubation
period, as transport was determined to be linear up to 4 s
[17]. Uptakes were measured in duplicate or triplicate.
To assess the activity of any potential facilitative glucose
transporter, the initial rate of uptake of 1 m
M
2-deoxy-
D
-
glucopyranoside, a specific substrate of Na
+
-independent
D
-glucose transporter isoforms, was determined at 39 °Cin
incubation medium consisting of 300 m
M
mannitol, 20 m
M
Hepes/Tris, pH 7.4, 0.1 m
M
MgSO
4
, and 0.02% (w/v)
NaN

3
in the presence and absence of 50 l
M
cytochalasin B,
as described [20].
Competition studies were carried out by determining the
initial rate of uptake of 0.1 m
MD
-glucose in the presence of
1m
M
competitor, using a standard technique as described
previously [21].
Immunodetection of SGLT1
Quantitative Western blotting. The abundance of SGLT1
protein was measured by quantitative Western blotting as
described previously [22]. The BBMV protein contents were
separated on an 8% polyacrylamide gel containing 0.1%
(w/v) SDS and were electrotransferred to nitrocellulose
membrane (TransBlot; Bio-Rad).
A standard calibration curve was constructed by slot-
blotting the synthetic peptide (amino acids 402–420 of the
ovine SGLT1 sequence, to which the antibody was raised)
on to nitrocellulose membrane, and this was probed
concurrently with the BBMV samples. The specific immu-
noreactive band was blocked when antibodies were pre-
incubated with the immunizing peptide. The membranes
were developed using the ECL system (Amersham-
Pharmacia, Little Chalfont, Bucks., UK), and exposed to
film (XOMAT-LS; Kodak).

The intensity of the immunoreactive bands detected in the
BBMVs and the peptide standard samples was quantified
using scanning densitometry (Phoretix 1D; Non-linear
Ó FEBS 2003 Intestinal glucose sensing (Eur. J. Biochem. 270) 3379
Dynamics Ltd, Newcastle upon Tyne, Tyne and Wear, UK),
and the abundance of SGLT1 protein per mg of BBMV
protein was calculated from the peptide standard curve.
Immunodetection of SGLT1 in STC-1 cells was carried
out as described previously [8]. Cells were washed with ice-
cold NaCl/P
i
and then lysed in 300 lL buffer containing
150 m
M
NaCl, 1% (w/v) SDS, 10 m
M
EDTA and 10 m
M
Hepes/Tris, pH 7.4, with protease inhibitor cocktail (Boeh-
ringer-Mannheim) and 0.2 m
M
phenylmethanesulfonyl
fluoride. Cells were scraped from the dish with a Ôrubber
policemanÕ and homogenized by 10 passages through a
syringe fitted with a 21-gauge needle. Protein (15 lgper
lane) was separated on 8% (w/v) polyacrylamide gels
containing 0.1% (w/v) SDS. After electrotransfer to
poly(vinylidene difluoride) (0.2 lm), the membrane was
blocked for 30 min in buffer containing 150 m
M

NaCl,
10 m
M
Tris/HCl, pH 7.4, 0.05% (v/v) Tween 20, 0.5% (w/v)
skimmed milk powder. Primary and secondary antibody
incubations were 1 h at room temperature; subsequently
membranes were washed three times with buffer for 10 min.
Detection and quantification were as described above.
Immunohistochemistry. Tissue sections (5 lm thick) were
cut using a cryostat at )20 °C, air-dried on to gelatin-coated
microscope slides, and fixed in methanol for 15 min at
)15 °C. Sections were washed in NaCl/P
i
/Tween (10 m
M
phosphate buffer, 150 m
M
NaCl, 0.05% Tween 20, pH 7.4)
and incubated for 30 min at 37 °CinNaCl/P
i
SuperBlock
(Pierce and Warriner, Chester, UK). Sections were washed
in six changes of NaCl/P
i
/Tween over 20 min at room
temperature and incubated for 60 min at 37 °Cin
5 lgÆmL
)1
SGLT1 antiserum in NaCl/P
i

/Tween containing
0.1% acetylated BSA (Aurion, Wageningen, the Nether-
lands). Sections were washed in six changes of NaCl/P
i
/
Tween over 20 min and incubated for 30 min at 37 °Cwith
a mouse monoclonal anti-rabbit IgG (clone RG-96; Sigma)
at a dilution of 1 : 400 in NaCl/P
i
/Tween containing 0.1%
acetylated BSA. Sections were washed in six changes of
NaCl/P
i
/Tween over 20 min and incubated for 30 min at
37 °Cin5lgÆmL
)1
Oregon Green goat anti-mouse IgG
(Molecular Probes, Cambridge Bioscience, Cambridge,
UK) in NaCl/P
i
/Tween containing 0.1% acetylated BSA.
Sections were washed in six changes of NaCl/P
i
/Tween, and
then mounted in Vectorshield antifading mountant (Vector
Laboratories) before examination by incident light fluores-
cence microscopy on a Zeiss Axioscope microscope.
Control sections were subjected to the same protocol except
that 0.1 lgÆmL
)1

peptide (amino acids 402–420 of the ovine
SGLT1) was added to the SGLT1 antibody, and the
peptide/antibody mixture was preincubated for 60 min at
37 °Cbeforeuse.
Immunodetection of G-proteins
The presence of G-proteins (G
a
subunits) on the intestinal
brush-border membrane of preruminant lambs, ruminant
sheep and age-matched glucose-infused ruminant sheep, as
well as in STC-1 cells, was determined by Western
blotting with a broad-range affinity-purified rabbit G
a
polyclonal antibody raised against the conserved GTP-
binding domain (Sigma) at 1 : 1000 dilution, as described
above for SGLT1.
Analysis of SGLT1 promoter function
The ovine SGLT1 promoter fragment used in these studies
was generated, and assayed, as described previously [8,13].
STC-1 cells were seeded into 12-well plates (0.5 · 10
6
cells)
containing 1 mL medium, and incubated for 24 h at 37 °C,
5% CO
2
. Cells were then transiently transfected using the
cationic lipid reagent Transfast (Promega) at a DNA/lipid
ratioof1 :1.Asecondplasmid,pRL-SV40(Promega),was
cotransfected (0.019 pmol) as an internal control. The cells
were incubated for 1 h at 37 °C, 5% CO

