Mapping the urea channel through the rabbit Na⁺-glucose cotransporter SGLT1

Abstract

1. The rabbit Na⁺-glucose cotransporter rbSGLT1 and its carboxy-terminal part, C5, which contains transmembrane helices 10-14 of SGLT1 and functions as a low affinity glucose uniporter, were expressed as individual proteins in Xenopus oocytes. Transport of 55 μm urea, ethylene glycol, mannitol and α-methyl-d-glucopyranoside (αMDG) by control oocytes and by oocytes expressing SGLT1 and C5 was studied by uptake measurements of the ¹⁴C-labelled substrates.

2. There was a 5 to 6-fold increase in urea transport mediated by C5, compared with control oocytes. Similar to SGLT1, the C5-urea uptake was cation independent, linear in time and with increasing urea concentration, and blocked with the same sensitivity by the inhibitor phloretin (K_i≈ 1 mm). Like SGLT1 in choline buffer, the C5-mediated uptake was insensitive to phlorizin.

3. Mannitol was transported by C5 but not by SGLT1 or control oocytes.

4. The activation energy (E_a) for urea transport through C5 was low (5 ± 3 kcal mol⁻¹) compared with that of non-injected oocytes (16 ± 0.5 kcal mol⁻¹) and comparable with the E_a of passive urea or water transport through intact SGLT1.

5. The urea influx through C5 increased in the presence of αMDG, but not in the presence of the same concentration of mannitol.

6. We conclude that the five carboxy-terminal transmembrane helices of SGLT1 form a channel for the permeation of small molecules such as urea and water.

The Na⁺-glucose cotransporter SGLT1 is a transmembrane protein with complex function. Its major responsibility is to accumulate sugar against its concentration gradient into intestinal or renal epithelial cells by exploiting the energy from the Na⁺ electrochemical potential gradient across the cell membrane (for reviews see Wright & Loo, 2000; Wright, 2001). In addition, SGLT1 includes Na⁺ uniport, water and urea channel activity (Loo et al. 1996, 1999; Leung et al. 2000) and water and urea influx coupled to its cotransport cycle (Loo et al. 1996; Meinild et al. 1998a; Zeuthen et al. 2001). Urea and water transport through SGLT1 was also expanded to other transporters, such as the rat Na⁺-iodide cotransporter rNIS and the human Na⁺-Cl⁻-GABA transporter hGAT1 (Leung et al. 2000). Since all transported water, it is possible that the water and urea activities may be interrelated.

Since no structural information is available about the intramolecular localization of the urea channel activities of SGLT1, this study presents an attempt to localize the urea transport function in a particular domain of the transporter. A previously well characterized, functional truncated form of SGLT1, the C5 protein, which contains the sugar permeation pathway of SGLT1 (Panayotova-Heiermann et al. 1997, Panayotova-Heiermann 1999) was expressed in Xenopus oocytes and its transport characteristics for different substrates were compared and contrasted with those of expressed SGLT1 and native oocytes.

METHODS

Harvesting and injection of oocytes

All procedures related to maintaining the frogs and harvesting the oocytes were performed in accordance with the UCLA's Chancellors Animal Protection Committee and NIH guidelines. Before surgery the selected Xenopus laevis female frog was anaesthetized by immersion in 0.1 % Tricaine (3-aminobenzoic acid ethyl ester; Sigma, St Louis, MO, USA) for 5-10 min, then transferred into an ice slush for 10 min.

Stage V-VI oocytes were isolated and treated with 0.32 % collagenase (Boeringer Mannheim, Indianapolis, IN, USA) for 1 h at room temperature and defolliculated in K₂HPO₄ buffer for 1 h. The oocytes were extensively rinsed, placed in Barth's medium and stored at 18 °C.

The C5 construct containing the last five transmembrane helices of rabbit SGLT1, amino acids 407-662 (Panayotova-Heiermann et al. 1997), was linearized with EcoR I and the corresponding capped cRNA was synthesized in vitro by using the SP6 MEGAscript transcription kit from Ambion (Austin, TX, USA). Oocytes were injected with 50 nl of rabbit (rb)SGLT1 or rbC5 cRNA (1 μg μl⁻¹) and kept for 3-6 days at 18 °C to allow protein expression.

