Studies on membrane vesicles from the kidney of Leucoraja erinacea suggested the sole presence of a sodium-d-glucose cotransporter type 1 involved in renal d-glucose reabsorption. For molecular characterization of this transport system, an mRNA library was screened with primers directed against conserved regions of human sglt1. A cDNA was cloned whose nucleotide and derived amino acid sequence revealed high homology to sodium glucose cotransporter 1 (SGLT1). Xenopus laevis oocytes injected with the respective cRNA showed sodium-dependent high-affinity uptake of d-glucose. Many positions considered functionally essential for sodium glucose cotransporter 1 (SGLT1) are also found in the skate protein. High conservation preferentially in transmembrane helices and small linking loops suggests early appearance and continued preservation of these regions. Larger loops, especially loop 13, which is associated with phlorizin binding, were more variable, as is the interaction with the specific inhibitor in various species. To study the intrarenal distribution of the transporter, a skate SGLT1-specific antibody was generated. In cryosections of skate kidney, various nephron segments could be differentiated by lectin staining. Immunoreaction with the antibody was observed in the proximal tubule segments PIa and PIIa, the early distal tubule, and the collecting tubule. Thus Leucoraja, in contrast to the mammalian kidney, employs only SGLT1 to reabsorb d-glucose in the early, as well as in the late segments of the proximal tubule and probably also in the late distal tubule (LDT). Thereby, it differs also partly from the kidney of the close relative Squalus acanthias, which uses SGLT2 in more distal proximal tubule segments but shows also expression in the later nephron parts.
- sodium glucose cotransporter 1
- intrarenal localization
secondary active transporters use the energy of electrochemical gradients of ions like Na+ to transport a substrate against its own concentration gradient. In the kidney, for example, sodium/d-glucose cotransporters (SGLT) are responsible for the reabsorption of d-glucose from the primary urine. In the mammalian kidney, the low-affinity and high-capacity SGLT2 is operating in the early proximal tubule segments, whereas in the later proximal tubule segments, the high-affinity and low-capacity SGLT1 mediates glucose reabsorption at low intraluminal glucose concentrations (19).
In early vertebrates like fish, sodium-dependent d-glucose transport has been observed in brush-border membrane fractions derived from hagfish, toadfish, flounder, trout, shark, and skate kidney (8). The latter studies revealed that in the skate kidney d-glucose transport exhibits a 2:1 stoichiometry (Na+:d-glucose) and a high d-glucose affinity (0.12 mM). Therefore, it was suggested that in the skate kidney, only an SGLT1-type transporter is operating. These observations differed from those with spiny dogfish shark (Squalus acanthias) kidney brush-border membrane vesicles, in which a 1:1 stoichiometry, a lower affinity to sugars, and phlorizin and different substrate specificity were observed (9).
In addition, it was found that in the closely related shark renal d-glucose reabsorption involves exclusively an SGLT2-type transporter (1). This prompted the current investigation to seek genomic evidence for the differences in the two transporters, as exhibited in the functional studies in vitro.
Furthermore, we found in the shark some striking differences in the intrarenal distribution of the transporter compared with the mammalian kidney (1). Thus, the second aim of the study was to create a specific antibody for use in immunohistochemistry to find out whether this phenomenon is restricted to the shark or whether it also occurs in the skate.
The current study demonstrates that the sodium-d-glucose cotransporter in the skate kidney belongs indeed to the SGLT1 subfamily. Skate SGLT1 is mainly expressed in the PIIa segment and the early distal tubule, which constitute ∼60–65% of the nephron mass but also in the collecting tubule. Thus, its intrarenal distribution differs significantly both from that found in the shark and in the mammalian kidney. The higher stoichiometry of transport and the high abundance may also explain that in the skate, but not in the shark, d-glucose is completely reabsorbed even at high glucose loads.
MATERIALS AND METHODS
Collection of tissue.