2
,andthen1mL
complete medium was added. After a further 48 h, the cells
were recovered and assayed for luciferase activity using the
Dual-Luciferase Reporter assay system (Promega) on a
Lumat LB9501 luminometer (Perkin–Elmer). Values are
presented as a ratio of the firefly luciferase to Renilla
luciferase activity.
Intracellular cAMP determination
The cAMP levels in cellular homogenates were measured
using a commercially available RIA kit (Amersham-Phar-
macia), according to the manufacturer’s instructions. STC-1
cells (1 · 10
6
per well) were harvested and homogenized,
using a Polytron at setting 5 for 30 s, in 0.4 mL buffer
containing 100 m
M
mannitol, 2 m
M
Hepes/Tris, pH 7.4,
5m
M
EDTA, 0.2 m
M
phenylmethanesulfonyl fluoride,
protease inhibitor cocktail (Boehringer-Mannheim) and
1m
M
3-isobutyl-1-methylxanthine, a phosphodiesterase

inhibitor. The homogenates were then deproteinated by
heating in a boiling water bath for 10 min, followed by
centrifugation at 15 000 g for 20 min at 4 °C to sediment
denatured proteins. Supernatants were transferred to fresh
1.5 mL tubes, placed on ice, and assayed for cAMP
following the instructions of the manufacturer. All proce-
dures were carried out at 4 °C.
Statistical analysis
Data are expressed as mean ± SEM. Statistical compari-
sons are made using Student’s t test, and results are
considered significant if P <0.05.
Results
Synthesis of di(glucos-6-yl)PEG
600
In this molecule we chose to link the glucose and PEG by
ether bonds, which made it improbable that enzymatic
hydrolysis would occur in vivo. On the basis of our previous
investigations on the stereoselectivity of the glucose sensor
[4,12], we chose the ether linkage to O6 of glucose; being a
primary alcohol, the 6-OH is the most reactive once the
anomeric position is blocked.
Bromination of primary alcohols using tetrabromometh-
ane/triphenylphosphine usually gives good yields of alkyl
bromides [23], but with PEG
600
as the alcohol, bromination
was accompanied by chlorination. However, the dihalogen-
ated products were all eluted in the same fraction from a
Sephadex G-15 column, and the mixture reacted fully with
methyl-a,

D
-glucopyranoside in dimethyl sulfoxide to give
good yields of the glucoside, provided that KOH was used
3380 J. Dyer et al.(Eur. J. Biochem. 270) Ó FEBS 2003
as base. The presence of two 6-O-glucosyl units per unit of
PEG was demonstrated by oxidation with periodate;
 4 mol was consumed per mol di(methylglucosyl)PEG
600
,
with production of formic acid. Hydrolysis of the glucoside
with 0.5
M
H
2
SO
4
was used to remove the methyl groups in
quantitative yield, giving the target compound (Fig. 1).
Stability of di(glucos-6-yl)PEG
600
The synthesized di(glucos-6-yl)PEG
600
was assayed for any
potential free glucose using a commercial kit (Boehringer-
Mannheim). The results indicated total absence of free
glucose, and this was confirmed by TLC. To determine
biological stability, di(glucos-6-yl)PEG
600
(30 m
M

)was
incubated with ovine intestinal digesta (10 mL) at 39 °Cfor
24 h, and fractions were removed periodically and assayed
for free glucose. Glucose was not detected in the samples up
to 8 h, indicating that the compound was not hydrolysed and
would remain intact over the infusion period. Similarly
glucose was not detected when di(glucos-6-yl)PEG
600
was
incubated with either the ovine intestinal mucosal homogen-
ate or purified BBMVs. Results indicate that free glucose
would not be present during the infusion period, or during the
passage of the compound through the small intestine. This
eliminated the possibility that the induction of SGLT1
expression may be due to the glucose released in the small
intestine as a result of the breakdown of the compound.
Induction of SGLT1 by intestinal infusion
Western blot analysis. The protein component of BBMVs
isolated from the jejunum of control ruminant sheep and
ruminant sheep the intestines of which were infused with
PEG
600
,
D
-glucose or di(glucos-6-yl)PEG
600
were separated
by SDS/PAGE and electrotransferred to nitrocellulose.
Samples were immunoblotted to determine the presence of
SGLT1 protein using an affinity-purified polyclonal peptide

antibody, as described previously [22]. The results are
presented in Fig. 2. The antibody recognizes a single protein
with an apparent molecular mass of 75 kDa in the BBMVs
isolated from the
D
-glucose and di(glucos-6-yl)PEG
600
-
infused animals, but not the PEG
600
-infused animals or
controls, indicating that the presence of
D
-glucose or the
nontransportable, membrane-impermeable, di(glucos-6-yl)-
PEG
600
in the lumen of the intestine induces expression of
SGLT1. The abundance of SGLT1 protein in the BBMVs
isolated from the intestine of sheep infused with
D
-glucose
or with di(glucos-6-yl)PEG
600
(Fig. 2) are 12.6 ± 1.3 and
13.1 ± 1.2 pmolÆ(mg protein)
)1
, respectively.
Transport of
D

-glucose. To determine if the sugar-induced
SGLT1 is capable of transport, the ability of the BBMVs to
transport glucose in a Na
+
-dependent manner was assessed.
The initial rates of 0.1 m
MD
-glucose uptake into BBMVs
isolated from the intestine of control and infused ruminant
sheep are presented in Fig. 3. The initial rates of Na
+
-
dependent
D
-glucose transport were 108.8 ± 10.5 and
111 ± 6.4 pmolÆs
)1
Æ(mg protein)
)1
in BBMVs isolated
from the jejuna for
D
-glucose and di(glucos-6-yl)PEG
600
,
respectively. The initial rate of uptake in vesicles isolated
from PEG
600
-infused sheep was 3.6 ± 1.2 pmolÆs
)1

