INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

NEUTRAL -D-glucosidase is a membrane-bound disaccharidase that catalyzes maltose hydrolysis into two glucose residues (26). It has been suggested that the products of hydrolysis derived from disaccharidase activities have a "kinetic advantage" for entry into cells over the same free hexoses added to the medium (17). Moreover, these brush-border membrane hydrolases would act on the substrate at the outer surface of the membrane and transfer their hydrolysis products directly across the membrane independently of the normal hexose transport process (17, 23). Lee and Forstner (16) proposed an additional function in the monosaccharide transport for the large hydrophobic domain (14 kDa) of rat intestinal maltase-glucoamylase. Other investigations support the hypothesis that the hydrolysis products cross the membranes via specific transporters that seem to be structurally and functionally linked to the enzyme in an enzyme-transport complex (28, 29). It has also been postulated that disaccharide hydrolysis occurs at the outer surface of the membrane but close to the transport carrier systems; the newly generated monosaccharides would thus have a better chance of being transported by the membrane transport systems (8, 25). Parsons and Prichard (21) concluded that this kinetic advantage was due to the presence of a "hexose unit pool" near the outer face of the membrane, which would also be accessible to the free glucose present in the intestinal lumen. However, no evidence of the existence of this pool has been reported. Moreover, these experiments were performed with intestine fragments, and the contribution of other components, such as glycocalyx and mucus, to hexose uptake or modification of the local membrane environment cannot be excluded (1).

The aim of this study was to investigate whether neutral -D-glucosidase participates in the transport of its hydrolysis products, alone or in combination with the sodium-glucose cotransporter present in brush-border membranes. Because the possible role of neutral -D-glucosidase in transport of its hydrolysis products is necessarily associated with insertion of this enzyme in the bilayer membrane, we propose an artificial membrane model composed of lecithin vesicles that allows incorporation of neutral -D-glucosidase and other membrane proteins in the same topology as in native membranes. Use of these artificial membrane vesicles also permits specific modifications in the topology of neutral -D-glucosidase without changing the surface arrangement of the other membrane proteins. Thus this model was used to determine whether modification of the topology of neutral -D-glucosidase affects the uptake of free glucose and glucose released by maltose hydrolysis.

We previously demonstrated the similarities between the molecular and immunologic properties of renal neutral -D-glucosidase and intestinal glucoamylase (9). Consequently, this work was performed on brush-border membrane vesicles purified from horse kidney cortex, rather than small intestine, for several reasons: 1) the absence of phlorizin hydrolase on kidney brush-border membranes allows the use of phlorizin, a potent inhibitor of neutral -D-glucosidase (13) and sodium-glucose cotransporter (30), 2) intestinal sodium-glucose cotransport is functionally inactivated when reconstituted into liposomes, and 3) the choice of kidney avoids any interference with mucus and glycocalyx (1, 2).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Membrane preparations. Brush-border membrane vesicles from horse kidney cortex prepared according to Boudouard et al. (3) were suspended in 150 mM NaCl-10 mM Tris · HCl (pH 8.0). Membrane proteins (10 mg/ml) were solubilized with Triton X-100 in the same buffer for 1 h at 4°C at a 1:1 detergent-to-protein weight ratio. After centrifugation at 105,000 g for 60 min, the supernatant containing solubilized proteins was dialyzed overnight against 100 mM sodium phosphate buffer (pH 6.8) containing 0.1% Triton X-100.

Neutral -D-glucosidase preparations. Neutral -D-glucosidase was purified in its "integral form" and "hydrophilic form." The difference between these forms of the enzyme was the proteolytic elimination from the integral form of the hydrophobic peptide.

Preparation of proteoliposomes. Proteoliposomes were prepared as previously described (4) using the bulk of solubilized brush-border membrane proteins for reconstitution. A specific reconstitution protocol was used for transport experiments (see Transport experiments). The pellets containing the artificial membrane vesicles were resuspended in an appropriate buffer, washed, then adjusted to a protein concentration of 2-5 mg/ml. When necessary, the bulk of solubilized brush-border membrane proteins was reconstituted with phosphatidylcholine containing a trace of [¹⁴C]phosphatidylcholine.

