HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 12, 10683-10690, March 22, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/12/10683    most recent M109357200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Alfalah, M. Articles by Naim, H. Y. Search for Related Content PubMed PubMed Citation Articles by Alfalah, M. Articles by Naim, H. Y. Social Bookmarking                      What's this?

Intestinal Dipeptidyl Peptidase IV Is Efficiently Sorted to the Apical Membrane through the Concerted Action of N- and O-Glycans as Well as Association with Lipid Microdomains^*

Marwan Alfalah, Ralf Jacob, and Hassan Y. Naim

From the Department of Physiological Chemistry, School of Veterinary Medicine Hannover, Hannover D-30559, Germany

Received for publication, September 27, 2001, and in revised form, December 7, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The apical sorting of human intestinal dipeptidyl peptidase IV (DPPIV) occurs through complex N-linked and O-linked carbohydrates. Inhibition of O-linked glycosylation by benzyl-N-acetyl--D-galactosaminide affects significantly the sorting behavior of DPPIV in intestinal Caco-2 and HT-29 cells. However, random delivery to the apical and basolateral membranes and hence a more drastic effect on the sorting of DPPIV in both cell types is only observed when, in addition to O-glycans, the processing of N-glycans is affected by swainsonine, an inhibitor of mannosidase II. Together the data indicate that both types of glycosylation are critical components of the apical sorting signal of DPPIV. The sorting mechanism of DPPIV implicates its association with detergent-insoluble membrane microdomains containing cholesterol and sphingolipids, whereas an efficient association largely depends on the presence of a fully complex N- and O-linked glycosylated DPPIV. Interestingly, cholesterol is a more critical component in this context than sphingolipids, because cholesterol depletion by -cyclodextrin affects the detergent solubility and the sorting behavior of DPPIV more strongly than fumonisin, an inhibitor of sphingolipid synthesis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Fundamental functions of columnar epithelial cells are maintained through the segregation of particular membrane proteins and lipids to the two domains, basolateral and apical, and thus characterize the structural and functional polarity of these cells. Proteins destined for these surfaces possess various types of structures or motifs that are recognized by specific sorting mechanisms in the trans-Golgi network (TGN)¹ and are subsequently delivered via separate vesicles to the correct surface domain (1, 2). The failure to sort proteins to the correct domain often results in pathological conditions (3, 4). Sorting is therefore a critical mechanism by which the function of sorted proteins is maintained. Sorting signals are diverse and operate hierarchically (5). Basolateral signals are usually located in the cytoplasmic tails of transmembrane proteins and are capable of targeting proteins in epithelial cells (6-12) or the cell body of neurons (13-17). Apical sorting signals are not as unique, are less well defined, and can be located in the luminal, membrane, or cytosolic domains of the targeted protein. Structurally diverse apical signals are known that are encoded in the proteinaceous, carbohydrate, or lipid moieties of the sorted proteins. Glycophosphatidylinositol (GPI) membrane anchors direct proteins to the apical surface of several types of epithelial cells (18, 19) via association with detergent-insoluble membrane domains enriched in glycosphingolipids and cholesterol (20) in the TGN (21, 22). N-Linked glycans on some secreted proteins may mediate apical transport (23), although this mechanism does not apply to all secreted proteins (24-27). Another type of glycosylation, O-glycosylation, has been demonstrated to be important in directing the neurotrophin receptor and intestinal sucrase-isomaltase (SI) to the apical membrane (28-30). Finally, carbohydrate-independent sorting signals, most likely protein subdomains, are implicated in the sorting of intestinal lactase-phlorizin hydrolase and many other proteins (31).

In this paper we investigate the role of glycans in the sorting of a heavily N- and O-glycosylated glycoprotein, dipeptidyl peptidase IV. This protein is usually sorted with high fidelity to the apical membrane of intestinal cells (32). By virtue of its convenient structural features DPPIV serves as an exquisite model protein for analyses aimed at elucidating the sorting behavior and sorting signals of proteins.

DPPIV is a type II membrane glycoprotein that is synthesized with an uncleavable signal sequence that functions as a membrane anchoring domain (33, 34). Processing of the 100-kDa mannose-rich DPPIV species to the 124-kDa complex glycosylated mature form includes an extensive O-glycosylation event, mainly in a Ser/Thr-rich stalk domain adjacent to the membrane anchor (35). After maturation in the Golgi apparatus, DPPIV is transported to the apical membrane, either directly from the TGN or along the transcytotic pathway through the basolateral membrane (32). In Caco-2 cells at least 85% of newly synthesized DPPIV is expressed at the apical membrane and in steady state determinations and in pulse-chase experiments DPPIV is entirely found at the apical membrane after prolonged chase periods (32). These observations indicate that the transcytotic pathway is not the major sorting event, otherwise substantially more DPPIV would be present at the basolateral membrane.

The identity and location of the sorting signals of DPPIV as well as the sorting mechanism have not yet been investigated in great detail. Recent data have proposed O-glycosylation to be crucial for apical sorting (36). A possible role of N-linked glycosylation remains undetermined, and recent data have excluded sialic acids on N-linked glycans being implicated in the sorting event of DPPIV (37).

The main purpose of this paper is to delineate the mechanism by which DPPIV is sorted to the apical membrane in Caco-2 cells. We show that O-glycans play a role in the sorting event. However, a high sorting fidelity is substantially strengthened by the presence of complex glycosylated but not mannose-rich N-linked glycans. Furthermore, DPPIV associates on its way to the brush border membrane with Triton X-100-insoluble cholesterol-rich microdomains, although sphingolipids do not constitute a critical structure of these domains.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Biosynthetic Labeling and Treatment of Cells-- Caco-2 and HT-29 (38) cells were cultured in 10-cm Petri dishes or on membrane filters (Falcon, Plymouth, UK), as described previously (36), and were usually labeled biosynthetically at day 6 after confluence. When used, benzyl-GalNAc, an inhibitor of O-glycosylation (39), swainsonine, an inhibitor of mannosidase II (40), fumonisin, an inhibitor of sphingolipid biosynthesis (41), or -cyclodextrin for cholesterol depletion (42) were present in the culture medium during preincubation of the cells in methionine-deficient medium and during the labeling intervals essentially as described before (36). The final concentration of benzyl-GalNAc was 4 mM, that of swainsonine was 4 µg/µl, that of fumonisin was 10 µM, and that of -cyclodextrin was 50 mM (all reagents were obtained from Sigma).

