High glucose reduces megalin-mediated albumin endocytosis in renal proximal tubule cells through protein kinase B O-GlcNAcylation

The role of albumin reabsorption in proximal tubule (PT) cells has emerged as an important factor in the genesis of albuminuria observed in the early stages of diabetes. Evidence has shown that a decrease in megalin expression could be the key mechanism in this process. In the present work, we investigated the molecular mechanism underlying the modulation of albumin endocytosis in LLC-PK1 cells, a model of PT cells. High glucose concentrations (HG) inhibited megalin expression and albumin endocytosis after 48 h of incubation. This inhibitory effect involves the entrance of glucose into PT cells through SGLT located at the luminal membrane. Once inside PT cells, glucose is diverted to the hexosamine biosynthetic pathway (HBP) increasing O-GlcNAcylation of several intracellular proteins, including PKB. This process promotes the inhibition of PKB activity measured by its phosphorylation at Thr-308 and Ser-473 and phosphorylation of specific substrates, glycogen synthase kinase 3β (GSK3β) and tuberous sclerosis complex 2. The decrease in PKB activity led to a decrease in megalin expression and, consequently, reducing albumin endocytosis in LLC-PK1 cells. HG did not change mammalian target of rapamycin (mTOR) C2 activity, responsible for phosphorylated PKB at Ser-473. In addition, HG activated the mTORC1/S6K pathway, but this effect was not correlated to the decrease in megalin expression or albumin endocytosis. Taken together, our data help to clarify the current understanding underlying the genesis of tubular albuminuria induced by hyperglycemia in the early stage of diabetes pathogenesis.

Diabetes mellitus is a major cause of progression to end-stage renal disease, and albuminuria is a good marker of this process (1,2). The role of albumin reabsorption in proximal tubule (PT) 2 cells has emerged as an important factor in the genesis of albuminuria observed in the early stages of diabetes (3)(4)(5). Albumin reabsorption in these cells involves receptor-mediated endocytosis, with megalin as a central receptor (6,7). It was observed that a high glucose concentration (HG) reduces megalin expression and albumin reabsorption in a streptozotocin-induced diabetes animal model as well as in cultured PT cells (8 -10). In addition, Wang et al. (11) showed that an intravenous infusion of HG in Sprague-Dawley rats induced microalbuminuria and reduced megalin expression in PT cells without changes in the glomerular flow rate.
Megalin/LRP2 is a scavenger receptor belonging to the lowdensity lipoprotein receptor family located at the PT luminal membrane (6,7,12,13). Megalin-mediated endocytosis is a fine-tuning process that involves protein-protein interaction and phosphorylation by specific kinases (12)(13)(14)(15)(16)(17). In this context, the phosphatidylinositide 3-kinase/PKB pathway has been shown to play an important role in modulating albumin endocytosis in PT cells (14). In a previous work, our group (15)(16)(17) showed that PKB binds to megalin at the luminal membrane of PT cells, promoting modulation of albumin endocytosis associated with cell survival. In addition, PKB is a central molecule linking the activation of mammalian target of rapamycin (mTOR) complexes, mTORC2 and mTORC1 (18 -20).
O-GlcNAcylation is mainly found in serine (Ser) or threonine (Thr) hydroxyl moieties on nuclear and cytoplasmic proteins. Similar to phosphorylation, O-GlcNAcylation is a dynamic post-translational modification that regulates protein stability, subcellular localization, and activity (28 -30). Interestingly, PKB was described containing few O-GlcNAcylation sites including residues Thr-308 (31) and Ser-473 (27). It is wellestablished that HG induces elevated O-GlcNAcylation by increasing glucose flux through the hexosamine biosynthetic pathway (HBP) (32)(33)(34). The end product of the HBP is the UDP-GlcNAc, which is used by O-GlcNAc transferase (OGT) to transfer the GlcNAc residue to proteins. O-GlcNAcase (OGA) is the enzyme responsible for its removal.
Based on these observations, we decided to investigate the modulation of albumin endocytosis in PT cells by HG, identifying the possible correlation among O-GlcNAcylation, PKB, and mTOR pathways. We used LLC-PK1 cells, a model of PT cells, incubated with a normal glucose concentration (NG) or HG. We observed that HG increased O-GlcNAcylation in these cells, and this phenomenon was correlated to the inhibition of albumin endocytosis. The molecular mechanism underlying this effect involves the inhibition of PKB activity as a result of its O-GlcNAcylation, leading to a decrease in megalin expression and, consequently, to the decrease in megalin-mediated albumin endocytosis.