Proteolytic treatments. Brush-border membrane vesicles and proteoliposomes reconstituted with the bulk of solubilized brush-border membrane proteins, including neutral -D-glucosidase, were digested with papain, as previously described (13). Proteinase K treatment was performed by addition of 5% protease (wt/wt) to the membrane or proteoliposome suspensions in 100 mM NaCl-50 mM Tris · HCl (pH 7.0). For both proteolytic treatments, the mixtures were incubated at 37°C for various times. Aliquots (100 µl) were removed, diluted 10-fold with cold 100 mM sodium phosphate buffer (pH 6.2) containing 5 mM phenylmethylsulfonyl fluoride, and centrifuged at 105,000 g for 45 min for membranes and 2 h for proteoliposomes. Neutral -D-glucosidase activity was then determined in each supernatant. No inhibition of neutral -D-glucosidase activity was observed after proteolytic treatment or addition of protease inhibitors. Total (100%) initial enzyme activity was defined as the enzyme activity of brush-border membrane vesicles or proteoliposomes treated with the proteinases without centrifugation. Proteoliposomes were characterized before and after proteolytic treatment by centrifugation on a linear sucrose gradient (0-1.5 M), as described elsewhere (4).

Immunotitration studies. The antiserum, raised against only the hydrophilic form of the enzyme (13), was diluted 60-fold in PBS (pH 7.4), dialyzed against the same buffer, and centrifuged at 105,000 g for 30 min before use. Immunotitrations were performed in a 250-µl volume in 0.75-ml capped centrifuge tubes according to Tsuboi et al. (27). Briefly, the hydrophilic form of the enzyme (0.015 IU) was incubated for 24 h at 4°C with increasing amounts of antiserum. After centrifugation at 18,000 g for 45 min, titration curves for the hydrophilic form of the enzyme were obtained by determining the neutral -D-glucosidase activity remaining in the supernatant. Twice the amount of specific antiserum, which was used for free enzyme titration, was first depleted for 24 h at 4°C with membrane-bound neutral -D-glucosidase (0.015 IU) from native brush-border membrane vesicles and from proteoliposomes constituted with the bulk of solubilized brush-border membrane proteins. After centrifugation at 105,000 g (45 min for membranes, 120 min for proteoliposomes), the supernatants were completely transferred to new tubes. The remaining antibody was then backtitrated by addition of a constant amount (0.015 IU) of the hydrophilic form of the neutral -D-glucosidase. After incubation for 24 h at 4°C and centrifugation at 18,000 g for 45 min, the supernatants were removed and the neutral -D-glucosidase activity was assayed. The immunotitration curves represent the relationship between the amount of antibody added and immunoprecipitation with 0.015 IU of the hydrophilic form of neutral -D-glucosidase alone (see Fig. 2A) or with 0.015 IU of enzyme integrated in native membrane (see Fig. 2B) or proteoliposomes reconstituted with the bulk of solubilized brush-border membrane proteins (see Fig. 2C). Immunotitration curves for neutral -D-glucosidase bound to native or artificial membranes (dashed curves in Fig. 2, B and C) were obtained by calculating the difference between the titration curves obtained with the free enzyme alone and those obtained with the free enzyme combined with equivalent amounts of enzyme bound to native or artificial membranes. In all conditions, no immunoprecipitation could be observed when nonimmune rabbit serum was substituted for the neutral -D-glucosidase antiserum.

Transport experiments. Three artificial membrane vesicle preparations were used for transport studies. In preparation 1 the purified integral form of neutral -D-glucosidase was added to the nonadsorbed protein from the affinity chromatography column to obtain the same specific activity as that measured before this step, and the protein mixture was then reconstituted in proteoliposomes. In preparation 2, only the proteins that were not retained by the affinity chromatography column were reconstituted; the purified hydrophilic form of neutral -D-glucosidase was then added to the medium to obtain the same specific activity as in proteoliposome preparation 1. In preparation 3, only the proteins not retained by the affinity chromatography step were reconstituted. All three of these artificial membrane vesicle preparations were then adjusted to a protein concentration of 1-1.5 mg/ml.