Leucoraja erinacea were caught in Frenchman Bay, ME, between June and August and held in seawater tanks at the Mount Desert Island Biological Laboratory, Salisbury Cove, ME, until use. For our experiments, females were used preferentially, as female kidneys are easier to dissect. Animals were killed by decapitation and pithing, and tissues were prepared nonperfused, but washed in shark Ringer solution (in mM): 330 urea, 280 NaCl, 72 trimethyl-amine-N-oxide, 5 glucose, 0.5 Na2HPO4, 1.0 NaH2PO4, 8 NaHCO3, 3 MgCl2, 6 KCl, 5 CaCl2, pH 7.4. All samples were freshly prepared on ice and either stored in an appropriate volume of Ambion RNAlater solution (Huntingdon, UK) at +4°C or directly homogenized for RNA or protein extraction. Samples for immunohistochemistry were rinsed with ice-cold 4% paraformaldehyde in 3 × PBS during preparation and immediately processed as described below.
RNA extraction and cDNA synthesis.
Total RNA was extracted with Tri Reagent (TRIzol) from Sigma (Deisenhofen, Germany) following the manufacturer's protocol. The tissue was homogenized using either a Polytron homogenizer (Kinematica AG, Switzerland) or a loose and a tight Dounce homogenizer. Possible DNA contamination was removed with DNaseI digestion with the DNA-free kit (Ambion) before use of the extracted RNA. RNA integrity was checked by formaldehyde agarose gel electrophoresis (1.2% agarose). To synthesize cDNA from a total RNA template, we used oligo-dT15 primer (Promega, Mannheim, Germany) and SuperScript II reverse transcriptase (GIBCO BRL, Inchinnan, UK). After first-strand synthesis, all samples were treated with RNase H (GibcoBRL) to remove the RNA template.
Preparation of cDNA library.
cDNA library was prepared with the SuperScript kit for cDNA libraries from GIBCO BRL. For integration into plasmid pSPORT1, first strands were synthesized from total RNA with oligo-dT15 with Not I adapter and labeled with [α32P]dCTP. To introduce asymmetry, Sal I adapters were ligated to the cDNA after second-strand synthesis. Subsequent digestion with Not I exposed the 5′ extensions of Not I sites for ligation. cDNA was purified by size fractionation by column chromatography. The size-fractionated cDNA was then ligated into pSPORT1 vectors and transformed into Escherichia coli using electroporation.
Southern blot analysis.
Probes for Southern blot (dot-blot) were labeled with DIG-dUTP in a random priming process with the DIG DNA labeling kit (Boehringer Mannheim, Mannheim, Germany). High stringent screening included hybridization of the probe with DIG Easy Hyb Granules (Boehringer) at 68°C for 16 h, washing once with 2× SSC + 0.1% SDS for 5 min at room temperature, two times with 0.2× SSC + 0.1% SDS at 50°C for 15 min and once with 0.2× SSC + 0.1% SDS at 60°C for 15 min. Low-stringency conditions meant incubation at 40°C and the last two washing steps with 0.2× SSC at 50°C. Signals were detected with the CSPD chemiluminescence phosphatase system from Boehringer.
Cloning of SGLT from the kidney of L. erinacea.
We previously reported the successful cloning of an SGLT2 from the kidney of S. acanthias by screening a cDNA library with a probe derived from trout (O. mykiss) kidney SGLT (1). In a similar approach, primers against conserved regions of mammalian sglt1 genes were designed and used in RT-PCR on trout kidney mRNA. A combination of 5′-AAT AAT GGC ACA GTG GCA GTC AAC-3′ and 5′-ACC TCA TTG AAG GCA TAA CCC ATG-3′ gave rise to a fragment of 650 bp. As this fragment showed similarity to sglt1 genes, it was named sglt1tk. The fragment was labeled and used as a probe to screen a library derived from L. erinacea kidney. Thereby, we isolated several clones that were subsequently subjected to restriction mapping, subcloning, and sequencing. Analysis showed that they were carrying a cDNA with high homology to sglt1 genes, as revealed by BLASTn search in the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov) (2) nucleotide database. In a screen with the 1,573 bp probe containing sglt2tk (1), we detected the same 2,468 bp cDNA as with sglt1tk.
To isolate plasmids from E. coli, we used the Plasmid kit from Qiagen (Hilden, Germany).
DNA was sequenced by labeling with BigDye terminator and reading on an ABI PRISM 3100 (Applied Biosystems).