Correlation between inhibition of albumin endocytosis in LLC-PK1 cells by high glucose concentration and O-GlcNAcylation
Initially, we investigated the effect of HG on albumin endocytosis in LLC-PK1 cells, a model of PT cells. The cells were incubated with a normal glucose concentration (5.5 mM, NG) or a high glucose concentration (30.0 mM, Fig. S1). Fig. 1A shows one representative experiment of albumin endocytosis measured with fluorescence microscopy using albumin-FITC. The cell membranes were marked with ZO-1 (red staining). HG inhibited albumin endocytosis after 48 h of incubation (Fig. 1, A  and B). To further characterize this effect, the cells were incubated with HG for different periods of incubation (from 1 to 48 h), and albumin endocytosis was measured by cell-associated fluorescence using albumin-FITC as a tracer. Albumin endocytosis was inhibited by HG after 2 h of incubation and was maintained up to 48 h (Fig. 1C). Cell viability, measured by lactate dehydrogenase and propidium iodide staining assays, was not changed under these experimental conditions (Fig. S2, This effect was specific for HG once the incubation of cells with 30 mM mannitol did not change the albumin endocytosis (Fig. S3A). In another experimental set, we grew the cells in Transwell inserts and the HG effect was added to the basolateral or luminal membranes (Fig. 1D). HG at the basolateral membrane did not change albumin endocytosis, but did inhibit it at the luminal membrane. Usually, entry of glucose into PT epithelial cells (PTECs) is mediated by the Na ϩ /glucose cotransporters, SGLT1 and SGLT2 (35). Co-incubation of the cells with 100 M phlorizin (PLZ), an SGLT inhibitor, for 48 h abolished the inhibitory effect of HG on albumin endocytosis (Fig. 1E). Based on these results, we postulate that the inhibition of albumin endocytosis in LLC-PK1 cells by HG involves the entry of glucose through SGLT cotransporters.
We then investigated if the inhibitory effect of HG on albumin endocytosis is due to an increase in the O-GlcNAcylation level. Cells were incubated with HG (from 1 to 48 h) and the O-GlcNAcylation level was assessed by immunoblotting using mAb (clone CTD110.6). HG increased the total O- . D, effect of HG on albumin endocytosis when added to either the luminal side or the basolateral side (n ϭ 3). E, the effect of 100 M PLZ (an SGLT inhibitor) on HG-inhibited albumin endocytosis (n ϭ 6). In C-E, albumin endocytosis was measured by cell-associated fluorescence using albumin-FITC. The results are shown as mean Ϯ S.E. *, p Ͻ 0.05 versus NG (normal glucose condition used as control).

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
GlcNAcylation level after 1 h of incubation and this was maintained up to 48 h (Fig. 2, A and B) as observed with albumin endocytosis. Simultaneous incubation of the cells with HG and 100 M PLZ reduced the O-GlcNAcylation induced by HG (Fig. 2, C and D).
The rate-limiting step in the production of UDP-GlcNAc, a substrate for O-GlcNAcylation, in the HBP is the conversion of fructose 6-phosphate into glucosamine 6-phosphate mediated by glutamine fructose-6-phosphate amidotransferase (GFAT) (26,33). As expected, the incubation of HG cells with 10.0 M 6-diazo-5-oxo-L-norleucine (DON), an inhibitor of GFAT, reduces the O-GlcNAcylation level (Fig. 2, E and F). Interestingly, the reduction of O-GlcNAcylation impaired the inhibitory effect on albumin endocytosis (Fig. 2G). Furthermore, the incubation of LLC-PK1 cells with Thiamet G (TMG), an inhibitor of OGA, for 48 h increased the O-GlcNAcylation level (Fig.  S3, B and C) and mimicked the inhibitory effect of HG on albumin endocytosis (Fig. 2H).
So far, these results show that HG increases cellular O-GlcNAcylation, leading to a reduction in albumin endocytosis in PT cells. However, how the increase in the O-GlcNAcylation level is correlated to the decrease in albumin endocytosis is an open question.