Cell Surface Immunoprecipitation-- Metabolic labeling of Caco-2 or HT-29 cells grown on filters or plated in 6-well culture dishes was performed as described previously (28). The cells were labeled for 4 h with 100 µCi of [³⁵S]methionine (Amersham Biosciences). In pulse-chase experiments, the cells were pulsed with 100 µCi of [³⁵S]methionine for 15 min followed by a chase with cold methionine for different intervals. The cells were solubilized using a combination of Triton X-100 and sodium deoxycholate in the presence of protease inhibitors, and the detergent extracts were immunoprecipitated with mouse mAb anti-DPPIV (hybridoma HBB 3/153, Ref. 42) as described previously (36). Cell surface antigens were immunoprecipitated from intact cells on filters by addition of anti-DPPIV antibody to either the apical or basolateral compartments. The immunoprecipitates were analyzed by SDS-PAGE on 6% polyacrylamide gels according to the method of Laemmli (43). After electrophoresis the gels were visualized using a PhosphorImager (Bio-Rad).

Detergent Extractability of DPPIV-- Caco-2 or HT-29 cells were grown 6 days after confluence and were biosynthetically labeled for 4 h with [³⁵S]methionine. The cells were solubilized in the cold for 2 h with 1% Triton X-100 in 25 mM Tris-HCl, pH 8.0, 50 mM NaCl. The detergent extracts were centrifuged at 100,000 × g for 1 h at 4 °C, and the supernatant was immunoprecipitated with mAb anti-DPPIV. The pellet was dissolved by boiling in 1% SDS for 10 min. Thereafter, a 10-fold volume of buffer containing 1% Triton X-100 was added. These extracts were centrifuged, and the supernatant was also immunoprecipitated with mAb anti-DPPIV that recognizes native and denatured forms of DPPIV (44). In parallel, lipid microdomains were isolated on a 5-35% sucrose gradient as described previously (28). DPPIV was immunoprecipitated from the individual gradient fractions with mAb anti-DPPIV, and the immunoprecipitates were analyzed by SDS-PAGE on 6% gels.

Protease Sensitivity Assays-- Caco-2 cells were labeled with [³⁵S]methionine for 15 min and chased for 4 h with cold methionine. The pulse and chase were performed in the presence or absence of benzyl-GalNAc, swainsonine, or both using similar final concentrations as described above. The cells were solubilized in the cold for 2 h with 1% Triton X-100 in 25 mM Tris-HCl, pH 8.0, 50 mM NaCl, and the detergent extracts were immunoprecipitated with mAb anti-DPPIV. The immunoprecipitates were treated or not treated with 50 µg of trypsin in phosphate-buffered saline (1 mg/ml) for 30 min at 37 °C, after which time the reaction was terminated by boiling in 3-fold SDS-PAGE Laemmli buffer. The treatment of immunoprecipitates with 5 units of the V8 protease was performed for 90 min in phosphate-buffered saline at 37 °C (45) followed by a similar treatment as in the case of trypsin. The treated and non-treated DPPIV glycoforms were subjected to analysis by SDS-PAGE.

Endoglycosidase Treatment and SDS-PAGE-- Digestion of [³⁵S]methionine-labeled immunoprecipitates with endo--N-acetylglucosaminidase F/glycopeptidase F (endo F) was performed as described previously (46). Analysis of proteins was performed on 6% slab gels according to Laemmli (43).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Inhibition of O-Glycosylation by Benzyl-GalNAc-- By using enzymatic and chemical deglycosylation procedures, it has been convincingly demonstrated that DPPIV is a heavily O-glycosylated glycoprotein (36). Moreover, an association between trimming of mannose-rich N-linked glycans of DPPIV and the onset of O-glycosylation in the cis-Golgi could be demonstrated in the presence of deoxymannojirimycin (dMM), an inhibitor of cis-Golgi mannosidase I. In fact, O-glycosylation of DPPIV is drastically reduced when trimming was inhibited. The polarized sorting of DPPIV was substantially affected in the presence of dMM implicating O-glycans in the sorting event. However, dMM inhibits processing of mannose-rich chains, and it is therefore unclear whether complex N-glycans that do not occur in the presence of dMM are involved in sorting. Here we employed benzyl-GalNAc, a potent inhibitor of O-linked glycosylation, which competes with N-acetylgalactosamine in binding to the potential O-glycosylation sites, Ser or Thr. This inhibitor has been shown not to affect complex N-glycosylation (39). The effect of benzyl-GalNAc on structural features of DPPIV was then examined in Caco-2 cells that were biosynthetically labeled with [³⁵S]methionine for 4 h in the presence or absence of the inhibitor. Fig. 1 depicts the results obtained in these experiments. DPPIV appeared predominantly as a 124-kDa complex glycosylated species within 4 h of labeling (denoted DPPIV_c) in the absence of benzyl-GalNAc as shown previously (36) in intestinal biopsy samples. It shifted to a diffuse O-glycosylated band upon endo F/GF treatment. A substantial reduction in the size of the complex glycosylated glycoform of DPPIV to about 115-kDa (denoted DPPIV_c/benzyl) was discerned in benzyl-GalNAc-treated Caco-2 cells. To examine whether O-glycosylation has been affected, we made use of the observation that N-deglycosylation of complex glycosylated wild type DPPIV_c with endo F/GF results in a polypeptide that is O-glycosylated. Alterations in the O-glycosylation pattern of DPPIV_c/benzyl could be readily monitored in their N-deglycosylated forms, and a change in the size of the N-deglycosylated forms is reminiscent of different O-glycosylation. Fig. 1 shows that the endo F/GF product of DPPIV_c/benzyl was substantially smaller than its DPPIV_c counterpart which is O-glycosylated. It is evident therefore that O-glycosylation has been directly affected and is markedly reduced in DPPIV_c/benzyl.

