Tripeptides of RS1 (RSC1A1) Inhibit a Monosaccharide-dependent Exocytotic Pathway of Na+-d-Glucose Cotransporter SGLT1 with High Affinity*

The human gene RSC1A1 codes for a 67-kDa protein named RS1 that mediates transcriptional and post-transcriptional regulation of Na+-d-glucose cotransporter SGLT1. The post-transcriptional regulation occurs at the trans-Golgi network (TGN). We identified two tripeptides in human RS1 (Gln-Cys-Pro (QCP) and Gln-Ser-Pro (QSP)) that induce posttranscriptional down-regulation of SGLT1 at the TGN leading to 40–50% reduction of SGLT1 in plasma membrane. For effective intracellular concentrations IC50 values of 2.0 nm (QCP) and 0.16 nm (QSP) were estimated. Down-regulation of SGLT1 by tripeptides was attenuated by intracellular monosaccharides including non-metabolized methyl-α-d-glucopyranoside and 2-deoxyglucose. In small intestine post-transcriptional regulation of SGLT1 may contribute to glucose-dependent regulation of liver metabolism and intestinal mobility. QCP and QSP are transported by the H+-peptide cotransporter PepT1 that is colocated with SGLT1 in small intestinal enterocytes. Using coexpression of SGLT1 and PepT1 in Xenopus oocytes or polarized Caco-2 cells that contain both transporters we demonstrated that the tripeptides were effective when applied to the extracellular compartment. After a 1-h perfusion of intact rat small intestine with QSP, glucose absorption was reduced by 30%. The data indicate that orally applied tripeptides can be used to down-regulate small intestinal glucose absorption, e.g. in diabetes mellitus.

Glucose absorption in the small intestine is mediated by the combined action of two glucose transporters in the enterocytes: the sodium-dependent D-glucose cotransporter SGLT1 in the brush-border membrane and the sodium-independent glucose transporter GLUT2 in the basolateral membrane (1,2). It plays a pivotal role for maintenance of blood glucose concentration (3) and glucose and fat metabolism in the liver (4,5). In addi-tion, small intestinal glucose uptake triggers glucose-dependent mechanisms that are mediated by glucose sensing neurons in the intestinal wall, for example, the regulation of intestinal motility (6 -9).
Previously, we identified the intracellular protein RS1 (human gene RSC1A1) that participates in the transcriptional and post-transcriptional regulation of SGLT1 (21,(23)(24)(25)(26)(27)(28). RSC1A1 is an intronless single copy gene that was only detected in mammals and codes for 67-68-kDa proteins in human, pig, rabbit, and mouse that exhibit about 70% amino acid identity. RS1 has a broad tissue distribution, including renal proximal tubules, small intestinal epithelial cells, hepatocytes, and neurons (23,28,29). In LLC-PK 1 cells RS1 was located to the intracellular side of the plasma membrane, at the trans-Golgi network (TGN) and within the nucleus (30). Importantly, the amount and distribution of RS1 were dependent on the state of confluence (30). Whereas subconfluent LLC-PK 1 cells contained large amounts of RS1 protein and exhibited pronounced nuclear location of RS1, the amount of RS1 was decreased in confluent LLC-PK 1 cells and RS1 did not distribute into nuclei (30). The observed changes in distribution correlated with functional data showing that RS1 down-regulates the transcrip-* This work was supported by Deutsche Forschungsgemeinschaft Grant SFB tion of SGLT1 in subconfluent LLC-PK 1 cells (25). In Xenopus oocytes, expression of SGLT1 was inhibited when human RS1 was coexpressed (23,24,27) or when purified RS1 protein was injected into SGLT1 expressing oocytes shortly before uptake measurements (21). This indicated that hSGLT1 was inhibited by hRS1 on the posttranscriptional level. We showed that this posttranscriptional inhibition occurred within 30 min and was due to blockage of the release of SGLT1 containing vesicles from the TGN (21,30). This posttranscriptional down-regulation of SGLT1 by RS1 was increased by PKC and modulated by intracellular methyl-␣-D-glucopyranoside (AMG) (21). Interestingly RS1 was found to be associated with a 28-kDa protein called the IRIP protein that is up-regulated in kidneys after ischemia and reperfusion (31). The physiological and potential biomedical importance of RS1 was demonstrated by targeted disruption of the Rsc1A1 gene in mice (26). Mice without RS1 showed an increased amount of SGLT1 protein in small intestinal brush-border membranes that was correlated with an increased capacity for small intestinal glucose absorption. Furthermore, mice without RS1 developed an obese phenotype.
In the present study we tried to identify the domain of human RS1 (hRS1) that is responsible for inhibition of the release of SGLT1 containing vesicles from TGN. Two tripeptides of hRS1 were identified that are high affinity inhibitors of hSGLT1 at the TGN. Down-regulation of hSGLT1 by the tripeptides was modulated by intracellular monosaccharides including AMG that is not metabolized, and 2-deoxyglucose, which does not interact with hSGLT1. We showed that the tripeptides are substrates of the small intestinal peptide transporters PepT1 (32) and that they were capable to down-regulate SGLT1 in rat small intestine after application to the intestinal lumen.

Animals
Male rats (Wistar, Crl:WI) purchased from Charles River (Sulzfeld, Germany) were housed in groups of 3-4 animals in makrolon cages with free access to Standard Diet 3433 (Provimi Kliba, Kaiseraugst, Switzerland) and drinking water. The animals were kept under 12-h light/dark cycle at room temperature of 22 Ϯ 2°C and a relative humidity of 45-75%. Body weight at the day of surgery was 473 Ϯ 17 g (n ϭ 7). Xenopus laevis frogs were obtained from Nasco (Ft. Atkinson, WI). Animals were handled in compliance with institutional guidelines and German laws.

Antibody Against Human SGLT1 (hSGLT1)
Polyclonal antibody against a peptide of hSGLT1 (34) (amino acids 581-599, QEGPKETIEIETQVPEKKK-C) was raised in rabbits. The peptide with the C-terminal cysteine was coupled to ovalbumin with m-maleimidobenzoyl-N-hydroxysuccimide ester. Affinity purification of the antisera was performed on the antigenic peptide coupled to polyacrylamide beads by using the Sulfolink kit from Pierce. The specificity of the antibody was verified by Western blots on cells that were transfected with hSGLT1 and by immunohistochemistry on human small intestine.