Free D-glucose uptake. Free D-glucose uptake by proteoliposome preparations 1 and 2 was measured by an equilibrium tracer exchange procedure, as reported in previous studies (22). Briefly, 30-45 µg of proteoliposome preparation (30 µl) were preincubated for 2 h at 4°C in 100 mM KCl or NaCl, 100 mM mannitol, and 10 mM HEPES · Tris (pH 7.4) containing 1.5 mM unlabeled D-glucose (60 µl). Isotope exchange was initiated by addition of 15 µl of 1 mM radioactive sugar in the same medium (activity of D-[³H]glucose = 2.5 × 10⁵ cpm/assay). Transport was terminated with 1 ml of cold stop solution (KCl or NaCl medium supplemented with 1 mM phlorizin, a potent inhibitor of sodium-glucose cotransporter and neutral -D-glucosidase activity). In all the transport experiments, after addition of the stop solution, the proteoliposomes were retained by filtering the resulting suspensions through a 0.65-µm presoaked Sartorius filter, then washed twice with 2 ml of cold stop solution. Elapsed time for the whole procedure did not exceed 15 s. The filtrates were retained, and the filters were dried and counted for radioactivity in a Packard Spectrometer after addition of scintillation liquid (Pico-Fluor 30, Packard).

Uptake of D-glucose produced by maltose hydrolysis. The uptake of D-glucose produced by maltose hydrolysis was determined using 1) the three proteoliposome preparations (30-45 µg protein/assay) and 2) proteoliposomes constituted with liposomes in which only the purified neutral -D-glucosidase was inserted and with liposomes (without any protein); in this case, the purified hydrophilic form of neutral -D-glucosidase was added to the medium to obtain the same enzyme activity as in the proteoliposomes containing only the disaccharidase. Both proteoliposome preparations were preincubated for 2 h at 25°C in 100 mM NaCl, 100 mM mannitol, and 10 mM HEPES · Tris (pH 7.4). Then 12.5 mM radiolabeled maltose (activity >2 × 10⁶ cpm/assay) in the same medium was added, and liposomes were incubated for 3 min at 25°C. The reaction was terminated with 1 ml of cold stop solution (KCl medium supplemented with 1 mM phlorizin).

Assays. Proteins were determined in the presence of 3% SDS, as described by Gerritsen et al. (12); neutral -D-glucosidase activity was measured as previously described (9); glucose produced by maltose hydrolysis was assayed in each filtrate of the uptake experiments according to the Dahlqvist method (6).

Electrophoresis. SDS-PAGE was performed following the procedure described by O'Farrell (19) after treatment of samples (80 µg) with 2% SDS. The running gel was a 5-12% (wt/vol) linear polyacrylamide gradient. After electrophoresis the gels were stained for protein by Coomassie Blue R250.

Statistics. Values are means ± SE and were evaluated for statistical significance using Student's unpaired t-test and considered significant at P < 0.05.

Chemicals. L--Phosphatidylcholine (type VE), phenylmethylsulfonyl fluoride, and phlorizin were obtained from Sigma Chemical, Biobeads SM-2 from Bio-Rad, and proteinase K from Boehringer. 1,2-Di-[1-¹⁴C]palmitoyl-L-3-phosphatidylcholine, D-[6-³H]glucose, L-[1-¹⁴C]glucose, and [U-¹⁴C]maltose were purchased from Amersham. This disaccharide was free of any glucose contamination, as confirmed by TLC on silica gel plates (from Merck) using 60:40:30 (vol/vol/vol) n-butanol-pyridine-distilled water. All other reagents have been previously described (13) or were of the best available purity.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Previous studies on proteoliposomes obtained from solubilized kidney brush-border membranes have shown that 70% of membrane proteins may be extracted by 1% Triton X-100 and that 32% of solubilized proteins are inserted in liposomes (4, 22). About 75-85% of neutral -D-glucosidase was incorporated under the same conditions at the lipid-to-protein ratio used (Table 1). To study the orientation of this enzyme in reconstituted membrane vesicles, neutral -D-glucosidase activity was measured before and after proteoliposome solubilization by 2% Triton X-100 for 18 h at 4°C. As seen in Table 1, Triton X-100 treatment increased neutral -D-glucosidase activity only 2-5%, consistent with asymmetric "right-side-out" integration of this disaccharidase in proteoliposomes.