Æ(mg
protein)
)1
, a rate identical with that measured in adult
ruminant control BBMVs [11,22]. The results indicate that
the SGLT1 protein that is expressed in response to lumenal
infusion is functional. There was no cytochalasin B-sensitive
2-deoxy-
D
-glucose transport detected in any of the BBMV
samples (data not shown), indicating the absence of GLUT2
from the BBMVs and therefore any basolateral membrane
contamination.
To investigate any interaction between the di(glucos-6-
yl)PEG
600
and SGLT1 function, the ability of this com-
pound to inhibit Na
+
-dependent
D
-glucose transport into
BBMVs was investigated. Concurrently the effect on
SGLT1 activity of other glucose analogues was also
determined. The results are presented in Fig. 4. The initial
rate of uptake of 0.1 m
MD
-glucose into lamb jejunal
BBMVs was reduced in the presence of 1 m
M

concentra-
tions of inducers of SGLT1 expression such as
D
-glucose,
Fig. 2. Abundance of SGLT1 protein in the intestinal brush-border
membrane of ruminant sheep after intestinal infusion with various solutes.
Ruminant sheep (3 years old) had their intestines infused with 30 m
M
solutions of PEG
600
,
D
-glucose, or di(glucos-6-yl)PEG
600
through
duodenal cannulae. BBMVs were prepared from the intestine of these
animals, and brush-border membrane proteins (20 lg per lane) were
separated on 8% polyacrylamide gels containing 0.1% SDS. Separated
proteins were electrotransferred to nitrocellulose membranes and
blotted for the presence of SGLT1, as described previously [8,22]. The
abundance of SGLT1 protein in the brush-border membrane samples
was quantified using the peptide antigen as a standard [22]. N.D., Not
detected.
Fig. 3. Initial rate of Na
+
-dependent glucose uptake in ovine intestine
BBMVs after intestinal infusion. The initial (3 s) rate of 0.1 m
M
D
-glucose uptake into BBMVs (0.1 mg protein) was measured at 39 °C

inthepresenceof100m
M
NaSCN, as described in Materials and
methods. Results are presented as mean ± SEM (n ¼ 3).
Ó FEBS 2003 Intestinal glucose sensing (Eur. J. Biochem. 270) 3381
D
-galactose (natural SGLT1 substrates), a-methylglucose or
3-O-methylglucose (nonmetabolisable SGLT1 substrates).
However,
D
-glucose uptake was unchanged when
D
-fructose
or di(glucos-6-yl)PEG
600
(not transported by SGLT1 but
induce SGLT1 expression),
L
-glucose or PEG
600
(not
transported by SGLT1 and do not induce SGLT1 expres-
sion) were included in the incubation medium. The results
suggest that there is no interaction between di(glucos-
6-yl)PEG
600
and SGLT1 protein that affects SGLT1
function.
Immunofluorescence localization of SGLT1. The distribu-
tion of SGLT1 protein along the crypt–villus axes of the

intestine of the infused ruminant sheep was also determined
by immunohistochemistry. Typical results, presented in
Fig. 5, indicate that infusion of the intestine of ruminant
sheep with either
D
-glucose or di(glucos-6-yl)PEG
600
results
in induction of SGLT1 protein expression. Infusion of the
intestine with PEG
600
has no effect (Fig. 5A). Immuno-
fluorescence localization of SGLT1 protein along the crypt–
villus axes of the ruminant sheep, with the intestinal infusion
of either di(glucos-6-yl)PEG
600
or
D
-glucose, shows labelling
on the entire brush-border surface, including the lower region
of the villus (Fig. 5B,C). The distribution of SGLT1 protein
in infused ruminant sheep (Fig. 5B,C)is similar to that seen in
the intestine of thepreruminant lamb (Fig. 5D). The labelling
is specific, as it was blocked when the primary antibody was
preincubated with the peptide antigen (Fig. 5E).
Effect of cAMP and PKA on ovine SGLT1 promoter
activity and SGLT1 expression
We have demonstrated that di(glucos-6-yl)PEG
600
, a mem-

brane-impermeable glucose analogue, enhances the level of
SGLT1. Furthermore, we have determined that SGLT1
function is not inhibited in the presence of
D
-fructose,
2-deoxy-
D
-glucose and di(glucos-6-yl)PEG
600
, compounds
that induce the expression of functional SGLT1. We
conclude therefore that a glucose sensor, with different
sugar specificity from SGLT1, is located on the external face
of the intestinal brush-border membrane. The sensor would
detect changes in the lumenal sugar concentration and
initiate signalling pathways, leading to modulations in the
expression of functional SGLT1. Using the intestinal cell
line, STC-1, as an in vitro expression system [8], we assessed
the potential role of cAMP/PKA in the transcriptional
regulation of the )66/+21-bp SGLT1 glucose-responsive
promoter fragment [8]. To this end, we used (a) 8Br-cAMP,
a membrane-permeable cAMP analogue and a PKA
agonist, and (b) H-89, a PKA antagonist.
Cells were cultured in medium containing 5 m
M
glucose,
transfected with the ovine SGLT1 promoter fragment, and
then either maintained in the same medium or transferred to
one containing 25 m
M

glucose, in the presence or absence of
(a) 0.5 m
M
8Br-cAMP or (b) H-89. The results are shown in
Fig. 6 and Fig. 7, respectively. Figure 6 shows that SGLT1
promoter activity increased twofold, after the addition of
25 m
M
glucose, in agreement with our previous data [8].
Cells maintained in low-glucose (5 m
M
) medium, but
treated with 0.5 m
M
8Br-cAMP, also showed a significant
increase in promoter activity compared with controls. 8Br-
cAMP also augmented the increase in promoter activity
observed in response to high-glucose medium by a further
30%. When the reporter gene construct was placed in the
reverse orientation, neither glucose nor 8Br-cAMP had any
effect. We conclude that an increase in intracellular cAMP
results in the activation of SGLT1 promoter function.
Glucose-induced SGLT1 promoter activity was inhibited,
in a dose-dependent manner, in response to increasing
concentrations of H-89 (0.1, 0.5, 1.0 l
M
; Fig. 7). In cells
switched to 25 m
M
glucose, in the presence of 1 l