View larger version (83K): [in this window] [in a new window]   Fig. 1.   Effect of benzyl-GalNAc on the O-glycosylation of intestinal DPPIV. Caco-2 cells were biosynthetically labeled with [35S]methionine for 4 h in the presence (denoted benzyl) or absence (denoted control) of the inhibitor of O-glycosylation of benzyl-GalNAc (benzyl). DPPIV was isolated by immunoprecipitation, and the immunoprecipitates were treated or not treated with endo F followed by SDS-PAGE analysis.

O-Linked Glycosylation Is Not Sufficient for an Efficient Apical Sorting of DPPIV-- Having assessed the effect of benzyl-GalNAc on the O-glycosylation of DPPIV, we examined the sorting behavior of the modified glycoform DPPIV_c/benzyl. For this, cell surface immunoprecipitations with mAb anti-DPPIV were performed on Caco-2 cells that have been pulse-labeled with [³⁵S]methionine and chased for several time points in the presence or absence of benzyl-GalNAc (Fig. 2). 1 h into the chase DPPIV appeared predominantly at the apical surface of Caco-2 cells that were labeled in the absence of benzyl-GalNAc. This sorting pattern persisted throughout 4 h of chase. In the presence of benzyl-GalNAc a shift in the sorting behavior of the modified DPPIV_c/benzyl glycoform to the basolateral membrane could be discerned (Fig. 2). At 1 h of chase almost 25% of the DPPIV_c/benzyl glycoform was found at the basolateral membrane, and the proportion of basolaterally located DPPIV_c/benzyl increased to almost 40% at the 2- and 4-h chase time points. This result indicates that modification of the O-linked glycosylation is compatible with alterations in the sorting event and points to a direct implication of O-linked glycans as an apical sorting signal. On the other hand, the variation in the sorting behavior is not as drastic as in the case of the heavily O-glycosylated sucrase-isomaltase, where almost equal distribution of this protein could be observed on both sides of the membrane when O-glycosylation is inhibited by benzyl-GalNAc (28). It is possible that components other than O-glycans are implicated in the sorting of DPPIV.

View larger version (47K): [in this window] [in a new window]   Fig. 2.   Implication of N- and O-linked glycans in the sorting of DPPIV. A, Caco-2 cells were grown on transmembrane filters and biosynthetically labeled with [35S]methionine for 15 min and chased for the indicated times with cold methionine. The pulse-chase experiments were performed in the presence or absence of benzyl-GalNAc or benzyl-GalNAc and swainsonine. Cell-surface immunoprecipitation of DPPIV from the apical (a) or basolateral (b) membranes was performed with mAb anti-DPPIV followed by SDS-PAGE analysis. B-D, quantification of the proportions of DPPIV appearing at the apical and basolateral membranes shown in A.

Inhibition of the Processing of N- and O-Linked Glycosylation Results in a Random Distribution of DPPIV Over the Apical and Basolateral Membranes-- Previous work from our laboratory (36) has shown that polarized sorting of DPPIV is not affected when complex N-linked glycosylation was modulated at the site of action of swainsonine, an inhibitor of Golgi mannosidase II. Additionally, swainsonine had neither a direct nor an indirect effect on O-glycosylation of DPPIV, whereas the size of the N-linked glycans was smaller than the normally glycosylated DPPIV species. By contrast dMM which inhibits the processing of the outermost mannose-rich mannoses by mannosidase I has markedly shifted the high polarity of DPPIV to an almost randomly distributed protein on both sides of the membrane. At this stage of investigation the effect was attributed to the absence of O-glycans, the processing of which has been indirectly affected by the presence of the unprocessed mannose-rich chains (see Ref. 36 for more details). The sorting pattern of DPPIV was not as drastically altered upon inhibition of O-glycosylation by benzyl-GalNAc as it was in the presence of dMM, and this is explained as follows. First, only partial O-glycosylation occurred in the presence of benzyl-GalNAc but was sufficient to prevent a complete depolarization of the targeting pattern of DPPIV. Second, additional components to O-glycans are implicated in the sorting of DPPIV. Possible candidates are mature complex glycosylated N-linked glycans. The following observations lend support to this hypothesis. In the presence of dMM the processing of mannose-rich N-linked glycans of DPPIV to the mature form does not take place, O-glycosylation is impaired, and targeting of DPPIV follows a random pattern. On the other hand, in the presence of swainsonine, which inhibits mannosidase II, an incompletely processed but nevertheless endo H-resistant N-linked and O-glycosylated DPPIV is correctly sorted to the apical membrane (36). Altogether a possible role for complex glycosylated N-linked glycans in conjunction with O-linked glycans could be hypothesized. To examine this possibility, we labeled cells in the presence of benzyl-GalNAc and swainsonine to affect both N- and O-linked glycosylation, and we analyzed the sorting of DPPIV. Under these conditions DPPIV shifted further to a smaller protein (DPPIV_c/benzyl/swa) as compared with the glycoforms obtained in the presence of benzyl-GalNAc or swainsonine alone. The sorting behavior of this DPPIV glycoform was next investigated using a pulse-chase protocol of biosynthetically labeled Caco-2 cells. Fig. 2 demonstrates that a further alteration in the trafficking of DPPIV as compared with DPPIV_c/benzyl has occurred. At 1 h of chase only 55% of DPPIV_c/benzyl/swa was located at the apical membrane. With increasing chase DPPIV_c/benzyl/swa became almost equally distributed at the apical and basolateral membranes. The results demonstrate that the type and structure of N- and O-linked complex glycans are together critical elements in the apical targeting of DPPIV. It is also obvious that O-linked glycans alone are not sufficient for a sorting of DPPIV with high fidelity to the apical membrane.