View this table: [in this window] [in a new window]   Table 1.   Effect of lipid-to-protein ratios on neutral -D-glucosidase incorporation and latency of enzyme activity

Native brush-border membranes and proteoliposomes obtained from the bulk of Triton X-100-solubilized membrane proteins were digested by papain and proteinase K for various times. After centrifugation at 105,000 g, neutral -D-glucosidase activity was measured in the supernatants. The kinetics of solubilization of neutral -D-glucosidase from both membrane preparations were similar after treatment by papain (Fig. 1A) and proteinase K (Fig. 1B). Nearly 95% of the maltase activity was solubilized after 60 min of proteolysis; this complete solubilization was confirmed by density gradient analysis of proteoliposomes digested for 1 h by these two proteinases (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Kinetic curves of proteolytic solubilization of neutral -D-glucosidase from membrane vesicles and proteoliposomes. Neutral -D-glucosidase solubilization from native membranes () and proteoliposomes () is shown with papain (A) and proteinase K (B). Values are means ± SE of 3 independent experiments.

Further information about the topology of neutral -D-glucosidase in native membrane vesicles or inserted in proteoliposomes prepared as mentioned above is provided by the immunotitration curves of the neutral -D-glucosidase plotted in Fig. 2. The titration curve of the hydrophilic form (or free form) of neutral -D-glucosidase (0.015 IU) is shown in Fig. 2A. All enzyme activity could be precipitated after addition of 50 µl of anti-neutral -D-glucosidase. Figure 2, B and C, illustrates the relationships between the addition of antiserum and precipitation of 0.015 IU of free enzyme in combination with an equivalent 0.015 IU of native (Fig. 2B) and artificial (Fig. 2C) membrane-bound enzyme. Similar titration curves were obtained when free neutral -D-glucosidase was combined with the enzyme bound to the native and artificial membranes. Immunotitration curves for neutral -D-glucosidase bound to native or artificial membranes can be obtained by calculating the difference between the titration curves obtained with the free enzyme alone (Fig. 2A) and those obtained with the free enzyme combined with equivalent amounts of the enzyme bound to native or artificial membranes. The immunotitration curves for neutral -D-glucosidase bound to the native (dashed curve in Fig. 2B) or artificial (dashed curve in Fig. 2C) membranes obtained in this manner were similar and were not significantly different from the titration curve obtained with the free enzyme alone. The antibodies thus recognized the same antigenic determinants of the enzyme inserted in both types of membranes. In addition, the whole enzyme, buried in native or artificial membranes, displayed the same immunotitration curves as the hydrophilic form alone. No significant precipitation of the enzyme activity was observed when a nonimmune serum was substituted for the neutral -D-glucosidase antiserum (Fig. 2A).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Immunotitration curves of free and membrane-bound neutral -D-glucosidase. A: immunotitration curves of 0.015 IU of free neutral -D-glucosidase (); B: membrane-bound enzyme (0.015 IU) plus an equivalent 0.015 IU of free enzyme (); C: 0.015 IU of neutral -D-glucosidase reintegrated, in combination with other integral membrane proteins in proteoliposomes, plus an equivalent 0.015 IU of free enzyme (). Immunotitration curves of neutral -D-glucosidase bound to native (dashed curves in B) or artificial membranes (dashed curves in C) were obtained as described in MATERIALS AND METHODS. , Nonimmune serum.