M
H-89,
SGLT1 promoter activity was reduced to the level detected
in cells maintained in 5 m
M
glucose. Promoter function was
not inhibited by H-89 in cells maintained in low glucose, or
in cells transfected with the reporter gene construct in the
reverse orientation (not shown). Therefore, the inhibitory
action of H-89 appears to be specific to the glucose-induced
SGLT1 promoter activity. These data suggest that PKA has
an important role in the transcriptional activation of the
ovine SGLT1 promoter.
We also examined the effects of both 8Br-cAMP and
H-89 on the level of endogenous SGLT1 protein expressed
in the STC-1 cells (Fig. 8A,B). Cells transferred from 5 m
M
to 25 m
M
glucose showed a (2.88 ± 0.22)-fold increase in
the abundance of SGLT1 (Fig. 8A, lanes 1 and 3, and
Fig. 8B, lanes 1 and 2), consistent with our previous findings
[8]. In cells cultured in 5 m
M
glucose in the presence of
0.5 m
M
8Br-cAMP there was a (4.49 ± 0.56)-fold enhance-
ment in SGLT1 abundance, compared with controls
(Fig. 8A, lanes 1 and 2). Cells switched to medium

containing 25 m
M
glucose and 8Br-cAMP showed a further
(1.81 ± 0.71)-fold enhancement in SGLT1 protein abun-
dance (Fig. 8A, lane 4). Cells maintained throughout in
25 m
M
glucose medium did not respond to 8Br-cAMP (not
shown). 8-Br-cAMP had no effect on the levels of b-actin
(Fig. 8A). Treatment of STC-1 cells with 0.1 l
M
H-89
resulted in 34.1 ± 3.3% reduction in glucose-induced
SGLT1 protein abundance (Fig. 8B, lane 3), and increasing
the concentration of H-89 to 1 l
M
had no further effect
(Fig. 8B, lanes 4 and 5), suggesting a role for PKA in the
regulatory process.
Interestingly, cAMP concentrations measured in depro-
teinated homogenates of cells cultured in 5 m
MD
-glucose
were 47% lower than cAMP levels detected in cells
transferred to 25 m
M
glucose [0.19 ± 0.04 vs. 0.28 ±
0.03 pmolÆ(mg protein)
)1
;mean±SEM,n ¼ 3], implying

Fig. 4. Competition studies. The initial rate of the Na
+
-dependent
uptake of 0.1 m
MD
-glucose into preruminant lamb jejunal BBMVs
was measured at 39 °C in the presence of of the indicated competitor
(1 m
M
). Results are expressed as percentage of control and are
means ± SEM (n ¼ 3).
3382 J. Dyer et al.(Eur. J. Biochem. 270) Ó FEBS 2003
Fig. 5. Immunofluorescence localization of SGLT1 protein along the crypt–villus axes of ruminant sheep after intestinal infusion with various solutes.
Typical immunofluorescence images are presented showing localization of SGLT1 protein on the jejunal villi of 3-year-old sheep after intestinal
infusion of 30 m
M
solutions of (A) PEG
600
, (B) di(glucos-6-yl)PEG
600
and (C)
D
-glucose. SGLT1 localization in preruminant lamb jejunum is also
shown (D), with signals blocked by preincubation of the antibody with the immunizing peptide antigen (E). Labelling over the entire villus (V)
brush-border surface is shown. Scale bar represents 100 lm.
Ó FEBS 2003 Intestinal glucose sensing (Eur. J. Biochem. 270) 3383
that there is an increase in intracellular cAMP levels in
response to increasing levels of medium
D
-glucose.

Involvement of a GPCR in intestinal glucose sensing
One mechanism for sensing glucose in yeast (Saccharomyces
cerevisiae) is through the plasma membrane GPRC, Gpr1,
which initiates a cAMP signal-transduction cascade cul-
minating in the expression of hexose transporter genes [24–
29]. To investigate the potential role of a G-protein in
glucose sensing, we used the G-protein (G
i
)-specific inhi-
bitor, pertussis toxin.
In STC-1 cells cultured in 5 m
M
glucose, the abundance
of SGLT1 protein increased in a dose-dependent manner
(1.83 ± 0.42-fold, 1.70 ± 0.20-fold and 2.47 ± 0.84-fold)
in response to increasing concentrations of pertussis toxin
(100, 250 and 500 ngÆmL
)1
;Fig.9lanes1,2,3and4).In
cells transferred from 5 m
M
to 25 m
M
glucose, SGLT1
expression was up-regulated twofold, as expected (lanes 1
and 5), and pertussis toxin had no effect on this response
(lanes 6, 7 and 8). Pertussis toxin had no effect on the levels
of b-actin.
The presence of G-proteins (G
a