Quantitative real-time PCR was performed in a GeneAmp 5700 Sequence Detection System from Applied Biosystems (Weiterstadt, Germany). Applied Biosystems also supplied all reagents used, including primers, probes, and disposables for real-time PCR. Fifty-microliter reactions contained 2× TaqMan Universal PCR Master Mix, 10 pmol TaqMan probe (labeled with carboxyfluorescein-reporter and carboxytetramethylrhodamine-quencher), 45 pmol of forward and reverse primer and 0–2 μl cDNA (20 ng) or plasmid solution as template. The standard thermal protocol was as following: 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, and 60°C for 1 min. The threshold of fluorescence signal was set to 0.06; the baseline was defined during seven or eight cycles in a range from cycle 7 to 17. Four replicates for each value were made. According to a standard curve derived from 7.50 pg, 1.88 pg, 0.47 pg, 0.12 pg, and 0.03 pg plasmid (141000 copies/pg) cloned from skate SGLT1, the control software of the GeneAmp 5700 calculated the quantity (in picograms) from the measured Ct-value in each well. Replicates with the highest deviation were taken out of the analysis. Statistics and conversion to the number of copies were calculated with Microsoft Excel. All values are presented as number of start copies in 20 ng cDNA with SE.
On the basis of hydrophobicity [as predicted by TMHMM2.0; www.cbs.dtu.dk (10, 14)], a tentative topology model was generated. Furthermore, we used the PROSITE database (www.expasy.ch) (4) to predict possible sites for N-glycosylation and phosphorylation.
Uptake studies on Xenopus laevis oocytes.
The library vector pSPORT is also suitable for in vitro transcription. Thus, the vector was linearized by HindIII digestion, and cRNA was synthesized using Ambion's T7 mMESSAGE mMACHINE kit. cRNA coding for skate kidney SGLT or a corresponding volume of water as negative control was injected into oocytes on the same or one day after preparation. To allow expression from the injected cRNA oocytes were incubated in Barth's solution supplemented with gentamycin under standard conditions for 3–5 days. Before uptake, the oocytes were washed 2× with Barth's solution and subsequently preincubated in Na+-free uptake solution for 3 min. Uptake solution for Na+-dependent uptake was composed of 100 mM NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 10 mM HEPES/Tris, pH 7.5, 10 μM cytochalasin B, d-glucose in the indicated concentration, and 40 μCi/ml d-[3H]glucose. In the solution for Na+-independent uptake NaCl was replaced with 100 mM N-methylglucamine. Oocytes were placed in the respective uptake media for 20 min at room temperature after which uptake of d-[3H]glucose was terminated by removing the medium and washing 3× with cold stop solution (composition as for Na+-independent uptake). To measure the content of d-[3H]glucose, each oocyte was placed in a single vial and lysed in 10% SDS for 1 h. After this, ReadyProtein+ scintillation solution (Beckman Coulter, Fullerton, CA) was added, and counting was performed in a Beckman Coulter LS 6500. For each condition, 8–10 oocytes were tested, and experiments were performed twice with independent sets of oocytes. For analysis, data of the uptake experiments were pooled and analyzed employing Microsoft Excel and Microcal Origin 6.0 (nonlinear regression).
Generation of antibodies.
As loop 13 has a quite unique primary structure for each isoform of SGLT, amino acids 553–634 from this region were selected as epitope for antigen production. The respective DNA fragment was amplified from the cloned cDNA and tagged with recombination adaptors for Invitrogen's Gateway cloning system (Karlsruhe, Germany). The sequence-specific parts of the primers were as follows: 5′-GAG GAT AAG CAC CTC CAC CGA-3′ and 5′-TGA GAT GTC GGT CAG CTT CTT CTG-3′. After integration of the PCR product into the pGEX-5x-3 expression vector, the sequence information was checked by DNA sequencing. The peptide was produced in BL21 E. coli as glutathione S-transferase fusion protein, extracted, purified with glutathione-sepharose (Amersham Pharmacia Biotechnology, Piscataway, NJ), and sent to Biotrend (Köln, Germany) for immunization of rabbits. The immune serum was purified by Biotrend with protein A to obtain the IgG fraction.