In an initial step, free D-glucose uptake by proteoliposomes containing the neutral -D-glucosidase (preparation 1) was compared with uptake in proteoliposomes without the enzyme incorporated in their structure but present in the hydrophilic form in the incubation medium (preparation 2). D-Glucose entry into both proteoliposome preparations was measured by the equilibrium tracer exchange procedure, a suitable method for determining the uptake of D-glucose into proteoliposomes (22) and for elucidating the mechanism of multireactant systems (14). In artificial membrane preparations 1 and 2 the uptake of labeled D-glucose was increased in vesicles suspended in NaCl medium compared with vesicles prepared in KCl medium (Fig. 3). Moreover, as previously demonstrated by this procedure, glucose uptake in such proteoliposomes was inhibited by phlorizin (22); the equilibrium value of the glucose isotope in internal and external spaces is reached after 2 h of incubation. D-Glucose was accumulated with the same velocity in both proteoliposome preparations, and the equilibrium levels reached after 2 h of incubation revealed similar solute concentrations in the intravesicular space (0.600 ± 0.015 nmol/mg protein). Replacement of mannitol by Tris in the incubation medium resulted in 82 ± 3% inhibition of neutral -D-glucosidase activity, regardless of the location of the enzyme, with no effect on the kinetics of D-glucose accumulation (Fig. 3).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   D-Glucose equilibrium exchange in proteoliposomes. Proteoliposome preparation 1, prepared from bulk of integral brush-border membrane proteins (A), and preparation 2, prepared from bulk of integral brush-border membrane proteins without neutral -D-glucosidase, which was added in its "hydrophilic form" in the incubation medium (B), were preincubated for 2 h at 4°C in 100 mM NaCl with () and without () 0.1 mM phlorizin or 100 mM KCl (), 100 mM mannitol, and 10 mM HEPES · Tris (pH 7.4) containing 1 mM D-glucose. Isotope exchange was initiated by addition of 1:7 volume of 1 mM radioactive D-glucose in same buffer used during preincubation. Uptake measurements were performed as described in MATERIALS AND METHODS. Same experiments were also performed in presence of 100 mM NaCl, 100 mM Tris, and 10 mM HEPES · Tris (pH 7.4) containing 1 mM glucose (). Values are means ± SE of 3 independent experiments performed in quadruplicate.

The uptake of D-glucose produced by maltose hydrolysis was measured in the three artificial membrane vesicle preparations, as described in MATERIALS AND METHODS. The maltose diffusion across the artificial membrane vesicles was determined using a proteoliposome preparation similar to proteoliposome preparation 2 without neutral -D-glucosidase and named proteoliposome preparation 3. As shown in Fig. 4, no neutral -D-glucosidase was seen in this proteoliposome preparation after SDS-PAGE analysis (lane 4) compared with the patterns of proteoliposome preparations 1 and 2 (lanes 2 and 3, respectively) and Triton X-100 supernatant from brush-border membrane vesicles (lane 1).

View larger version (110K): [in this window] [in a new window]   Fig. 4.   Pattern of solubilized proteins from native and reconstituted membranes. Left: molecular weight standards. Lane 1, Triton X-100 supernatants of native membrane vesicles; lane 2, proteoliposome preparation 1; lane 3, proteoliposome preparation 2; lane 4, proteoliposome preparation 3, which was prepared from bulk of integral brush-border membrane proteins without neutral -D-glucosidase. Arrowhead, neutral -D-glucosidase.

Under the experimental conditions, the hydrolytic rates remained constant during 3 min and ~5-7% of maltose was hydrolyzed, corresponding to production of 1.25-1.8 mM glucose. Glucose produced by maltose hydrolysis and uptake by proteoliposome preparations 1 and 2 can easily be obtained by subtraction of the maltose diffusion rate, given by proteoliposome preparation 3. Maltose diffusion was 0.124 ± 0.024 nmol · mg¹ · 3 min¹.

As shown in Fig. 5A, after 3 min of incubation and correction for maltose diffusion, no significant variation was observed (P > 0.05) in the amount of glucose produced by maltose hydrolysis and accumulated in the intravesicular space of proteoliposome preparations 1 and 2: 0.349 ± 0.027 and 0.387 ± 0.036 nmol · mg¹ · 3 min¹, respectively.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Incorporation of D-glucose produced from maltose hydrolysis into proteoliposomes with (A) and without (B) all reintegrated brush-border membrane "intrinsic" proteins. 1, Proteoliposome preparation 1; 2, proteoliposome preparation 2; 3, proteoliposomes containing only "integrated" neutral -D-glucosidase; 4, liposomes with "free" form of neutral -D-glucosidase were preincubated for 2 h at 4°C in 100 mM NaCl, 100 mM mannitol, and 10 mM HEPES · Tris (pH 7.4) without sugar, then incubated for 3 min at 25°C with 12.5 mM radiolabeled maltose in same buffer. Uptake conditions are described in MATERIALS AND METHODS.