Decrease in megalin expression mediates the inhibitory effect of high glucose concentration on albumin endocytosis in PTECs
HG decreased megalin expression in LLC-PK1 cells measured by immunoblotting and immunofluorescence (Fig. 3, A-F). In agreement with a previous result, this effect of HG was only observed when HG was added at the luminal membrane (Fig. 3, C and D). Furthermore, co-incubation of the cells with 100 M PLZ or 10 M DON abolished the inhibitory effect of HG on megalin expression (Fig. 3, E and F). In addition to megalin, it has been proposed that CD36 present at the luminal membrane of PT cells could also be involved in albumin endocytosis in PT cells (36). Thus, the effect of HG on CD36 expression was assessed (Fig. 3, G and H). Contrary to the effect on megalin expression and albumin endocytosis, HG increased CD36 expression. These results indicate that the inhibitory effect of HG on albumin endocytosis should be correlated to the decrease in megalin expression due to the increase in the O-GlcNAcylation level.

Activation of the mTORC1/S6K pathway by high glucose concentration is not involved in the inhibition of albumin endocytosis
Here, it was observed that HG increased mTORC1 activity measured by phosphorylation of mTOR at Ser-2448, a marker of mTORC1 activation, and phosphorylation of S6K at Thr-389, a specific substrate for mTORC1, in LLC-PK1 cells (Fig. 4,  A and B). The co-incubation of the cells with 100 M PLZ or 10 M DON abolished the phosphorylation of both residues induced by HG (Fig. 4, C and D), indicating the role of O-GlcNAcylation in this activation. mTOR inhibitors, rapamycin and WYE-354, were used to study the possible correlation of the mTORC1/S6K pathway with the inhibitory effect of HG on albumin endocytosis. Coincubation of the cells with 10 Ϫ9 M rapamycin and 10 Ϫ6 M WYE-354 did not change the inhibitory effect of HG on albumin endocytosis (Fig. 5, A-C). In agreement, these mTOR inhibitors also did not block the inhibitory effect of HG on megalin expression (Fig. 5, D and E). Thereby, the inhibitory

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
effect of HG on albumin endocytosis does not involve activation of the mTORC1/S6K pathway.

PKB mediates the inhibitory effect of HG on albumin endocytosis
In previous work (15,16), it was shown that PKB anchors at the plasma membrane through megalin binding, and this process is important for albumin endocytosis. Thus, we investigated the possible involvement of PKB in the inhibitory effect of HG on albumin endocytosis.
Initially, PKB activity was measured by phosphorylation of Ser-473 and Thr-308 residues and phosphorylation of specific substrates GSK3␤ and TSC2 at Ser-9 and Thr-1462, respectively. It was observed that PKB activity was decreased by HG (Fig. 6, A and B). Then, to correlate the inhibition of PKB activity with the inhibitory effect of HG on albumin endocytosis, the cells were incubated with 10 nM MK-2206, a PKB inhibitor.
Interestingly, the PKB inhibitor mimicked the inhibitory effect of HG on albumin endocytosis (Fig. 6C) and on megalin expression (Fig. 6, D and E). Furthermore, it was observed that incubation of the cells with 10% FBS, which activates PKB, blocked the inhibitory effect of HG on albumin endocytosis (Fig. 6C) as well as on megalin expression (Fig. 6, D-G). Together, our results show that HG inhibited PKB activity, and this effect is involved in the inhibition of albumin endocytosis in PT cells.