Immediately after the animals were killed and the kidneys were exposed, the tissue was fixed in situ by dripping ice-cold fixation fluid (4% paraformaldehyde in 3× PBS; ∼900 mOsmol/l) onto the samples. The tissue was dissected under a constant drip of fixation fluid. Small pieces from the middle part of the organ were transferred into fresh fixative and incubated overnight at 4°C. Afterward, the tissue was washed overnight at 4°C in PBS. The specimens were embedded in Tissue-Tek OCT compound (Sakura Finetek USA, Torrance, CA) and frozen in liquid nitrogen; sections of 5–7 μm were obtained with a Leica Microsystems cryostat (Bensheim, Germany). After thawing and briefly air-drying the sections on glass slides, the sections were stored at −80°C until use. Unspecific immunoreactive binding sites were blocked with 3% nonfat dry milk in PBS during a 1-h incubation at room temperature. Subsequently, the respective primary antibodies were diluted 1:300 in PBS + 1% Triton X-100 and incubated on the slides for 1 h at room temperature. In antigen competition experiments, 50 μg of antigenic peptide were added to the antibody dilution before application. Immuno signals were detected with Alexa-Fluor 488 donkey anti-rabbit IgG or Alexa-Fluor 555 goat anti-rabbit IgG secondary antibodies (Molecular Probes, Eugene, OR) diluted 1:300, and incubated for 1 h at room temperature. The following fluorescein-labeled lectins were used to stain distinct nephron segments: Concanavalin A (Con A), Jacalin (JAC), Lens culinaris agglutinin (LCA), Lycopersicon esculentum lectin, Pisum sativum agglutinin (PSA), Peanut agglutinin (PNA), Phaseolus vulgaris erythroagglutinin (PHA-E), Ricinus communis agglutinin I, Solanum tuberosum lectin (STL), and succinylated wheat germ agglutinin (SWGA) (Vector Laboratories, Burlingame, CA). The dilution was 1:150 in PBS, and the incubation was performed for 20 min at room temperature. Nuclei were counterstained with the DNA dye 4′,6-diamidino-2-phenylindole, which was included in the ProLong antifade Gold solution (Molecular Probes) used for mounting the sections. Multichannel analysis of the immunostainings was performed with an Axiovert 200M fluorescence microscope equipped with Plan-Neofluar lenses 16×/0.5, 25×/0.8 and F-Fluar 40×/1.3. Images were acquired with an Axiocam MR CCD camera Apotome (multidimensional images).
Sequence of the cloned cDNA.
By screening a cDNA library from kidney of L. erinacea, a cDNA was cloned whose sequence has been submitted to EMBL-Bank as AM183563. Both the nucleotide sequence, as well as the amino acid sequence, derived from an open reading frame from ATG at 32 to TAA at 2,018 showed high homology to SGLT in a BLAST search. In a dendrogram of the amino acid comparison, the skate sequence was sorted into one branch with the mammalian SGLT1 and SGLT3 isoforms (Fig. 1). The highest identity was with SGLT1 with an average value of 71.2%. Identity with SGLT3 was 66.0% and with SGLT2 only 63.0%. More specifically, skate SGLT1 and shark SGLT2 showed only 64.1% identity, and major differences in the extramembraneous loops (Fig. 2). Hydrophobicity plot predicted 14 transmembrane domains and an identical membrane topology. The similarities extended further to possible posttranslational modifications (as predicted by computer analysis) and to amino acids previously associated with the function of the protein (Na+, glucose, and phlorizin binding) (1).
Expression of the gene in Xenopus laevis oocytes.
cRNA or a corresponding volume of nuclease-free water was injected into Xenopus oocytes, and uptake of d-[3H]glucose was determined on the 4th day after injection. cRNA-injected oocytes showed significant (P < 0.03) sodium-dependent uptake of d-glucose (Fig. 3A). This was not observed in control oocytes. Kinetics of this sodium-dependent uptake into SGLT-expressing oocytes was determined for several glucose concentrations (Fig. 3B). Total uptake in the presence of sodium was saturable at high sugar concentrations. Passive uptake in the absence of sodium increased in a linear way with glucose concentration in the medium. The difference between total and passive uptake was the secondary-active, Na+-dependent and SGLT-mediated part. This part was again saturable with a Vmax of 207 ± 16 pmol/20 min and an apparent Km of 0.11 ± 0.04 mmol/l (n = 2 experiments with 8–10 oocytes per glucose concentration).