Modification of N- and O-Linked Glycans Does Not Affect the Folding Pattern of DPPIV-- The drastic effects of N- and O-linked glycosylation on the sorting of DPPIV raise the question whether carbohydrates are directly implicated in the sorting event or whether their modification has altered the folding of a protein epitope that comprises the sorting signal. To examine the folding pattern of the various glycoforms of DPPIV, we analyzed the sensitivity of these forms toward trypsin and the V8 protease. The rationale is that a change in the folding pattern would be associated with an altered accessibility of the protein toward the proteases. As shown in Fig. 3 the various glycoforms of DPPIV did not reveal differences in their band pattern upon treatment with trypsin, and in all these cases no additional bands were discerned after trypsin treatment of DPPIV_c/benzyl or DPPIV_c/benzyl/swa. This result strongly suggests that gross structural alterations did not occur in DPPIV after modification of the carbohydrate chains in the Golgi, the site of action of benzyl-GalNAc and swainsonine. These findings were corroborated by employing another protease, the V8 protease. Fig. 3 shows that the fully glycosylated DPPIV, and the underglycosylated DPPIV glycoforms were cleaved in the same way by the protease. In fact, in each case two major polypeptides were revealed that varied in their apparent molecular masses depending on their underglycosylation pattern. Moreover, the labeling intensities of these glycoforms were in all cases comparable, reminiscent of similar susceptibilities of the various DPPIV glycoforms to the digesting protease. Altogether, these data strongly suggest that the folding pattern of DPPIV was not altered due to the variations in the glycosylation pattern in the presence of benzyl-GalNAc, swainsonine, or both. Furthermore, the results are in line with current concepts of protein folding, which have established that major folding events of membrane and secretory proteins including acquisition of a mature, transport-competent form occur in the endoplasmic reticulum (47).

View larger version (37K): [in this window] [in a new window]   Fig. 3.   Sensitivity of the various glycoforms of DPPIV toward trypsin and V8 protease. Caco-2 cells were biosynthetically labeled with [35S]methionine for 15 min and chased with cold methionine for 4 h in the presence or absence of benzyl-GalNAc, swainsonine, or both. Detergent extracts of the labeled cells were immunoprecipitated with mAb anti-DPPIV, and each immunoprecipitate was treated or not treated with 50 µg of trypsin for 30 min at 37 °C (A) or with 5 units of V8 protease for 90 min at 37 °C (B). The reactions were stopped by boiling in SDS-PAGE sample buffer followed by SDS-PAGE analysis on 8% slab gels.

Sorting of DPPIV Implicates Its Association with Lipid Rafts-- One type of signals for sorting proteins to the apical surface of epithelial cells, or axon of neurons, is that of a glycolipid anchor. These anchors have been demonstrated to direct proteins to the apical surface of several types of epithelial cells (48-50), apparently by associating in the TGN with detergent-insoluble membrane or lipid microdomains enriched in glycosphingolipids and cholesterol (20). Although initially observed with GPI-anchored proteins, association with lipid microdomains as a possible mechanism for apical sorting appears to also occur with proteins that possess a transmembrane domain. Localization of DPPIV to membrane microdomains in Caco-2 cells has been shown in pig intestinal explants. We investigated here the specific association of DPPIV with lipid rafts as a possible sorting mechanism and the link between N- and O-linked glycosylation and the sorting process.

DPPIV Is Associated with Lipid Microdomains-- Sphingolipids and cholesterol are the major lipid constituents of Triton X-100-insoluble lipid microdomains. Here we analyzed the possible association of DPPIV and the various underglycosylated DPPIV forms with these domains. Detergent extracts of cells that were biosynthetically labeled in the presence or absence of inhibitors of glycosylation were subjected to density gradient centrifugation, and the presence of DPPIV in the different fractions of these gradients was assessed by immunoprecipitation. Fig. 4 demonstrates that at 4 h of chase the complex glycosylated DPPIV_c was found in the floating fractions at low buoyant density of the gradient characteristic of detergent-insoluble membrane microdomains. On the other hand, the high density fractions contained DPPIV_c as well as its mannose-rich form. Together, the results indicate that DPPIV associates with Triton X-100 membrane microdomains on its way to the brush border membrane and that this association occurs when the protein has acquired its complex glycosylated mature form. When O-glycosylation was affected by benzyl-GalNAc, a decrease in the labeling intensity of DPPIV in the floating fractions was revealed. A similar effect was also obtained when the N-glycosylation pattern was affected by swainsonine. In the presence of both inhibitors, i.e. benzyl-GalNAc and swainsonine, the floating fractions were completely devoid of the corresponding DPPIV glycoform.

View larger version (142K): [in this window] [in a new window]   Fig. 4.   Apical sorting of DPPIV involves its association with lipid rafts. Caco-2 cells were biosynthetically labeled for 4 h with [35S]methionine in the presence or absence of benzyl-GalNAc, swainsonine, or both. The cells were solubilized with Triton X-100, and microdomains were prepared by loading the detergent extracts on 5-35% sucrose gradients. DPPIV was immunoprecipitated from the individual gradient fractions, and the immunoprecipitates were analyzed by SDS-PAGE.

The data indicate that variations in the glycosylation pattern are associated with an increase in the detergent extractability of DPPIV with Triton X-100.

The association of DPPIV and other proteins with cholesterol/sphingolipid-rich membrane microdomains could be also examined by immunoprecipitation of the detergent-insoluble pellet after ultracentrifugtaion of the cellular detergent extracts (Fig. 5). As shown in Fig. 5A the complex glycosylated DPPIV was recovered in the Triton X-100-insoluble pellet supporting the results obtained above using gradient centrifugations. As a control we used intestinal brush border sucrase-isomaltase that is expressed in Caco-2 cells and is sorted to the apical membrane through O-glycans and association with lipid microdomains (28). Here too complex glycosylated sucrase-isomaltase was recovered in the pellet. We next assessed the specificity of the interaction of DPPIV and the two major constituents of the membrane microdomains, sphingolipids and cholesterol. For this the synthetic levels of these lipids were regulated by using inhibitors of key reactions along their biosynthetic pathways. Fumonisin inhibits sphingosine N-acyltransferase and therefore blocks sphingolipid synthesis (41), whereas cyclodextrin specifically removes plasma membrane cholesterol as shown in biochemical and cytochemical studies (42). Fig. 5 demonstrates that inhibition of sphingolipid synthesis has affected the association of DPPIV with microdomains due to the reduced proportion of DPPIV in the pellet fraction. This indicates that sphingolipids are essential components of the membrane microdomains containing DPPIV. The role of cholesterol was next investigated by inhibition of its synthesis using -cyclodextrin. Fig. 5 shows that the Triton X-100-insoluble pellet was almost devoid of DPPIV, and the protein was recovered in the detergent-soluble supernatant. Inhibition of cholesterol synthesis has therefore abolished the detergent insolubility of DPPIV and disrupted its association with membrane microdomains. The fact that depletion of cholesterol affects the association of DPPIV with lipid microdomains more significantly than the inhibition of the synthesis of sphingolipids suggests that cholesterol is the major constituent of the DPPIV-containing lipid microdomains.