Cloning of Truncated Variants of hRS1
Variants of hRS1 were made on the basis of wild type hRS1 cloned into ApaI and XhoI sites of the oocyte expression vector pRSSP (21).
Variant 1-251-PCR was performed using pRSSP plasmidspecific forward primer and a reverse primer at the corresponding position of hRS1 containing stop codon and the XhoI restriction site. The amplificate was digested with ApaI and XhoI and inserted into pRSSP vector.
Variants 326 -617, 407-617, and 534 -617-PCR was performed using pRSSP plasmid-specific reverse primer and forward primers at corresponding positions of hRS1. Forward primers were provided with initiation codon and ApaI site. After ApaI/XhoI digestion isolated fragments were inserted into the pRSSP vector.
Variants 350 -425 and 396 -425-PCR was performed using primers to the corresponding positions of hRS1. Forward primers contained initiation codon and ApaI site, whereas reverse primers contained stop codon and the XhoI site. PCR amplified fragments were cloned into ApaI/XhoIdigested pRSSP vector.
Variant 251-617-Plasmid pRSSP/hRS1 was cut with ApaI, the adaptor ApaI/ClaI containing the initiation codon was added, and digestion with ClaI was performed to remove excess ligated adaptor and to cut hRS1 at the ClaI site of the coding region. The 5Ј terminal fragment of hRS1 was removed and the plasmid was self-ligated resulting in construct 251-617.
Variant 251-488-Variant 251-617 was cut with Eco81I and XhoI, the 5Ј overhangs were filled in using Klenow fragment of DNA polymerase I, and the adaptor SpeI provided with the stop codon was added. After ligation the mixture was digested with SpeI to remove excess ligated adaptor, the 3Ј terminal fragment of hRS1 was removed, and the plasmid was self-ligated resulting in construct 251-488.
Variant 407-415-Two complementary phosphorylated oligonucleotides were designed to cover the desired region of hRS1: 5Ј-pCATGCAGAATGAACAGTGTCCACAAGTCTC-ATGAC-3Ј and 5Ј-pTCGAGTCATGAGACTTGTGGACAC-TGTTCATTCTGCATGGGCC-3Ј. The oligonucleotides contained the initiation and stop codon (bold letters) and were provided with ApaI and XhoI cohesive ends (underlined). The oligonucleotides were annealed and ligated into ApaI/XhoIdigested pRSSP vector. All PCR-derived parts of the truncated variants were sequenced to rule out PCR errors.

Expression of Transporters and hRS1 in Oocytes of X. laevis
Mature female X. laevis were anesthetized by immersion in fresh water containing 0.1% 3-aminobenzoic acid ethyl ester. Oocytes at the stages V and VI were obtained by partial ovariectomy and treated overnight with 1 mg ml Ϫ1 collagenase I in Ori buffer (5 mM HEPES-Tris, pH 7.6, 110 mM NaCl, 3 mM KCl, 1 mM MgCl 2 , and 2 mM CaCl 2 ). The oocytes were washed twice with Ca 2ϩ -free Ori buffer and kept at 16°C in modified Barth's solution (15 mM HEPES, pH 7.6, 88 mM NaCl, 1 mM KCl, 0.3 mM Ca(NO 3 ) 2 , 0.4 mM CaCl 2 , 0.8 mM MgSO 4 ) containing 12.5 g ml Ϫ1 gentamycin. Selected oocytes were injected with 25 nl of water containing cRNAs (2.5 ng of hSGLT1, 10 ng of hPepT1, 7.5 ng of hRS1 or fragments of hRS1). For expression, injected oocytes were kept for 3 days at 16°C in modified Barth's solution with gentamycin. Non-injected oocytes served as controls.
In some experiments PKC was stimulated by incubation of the oocytes for 2 min with 1 M phorbol 12-myristate 13-acetate (PMA).

Injection of Peptides, Monosaccharides, and Biochemicals into Oocytes
Three days after cRNA injections into oocytes before flux experiments were started, 25 nl of K-Ori buffer (5 mM HEPES, pH 7.6, 100 mM KCl, 3 mM NaCl, 2 mM CaCl 2 , 1 mM MgCl 2 ) containing peptides, 0.5 pmol of sn-1,2-dioctanoylglycerol (stimulation of PKC), 2 ng of BTXB (inhibition of exocytosis), and/or 0.5 pmol of BFA (inhibition of vesicle budding from the TGN) were injected. Oocytes were incubated for 1 h at room temperature in Ori buffer and tracer uptake measurements were performed. To investigate the influence of intracellular monosaccachride on tripeptide-dependent down-regulation of hSGLT1, different concentrations of sugars or polyacohols with or without 75 pmol tripeptides were injected into hSGLT1 expressing oocytes 1 h before uptake measurements were started. Intracellular concentrations of injected compounds were estimated by assuming an internal distribution volume of 0.4 l (38).

Uptake Measurements in Oocytes
For AMG uptake measurements, hSGLT1 expressing oocytes or non-cRNA-injected control oocytes were incubated for 15 min at room temperature in Ori buffer containing 1.75 or 50 M [ 14 C]AMG. Oocytes were washed four times in ice-cold Ori buffer containing 1 mM phlorizin, and single oocytes were solubilized in 5% (w/v) SDS and analyzed for radioactivity by scintillation counting. Uptake of [ 14 C]AMG was corrected for uptake measured in non-cRNA-injected control oocytes.

Current and Capacitance Measurements
Parallel measurements of membrane current (I m ) and membrane capacitance (C m ) were performed using an EPC-05 amplifier controlled by PULSE and X-Chart software provided by HEKA Electronics (Lambrecht, Germany). I m measurements were performed using the two-electrode voltage clamp method and C m was measured using a previously described paired ramps approach (39). For I m and C m measurements in oocytes expressing hSGLT1, oocytes were superfused at room temperature with Ori buffer and clamped to Ϫ50 mV. For I m measurements in oocytes expressing hPepT1, oocytes were superfused with modified Barth's solution adjusted to pH 6.5, and clamped to Ϫ60 mV.