Frogs were killed with an overdose of anaesthetic (Nembutal, Sigma) after the final collection of oocytes.

Radioactive transport assays

All flux measurements were performed at 22 °C as described by Ikeda et al. (1989). The ¹⁴C-labelled substrates were used at a final concentration of 55 μm in a buffer containing (mm): 100 NaCl, 2 KCl, 1 MgCl₂, 1 CaCl₂ and 10 Hepes-Tris, pH 7.5. Experiments were repeated at least twice with oocytes from different donor frogs. Each bar in the graphs represents the mean of the substrate influx from a group of seven to eight individual oocytes expressed as the mean ±s.e.m. A p value of < 0.05 was assumed to indicate a statistically significant difference (Student's t test).

¹⁴C-labelled methyl-α-d-glucopyranoside (316 mCi mmol⁻¹), urea (55 mCi mmol⁻¹) and mannitol (57 mCi mmol⁻¹) were purchased from Amersham Life Science Inc. (Arlington Heights, IL, USA). ¹⁴C-labelled ethylene glycol (51 mCi mmol⁻¹), non-radioactive sugars and phloretin were purchased from Sigma.

Solute permeability

The diffusive solute permeability coefficient P (cm s⁻¹) was determined by using the relation: P= J/(AΔc), where J is the radiotracer uptake (mol oocyte⁻¹ s⁻¹); A is the oocyte membrane area (0.4 cm², Zampighi et al. 1995); and Δc is the concentration difference of the solute across the membrane (mol cm⁻³).

RESULTS

The passive endogenous permeability of control Xenopus oocytes for organic solutes was studied by radiotracer experiments. Groups of eight oocytes were incubated over 30 min in 55 μm¹⁴C-labelled urea, mannitol or ethylene glycol. Whereas urea (1 × 10⁻⁷ cm s⁻¹, n = 18; where n is the number of experiments) and mannitol (0.8 × 10⁻⁷ cm s⁻¹, n = 6) had similar low permeabilities through the oocyte plasma membrane, the permeability for ethylene glycol was higher (1.6 × 10⁻⁶ cm s⁻¹, n = 5). In control oocytes the permeability for urea, mannitol and ethylene glycol was linear with increasing concentration. For oocytes from the same batch, the urea uptake was 1.3 ± 0.2 and 2900 ± 600 pmol oocyte⁻¹ (30 min)⁻¹, the mannitol uptake was 1.1 ± 0.1 and 2000 ± 180 pmol oocyte⁻¹ (30 min)⁻¹, and the ethylene glycol uptake was 20 ± 0.3 and 36 300 ± 900 pmol oocyte⁻¹ (30 min)⁻¹ at 55 μm and 100 mm, respectively.

As illustrated in Fig. 1 the uptakes increased in SGLT1 or C5-expressing oocytes. Shown is one specific experiment performed 4 days post-cRNA injection when the expression level of rbSGLT1 measured by the uptake in 55 μmαMDG was 156 ± 12 pmol oocyte⁻¹ (30 min)⁻¹. The total urea uptake by rbSGLT1 cRNA-injected oocytes from the same batch was 4 ± 0.5 vs. 0.8 ± 0.1 pmol oocyte⁻¹ (30 min)⁻¹ in control oocytes. The urea uptake by C5-expressing oocytes was 6.3 ± 1.4 pmol oocyte⁻¹ (30 min)⁻¹. After subtraction of the background urea permeability of control oocytes from the total urea uptakes in SGLT1or C5-expressing oocytes, the specific urea uptake was 3.2 ± 0.5 and 5.5 ± 1.5 pmol oocyte⁻¹ (30 min)⁻¹ for SGLT1 and C5, respectively. The 55 μm urea or αMDG uptakes through C5 were similar. In three independent experiments the C5-mediated urea influx was 7.0 ± 1.6, 2.3 ± 0.3 and 6.45 ± 1.4 pmol oocyte⁻¹ (30 min)⁻¹, and the C5-mediated αMDG influxes was 4 ± 0.4, 6.6 ± 2.7 and 3.6 ± 1.4 pmol oocyte⁻¹ (30 min)⁻¹.