View larger version (64K): [in this window] [in a new window]   Fig. 5.   Implication of sphingolipids, cholesterol, and N- and O-glycans in the association of DPPIV with lipid microdomains. A, Caco-2 cells were biosynthetically labeled with [35S]methionine for 4 h in the presence or absence of fumonisin, -cyclodextrin, benzyl-GalNAc, swainsonine, or a combination of benzyl-GalNAc and swainsonine. Triton X-100 extracts were centrifuged at 100,000 × g for 1 h at 4 °C, and the supernatants (S) and pellets (P) were immunoprecipitated with mAb anti-DPPIV and analyzed by SDS-PAGE. SI was used as a control of a protein that is associated with lipid microdomains. B, quantification of the proportions of DPPIV detected in the supernatants or pellets shown in A.

Relationship between N- and O-Linked Glycosylation and Association of DPPIV with Cholesterol-rich Microdomains-- To assess the sorting mechanism further, the association of DPPIV with lipid microdomains was examined after modulation of the posttranslational N- and O-linked glycosylation events with benzyl-GalNAc, swainsonine, or both. Fig. 5 shows that the association of DPPIV with lipid microdomains was only slightly affected when N-linked glycosylation was modified in the presence of swainsonine. This together with the observation that swainsonine has no influence on the sorting behavior of DPPIV (36) imply that O-glycans and the complex N-glycans present in this glycoform are sufficient for its association with microdomains and subsequent polarized sorting. On the other hand, the proportion of DPPIV associated with lipid microdomains decreases upon inhibition of O-linked glycosylation with benzyl-GalNAc (Fig. 5). Almost 20% are found now associated with lipid microdomains as compared with 32% of fully N- and O-glycosylated DPPIV. Substantial alteration in the detergent extractability of DPPIV to ~92% solubility is obtained when both types of glycosylation were modified in the presence of swainsonine and benzyl-GalNAc. Complete dissociation of DPPIV was, however, not achieved. In comparison, SI is completely converted into a detergent-soluble protein upon inhibition of O-glycosylation by benzyl-GalNAc, and it became randomly transported to the apical and basolateral membranes (28). A similar role of O-glycans on the apical sorting of DPPIV was not observed, because the protein has maintained its targeting pattern to the apical membrane, albeit to a lesser extent (Fig. 2). Only when both N- and O-glycosylation were affected by swainsonine and benzyl-GalNAc was the random transport of DPPIV to both membranes discerned. The data could be summarized as follows. Complex N- and O-linked glycosylation act in concert to maintain the association of DPPIV with membrane microdomains, and this association was indispensable for apical sorting.

The Effects of Fumonisin and -Cyclodextrin on the Apical Sorting of DPPIV_c-- As shown above, the detergent extractabilities of DPPIV_c with Triton X-100 were altered upon inhibition of sphingolipid synthesis by fumonisin and cholesterol depletion by -cyclodextrin. We examined the sorting behavior of DPPIV_c in the presence of these two reagents using Caco-2 cells grown on membrane filters and biosynthetically labeled in the presence or absence of fumonisin or -cyclodextrin. Cell surface immunoprecipitation revealed that DPPIV_c partially shifted to the basolateral membrane in the presence of fumonisin (Fig. 6). A higher effect on the targeting of DPPIV_c was obtained when cholesterol was depleted in the presence of -cyclodextrin. Here, more than 35% of DPPIV_c was found at the basolateral membranes, supporting the notion that cholesterol plays a more decisive role in the trafficking mechanism of DPPIV than sphingolipids.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   The sorting of DPPIV is strongly affected by cholesterol depletion. A, Caco-2 cells were pulse-labeled for 4 h in the presence or absence of -cyclodextrin or fumonisin. DPPIV was precipitated from the apical (a) or basolateral (b) membranes by cell-surface immunoprecipitation with mAb anti-DPPIV. The immunoprecipitates were finally analyzed by SDS-PAGE. B, quantification of the data shown in A.

The Effect of N- and O-Glycosylation on the Sorting Behavior of DPPIV and Its Membrane Association in HT-29 Cells-- To determine whether the effects observed in Caco-2 are not cell-specific, we used another human colon cell line, HT-29, that expresses at confluence and differentiation adequate levels of endogenous DPPIV at the apical surface (51). Similar experimental protocols were utilized to examine the influence of swainsonine and benzyl-GalNAc on the trafficking and membrane association of DPPIV. Fig. 7A demonstrates that DPPIV is predominantly delivered to the apical surface (almost 80%) of HT-29 cells and is only partially found at the basolateral membrane thus confirming previous data (51). In the presence of benzyl-GalNAc O-glycosylation was substantially affected as assessed by the marked reduction in the apparent molecular weight of DPPIV (Fig. 7B). The apical sorting of this DPPIV glycoform was partially affected, because almost 65% were now transported to the apical membrane. In the presence of swainsonine a reduction in the apparent molecular weight was also observed due to inhibition of the processing of N-linked glycans (Fig. 7B). As in the case of benzyl-GalNAc a partial effect on the apical sorting of this DPPIV glycoform was observed, because more than 35% of this species was found at the basolateral membrane. In the presence of both inhibitors and in a fashion similar to Caco-2 cells, the strong sorting fidelity of DPPIV to the apical membrane was further impaired. Here, more than 42% of the DPPIV molecules were found now at the basolateral membrane. It is obvious that the role of N-glycans in the sorting of DPPIV in HT-29 cells is more pronounced than in Caco-2 cells. However, similar to Caco-2 cells the sorting of DPPIV in HT-29 cells requires intact structures of both types of glycosylation.