Cultivation of Caco-2 Cells, Incubation with Tripeptides, and Uptake Measurements
The cell line Caco-2 was acquired from the German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). Cells were grown on Petri dishes (37°C, 5% CO 2 ) in complete Dulbecco's modified Eagle's medium containing 25 mM D-glucose. Nine days after reaching confluence cells were detached by incubation with DPBS containing 0.5 mM EDTA and 20 mM D-glucose, collected by 10-min centrifugation at 1000 ϫ g, and washed with DPBS containing 20 mM D-glucose (DPBS-G). To test effects of tripeptides, cells were incubated for 1 h at 37°C in DPBS-G that was adjusted to pH 6.5. Incubation at pH 6.5 was performed in the absence and presence of tripeptides. Before transport measurements were started cells were washed two times with DPBS. For uptake measurements cells were incubated for 10 min at 37°C in DPBS containing 50 M [ 14 C]AMG or 50 M [ 14 C]AMG plus 0.5 mM phlorizin. Uptake was terminated by the addition of ice-cold DPBS containing 1 mM phlorizin (stop solution). Cells were washed 3 times with stop solution, solubilized, and analyzed for radioactivity. For each experiment, five measurements were performed per experimental condition.

Measurement of Glucose Absorption in Rat Small Intestine
A modified version of a previously described perfusion model was used (40). Fasted male rats were anesthetized with pentobarbital (Narcoren, Merial GmbH, Halbergmoos, Germany, 60 mg/kg intraperitoneal). Their body temperature was maintained with a heated pad. Following a laparotomy, a 25-cm long segment of the small intestine starting 1 cm distally of the stomach was isolated and cannulated at both ends. For pretreatment with tripeptide, the small intestinal segments were perfused for 1 h (recirculating system, 4 ml/min, 38°C) with DPBS, pH 6.5. The perfusion was performed in the absence and presence of 15 mM QSP. Thereafter, perfusion was switched to single pass perfusion with DPBS (pH 7.4, 38°C) containing 10 M D-glucose, and after 20 min to DPBS containing 50 M D-glucose. Intestinal efflux samples were collected at 2.5-min intervals. D-Glucose determination in the efflux samples was performed using the Amplex Red Glucose GO assay kit (Invitrogen) adapted for measurements of low D-glucose concentrations.

Isolation of Plasma Membrane from X. laevis Oocytes
Non-injected oocytes and oocytes injected with hSGLT1 cRNA were incubated for 3 days and oocytes without follicular epithelial cells were selected. hSGLT1 expressing oocytes were injected with 25 nl of K-Ori buffer without and with 75 pmol of acetylated QCP (QCPac) and kept for 1 h at room temperature in modified Barth's solution. Thereafter oocytes were incubated for 15 min at room temperature in Ori buffer containing 50 M AMG, and washed twice with Ori buffer. Isolation of plasma membranes from oocytes was performed as described (41). For each experimental condition experiments were performed with 10 oocytes. Defolliculated oocytes were rotated for 30 min at 4°C in MES buffer (20 mM MES, pH 6.0, 80 mM NaCl) containing 1% (w/v) colloidal silica (Ludox Cl from Sigma). Then oocytes were washed two times with MES buffer, rotated 30 min at 4°C in MES buffer containing 0.1% polyacrylic acid (Sigma), and washed two times with Ori buffer. Oocytes were homogenized in 1.5 ml of homogenization buffer (HbA) (20 mM Tris-HCl, pH 7.4, 5 mM MgCl 2 , 5 mM NaH 2 PO 4 , 1 mM EDTA, 80 mM sucrose) and centrifuged at 10 ϫ g for 1 min at 4°C. 1.3 ml of the top of the sample was removed, 1 ml of HbA buffer was added, and the sample was mixed. Centrifugation, removal of 1 ml from the top, addition of 1 ml of HbA buffer, and mixing were repeated 4 times but the centrifugation was performed twice at 10 ϫ g, once at 20 ϫ g, and once at 40 ϫ g. After the last step, plasma membranes were spun down by a 30-min centrifugation at 16,000 ϫ g (4°C). The resulting pellet was resuspended in 4 l/oocyte of sample buffer for SDS-PAGE (SDS-PAGE sample buffer: 60 mM Tris HCl, pH 6.8, 100 mM dithiothreitol, 2% (w/v) SDS, and 7% (v/v) glycerol, 0.1% (w/v) bromphenol blue).

SDS-PAGE and Western Blotting
Protein concentration was determined according to Bradford (42). For SDS-PAGE, membrane protein samples were pretreated for 45 min at 37°C in SDS-PAGE sample buffer. Electrophoresis and Western blotting were performed as described (33). Proteins separated by SDS-PAGE were trans-ferred by electroblotting to polyvinylidene difluoride membrane. For Western blot analysis, the blots were blocked with blocking buffer (DPBS, pH 7.4, 2% bovine serum albumin, 0.1% Tween 20), followed by a 2-h incubation at room temperature with affinity purified antibody against hSGLT1 diluted 1:20,000 in blocking buffer. After washing, the blots were incubated for 2 h at room temperature with horseradish peroxidase-conjugated protein G (Bio-Rad) diluted in blocking buffer. The immunoreaction was visualized by enhanced chemiluminescence using the ECL system (Amersham Biosciences Europe).

Calculation and Statistics
Values are given as mean Ϯ S.E. One way analysis of variance test with post hoc Tukey comparison was used to test for significance of difference between mean values. Unpaired Student's t test was used to assess differences between two groups. IC 50 values were determined by fitting the Hill equation to the data after the base of noninhibitable activity was subtracted. The K m values for uptake of QSP or QCP by hPepT1 were obtained by fitting the Michaelis-Menten equation to the data. The band intensity in Western blots was quantified by densitometry (25).