subunits) on the intestinal
brush-border membrane of preruminant lambs, ruminant
sheep and age-matched glucose-infused ruminant sheep, as
well as in STC-1 cells, was determined by Western blotting
with a broad-range G
a
subunit antibody. The abundance of
SGLT1 and b-actin were also determined in the same
samples (Fig. 10A,B). G
a
subunits were detected, as an
immunoreactive band of  41 kDa, in the intestinal
brush-border membrane of lambs, ruminant sheep, and
Fig. 6. Effect of 8Br-cAMP on SGLT1 promoter activity. STC-1 cells
were cultured in medium containing 5 m
MD
-glucose and transfected
with the )66/+21-bp ovine SGLT1 promoter construct, as described
in the Methods section. After transfection, the cells were incubated in
medium containing either 5 m
M
or 25 m
MD
-glucose with (j)or
without (h)0.5m
M
8Br-cAMP, for a further 48 h, before assaying.
Values are the means ± SEM from four determinations.
Fig. 7. Effect of PKA inhibitor, H-89, on SGLT1 promoter activity.
STC-1 cells were treated as described in the legend to Fig. 6. After

transfection with the promoter construct, the cells were incubated in
medium containing either 5 m
M
or 25 m
MD
-glucose with increasing
concentrations of H-89 (0.1, 0.5, 1.0 m
M
), for a further 48 h, before
assay. Values are the means ± SEM from three to seven determina-
tions.
Fig. 8. Effect of 8Br-cAMP and H-89 on the levels of SGLT1 expressed
endogenously in STC-1 cells. STC-1 cells were cultured in 5 m
M
D
-glucose medium and then exposed to medium containing either
5m
M
or 25 m
MD
-glucose, in the presence of (A) 0.5 m
M
8-Br-cAMP,
or (B) increasing concentrations of H-89 (0.1, 0.5 or 1.0 l
M
)fora
further 48 h. Cell lysates were then prepared for Western blotting.
Equal amounts of protein (15 lg) were loaded per lane. Immuno-
detection of SGLT1 and b-actin were carried out using the SGLT1
antibody (1 : 5000 dilution) and a mouse monoclonal b-actin antibody

(1 : 10 000 dilution), respectively. Data shown are representative of
three experiments.
3384 J. Dyer et al.(Eur. J. Biochem. 270) Ó FEBS 2003
age-matched glucose-infused animals, as well as in lysates
from STC-1 cells cultured under low or high glucose
conditions. The abundance of SGLT1 protein in BBMVs
isolated from the intestinal tissues and the STC-1 cell lysates
was a good representation of the glucose induction of
SGLT1 [22]. b-Actin abundance was constant in all samples.
These observations confirm the presence of G
a
subunits in
the ovine intestinal lumenal membrane and support the
potential involvement of a GPCR in the signalling pathway
for glucose regulation of SGLT1 expression.
Discussion
Glucose is a major source of energy for most eukaryotic cells,
and has significant and varied effects on cell function.
Consequently maintenance of glucose homoeostasis is of
great importance to many organisms. Interest in identifying
mechanisms by which cells sense and respond to variations in
glucose concentration has increased recently, and promising
advances have been made [30]. It has been shown that
different eukaryotic cells use specific mechanisms to sense the
presence of glucose, and that the physiological roles of these
mechanisms are dependent on the particular cell type [27].
Yeast cells, S. cerevisiae, have a remarkable preference
for glucose as a carbon source [27] and have evolved
mechanisms for sensing and responding to wildly fluctu-
ating levels of extracellular glucose. These mechanisms

involve a large family of hexose transporters (HXT
proteins), and the glucose transporter homologues Snf3
and Rgt2. Snf3 and Rgt2 are plasma membrane glucose-
sensing proteins with no detectable transport activity. They
Fig. 9. Effect of pertussis toxin on SGLT1 protein abundance. STC-1
cells were treated as described in the legend to Fig. 8. They were then
transferred to medium containing either 5 m
M
or 25 m
MD
-glucose, in
thepresenceof100,250or500ngÆmL
)1
pertussis toxin for a further
48 h. Cell lysates were then prepared for Western blotting, and equal
amounts of protein (15 lg) were loaded per lane. Immunodetection of
SGLT1 and b-actin were carried out using the SGLT1 antibody
(1 : 5000 dilution) and a mouse monoclonal b-actin antibody
(1 : 10 000 dilution), respectively. Data shown are representative of
three experiments.
Fig. 10. Immunodetection of G
a
subunits in ovine BBMV and STC-1
cells. BBMVs were prepared from the jejunal mucosal scrapings of
preruminant lambs, 3-year-old-adult sheep, and age-matched sheep
after glucose infusion. STC-1 cells were cultured in the presence of
medium containing either 5 m
M
or 25 m
MD

-glucose for 48 h and then
cell lysates prepared. Equal amounts of protein (15 lg), from (A)
intestinal BBMVs or (B) STC-1 cell lysates were loaded per lane.
Immunodetection of SGLT1 was carried out using the SGLT1 anti-
body (1 : 5000 dilution). Immunodetection of G
a
subunits was per-
formed using an affinity-purified rabbit G
a
polyclonal antibody raised
against the conserved GTP-binding domain at 1 : 1000 dilution.
Ó FEBS 2003 Intestinal glucose sensing (Eur. J. Biochem. 270) 3385
sense the extracellular glucose and generate an intracellular
glucose signal that triggers the induction of HXT gene
expression [31]. Using Snf3/Rgt2 double mutants it was
shown that, in yeast, glucose-sensing and signalling are
receptor-mediated processes and are independent of glucose
metabolism [31]. In addition, a novel GPCR, Gpr1, which
senses external medium glucose, has also been identified.
Gpr1 acts via the G-protein, Gpa2, to initiate a cAMP/PKA
signalling cascade [25,26,29]. Gpr1 is activated by glucose
and transmits a signal, via Gpa2, to adenylate cyclase.
Glucose activation of cAMP synthesis requires active sugar
phosphorylation but no further metabolism of sugar. The
glucose-sensing Gpr1–Gpa2 system for activation of the
cAMP pathway in S. cerevisiae appears to be the first
example of a nutrient-sensing GPCR system. If nutrient
sensing GPCRs were common to eukaryotic cells, they
would provide a means of regulating major signal-trans-
duction pathways by the nutrient status of the cellular