PKB O-GlcNAcylation induced by HG is responsible for reduced PKB activity
Because phosphorylation of PKB at Ser-473 is mediated by mTORC2, we decided to investigate if modulation of mTORC2 activity mediates the inhibitory effect of HG on PKB activity. Interestingly, HG did not change the phosphorylation of mTOR at Ser-2481, a marker of mTORC2 activity (Fig. 7A) or the phosphorylation of PKB at Thr-450, a specific substrate for

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
mTORC2 (Fig. 7B). Furthermore, it has been shown that the phosphorylation of RICTOR at Thr-1135 is a negative modulator of mTORC2 activity (20). Here, we observed that this phosphorylation is not changed by HG (Fig. 7C). These results indicate that mTORC2 activity is not changed by HG. Based on these results, we investigated if O-GlcNAcylation could be responsible for PKB inhibition induced by HG.
Co-incubation of the cells with 100 M PLZ (Fig. 8A) 8C) and enzymatic labeling (Fig. 8D). Taken together these results indicate that PKB inactivation is probably induced by its O-GlcNAcylation. This effect leads to a decrease in PKB activity, megalin expression, and PT albumin uptake.

Discussion
Albuminuria is observed in the early stages of diabetes even before the onset of chronic kidney disease (1, 2). The role of albumin reabsorption in PT cells has been highlighted as an important mechanism in this process and a possible therapeutic target to halt the progression of renal disease (5,16,(37)(38)(39). Some groups showed that HG decreases megalin expression and albumin endocytosis in PT cells (8 -10). In this present work, using LLC-PK1 cells, we showed that the mechanism underlying the effect of HG involves the entry of glucose through luminal SGLT, O-GlcNAcylation of PKB, and its inhibition. These effects lead to a decrease in megalin expression and, consequently, in albumin reabsorption in PT cells. Interestingly, the modulatory effect of HG on the mTORC1/S6K pathway is not involved in this process. These results help to clarify the current understanding underlying the genesis of tubular albuminuria induced by HG in the early stage of diabetes pathogenesis.
One possible question concerns the concentration of HG used. The blood glucose concentration in the streptozotocininduced diabetes animal model ranges from 360 to 600 mg/dl, corresponding to 20 -33 mM (8,9,40,41). This blood glucose concentration is close to the concentration used in the present study. Furthermore, the concentration of HG in the present work follows the same protocol used by different authors to study the effect of HG on PT cells (43)(44)(45)(46)(47)(48). In addition, Ishi-

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
bashi (43) showed that incubation of LLC-PK1 cells for 6 days with 16 mM glucose inhibits albumin uptake in LLC-PK1 cells at the same magnitude as 27 mM glucose.
In agreement with our results, Tojo et al. (8), using an earlystage diabetic nephropathy animal model, observed glucosuria and albuminuria without changes in the glomerular flow rate. They observed a decrease in megalin expression leading to a decrease in PT albumin reabsorption. In addition, it has been shown that PT PKB activity is inhibited in the streptozotocininduced diabetes animal model (49). Here, we observed that HG decreases megalin expression and albumin uptake and inhibits PKB activity in LLC-PK1 cells. These observations indicate that the results obtained for LLC-PK1 cells are compatible with those obtained in the animal model.
It is well-known that the genesis of diabetic nephropathy (DN) comprises deleterious effects of hyperglycemia in renal glomerular and tubular structures (1,2,22,23,25,26,50,51). During the early stage of diabetes, there is an increase in serum glucose, glomerular glucose filtration, and, consequently, luminal glucose concentration (1,2,52). Glucose is reabsorbed at PT cells through coordinate transporter mechanisms located at the luminal and basolateral membranes, avoiding loss of glucose in the urine. At the early stage of diabetes, this mechanism could lead to a dangerous loop aggravating hyperglycemia (1,2,52). Therefore, PT cells are exposed to HG at both luminal and basolateral membranes.
One important issue to be considered is the laterality of the HG effect. Here, we observed that the inhibitory effect of HG on albumin endocytosis depends on its entry at the luminal membrane by SGLT decreasing megalin expression and, consequently, inhibition of albumin endocytosis in LLC-PK1 cells. In agreement, some authors showed that the treatment of patients and animal models of DN with SGLT2 inhibitors ameliorates proteinuria, tubule-interstitial injury, and consequently, slows the progression of renal disease (52)(53)(54)(55). We showed that glucose inside the LLC-PK1 cells increases O-GlcNAcylation, which is correlated with inhibition of albumin endocytosis, opening the possibility that this mechanism could be involved in albuminuria in the early stage of DN. In agreement, Akimoto et al. (56) showed increased O-GlcNAcylation in specific PT cytoskeleton proteins in a T2DM animal model. Ji et al. (57) suggested that abnormal O-GlcNAcylation of actin and tubulin alters their polymerization, reducing microtubule-dependent reabsorption mechanism in PT cells. Gellai et al. (58) also showed that incubation of HK2 (human kidney cells, a model of PT cells) with HG also increased O-GlcNAcylation, which was correlated to an increase in DN-induced fibrosis.
Some authors have proposed that changes in the albumin endocytosis machinery induce a proinflammatory phenotype in the PT environment, contributing to tubule-interstitial injury (16,39). It was observed that the decrease in megalin expression involves transforming growth factor-␤ secretion (59 -61). In a previous work (16), our group showed that the decrease in megalin expression induced by a high albumin concentration is associated with apoptosis in LLC-PK1 cells. Thus, it is possible to postulate that the decrease in megalin expres-