Figure 1. Organic solute transport through SGLT1 and C5.

A, passive urea transport by rbSGLT1 and rbC5. Uptakes of 55 μm¹⁴C-urea into non-injected oocytes (NI) and oocytes from the same batch injected with 50 ng of rbSGLT1 cRNA (SGLT1) or rbC5 cRNA (C5) were followed for 30 min and measured 0.8 ± 0.1, 4 ± 0.5 and 6.3 ± 1.4 pmol oocyte⁻¹, respectively (left). The uptake of ¹⁴C-αMDG into rbSGLT1-expressing oocytes was 156 ± 12 pmol oocyte⁻¹ (30 min)⁻¹ (right). Error bars are s.e.m. between the oocytes in each group, n = 8. The experiment was performed 4 days post-cRNA injection. B, mannitol transport. The total 55 μm¹⁴C-mannitol uptake measured over 30 min in non-injected (NI), rbSGLT1-(SGLT1), or rbC5 cRNA-injected oocytes (C5) was 0.8 ± 0.1, 0.9 ± 0.1 and 2.5 ± 0.6 pmol oocyte⁻¹, respectively. The SGLT1 uptake in 55 μmαMDG was 178 ± 11 pmol oocyte⁻¹ (30 min)⁻¹. The experiment was performed 4 days post-cRNA injection. C, ethylene glycol transport. The 55 μm total ¹⁴C-ethylene glycol uptake in non-injected, SGLT1 and C5-expressing oocytes measured 18.5 ± 0.5, 27 ± 1.5 and 22.5 ± 1.5 pmol oocyte⁻¹ (30 min)⁻¹, respectively. In this batch the level of SGLT1-mediated uptake in 55 μmαMDG was 315 ± 23 pmol oocyte⁻¹ (30 min⁻¹).

Similar to the increase in urea uptake, the mannitol and ethylene glycol uptakes also increased in C5 (or SGLT1)-expressing oocytes (Fig. 1B and C), compared with those in control oocytes. The uptake of 55 μm mannitol was 0.8 ± 0.1 pmol oocyte⁻¹ (30 min)⁻¹ in control oocytes and 2.5 ± 0.6 pmol oocyte⁻¹ (30 min)⁻¹ in C5-expressing oocytes (p value 7 × 10⁻⁵). SGLT1 expressing oocytes did not show a significant increase in the mannitol uptake (0.9 ± 0.1 pmol oocyte⁻¹ (30 min)⁻¹) vs. control oocytes (p value 0.04). The urea-specific uptake for SGLT1 and C5 in this batch was 5.5 ± 0.3 and 6.7 ± 3 pmol oocyte⁻¹ (30 min)⁻¹, respectively, and the αMDG uptake by SGLT1 at 55 μm was 178 ± 11 pmol oocyte⁻¹ (30 min)⁻¹. The mean values from six mannitol uptake experiments were 1.1 ± 0.2 and 1.1 ± 0.1 pmol oocyte⁻¹ (30 min)⁻¹ for SGLT1 and control oocytes, respectively. Comparison of the uptake in mannitol and urea showed that the specific C5 mannitol uptake was ∼46 ± 8 % of the specific C5 urea uptake (five experiments). The C5-specific mannitol influx was linear from 55 μm to 100 mm, e.g. in one experiment it was 4 ± 1 pmol oocyte⁻¹ (30 min)⁻¹ at 55 μm and 6727 ± 1820 pmol oocyte⁻¹ (30 min)⁻¹ at 100 mm mannitol.