View larger version (38K): [in this window] [in a new window]   Fig. 7.   Effects of benzyl-GalNAc and swainsonine on the sorting and membrane association of DPPIV in HT-29 cells. A, sorting of the DPPIV glycoforms in HT-29 cells. HT-29 cells were grown on transmembrane filters and biosynthetically labeled with [35S]methionine for 3 h in the presence or absence of benzyl-GalNAc or benzyl-GalNAc and swainsonine. Cell-surface immunoprecipitation of DPPIV from the apical or basolateral membranes was performed with mAb anti-DPPIV followed by SDS-PAGE analysis. The amount of complex glycosylated DPPIV on the apical or basolateral surface was quantified from two separate experiments and indicated as vertical bar chart. B, association of the DPPIV glycoforms with membrane microdomains. HT-29 cells were biosynthetically labeled as in A. Triton X-100 extracts were ultracentrifuged, and the detergent-insoluble microdomain-containing pellets were solubilized with 0.1% SDS followed by dilution with 10-fold volume of Triton X-100. The solubilized pellets (P) and the supernatants (S) of the ultracentrifugation were immunoprecipitated with mAb anti-DPPIV. The immunoprecipitates were analyzed by SDS-PAGE. C, quantification of the proportions of DPPIV appearing in the supernatant or pellet fractions shown in B.

In a similar fashion to the effect on the apical delivery of DPPIV, the inhibitors also exploited specific effects on the association of this protein with lipid microdomains. Thus, in the presence of benzyl-GalNAc or swainsonine, an increase in the Triton X-100 detergent solubility of DPPIV and thus reduction in its association with membrane microdomains were observed (Fig. 7, B and C). Compatible with the sorting behavior, however, a stronger effect on the detergent extractability of DPPIV was generated when both inhibitors were present in the labeling medium of HT-29 cells. Here almost 90% of the DPPIV glycoform was now recovered in the supernatant as compared with 75% in the control cells (Fig. 7, B and C).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Current concepts have proposed either N-linked or O-linked glycans as sorting signals for apical proteins. This study demonstrates for the first time and in two different cell lines that complex N- as well as O-linked glycans constitute the basis of an apical sorting signal. The sorting of human intestinal DPPIV to the microvillar membrane in Caco-2 and HT-29 cells requires the concerted action of both types of glycosylation. Moreover, the mechanism required for the sorting implicates an association of the protein with detergent-insoluble cholesterol- and sphingolipid-rich microdomains.

An important tool in these studies is the utilization of benzyl-GalNAc, a potent inhibitor of the extension of O-linked glycans (39), and swainsonine, a reagent that inhibits the function of Golgi mannosidase II (40). Benzyl-GalNAc competes with GalNAc in binding to potential O-glycosylated Ser/Thr sites thus preventing the addition of Gal to GalNAc-O-Ser/Thr by galactosyltransferase (52). It has been shown that the optimal substrate of galactosyltransferase is GalNAc-O-Ser/Thr as compared with benzyl-GalNac-O-Ser/Thr (53). Our data are in line with this mode of action because the size of the DPPIV_benzyl glycoform is substantially reduced upon depletion of its N-linked glycans to a species smaller than the N-deglycosylated but O-glycosylated control DPPIV. This indicates that DPPIV_benzyl contains either no O-linked Gal residues or has a drastically reduced content of these chains. Another possible function of benzyl-GalNAc constitutes the competitive inhibition of 2,3-sialyltransferase by the disaccharide Gal-benzyl-GalNAc which is metabolized in the cell instead of Gal-GalNAc (51). This mechanism applies particularly to mucins in HT-29 cells, because proteins in Caco-2 cells, which do not express 2,3-sialyltransferase, are also affected by benzyl-GalNAc (this paper and Ref. 54) suggesting that not only sialylation but rather an earlier glycosylation step is affected in these cells. The common aspect that emerges from these studies is that benzyl-GalNAc effect is restricted to O-glycans (28, 51). A recent report has proposed that benzyl-GalNAc affects only sialylation of N-linked carbohydrates in DPPIV, and moreover DPPIV is not O-glycosylated. However, this is not in line with many studies that demonstrated that DPPIV is an O-glycosylated glycoprotein (36, 51, 55).

Complementary to investigating the effect of O-glycans, modulation of complex N-linked glycans by swainsonine, an inhibitor of Golgi mannosidase II (40), should provide important clues on a possible implication of N-glycosylation together with O-glycans in the sorting mechanism of DPPIV. The effect of swainsonine is exclusively restricted to N-linked glycans, whereas O-glycans are not affected. DPPIV glycoforms generated in the presence or absence of swainsonine are similarly O-glycosylated but contain a substantially different content of mature sialylated endo H-resistant N-linked glycans (36). The common effects of swainsonine and benzyl-GalNAc on DPPIV would be reflected by a reduced N-glycosylation and a drastic reduction or an inhibition of O-glycans. This appears to be the case due to the further reduction in the apparent molecular mass of DPPIV as compared with swainsonine- or benzyl-GalNAc-treated cells. The variable effects of these reagents, individually or together, provide a strong means to analyze the role of N- and O-linked glycans or their common effects on the polarized sorting of DPPIV.