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
sion due to HG-induced O-GlcNAcylation is also involved with the development of tubule-interstitial injury.
Besides megalin, some evidence indicates that CD36 could also contribute to albumin endocytosis in PTECs (36,62). However, the significance of this process on total albumin endocytosis is still an open matter. Significant albuminuria was observed in T2DM patients, even with increasing CD36 expression in PT cells (36). We also observed an increase in CD36 expression induced by HG despite the decrease in albumin endocytosis in LLC-PK1 cells. Taken together, these observations indicate that the increase in CD36 expression could be a compensatory mechanism avoiding more severe proteinuria and/or it could be involved in tubule-interstitial injury induced by HG contributing to the progression of DN. Reinforcing the second idea, it has been proposed that CD36 mediates injury in PT cells, including apoptosis of these cells in DN (62).

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
The idea that mTOR complexes could be involved in the effects of HG on renal function is supported by enhancement of mTORC1 activity under these conditions. This process modifies the fine balance between mTORC1 and mTORC2 activities, which is required to maintain cell homeostasis (18,(21)(22)(23). In podocyte cells, the activation of mTORC1 by hyperglycemia mediates the podocyte injury observed in DN (22,23). In recent work, Grahammer et al. (21), using specific raptor knockout mice in renal tubular cells, observed significant proteinuria due to inhibition of albumin endocytosis in PT cells, although without change in megalin expression. Similarly, we observed that the inhibition of mTORC1 activity decreased albumin endocytosis in PT cells. However, we observed that despite activation of the mTORC1/S6K pathway, HG inhibited albumin endocytosis, indicating that these effects involve independent pathways. Palm et al. (63), using mouse embryonic fibroblasts, showed that mTORC1 activation inhibits albumin proteolysis in lysosomes.
Interestingly, we observed that HG did not change the mTORC2 activity measured by phosphorylation of PKB at Thr-450, a specific site for mTORC2 (18,27). Supporting our observation, Grahammer et al. (21) showed that inhibition of mTORC2, using a specific RICTOR knockout in renal tubular cells, did not change albumin reabsorption in PTECs. In a previous work, we showed that overactivation of the mTORC1/ S6K pathway by higher albumin concentration promotes the inhibition of mTORC2 activity due to the increase in phosphorylation of RICTOR at Thr-1135 (20). Here, it was observed that the phosphorylation of this residue was not changed by HG, even when the mTORC1/S6K pathway was activated. Despite these results, we cannot rule out the possible modulation of albumin endocytosis in PT cells by mTORC2. It has been shown that simultaneous inhibition of mTORC1 and mTORC2 could have additional deleterious effects on podocyte cells and albumin endocytosis in PT cells (21,22). This view agrees with a proposal that a fine balance between mTORC1 and mTORC2 activities is required to maintain the fine-tuning in cell homeostasis.
The following data obtained in our study show that the inhibition of PKB activity mediates the inhibitory effect of HG on megalin expression and, consequently on albumin endocytosis: 1) HG inhibits PKB activity; 2) PKB inhibitor mimics the effects of HG on albumin endocytosis and megalin expression; 3) serum reconstitution, an activator of PKB, reversed the effects of HG. In agreement with this result, a strict relationship between albumin endocytosis and PKB activity has been shown  (14 -17). Here, we show that the mechanism underlying the inhibitory effect of HG on PKB activity involves O-GlcNAcylation. The increase of O-GlcNAc levels induced by HG often results in deregulation of signaling pathways that are important for the development of pathologies such diabetes, chronic kidney disease, cancer, among others (33, 64 -69). Recent evidence indicates that O-GlcNAcylation of kinases play an important role in regulating cell signaling (70). In accordance, it has been shown that O-GlcNAcylation of PKB in different cell types decreases phosphorylation at Ser-473 and Thr-308 by steric hindrance or competition, and consequently, inhibition of PKB (27,31). Once O-GlcNAcylated, PKB is inhibited, leading to a decrease in total megalin level. How this happens is still an open matter. One possibility is the change in the half-life of megalin due to its degradation (71). Another possibility could be a decrease in megalin synthesis. Zhou et al. (10) showed that the incubation of NRK52E cells, a tubule epithelial cell line, with HG decreases mRNA for megalin, as well as megalin synthesis. In addition, some studies have shown that PKB can modulate transcription factors such as FOXO1 and nuclear factor-B, changing the protein synthesis (18,72). However, more experiments will be necessary to clarify this issue.