The highest endogenous solute permeability of control oocytes was observed with ethylene glycol. Whereas the ethylene glycol uptake in SGLT1-expressing oocytes increased from the endogenous background (18.5 ± 0.5 pmol oocyte⁻¹ (30 min)⁻¹) to 27 ± 1.5 pmol oocyte⁻¹ (30 min)⁻¹ (p value 4 × 10⁻⁵), the C5-mediated ethylene glycol uptake was 22 ± 1.5 pmol oocyte⁻¹ (30 min)⁻¹, and differed only slightly from the ethylene glycol background (p value 0.04; Fig. 1C).

We proceeded further with a more detailed characterization of the urea influx mediated by C5. The C5 urea uptake increased linearly with increasing urea concentration (from 55 μm to 100 mm). For example the specific urea influx in C5-expressing oocytes increased from 13 ± 2 to 21 300 ± 3600 pmol oocyte⁻¹ (30 min)⁻¹ when the concentration was increased from 55 μm to 100 mm urea. Increasing the incubation time from 30 to 60 min doubled the increase in the C5 urea uptake from 7 ± 2 to 11 ± 2 pmol oocyte⁻¹. Figure 2A shows that the C5 urea uptake was Na⁺ independent (endogenous urea permeability was also Na⁺ independent). The C5 urea influx was 5.2 ± 0.7 pmol oocyte⁻¹ 30 min⁻¹ in 100 mm choline, and 4.2 ± 0.3 pmol oocyte⁻¹ (30 min)⁻¹ in 100 mm Na⁺. The specific urea uptake by rbSGLT1 was 7 ± 1 and 6 ± 0.5 pmol oocyte⁻¹ (30 min)⁻¹ in choline and Na⁺, respectively. Figure 2B illustrates that 1 mm phloretin reduced the C5 urea uptake by ∼45 % (from 9.5 ± 0.5 to 5 ± 1 pmol oocyte⁻¹ (30 min)⁻¹, left panel). In two additional experiments the inhibition of the urea uptake was 33 and 28 %. The right panel in Fig. 2B shows that 1 mm phlorizin did not affect the C5 urea uptake; i.e. 4.1 ± 0.4 and 4.4 ± 0.5 pmol oocyte⁻¹ (30 min)⁻¹ in the absence and presence of phlorizin, respectively.

Figure 2. Characteristics of the urea transport by C5.

A, urea transport in choline and sodium. C5-expressing oocytes (C5) incubated for 30 min in 55 μm¹⁴C-urea in 100 mm choline chloride transported 5.2 ± 0.7 pmol oocyte⁻¹, and 4.2 ± 0.3 pmol oocyte⁻¹ when incubated in 55 μm¹⁴C-urea in 100 mm NaCl (5 days post-cRNA injection); both results were corrected for the endogenous oocyte urea transport. SGLT1-expressing oocytes (SGLT1) from the same batch transported under the same conditions 7 ± 1 pmol oocyte⁻¹ urea in choline chloride and 6 ± 0.5 pmol oocyte⁻¹ in NaCl. The endogenous urea uptake (NI) was the same in 100 mm Na⁺ or choline and measured 1.4 ± 0.2 pmol oocyte⁻¹ (30 min)⁻¹. B, phloretin (PT) and phlorizin (PZ) inhibition of the C5 urea transport. Since the left and right panels show results from representative experiments performed with different batches of oocytes, the absolute values were normalized to the amount of the C5-mediated uptakes in the absence of inhibitor, which was taken as 100 %. Left, 30 min uptake of 55 μm¹⁴C-urea into oocytes expressing C5 measured 6 days post-cRNA injection. The urea influx of 9.5 ± 0.5 pmol oocyte⁻¹ in 100 mm Na⁺ (100 ± 6 %) was reduced to 5 ± 0.7 pmol oocyte⁻¹ (53.2 ± 7.4 %) after addition of 1 mm phloretin. The endogenous urea uptake (NI) remained the same in the presence and absence of 1 mm phloretin (1 ± 0.1 pmol oocyte⁻¹). The rbSGLT1-mediated urea uptake in this experiment was 7.6 ± 0.5 pmol oocyte⁻¹ (30 min)⁻¹ (not shown). Right, 30 min uptakes of 55 μm¹⁴C-urea were studied in the presence and absence of 1 mm phlorizin 3 days post-cRNA injection. The oocytes expressing C5 transported 4.1 ± 0.4 pmol oocyte⁻¹ (100 ± 7 %) in the absence and 4.4 ± 0.5 pmol oocyte⁻¹ (107.3 ± 12.2 %) in the presence of 1 mm phlorizin. Urea uptake into non-injected (NI) oocytes was 1.1 ± 0.1 pmol oocyte⁻¹ (30 min)⁻¹ in the presence or in the absence of the inhibitor.