RESULTS
hRS1-derived Tripeptides Inhibit AMG Uptake Expressed by hSGLT1-Previously we described that uptake of AMG into oocytes of X. laevis expressing hSGLT1 was decreased when hRS1 was coexpressed or when purified hRS1 protein was injected into hSGLT1 expressing oocytes (21,24,27). We showed that hRS1 protein is a coat protein of the TGN and that it blocks the release of SGLT1 containing vesicles (21,30). In the present study we tried to identify domains of hRS1 that are involved in posttranscriptional inhibition of hSGLT1. We coinjected cRNAs encoding various fragments of hRS1 together with cRNA of hSGLT1 into oocytes, incubated them 3 days for expression, and measured the uptake of 50 M [ 14 C]AMG mediated by hSGLT1. Fig. 1a shows that AMG uptake was inhibited when various fragments of hRS1 were coexpressed. An inhibition by 40 -50% was observed when cRNA of intact hRS1 or cRNA fragments encoding amino acids (aa) 1-251 or 251-617 of hRS1 were expressed together with hSGLT1. This suggested that the N-terminal and C-terminal domains of hRS1 can mediate inhibition of hSGLT1, and that the inhibition by both domains is not additive. Focusing on the C-terminal part of hRS1 we performed additional truncations. Whereas cRNA fragments encoding aa 326 -617 or 407-617 showed similar inhibition compared with cRNA encoding total hRS1, the inhibition by a cRNA fragment encoding aa 534 -617 was significantly smaller although still significant. Because aa 572-611 of hRS1 represent an ubiquitin-associated domain (43) that may mediate various ubiquitin-related effects, we concentrated further analysis on RS1 fragments that do not contain the ubiquitin-associated domain. Performing further truncations we were able to attribute an inhibitory effect to a cRNA fragment encoding aa 407-415 of hRS1 (Fig. 1a).
The coexpression experiments described above do not differentiate between posttranscriptional short-term RS1 effects on SGLT1 and long-term effects of RS1 that may involve endog-enous oocyte proteins that participate in the regulation of SGLT1. For this reason, we expressed hSGLT1 in oocytes, injected 75 pmol of hRS1-derived peptides and measured uptake of 50 M [ 14 C]AMG 60 min later. Assuming that oocytes contain an internal aqueous volume of about 0.4 l (38), the intracellular concentrations of injected peptides was about 0.2 mM. In initial experiments acetylated peptides were used because they may be more stable and more soluble in the cytosol. Fig. 1b shows that acetylated nonapeptide QNEQCPQVSac comprising aa 407-415 of hRS1 inhibited SGLT1-expressed AMG uptake by about 50%, whereas the acetylated reversed control peptide SVQPCQENQac revealed no significant inhibition. Injection of acetylated hexapeptides derived from QNEQCPQVS that overlap by three amino acids showed the same inhibition as the nonapeptide. The same inhibition was also obtained after injection of acetylated tripeptide QCPac (aa 410 -412 of hRS1) representing the overlapping amino acids. In contrast, the acetylated reversed tripeptide PCQac was not effective. Identical degrees of inhibition were observed with both QCPac and nonacetylated QCP.
We wondered whether the cysteine in QCP is effective by forming a disulfide bridge to an intracellular protein and whether the tripeptides QSP and QPP that occur in the sequence of hRS1 are also effective. In the experimental series shown Fig. 1c we investigated various acetylated or nonacetylated tripeptides in which cysteine in the middle position was replaced by other amino acids. Under the employed experimental condition using an estimated intracellular peptide concentration of about 0.2 mM, significant inhibition of AMG uptake was not only observed for QCPac, QSPac, QCP, and QSP, but also for QTPac, QPPac, QGPac, QLPac, and QEP. The data suggest that the middle position in QCP is not critical for function, however, it may change affinity. Interestingly, hRS1 contains two QSP tripeptides in the N-terminal part (aa 19 -21 and 91-93) and QPP (aa 311-313) in the middle part of the sequence.
To investigate whether the decrease of hSGLT1-mediated AMG uptake by QCP and QSP is additive we injected hSGLT1 expressing oocytes with 75 pmol of QCPac, 75 pmol of QSPac, or 75 pmol of QCPac plus 75 pmol of QSPac, incubated the oocytes for 1 h, and measured the uptake of 50 M [ 14 C]AMG. AMG uptake was decreased by 34 Ϯ 7% by QCPac, 44 Ϯ 7% by QSPac, and 38 Ϯ 7% by QCPac plus QSPac (2 independent experiments). The data suggest that QCP and QSP activate the same mechanism to decrease AMG transport.  SEPTEMBER 28, 2007 • VOLUME 282 • NUMBER 39 QCP Reduces the Amount of Expressed hSGLT1 within the Plasma Membrane-We investigated whether QCP decreases SGLT1-mediated AMG uptake in the same way as has been proposed for total hRS1, namely by decreasing the amount of SGLT1 in the plasma membrane via inhibition of SGLT1 release from the TGN (21). We isolated oocyte plasma membranes from non-injected control oocytes, from oocytes expressing hSGLT1, and from oocytes expressing hSGLT1 that were injected with 75 pmol of QCP, and incubated for 1 h. Because the effect of QCP was shown to be glucose-dependent (see below and Ref. 21) we incubated the oocytes 15 min with 50 M AMG before the plasma membranes were purified. Separation of plasma membranes from intracellular vesicles was achieved using colloidal silica that binds to the glycocalix of plasma membranes (41). We observed some variation of purified plasma membrane proteins between different oocyte batches, however, within individual oocyte batches total plasma membrane protein was neither increased significantly after expression of hSGLT1 nor decreased significantly when QCPac was injected into hSGLT1 expressing oocytes. The total protein content of isolated plasma membranes normalized to that of non-injected oocytes was 1.01 Ϯ 0.05 in hSGLT1 expressing oocytes and 0.92 Ϯ 0.04 in SGLT1 expressing oocytes that had been injected with QCPac (five independent experiments). Western blots stained with an antibody against hSGLT1 showed that QCPac decreased the amount of hSGLT1 in the plasma membrane by about 50% (Fig. 2a).