In mammals SGLT1 is expressed not only in the kidney but also at high levels in the intestine. To test for the presence of the cloned SGLT in the skate intestine, we used quantitative real-time PCR with TaqMan probes. With this method we found about 97,100 ± 24,100 (n = 4) copies of SGLT1 in 20 ng cDNA from intestine, which is more than twice the cellular concentration of transcripts found in kidney [40,700 ± 2,300 (n = 4) copies].
Identification of renal structures and localization of SGLT1 protein in the skate kidney by immunohistochemistry.
In the kidney of elasmobranch fish like L. erinacea, three different zones can be distinguished: 1) The lateral bundle zone with the countercurrent system, 2) a small region where the glomeruli are located, and 3) the zone of mesial tissue (3, 6, 7). Fig. 4 gives a schematic overview of the meandering course of a single nephron. It also shows the binding sites of an antibody raised against the predicted loop 13 of skate SGLT1 used for immunostainings, as described below.
Figure 5 experiments are related to the specificity of the antibody reaction. With the serum obtained after immunization, the apical membrane for example of the early distal tubule was labeled (Fig. 5A, red), and the addition of the antigen to the immune serum abolished the binding (Fig. 5B). Incubation with the preimmune serum gave no signals either (Fig. 5C). Similar controls were performed for all stainings with identical (negative) results. Furthermore, in Western blot analysis of kidney homogenates, mainly a band of 73.5-kDa molecular weight reacted with the antibody, which is in the range expected from other species. This reaction was also not observed with preimmune serum or after the addition of the antigen (data not shown).
Glomeruli displayed binding of LCA to the basement membrane of the capillaries. Binding of the anti-SGLT1 antibody was not observed (Fig. 6A). Some erythrocytes in the capillaries of the glomerulus appear yellowish, probably due to autofluorescence; compare also Fig. 6B. Otherwise a yellow signal indicates an overlay of the red signal for SGLT1 and the green signal from the lectin.
The first segment of the nephron on its course into the lateral bundle is the neck segment. Near the glomerulus, the cells are almost cubic, many of them carrying several cilia. But the deeper into the bundle, the more the cells flatten and lose the cilia. In the posterior portions of the neck segment, where cells are flat, a few epithelial cells are marked by JAC at the apical region (Fig. 7A). SGLT1 was not detected (Fig. 6B).
Proximal tubule segment PI.
The early part of the proximal tubule (PIa) is characterized by low prismatic cells that slowly flatten. SWGA showed a strong reaction with the apical region of these cells (Fig. 6B). This is either unspecific because of the high concentration of the lectin or perhaps intracellular vesicle pools below the apical membrane. In the earlier part of this segment, we also observed the first weak signal with the anti-SGLT1 antibody on the luminal brush-border membrane (Fig. 6B). JAC gave a speckled signal in the later flatter cells (Fig. 7A), probably due to binding of the lectin to intracellular vesicles of the secretory pathway. The subsequent proximal tubule PIb is only found in the vicinity of the glomeruli. The brush-border membrane was specifically labeled by PHA-E (Fig. 7C). LCA was also bound to the luminal membrane but not the anti-SGLT1 antibody (Fig. 6C).
Proximal tubule segment PII.
The later part of the proximal tubule, which is located in the mesial tissue, is called PII. It is further differentiated into PIIa and PIIb. PIIa has large high prismatic cells with a central round nucleus but only a small lumen. A strong fluorescence signal for SGLT1 was observed in the brush-border membrane of this segment. Single cells also reacted with PNA. Overlay of the two signals resulted in a yellow color (Fig. 6D). In PIIb the cells are smaller while lumen of the tubule gets wider. Here, binding of neither lectins nor the antibody against SGLT1 was observed (Fig. 7A).
The intermediate segment (IS) connects the proximal tubule to the distal tubule. The cells are cubic to flat with large nuclei. Interspersed are multicilliary cells. Thus, it is similar to the neck segment. Usually, the intermediate segment starts in the zone of the glomeruli and extends into the lateral bundle, forming part of the countercurrent system. As shown in Fig. 7A, JAC labeled parts of the low epithelium of the IS, but no reaction with the anti-SGLT1 antibody was found.