The activation energy (E_a) of the passive C5 urea transport was estimated by measuring the 55 μm¹⁴C uptake of groups of eight oocytes over 30 min at 12, 22 and 30 °C. E_a values were determined from the slope of the corresponding Arrhenius plots (E_a=R× slope, where R is the gas constant). For C5-expressing oocytes the slope of -2.5 corresponds to an E_a of 5 ± 3 kcal mol⁻¹ and for control oocytes the slope of -8 corresponds to a higher E_a of 16 ± 0.5 kcal mol⁻¹.

The C5 urea uptake increased in the presence of αMDG (Fig. 3). The 55 μm C5 urea uptake doubled from 4 ± 0.4 to 9 ± 2 pmol oocyte⁻¹ (30 min)⁻¹ in the presence of 25 mmαMDG (left panel). In the same batch of oocytes the C5 αMDG uptake at 25 mmαMDG increased to 875 ± 212 from 7 ± 2 pmol oocyte⁻¹ (30 min)⁻¹ at 55 μm (right panel). Since mannitol was also transported by C5, we studied the effect of mannitol on the urea influx. Addition of 25 mm cold mannitol did not alter the C5 urea transport (15 ± 2.5 vs. 20 ± 2.5 pmol oocyte⁻¹ (30 min)⁻¹, p value 0.11), whereas the urea uptake in the presence of 25 mmαMDG was higher (22 ± 1 vs. 15 ± 2.5 pmol oocyte⁻¹ (30 min)⁻¹, p value 0.04). In the same batch of oocytes the C5-mediated uptake in 55 μm mannitol or αMDG was 8.4 ± 2 and 251 ± 11 pmol oocyte⁻¹ (30 min)⁻¹, respectively. The endogenous urea background did not increase with addition of cold mannitol (1.3 ± 0.1 vs. 1.6 ± 0.2 pmol oocyte⁻¹ (30 min)⁻¹).

Figure 3. C5 urea transport in the presence of αMDG.

The 55 μm¹⁴C-urea uptake in the absence and presence of 25 mmαMDG (100 mm Na⁺) was measured over 30 min in the same batch of non-injected (NI) or C5-expressing oocytes (C5) 4 days post-cRNA injection. The specific C5 urea uptake was 4 ± 0.4 pmol oocyte⁻¹ in the absence and 9 ± 2 pmol oocyte⁻¹ in the presence of 25 mmαMDG (left), whereas the uptake in the non-injected oocytes was the same for both conditions (1.1 ± 0.1 pmol oocyte⁻¹, not shown). The specific C5-mediated uptake in the presence of 55 μm¹⁴C-αMDG (right) was 7 ± 2 pmol oocyte⁻¹ and increased to 875 ± 212 pmol oocyte⁻¹ in 55 μm¹⁴C-αMDG containing 25 mm cold αMDG.

DISCUSSION

Permeability of native oocytes, and oocytes expressing channels and transporters

Water and polar, non-charged organic solute molecules have a low permeability through the lipid plasma membrane of cells. Indeed, the plasma membrane of native oocytes has a low permeability to water and some polar solutes such as urea and mannitol. These molecules appear to permeate via diffusion through the lipid membrane as the activation energy for water is E_a > 10 kcal mol⁻¹ (Wright & Loo, 2000) and the E_a for urea is ∼14 kcal mol⁻¹ (Leung et al. 2000).