environment. The latter is supported by a recent report
showing that a GPCR, GPR40, abundantly expressed in the
pancreas, functions as a receptor for long-chain free fatty
acids. The latter amplify glucose-stimulated insulin secretion
from pancreatic b-cells by activating GPR40 [32].
Intestinal epithelial cells are exposed from the lumenal
domain to an environment with continuous and massive
fluctuations in the level of dietary monosaccharides. This is
in contrast with other mammalian cells, which are exposed
to a relatively constant blood glucose concentration regu-
lated by endocrine hormones. Enterocytes therefore have to
sense and respond to the significant fluctuations in lumenal
sugars and regulate their function accordingly. Dietary
sugars have been shown to regulate the expression of the
intestinal lumenal membrane glucose transporter, the Na
+
/
glucose cotransporter (SGLT1), in a wide range of species
[1,2,4,5]. Using the sheep intestine as a model system, we
have shown that lumenal sugars regulate the expression of
SGLT1 at both transcriptional and post-transcriptional
level [3,8]. It was demonstrated, using nuclear run-on assays,
that the transcriptional activity of the ovine SGLT1 gene
increased 2–3-fold, in response to lumenal sugar. This
increase did not account entirely for the overall enhance-
ment in steady-state levels of SGLT1 mRNA determined by
Northern blot analysis [8,11].
Rumen development in sheep is a natural and efficient
way of ensuring a virtual block in the delivery of monosac-
charides to the small intestine. Nutrients, such as peptides,

amino acids and fats, enter the intestinal lumen, but
monosaccharides are selectively excluded [10]. Associated
with the decline in the levels of monosaccharides, there is a
> 50-fold decline in the levels of functional SGLT1 protein
and mRNA [8,11,22]. Introduction of monosaccharides,
D
-glucose,
D
-galactose, a-methyl-
D
-glucose, 3-O-methyl-
D
-
glucose,
D
-fructose and 2-deoxy-
D
-glucose into the lumenal
contents of ruminant sheep intestine, through duodenal
cannulae, resulted in increased expression of SGLT1 to the
levels detected in the preruminant lamb. We concluded that
induction of SGLT1 by lumenal sugar is independent of
glucose metabolism and that the inducing sugar need not be
a substrate of SGLT1 [4,6,22].
To determine if transport of sugar into the enterocyte is
required for SGLT1 induction, we set out to synthesize a
water-soluble, metabolically inert, membrane-impermeable
glucose analogue. Our overall objective was to join glucose
to a water-soluble polymer in such a way that the conjugate
would activate the glucose sensor, but would not liberate

free glucose by chemical or enzymatic reactions in the gut.
We decided to join the water-soluble polymer PEG
600
,
which is sufficiently large to be impermeable to the gut
plasma membrane [33], via an ether linkage to the 6-position
of glucose. The requirement for stability led us to reject a
glycoside linkage to PEG and also any ester link. We
prevented linkage at the aromatic position by using
methylglucose and anticipated that reaction with an
electrophilic derivative would occur largely at the primary
alcohol (O6) of glucose. Using spectroscopic, chromato-
graphic, chemical, and enzymatic analyses, we confirmed
the structure and stability of the compound di(glucos-6-
yl)PEG
600
. Competition studies indicated that this com-
pound did not inhibit Na
+
/glucose transport into intestinal
BBMVs, and the infusion of the intestine with di(glucos-6-
yl)PEG
600
, but not PEG
600
, led to induction of functional
SGLT1.
We conclude that the lumenal sugar is sensed by a glucose
sensor, with a sugar specificity different from that of
SGLT1, which is located on the external face of the

intestinal lumenal membrane. This initiates a signalling
pathway, independent of glucose metabolism, leading to
enhanced SGLT1 expression.
To identify the molecular components of the signalling
pathway, we used the intestinal cell line, STC-1, as an in vitro
system. We have shown previously that these cells respond
to medium glucose, and regulate SGLT1 expression, in a
manner similar to that shown in the native intestinal tissue
[8]. In STC-1 cells, SGLT1 abundance is up-regulated in
response to increased medium glucose concentration [8]
(Fig. 8). Measurements of intracellular cAMP under the
same experimental conditions, indicated that the levels are
also increased when cells are exposed to high-glucose
medium. Inclusion of 8Br-cAMP in the culture medium
resulted in an increase in SGLT1 protein abundance similar
to that observed in response to glucose. 8Br-cAMP also
enhanced the ovine SGLT1 promoter activity. The PKA-
specific inhibitor, H-89, completely abolished the glucose-
induced SGLT1 promoter activity, and significantly
inhibited the induction of endogenous SGLT1 expression
in the STC-1 cells. Therefore, we conclude that changes in
the intracellular cAMP level, and activation of PKA could
be mechanisms for glucose-responsive SGLT1 gene expres-
sion. These data are consistent with other reports showing
SGLT1 upregulation by elevated cAMP levels in the
porcine kidney-derived cell line, LLC-PK
1
[34]. Our previ-
ous studies indicated that induction of SGLT1 expression
requires de novo protein synthesis; there is no evidence for an

intracellular pool of SGLT1, and therefore the increase in
SGLT1 protein abundance is unlikely to be due to the
recruitment of the protein from intracellular stores [8].
Having shown that increases in intracellular cAMP
increased SGLT1 promoter activity, and SGLT1 expres-
sion, we investigated the possibility that glucose sensing may
be linked to a G-protein, analogous to Gpr1, the GPCR in
yeast. The presence of G
a
subunits was confirmed in the
BBMVs of lambs, adult ruminant sheep, and glucose-
infused ruminant sheep, as well as in STC-1 cells. The
addition of pertussis toxin, an inhibitor of the inhibitory
3386 J. Dyer et al.(Eur. J. Biochem. 270) Ó FEBS 2003
G-protein (G
i
), resulted in a twofold increase in SGLT1
abundance, identical with the effect of adding 8Br-cAMP.
Inhibition of G
i
by pertussis toxin results in hyperstimula-
tion of adenylate cyclase leading to increased cAMP levels.
Adenylate cyclase is under negative regulation by G
i
and
positive activation by G
as
. It is tempting to propose that the
G-protein, G
as