Inhibition of SGLT1 Expression by Tripeptides
To determine whether QCP decreases the number of functional active SGLT1 molecules, we performed capacitance measurements. Transport of D-glucose or AMG by SGLT1 or binding of phlorizin to SGLT1 is correlated with a decrease of capacitance that is proportional to the number of functional SGLT1 molecules (44). The left panel of Fig. 2b shows an experiment in which inward current and capacitance at Ϫ50 mV were measured when SGLT1 expressing oocytes were superfused with 5 mM AMG or 100 M phlorizin. AMG is transported by hSGLT1 together with sodium and induces both, inward current and decrease of capacitance. Phlorizin does not induce current because it is not transported, however, it decreases capacitance because it blocks potential-dependent charge movements within hSGLT1. Because superfusion with AMG alters the intracellular concentration of AMG that modulates the effect of QCP on hSGLT1 (see below) we measured the effect of QCPac on phlorizin-induced capacitance changes. The right panel of Fig. 2b shows experiments in which phlorizin-induced capacitance changes were measured in hSGLT1 expressing oocytes that had been injected with K-Ori buffer or with K-Ori buffer containing 75 pmol of QCPac and incubated for 1 h. QCPac decreased phlorizin-induced capacitance changes by about 50%. The data indicate that the short-term inhibition of AMG uptake observed after injection of QCPac is due to a decrease of the number of functional hSGLT1 molecules in the plasma membrane.
QCP Blocks the Exocytotic Pathway of SGLT1-We tested whether fusion of intracellular vesicles with the plasma membrane and release of vesicles from the TGN are required for inhibition of hSGLT1 by QCP. We expressed hSGLT1 in oocytes and injected K-Ori buffer or K-Ori buffer containing 75 pmol of QCPac, 2 ng of BTXB, 0.5 pmol of BFA, QCPac plus BTXB, or QCPac plus BFA. After a 1-h incubation we measured AMG uptake (Fig. 3). BTXB cleaves synaptobrevin located at the intracellular vesicles (45), whereas BFA inhibits guanosine nucleotide exchange factors that activate ADP-ribosylation factors, which regulate the assembly of coat complexes at the TGN and endosomes that are involved in protein sorting and release of vesicles (46). BTXB inhibited SGLT1-mediated AMG uptake by 60%. This indicates that BTXB was active in our experiments and that exocytosis participates significantly in short-term regulation of SGLT1. Importantly, QCPac did not inhibit hSGLT1mediated AMG uptake in the presence of BTXB. Fig. 3 shows also that BFA inhibited hSGLT1-mediated AMG uptake by FIGURE 2. QCP decreases the amount of hSGLT1 protein in the plasma membrane. a, quantification of hSGLT1 protein in the plasma membrane. Plasma membranes were isolated from non-injected, oocytes expressing hSGLT1, and oocytes expressing hSGLT1 that were injected with 75 pmol of QCP and incubated for 1 h. The left panel shows a Western blot from a typical experiment that was stained with affinity purified antibody against hSGLT1. 2 g of protein was applied to each lane. A densitometric quantification of five independent experiments is shown on the right panel. ***, p Ͻ 0.001. b, effect of QCP on phlorizin-induced capacitance changes in oocytes expressing hSGLT1. Oocytes expressing hSGLT1 were superfused with Ori buffer and clamped to Ϫ50 mV. Inward currents and capacitance were measured when oocytes were superfused with Ori buffer containing 5 mM AMG or 100 M phlorizin (left panel). Phlorizin-induced capacitance decrease in hSGLT1 expressing oocytes injected with K-Ori buffer (hSGLT1) or K-Ori buffer containing 75 pmol/oocyte of QCPac (hSGLT1, QCPac) is compared on the right. Mean Ϯ S.E. of 9 oocytes from three independent experiments are compared. ***, p Ͻ 0.001.
50%. This indicates that the release of SGLT1 containing vesicles from endosomes or the TGN participates in short-term regulation. In the presence of BFA no inhibition of AMG uptake by QCPac was detected. The data indicate that QCPac blocks the exocytotic pathway of SGLT1 that is critically involved in short-term regulation.
To investigate how long the inhibitory effect of QCP persists, we injected 75 pmol of QCPac into oocytes, incubated the oocytes between 1 and 9 h at room temperature, and measured the uptake of 50 M [ 14 C]AMG. The inhibitory effect of QCPac persisted throughout the entire investigated time period (data not shown).
Down-regulation of hSGLT1 by QCP and QSP Is Monosaccharide-dependent-Previously we reported that inhibition of hSGLT1-mediated AMG uptake in oocytes observed after injection of purified hRS1 protein was prevented when the oocytes were preincubated with D-glucose (21). We tested whether the protective effects of monosaccharides can be demonstrated for the inhibition of hSGLT1-mediated AMG uptake by QCP or QSP. We expressed hSGLT1 in oocytes and injected 25 nl of K-Ori buffer with and without 75 pmol of QCPac (Fig.  5, a, b, and e) or QSP (Fig. 5, c and d) containing various amounts of mannitol, sorbitol, D-glucose, D-fructose, AMG, or 2-deoxyglucose. Oocytes were incubated for 1 h at room temperature and uptake of 1.75 M [ 14 C]AMG was measured. Fig.  5a shows control experiments where we assessed potential effects of intracellular polyalcohols and monosaccharides on hSGLT1-mediated AMG uptake in the absence of the inhibitory tripeptides. AMG uptake was not influenced by intracellular sorbitol, mannitol, D-glucose, AMG, and 2-deoxyglucose up to a concentration of about 60 mM, and by intracellular D-fructose concentrations of 0.025, 0.25, 2.5, and 62.5 mM. At variance AMG uptake was 30% increased by 25 mM intracellular D-fructose. Inhibition of AMG uptake by QCPac was not changed by various intracellular concentrations of mannitol and sorbitol excluding osmotic effects (Fig. 5b). Protective effects of metabolized monosaccharides (D-glucose, D-fructose, and 2-deoxyglucose) and of the non-metabolized monosaccharide AMG on the inhibition of AMG uptake by QCPac and QSP were observed. For example, inhibition of AMG uptake by QCPac or QSP was significantly protected when the oocytes contained 0.25-62.5 mM D-glucose or D-fructose (Fig. 5, b and c). The monosaccharides D-fructose, AMG, and 2-deoxyglucose exhibited biphasic effects on the inhibition of hSGLT1-mediated AMG uptake by QCPac and QSP (Fig. 5, b and d). Inhibition of AMG uptake by tripeptides was most strongly protected by 0.25 mM AMG or 2-deoxyglucose. We determined the half-maximal concentrations for the protective high affinity effects (PC 50 ) of 2-deoxyglucose and AMG (Fig. 5e). Significantly different values of 11.9 Ϯ 0.8 M (2-deoxyglucose) and 53 Ϯ 1.7 M (AMG) (n ϭ 3 each, p Ͻ 0.001) were obtained. For the high affinity protective effect of D-glucose on inhibition of AMG uptake by QSP, a PC 50 value of 30 Ϯ 6.2 M (n ϭ 3) was determined (Fig.  5c). The data indicate that a high affinity monosaccharide binding site is involved in modulation of QCP/QSP-mediated down-regulation of hSGLT1.
Down-regulation of hSGLT1 by QCP Is Independent of PKC-We investigated whether the effect of QCP on hSGLT1-mediated AMG uptake is increased after stimulation of PKC as has been observed for the effect of total hRS1 protein (21). We injected K-Ori buffer, K-Ori buffer containing 75 pmol of QCPac, or K-Ori buffer containing 75 pmol of QCPac plus 0.5 pmol of dioctanoylglycerol (stimulator of PKC) into hSGLT1 expressing oocytes, incubated the oocytes for 60 min and measured uptake of 50 M [ 14 C]AMG. In the absence of QCP, dioctanoylglycerol increased hSGLT1-mediated AMG significantly as described earlier (21,27) (Fig. 6). Whereas dioctanoylglycerol increased the short-term inhibition of AMG uptake after injection of purified hRS1 protein (21), dioctanoylglycerol did not alter the inhibition of AMG uptake after injection of QCPac. The same result was obtained when PKC was stimulated by PMA. In absence of QCPac, hSGLT1-mediated AMG uptake was significantly increased when oocytes were incubated for 2 min with 1 M PMA, however, inhibition of AMG uptake after injection of QCPac was not changed. The data indicate that at variance to hRS1 the short-term down-regulation of hSGLT1 by QCP is independent of PKC. This suggests that the down-regulation of hSGLT1 by hRS1 is mediated via QCP and/or QSP sites in hRS1 that are modulated by PKC-dependent phosphorylation in hRS1.