Early distal tubule.
The early distal tubule segment (EDT) is found only in the lateral bundle zone. The EDT was stained by several lectins: LCA (Fig. 6C), JAC (Fig. 7A), and STL (Fig. 7B). There was also a strong signal for SGLT1 at the apical side of the cells, as evident from the yellow staining of the lumen of the early distal tubule in Fig. 6C and Fig. 7, A and B.
Late distal tubule.
On the way to the mesial tissue zone the early distal tubule changes to the LDT. Out of the tested lectins, none labeled the LDT, and no labeling with the anti-SGLT1 antibody was found (Fig. 6B).
Collecting tubule-collecting duct system.
After another turn in the mesial tissue back to the lateral bundles the cells of the late distal tubule flatten, spread more, and evolve into the collecting tubule. Back in the bundle zone, the collecting tubule eventually joins the collecting duct. JAC (Fig. 7A), and PHA-E (Fig. 7C) marked this nephron segment. Fig. 7, A and C shows also that in the CT, there was again strong colocalization with the antibody against SGLT1 at the luminal membrane.
Table 1 gives a semiquantitative comparison of the intensity of the immunoreactivity observed. It is noteworthy that the two segments that comprise the major mass of the tubule, the late proximal tubule PIIa with ∼50% of abundance and the early distal tubule (10–15% abundance), show strong to very strong signals. In addition the collecting tubule is clearly stained. These results suggest a very high expression of SGLT1 in the late tubular segments of the skate kidney.
Recently, we reported the identification of an SGLT2 from the kidney of S. acanthias, which was the first SGLT characterized in a nonmammalian vertebrate (1). Here, we now report the cloning and characterization of an SGLT from the kidney of L. erinacea, another marine elasmobranch.
Southern blot analysis screen of a skate kidney mRNA library resulted in the isolation of a cDNA that showed high homology to the mammalian SGLT1. The high homology to SGLT1 also extended to the derived amino acid sequence with 662 residues forming 14 transmembrane domains, a typical feature of this class of transporters (16) and containing sodium-symporter motifs. In a dendrogram of the comparison of the amino acid sequences, the skate transporter is sorted into the branch leading to the mammalian SGLT1 and SGLT3. One can now speculate that SGLT3 evolved after the separation of the chondrichthyes from the main path of evolution leading to the higher vertebrates.
There are several motifs (20) and single amino acids that are either conserved between the different members of the slc5a family or have been directly connected to the function of the protein (1). All of them are conserved in similar positions in the skate. The fact that all of these residues already appear in the early vertebrates underlines their functional relevance. However, some diversity among the SGLT sequences can be found. The NH2-terminus and the larger loops, especially loop 13, show more variation. Thus, this might explain the differences observed in the functional properties.
There are, however, some major differences between the skate and the shark protein: for example, the sequence of loop 13 is very different. This might be the reason for the different interaction of the skate and the shark transporter with phlorizin and alkylglucosides observed previously (9). However, further studies similar to ones performed with the isolated loop 13 of the rabbit SGLT1 (11, 12) are required to substantiate this assumption.
With regard to posttranslational modifications, there is also much homology (1). Computer prediction also found several of the possible phosphorylation sites in Leucoraja SGLT1 (Ser51, Ser304, Thr419, and Thr563). The threonine at 419 is no PKA site, as in the human SGLT1; instead, there is one at threonine 631. Several other sites are potential motifs for phosphorylation by casein kinase (four in loop 13). Recently, position Ser418 has been shown to be involved in the regulation of the transport activity of mammalian SGLT1 (S. Subramanian and R. K. Kinne, unpublished data). Investigation of the elasmobranch transporters might give additional information about which of these sites are of functional importance.
To prove the functionality of the cloned sequence and to investigate kinetic properties, cRNA was transcribed in vitro, injected into Xenopus laevis oocytes and uptake of d-[3H]glucose was measured. Oocytes injected with cRNA coding for skate kidney SGLT showed a sodium-dependent uptake of glucose with an apparent affinity of 0.11 mM. This is in agreement with the results obtained in vesicle studies where a Km of 0.12 mM had been determined (9) and is also in the range of mammalian SGLT1 (i.e., human SGLT1 Km ≈ 0.8 mM; Ref. 5).