, could be linked to the glucose-sensing
mechanism in the intestine. We suggest that glucose-induced
SGLT1 gene activation may be initiated through a GPCR,
via an adenylate cyclase–PKA pathway.
In summary, our data indicate that lumenal glucose is
sensed by a sugar sensor, probably distinct from SGLT1,
and located on the external face of the intestinal lumenal
membrane. The glucose sensor initiates a signalling path-
way, involving a GPCR linked to a cAMP–PKA pathway,
which eventually leads to enhancement of SGLT1 expres-
sion, resulting in an increase in the number of functional
intestinal Na
+
-dependent sugar transporters.
It would be intriguing to determine if enterocytes, which,
like S. cerevisiae, have to adjust their function to wildly
fluctuating levels of extracellular glucose, have developed
similar mechanisms to sense glucose and regulate their
function, and furthermore, if the glucose-sensing GPCR
system is a common, but unexplored, mechanism present in
other eukaryotic cells. In any case, the data from these
studies should allow a better understanding of the mech-
anisms of glucose sensing and glucose-induced signalling in
the control of intestinal glucose absorption.
Acknowledgements
We thank Drs Richard Simmonds and Kishore Bagga for their help in
the chemical synthesis and Dr Dennis Scott for his assistance with the
ovine intestinal infusions. Financial support from the Wellcome Trust is
gratefully acknowledged.
References

1. Solberg, D.H. & Diamond, J.M. (1987) Comparison of different
dietary sugars as inducers of intestinal sugar transporters. Am. J.
Physiol. 252, G574–G584.
2. Ferraris, R.P. & Diamond, J.M. (1989) Specific regulation of
intestinal nutrient transporters by their dietary substrates. Annu.
Rev. Physiol. 51, 125–141.
3. Lescale-Matys, L., Dyer, J., Scott, D., Freeman, T.C., Wright,
E.M. & Shirazi-Beechey, S.P. (1993) Regulation of the ovine
intestinal Na
+
/glucose co-transporter (SGLT1) is dissociated
from mRNA abundance. Biochem. J. 91, 435–440.
4. Shirazi-Beechey, S.P. (1996) Intestinal sodium-dependent
D
-glu-
cose co-transporter: dietary regulation. Proc. Nutr. Soc. 55,
167–178.
5. Dyer, J., Hosie, K.B. & Shirazi-Beechey, S.P. (1997) Nutrient
regulation of human intestinal sugar transporter (SGLT1)
expression. Gut 41, 56–59.
6. Shirazi-Beechey, S.P., Hirayama, B.A., Wang, Y., Scott, D.,
Smith, M.W. & Wright, E.M. (1991) Ontogenic development of
lamb intestinal sodium-glucose co-transporter is regulated by diet.
J. Physiol. (Lond.) 437, 699–708.
7. Freeman, T.C., Wood, I.S., Sirinathsinghji, D.J., Beechey, R.B.,
Dyer, J. & Shirazi-Beechey, S.P. (1993) The expression of the
Na
+
/glucose cotransporter (SGLT1) gene in lamb small intestine
during postnatal development. Biochim. Biophys. Acta. 1146,

203–212.
8. Vayro, S., Wood, I.S., Dyer, J. & Shirazi-Beechey, S.P. (2001)
Transcriptional regulation of the ovine intestinal Na
+
/glucose
co-transporter SGLT1 gene: the role of HNF-1 in glucose acti-
vation of promoter function. Eur. J. Biochem. 268, 5460–5470.
9. Wallace, R.J. (1995) Biochemistry and microbiology in the rumen.
In Physiological and Clinical Aspects of Short Chain Fatty Acids
(Cummings, J.H., Rombeau, J.L. & Sakarta, T., eds), pp. 57–71.
Cambridge University Press, Cambridge, UK.
10. Mills, J., France, J. & Dijkstra, J. (1999) A review of starch
digestion in the lactating dairy cow and proposal for a mechanistic
model. 1. Dietary starch characterization and ruminal starch
digestion. J. Anim. Feed. Sci. 8, 291–340.
11. Wood, I.S., Dyer, J., Hofmann, R.R. & Shirazi-Beechey, S.P.
(2000) Expression of the Na
+
/glucose co-transporter (SGLT1) in
the intestine of domestic and wild ruminants. Pflugers Arch. 441,
155–162.
12. Dyer, J., Scott, D., Beechey, R.B., Care, A.D., Abbas, S.K. &
Shirazi-Beechey, S.P. (1993) Dietary regulation of intestinal glu-
cose transport. In Mammalian Brush-Border Membrane Proteins,
Part II (Lentze, M.D. & Grand, R., eds), pp. 65–72. Thieme
Verlag, Stuttgart and New York.
13. Wood, I.S., Allison, G.G. & Shirazi-Beechey, S.P. (1999) Isolation
and characterization of a genomic region upstream from the ovine
Na
+