QCP and QSP Are Substrates of the Peptide Transporter
PepT1-We wanted to evaluate the possibility to decrease small intestinal absorption of D-glucose by down-regulation of hSGLT1 via orally applied QCP or QSP. Because the tripeptides act intracellularly, we tested whether the human H ϩ -peptide cotransporter hPepT1 in the brush-border membrane of small intestinal enterocytes transports QCP and/or QSP. We expressed hPepT1 in oocytes, superfused the oocytes with modified Barth's solution adjusted to pH 6.5, clamped them to various membrane potentials, and superfused them with saturating concentrations of the well transported dipeptide glycilglutamin (GQ) or with 5 mM QCPac, PCQac, or QSP. Significant inward currents were obtained with acetylated and nonacetylated tripeptides, indicating that QCPac (Fig. 7a), PCQac (data not shown), and QSP (Fig. 7a) are substrates of hPepT1. Measuring hPepT1-mediated inward currents at different concentrations of QCPac and QSP, Michaelis-Menten type activation curves were obtained (see e.g. Fig. 7b). At Ϫ60 mV, K m values of 21.4 Ϯ 6.6 and 4.2 Ϯ 0.2 mM (n ϭ 6, each) were obtained for QCPac and QSP, respectively.
To determine whether hSGLT1 expression can be down-regulated by extracellular application of QCP in cells, we expressed hSGLT1 either alone or together with hPepT1 in oocytes. We then incubated the oocytes for 1 h in either Ori buffer adjusted to pH 6.0 (control) or in Ori buffer, pH 6.0, that contained 150 M QCPac or 150 M PCQac. After a brief wash in Ori buffer, pH 7.6, we measured uptake of 50 M [ 14 C]AMG (Fig. 7c). In oocytes expressing hSGLT1 alone AMG uptake was not altered after incubation with extracellular QCP. However, in oocytes expressing hSGLT1 plus hPepT1, extracellular QCP decreased hSGLT1-mediated AMG uptake by 40%. At variance, the control peptide PCQac that is also transported by hPepT1 had no effect.
Phlorizin-inhibitable AMG Uptake into Caco-2 Cells Is Down-regulated after Incubation with QCP and QSP-To explore whether the hRS1-derived tripeptides supplied with the food are capable to downregulate hSGLT1 in human small intestine we investigated whether extracellular application of tripeptides QCP and QSP to confluent human colon carcinoma (Caco-2) cells inhibits AMG uptake by hSGLT1. At late confluence, Caco-2 cells exhibit similar morphological characteristics as differentiated small intestinal enterocytes (47). They express hSGLT1 and hPepT1 and mediate phlorizin-inhibitable AMG and proton-dependent uptake of dipeptides (37,48). We cultivated Caco-2 cells for 9 days after confluence, incubated them for 1 h at pH 6.0 without and with 1 mM QCPac or 1 mM QSP, washed them at pH 7.4, and measured uptake of 50 M [ 14 C]AMG in the absence and presence of 0.5 mM phlorizin (Fig. 8). 30 -50% of AMG uptake into Caco-2 cells was inhibited by phlorizin and was dependent on extracellular sodium (data not shown). This fraction of AMG uptake is mediated by hSGLT1. We observed that total AMG uptake into Caco-2 cells was significantly inhibited by 15-20% when cells had been preincubated with QCPac or QSP (Fig. 8). The phlorizin inhibitable fraction of AMG uptake was inhibited 55-65% by QCPac and QSP. The data strongly suggest that QCP and QSP are taken up by human enterocytes and inhibit hSGLT1-mediated uptake of glucose.
D-Glucose Absorption in Rat Small Intestine Is Decreased after Perfusion with QSP Containing Solution-To test the effect of QSP in intact small intestine, 25-cm long small intestinal segments of male rats were perfused for 1 h with DPBS adjusted to pH 6.5 in the absence or presence of 15 mM QSP (Fig. 9). Thereafter the segments were perfused for 20 min with DPBS, pH 7.4, containing 10 M D-glucose, and for another 20 min with DPBS containing 50 M D-glucose. During the perfu- sion with D-glucose the absorption of D-glucose was determined by measuring the D-glucose concentration in the effluate. In control animals the concentration of D-glucose in the effluate was about 50% smaller compared with the infused solution. After pretreatment with QSP the concentration of D-glucose in the effluate was about 30% higher compared with the controls. During perfusion with 50 M D-glucose this effect of QSP was statistically significant (p Ͻ 0.05). Assuming that the lower D-glucose concentration in the effluate compared with the infused solution was due to D-glucose absorption, the data indicate that QSP decreased small intestinal D-glucose absorption by more than 30%.