Figure 5B shows the uptake of glucose produced by maltose hydrolysis into proteoliposomes constituted only with neutral -D-glucosidase or into liposomes (in this condition, the disaccharidase was present in its hydrophilic form in the incubation medium). No significant difference (P > 0.05) was observed in glucose incorporation between artificial vesicle preparations. These findings suggest that neutral -D-glucosidase alone cannot assume the uptake of its hydrolysis product into vesicular space. Interestingly, the glucose concentration observed with these artificial vesicle preparations was lower (P < 0.05) than that measured with proteoliposome preparations 1 and 2 and represents glucose diffusion when the glucose transporter was inhibited by phlorizin addition (Fig. 3).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Evidence of neutral -D-glucosidase as a factor in the kinetic advantage for transport of glucose produced by maltose hydrolysis could be obtained using artificial membrane vesicles containing neutral -D-glucosidase inserted in the presence of the bulk of native integral brush-border membrane proteins. The prerequisite for use of such artificial membrane vesicles is verification that neutral -D-glucosidase has an identical topology in native and artificial membranes. Once this has been confirmed, the putative role of neutral -D-glucosidase in glucose transport can be investigated by studying the effects of modifications in the topology of the enzyme on glucose transport.

The reconstitution method used here was first developed for integration of the band 3 protein from human erythrocyte membrane into phospholipid vesicles (12). Slight modification of this technique has been successfully used for insertion of kidney integral brush-border membrane proteins (4) and sodium-glucose cotransporter (22) in liposomes. About 95-98% of neutral -D-glucosidase obtained before and after Triton X-100 solubilization (Table 1) is incorporated asymmetrically right side out in proteoliposomes regardless of the lipid-to-protein ratio.

Electron microscopy, after negative staining, represents the best protocol for determination of the quaternary structure of purified integral protein reintegrated alone in artificial membranes (5, 18). However, this method cannot be used if the protein is reintegrated in combination with the bulk of native integral membrane proteins. For this reason, we proposed two indirect alternative methods to study the topology of neutral -D-glucosidase in native and artificial membranes. First, we compared solubilization of neutral -D-glucosidase integrated in native membranes and proteoliposomes by various-sized proteolytic enzymes. After proteolysis by papain (5.0 × 3.7 × 3.7 nm) (10) and proteinase K (5.4 × 3.4 × 3.4 nm) (7), both membranes displayed similar solubilization kinetics of neutral -D-glucosidase activity; ~95% of the enzyme activity was solubilized after 60 min of incubation. These results confirm the asymmetric right-side-out integration of neutral -D-glucosidase into the proteoliposomes. By use of these two proteinases, similar conclusions concerning asymmetric integration of dipeptidyl peptidase IV have been reported (15).

In a second set of experiments, we tested the accessibility of neutral -D-glucosidase to specific antibodies after insertion in native and artificial membranes. Antiserum with a high specificity for the enzyme (13) was used for this purpose. Both membrane systems, which contained the same amount of enzyme protein, bound an equivalent amount of anti-neutral -D-glucosidase. Interestingly, this amount of antiserum was also equivalent to that required to precipitate a similar amount of neutral -D-glucosidase activity in its hydrophilic form. These findings suggest that antibody binding by both membrane systems is selective and involves only the protein against which it is directed, and the fact that equivalent amounts of neutral -D-glucosidase in native and artificial membranes or in the hydrophilic form bind equivalent amounts of antibodies suggests a very similar topology of the enzyme in both membrane preparations. In both types of membranes, neutral -D-glucosidase was located on the surface in a maximal immunoreactive form. Similar observations concerning membrane arrangement have been described for other disaccharidases from rat enterocytes (27). These results, which demonstrate the similar topology of neutral -D-glucosidase in native and artificial membrane vesicles, validate use of this model to study the possible role of the enzyme in the transport of glucose produced by maltose hydrolysis.