To our knowledge, our data provide the first evidence that complex N- and O-linked glycans act together as a sorting signal for an apical protein. Neither O-linked nor N-linked glycans individually are capable of driving DPPIV to the apical membrane. O-Linked glycans alone have been proposed to be involved directly in apical sorting of some integral membrane proteins (30), and our data show that O-linked glycans play a role in the apical sorting of DPPIV in the two cell lines, Caco-2 and HT-29. However, fully and properly processed O-linked glycan units are not sufficient per se for high sorting fidelity of DPPIV; the proportion of DPPIV at the basolateral membrane as a consequence of diminished or impaired N- or O-linked glycosylation is elevated when both types of glycosylation are affected. Interestingly, an impaired O-glycosylation of DPPIV in the presence of deoxymannojirimycin, an inhibitor of mannosidase I that indirectly affects O-glycans processing, results in an almost equal distribution of DPPIV on both sides of the membrane in Caco-2 (36). How could these data be accommodated with the current ones obtained in the presence of the specific O-glycosylation inhibitor benzyl-GalNAc? A major difference between the glycosylation pattern of DPPIV in the presence of dMM and benzyl-GalNAc is the structure of the N-linked glycans. Whereas DPPIV is revealed exclusively as an endo H-sensitive mannose-rich glycosylated polypeptide in the presence of deoxymannojirimycin, its pattern of N-linked glycosylation is of the complex endo H-resistant type when cells are incubated with benzyl-GalNAc. A particular role could be therefore attributed to complex glycosylated N-linked glycans as an additional determinant that contributes to the structure of the apical sorting signal of DPPIV complementary to O-linked glycans. Obviously, mannose-rich N-linked glycans do not contribute to the sorting signal of DPPIV as shown with the inhibitor deoxymannojirimycin (36).

The protease sensitivity assays strongly suggest that the modification of the N- and O-linked glycans is not associated with an alteration in the folding pattern of DPPIV, indicating that the loss of sorting fidelity of DPPIV is exclusively the consequence of an altered carbohydrate structure and not due to an affected protein epitope that may act as a sorting signal.

Our data further demonstrate that DPPIV does not harbor repressed basolateral sorting elements, because elimination of the apical signals does not lead to an exclusive and substantial shift in the targeting of the protein to the basolateral membrane. Repressed signals have been observed in apically sorted proteins, for example the neurotrophin receptor (30) and lactase-phlorizin hydrolase (56), and acquire functionality in the absence of the apical signal leading to basolateral localization.

The mechanism responsible for targeting DPPIV to the apical membrane implicates its association with detergent-insoluble microdomains. The specificity of this interaction is corroborated by the fact that inhibition of the synthesis of sphingolipids and cholesterol, the two major components of the microdomains, results in a significant alteration in the detergent solubility and in the sorting behavior of DPPIV from a highly polarized to a randomly delivered molecule to either membrane. Of particular interest is the importance of cholesterol in the association of DPPIV with lipid microdomains. Here, cholesterol depletion by -cyclodextrin alters this association as well as its sorting pattern more strongly than inhibition of sphingolipid synthesis by fumonisin. Similar effects could not be observed with SI, another intestinal brush border protein that is associated with detergent-insoluble microdomains (28).² This suggests the existence of structural differences between DPPIV- and SI-associated microdomains. The critical role of cholesterol in the micrdomains containing DPPIV may reflect a higher concentration of this lipid as compared with sphingolipids. Recent observations (57) on variations in the extractabilities of proteins with different detergents are in line with the notion that proteins may associate with different lipid structures in the same membrane bilayer.

Similar to the sorting behavior of DPPIV, association of this protein with lipid microdomains directly implicates N- and O-linked glycans, the modification of which by the concerted action of benzyl-GalNAc and swainsonine results in a detergent-soluble glycoform of DPPIV. Whereas modification of N-glycosylation alone in the presence of swainsonine remains without a marked effect on the detergent extractability and the sorting behavior of DPPIV, reduction of O-linked glycans results in a higher detergent solubility and reduction in the polarized sorting to the apical membrane. Complete detergent solubility and random sorting is virtually achieved when both types of glycosylation are affected. The synergy in the effects of modulators of lipid microdomains and N- and O-glycosylation on detergent solubility and apical sorting strongly proposes the existence of a sorting mechanism that implicates the association of DPPIV through its N- and O-linked glycans with lipid microdomains. A direct association with the lipid components cholesterol and sphingolipids is unlikely to occur perhaps due to charge constraints (58). One potential mechanism implicates the binding of N- and O-linked glycans of DPPIV to a lectin-like sorting protein in the TGN that is associated with lipid microdomains. The identity of a putative cellular lectin-like protein specific for binding these residues is unclear.

The insufficient function of O-linked glycans of DPPIV as an apical sorting signal is probably due to their spreading over the ectodomain of DPPIV. Proteins such as SI and the neurotrophin receptor possess Ser/Thr-rich stalk domains in which O-glycans are clustered permitting an efficient access to multivalent binding proteins, such as lectins (29, 30). A high binding affinity of a putative lectin receptor to the O-glycans of DPPIV would not be expected. This would also explain the auxiliary role and the requirement of the complex N-glycans containing galactose residues in the binding to the putative receptor.

The multiplicity and diversity of apical signals described so far, such as N- and O-linked glycans acting individually or together, membrane anchoring domains, ectodomains, or GPI anchors, favor the notion that one or more mechanisms function individually or in concert to ensure a strong association with the apical sorting machinery and consequently high fidelity of targeting. It is obvious here that O-linked glycans could not achieve a high fidelity of sorting of DPPIV and require for this the assistance of N-linked glycans. Regardless of the diversity of the signals, however, apical sorting mechanisms could be discriminated on the basis of the association of the sorted proteins with detergent-insoluble lipid microdomains. The association with lipid microdomains makes use of N- and O-linked glycans as a sorting signal, presumably via a lectin-like receptor and also membrane anchoring domains (59). In another mechanism proteins are sorted through proteinaceous elements and do not associate with a lectin-like receptor. Lactase-phlorizin hydrolase and aminopeptidase N (29, 36), for example, are heavily N- and O-glycosylated but are sorted by a different mechanism that employs non-glycan sorting signals located in their ectodomains (5, 60, 61). In fact, recent analysis of vesicular structures with green fluorescent protein-derived constructs of apical proteins that are sorted through association or non-association with lipid microdomains has provided an unequivocal evidence for the existence of two types of apical carriers, the analysis of which would be crucial for elucidating the components of the apical sorting machinery (62).