DISCUSSION
In the present study we identified the tripeptides QCP and QSP in sequence of hRS1 that act as high affinity posttranscriptional inhibitors of the Na ϩ -D-glucose cotransporter hSGLT1. Similar to total hRS1 protein (21), QCP and QSP inhibit an exocytotic pathway of hSGLT1. We showed that QCP decreases the amount of hSGLT1 protein in the plasma membrane and that the effect of QCP on hSGLT1-mediated AMG uptake was abolished when the release of vesicles from the TGN was blocked by BFA or when fusion of vesicles with the plasma membrane was blocked by BTXB. Recently we observed that hRS1 is located at the TGN and dissociates from the TGN after application of BFA (30). We concluded that hRS1 is a coat protein of the TGN. The similarities between the observed posttranscriptional effects of hRS1 and the tripeptides indicate that QCP and/or QSP are part of a posttranscriptionally active domain(s) of hRS1. The tripeptides are supposed to bind to a high affinity binding site of a protein that interacts with the TGN. hRS1 contains two QSP motifs in the N-terminal part (aa 19 -21 and 91-93) and one QCP motif in the middle part (aa 410 -  SEPTEMBER 28, 2007 • VOLUME 282 • NUMBER 39 JOURNAL OF BIOLOGICAL CHEMISTRY 28509 412). Because the effects of QCP and QSP were not additive both tripeptides may interact with the same tripeptide binding site. However, it is also possible that two binding sites at the same protein or at two proteins of the same regulatory pathway are involved. Whether both QSP motives in the N-terminal domain of hRS1 are effective, and whether the amino acids between these two sites that include two consensus sequences for the binding of protein 14-3-3 (aa 13-36 and 80 -85) (49) and two consensus sequences for PKC-dependent phosphorylation (serines 45 and 66), play a functional role, must be clarified by future experiments. Because endogenous and overexpressed RS1 protein in HEK 293 cells, Caco-2 cells, and LLC-PK 1 cells is rapidly degraded, 4 and we demonstrated inhibitory activity of small hRS1 fragments including tripeptides, it is possible that the posttranscriptional down-regulation of hSGLT1 by hRS1 is partially mediated by hRS1 fragments. However, the observation that the inhibition of hSGLT1 expression after injection of total hRS1 protein was increased when PKC was activated, whereas the inhibition by QCP was independent of PKC, suggests that the physiologically active domain of hRS1 contains a PKC site or is regulated by a distant PKC site in hRS1.