/
D
-glucose cotransporter (SGLT1) cDNA. Biochem. Biophys.
Res. Commun. 257, 533–537.
14. Krzeminski, L.F. & Landmann, W.A. (1963) Methyl yellow as a
spray reagent in the paper chromatograpy of chlorinated hydro-
carbon pesticides. J. Chromatogr. 10, 515–516.
15. Dittrich, K. (1986) Analyses by emission, absorption, and fluor-
escence of small molecules in the visible and ultraviolet range in
gaseous-phase. CRC Crit. Rev. Anal. Chem. 16, 223–279.
16. Rindi, G., Grant, S.G.N., Yiangou, Y., Ghatei, M.A., Bloom,
S.R., Bautch, V.L., Solcia, E. & Polak, L.M. (1990) Development
of neuroendocrine tumors in the gastrointestinal tract of trans-
genic mice. Heterogeneity of hormone expression. Am.J.Pathol.
136, 13449–13463.
17. Shirazi-Beechey, S.P., Davies, A.G., Tebbutt, K., Dyer, J., Ellis,
A., Taylor, C.J., Fairclough, P. & Beechey, R.B. (1990) Prepara-
tion and properties of brush-border membrane vesicles from
human small intestine. Gastroenterology 98, 676–685.
18. Tarpey, P.S., Wood, I.S., Shirazi-Beechey, S.P. & Beechey, R.B.
(1995) Amino acid sequence and the cellular location of the Na
+
-
dependent
D
-glucose symporters (SGLT1) in the ovine enterocyte
and the parotid acinar cell. Biochem. J. 312, 293–300.
19. Shirazi-Beechey, S.P., Kemp, R.B., Dyer, J. & Beechey, R.B.
(1989) Changes in the functions of the intestinal brush border
membrane during the development of the ruminant habit in

lambs. Comp. Biochem. Physiol. B 94, 801–806.
20. Dyer, J., Fernandez-Castan
˜
o Merediz, E., Salmon, K.S.H.,
Proudman, C.J., Edwards, G.B. & Shirazi-Beechey, S.P. (2002)
Molecular characterisation of carbohydrate digestion and
absorption in equine small intestine. Equine Vet. J. 34, 349–358.
21. Dyer, J., Wood, I.S., Palejwala, A., Ellis, A. & Shirazi-Beechey,
S.P. (2002) Expression of monosaccharide transporters in the
intestine of human diabetics. Am. J. Physiol. 282, G241–G248.
22. Dyer, J., Barker, P.J. & Shirazi-Beechey, S.P. (1997) Nutrient
regulation of the intestinal Na
+
/glucose co-transporter (SGLT1)
gene expression. Biochem. Biophys. Res. Commun. 230, 624–629.
23. Kocienski, P.J., Cemigliaro, G. & Feldstein, G. (1977) Pheromone
synthesis. 4. A synthesis of (+)-methyl n-tetradeca-trans-2,4,5-
trienoate, an allenic ester produced by the male dried bean beetle
Acanthoscelides obtectus (Say). J. Org. Chem. 42, 353–355.
24. Nakafuku, M., Obara, T., Kaibuchi, K., Miyajima, I., Miyajima,
A., Itoh, H., Nakamura, S., Arai, K I., Matsumoto, K. & Kaziro,
Y. (1988) Isolation of a second yeast Saccharomyces cerevisiae
Ó FEBS 2003 Intestinal glucose sensing (Eur. J. Biochem. 270) 3387
(GPA2) coding for guanine nucleotide-binding regulatory protein:
studies on its structure and possible functions. Proc. Natl Acad.
Sci. USA 85, 1374–1378.
25. Xue, Y., Batlle, M. & Hirsch, J.P. (1998) GPR1 encodes a putative
G-protein coupled receptor that associates with the Gpa2p Ga
subunit and functions in a Ras-independent pathway. EMBO
J. 17, 1996–2007.

26. Kraakman, L., Lemarie, K., Pingsheng, M., Teunissen,
A.W.R.H., Donaton, M.C., Van Dijck, P., Winderickx, J., de
Winde, J.H. & Thevelein, J.M. (1999) A Saccharomyces cerevisiae
G-protein coupled receptor, Gpr1, is specifically required for
glucose activation of the cAMP pahway during the transition to
growth on glucose. Mol. Microbiol. 32, 1002–1012.
27. Rolland, F., Winderickx, J. & Thevelein, J.M. (2001) Glucose-
sensing mechanisms in eukaryotic cells. Trends Biochem. Sci. 26,
310–317.
28. Lorenz, M.C., Pan, X., Harashima, T., Cardenas, M., Xue, Y.,
Hirsch, J.P. & Heitman, J. (2000) The G-protein-coupled receptor
Gpr1 is a nutrient sensor that regulates pseudohyphal differ-
entiation in Saccharomyces cerevisiae. Genetics 154, 609–622.
29. Colombo, S., Pingsheng, M., Cauwenberg, L., Winderickx, J.,
Crauwels, M., Teunissen, A., Nauwelaers, D., de Winde, J.H.,
Gorwa, M F., Colavizza, D. & Thevelein, J. (1998) Involvement
of distinct G-proteins, Gpa2 and Ras, in glucose-and intracellular
acidification-induced cAMP signalling in the yeast Saccharomyces
cerevisiae. EMBO J. 17, 3326–3341.
30. Johnston, M. (1999) Feasting, fasting and fermenting, glucose
sensing in yeast and other cells. Trends Genet. 15, 29–33.
31. O
¨
zcan, S., Dover, J. & Johnston, M. (1998) Glucose sensing and
signalling by two glucose receptors in the yeast Saccharomyces
cerevisiae. EMBO J. 17, 2566–2573.
32. Itoh, Y., Kawamata, Y., Harada, M., Kobayashi, M., Fujii, Y.,
Fukusumi, S., Ogi, K., Hosoya, M., Tanaka, Y., Uejima,
H., Tanaka, H., Maruyama, M., Satoh, R., Okubo, S., Kizawa,
H., Komatsu, H., Matsumura, F., Noguchi, Y., Shinohara, T.,

Hinuma, S., Fujisawa, Y. & Fujino, M. (2003) Free fatty acids
regulate insulin secretion from pancreatic b-cells through GPR40.
Nature (London) 422, 173–176.
33. Lloyd, J.B. (1998) Intestinal permeability to polyethyleneglycol
and sugars: a re-evaluation. Clin. Sci. 95, 107–110.
34. Yet, S.F., Kong, C.T., Peng, H. & Lever, J.E. (1994) Regulation of
Na
+
/glucose cotransporter (SGLT1) mRNA in LLC-PK1 cells.
J. Cell Physiol. 158, 506–512.
3388 J. Dyer et al.(Eur. J. Biochem. 270) Ó FEBS 2003