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
Taken together, our results allow us to build up a model describing the molecular mechanism underlying the inhibition of megalin-mediated albumin endocytosis triggered by HG (Fig. 9). Glucose enters PT cells through SGLT, increasing UDP-GlcNAc levels and O-GlcNAcylation of several intracellular proteins, including PKB, promoting its inhibition. This process leads to a decrease in megalin levels and, consequently a decrease in albumin endocytosis in PT cells.
The results obtained in the present study together with those previously published indicate that PT cells could play a more important role in DN than has been proposed previously. This hypothesis is well-supported in a recent revision published by Gilbert (73). The author pointed out the central role of PT cells in DN, suggesting that a paradigm shift in "glomerulocentric" is necessary and pointing out the role of the PT segment in the genesis and development of so-called diabetic kidney disease.

Cell culture
LLC-PK1 cells, a PT cell line, were cultured in six-well plates or in Transwell inserts when indicated. The cells were grown at 37°C in 5% CO 2 in DMEM (glucose 5.5 mM, NG) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin for 3 consecutive days, usually reaching 95-98% confluence (15, 16, 19, 20, 74 -76). When indicated, a high glucose concentration (final concentration, 30 mM) was added to the cells 24 h after seeding and then maintained for 48 h. The cells were serum starved in the last 12 h before they were used in the experiments. The scheme of the experimental design is indicated in Fig. S1. The serum starvation was done to keep several signaling proteins triggered by growth factors under a basal state of activation. After treatment, the cells were used to carry out different experimental approaches as shown below, such as albumin endocytosis, immunoprecipitation, immunoblotting, and immunofluorescence assays.

Albumin endocytosis assay
Albumin endocytosis was measured by BSA-FITC uptake assay as described previously (15,75). Briefly, cells were incubated with Ringer solution (20 mM Hepes-Tris, pH 7.4, 140 mM NaCl, 2.7 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM D(ϩ)glucose) containing 30 g/ml of BSA-FITC at 37°C for 30 min. Unbound BSA-FITC was removed by washing 11 times with ice-cold Ringer solution. Parallel experiments were performed on cells held at 4°C to abolish endocytosis and verify possible nonspecific bound BSA-FITC. BSA uptake was assessed by cell-associated fluorescence using albumin-FITC or fluorescence microscopy.
In the cell-associated fluorescence method, the cells were lysed using detergent solution (0.1% Triton X-100 in 20 mM MOPS) and the cell-associated fluorescence was measured using a microplate spectrofluorimeter (SpectraMax M2, Molecular Devices, Sunnyvale, CA). BSA-specific uptake was calculated as the difference between the BSA-FITC uptake in the absence and presence of 20 mg/ml of unlabeled BSA. The blank in the experiment was defined as the fluorescence measured in the absence of BSA-FITC; its value was less than 10% of the total fluorescence.
For fluorescence microscopy, the cells were fixed with 4% paraformaldehyde at 25°C for 15 min, permeabilized with 0.1% Triton X-100 for 1-2 min, and blocked with 3% nonfat milk in PBS 2ϩ for 30 min (15). To stain the plasma membrane, the cells were incubated with specific mouse primary antibodies, ZO-1 and Alexa Fluor 546-conjugated IgG. Then, the cells were