When aquaporins and secondary active transporters are integrated into the oocyte membrane as a result of recombinant expression, the water and urea permeability increase and this is accompanied by a decrease in E_a. For example, the E_a of water transport through oocyte membranes expressing CHIP28 (AQP1) water channels was only 3-4 kcal mol⁻¹ (Preston et al. 1992). Studies with expressed transporters have demonstrated also their ability to increase the water permeability of the plasma membrane and to act as low conductance water channels in the absence of their natural substrates. Some examples include the facilitated glucose transporter GLUT1 (Fishbarg et al. 1990), the cystic fibrosis transmembrane conductance regulator CFTR (Hasegawa et al. 1992, Schreiber et al. 1997), the Na⁺-Cl⁻-GABA GAT1 transporter (Loo et al. 1999), the Na⁺-carboxylate cotransporter NaDC-1 (Meinild et al. 2000) and the plant amino acid transporter AAP5 (Loo et al. 2000).

Permeability of SGLT1

An increased passive water permeability has also been observed in the case of the sodium-glucose cotransporter SGLT1 (Zampighi et al. 1995; Loo et al. 1996, 1999) and it also was accompanied by a decrease of the E_a from 10 to ∼5 kcal mol⁻¹ (Loo et al. 1999). This value is close to the energy required by water channels. Recent measurements of the passive urea permeability in SGLT1 determined that the E_a for urea transport across the oocyte membrane was reduced from 14 ± 3 kcal mol⁻¹ for non-injected oocytes to 6 ± 1 kcal mol⁻¹ for oocytes expressing SGLT1 (Leung et al. 2000).

In this study we found that another small organic molecule, ethylene glycol, could also permeate through SGLT1 (Fig. 1C) and that mannitol could not (Fig. 1B). This indicates that the radius of the channel through SGLT1 is < 4 Å (< 0.4 nm) and this size is in agreement with the formamide reflection coefficient of SGLT1, σ= 0.37 (Meinild et al. 1998a).

Permeability of C5

Despite the diverse data for an increased permeability of water and small polar organic molecules through SGLT1 and other transporters, there is no information about the structural localization of these channel-like pathways within the transporter molecules. In this study we have shown that the last five carboxy-terminal helices (the C5 protein) may constitute the passive urea channel of native SGLT1. C5 has previously been described as a low affinity glucose uniporter with stereospecific, selective and cation-independent sugar transport (K_0.5 for αMDG ∼50 mm), which was inhibited by phloretin (K_i≈ 1 mm) but not by phlorizin. It has been concluded that C5 contains the sugar permeation pathway of SGLT1 (Panayotova-Heiermann et al. 1997, 1999). In the present study, the C5 urea transport characteristics were found to be almost identical to these describing the passive urea transport through intact SGLT1 (Table 1): (i) low E_a of 5 ± 3 kcal mol⁻¹; (ii) linear to 100 mm; (iii) independent from the applied cation Na⁺ or choline (Fig. 2A), and (iv) inhibited by 1 mm phloretin, the classical inhibitor for urea transport (Macey, 1984) but not by phlorizin (Fig. 2B). The binding affinity of phlorizin (K_i) to SGLT1 is dependent on the conformational changes induced by the binding of Na⁺, but the urea transport in both SGLT1 and C5 is Na⁺ independent (Leung et al. 2000). In addition, the Na⁺-binding site is not intact in C5, which may explain the insensitivity of the urea transport to phlorizin. Similarly, the C5-mediated αMDG transport was only inhibited by phloretin but not phlorizin (Panayotova-Heiermann et al. 1997). Unfortunately, our attempts to express the nine amino-terminal helices of SGLT1 (N9) as an individual protein in order to provide a negative control for the urea transport through SGLT1 remained unsuccessful. The lack of an antibody for the N9 region of the transporter left the question open as to whether N9 is synthesized and then quickly degraded in the oocytes, or is misfolded and/or not integrated properly into the plasma membrane.

Table 1.