	ACKNOWLEDGEMENTS

We thank Dr. Hans-Peter Hauri, Biozentrum, University of Basel, and Dr. Erwin Sterchi, Institute of Biochemistry and Molecular Biology, University of Bern, for generous gifts of monoclonal anti-DPPIV antibodies. HT-29 and Caco-2 cells were kindly provided by Dr. Monique Rousset and Dr. Alain Zweibaum, INSERM, Villejuif, France.

	FOOTNOTES

^* This work was supported by Grant Na 331/1-2 from the Deutsche Forschungsgemeinschaft, Bonn, Germany (to H. Y. N.), and the Sonderforschungsbereich 280.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.

To whom all correspondence should be addressed: Dept. of Physiological Chemistry, School of Veterinary Medicine, Hannover, Bünteweg 17, D-30559 Hannover, Germany. Tel.: 49-511-9538780; Fax: 49-511- 9538585; E-mail: Hassan.Naim@tiho-hannover.de.

Published, JBC Papers in Press, December 28, 2001, DOI 10.1074/jbc.M109357200

² M. Alfalah, R. Jacob, and H. Y. Naim, unpublished data.

	ABBREVIATIONS

The abbreviations used are: TGN, trans-Golgi network; GPI, glycophosphatidylinositol; benzyl-GalNAc, benzyl-N-acetyl--D-galactosaminide; DPPIV, dipeptidyl peptidase IV; DPPIV_c/benzyl, complex form after treatment with benzyl-GalNAc; DPPIV_c/benzyl/swa, complex form after treatment with benzyl-GalNAc and swainsonine; SI, sucrase-isomaltase; dMM, deoxymannojirimycin; mAb, monoclonal antibody; endo F/GF, endoglycosidase F/N-glycopeptidase F.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				D. Delacour, A. Koch, W. Ackermann, I. E.-L. Parco, H.-P. Elsasser, F. Poirier, and R. Jacob Loss of galectin-3 impairs membrane polarisation of mouse enterocytes in vivo J. Cell Sci., February 15, 2008; 121(4): 458 - 465. [Abstract] [Full Text] [PDF]

				H. R. S. Tagaram, G. Wang, T. M. Umstead, A. N. Mikerov, N. J. Thomas, G. R. Graff, J. C. Hess, M. J. Thomassen, M. S. Kavuru, D. S. Phelps, et al. Characterization of a human surfactant protein A1 (SP-A1) gene-specific antibody; SP-A1 content variation among individuals of varying age and pulmonary health Am J Physiol Lung Cell Mol Physiol, May 1, 2007; 292(5): L1052 - L1063. [Abstract] [Full Text] [PDF]

				S. Parry, F. G. Hanisch, S.-H. Leir, M. Sutton-Smith, H. R. Morris, A. Dell, and A. Harris N-Glycosylation of the MUC1 mucin in epithelial cells and secretions Glycobiology, July 1, 2006; 16(7): 623 - 634. [Abstract] [Full Text] [PDF]

				M. Keiser, M. Alfalah, M. J. Propsting, D. Castelletti, and H. Y. Naim Altered Folding, Turnover, and Polarized Sorting Act in Concert to Define a Novel Pathomechanism of Congenital Sucrase-Isomaltase Deficiency J. Biol. Chem., May 19, 2006; 281(20): 14393 - 14399. [Abstract] [Full Text] [PDF]

				D. Castelletti, G. Fracasso, M. Alfalah, S. Cingarlini, M. Colombatti, and H. Y. Naim Apical Transport and Folding of Prostate-specific Membrane Antigen Occurs Independent of Glycan Processing J. Biol. Chem., February 10, 2006; 281(6): 3505 - 3512. [Abstract] [Full Text] [PDF]

				B. A. Potter, R. P. Hughey, and O. A. Weisz Role of N- and O-glycans in polarized biosynthetic sorting Am J Physiol Cell Physiol, January 1, 2006; 290(1): C1 - C10. [Abstract] [Full Text] [PDF]

				M. Alfalah, G. Wetzel, I. Fischer, R. Busche, E. E. Sterchi, K.-P. Zimmer, H.-P. Sallmann, and H. Y. Naim A Novel Type of Detergent-resistant Membranes May Contribute to an Early Protein Sorting Event in Epithelial Cells J. Biol. Chem., December 30, 2005; 280(52): 42636 - 42643. [Abstract] [Full Text] [PDF]

				D. Delacour, V. Gouyer, J.-P. Zanetta, H. Drobecq, E. Leteurtre, G. Grard, O. Moreau-Hannedouche, E. Maes, A. Pons, S. Andre, et al. Galectin-4 and sulfatides in apical membrane trafficking in enterocyte-like cells J. Cell Biol., May 9, 2005; 169(3): 491 - 501. [Abstract] [Full Text] [PDF]

				W. Popik and T. M. Alce CD4 Receptor Localized to Non-raft Membrane Microdomains Supports HIV-1 Entry: IDENTIFICATION OF A NOVEL RAFT LOCALIZATION MARKER IN CD4 J. Biol. Chem., January 2, 2004; 279(1): 704 - 712. [Abstract] [Full Text] [PDF]

				L. K. Nyasae, A. L. Hubbard, and P. L. Tuma Transcytotic Efflux from Early Endosomes Is Dependent on Cholesterol and Glycosphingolipids in Polarized Hepatic Cells Mol. Biol. Cell, July 1, 2003; 14(7): 2689 - 2705. [Abstract] [Full Text] [PDF]

				D. Hoekstra, O. Maier, J. M. van der Wouden, T. A. Slimane, and S. C. D. van IJzendoorn Membrane dynamics and cell polarity: the role of sphingolipids J. Lipid Res., May 1, 2003; 44(5): 869 - 877. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/12/10683    most recent M109357200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Alfalah, M. Articles by Naim, H. Y. Search for Related Content PubMed PubMed Citation Articles by Alfalah, M. Articles by Naim, H. Y. Social Bookmarking                      What's this?