PKB O-GlcNAcylation in megalin-mediated albumin endocytosis
mounted in Vectashield medium. To quantify albumin endocytosis, images of fluorescent-labeled markers were acquired by a Nikon 80i fluorescence microscope using a ϫ40 oil immersion objective lens, and the average intensity per cell was calculated using Image-Pro Plus software version 7.0.1.658 (Media Cybernetics, Rockville, MD) from multiple images after subtracting the background.

Immunoprecipitation and immunoblotting assay
The immunoprecipitation and immunoblotting assays were performed as described previously (20,74). Briefly, the cells were incubated in lysis buffer (20 mM Hepes, pH 7.4, 2 mM EGTA, 1% Triton TM X-100, 0.01% SDS, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 5 mM sodium orthovanadate, 10 mM sodium ␤-glycerophosphate, 1 mM PMSF, 2ϫ complete protease inhibitor) and cleared by centrifugation at 15,000 ϫ g for 10 min at 4°C. After centrifugation, the concentration of protein in the supernatant was determined by the Folin phenol method using BSA as standard (77). When indicated, PBS was immunoprecipitated using polyclonal PKB antibody according to the manufacturer's instructions. Then, proteins were then resolved on SDS-PAGE and transferred to PVDF membranes, according to the manufacturer's instructions. After antibody labeling, detection was performed using ECL TM prime.

Enzymatic labeling of O-GlcNAc sites
The enzymatic labeling of O-GlcNAc sites was determined as described previously (42). PKB was immunoprecipitated, washed with lysis buffer, and washed twice with reaction buffer containing 20 mM Hepes, pH 7.9, 50 mM NaCl, 1 M PUGNAc, and 5 mM MnCl 2 with protease and phosphatase inhibitors. Next, 3 l of Gal-T1 Y289L and 3 l of 0.5 mM UDP-GalNAz were added to a reaction volume of 30 l. The reaction was performed overnight at 4°C. The beads were washed twice with reaction buffer to remove excess UDP-GalNAz. The samples were then reacted with biotin alkyne according to the manufacturer's instructions. The enzymatic reactions were eluted with Laemmli buffer, resolved in SDS-PAGE. 12%, and then transferred to PVDF membranes according the manufacturer's instructions. After HRP-conjugated streptavidin labeling, detection was performed using ECL TM prime.

Immunofluorescence assay
The immunofluorescence assay was performed as described (16,20,74). Immunofluorescent staining for megalin, CD36, and ZO-1 was performed using specific antibodies as described by the manufacturer. Briefly, cells were grown in coverslips following the treatment protocols described above. After washing with PBS, the cells were fixed and permeabilized appropriately. The nonspecific binding sites for antibodies were blocked using 5% normal donkey or goat serum for 1 h, both at room temperature. Then, cells were incubated with the primary and secondary antibodies for 1 h at room temperature as indicated by the manufacturer's protocol. The nucleus was stained using DAPI. After staining, the coverslips were washed thoroughly in PBS and mounted with antiquenching medium (Vector Laboratories, Burlingame, CA), and then the slides were sealed. The fluorescence label was examined with a microscope (model Eclipse 80i, Nikon Corporation, Chiyoda-ku, Tokyo, Japan). Images were acquired using the manufacturer's software. Contrast and brightness settings were chosen to ensure that all pixels were within the linear range. Images were prepared for publication with Adobe Photoshop (Adobe Systems, San Jose, CA).

Statistical analysis
The results are expressed as mean Ϯ S.E. Statistical significance was assessed by one-way analysis of variance, considering the treatments as factors. The significance of the differences was verified by the Newman-Keuls test for multiple comparisons. Statistical analysis was performed using absolute values. Significance was determined as p Ͻ 0.05.