The total glucose incorporated by small intestinal and kidney brush-border membrane vesicles has been defined as the result of two components: simple diffusion and a specific sodium-dependent, phlorizin-sensitive event mediated by at least the specific SGLT1 protein (30). Our previous investigations of D-glucose uptake by proteoliposomes prepared by the same method using Triton X-100-solubilized kidney brush-border proteins provided evidence that the sodium-dependent, phlorizin-inhibitable glucose transporter was conserved (11, 22). These conclusions are confirmed here by the sodium-dependent nature of glucose uptake in proteoliposome preparations 1 and 2. The absence of any modification in the kinetics of free glucose uptake 1) whatever the neutral -D-glucosidase (inserted in artificial membrane or only in the hydrophilic form in the incubation medium) and 2) after inhibition of the enzyme activity by addition of Tris to the incubation medium demonstrates that free D-glucose uptake is totally independent of the presence of the disaccharidase in the membrane. The fact that the low-affinity glucose transporter is predominant in kidney, whereas the high-affinity glucose transporter is predominant in intestine, does not modify these conclusions.

The uptake of D-glucose produced by maltose hydrolysis was measured in artificial membrane vesicle preparations 1 and 2. These two preparations are characterized by the same intravesicular space, as evidenced by free D-glucose concentrations at the time of equilibrium, and they take up free glucose at the same velocity; they differ only in the topology of neutral -D-glucosidase. After correction for maltose diffusion, determined using proteoliposome preparation 3 (characterized by the absence of any form of neutral -D-glucosidase), no significant variation was observed (P > 0.05) in the amount of glucose produced by maltose hydrolysis and accumulated in both proteoliposome preparations: 0.349 ± 0.027 and 0.387 ± 0.036 nmol · mg¹ · 3 min¹ for preparations 1 and 2, respectively. The fact that glucose produced by maltose hydrolysis was incorporated similarly into proteoliposome intravesicular space, regardless of the location of the neutral -D-glucosidase, rules out the hypothesis whereby kinetic advantage would be due to the proximity of the enzyme and glucose carrier in the membrane (28). However, it could be argued from our experiments that the rate of glucose transport measured in proteoliposome preparations 1 and 2 was too high to allow detection of glucose uptake by the enzyme. Inconsistent with this hypothesis is the fact that no difference was observed in the uptake of glucose derived from maltose hydrolysis when liposomes were prepared in the presence of only purified neutral -D-glucosidase in the "integrated" or "free" form.

The concept of a disaccharidase-related transport system has been recently used to explain the joint absorption of fructose and glucose at >40 mM glucose (24, 29). It was suggested that the enhancement of fructose absorption by glucose might involve sucrase-isomaltase. When glucose and fructose are ingested simultaneously, this disaccharidase processes the two monosaccharides as if they were these products and are carried by two different membrane proteins closely associated in the intact membrane (28, 29). These results, obtained with perfused isolated intestines, were not replicated when everted sleeves of intestine or brush-border membrane vesicle preparations were used (29). To explain these differences, it was proposed that, under the physical conditions needed to prepare tissue for in vitro experiments, these protein associations would be disrupted (28, 29). Two alternative hypotheses could explain the disaccharidase-related transport system observed in perfused segments of isolated intestine. As previously suggested by Alvarado et al. (1), the kinetic advantage for glucose produced from disaccharide hydrolysis observed in all experiments performed with intestine segments may be due to the presence of glycocalyx and/or mucus at the periphery of the apical membrane of enterocytes. The second hypothesis, proposed by Pappenheimer (20), suggested that the excess of glucose that cannot be transported by sodium-glucose cotransporter increases the concentration of free glucose present in the brush-border microenvironment and produced by oligosaccharide hydrolysis through membrane disaccharidases, until a steady state is reached between the rate of new glucose formation and the rate of its removal by solvent drag through intercellular junctions.

Finally, our experiments clearly exclude any involvement of neutral -D-glucosidase in transport of free glucose or glucose produced by maltose hydrolysis. Hydrolytic activity and transport across the membrane would be two independent and sequential steps. These findings agree with the conclusions of previous studies on intestinal segments (1, 8, 25) and rule out the hypothesis of a functional enzyme-transport complex (28).

Perspectives

Our study does not take into account the possible presence of one or more additional proteins coupling neutral -D-glucosidase to glucose transport in small intestine. To resolve the problem it would be interesting to determine whether disaccharidases can closely associate with one or several molecules of sodium-D-glucose cotransporter SGLT1 or other intermediate proteins in native small intestinal brush-border membranes using cross-linker reagents in the presence of various sugar conditions.