The expression of skate SGLT1 in different tissues is similar to mammalian tissues, too. The transporter from Leucoraja was cloned from and is expressed in kidney. But more than twice, the cellular concentration of transcripts can be found in the intestine. Furthermore, expression of SGLT1 in this organ was confirmed by immunostaining of the tissue with the anti-skate SGLT1 antibody used in this study (13).
Despite the many similarities between the mammalian and the elasmobranch SGLT1 in regard to the molecular features, transport characteristics, and tissue distribution, we found marked differences in the expression pattern along the nephron. In the mammalian nephron, a serial system of SGLT2 and SGLT1 exists. That is, in the early segments of the proximal tubule (S1), the low-affinity but high-capacity SGLT2 performs the bulk reabsorption of d-glucose. In the later segments of the proximal tubule (S3), SGLT1 can remove small residual amounts of sugar, because of its high affinity (17). The later parts of the nephron like the early and late distal tubule and the collecting duct then serve for fine-tuning the water inorganic electrolyte content of the urine but no longer absorb organic compounds.
In L. erinacea, we detected SGLT1 predominantly in the later proximal tubule PIIa, and only a weak signal was observed in the early proximal tubule PIa. This corresponds to the situation observed in the mammalian kidney. The presence of SGLT1 in the early distal tubule and in the collecting tubule of skate kidney is, however, not observed in the nephron of mammals.
Compared to Squalus acanthias, in which we detected SGLT2 in the proximal tubule PIa, PIb, PIIb, and LDT (1), it seems as if the place of d-glucose reabsorption moved one segment closer to the glomerulus. A comparable phenomenon has been observed for the NaPi-II transporters. In fish like flounder and zebrafish, NaPi is found in the late nephron segments. In contrast, the rat expresses NaPi only in the early segments. The reason for this shift during evolution could be the much higher glomerular filtration rate of mammals that requires, in return, use of the full reabsorptive potential of the proximal tubule (18). Whether this holds true for skate and shark remains to be determined.
According to personal observations the final urine of Leucoraja is free of glucose, even after d-glucose infusion. Thus, the one-transporter system in the skate kidney seems to be as effective as the two-transporter system in the mammals in reabsorbing glucose. This is probably achieved by expressing SGLT1 for a longer distance along the nephron and using the full reabsorptive potential of the proximal tubule. Whether the SGLT is involved in d-glucose reabsorption in the collecting tubule, too, is not clear; it might be necessary to compensate for the generally low capacity of SGLT1 transporters to achieve complete glucose reabsorption. But an involvement of SGLT1 in the transport of urea and/or water is also possible (20).
To conclude, we have reported the isolation and characterization of an SGLT1-type transporter from the kidney of the early vertebrate L. erinacea. In immunohistochemistry, the intrarenal distribution of the transporter revealed a localization of the transporter partly differing from the mammalian homologs.
On the basis of the current data, it seems unlikely that the skate expresses another SGLT-type transporter. Although we cannot exclude the existence of an SGLT2 in skate kidney completely, the vesicle studies, the lack of cross reactivity of the antibody, as well as PCR experiments, indicate a homogenous population of SGLT1.
These findings are quite in contrast to those obtained in shark kidney, raising some interesting questions on the relation between the two species.
This research was supported by National Institute of Environmental Health Sciences Grant 1-P30-ESO 3828 to R. K.-H. Kinne.
The help of Birgit Hülseweh on the experiments is gratefully acknowledged. Daniel Scharlau helped with immunohistochemistry.
Parts of the data were presented in form of abstracts in the Bulletin of the Mount Desert Island Biological Laboratory in Maine. Parts of the data are part of the diploma thesis of Thorsten Althoff at Philipps-Universität Marburg, Germany, supervised by Rolf K.-H. Kinne and Jürgen Kunz.
Present address for T. Althoff: Max-Planck-Institut für Biophysik, Abteilung Strukturbiologie, Frankfurt, 60438, Germany.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society