Inhibition of SGLT1 Expression by Tripeptides
In the present study the previously described modulatory effect of AMG on posttranscriptional regulation by hRS1 (21) was further investigated. We observed similar effects of AMG, 2-deoxyglucose, D-glucose, and D-fructose on QCP-or QSPmediated down-regulation of hSGLT1. The monosaccharides attenuated the inhibition of hSGLT1 transport activity by the tripeptides. The observed effects of metabolized monosaccha-4 H. Koepsell and A. Filatova, unpublished data.  Oocytes were injected with cRNA of hPepT1 (a-c), cRNA of hSGLT1 (c) or cRNAs of hSGLT1 and hPepT1 (c) and incubated 3 days for expression. In a and b, oocytes were superfused with modified Barth's solution, pH 6.5, clamped to Ϫ60 mV, and inward currents induced by glycilglutamin (GQ), acetylated QCP (QCPac), or nonacetylated QSP were measured. a, current traces. b, concentration dependence of QSP induced inward currents (I QSP ). Mean Ϯ S.E. of six oocytes are shown that were normalized to maximal GQ-induced currents measured after superfusion with 5 mM GQ (I GQmax ). The Michaelis-Menten equation was fitted to the data. c, down-regulation of hSGLT1 after incubation of oocytes expressing hSGLT1 plus hPepT1 with QCP. Oocytes expressing hSGLT1 or hSGLT1 plus hPepT1 were incubated for 1 h at pH 6.5 in the absence of peptides (control), or in the presence of 5 mM acetylated QCP or acetylated PCQ. Oocytes were washed and uptake of 50 M [ 14 C]AMG was measured. Mean Ϯ S.E. of three independent experiments are presented. **, p Ͻ 0.01. rides (2-deoxyglucose, D-glucose, and D-fructose) and the nonmetabolized monosaccharide AMG, indicate the involvement of a hexokinase independent pathway. This pathway is triggered by a monosaccharide side with a K d of 30 M for D-glucose. For 2-deoxyglucose a lower apparent K d value was determined, however, we did not distinguish whether this is due to accumulation of 2-deoxyglucose 6-phosphate. The detected high affinity monosaccharide binding site cannot be located on hRS1 itself because monosaccharide effects were observed with hRS1-derived tripeptides, and tripeptides are supposed to be too small to form a monosaccharide binding site. This site has a different sugar specificity than SGLT1-mediated monosaccharide transport because 2-deoxyglucose was effective and SGLT1 does not translocate 2-deoxyglucose or 2-deoxyglucose 6-phosphate (1). However, it has been shown recently that rabbit SGLT1 contains an additional monosaccharide binding site with a different sugar specificity than the transport site that interacts with 2-deoxyglucose (50). Thus we cannot exclude that the observed modulation of tripeptide-mediated posttranscriptional down-regulation of hSGLT1 by monosaccharides is mediated via binding of the monosaccharides to SGLT1. However, this possibility is con-sidered to be highly improbable, because hRS1 protein and QCP also down-regulate the expression of the human organic cation transporter 2 in a monosaccharide-dependent manner (21). 5 The modulatory monosaccharide binding site in the RS1 regulation pathway is supposed to be localized on a protein different from SGLT1 that interacts with hRS1 at the TGN.
Previous data suggested monosaccharide-dependent posttranscriptional regulation of SGLT1. One study with rat jejunum showed that V max of phlorizin inhibitable D-glucose uptake was increased 30 min after incubation with 25 mM D-glucose (51). In this work the authors did not distinguish whether the SGLT1 protein in the plasma membrane or the turnover of SGLT1 was increased. In another study in which Caco-2 cells were preincubated for 1 h with 100 mM D-glucose, a 45% decrease of phlorizin inhibitable AMG uptake was observed (20). The authors proposed that this down-regulation is due to increased endocytosis, however, no significant changes of the intracellular distribution of SGLT1 could be detected. In small intestine, monosaccharide-dependent regulation of SGLT1 has been also observed on the level of transcription (10,18). In yeasts, monosaccharide-dependent posttranscriptional regulation of several hexose transporters has been described (52). Recently, it has been reported that the monosaccharide H ϩ -cotransporter VvHT1 in grape is post-5 M. Veyhl, A. Vernaleken, and H. Koepsell, unpublished data.  transcriptionally regulated by the non-metabolized monosaccharide 3-O-methyl-D-glucose (53).
Do mammals need monosaccharide-dependent posttranscriptional and transcriptional regulation of SGLT1 in addition to regulation by enterohormones (14,15,54) and the autonomous nervous system (13)? Apparently mammals developed a highly complex regulatory network in which small intestinal sugar absorption steers the regulation of 1) gastric and intestinal mobility (6,7,55), 2) liver metabolism (4,56), and 3) insulin secretion (54,57). Both, the glucose concentration in the small intestinal lumen and the activity of glucose uptake systems SGLT1 and GLUT2, determine the D-glucose concentration in the small intestinal submucosa and the portal blood. Glucose sensors in small intestinal ganglia cells (8,55), in the portal vein, and/or liver (58,59) transform the signal to neuronal activity. Because SGLT1 in the brush-border membrane of enterocytes mediates the first step in glucose absorption, monosaccharidedependent regulation at this step may be very effective. After high glucose loading of small intestine GLUT2 distributes from the basolateral membrane of the enterocytes to the brushborder membrane (2). This redistribution of GLUT2 is supposed to be triggered by intracellular glucose and is consequently dependent on the function of SGLT1. Considering the described functions of SGLT1, the physiological importance of monosaccharide-dependent posttranscriptional regulations of SGLT1 for short-term adaptations is obvious. In addition, monosaccharide-dependent transcriptional regulation of SGLT1 expression is important for long-term adaptation of glucose absorption to diet.
The identification of tripeptides that down-regulate the expression of hSGLT1 and are substrates of the H ϩ -peptide cotransporter PepT1 opens new possibilities for treatment of obesity and diabetes type 2. PepT1 is expressed in enterocytes of the small intestine epithelial cells (32) and we showed that in cells expressing PepT1, the intracellular concentrations of QCP or QSP were high enough to down-regulate SGLT1 when the tripeptides were added extracellularly. This was demonstrated in oocytes, in polarized Caco-2 cells that are widely used as a model for human enterocytes (47), and in isolated and perfused rat small intestine. These experiments showed that QCP or QSP added to the food may down-regulate small intestinal D-glucose and D-galactose absorption by 20 -50%. In addition to reducing energy supply this is beneficial for the prevention and treatment of diabetes because it alleviates postprandial glycemic excursions as has been shown for the ␣-glucosidase inhibitors acarbose (60,61). Glycemic peaks during diabetes are responsible for the desensitization of insulin receptors and cause vascular and renal complications. It has been shown that acarbose treatment of patients with impaired glucose tolerance was correlated with decreased incidence of diabetes type 2 (61). QSP and related compounds are supposed to have a similar beneficial effect but two advantages compared with acarbose. Two daily oral doses should be enough to obtain down-regulation of glucose absorption, and overdosage does not lead to side effects due to sugar malabsorption.
In future experiments we want to elucidate whether QSP or related compounds may also be useful to decrease postprandial glucose peaks in the blood via inhibition of renal glucose reab-sorption. Several pharmaceutical companies focus on this strategy using analogs of phlorizin that are selective for hSGLT2 in comparison to hSGLT1 (62)(63)(64). SGLT2-specific inhibitors are preferred for two reasons. First, the bulk of filtrated glucose is reabsorbed by the low affinity transporter SGLT2 that is expressed in early proximal tubules, whereas the high affinity transporter SGLT1 in late proximal tubules mediates fine adjustment of glucose reabsorption (1). Second, human SGLT1 is highly expressed in heart (65,66) and inhibitors of SGLT1 may impair energy supply of cardial myocytes. Preliminary experiments showed that QCP and QSP also mediate shortterm down-regulation of hSGLT2 and are also transported by the H ϩ /peptide cotransporter PepT2. 5 Using a QSP-related compound that is absorbed in small intestine or applied intravenously, it may be possible to down-regulate hSGLT1 and hSGLT2 in the kidney. After glomerular filtration, such a compound will be taken up by proximal tubules via PepT1 and PepT2 (32). Because QSP-related compounds only act intracellularly, they should not down-regulate hSGLT1 in heart, where no H ϩ /peptide cotransporters are expressed (65,66).
In summary, we identified tripeptides in hRS1 that mediate posttranscriptional down-regulation of hSGLT1 at the TGN. We indicated that the tripeptides activate an inhibitory pathway that is modulated by intracellular monosaccharides. These tripeptides and related compounds are potential drugs for treatment of obesity and diabetes type 2.