Summary of the passive urea and water transport characteristics of expressed SGLT1 and C5, and those of native (non-injected) oocytes (NI)

	NI		SGLT1			C5(this study)
	Passive urea transport	Passive water transport	Na+ uniport	Passive water transport	Passive urea transport	Passive urea transport
Activation energy(kcal mol−1)	14 ± 3(Leung et al. 2000 and this study)	12 ± 2(Loo et al. 1996)	19 ± 1(Loo et al. 1999)	5 ± 1(Loo et al. 1999)	6 ± 1(Leung et al. 2000)	5 ± 3
Cation dependence	Independent(this study)	Independent(D. D. F. Loo*)	Dependent(P.-Heiermann et al. 1998)	Independent(Loo et al. 1999)	Independent(Leung et al. 2000)	Independent
Concentration dependency	Linear(Leung et al. 2000 and this study)	Linear(Meinild et al. 1998b)	Saturable(P.-Heiermann et al. 1998)	Linear(Meinild et al. 1998a)	Linear(Leung et al. 2000)	Linear
Phlorizin inhibition	Not inhibited (Leung et al. 2000 and this study)	Independent(D. D. F. Loo*)	Ki≈ 1.5 μM (P.-Heiermann et al. 1998)	Ki≈ 5 μM(no inhibition in choline, Loo et al. 1999)	Ki≈ 1 μM(no inhibition in choline, Leung et al. 2000)	Not inhibited
Phloretin inhibition	Not inhibited(Leung et al. 2000)	Not inhibited(Loo et al. 1999)	Ki≈ 100 μM(Loo et al. 1999)	Ki > 100 μM(Loo et al. 1999)	Ki≈ 1 mM(Leung et al. 2000)	Ki≈ 1 mM

^*Unpublished observations.

The urea and water channel activity of SGLT1 is inhibited by phlorizin with the same affinity (1-5 μm, Table 1), and the existence of one inhibitor for both pathways suggests that SGLT1 possesses one channel for the passive permeation of both substrates (water and urea). In contrast, the urea and water channel activity is dissociated functionally from the Na⁺ uniport activity by these same characteristics, activation energy (19 kcal mol⁻¹ for Na⁺ uniport vs. 5-6 kcal mol⁻¹ for water or urea transport by SGLT1 or C5), concentration dependency and phlorizin/phloretin inhibition (Table 1). The Na⁺ insensitivity (or cation independence) may suggest that the passive water flow through the SGLT1 and C5 urea channel differ structurally from the Na⁺ uniport pathway.

The molecular radius of the C5 urea channel is probably larger than that of SGLT1 since C5 transported mannitol, and SGLT1 did not (Fig. 1B). The permeability of C5 was highest in urea > mannitol > ethylene glycol (Fig. 1), where mannitol transport was almost two times lower than that of urea at 55 μm. The results for the ethylene glycol transport were ambiguous because of the high endogenous background of native oocytes, but small amounts seemed to be transported by SGLT1 (Fig. 1C).

Cotransporters, in addition to acting as channels for urea and water also behave as water and urea cotransporters (Loo et al. 1996; Meinild et al. 1998a; Leung et al. 2000; Zeuthen et al. 2001), i.e. the addition of substrates, e.g. glucose in the case of SGLT1, increase the transport of urea and water. Our explanation for this phenomenon is that the turnover of the transport cycle is accompanied by large asymmetrical changes in cotransporter conformation. Our present study also shows that the addition of substrate increases C5-mediated urea transport (Fig. 3). αMDG, but not mannitol, stimulated urea uptake into oocytes expressing C5. This suggests that the conformational changes that underlie sugar uniport by the truncated protein also increase urea transport, and implies that the pathway for urea (and possibly also water) cotransport through SGLT1 is formed by the C-terminal five transmembranes helices.

In summary, our results demonstrated that the passive water and urea transport by SGLT1 and the passive urea transport by C5 share the same transport characteristics, and it appears that transmembrane segments 10-14, which comprise the C5 protein, are the relevant part of SGLT1 that contains the passive water and urea pathways. In contrast, the SGLT1 Na⁺ uniport pathway is distinct from them based on activation energy, cation dependence and saturability. This suggests that the passive water and urea channel pathways are localized in C5 and do not structurally overlay the Na⁺ uniport pathway.